estudogeral.uc.pt · 2021. 1. 7. · Agradecimentos “O agradecimento é a mais alta forma de...

1

António Campos de Figueiredo

STUDY ON THE CONTRIBUTION OF THE

CHOROID TO THE PATHOPHYSIOLOGY OF

DIABETIC RETINOPATHY

Tese no âmbito do Programa de Doutoramento em Ciências da Saúde – ramo de Medicina,

orientada pelos Professores Doutores Rufino Martins Silva e António Francisco Ambrósio, e

apresentada à Faculdade de Medicina da Universidade de Coimbra

Abril de 2020

António Campos de Figueiredo

STUDY ON THE CONTRIBUTION OF THE

CHOROID TO THE PATHOPHYSIOLOGY OF

DIABETIC RETINOPATHY

Tese no âmbito do Programa de Doutoramento em Ciências da Saúde – ramo de Medicina,

orientada pelos Professores Doutores Rufino Martins Silva e António Francisco Ambrósio, e

apresentada à Faculdade de Medicina da Universidade de Coimbra

Abril de 2020

The clinical study was conducted at Centro Hospitalar de Leiria, Ophthalmology

Department, Leiria, Portugal. The experimental research was performed at the Retinal

Dysfunction and Neuroinflammation Lab, Coimbra Institute for Clinical and Biomedical

Research (iCBR), Faculty of Medicine, University of Coimbra, Coimbra, Portugal.

Agradecimentos

“O agradecimento é a mais alta forma de pensamento e a gratidão é a felicidade duplicada

pelo espanto”, G. K. Chesterton.

Aos meus pais, onde me dizem que tudo começou.

Aos Professores Doutores Rufino Silva e Francisco Ambrósio, pela orientação e revisões,

pela disponibilidade, proximidade, cumplicidade e pela paciência.

Ao Professor Doutor João Paulo Castro de Sousa, porque me ajudou numa altura em que

eu necessitava.

Ao Professor Doutor Joaquim Murta, pelo seu encorajamento e simpatia.

Aos Doutores Elisa e João, por todo o apoio, ajuda e cumplicidade; pela sua simpatia e

bondade. Sem eles o trabalho experimental teria sido bastante improvável.

À Dra Dulce Castanheira, que me motivou.

À Ann, pelo que me trouxe da mentalidade inglesa.

À minha família próxima e aos amigos, que são como os irmãos que não tive, pela minha

ausência.

Aos meus filhos, que são uma das maravilhas da minha vinda a este mundo.

À Anália, pela bondade, pelo mérito, pela humildade, pela beleza e pelo amor. A minha

casa.

A todos aqueles a quem não agradeci.

Aos meus amigos que se sentam nas estantes da sala de jantar: Hugo, Dickens,

Doistoiévski, Shakespeare, Wittgenstein e Chesterton. Foram um amparo durante todos

estes anos. Expressões como “reviver uma dor do passado no presente, é fazer outra dor

e sofrer novamente”, “sobre aquilo que não sabemos, o melhor é estar calado”, “nenhuma

corrente é mais forte que o seu elo mais fraco” e “quem acende uma luz é o primeiro a

beneficiar da claridade”, foram reconfortantes e inspiradoras.

A Maria, Mãe, que nunca me deixou experimentar a solidão.

A Jesus Cristo, a minha “raison d’être”.

Table of Contents

Resumo .......................................................................................................................... I

Abstract ......................................................................................................................... V

Publications ................................................................................................................ VIII

Communications .......................................................................................................... IX

List of abbreviations ...................................................................................................... X

List of figures .............................................................................................................. XIII

List of tables ............................................................................................................. XXV

List of supplementary videos .................................................................................. XXVII

Thesis outline .......................................................................................................... XXIX

1. Introduction ................................................................................................................ 1

1.1. Diabetic retinopathy and diabetic macular edema ............................................... 2

1.2. Anatomy and physiology of the choroid .............................................................. 4

1.3. Optical coherence tomography (OCT) ...............................................................14

1.4. OCT and the choroid .........................................................................................15

1.5. Choroid and diabetes .........................................................................................17

1.5.1. Choroidal thickness in diabetes without retinopathy ....................................18

1.5.2. Choroidal thickness and retinopathy progression ........................................19

1.5.3. Choroidal thickness in diabetic macular edema ..........................................23

1.6. The influence of treatment on choroidal thickness .............................................24

1.6.1. Panretinal photocoagulation .......................................................................24

1.6.2. Intravitreal therapy ......................................................................................25

1.7. Choroidal thickness as a biomarker of progression or treatment response ........28

1.8. OCT angiography and the choriocapillaris .........................................................30

1.9. From bedside to bench ......................................................................................31

1.9.1. The animal as a model of man ....................................................................31

1.9.2. OCT in the animal .......................................................................................33

1.9.3. Visualization of the choroidal vasculature ...................................................35

1.9.4. Regulation, remodelling and inflammation .................................................37

1.9.4.1. Pericytes and mural cells ........................................................................37

1.9.4.2. Glia and the neuro-vascular unit .............................................................39

1.6. Objectives of the present study ..........................................................................41

1.7. References ........................................................................................................42

2. Choroidal thickness stratified by outcome in diabetic macular edema .......................53

2.1. Abstract .................................................................................................................54

2.2. Introduction ...........................................................................................................55

2.3. Methods ................................................................................................................55

2.4. Results ..................................................................................................................59

2.5 Discussion ..............................................................................................................66

2.6. References ........................................................................................................71

2.7. Supplementary files ...........................................................................................73

3. Markers of outcome in real-world treatment of diabetic macular edema ....................77

3.1. Abstract .................................................................................................................78

3.2. Introduction ...........................................................................................................79

3.3. Patients and Methods ............................................................................................80

3.4. Results ..................................................................................................................84

3.5. Discussion .............................................................................................................93

3.6. Conclusion ............................................................................................................96

3.7. References ........................................................................................................97

3.8. Supplementary files ...........................................................................................99

4. Inflammatory cells proliferate in the choroid and retina without choroidal thickness

change in Type 1 diabetes ............................................................................................. 105

4.1. Abstract ............................................................................................................... 106

4.2. Introduction ......................................................................................................... 107

4.3. Materials and Methods ........................................................................................ 108

4.4. Results ................................................................................................................ 113

4.5. Discussion ....................................................................................................... 124

4.6. References ...................................................................................................... 129

4.7. Supplementary files ......................................................................................... 132

5. Choroidal and retinal structural, cellular and vascular changes in Type 2 diabetes .... 145

5.1. Abstract ............................................................................................................... 146

5.2. Introduction ......................................................................................................... 147

5.3. Materials and Methods ........................................................................................ 148

5.4. Results ............................................................................................................ 153

5.5. Discussion ........................................................................................................... 164

5.6. References ...................................................................................................... 167

5.7. Supplementary files ......................................................................................... 171

6. Discussion and future perspectives......................................................................... 179

7. Conclusions ............................................................................................................ 191

Resumo

I

Resumo

Introdução: Estudos recentes indicam que a espessura da coroide pode ser

considerada fator de prognóstico na retinopatia diabética (RD) e no edema macular

diabético (EMD), embora os resultados sejam contraditórios.

A coroidopatia diabética e a natureza da RD, incluindo do EMD, têm características

complexas, que incluem um componente inflamatório. Alterações na coroide, como a

renovação vascular e a depleção capilar a nível da coriocapilar, e alterações na retina, tais

como a ativação e migração de células da glia, foram descritas em ratos diabéticos. No

entanto, o papel da espessura basal da coroide como fator de prognóstico no EMD não é

consensual. Em modelos animais de diabetes, desconhece-se como varia a espessura da

coroide e se existem alterações celulares e moleculares que ocorrem simultaneamente na

coroide e na retina.

Objetivos: Determinar o valor prognóstico da espessura basal da coroide e

pesquisar outros fatores de prognóstico em doentes com EMD.

Avaliar a espessura da coroide e alterações celulares e moleculares na coroide e

na retina em modelos animais de diabetes tipo 1 (T1D) e tipo 2 (T2D).

Métodos: Cento e vinte e seis olhos de 126 doentes com EMD foram incluídos num

estudo prospetivo, para avaliar o valor prognóstico da espessura coroideia subfoveal

(ECSF) inicial, definido como anatómico (baixa da espessura basal central da retina ≥

10%,) e como funcional (ganho na melhor acuidade visual corrigida, MAVC, basal ≥ 5 letras

ETDRS), na resposta ao tratamento com ranibizumab ou aflibercept, ao final de 3 e 6

meses. Para determinar o valor da ECSF como indicador de espessura coroideia

comparou-se a ECSF com espessuras da coroide à volta da fovea.

Resumo

II

Adicionalmente, 122 olhos de 122 doentes com EMD foram prospetivamente

incluídos, para determinar outros fatores de prognóstico no EMD recente sob tratamento

com anti-angiogénicos.

Relativamente à diabetes experimental, utilizaram-se dois modelos de ratos

diabéticos. Ratos Wistar em que foi induzida T1D através de uma injeção de

estreptozotocina (STZ, às 8 semanas de idade, com mais 8 semanas de duração da

diabetes) e ratos Goto-Kakizaki (GK) com um ano de idade, como modelo de T2D.

A espessura da coroide foi avaliada in vivo por tomografia de coerência ótica (OCT)

em ambos os modelos animais. A densidade vascular da coriocapilar e da coroide vascular,

média e externa, foi quantificada em explantes esclero-coroideus de olhos perfundidos por

perclorato de 1,1’-dioctadecyl-3,3,3’,3’-tetramethilindocarbocianina (DiI). As

imunorreatividades do fator de crescimento do endotélio vascular (VEGF) e do seu recetor

2 (VEGFR2), assim como a da vimentina (marcador das células da macroglia), de Iba1 e

MHC II (marcadores da microglia/macrófagos não ativados e ativados, respetivamente), e

de NG2 (marcador de pericitos e células murais peri-vasculares), foram determinadas por

imuno-histoquímica na coroide e na retina, em criosecções e em explantes esclero-

coroideus.

As imagens foram adquiridas por microscopia de fluorescência ou confocal e a

imunofluorescência foi quantificada pelo ImageJ. Procedeu-se também à contagem de

células positivas para Iba1, MHC II e NG2.

Resultados: A espessura coroideia subfoveal diminuiu com o tratamento do EMD,

mas não revelou possuir valor prognóstico, quer anatómico quer funcional, precoce ou

tardio. A ECSF revelou-se um bom marcador da espessura coroideia, possuindo uma boa

correlação com os outros parâmetros de espessura coroideia. A existência basal de fluido

sub-retiniano revelou-se fator de bom prognóstico anatómico, enquanto que uma zona

elipsoide íntegra e um bom equilíbrio metabólico se revelaram fatores de bom prognóstico

funcional, quer precoces quer tardios.

Resumo

III

Nos modelos experimentais, observaram-se diferenças significativas entre os ratos

com diabetes induzida pela STZ (desequilíbrio metabólico acentuado) e os ratos GK (maior

duração de diabetes com desequilíbrio metabólico ligeiro/moderado). Observou-se um

aumento da espessura da coroide e uma diminuição da densidade vascular da coriocapilar,

in vivo, apenas em ratos GK. A imunorreactividade para o VEGFR2 aumentou na retina

dos ratos GK e diminuiu na retina dos ratos STZ. O número de células Iba1+ aumentou na

retina externa em ambos os modelos animais, embora apenas nos ratos STZ se

encontrasse aumentado no estroma da coroide. O número de células MHC II+ também

aumentou apenas na coroide de ratos STZ. Estes resultados indicam que o incremento das

células inflamatórias na coroide depende do estado metabólico e da duração da doença.

Além disso, também se observaram sinais de rarefação de pericitos a nível da coriocapilar

em ambos os modelos, embora essa alteração fosse mais evidente em ratos GK.

Conclusões: Embora a ECSF diminua no EMD sob tratamento, não se revelou um

fator de prognóstico para o EMD. Revelou-se apenas um indicador de duração de ação do

anti-angiogénico e um bom índice de espessura coroideia em geral. Um bom controlo

metabólico e uma zona elipsoide íntegra revelaram-se fatores de bom prognóstico

funcional, enquanto que o fluido subretiniano se revelou fator de bom prognóstico

anatómico.

A espessura coroideia aumentou e a densidade vascular da coroide diminuiu

apenas no modelo animal de T2D. O número de células Iba1+ e MHC II+ encontrava-se

aumentado na coroide e na retina dos ratos T1D e T2D, mas esse aumento variou com o

desequilíbrio metabólico e com a duração da doença. A imunorreactividade do VEGFR2

encontrava-se aumentada quando a duração da diabetes era mais prolongada e quando

existia apenas um desequilíbrio metabólico ligeiro/moderado. Pelo contrário, a

imunorreactividade do VEGFR2 revelou-se diminuída quando o desequilíbrio metabólico

era acentuado. A rarefação e o aumento da renovação vasculares a nível da coriocapilar

era uma característica que se acentuava numa situação de doença prolongada.

Resumo

IV

Palavras-chave:

Coroide, retina, espessura coroide, diabetes, edema macular, inflamação, glia,

pericitos, VEGFR2.

Abstract

V

Abstract

Background: Recent studies have reported that the choroidal thickness may be a

prognostic factor for diabetic retinopathy and diabetic macular edema (DME), but there are

conflicting results in the literature. Moreover, diabetic choroidopathy and the nature of

diabetic retinopathy, including macular edema, have been recognized as complex traits,

with an inflammatory component. Alterations in the choroid, such as vascular remodelling

and capillary depletion of the choriocapillaris, and in the retina, as glial cell reactivity and

migration, have been described in diabetic rats. However, there is no consensus about the

role of baseline choroidal thickness as a prognostic factor in DME under treatment.

Likewise, it is unknown how the choroidal thickness changes in animal models of diabetes,

as well as the cellular and molecular alterations occurring simultaneously in the choroid and

retina in diabetes.

Purpose: To determine the prognostic value of choroidal thickness and to search

other prognostic factors in patients with DME.

To evaluate the choroidal thickness and changes in cellular and molecular

signatures in the choroid and retina in the course of diabetes, in animal models of Type 1

and Type 2 diabetes.

Methods: In a prospective study, 126 eyes of 126 patients with DME were enrolled

to assess the anatomical (central retinal thickness, CRT, decrease ≥ 10% from baseline)

and functional (best corrected visual acuity, BCVA, gain ≥ 5 ETDRS letters from baseline)

prognostic value of baseline subfoveal choroidal thickness (SFCT) on anti-vascular

endothelial growth factor (anti-VEGF), ranibizumab or aflibercept, treatment response after

3 (early outcome) and 6 months (late outcome). A comparison was made between SFCT

and other choroidal thicknesses collected at different locations from the fovea to establish

the value of SFCT as a surrogate of the choroidal thickness.

Abstract

VI

In addition, 122 eyes of 122 patients were prospectively enrolled to search for

anatomical (CRT) and functional (BCVA) baseline prognostic factors, other than SFCT, for

recent onset DME under anti-VEGF agents’ treatment.

Furthermore, two rat models of diabetes, streptozotocin (STZ)-induced Type 1

diabetes in Wistar rats (8 weeks-old; with further 8 weeks of diabetes duration) and Goto-

Kakizaki (GK) Type 2 diabetes rats (1 year old) were used. In vivo choroidal thickness was

evaluated in both models by optical coherence tomography (OCT). Vascular density of the

choriocapillaris and middle/outer choroid was quantified in sclerochoroidal whole mounts of

eyes perfused by 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate (DiI).

The immunoreactivity of vascular endothelial growth factor (VEGF), VEGF-receptor 2

(VEGFR2), as well as the immunoreactivity of vimentin (marker of macroglial cells), Iba1

and MHC II (markers of non-activated and activated microglial cells/macrophages,

respectively) and NG2 (marker of pericytes and perivascular mural cells), were assessed

by immunohistochemistry, in the choroid and retina, in eye cryosections and in

sclerochoroidal whole mounts. Images were acquired by fluorescence and confocal

microscopy and the immunofluorescence was quantified by ImageJ. Moreover, Iba1, MHC

II and NG2 positive cells were counted.

Results: In diabetic patients, treatment of DME with anti-VEGF agents, ranibizumab

and aflibercept, decreased the choroidal thickness. However, the SFCT was not a predictor

of the anatomical or functional outcomes. SFCT was an excellent surrogate of the choroidal

thickness, showing an excellent correlation with the other choroidal thickness parameters

evaluated. The subretinal fluid was a predictor of the anatomical outcome, whereas the

ellipsoid zone status and a good metabolic control were predictors of functional outcome,

regardless of being early or late outcomes.

In experimental models, significant differences between STZ (serious metabolic-

imbalance) and GK (longer lasting diabetes and light metabolic-imbalance) rats were found.

In vivo choroidal thickness increased in GK rats only and the choriocapillaris vascular

Abstract

VII

density decreased in GK rats only, as well. Moreover, VEGFR2 immunoreactivity was

upregulated in the retina of GK rats, being downregulated in the retina of STZ rats. The

number of Iba1+ cells increased in the outer retina of both animal models. However, in the

choroid, the number of Iba1+ cells and MHC II+ cells increased in STZ rats only. The

aforementioned results for Iba1+ and MHC II+ cells indicate that the degree of such increase

may depend on metabolic status and/or disease duration. Signs of pericyte depletion at the

choriocapillaris were present in both models, being more evident in GK rats.

Conclusions: Although there were alterations in the SFCT in DME under anti-VEGF

treatment, the baseline SFCT was not a useful prognostic tool for DME. It was an indicator

of time-dependent anti-VEGF’s subsiding effect on the choroid instead, and a good

surrogate of the choroidal thickness as such. Good metabolic control and an intact ellipsoid

zone were predictors of functional outcome while subfoveal neuroretinal detachment was a

predictor of anatomic outcome only.

The number of Iba1+ cells and MHC II+ cells increased in the choroid and retina in

diabetic rats but the magnitude of such increase changed considerably when the metabolic

status was seriously imbalanced. VEGFR2 immunoreactivity increased in the retina in

longer diabetes duration and slighter metabolic imbalance. Conversely, VEGFR2

immunoreactivity decreased when there was a serious metabolic imbalance. Vascular

remodelling or vascular depletion at the choriocapillaris was also a trait of the long lasting

disease.

Keywords:

Choroid, retina, choroidal thickness, diabetes, macular edema, inflammation, glia,

pericytes, VEGFR2.

Publications and communications

VIII

Publications

1. António Campos; Elisa J. Campos; João Martins; António Francisco Ambrósio; Rufino

Silva. Viewing the choroid: where we stand, challenges and contradictions in diabetic

retinopathy and diabetic macular oedema. Acta Ophthalmol. 2017; 95:446-459.

https://www.ncbi.nlm.nih.gov/pubmed/27545332. DOI: 10.1111/aos.13210.

2. António Campos; Elisa J. Campos; Anália do Carmo; Miguel Patrício; João P. Castro

de Sousa; António Francisco Ambrósio; Rufino Silva. Choroidal thickness changes

stratified by outcome in real-world treatment of diabetic macular edema. Graefes Arch

Clin Exp Ophthalmol. 2018; 256 (10):1857-1865.

https://www.ncbi.nlm.nih.gov/pubmed/30039271. DOI: 10.1007/s00417-018-4072-z.

3. António Campos; Elisa J Campos; Anália do Carmo; Francisco Caramelo; João Martins;

João P Sousa; António Francisco Ambrósio; Rufino Silva. Evaluation of markers of

outcome in real-world treatment of diabetic macular edema. Eye Vis (Lond). 2018; 11;

5:27. https://www.ncbi.nlm.nih.gov/pubmed/30386806. DOI: 10.1186/s40662-018-

0119-9.

4. António Campos, Elisa J. Campos, Anália do Carmo, Rufino Silva. Response to:

Choroidal thickness changes stratified by outcome in real-world treatment of diabetic

macular edema. Graefes Arch Clin Exp Ophthalmol. 2019; 257:243.

https://www.ncbi.nlm.nih.gov/pubmed/30191300. DOI: 10.1007/s00417-018-4128-0.

5. António Campos, Elisa J Campos, João Martins, Flávia SC Rodrigues, Rufino Silva,

António Francisco Ambrósio. Inflammatory cells proliferate in the choroid and retina

without choroidal thickness change in early Type 1 diabetes. Exp Eye Res. 2020

(submitted).

https://www.ncbi.nlm.nih.gov/pubmed/30039271

Publications and communications

IX

6. António Campos, João Martins, Elisa J. Campos, Rufino Silva, António Francisco

Ambrósio. Choroidal and retinal structural, cellular and vascular changes in a rat model

of Type 2 diabetes. Plos One. 2020 (submitted).

Communications

1. A. Campos; J. Sousa, A. do Carmo, E. Campos, F. Caramelo, A. F. Ambrosio, R. Silva.

Evaluation of markers of outcome in diabetic macular edema. 18th EURETINA

Congress, Vienna, September 20-23, 2018.

2. A. Campos, E. Campos, J. Martins, R. Silva, A. F. Ambrósio. Alterações da Coroide e

Retina num modelo experimental de Diabetes. Jornadas de Investigação do 62º

Congresso Português de Oftalmologia, Vilamoura, Dezembro 5, 2019.

List of abreviations

X

List of abbreviations

3M-EZ Re-rating of the ellipsoid zone at 3 months

ACs Amacrine cells

AFL Aflibercept

AGEs Advanced glycated end-products

AL Axial lenght

AMD Age related macular degeneration

Anti-VEGF Anti-vascular endothelial growth factor

ARVO Association for research in vision and ophthalmology

AUC Area under the curve

B Regression coefficient

BCs Bipolar cells

BCVA Best corrected visual acuity

BCVAi Baseline best corrected visual acuity

BRB Blood retinal barrier

CI-CSME Central involving clinical significant macular edema

CMT Central macular thickness

CNV Choroidal neovascularization

CRT Central retinal thickness

CSC Central serous chorioretinopathy

CSME Clinical significant macular edema

CT Choroidal thickness

CT1750i Choroidal thickness at 1750 µm inferior to the center of the fovea

CT1750n Choroidal thickness at 1750 µm nasal to the center of the fovea

CT1750s Choroidal thickness at 1750 µm superior to the center of the fovea

CT1750t Choroidal thickness at 1750 µm temporal to the center of the fovea

CT3500 Choroidal thickness of the area defined from 1750 µm temporal to 1750 µm nasal from the center of the fovea

D Diopters

DAPI 4’,6-diaminophenylindole

DBP Diastolic blood pressure

DC Diabetic choroidopathy

DCP Deep capillary plexus

Dif T0_T1 Difference between baseline and 3 months

DiI 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine perchlorate

DM Diabetes mellitus

DME Diabetic macular edema

DR Diabetic retinopathy


XI

DRIL Disruption of the inner retinal layers

EDI Enhanced deep imaging

ELM External limiting membrane

ER Early responders

ETDRS Early treatment diabetic retinopathy study

EZ Ellipsoid zone

GCL Ganglion cell layer

GK Goto-Kakizaki

Hb A1c Glycated hemoglobina A1c

HCs Horizontal cells

HR High resolution

HRS Hyper-reflective spots

I Inferior

Iba1 Ionized calcium-binding adapter molecule 1

ICC Intra-class correlations

ICP Intermediate capillary plexus

ILM Inner limiting membrane

INL Inner nuclear layer

IPL Inner plexiform layer

L Letters

LR Late responders

LSM Laser scanning microscope

M Month

MAP Mean arterial blood pressure

MC Müller cells

MG Müller glia

MHC II Major histocompatibility complex II

N Nasal

n Number

NG2 Neuron-glial antigen 2

NG2/Cspg4 Neuron-glial antigen 2/chondroitin sulphate proteoglycan 4

NO Nitric oxide

NPDR Non-proliferative diabetic retinopathy

NR Non-responders

OCT Optical coherence tomography

OCTA Optical coherence tomography angiography

OLM Outer limiting membrane

OLMPr Outer limiting membrane and photoreceptor outer segments

ONH Optic nerve head

ONL Outer nuclear layer

OPL Outer plexiform layer

OR Odds ratio


XII

PBS Phosphate buffer saline

PDR Proliferative diabetic retinopathy

PFA Paraformaldehyde

PIGF Placental growth factor

PRN Pro re nata

PRP Panretinal photocoagulation

PRs Photoreceptors

RECA-1 Rat endothelial cell antibody 1

RGC Retinal ganglion cells

RMG Retinal Müller glial cells

RNFL Retinal nerve fiber layer

RNZ Ranibizumab

ROC Receiving operating characteristic curve

ROS Reactive oxygen species

RPE Retinal pigment epithelium

S Superior

SBP Systolic blood pressure

SCP Superficial capillary plexus

SD Square deviation

SD-OCT Spectral domain optical coherence tomography

SE Standard error

SEM Standard error mean

SFCT Subfoveal choroidal thickness

SND Subfoveal neuroretinal detachment

SPCA Short posterior ciliary arteries

SPSS Statistical package for social sciences

SS-OCT Swept source optical coherence tomography

STZ Streptozotocin

T Temporal

T1D Type 1 diabetes

T2D Type 2 diabetes

VEGF Vascular endothelial growth factor

VEGF-A Vascular endothelial growth factor A

VEGFR1 Vascular endothelial growth factor receptor 1

VEGFR2 Vascular endothelial growth factor receptor 2

y years

List of figures

XIII

List of figures

Figure 1.1. Anatomy of the choroid: In this diagram the anatomy of the choroid is depicted

from the Bruch’s membrane outwards to the suprachoroid. Artwork, courtesy from Anália

do Carmo, MD, MsC, PhD.

Figure 1.2. Expression of rat endothelial cell antibody 1 (RECA-1, Abcam 9774) in the rat

choriocapillaris, viewed from the choroidal side, visualized by laser scanning confocal

microscope LSM 710 (Zeiss, Germany), magnification 200x. A. Network of capillaries in the

rat choroid and B. Inward view with hexagonal shaped RPE cells (thin arrow) and capillaries

derived from a Sattler’s pre-terminal artery (thick arrow) [12].

Figure 1.3. The choriocapillaris and choroidal feeding vessels. A. Scanning electron

micrograph of the choriocapillaris network from vascular casts. Choroidal arteries (A) and

veins (V) beneath the coriocapillaris network (adapted from from Risco and Nopanitaya

[58]). Reproduced with permission. B. Histology of the choriocapillaris. Each feeder arteriole

at the Sattler’s layer supplies a hexagonally-shaped area of capillaries (adapted from

Forrester et al. [57]). Reproduced with permission.

Figure 1.4. The choriocapillaris structure and fenestrations. A. Scanning electron

micrographs of the endothelial surface of the choriocapillaris with numerous and uniformly

arranged fenestrations at the RPE-facing side. Scale bar: 500 nm (adapted from

Shimomura et al [60]). Reproduced with permission. B. Posterior pole coriocapillaris viewed

from the retinal aspect. The inferior retinal vessels are still visible and further deep the

lobular arrangement of the coriocapillaris may be identified though difficult to distinguish.

Scale bar: 250 µm (adapted from Olver, J.M. [56]). Reproduced with permission.

Figure 1.5. The lobular structure of the choriocapillaris in man and in the rat. A. Scanning

electron micrograph of the choriocapillaris viewed from retinal aspect. Lobular pattern is

apparent. Terminal parts of venules are visible as dilated channels (large fan-shaped) in the

periphery of lobules, in the choriocapillaris plane (black asterisks). Terminal arteries are

difficult to be viewed in the center of lobules from the retinal aspect. Scale bar: 250 µm.

Adapted from Olver, J.M. [56]. Reproduced with permission. B. Rat choriocapillaris

immunollabed with RECA-1 viewed from retinal aspect by confocal microscopy. Pre-venular

dilated sinuses at the same plane of the choriocapillaris are visible (white asterisks), but a

clear lobular arrangement of the choriocapillaris is absent. Scale bar: 150 µm.

List of figures

XIV

Figure 1.6. Diagrammatic representation of choriocapillaris and its watershed zones. A,

choroidal arteriole; V, choroidal vein (adapted from Hayreh SS [65]). Reproduced with

permission.

Figure 1.7. Functional end-arterial model of the choroidal circulation. Red and gray vascular

channels represent perfused and non-perfused status, respectively: (a) When a terminal

arteriole is obstructed, the choriocapillaris lobule becomes ischemic because the blood is

drained through the venous channel in the periphery of the lobule; (b) A sector becomes

ischemic when the posterior ciliary artery is occluded, as there is no arteriolar anastomosis

among the other sectors; (c) When triolein embolus flows down to a small arteriole,

perfusion of the choriocapillaris is restored owing to the extensive arteriolar anastomosis

within the sector (adapted from Lee JE et al. [71]). Reproduced with permission.

Figure 1.8. Three-dimensional drawing of the space between photoreceptor outer

segments (rods) and cells of the retinal pigment epithelium (RPE). Thick sheaths (a) of RPE

enclose external portions of rod outer segments without intercellular junctions of any sort

(b). RPE finger-like villous processes (c) are found between photoreceptors and contain

pigment granules (d). Apical portion of RPE layer of cells at bottom contains numerous

pigment granules (e); mitochondria (f); a well-developed, smooth-surfaced endoplasmic

reticulum (g); a poorly developed, rough-surfaced endoplasmic reticulum (h); and scattered

free ribosomes. Stacks of rod outer segment discs are depicted in longitudinal section (i)

and in cross section (j). Periphery of discs shows scalloping (k). Microtubules originating in

basal body of rod cilium extend externally into the outer segment (l) (adapted from Hogan

et al. [68]). Reproduced with permission.

Figure 1.9. Inter-individual variability in choroidal thickness (CT) evaluated by SD-

Spectralis OCT, horizontal scans encompassing the fovea. Eyes from two different patients,

same age and refractive error difference of 0.50 D, were collected. Both OCTs were

obtained during the morning. Measures of total CT were taken from the outer limit of the

RPE to the choroid-scleral junction (yellow line in the right panel and purple line in the left

panel). Note that variability in CT may be about two fold from one another.

Figure 1.10. The structure of the choroid, SD-Spectralis OCT (horizontal scan

encompassing the upper macular ETDRS grid area). White squares signal large choroidal

vessels, with diameter ≥ 100 μm, belonging to the Haller’s layer. Large vessels push the

choriocapillaris/Sattler’s complex layer upwards leaving thick fingertip-like areas of the

List of figures

XV

choriocapillaris/Sattler’s layer in-between (red arrows). Yellow arrows signal the ‘cone of

shadow’ from the large vessels of the choroid that may be a confounding factor when

determining the choroid-scleral junction.

Figure 1.11. Sketches of a primate eye showing position of fovea and macula and the

peripheral retina. A. Central region indicating diameters of the foveola (the foveal pit), fovea,

and macula. At the foveola there is displacement of ganglion cell layer (GCL) and inner

nuclear layer (INL) cells. B. Sketch of peripheral retina showing its major cell classes -

photoreceptors (PRs), horizontal cells (HCs), bipolar cells (BCs), amacrine cells (ACs),

retinal ganglion cells (RGCs) and Müller glia (MG), outer and inner plexiform (synaptic)

layers (OPL and IPL), outer and inner nuclear layers (ONL and INL), and ganglion cell layer

(GCL), (adapted from Peng et al., 2019 [172]). Reproduced with permission.

Figure 1.12. Choroidal thickness collected by SD-OCT in the rat using the EDI technique.

On the left panel, inverted image of the retina (A) and choroid (C) separated by the bright

hyper-reflective line of the RPE (B). While the inner boundary of the choroid follows the

linear outline of the RPE, the outer boundary swings in and out according to the vascular

profile of the large vessels. On the right panel, image shows the position where the scan

was acquired.

Figure 1.13. Automatic determination of the choroid boundaries from the segmentation

software InSight (Phoenix Research Labs) with manual correction. On the left side image,

the green line marks the inner limits of the choroid, while the blue line marks its outer limits,

that is, the choroid-scleral border. Top right side image is the image of the fundus where

the scan was collected. Bottom right side image shows a green line of choroidal thickness

magnitude resulting from collection of the 1024 raster scans all along the raster scan length.

Figure 1.14. Terminal choroidal arterioles visualized by DiI (red) in a 16-week Wistar rat. A.

Arteriole dividing in pre-terminal arteries that originate the multiple lobular network of the

choriocapillaries. Scale bar: 100 µm. 10x Full size: x: 850.19 µm, y: 850.19 µm. B. Pre-

terminal artery giving rise to the choriocapillaris at a right angle after a short trajectory as

previously described in vascular casts [56, 61]. Scale bar: 50 µm. 20x Full size: x: 425.1

µm, y: 425.1 µm. C. The lobular pattern is not self-evident in the coriocapillaris. Instead, a

honeycomb-like pattern is depicted. Scale bar: 50 µm. 20x Full size: x: 425.10 µm, y: 425.10

µm.

Figure 1.15. Composed visualization of the whole choroid in 16-week Wistar rat by DiI. The

List of figures

XVI

vascular profile exhibited is arterial-type with absence of the collecting vortex veins. A.

Orientation of the choroid: S = superior, N = nasal, I = inferior, T = temporal. B. Display of

the whole choroidal circulation resulting from gathering all orientated flat mounts. The

choroidal circulation is arterial end-terminal with pre-terminal arterial-arterial anastomoses,

branching in a bronchiolar pattern. C. Detail shows that anastomosis are frequent before

the emergence of the pre-terminal arteries (white arrows). Scale bar: A; 2mm, B; 1 mm and

C; 0.2 mm.

Figure 1.16. Pericytes and perivascular mural cells in the choriocapillaris and middle

choroid of a 16-week Wistar rat show distint morphology distribution. A. Perivascular mural

cells immunostained by desmin wrap around choroidal vessels, while they assume a linear

or stellate configuration at the choriocapillaris, corresponding to the scanty non-

circumferential distribution of pericytes (yellow arrows). B. Linear immunomarking of

pericytes by desmin near the hexagonal RPE cells’ plane (yellow thick arrows) show a

scanty non-circumferential distribution. C. Distinct morphology of mural cells

immunomarked by desmin wrapping around choroidal vessels while pericytes show a linear

morphology and scanty non-circumferential distribution at the choriocapillaris level (yellow

thick arrow). Scale bar: 50 µm. 10x Full size: x: 850.19 µm, y: 850.19 µm.

Figure 1.17. Cellular components of the inner blood-retinal barrier. Schematic

representation of the neuro-glio-vascular unit forming the inner blood-retinal barrier,

composed by vascular endothelial cells, pericytes (p), retinal Müller glial (RMG) cells,

astrocytes (a), microglia (mc). RMG cell projections are present at the level of all retinal

vascular plexuses (superficial, SCP; intermediate, ICP and deep, DCP), while astrocytes

are only present at the level of the superficial plexus. Adapted from Daruich et al. 2019

[188]. Reproduced with permission.

Figure 1.18. Presence of glial cells around vessels in the retina and in the choroid. A.

Retinal projection showing vessels RECA1+ (green) and Iba+ cells/microglia (red) mostly in

vessels’ vicinity. B. Choroid single confocal plane visualized from the RPE side, showing

choroidal medium-sized and large vessels surrounded by Iba+ cells (white arrows). Laser

scanning microscope LSM 710 (Zeiss), objective lens: 20x, numerical aperture 0.8,

magnification 200x. Scale bar = 50 µm.

Figure 2.1. Choroidal thickness manually measured in the central 3500-μm area

underneath the RPE line, subfoveal (SFCT) and at 1750 μm nasal (CT1750n) and temporal

(CT1750t) from the center, in the plane defined by the horizontal line scan encompassing

List of figures

XVII

the fovea (Image J software, version 1.48, National Institutes of Health, USA). A similar

procedure was done in the plane defined by the vertical line encompassing the fovea to

obtain the superior (CT1750s) and inferior (CT1750i) choroidal thicknesses.

Figure 2.2. Evolution of mean central retinal thickness (CRT), mean subfoveal choroidal

thickness (SFCT), and mean best-corrected visual acuity (BCVA) scored in ETDRS letters

collected from ETDRS charts with time in eyes with DME under anti-VEGF treatment. Note

that the evolution of the SFCT curve does not have the same profile as those of the CRT

and BCVA. Until 6M when the stratification by outcome was done, the slopes of the BCVA

curve and that of the CRT curve are also different, expressing a poor correlation between

anatomic and functional outcome as depicted in Tables 2.2 and 2.3.

Figure 2.3. Example on how variability in SFCT may introduce bias when dealing with small

samples: Scatterplot depicting the negative weak correlation between age and baseline

SFCT (μm), r = −0.328, p < 0.001. Though a wide inter-individual variability can be

observed, the average decrease of SFCT per decade was seen to be 25.45 μm.

Figure 2.4. Radar chart displaying the percentage of participants with decrease in the CT

parameters at different time points. SFCT subfoveal choroidal thickness (dotted black line);

CT temp choroidal thickness 1750 μm temporal to the fovea, in the plane defined by the

horizontal line scan encompassing the fovea (orange line); CT nasal same as the previous

but 1750 μm nasal to the fovea (purple line); CT sup choroidal thickness 1750 μm superior

to the fovea, in the plane defined by the vertical line scan encompassing the fovea (red

line); CT inf same as previous but 1750 μm inferior to the fovea (green line); CT area area

of choroidal thickness from 1750 μm nasal to 1750 μm temporal to the fovea, in the plane

defined by the horizontal line scan encompassing the fovea (blue line) (Image J software,

version 1.48, National Institutes of Health, USA). 3M = 3-month endpoint, 6M = 6-month

endpoint, 12M = 12- month endpoint, 18M = 18-month endpoint, 24M = 24-month endpoint.

Figure 2.5. ROC curve analysis comparing the decrease in SFCT from baseline to 3M with

a ≥ 5 L gain at EM (early functional response). Dif T0_T1 is the difference between mean

baseline SFCT and mean 3M SFCT; Sig is the P value. The Area (area under the curve,

AUC) of 0.529 means that the change in baseline SFCT at 3M does not predict visual gain.

Figure 2.6. Boxplot or figure of extremes and quartiles of the difference found in SFCT from

baseline to 3M in functional responders at 3M (gain of 5 L or more) and non-responders.

T0-T1 is the difference between baseline SFCT and 3M SFCT; NR + LR is the group of

List of figures

XVIII

non-responders at 3M (non-responders at 6M and late functional responders); ER is the

group of responders at 3M (early functional responders). The distribution of the Baseline -

3M SFCT of either group mostly overlaps, meaning that the decrease in the SFCT at 3M is

not a useful tool to mark functional gain.

Supplementary Figure 2.1. Descriptive values during follow-up for mean best corrected

visual acuity (BCVA), central retinal thickness (CRT) and choroidal thickness. CRT = 1 mm

central retinal thickness; SFCT= subfoveal choroidal thickness; CT = choroidal thickness

measured at 1750 µm, nasal (CT1750n) and temporal (CT1750t) from the fovea in the plane

defined by the horizontal line scan encompassing the fovea; superior (CT1750s) and inferior

(CT1750i) from the fovea in the plane defined by the vertical line scan encompassing the

fovea; CT3500 µm (area) = choroidal thickness area underneath the fovea measured from

1750 µm nasal to 1750 µm temporal from the central fovea in the plane defined by the

horizontal line scan encompassing the fovea. 3M = 3 months, after the 3 injection loading

dose; 6M = 6 months; 12M = 12 months; 18M = 18 months; 24M = 24 months. All the

measures were performed at baseline (black bar), 3M (red bar), 6M (green bar), 12M (blue

bar), 18M (purple bar) and 24M (grey bar) after baseline. All measures were performed by

two independent graders and subsequently reached by consensus. Results related to areas

were analysed using the ImageJ software, version 1.48, National Institutes of Health, USA.

Results are presented as mean ± SD.

Supplementary Figure 2.2. Differences in the variables considered between each endpoint

and baseline in the laser-naives (white bars) and in the laser-treated subgroups (black

bars). Comparison between the differences exhibited in either group. CRT = 1 mm central

retinal thickness; SFCT= subfoveal choroidal thickness; CT = choroidal thickness measured

at 1750 µm, nasal (CT1750n) and temporal (CT1750t) from the fovea, in the plane defined

by the horizontal line scan encompassing the fovea; superior (CT1750s) and inferior

(CT1750i) from the fovea in the plane defined by the vertical line scan encompassing the

fovea; CT3500 µm (area) = choroidal thickness area underneath the fovea manually

measured from 1750 µm nasal to 1750 µm temporal from the central fovea in the plane

defined by the horizontal line scan encompassing the fovea (ImageJ software, version 1.48,

National Institutes of Health, USA). BCVA = best corrected visual acuity; 3M = 3 months,

after the 3 injection loading dose; 6M = 6 months; 12M = 12 months; 18M = 18 months;

24M = 24 months. Results are presented as mean ± SD. * p<0.05, correspond to

comparisons between the laser-naives and in the laser-treated groups and were obtained

using independent samples t-Student or Mann-Whitney tests.

List of figures

XIX

Figure 3.1. Re-rating the ellipsoid zone (EZ) after 3 injections of anti-VEGF. ETDRS grid

from the caliper tool set in place centered at the fovea. A. Horizontal scan, 500 μm each

side of the fovea to evaluate the EZ. Note that laser dots are outside the 1500 μm radius

(second circle of the ETDRS grid has a radius of 1750 μm) from the center of the foveola.

B. ETDRS grid set in place centered at the foveola. Vertical scan, 500 μm each side of the

fovea to evaluate the EZ.

Figure 3.2. Examples of the difficulties in rating the ellipsoid zone (EZ) at baseline and after

the 3-monthly injection of anti-VEGF. A. A small subfoveal neuroretinal detachment and in

the shadowing cone effect of a retinal cyst makes the rating of the EZ difficult. In this case

the EZ was rated as ‘disrupted’ by consensus. B. The EZ seems to be disrupted with an

intact external limiting membrane (ELM). C. and D. Eyes shown in A and B after the loading

dose. The EZ is now clearly visible, rated as ‘intact’ by both graders. E. EZ after the loading

dose being rated as ‘disrupted’.

Supplementary Figure 3.1. ETDRS grid in place centered at the fovea. Note that ETDRS

grid plotted (7.2 mm in diameter) is larger than the OCT-modified ETDRS grid (6 mm in

diameter) plotted to access central retinal thickness CRT. A. and B. HR horizontal scans

used to measure the SFCT. ETDRS grid inner circle is 1200 μm (a) and middle circle is

3600 μm wide (b). C. HR vertical scan with SFCT measured underneath the fovea.

Figure 4.1. Effect of Type 1 diabetes on the choroidal thickness. (A) Representative OCT

images of the retina and choroid of control and diabetic rats, at the beginning of the study

(baseline) and at 8 weeks after diabetes onset. Scale bar: 50 µm. (B) Choroidal thickness

of control and diabetic rats, at the baseline and 8 weeks after diabetes onset, based on in

vivo OCT line scans. Bars represent mean ± SEM (control rats, n = 12; diabetic rats, n =

16). (C) Vascular density analysis at the inner (≤10 µm from the outer RPE plane) and

middle + outer choroid (>10 µm from the outer RPE plane) in control and diabetic rats

assessed in sclerochoroidal wholemounts, at 8 weeks after diabetes onset, based on Dil

labelling of choroidal vessels. Bars represent mean ± SEM (control animals, n = 5; diabetic

animals, n = 7).

GCL: ganglion cell layer; INL: inner nuclear layer; OPL: outer plexiform layer; ONL: outer

nuclear layer; RPE: retinal pigment epithelium.

Figure 4.2. Effect of T1D on the localization of mural and endothelial cells in the choroid

and retina. Representative images showing the immunolabelling pattern of (A) NG2 and (B)

RECA-1 in the choroid and retina of control and diabetic rats, 8 weeks after diabetes onset

List of figures

XX

(control animals, n = 7; diabetic animals, n = 9). Polarization in the disposition mural cells

is highlighted (white arrows). Scale bar: 100 µm.



Figure 4.3. Effect of early T1D on the immunoreactivity of VEGF and VEGFR2 in the

choroid and retina. Representative images showing the immunoreactivity of (A) VEGF and

(B) VEGFR2 in the choroid and retina of control and diabetic rats, at 8 weeks after diabetes

onset. Scale bar: 100 µm. (C) VEGF and VEGFR2 immunoreactivity in eye cryosections

based on 12 independent specimen counts per eye. VEGF and VEGFR2 immunoreactivity

was quantified as fluorescence intensity/area per layer. Counting was done for the right eye

only, in all animals. Bars represent mean ± SEM (control animals, n = 7; diabetic animals,

n = 9). Significance: *p < 0.05.

GCL + RNFL: ganglion cell layer and retinal nerve fibre layer; IPL: inner plexiform layer;

INL: inner nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear layer; OLM: outer

limiting membrane; RPE: retinal pigment epithelium.

Figure 4.4. Co-localization of the immunoreactivity of vimentin and VEGF in the retina.

Representative eye cross-sections of (A) control and (B) diabetic rats, at 8 weeks after

diabetes onset, immunolabelled against vimentin and VEGF. Scale bar: 100 µm.



Figure 4.5. Effect of Type 1 diabetes on microglial cell counts and reactivity in the choroid

and retina. Representative images showing the immunoreactivity of (A) Iba1+ and (B) MHC

class II+ cells in the choroid and retina of control and diabetic rats, at 8 weeks after diabetes

onset. Scale bar: 100 µm. (C) Iba1+ cell counts in the retina, and (D) Iba1+ and MHC class

II+ cell counts in the choroid, collected from immunolabelling of eye cross-sections. Bars

represent mean ± SEM (control animals, n = 7; diabetic animals, n = 9). (E) Iba1+ and MHC

class II+ cells density in the choroid collected from sclerochoroidal wholemounts. Bars

represent mean ± SEM (control animals, n = 5; diabetic animals, n = 7). Significance: *p <

0.05; **p < 0.01.

MHC class II: major histocompatibility complex class II; IPL: inner plexiform layer; INL:

inner nuclear layer; OPL: outer plexiform layer.

Supplementary Figure 4.1. SD-OCT scan acquisition for choroidal thickness evaluation.

(A) Linear scans (blue line) were acquired above the optic nerve head (ONH), in the area

List of figures

XXI

within 1 to 3 ONH diameter from the optic disc. (B) Segmentation lines were manually drawn

at the inner (retinal pigment epithelium, RPE; green line) and outer (choroidal-scleral border;

yellow line) boundaries of the choroid. Scale bar: 50 µm.

GCL: ganglion cell layer; INL inner nuclear layer; OPL: outer plexiform layer; ONL: outer


Supplementary Figure 4.2. Quantification of Iba1+ and MHC class II+ cells in

sclerochoroidal whole mounts (n = 5 controls; n = 7 diabetics). (A) Iba1+ cells were counted

in all depth planes of the choroid. (B) MHC class II+ cells marked using yellow dots to avoid

duplicate counting. Scale bar: 100 µm.

Supplementary Figure 4.3. Quantification of the vascular density in sclerochoroidal

wholemounts (defined as the percentage of total area covered by choriocapillaris vessels)

using the ‘image>adjust>threshold’ window tool of ImageJ to obtain the percentage of

vascular coverage (n = 5 controls; n = 7 diabetics). (A) at the inner choroid/choriocapillaris

(z-stacks collected at ≤10 µm from the posterior RPE cell plane), and (B) at middle and

outer choroid/medium and large vessels (z-stacks collected at >10 µm from the posterior

RPE cell plane). Scale bar: 50 µm.

Supplementary Figure 4.4. Body weight and glycaemia values of animals, since diabetes

onset. (A) At diabetes onset (0 weeks), both control (254.9 ± 7.5 g) and diabetic (267.8 ±

7.0 g) groups did not differ significantly for body weight; at 8 weeks after diabetes onset,

control rats were significantly heavier (control: 398.7 ± 8.5 g vs diabetic: 266.7± 0.8 g). (B)

Glycaemia was significantly higher in the diabetic rats, at diabetes onset (control: 109.6 ±

3.4 g/dL vs diabetic: 502.3 ± 27.7 g/dL), and 8 weeks later (control: 100.0 ± 2.1 g/dL vs

diabetic: 537.8 ± 17.1 g/dL). HbA1c was significantly higher in diabetic rats, at 8 weeks after

diabetes onset (control: 6.2 ± 0.1% vs diabetic: 8.3 ± 0.3%, t(13)=4.486, p = 0.001). Bars

represent mean ± SEM (control animals, n = 12; diabetic animals, n = 16). Significance: ***p

< 0.001.

Supplementary Figure 4.5. Interconnecting vessels between retinal plexuses were

observed in eye cross-sections immunolabelled against RECA-1: superficial and middle

plexuses (white arrow); middle and deep plexuses (yellow arrow); deep and superficial

plexuses (red arrow). Scale bar: 100 µm.



List of figures

XXII

Figure 5.1. Choroidal thickness (CT) of GK and age-matched control Wistar Han rats (52

W). (A) Images resulting from five frames averaged, collected as inverted images using

OCT raster scans obtained in 1024 continuous points, by approaching the device to the

zero delay line. Horizontal raster scan line encompasses an area within 1 to 3 disk

diameters from the optic nerve head. Choroidal layer obtained by automatic segmentation

was manually corrected. A mean of three independent scores obtained from 3 different

located sets “five frames averaged” were used as the CT value per eye per time-point. (B)

CT values from all eyes of GK (n = 36) and age-matched control (n = 26) rats. Data are

expressed as mean ± SEM. Scale bar: 100 µm. Significance: ***p < 0.001.

GCL = ganglion cell layer, INL = inner nuclear layer, ONL = outer nuclear layer, RPE =

retinal pigment epithelium.

Figure 5.2. Vascular and cell profiles of the choroid of GK and age-matched control Wistar

Han rats (52 W). (A) Representative images of vascular density in the inner choroid (≤ 10

µm). (B) Choroidal Iba1+ and MHC II+ cells. Ramified cells (green arrows) and round cells

(red arrows) were highlighted. (C) Quantification of the choroidal vascular density in the

inner and outer choroid using the ‘image>adjust>threshold’ window tool of ImageJ to obtain

the percentage of vascular coverage, obtained from z-stacks collected at ≤ 10 µm or > 10

µm from the outer RPE plane, respectively. (D) Iba1+ and MHC II+ cell number in the choroid

in all in-depth z-stacks. Images were collected with Zeiss EC Plan-Neofluor 40x oil objective

lens, NA 1.3. Quantitative analyses were performed based on 14 independent counts per

eye in each and all in-depth z-stacks per specimen. Data are expressed as mean ± SEM (n

= 8, control group; n = 10, GK group). Scale bar: 50 µm. Significance: *p < 0.05, **p < 0.01.

Figure 5.3. Microglial cells in the retina and choroid of GK and age-matched control Wistar

Han rats (52 W). (A) Representative eye cross-sections immunolabelled against Iba1 (left

panels), MHC-II (middle panels) and merge (right panels). Iba1+ cells are located in the

superficial and plexiform layers of the retina, mainly. Iba1+ cells located in the OPL of the

GK cohort only (green arrows). In GK rats, Iba1+ cells migrate from the IPL to the OPL,

crossing the INL (red arrow). (B) Quantification of Iba1+ and MHC II+ cell density of GK (n =

8) and age-matched control (n = 5) rats based on 12 independent specimen counts per eye.

Counting was done for the right eye only, in all animals. Data are expressed as mean ±

SEM. Scale bar: 100 µm. Significance: **p < 0.01.

GCL = ganglion cell layer, IPL = inner plexiform layer, INL = inner nuclear layer, OPL =

outer plexiform layer, ONL = outer nuclear layer, RPE = retinal pigment epithelium.

Figure 5.4. Localization of mural and endothelial cells in the retina and choroid of GK and

List of figures

XXIII

age-matched control Wistar Han rats (52 W). (A) Representative eye cross-sections

immunolabelled against RECA-1(left panels), NG2 (middle panels) and merge (right

panels). The 3 plexuses of the retina are evidenced by RECA-1 immunostaining (white

arrows). Communications between (i) the superficial and middle plexuses of the retina (red

arrow), (ii) the middle and deep plexuses and (iii) the deep and superficial plexuses (green

arrow), are visible. RECA-1 immunoreactivity relates to the presence of endothelial cells

and fluorescence of the choriocapillaris endothelial cells is continuous with the RPE cell

plane (left panels). Conversely, NG2 immunostaining of pericyte/mural cells leaves focal

gaps between the RPE and the inner choroid, drawing a jagged pattern, more pronounced

in GK rats (blue arrows). (B) Quantification of RECA-1 and NG2 immunoreactivity in the

retina and choroid of GK (n = 8) and age-matched control (n = 5) rats. RECA-1 and the

proteoglycan NG2/Cspg4 (NG2) immunoreactivities were scored as fluorescence intensity

per area selected in the choroid (reference area selected of 10,737.08 ± 6,306.11 µm2),

while NG2+ cells and RECA-1 focal immunostaining were manually counted in the retina.

Counting was done for the right eye only, in 12 independent specimen counts per eye. Data

are expressed as mean ± SEM. Scale bar: 100 µm.


outer plexiform layer, ONL = outer nuclear layer, RPE = retinal pigment epithelium.

Figure 5.5. Immunoreactivity of VEGF and VEGFR2 of GK and age-matched control Wistar

Han rats (52 W). (A) Representative eye cross-sections immunolabelled against VEGF (left

panels), VEGFR2 (middle panels) and merge (right panels). VEGF immunoreactivity

spreads throughout the retina, increasing in the OPL, OLM, RPE and choroid. VEGFR2

immunoreactivity is higher in the innermost retina (retinal nerve fiber layer) and very low or

absent in the RPE and choroid. VEGFR2 immunoreactivity is still visible as a faint coloration

in the retinal layers other than the retinal nerve fiber layer of GK rats only (white arrows).

(B) Quantification of the VEGF and VGFR2 immunoreactivity in the retina and choroid of

GK rats (n = 8) and age-matched controls (n = 5) based on 12 independent specimen counts

per eye. VEGF and VEGFR2 immunoreactivities were quantified as fluorescence

intensity/area per layer. Counting was done for the right eye only, in all animals. Data are

expressed as mean ± SEM. Scale bar: 100 µm. Significance: *p < 0.05.


outer plexiform layer, ONL = outer nuclear layer, OLM = outer limiting membrane, RPE =

retinal pigment epithelium.

Supplementary Figure 5.1. Iba1+ and MHC II+ cell count by ImageJ. (A) Iba1+ cells were

quantified in all-dept planes. (B) For each plane, MHC II+ cells were marked with yellow

List of figures

XXIV

dots to avoid duplication.

Supplementary Figure 5.2. Quantification of the choroidal vascular density in the inner

and outer choroid (defined as the percentage of total area covered by choriocapillaris

vessels) using the ‘image>adjust>threshold’ window tool of ImageJ to obtain the percentage

of vascular coverage. (A) z-stacks collected at ≤ 10 µm (choriocapillaris). (B) z-stacks

collected at > 10 µm from the outer RPE plane (medium and large vessels).

Supplementary Figure 5.3. Body weight and glycemia in GK and in age-matched control

Wistar Han rats (52 W).

List of tables

XXV

List of tables

Table 1.1. Choroidal thickness and diabetic retinopathy stage

Table 1.2. Changes in the choroidal thickness with treatment of diabetic retinopathy or

diabetic macular edema

Table 2.1. Differences in the variables considered between each endpoint and baseline

Table 2.2. Comparison of outcome measures between anatomic responders and non-

responders at baseline, 3M and 6M

Table 2.3. Stratification of the population by functional outcome

Supplementary Table 2.1. Demographic and ocular characteristics

Table 3.1. Demographic and ocular characteristics

Table 3.2. Comparison of OCT baseline characteristics and outcome measures between

functional responders and non-responders

Table 3.3. Results of the multivariate linear regression model obtained using twelve

predictors of the increase of BCVA from baseline as independent variables

Supplementary Table 3.1. Baseline values for BCVA, CRT and SFCT. Differences in

BCVA, CRT and SFCT between endpoints and baseline, and number of injections given

Supplementary Table 3.2. Comparison of outcome measures between anatomic

responders and non-responders at baseline, 3 months and 6 months

Supplementary Table 3.3. Comparison of outcome measures between anatomic

responders and non-responders using a cut-off for CRT of 350 μm

Supplementary Table 3.4. Demographic characteristics of functional responders and non-

responders

List of tables

XXVI

Supplementary Table 3.5. Logistic regression model using all predictors of BCVA increase

as independent variables entering the interaction between duration of diabetes and laser

treatment

Supplementary Table 4.1. Primary antibodies

Supplementary Table 4.2. Secondary antibodies

SupplementaryTable 4.3. Choroidal vascular density

Supplementary Table 4.4. NG2 and RECA-1 immunoreactivity in the retina and choroid

Supplementary Table 4.5. VEGF and VEGFR2 immunoreactivity in the retina and choroid

Supplementary Table 4.6. Iba1+ and MHC class II+ cell counts in the retina and choroid

based on immunolabelling of eye cryosections

Supplementary Table 4.7. Iba1+ and MHC class II+ cells density in the choroid based on

immunolabelling of sclerochoroidal whole mounts, after labelling of choroidal blood vessels

by cardiac perfusion with DiI

Supplementary Table 5.1. Primary antibodies

Supplementary Table 5.2. Secondary antibodies

Supplementary Table 5.3. Choroidal vascular density

Supplementary Table 5.4. Quantification of Iba1+ and MHC II+ cells in whole mounts of the

choroid

Supplementary Table 5.5. Quantification of Iba1+ and MHC II+ cells in cryosections of the

retina and choroid.

Supplementary Table 5.6. NG2 and RECA-1 immunoreactivity in cryosections of the retina

and choroid

Supplementary Table 5.7. VEGF and VEGFR2 immunoreactivity in cryosections of the

retina and choroid

List of videos

XXVII

List of supplementary videos

Video 4.1. Sequenced images showing the localization of Iba1+ cells (green) sparing the

innermost choroid outwards the RPE cell plane of a control rat, aged 16 weeks, at 8 weeks

after diabetes onset in diabetic rats. Slice thickness: 14.8 μm. AC and JM authored the

video: 18’’; 9,982 KB.

Video 4.2. Sequenced images showing the localization of Iba1+ cells (green) outwards the

inner choroidal vascular network (red) in a control rat aged 16 weeks, at 8 weeks after

diabetes onset in diabetic rats. Slice thickness: 23.1 μm. AC and JM authored the video:

20’’; 10,925 KB.


inner choroidal vascular network perfused using DiI (red) in a diabetic rat, at 8 weeks after

diabetes onset. Slice thickness: 23.1 μm. AC and JM authored the video: 13’’; 7,188 KB.

Video 4.4. Sequenced images in the same location as in video 4.3, showing the localization

of MHC class II+ cells (purple) outwards the choriocapillaris perfused by DiI (red) in a

diabetic rat, at 8 weeks after diabetes onset. Slice thickness: 23.1 μm. AC and JM authored

the video: 14’’; 7,433 KB.

Video 4.5. Sequenced images in the same location as in videos 4.3 and 4.4, showing the

co-localization of Iba1+ cells (green) and MHC class II+ cells (purple) outwards the inner

choroidal vascular network perfused using DiI (red) in a diabetic rat, at 8 weeks after

diabetes onset. Slice thickness: 23.1 μm. AC and JM authored the video: 13’’; 7,107 KB.


innermost choroid (red) outwards the RPE cell plane of a 52-week-old GK rat. Slice

thickness: 27.2 μm. AC and JM authored the video: 11’’; 40,221 KB.

Video 5.2. Sequenced images in the same location as in video 5.1, showing the localization

of MHC class II+ cells (cyan) outwards the choriocapillaris perfused by DiI (red) in a 52-

week-old GK rat. Slice thickness: 23.8 μm. AC and JM authored the video: 10’’; 25,306 KB.


co-localization of Iba1+ cells (green) and MHC II+ cells (cyan) outwards the choriocapillaris

List of videos

XXVIII

perfused by DiI (red) in a 52-week-old GK rat. Slice thickness: 30.6 μm. AC and JM authored

the video: 17’’; 30,465 KB.


choriocapillaris (red) in a 52-week-old control Wistar Han rat. Slice thickness: 48 μm. AC

and JM authored the video: 16’’; 51,062 KB.

Video 5.5. Sequenced images showing the localization of MHC II+ cells (cyan) outwards

the inner choroidal vascular network perfused by DiI (red) in a 52-week-old control Wistar

Han rat. Slice thickness: 45 μm. AC and JM authored the video: 16’’; 49,140 KB.


co-localization of Iba1+ cells (green) and MHC II+ cells (cyan) outwards the choriocapillaris

perfused by DiI (red) in a 52-week-old control Wistar Han rat. Slice thickness: 45 μm. AC

and JM authored the video: 16’’; 63,560 KB.

Suplementary videos available at:

https://drive.google.com/drive/folders/1YWm9HQ8ijOKu0XWj6bytF7TpG7Lp_FDo?usp=s

haring

Thesis outline

XXIX

Thesis outline

This thesis is divided in six sections, defined by numerals, and conclusions.

The first section includes a general introduction to the thesis, covering fundamental

aspects of the anatomy and physiology of the human choroid, the role of the choroidal

thickness in diabetic retinopathy and diabetic macular edema, the use of OCT in rats, the

vascular profile of the choroid in the rat model, and the description molecular and cellular

signatures in the retina and choroid of the diabetic rat.

The research purpose, which states for the motivation and scope of the thesis,

concludes this section.

The second and third sections gather the data from clinical investigation. They

include two published clinical prospective studies about the role of the baseline choroidal

thickness as a prognostic factor for diabetic macular edema (section 2) and the search for

prognostic factors of anatomic or functional outcomes in diabetic macular edema (section

3).

The fourth and fifth sections are experimental. The vascular density of the

choriocapillaris and remaining choroid, the presence of Iba+ and MHC II+ cells, alteration of

pericytes and expression of VEGF and VEGFR2 in the choroid and retina of Type 1

streptozotocin-induced 16 weeks-old diabetic rats (section 4) and of 52 weeks-old Type 2

Goto-Kakizaki diabetic rats (section 5) are considered.

The sixth section includes an integrated conclusion arising from the main results and

presents an outlook into possible future directions in this field of research.

The thesis ends with bullet-point conclusions.

Choroid and diabetes Introduction

1

1. Introduction1

1 The first part of section 1 is based on the article: Campos et al. Viewing the choroid: where we stand,

challenges and contradictions in diabetic retinopathy and diabetic macular oedema. Acta Ophthalmol. 2017;

95:446-459.


2

1.1. Diabetic retinopathy and diabetic macular edema

Diabetes mellitus (DM) has become one of the most dramatic challenges worldwide

[1]. Sedentary life, lack of exercise and overweight, are risk factors for diabetes and its

complications, including diabetic retinopathy (DR). The prevalence of Type 1 diabetes (T1D)

in the U.S. is 0.55% whereas that of Type 2 (T2D) is 8.6% (T2D encloses 94% of all

diabetics) [2]. In developed countries, DR is the leading cause of blindness in the active

population [3] and it became a burden on healthcare facilities [4, 5]. The overall prevalence

of DR is 35% and the 5-year cumulative incidence of DR for diabetics with no DR at baseline

ranges from 4% in T2D to 50% in T1D [6]. There are not many available studies assessing

the incidence of DR, but in a recent review the annual incidence of DR ranged from 2.2%

to 12.7% and the annual progression from 3.4% to 12.3% [7]. Progression is faster and

severity worse in T1D [6, 8, 9].

Progression of DR causes microvascular damage leading to increased permeability,

retinal ischemia, macular edema and neovascularization [8, 10]. Since DR may lead to loss

of vision, it is extremely important to evaluate the stage of DR in order to establish an

adequate follow-up and therapy [11]. Diabetic macular edema (DME) is the leading cause

of visual loss in patients with DR [12]. The prevalence of DME increases from 0% to 3% in

individuals recently diagnosed up to 28%-29% in those with diabetes duration of over 20

years [13]. In a UK diabetic population an overall prevalence of DME has been estimated

as 13.9% [14].

The breakdown of the inner blood–retinal barrier (BRB) is believed to be the initial

event in the development of DR [15]. However, experimental studies demonstrated that the

BRB breakdown occurs at both the inner and outer BRB [16]. Inner and outer BRB

disruption leads to the accumulation of fluid, exudation and hemorrhages, thickening of the

macular region, resulting in DME [15].

DM is a complex metabolic disorder, characterized by chronic hyperglycemia along

with dyslipidemia, hypoinsulinemia and hypertension. For a long time, DR was considered


3

to be a pure microvascular complication, but the retinal microvasculature is intimately

associated with and governed by local neurons and glia, which may be affected prior to

clinically detectable vascular lesions. DR and its complications are commonly treated with

anti-vascular endothelial growth factor (VEGF) agents [17, 18] and the greater focus has

been put on the role of VEGF on the pathogenesis of DR [19-21]. In fact, retinal hypoxemia

has been also related to the pathogenesis of DR [22, 23]. VEGF-driven BRB breakdown at

the venule side of the superficial retinal vasculature has been pointed as the earliest event

in DR [19] and it was related to pericyte loss and hypoxemia [23]. Nevertheless, retinal

hypoxemia was reported to be absent in the early stages of diabetes in rats [24]. Actually,

VEGF is not increased in the vitreous of all patients with DME while pro-inflammatory

markers were found to be increased [25]. Accordingly, about one third of patients with DME

fail to respond to anti-VEGF therapy [26, 27]. Furthermore, steroids proved to be effective

in treating post-surgical cystoid macular edema and DME [28, 29]. Those facts pointed DR

to be an inflammatory condition and that was confirmed experimentally [30-32].

Inflammation is a key player in DR [31], either in T1D or T2D [33, 34]. Retinal neuropathy

with evidence of neural apoptosis associated with DM has been reported [35, 36].

Mitochondrial superoxide production in response to hyperglycemia, rather than

hypoxemia, may be the first event to dislodge pericytes from the capillary wall. The retinal

resident innate immune system, which is primarily composed of tissue-resident

macrophage-like microglial cells, become activated and start to produce pro-inflammatory

mediators [32]. Overproduction of reactive oxygen species (ROS) leads to the increased

formation of advanced glycated end-products (AGEs), activation of protein kinase C, aldose

reductase, and nuclear factor kB, leading to pericyte loss, exposure of endothelial junction

proteins to VEGF, BRB breakdown and diabetic microangiopathies [37]. Despite the

alterations in the BRB are believed to be the main responsible for the development of DR

and DME, several studies indicate the choroid as an important player in the pathophysiology

of DR and DME, since the choroid nourishes the macula and the outer one third of the retina

[38-40]. The perspective of DR as inflammatory disease brought new attention on previous


4

Medium-size vessel layer

works describing choroidal inflammatory alterations in diabetes, named as ‘diabetic

choroidopathy’, including Brüch’s membrane deposits and increased thickness, and

choriocapillaris dropout [39, 41, 42]. Since the advent of optical coherence tomography

(OCT) [43], the choroidal thickness (CT) [44] has been sought as a surrogate of choroidal

flux, diabetic choroidopathy or DR, but the results are conflicting and disappointing to some

extent [45-49].

1.2. Anatomy and physiology of the choroid

The choroid is a highly vascularized and pigmented structure localized between the

lamina fusca of the sclera and the retinal pigment epithelium (RPE), extending anteriorly

from the ora serrata to the optic nerve posteriorly. The choroid is composed of the

choriocapillaris, the basal membrane of which forms the outer part of the 5-laminar structure

of Bruch’s membrane, the middle layer of medium-sized vessels (Sattler’s layer), the outer

layer of large vessels (Haller’s layer) originating from the short posterior ciliary arteries

(SPCA), and the suprachoroid, limited externally by the lamina fusca (Figure 1.1).

Figure 1.1. Anatomy of the choroid: In this diagram the anatomy of the choroid is depicted from the

Bruch’s membrane outwards to the suprachoroid. Artwork, courtesy from Anália do Carmo, MD,

MsC, PhD.


5

The medium-sized arteries of the Sattler’s layer give rise to the patch-like structure

of the choriocapillaris (Figure 1.2 A and B) [50].

From the Sattler’s layer, at the level of the outer choriocapillaris, there are columns

of collagen fibres running between the capillaries and attaching to the outer fibrous layer of

the Bruch’s membrane probably supporting the capillaries network [51]. The Bruch’s

membrane is a five layered structure comprehending the basement membrane of the

choriocapillaris, an outer collagenous zone, an elastic layer, an inner collagenous zone, and

the basement membrane of the RPE [52].

Figure 1.2. Expression of rat endothelial cell antibody 1 (RECA-1, Abcam 9774) in the rat

choriocapillaris, viewed from the choroidal side, visualized by laser scanning confocal microscope

LSM 710 (Zeiss, Germany), magnification 200x. A. Network of capillaries in the rat choroid and B.

Inward view with hexagonal shaped RPE cells (thin arrow) and capillaries derived from a Sattler’s

pre-terminal artery (thick arrow) [12].

The Bruch´s membrane is structurally analogous to the renal glomerulus, with a

vascular intima, a subendothelial extracellular matrix and an elastic layer equivalent to the

internal elastic layer of blood vessels. Thus, it possesses a basal membrane in its abluminal

surface (RPE’s basal membrane, a parallel to Bowman’s capsule visceral layer) and a

fenestrated vascular endothelium with its own luminal basal lamina. As in the glomerulus,


6

transport and filtration are the most important functions of Bruch’s membrane [53]. The

suprachoroid lies between the choroid and the sclera, containing fibroblasts, collagen fibres

and melanocytes. The suprachoroid has large endothelial-lined spaces receiving fluid via

the uveoscleral route and from the remaining choroid due to an oncotic gradient, and

emptying into veins [54]. The 30 μm thick outmost layer of the suprachoroid is the lamina

fusca, consisting of several layers of melanocytes and fibroblast-like cells disposed in

plates, with bundles of myelinated axons [38].

Earlier experiences in guinea pigs, rats and post-mortem examinations of newborn

infants, using Indian ink perfusions and occlusion of the short posterior ciliary arteries,

showed they behave as terminal arteries, despite the existence of numerous anastomoses

revealed by pathology studies [55]. Therefore, the choriocapillaris is not an anastomotic

network but works as a group of independent lobular units. Actually, the choriocapillaris is

a single-layered network of fine fenestrated channels about 10-20 μm thick at the fovea,

thinning to about 8 μm at the periphery [56], arranged in a lobular hexagonal pattern (Figure

1.3 A and B) [57].

Figure 1.3. The choriocapillaris and choroidal feeding vessels. A. Scanning electron micrograph of

the choriocapillaris network from vascular casts. Choroidal arteries (A) and veins (V) beneath the

coriocapillaris network (adapted from from Risco and Nopanitaya [58]). Reproduced with permission.

B. Histology of the choriocapillaris. Each feeder arteriole at the Sattler’s layer supplies a hexagonally-

shaped area of capillaries (adapted from Forrester et al. [57]). Reproduced with permission.

A B


7

The pores of those capillaries are mainly facing the RPE and are permeable to

proteins, originating a high oncotic pressure in the stroma, directing the movement of fluids

out of the retina into the choroid (Figure 1.4 A) [56, 59-61].

Nowadays, the structure of the choroidal lobules, along with arterial-arterial and

venular-venular anastomoses, and blood flowing in the same direction in arteries and veins

has been confirmed by vascular casting studies (Figure 1.4 B) [56, 62].

Figure 1.4. The choriocapillaris structure and fenestrations. A. Scanning electron micrographs of the

endothelial surface of the choriocapillaris with numerous and uniformly arranged fenestrations at the

RPE-facing side. Scale bar: 500 nm (adapted from Shimomura et al [60]). Reproduced with

permission. B. Posterior pole coriocapillaris viewed from the retinal aspect. The inferior retinal

vessels are still visible and further deep the lobular arrangement of the coriocapillaris may be

identified though difficult to distinguish. Scale bar: 250 µm (adapted from Olver, J.M. [56]).

Reproduced with permission.

The position of the artery at the periphery [55, 63, 64] or at the center of the lobule

[61, 65] was disputed. Vascular cast studies in man evidenced a lobular arrangement of the

choriocapillaris with pre-venular sinuses at the same plane of the choriocapillaris (Figure

1.5 A), while central arteries were not visible from the retinal aspect, but only from its

choroidal aspect [56, 62]. Pre-venular sinuses at the plane of the choriocapillaris are

present, but the lobular structure of the choriocapillaris is not well defined in the rat (Figure

1.5 B) [66].


8

Angiographic studies in humans and monkeys made by Hayreh, confirmed that the

SPCAs and their branches, as well as the vortex veins, have a segmental distribution in the

choroid, and that the choroidal arteries are, in fact, end-arteries. Each piece of the jig-saw

pattern has a well-defined margin which forms the watershed zone between the adjacent

SPCAs [65].

Figure 1.5. The lobular structure of the choriocapillaris in man and in the rat. A. Scanning electron

micrograph of the choriocapillaris viewed from the retinal aspect. The lobular pattern is evidenced.

Terminal parts of venules are visible as dilated channels (large fan-shaped) in the periphery of

lobules, in the choriocapillaris plane (black asterisks). Terminal arteries are difficult to be viewed in

the center of lobules from the retinal aspect. Scale bar: 250 µm. Adapted from Olver, J.M. [56].

Reproduced with permission. B. Rat choriocapillaris immunollabed with RECA-1 viewed from retinal

aspect by confocal microscopy. Pre-venular dilated sinuses at the same plane of the choriocapillaris

are visible (white asterisks), but a clear lobular arrangement of the choriocapillaris is absent. Scale

bar: 150 µm.

A watershed zone is the border between the territories of distribution of any two end-

arteries. The significance of the watershed zones is that in the event of a decrease in the

perfusion pressure in the vascular bed of one or more of the end arteries, the watershed

zone, being an area of comparatively poor vascularity, is most vulnerable to ischemia [67].

These studies also evidenced that the location of the arteriole in the choroidal lobules was

at the center, while the venules were located at the periphery (Figure 1.6).

A

*

* *

*

B

*

*


9

Figure 1.6. Diagrammatic representation of choriocapillaris and its watershed zones. A, choroidal

arteriole; V, choroidal vein (adapted from Hayreh SS [65]). Reproduced with permission.

However, conflicting results remain as to whether the choroidal circulation is of end-

arterial nature, echoing from the early days. In the choroid, as stated by Hogan et al.,

'extensive anastomoses exist between the various branches of all the short ciliary arteries,

so that occlusion of one vessel ordinarily does not produce infarction of the choroid' [68].

Conversely, Duke-Elder commented in 1961, that 'the tendency for inflammatory and

degenerative diseases of the choroid to show a considerable degree of selective

localization, despite the fact that anatomically the vessels would appear to form a

continuous network, has given rise to speculations regarding the anatomical isolation of

specific choroidal areas' [69].

Angiographic studies support the end-arterial theory [70], while post-mortem studies

revealed that the choroidal circulation has multiple anastomosis at various levels [62]. An

end-arterial functional model was proposed after experiments in cats, reconciling the post-

mortem and the angiographic findings, featuring the choroid as composed of multiple

sectors, with the presence of anastomoses within sector, but with the continuous

coriocapillaris bed being end-arterial in nature and the drainage towards the collecting


10

venules for each lobule preventing blood from crossing the boundaries among lobules

(Figure 1.7) [61, 71].

Figure 1.7. Functional end-arterial model of the choroidal circulation. Red and gray vascular

channels represent perfused and non-perfused status, respectively: (a) When a terminal arteriole is

obstructed, the choriocapillaris lobule becomes ischemic because the blood is drained through the

venous channel in the periphery of the lobule; (b) A sector becomes ischemic when the posterior

ciliary artery is occluded, as there is no arteriolar anastomosis among the other sectors; (c) When

triolein embolus flows down to a small arteriole, perfusion of the choriocapillaris is restored owing to

the extensive arteriolar anastomosis within the sector (from Lee JE et al. [71]). Reproduced with

permission.

The RPE is so intimately related with the choroid and choriocapillaris under

homeostasis and disease that photoreceptors (PRs), RPE, Brüch’s membrane and

choriocapillaris, may be considered as the tapetoretinal unit [72]. The choriocapillaris

originates from the mesenchyme but needs to be in contact with the developing RPE,

derived from the neural crest, in order to differentiate [73]. It is the VEGF secreted by the

RPE that promotes the coriocapillaris fenestrations, essential for the nourishment of the

RPE and outer retina, including the macula. The embryonic retina releases retinoic acid

which in turn promotes the RPE differentiation. The tyrosinase promoter elaborated from

PRs contributes to melanogenesis within the RPE cells which is crucial for RPE maturation.

The 11-cis-retinal elaborated by the RPE is crucial for the PRs outer segment growing.


11

Primordial PRs start to extend their outer segments and the RPE responds by elongating

its apical microvilli into the subretinal space [74]. The connection between the retina and

the RPE at the level of the outer segments of the PRs and of the microvilli of the RPE is not

mediated by any type of intercellular junctions, but only by metabolic, oncotic, hydrostatic

and electrostatic gradients (Figure 1.8).

The homeostasis and nourishment of the RPE and outer retina is intimately under

choroidal mediation as demonstrated by the age-related choroidal atrophy that goes along

with age-related macular degeneration (AMD) [75]. Histopathologic comparisons of eyes

with early AMD to age-matched controls have shown correlations between choriocapillaris

loss and drusen density [76].

Figure 1.8. Three-dimensional drawing

of the space between photoreceptor

outer segments (rods) and cells of the

retinal pigment epithelium (RPE). Thick

sheaths (a) of RPE enclose external

portions of rod outer segments without

intercellular junctions of any sort (b).

RPE finger-like villous processes (c) are

found between photoreceptors and

contain pigment granules (d). Apical

portion of RPE layer of cells at bottom

contains numerous pigment granules

(e); mitochondria (f); a well-developed,

smooth-surfaced endoplasmic reticulum

(g); a poorly developed, rough-surfaced

endoplasmic reticulum (h); and scattered

free ribosomes. Stacks of rod outer

segment discs are depicted in

longitudinal section (i) and in cross

section (j). Periphery of discs shows

scalloping (k). Microtubules originating in basal body of rod cilium extend externally into the outer

segment (l) (adapted from Hogan et al. [68]). Reproduced with permission.


12

The main physiologic function of the choroid is to provide oxygen and nutrients to

the highly metabolic outer retinal layers, namely the central avascular fovea and the

prelaminar portion of the optic nerve [77].

The choroid provides most of the blood supply the retina needs. The choroid

receives more than 70% of the ophthalmic artery blood supply, which is the highest rate per

unit weight in any tissue, while only 2% enters the retinal vessels [61]. Due to the large area

of the coriocapillaris, the speed of the flow is slowed to 77% of the flow speed in the

capillaries of the retina [78].

PRs use about 90% of oxygen delivered to the retina, mainly under mesopic and

scotopic environments. In order to bypass the Bruch’s membrane and the RPE, specific

adaptations were needed: a blood flow in the choroid ten-fold higher than in the brain, a

high oxygen tension in the choroid (arterial/venous difference of about 3% compared with

38% in the retinal circulation) and the coriocapillaris pores disposed mainly on the Bruch’s

membrane side [64, 79-82]. Furthermore, a higher hemoglobin oxygen saturation level in

the vessels of the choroid, when compared with the vessels of the retina, and further rise in

that difference with inhalation of 100% oxygen, was demonstrated in vivo in human subjects

using a non-invasive spectrophotometric oximeter [83]. Capillary fenestration dependent on

VEGF secreted by the RPE [84], allows prompt delivery of oxygen and nutrients to the outer

retina and macula and is reversed by anti-VEGF agents [60]. In addition, the choroid is of

outmost importance in temperature regulation by conveying heat [38], accumulated due to

the focused light onto the macula and due to the high metabolism of the tapetoretinal unit

[85, 86].

Despite choroidal blood flow in the fovea compensates better for an increase in

arterial blood pressure than for an increase in intraocular pressure [87], a fundamental role

of the choroid in angle closure glaucoma mediated by choroidal expansion has been

demonstrated [88]. The choroid also plays a role in the drainage of the aqueous humour via

the uveoscleral pathway. This drainage is about 35% of the total aqueous drainage and is

enhanced by atropine, epinephrine and latanoprost, but blocked by pilocarpine.


13

Additionally, non-vascular smooth muscle cells in the lamellae of the suprachoroid are

responsive to neurogenic stimulation and may account for variations in the choroidal

thickness (CT) as well as for the stabilization of the position of the fovea during

accommodation. Since there is no evidence hitherto of classic lymphatic vessels in the adult

human choroid, clearance of toxins and debris of metabolism might go through the

suprachoroid to the episcleral veins and to the vortex vein system [89].

The positive oncotic pressure at the level of the Brüch’s membrane, created by the

extravascular accumulation of large molecules, allows fluid to flow out of the retina into the

choroidal stroma and suprachoroid [90, 91]. Thus, one possible mechanism accounting for

the CT changes is the expansion of the lacunae in the suprachoroid. The lacunae expansion

is mediated by the synthesis of large proteoglycans [92], by the modulation of the size and

number of fenestrations in the choriocapillaris [93], by changes in the flux of the uveoscleral

pathway [54], by altered transport from the retina across the RPE [94] and by changes in

the tonus of the non-vascular smooth muscle of the suprachoroid [38]. Nitric oxide (NO)

synthase-positive axon terminals are found in the nonvascular smooth muscle cells of the

choroid, suggesting a role of NO in the regulation of the CT, by reducing the degree of

contraction of these cells [38].

Currently, it is considered that the choroid may contribute to the pathogenesis of

several retinal diseases. Choroidal atrophy seems to be associated with high myopic retinal

degeneration [95] and with atrophic AMD [75]. Several inflammatory conditions of the retina

are in fact choroiditis [96, 97]. It was also reported a general disturbance in the choroidal

blood flow of both eyes in central serous choroidopathy (CSC), with increased CT and

subfoveal neuroretinal detachment (SND), probably dependent on a mineralocorticoid

receptor mediation [98-100]. In diabetes, the choroid behaves as a pro-inflammatory

environment. In fact, inflammation, glial cell activation and cell migration from the retina to

the choroid, are involved in the pathogenesis of diabetic retinopathy [101, 102].


14

1.3. Optical coherence tomography (OCT)

Optical coherence tomography (OCT) was introduced as a non-invasive modality for

imaging transparent and translucent samples and tissues with a resolution of a few μm [43].

Since the eye is essentially transparent, it provides easy optical access to the retina and

therefore OCT was first investigated in ophthalmology [43, 103].

The introduction of the spectral domain (SD) principle changed the paradigm in the

OCT technology. Its improved sensitivity enabled to operate at 800 - 870 nm with a higher

acquisition speed, with light suffering minimal optical attenuation and scattering [73]. The

resolution of about 5 μm provides high-quality images of the retina, from the inner limiting

membrane (ILM) down to the RPE [104]. This wavelength is suitable to resolve all main

intra-retinal layers, but its penetration depth is limited by absorption and scattering at the

RPE level. Since the absorption of light by melanin is strongly wavelength-dependent, the

use of longer wavelengths as in Swept Source OCT (SS-OCT) may improve the penetration

depth into the choroid [105]. Unfortunately, when going towards longer wavelengths water

absorption of light increases. The advantage of a better tissue penetration, using longer

wavelength devices, is overshadowed by a poorer resolution in an organ consisting mostly

of water [73]. The signal double passing through the ocular media to the retina is

significantly attenuated (up to 50%) [106]. In addition, SS-OCT devices are large and not

easy to handle [107].

The enhanced deep imaging (EDI) is the approach developed to overcome these

difficulties. The 870 nm SD-OCT device is positioned closer to the eye in order to change

the focal point backwards to the choroid. This avoids the loss of signal caused by the RPE

[44, 108]. Moreover, a correlation has been found between data collected from the choroid

with the longer wavelength devices and the EDI procedure with the 870 nm SD-

SPECTRALIS® (Heidelberg Engineering GmbH, Heidelberg, Germany) device [109-112].


15

1.4. OCT and the choroid

Choroidal thickness was first evaluated in a focal fashion from the posterior edge of

the RPE to the choroid/sclera junction, at 500 μm intervals up to 2500 μm temporal and

nasal to the fovea [44]. The choroid is thicker in the subfoveal area and decreases to the

nasal and temporal choroid [109, 113], with the nasal CT being usually thinner [44, 114].

Margolis et al. reported a subfoveal choroidal thickness (SFCT) of 287 ± 76 μm,

using a 870 nm device in the EDI mode [44] and others reported a SFCT up to 10-20 μm

thicker, using longer wavelength devices, and found reproducibility between one another

[109, 115]. There is great variability in the CT according to age [75], refraction, and even

the time of day [44]. Previous studies, using both OCT and histologic findings, have found

statistically significant negative correlations between CT and age (decreasing CT with

increasing age) [44, 109, 116]. It was reported that SFCT decreased 1.56 - 1.95 μm for

each additional year of age [44, 117], or 15.6 μm for each decade of life [44]. Therefore, in

a 80-year lifespan, the choroid loses approximately 1/3 of its SFCT. Based on histologic

evaluation, a decrease in CT of 1.1 μm per year of age was found, which represents a rough

estimation of the actual in vivo CT [116]. SFCT also decreases with increasing axial length

(AL), 31.96 μm for each 1-mm increase in AL [117]. In addition, a person with a normal

choroid may manifest differences in thickness at intervals of a few hours or days. Unlike the

retinal thickness, it has been reported that CT shows a diurnal variation of about 30 μm

(thinnest at 6 p.m. and thickest at 3 a.m.), decreasing roughly 8% from 9 a.m. to 5 p.m.

[118, 119]. Moreover, there is inter-individual variability in CT, independently of age, AL or

time of the day, which must be taken into account when including both eyes of the same

patient (Figure 1.9).


16

Figure 1.9. Inter-individual variability in choroidal thickness (CT) evaluated by SD-Spectralis OCT,

horizontal scans encompassing the fovea. Eyes from two different patients, same age and refractive

error difference of 0.50 D, were collected. Both OCTs were obtained during the morning. Measures

of total CT were taken from the outer limit of the RPE to the choroid-scleral junction (yellow line in

the right panel and purple line in the left panel). Note that variability in CT may be about two fold from

one another.

The layers of the choroid do not behave as the layers of the retina. Unlike the retina,

wherein layers are densely stacked as regular sheets upon one another, almost like pilled

sheets of paper, in the choroid visualized by OCT, the large vessels from underneath

(Haller’s layer) press upwards the layers above (Sattler’s and choriocapillaris layers),

leaving thicker areas in-between (Figure 1.10).

Figure 1.10. The structure of the choroid, SD-Spectralis OCT (horizontal scan encompassing the

upper macular ETDRS grid area). White squares signal large choroidal vessels, with diameter ≥ 100

μm, belonging to the Haller’s layer. Large vessels push the choriocapillaris/Sattler’s complex layer

upwards leaving thick fingertip-like areas of the choriocapillaris/Sattler’s layer in-between (red

arrows). Yellow arrows signal the ‘cone of shadow’ from the large vessels of the choroid that may be

a confounding factor when determining the choroid-scleral junction.


17

This results in a generalized intertwined network of ‘hills and valleys’. This

fingerprint-like structure of the choroidal layers and the presence of the suprachoroid, make

the actual evaluation of the sublayers of the choroid and the correct identification of the

choroid-scleral border occasionally difficult.

1.5. Choroid and diabetes

Clinical and experimental findings suggest that a choroidal vasculopathy in DM may

play a role in the pathogenesis of DR. Diabetic choroidopathy (DC) was defined in DM using

indocyanine green angiography. Late phase choroidal hypoperfusion along with an inverted

inflow phenomenon have been related to DR severity and choroidal vascular resistance has

been related to retinal ischemia [40]. Large hyperfluorescent spots were associated with

high glycated hemoglobin (Hb A1c) levels and might be an indicator for choroidal

microangiopathy [120].

Histopathological studies of eyes in T2D reported decreased alkaline phosphatase

activity in the choriocapillaris, loss of viable endothelial cells, degeneration of the

choriocapillaris, obstruction and choroidal aneurysms, Brüch’s membrane degenerative

changes and choroidal neovascularization [121]. Lutty and McLeod demonstrated that the

decrease in the alkaline phosphatase enzyme activity is related to choriocapillaris loss in

DC [122]. NO synthase expression is increased in the retina even in early onset T1D [123]

or T2D [124]. The neuronal NO release in the parasympathetic perivascular nerve fibres of

the choroid may also result in a diabetes-induced neuronal damage. Therefore, DC may

encompass a microangiopathy along with a DR [125]. In fact, a previous study hypothesized

that the unexplained loss of visual acuity in diabetic patients, regardless of the inexistence

of retinopathy, might be due to DC [126]. Choroidal inflammation and ischemia may disturb

the outer BRB, leading to the accumulation of subretinal fluid or SND [127]. Indeed, there

is growing evidence indicating that the choroid is implicated in the onset of SND [40]. As

the nourishment of the macula depends on the choroid, DME with SND may be associated


18

with macular ischemia in some cases [128]. Furthermore, SND is a main feature of CSC, a

disease located primarily in the choroid [129]. Once disturbed the outer BRB, the inner BRB

would be further unbalanced since there is cross-talk between one another [16]. Despite

the alterations reported above, the cross-talk between the choriocapillaris and the BRB is

not fully understood yet.

1.5.1. Choroidal thickness in diabetes without retinopathy

When comparing CT between diabetic eyes without DR and controls, the results are

contradictory. Some studies reported that there is no difference [130-133], while other

studies reported either a significant increase [134] or, more commonly, a significant

decrease [115, 135-137].

In the Beijing Eye Study no correlation was found between DM and SFCT. However,

the range of CT reported was too wide (from 8 to 854 μm), which does introduce a bias

towards axial length (AL) dependent parameters [138]. A report issued later, using the same

sample corrected for AL and age, found DM to be related with choroidal thickening, but

added no additional risk to the DR stage. However, diabetics were only 12% of the total

number of subjects enrolled and 23 diabetic eyes only with DR (0.7%) with scarce numbers

in the different subgroups of DR [134].

Microalbuminuria has been correlated with a thinner choroid in eyes with early stage

DR [139]. Unfortunately, the study enrolled a small sample size, included both eyes, and

considered a small cut-off as significantly different for CT (15 μm). Time of day was not

taken into account and the difference of 2 diopters of refractive error between the comparing

groups might actually result in a difference in CT as great as 50 μm [140]. Considering small

differences as significant, the sensibility is overweighted at the expense of specificity. Small

differences in CT may not be independent from individual variability in CT, duration of

diabetes and hour of day for the collection of the OCT data.


19

Therefore, CT seems to remain unchanged or to decrease in diabetic eyes without

DR, but consensus is lacking.

1.5.2. Choroidal thickness and retinopathy progression

A decrease in CT in diabetic eyes unrelated to the DR stage was reported in some

studies [115, 135, 137], while others found a decrease in CT related to progression of DR,

but not with DM without DR [130-133]. CT decrease has been associated with advanced

DR stages only, either DME or proliferative diabetic retinopathy (PDR) [113].

By opposition, Kim et al. related DR progression to an increase in CT [136]. This

report had important drawbacks, therefore evidence remains elusive (Table 1.1).


20

Table 1.1. Choroidal thickness and diabetic retinopathy stage

Author N diabetic

eyes/controls

Study type Device CT and DR

staging

CT and

DME

Strength Drawbacks

Kim et al. 2013

[136]

235/36

5 subgroups

Retrospective

Cross-

sectional

SD-

Spectralis

↑ ↑ Positive correlation HbA1c/CT

195 naïves

Excluded PRP treated eyes in the

latest year

Profile of the DME: SND type

No separation between T1D and T2D

Wide age range in groups with advanced disease

status. Wide CT range within group

Inclusion of both eyes

Recent onset DME, unbalanced DM status

Esmaelpour et

al. 2011 [115]

63/16 eyes

4 subgroups

Prospective,

Cross-

sectional

1060 nm

OCT

↓ in all DM

0 for DR

staging

0 Long wave OCT

CT maps

Wide age range

Small sample per group

Both eyes included

Querques et al.

2012 [135]

63/21

3 subgroups

Prospective

Cross-

sectional

SD-

Spectralis

↓ in all DM

0 for DR

staging

↓ One eye per patient

Naïve eyes

T2D only

Small sample size

Vujosevic et al.

2012 [131]

102/48

3 subgroups

Observational

Cross-

sectional

SD-Nidek ↓ 0 Corrected refraction above 3 D

No prior treatment

No EDI

Both Type 1 and Type 2

Wide age range


21

Regatieri et al.

2012 [113]

49 /24

3 subgroups

Retrospective

Cross-

sectional

Cirrus-HD ↓ ↓ T2D only No refraction or AL correction

No EDI procedure

Small number of patients per group

Lee HK et al.

2013 [130]

203/48

4 subgroups

Cross-

sectional,

institutional

SD -

Spectralis

↓ 0 Uncontrolled hypertension

excluded

Hb A1c

Age was a confounding factor, the authors were dealing

with an aged sample and a wide age range

No mention to diabetes Type

Unsal et al.

2014 [132]

151/40

3 subgroups

Retrospective

Cross-

sectional

Optovue

RT Vue

100-2

↓ ↓ Attempt to correlate CRT with CT

and Hb A1c with CRT and CT

Both eyes included in controls and in some diabetics

No EDI

Inner limit of the choroid was clearly mismarked

Absence of naïve DME eyes

Gerendas et al

2014 [107]

284/20 Cross-

sectional

SD OCT

Cirrus

# ↓ Automated segmentation with

manual correction

Volume measurements

Reading centers (n = 2)

Small control sample

No correction for age or refraction

No correlation searched for Hb A1c

Vascular profile excludes the suprachoroid

Rewbury et al

2016 [141]

145/0

6 subgroups

Retrospective

Cross-

sectional

SD-

Spectralis

↑ ↑ NS T2D only

Treatment naïves

Both eyes included

No mention of correction for age, refraction or Hb A1c


Case et al 2016

[142]

172/57

5 subgroups

Retrospective

Cross-

sectional

SD OCT

Cirrus

↑ 0 Relation to systemic treatment

Treatment naïves

Both eyes included

T1D and T2D

Small samples in the DR late stages


22

Lains et al

2017 [46]

160/50

5 subgroups

Cross-

sectional

observational

SS - DRI

OCT-1

Atlantis

↓ 0 ETDRS Grid Sectors Both eyes included, multilevel mixed models

Two different populations included


Previous PRP and anti-VEGF treatment included

Gupta et al

2018 [48]

82/86 Cross-

sectional

SD-

Spectralis

↑ NS ↑ naïves Both eyes included

No correction for age or refraction

Mohamed et al

2019 [47]

60/30

3 subgroups

Cross-

sectional

Institutional

RS-3000

advance;

NIDEK

0 0 One eye per patient

Data collected 8 - 10 a. m.

No mention to refraction

Wide age range (30-60 yo)


Endo et al.

2020 [49]*

100/318

4 subgroups

Retrospective

Institutional

Cirrus

HD–OCT

# ↑ Hb A1c

Double organ bias, highly biased*

Small subgroups

(↑), Increase; (0), no change, (↓), decrease; AL = axial length; CT, choroidal thickness; D, diopters; DME, diabetic macular edema; DR, diabetic retinopathy; EDI, enhanced deep imaging;

NPDR, non-proliferative diabetic retinopathy; OCT, optical coherence tomography, PDR, proliferative diabetic retinopathy; PRP, panretinal photocoagulation; SND, subfoveal neuroretinal

detachment; ↓ in the fellow eye of the DME eye; NS, not statistically significant. # not searched. *Most bias were commented by me on journal site, available at:

https://journals.plos.org/plosone/article/comment?id=10.1371/annotation/a880fd21-6bf1-477a-859b-212d56395a75.


23

1.5.3. Choroidal thickness in diabetic macular edema

Most authors relate DME with decreased CT or decreased choroidal circulation

(Table 1.1) [107, 113, 132, 133, 135, 143].

Others do not confirm this finding, reporting no independent association between

DME and CT [46, 47, 115, 130, 131].

By opposition, some found that CT increases in eyes with DME and with the severity

of DR [40, 48, 136, 141]. Kim et al. found that the SFCT decreases in diabetic eyes with no

DR and with nonproliferative diabetic retinopathy (NPDR). Ischemia of the choriocapillaris

in early DC would be responsible for such decrease. Nevertheless, in the more advanced

stages of DR, where the authors found an increase in SFCT, ischemia of the choriocapillaris

is likely to be present as well. The authors theorized that an increased secretion of VEGF

would account for the increased SFCT in the more advanced stages of DR. However, such

increased secretion of VEGF is also likely to occur in the early stages of ischemia where

SFCT was reduced. Perhaps more consistent, is the possibility that the increase in the

SFCT in the more advanced stages of DR was due to the presence of SND. SND might be

related with increased CT, increased choriocapillaris permeability and outer BRB

dysfunction. These alterations occur more frequently in naïve eyes with recent DME, where

SND is more likely to occur as well.

Therefore, SND may be related to, or result from, an increased thickness of the

choroid in early onset DME. Furthermore, SND is present in CSC where CT is increased

and patients are younger [129]. Unfortunately, the work of Kim et al. has important

drawbacks (Table 1.1). Heterogeneity is clearly expressed in the wide range of variation of

CT in all groups. The CT differences within cohort (± 58 to ± 108 μm) are greater than the

differences between cohorts (14 to 73 μm), except for the panretinal photocoagulation

(PRP) group (124 μm) [136].

The role of prior focal laser photocoagulation should be considered as a confounding

factor when comparing the effect on CT caused by DME or by DME treatment. The effect

of focal laser on CT has been discarded, but there were shortcomings: small cohort, no EDI


24

procedure, scans only within 500 μm away from the fovea where laser burns would be

unlikely to be present, short follow-up and set of data from focal scans only, without

determination of an area of CT [144].

In conclusion, most of the evidence hitherto available is mainly contradictory and the

studies have important shortcomings.

1.6. The influence of treatment on choroidal thickness

1.6.1. Panretinal photocoagulation

PRP alters choroidal blood flow in patients with PDR and it was reported to increase

SFCT [145], whereas in most studies PRP was associated with a decrease in SFCT, (Table

1.2) [113, 132, 136, 146, 147].

A study evaluating choroidal blood flow in the foveal area one month after PRP,

using a laser doppler flowmetry technique, showed that PRP induces an increase in both

choroidal blood flow and choroidal blood volume shortly after PRP [148]. An increase in

SFCT can be interpreted as either an increase in choroidal blood flow due to vasodilation

or choroidal effusion, induced by choroidal vascular obstruction from laser

photocoagulation. Another study found SFCT thickening after PRP, but OCT measurements

where made as soon as one week after PRP, and may reflect the effect of choroidal effusion

rather than a real vascular thickening of the choroid [145].

Evidence for a decrease in CT long after PRP, unbiased by choroidal effusion is

more consistent. It was found a significant SFCT thinning one month after either PRP or

anti-VEGF treatment [146]. These data relating choroidal thinning with the PRP aftermath

were confirmed by other studies [113, 132, 136].

A plausible explanation for this contradiction may be the time of measuring SFCT

after PRP. Studies where SFCT is found to be thinner after PRP had treatment completed

3 months before the study, while studies where SFCT was found to be increased collected

the measurements just 1 week to 1 month after PRP. Thus, increased blood flow,

vasodilation and effusion in the retina and choroid might account for such increase. A


25

possible confounding factor is that PDR may reduce CT along with PRP, thus contributing

to SFCT reduction attributed to PRP [46, 132, 133]. However, PRP often follows shortly

after the diagnosis of PDR, so it is unlikely that PDR masks the effect of PRP on CT. Long

term follow-up of PRP-treated eyes is crucial to evaluate the effect of PRP on CT. Indeed,

a longitudinal study showed the pattern described by other studies: an increase in CT one

week after PRP with a decrease in CT beneath the baseline at 12 weeks [147].

Unchanged CT after PRP was also reported, but there were some shortcomings in

the study: small sample, the eyes were previously treated for DME, and the period of follow-

up after PRP was very short (one month) [149]. Since the first endpoint to evaluate whether

neovascularization regresses in PDR eyes treated with PRP is 3 to 4 months [150, 151], it

would be more appropriate to check for a possible alteration in the CT after PRP at least

after a similar endpoint.

Overall, most studies indicate that PRP decreases CT in the long term (after 3

months of treatment).

1.6.2. Intravitreal therapy

Most studies correlate the use of anti-VEGF agents to treat DME with a decrease in

CT (Table 1.2) [113, 132, 136, 146, 152-154]. Nevertheless, there are some few studies

contradicting this finding.

One study found that anti-VEGF therapy does not affect SFCT. However, the

number of eyes enrolled was small, the EDI protocol was not used and the choroidoscleral

interface was not clearly identified in 36% of the eyes with DME [133].


26

Table 1.2. Changes in the choroidal thickness with treatment of diabetic retinopathy or diabetic macular edema

Authors

Takahashi et

al 2008 [148]

Regatieri et

al 2012

[113]

Adhi et al

2013 [133]

Kim et al

2013 [136]

Cho et al

2013 [145]

Hwang et al

2014 [155]

Zhang et al

2015 [156]

Lains et al

2014 [152]

Lee SH et al

2014 [146]

Unsal et

al 2014

[132]

Yiu et al

2014 [153]

Sonoda et al

2014 [157]

Treatments and

Drawbacks

PRP ↑ ↓ ↓ ↓ ↑ 0 ↓ # ↓ ↓ # #

Anti-VEGF # ↓ 0 ↓ # # # ↓ ↓ ↓ ↓ 0

Steroids # # # # # # # # # # # ↓

Drawbacks Data from 1

week to 1

month after

PRP

See Table

1.1

Small

sample for

anti-VEGF

treated

eyes

As in

Table 1.1

See

Table 1.1

Short follow

up (1 week

after PRP)

Short follow up

(1 month after

PRP)

Small Sample

Eyes with

previous

treatment for

DME included

Small

sample size

3 month

follow up

No control

group

Selection of

the worst

eye to treat

See

Table 1.1

See

Table 1.1

Small sample

Eyes until -6D

Non-uniform

treatment

regimen

Non-naïve eyes

Small sample

size

3 month follow

up

Only one

injection of

either drug

(↑), Increase; (0), no change, (↓), decrease; AL = axial length; CT, choroidal thickness; D, diopters; DR, diabetic retinopathy; DME, diabetic macular edema; PRP, panretinal photocoagulation; VEGF, vascular

endothelial growth factor; D, diopters. # not searched.


27

Another study did not correlate CT with the treatment of DME, either with focal

laser or with anti-VEGF agents, but only with age. As this study found no correlation

between the severity of DME and the changes in CT, the authors concluded that the CT

is not an important factor in the pathophysiology of DME. However, this study was

retrospective, lacked a control group, included both eyes and enrolled small numbers in

each subgroup [158]. The finding of a not significant change in CT with anti-VEGF

treatment was also described in an uncontrolled, prospective, longitudinal study, with a

12-month follow-up period. Unfortunately, the number of eyes enrolled was rather small

(n = 23) and there was no available data on disease duration or previous treatments

[159].

A prospective study with a follow-up of 3 months correlated a decrease of CT

while treating DME with steroids but not with bevacizumab. However, only one injection

of either drug was used, the sample was small (n = 25 in the triamcinolone group and n

= 26 in the bevacizumab group) and the study was short-lasting [157].

It was described that the CT thins with anti-VEGF agents although with no

cumulative effect of the number of injections given [153]. Despite this study had a

longitudinal profile, with a follow-up of six months, it had some drawbacks. Eyes were

included until -6D without further correction for CT, and eyes were treated until three

months with different treatments, including PRP, focal laser and intravitreal steroids.

Additionally, it had a non-uniform treatment regimen with an average of 2.73 (range 1-6)

anti-VEGF injections over the 6 months of follow-up and a small number of eyes treated

(n = 33).

In summary, there seems to be plenty of agreement that the use of anti-VEGF

agents to treat DME decreases CT [146, 152-154], but there is some inconsistency with

respect to whether baseline CT may be taken as a prognostic factor of treatment

response.


28

1.7. Choroidal thickness as a biomarker of progression

or treatment response

There are contradictory arguments as to CT may be taken as a biomarker for DR

progression or treatment response.

Data correlating SFCT thinning with anatomic and functional outcome was

reported in a study involving DME-naïve eyes. Unfortunately there were important

drawbacks: the use of Snellen charts, exquisite small numbers (standard error used and

not standard deviation, instead), wide age range, double organ bias, an odd criterion for

anatomical outcome, inclusion of T1D and T2D, short follow-up (3 months) and no

stratification for outcome [154]. Most of this drawbacks were present in a posterior study

pointing out in the same direction [160].

The study of Yiu et al. compared DME-eyes treated with anti-VEGF versus DME

non-treated eyes and found no association between the decrease in CT with the number

of anti-VEGF injections or with the anatomic outcome in the retina. Hence, the decrease

in CT did not seem to modulate or to be a good marker of DME response to treatment,

seeming to be a side effect of treatment instead [153]. Baseline DME was more severe

in treated than non-treated eyes, that is, higher baseline central retinal thickness (CRT)

and lower baseline best corrected visual acuity (BCVA), while the CT was similar.

Therefore, CT could not be taken as a biomarker for the severity of DME. The

choroidal thinning resulting from the anti-VEGF treatment neither correlated with the

number of anti-VEGF injections nor with the improvement in BCVA or CRT. Hence, the

authors concluded that CT was not a good marker for response to treatment either. The

thinning of the choroid after the anti-VEGF therapy might just be a side effect rather than

a modulation of DME. Moreover, eyes with DME left untreated did not progress in

severity for six months, which probably was related to a less severe disease at baseline,

despite no differences in CT. More pronounced DME with higher retinal thickness was

an obvious choice to treat, but there was no difference in the baseline CT related to DME


29

severity. There were, however, confounding factors involved: none of the eyes had

history of previous anti-VEGF treatment, but several had a history of previous steroid

treatment, focal laser, PRP, or a combination, which is related to long-standing DME.

Additionally, the fact that cumulative injections did not cause further choroid thinning may

indicate the presence of long-standing DME [161].

A study by Lee et al. showed a decrease in CT with anti-VEGF therapy for DME

or DME+PDR but no correlation between CT changes and anatomic (CRT) or functional

(BCVA) outcome or the number of injections given. A correlation was found between the

decrease in CT and the first intravitreal injection of anti-VEGF, with no further reduction

after additional injections [146]. This was different from the floor effect after 3 injections

reported in AMD [162] or after 4 injections in DME [153]. As CRT increased and CT

decreased after PRP, whilst CRT and CT decreased after anti-VEGF therapy, the

authors concluded that CT was not a good biomarker of response to treatment. The study

had important drawbacks as well, including a non-uniform treatment regimen (1-3

injections), a small number of eyes in the anti-VEGF treated arm (n = 31), short follow-

up (3 months) and a low cut-off value for a significant CT decrease (5-14%).

A longitudinal study involving PRP-treated eyes, with a follow-up of 3 months,

confirmed that CT was not a good biomarker of anatomic or functional outcome shortly

after PRP. CRT increased very rapidly after PRP but took up to 12 weeks to decrease

down to the baseline level while CT increased at one week, returned to normal levels

after 4 weeks and decreased significantly at 12 weeks [147].

DME is not a homogeneous entity, as revealed by the RISE and RIDE and the

RESTORE studies. Eyes with long-standing DME (≥ 2 years) failed to reach the same

gain achieved by DME-eyes treated with anti-VEGFs right from the start [161]. In

addition, a treatment delay of one year took two additional years of treatment to close

the gap from the promptly treated [163]. Henceforth, the duration of DME will be of crucial

importance in considering outcomes and a thicker choroid might be related to early onset

DME and younger age.


30

Therefore, a thicker choroid at baseline, may not be a biomarker of treatment

response, but a biomarker of younger age or short-standing DME. The evidence hitherto

available, is mostly contradictory. Additional effort is needed to cross-check the available

contradictory evidence.

1.8. OCT angiography and the choriocapillaris

OCT angiography (OCTA) calculates the differences between B-scans that arise

from red blood cells’ movement inside vessels when the backscattering from retinal

tissue outside the vessels remains static. There are two ways to detect change in

amplitude between B-scans: (i) speckle (or intensity) decorrelation, which detects

intensity changes in OCT structural images, and (ii) phase variance, which assesses

changes in the phase of a light wave. Based on the signals, OCTA can be full-spectrum

or split-spectrum [164]. OCTA depth resolved capability and high spatial resolution (~15-

20 µm laterally, ~6 µm axially) is suitable for vasculature quantification and in vivo

imaging of the choriocapillaris. Binarization of images allows to separate choriocapillaris

stroma from the vascular bed and, therefore is most useful to study choriocapillaris

remodelling or permanent damage [165, 166]. The ability to quantify the vasculature is

attributed to the power of OCTA to resolve the microvascular networks of the retina

because the inter-capillary distances are generally larger (71.30 ± 5.17 μm) than the

system’s lateral resolution (~15-20 μm). However, we should be aware that if two vessels

are separated with a distance similar to or less than the system’s lateral resolution, then

OCTA would not be able to tell them apart. Unlike the retinal microvasculature, the

choriocapillaris is a much denser capillary network (5-20 μm of inter-capillary distance in

the posterior pole). OCTA is able to examine nonperfusion but not detailed morphological

vascular patterns, since the choriocapillaris is as thin as 10-20 µm [167]. Besides, as

OCTA imaging is based on movement of erythrocytes inside vessels, it gives no

information about the suprachoroid [168].


31

1.9. From bedside to bench

1.9.1. The animal as a model of man

While the eye in mice and rats differs from the human eye, notably by the absence

of macula, a primate-specific structure in the mammalian, animal models have

extensively been used in translational science, in the preclinical research and testing of

new drugs, due to coexisting similarities. By opposition to primates and men, the mouse

retina has a single ganglion cell layer and the average number of retinal ganglion cells

(RGCs) goes from 50.000 in mice [169] and 100.000 in rats [170] to 500,000 – 1,000,000

in the primate retina [171, 172] and 700,000 – 1,500,000 in human’s, with an the overall

cone:ganglion cell ratio ranging from 2.9 to 7.5 [173, 174]. This huge difference is

explained in a great extent by the fact that half of the RGCs are related to the fovea, a

structure absent in rats. The fovea, mediates high acuity vision. Although it occupies

<1% of the retinal surface, it provides 50% of the input to the visual cortex (Figure 1.11)

[172, 175]. As for rat and man, the fovea and the peripheral retina relate one another in

many ways. Cell types from the fovea and periphery are genetically related, therefore it

is cell proportion and not type, that makes the greater difference between central and

peripheral retina [172]. Peripheral RGCs are supplied by dozens to hundreds of PRs

(rods and cones) each, enhancing sensitivity, whereas most RGCs in the fovea are

supplied by a single PR, which maximizes acuity [174]. While cell type conservation is

present between mouse and macaque at the level of the PRs with genetic analogy, the

major RGC types in primates lack clear counterparts in rats. Similarly, the molecular

distinctions between the fovea and the periphery are higher for RGCs than for PRs. This

profile suggests that the outer retina may comprise a conserved set of information

processors, while adaptation to species- and region-specific visual-processing

requirements begin at the level of RGCs [172]. Moreover, basic aspects of retinal

structure and function are shared by all retinal regions in all vertebrate species. PRs

convert light into electrical signals and pass them to interneurons (HCs, horizontal cells;


32

BCs, bipolar cells; ACs, amacrine cells). Interneurons process the information and

transfer it to RGCs, whose axons extend to the brain [176]. Selective type-specific

connectivity among interneurons and RGC types make each RGC type responsive to a

specific visual feature (Figure 1.11) [177].

Figure 1.11. Sketches of a primate eye showing position of fovea and macula and the peripheral

retina. A. Central region indicating diameters of the foveola (the foveal pit), fovea, and macula. At

the foveola there is displacement of ganglion cell layer (GCL) and inner nuclear layer (INL) cells.

B. Sketch of peripheral retina showing its major cell classes - photoreceptors (PRs), horizontal

cells (HCs), bipolar cells (BCs), amacrine cells (ACs), retinal ganglion cells (RGCs) and Müller

glia (MG), outer and inner plexiform (synaptic) layers (OPL and IPL), outer and inner nuclear

layers (ONL and INL), and ganglion cell layer (GCL), (adapted from Peng et al., 2019 [172]).

Reproduced with permission.

Although rats are not blueprints for the diseased humans, they provide excellent

insights into the pathogenesis of diabetes [178]. Furthermore, rats are recognized as the

preeminent model for studying the choroid [52, 179]. Rats are the most used model for

studying human diseases and their choroidal structure is closest to human’s than

rabbits’, cats’ or guinea pigs’ [52, 179].

A B


33

1.9.2. OCT in the animal

OCT has been used in animals to check for retinal changes in diabetes, but not

for changes in the choroid [180]. OCT thickness measurements of the retinal layers have

been previously performed with 830 - 840 nm SD-OCTs customized for retinal imaging

of small animals, mainly in mice and rats [181, 182]. The choroid is better visualized and

the CT can be measured in vivo in the anesthetized animal by making a manual EDI

technique as described in section 1.3 [108]. The OCT is placed closer to the eye such

that an inverted image is obtained and the deeper structures are placed closer to zero-

delay (Figure 1.12).

Figure 1.12. Choroidal thickness collected by SD-OCT in the rat using the EDI technique. On the

left panel, inverted image of the retina (A) and choroid (C) separated by the bright hyper-reflective

line of the RPE (B). While the inner boundary of the choroid follows the linear outline of the RPE,

the outer boundary swings in and out according to the vascular profile of the large vessels. On

the right panel, image shows the position where the scan was acquired.

The 830nm SD-OCT Imagine System (Phoenix Micron IV, Phoenix Research

Labs, Pleasanton, CA), contains an OCT engine, a scan head, and a computer with

software to detect and photograph the retina. The SD-OCT engine has a spectrometer

covering 740 – 920 nm and is combined with a broadband super-luminescent diode with

a 3 dB bandwidth of >150 nm. The spectrometer runs an A-scan rate of 40,000 lines per

second and contains 2048 pixels. The scan mode length is about 1.8 mm on the X and

Y axes in rats and may be displaced or rotated within the entire scan region. The SD-

A

BC


34

OCT system is able to capture 10,000 – 20,000 A scans per second with an axial

resolution of 2 µm and a transverse resolution of 4 µm.

Thinning of the retina associated with a pre-diabetic state in the animal has been

reported [183], but research about CT in the animal with DM is lacking. For each location

the device collects a set of 1024 raster scans along the scan length. By averaging a set

of five images most of the noise observed on individual images is dramatically reduced.

Automatic measurements of the retinal layers or of the inner and outer choroid

boundaries are possible with the automatic segmentation software InSight (Phoenix

Research Labs), making manual corrections of the boundary lines when necessary

(Figure 1.13).

Figure 1.13. Automatic determination of the choroid boundaries from the segmentation software

InSight (Phoenix Research Labs) with manual correction. On the left side image, the green line

marks the inner limits of the choroid, while the blue line marks its outer limits, that is, the choroid-

scleral border. Top right side image is the image of the fundus where the scan was collected.

Bottom right side image shows a green line of choroidal thickness magnitude resulting from

collection of the 1024 raster scans all along the raster scan length.


35

1.9.3. Visualization of the choroidal vasculature

Current methods to readily visualize the choroidal vessels in rats present

challenges as they result in incomplete labeling of choroidal capillaries, have nonspecific

staining, or are outside the visible spectrum. Recently, the fluorescent DyLight-594

conjugated tomato lectin (Lycopersicon esculentum agglutinin) has been used as a

surrogate markers for endothelial cells in mice and rats [184]. However, direct labelling

of blood vessels in the retina or in the choroid, by cardiac perfusion using a specially

formulated aqueous solution containing 1,1’-dioctadecyl-3,3,3’,3’-

tetramethylindocarbocyanine perchlorate (DiI), a lipophilic carbocyanine dye, which

incorporates into endothelial cell membranes upon contact, has an easier protocol and

achieves higher signal intensities that can be achieved by indirect labelling methods

[185] (Figures 1.14 and 1.15).

Figure 1.14. Terminal choroidal arterioles visualized by DiI (red) in a 16-week Wistar rat. A.

Arteriole dividing in pre-terminal arteries that originate the multiple lobular network of the

choriocapillaries. Scale bar: 100 µm. 10x Full size: x: 850.19 µm, y: 850.19 µm. B. Pre-terminal

artery giving rise to the choriocapillaris at a right angle after a short trajectory as previously

described in vascular casts [56, 61]. Scale bar: 50 µm. 20x Full size: x: 425.1 µm, y: 425.1 µm.

C. The lobular pattern is not self-evident in the coriocapillaris. Instead, a honeycomb-like pattern

is depicted. Scale bar: 50 µm. 20x Full size: x: 425.10 µm, y: 425.10 µm.

A B C


36

Figure 1.15. Composed visualization of the whole choroid in a 16-week Wistar rat by DiI. The

vascular profile exhibited is arterial-type with absence of the collecting vortex veins. A. Orientation

of the choroid: S = superior, N = nasal, I = inferior, T = temporal. B. Display of the whole choroidal

circulation resulting from gathering all orientated flat mounts. The choroidal circulation is arterial

end-terminal with pre-terminal arterial-arterial anastomoses, branching in a bronchiolar-like

pattern. C. Detail shows that anastomosis are frequent before the emergence of the pre-terminal

arteries (white arrows). Scale bar: A; 2mm, B; 1 mm and C; 0.2 mm.

S

N T

I

A

B

C


37

1.9.4. Regulation, remodelling and inflammation

1.9.4.1. Pericytes and mural cells

Perivascular mural cells, either smooth muscle cells or pericytes, contribute to

regulation and homeostasis of endothelial functions, sometimes responding differently

to vasoactive elements, such as adenosine, epinephrine, phenylephrine, aspartate,

glutamate, insulin, NO, endothelins, prostaglandins and others [186]. As there is a lack

of strictly pericyte-specific markers, the unambiguous identification of these cells often

requires immunostaining of multiple antigens or careful analysis of morphological criteria.

Pericytes directly contact capillary endothelial cells sharing a common basement

membrane. Pericytes share certain molecular markers, such as expression of the

proteoglycan NG2/Cspg4 or the intermediate filament protein desmin, with perivascular

smooth muscle cells. The latter, however, cover larger calibre arteries and veins, and

are separated by the subendothelial basement membrane from the underlying

endothelium monolayer. Pericytes emerge as important regulators of endothelial

sprouting and branch formation, mediated by the modulation of VEGF-A/vascular

endothelial receptor (VEGFR) 2 signaling activity via the expression of the receptor 1

(VEGFR1), mostly a decoy receptor [187]. There is evidence that brain, retinal and

choroidal pericytes derive, at least in part, from neural crests. VEGF-A and the related

placental growth factor (PlGF) have been shown to induce pericyte ablation. Their role

in the development and maintenance of the BRB has been demonstrated, with pericyte

depletion being associated with endothelial hyperplasia, microaneurism formation and

increased permeability [188]. Recently, pericytes were associated with blood flow control

within the deep and intermediate retinal capillary plexuses in a very similar modus as

their regulation of vasa recta in the kidney [189].

Disposition of pericytes at the choriocapillaris was described in intact

sclerochoroidal whole mounts of albino transgenic mice to be scanty and at its scleral

surface only (polarized distribution) and focal pericyte depletion has been related to


38

vascular remodeling. At the Sattler and Haller layers, pericytes or mural cells wrap

around vessels, showing contractile properties (Figure 1.16) [190].

Figure 1.16. Pericytes and perivascular mural cells in the choriocapillaris and middle choroid of

a 16-week Wistar rat show distint morphology distribution. A. Perivascular mural cells

immunostained by desmin wrap around choroidal vessels, while they assume a linear or stellate

configuration at the choriocapillaris, corresponding to the scanty non-circumferential distribution

of pericytes (yellow arrows). B. Linear immunomarking of pericytes by desmin near the hexagonal

RPE cells’ plane (yellow thick arrows) show a scanty non-circumferential distribution. C. Distinct

morphology of mural cells immunomarked by desmin wrapping around choroidal vessels while

pericytes show a linear morphology and scanty non-circumferential distribution at the

choriocapillaris level (yellow thick arrow). Scale bar: 50 µm. 10x Full size: x: 850.19 µm, y: 850.19

µm.

Although pericyte coverage of human capillaris decreases from 90% in the retina

to 11% in the choriocapillaris, perivascular mural cells located in close proximity to

endothelial cells of choroidal vessels are thought to play a role in regulating blood flow

and hence controlling metabolic supply to the outer retina. Inadequacy of this metabolic

supply (otherwise termed ‘‘choroidal insufficiency’’) has been hypothesized to be present

in retinal diseases [191], including in the formation of choroidal neovascularization (CNV)

[192], in the resistance to anti-VEGF agents [190] and in the impaired development of

the outer nuclear layer (ONL) in embryonic days [193].

A Desmin DesminB CDesmin NG2A B C


39

SCP

1.9.4.2. Glia and the neuro-vascular unit

The retinal neurovascular unit is composed of vasculature (namely the capillary

endothelia), neurons, glial cells, pericytes, and smooth muscle cells. The overriding

purpose of the neurovascular unit is to optimize regional metabolic efficiency (e.g., to

increase blood flow to sites of increased cellular activity and to preserve flow at sites of

low activity) [189]. There are three types of glial cells in the retina: astrocytes located in

the inner retina, vitreous surface and around vessels of the superficial vascular plexus;

microglial ramified cells located only in the inner retina and around vessels and Müller

cells (MC), also named Müller glia (MG) or retinal Müller glial cells (RMG), that are the

only cells that span the entire thickness of the sensory retina, from the ILM, which is the

basal membrane of the MC, to the outer limiting membrane (OLM), formed by cellular

connexions between MC and PRs inner segments (Figure 1.17).

Figure 1.17. Cellular components of the inner blood-retinal barrier. Schematic representation of

the neuro-glio-vascular unit forming the inner blood-retinal barrier, composed by vascular

endothelial cells, pericytes (p), retinal Müller glial (RMG) cells, astrocytes (a), microglia (mc). RMG

cell projections are present at the level of all retinal vascular plexuses (superficial, SCP;

intermediate, ICP and deep, DCP), while astrocytes are only present at the level of the superficial

plexus. Adapted from Daruich et al. 2019 [188]. Reproduced with permission.


40

Microglia cells are distinguished in “surveillant” phenotype with highly ramified

units and “activated” (or ameboid) state with larger cell bodies and thicker processes.

They are the main players in neuro-vascular coupling in the retina (Figure 1.18). They

have a fundamental role in homeostasis, keeping a fine tuned environment for neuronal

survival in the retina, via regulating the blood flow through cytokines and pericytes, and

by removing neuronal debris [32, 189, 194].

Figure 1.18. Presence of glial cells around vessels in the retina and in the choroid. A. Retinal

projection showing vessels RECA1+ (green) and Iba+ cells/microglia (red) mostly in vessels’

vicinity. B. Choroid single confocal plane visualized from the RPE side, showing choroidal

medium-sized and large vessels surrounded by Iba+ cells (white arrows). Laser scanning

microscope LSM 710 (Zeiss), objective lens: 20x, numerical aperture 0.8, magnification 200x.

Scale bar = 50 µm.

Under the low level of chronic inflammation in the diabetic retina due to

hyperglycemia, dyslipidemia and oxidative stress, tissue-resident macrophage-like

microglia cells, become activated, changing their morphologic appearance from

‘dendritic’ to ‘ameboid’, and start to produce proinflammatory mediators [32]. Glutamate,

metalloproteases and nitrous oxide produced by activated microglia, lead to neuronal

cell dysfunction, damage capillary pericytes and endothelial cells, exposing endothelia

to VEGF [195]. Migration of activated microglial cells from the retina to the choroid by

A B


41

transcytosis has been demonstrated in diabetes [102, 196]. However, the role of glia on

vascular remodelling of the choriocapillaris, on regulation of the choroidal blood flow and

interaction with choroidal pericytes/perivascular mural cells remains to be established.

1.6. Objectives of the present study

There are controversial data regarding the value of the CT as a marker of

prognosis in DME. Cellular and molecular signatures occurring simultaneously in the

choroid and retina in diabetes are not fully understood. The aims of this clinical and

experimental work are:

1 – In human subjects, prospectively, to find whether the CT might be a marker of

outcome of DME under anti-VEGF treatment and to test the reliability of SFCT as a

marker of CT (chapter 2).

2 - In human subjects, prospectively, to find markers of DME outcome other than CT or

SFCT, including independent demographic, metabolic and OCT parameters (chapter 3).

3 – To test whether there are CT changes in vivo, in two animal models of diabetes, T1D

(chapter 4) and T2D (chapter 5).

4 – To evaluate the impact of experimental T1D (Chapter 4) and T2D (Chapter 5) on glial

cells, pericytes, endothelial cells, VEGFR 2 and VEGF in the choroid and retina.

5 – To evaluate the impact of experimental T1D (chapter 4) and T2D (chapter 5) on

vascular density (relation vessel/stroma) in the choroid of sclerochoroidal whole mounts

assessed by confocal microscopy.

Acknowledgements: I am extremely grateful to Anália Carmo, MD, PhD, for drawing

Figure 1.1 for me and to Elisa J. Campos, PhD, for the adaptation of Figure 1.17.


42

1.7. References

1. Wild, S., et al., Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care, 2004. 27(5): p. 1047-53.

2. Bullard, K.M., et al., Prevalence of Diagnosed Diabetes in Adults by Diabetes Type - United States, 2016. MMWR Morb Mortal Wkly Rep, 2018. 67(12): p. 359-361.

3. Williams, R., et al., Epidemiology of diabetic retinopathy and macular oedema: a systematic review. Eye (Lond), 2004. 18(10): p. 963-83.

4. Brown, M.M., et al., Utility values and diabetic retinopathy. Am J Ophthalmol, 1999. 128(3): p. 324-30.

5. Javitt, J.C. and L.P. Aiello, Cost-effectiveness of detecting and treating diabetic retinopathy. Ann Intern Med, 1996. 124(1 Pt 2): p. 164-9.

6. Ting, D.S., G.C. Cheung, and T.Y. Wong, Diabetic retinopathy: global prevalence, major risk factors, screening practices and public health challenges: a review. Clin Exp Ophthalmol, 2016. 44(4): p. 260-77.

7. Sabanayagam, C., et al., Incidence and progression of diabetic retinopathy: a systematic review. Lancet Diabetes Endocrinol, 2019. 7(2): p. 140-149.

8. Klein, R., et al., The Wisconsin Epidemiologic Study of diabetic retinopathy. XIV. Ten-year incidence and progression of diabetic retinopathy. Arch Ophthalmol, 1994. 112(9): p. 1217-28.

9. Romero-Aroca, P., et al., Differences in incidence of diabetic retinopathy between type 1 and 2 diabetes mellitus: a nine-year follow-up study. Br J Ophthalmol, 2017. 101(10): p. 1346-1351.

10. Engerman, R.L. and T.S. Kern, Retinopathy in animal models of diabetes. Diabetes Metab Rev, 1995. 11(2): p. 109-20.

11. Klein, R., B.E. Klein, and S.E. Moss, Visual impairment in diabetes. Ophthalmology, 1984. 91(1): p. 1-9.

12. Moss, S.E., R. Klein, and B.E. Klein, Ten-year incidence of visual loss in a diabetic population. Ophthalmology, 1994. 101(6): p. 1061-70.

13. Klein, R., et al., The Wisconsin epidemiologic study of diabetic retinopathy. IV. Diabetic macular edema. Ophthalmology, 1984. 91(12): p. 1464-74.

14. Keenan, T.D., et al., United Kingdom National Ophthalmology Database Study: Diabetic Retinopathy; Report 1: prevalence of centre-involving diabetic macular oedema and other grades of maculopathy and retinopathy in hospital eye services. Eye (Lond), 2013. 27(12): p. 1397-404.

15. Cunha-Vaz, J., J.R. Faria de Abreu, and A.J. Campos, Early breakdown of the blood-retinal barrier in diabetes. Br J Ophthalmol, 1975. 59(11): p. 649-56.

16. Bento, C.F., et al., Methylglyoxal-induced imbalance in the ratio of vascular endothelial growth factor to angiopoietin 2 secreted by retinal pigment epithelial cells leads to endothelial dysfunction. Exp Physiol, 2010. 95(9): p. 955-70.

17. Wells, J.A., et al., Aflibercept, Bevacizumab, or Ranibizumab for Diabetic Macular Edema: Two-Year Results from a Comparative Effectiveness Randomized Clinical Trial. Ophthalmology, 2016. 123(6): p. 1351-9.

18. Gross, J.G., et al., Five-Year Outcomes of Panretinal Photocoagulation vs Intravitreous Ranibizumab for Proliferative Diabetic Retinopathy: A Randomized Clinical Trial. JAMA Ophthalmol, 2018. 136(10): p. 1138-1148.

19. Qaum, T., et al., VEGF-initiated blood-retinal barrier breakdown in early diabetes. Invest Ophthalmol Vis Sci, 2001. 42(10): p. 2408-13.


43

20. Baharivand, N., et al., Relationship between vitreous and serum vascular endothelial growth factor levels, control of diabetes and microalbuminuria in proliferative diabetic retinopathy. Clin Ophthalmol, 2012. 6: p. 185-91.

21. Shimada, H., et al., Concentration gradient of vascular endothelial growth factor in the vitreous of eyes with diabetic macular edema. Invest Ophthalmol Vis Sci, 2009. 50(6): p. 2953-5.

22. Arden, G.B. and S. Sivaprasad, Hypoxia and oxidative stress in the causation of diabetic retinopathy. Curr Diabetes Rev, 2011. 7(5): p. 291-304.

23. Hammes, H.P., et al., Pericytes and the pathogenesis of diabetic retinopathy. Diabetes, 2002. 51(10): p. 3107-12.

24. Lau, J.C. and R.A. Linsenmeier, Increased intraretinal PO2 in short-term diabetic rats. Diabetes, 2014. 63(12): p. 4338-42.

25. Funatsu, H., et al., Vitreous levels of interleukin-6 and vascular endothelial growth factor are related to diabetic macular edema. Ophthalmology, 2003. 110(9): p. 1690-6.

26. Diabetic Retinopathy Clinical Research, N., et al., Randomized trial evaluating ranibizumab plus prompt or deferred laser or triamcinolone plus prompt laser for diabetic macular edema. Ophthalmology, 2010. 117(6): p. 1064-1077 e35.

27. Gonzalez, V.H., et al., Early and Long-Term Responses to Anti-Vascular Endothelial Growth Factor Therapy in Diabetic Macular Edema: Analysis of Protocol I Data. Am J Ophthalmol, 2016. 172: p. 72-79.

28. Campos, A., et al., Deferred intravitreal triamcinolone in diabetic eyes after phacoemulsification. J Ocul Pharmacol Ther, 2014. 30(9): p. 717-28.

29. Bellocq, D., et al., The Pattern of Recurrence in Diabetic Macular Edema Treated by Dexamethasone Implant: The PREDIAMEX Study. Ophthalmol Retina, 2018. 2(6): p. 567-573.

30. Adamis, A.P., Is diabetic retinopathy an inflammatory disease? Br J Ophthalmol, 2002. 86(4): p. 363-5.

31. Rubsam, A., S. Parikh, and P.E. Fort, Role of Inflammation in Diabetic Retinopathy. Int J Mol Sci, 2018. 19(4).

32. Sorrentino, F.S., et al., The importance of glial cells in the homeostasis of the retinal microenvironment and their pivotal role in the course of diabetic retinopathy. Life Sci, 2016. 162: p. 54-9.

33. Clark, M., C.J. Kroger, and R.M. Tisch, Type 1 Diabetes: A Chronic Anti-Self-Inflammatory Response. Front Immunol, 2017. 8: p. 1898.

34. Wellen, K.E. and G.S. Hotamisligil, Inflammation, stress, and diabetes. J Clin Invest, 2005. 115(5): p. 1111-9.

35. Antonetti, D.A., R. Klein, and T.W. Gardner, Diabetic Retinopathy. New England Journal of Medicine, 2012. 366(13): p. 1227-1239.

36. Barber, A.J., et al., Neural apoptosis in the retina during experimental and human diabetes. Early onset and effect of insulin. J Clin Invest, 1998. 102(4): p. 783-91.

37. Nishikawa, T., et al., Normalizing mitochondrial superoxide production blocks three pathways of hyperglycaemic damage. Nature, 2000. 404(6779): p. 787-790.

38. Nickla, D.L. and J. Wallman, The multifunctional choroid. Prog Retin Eye Res, 2010. 29(2): p. 144-68.

39. Hidayat, A.A. and B.S. Fine, Diabetic choroidopathy. Light and electron microscopic observations of seven cases. Ophthalmology, 1985. 92(4): p. 512-22.

40. Hua, R., et al., Imaging evidence of diabetic choroidopathy in vivo: angiographic pathoanatomy and choroidal-enhanced depth imaging. PLoS One, 2013. 8(12): p. e83494.


44

41. Fryczkowski, A.W., S.E. Sato, and B.L. Hodes, Changes in the diabetic choroidal vasculature: scanning electron microscopy findings. Ann Ophthalmol, 1988. 20(8): p. 299-305.

42. Lutty, G.A., Effects of diabetes on the eye. Invest Ophthalmol Vis Sci, 2013. 54(14): p. ORSF81-7.

43. Huang, D., et al., Optical coherence tomography. Science, 1991. 254(5035): p. 1178-81. 44. Margolis, R. and R.F. Spaide, A pilot study of enhanced depth imaging optical coherence

tomography of the choroid in normal eyes. Am J Ophthalmol, 2009. 147(5): p. 811-5. 45. Campos, A., et al., Viewing the choroid: where we stand, challenges and contradictions

in diabetic retinopathy and diabetic macular oedema. Acta Ophthalmol, 2017. 95(5): p. 446-459.

46. Lains, I., et al., Choroidal Thickness in Diabetic Retinopathy Assessed with Swept-Source Optical Coherence Tomography. Retina, 2018. 38(1): p. 173-182.

47. Mohamed, D.M.F., et al., Subfoveal choroidal thickness in diabetic macular edema. Clin Ophthalmol, 2019. 13: p. 921-925.

48. Gupta, C., et al., Choroidal structural analysis in eyes with diabetic retinopathy and diabetic macular edema-A novel OCT based imaging biomarker. PLoS One, 2018. 13(12): p. e0207435.

49. Endo, H., et al., Relationship between diabetic macular edema and choroidal layer thickness. PLoS One, 2020. 15(1): p. e0226630.

50. Hayreh, S.S., Segmental nature of the choroidal vasculature. Br J Ophthalmol, 1975. 59(11): p. 631-48.

51. Krebs, W. and I.P. Krebs, Ultrastructural evidence for lymphatic capillaries in the primate choroid. Arch Ophthalmol, 1988. 106(11): p. 1615-6.

52. Castro-Correia, J., Understanding the choroid. Int Ophthalmol, 1995. 19(3): p. 135-47. 53. Curcio, C.A. and M. Johnson, Structure, Function, and Pathology of Bruch's Membrane

in Retina, S.J. Ryan, Editor. 2013, Elsevier: Philadelphia, PA. p. 465–81. 54. Alm, A. and S.F. Nilsson, Uveoscleral outflow--a review. Exp Eye Res, 2009. 88(4): p. 760-

8. 55. Castro Correia, J., Anátomo-fisiologia da coroide. Thesis Doctorat, 1958: p. 52-126. 56. Olver, J.M., Functional anatomy of the choroidal circulation: methyl methacrylate

casting of human choroid. Eye (Lond), 1990. 4 ( Pt 2): p. 262-72. 57. Forrester J. V., et al., Anatomy of the eye and orbit, in The Eye (Fourth Edition), W.B.

Saunders, Editor. 2016, Saunders/Elsevier Edinburgh; New York. p. 1-102. 58. Risco, J.M. and W. Nopanitaya, Ocular microcirculation. Scanning electron microscopic

study. Invest Ophthalmol Vis Sci, 1980. 19(1): p. 5-12. 59. McLeod, D.S., et al., Relationship between RPE and choriocapillaris in age-related

macular degeneration. Invest Ophthalmol Vis Sci, 2009. 50(10): p. 4982-91. 60. Shimomura, Y., et al., Changes in choriocapillaris fenestration of rat eyes after

intravitreal bevacizumab injection. Graefes Arch Clin Exp Ophthalmol, 2009. 247(8): p. 1089-94.

61. Torczynski, E. and M.O. Tso, The architecture of the choriocapillaris at the posterior pole. Am J Ophthalmol, 1976. 81(4): p. 428-40.

62. Yoneya, S., M.O. Tso, and K. Shimizu, Patterns of the choriocapillaris. A method to study the choroidal vasculature of the enucleated human eye. Int Ophthalmol, 1983. 6(2): p. 95-9.

63. Krey, H.F., Segmental vascular patterns of the choriocapillaris. Am J Ophthalmol, 1975. 80(2): p. 198-202.

64. McLeod, D.S. and G.A. Lutty, High-resolution histologic analysis of the human choroidal vasculature. Invest Ophthalmol Vis Sci, 1994. 35(11): p. 3799-811.


45

65. Hayreh, S.S., The choriocapillaris. Albrecht Von Graefes Arch Klin Exp Ophthalmol, 1974. 192(3): p. 165-79.

66. Bhutto, I.A. and T. Amemiya, Microvascular architecture of the rat choroid: corrosion cast study. Anat Rec, 2001. 264(1): p. 63-71.

67. Hayreh, S.S., Posterior ciliary artery circulation in health and disease: the Weisenfeld lecture. Invest Ophthalmol Vis Sci, 2004. 45(3): p. 749-57; 748.

68. Hogan, M.J., J.A. Alvarado, and J.E. Weddell, Histology of the Human Eye — An Atlas and Textbook 1971, Philadelphia, London, Toronto: Philadelphia: W. B. Saunders. 403-497.

69. Duke-Elder, S., System of Ophthalmology. Vol. 2. 1961, London: Kimpton. 339, 346, 351. 70. Hayreh, S.S., In vivo choroidal circulation and its watershed zones. Eye (Lond), 1990. 4 (

Pt 2): p. 273-89. 71. Lee, J.E., et al., Functional end-arterial circulation of the choroid assessed by using fat

embolism and electric circuit simulation. Sci Rep, 2017. 7(1): p. 2490. 72. Dzhemileva, L.U., et al., [Molecular-genetic basis of tapetoretinal degeneration]. Mol

Biol (Mosk), 2008. 42(1): p. 3-11. 73. Mrejen, S. and R.F. Spaide, Optical coherence tomography: imaging of the choroid and

beyond. Surv Ophthalmol, 2013. 58(5): p. 387-429. 74. Strauss, O., The retinal pigment epithelium in visual function. Physiol Rev, 2005. 85(3):

p. 845-81. 75. Spaide, R.F., Age-related choroidal atrophy. Am J Ophthalmol, 2009. 147(5): p. 801-10. 76. Sohrab, M., K. Wu, and A.A. Fawzi, A pilot study of morphometric analysis of choroidal

vasculature in vivo, using en face optical coherence tomography. PLoS One, 2012. 7(11): p. e48631.

77. Hayreh, S.S., Blood flow in the optic nerve head and factors that may influence it. Prog Retin Eye Res, 2001. 20(5): p. 595-624.

78. Wajer, S.D., et al., Velocity measurements of normal and sickle red blood cells in the rat retinal and choroidal vasculatures. Microvasc Res, 2000. 60(3): p. 281-93.

79. Linsenmeier, R.A. and R.D. Braun, Oxygen distribution and consumption in the cat retina during normoxia and hypoxemia. J Gen Physiol, 1992. 99(2): p. 177-97.

80. Yu, D.Y. and S.J. Cringle, Oxygen distribution and consumption within the retina in vascularised and avascular retinas and in animal models of retinal disease. Prog Retin Eye Res, 2001. 20(2): p. 175-208.

81. Spitznas, M. and E. Reale, Fracture faces of fenestrations and junctions of endothelial cells in human choroidal vessels. Invest Ophthalmol, 1975. 14(2): p. 98-107.

82. Federman, J.L., The fenestrations of the choriocapillaris in the presence of choroidal melanoma. Trans Am Ophthalmol Soc, 1982. 80: p. 498-516.

83. Kristjansdottir, J.V., et al., Choroidal oximetry with a noninvasive spectrophotometric oximeter. Invest Ophthalmol Vis Sci, 2013. 54(5): p. 3234-9.

84. Blaauwgeers, H.G., et al., Polarized vascular endothelial growth factor secretion by human retinal pigment epithelium and localization of vascular endothelial growth factor receptors on the inner choriocapillaris. Evidence for a trophic paracrine relation. Am J Pathol, 1999. 155(2): p. 421-8.

85. Wangsa-Wirawan, N.D. and R.A. Linsenmeier, Retinal oxygen: fundamental and clinical aspects. Arch Ophthalmol, 2003. 121(4): p. 547-57.

86. Parver, L.M., C. Auker, and D.O. Carpenter, Choroidal blood flow as a heat dissipating mechanism in the macula. Am J Ophthalmol, 1980. 89(5): p. 641-6.

87. Polska, E., et al., Regulation of choroidal blood flow during combined changes in intraocular pressure and arterial blood pressure. Invest Ophthalmol Vis Sci, 2007. 48(8): p. 3768-74.

88. Quigley, H.A., Angle-closure glaucoma-simpler answers to complex mechanisms: LXVI Edward Jackson Memorial Lecture. Am J Ophthalmol, 2009. 148(5): p. 657-669 e1.


46

89. Schroedl, F., et al., Consensus statement on the immunohistochemical detection of ocular lymphatic vessels. Invest Ophthalmol Vis Sci, 2014. 55(10): p. 6440-2.

90. Saint-Geniez, M., et al., An essential role for RPE-derived soluble VEGF in the maintenance of the choriocapillaris. Proc Natl Acad Sci U S A, 2009. 106(44): p. 18751-6.

91. Marmor, M.F., A.S. Abdul-Rahim, and D.S. Cohen, The effect of metabolic inhibitors on retinal adhesion and subretinal fluid resorption. Invest Ophthalmol Vis Sci, 1980. 19(8): p. 893-903.

92. Wallman, J., et al., Moving the retina: choroidal modulation of refractive state. Vision Res, 1995. 35(1): p. 37-50.

93. Pendrak, K., et al., Choroidal vascular permeability in visually regulated eye growth. Exp Eye Res, 2000. 70(5): p. 629-37.

94. Liang, H., et al., Structural and elemental evidence for edema in the retina, retinal pigment epithelium, and choroid during recovery from experimentally induced myopia. Invest Ophthalmol Vis Sci, 2004. 45(8): p. 2463-74.

95. Wang, S., et al., Choroidal thickness and high myopia: a cross-sectional study and meta-analysis. BMC Ophthalmol, 2015. 15(1): p. 70.

96. Jampol, L.M., et al., Subretinal Neovascularization with Geographic (Serpiginous) Choroiditis. American Journal of Ophthalmology. 88(4): p. 683-689.

97. Campos, J., et al., Punctate inner choroidopathy: a systematic review. Med Hypothesis Discov Innov Ophthalmol, 2014. 3(3): p. 76-82.

98. Zhao, M., et al., Mineralocorticoid receptor is involved in rat and human ocular chorioretinopathy. J Clin Invest, 2012. 122(7): p. 2672-9.

99. Imamura, Y., et al., Enhanced depth imaging optical coherence tomography of the choroid in central serous chorioretinopathy. Retina, 2009. 29(10): p. 1469-73.

100. Scheider, A., J.E. Nasemann, and O.E. Lund, Fluorescein and indocyanine green angiographies of central serous choroidopathy by scanning laser ophthalmoscopy. Am J Ophthalmol, 1993. 115(1): p. 50-6.

101. Grigsby, J.G., et al., The Role of Microglia in Diabetic Retinopathy. Journal of Ophthalmology, 2014. 2014: p. 15.

102. Omri, S., et al., Microglia/macrophages migrate through retinal epithelium barrier by a transcellular route in diabetic retinopathy: role of PKCzeta in the Goto Kakizaki rat model. Am J Pathol, 2011. 179(2): p. 942-53.

103. Fercher, A.F., et al., In vivo optical coherence tomography. Am J Ophthalmol, 1993. 116(1): p. 113-4.

104. Geitzenauer, W., C.K. Hitzenberger, and U.M. Schmidt-Erfurth, Retinal optical coherence tomography: past, present and future perspectives. Br J Ophthalmol, 2011. 95(2): p. 171-7.

105. Drexler, W., Ultrahigh-resolution optical coherence tomography. J Biomed Opt, 2004. 9(1): p. 47-74.

106. Unterhuber, A., et al., In vivo retinal optical coherence tomography at 1040 nm - enhanced penetration into the choroid. Opt Express, 2005. 13(9): p. 3252-8.

107. Gerendas, B.S., et al., Three-dimensional automated choroidal volume assessment on standard spectral-domain optical coherence tomography and correlation with the level of diabetic macular edema. Am J Ophthalmol, 2014. 158(5): p. 1039-48.

108. Spaide, R.F., H. Koizumi, and M.C. Pozzoni, Enhanced depth imaging spectral-domain optical coherence tomography. Am J Ophthalmol, 2008. 146(4): p. 496-500.

109. Ikuno, Y., et al., Reproducibility of retinal and choroidal thickness measurements in enhanced depth imaging and high-penetration optical coherence tomography. Invest Ophthalmol Vis Sci, 2011. 52(8): p. 5536-40.


47

110. Tan, C.S., W.K. Ngo, and K.X. Cheong, Comparison of choroidal thicknesses using swept source and spectral domain optical coherence tomography in diseased and normal eyes. Br J Ophthalmol, 2015. 99(3): p. 354-8.

111. Waldstein, S.M., et al., Comparison of penetration depth in choroidal imaging using swept source vs spectral domain optical coherence tomography. Eye (Lond), 2015. 29(3): p. 409-15.

112. Sala-Puigdollers, A., et al., Repeatability and reproducibility of retinal and choroidal thickness measurements in Diabetic Macular Edema using Swept-source Optical Coherence Tomography. PLoS One, 2018. 13(7): p. e0200819.

113. Regatieri, C.V., et al., Choroidal thickness in patients with diabetic retinopathy analyzed by spectral-domain optical coherence tomography. Retina, 2012. 32(3): p. 563-8.

114. Noori, J., et al., Choroidal mapping; a novel approach for evaluating choroidal thickness and volume. J Ophthalmic Vis Res, 2012. 7(2): p. 180-5.

115. Esmaeelpour, M., et al., Mapping choroidal and retinal thickness variation in type 2 diabetes using three-dimensional 1060-nm optical coherence tomography. Invest Ophthalmol Vis Sci, 2011. 52(8): p. 5311-6.

116. Ramrattan, R.S., et al., Morphometric analysis of Bruch's membrane, the choriocapillaris, and the choroid in aging. Invest Ophthalmol Vis Sci, 1994. 35(6): p. 2857-64.

117. Ouyang, Y., et al., Spatial distribution of posterior pole choroidal thickness by spectral domain optical coherence tomography. Invest Ophthalmol Vis Sci, 2011. 52(9): p. 7019-26.

118. Usui, S., et al., Circadian changes in subfoveal choroidal thickness and the relationship with circulatory factors in healthy subjects. Invest Ophthalmol Vis Sci, 2012. 53(4): p. 2300-7.

119. Tan, C.S., et al., Diurnal variation of choroidal thickness in normal, healthy subjects measured by spectral domain optical coherence tomography. Invest Ophthalmol Vis Sci, 2012. 53(1): p. 261-6.

120. Shiragami, C., et al., Risk factors for diabetic choroidopathy in patients with diabetic retinopathy. Graefes Arch Clin Exp Ophthalmol, 2002. 240(6): p. 436-42.

121. Fukushima, I., D.S. McLeod, and G.A. Lutty, Intrachoroidal microvascular abnormality: a previously unrecognized form of choroidal neovascularization. Am J Ophthalmol, 1997. 124(4): p. 473-87.

122. Lutty, G.A. and D.S. McLeod, Phosphatase enzyme histochemistry for studying vascular hierarchy, pathology, and endothelial cell dysfunction in retina and choroid. Vision Res, 2005. 45(28): p. 3504-11.

123. do Carmo, A., et al., Nitric oxide synthase activity and L-arginine metabolism in the retinas from streptozotocin-induced diabetic rats. Gen Pharmacol, 1998. 30(3): p. 319-24.

124. Carmo, A., et al., Nitric oxide synthase activity in retinas from non-insulin-dependent diabetic Goto-Kakizaki rats: correlation with blood-retinal barrier permeability. Nitric Oxide, 2000. 4(6): p. 590-6.

125. Sakurai, M., et al., Characterization and diabetes-induced impairment of nitric oxide synthase in rat choroid. Curr Eye Res, 2002. 24(2): p. 139-46.

126. Cao, J., et al., Choriocapillaris degeneration and related pathologic changes in human diabetic eyes. Arch Ophthalmol, 1998. 116(5): p. 589-97.

127. Kaur, C., W.S. Foulds, and E.A. Ling, Blood-retinal barrier in hypoxic ischaemic conditions: basic concepts, clinical features and management. Prog Retin Eye Res, 2008. 27(6): p. 622-47.

128. Koleva-Georgieva, D. and N. Sivkova, Assessment of serous macular detachment in eyes with diabetic macular edema by use of spectral-domain optical coherence tomography. Graefes Arch Clin Exp Ophthalmol, 2009. 247(11): p. 1461-9.


48

129. Gemenetzi, M., G. De Salvo, and A.J. Lotery, Central serous chorioretinopathy: an update on pathogenesis and treatment. Eye (Lond), 2010. 24(12): p. 1743-56.

130. Lee, H.K., J.W. Lim, and M.C. Shin, Comparison of choroidal thickness in patients with diabetes by spectral-domain optical coherence tomography. Korean J Ophthalmol, 2013. 27(6): p. 433-9.

131. Vujosevic, S., et al., Macular and peripapillary choroidal thickness in diabetic patients. Retina, 2012. 32(9): p. 1781-90.

132. Unsal, E., et al., Choroidal thickness in patients with diabetic retinopathy. Clin Ophthalmol, 2014. 8: p. 637-42.

133. Adhi, M., et al., Analysis of morphological features and vascular layers of choroid in diabetic retinopathy using spectral-domain optical coherence tomography. JAMA Ophthalmol, 2013. 131(10): p. 1267-74.

134. Xu, J., et al., Subfoveal choroidal thickness in diabetes and diabetic retinopathy. Ophthalmology, 2013. 120(10): p. 2023-8.

135. Querques, G., et al., Enhanced depth imaging optical coherence tomography in type 2 diabetes. Invest Ophthalmol Vis Sci, 2012. 53(10): p. 6017-24.

136. Kim, J.T., et al., Changes in choroidal thickness in relation to the severity of retinopathy and macular edema in type 2 diabetic patients. Invest Ophthalmol Vis Sci, 2013. 54(5): p. 3378-84.

137. Esmaeelpour, M., et al., Choroidal thinning in diabetes type 1 detected by 3-dimensional 1060 nm optical coherence tomography. Invest Ophthalmol Vis Sci, 2012. 53(11): p. 6803-9.

138. Wei, W.B., et al., Subfoveal choroidal thickness: the Beijing Eye Study. Ophthalmology, 2013. 120(1): p. 175-80.

139. Farias, L.B., et al., Choroidal Thickness in Patients with Diabetes and Microalbuminuria. Ophthalmology. 121(10): p. 2071-2073.

140. Tan, C.S., K.X. Cheong, and L.W. Lim, Re: Farias et al.: Choroidal thickness in patients with diabetes and microalbuminuria (Ophthalmology 2014;121:2071-3). Ophthalmology, 2015. 122(7): p. e42-3.

141. Rewbury, R., et al., Subfoveal choroidal thickness in patients with diabetic retinopathy and diabetic macular oedema. Eye (Lond), 2016. 30(12): p. 1568-1572.

142. Kase, S., et al., Choroidal thickness in diabetic retinopathy in relation to long-term systemic treatments for diabetes mellitus. Eur J Ophthalmol, 2016. 26(2): p. 158-62.

143. Nagaoka, T., et al., Alteration of choroidal circulation in the foveal region in patients with type 2 diabetes. Br J Ophthalmol, 2004. 88(8): p. 1060-3.

144. Adhi, M., A.A. Alwassia, and J.S. Duker, Analysis of choroidal thickness in eyes treated with focal laser photocoagulation using SD-OCT. Can J Ophthalmol, 2013. 48(6): p. 535-8.

145. Cho, G.E., H.Y. Cho, and Y.T. Kim, Change in subfoveal choroidal thickness after argon laser panretinal photocoagulation. Int J Ophthalmol, 2013. 6(4): p. 505-9.

146. Lee, S.H., et al., Changes of choroidal thickness after treatment for diabetic retinopathy. Curr Eye Res, 2014. 39(7): p. 736-44.

147. Zhang, Z., et al., Changes in Choroidal Thickness After Panretinal Photocoagulation for Diabetic Retinopathy: A 12-Week Longitudinal StudyChanges in Choroidal Thickness After PRP for DR. Investigative Ophthalmology & Visual Science, 2015. 56(4): p. 2631-2638.

148. Takahashi, A., et al., Effect of panretinal photocoagulation on choroidal circulation in the foveal region in patients with severe diabetic retinopathy. Br J Ophthalmol, 2008. 92(10): p. 1369-73.


49

149. Hwang, D., S. Woo, and K. Park, Changes in Choroidal Thickness and Volume in Patients with Diabetic Retinopathy after Panretinal Photocoagulation by Using a Choroidal Thickness Map. J Diabetes Metab, 2014. 5(405): p. 2.

150. Bressler, N.M., R.W. Beck, and F.L. Ferris, Panretinal Photocoagulation for Proliferative Diabetic Retinopathy. New England Journal of Medicine, 2011. 365(16): p. 1520-1526.

151. Mohamed, Q., M.C. Gillies, and T.Y. Wong, Management of diabetic retinopathy: a systematic review. JAMA, 2007. 298(8): p. 902-16.

152. Lains, I., et al., Choroidal thickness in diabetic retinopathy: the influence of antiangiogenic therapy. Retina, 2014. 34(6): p. 1199-207.

153. Yiu, G., et al., Effect of anti-vascular endothelial growth factor therapy on choroidal thickness in diabetic macular edema. Am J Ophthalmol, 2014. 158(4): p. 745-751 e2.

154. Rayess, N., et al., Baseline choroidal thickness as a predictor for response to anti-vascular endothelial growth factor therapy in diabetic macular edema. Am J Ophthalmol, 2015. 159(1): p. 85-91 e1-3.

155. Hwang DJ, Woo SJ, and P. KH, Changes in Choroidal Thickness and Volume in Patients with Diabetic Retinopathy after Panretinal Photocoagulation by Using a Choroidal Thickness Map. J Diabetes Metab, 2014. 5(7): p. 1-7.

156. Zhang, Z., et al., Changes in Choroidal Thickness After Panretinal Photocoagulation for Diabetic Retinopathy: A 12-Week Longitudinal Study. Invest Ophthalmol Vis Sci, 2015. 56(4): p. 2631-8.

157. Sonoda, S., et al., Effect of intravitreal triamcinolone acetonide or bevacizumab on choroidal thickness in eyes with diabetic macular edema. Invest Ophthalmol Vis Sci, 2014. 55(6): p. 3979-85.

158. Manjunath, V., et al., Effect of Focal Laser and anti-VEGF Therapy on Choroidal Thickness in Diabetic Macular Edema as measured by Enhanced-Depth Imaging Optical Coherence Tomography. Investigative Ophthalmology & Visual Science, 2013. 54(15): p. 2363-2363.

159. Giorno, P., et al., Retinal-choroidal changes after intravitreal ranibizumab in diabetic macular edema:12 months follow-up. Investigative Ophthalmology & Visual Science, 2014. 55(13): p. 576-576.

160. Nourinia, R., et al., Changes in Central Choroidal Thickness after Treatment of Diabetic Macular Edema with Intravitreal Bevacizumab Correlation with Central Macular Thickness and Best-Corrected Visual Acuity. Retina, 2018. 38(5): p. 970-975.

161. Brown, D.M., et al., Long-term outcomes of ranibizumab therapy for diabetic macular edema: the 36-month results from two phase III trials: RISE and RIDE. Ophthalmology, 2013. 120(10): p. 2013-22.

162. Yamazaki, T., et al., Subfoveal choroidal thickness after ranibizumab therapy for neovascular age-related macular degeneration: 12-month results. Ophthalmology, 2012. 119(8): p. 1621-7.

163. Schmidt-Erfurth, U., et al., Three-year outcomes of individualized ranibizumab treatment in patients with diabetic macular edema: the RESTORE extension study. Ophthalmology, 2014. 121(5): p. 1045-53.

164. Conti, F.F., et al., Choriocapillaris and retinal vascular plexus density of diabetic eyes using split-spectrum amplitude decorrelation spectral-domain optical coherence tomography angiography. Br J Ophthalmol, 2019. 103(4): p. 452-456.

165. Nicolo, M., et al., Choroidal Vascular Flow Area in Central Serous Chorioretinopathy Using Swept-Source Optical Coherence Tomography Angiography. Invest Ophthalmol Vis Sci, 2017. 58(4): p. 2002-2010.

166. Carnevali, A., et al., Optical coherence tomography angiography analysis of retinal vascular plexuses and choriocapillaris in patients with type 1 diabetes without diabetic retinopathy. Acta Diabetol, 2017. 54(7): p. 695-702.


50

167. Chu, Z., et al., Quantification of Choriocapillaris with Optical Coherence Tomography Angiography: A Comparison Study. Am J Ophthalmol, 2019. 208: p. 111-123.

168. Borrelli, E., et al., OCT angiography and evaluation of the choroid and choroidal vascular disorders. Prog Retin Eye Res, 2018. 67: p. 30-55.

169. Jeon, C.J., E. Strettoi, and R.H. Masland, The major cell populations of the mouse retina. J Neurosci, 1998. 18(21): p. 8936-46.

170. Danias, J., et al., Cytoarchitecture of the retinal ganglion cells in the rat. Invest Ophthalmol Vis Sci, 2002. 43(3): p. 587-94.

171. Spear, P.D., et al., Relationship between numbers of retinal ganglion cells and lateral geniculate neurons in the rhesus monkey. Vis Neurosci, 1996. 13(1): p. 199-203.

172. Peng, Y.R., et al., Molecular Classification and Comparative Taxonomics of Foveal and Peripheral Cells in Primate Retina. Cell, 2019. 176(5): p. 1222-1237 e22.

173. Curcio, C.A. and K.A. Allen, Topography of ganglion cells in human retina. J Comp Neurol, 1990. 300(1): p. 5-25.

174. Bringmann, A., et al., The primate fovea: Structure, function and development. Prog Retin Eye Res, 2018. 66: p. 49-84.

175. Provis, J.M., et al., Adaptation of the central retina for high acuity vision: cones, the fovea and the avascular zone. Prog Retin Eye Res, 2013. 35: p. 63-81.

176. Masland, R.H., The neuronal organization of the retina. Neuron, 2012. 76(2): p. 266-80. 177. Gollisch, T. and M. Meister, Eye smarter than scientists believed: neural computations in

circuits of the retina. Neuron, 2010. 65(2): p. 150-64. 178. Portha, B., et al., The GK rat: a prototype for the study of non-overweight type 2 diabetes.

Methods Mol Biol, 2012. 933: p. 125-59. 179. Sengupta, P., The Laboratory Rat: Relating Its Age With Human's. Int J Prev Med, 2013.

4(6): p. 624-30. 180. Masser, D.R., et al., Functional changes in the neural retina occur in the absence of

mitochondrial dysfunction in a rodent model of diabetic retinopathy. J Neurochem, 2017. 143(5): p. 595-608.

181. Berger, A., et al., Spectral-domain optical coherence tomography of the rodent eye: highlighting layers of the outer retina using signal averaging and comparison with histology. PLoS One, 2014. 9(5): p. e96494.

182. Lin, C.H., et al., Long-term Fluorometholone Topical Use Induces Ganglion Cell Damage in Rats Analyzed With Optical Coherence Tomography. Toxicol Sci, 2015. 147(2): p. 317-25.

183. Alves, M.R.P., et al., Subtle thinning of retinal layers without overt vascular and inflammatory alterations in a rat model of prediabetes. Mol Vis, 2018. 24: p. 353-366.

184. Sohn E. H., et al., Visualization of mouse choroidal and retinal vasculature using fluorescent tomato lectin perfusion. Invest Ophthalmol Vis Sci, 2019. 60(9): p. 1252.

185. Li, Y., et al., Direct labeling and visualization of blood vessels with lipophilic carbocyanine dye DiI. Nat Protoc, 2008. 3(11): p. 1703-8.

186. Yu, D.Y., et al., Functional and morphological characteristics of the retinal and choroidal vasculature. Prog Retin Eye Res, 2014. 40: p. 53-93.

187. Eilken, H.M., et al., Pericytes regulate VEGF-induced endothelial sprouting through VEGFR1. Nat Commun, 2017. 8(1): p. 1574.

188. Daruich, A., et al., Mechanisms of macular edema: Beyond the surface. Prog Retin Eye Res, 2018. 63: p. 20-68.

189. Chandrasekera, E., et al., Three-Dimensional Microscopy Demonstrates Series and Parallel Organization of Human Peripapillary Capillary Plexuses. Invest Ophthalmol Vis Sci, 2018. 59(11): p. 4327-4344.


51

190. Condren, A.B., et al., Perivascular mural cells of the mouse choroid demonstrate morphological diversity that is correlated to vasoregulatory function. PLoS One, 2013. 8(1): p. e53386.

191. Lutty, G.A., Diabetic choroidopathy. Vision Res, 2017. 139: p. 161-167. 192. Motiejunaite, R. and A. Kazlauskas, Pericytes and ocular diseases. Exp Eye Res, 2008.

86(2): p. 171-7. 193. Morita, A., et al., Establishment of an abnormal vascular patterning model in the mouse

retina. J Pharmacol Sci, 2018. 136(4): p. 177-188. 194. Metea, M.R. and E.A. Newman, Glial cells dilate and constrict blood vessels: a

mechanism of neurovascular coupling. J Neurosci, 2006. 26(11): p. 2862-70. 195. Ferrara, N., H.P. Gerber, and J. LeCouter, The biology of VEGF and its receptors. Nat Med,

2003. 9(6): p. 669-76. 196. Omri, S., et al., PKCzeta mediates breakdown of outer blood-retinal barriers in diabetic

retinopathy. PLoS One, 2013. 8(11): p. e81600.

Choroid and diabetes Choroidal thickness changes in DME

53

2. Choroidal thickness stratified by outcome in

diabetic macular edema2

2 Section 2 is based on two articles: Campos A. et al., Choroidal thickness changes stratified by outcome in

real-world treatment of diabetic macular edema. Graefes Arch Clin Exp Ophthalmol, 2018. 256 (10):1857-

1865 and Campos A. et al., Response to: Choroidal thickness changes stratified by outcome in real-world

treatment of diabetic macular edema. Graefes Arch Clin Exp Ophthalmol, 2019. 257 (1):243-244.


54

2.1. Abstract

Purpose: To evaluate subfoveal choroidal thickness (SFCT) as a marker of outcome

while treating diabetic macular edema (DME) and to correlate different choroidal

thickness (CT) parameters.

Methods: Prospective interventional case series included a total of 126 eyes from 126

patients with recently diagnosed DME treated with a 3-monthly loading dose of

ranibizumab or aflibercept and PRN thereafter until 24 months (M). SD OCT was used

to measure CT and central retinal thickness (CRT). CT was manually measured, in the

central 3500 μm area, subfoveally (SFCT) and at 1750 μm right and left from the center,

in the horizontal plane, and 1750 μm up and down from the center in the vertical plane.

Anatomic (10% decrease in CRT) and functional (BCVA gain ≥ 5 letters) responses were

assessed using univariate and multivariate analyses. The areas under ROC curves were

used to assess whether baseline SFCT was a predictor of outcome.

Results: CT significantly decreased in all follow-ups (3 months after the 3 injections’

loading dose (3M), 6 months (6M), 12 months (12M), 18 months (18M), 24 months

(24M)). SFCT and other CT parameters are correlated. SFCT decrease from baseline

was related with treatment (p = 0.003 to p < 0.001) but not with anatomic (3M, p = 0.858;

6M p = 0.762) or functional response (3M, p = 0.746; 6M, p = 0.156). SFCT was not

found to be predictive of anatomic (AUC = 0.575, p = 0.172) or functional (AUC = 0.515,

p = 0.779) outcome.

Conclusions: SFCT is a reliable marker of choroid thickness. Baseline SFCT decreased

with anti-VEGF treatment but did not predict DME outcome.

Keywords: diabetic macular edema; choroidal thickness; subfoveal choroidal thickness;

outcome.


55

2.2. Introduction

Treatment of diabetic macular edema (DME) shifted from laser to anti-VEGF

agents [1]. Most patients respond well to therapy while others do not so well [2]. Several

attempts have been made to find markers of prognosis or predictors of treatment

response in DME. The choroidal thickness (CT) or the subfoveal choroidal thickness

(SFCT) were suggested as a predictors of treatment response in DME, but the use of

the SFCT as a marker of the CT needs further evidence [2]. A diabetic choroidopathy

has been demonstrated [3], therefore the choroid has been looked out in search for

markers of the pathogenesis of diabetic retinopathy (DR) or of the response to treatment.

The availability of the OCT was a step forward in order to easily assess and monitor CT

[4]. Despite growing evidence demonstrating alterations of the SFCT in DME and with

DME treatment [5], its use as a marker of prognosis and of response to anti-VEGF

therapy has not been fully elucidated. Previous studies used different approaches that

originated contradictory data [4].

The present study was designed to investigate whether baseline SFCT is a

predictor of anatomic or functional response to anti-VEGF therapy and whether

measuring the choroid at different locations from the center correlates with SFCT.

2.3. Methods

In a prospective interventional case series, Type 2 diabetic patients with NPDR,

diagnosed with recent onset DME in at least one eye, naïve to intravitreal treatment,

were included after approval from the Ethical Committee of the Leiria Hospital. Informed

consent was given before inclusion. The study adhered to the tenets of the Declaration

of Helsinki and to the standards of Good Clinical and Scientific Practice of the Faculty of

Medicine of the University of Coimbra.

Diagnostic criteria for DR and DME were based on past ophthalmic history and

complete ophthalmic evaluation, including a dilated fundus examination, fundus


56

photography, OCT imaging and fluorescein angiography in selected cases. Data

included patient age, sex, blood pressure, duration of diabetes, glycated hemoglobin

level (Hb A1c), best-corrected visual acuity (BCVA), biomicroscopic examination, past

laser therapy and length of follow-up. DME was considered when clinical significant

macular edema (CSME) involving the central macula (CI-CSME) [6] or a central (1 mm

central subfield thickness in the OCT-modified ETDRS grid) retinal thickness (CRT) ≥

300 μm, were present. Eyes were included when baseline BCVA ranged from 24 to 78

ETDRS letters (Snellen equivalent 20/320-20/32, logMAR 1.22-0.14). When DME was

bilateral at baseline, the right eye was included in patients whose year of birth was an

even number and the left eye was included when the year of birth was an odd number

[7]. Patients were excluded if they had any other treatments related to their NPDR,

except for focal laser for more than 6 months. Individuals were excluded if they had any

ocular diseases aside from DME in the treated eye. Eyes with a myopic refractive error

of greater than 4 diopters (D) were also excluded [4, 8].

Eyes were treated with a 3-monthly loading dose of ranibizumab or aflibercept and

on a pro re nata (PRN) regimen thereafter. Eyes that developed proliferative diabetic

retinopathy and needed panretinal photocoagulation (PRP), eyes that were rescued with

focal/grid laser and eyes that were switched to intravitreal steroids were discontinued

from follow-up. Eyes were allowed to be switched from ranibizumab to aflibercept. Data

from BCVA using the ETDRS standardized chart and SD-OCT were collected in every

visit (baseline, 3M, 6M, 12M, 18M and 24M). Top score allowed in the ETDRS chart was

85L (20/20).

Enhanced depth imaging (EDI) optical coherence tomography mode was selected

(Spectralis; Heidelberg Engineering, Heidelberg, Germany) and a 6-mm x 6-mm macular

cube scan was performed using the high-resolution (HR) scanning mode. HR star scan

mode (6 scans, each made up of 1536 A scans, 30° apart from each other cutting through

the fovea) and HR map scan mode (comprising 61 horizontal B-scans, 120 μm apart

from each other, each made up of 1536 A-scans, 1536 × 1536 pixels, lateral resolution


57

of 6 μm/pixel) were acquired. A signal strength greater than 20 was required for all scans.

CT was evaluated manually after plotting a 7.2 mm-ETDRS grid centered at the fovea

and CRT was evaluated automatically using a 6 mm-OCT modified ETDRS grid. The

fovea was always checked and the center of the OCT star mode was re-centered at the

fovea whenever needed before performing the scans and thereafter. CT was manually

measured using the digital caliper tool in the Heidelberg Eye Explorer software, from the

hyperreflective line of the Bruch’s membrane to the hyperreflective line of the

choroidoscleral interface, in the central 3500 μm area, subfoveally, and at a distance of

1750 μm right and left from the center in the horizontal plane, and 1750 μm up and down

from the center in the vertical plane. The CT area underneath the 3500 μm central

macula (square inches) in the plane defined by the horizontal line scan encompassing

the fovea, was calculated manually from the RPE hiperreflective line to the

choroidoscleral junction, by Image J software (version 1.48, National Institutes of Health,

USA), (Figure 2.1).

Figure 2.1. Choroidal thickness manually measured in the central 3500-μm area underneath the

RPE line, subfoveal (SFCT) and at 1750 μm nasal (CT1750n) and temporal (CT1750t) from the

center, in the plane defined by the horizontal line scan encompassing the fovea (Image J software,

version 1.48, National Institutes of Health, USA). A similar procedure was done in the plane

defined by the vertical line encompassing the fovea to obtain the superior (CT1750s) and inferior

(CT1750i) choroidal thicknesses.


58

Two independent raters (AC, do Carmo) measured the scans masked to the

subject’s outcome in a prospective way and final measures were the mean of the two

scored for each location and follow-up period. Both raters retrospectively reviewed all

scans masked to subjects’ outcome at the end of the study and when the gap from one

another was greater, final measures were reached by consensus. All scans were

performed from 9.00 a.m. to 1.00 p.m. To evaluate whether baseline SFCT might predict

clinically relevant response to treatment, we defined anatomic responders as eyes

having a ≥ 10% decrease in baseline CRT (≥ 300 µm). Eyes with BCVA gains of ≥ 5

letters from baseline at 3M were defined as early functional responders, while eyes with

BCVA gains of ≥ 5 letters from baseline at 6M only, were defined as late functional

responders. Only the eyes followed for 6M were considered for these calculations, n =

122 eyes.

Statistical analysis

Nominal data were described by absolute and relative frequencies. Quantitative

data were described by using the mean, standard deviation, median, minimum and

maximum in the sample characterization. Median, minimum and maximum were omitted

in the tables for convenience. Quantitative variables were assessed for normality with

Shapiro-Wilk tests. Comparisons between two measures, in different time points, of the

same variable, were performed resorting to paired sample t-tests or Wilcoxon tests,

taking normality requirements into account. Comparisons of quantitative variables

between two groups were performed with t-Student or Mann-Whitney tests, as

applicable. Intra-class correlations (ICC) were used to compare CT measurements

between raters at all locations and time points. A ROC analysis was undertaken to

assess how accurately SFCT baseline values could be used to predict anatomic or

functional response. Correlations between quantitative variables were assessed by

computing Pearson or Spearman correlation coefficients, depending on whether


59

normality requirements were met or not. The statistical analyses were performed on IBM

SPSS Statistics 24 and on statistic platform R v3.3.2, The R Foundation Vienna, Austria.

The level of significance adopted was 0.05.

2.4. Results

From June 2014 to November 2017, a total of 126 eyes from 126 patients were

prospectively included. A total of 113 eyes were treated with a loading dose of 0.5 mg

ranibizumab, 13 eyes were treated with a loading dose of 2 mg aflibercept and 18 were

switched from ranibizumab to aflibercept. Baseline demographic and ocular

characteristics are outlined in Supplementary Table 1.1. A total of 126 eyes were

followed by 3 months, 122 eyes were followed by 6 months, 60 eyes were followed by

12 months, 29 eyes were followed by 18 months and 26 eyes were followed by 24

months.

The mean number of injections given was 3.0 at 3M, 4.6 ± 1.3 (3-7, median 4) at

6M, 5.3 ± 2.0 (3-10, median 6) at 12M, 7.6 ± 2.5 (3-12, median 8) at 18M and 8.0 ± 4.0

(3-15, median 7.5) at 24M. The mean baseline BCVA improved significantly, the mean

baseline CRT and the mean baseline SFCT decreased significantly (Table 2.1 and

Figure 2.2).


60

Table 2.1. Differences in the variables considered between each endpoint and baseline

Baseline

(n = 126)

3M

(n = 126)

6M

(n = 122)

12M

(n = 60)

18M

(n = 29)

24M

(n = 26)

CRT 432.4 ± 107.0 -92.8 ± 103.9 -95.7 ± 108.6 -83.9 ± 96.0 -78.0 ± 85.3 -81.5 ± 112.4

<0.001 <0.001 <0.001 <0.001 0.002

SFCT 346.6 ± 75.6 -22.5 ± 35.8 -25.6 ± 44.8 -31.8 ± 41.7 -33.7 ± 45.9 -31.2 ± 53.4

<0.001 <0.001 <0.001 0.001 0.006

CT1750t 307.9 ± 71.5 -16.5 ± 38.3 -13.5 ± 44.5 -19.3 ± 51.4 -14.7 ± 48.2 -20.2 ± 54.2

<0.001 0.005 0.005 0.126 0.070

CT1750n 253.1 ± 71.2 -15.7 ± 39.6 -15.9 ± 41.2 -23.2 ± 43.8 -31.3 ± 47.1 -25.9 ± 40.1

<0.001 <0.001 <0.001 0.002 0.003

CT1750s 320.1 ± 66.7 -10.2 ± 40.4 -14.2 ± 44.7 -29.1 ± 53.9 -26.3 ± 37.8 -19.8 ± 65.1

0.006 0.003 <0.001 0.001 0.134

CT1750i 283.4 ± 71.3 -11.0 ± 36.6 -13.0 ± 40.4 -21.6 ± 41.8 -32.1 ± 54.5 -10.9 ± 53.5

0.001 0.003 <0.001 0.005 0.308

CT3500a 5791.4 ± -344.4 ± 621.8 -338.8 ± 659.2 -440.1 ± 796.2 -416.6 ± 760.7 -368.5 ± 941.2

± 1272.5 <0.001 <0.001 <0.001 0.0010 0.057

BCVA 63.2 ± 12.7 5.9 ± 7.1 9.5 ± 7.9 7.1 ± 9.5 7.3 ± 9.1 8.4 ± 9.2

<0.001

60.6%*

<0.001

77.9%*

<0.001

61.7%*

0.001

65.4%*

<0.001

65.4%*

N Injections 3.0 ± 0.0 4.6 ± 1.3 5.3 ± 2.0 7.6 ± 2.5 8.0 ± 4.0

(3.0 – 3.0) (3.0 – 7.0) (3.0 – 10.0) (3.0 - 12.0) (3.0 – 15.0)

Abbreviations: CRT = 1 mm central retinal thickness; SFCT= subfoveal choroidal thickness; CT = choroidal thickness measured at 1750 µm, nasal (CT1750n) and temporal

(CT1750t) from the fovea in the plane defined by the horizontal line scan encompassing the fovea and superior (CT1750s) and inferior (CT1750i) from the fovea in the plane defined

by the vertical line scan encompassing the fovea; CT3500a = choroidal thickness area underneath the fovea measured from 1750 µm nasal to 1750 µm temporal from the central

fovea in the plane defined by the horizontal line scan encompassing the fovea; BCVA = best corrected visual acuity collected from ETDRS charts; 3M = 2-3 months after the 3

injections’ loading dose; 6M = 6 months; 12M = 12 months; 18M = 18 months; 24M = 24 months. Results are presented as mean ± SD. For each variable except for BCVA, the

percentages relate to the proportion of individuals for which a decrease in the variable was observed when compared to the baseline value. * For BCVA, the percentages relate

to the proportion of eyes that displayed an increase of ≥ 5 letters when compared to the baseline.


61

Figure 2.2. Evolution of mean central retinal thickness (CRT), mean subfoveal choroidal

thickness (SFCT), and mean best-corrected visual acuity (BCVA) scored in ETDRS letters

collected from ETDRS charts with time in eyes with DME under anti-VEGF treatment. Note that

the evolution of the SFCT curve does not have the same profile as those of the CRT and BCVA.

Until 6M when the stratification by outcome was done, the slopes of the BCVA curve and that of

the CRT curve are also different, expressing a poor correlation between anatomic and functional

outcome as depicted in Tables 2.2 and 2.3.

The distribution of baseline SFCT showed a wide range of variability although it

decreased by 25.45 μm per each decade of life (Figure 2.3).

Figure 2.3. Example on how variability in SFCT may introduce bias when dealing with small

samples: Scatterplot depicting the negative weak correlation between age and baseline SFCT

(μm), r = −0.328, p < 0.001. Though a wide inter-individual variability can be observed, the

average decrease of SFCT per decade was seen to be 25.45 μm.


62

The mean CT decreased at all locations in all follow-up intervals (Supplementary

Figure 2.1). The percent decrease in the SFCT and in the other CT parameters were

correlated in all follow-up intervals, more prominently between the SFCT and CT areas

until 6M (Figure 2.4). ICC between raters was 0.98, ranging from 0.94 to 0.98, according

to the locations. The mean inter-observer difference was 6.8 μm.

From the 126 eyes enrolled, 64 eyes (50.8%) had past history of macular

photocoagulation, but past laser history did not significantly affect the CT parameters at

any follow-up interval (Supplementary Figure 2.2). A total of 98 eyes from the 119 eyes

with baseline CRT ≥ 300 μm decreased CRT 10% from baseline (82.4%), and were

considered anatomic responders. A total of 74 eyes (60.7%) at 3M and an additional 22

eyes (18.0%) at 6M were functional responders (96 eyes out of 122, 78.7%).

Figure 2.4. Radar chart displaying the percentage of participants with decrease in the CT

parameters at different time points. SFCT subfoveal choroidal thickness (dotted black line); CT

temp choroidal thickness 1750 μm temporal to the fovea, in the plane defined by the horizontal

line scan encompassing the fovea (orange line); CT nasal same as the previous but 1750 μm

nasal to the fovea (purple line); CT sup choroidal thickness 1750 μm superior to the fovea, in the

plane defined by the vertical line scan encompassing the fovea (red line); CT inf same as previous

but 1750 μm inferior to the fovea (green line); CT area area of choroidal thickness from 1750 μm

nasal to 1750 μm temporal to the fovea, in the plane defined by the horizontal line scan

encompassing the fovea (blue line) (Image J software, version 1.48, National Institutes of Health,

USA). 3M = 3-month endpoint, 6M = 6-month endpoint, 12M = 12- month endpoint, 18M = 18-

month endpoint, 24M = 24-month endpoint.


63

SFCT was not significantly different when comparing either anatomic responders versus

non-responders (Table 2.2) or functional responders versus non-responders (Table 2.3).

Table 2.2. Comparison of outcome measures between anatomic responders and non-responders

at baseline, 3M and 6M

Anatomic

non-responders

(n = 21)

Anatomic responders

(n = 98)

p-value

BCVA Baseline 67.4 ± 11.1 61.9 ± 12.9 0.037

3M 72.7 ± 11.5 68.2 ± 12.6 0.088

6M 75.7 ± 10.7 71.9 ± 11.6 0.109

Percentage increasing ≥ 5L 61.9% 62.2% 1.000

CRT

Baseline 367.8 ± 58.2 450.6 ± 108.3 <0.001

3M 366.2 ± 71.8 336.4 ± 72.3 0.026

6M 367.2 ± 71.2 332.1 ± 74.3

SFCT

Baseline 350.6 ± 73.7 347.3 ± 76.1 0.859

3M 322.9 ± 72.0 326.0 ± 82.3

0.858

6M 326.7 ± 70.6 321.0 ± 78.9 0.762

p-value 3M 0.003 <0.001

p-value 6M 0.036 <0.001

Laser

Yes 14 (66.7%) 51 (52.0%) 0.239

No 7 (33.3%) 47 (48.0%)

Number of injections 4.0 ± 1.2 4.8 ± 1.3 0.016

Abbreviations: BCVA = best corrected visual acuity scored using the ETDRS letter scale: 62 letters (L)

are Snellen 20/58, 67L (20/46), 68L (20/44), 72L (20/36), 73L (20/35) and 76L (20/30); 3M = 3 month

endpoint after the loading dose; 6M = 6 month endpoint; CRT = 1 mm central retinal thickness; SFCT =

subfoveal choroidal thickness; SND = subfoveal neuroretinal detachment. For anatomic responders’

calculation, only eyes with baseline CMT ≥300 μm were considered, n = 119 eyes. Eyes were considered

responders if they had a 10% decrease from baseline CRT. The difference in vision gain between

anatomic responders and non-responders was not statistically significant. A higher mean baseline CRT

correlated with anatomic response. The mean baseline SFCT decreased significantly with treatment in

both groups with no statistically significant difference between them.


64

Table 2.3 – Stratification of the population by functional outcome

Functional non-responders

(n = 26)

Early functional responders

(n = 74)

Late functional responders

(n = 22)

p-valuea

BCVA Baseline 65.3 ± 10.5 62.3 ± 13.2 63.8 ± 13.4 0.535

3M 63.9 ± 13.4 72.2 ± 11.0 65.1 ± 12.6 0.009

6M 64.7 ± 12.0 75.5 ± 9.6 72.7 ± 11.8 <0.001

CRT Baseline 420.7 ± 99.1 435.8 ± 109.5 434.9 ± 111.1 0.563

3M 346.9 ± 90.3 334.4 ± 62.0 348.3 ± 86.7 0.712

6M 370.4 ± 111.4 332.3 ± 61.5 311.7 ± 44.4 0.455

SFCT Baseline

339.3 ± 63.4 343.7 ± 82.6 365.3 ± 62.5 0.532

3M 328.2 ± 71.9 319.3 ± 76.2 335.4 ± 57.5 0.746

6M 303.6 ± 66.4 326.9 ± 83.2 321.7 ± 66.2 0.156

Laser Yes 22 (84.6%) 36 (48.6%) 10 (45.5%) <0.001

No 4 (15.4%) 38 (51.4%) 12 (54.5%)

N Injections

4.4 ± 1.3 4.7 ± 1.3 4.8 ± 1.3 0.267

Abbreviations: BCVA = best corrected visual acuity scored using the ETDRS letters chart. ETDRS 62 letters

(L) are Snellen 20/58; 64L (20/53); 65L (20/50); 72L (20/36); 73L (20/35) and 76L (20/30). 3M = 3 month

endpoint after the loading dose; 6M = 6 month endpoint. CRT = 1 mm central retinal thickness; SFCT =

subfoveal choroidal thickness. Early functional responders is the group of eyes gaining ≥ 5 letters at 3M and

late functional responders is the group of eyes gaining ≥ 5 letters at 6M only. aComparison responders vs

non-responders. CRT changes from baseline do not show a statistically significant difference between

responders and non-responders, displaying the poor correlation between functional response and anatomic

response. SFCT does not show a statistically significant difference between responders and non-

responders, meaning that while decreasing with anti-VEGF treatment it is not a marker of outcome, once

the population is stratified by outcome.


65

The ROC analyses showed that baseline SFCT was not found to be a statistically

significant predictor of being an anatomic responder, (area under the curve, AUC =

0.575, p = 0.172) nor of being a functional responder (sorting out early functional

responders from non-responders) (AUC = 0.515, p = 0.779). Since baseline SFCT was

not found to be a marker of outcome at 6M but changed more prominently in the first 3

months, a new ROC curve was calculated to correlate the early SFCT decrease and

early functional outcome (Figure 2.5).

Area Under the Curve

Test Result Variable(s): difT0 T1

Area Std. Errora Asymptotic Sig.b

Asymptotic 95% Confidence

Interval

Lower Bound Upper Bound

0.529 0.055 0.589 0.422 0.636

Figure 2.5. ROC curve analysis comparing the decrease in SFCT from baseline to 3M with a ≥5

L gain at EM (early functional response). Dif T0_T1 is the difference between mean baseline

SFCT and mean 3M SFCT; Sig is the p value. The Area (area under the curve, AUC) of 0.529

means that the change in baseline SFCT at 3M does not predict visual gain.

0.0 0.2 0.4 0.6 0.8 1.0

1 - Specificity

1.0

0.8

0.6

0.4

0.2

0.0

Sen

siti

vit

y


66

The baseline SFCT was not found to be a marker of early outcome either (AUC

= 0.529, p = 0.589). This absence of a significant difference between SFCT decreases

between functional responders and non-responders from baseline to 3M is depicted in a

boxplot (Figure 2.6).

Figure 2.6. Boxplot or figure of extremes and quartiles of the difference found in SFCT from

baseline to 3M in functional responders at 3M (gain of 5 L or more) and non-responders. T0-T1

is the difference between baseline SFCT and 3M SFCT; NR + LR is the group of non-responders

at 3M (non-responders at 6M and late functional responders); ER is the group of responders at

3M (early functional responders). The distribution of the Baseline - 3M SFCT of either group

mostly overlaps, meaning that the decrease in the SFCT at 3M is not a useful surrogate of

functional gain.

2.5 Discussion

The aim of this study was to evaluate whether the mean baseline SFCT is a

predictor of outcome in DME. Additional outcomes were to study how the CT changes

with anti-VEGF agents and to correlate SFCT with other CT data (collected focally

around the center of the fovea or from an area underneath the fovea).

As previously reported, the choroid thinned under anti-VEGF treatment at variable

degrees with time [4, 9]. The mean SFCT and the mean CRT decreased while the mean

BCVA increased. Therefore, it is mathematically possible to find a correlation between


67

the evolution of the mean SFCT, the mean CRT and the mean BCVA [5]. However, when

the eyes are stratified by outcome, that association is actually accidental, reflecting not

a prognostic value but only the effectiveness of the anti-VEGF therapy.

A similar behavior was observed between SFCT, which is the CT marker most

commonly used, and the other CT data collected at other locations around the fovea

(Supplementary Figure 2.1 and Figure 2.4).

There were previous retrospective studies pointing out in different directions, while

considering the value of SFCT as a marker of outcome [10-12]. All had considerable

drawbacks: inclusion of both Type 1 and Type 2 diabetic patients, eyes with proliferative

retinopathy included, short follow-up [10-12], the use of Snellen charts to assess BCVA,

an odd definition of anatomic response and double organ bias [7, 12]. Snellen charts are

not logarithmic and do not evaluate BCVA accurately bellow 20/50, and the choroid is

actually thicker in Type 1 diabetics [4].

About half of the eyes included had previous focal laser treatment, but focal laser

does not seem to change SFCT [13]. In the population included, previous history of laser

did not change the CT parameteres’ profile with time, and the differences found with

laser-naïves were not significantly different from laser-treated (Supplementary Figure

2.2).

Outcome calculations were done at 6 M. At this time point, none of the eyes were

excluded from follow-up due to laser rescue, need of PRP or switch to steroids. Two

eyes were treated with PRP after the 12M of follow-up and 6 were treated with steroids

(2 after 12M and 4 after 18 M, data not shown). None was rescued with focal laser, until

12 M, because, in our department, rescues apply only after 12 M of anti-VEGF treatment,

since the role of laser rescue is questionable [14]. Less patients were included in the

longer follow-ups due to the prospective profile of the study, with the later patients

included having a shorter follow-up, but that has not biased the assessment, once the

longest outcome evaluated was the 6 months’. We included eyes treated with aflibercept

and ranibizumab since they are very similar drugs in the treatment of DME. It is known


68

that differences in the outcome were observed for patients with BCVA ≤ 20/50 at year 1

only [15]. If aflibercept treated eyes were expected to have a better outcome and if SFCT

would be a marker of outcome, it would be pointless to this relative relation which drug

was to be used. Moreover, there is some controversy about the first year results of the

Protocol T study and patients with BCVA ≤ 20/50 had better functional outcome but had

no significant differences in the baseline CRT [16]. This issue will be addressed in detail

in Chapter 3 wherein prognostic factors of DME will be tested with linear multivariate

regression models.

As expected, the baseline SFCT decreased with age [17]. Nevertheless, the

distribution was widespread and there were several outliers, indicating that small

samples or double organ inclusion may mislead data when studying the choroid in cross

sectional studies (Figure 2.3).

A recovery of the mean BCVA and mean CRT that fluctuates over time was

observed, as that is the hallmark of the reactive regimens such as PRN (in this study 4.6

± 1.3 injections at 6M and 5.3 ± 2.0 injections at 12M). However, the evolution of SFCT

did not mimic that profile (Figure 2.2). The thickness of the retina decreases more sharply

than the thickness of the choroid (Figure 2.2). In the retina, the edema is mostly

composed of liquid material, at least in its early phases, and may increase the CRT up

to two- or threefold. On the other hand, the choroid is mostly composed of medium and

large vessels (Sattler and Haller layers), the choriocapillaris being less than 10% of the

total choroid thickness. Therefore, the decrease in the thickness of the choroid mediated

by the anti-VEGF use is probably due to liquid reabsorption, mainly from the stroma and

from the lamellae of the suprachoroidal space [4]. Anti-VEGF agents act upon blood

vessel permeability, decreasing retinal edema, by improving inner blood-retinal barrier

and probably by restoring outer retinal barrier homeostasis. RPE under homeostasis

produces VEGF to assure the choriocapillaris fenestrations, needed to establish the

routes for RPE and outer retina nutrition and for water clearance from the retina [4]. One

may speculate that in diabetes, an overproduction of VEGF by a RPE under metabolic


69

stress may lead to an increased permeability of the outer blood retinal barrier,

contributing to DME, sometimes with subfoveal neuroretinal detachment (SND) in the

acute phases. These changes can be reversed by the use of anti-VEGF agents. As the

choroid thins under anti-VEGF administration, a link between the decrease in CRT, visual

restoration and a decrease in SFCT, seemed unquestionable [5]. Indeed, the idea that

the thinning of the choroid under the action of anti-VEGF may predict the homeostasis

of the retina, the water clearance from the retina, resolution of the edema, and visual

recovery, is appealing. Unfortunately, the thickening of the choroid in diabetic macular

edema is controversial [4]. More importantly, DME behaves more like a group of

diseases than as a single well-defined disease. As we pointed out in a previous review

[4], and as stated by another group [18], the presence of SND probably signals an acute

form of DME with outer blood retinal barrier dysfunction, very prone to be a good

anatomic responder. It is possible that in such cases, the thinning of the choroid mirrors

somehow the restoration of the outer blood retinal barrier homeostasis. It is doubtful,

though, that this applies to all forms of DME, most notably those forms with persistent

cysts or retinal thickening where there is only limited or poor response to anti-VEGFs

and nevertheless the choroid thins. Furthermore, it is controversial that the SFCT

decrease would differentiate those eyes that would recover vision or retinal edema from

those that would not. When stratified by outcome, the mean SFCT decrease seems to

be a side effect of the anti-VEGFs only. Nevertheless, data from area measurements

seem to indicate that the percentage of the choroidal thinning may be an indicator of

undertreatment in non-fixed regimens (Figure 2.4). In Table 2.3, we see that there is a

poor correlation of baseline CRT and functional outcome. The decrease in the CRT is

not significantly different in functional responders and non-responders. This poor

correlation between anatomic response and functional response has been previously

reported [19].

HR scan mode was used as it gives a higher quality image and allows a good

visualization of the choroid-scleral border. Unlike the automatic segmentation modes that


70

only include the choroidal vessels [20], this method includes the suprachoroidal space,

most prone to change, and gives higher scores for the CT, even when dealing with Type

2 diabetics only. Furthermore, SD-Spectralis OCT collects an amplified image and has

a higher normative data than other devices [21]. Automatic segmentation was not used

because it was not available in the EDI mode of the software version 6.9.4.0 we used.

The central 1 mm cube automatic acquired volume for the choroid is not available either,

even while using the normal non-EDI mode. However, as long as automatic

segmentation does not overpass its most relevant shortcoming, that is, the need for

manual correction of the border lines, its relevance is questionable.

The results indicating that the baseline SFCT does not predict outcome in DME do

not exclude a role of the choroid in the pathogenesis of DME or DR, for which there is

plenty of evidence [3, 4].

The biggest limitation of this study is not to be a randomized controlled trial or

involving multiple centers and the small number of eyes in the longer follow-ups.

The strengths of this study are related to its prospective profile, the stratification of

the eyes treated by outcome, the correlation of the SFCT with outcome, the inclusion of

one eye per patient only [22], the inclusion of Type 2 diabetics only, the use of the HR

scan mode, the multiple locations for CT evaluation and the use of the ETDRS charts for

the evaluation of BCVA.

In conclusion: the choroidal thins with anti-VEGF therapy but SFCT is not a

predictor of outcome in DME. The amount of the choroid thinning may be an indicator of

undertreatment in non-fixed regimens.

Funding

This work was funded by Portuguese Foundation for Science and Technology (Fundacao

para a Ciencia e a Tecnologia, FCT), Strategic Project (UID/NEU/04539/2013), and

COMPETE-FEDER (POCI-01-0145-FEDER-007440). EJC was financially supported by


71

the FCT Postdoctoral Fellowship SFRH/BPD/93672/2013, through European Union and

National funds and co-funded by Human Capital Operating Programme (Programa

Operacional do Capital Humano, POCH).

The sponsor had no role in the design or conduct of this research.

Conflict of Interest: All authors certify that they have no affiliations with or involvement in

any organization or entity with any financial interest (such as honoraria; educational

grants; participation in speakers' bureaus; membership, employment, consultancies,

stock ownership, or other equity interest; and expert testimony or patent-licensing

arrangements), or non-financial interest (such as personal or professional relationships,

affiliations, knowledge or beliefs) in the subject matter or materials discussed in this

manuscript.

Ethical approval

All procedures performed in studies involving human participants were in accordance

with the ethical standards of the Ethical Committee of the Leiria Hospital and with the

1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent: Informed consent was obtained from all individual participants

included in the study.

2.6. References


2. Ashraf, M., A. Souka, and R. Adelman, Predicting outcomes to anti-vascular endothelial growth factor (VEGF) therapy in diabetic macular oedema: a review of the literature. Br J Ophthalmol, 2016. 100(12): p. 1596-1604.

3. Lutty, G.A., Diabetic choroidopathy. Vision Res, 2017. 139: p. 161-167. 4. Campos, A., et al., Viewing the choroid: where we stand, challenges and contradictions



72

5. Nourinia, R., et al., Changes in Central Choroidal Thickness after Treatment of Diabetic Macular Edema with Intravitreal Bevacizumab Correlation with Central Macular Thickness and Best-Corrected Visual Acuity. Retina, 2018. 38(5): p. 970-975.

6. Photocoagulation for diabetic macular edema. Early Treatment Diabetic Retinopathy Study report number 1. Early Treatment Diabetic Retinopathy Study research group. Arch Ophthalmol, 1985. 103(12): p. 1796-806.

7. Esen, F., et al., Double-Organ Bias in Published Randomized Controlled Trials of Glaucoma. J Glaucoma, 2016. 25(6): p. 520-2.

8. Meng, W., et al., Axial length of myopia: a review of current research. Ophthalmologica, 2011. 225(3): p. 127-34.

9. Lains, I., et al., Choroidal thickness in diabetic retinopathy: the influence of antiangiogenic therapy. Retina, 2014. 34(6): p. 1199-207.

10. Lee, S.H., et al., Changes of choroidal thickness after treatment for diabetic retinopathy. Curr Eye Res, 2014. 39(7): p. 736-44.

11. Yiu, G., et al., Characterization of the choroid-scleral junction and suprachoroidal layer in healthy individuals on enhanced-depth imaging optical coherence tomography. JAMA Ophthalmol, 2014. 132(2): p. 174-81.


13. Adhi, M., A.A. Alwassia, and J.S. Duker, Analysis of choroidal thickness in eyes treated with focal laser photocoagulation using SD-OCT. Can J Ophthalmol, 2013. 48(6): p. 535-8.

14. Regnier, S., et al., Efficacy of anti-VEGF and laser photocoagulation in the treatment of visual impairment due to diabetic macular edema: a systematic review and network meta-analysis. PLoS One, 2014. 9(7): p. e102309.


16. Sivaprasad, S., et al., Using Patient-Level Data to Develop Meaningful Cross-Trial Comparisons of Visual Impairment in Individuals with Diabetic Macular Edema. Adv Ther, 2016. 33(4): p. 597-609.

17. Margolis, R. and R.F. Spaide, A pilot study of enhanced depth imaging optical coherence tomography of the choroid in normal eyes. Am J Ophthalmol, 2009. 147(5): p. 811-5.

18. Vujosevic, S., et al., Diabetic Macular Edema With and Without Subfoveal Neuroretinal Detachment: Two Different Morphologic and Functional Entities. Am J Ophthalmol, 2017. 181: p. 149-155.

19. Diabetic Retinopathy Clinical Research, N., et al., Relationship between optical coherence tomography-measured central retinal thickness and visual acuity in diabetic macular edema. Ophthalmology, 2007. 114(3): p. 525-36.

20. Gerendas, B.S., et al., Three-dimensional automated choroidal volume assessment on standard spectral-domain optical coherence tomography and correlation with the level of diabetic macular edema. Am J Ophthalmol, 2014. 158(5): p. 1039-48.

21. Grover, S., et al., Normative data for macular thickness by high-definition spectral-domain optical coherence tomography (spectralis). Am J Ophthalmol, 2009. 148(2): p. 266-71.

22. Armstrong, R.A., Statistical guidelines for the analysis of data obtained from one or both eyes. Ophthalmic Physiol Opt, 2013. 33(1): p. 7-14.


73

2.7. Supplementary files


Demographic characteristics

(n = 126 eyes of 126 patients; 68 RE and 58 LE )

Age (y)

Mean ± SD 66.2 ± 9.4

Median (range) 67 (46-85)

Sex

Female 60 (47.6%)

Male 66 (52.4%)

Duration of diabetes (y)

1 - 15 58 (46.0%)

16 - 25 52 (41.3%)

> 25 16 (12.7%)

Hb A1c (%)

≤ 7 33 (26.2%)

> 7 and ≤ 8 46 (36.5%)

> 8 47 (37.3%)

Hypertensiona

Yes 79 (62.7%)

No 47 (37.3%)

Insulin

Yes 67 (53.2%)

No 59 (46.8%)

Ocular characteristics (n = 126 eyes)

Lens status

Phakic 78 (61.9%)

Pseudophakic 48 (38.1%)

Injection received

RNZ only 95 (75.4%)

AFL only 13 (10.3%)

Both 18 (14.3%)

Laser

Yes 68 (55.7%)

No 54 (44.3%)

Baseline BCVA

Mean ± SD 63.2 ± 12.7

Median (range) 65.0 (24.0 -78.0)


74

Baseline CRT

Mean ± SD 432.4 ± 107.0

Median (range) 381 (291.0 - 749.0)

Baseline SFCT

Mean ± SD 346.6 ± 75.6

Median (range) 346.0 (124.0 - 552.0)

Baseline CT area

Mean ± SD 5791.4 ± 1272.5

Median (range) 5759.5 (2306.0 - 9576.0)

Abbreviations: RE = right eye; LE = left eye; RNZ = ranibizumab 0.5 mg; AFL = aflibercept 2 mg;

BCVA = best corrected visual acuity collected from ETDRS charts; CRT = central retinal thickness;

SFCT = subfoveal choroidal thickness; CT area = choroidal thickness area underneath the fovea

measured from 1750 µm nasal to 1750 µm temporal from the central fovea in the plane defined by

the horizontal line scan encompassing the fovea. aSystolic (SBP) and diastolic (DBP) blood pressures

were collected in every visit to the hospital. The mean arterial blood pressure (MAP) was calculated

according with the formula DBP + 1/3 (SBP – DBP). The patient was rated as hypertensive whenever

2 MAP values above 110 mmHg were recorded in two separate visits to the hospital.


75

Figure 2.1. Descriptive values during follow-up for mean best corrected visual acuity (BCVA), central retinal thickness (CRT) and choroidal thickness. CRT = 1 mm central retinal

thickness; SFCT= subfoveal choroidal thickness; CT = choroidal thickness measured at 1750 µm, nasal (CT1750n) and temporal (CT1750t) from the fovea in the plane defined

by the horizontal line scan encompassing the fovea; superior (CT1750s) and inferior (CT1750i) from the fovea in the plane defined by the vertical line scan encompassing the

fovea; CT3500 µm (area) = choroidal thickness area underneath the fovea measured from 1750 µm nasal to 1750 µm temporal from the central fovea in the plane defined by the

horizontal line scan encompassing the fovea. 3M = 3 months, after the 3 injection loading dose; 6M = 6 months; 12M = 12 months; 18M = 18 months; 24M = 24 months. All the

measures were performed at baseline (black bar), 3M (red bar), 6M (green bar), 12M (blue bar), 18M (purple bar) and 24M (grey bar) after baseline. All measures were performed

by two independent graders and subsequently reached by consensus. Results related to areas were analysed using the ImageJ software, version 1.48, National Institutes of

Health, USA. Results are presented as mean ± SD.


76

Figure 2.2. Differences in the variables considered between each endpoint and baseline in the laser-naives (white bars) and in the laser-treated subgroups (black bars).

Comparison between the differences exhibited in either group. CRT = 1 mm central retinal thickness; SFCT= subfoveal choroidal thickness; CT = choroidal thickness measured

at 1750 µm, nasal (CT1750n) and temporal (CT1750t) from the fovea, in the plane defined by the horizontal line scan encompassing the fovea; superior (CT1750s) and inferior

(CT1750i) from the fovea in the plane defined by the vertical line scan encompassing the fovea; CT3500 µm (area) = choroidal thickness area underneath the fovea manually

measured from 1750 µm nasal to 1750 µm temporal from the central fovea in the plane defined by the horizontal line scan encompassing the fovea (ImageJ software, version

1.48, National Institutes of Health, USA). BCVA = best corrected visual acuity; 3M = 3 months, after the 3 injection loading dose; 6M = 6 months; 12M = 12 months; 18M = 18

months; 24M = 24 months. Results are presented as mean ± SD. * p<0.05, correspond to comparisons between the laser-naives and in the laser-treated groups and were

obtained using independent samples t-Student or Mann-Whitney tests.

Choroid and diabetes Evaluation of markers of outcome

77

3. Markers of outcome in real-world treatment of

diabetic macular edema3

3 Section 3 is based on the article: Campos A et al., Evaluation of markers of outcome in real-world treatment of diabetic macular edema. Eye Vis (Lond), 2018. 5:27.


78

3.1. Abstract

Purpose: To evaluate short-term markers of outcome in diabetic macular edema (DME).

Methods: Prospective interventional case series included 122 eyes of 122 patients with

recently diagnosed DME. Eyes were treated with a 3-monthly loading dose of

ranibizumab or aflibercept and pro re nata thereafter. Serial enhanced deep imaging SD-

OCT high resolution scans were used to measure subfoveal choroidal thickness (SFCT)

and central retinal thickness (CRT). Anatomic (≥ 10% CRT decrease) and functional

responses (best corrected visual acuity, BCVA gain ≥ 5 letters) were assessed at 3

months and 6 months using univariate and multivariate analyses. Parameters tested

were gender, duration of diabetes, HbA1c, hypertension, CRT, SFCT, BCVA, ellipsoid

zone (EZ) status, subfoveal neuroretinal detachment (SND), anti-VEGF used and laser

naivety. A logistic regression model was applied to find independent markers of outcome.

Results: BCVA increased, CRT and SFCT decreased at 3M and 6M. Good metabolic

control (p = 0.003), intact baseline EZ (p = 0.030), EZ re-grading at 3M (p < 0.001) and

laser naivety (p = 0.001) were associated with better functional outcome. The regression

model showed that baseline SND and CRT are predictors of anatomic response, while

lower baseline BCVA and intact EZ are predictors of functional response.

Conclusions: The presence of SND predicts anatomic response only, while an intact

EZ is critical to achieve a good functional outcome in DME.

Keywords: diabetic macular edema, outcome factors, spectral-domain optical

coherence tomography, anti-vascular endothelial growth factor, laser.


79

3.2. Introduction

Diabetic macular edema (DME) is the leading cause of blindness in patients with

diabetic retinopathy (DR) worldwide [1]. Blood retinal barrier dysfunction, inflammation

and choroidopathy seem to contribute to DME pathogenesis [2]. Optical coherence

tomography (OCT) became the most useful tool for the evaluation and follow-up of DME

and enhanced deep imaging spectral domain optical coherence tomography (EDI SD-

OCT) was successfully used in evaluating choroidal thickness [2]. Treatment of DME

shifted from laser photocoagulation to anti-VEGF therapy. However, DME exhibits wide

variability and heterogeneity [3, 4], as well as different patterns of response to anti-VEGF

treatment [5]. In Protocol T, up to half of the eyes treated were rescued with laser after

24 weeks of treatment [6]. Several attempts have been made to find markers of

prognosis or predictors of treatment response in DME. The ellipsoid zone (EZ) [7],

external limiting membrane (ELM) [8], disruption of the inner retinal layers (DRIL) [9],

hyper-reflective retinal spots (HRS) [3], subfoveal neuroretinal detachment (SND) [4],

central retinal thickness (CRT), subfoveal choroidal thickness (SFCT) [10], among others

[5], have been suggested as predictors. However, some of these reports have limitations,

including retrospective profiles, small sample sizes, symmetry bias and the inclusion of

both Type 1 and Type 2 diabetic patients while evaluating the choroid [2].

The present study attempted to avoid those limitations and was designed to

evaluate some of the predictors of outcome in eyes with recent onset DME, with special

focus on SND, EZ, metabolic control, hypertension, SFCT, CRT, baseline best corrected

visual acuity (BCVA), gender, duration of diabetes and history of macular laser.


80

3.3. Patients and Methods

After approval from the Ethical Committees of the Faculty of Medicine of the

University of Coimbra and of the Leiria Hospital, Type 2 diabetic patients with NPDR and

recent onset DME, naive to intra-vitreal treatment, were included consecutively in a

prospective, institutional study, from June 2014 to June 2017. Each patient gave

informed consent before inclusion in the study.

The study adhered to the tenets of the Declaration of Helsinki and the standards

of Good Scientific Practice of the Faculty of Medicine of the University of Coimbra.

Patients were included either to be treated with ranibizumab 0.5 mg or with aflibercept 2

mg, depending on the availability of aflibercept (June 2015) and baseline BCVA

according to the results at one year of the Protocol T study [6].

Diagnostic criteria for DR and DME were based on past ophthalmic history and

ophthalmic evaluation, including a dilated fundus examination, fundus photography, SD-

OCT, and fluorescein angiography in selected cases.

Patient data including age, gender, blood pressure, duration of diabetes, baseline

glycated hemoglobin (HbA1c) level and previous focal laser therapy were recorded.

DME was considered when clinically significant macular edema (CSME) involving

the central macula (CI-CSME) or a CRT (1 mm central subfield thickness in the OCT-

modified ETDRS grid) ≥ 300 μm was present. Eyes were included when baseline BCVA

ranged from 24 to 78 ETDRS letters (L) (Snellen equivalent 20/320-20/32, LogMAR 1.22-

0.14). When both eyes had DME, only one eye per patient was included [11]. The right

eye was included in patients whose year of birth was an even number and the left eye

was included when the year of birth was an odd number. Eyes with prior focal laser

treatment were not excluded, as long as laser treatment was dated more than six months

prior to enrollment and laser burns did not involve the fovea.


81

Patients were excluded if they had any other previous DR treatment other than

focal photocoagulation or any ocular diseases aside from NPDR in the treated eye. Eyes

with a myopic refractive error of greater than 4 diopters (D) were also excluded [12].

Patients whose eyes had visually significant cataract graded at more than N03 or

NC3 according to the Lens Opacity Classification Scheme were excluded.

Follow-up included baseline, 3 months (3M) and 6 months (6M). BCVA was

measured at every visit using the ETDRS standardized scale at 4 meters distance. Top

score allowed for records in the ETDRS scale was 85 L (Snellen 20/20). Patients were

treated with a monthly 3 injections’ loading dose and on a pro re nata (PRN) regimen

thereafter.

Imaging

EDI SD-OCT (Spectralis; Heidelberg Engineering, Heidelberg, Germany) scans

were performed monthly in all eyes included and guided PRN decision to treat, after the

loading dose. For each study eye, a 6 mm × 6 mm macular cube scan was performed

using the high resolution (HR) posterior pole scanning mode comprising 61 horizontal B-

scans, 120 µm apart from each other, each made up of 1536 A-scans, and a 6-line star

scan, each made up of 1536 A scans, 30º apart from each other, cutting through the

fovea. Two independent raters (AC, do Carmo) measured all scans in a prospective way

and reviewed all of them at the end of the study being masked to the subjects’ outcomes,

and definitive measures were reached by consensus.

The average thickness of the central 1 mm field of the 6 mm OCT-modified

ETDRS grid was used to evaluate changes in the CRT over time. The presumed fovea

was considered as the region with the photoreceptor layer alone and was checked again

retrospectively using the device’s automatic follow-up tool. SFCT was measured using

the horizontal scan of the star scan mode centered at the fovea. Scans were evaluated

by the two scorers after marking the choroid-scleral border (1 scan × rater × follow-up


82

period). SFCT was manually measured from the hyperreflective line of the Bruch’s

membrane to the hyperreflective line of the choroid-scleral interface (Supplementary

Figure 3.1) using the digital caliper tool in the Heidelberg Eye Explorer software.

Whenever there were doubts about the choroid-scleral border, measurements were

compared with the horizontal line scan bypassing the fovea of the macular cube scan.

The integrity of the EZ was evaluated at baseline and after the loading dose, in the central

500 μm in either direction of the fovea (Figure 3.1 A and 3.1 B).

Figure 3.1. Re-rating the ellipsoid zone (EZ) after 3 injections of anti-VEGF. ETDRS grid from

the caliper tool set in place centered at the fovea. A. Horizontal scan, 500 μm each side of the

fovea to evaluate the EZ. Note that laser dots are outside the 1500 μm radius (second circle of

the ETDRS grid has a radius of 1750 μm) from the center of the foveola. B. ETDRS grid set in

place centered at the foveola. Vertical scan, 500 μm each side of the fovea to evaluate the EZ.

The EZ was considered disrupted when there was any focal absence of the second

hyperreflective band in the central 1000 µm either in the horizontal or in the vertical line

scans centered at the fovea of the star scan mode and that could not be attributed to the

A

B

503 µm

503 µm

503 µm

503 µm


83

shadowing effect of cysts or retinal vessels [13]. Whenever the raters did not agree, the

ellipsoid zone was considered unreadable. All scans were performed from 9.00 a.m. to

1.00 p.m.

Evaluation of outcome

For anatomic responders’ calculation, only baseline CRT values ≥ 300 μm were

considered. Eyes with CSME or cysts in the central 1000 μm, but with CRT < 300 μm,

were included in this study but were not considered when checking for anatomic

responders. We defined anatomic responders as eyes having a 10% reduction in the

baseline CRT either at 3M (early) or at 6M (late). Functional responders were also

divided into early and late functional responders. Eyes with BCVA gains of ≥ 5 L from

baseline at 3M were defined as early responders, while eyes with BCVA gains of ≥ 5 L

from baseline at 6M only, were defined as late responders [5].


Nominal data were described by absolute and relative frequencies. Quantitative

data were described using the mean, standard deviation, median, minimum and

maximum in the sample characterization. For other quantitative data, median, minimum

and maximum were calculated but were omitted in the tables for convenience.

Quantitative variables were assessed for normality with Shapiro-Wilk test.

Comparisons between two measures, in different time points, of the same variable, were

analyzed using the paired sample t-test or Wilcoxon test, taking normality requirements

into account. Comparisons of quantitative variables between two groups were performed

with t-Student or Mann-Whitney tests, as applicable. The association between

categorical variables was assessed with Fisher’s test.


84

Linear multivariate regression models, where being an anatomic responder or

being a functional responder were the dependent variables (baseline compared to 6M),

were built up using 12 predictors as independent variables: male gender, baseline BCVA

< 65L (Snellen < 20/50), intact baseline EZ, laser non-naivety, HbA1c, hypertension,

baseline CRT, baseline SFCT, baseline SND, ranibizumab or aflibercept use and

duration of diabetes. In both cases, the predictors were those that bore clinical

significance in addition to those variables found to be statistically relevant (the criterion

was p < 0.1). To evaluate whether laser treatment as a predictor of functional response

was associated with the duration of diabetes, a Fisher’s test was employed. An

interaction variable between diabetes duration and laser treatment was constructed and

a logistic regression model was performed. The interaction variable was built with four

different categories, which were DM ≤ 15 years and no laser treatment, DM > 15 years

and no laser treatment, DM ≤ 15 years and laser treatment, DM > 15 years and laser

treatment. This interaction variable entered in the regression model as a set of three

dummy variables representing the last three categories described before (dummy 1 =

‘DM > 15 years and no laser treatment’, dummy 2 = ’DM ≤ 15 years and laser treatment’,

dummy 3 = ’DM > 15 years and laser treatment’). The assumptions of the model

regarding residuals were observed as well as collinearity. Correlations between

quantitative variables were assessed by computing Pearson’s or Spearman’s correlation

coefficients, depending on whether normality requirements were met or not. The

statistical analyses were performed on the IBM SPSS Statistics 24 and on statistic

platform R v3.3.2, The R Foundation Vienna, Austria. The level of significance adopted

was 0.05.

3.4. Results

From June 2014 to June 2017, 122 eyes from 122 patients were prospectively

included and were followed for 6M. Baseline demographic and ocular characteristics are


85

outlined in Table 3.1. Baseline SND was present in 27 eyes (22.1%). Baseline EZ was

intact in 80 eyes (65.5%), disrupted in 41 eyes (33.6%) and declared unreadable in 1

eye (0.8%). Graders disagreed in 14 eyes (11.1%) and a final decision was reached by

consensus (Figure 3.2 A and 3.2 B). The EZ was graded again at 3M. It was graded as

intact in 89 eyes (73.0%, Figure 3.2 C and 3.2 D) and disrupted in 33 eyes (27.0%, Figure

3.2 E).


Demographic characteristics (n = 122 patients; 68 RE and 54 LE )

Age (y)

Mean ± SD 65.2 ± 8.9

Median (range) 66 (46-85)

Sex Male 66 (54.1%)

Female 56 (45.9%)

Duration of diabetes (y)

1-15 55 (45.1%)

16-25 52 (42.6%)

>25 15 (12.3%)

HbA1c (%)

≤7 31 (25.4%)

>7 and <8 44 (36.1%)

≥8 47 (38.5%) Hypertensiona

Yes 75 (61.5%)

No 47 (38.5%)

Insulin

Yes 64 (52.5%)

No 58 (47.5%)

Ocular characteristics (n = 122 eyes)

Lens status

Phakic 80 (65.6%)

Pseudophakic 42 (34.4%)


86

Intravitreal injection received

RNZ only 93 (76.2%)

AFL only 14 (11.5%)

Both 15 (12.3%)

Laser

Yes 68 (55.7%)

No 54 (44.3%)

Baseline BCVA (ETDRS letters)

Mean ± SD 63.2 ± 12.7

Median (range) 67.0 (24.0 - 78.0)

Baseline CRT

Mean ± SD 432.4 ± 107.0

Median (range) 400.5 (289.0 - 776.0)

Baseline SFCT

Mean ± SD 346.6 ± 75.6

Median (range) 345.0 (124.0 - 580.0)

Baseline EZ

Intact 80 (65.6%)

Disrupted

Unreadable

41 (33.6%)

1 (0.8%)

3M EZ

Intact 89 (73.0%)

Disrupted 33 (27.0%)

Baseline SND

Yes 27 (22.1%)

No 95 (77.9%)

Abbreviations: RE = right eye; LE = left eye; Hb A1c = level of glycated haemoglobin; RNZ =

ranibizumab; AFL = aflibercept; BCVA = best corrected visual acuity collected from ETDRS charts;

CRT = central retinal thickness; SFCT = subfoveal choroidal thickness; EZ = ellipsoid zone; 3M EZ =

re-rating of the EZ after the loading dose; SND = subfoveal neuroretinal detachment. SBP = systolic

blood pressure; DBP = diastolic blood pressure; MAP = mean arterial blood pressure. aSBP and DBP

were measured at baseline and whenever coming back to the hospital, including visits and injections.

MAP was determined using the formula: MAP = DBP + 1/3 (SBP – DBP). The patient was rated as

hypertensive whenever 2 MAP values above 110 mmHg were recorded in two separate visits to the

hospital.


87

Figure 3.2. Examples of the difficulties in rating the ellipsoid zone (EZ) at baseline and after the

3-monthly injection of anti-VEGF. A. A small subfoveal neuroretinal detachment and in the

shadowing cone effect of a retinal cyst makes the rating of the EZ difficult. In this case the EZ

was rated as ‘disrupted’ by consensus. B. The EZ seems to be disrupted with an intact external

limiting membrane (ELM). C. and D. Eyes shown in A and B after the loading dose. The EZ is

now clearly visible, rated as ‘intact’ by both graders. E. EZ after the loading dose being rated as

‘disrupted’.

A

B

C

D

E

502 µm

503 µm

503 µm

502 µm

503 µm

502 µm

503 µm

502 µm

503 µm

503 µm


88

At 3M, graders agreed totally in EZ grading. The mean number of injections given

was 3.0 at 3M and 4.6 ± 1.3 (3 – 7) at 6M.

The mean baseline BCVA improved significantly, while the mean baseline CRT

and the mean baseline SFCT decreased significantly, at 3M and 6M (Supplementary

Table 3.1).

A total of 119 eyes out of 122 had baseline CRT values ≥ 300 μm and were

considered for anatomic response calculation. Mean baseline CRT for a total of 98 eyes

(82.4%) decreased significantly (anatomic responders) while 21 (17.6%) were non-

responders. Baseline BCVA was significantly higher and CRT was significantly lower in

anatomic non-responders. The mean number of injections given was also lower in the

anatomic non-responders (Supplementary Table 3.2).

To test whether these differences might be attributable to eyes with baseline

lower CRTs and better BCVA owing to a ‘floor effect’ that decreased CRT less than 10%,

we re-calculated anatomic response to include eyes with CRT above 350 µm only.

Thereby, allowing for a 10% decrease until 315 µm, which was previously suggested as

the cut-off value for Spectralis SD-OCT [14]. Following this criterion, the differences in

baseline BCVA and CRT between anatomic responders and non-responders did not

stand (Supplementary Table 3.3).

A total of 74 eyes (60.7%) at 3M and an additional 22 eyes (18.0%) at 6M

improved BCVA ≥ 5L and were considered functional responders (96 eyes, 78.7%). A

total of 26 eyes (21.3%) were functional non-responders (Table 3.2).

Metabolic control (HbA1c) was significantly better in functional responders (p =

0.003), whereas the duration of diabetes was not significantly different (p = 0.432)

(Supplementary Table 4).

Laser naivety was strongly associated with being a functional responder (Table

3.2 and Supplementary Table 3.4). Intact baseline EZ was associated with early


89

functional response and EZ re-rating at 3M was even more significantly associated with

being a functional responder.

Table 3.2. Comparison of OCT baseline characteristics and outcome measures between functional

responders and non-responders

Functional

nonresponders

(n = 26)

Early

functional

responders

(n = 74)

Late functional

responders

(n = 22)

p-

valuea

p-

valueb

BCVA

Baseline 65.3 ± 10.5 62.3 ± 13.2 63.8 ± 13.4 0.535 0.447

3M 63.9 ± 13.4 72.2 ± 11.0 65.1 ± 12.6 0.009 0.001

p-value 3M 0.506 <0.001 0.13

6M 64.7 ± 12.0 75.5 ± 9.6 72.7 ± 11.8 <0.001 <0.001

p-value 6M 0.847 <0.001 <0.001

CRT

Baseline 420.7 ± 99.1 435.8 ± 109.5 434.9 ± 111.1 0.563 0.553

3M 346.9 ± 90.3 334.4 ± 62.0 348.3 ± 86.7 0.712 0.792

p-value 3M <0.001 <0.001 <0.001

6M 370.4 ± 111.4 332.3 ± 61.5 311.7 ± 44.4 0.455 0.460

p-value 6M 0.001 <0.001 <0.001

SFCT

Baseline 339.3 ± 63.4 343.7 ± 82.6 365.3 ± 62.5 0.532 0.783

3M 328.2 ± 71.9 319.3 ± 76.2 335.4 ± 57.5 0.746 0.597

p-value 3M 0.061 <0.001 0.003

6M 303.6 ± 66.4 326.9 ± 83.2 321.7 ± 66.2 0.156 0.157

p-value 6M <0.001 0.001 <0.001

Baseline SND

Yes 3 (11.5%) 21 (28.4%) 3 (13.6%) 0.187 0.111

No 23 (88.5%) 53 (71.6%) 19 (86.4%)

Baseline EZ


90

Intact 13 (50.0%) 54 (72.9%) 13 (59.1%) 0.063 0.030

Disrupted 13 (50.0%) 19 (25.7%)c 9 (40.9%)

3M EZ

Intact 12 (46.2%) 62 (83.8%) 15 (68.2%) 0.001 <0.001

Disrupted 14 (53.8%) 12 (16.2%) 7 (31.8%)

Laser

Yes 22 (84.6%) 36 (48.6%) 10 (45.5%) <0.001 0.001

No 4 (15.4%) 38 (51.4%) 12 (54.5%)

Number of Injections

4.4 ± 1.3 4.7 ± 1.3 4.8 ± 1.3 0.267 0.334

Abbreviations: BCVA = best corrected visual acuity scored using the ETDRS letters chart. ETDRS 62 letters (L) are

Snellen 20/58; 64L (20/53); 65L (20/50); 72L (20/36); 73L (20/35) and 76L (20/30). 3M = 3 month endpoint after the

loading dose; 6M = 6 month endpoint. CRT = 1 mm central retinal thickness; SFCT = subfoveal choroidal thickness;

SND = subfoveal neuroretinal detachment; EZ = ellipsoid zone; 3M EZ = re-rating of the ellipsoid zone after the loading

dose. aComparison responders vs non-responders; bcomparison early responders vs non-responders. cOne eye with

unreadable EZ. BCVA increased significantly only in functional responders. CRT and SFCT changes from baseline do

not show a statistically significant difference between responders and non-responders displaying the poor correlation

between functional response and anatomic response. An intact EZ at baseline was present in a higher proportion

among functional responders and that was even more significant with the 3M re-rating. Laser naivety was more

commonly found in functional responders.

Baseline predictors for anatomic responders

According to the multivariate linear regression model, CRT and SND were found

to significantly contribute as predictors. Indeed, following this criterion, the anatomic non-

responders displayed a mean baseline CRT of 367.8 ± 58.2 µm, whereas the anatomic

responders displayed a mean baseline CRT of 450.6 ± 108.3 µm (p < 0.001). SND was

absent in all anatomic non-responders. The model was statistically significant (𝑂𝑚𝑛𝑖𝑏𝑢𝑠

𝑡𝑒𝑠𝑡 χ2(2) = 33.27, p < 0.001), the variance explained was 47%, 𝑅2𝑁𝑎𝑔𝑢𝑒𝑙𝑘𝑒𝑟𝑘𝑒 = 0.470.

To evaluate the model, a ROC analysis was performed, obtaining an area under the

curve (AUC) equal to 0.913 (p < 0.001, CI 95% [0.861; 0.965]).


91

Baseline predictors for functional responders

The multivariate linear regression model indicated low baseline BCVA, laser

naivety, lower HbA1c and intact baseline EZ as strong predictors of being a functional

responder (Table 3.3). Since laser treatment was found to be associated with the

duration of diabetes (Fisher’s test, p = 0.045, OR = 2.150), an interaction variable

between diabetes duration and laser treatment was constructed and a logistic regression

model was performed. According to the new model, laser naivety was not found to be

independent as a predictive factor (Supplementary Table 3.5).

Considering the proportion of eyes that attained higher BCVA scores at 6M, from

the 41 eyes with disrupted baseline EZ, only 11 (26.8%) attained a BCVA ≥ 75 L (Snellen

20/32). However, from the 80 eyes with intact baseline EZ, 61 (76.3%) attained a BCVA

≥ 75 L (p < 0.001), a probability 9 times greater (OR = 8.8, CI 95% [3.7, 20.7]).

As expected, low baseline BCVA (< 65 L, Snellen < 20/50) suggested poor

chances of attaining a high BCVA (≥ 75 L, Snellen ≥ 20/32) at 6M. Indeed, out of the 50

eyes for which low BCVA values were observed, only 14 (28.0%) attained high BCVA

scores. Furthermore, out of the 72 eyes for which BCVA values at baseline were ≥ 65L,

58 (80.6%) attained high BCVA scores, a probability 11 times greater (OR = 10.7, CI

95% [4.6, 24.9]) to achieve a BCVA ≥ 75 L at 6M.


92

Table 3.3. Results of the multivariate linear regression model obtained using twelve predictors of the increase of BCVA from baseline as independent variables

Unstandartized Coefficients Standartized Coefficients 95% Confidence Interval for B

Model B SE Beta t p Lower Bound Upper Bound

(Constant) 38.773 4.219 9.190 0.000 30.414 47.131

BCVAi (L) -0.386 0.056 -0.611 -6.862 0.000 -0.497 -0.274a

Baseline EZ 4.291 1.407 0.257 3.049 0.003 1.502 7.079b

Laser -2.457 1.212 -0.155 -2.027 0.045 -4.858 -0.055c

HbA1c_bin -4.188 1.350 -0.232 -3.102 0.002 -6.862 -1.513d

DM_bin -0.960 1.193 -0.061 -0.805 0.423 -3.323 1.403

RNZ -1.523 2.469 -0.062 -0.617 0.538 -6.415 3.368

AFL -1.588 1.811 -0.086 -0.877 0.382 -5.176 1.999

Abbreviations: BCVAi (L) = baseline best corrected visual acuity in ETDRS letters; EZ = ellipsoid zone; HbA1c_bin = glycated haemoglobin level entered as a dichotomous variable (≤ 7 versus >

7); DM_bin = diabetes duration entered as a dichotomous variable (≤ 15 years versus > 15 years); RNZ = ranibizumab; AFL = aflibercept; B = regression coefficient; SE = standard error for B.

aBaseline BCVA: for each unity of increase in the baseline BCVA there is an average decrease of 0.386 letters in the dependent variable (increase of BCVA after 6 months); bBaseline EZ: eyes

with intact EZ have an average increase of 4.291 letters in the BCVA after 6 months; cLaser: eyes with history of macular photocoagulation have an average decrease of 2.457 letters in the

dependent variable (increase of BCVA after 6 months); dHbA1c: eyes with HbA1c > 7 have an average decrease of 4.188 letters in the dependent variable. In this model, diabetes duration and the

use of ranibizumab or aflibercept were not statistically significant. The regression model obtained was statistically significant (F(4,116) = 14.791, p < 0.001) and the variables explained about 32%

of the variance (R_adj = 0.315). The assumptions of the model regarding residuals were observed as well as collinearity.


93

3.5. Discussion

The aim of this study was to evaluate markers of outcome in DME. Baseline CRT

and baseline SND were predictors of anatomic response to treatment. An intact EZ, good

metabolic control and lower BCVA were found to be baseline predictors of a better

functional response. Moreover, laser naivety was found to be an indicator of better

functional response.

Low baseline BCVA was predictive of having a large recovery (larger number of

letters gained) but not of getting higher final BCVA scores. Therefore, the lower the

baseline BCVA is, the better is the chance of getting a higher recovery in letters

(functional response). However, due to the ‘ceiling effect’ existing in eyes with higher

baseline BCVA, a higher baseline BCVA has a smaller chance of closing a wider gap in

the recovery of letters.

The presence of an intact EZ and better BCVA at baseline, were important for

attaining higher final BCVA scores. The 3M re-rating of the EZ strongly correlated with

being a functional responder. It is not clear whether the improvement of the EZ at 3M

was due to re-arrangement of the photoreceptors, true neuronal regeneration or just

better definition of the OCT scan.

Laser naivety was found to be a predictor of better functional outcome, using the

multivariate linear regression model. This is an important issue since laser was widely

used in most randomized clinical trials [6]. Laser rescue seems to decrease the number

of injections and CRT, at a cost of a lesser gain in the BCVA [15], making the role of

laser rescue questionable [16]. However, when using the logistic regression model, laser

naivety was not independent from the duration of diabetes. This association of a factor

that indicates better prognosis (laser naivety, using the multivariate regression model)

with a factor that does not (duration of diabetes) may be attributed to the shift in DME

treatment, from laser to anti-VEGFs, where laser naivety would be a real prognostic


94

factor indeed or, on the other hand, may indicate that eyes with a history of laser

photocoagulation had prior history of DME, therefore worsening the prognosis [17].

Good metabolic control was associated with being a functional responder, and it

was confirmed to be an independent marker when using the linear multivariate

regression model. These data enhance the importance of good metabolic control when

using non-fixed regimens of treatment. Our results partially agree with the results of a

previous retrospective study in DME patients whose eyes were treated with

bevacizumab, where previous macular laser was correlated with poor functional

response [18].

SND and CRT were powerful markers of anatomic response. Furthermore, as

previously described [19], we found a poor association between anatomic and functional

response that withstood even when the milder forms of DME were withdrawn

(Supplementary Table 3.3). These data correlate with the fact that SND is a marker of

anatomic response only, while an intact EZ is a marker of functional response.

Furthermore, SND probably is a marker of very recent or acute onset DME, particularly

prone to a swift response to treatment, but the final BCVA lies beyond the resolution of

the retinal edema, on photoreceptor integrity [2, 4, 13]. Similar to the study by Vujosevic

et al., we also did not find SND to be a marker of functional outcome [4].

Our results do not agree with previous results that pointed baseline SFCT and

CRT as predictors of outcome [10]. We found baseline SFCT lacking value as a predictor

of outcome using the multivariate linear regression model and analyzed this factor in

detail in a recent report [20].

A 3-monthly injections’ loading dose protocol was used, yet anatomical

responders and early functional responders were evaluated at 3M, where all eyes were

treated alike. Moreover, most of the improvement in BCVA occurs until 3M and this

improvement predicts BCVA in the long term [21]. We used two different anti-VEGFs,

ranibizumab and aflibercept. Available data suggest that these two drugs are mostly


95

similar [22, 23]. According to the multivariate linear regression model, there was no

difference between the two anti-VEGFs (Table 3.3).

The cut-off definition of DME by OCT is elusive or variable in most of trials and

about 1 of 5 cases of DME may be missed if the diagnosis is supported by OCT thickness

measurements only [24]. This is why we used the ETDRS funduscopic criteria of CI-

CSME (hard exudates or hemorrhages within 500 μm of the fovea) to include cases of

DME whose CRT was less than 300 μm. Only Type 2 diabetics were included because

we wanted to check the value of SFCT as a prognostic marker and Type 1 diabetics

have thicker choroids [2]. HR scan mode was used as it gives a higher quality image and

allows a better visualization of the choroidoscleral border and of the EZ.

ELM was not evaluated, since it largely parallels the prognostic profile of the EZ

[8]. Other possible prognostic markers such as cysts and DRIL were not evaluated,

mainly because of the limitation in the input imposed by the multivariate linear regression

models. However, a recent study by one of the authors of this study compared those

factors and the EZ and concluded that an intact EZ was the most reliable OCT marker

of them all [25].

The biggest limitation of this study is that it is not a randomized controlled trial,

does not involve multiple centers, and that factors such as DRIL, HRS or cysts were not

evaluated.

The strengths of this study are related to the fact that it is a real-world study with

a prospective profile, including one eye per patient only, to avoid a Type 1 error (when

comparing two means, concluding the means were different when in reality they were

not different – the rejection of a true null hypothesis) [26], the inclusion of Type 2 diabetics

only, the use of the HR scan mode and the use of the ETDRS charts to evaluate BCVA.


96

3.6. Conclusion

What was known before

An intact EZ and lower baseline BCVA are predictors of functional outcome and

higher baseline CRT is a predictor of anatomic outcome.

What this study adds

SND is a predictor of anatomic outcome but does not predict the functional

outcome.

Neither CRT nor SFCT are predictors of functional outcome.

Good metabolic control is a predictor of outcome in non-fixed regimens.

Laser naivety is associated with being a functional responder but needs further

research since it did not prove to be an independent predictor in the logistic regression

model.

Declarations

Ethics approval and consent to participate: This study was developed after approval from

the Ethical Committees of the Faculty of Medicine of the University of Coimbra and of

the Leiria Hospital.

Availability of data and materials: Most of data generated or analyzed during this study

are included in this published article and in its supplementary information files. The

remaining datasets are available from the corresponding author upon reasonable

request.

Conflict of interests: The authors declare that they have no competing interests.

Funding/support: Grant by the Portuguese Foundation for Science and Technology,

Strategic Project (UID/NEU/04539/2013) and COMPETE-FEDER (POCI-01-0145-

FEDER-007440). EJC was financially supported by the FCT Postdoctoral Fellowship

SFRH/BPD/93672/2013, through European Union and National funds and co-funded by


97

Human Capital Operating Program (Programa Operacional do Capital Humano, POCH).

JM was financially supported by an unrestricted grant from Novartis. The funding

organizations had no role in the design or conduct of this research.

3.7. References

1. Das, A., P.G. McGuire, and S. Rangasamy, Diabetic Macular Edema: Pathophysiology and Novel Therapeutic Targets. Ophthalmology, 2015. 122(7): p. 1375-94.

2. Campos, A., et al., Viewing the choroid: where we stand, challenges and contradictions in diabetic retinopathy and diabetic macular oedema. Acta Ophthalmol, 2017. 95(5): p. 446-459.

3. Vujosevic, S., et al., Imaging retinal inflammatory biomarkers after intravitreal steroid and anti-VEGF treatment in diabetic macular oedema. Acta Ophthalmol, 2017. 95(5): p. 464-471.

4. Vujosevic, S., et al., Diabetic Macular Edema With and Without Subfoveal Neuroretinal Detachment: Two Different Morphologic and Functional Entities. Am J Ophthalmol, 2017. 181: p. 149-155.

5. Ashraf, M., A. Souka, and R. Adelman, Predicting outcomes to anti-vascular endothelial growth factor (VEGF) therapy in diabetic macular oedema: a review of the literature. Br J Ophthalmol, 2016. 100(12): p. 1596-1604.

6. Diabetic Retinopathy Clinical Research, N., et al., Aflibercept, bevacizumab, or ranibizumab for diabetic macular edema. N Engl J Med, 2015. 372(13): p. 1193-203.

7. Tao, L.W., et al., Ellipsoid zone on optical coherence tomography: a review. Clin Exp Ophthalmol, 2016. 44(5): p. 422-30.

8. Muftuoglu, I.K., et al., Integrity of Outer Retinal Layers after Resolution of Central Involved Diabetic Macular Edema. Retina, 2017. 37(11): p. 2015-2024.

9. Sun, J.K., et al., Disorganization of the retinal inner layers as a predictor of visual acuity in eyes with center-involved diabetic macular edema. JAMA Ophthalmol, 2014. 132(11): p. 1309-16.


11. Esen, F., et al., Double-Organ Bias in Published Randomized Controlled Trials of Glaucoma. J Glaucoma, 2016. 25(6): p. 520-2.

12. Meng, W., et al., Axial length of myopia: a review of current research. Ophthalmologica, 2011. 225(3): p. 127-34.

13. Maheshwary, A.S., et al., The association between percent disruption of the photoreceptor inner segment-outer segment junction and visual acuity in diabetic macular edema. Am J Ophthalmol, 2010. 150(1): p. 63-67 e1.

14. Grover, S., et al., Normative data for macular thickness by high-definition spectral-domain optical coherence tomography (spectralis). Am J Ophthalmol, 2009. 148(2): p. 266-71.

15. Schmidt-Erfurth, U., et al., Three-year outcomes of individualized ranibizumab treatment in patients with diabetic macular edema: the RESTORE extension study. Ophthalmology, 2014. 121(5): p. 1045-53.


98

16. Regnier, S., et al., Efficacy of anti-VEGF and laser photocoagulation in the treatment of visual impairment due to diabetic macular edema: a systematic review and network meta-analysis. PLoS One, 2014. 9(7): p. e102309.

17. Brown, D.M., et al., Long-term outcomes of ranibizumab therapy for diabetic macular edema: the 36-month results from two phase III trials: RISE and RIDE. Ophthalmology, 2013. 120(10): p. 2013-22.

18. Joshi, L., et al., Intravitreal bevacizumab injections for diabetic macular edema - predictors of response: a retrospective study. Clin Ophthalmol, 2016. 10: p. 2093-2098.

19. Diabetic Retinopathy Clinical Research, N., et al., Relationship between optical coherence tomography-measured central retinal thickness and visual acuity in diabetic macular edema. Ophthalmology, 2007. 114(3): p. 525-36.

20. Campos, A., et al., Choroidal thickness changes stratified by outcome in real-world treatment of diabetic macular edema. Graefes Arch Clin Exp Ophthalmol, 2018.



23. Sivaprasad, S., et al., Using Patient-Level Data to Develop Meaningful Cross-Trial Comparisons of Visual Impairment in Individuals with Diabetic Macular Edema. Adv Ther, 2016. 33(4): p. 597-609.

24. Virgili, G., et al., Optical coherence tomography (OCT) for detection of macular oedema in patients with diabetic retinopathy. Cochrane Database Syst Rev, 2015. 1: p. CD008081.

25. Santos, A.R., et al., Optical coherence tomography baseline predictors for initial best-corrected visual acuity response to intra-vitreal anti-vascular endothelial growth factor treatment in eyes with diabetic macular edema: The Chartres Study. Retina, 2018. 38(6): p. 1110-1119.

26. Armstrong, R.A., Statistical guidelines for the analysis of data obtained from one or both eyes. Ophthalmic Physiol Opt, 2013. 33(1): p. 7-14.


99


Figure 3.1. ETDRS grid in place centered at the fovea. Note that ETDRS grid plotted (7.2 mm in

diameter) is larger than the OCT-modified ETDRS grid (6 mm in diameter) plotted to access

central retinal thickness CRT. A. and B. HR horizontal scans used to measure the SFCT. ETDRS

grid inner circle is 1200 μm (a) and middle circle is 3600 μm wide (b). C. HR vertical scan with

SFCT measured underneath the fovea.

A

B

C


100

Table 3.1. Baseline values for BCVA, CRT and SFCT. Differences in BCVA, CRT and SFCT

between endpoints and baseline, and number of injections given

Baseline

(n = 122)

3M - Baseline

(n = 122)

6M - Baseline

(n = 122)

63.2 ± 12.7 5.9 ± 7.1 9.5 ± 7.9

BCVA (L) <0.001 <0.001

60.6%a 77.9%a

432.4 ± 107.0 -92.8 ± 103.9 -95.7 ± 108.6

CRT (µm) <0.001 <0.001

346.6 ± 75.6 -22.5 ± 35.8 -25.6 ± 44.8

SFCT (µm) <0.001 <0.001

n Injections 3.0 ± 0.0

(3.0–3.0)

4.6 ± 1.3

(3.0–7.0)

Abbreviations: BCVA (L) = best corrected visual acuity scored using the ETDRS letters (L) chart: 63L are equivalent

to LogMAR 0.44 or Snellen 20/55; CRT = 1 mm central retinal thickness; SFCT = subfoveal choroidal thickness; n

injections = number of intra-vitreal injections given; 3M = after the 3-monthly injection loading dose; 6M = 6 months;

N injections = number of injections given at each endpoint. Results are presented as mean ± SD and range for

injections. aProportion of eyes that displayed an increase of 5L or more when compared to the baseline.


101


at baseline, 3 months and 6 months

Anatomic non-

responders

(n = 21)

Anatomic responders

(n = 98) p-value

BCVA (L)

Baseline 67.4 ± 11.1 61.9 ± 12.9 0.037

3M 72.7 ± 11.5 68.2 ± 12.6 0.088

6M 75.7 ± 10.7 71.9 ± 11.6 0.109

Percentage increasing ≥5L 61.9% 62.2% 1.000

p-value 3M <0.001 <0.001

p-value 6M <0.001 <0.001

CRT (µm)

Baseline 367.8 ± 58.2 450.6 ± 108.3 <0.001

3M 366.2 ± 71.8 336.4 ± 72.3 0.026

6M 367.2 ± 71.2 332.1 ± 74.3

p-value 3M 0.832 <0.001

p-value 6M 0.941 <0.001

SFCT (µm)

Baseline 350.6 ± 73.7 347.3 ± 76.1 0.859

3M 322.9 ± 72.0 326.0 ± 82.3 0.858

6M 326.7 ± 70.6 321.0 ± 78.9 0.762

p-value 3M 0.003 <0.001

p-value 6M 0.036 <0.001

Baseline SND

Yes 0 (0.0%) 27 (27.6 %) 0.003

No 21 (100%) 71 (72.4 %)

Baseline EZ

Intact 18 (85.7%) 59 (60.8%) 0.042

Disrupted 3 (14.3%) 38 (39.2%)

Laser

Yes 14 (66.7%) 51 (52.0%) 0.239

No 7 (33.3%) 47 (48.0%)


Abbreviations: BCVA = best corrected visual acuity scored using the ETDRS letters (L) chart: 62L are Snellen 20/58,

67L (20/46), 68L (20/44), 72L (20/36), 73L (20/35) and 76L (20/30); 3M = 3 month endpoint after the loading dose;

6M = 6 month endpoint; CRT = 1 mm central retinal thickness; SFCT = subfoveal choroidal thickness; SND = subfoveal

neuroretinal detachment; EZ = ellipsoid zone. For anatomic responders’ calculation, only eyes with baseline CRT

≥300 μm were considered, n = 119 eyes. Eyes were considered responders if they had a 10% decrease from baseline

CRT. The difference in vision gain between anatomic responders and non-responders was not statistically significant.

A higher mean baseline CRT and baseline SND correlated with anatomic response. The mean baseline SFCT

decreased significantly with treatment in both groups with no statistically significant difference between them.


102


using a cut-off for CRT of 350 μm

Anatomic

responders

(n = 10)

Anatomic

nonresponders

(n = 84)

p-value

BCVA (L)

Baseline 63.0 ± 13.5 60.2 ± 13.0 0.450

3M 69.3 ± 14.8 67.1 ± 12.2 0.407

6M 71.7 ± 14.0 70.9 ± 11.6 0.572

p-value 3M 0.021 <0.001

p-value 6M <0.001 <0.001

CRT (µm)

Baseline 407.0 ± 64.3 471.8 ± 102.3 0.029

3M 401.6 ± 80.0 343.8 ± 74.5 0.006

6M 404.1 ± 85.8 340.3 ± 76.9 0.007

p-value 3M 0.432 <0.001

p-value 6M 0.415 <0.001


Abbreviations: BCVA = best corrected visual acuity scored using the ETDRS letters (L) chart: 60L are Snellen 20/63,

63L (20/55), 67L (20/46), 69L (20/42), 71L (20/38), and 72L (20/36); 3M = 3-month endpoint after the loading dose;

6M = 6-month endpoint; CRT = 1 mm central retinal thickness. For anatomic responders’ calculation, only eyes with

baseline CRT ≥350 μm were considered, n = 94 eyes. Unlike the former criterion using a CRT cut-off of 300 μm for

calculating anatomic outcome displayed in Supplementary Table 2.2, the differences in the number of injections given

and in the baseline BCVA did not withstand with a CRT cut-off of 350 μm. Only baseline CRT was significantly higher

in the anatomic responders.


103

Table 3.4. Demographic characteristics of functional responders and non-responders

Functional

non-responders (n = 26)

Functional responders

(n = 96)

p-value

Age (years)

Mean ± SD 64.3 ± 8.6 65.4 ± 9.0 0.562

Median (range) 66 (47-78) 66 (46–85)

Sex

Male 11 (42.3%) 55 (7.3%) 0.190

Female 15 (57.7%) 41 (42.7%)

Duration of diabetes (years)

1-15 9 (34.6%) 46 (47.9%) 0.432

16-25 14 (53.9%) 38 (39.6%)

>25 3 (11.5%) 12 (12.5%)

HbA1c (%)

<7 1 (3.8%) 30 (31.3%) 0.003

>7 and <8 15 (57.7%) 29 (30.2%)

>8 10 (38.5%) 37 (38.5%)

Hypertensiona

Yes 19 (73.1%) 56 (58.3%) 0.256

No 7 (26.9%) 40 (41.7%)

Insulin

Yes 15 (57.7%) 49 (51.0%) 0.659

No 11 (42.3%) 47 (49.0%)

Laser

Yes 22 (84.6%) 46 (47.9%) <0.001

No 4 (15.4%) 50 (52.1%)

Abbreviations: HbA1c = level of glycated hemoglobin (percentage); SBP = systolic blood pressure; DBP = diastolic

blood pressure; MAP = mean arterial blood pressure. MAP was determined using the formula: MAP = DBP + 1/3 ×

(SBP - DBP). aThe patient was rated as hypertensive whenever two MAP values above 110 mmHg were recorded in

two separate visits to the hospital. Baseline demographic characteristics show a statistically significant difference

between functional responders and non-responders for metabolic control and laser treatment. Interestingly, the

duration of diabetes is not a factor influencing the prognosis when defining functional response as a gain of 5 letters.


104

Table 3.5. Logistic regression model using all predictors of BCVA increase as independent

variables entering the interaction between duration of diabetes and laser treatment

Predictor (reference) B SE p-value 95% CI for B

Baseline BCVA (L) -0.397 0.055 <0.001 (-0.505, -0.289)

Baseline EZ 4.917 1.453 0.001 (2.040, 7.795)

Dummy 1 1.136 1.808 0.531 (-2.444, 4.717)

Dummy 2 -0.936 1.794 0.603 (-4.490, 2.617)

Dummy 3 -3.591 1.567 0.024 (-6.694, -0.488)

Constant 32.590 3.247 <0.001 (26.158, 39.022)

Abbreviations: BCVA (L) = best corrected visual acuity scored using the ETDRS letters (L) chart; EZ = ellipsoid zone;

B = regression coefficient; SE = standard error for B; 95% CI for B is the 95% confidence interval for the regression

coefficient; DM = duration of diabetes. The interaction variable between diabetes duration and laser treatment has

four different categories: DM ≤15 years and no laser treatment; DM >15 years and no laser treatment; DM ≤15 years

and laser treatment; DM >15 years and laser treatment. This interaction variable entered in the regression model as

a set of three dummy variables representing the last three categories described before (Dummy 1 = DM >15 years

and no laser treatment; Dummy 2 = DM ≤15 years and laser treatment; Dummy 3 = DM >15 years and laser treatment).

In this model, the variables Dummy 1 and Dummy 2 were not statistically significant. The meaning of the regression

coefficient is similar to the first model (Table 3.3) except for the variable Dummy 3. In this case, the B value means

that eyes of patients being diabetic for more than 15 years and undergone laser treatment present on average a

decrease of 3.591 letters in the dependent variable (increase of BCVA after 6 months) when compared to the eyes of

patients that did not have laser treatment and have diabetes for less than 16 years. The model attained was statistically

significant (F(5,115) = 12.624, p <0.001) and the variables explained about 33% of the variance (R_adj = 0.326). The

assumptions of the model regarding residuals were observed as well as collinearity.

Choroid and diabetes Choroid and retina in T1D

105

4. Inflammatory cells proliferate in the choroid and

retina without choroidal thickness change in

Type 1 diabetes4

4 Section 4 is based on an article submitted to Experimental Eye Research.


106

4.1. Abstract

Purpose: Increasing evidence points to inflammation as a key factor in the pathogenesis

of diabetic retinopathy (DR). Choroidal inflammatory changes in diabetes have been

reported and in vivo choroidal thickness (CT) has been searched as a marker of

retinopathy with contradictory results. We aimed to investigate the early stages in the

choroid and retina in an animal model of Type 1 diabetes.

Methods: Type 1 diabetes was induced in male Wistar rats via a single i.p.

streptozotocin injection. At 8 weeks after disease onset, CT, choroidal vascular density,

VEGF and VEGFR2 expression, microglial cell and pericyte distribution were evaluated.

Results: Diabetic rats showed no significant change in CT and choroidal vascular

density. A widened pericyte-free gap between the retinal pigment epithelium and the

choroid was observed in diabetic rats. The immunoreactivity of VEGFR2 was decreased

in the retina of diabetic rats, despite no statistically significant difference in the

immunoreactivity of VEGF. The density of microglial cells significantly increased in the

choroid and retina of diabetic rats. Reactive microglial cells were found to be more

abundant in the choroid of diabetic rats. Evidences of the interconnection between the

superficial, intermediate, and deep plexuses of the retina were also observed.

Conclusions: At early stages, Type 1 diabetes does not affect choroidal thickness and

choroidal vascular density. Proliferation and reactivity of microglial cells occurs in the

choroidal stroma and the retina. The expression of VEGFR2 decreases in the retina.

Keywords: Choroid; retina; Type 1 diabetes; choroidal thickness; microglia.


107

4.2. Introduction

Diabetic retinopathy (DR) is the most common complication of diabetes and is

characterized by alterations in the blood-retinal barrier (BRB), inflammation and

choroidopathy [1]. Although Type 1 diabetes (T1D) accounts for less than 10% of all

cases of diabetes, DR progression is faster and more severe than in Type 2 diabetes [2].

DR was for long considered to be a pure microvascular disease, but increasing

evidences point to inflammation as a key factor in the pathogenesis of DR [3].

Inflammatory events develop either in the retina and the choroid during the course of DR

[4, 5]. The choroid is essential for the nutrition and water clearance of the outer retina

[6]. A diabetic choroidopathy has long been described, although clear evidence of the

role of the choroid in the pathophysiology of DR is lacking.

Several attempts have been made to establish a link between alterations in the

choroidal thickness (CT), assessed by optical coherence tomography (OCT), and the

progression of DR in humans. However, CT as a surrogate of choroidal flux, DR or

choroidal inflammation in studies with human diabetic subjects failed to be reliable [7].

OCT has been used to search for retinal changes in streptozotocin (STZ)-induced

diabetic rats [8, 9]. However, in vivo CT has not been assessed in animal models of

diabetes.

While involved in the regular homeostasis of the retina, under the low level

chronical inflammation in the diabetic retina, due to hyperglycemia, dyslipidemia and

oxidative stress, the retinal resident innate immune cells, microglial cells, become

activated, change their morphologic appearance from “dendritic” to “ameboid”, and start

to produce pro-inflammatory mediators [4]. These mediators are known to lead to

neuronal cell dysfunction, and damage capillary pericytes and endothelial cells, exposing

the endothelia to vascular endothelial growth factor (VEGF), resulting in inner BRB


108

breakdown [10, 11]. Furthermore, migration of activated microglial cells from the retina

to the choroid by transcytosis has been demonstrated in diabetes [1, 12].

VEGF-A is the VEGF isoform related to vasculogenesis and angiogenesis in

homeostasis and disease. VEGF-A triggers its effects in the retina and choroid through

complex receptor-mediated signaling cascades, involving tyrosine kinase VEGF receptor

2 (VEGFR2), which is the major mediator of mitogenesis, angiogenesis and

microvascular permeability [11]. VEGF secreted at the basolateral side of retinal RPE

cells contribute to keep the fenestrations of the choriocapillaris endothelium, primordial

for the nourishment of the outer retina, pan-retinal dehydration and outer BRB

homeostasis. Nevertheless, increased VEGF levels may lead to junction-protein loss,

altered polarization and loss of function of the RPE cells, with disruption of the outer BRB

and macular edema [1]. Comparison of VEGF and VEGFR2 profiles throughout the

retina and choroid in T1D has not been assessed yet.

Herein, we aimed to evaluate the impact of the early stages of T1D on the choroid

and retina, searching for molecular and cellular signatures of the normal physiological

functioning, as well as of pathological change in disease.

4.3. Materials and Methods

Animals

All procedures got previous approval by the Animal Welfare Committee of the

Coimbra Institute for Clinical and Biomedical Research (iCBR), Faculty of Medicine,

University of Coimbra. The animals received humane care according to the criteria

outlined in the Guide for the Care and the Use of Laboratory Animals prepared by EU

Directive 2010/63/EU for animal experiments and with the Association for Research in

Vision and Ophthalmology (ARVO) statement for animal use.

Male Wistar Han rats (8-weeks old) were randomly assigned to T1D (n = 16) and

control (n = 12) experimental groups. T1D was induced with a single intra-peritoneal


109

injection of STZ (65 mg/kg), freshly dissolved in 10 mM sodium citrate buffer, pH 4.5,

(Sigma, St. Louis, MO, USA) [13]. Hyperglycemic status (glycemia > 250 mg/dL) was

confirmed two days later using a glucometer (Ascensia ELITE™, Bayer Corporation,

Mishawaka, IN, USA). The rats were weighted, glycemia was measured, blood samples

were collected (i) at the beginning of the study, i.e., just before STZ injection (these

values were omitted for convenience), (ii) 2 days after STZ injection (diabetes onset),

and (iii) 8 weeks after diabetes onset. Hemoglobin A1c (Hb A1c) was measured at 8

weeks after diabetes onset only.

Optical coherence tomography (OCT)

The CT was evaluated by spectral domain optical coherence tomography (SD-

OCT, Phoenix Micron IV, Phoenix Research Labs, Pleasanton, CA, USA) [14]. SD-OCT

scans were performed in both eyes of all rats, (i) at the beginning of the study, i.e., just

before STZ injection in the animals randomised for the diabetic cohort, and (ii) 8 weeks

after diabetes onset. The animals were anesthetised via i.p. injection with ketamine 80

mg/kg+xylazine 5 mg/kg (80:5, for short; Imalgene 1000, Merial, Lyon, France, and

Rompum, Bayer, Leverkusen, Germany, respectively), and cornea anesthesia (4

mg/mL oxybuprocaine hydrochloride; Anestocil®, Laboratório Edol, Carnaxide,

Portugal), pupils dilation (1%, Tropicil®, Laboratório Edol, Carnaxide, Portugal) and

corneal hydration (2% Methocel™, Dávi II Farmacêutica S.A., Barcarena, Portugal) was

kept during procedure. The lens was placed closer to the rat’s eye, such that an inverted

image was obtained, and the deeper structures were placed closer to zero-delay [15].

SD-OCT scans were obtained above the optic nerve head (ONH), in both eyes

of all rats, in the area within 1 to 3 ONH diameter from the optic disc. The CT was

automatically measured using the InSight image segmentation software (v.1; Voxeleron

LLC - Image analysis solutions, Chabot Drive, CA, USA). Lines of automatic

segmentation of the choroid delivered by the software were manually replaced at the


110

boundary lines, i.e., at the RPE and the choroid-scleral border (Supplementary Figure

4.1). The mean CT per scan was determined and the final CT value was obtained by

averaging the results of three scans from distinct locations.

Tissue preparation for cryosections

The animals were anesthetised, as described above, and intracardially perfused

with 0.1 M phosphate buffer saline (PBS, 137 mM NaCl, 2.7 mM KCl, 1.8 mM KH2PO4,

10 mM NaH2PO4, pH 7.4), followed by 4% paraformaldehyde (PFA) in 0.1 M PBS. The

enucleated eyes of diabetic rats (n = 9) and controls (n = 7) were post-fixed in 4% PFA,

for 1 h. The eyes were washed in successive solutions of PBS, and immersed in solutions

containing 15% and 30% of sucrose in PBS, for 1 h in each solution. Afterwards, they

were embedded in a 1:1 30% sucrose and embedding resin (Shandon™ Cryomatrix™,

Thermo Fisher Scientific, Waltham, MA, USA) solution, before freezing in dry ice. Eyeball

sections, 14 µm-thick, from both right and left eyes, were obtained using a cryostat (Leica

CM3050S, Nussloch, Germany), at -22ºC, and mounted on adhesive slides (Superfrost

Plus™, Thermo Fisher Scientific). A total of four sections (140 µm apart) were collected

per slide.

Immunofluorescence

Eye cryosections were labelled to assess the retinal and choroidal structure,

following a procedure described previously [16]. The cryosections were rehydrated twice

in PBS for 5 min, followed by blocking and permeabilization for 1 h in 10% goat serum

and 0.5% Triton X-100 in PBS. The cryosections were then incubated overnight with

primary antibodies (Supplementary Table 4.1) diluted in 0.5% Triton X-100, at 4ºC. After

washing with PBS, the crysections were incubated with corresponding secondary

antibodies (Supplementary Table 4.2) diluted in 0.5% Triton X-100, for 1 h. After

washing, the cryosections were incubated with 1:5,000 4’,6-diamidino-2-phenylindole


111

(DAPI, Invitrogen™), and coverslipped using mounting medium (Glycergel, Dako,

Carpinteria, CA, USA).

Digital images were captured using an inverted fluorescence microscope (Axio

Observer.Z1, Zeiss, Carl Zeiss Meditec AG, Jena, Germany), using a Plan-Apochromat

20×/0.8 objective), and a laser scanning confocal inverted microscope (LSM 710 Axio

Observer, Zeiss), using a Plan-Apochromat 20×/0.8 objective.

Eight bit images were analyzed using the ImageJ software (version 1.48, National

Institutes of Health, USA) [17]. Iba1+ and MHC class II+ cells were manually counted in

the choroid and retina. Data were expressed as number of cells/mm of choroidal or

retinal length, respectively. Rat endothelial cell antigen 1 (RECA-1) and NG2

immunoreactivity was scored in the choroid as mean fluorescence intensity per area

selected (reference area selected of 10,737.08 ± 6,306.11 µm2), while fluorescent NG2+

cells and RECA-1 focal immunostaining were manually counted in the retina. VEGF and

VEGFR2 immunoreactivity were quantified as mean fluorescence intensity/area for each

layer analysed. All results were expressed as the mean count of 12 slices per eye (right

eyes only).

Direct labelling and visualization of choroidal vessels in

sclerochoroidal whole mounts

Intracardiac perfusion with PBS was performed in rats under anesthesia (i.p.;

ketamine:xylazine 160:30). Intra-cardiac perfusion of 1,1’-dioctadecyl-3,3,3’,3’-

tetramethylindocarbocyanine perchlorate (DiI, Cat. #D-282, Invitrogen/Molecular

Probes, Carlsbad, CA, USA) 0.120 mg/mL in 1% glucose in PBS was applied, following

perfusion with 4% PFA. The eyes were enucleated, and sclerochoroidal whole mounts

were prepared. The whole mounts were fixed in 4% PFA for 15 min, washed with PBS,

and blocked with 10% goat serum in 0.3% Tween in PBS for 1 h. Samples were

incubated with the primary antibodies diluted in 3% goat serum in PBS for 3 days at 4ºC

(Supplementary Table 4.1). After washing overnight with PBS, whole mounts were


112

incubated with secondary antibodies (Supplementary Table 4.2) overnight at 4ºC. After

washing, explants were flat mounted onto glass slides using mounting medium.

The images were obtained by confocal microscopy using Plan-Apochromat 20×

objective lens NA0.8 or EC Plan-Neofluor 40x oil objective lens NA1.3, and were

analysed using the ImageJ software. Iba1+ and MHC class II+ cells were classified as

elongated or round-shaped cells, and were manually counted in the choroid in all in-

depth planes of the slice (Supplementary Figure 4.2). Z-stacks were 34 µm-thick and

were imaged in 11 in-depth slices. The choroidal vascular density (defined as the

percentage of total area covered by choroidal vessels) [18] was determined at ≤10 µm

(choriocapillaris) and >10 µm outwards from the RPE plane (medium and large vessels)

in a selected area (213 × 213 µm, Supplementary Figure 4.3). Results were expressed

as the mean of 14 counts per eye (n = 5 controls; n = 7 diabetics).


The statistical analysis was performed using SPSS (Version 25.0, IBM Corp.,

Armonk, NY, USA). The Shapiro-Wilk test was used to assess the normality of data (p >

0.05). The normally distributed data were evaluated concerning the homogeneity of

variance, using the Levene’s test (p > 0.05). The independent Student’s t-test was used

to compare the means between two experimental groups for the same variable.

Statistical significance was defined as p < 0.05. Values were presented as mean ±

standard error of the mean (SEM).


113

4.4. Results

Choroidal thickness and vascular density do not change in Type 1

diabetes

T1D was induced in 16 rats while 12 animals were in the control group. Diabetic

rats exhibited significantly decreased body weight (p < 0.001), hyperglycemia (p <

0.001), and increased HbA1c (control: 6.2 ± 0.1% versus T1D: 8.3 ± 0.3%, t(13) = 4.486,

p = 0.001), at 8 weeks after diabetes onset (Supplementary Figure 4.4).

CT was assessed in both eyes of all rats by OCT (Figure 4.1A). At the beginning

of the study, CT was similar in both experimental groups (38.00 ± 1.18 µm, in the control

group, and 39.13 ± 0.88 µm, in the diabetic group; t(52) = - 0.773, p = 0.443). Likewise,

8 weeks after diabetes onset, no significant differences were detected between both

groups (39.83 ± 0.91 µm, in the control group, and 39.44 ± 0.62 µm, in the diabetic group;

t(52) = 0.369, p = 0.714) (Figure 4.1B).

Twelve rats were assigned for choroidal whole mounts analyses (n = 5 controls

and n = 7 diabetics). A trend towards increased vascular density was observed in the

middle+outer choroid of diabetic rats. However, no statistically significant differences in

vascular density were found when comparing the choroidal vasculature between both

experimental groups (Figure 4.1C and Supplementary Table 4.3).


114

Figure 4.1. Effect of Type 1 diabetes on the choroidal thickness. (A) Representative OCT images

of the retina and choroid of control and diabetic rats, at the beginning of the study (baseline) and

at 8 weeks after diabetes onset. Scale bar: 50 µm. (B) Choroidal thickness of control and diabetic

rats, at the baseline and 8 weeks after diabetes onset, based on in vivo OCT line scans. Bars

represent mean ± SEM (control rats, n = 12; diabetic rats, n = 16). (C) Vascular density analysis

at the inner (≤ 10 µm from the outer RPE plane) and middle + outer choroid (> 10 µm from the

outer RPE plane) in control and diabetic rats assessed in sclerochoroidal wholemounts, at 8

weeks after diabetes onset, based on Dil labelling of choroidal vessels. Bars represent mean ±

SEM (control animals, n = 5; diabetic animals, n = 7).

GCL: ganglion cell layer; INL: inner nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear

layer; RPE: retinal pigment epithelium.

B

A

Control

Ba

seli

ne

8

we

ek

s

Diabetic

INL

GCL

ONL

RPEChoroid

Control Diabetic

0

20

40

60

80

100

inner choroid outer choroid

Ch

oro

idal va

scu

lar

den

sity

(%)

C

Control Diabetic

0

10

20

30

40

50

0 8

Ch

oro

idal th

ickn

ess (

µm

)

Time (week)

baseline

inner middle + outer

INL

GCL

ONL

RPEChoroid


115

Type 1 diabetes changes pericyte cells distribution in the inner

choroid

Sixteen rats (7 controls and 9 diabetics) were assigned for immunoreactivity

experiments in eye cryosections. Immunoreactivity of NG2 evidenced polarity in

distribution of perivascular mural cells (pericytes) at the innermost choroidal area. NG2+

cells were found to leave unmarked ‘blank’ gaps behind, at the area between the RPE

posterior cell line and the inner choroid, drawing a jagged pattern, corresponding to

multifocal decreases in mural cells. This pattern was more evident in diabetic rats (Figure

4.2A). RECA-1 immunoreactivity at the choriocapillaris showed continuity with the RPE

line (Figure 4.2B). Evidences of interconnecting vessels between retinal plexuses were

visualised by immunolabelling RECA-1 (Supplementary Figure 4.5), between (i) the

superficial and intermediate plexuses, (ii) the intermediate and deep plexuses, and (iii)

the superficial and deep plexuses. There were no statistically significant differences in

the immunoreactivity of NG2 and RECA-1 in the whole choroidal or retinal areas,

between both experimental groups (Supplementary Table 4.4).


116

Figure 4.2. Effect of T1D on the localization of mural and endothelial cells in the choroid and

retina. Representative images showing the immunolabelling pattern of (A) NG2 and (B) RECA-1

in the choroid and retina of control and diabetic rats, 8 weeks after diabetes onset (control

animals, n = 7; diabetic animals, n = 9). Polarization in the disposition mural cells is highlighted

(white arrows). Scale bar: 100 µm.



NG

2

Control Diabetic

RE

CA

-1

A

B

Control Diabetic

GCL

INL

Choroid

ONL

RPE

GCL

INL

Choroid

ONL

RPE

GCL

INL

Choroid

ONL

RPE

GCL

INL

Choroid

ONL

RPE


117

Type 1 diabetes decreases the expression of VEGR2 in the retina

VEGF immunoreactivity was higher at the outer limiting membrane (OLM) and

RPE, being still important at the retinal nerve fibre layer and choroid (Figure 4.3A and

C). There was a trend for an increase in the VEGF immunoreactivity in diabetic rats, but

this difference was not statistically significant at any of the locations considered (Figure

4.3C and Supplementary Table 4.5).

VEGFR2 immunoreactivity was of overall much lower magnitude of fluorescence,

being observed mainly in the retinal nerve fibre and ganglion cell layers (GCL), outer

plexiform layer (OPL) and OLM, in control and diabetic rats (Figure 4.3B and D).

Conversely to VEGF’s, the VEGFR2 immunoreactivity significantly decreased in diabetic

rats in the inner nuclear layer (INL) (t(13) = 2.337, p = 0.036) and outer nuclear layer

(ONL) (t(13) = 2.465, p = 0.028) (Figure 4.3D and Supplementary Table 4.5).


118


119

Figure 4.3. Effect of early T1D on the immunoreactivity of VEGF and VEGFR2 in the choroid and

retina. Representative images showing the immunoreactivity of (A) VEGF and (B) VEGFR2 in the

choroid and retina of control and diabetic rats, at 8 weeks after diabetes onset. Scale bar: 100

µm. (C) VEGF and (D) VEGFR2 immunoreactivity in eye cryosections based on 12 independent

specimen counts per eye. VEGF and VEGFR2 immunoreactivity was quantified as fluorescence

intensity/area per layer. Counting was done for the right eye only, in all animals. Bars represent

mean ± SEM (control animals, n = 7; diabetic animals, n = 9). Significance: *p < 0.05.

GCL + RNFL: ganglion cell layer and retinal nerve fibre layer; IPL: inner plexiform layer; INL: inner

nuclear layer; OPL: outer plexiform layer; ONL: outer nuclear layer; OLM: outer limiting

membrane; RPE: retinal pigment epithelium.

C

D

Control Diabetic

0

500

1000

1500

2000

2500

Flu

ore

sce

nce

in

ten

sity/a

rea

0

100

200

300

400

500

600

700

Flu

ore

sce

nce

in

ten

sity/a

rea

Control Diabetic

**


120

Interestingly, VEGF immunoreactivity across the retina co-localises with MG

marker, vimentin (Figure 4.4). Maximum immunoreactivity of both, VEGF and vimentin,

was found at the GCL in its innermost area and retinal nerve fibre layer, limited by the

inner limiting membrane or MG basement membrane, peaking again at the plexiform

layers until the OLM, where MG end feet intertwine with PRs outer segments.

Figure 4.4. Co-localization of the immunoreactivity of vimentin and VEGF in the retina.

Representative eye cross-sections of (A) control and (B) diabetic rats, at 8 weeks after diabetes

onset, immunolabelled against vimentin and VEGF. Scale bar: 100 µm.



A Control

Vimentin VEGF Vimentin VEGF DAPI

DA

PI

DA

PI

B Diabetic

Vimentin VEGF Vimentin VEGF DAPI

GCL

INL

Choroid

ONL

RPE

GCL

INL

Choroid

ONL

RPE


121

Type 1 diabetes increases microglial cells’ density and reactivity in

the choroid and retina

In eye cryosections, we observed a trend to higher Iba1+ cells’ density in the

choroid of diabetic rats. In the retina of diabetic rats, Iba1+ cells’ density significantly

increased (t(9.293) = - 2.785, p = 0.021; Figure 4.5A and C and Supplementary Table

4.6). Iba1+ cells in the retina of control rats were found mostly in the inner retinal layers.

Conversely, in diabetic rats, Iba1+ cells were observed throughout the retina, including

in the outer retinal layers, ONL and OPL.

MHC class II+ cells were significantly increased in the choroid of diabetic rats (t(14)

= - 2.669, p = 0.018). No MHC class II+ cells were observed in the retina of both

experimental groups (Figure 4.5B and D and Supplementary Table 4.6).

Iba1 and MHC class II immunolabelling was further performed in sclerochoroidal

whole mounts, after cleaning and labelling choroidal blood vessels by cardiac perfusions

with PBS and DiI (n = 5, controls; n = 7 diabetics). Iba1+ cells density in the choroidal

stroma were statistically significantly increased in diabetic rats (Figure 4.5E and

Supplementary Table 4.7), which is in line with the aforementioned trend obtained in eye

cryosections. Statistically significant differences were observed for elongated cell counts

(t(10) = -3.201, p = 0.009). Likewise, a trend towards increased choroidal round MHC

class II+ cell counts in T1D rats was observed (Figure 4.5E and Supplementary Table

4.7).

Interestingly, Iba1+ and MHC class II+ cells are located increasingly outwards in the

choroidal stroma, respecting a polarity in distribution of cells as observed with pericytes,

leaving a cell-free space behind at the innermost choroidal area (Supplementary videos

4.1, 4.2, 4.3, 4.4 and 4.5).


122

Iba1

MH

C c

lass II

A

B

Control Diabetic

Control Diabetic

GCL

INL

Choroid

ONL

RPE

GCL

INL

Choroid

ONL

RPE

GCL

INL

Choroid

ONL

RPE

GCL

INL

Choroid

ONL

RPE


123

Figure 4.5. Effect of Type 1 diabetes on microglial cell counts and reactivity in the choroid and

retina. Representative images showing the immunoreactivity of (A) Iba1+ and (B) MHC class II+

cells in the choroid and retina of control and diabetic rats, at 8 weeks after diabetes onset. Scale

bar: 100 µm. (C) Iba1+ cell counts in the retina, and (D) Iba1+ and MHC class II+ cell counts in the

choroid, collected from immunolabelling of eye cross-sections. Bars represent mean ± SEM

(control animals, n = 7; diabetic animals, n = 9). (E) Iba1+ and MHC class II+ cells density in the

choroid collected from sclerochoroidal wholemounts. Bars represent mean ± SEM (control

animals, n = 5; diabetic animals, n = 7). Significance: *p < 0.05; **p < 0.01.

MHC class II: major histocompatibility complex class II; IPL: inner plexiform layer; INL: inner

nuclear layer; OPL: outer plexiform layer.

0

10

20

30

40

elongated round elongated round

Iba1+ cells MHC class II+ cells

Num

ber

of

cells

all

in-d

epth

pla

nes

C

Control Diabetic

D

Control Diabetic

0

5

10

15

20

25

30

35

IPL OPL total retina

Iba1

+cells

/mm

0

5

10

15

20

25

30

Iba1 MHC class II

Positiv

e c

ells

/mm

*

*

Control Diabetic

E

**


124

4.5. Discussion

CT has been used as a surrogate for choroidal blood flow, DR, diabetic

choroidopathy and diabetic macular edema (DME) in diabetic patients. Most studies

found CT to decrease after panphotocoagulation for proliferative DR and after anti-VEGF

treatment for DME. However, different studies led to different conclusions about CT

changes with diabetes, DR staging and DME [6, 19, 20]. Most reasons are related to

bias in collecting and treating data, but other reasons are linked to CT inter- and intra-

individual variability and difficulty in evaluating the suprachoroid [7]. CT decreases with

age and T1D patients are usually younger with thicker choroids. The evidences of

diabetes as an inflammatory disease and the findings that one third of patients with DME

are resistant to anti-VEGF therapy and responsive to steroids [10, 21-24] brought the

focus back on diabetic choroidopathy.

OCT has been used in animals to check for retinal changes in diabetes, but not

for changes in the choroid [9]. Rats are the most used model for studying human

diseases and their choroidal structure is closest to human’s than other models’ [25, 26].

In this work, to the best of our knowledge, in vivo CT was evaluated by OCT for the first

time in STZ-induced T1D rats. No significant differences in CT were observed between

control and diabetic rats. Perhaps the duration of diabetes (8 weeks) might not be long

enough to cause significant differences in the CT. Nonetheless, this was an aggressive

model of T1D, with hyperglycemia, body weight loss and mean Hb A1c of 8.3 ± 0.3%.

DiI-perfused sclerochoroidal whole mounts observed by confocal microscopy

revealed no significant differences in vessel density in the inner or outer choroid between

control and diabetic rats. Indeed, the choriocapillaris analysis failed to reveal significant

differences in density between control and diabetic rats, paralleling similar findings in

humans in histopathological studies [27] and OCT angiography (OCTA) [28]. Subtle,

transient, reversible changes in the choriocapillaris detected by OCTA, like vascular


125

remodeling, dependent on, or causing, RPE cell stress and disease, may be more

accountable in humans. Nevertheless, our data agree with previous data from electron

microscopy, hemodynamic and histopathological studies of the choriocapillaris in T1D

rats, where no differences in the capillary diameter, despite reduced blood flow, and no

changes in the luminal surface area or in the area of intervessel stroma, were reported

[29]. Vascular remodeling or focal alterations were suggested, since there was evidence

of cellular debris in the stroma and of migrating endothelial cells in the choriocapillaris.

Mural cells, essentially pericytes, were reported to have a sparse,

noncircumferential, polarized distribution, leaving a gap free of cells at the

choriocapillaris side facing the RPE, a polarized disposition [30]. We observed this

polarized disposition in the form of multifocal gaps between the RPE and the inner

choroid by NG2 immunolabelling, drawing a jagged pattern. Enhancement of this jagged

pattern to some degree was observed in diabetic rats, where pericyte cell disposition

leaves slightly wider cell-free gaps between the choroid and the RPE. The total

expression of NG2 in the choroid was not different between control and diabetic rats,

suggesting that there is not a significant loss of perivascular mural cells that wrap around

medium and large vessels. However, if depletion of pericytes at the choriocapillaris is

related with vascular remodeling, the increased jagged pattern observed in diabetic rats

might just be an indirect clue of increased vascular remodeling at the choriocapillaris

level in T1D. Interestingly, polarized disposition of cells was also observed for

inflammatory cells in the choroidal stroma via Iba1 and MHC class II immunolabelling in

sclerochoroidal whole mounts (Supplmentary videos 1-5).

VEGF immunoreactivity was higher at the retinal nerve fibre layer, GCL and RPE,

being still relevant at the choroid. VEGFR2 immunoreactivity peaked at the innermost

retina but was negligible at the RPE and choroid. Moreover, there was a trend of the

immunoreactivity of VEGF to be increased in diabetic rats while the immunoreactivity of

VEGFR2 was higher in control rats. These results agree with data previously described,


126

where VEGF and VEGFR2 were found to be present at the same locations in normal rat

eyes as constitutive survival factors and VEGF was increased in diabetic eyes [31].

Conversely, they do not agree with results reporting VEGFR2 to be increased in the

retina and choroid of early diabetic rats, where VEGFR2 expression was mainly located

in the capillaries [32]. Our data seem to be more in line with previous results relating

VEGFR2 expression to neurons than to blood vessels [33]. These differences may

depend on a shorter duration of diabetes, on a distinct animal model, on a different

antibody, or on a different locus of probe or antibody linkage to the tyrosine kinase

receptor 2. The profile of VEGF immunoreactivity is consistent with VEGF secretion by

Müller cells and the RPE, as reported by others [34, 35].

A statistically significant increase of elongated Iba1+ cells and a trend towards

increased MHC class II+ cell number was observed in diabetic choroidal stroma. Cell

migration from the retina into the choroid has been described before [12]. We

demonstrated that the Iba1+ and MHC class II+ cells were actually present in the

choroidal stroma and not just passing by within the vessel lumina, since the choroidal

vasculature had been previously cleaned by perfusions. These data were confirmed by

the analysis of eye cryosections, where Iba1+ cells in the retina and MHC class II+ cells

in the choroid were statistically significantly increased in diabetic rats. The presence of

Iba1+ cells in the outer retina layers was a distinct feature of diabetic retinas, as

previously described [36], including in the outer retinal layers [10, 37].

The accumulation of inflammatory cells in the outer retina may be due to

increased cell traffic from the retina into the choroid [12] or to increased metabolic-driven

hypoxemia at the outer retina [38] with VEGF-mediated recruitment of inflammatory cells

[11, 39]. Increased number and activity of inflammatory cells in the retina displace

pericytes from the endothelium and increase endothelial leakage via increased VEGF

secretion and rupture junction proteins [10, 40], resulting in inner BRB breakdown. These


127

are features that confirm the association of T1D with increased inflammation in the retina

and in the choroid, including of choriocapillaris remodeling [4, 5].

We sought for molecular, inflammatory, and mural cells signatures in T1D,

because diabetes is not just a VEGF-mediated but it is also an inflammatory condition

[3, 41].

VEGF-driven inner BRB breakdown at the venule side of the superficial retinal

vasculature has been pointed as the earliest event in diabetes [42] and it was related to

pericyte loss and retinal hypoxemia [43, 44]. However, retinal hypoxemia may not be

present in the early stages of T1D in rats [45]. Microglial cell migration throughout the

retina has been described in T1D [10, 12], and, particularly, microglia cell activation has

been observed since the early stages of the disease [36]. Pericytes protect the

endothelial cells from exposition to VEGF [11]. Mitochondrial superoxide production in

response to hyperglycaemia, rather than hypoxemia, may be the first event to dislodge

pericytes from the capillary wall [46].

In this work we found important simultaneous events occurring at the choroid and

retina in T1D. Accumulation of microglial cells in the outer retina and of

monocyte/macrophage cells in the choroid, along with loss of the polarized distribution

of choroidal pericytes, widening the gap between the RPE and the inner choroid, may

significantly disturb the function of the RPE. Iba1+ and MHC class II+ cells showed the

same kind of polarized distribution in the inner choroid. We hypothesise that the

increased burden of inflammatory cells in the T1D choroid may change this polarized

distribution. All these findings point to a role of choriocapillaris remodeling and outer BRB

in the initial stages of T1D retinopathy.

In vivo CT characterization by OCT and the cell signatures in whole thickness

tissues, such as the choroid and retina, was also performed in T1D rats. This approach

was important to understand the role of the choroid in DR while looking at events that


128

are simultaneously taking place in the retina. Immunofluorescence colocalization studies

allowed the identification of several types of cells involved in the choroid and retina of

T1D rats, as well as changes in their number and disposition. These studies also allowed

to map and compare the expression of VEGF and VEGFR2 in the choroid, RPE and

retina. T1D rats used in our experiments had a significant metabolic imbalance (Hb A1c

8.3 ± 0.3%) and 8 weeks of ongoing DM left untreated in rat, is roughly equivalent to six

years of unbalanced diabetes in a human being without any treatment at all [25].

Although we refer often to this model as “early T1D”, it does not match early diabetes in

man, since T1D in human beings is seldom left untreated for so long. Therefore, it is not

surprising that some features found in this work correspond to a longer disease duration

in T1D patients.

In summary, the choroidal thickness measured by OCT is probably a doubtful

surrogate of choroid flux and of an ongoing T1D choroidopathy and retinopathy in rats

as in humans, leading to confusing and contradictory results. We showed evidence that

an inflammatory condition develops in the choroid of rats in early T1D, as previously

reported in humans, even before gross histopathological alterations are observed. We

have shown that inflammatory cells are more abundant in the choroid and retinal outer

layers in T1D. We further evidenced polarized disposition of pericytes, Iba1+ and MHC

class II+ cells in the choroid.

Funding

This work was supported by the Portuguese Foundation for Science and Technology

(UID/NEU/04539/2013, UID/NEU/04539/2019, UIDB/04539/2020 and

UIDP/04539/2020), COMPETE-FEDER (POCI-01-0145-FEDER-007440), Centro 2020

Regional Operational Programme (CENTRO-01-0145-FEDER-000008: BrainHealth

2020) and Novartis. JM was financially supported by an unrestricted grant from Novartis.


129

4.6. References


2. Romero-Aroca, P., et al., Differences in incidence of diabetic retinopathy between type 1 and 2 diabetes mellitus: a nine-year follow-up study. British Journal of Ophthalmology, 2017. 101(10): p. 1346.

3. Rübsam, A., S. Parikh, and P.E. Fort, Role of inflammation in diabetic retinopathy. International Journal of Molecular Sciences, 2018. 19(4): p. 942.

4. Sorrentino, F.S., et al., The importance of glial cells in the homeostasis of the retinal microenvironment and their pivotal role in the course of diabetic retinopathy. Life Sciences, 2016. 162: p. 54-59.

5. Lutty, G.A., Diabetic choroidopathy. Vision Research, 2017. 139: p. 161-167.



8. He, M., et al., ALDH2 attenuates early-stage STZ-induced aged diabetic rats retinas damage via Sirt1/Nrf2 pathway. Life Sci, 2018. 215: p. 227-235.

9. Masser, D.R., et al., Functional changes in the neural retina occur in the absence of mitochondrial dysfunction in a rodent model of diabetic retinopathy. J Neurochem, 2017. 143(5): p. 595-608.


11. Ferrara, N., H.P. Gerber, and J. LeCouter, The biology of VEGF and its receptors. Nat Med, 2003. 9(6): p. 669-76.


13. Baptista, F.I., et al., Diabetes induces changes in KIF1A, KIF5B and dynein distribution in the rat retina: implications for axonal transport. Exp Eye Res, 2014. 127: p. 91-103.



16. Fernandez-Sanchez, L., et al., Loss of outer retinal neurons and circuitry alterations in the DBA/2J mouse. Invest Ophthalmol Vis Sci, 2014. 55(9): p. 6059-72.


130

17. Schneider, C.A., W.S. Rasband, and K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis. Nature Methods, 2012. 9(7): p. 671-675.

18. Kumar, A., et al., Vascular associations and dynamic process motility in perivascular myeloid cells of the mouse choroid: implications for function and senescent change. Invest Ophthalmol Vis Sci, 2014. 55(3): p. 1787-96.







25. Sengupta, P., The Laboratory Rat: Relating Its Age With Human's. Int J Prev Med, 2013. 4(6): p. 624-30.

26. Castro-Correia, J., Understanding the choroid. Int Ophthalmol, 1995. 19(3): p. 135-47.

27. Lutty, G.A., Diabetic choroidopathy. Vision Res, 2017. 139: p. 161-167.

28. Carnevali, A., et al., Optical coherence tomography angiography analysis of retinal vascular plexuses and choriocapillaris in patients with type 1 diabetes without diabetic retinopathy. Acta Diabetol, 2017. 54(7): p. 695-702.

29. Braun, R.D., C.A. Wienczewski, and A. Abbas, Erythrocyte flow in choriocapillaris of normal and diabetic rats. Microvascular Research, 2009. 77(3): p. 247-255.


31. Kim, I., et al., Constitutive Expression of VEGF, VEGFR-1, and VEGFR-2 in Normal Eyes. Investigative Ophthalmology & Visual Science, 1999. 40(9): p. 2115-2121.

32. Sun, D., et al., Molecular imaging reveals elevated VEGFR-2 expression in retinal capillaries in diabetes: a novel biomarker for early diagnosis. FASEB J, 2014. 28(9): p. 3942-51.


131

33. Okabe, K., et al., Neurons limit angiogenesis by titrating VEGF in retina. Cell, 2014. 159(3): p. 584-96.

34. Le, Y.Z., VEGF production and signaling in Muller glia are critical to modulating vascular function and neuronal integrity in diabetic retinopathy and hypoxic retinal vascular diseases. Vision Res, 2017. 139: p. 108-114.

35. Strauss, O., The retinal pigment epithelium in visual function. Physiol Rev, 2005. 85(3): p. 845-81.

36. Rungger-Brandle, E., A.A. Dosso, and P.M. Leuenberger, Glial reactivity, an early feature of diabetic retinopathy. Invest Ophthalmol Vis Sci, 2000. 41(7): p. 1971-80.

37. Lechner, J., O.E. O'Leary, and A.W. Stitt, The pathology associated with diabetic retinopathy. Vision Res, 2017. 139: p. 7-14.

38. Yu, D.Y., et al., Retinal capillary perfusion: Spatial and temporal heterogeneity. Prog Retin Eye Res, 2019. 70: p. 23-54.

39. Holmes, D.I. and I. Zachary, The vascular endothelial growth factor (VEGF) family: angiogenic factors in health and disease. Genome Biol, 2005. 6(2): p. 209.

40. Cebe-Suarez, S., A. Zehnder-Fjallman, and K. Ballmer-Hofer, The role of VEGF receptors in angiogenesis; complex partnerships. Cell Mol Life Sci, 2006. 63(5): p. 601-15.

41. Adamis, A.P., Is diabetic retinopathy an inflammatory disease? British Journal of Ophthalmology, 2002. 86(4): p. 363.







132


Figure 4.1. SD-OCT scan acquisition for choroidal thickness evaluation. (A) Linear

scans (blue line) were acquired above the optic nerve head (ONH), in the area within 1

to 3 ONH diameter from the optic disc. (B) Segmentation lines were manually drawn at

the inner (retinal pigment epithelium, RPE; green line) and outer (choroidal-scleral

border; yellow line) boundaries of the choroid. Scale bar: 50 µm.

GCL: ganglion cell layer; INL inner nuclear layer; OPL: outer plexiform layer; ONL: outer


A B

INL

GCL

ONL

RPE

Choroid


133

Figure 4.2. Quantification of Iba1+ and MHC class II+ cells in sclerochoroidal whole

mounts (n = 5 controls; n = 7 diabetics). (A) Iba1+ cells were counted in all depth planes

of the choroid. (B) MHC class II+ cells marked using yellow dots to avoid duplicate

counting. Scale bar: 100 µm.


134

Figure 4.3. Quantification of the vascular density in sclerochoroidal wholemounts

(defined as the percentage of total area covered by choriocapillaris vessels) using the

‘image>adjust>threshold’ window tool of ImageJ to obtain the percentage of vascular

coverage (n = 5 controls; n = 7 diabetics). (A) At the inner choroid/choriocapillaris (z-

stacks collected at ≤10 µm from the posterior RPE cell plane), and (B) at middle and

outer choroid/medium and large vessels (z-stacks collected at >10 µm from the posterior

RPE cell plane). Scale bar: 50 µm.

A B


135

Figure 4.4. Body weight and glycaemia values of animals, since diabetes onset. (A) At

diabetes onset (0 weeks), both control (254.9 ± 7.5 g) and diabetic (267.8 ± 7.0 g) groups

did not differ significantly for body weight; at 8 weeks after diabetes onset, control rats

were significantly heavier (control: 398.7 ± 8.5 g vs diabetic: 266.7± 0.8 g). (B) Glycaemia

was significantly higher in the diabetic rats, at diabetes onset (control: 109.6 ± 3.4 g/dL

vs diabetic: 502.3 ± 27.7 g/dL), and 8 weeks later (control: 100.0 ± 2.1 g/dL vs diabetic:

537.8 ± 17.1 g/dL). HbA1c was significantly higher in diabetic rats, at 8 weeks after

diabetes onset (control: 6.2 ± 0.1% vs diabetic: 8.3 ± 0.3%, t(13)=4.486, p = 0.001). Bars

represent mean ± SEM (control animals, n = 12; diabetic animals, n = 16). Significance:

***p < 0.001.


136

Figure 4.5. Interconnecting vessels between retinal plexuses were observed in eye

cross-sections immunolabelled against RECA-1: superficial and middle plexuses (white

arrow); middle and deep plexuses (yellow arrow); deep and superficial plexuses (red

arrow). Scale bar: 100 µm.




137

Table 4.1. Primary antibodies

Antigen Target Host Supplier Cat. N.er Dilution

Iba1 Microglia Rabbit Wako Chemicals Inc.,North Chesterfield, VA, USA 019-19741 1:250

MHC II Activated microglia Mouse Bio-RAD Laboratories, Hercules, CA, USA MCA46R 1:200

NG2 Cell membrane chondroitin sulfate proteoglycan Rabbit Merck, Darmstadt, Germany AB530 1:200

RECA-1 Endothelial cells Mouse Abcam Inc., Cambridge, MA, USA ab9774 1:200

VEGF-A Signal growth factor protein Mouse Abcam Inc., Cambridge, MA, USA ab1316 1:200

VEGFR2 VEGF receptor 2 Rabbit Abcam Inc., Cambridge, MA, USA ab131241 1:200

Iba1: calcium binding adapter molecule 1; MHC II: major histocompatibility complex II; RECA-1: rat endothelial cell antigen; NG2: proteoglycan NG2/Cspg4, neuroglial antigen

2; VEGF: vascular endothelial growth factor; VEGFR2: vascular endothelial growth factor receptor 2.

Table 4.2. Secondary antibodies

Fluorophore Target Host Supplier Cat. N.er Dilution

Alexa Fluor® 488 Mouse Ig G Goat IntrovitrogenTM, Thermo Fisher Scientific, Waltham, MA, USA A-11001 1:500

Alexa Fluor® 568 Rabbit Ig G Goat IntrovitrogenTM, Thermo Fisher Scientific, Waltham, MA, USA A-110036 1:500

Abbreviations: Ig G, immunoglobulin G.


138

Table 4.3. Choroidal vascular density

Choroidal vessels area/total area

Control Diabetic p

Inner choroid 93.96 ± 0.75 94.49 ± 0.72 0.631

Middle + outer

choroid

71.98 ± 2.29 79.72 ± 2.82 0.080

Quantification of the choroidal vascular density in the inner and outer choroid, after labelling choroidal

blood vessels by cardiac perfusion with DiI (defined as the percentage of total area covered by

choriocapillaris vessels), using the ‘image>adjust>threshold’ window tool of ImageJ to obtain the

percentage of vascular coverage. Inner choroid: z-stacks collected at ≤ 10 µm (choriocapillaris). Middle +

outer choroid: z-stacks collected at > 10 µm from the outer RPE plane (medium and large vessels).

Quantitative analyses were performed based on 14 independent counts per eye in each and all in-depth z-

stacks per specimen. Data are expressed as mean ± SEM (n = 5, control group; n = 7, diabetic group).

Significance: p < 0.05.


139

Table 4.4. NG2 and RECA-1 immunoreactivity in the retina and choroid

NG2 RECA-1

Control Diabetic p Control Diabetic p

Positive cells/mm

GCL 9.439 ± 0.38 10.287 ± 0.49 0.204

OPL 13.755 ± 1.04 13.199 ± 1.82 0.796

Retina 10.04± 0.53 9.85 ± 1.79 0.826

Fluorescence intensity/area

Choroid 617.30 ± 42.69 591.13 ± 39.83 0.661 857.18 ± 55.0 812.89 ± 31.91 0.485

Quantification of the immunoreactivity of NG2 and RECA-1 in the retina and choroid scored as (i) manually

counted positive cells, in the retina; (ii) fluorescence intensity per selected area (reference area selected

of 10,737.08 ± 6,306.11 µm2), in the choroid. Counting was done for the right eye only, in 12 independent

specimen counts per eye. Values are expressed as mean ± SEM (control animals, n = 7; diabetic animals,

n = 9). Significance: p < 0.05. GCL: ganglion cell layer; OPL: outer plexiform layer.


140

Table 4.5. VEGF and VEGFR2 immunoreactivity in the retina and choroid

VEGF VEGFR2

Fluorescence intensity/area


GCL+ RNFL

725.47 ± 31.24 752.05 ± 84.25 0.549 574.42 ± 27.76 537.61 ± 33.97 0.425

IPL 682.85 ± 46.30 757.25 ± 57.02 0.339 279.71 ± 20.33 228.75 ± 23.96 0.135

INL 561.14 ± 29.21 581.25 ± 42.03 0.701 240.29 ± 16.87 184.88 ± 16.55 0.036*

OPL 883.26 ± 56.61 834.00 ± 46.94 0.511 349.02 ± 11.41 321.77 ± 18.17 0.242

ONL 678.00 ± 58.12 734.75 ± 66.21 0.536 220.71 ± 11.59 170.38 ± 16.14 0.028*

OLMPr 1426.29 ± 189.01 1791.50 ± 237.79 0.260 303.12 ± 12.27 267.25 ± 25.36 0.246

RPE 1070.32 ± 52.03 1208.12 ± 56.22 0.098 155.20 ± 10.46 165.07 ± 6.22 0.418

Choroid 705.37 ± 39.19 769.02 ± 21.52 0.164 133.88 ± 9.02 135.52 ± 6.23 0.881

Quantification of the VEGF and VGFR2 immunoreactivities in the retina and choroid based on 12

independent specimen counts per eye. VEGF and VEGFR2 immunoreactivities were quantified as

fluorescence intensity/area per layer. Counting was done for the right eye only, in all animals. The

expression of VEGF tends to be higher in diabetic animals, but the difference is not statistically significant.

There is a tendency to higher expression of VEGFR2 in control animals; statistically significant differences

in the retina were observed only at the level of the INL and ONL though. Values are expressed as mean

± SEM (control animals, n = 7; diabetic animals, n = 9). Significance: *p < 0.05.

VEGF: vascular endothelial growth factor; VEGFR2: VEGF receptor 2; GCL + RNFL: ganglion cell layer

and retinal nerve fibre layer; IPL: inner plexiform layer; INL: inner nuclear layer; OPL: outer plexiform

layer; ONL: outer nuclear layer; OLMPr: outer limiting membrane and photoreceptor inner segments;

RPE: retinal pigment epithelium.


141

Table 4.6. Iba1+ and MHC class II+ cell counts in the retina and choroid

based on immunolabelling of eye cryosections

Iba1+ cells/mm MHC class II+ cells/mm


IPL 7.73 ± 0.32 8.13 ± 0.71 0.652 - - -

OPL 0.37 ± 0.17 2.74 ± 1.06 0.056 - - -

Retina 18.57 ± 0.8 26.64 ± 2.78 0.021* - - -

Choroid 23.11 ± 1.99 26.42 ± 1.13 0.148 15.28 ± 0.1 18.39 ± 0.68 0.018*

Quantification of Iba1+ and MHC II+ cells in eye cryosections based on 12 independent specimen counts

per eye. Counting was made as the number of cell/mm of choroidal or retinal length, respectively.

Counting was done for the right eye only, in all animals. Data are expressed as mean ± SEM. Significance:

*p < 0.05. Iba1+ cells are significantly increased in the retina of Type 1 diabetic animals. MHC class II+

cells are significantly increased in the choroid of Type 1 diabetic animals. Values are expressed as mean

± SEM (control animals, n = 7; diabetic animals, n = 9).

MHC class II: major histocompatibility complex class II; IPL: inner plexiform layer; OPL: outer plexiform

layer.


142

Table 4.7. Iba1+ and MHC class II+ cells’ density in the choroid

Positive cells/mm

Marker Cell shape Control Diabetic p

Iba1 elongated 23.40 ± 0.76 29.74 ± 1.57 0.009**

round 2.53 ± 0.36 2.91 ± 0.37 0.502

MHC class II elongated 1.83 ± 0.67 1.89 ± 0.61 0.954

round 18.98 ± 1.64 25.17 ± 2.36 0.116

Iba1+ and MHC II+ in the choroid of control and diabetic rats in all in-depth z-stacks of the choroid, based

on immunolabelling of sclerochoroidal whole mounts, after labelling choroidal blood vessels by

cardiac perfusion with DiI. Iba1+ and MHC II+ cells were manually counted in the choroid. Quantitative

analyses were performed based on 14 independent counts per eye in each and all in-depth z-stacks per

specimen. Images were collected by confocal microscopy with Zeiss EC Plan-Neofluor 40x oil objective

lens, NA 1.3. Significance: **p < 0.01. Elongated Iba1+ cells are significantly increased in the choroid of

diabetic animals. Values are expressed as mean ± SEM (control animals, n = 5; diabetic animals, n = 7).

MHC class II: major histocompatibility complex class II.


143

Video 4.1. Sequenced images showing the localization of Iba1+ cells (green) sparing

the innermost choroid outwards the RPE cell plane of a control rat, aged 16 weeks, at 8

weeks after diabetes onset in diabetic rats. Slice thickness: 14.8 μm. AC and JM

authored the video: 18’’; 9,982 KB.

Video 4.2. Sequenced images showing the localization of Iba1+ cells (green) outwards

the inner choroidal vascular network (red) in a control rat aged 16 weeks, at 8 weeks

after diabetes onset in diabetic rats. Slice thickness: 23.1 μm. AC and JM authored the

video: 20’’; 10,925 KB.


the inner choroidal vascular network perfused using DiI (red) in a diabetic rat, at 8 weeks

after diabetes onset. Slice thickness: 23.1 μm. AC and JM authored the video: 13’’; 7,188

KB.

Video 4.4. Sequenced images in the same location as in video 4.3, showing the

localization of MHC class II+ cells (purple) outwards the choriocapillaris perfused by DiI

(red) in a diabetic rat, at 8 weeks after diabetes onset. Slice thickness: 23.1 μm. AC and

JM authored the video: 14’’; 7,433 KB.

Video 4.5. Sequenced images in the same location as in videos 4.3 and 4.4, showing

the co-localization of Iba1+ cells (green) and MHC class II+ cells (purple) outwards the

inner choroidal vascular network perfused using DiI (red) in a diabetic rat, at 8 weeks

after diabetes onset. Slice thickness: 23.1 μm. AC and JM authored the video: 13’’; 7,107

KB.


https://drive.google.com/drive/folders/1YWm9HQ8ijOKu0XWj6bytF7TpG7Lp_FDo?usp

=sharing


145

5. Choroidal and retinal structural, cellular and

vascular changes in Type 2 diabetes5

5 Section 5 is based on an article submitted to Plos One.


146

5.1. Abstract

Purpose: Increasing evidences point to inflammation as a key factor in the pathogenesis

of diabetic retinopathy (DR). Choroidal changes in diabetes have been reported and

several attempts were made to validate in vivo choroidal thickness (CT) as a marker of

retinopathy. We aimed to study choroidal and retinal changes associated to retinopathy

in an animal model for spontaneous Type 2 diabetes, Goto-Kakikazi (GK) rats.

Methods: Sclerochoroidal whole mounts and cryosections were prepared from 52-week

GK rats and age-matched Wistar Han controls. CT was measured by optical coherence

tomography. Microglia reactivity, pericyte and endothelial cells distribution, and

immunoreactivity of vascular endothelial growth factor (VEGF) and VEGF receptor 2

(VEGFR2) were evaluated by immunofluorescence. Choroidal vessels were visualized

by direct perfusion with 1,1’-dioctadecyl-3,3,3’,3’-tetramethylindocarbocyanine

perchlorate (DiI). Choroidal vascular density was evaluated by fluorescence microscopy.

Results: GK rats had increased CT (58.40 ± 1.15 µm versus 50.90 ± 1.58 µm, p < 0.001),

reduced vascular density of the choriocapillaris (p = 0.045), increased Iba1+ cells density

in the outer retina (p = 0.003) and increased VEGFR2 immunoreactivity in most retinal

layers (p = 0.021 to 0.037). Choroidal microglial cells and pericytes showed polarity in

their distribution, sparing the innermost choroid. This cell-free gap at the inner choroid

was more pronounced in GK rats.

Conclusions: GK rats have increased CT with decreased vascular density at the

innermost choroid, increased VEGFR2 immunoreactivity in the retina and Iba1+ cells

density at the outer retina.

Keywords: Type 2 diabetes; choroid; retina; choroidal thickness; microglia; VEGFR2.


147

5.2. Introduction

Type 2 diabetes (T2D) accounts for more than 90% of all cases of diabetes and is

related, mostly, with age, sedentary life and diet overload [1, 2]. Diabetic retinopathy

(DR) and its complications are commonly treated with anti-VEGF agents [3, 4] and the

greater focus has been put on the role of VEGF on the pathogenesis of DR [5-7]. In fact,

retinal hypoxemia has been also related to the pathogenesis of DR [8, 9]. VEGF-driven

inner blood-retinal barrier (BRB) breakdown in the venule side of the superficial retinal

vasculature has been pointed as the earliest event in DR [5] and it was related to pericyte

loss and hypoxemia [9]. Nevertheless, retinal hypoxemia was reported to be absent on

the early stages of diabetes in rats [10].

Actually, VEGF is not increased in the vitreous of all patients with diabetic macular

edema (DME) while pro-inflammatory markers were found to be increased [11].

Accordingly, about one third of patients with DME fail to respond to anti-VEGF therapy

[12, 13]. Furthermore, steroids proved to be effective in treating post-surgical cystoid

macular edema and DME [14, 15]. Those facts pointed DR to be an inflammatory

condition and that was confirmed experimentally [16-18].

Mitochondrial superoxide production in response to hyperglycemia may be the first

event to dislodge pericytes from the capillary wall. The retinal resident innate immune

system, macrophage-like microglial cells, become activated and start to produce

proinflammatory mediators [18]. Overproduction of reactive oxygen species (ROS) leads

to the increased formation of advanced glycated end-products (AGEs), activation of

protein kinase C, aldose reductase, and nuclear factor kB, leading to pericyte loss,

exposure of endothelial junction proteins to VEGF, BRB breakdown and diabetic

microangiopathies [19].

The perspective of DR as inflammatory disease brought new attention on previous

works describing choroidal inflammatory alterations in diabetes, named as ‘diabetic


148

choroidopathy’, including Brüch’s membrane deposits and increased thickness, and

choriocapillaris dropout [20-22].

Since the advent of optical coherence tomography (OCT) [23], the choroidal

thickness (CT) [24] has been sought as a surrogate of choroidal flux, diabetic

choroidopathy or DR, but the results are conflicting and disappointing [25-29]. The

thinning of the choroid with the anti-VEGF treatment increased the assumption that the

choroid thickens in DR and DME in an exclusive VEGF-dependent manner [29].

Nevertheless, the thinning of the choroid under anti-VEGF treatment seems to be only a

side effect with poor prognostic value [30], focusing back the attention on cellular and

molecular signatures that might take place in diabetic choroidopathy.

Goto-Kakizaki (GK) rats present some features found in patients with DR and they

are a model to study the kinetics and events of T2D [31]. Increased NO production, early

inner BRB breakdown and migration of activated microglial cells from the retina to the

choroid by transcytosis has been demonstrated in GK rats [32-34], but the breakdown of

the outer BRB and Brüch’s membrane permeability in diabetes is not completely

understood yet [35, 36].

We investigated the impact of T2D on the CT and choroidal vascular density, as

well as on endothelial cells and pericytes, microglial cell reactivity, VEGF and VEGFR2

immunoreactivity, in the retina and choroid of GK (52 weeks old, 52 W) and age-matched

control Wistar Han rats.

5.3. Materials and Methods

Animals

GK rats are characterized by an early increase in serum insulin and also by mild

hyperglycemia, insulin resistance and mild T2D [34]. All procedures were approved by


149

the Animal Welfare Committee of the Coimbra Institute for Clinical and Biomedical

Research (iCBR), Faculty of Medicine, University of Coimbra. The animals received

humane care according to the criteria outlined in the Guide for the Care and the Use of

Laboratory Animals prepared by EU Directive 2010/63/EU for animal experiments and

with the Association for Research in Vision and Ophthalmology (ARVO) statement for

animal use. At 52 W, the animals were weighted, blood samples were collected for

measurement of glycemia and Hb A1c (Vantage® Analyzer, Siemens Healthinners,

Erlanger, Germany).

Optical coherence tomography

The animals were anesthetized via i.p. injection with ketamine 80 mg/kg + xylazine

5 mg/kg (80:5, for short; Imalgene 1000, Merial, Lyon, France, and Rompum, Bayer,

Leverkusen, Germany, respectively), and cornea anesthesia (4 mg/mL oxybuprocaine

hydrochloride; Anestocil®, Laboratório Edol, Carnaxide, Portugal), pupils dilation (1%,

Tropicil®, Laboratório Edol, Carnaxide, Portugal) and corneal hydration (2% Methocel™,

Dávi II Farmacêutica S.A., Barcarena, Portugal) was kept during procedure.

The choroid was evaluated in the animals using SD-OCT at 52 W. The SD-OCT

system is able to capture 10,000 – 20,000 A scans per second with an axial resolution

of 2 µm and a transverse resolution of 4 µm. OCT was performed in both eyes of all

animals at 52 W, just before being euthanized.

The 830 nm SD-OCT Imagine System (Phoenix Micron IV, Phoenix Research

Labs, Pleasanton, CA, USA) [37] was placed closer to the eye such that an inverted

image was obtained and the deeper structures were placed closer to zero-delay [38].

Scans were obtained superior to the optic nerve head/optic disc (ONH), in both eyes of

all animals, in the area within 1 to 3 ONH diameter from the optic disc. For each location,

the device collected a set of 1024 raster scans along the scan length. Upon collection, a


150

set of five images was converted into a single image to reduce the noise observed on

individual images. The CT was measured using InSight image segmentation software

(v.1, Voxeleron LLC - Image analysis solutions, Chabot Drive, CA, USA). An average of

three independent scores obtained from 3 sets “five frames averaged” was used as the

CT value per eye per time-point.

Tissue preparation for cryosections

The animals were anesthetized, as described above, and intracardially perfused

with 0.1 M phosphate buffer saline (PBS, 137 mM NaCl, 2.7 mM KCl, 1.8 mM KH2PO4,

10 mM NaH2PO4, pH 7.4), followed by 4% paraformaldehyde (PFA) in 0.1 M PBS, pre-

warmed at 37ºC.

The enucleated eyes of GK (n = 8) and age-matched control Wistar Han (n = 5)

rats were post-fixed in 4% PFA for 1 h. The eyes were washed in successive solutions

of PBS and immersed in solutions containing 15% and 30% of sucrose in PBS, for 1 h

in each solution. They were embedded in a 1:1 30% sucrose and embedding resin

(Shandon™ Cryomatrix™, Thermo Fisher Scientific, Waltham, MA, USA) solution,

before freezing in dry ice. The samples were stored at -80ºC, until further use. Eye ball

sections 14 µm-thick from both right and left eyes were obtained using a cryostat (Leica

CM3050S, Nussloch, Germany), at -22ºC, and mounted on adhesive slides (Superfrost

Plus™, Thermo Fisher Scientific). A total of four sections (14 µm apart) were collected

per slide.

Immunofluorescence

Eye sections were labeled to assess the retinal and choroidal structure, following

a procedure described previously [39]. The cryosections were rehydrated twice in PBS

for 5 min, followed by blocking and permeabilization for 1 h in 10% goat serum and 0.5%

Triton X-100 in PBS. The sections were then incubated overnight with primary antibodies


151

(Supplementary Table 5.1) diluted in 0.5% Triton X-100, at 4ºC. After washing with PBS,

sections were incubated with corresponding secondary antibodies (Supplementary

Table 5.2) diluted in 0.5% Triton X-100, for 1 h. After washing, the sections were

incubated with 1:5,000 4’,6-diamidino-2-phenylindole (DAPI, Invitrogen™), and

coverslipped using mounting medium (Glycergel, Dako, Carpinteria, CA, USA).

Digital images were captured using an inverted fluorescence microscope (Axio

Observer.Z1, Zeiss, Carl Zeiss Meditec AG, Jena, Germany), using a Plan-Apochromat

20×/0.8 objective), and a laser scanning confocal inverted microscope (LSM 710 Axio

Observer, Zeiss), using a Plan-Apochromat 20×/0.8 objective.

Eight bit images were analyzed using the ImageJ software (version 1.48,

National Institutes of Health, USA) [40]. Iba1+ and MHC II+ cells were manually

counted in the choroid and retina. Data were expressed as number of cell/mm of

choroidal or retinal length, respectively. Rat endothelial cell antigen 1 (RECA-1)

and the proteoglycan NG2/Cspg4 (NG2) immunoreactivities were scored in the

choroid as mean fluorescence intensity per area selected (reference area

selected of 10,737.08 ± 6,306.11 µm2), while fluorescent NG2+ cells and RECA-

1 focal immunostaining were manually counted in the retina. VEGF and VEGFR2

immunoreactivities were quantified as mean fluorescence intensity/area for each

layer analyzed. All results were expressed as the mean count of 12 slices per eye

(right eyes only).

Direct labeling and visualization of choroidal vessels in

sclerochoroidal whole mounts

Intracardiac perfusion with PBS was performed in rats under anesthesia (i.p.;

ketamine:xylazine 160:30). DiI (Cat. #D-282, Invitrogen/Molecular Probes, Carlsbad, CA,

USA) 0.120 mg/mL in 1% glucose in PBS was applied via cardiac perfusion, following


152

perfusion with 4% PFA. The eyes were enucleated, and choroidoscleral whole mounts

were prepared. The whole mounts were fixed in 4% PFA for 15 min, washed with PBS,

and blocked with 10% goat serum in 0.3 % Tween in PBS for 1 h. Samples were

incubated with the primary antibodies diluted in 3% goat serum in PBS for 3 days at 4ºC

(Supplementary Table 5.1). After washing overnight with PBS, whole mounts were

incubated with secondary antibodies (Supplementary Table 5.2) overnight at 4ºC. After

washing, samples were flat mounted onto glass slides using mounting medium.

The images were obtained by laser scanning confocal microscope LSM 710

(Zeiss), using Zeiss EC Plan-Neofluor 40x oil objective lens, NA 1.3. A series of z-stacks

were captured from the retinal pigment epithelium (RPE) outer surface, to the outer

choroid. Each z-stack consisted of a depth of optical sections, 3 µm apart, along the z-

axis. Iba1+ and MHC II+ cells were classified as ramified or round cells and were manually

counted in the choroid in all in-depth planes of the slide (Supplementary Figure 5.1).

The choroidal vascular density (defined as the percentage of total area covered

by choriocapillaris vessels) [41] was determined from independent z-stacks collected at

≤ 10 µm (choriocapillaris) and > 10 µm outwards from the RPE plane (medium and large

vessels) in a selected area (213 × 213 µm, Supplementary Figure 5.2). Results were

expressed as the mean of 14 counts per eye.


The statistical analysis was performed using SPSS (Version 25.0, IBM Corp.,

Armonk, NY, USA). The Shapiro-Wilk test was used to assess the normality of data (p >

0.05). The normally distributed data were evaluated concerning the homogeneity of

variance, using the Levene’s test (p > 0.05). The independent Student’s t-test was used

to compare the means between two experimental groups for the same variable.


153

Statistical significance was defined as p < 0.05. Values were presented as mean +

standard error of the mean (SEM).

5.4. Results

Diabetic animals exhibit decreased body weight and

hyperglycemia

A total of 31 animals (52 W) were enrolled in the study (18 GK and 13 age-matched

control Wistar Han rats). The weight of GK rats was significantly lower than that of age-

matched control rats (416.31 ± 7.0 g and 462.5 ± 9.24 g, respectively, p < 0.001) and

glycemia was significantly higher (227.75 ± 12.1 mg/dL versus 107.0 ± 0.94 mg/dL, p <

0.001, Supplementary Figure 5.3).

Increased choroidal thickness in GK rats

CT was assessed in both eyes of all GK and age-matched control Wistar Han

rats at 52 W (n = 62 eyes), using OCT. CT was higher in GK rats (58.40 ± 1.15 µm versus

50.90 ± 1.58 µm, p < 0.001, Figures 5.1A and B).


154

Figure 5.1. Choroidal thickness (CT) of GK and age-matched control Wistar Han rats (52 W). (A)

Images resulting from five frames averaged, collected as inverted images using OCT raster scans

obtained in 1024 continuous points, by approaching the device to the zero delay line. Horizontal

raster scan line encompasses an area within 1 to 3 disk diameters from the optic nerve head.

Choroidal layer obtained by automatic segmentation was manually corrected. A mean of three

independent scores obtained from 3 different located sets “five frames averaged” were used as

the CT value per eye per time-point. (B) CT values from all eyes of GK (n = 36) and age-matched

control (n = 26) rats. Data are expressed as mean ± SEM. Scale bar: 100 µm. Significance: ***p

< 0.001.

GCL = ganglion cell layer, INL = inner nuclear layer, ONL = outer nuclear layer, RPE = retinal

pigment epithelium.

B

Control GK

0

20

40

60

80

0

Choro

idal th

ickness (

µm

) ***

INLGCL

ONL

RPEChoroid

A

Control GK


155

Choroidal blood vascular density is reduced in the inner choroid of

GK rats

Eighteen animals (10 GK and 8 age-matched control Wistar Han rats) were assigned

to perfusion with DiI and sclerochoroidal whole mounts were prepared and assessed by

confocal microscopy.

The choroidal vascular density in the innermost choroid/choriocapillaris (≤ 10 µm)

was significantly decreased in GK rats (p = 0.045, Figures 5.2A and C and

Supplementary Table 5.3). Choroidal Iba1+ and MHC II+ cell number was not significantly

different between GK and age-matched control rats (Figures 5.2B and D and

Supplementary Table 5.4).

Interestingly, Iba1+ cells and MHC II+ cells topographic disposition in the choroid

spare the innermost choroid, where they are notably rare or absent, being preferentially

present in the middle and outer choroidal stroma, either in GK or age-matched control

rats (Supplementary videos 5.1 - 5.6).


156

A

Control GK

DiI

B

Control GK

Iba1

MH

C II

*


157

Figure 5.2. Vascular and cell profiles of the choroid of GK and age-matched control Wistar Han

rats (52 W). (A) Representative images of vascular density in the inner choroid (≤ 10 µm). (B)

Choroidal Iba1+ and MHC II+ cells. Ramified cells (green arrows) and round cells (red arrows)

were highlighted. (C) Quantification of the choroidal vascular density in the inner and outer choroid

using the ‘image>adjust>threshold’ window tool of ImageJ to obtain the percentage of vascular

coverage, obtained from z-stacks collected at ≤ 10 µm or > 10 µm from the outer RPE plane,

respectively. (D) Iba1+ and MHC II+ cell number in the choroid in all in-depth z-stacks. Images

were collected with Zeiss EC Plan-Neofluor 40x oil objective lens, NA 1.3. Quantitative analyses

were performed based on 14 independent counts per eye in each and all in-depth z-stacks per

specimen. Data are expressed as mean ± SEM (n = 8, control group; n = 10, GK group). Scale

bar: 50 µm. Significance: *p < 0.05, **p < 0.01.

C

Control GK

0

20

40

60

80

100

inner choroid outer choroid

Blo

od v

ascula

r density in t

he

choro

id

*

D

Control GK

0

10

20

30

40

ramified round ramified round

Iba1-positive cells MHC II-positive cells

Num

ber

of

cells

in t

he c

horo

id

Iba1+ cells MHC II+ cells

**


158

Increased immunoreactivity of microglial cell markers in the outer

retina of GK rats

Iba1+ cells increased in the retina of GK rats, being statistically significantly

increased in the outer plexiform layer (OPL, t(11.000) = -3.872, p = 0.003). Iba1+ (p =

0.195) and MHC II+ cells density was not statistically significantly different in the choroid

of GK rats (t(11.000) = 2.190, p = 0.051, Figure 5.3 and Supplementary Table 5.5).

Interestingly, Iba1+ cells were distributed throughout the retina of GK rats paralleling

the topographic distribution of the 3 vascular plexuses of the retina. Iba1+ cell extensions

from the inner plexiform layer (IPL) to the OPL resemble the communications between

the deep and middle vascular plexuses of the retina, suggesting that migrating glial cells

towards the outer retina may use the communicating inter-plexuses retinal capillaries as

a scaffold (Figure 5.3A, bottom left panel). Communications between plexus were

evidenced.


159

Figure 5.3. Microglial cells in the retina and choroid of GK and age-matched control Wistar Han

rats (52 W). (A) Representative eye cross-sections immunolabelled against Iba1 (left panels),

MHC-II (middle panels) and merge (right panels). Iba1+ cells are located in the superficial and

plexiform layers of the retina, mainly. Iba1+ cells located in the OPL of the GK cohort only (green

arrows). In GK rats, Iba1+ cells migrate from the IPL to the OPL, crossing the INL (red arrow). (B)

Quantification of Iba1+ and MHC II+ cell density of GK (n = 8) and age-matched control (n = 5) rats

based on 12 independent specimen counts per eye. Counting was done for the right eye only, in

all animals. Data are expressed as mean ± SEM. Scale bar: 100 µm. Significance: **p < 0.01.

GCL = ganglion cell layer, IPL = inner plexiform layer, INL = inner nuclear layer, OPL = outer

plexiform layer, ONL = outer nuclear layer, RPE = retinal pigment epithelium.

B

Control GK

0

10

20

30

40

50

INL IPL OPL retina choroid choroid

Iba1 MCH II

Po

sitiv

e

ce

lls/m

m

**

A

Co

ntr

ol

GK

Iba1 MHC II

GCL

INL

Choroid

ONL

RPE

IPL

OPL

Iba1 MHC II DAPI

GCL

INL

Choroid

ONL

RPE

IPL

OPL


160

Pericytes are rare in the innermost choroid of GK rats

There were no statistically significant differences in the immunoreactivity of NG2

and RECA-1 between GK and age-matched control Wistar Han rats, whatever the

location considered (Figure 5.4 and Supplementary Table 5.6). Nevertheless, there was

a trend to increased NG2 immunostaining in the choroid of GK rats. Combined

immunoreactivity of RECA-1 and NG2 evidenced a rarefaction of pericytes in the

innermost choroid of GK rats. RECA-1 co-localized with the choriocapillaris layer, just

underneath and in close continuity with the RPE cell plane (Figure 5.4A, left panels).

NG2 immunoreactivity was absent in some areas of the choriocapillaris plane, leaving

fluorescent-free gaps between the RPE and the inner choroid, drawing a typical jagged

pattern, corresponding to the absence of pericytes/mural cells in the innermost choroid

(Figure 5.4A, middle panels, blue arrows). This ‘polarity’ in choroidal vascular regulatory

cells disposition evidenced by NG2, corresponding to a rarefaction of pericytes in the

innermost choroid, was more enhanced in GK rats.

The three retinal plexuses (superficial, in the retinal nerve fiber and ganglion cell

(GCL) layers; middle, in the IPL and deep, in the OPL) were visualized by RECA-1

immunolabelling (Figure 5.4A, top panels, left and right).


161

Figure 5.4. Localization of mural and endothelial cells in the retina and choroid of GK and age-

matched control Wistar Han rats (52 W). (A) Representative eye cross-sections immunolabelled

against RECA-1(left panels), NG2 (middle panels) and merge (right panels). The 3 plexuses of

the retina are evidenced by RECA-1 immunostaining (white arrows). Communications between

(i) the superficial and middle plexuses of the retina (red arrow), (ii) the middle and deep plexuses

and (iii) the deep and superficial plexuses (green arrow), are visible. RECA-1 immunoreactivity

relates to the presence of endothelial cells and fluorescence of the choriocapillaris endothelial

cells is continuous with the RPE cell plane (left panels). Conversely, NG2 immunostaining of

pericyte/mural cells leaves focal gaps between the RPE and the inner choroid, drawing a jagged

pattern, more pronounced in GK rats (blue arrows). (B) Quantification of RECA-1 and NG2

immunoreactivity in the retina and choroid of GK (n = 8) and age-matched control (n = 5) rats.

RECA-1 and the proteoglycan NG2/Cspg4 (NG2) immunoreactivities were scored as

fluorescence intensity per area selected in the choroid (reference area selected of 10,737.08 ±

6,306.11 µm2), while NG2+ cells and RECA-1 focal immunostaining were manually counted in the

retina. Counting was done for the right eye only, in 12 independent specimen counts per eye.

Data are expressed as mean ± SEM. Scale bar: 100 µm.


plexiform layer, ONL = outer nuclear layer, RPE = retinal pigment epithelium.

B

Control GK

0

5

10

15

20

GCL OPL retina

RECA-1 NG2P

ositiv

e

ce

lls in

th

e r

etin

a/m

m

0

200

400

600

800

RECA-1 NG2

Flu

ore

sce

nce

in

ten

sity in

th

e

ch

oro

id/a

rea


162

VEGR2 immunoreactivity was significantly increased in GK rats

VEGF immunoreactivity was more intense in the level of the OLM, RPE and

choroid, while VEGFR2 immunoreactivity was negligible at those locations. VEGFR2

immunoreactivity was more intense in the inner retina. Higher VEGFR2 immunoreactivity

was observed in GK rats, significantly in the IPL, INL, OPL, ONL and OLM, when

comparing with age-matched control rats (Figure 5.5 and Supplementary Table 5.7).

A

Co

ntr

ol

GK

VEGF VEGFR2

GCL

INL

Choroid

ONL

RPE

IPL

OPL

VEGF VEGFR2 DAPI

GCL

INL

Choroid

ONL

RPE

IPL

OPL

GCL

INL

Choroid

ONL

RPE

IPL

OPL

OLM

OLM


163

Figure 5.5. Immunoreactivity of VEGF and VEGFR2 of GK and age-matched control Wistar Han

rats (52 W). (A) Representative eye cross-sections immunolabelled against VEGF (left panels),

VEGFR2 (middle panels) and merge (right panels). VEGF immunoreactivity spreads throughout

the retina, increasing in the OPL, OLM, RPE and choroid. VEGFR2 immunoreactivity is higher in

the innermost retina (retinal nerve fiber layer) and very low or absent in the RPE and choroid.

VEGFR2 immunoreactivity is still visible as a faint coloration in the retinal layers other than the

retinal nerve fiber layer of GK rats only (white arrows). (B) Quantification of the VEGF (top) and

VGFR2 (bottom) immunoreactivities in the retina and choroid of GK rats (n = 8) and age-matched

controls (n = 5) based on 12 independent specimen counts per eye. VEGF and VEGFR2

immunoreactivities were quantified as fluorescence intensity/area per layer. Counting was done

for the right eyes only, in all animals. Data are expressed as mean ± SEM. Scale bar: 100 µm.

Significance: *p < 0.05.


plexiform layer, ONL = outer nuclear layer, OLM = outer limiting membrane, RPE = retinal pigment

epithelium.


164

5.5. Discussion

CT has been searched as a surrogate of DR, DME, choroidal flux and diabetic

choroidopathy in T2D patients, but contradictory results have been reported [25-29].

Although GK rats are not blueprints for the diseased T2D humans, they provide

excellent insights into the pathogenesis of T2D [42]. Furthermore, rats are recognized

as the preeminent model for studying the choroid [43, 44]. We found that CT was

significantly increased in GK rats when comparing with age-matched control Wistar Han

rats. To our knowledge, this is the first time that the CT is assessed in a rat model of T2D

[45].

The vascular density was reduced in the inner choroid of GK rats, corresponding

to the choriocapillaris. This observation is in accordance with choriocapillaris

degeneration observed in T2 humans, either by post-mortem specimens or by OCT-

angiography (OCTA) [46-48]. The findings of no significant difference in the outer choroid

vascular density combined with a significantly higher CT in GK rats, enhances the role

of the outer choroid layer, the suprachoroid, as an important contributor to total CT and

may explain why available data on human CT are so contradictory [27-29, 49, 50]. The

suprachoroid is the choroidal layer most prone to change, but it is the most difficult to

evaluate correctly, mainly when not using high definition devices as Swept Source OCT

and the high resolution mode EDI of the SD-Spectralis OCT [25, 30]. Unfortunately,

devices based on red blood cells’ movement as the OCTA are not expected to help much

in this issue either [51].

The normal distribution of inflammatory cells in the choroidal stroma, sparing the

innermost choroid, suggests that under inflammatory stress a dramatic increase in the

number of inflammatory cells may result in packing of cells at this area, leading to

disturbances of the PRs/RPE/Bruch´s membrane/choriocapillaris tapetoretinal unit


165

(Supplementary videos 5.1-5.3) [52]. Nevertheless, we did not find inflammatory cells to

be significantly increased in the choroid of GK rats. It is possible that this finding

correlates with the low level of diabetes in this animal model. If that was to be so, one

might expect to find increased inflammatory cells in the choroid of the more aggressive

diabetic animal models such as the STZ-induced Type 1. The number of inflammatory

cells were increased in the retina, as previously described, mainly in the OPL [18, 33].

Interestingly, Iba1+ cells disposition parallels the three retinal plexuses’, and the

migration of cells from the inner to the outer retina resembles the capillary

communications between plexuses, suggesting that vessels are a scaffold for glial cell

migration.

The VEGF levels were not significantly increased in the retina, RPE or choroid, in

GK rats. VEGFR2 expression was negligible in the RPE or choroid in GK and age-

matched control Wistar Han rats. Conversely, VEGFR2 immunoreactivity peaked in the

retinal nerve fiber layer of both GK and control rats, and it was increased throughout the

retina of GK rats. The precise cellular background for this increased expression is not

clear, but the combination of maximum expression in the innermost retina and

expression throughout the retina resembles the distribution of Müller cells and neurons

within the retina. Nevertheless, a previous report in an animal model of T1D showed

VEGFR2 to be expressed mainly in the capillaries of the retina and, in opposition to our

findings, in the choriocapillaris [53]. Conversely, other data correlated VEGFR2 more

strikingly with neurons and Müller cells rather than with endothelial cells [54].

Interestingly, the later correlated VEGFR2 expression by neurons and Müller cells

combined with the action of MHC II cells, to be related with capillary vertical sprouting

and the formation of the deep retinal plexus, an example of neurovascular crosstalk or

coupling. The relationship between increased VEGFR2 and normal VEGF

immunoreactivity is not clear, if any at all. If that results from a milder diabetes status in


166

our model, VEGFR2 would be among the first molecular biomarkers to change in DR,

rather than VEGF.

Disposition of pericytes in the choriocapillaris was previously described in mice to

be at its scleral surface only (polarized distribution) and focal pericyte depletion has been

related to vascular remodeling [55]. We found distribution of NG2+ cells in the innermost

choroid of rats to draw a jagged pattern, when comparing with RECA-1 endothelial

immunostaining. These focal ‘gaps’, suggesting the existing of cell-free spaces in the

innermost choroid, may be related to vascular remodeling of the choriocapillaris and not

necessarily to permanent alterations, and were more pronounced in T2D. Interestingly,

this disposition is shared by stromal Iba1+ cells in the same location (Supplementary

videos 5.1-5.6).

The finding of indirect signs of vascular remodeling in the choriocapillaris suggests

that some alterations reported by OCTA-based studies may be only transient. OCTA is

able to examine nonperfusion but not detailed morphological vascular patterns, since the

choriocapillaris being as thin as 10-20 µm is not resolved by a system that has a

maximum lateral resolution within that range [56].

Perivascular mural cells in the middle and outer choroid were previously reported

to show contractile properties that were related to vasoregulatory functions [55]. We did

not find a significant difference in immunostaining intensity for NG2 between GK and

age-matched control Wistar Han rats. Perhaps, a more severe model of T2D and

individual cell counting using combined immunomarkers of perivascular mural cells in a

3D imaging system may eventually reveal such differences in future studies [55].

In conclusion, we found an increase in CT, a decrease in the inner choroidal

vascular density, increased Iba1+ cells density in the outer retina, and increased

VEGFR2 immunoreactivity in the retina, in GK rats. Disposition of pericytes and Iba1+


167

cells in the choroidal stroma spare the innermost choroid, either in GK or age-matched

control Wistar Han rats.

5.6. References




4. Gross, J.G., et al., Five-Year Outcomes of Panretinal Photocoagulation vs Intravitreous Ranibizumab for Proliferative Diabetic Retinopathy: A Randomized Clinical Trial. JAMA Ophthalmol, 2018. 136(10): p. 1138-1148.


6. Baharivand, N., et al., Relationship between vitreous and serum vascular endothelial growth factor levels, control of diabetes and microalbuminuria in proliferative diabetic retinopathy. Clin Ophthalmol, 2012. 6: p. 185-91.

7. Shimada, H., et al., Concentration gradient of vascular endothelial growth factor in the vitreous of eyes with diabetic macular edema. Invest Ophthalmol Vis Sci, 2009. 50(6): p. 2953-5.




11. Funatsu, H., et al., Vitreous levels of interleukin-6 and vascular endothelial growth factor are related to diabetic macular edema. Ophthalmology, 2003. 110(9): p. 1690-6.




168









22. Lutty, G.A., Effects of diabetes on the eye. Invest Ophthalmol Vis Sci, 2013. 54(14): p. ORSF81-7.

23. Huang, D., et al., Optical coherence tomography. Science, 1991. 254(5035): p. 1178-81.

24. Margolis, R. and R.F. Spaide, A pilot study of enhanced depth imaging optical coherence tomography of the choroid in normal eyes. Am J Ophthalmol, 2009. 147(5): p. 811-5.








169

31. Omri, S., et al., PKCzeta mediates breakdown of outer blood-retinal barriers in diabetic retinopathy. PLoS One, 2013. 8(11): p. e81600.



34. Carmo, A., et al., Nitric oxide synthase activity in retinas from non-insulin-dependent diabetic Goto-Kakizaki rats: correlation with blood-retinal barrier permeability. Nitric Oxide, 2000. 4(6): p. 590-6.

35. Moore, D.J. and G.M. Clover, The effect of age on the macromolecular permeability of human Bruch's membrane. Invest Ophthalmol Vis Sci, 2001. 42(12): p. 2970-5.

36. Sorensen, N.B., et al., Bruch's membrane allows unhindered passage of up to 2mum latex beads in an in vivo porcine model. Exp Eye Res, 2019. 180: p. 1-7.



39. Fernandez-Sanchez, L., et al., Loss of outer retinal neurons and circuitry alterations in the DBA/2J mouse. Invest Ophthalmol Vis Sci, 2014. 55(9): p. 6059-72.

40. Schneider, C.A., W.S. Rasband, and K.W. Eliceiri, NIH Image to ImageJ: 25 years of image analysis. Nature Methods, 2012. 9(7): p. 671-675.

41. Kumar, A., et al., Vascular associations and dynamic process motility in perivascular myeloid cells of the mouse choroid: implications for function and senescent change. Invest Ophthalmol Vis Sci, 2014. 55(3): p. 1787-96.

42. Portha, B., et al., The GK rat: a prototype for the study of non-overweight type 2 diabetes. Methods Mol Biol, 2012. 933: p. 125-59.


44. Castro-Correia, J., Understanding the choroid. Int Ophthalmol, 1995. 19(3): p. 135-47.

45. Wild, S., et al., Global prevalence of diabetes: estimates for the year 2000 and projections for 2030. Diabetes Care, 2004. 27(5): p. 1047-53.


47. Fryczkowski, A.W., B.L. Hodes, and J. Walker, Diabetic choroidal and iris vasculature scanning electron microscopy findings. Int Ophthalmol, 1989. 13(4): p. 269-79.



170

49. Abalem, M.F., et al., The effect of glycaemia on choroidal thickness in different stages of diabetic retinopathy. Ophthalmic Res, 2020.

50. Wang, H. and Y. Tao, Choroidal structural changes correlate with severity of diabetic retinopathy in diabetes mellitus. BMC Ophthalmol, 2019. 19(1): p. 186.


52. Ma, W., et al., Microglia in the mouse retina alter the structure and function of retinal pigmented epithelial cells: a potential cellular interaction relevant to AMD. PLoS One, 2009. 4(11): p. e7945.

53. Sun, D., et al., Molecular imaging reveals elevated VEGFR-2 expression in retinal capillaries in diabetes: a novel biomarker for early diagnosis. FASEB J, 2014. 28(9): p. 3942-51.

54. Okabe, K., et al., Neurons limit angiogenesis by titrating VEGF in retina. Cell, 2014. 159(3): p. 584-96.




171


Figure 5.1. Iba1+ and MHC II+ cell count by ImageJ. (A) Iba1+ cells were quantified in

all-dept planes. (B) For each plane, MHC II+ cells were marked with yellow dots to avoid

duplication.

Figure 5.2. Quantification of the choroidal vascular density in the inner and outer choroid

(defined as the percentage of total area covered by choriocapillaris vessels) using the

‘image>adjust>threshold’ window tool of ImageJ to obtain the percentage of vascular

1 32A

B

BA


172

coverage. (A) z-stacks collected at ≤ 10 µm (choriocapillaris). (B) z-stacks collected at >

10 µm from the outer RPE plane (medium and large vessels).

Figure 5.3. Body weight and glycemia in GK and in age-matched control Wistar Han rats

(52 W).

0

100

200

300

400

500

600

1 2

Body w

eig

ht

(g)

0

50

100

150

200

250

300

1 2

Gly

cem

ia (

mg/d

l)***

Control GK

0

50

100

150

200

250

300

0

50

100

150

200

250

300

1 2

Gly

cem

ia (

mg/d

l)

*** Gly

cem

ia (m

g/d

l)


173

Table 5.1. Primary antibodies

Antigen Target Host Supplier Cat. N.er Dilution

Iba1 Microglia Rabbit Wako Chemicals Inc.,North Chesterfield, VA, USA 019-19741 1:250

MHC II Activated microglia Mouse Bio-RAD Laboratories, Hercules, CA, USA MCA46R 1:200

NG2 Cell membrane chondroitin sulfate proteoglycan Rabbit Merck, Darmstadt, Germany AB530 1:200

RECA-1 Endothelial cells Mouse Abcam Inc., Cambridge, MA, USA ab9774 1:200

VEGF Signal growth factor protein Mouse Abcam Inc., Cambridge, MA, USA ab1316 1:200

VEGFR2 VEGF receptor 2 Rabbit Abcam Inc., Cambridge, MA, USA ab131241 1:200

Abbreviations: Iba1, calcium binding adapter molecule 1; MHC II, major histocompatibility complex II; RECA-1, rat endothelial cell antigen; NG2, neuroglial antigen 2; VEGF,

vascular endothelial growth factor; VEGFR-2, vascular endothelial growth factor 2.

Table 5.2. Secondary antibodies

Fluorophore Target Host Supplier Cat. N.er Dilution

Alexa Fluor® 488 Mouse Ig G Goat IntrovitrogenTM, Thermo Fisher Scientific, Waltham, MA, USA A-11001 1:500

Alexa Fluor® 568 Rabbit Ig G Goat IntrovitrogenTM, Thermo Fisher Scientific, Waltham, MA, USA A-110036 1:500

Abbreviations: Ig G, immunoglobulin G.


174

Table 5.3. Choroidal vascular density

Vessel area/total area

Control GK p

Inner Choroid 90.95 ± 0.65% 87.04 ± 1.65% 0.045*

Middle-outer choroid 67.75 ± 1.83% 63.57 ± 1.53% 0.121

Quantification of the mean choroidal vascular density in the inner and outer choroid (defined

as the percentage of total area covered by choriocapillaris vessels) using the

‘image>adjust>threshold’ window tool of ImageJ to obtain the percentage of vascular coverage.

Inner choroid: z-stacks collected at ≤ 10 µm (choriocapillaris). Outer choroid: z-stacks collected

at > 10 µm from the outer RPE plane (medium and large vessels). Quantitative analyses were

performed based on 14 independent counts per eye in each and all in-depth z-stacks per

specimen. Data are expressed as mean ± SEM (n = 8, control group; n = 10, GK group).

Significance: *p < 0.05.

Table 5.4. Quantification of Iba1+ and MHC II+ cells in whole mounts of the

choroid

Marker Cell shape Positive cells

Control GK p

Iba1 ramified 31.61 ± 1.42 31.87 ± 1.52 0.903

round 2.73 ± 0.30 1.00 0.085

MHC II ramified 2.96 ± 0.55 0 0.003**

round 26.60 ± 2.30 26.69 ± 1.39 0.973

Mean Iba1+ and MHC II+ in the choroid of control and GK rats in all in-depth z-stacks of the

choroid. Iba1+ and MHC II+ cells were manually counted in the choroid. Quantitative analyses

were performed based on 14 independent counts per eye in each and all in-depth z-stacks per

specimen. Images were collected by confocal microscopy with Zeiss EC Plan-Neofluor 40x oil

objective lens, NA 1.3. Data are expressed as mean ± SEM (n = 8, control group; n = 10, GK

group). Significance: **p < 0.01.


175

Table 5.5. Quantification of Iba1+ and MHC II+ cells in cryosections of the

retina and choroid.

Iba1+ cells/mm MHC II+ cells/mm

Control GK p Control GK p

Choroid 38.44 ± 2.57 32.70 ± 2.85 0.195 25.33 ± 2.06 19.77 ± 1.54 0.051

Retina 34.62 ± 1.97 37.91 ± 4.54 0.523 0 0

INL 0.53 ± 0.17 1.37 ± 0.30 0.065 0 0

IPL 14.33 ± 1.23 10.30 ± 2.07 0.124 0 0

OPL 3.48 ± 0.70 11.41 ± 1.54 0.003** 0 0

Quantification of mean Iba1+ and MHC II+ cells in GK rats (n = 8) and age-matched controls (n

= 5) based on 12 independent specimen counts per eye. Counting was made as the number

of cell/mm of choroidal or retinal length, respectively. Counting was done for the right eye only,

in all animals. Data are expressed as mean ± SEM. Significance: **p < 0.01.

INL = inner nuclear layer, IPL = inner plexiform layer, OPL = outer plexiform layer.


176

Table 5.6. NG2 and RECA-1 immunoreactivity in cryosections of the retina

and choroid

NG2 RECA-1


Choroid 473.66 ± 52.17 558.01 ± 42.78 0.241 675.81 ± 54.97 658.71 ± 30.94 0.774

Retina 8.86 ± 0.34 8.98 ± 0.43 0.845 0 0

GCL 0 0 9.94 ± 0.36 8.88 ± 0.45 0.130

OPL 0 0 16.76 ± 0.92 14.68 ± 0.89 0.151

Quantification of mean RECA-1 and NG2 immunoreactivities in the retina and choroid in GK

rats (n = 8) and age-matched controls (n = 5). RECA-1 and the NG2 immunoreactivity was

scored as mean fluorescence intensity per area selected in the choroid (reference area

selected of 10,737.08 ± 6,306.11 µm2), while NG2+ cells and RECA-1 focal immunostaining

were manually counted in the retina. Counting was done for the right eye only, in 12

independent specimen counts per eye. Data are expressed as mean ± SEM. Significance: p <

0.05. GCL = ganglion cell layer, OPL = outer plexiform layer.


177

Table 5.7. VEGF and VEGFR2 immunoreactivity in cryosections of the

retina and choroid

VEGF VEGFR2


GCL+RNFL 766.67 ± 33.82

827.54 ± 61.08

0.403 323.15 ± 17.16

418.33 ± 52.84

0.123

IPL 769.40 ± 74.70

809.86 ± 96.18

0.746 157.00 ± 16.53

267.75 ± 36.37

0.021*

INL 691.20 ± 77.01

714.88 ± 78.89

0.844 145.80 ± 16.37

264.38 ± 41.34

0.026*

OPL 828.12 ± 83.86

875.01 ± 61.45

0.655 207.43 ± 10.23

290.69 ± 35.04

0.050*

ONL 730.40 ± 102.38

641.00 ± 57.42

0.425 136.60 ± 12.64

242.88 ± 37.55

0.026*

OLM 1546.80 ± 231.61

1142.38 ± 121.76

0.116 184.80 ± 19.63

320.25 ± 51.54

0.037*

RPE 1168.45 ± 60.60

1284.52 ± 88.12

0.364 118.67 ± 5.05

132.59 ± 9.70

0.232

Choroid 771.66 ± 32.21

867.25 ± 61.69

0.200 95.42 ± 3.57

110.81 ± 8.34

0.123

Quantification of the mean VEGF and VGFR2 immunoreactivities in the retina and choroid in GK

rats (n = 8) and age-matched controls (n = 5) based on 12 independent specimen counts per eye.

VEGF and VEGFR2 immunoreactivities were quantified as mean fluorescence intensity/area per

layer. Counting was done for the right eye only, in all animals. Data are expressed as mean ±

SEM. Significant values: *p < 0.05.

RPE = retinal pigment epithelium, GCL+RNFL = ganglion cell and nerve fiber layers, IPL = inner

plexiform layer, INL = inner nuclear layer, OPL = outer plexiform layer, ONL = outer nuclear layer,

OLM = outer limiting membrane (and photoreceptor inner segments); VEGF = vascular

endothelial growth factor, VEGFR2 = VEGF receptor 2.


178


innermost choroid (red) outwards the RPE cell plane of a 52-week-old GK rat. Slice

thickness: 27.2 μm. AC and JM authored the video: 11’’; 40,221 KB.

Video 5.2. Sequenced images in the same location as in video 5.1, showing the

localization of MHC class II+ cells (cyan) outwards the choriocapillaris perfused by DiI

(red) in a 52-week-old GK rat. Slice thickness: 23.8 μm. AC and JM authored the video:

10’’; 25,306 KB.


the co-localization of Iba1+ cells (green) and MHC II+ cells (cyan) outwards the

choriocapillaris perfused by DiI (red) in a 52-week-old GK rat. Slice thickness: 30.6 μm.

AC and JM authored the video: 17’’; 30,465 KB.


the choriocapillaris (red) in a 52-week-old control Wistar Han rat. Slice thickness: 48 μm.

AC and JM authored the video: 16’’; 51,062 KB.

Video 5.5. Sequenced images showing the localization of MHC II+ cells (cyan) outwards

the inner choroidal vascular network perfused by DiI (red) in a 52-week-old control Wistar

Han rat. Slice thickness: 45 μm. AC and JM authored the video: 16’’; 49,140 KB.


the co-localization of Iba1+ cells (green) and MHC II+ cells (cyan) outwards the

choriocapillaris perfused by DiI (red) in a 52-week-old control Wistar Han rat. Slice

thickness: 45 μm. AC and JM authored the video: 16’’; 63,560 KB.


https://drive.google.com/drive/folders/1YWm9HQ8ijOKu0XWj6bytF7TpG7Lp_FDo?usp

=sharing

Choroid and diabetes Discussion and future perspectives

179

6. Discussion and future perspectives


180

Progression of microvascular complications in diabetes depends on diabetes

type and metabolic balance [1-3]. Type 1 diabetes (T1D) accounts for less than 10% of

all cases of diabetes, but affects younger people and when retinopathy develops,

progression is faster and more severe than in Type 2 (T2D) [4-6]. Diabetic retinopathy

(DR) was for long considered to be a pure microvascular disease, but increasing

evidence points to inflammation as a key factor in the pathogenesis of DR [7-9].

Inflammatory events develop both in the retina and choroid during the course of

disease [10, 11]. The choroid is essential for the nutrition and water clearance of the

outer retina [12]. These facts lead investigators to focus back on earlier evidence of

diabetic choroidopathy [13, 14] and on the role of the outer blood retinal barrier in DR

[15].

Several attempts have been made to establish a link between alterations in the

choroidal thickness (CT), assessed by optical coherence tomography (OCT), and the

progression of DR in humans. However, CT as a surrogate for choroidal blood flow,

metabolic status, DR, diabetic macular edema (DME), choroidal inflammation or diabetic

choroidopathy, in studies with human diabetic subjects, failed to be reliable [16-21]. We

found that most contradictions are related to bias in collecting and treating data, but other

reasons are linked to CT inter- and intra-individual variability and to the difficulty in

evaluating the suprachoroid [12, 16]. The suprachoroid is the choroidal layer most prone

to change but its posterior location next to the sclera and its non-vascular nature makes

the evaluation difficult when not using high resolution modes in SD-OCT or swept-source

OCT [16, 18]. Even though after eliminating most of the reasons that might lead to bias,

the CT did not prove to have prognostic value in DME, even though it decreased with

anti-VEGF agents’ administration [3, 22, 23].

Conversely, we found that good metabolic control (Hb A1c ≤ 7%) and intact

ellipsoid zone were markers of good functional outcome in DME, while subfoveal


181

neuroretinal detachment, also known as subretinal fluid, was a marker of anatomic

outcome only.

Optical coherence tomography angiography (OCTA) does not detect structures

with no significant erythrocyte movement as the suprachoroid, therefore it does not add

much to OCT high resolution modes or swept source OCT in the evaluation of CT. OCTA

has been invaluable in the evaluation of the choriocapillaris in human subjects [24].

Alterations in the diabetic choriocapillaris were reported in OCTA studies, although it is

not clear if they are permanent or if they are transient and reversible, mirroring increased

choriocapillaris remodeling in diabetes reported in post-mortem studies [25-28]. Being

able to resolve the microvascular network of the retina and nonperfusion areas at the

choriocapillaris, OCTA does not evaluate hitherto detailed morphological vascular

patterns of the later, since the choriocapillaris being as thin as 10-20 µm is not resolved

by a system that has a maximum lateral resolution within that range (~15-20 μm) [29].

Rats are widely used models for studying human diseases and their choroidal

structure is closest to human’s than rabbits’, cats’ or guinea pigs’ [30, 31]. Recently, a

novel imaging platform combining photoacoustic microscopy and SD-OCT was

suggested to be valuable in evaluating the CT in rabbits [32]. We evaluated for the first

time in vivo CT by OCT in T1D (streptozotocin-induced diabetes) and T2D (Goto-

Kakizaki, GK) rats. No significant differences in CT were observed between control and

16 weeks old T1D rats, but one year old T2D rats had larger CT.

One may hypothesize that the shorter duration of diabetes (8 weeks) in T1D

accounts for the absence of differences or, conversely, that differences in T2D rats

depend on the model and not on T2D disease. Future perspectives to shed light on these

issues may follow two complementary pathways: in vivo CT evaluation in T1D rats with

longer disease duration, under metabolic control with insulin, and in vivo CT evaluation

in younger T2D GK rats (16 weeks old), to compare with T1D rats data.


182

Vascular remodeling of the choriocapillaris has been related to the depletion of

pericytes [33]. We found a reduced vascular density at the inner choroid of T2D rats,

corresponding to the choriocapillaris. Similar findings were reported in human T2D

individuals, in vivo and post-mortem [24, 28, 34]. The findings of a significantly higher

CT in T2D rats, enhances the role of the suprachoroid and may explain why available

data on human in vivo CT are so contradictory [17, 19, 21, 26, 35].

Glial cells are associated with pericyte removal from the perivascular wall and to

increased exposure of the vascular endothelia to vascular endothelial growth factor

(VEGF) [10, 36]. Increased number and activity of inflammatory cells in the retina

displace pericytes from the endothelium and increase endothelial leakage via increased

VEGF secretion and tight junction proteins downregulation and disorganization [10, 37].

It is possible that the vascular remodeling at the choriocapillaris follows the same

direction: increased glial cell activity, increased focal pericyte depletion, increased

endothelial exposure to VEGF, leading not to leakage as in the retinal endothelium, but

to remodeling [36]. We found pericytes’ disposition in the choroid sparing the innermost

choroid, which has been previously named as ‘polarized disposition’ [38]. When

comparing the coriocapillaris immunostaining (RECA-1) with the pericyte

immunostaining (NG2), focal gaps were observed at the inner choroid of control, T1D

and T2D rats. These gaps were wider in diabetic rats, suggesting increased vascular

remodeling. These data should be assessed in future confocal microscopy studies

combining several mural cells’ markers, such as NG2, α-smooth muscle actin and

desmin immunostaining at the inner choroid. Electron microscopy may be useful as well

[39].

We found differences in VEGF/VEGFR2 immunoreactivity between T1D and T2D

rats. VEGFR2 immunoreactivity was increased in T2D. One possible explanation to be

cross-checked in the future is whether VEGFR2 upregulation is a marker of a low level

of diabetes imbalance in T2D, while decreased VEGFR2 immunoreactivity in T1D is a


183

marker of a significant metabolic imbalance. Data from perfusion recovery experiments

in the hind limb muscles of T1D mice, show that VEGFR2 expression is decreased in

the endothelial cells in a hyperglycemia-depend gradient mode via increased

ubiquitination of VEGFR2. Such decrease of VEGFR2 expression was related to poor

adaptation to ischemia and was reversed by improving the metabolic control with insulin

[40]. T1D rats used in our experiments had a significant metabolic imbalance (Hb A1c

8.3 ± 0.3%) and the diabetes duration of 8 weeks is roughly equivalent to six years of

unbalanced diabetes left untreated in a human being [30]. Therefore, one may speculate

that a decreased VEGFR2 immunoreactivity in the T1D rat retina follows the same

pathway, enhancing the role of metabolic control in the development and progression of

diabetic retinopathy, as shown by the results herein displayed in human subjects.

The distribution of the VEGFR2 immunoreactivity in the retina and the co-

localization of the VEGF immunoreactivity with Müller cell marker, vimentin, suggest that

the regulation of VEGFR2 expression is in the endothelium of the three retinal plexuses,

while the regulation of VEGF expression is, at least in part, in Müller cells.

Inflammatory cells were increased in the outer retina of T1D and T2D rats and in

choroid of T1D rats only. Iba1+ and MHC class II+ cells were increased in the choroidal

stroma of T1D rats, and not just passing by within the vessel lumina, since the choroidal

vasculature had been previously cleaned by perfusions. Conversely, they were not

increased in T2D, suggesting that intensity and not duration of diabetes, is paramount in

inflammatory cell expression in the choroid.

Glia migration throughout the retina has been described in normal homeostasis

and diabetes [10, 41-43], but its role in the outer BRB breakdown and in Bruch’s

membrane permeability status in diabetes remains to be established [44, 45]. We found

that the increased presence of Iba1+ cells in the retina, and in particular in the outer

retinal layers, is a distinct feature of diabetic retinas, as previously described by others


184

[10, 46, 47]. The distribution of glial cells around the retinal microvascular network

suggest that they may use vessels as a scaffold for migration. The accumulation of

inflammatory cells in the outer retina may be due to increased cell traffic from the retina

into the choroid [43], to increased metabolic-driven hypoxemia at the outer retina [48]

with VEGF-mediated recruitment of inflammatory cells [36, 49], to increased Brüch’s

membrane impermeabilization [44, 45, 50] or to transcytosis arrest through the RPE cells

[43]. Since we found inflammatory cells to be increased in the choroid and outer retina

of T1D, it will be interesting in the future to find a link between these two events, although

some available evidence seems to indicate that some of the inflammatory cells present

in the choroid, including in the choriocapillaris, actually come from the retina [43].

Interdependence of glial cell proliferation, inner choroid pericyte disposition,

Bruch’s membrane inflammatory-driven changes and choriocapilaris remodeling may be

pathways for future studies.

The localization of arteries at the choriocapillaris lobules, e. g., center [31, 51, 52]

versus periphery [53-58] has been subject to debate, as well as the end-arterial nature

of the choroidal circulation. Angiographic studies support the end-arterial theory [59],

while post-mortem studies revealed that the choroidal circulation has multiple

anastomosis at various levels [54]. An end-arterial functional model was proposed

reconciling the post-mortem and the angiographic findings. The choroidal vascular

system has been proposed as composed of multiple sectors, with the presence of

anastomoses within sector, before the emergence of the last pre-terminal arteriole.

Therefore, the coriocapillaris bed would be end-arterial [60]. By DiI perfusion, we found

a lobar structure of the choroid very similar to the bronchiolar tree in lungs and by RECA-

1 immunostaining a lobular structure of the choriocapillaris as described by others using

different techniques [52, 56, 58]. In the choroid, the blood flows in arteries and veins in

the same direction, contrariwise to most organs [55, 57, 61]. After labeling all sectors,

we mounted the whole choroid of the rat perfused by DiI and found no staining of the


185

vortex veins, which we interpreted as an indirect signal that Dil behaves as an arterial

and choriocapillaris preferential stain. We found evidence of pre-terminal arterial-arterial

anastomoses and of central position of the artery in choroidal lobules.

Capillary communications between the superficial and middle, middle and deep,

and deep and superficial retinal capillary plexuses, were found, as described before by

OCTA in humans and rats [62-65]. Two different conceptions about the organization of

the retinal plexuses were described, always with arterioles more superficial than veins

[66]. A serial organization of the retinal vasculature with the venous drainage draining

primarily into the deep retinal capillary plexus [63] versus an alternative parallel or

‘‘hammock’’ model, wherein each neurovascular/capillary layer operates as an

independent unit with its own arteriolar supply and venous drainage [62]. This model

would support the generally accepted neurovascular unit, with independent control of the

capillary layers that match their neuronal needs. Mixed or combined models were

purposed as well [65, 66]. Whatever model is closest to the real architecture of the retinal

vascular plexuses, undoubtedly the deep vascular plexus drives in a low oxygen

environment and seems to be critical in permanent damage from retinal diseases,

including DR [48, 67-71].

Future studies of arterial versus venous phenotyping or immunolabelling with

three dimensional analyses by confocal microscopy may help to better characterize the

choroidal and retinal microvasculature [72-76].

References

1. Varma, R., et al., Prevalence of and risk factors for diabetic macular edema in the United States. JAMA Ophthalmol, 2014. 132(11): p. 1334-40.

2. Stratton, I.M., et al., Association of glycaemia with macrovascular and microvascular complications of type 2 diabetes (UKPDS 35): prospective observational study. BMJ, 2000. 321(7258): p. 405-12.


186

3. Campos, A., et al., Evaluation of markers of outcome in real-world treatment of diabetic macular edema. Eye Vis (Lond), 2018. 5: p. 27.



6. Romero-Aroca, P., et al., Differences in incidence of diabetic retinopathy between type 1 and 2 diabetes mellitus: a nine-year follow-up study. Br J Ophthalmol, 2017. 101(10): p. 1346-1351.


8. Lampeter, E.R., et al., Elevated levels of circulating adhesion molecules in IDDM patients and in subjects at risk for IDDM. Diabetes, 1992. 41(12): p. 1668-71.



11. Lutty, G.A., Diabetic choroidopathy. Vision Res, 2017. 139: p. 161-167. 12. Campos, A., et al., Viewing the choroid: where we stand, challenges and contradictions




15. Bento, C.F., et al., Methylglyoxal-induced imbalance in the ratio of vascular endothelial growth factor to angiopoietin 2 secreted by retinal pigment epithelial cells leads to endothelial dysfunction. Exp Physiol, 2010. 95(9): p. 955-70.





20. Torabi, H., et al., Choroidal Thickness and Hemoglobin A1c Levels in Patients with Type 2 Diabetes Mellitus. J Ophthalmic Vis Res, 2019. 14(3): p. 285-290.

21. Abalem, M.F., et al., The effect of glycaemia on choroidal thickness in different stages of diabetic retinopathy. Ophthalmic Res, 2020.

22. Campos, A., et al., Choroidal thickness changes stratified by outcome in real-world treatment of diabetic macular edema. Graefes Arch Clin Exp Ophthalmol, 2018. 256(10): p. 1857-1865.

23. Campos, A., et al., Response to: Choroidal thickness changes stratified by outcome in real-world treatment of diabetic macular edema. Graefes Arch Clin Exp Ophthalmol, 2019. 257(1): p. 243-244.


187



26. Wang, H. and Y. Tao, Choroidal structural changes correlate with severity of diabetic retinopathy in diabetes mellitus. BMC Ophthalmol, 2019. 19(1): p. 186.

27. Caldwell, R.B. and M.E. Fitzgerald, The choriocapillaris in spontaneously diabetic rats. Microvasc Res, 1991. 42(3): p. 229-44.




31. Castro-Correia, J., Understanding the choroid. Int Ophthalmol, 1995. 19(3): p. 135-47. 32. Tian, C., et al., Retinal and choroidal imaging in vivo using integrated photoacoustic

microscopy and optical coherence tomography. Proc SPIE Int Soc Opt Eng, 2018. 10474. 33. Manche, E.E. and G.E. Korte, Ultrastructural evidence of remodelling in the

microvasculature of the normal rabbit and human eye. Acta Anat (Basel), 1990. 138(2): p. 89-96.

34. Fryczkowski, A.W., B.L. Hodes, and J. Walker, Diabetic choroidal and iris vasculature scanning electron microscopy findings. Int Ophthalmol, 1989. 13(4): p. 269-79.


36. Ferrara, N., H.P. Gerber, and J. LeCouter, The biology of VEGF and its receptors. Nat Med, 2003. 9(6): p. 669-76.

37. Cebe-Suarez, S., A. Zehnder-Fjallman, and K. Ballmer-Hofer, The role of VEGF receptors in angiogenesis; complex partnerships. Cell Mol Life Sci, 2006. 63(5): p. 601-15.


39. Tilton, R.G., et al., Absence of degenerative changes in retinal and uveal capillary pericytes in diabetic rats. Investigative Ophthalmology & Visual Science, 1986. 27(5): p. 716-721.

40. Dokun, A.O., et al., Glycaemic control improves perfusion recovery and VEGFR2 protein expression in diabetic mice following experimental PAD. Cardiovasc Res, 2014. 101(3): p. 364-72.


42. Omri, S., et al., PKCzeta mediates breakdown of outer blood-retinal barriers in diabetic retinopathy. PLoS One, 2013. 8(11): p. e81600.


44. Moore, D.J. and G.M. Clover, The effect of age on the macromolecular permeability of human Bruch's membrane. Invest Ophthalmol Vis Sci, 2001. 42(12): p. 2970-5.

45. Sorensen, N.B., et al., Bruch's membrane allows unhindered passage of up to 2mum latex beads in an in vivo porcine model. Exp Eye Res, 2019. 180: p. 1-7.


188

46. Rungger-Brandle, E., A.A. Dosso, and P.M. Leuenberger, Glial reactivity, an early feature of diabetic retinopathy. Invest Ophthalmol Vis Sci, 2000. 41(7): p. 1971-80.

47. Lechner, J., O.E. O'Leary, and A.W. Stitt, The pathology associated with diabetic retinopathy. Vision Res, 2017. 139: p. 7-14.

48. Yu, D.Y., et al., Retinal capillary perfusion: Spatial and temporal heterogeneity. Prog Retin Eye Res, 2019. 70: p. 23-54.

49. Holmes, D.I. and I. Zachary, The vascular endothelial growth factor (VEGF) family: angiogenic factors in health and disease. Genome Biol, 2005. 6(2): p. 209.

50. Curcio, C.A. and M. Johnson, Structure, Function, and Pathology of Bruch's Membrane in Retina, S.J. Ryan, Editor. 2013, Elsevier: Philadelphia, PA. p. 465–81.

51. Krey, H.F., Segmental vascular patterns of the choriocapillaris. Am J Ophthalmol, 1975. 80(2): p. 198-202.

52. Castro Correia, J., Anátomo-fisiologia da coroide. Thesis Doctorat, 1958: p. 52-126. 53. Hayreh, S.S., The choriocapillaris. Albrecht Von Graefes Arch Klin Exp Ophthalmol,

1974. 192(3): p. 165-79. 54. Yoneya, S., M.O. Tso, and K. Shimizu, Patterns of the choriocapillaris. A method to

study the choroidal vasculature of the enucleated human eye. Int Ophthalmol, 1983. 6(2): p. 95-9.

55. Torczynski, E. and M.O. Tso, The architecture of the choriocapillaris at the posterior pole. Am J Ophthalmol, 1976. 81(4): p. 428-40.

56. Olver, J.M., Functional anatomy of the choroidal circulation: methyl methacrylate casting of human choroid. Eye (Lond), 1990. 4 ( Pt 2): p. 262-72.

57. Hogan, M.J., J.A. Alvarado, and J.E. Weddell, Histology of the Human Eye — An Atlas and Textbook 1971, Philadelphia, London, Toronto: Philadelphia: W. B. Saunders. 403-497.

58. Risco, J.M. and W. Nopanitaya, Ocular microcirculation. Scanning electron microscopic study. Invest Ophthalmol Vis Sci, 1980. 19(1): p. 5-12.

59. Hayreh, S.S., In vivo choroidal circulation and its watershed zones. Eye (Lond), 1990. 4 ( Pt 2): p. 273-89.

60. Lee, J.E., et al., Functional end-arterial circulation of the choroid assessed by using fat embolism and electric circuit simulation. Sci Rep, 2017. 7(1): p. 2490.

61. Federman, J.L., The fenestrations of the choriocapillaris in the presence of choroidal melanoma. Trans Am Ophthalmol Soc, 1982. 80: p. 498-516.

62. Campbell, J.P., et al., Detailed Vascular Anatomy of the Human Retina by Projection-Resolved Optical Coherence Tomography Angiography. Sci Rep, 2017. 7: p. 42201.

63. Garrity, S.T., et al., Considerations in the Understanding of Venous Outflow in the Retinal Capillary Plexus. Retina, 2017. 37(10): p. 1809-1812.

64. Pi, S., et al., Rodent retinal circulation organization and oxygen metabolism revealed by visible-light optical coherence tomography. Biomed Opt Express, 2018. 9(11): p. 5851-5862.

65. Leahy, C., et al., Mapping the 3D Connectivity of the Rat Inner Retinal Vascular Network Using OCT Angiography. Invest Ophthalmol Vis Sci, 2015. 56(10): p. 5785-93.

66. Nesper, P.L. and A.A. Fawzi, Human Parafoveal Capillary Vascular Anatomy and Connectivity Revealed by Optical Coherence Tomography Angiography. Invest Ophthalmol Vis Sci, 2018. 59(10): p. 3858-3867.

67. Hasegawa, N., et al., New Insights Into Microaneurysms in the Deep Capillary Plexus Detected by Optical Coherence Tomography Angiography in Diabetic Macular Edema. Invest Ophthalmol Vis Sci, 2016. 57(9): p. OCT348-55.

68. Garrity, S.T., V.L. Tseng, and D. Sarraf, Paracentral Acute Middle Maculopathy in a Perivenular Fern-Like Distribution with En Face Optical Coherence Tomography. Retin Cases Brief Rep, 2018. 12 Suppl 1: p. S25-S28.


189

69. Yu, S., et al., The spectrum of superficial and deep capillary ischemia in retinal artery occlusion. Am J Ophthalmol, 2015. 159(1): p. 53-63 e1-2.

70. Spaide, R.F., J.M. Klancnik, Jr., and M.J. Cooney, Retinal vascular layers in macular telangiectasia type 2 imaged by optical coherence tomographic angiography. JAMA Ophthalmol, 2015. 133(1): p. 66-73.

71. Moon, B.G., et al., Correlation between Deep Capillary Plexus Perfusion and Long-Term Photoreceptor Recovery after Diabetic Macular Edema Treatment. Ophthalmol Retina, 2018. 2(3): p. 235-243.

72. Crist, A.M., C. Young, and S.M. Meadows, Characterization of arteriovenous identity in the developing neonate mouse retina. Gene Expr Patterns, 2017. 23-24: p. 22-31.

73. Poissonnier, L., et al., Egfl7 is differentially expressed in arteries and veins during retinal vascular development. PLoS One, 2014. 9(3): p. e90455.

74. Corada, M., M.F. Morini, and E. Dejana, Signaling pathways in the specification of arteries and veins. Arterioscler Thromb Vasc Biol, 2014. 34(11): p. 2372-7.

75. dela Paz, N.G. and P.A. D'Amore, Arterial versus venous endothelial cells. Cell Tissue Res, 2009. 335(1): p. 5-16.

76. Rosa, S., et al., Functional characterization of iPSC-derived arterial- and venous-like endothelial cells. Sci Rep, 2019. 9(1): p. 3826.

Choroid and diabetes Conclusions

191

7. Conclusions


192

Studies on human subjects

1. In a prospective study lasting 2 years, avoiding the most common biases that

are actually present in many studies, the choroidal thickness was found to

decrease with anti-VEGF treatment for diabetic macular edema, but without

prognostic value for anatomic or functional outcomes, at the 3 and 6 months

end-points. Besides, the subfoveal choroidal thickness was found to be a

reliable surrogate of the choroidal thickness as such.

2. In a prospective study lasting 6 months for each subject enrolled, comparing

prognostic factors for anatomic and functional recovery of diabetic macular

edema under anti-VEGF treatment, metabolic control and the ellipsoid zone

status were found to be factors of functional outcome, while subretinal fluid

was found to be a factor of anatomic outcome only.

Studies on animal models of diabetes

3. In streptozotocin-induced Type 1 diabetic (T1D) Wistar rats (16 weeks old; 8

weeks of ongoing T1D):

3.1. To the best of our knowledge, we evaluated for the first time in vivo

choroidal thickness assessed by OCT in T1D rats. The combined findings

of no change in the choroidal thickness and in the choroidal vascular

density suggest that these alterations may be unnoticeable until the late

stages of disease, even when the metabolic control is poor;


193

3.2. The combined findings of no change in the density of mural cells that

wrap around choroidal vessels and the existence of multifocal pericyte

cell-free areas at the choriocapillaris, suggest that most of the vascular

regulatory cells in the choroidal vasculature are present, but vascular

remodeling at the choriocapillaris is increased. Taken together, such

findings and data aforementioned in 3.1 may explain some contradictions

or limitations reported by OCTA in the choriocapillaris;

3.3. Decreased VEGFR2 immunoreactivity in the retina of T1D rats suggests

that VEGFR2 immunoreactivity decreases in a condition of severe

hyperglycemia, probably due to increased VEGFR2 ubiquitination in

endothelial cells, as reported before for hind limb muscles in T1D mice;

VEGF immunostaining co-localized with Müller cell marker vimentin,

suggesting that the regulation of VEGF is significantly dependent on

Müller cells;

3.4. The combined findings of Iba1+ and MHC+ cells sparing the inner choroid

with the increased number of inflammatory cells in the choroidal stroma

and in the retina, including the outer retina, suggest a potential

mechanism of injury at the choriocapillaris in diabetes, mainly when the

metabolic control is poor.

4. In Goto Kakizaki Type 2 diabetes (T2D) rats (52 weeks old):

4.1. To the best of our knowledge we evaluated for the first time in vivo


194

choroidal thickness assessed by OCT in T2D rats. Increased choroidal

thickness and decreased vascular density at the choriocapillaris, suggest

that in the long lasting disease substantial alterations occur in the

choroid, including vascular remodeling and permanent damage, and that

such alterations may be assessed by OCT and OCTA with a high

potential agreement;

4.2. As for point 3.2, the combined findings of no difference in perivascular

mural cells immunoreactivity in the choroid and pericytes’ rarefaction at

the inner choroid, suggest that the disease has more impact on vascular

remodeling at the choriocapillaris than on regulatory peri-vascular mural

cells;

4.3. Increased VEGFR2 immunoreactivity in the retina with no difference for

VEGF immunoreactivity, suggests that in the long lasting disease, with

slight metabolic imbalance, VEGFR2, rather than VEGF, may be the best

molecular signature of disease status or the best surrogate of the

homeostasis mechanisms.

The retinal layers where VEGFR2 immunoreactivity increased (Figures

4.3 and 5.5) suggest a co-localization with the three retinal plexuses and

with Müller cells;

4.4. Increased number of inflammatory cells in the outer retina, but not in the

choroid, where Iba1+ and MHC+ spare the inner choroid, suggests that

even in the long lasting disease, as long as there is a reasonable

metabolic control, the signs of inflammation are mainly confined to the

retina.


195

5. The combined findings in T1D and T2D rats suggest that:

5.1. Choriocapillaris remodeling was increased and vascular density at the

choriocapillaris was decreased in T2D rats. The choroidal thickness was

increased in T2D rats only.

These results suggest that OCT and OCTA findings on the human

choroid, respecting choroidal thickness or vascular indexes of the

choriocapillaris, may be more close to consensus between studies

whenever checking the long lasting disease;

5.2. VEGFR2 immunoreactivity in the retina decreased when the metabolic

control was poor in T1D. Conversely, VEGFR2 immunoreactivity in the

retina increased in the presence of mild hyperglycemia in T2D.

These resuts may be related to a decreased VEGFR2 expression in the

presence of sustained high glucose levels in the retinal vascular

endothelium;

5.3. Higher immunoreactivity of VEGF and of VEGFR2 at the innermost

retina, namely at the retinal nerve fiber layer, may explain why the

intravitreal administration of anti-VEGF agents is so effective. Anti-

VEGFs readily sequester VEGF in the vitreous and, by rapidly diffusing

to the inner retina, they prevent VEGF-A to bind to VEGFR2, the VEGF

receptor with the highest tyrosine kinase activity, either in the superficial

vascular plexus’ endothelium or in Müller cells;

5.4. The increased presence of inflammatory cells in the choroid of T1D rats,

suggests that when the metabolic control is poor, inflammatory cells


196

may pack at the innermost choroid, potentially impairing the

choriocapillaris or causing permanent damage.

The perivascular distribution of Iba1+ cells in the retina suggests that

inflammatory cells may use the retinal capillaris network as a scaffold

for migration;

5.5. These inflammatory-like alterations demonstrated in the choroid may

explain, at least in part, the success of intravitreal steroids in DME, even

when nonresponsive to anti-VEGFs, either by acting at the Müller cells

mechanisms of water clearance, or by reducing microglial cell migration

to the outer retina or, in addition, by reducing the inflammatory

microenvironment in the choroidal stroma.

estudogeral.uc.pt · 2021. 1. 7. · Agradecimentos “O agradecimento é a mais alta forma de...

Documents

Transcript of estudogeral.uc.pt · 2021. 1. 7. · Agradecimentos “O agradecimento é a mais alta forma de...