REGULATION OF PXDN IN EYE DEVELOPMENT AND PXDN GENE VARIANT 

SCREENING WITHIN A SOUTH AFRICAN COHORT OF PATIENTS PRESENTING 

WITH ANTERIOR SEGMENT DYSGENESIS 

By 

TEBOGO RECTOR MARUTHA 

 

Thesis 

Submitted in fulfilment of the requirements for the degree 

Doctor of Philosophy 

in  

Molecular and Cell Biology 

in the Faculty of Science, University of the Witwatersrand, Johannesburg, South Africa 

 

Supervisor: Prof D. Mavri-Damelin 

Co-supervisor: Dr N. Carstens 

 

December 2022 

 

  



i 
 

Declaration 

I, Tebogo Rector Marutha (1111612), am a student registered for the degree of Doctor of 

Philosophy in the academic year 2022.  

I hereby declare the following:  

 I am aware that plagiarism (the use of someone else’s work without their permission 

and/or without acknowledging the original source) is wrong.  

 I confirm that the research proposal submitted for assessment for the above degree 

is my own unaided work except where explicitly indicated otherwise and 

acknowledged.  

 I have followed the required conventions in referencing the thoughts and ideas of 

others.  

 I understand that the University of the Witwatersrand may take disciplinary action 

against me if there is a belief that this is not my own unaided work or that I have 

failed to acknowledge the source of the ideas or words in my writing.  

 

 Signature_____   07th day of December 2022_ YEAR 

  



ii 
 

 

 

 

 

 

 

 

 

I would like to dedicate this work to my loving parents and family for all the support they 

have given me. 

  



iii 
 

Abstract 

Peroxidasin (PXDN) is an extracellular matrix-associated haem-peroxidase predominantly 

expressed in the vasculature and eye. PXDN crosslinks collagen IV through sulfilimine bond 

formation in the basement membrane. Aberrant PXDN expression is associated with fibrosis, 

heart failure and cancer, and various pathologies of the eye, where PXDN likely provides 

structural support via basement membrane synthesis in the cornea and lens during eye 

development, as well as protect the lens, trabecular meshwork and cornea against oxidative 

damage. Furthermore, PXDN pathogenic variants have been associated with anterior segment 

dysgenesis (ASD), congenital cataracts and corneal opacity. To further understand the role of 

PXDN in the eye, first we aimed to identify PXDN as a novel target of key transcriptional 

regulators of eye development, namely PAX6, FOXC1 and PITX2, and second, to screen a 

cohort of South African patients with ASD to look for pathogenic variants in PXDN and other 

ASD genes. Protein expression of PAX6, FOXC1, PITX2 and PXDN, in response to Fibroblast 

Growth Factor-2 (FGF-2) were quantified by western blotting and localisation visualised using 

immunofluorescence confocal microscopy. Chromatin immunoprecipitation-PCR and 

luciferase assays were employed to detect transcription factor-PXDN promoter interactions. 

Expression data established that PXDN, PAX6, FOXC1 and PITX2 were induced by FGF2 at 

varying time points. Putative binding sites for all three transcription factors were identified in 

the PXDN promoter and ChIP-PCR confirmed that PAX6, FOXC1 and PITX2 interact with 

various regions of the promoter. Luciferase reporter assays are currently underway. Next 

Generation Sequencing of genomic DNA from South African patients exhibiting ASD 

disorders. identified disease causing variants in PAX6 and GJA8. Variants of uncertain 

significance were identified in PXDN, BCOR, EPHA2 and LTBP2 genes and are being 

investigated further. In conclusion, we identified PXDN as a novel target of PAX6, FOXC1 

and PITX2 that further supports for the integral role of PXDN in eye development. 

  



iv 
 

Research output 

Original Publication: Marutha, T.R., Williams, S., Novellie, M., Dillon, B., Carstens, N., and Mavri-

Damelin, D., 2022. Exome sequencing in South African patients presenting with Anterior segment 

dysgenesis. Submitted manuscript to Ophthalmic Genetics Journal, ISSN: 1744-5094.  

Conferences: 

Marutha, T.R., Williams, S., Novellie, M., Dillon, B., Carstens, N., and Mavri-Damelin, D., 2018. 

Regulation of PXDN in eye development and PXDN gene variant screening within a South African 

cohort of patients presenting with anterior segment dysgenesis. Life sciences imaging facility research 

day, University of the Witwatersrand, Johannesburg, Oral presentation. 

Marutha, T.R., Williams, S., Novellie, M., Dillon, B., Carstens, N., and Mavri-Damelin, D., 2019. 

Regulation of PXDN in eye development and PXDN gene variant screening within a South African 

cohort of patients presenting with anterior segment dysgenesis. University of the Witwatersrand 10th 

Cross-Faculty postgraduate symposium, Johannesburg, Poster presentation. 

Marutha, T.R., Williams, S., Novellie, M., Dillon, B., Carstens, N., and Mavri-Damelin, D., 2020. 

Regulation of PXDN in eye development and PXDN gene variant screening within a South African 

cohort of patients presenting with anterior segment dysgenesis. University of the Witwatersrand 11th 

Cross-Faculty postgraduate symposium, Johannesburg, Grad-flash presentation. 

Marutha, T.R., Williams, S., Novellie, M., Dillon, B., Carstens, N., and Mavri-Damelin, D., 2022. 

Regulation of PXDN in eye development and PXDN gene variant screening within a South African 

cohort of patients presenting with anterior segment dysgenesis. South African Society of Biochemistry 

and Molecular Biology (SASBMB), Virtual event, Poster presentation. 

Marutha, T.R., Williams, S., Novellie, M., Dillon, B., Carstens, N., and Mavri-Damelin, D., 2023. 

Regulation of PXDN in eye development and PXDN gene variant screening within a South African 

cohort of patients presenting with anterior segment dysgenesis. 14th International Congress of Human 

Genetics (ICHG2023), Poster presentation.  



v 
 

Acknowledgements 

This project has been one of the most challenging journeys in my quest for scientific knowledge 

and as such I would like to thank the following individuals as well as organisations for their 

contribution towards the success of this study. 

I would like to thank my supervisor, Prof Demetra Mavri-Damelin, first for designing the 

project and setting it in motion by granting me the opportunity to join her research lab. 

Secondly for her support, encouragement and guidance which enabled me to achieve more than 

I thought possible in the years spent in the Functional Genetics Research Laboratory. It was 

indeed a blessing to have been under her guidance for the duration of this research project. 

I would also like to thank my co-supervisor Dr Nadia Carstens, for collaborating with us on 

this project which enabled me to learn even more techniques and acquire more skills. Thank 

you for the support with analysis of the NGS data, exposure to various bioinformatics tools and 

the co-supervision. 

I am grateful to the funding bodies that supported me, and this work, financially, through the 

provision of bursaries and awards. This work was funded by scholarship from the National 

Research Foundation. I would also like to thank the National Health Laboratory services for 

technical assistance with Next generation sequencing and the University of the Witwatersrand. 

I would like to thank the participants as well as their families for their participation in this 

study. Dr Susan Williams, Dr Michael Novellie, Dr Bronwyn Dillon and Mrs Prescilla January 

for recruitment of patients, diagnosis, as well as sample collection. 

To my colleagues in the Functional Genetics Research Laboratory whom have become very 

close and precious friends, Kayleen Jegels, Thokozile Makhanya, Lebogang Moshupya, 

Jemma Falkov, Eunice Niyobuntu and Ntakadzeni Nemuvumoni thank you for all the 

memories, support, encouragement advice and most importantly the laughs in both rainy and 

sunny days. 

I would like to thank Dr. Clement Penny and Dr. Aurelie Deroubaix for so much assistance 

with the Carl Zeiss LSM 780 confocal microscope at the Life Sciences Imaging Facility at the 

University of the Witwatersrand Medical School. Dr Angela Botes for access to the Qsonica 

probe sonicator, Daniesha Govender and Bronwyn Mol for access to HEK293 cell line, finally 

Dr Kerry Hanmer and Dr Boitumelo Sitole for optimisation of some section of the methods. 



vi 
 

To my parents, Headman Marutha and Agnes Marutha thank you for all your unwavering 

support throughout my academic journey and life in general. Each of you in your own unique 

ways have imparted wisdom to me that has strengthened me through difficult times of this 

journey. 

And over and above I thank the all-mighty God for everything he has made possible.    



vii 
 

Table of contents 

Declaration.............................................................................................................................................. i 

Abstract ................................................................................................................................................. iii 

Research output ................................................................................................................................... iv 

Acknowledgements ............................................................................................................................... v 

Table of contents ................................................................................................................................. vii 

List of figures ........................................................................................................................................ xi 

List of tables........................................................................................................................................ xiv 

List of abbreviations ........................................................................................................................... xv 

CHAPTER 1: Introduction ................................................................................................................ xx 

1.1 Peroxidasin overview of structure and functions .................................................................... 1 

1.2 Functions of other haem peroxidase-cyclooxygenases ........................................................... 3 

1.3 Peroxidasin Expression ........................................................................................................... 5 

1.3.1 PXDN in the cardiovascular system ............................................................................... 5 

1.3.2 PXDN in the eye ............................................................................................................. 6 

1.4 Physiological Functions of Peroxidasin .................................................................................. 7 

1.4.1 Extracellular Matrix Formation in Basement Membrane ............................................... 7 

1.4.2 Collagen IV and Reaction Mechanism of PXDN ........................................................... 9 

1.4.3 Role in Host Immune Defence ...................................................................................... 12 

1.5 Peroxidasin in Pathophysiological conditions ...................................................................... 13 

1.5.1 Cardiovascular Disorders .............................................................................................. 13 

1.5.2 Cancer ........................................................................................................................... 14 

1.5.3 Kidney Disorders .......................................................................................................... 15 

1.5.4 Eye Disorders ................................................................................................................ 16 

1.6 Anatomy of the human eye ................................................................................................... 17 

1.6.1 The anterior segment of the eye .................................................................................... 17 

1.6.2 Anterior segment morphogenesis .................................................................................. 19 

1.6.2.1 Optic cup morphogenesis .............................................................................................. 21 

1.6.2.2 Lens morphogenesis ...................................................................................................... 22 

1.6.2.3 Cornea morphogenesis .................................................................................................. 23 

1.6.2.4 Iris and Ciliary body morphogenesis ............................................................................ 24 

1.6.2.5 Anterior chamber morphogenesis ................................................................................. 25 

1.6.3 Regulatory networks involved in the anterior segment of the eye. ............................... 25 

1.6.4 Anterior segment dysgenesis ........................................................................................ 29 

1.6.4.1 Axenfeld Rieger syndrome ........................................................................................... 30 



viii 
 

1.6.4.2 Peters anomaly and Peters-plus syndrome .................................................................... 30 

1.6.4.3 Aniridia ......................................................................................................................... 30 

1.6.4.4 Sclerocornea .................................................................................................................. 31 

1.6.4.5 Primary congenital glaucoma ........................................................................................ 31 

1.6.4.6 Microphthalmia ............................................................................................................. 32 

1.6.5 The genetics of anterior segment dysgenesis ................................................................ 33 

1.6.6 Eye disorders in South Africa ....................................................................................... 45 

1.7 Aims and Objectives ........................................................................................................... 46 

CHAPTER 2: Regulation of PXDN by PAX6, FOXC1 and PITX2 ............................................... 47 

2.1 Introduction ........................................................................................................................... 48 

2.1.1 PXDN in eye development ........................................................................................... 48 

2.1.2 PAX6, FOXC1 and PITX2 in development and maintenance of the eye ..................... 48 

2.2 Methodology ......................................................................................................................... 51 

2.2.1 Reagents ........................................................................................................................ 51 

2.2.2 Cell culture .................................................................................................................... 52 

2.2.3 Immunofluorescence confocal microscopy ................................................................... 53 

2.2.4 Protein quantification and Western blot ........................................................................ 54 

2.2.4.1 Cell lysate preparation .................................................................................................. 55 

2.2.4.2 Protein estimation using the Bramhall assay ................................................................ 55 

2.2.4.3 SDS-PAGE ................................................................................................................... 56 

2.2.4.4 Electrophoretic Transfer and Antibody Probing ........................................................... 57 

2.2.4.5 Protein Visualisation and Quantification ...................................................................... 58 

2.2.5 Chromatin immunoprecipitation assay ......................................................................... 58 

2.2.5.1 Analysis of the PXDN promoter region and ChIP-PCR primer design ........................ 58 

2.2.5.2 Phenol-Chloroform gDNA extraction and optimisation of primers .............................. 62 

2.2.5.3 ChIP protocol ................................................................................................................ 63 

2.2.5.4 ChIP-PCR ..................................................................................................................... 64 

2.2.6 Luciferase reporter assay .............................................................................................. 65 

2.2.6.1 Vector constructs preparation for luciferase ................................................................. 65 

2.2.6.2 Competent E. coli cells preparation .............................................................................. 67 

2.2.6.3 Transformation and Colony PCR .................................................................................. 68 

2.2.6.4 Alkaline Lysis ............................................................................................................... 68 

2.2.6.5 Luciferase reporter assay .............................................................................................. 69 

2.2.7 Statistical analysis ......................................................................................................... 70 



ix 
 

2.3 Results ................................................................................................................................... 71 

2.3.1 Expression of PXDN, PAX6, FOXC1 and PITX2 in response to FGF-2 ..................... 71 

2.3.2 Quantification of PXDN, PAX6, FOXC1 and PITX2 proteins in response to FGF-2 .. 74 

2.3.3 PAX6, FOXC1 and PITX2 interaction with PXDN promoter ...................................... 78 

2.3.4 PAX6-PXDN luciferase reporter gene assay ............................................................ 83 

2.4 Discussion ............................................................................................................................. 85 

CHAPTER 3: Variant screening in eye development genes ........................................................... 94 

3.1 Introduction ........................................................................................................................... 95 

3.1.1 Genetics of eye development ........................................................................................ 95 

3.1.2 Genetics of eye disorders .............................................................................................. 95 

3.2 Methods................................................................................................................................. 96 

3.2.1 Phenotypic data collection and eye disorders diagnosis ............................................... 96 

3.2.2 Sample collection and DNA extraction......................................................................... 96 

3.2.3 Target gene selection and custom gene panel design .................................................... 96 

3.2.4 Next generation sequencing of target genes .................................................................. 97 

3.2.4.1 Library preparation of DNA samples ............................................................................ 98 

3.2.4.2 DNA Templating........................................................................................................... 99 

3.2.4.3 Ion torrent sequencing of DNA samples ....................................................................... 99 

3.2.5 Analysis of genetic variants data ................................................................................ 100 

3.2.5.1 Post sequencing metrics and Read alignment ............................................................. 100 

3.2.5.2 Variant calling and interpretation ............................................................................... 101 

3.2.5.3 Annotation of variants ................................................................................................. 101 

3.2.5.4 Prioritisation of variants .............................................................................................. 102 

3.2.5.5 Variant classification ................................................................................................... 103 

3.2.5.6 Integrative genomics viewer ....................................................................................... 103 

3.3 Results ................................................................................................................................. 104 

3.3.1 Patient’s clinical data ...................................................................................................... 104 

3.3.2 Quality control of Ion torrent sequencing data ............................................................... 107 

3.3.3 Identified variants from NGS .......................................................................................... 110 

3.3.4 Putative disease-causing variants .................................................................................... 111 

3.3.4.1 Paired box protein 6 (PAX6) ...................................................................................... 113 

3.3.4.2 Gap junction alpha 8 (GJA8) ...................................................................................... 115 

3.3.5 Variants of uncertain significance (VUS) ....................................................................... 117 

3.3.5.1 BCL6 corepressor (BCOR) ......................................................................................... 120 



x 
 

3.3.5.2 EPH receptor A2 (EPHA2) ......................................................................................... 124 

3.3.5.3 Latent transforming growth factor beta binding protein 2 (LTBP2) ........................... 126 

3.3.5.4 Peroxidasin (PXDN) .................................................................................................... 131 

3.4 Discussion ........................................................................................................................... 136 

3.4.1 Pathogenic PAX6 variant ............................................................................................ 137 

3.4.2 Likely pathogenic GJA8 variant ................................................................................. 139 

3.4.3 Varsome pathogenic BCOR and EPHA2 but VUS in this study ................................. 143 

3.4.4 Variants of uncertain significance: LTBP2 and PXDN ............................................... 146 

CHAPTER 4: Overall conclusions and future experiments ......................................................... 152 

4.1 Summary conclusions ......................................................................................................... 153 

4.2 Future experiments .............................................................................................................. 155 

Appendices ......................................................................................................................................... 195 

Appendix A ..................................................................................................................................... 195 

Appendix B ..................................................................................................................................... 196 

Appendix C ..................................................................................................................................... 201 

Appendix D ...................................................................................................................................... 202 

Appendix E ..................................................................................................................................... 214 

 

  



xi 
 

List of figures 

Figure 1.1 The structure of both PXDN and PXDNL…………………………...........2 

Figure 1.2 Evolution of PXDN………………………………………………………..3 

Figure 1.3 Two subtypes of ECM……………………………………………………..8 

Figure 1.4 Crosslinking of collagen IV is mediated by PXDN………………………..9 

Figure 1.5 PXDN catalyses of HOBr to produce sulfilimine (S=N)…………………..11 

Figure 1.6 Anatomical structure of the human eye…………………………………….18 

Figure 1.7 Foetal development of the anterior segment of the human eye…………….20 

Figure 1.8 Optic cup development…………………………………………..................21 

Figure 1.9 Cornea development……………………………………………..................23 

Figure 1.10 Genes that overlap in clinical features of congenital eye disorders………..33 

Figure 1.11 Various genes activated at different stages of oculogenesis…….................34 

Figure 2.1 Multiple cloning site of pGL4.10 vector.……………………….................66 

Figure 2.2 PXDN localisation in response to FGF-2……………………….................71 

Figure 2.3 FOXC1 localisation in response to FGF-2……………………....................72 

Figure 2.4 PAX6 localisation in response to FGF-2………………………..................72 



xii 
 

Figure 2.5 PITX2 localisation in response to FGF-2………………………..................73 

Figure 2.6 Western blot analyses of PXDN in response to FGF-2…………………….74 

Figure 2.7 Western blot analyses of FOXC1 in response to FGF-2…………………...75 

Figure 2.8 Western blot analyses of PAX6 in response to FGF-2……………………..76 

Figure 2.9 Western blot analyses of PITX2 in response to FGF-2…………………….77 

Figure 2.10 Western blot analyses of ACTB in response to FGF-2…………………….77 

Figure 2.11 Sonicated DNA used in ChIP-PCR………………………………………..78 

Figure 2.12 FOXC1 interaction with PXDN……………………………………………80 

Figure 2.13 PAX6 interaction with PXDN……………………………………………..81 

Figure 2.14 PITX2 interaction with PXDN……………………………………………82 

Figure 2.15 Successful cloning of oligonucleotides into pGL4.10….…………………84 

Figure 3.1 Diagram illustrating the principle of whole exome sequencing…………..98 

Figure 3.2 Quality metrics data for NGS pre and post sequence alignment………….108 

Figure 3.3 Quality metrics data for NGS pre and post sequence alignment………….109 

Figure 3.4 IGV plot for PAX6 c.760C>T variant…………………………………….114 

Figure 3.5 IGV plot for GJA8 c.22G>A variant………………………………………116 



xiii 
 

Figure 3.6 IGV plot for BCOR frameshift variant patient (ID: 6)…………................121 

Figure 3.7 IGV plot for BCOR frameshift variant parent (ID: 11)…………………...122 

Figure 3.8 IGV plot for BCOR frameshift variant patient (ID: 15)…………………..123 

Figure 3.9 IGV plot for EPHA2 frameshift variant patient (ID: 13)……….................125 

Figure 3.10 IGV plot for LTBP2 c.1577G>A patient (ID: 12)…………………………127 

Figure 3.11 IGV plot for LTBP2 c.1577G>A patient (ID: 11)…………………………128 

Figure 3.12 IGV plot for LTBP2 c.3527-3C>A patient (ID: 1)………………………...130 

Figure 3.13 IGV plot for PXDN c.2827C>T patient (ID: 12)………………………….132 

Figure 3.14 IGV plot for PXDN c.2827C>T parent (ID: 10)…………………………..133 

Figure 3.15 IGV plot for PXDN c.2827C>T patient (ID: 14)………………………….134 

Figure 3.16 IGV plot for PXDN c.1112C>T patient (ID: 14)………………………….135 

  



xiv 
 

List of tables 

 

Table 1.1 List of genes associated with Anterior segment dysgenesis disorders............37-42 

Table 1.2 Studies that employed NGS in the interrogation of various ASD genes……43 

Table 1.3 Pathogenic variants detected in patients presenting with ASD………….......44 

Table 2.1 Antibodies used for the immunochemistry techniques in this study…….......51 

Table 2.2 Putative binding sites for PAX6, FOXC1 and PITX2 on PXDN……………59 

Table 2.3 ChIP-PCR primer sequences and annealing temperature……………….......61 

Table 2.4 PAX6 and WNT5A oligonucleotides sequences for cloning………………..67 

Table 3.1 Patient demographic and clinical data………………………………………105-106 

Table 3.2 Total number of variants detected per patient pre-filtration…………….......110 

Table 3.3 Pathogenic Variants detected in this study………………………………….112 

Table 3.4 Variants of Uncertain significance in this study…………………………….118-119 

   



xv 
 

List of abbreviations 

 

A Adenine 

ACMG American college of medical genetics and genomics 

ARS Axenfeld-Rieger syndrome 

ASD Anterior Segment Dysgenesis 

ASD Anteror segment dysgenesis 

BAM Binary annotation map 

BCOR BCL6 corepressor 

BM Basement Membrane 

Br- Bromine 

C Cytosine 

CADD Combined annotation-dependent depletion 

ChIP Chromatin immunoprecipitation 

ChIP Chromatin Immunoprecipitation 

Cl- Chlorine 

COL4A Collagen IV 

COPD Chronic Obstructive Pulmonary Disease 

Cx50 Connexin50 

DANN Deleterious annotation of Genetic Variants using Neural 

Network 

DAP 4',6-Diamidino-2-Phenylindole 

DMEM Dulbecco's Modified Eagle's Medium 



xvi 
 

DUOX Dual Oxidase 

ECL1 Extracellular loop one 

ECL2 Extracellular loop two 

ECM Extracellular Matrix 

EMT Epithelial-Mesenchymal Transition 

EPHA2 Ephrin receptor tyrosine kinase A2 

EPO Eosinophil Peroxidase 

ExAC Exome Aggregation Consortium 

FATHMM Functional Analysis through Hidden Markov Models 

FBN1 Fibrillin 

FBS Foetal Bovine Serum 

FOXC1 Forkhead box c1 

G Guanine 

GBM Glomerular Basement Membrane 

GJA8 Gap junction alpha 8 

gnomAD Genome aggregation database 

H2O2 Hydrogen Peroxide 

HO-1 Heme Oxygenase 1 

HOBr Hypobromous Acid 

HOCI Hypochlorus Acid 

HOSCN Hypothiocyanous Acid 

IGV Integrative genomics viewer 



xvii 
 

IOP Intraocular pressure 

IOP Intra-Ocular Pressure 

IR Ischemia Reperfusion 

LPO Lactoperoxidase 

LTBP2 Latent Transforming Growth Factor Beta Binding Protein 2 

MAC Microphthalmia, Anophthalmia and Coloboma 

MAF Minor allele frequency 

Met Methionine 

MG50 Melanoma Associated Gene 50 

MPO Myeloperoxidase 

NaCl Sodium Chloride 

NADPH Nicotinamide Adenine Dinucleotide Phosphate 

NC1 Non-Collagenous Domain 1 

NCBI National Center for Biotechnology Information 

NGS Next generation sequencing 

NGS Next Generation Sequencing 

NOX NADPH Oxidase 

NRF2 Nuclear factor-erythroid factor 2-related factor 2 

NT N-terminal domain 

O2 Oxygen 

O2- Superoxide 

OCFD Oculofaciocardiodental 



xviii 
 

PAX6 Paired Box 6 

PCG Primary Congenital Glaucoma 

PH Pulmonary Hyperfusion 

PITX2 Paired-like homeodomain transcription factor 2 

Polyphen2 Polymorphism Phenotyping v2 

PRG2 P53-Responsive Gene 2 Protein 

PST Proline serine threonine 

PUFD PCGF Ub-like fold discriminator domain 

PXDN Peroxidasin 

PXDNL Peroxidasin Like  

ROS Reactive Oxygen Species 

SNAI1 Snail Family Transcriptional Repressor 1 

SCN- Thiocyanate 

SDS Sodium Dodecyl Sulfate 

SIFT Sorting intolerant from tolerant  

SNP Single Nucleotide Polymorphism 

SNV Single nucleotide variation 

T Thymine 

TGF-β1 Transforming Growth Factor Beta 1 

TM1 Transmembrane one 

TM2 Transmembrane two 

TPO Thyroid Peroxidase 



xix 
 

URE Uncorrected Refractive Errors 

VCF Variant caller file type 

VPO1 Vascular Peroxidasin 1  

VUS Variant of uncertain significance 

vWF Von Willebrand Factor 

WES Whole exome sequencing 

  



xx 
 

 

 

 

 

CHAPTER 1: 

Introduction



1 
 

1.1 Peroxidasin overview of structure and functions 

Peroxidasin (PXDN) was first isolated from Drosophila, as being secreted by 

mesoderm derived haemocytes (Nelson et al., 1994).  Since then, the human gene 

has been identified to encode for a secreted enzyme that contains both a peroxidase 

catalytic domain and multiple motifs found in extracellular matrix (ECM) proteins 

(Nelson et al., 1994; Péterfi and Geiszt, 2014). In humans, PXDN is predominantly 

expressed in the cardiovascular system by cardiomyocytes (Zhang et al., 2012) and 

it is also expressed in the developing eye, more specifically in the layers of lens and 

corneal epithelium (Khan et al., 2011). Some suggested general functions of PXDN 

include innate immune defence (Li et al., 2012), catalysis of oxidative reactions 

using hydrogen peroxide (H2O2) (Ma et al., 2013), and in the biogenesis of basement 

membrane in cells and tissues (Bhave et al., 2012). PXDN belongs to the haem 

containing cyclooxygenase peroxidase family that consists of: thyroid peroxidase 

(TPO), lactoperoxidase (LPO), eosinophil peroxidase (EPO) and myeloperoxidase 

(MPO) (Davies et al., 2008; Soudi et al., 2012; Lázár et al., 2015). In addition, a 

peroxidasin like protein (PXDNL) has also been identified, which was found to have 

a high level of sequence similarity to PXDN but lacks peroxidase activity (Figure 

1.1); to date it has only been found to be expressed in humans (Peterfi et al., 2014).  

The PXDN gene is located at chromosome 2 (2p25.3) and consists of 23 exons that 

encode a 1479 amino acid protein. PXDN has also been referred to in the literature 

as MG50 (melanoma-associated gene 50) (Weiler et al., 1994; Mitchell et al., 2002), 

PRG2 (p53-responsive gene 2 protein) (Horikoshi et al., 1999) and VPO1 (vascular 

peroxidase 1) (Cheng et al., 2008). Besides containing a peroxidase domain, which 

is highly homologous to other animal peroxidases (Figure 1.1), PXDN also contains 

domains with characteristics of proteins of the ECM, these include: Ig-like C2 

domains, which have been found to have various biological functions with the most 

common function being pattern recognition and cell adhesion (Von Castelmur et al., 

2008); von Willebrand Factor domain (vWF), which aids in oligomerisation of 

proteins as well as regulation of bone morphogenetic binding (Sadler, 2009) and 

leucine-rich repeats which have been found in numerous proteins, which seem to be 

involved in protein to protein interactions including ECM assembly, RNA 



2 
 

processing, cell adhesion, platelet aggregation, neural development and signal 

transduction (Bella et al., 2008) (Figure 1.1). These motifs often function in 

interactions between proteins although the exact role of each domain in PXDN 

remains largely un-investigated (Péterfi and Geiszt, 2014). In a study by Ero-Tolliver 

etal., (2015) it was found that the catalytic and Ig domains were required for 

successful sulfilimine bond formation between collagen IV molecules. The leucine-

rich repeats domain was also found to be required for efficient binding of laminin in 

basement membrane (BM) (Sevcnikar et al., 2020). 

 
 

Figure 1.1: Comparison of PXDN and PXDNL structure. The structure of both 

PXDN and PXDNL show similar domains (vWF C type, Leucine rich repeats, 

Peroxidase, Ig like and N terminal secretory signal) and share homology with other 

haem peroxidases within the peroxidase domain (Modified from Péterfi and Geiszt, 

2014). 

 

Among the haem containing cyclooxygenase family of peroxidases, PXDN has a 

larger molecular size of approximately 165 Kilodalton (kDa) and shares greatest 

86% similarity with MPO (Cheng et al., 2008). Despite the additional domains of 

PXDN, the amino acid residues that form the core of PXDN within the peroxidase 



3 
 

domain are highly conserved, which suggest that the PXDN haem catalytic site 

functions in a similar manner (Cheng et al., 2011). 

 

1.2 Functions of other haem peroxidase-cyclooxygenases 

Peroxidase cyclooxygenases is a family of haem containing peroxidases with a 

central structure of coordinated haem iron which is used as a redox cofactor and 

H2O2 as the electron acceptor in the catalysis of various oxidative reactions 

(Zámocký et al., 2015). Peroxidase-cyclooxygenases can also use haem to catalyse 

electron oxidation of other molecules such as proteins, anions, cations as well as 

tyrosine (Zámocký et al., 2015). PXDN has been found to likely be the ancestral 

member of the haem containing peroxidase family conserved to the basal phylum 

Cnidaria (Figure 1.2), which reflects its function in the critical aspect of tissue 

formation in animals and BM assembly (Soudi et al., 2012).  

 
Figure 1.2: A diagrammatic representation of the evolution of PXDN originally in 

Cnidaria then to invertebrates to vertebrates through a series of gene duplication, loss 

of non-catalytic domain and loss of TM domain to give rise to EPO, MPO and LPO 

(Ero-Tolliver et al., 2015). 

 



4 
 

Through gene duplication, loss of the non-catalytic domains and merging with 

transmembrane domain, PXDN was observed to have likely given rise to TPO in 

early Chordates with thyroid hormone signalling which may be associated with 

endothermy (Little and Seebacher, 2014; Ero-Tolliver et al., 2015). After the loss of 

the transmembrane domain in TPO, coupled with successive gene duplication, this 

may have likely given rise to LPO, which is primarily expressed by the glandular 

epithelium; MPO, which appears in vertebrates for the development of neutrophils 

and EPO, which appears in eosinophils (Zámocký et al., 2008). The human genome 

location of TPO and PXDN on chromosome 2 adjacent to one another, and MPO, 

LPO and EPO forming a cluster on chromosome 17, reflects the evolutionary 

ontogeny of animal haem peroxidases (Ero-Tolliver et al., 2015). 

 

Much of the activity of this family of enzymes is used in antimicrobial defense. The 

antimicrobial toolbox of neutrophils comprises MPO as one of its key components, 

which enables neutrophils in both adaptive and innate immunity to serves as effector 

cells (Nauseef and Borregaard, 2014). MPO which is expressed by activated 

neutrophils, is released at sites of tissue injury to phagolysosomes containing 

engulfed pathogens. This is achieved by infiltrating inflammatory cells to provide 

defence against invading pathogenic microbes via assembly of nicotinamide adenine 

dinucleotide phosphate (NADPH) oxidases (NOX) membrane which produces 

superoxide that dismutase H2O2 to provide a substrate for production of 

hypochlorous acid (HOCl). MPO may also be secreted extracellularly in the normal 

degranulation process, packaged within neutrophil extracellular traps, or released 

after neutrophil cell death (Van der Veen et al., 2009; Davies, 2011). Lower 

concentrations of MPO are also secreted by macrophages and monocytes which are 

required for host defense (McMillen et al., 2005).  

LPO is released from mucosal glands and found in secretions like saliva, milk or 

tears where it exhibits antimicrobial activity due to the ability to catalyse oxidative 

reactions with H2O2 as a substrate and a halide ion such as SCN- in the above-

mentioned tissues (Kussendrager and Hooijdonk, 2000; Köksal et al., 2016).   



5 
 

EPO is also part of the antimicrobial arsenal, which is secreted by eosinophils tasked 

to destroy viral and bacterial infections as well as parasitic-worms or helminths. 

Granules containing EPO are secreted towards the target rather than phagocytise 

their target pathogens since parasites are bigger in size and this is a primary function 

unique to EPO (Furtmüller et al., 2006; Davies et al., 2008). 

TPO is expressed in the thyroid gland and aids with the synthesis of thyroid hormone 

via initiation of the biosynthetic pathway for thyroxine, which involves the detection 

of low concentrations in millimolar range and leading to iodide oxidation to modify 

thyroglobulin (Taurog et al., 1996; Ruf and Carayon, 2006; Le et al., 2015). More 

recently this class of peroxidases have been discovered to also regulate the 

development of the ECM (DeNichilo et al., 2015), initiation of angiogenesis 

(Panagopoulos et al., 2015) and are therefore involved in tissue regeneration (Péterfi 

and Geiszt, 2014; DeNichilo et al., 2015). The addition of PXDN to this family has 

helped to clarify the role of peroxidases in this context. 

  

1.3 Peroxidasin Expression 

1.3.1 PXDN in the cardiovascular system 

PXDN was originally found to be highly expressed in the vasculature, including 

cardiomyocytes, and due to this distribution, the protein was termed VPO1 (Cheng 

et al., 2008; Zhang et al., 2012; Ma et al., 2013). Cardiovascular cells express PXDN 

and secrete the protein into the blood stream (Cheng et al., 2011). The normal 

physiological functions of PXDN circulating in plasma are largely unknown. In 

general, oxidative reactions involve catalysis of H2O2 as a substrate to produce 

HOCl, with the dissociation of H2O2 which generates water (Battistuzzi et al., 2010). 

PXDN contains essential features that are not found in MPO and might aid in its 

ability to confine to exact vascular sites where it performs biological functions such 

as ECM modification (Cheng et al., 2011). Whether animal peroxidases assist in 

cross linking of dityrosine molecules during the development of the ECM in the 

vascular tissue is largely unknown, while the data presented by Cheng et al., (2011) 

make evident the capability of PXDN to promote such modifications, which was 



6 
 

further confirmed by Bathish et al., (2020) who showed that PXDN mediates 

bromination of tyrosine residues in ECM. Dityrosine cross links result when tyrosine 

molecules are oxidised therefore producing tyrosine radicals which when joined 

together form dityrosine (Cheng et al., 2011). 

1.3.2 PXDN in the eye 

PXDN is believed to contribute in providing support of the lens and cornea 

framework during their construction in the early stages of eye development (Khan et 

al., 2011; Choi et al., 2015). In-situ hybridisation found that PXDN is expressed in 

the eye forming region and in particular the lens of Xenopus tropicalis (Tindall et 

al., 2005). In mice, expression was observed in the lens and retina and during 

development of eye of the mouse, PXDN expression was detected from embryonic 

age (E11.5) onwards which is equivalent to 33 days in human embryology, and 

furthermore PXDN knock-out animals showed absence of collagen IV crosslinking 

(Homma et al., 2009; Lázár et al., 2015). As such, PXDN is understood to be 

essential for normal growth of eye structures such as the lens epithelium and cornea 

via collagen IV crosslinking, which promotes consolidation of ocular BM (Yan et 

al., 2014).  

Evidence for a role of PXDN in human eye development has been found in pathology 

context with various genetic variants in PXDN being direct cause of various 

disorders affecting eye development; currently there are only three papers (Khan et 

al., 2011; Choi et al., 2015; Michel et al., 2016). Aberrant PXDN expression causes 

loss of structural integrity of the lens capsule and lens molecules such as g-crystallin 

seem to be extruded into the anterior and posterior chamber of the eye (Yan et al., 

2014). These results presented by Yan et al. (2014) confirmed that indeed PXDN 

plays a pivotal role in consolidation of the BM and that PXDN is required for normal 

BM integrity along with adhesion of cells during embryonic eye development in 

higher organisms. Studies that have investigated PXDN gene mutations in 

association with ocular phenotypes in humans are discussed further in section 1.6.5. 

 

 



7 
 

1.4 Physiological Functions of Peroxidasin 

1.4.1 Extracellular Matrix Formation in Basement Membrane  

The extracellular matrix (ECM) is a non‐ cellular, organised network of proteins that 

provides a structural framework for cells. This framework provides for both 

chemical and mechanical signals between cells. Surface cell receptors such as 

integrins detect these signals and relay them intracellularly (Hynes, 2009). This 

relationship between the ECM and cells enables the ECM to modify behaviour of 

cells and for cells to modify their surrounding ECM (Hynes, 2009; Bonnans et al., 

2014). Biological processes such as tissue development, response to injury and 

maintenance are critically dependent on the interaction of cells with the ECM 

(Rosario and DeSimone, 2010; Clause and Baker, 2013). ECM composition varies 

in specific tissues and the composition changes during and after development and 

upon tissue injury (Frantz et al., 2010; Rosario and DeSimone, 2010).  

The ECM is divided into two subtypes namely, basement membrane (BM) and 

interstitial matrix (Frantz et al., 2010; Bonnans et al., 2014). The BM and interstitial 

matrix form separate but adjacent structures in vivo (Laurila and Leivo, 1993). The 

structural support and protection against compression is provided by the interstitial 

matrix, which surrounds mesenchymal cells (Yurchenco, 2011). The BM is a 

specialised sheet‐ like network arrangement of the ECM (Koshnoodi et al., 2008) 

and it underlies cells of the muscle, epithelial and endothelial tissues, and separates 

them from underlying stroma (Figure 1.3) (Yurchenco, 2011). The proliferation, 

differentiation, migration and polarity of primary cells in almost all tissues is 

regulated by the BM (Yurchenco, 2011). Consequently, the BM is essential for the 

maintenance and development of tissues and plays a critical role in normal 

physiology and pathogenesis of disease. 



8 
 

 
Figure 1.3: Two subtypes of ECM. The diagram depicts two main types of ECM. 

The BM underlies cells of various tissues and comprises collagen IV, nidogen, 

laminin, collagen XVIII and perlecan. The interstitial matrix which consists of 

elastin, proteoglycans, collagen I and glycosaminoglycans, surrounds cells (Colon et 

al., 2017).  

 

The core BM proteins are collagen IV, laminins, proteoglycans, and nidogens 

(Yurchenco, 2011). Recently, PXDN has been found to be responsible for 

consolidating collagen IV (Bhave et al., 2012). Specifically, PXDN generates 

hypohalous acids to cross‐ link collagen IV protomers (Figure 1.4) (Cheng et al., 

2011; Ma et al., 2013), which are key to BM formation and maintaining the integrity 

of tissues during formation and development (Bhave et al., 2012; Péterfi and Geiszt, 

2014; Ero-Tolliver et al., 2015). The mechanism of how PXDN cross-links collagen 

IV promoters is discussed in detail in the section below. 



9 
 

 
Figure 1.4: Crosslinking of collagen IV is mediated by PXDN which utilises H2O2 

and a halide ion to produce a sulfilimine (S = N) bond (Modified from Péterfi and 

Geiszt, 2014). 

 

1.4.2 Collagen IV and Reaction Mechanism of PXDN 

The most predominant proteins in animals making up approximately 30% of total 

protein are collagens. Collagens are mainly characterised by the presence of a triple 

helical structure, a collagenous domain, and an amino acid repeat structure in which 

glycine residues occur every third residue (Shoulders and Raines, 2009; Ricard-

Blum, 2011). Collagen IV is the predominant collagen found in BM and is comprised 

of six alpha chains that form heterotrimeric triple helical molecules. There are three 

distinct domains on each alpha chain, namely a C‐ terminal non‐ collagenous 

domain (NC1), a triple helical collagenous domain and an N‐ terminal 7S domain 

(Söder and Pöschl, 2004). Collagen IV also contains protomers, which exist in 

mammals only [α1α1α2 (α112), α3α4α5 (α345) and α5α5α6 (α556)], which further 

assemble into three major networks (α112:α112, α112:α556, α345:α345) (Borza et 

al., 2001). In nearly all mammalian tissues the α112 network is the most predominant 

(Koshnoodi et al., 2008; Yurchenco, 2011). Conversely, the α345 and α1256 

networks have restricted tissue localization, for example, the α345 network is 

restricted to the glomerular BM (GBM) of kidney, the lens epithelial BM, cochlear 

BM, and in small quantities in lung alveolar and testicular BM (Gunwar et al., 1991; 

Kahsai et al., 1997; Kalluri et al., 1998, Abrahamson et al., 2009; Saito et al., 2011).   

https://pubmed.ncbi.nlm.nih.gov/?term=P%C3%B6schl+E&cauthor_id=15522229


10 
 

In order to form the BM, the collagen protomers self‐ oligomerise to form three 

sheet‐ like networks (Koshnoodi et al., 2008). Oligomerisation enables the networks 

to assemble and these are further stabilised by sulfilimine chemical bonds (S = N). 

To date, the sulfilimine bond is considered to be exclusive to collagen IV molecules 

in living organisms (Vanacore et al., 2009). Methionine at position 93 (Met93) and 

hydroxylysine at position 211 (Hyl211) are joined covalently and stabilise the edges 

of adjacent NC1 in the protomers of collagen IV (Risteli et al., 1980; Bhave et al., 

2012; Ronsein et al., 2014; Anazco et al., 2016). The formation of these crosslinks 

in collagen IV has recently been assigned to PXDN (Bhave et al., 2012). 

PXDN catalyses the crosslinking of collagen IV using hypohalous acids. Firstly, 

hypohalous acids (hypoiodous acid (HOI), hypobromous acid (HOBr), hypochlorous 

acid (HOCl), and hypothiocyanous acid (HOSCN)) are produced by haem 

peroxidases in the presence of H2O2, in this case PXDN, via oxidation of halide ions 

such as bromide (Br-), thiocyanate (SCN-) and chloride (Cl-) (Figure 1.5) (Ronsein 

et al., 2007; Bhave et al., 2012; Colon et al., 2017). The halide from the hypochlorous 

acid is donated to the sulfur atom from methionine producing a halosulfonium cation 

that in turn links with the lysine amino group of the subsequent collagen IV (; 

Vanacore et al., 2009; Bhave et al., 2012; Ronsein et al., 2014). The source of 

extracellular H2O2 is currently unknown, but NOX, which are membrane proteins 

have been suggested to be the source of H2O2 in bronchial epithelial cells (Geiszt et 

al., 2003). NOX consist of NOX (1 to 5) and dual oxidase (DUOX) 1 to 2: NOX1, 

NOX2, NOX3 and NOX5 have the ability to reduce extracellular oxygen (O2) to 

superoxide (O2
-) and NOX4, DUOX1 and DUOX2 to hydrogen peroxide, while 

oxidizing NADPH intracellularly (Lamberth and Neish, 2014). Superoxide, 

spontaneously or via superoxide dismutase, forms the hydrogen peroxide required 

for the reaction (Fukai and Ushio-Fukai, 2011). The role of PXDN in collagen IV 

and BM consolidation is still undergoing investigation with a recent study by 

Sirokmány et al. (2018), finding that PXDN catalysed collagen IV synthesis may be 

independent of H2O2 generated by the NOX/DUOX system. This was done by testing 

crosslinks of collagen IV in mouse models deficient of the NOX/DUOX isoforms, 

and the results were in support of the view that other ECM proteins might act as the 

source of H2O2 (Sirokmány et al., 2018). Lévigne et al. (2016) also showed that mice 



11 
 

with wounds and deficient in NOX4 healed significantly slower as compared to mice 

with wounds and active NOX4, this was thought to be because of insufficient 

dityrosine cross-links in the ECM (Lévigne et al., 2016). 

 

Figure 1.5: PXDN catalysis of HOBr to produce sulfilimine (S=N) cross‐ links 

in collagen IV. PXDN is secreted into the ECM by activated fibroblasts, while 

activated infiltrating leukocytes release MPO and EPO. The NADPH oxidases 

generate superoxide (O2
.‐ ), which dismutates into H2O2, and along with Br‐  and Cl‐  

halides or SCN‐  anions, lead to the generation of hypohalous acids by PXDN during 

sulfilimine bond formation in the BM (Colon et al., 2017). 



12 
 

1.4.3 Role in Host Immune Defence 

Phagocytes such as the neutrophils which have the ability to move from blood into 

various organs, provide the key mechanism for upkeep of a sterile environment in 

tissues by employing both oxidative and non-oxidative mechanisms to kill invading 

microbial agents (Rada and Leto, 2008). The primary mechanism usually involves 

the NADPH oxidase system by using NOX2 (Rada and Leto, 2008). For example, 

MPO, by combining H2O2 and Cl-, synthesises HOCl; this reaction was previously 

thought to be exclusive to MPO and EPO but is now also attributed to PXDN 

(Klebanoff et al., 2005; Li et al., 2012). LPO, MPO and EPO also predominantly 

mediate host defence functions but are restricted to specific cell types, for instance 

in the body fluids saliva and breast milk this is performed by LPO and in phagocytic 

cells this is carried out by MPO and EPO (Klebanoff et al., 2005). MPO is usually 

retained inside phagocytes, and PXDN is secreted into plasma by vascular 

endothelial cells (Cheng et al., 2011), where it is understood to demonstrate a role in 

host defence (Li et al., 2012). It was initially thought that since PXDN lacks a 

methionine (Met) at position 981 (corresponding to position Met409 in MPO), it was 

thought to render the catalysis of HOCl by PXDN impossible, since PXDN has 

glutamine (Gln981) at that position (Li et al., 2012). The methionine at position 409 

is a distinctive feature of MPO that has been suggested to be a main factor of its 

unique capability to produce HOCl (Klebanoff et al., 2005). However, Li et al. 

(2012), showed that PXDN like MPO, can synthesise HOCl and further hypothesised 

that PXDN may be localised at infected and injured tissues. HOCl is a very potent 

microbicidal agent that can kill microorganisms, for example, the incubation of 

Escherichia coli (E. coli) bacteria with PXDN supplemented with H2O2 and Cl- 

halide completely killed the E. coli via HOCl synthesis (Cheng et al., 2011; Li et al., 

2012). As such, the findings by Li et al. (2012) propose that methionine at position 

409 of MPO is not compulsory but its presence can augment the synthesis of HOCl 

by peroxidases. 

  



13 
 

1.5 Peroxidasin in Pathophysiological conditions 

PXDN with its various functions has been implicated in various diseases, most 

notably those affecting the cardiovascular system (Cheng et al., 2008; Zhang et al., 

2012; Ma et al., 2013), cancer (Desmond et al., 2007; Tauber et al., 2010; Liu et al., 

2010; Jayachandran et al., 2016; Sitole and Mavri-Damelin, 2018), fibrosis (Peterfi 

et al 2009), and the eye (Khan et al., 2011, Choi et al., 2015; Micheal et al., 2016). 

1.5.1 Cardiovascular Disorders 

Oxidative stress refers to an imbalance between free radicals and antioxidants 

resulting from oxidation within cells, tissues or organs and is caused by increased 

reactive oxygen species (ROS) (Lu et al., 2011). This can cause pathophysiological 

conditions such as atherosclerosis (Lu et al., 2011), heart failure and heart attack 

(Strobel et al., 2011). PXDN expression may be altered under conditions of oxidative 

stress and may therefore have a role to play in these conditions (Ma et al., 2013). 

The HOCl derived from reactions catalysed by PXDN has been implicated in injury 

of the cardiovascular system since it is a stronger oxidant than to H2O2 and O2
- (Goud 

et al., 2008). HOCl has been found to induce direct oxidative damage to lipids, 

deoxyribonucleic acid (DNA) and proteins (Kang and Neidigh, 2008) which can lead 

to apoptosis. Oxidation of lipids by HOCl has been suggested as a possible cause of 

atherosclerosis, since it produces chlorinated products which may form plaques 

(Ford, 2010; Yang et al., 2013). PXDN generated HOCl has been found to also 

activate apoptosis in cardiac and endothelial cells (Bai et al., 2011).  Bai et al. (2011) 

hypothesised that the activation of apoptosis in endothelial cells might be induced 

by oxidised low-density lipoproteins (ox-LDL), which is the key constituent of 

serum lipids and is a critical risk factor for atherosclerosis and endothelial 

dysfunction (Kita et al., 2001). Ox-LDL further upregulates the expression of 

NADPH oxidase, which results in elevated ROS, PXDN and HOCl, thereby forming 

a positive loop between PXDN (which forms HOCl) and NADPH oxidase (to form 

ROS). HOCl production was inhibited when PXDN was silenced. This loop is 

thought to activate the p38 MAPK/caspace-3-dependent signaling pathway, which 

in turn induces ox-LDL endothelial cell apoptosis (Bai et al., 2011). 



14 
 

 

Other studies using rat model/tissue also provided evidence showing that PXDN 

mediates oxidative stress causing oxidative injury (Shi et al., 2011; Zhang et al., 

2012; Li et al., 2012). In a rat model of myocardial ischemia reperfusion (IR) injury, 

it was found that PXDN expression was up regulated and was accompanied by 

oxidative injury. When NOX activity was inhibited, PXDN expression was 

suppressed thereby attenuating the IR injury.  Shi et al. (2011) found that arterial 

tissue of spontaneously hypertensive rats had elevated PXDN, which was 

accompanied by vascular remodelling. In a case-control study by Zhuan et al. (2017) 

patients diagnosed with chronic obstructive pulmonary disease (COPD) combined 

with Pulmonary hypertension (PH) had significantly high levels of plasma PXDN, 

NADPH oxidase, ROS activity and NF-κB as compared to the controls, thus 

suggesting that PXDN and the NADPH oxidase/ROS/NF-κB pathway may be 

actively involved in the pathogenesis of COPD combined with PH via the resulting 

oxidative stress (Zhuan et al., 2017). 

1.5.2 Cancer 

PXDN displays irregular expression patterns in different cancers affecting different 

tissues such as melanoma, colon, breast, ovarian cancers along with metastatic 

gliomas as well as renal carcinoma, (Mitchell, 2000; Desmond et al., 2007; Tauber 

et al., 2010; Liu et al., 2010; Jayachandran et al., 2016; Sitole and Mavri-Damelin, 

2018; Paumann-Page et al., 2021;). The precise function of PXDN in cancer remains 

unclear but has been associated with changes to cell attachment, migration and 

invasion of cancer cells. Two studies characterised PXDN as being involved in 

melanoma development (Paumann-Page et al., 2021), but one did not study the 

normal physiological function of PXDN (Mitchel et al., 2000). In another study by 

Tauber et al. (2010), gene expression profiling technique was employed with an aim 

to pinpoint a group of genes associated with haem oxygenase 1 (HO-1) in 14 

different types of cancer. PXDN was found to be a key player in adhesion of 

neoplastic cells, which are dependent on HO-1, and PXDN silencing stopped the 

increase in adhesion in response to HO-1 in various cell lines, whilst cells deficient 

of HO-1 and reduced PXDN expression did not affect cell adhesion. 



15 
 

Madden et al. (2004) previously showed a 17-fold increase in PXDN expression in 

glioma endothelial cells as compared to non-neoplastic endothelial brain cells. 

Subsequently, PXDN was found to be selectively upregulated in a set of endothelial 

marker genes in the microvasculature of metastatic brain tumours glioblastoma and 

pilocytic astrocytomas, while it was absent or detected at low quantity in the controls 

(non-neoplastic temporal lobe) specimens (Liu et al., 2010). PXDN protein was 

localised in microvascular endothelial cells, which was also reported by Castronovo 

et al. (2006) when it was found that PXDN localised in vascular structures in kidney 

cancer, thus it was hypothesised that PXDN may be playing a role in tumour 

angiogenesis in various cancers. One other major feature of cancer is the ability of 

cells to undergo epithelial-to-mesenchymal transition (EMT), a process that enables 

cancer cells to typically acquire invasive ability (Kalluri and Weinberg, 2009). As a 

result of EMT, immobile cells of the epithelial phenotype attain a mesenchymal 

phenotype as well as the capability to invade and migrate during tumour metastasis 

and developmental morphogenesis (Nieto, 2013; Lamouille et al., 2014). 

Jayachandran et al. (2016) investigated genes that regulate cellular invasion with the 

aim of inhibiting their effect in melanoma invasion. In mesenchymal like melanoma 

cells PXDN was found to be upregulated and silencing of PXDN reduced melanoma 

invasion in vitro. As to how PXDN exerts its effect in the development of cancer is 

not known and the studies discussed above shows that the role of PXDN may differ 

per cancer type. Studies in our laboratory have found that Snai1 which is a key player 

in EMT modification regulates PXDN in cervical carcinoma cell lines (Sitole and 

Mavri-Damelin, 2018).  

1.5.3 Kidney Disorders 

Almost every organ can undergo fibrosis and it is predominantly observed in 

diseases affecting the liver, heart, kidney and lung tissues (Wynn, 2007). It has been 

observed that PXDN is also expressed in dermal fibroblasts (Peterfi et al., 2009). 

Elevated expression of PXDN was found in myofibroblasts that were undergoing 

differentiation induced by TFG-β1. The fibrillar like structure forming the ECM was 

formed by PXDN secreted from myofibroblasts (Peterfi et al., 2009). In a normal 

healthy kidney PXDN was undetectable, while in a mouse model with kidney 



16 
 

fibrosis PXDN expression was found to be elevated and the expression paralleled 

fibrotic kidney remodelling (Peterfi et al., 2009).  

Goodpasture (GP) is an autoimmune disease which affect the kidneys when 

circulating autoantibodies bind to regions within the NC1 of collagen IV which is 

required for BM synthesis under normal physiological functions (Pedchenko et al., 

2010). The formation of sulfilimine bonds which crosslinks collagen IV molecules 

in glomerular basement membrane (GBM) confers protection against GP given that 

the binding of circulating autoantibodies to NC1 domain of collagen IV prevents 

normal formation of GBM (Pedchenko et al., 2010). A study by McCall et al., 2018 

showed that approximately 46% of patients diagnosed with GP disease show reduced 

HOBr syntheses due to autoantibodies against PXDN. Alteration of collagen IV 

crosslinking has been proposed as key pathogenic mechanism for GP via anti-PXDN 

autoantibodies, and this was based on in vitro data that showed autoantibodies had 

the ability to inhibit HOBr syntheses (McCall et al., 2018). The resulting reduced 

crosslinking of collagen IV in GBM is vital for the pathogenesis of GP disease 

(McCall et al., 2018). PXDN has also been discovered to be a novel target of 

autoantibodies for Lupus nephritis (Manral et al., 2019).  

1.5.4 Eye Disorders 

Various mutations in PXDN have been found to be present in patients diagnosed with 

developmental eye disorders, specifically anterior segment dysgenesis (ASD), 

lending support to the significance of PXDN in eye development (Khan et al., 2011; 

Choi et al., 2015). Whilst the precise functions of PXDN in the eye are as yet 

unknown, PXDN is understood to be involved in supporting the cornea as well as 

lens framework during development of the eye as well as acting as an antioxidant 

enzyme to protect various eye structures from oxidative damage (Khan et al., 2011; 

Yan et al., 2014; Choi et al., 2015).  The protein seems to be involved in number of 

pathways, yet the exact route through which it exerts its function is not well 

understood especially with regards to development of the eye. Support from an 

animal model showed that a mutant KTA048 mouse model Yan et al. (2014), 

displayed similar phenotypes observed in individuals harbouring mutations in 

PXDN. The mutant mouse model was induced via N-ethyl-N-nitrosourea of which 



17 
 

upon sequencing revealed a recessive mutation (c.T3816A) causing a premature stop 

codon in the peroxidase domain, which resulted in development of ASD as well as 

microphthalmia (Yan et al., 2014). No precise mechanisms are known currently as 

to how mutations discovered in PXDN cause ASD disorders, which include 

glaucoma, abnormal lens and cornea development. Few possible mechanisms have 

been suggested: (1) the lack of PXDN protein in the cornea as well as in the epithelial 

layers of the lens causes accumulation of ROS that generally oxidise lipids also 

proteins, producing aggregates that are insoluble leading to cataractogenesis and 

corneal clouding and (2) the accumulation of ROS in the aqueous humour can affect 

the differentiation also development of the structures found in the anterior segment 

which is involved in the development of glaucoma (Izzotti et al., 2006, Khan et al., 

2011). For instance, aqueous humour of glaucoma patients contains elevated 

enzymes induced by oxidative stress and reduced antioxidants which promotes 

oxidative stress (Ferreira et al., 2004) and the trabecular meshwork is very sensitive 

to damage by oxidation (Khan et al., 2011). The specific mutations detected in 

PXDN are discussed in section 1.6.5 that follows along with other genes that may 

interact with PXDN to bring about various ocular phenotypes.  

 

The primary focus of this study is to understand the regulation of PXDN in eye 

development. Formation of the human eye is discussed in detail in the sections that 

follow to create a link between successful eye development and PXDN requirement. 

 

1.6 Anatomy of the human eye 

1.6.1 The anterior segment of the eye 

The human eye is a unique and complex part of the human body which is broadly 

divided into two segments (Figure 1.6) namely the anterior and posterior segments 

(Addo et al., 2016). The anterior chamber and posterior chamber are filled with 

aqueous humour which provides nutrients to tissues in the anterior segment of the 

eye. The anterior segment of the eye comprises of trabecular meshwork, cornea, lens, 

iris, ciliary body and Schlemm’s canal while the retina and optic nerve form the 

posterior segment of the eye (Ito and Walter, 2014). In order for the eye to form a 



18 
 

clear image, different structures within the eye contribute to the process. The cornea 

as a curved membrane which encompasses the front section of the eye allows light 

to pass through which leads to the light rays bending and focusing as they pass 

through the iris on the way to the lens. The dilator muscles and pupillary sphincter 

generate a synchronised action that helps the pupil to constrict and dilate in order to 

regulate the amount of incoming light. The lens with the aid of ciliary muscles 

focuses the light which will travel through the vitreous humour then straight to the 

retina (Kaplan, 2007). 

 
Figure 1.6: Anatomical structure of a fully developed human eye. In this study only 

the disorders affecting the anterior segment structures such as iris, cornea, lens, 

ciliary body and Schlemm’s canal were discussed (Addo et al., 2016). 

 

In humans, the eye starts developing within 21 days of gestation through a highly 

organised process, which is tightly regulated by various transcription factors and 

multiple cell signalling pathways which can lead to development of various eye 

disorders if perturbed (Dash et al., 2016; Budak et al., 2018). 

 

 



19 
 

1.6.2 Anterior segment morphogenesis 

Starting from the sixth week of human development, morphogenesis of various eye 

structures is initiated, such as the embryonic optic cup (bi-layered) from 

neuroectoderm of the forebrain and the lens vesicle is invaginated thus separating 

from the ectoderm (Cvekl and Tamm, 2004; Sowden, 2007). Figure 1.7a show a 

rudimentary diagram of the eye which is surrounded by progenitor cells 

(mesenchyme) most of which originate from the neural crest migrating anteriorly 

(Sowden, 2007). Tissue differentiation and morphogenesis driven by the progenitor 

cells along with those from the surface ectoderm and optic cup peripheral region will 

give rise to the cornea, iris and drainage structures of the irido-cornea (Cvekl and 

Tamm, 2004). The normal development of the anterior segment of the eye is 

dependent on the correct differentiation and specification of the progenitor cells 

(mesenchymal) in early embryogenesis (Gage et al., 2005). Figure 1.7b shows that 

the cornea is formed first from the primitive endothelium, then the trabecular 

meshwork which is situated posterior to the surface ectoderm (Sowden, 2007).  



20 
 

 
Figure 1.7.: Anterior segment development of the human (foetal) eye. (a) 5th 

week of embryogenesis, the optic cup stage (b) 5th month of development, formation 

of the anterior chamber (c) fully developed anterior segment showing all structures 

including the iris, lens, cornea. Key colours indicate two important genes expressed 

during anterior segment development and origin of cells that form the anterior 

segment tissues (Sowden, 2007). 

 

 

Collagen, which is usually in a lamellar arrangement, is synthesised when the corneal 

(epithelium and endothelium) cells form corneal stroma (Beebe and Coats, 2000). In 

the 5th month of gestation, the cornea, iris and the anterior chamber will be well 

developed. Further development, as well as maturation of the tissues, involves more 

differentiation where the sclera spur separates the iris root and ciliary body from the 

trabecular meshwork (Gage et al., 2005; Sowden, 2007). The trabecular meshwork 

will be situated anterior to the iris root as well as visible to the aqueous humour as 

shown in Figure 1.7c (Sowden, 2007). 



21 
 

1.6.2.1 Optic cup morphogenesis 

 

Figure 1.8: Morphogenesis of the optic cup and involvement of the ECM 

proteins through the stages of evagination, lens placode formation and 

invagination. The ECM components collagen IV (purple), fibronectin (orange) and 

laminin (blue) when labelled in red colour it means the component is required for 

that specific stage and when in grey colour it is not required for that specific stage 

of eye development (Kwan, 2014). 

Optic cup formation initiates establishment of the basic structures of the eye which 

are derived from various origins of embryonic nature such as mesenchyme, surface 

ectoderm and neuroectoderm (Graw, 2003). Optic vesicle evagination from either 

side of the brain neuroepithelium signals the commencement of optic cup 

development with formation of single eye fields in the anterior neural plate median 

region (Figure 1.8). Adjacent to the surface ectoderm, the single eye field will then 

move laterally, separate and form the optic vesicle (Yang, 2004). The optic cup is 

generated by this out-pocketing of tissue which undergoes a series of cell and tissue 

rearrangements which further leads to formation of retinal pigmented epithelium, 

neural retina and lens (Chow and Lang, 2001; Fuhrmann, 2010). The ECM 

components collagen IV, laminin and fibronectin are present from evagination, lens 

placode formation and invagination stages of eye development with laminin and 

fibronectin required in the first early stages and all three required for the last stages 

including collagen IV (Figure 1.8) (Kwan, 2014). 

  



22 
 

1.6.2.2 Lens morphogenesis 

At day 33 of gestation post optic cup formation, formation of the lens placode will 

be the next step during human eye development (Graw, 2003). A region of contact 

with the overlying ectoderm is established with the developing optic vesicle 

following evagination, lens placode development will begin. The ECM is found in 

abundance in the region between lens placode and prospective retina separating the 

tissues (Kwan, 2014). The lens will be formed by the mesenchymal cells as they also 

differentiate to form melanocytes and fibroblasts of the iris stroma (Gage et al., 

2005). The iris epithelium is formed when cells within the two layers of peripheral 

optic cup spread inwardly after proliferation amid the iris stroma as well as the lens 

(Beebe and Coats, 2000). The lens placode composed of lens progenitor cells will 

invaginate to form the lens cup which will close up and form the lens vesicle by 

separating from the surface ectoderm. The lens vesicle and surface ectoderm will be 

connected by a temporary link from a lens stalk which will be retracted later on with 

the surface ectoderm giving rise to epithelium (Figure 1.7b) that will form the 

cornea at later stage (Graw, 2003; Snowden, 2007).  

A spherical shape with an empty cavity is representative of a lens vesicle, with the 

cuboidal lens epithelium formed from the anterior cells while elongation of the 

posterior epithelial cells via filling of the lumen will give rise to primary lens fibre 

cells which will in turn form embryonic nucleus (Graw, 2010). This process is 

initiated at the 44th day of human eye development, with much of the lens formation 

driven by the posterior epithelial cells in the first two months of lens formation. The 

secondary lens fibre cells are formed from the actively dividing anterior progenitor 

cells through migration from the central to the equatorial region, proliferation and 

differentiation which will be displaced inwards flanked by embryonic nucleus and 

the capsule (Cvekl and Ashery-Padan 2014). As humans continue to grow, the 

epithelial cells (anterior) mitotically continue to divide to form secondary lens fibre 

cells and only stop when they are displaced into the inner layer since by then they 

will be mature and start losing their differentiation ability as well as their organelles 

(Graw, 2010). Lens fibres at the periphery are therefore newly formed young cells, 



23 
 

as the lens fibres are successively displaced to the inner cortex (Cvekl and Ashery-

Padan 2014). 

1.6.2.3 Cornea morphogenesis 

The epithelium (anterior and posterior) forms the components of the cornea along 

with the corneal stroma in between them. The optic vesicle along with lens vesicle 

induce the formation of the cornea via the continued proliferation of the surface 

ectoderm to form the corneal epithelium even after the lens vesicle has completely 

separated from the surface ectoderm. The corneal stroma formation results from 

production of glycosaminoglycans and collagen fibres by corneal epithelium in the 

basal lamina (Graw, 2010). 

 

Figure 1.9: Diagram illustration of development of the cornea during embryogenesis 

(Graw, 2010). 

The space between the lens vesicle and surface ectoderm will be filled by 

mesenchymal cells of the neural crest origin in three waves (Figure 1.9) (Graw, 

2010). The first wave gives rise to corneal endothelium, while the second wave of 

cells condense and contribute to various corneal structures and the third wave form 

keratocytes in the stroma via migration between epithelium and endothelium (Cvekl 

and Tamm 2004). The secondary corneal stroma is formed from the organisation of 

type 1 collagen and proteoglycans (synthesised by keratocytes) as lamellae while the 

basal membrane derives the formation of Descemet’s membrane (Cvekl and Tamm 

2004). 

  



24 
 

1.6.2.4 Iris and Ciliary body morphogenesis  

The neuroectoderm which comprises of two layers, the outer pigmented and inner 

non pigmented epitheliums is formed when the optical cup distal tips proliferate and 

extend in parallel to the formation of the lens as well as the cornea which leads to 

formation of iris and ciliary body (Graw, 2010). Anterior to the iris is the iris stroma 

and the iris muscles, while the posterior layer has the iris pigmented epithelium 

(Figure 1.7c) and the base of the iris is attached to the ciliary body in addition to 

irido-corneal angle (Davis-Silberman and Ashery-Padan 2008).  

The formation of the iris results from derivation of two embryonic origins: Periocular 

mesenchyme which forms iris stroma and neuroectoderm which forms iris sphincter, 

dilator muscles as well as the iris pigmented epithelium. Pupillary membrane is 

formed when the aforementioned second wave of mesenchymal cells condensed, 

while the two distal tips of the optic cup extend to reach the pupillary membrane 

fusing to give rise to the iris (Davis-Silberman and Ashery-Padan 2008; Graw 2010). 

Myofibrils, which result from differentiating iris pigmented epithelium, form the iris 

sphincter and dilator muscles; these regulate (in an antagonist manner) the 

constriction and dilation of the pupil in response to light (Graw, 2003). The iris 

stroma is composed of melanocytes as well as fibroblast derived from periocular 

mesenchymal cells. When migrating to the anterior section of the lens while in 

peripheral part of the pupillary membrane these cells secrete collagen fibrils in ECM 

which aids in formation of the anterior stroma of the iris. Different melanin quantity 

in the iris stroma results in various colours of the iris in different individuals (Davis-

Silberman and Ashery-Padan 2008). 

The proximal non pigmented layers result in the formation of ciliary body when the 

distal tips of the optic cup neuroectoderm form the iris. The ciliary body muscles are 

formed from proliferating periocular mesenchyme myofilament, while the stroma of 

the ciliary body is formed when the periocular mesenchyme condenses at the anterior 

section of the ciliary body (Graw, 2003). The pupil develops from the degeneration 

of the pupillary membrane when gestation reaches the 8th month, while the ciliary 



25 
 

body, iris stroma and dilator muscle continue to develop even after birth since they 

would have been immature after birth (Davis-Silberman and Ashery-Padan 2008). 

1.6.2.5 Anterior chamber morphogenesis 

During the 3rd month of gestation between the iris epithelium and corneal 

endothelium a presumptive anterior chamber can be observed (Graw, 2010). Within 

the anterior chamber the mesenchyme forms the trabecular meshwork while the 

venous canaliculi form the Schlemm’s canal from the small plexus (Sampaolesi et 

al., 2009). The ciliary body muscle continues to develop even after 7 months of 

gestation which enables the extension of Schlemm’s canal in the direction of the 

anterior chamber while the anterior chamber angle continues to recede as well as 

widen in a triangular shape. The anterior chamber continues to further deepen as well 

as surpasses the spur even after birth and the anterior chamber angle also continues 

to develop to maturity a year after birth (Sampaolesi et al., 2009). 

1.6.3 Regulatory networks involved in the anterior segment of the eye. 

The regulatory programmes that control eye development are complex and involves 

a myriad of both transcriptional and post transcriptional events (Dash et al., 2016; 

Budak et al., 2018). Master control genes of eye development refers to the genes that 

drive eye development at the early stages, and most of which code for transcription 

factors and signalling molecules (Graw, 2010). Transcription factors such as Sry-

box 2 (SOX2), sine oculis homeobox homolog 3 (SIX3), retina and anterior neural 

fold homeobox (RAX), sonic hedgehog (SHH) and Paired box 6 protein (PAX6) are 

at the top of the hierarchy of master control genes for eye development since they 

are responsible for the formation of the optic vesicle through patterning and splitting 

of the single eye field (Graw, 2003). PAX6, Paired-like homeodomain 2 (PITX2) 

and Forkhead box C1 (FOXC1) are among the key role players in the development 

of the anterior segment by directing the proper differentiation of ocular mesenchyme 

(Baulmann et al., 2002; Graw, 2003).  

 

PAX6 like many other transcription factors induces embryonic differentiation along 

major body axes in response to the concentration gradient of other regulatory 



26 
 

proteins or transcription factors, which bind to DNA sequences of other genes and 

regulate their expression (Friedman, 1998). PAX6 belongs to the PAX multigene 

family of transcription factors that aid in the regulation of embryonic differentiation 

(Friedman, 1998). PAX6 is termed a master control gene since mutations in the 

highly conserved transcription factor interrupts development of the eye in both 

insects and mammals (Hill et al., 1991; Quiring et al., 1994). The study of the 

transcription factor PAX6 has improved the understanding of how ocular tissues 

develop and mis-expression of PAX6 induces ectopic eyes (Zuber et al., 2003; 

Kenyon et al., 2003; Kozmik, 2005). In vertebrates, this factor is essential for normal 

development of several organs including the brain, pancreas and the eye (Kenyon et 

al., 2003). PAX6 as a transcription factor, leads the cascade of eye development 

pathways and it initiates development of the eye in almost any tissue in which it is 

ectopically expressed (Ashery-Padan and Gruss, 2001). Lens differentiation and 

proliferation is dependent on PAX6 expression, while aberrant expression causes 

disruption of lens formation (Yan et al., 2014). PAX6 is expressed from the early 

stages of eye development, on the surface of neuroectoderms, and then expressed in 

the differentiating cells in the retina, lens, ciliary body and cornea throughout 

development (Nishina et al., 1999; Chanas et al., 2009). Expression of PAX6 is very 

precise within the different tissues, therefore normal eye development requires the 

correct dosage of PAX6. Reduced expression of PAX6, due to mutations, results in 

clinical phenotypes such as small eye in mice and rats (Kroeber et al., 2010), whereas 

in humans it induces Aniridia (Yokoi et al., 2016; Lim et al., 2017). Overexpression 

in mice was found to also cause severe eye phenotypes (Schedl et al., 1996; Chanas 

et al., 2009). The functional studies conducted on PAX6 and the observed expression 

pattern which is conserved have implicated this protein in various pathways essential 

for normal eye development (Ashery-Padan and Gruss, 2001).  

 

The transcription factor SIX3 has been found to be a regulator of PAX6 in the 

development of the lens (Liu et al., 2006). SIX3 is present during early eye 

development in the optic stalk, in the presumptive retina and in the lens vesicle and 

later also in the lens (Oliver et al., 1995). In SIX3 mutant embryos it was found that 

PAX6 was downregulated while SOX2 was absent (Liu et al., 2006). PAX6 was 



27 
 

found to be interacting with SOX2 gene promoter therefore initiating expression of 

the SOX2 protein (Liu et al., 2006; Lengler et al., 2005), while SIX3 was found to 

not interact with the SOX2 gene promoter (Lengler et al., 2005). Meis homeobox 1 

(MEIS1) and homeobox 2 (MEIS2) were found to bind upstream of the PAX6 gene 

promoter and induce expression in vertebrate lens ectoderm (Zhang et al., 2002). 

Both MEIS1 and MEIS2 are expressed throughout development in a similar manner 

like PAX6, but their expression occurs independently of PAX6. In the lens of mice, 

it was found that MEIS2 overexpression caused increased expression of PAX6 

(Zhang et al., 2002). 

 

PITX (Paired-like homeodomain), MAF (bZIP transcription factor) and FOX 

(Forkhead box) transcription factor genes act downstream of the master control 

genes such as PAX6 in specific regions of the anterior segment of the eye such as the 

cornea, lens and iris (Graw, 2003). A wide variety of genes are regulated by the FOX 

class of transcription factors, with FOXC1 and FOXE3 being the most critical 

transcription factors required for development of the anterior segment of the eye 

(Graw, 2010). 

 

FOXC1 is composed of a Forkhead domain which enables DNA-binding and has 

110 amino acid protein which is highly conserved. The transcription factor has a 

single exon that codes for 553 amino acid protein and is located on the short arm of 

chromosome 6 band 25 (6p25). In both various ocular and non-ocular structures 

FOXC1 protein has been observed to be expressed (Erickson, 2001; Mathew et al., 

2002; Walter, 2003; Seo and Kume, 2006). The periocular mesenchyme cells express 

FOXC1 protein which enables the cells to give rise to anterior segment structures 

such as trabecular meshwork, iris and cornea (Wang et al., 2001; Sowden, 2007).   

 

PITX2 encodes a 33 kDa protein which comprises of homeodomain with 60 amino 

acids at the N terminus as well as a 14 amino acid C terminal domain which enables 

protein to protein interaction (Amendt et al., 2000). The involvement of PITX2 

protein in regulation of cellular differentiation in optic nerves, heart, nerves, 

pituitary, midbrain neurons and anterior segment of the eye indicates the various 



28 
 

roles the protein plays during development (Gage et al., 1999; Martin et al., 2004; 

Houssaye et al., 2006). 

 

Figure 1.7 shows that during development of anterior segment of the eye both 

FOXC1 and PITX2 are highly expressed (Sowden, 2007), which is also the case in 

animal models such as the mouse; in mouse eye development PITX2 mRNA is 

expressed in the mesenchyme that surrounds the optic eminence as early as E8.5 in 

preparation for the formation of the lens vesicle and optic cup (Mucchielli et al., 

1997). At the stage of E11.5 after both the optic cup and lens vesicle have developed, 

the migrating periocular mesenchyme highly expresses PITX2 (Gage and Camper, 

1997; Hjalt et al., 2000). Corneal endothelium and trabecular meshwork which are 

mesenchymal derivatives retains the expression of PITX2 and by E18.5 PITX2 

becomes restricted to the angle (Kidson et al., 1999). Like PITX2, FOXC1 

expression in mouse eye development is similar in that it is initiated in the periocular 

mesenchyme by E11.5 which is prior to the commencement of mesenchymal cell 

migration (Hjalt et al., 2000; Wang et al., 2017). FOXC1 continues to be expressed 

in the hyaloid plexus at E12.5, conjunctiva and trabecular meshwork which persist 

until parturition and potentially thereafter and also in the precorneal mesenchyme 

(Hjalt et al., 2000; Sowden, 2007). 

 

A study by Berry et al., 2006 showed that PITX2 and FOXC1 are co-expressed 

during eye development in discreet cell types and that they share a subnuclear 

compartment as well as interact with each other physically. The data further showed 

that both FOXC1 and PITX2 are crucial for the proper development as well as 

function of the anterior segment of the eye as they form a higher order transcription 

factor complex (Berry et al., 2006). Signals which are essential for proper migration 

and condensation into endothelia such as those of trabecular meshwork as well as 

cornea may be initiated by FOXC1 and PITX2 (Berry et al., 2006). Both 

transcription factors are likely to be the ones to initiate elaboration of the anterior 

chamber angle along with other mechanical messages during migration transduced 

by the ECM and cytoskeleton (Berry et al., 2006). 

 



29 
 

The interaction between the genes that regulate the development of the anterior 

segment of the eye is not fully understood especially PAX6, FOXC1, PITX2 

interaction with PXDN. Even though specific functions of PXDN are yet unknown 

in eye development, it is understood that PXDN is involved in supporting the cornea 

as well as lens by syntheses of BM which provides a structural framework and acting 

as an antioxidant enzyme to protect various eye structures from oxidative damage 

(Khan et al., 2011; Yan et al., 2014; Choi et al., 2015). Mutations in the genes that 

facilitate the development of the anterior segment of the eye results in poor 

development of the structures such as iris, cornea, lens, trabecular meshwork etc 

which leads to impaired vision. Eye development disorders resulting from mutations 

and/or aberrant expression of the genes discussed above (FOXC1, PITX2, PAX6 and 

PXDN, with relevance to the current study) are discussed in detail in below. 

1.6.4 Anterior segment dysgenesis 

Anterior segment dysgenesis (ASD) defines a collection of various congenital eye 

disorders that affect structures within the anterior segment of the eye such as the 

cornea, iris, trabecular meshwork, and ciliary body (Sowden, 2007; Reis and Semina, 

2011). ASD is comprised of disorders such as Peter’s anomaly, Axenfeld-Rieger 

Syndrome (ARS), Sclerocornea and microphthalmia, while ARS, Peter’s anomaly 

and sclerocornea can lead to cataract formation, corneal opacity and glaucoma (Reis 

and Semina, 2011).  

ASD congenital disorders are usually accompanied by a 50% increased 

predisposition to glaucoma (Alward, 2000; Sowden, 2007; Ito and Walter, 2014). 

The congenital disorders typically involve modifications to adhesions between the 

lens and cornea or cornea and iris, corneal opacity, iris hypoplasia and posterior 

embryotoxon, these can present themselves separately or in combination (Reis and 

Semina; 2011). The most common disorders associated with ASD are (1) ARS which 

is usually noted by a combination of adhesion of the cornea, iris hypoplasia and 

polycoria (more than one pupillary opening in the iris), (2) Peter’s anomaly which 

denotes the trio of malformations in the layers of the cornea, corneal opacity and 

adhesion of the cornea. (3) Sclerocornea is characterised by the anterior extension of 

scleral tissue to replace the peripheral cornea (Ito and Walter, 2014).  



30 
 

1.6.4.1 Axenfeld Rieger syndrome 

ARS was first described by Axenfield in 1920 in patients presenting with iris strands 

adhesion, iris atrophy, pseudopolycoria and further described a sequence of 

mesodermal dysgenesis of the cornea and iris (Stahl, 2014). The syndrome at times 

it is accompanied by systemic complications such as facial bone defects, dental 

abnormalities, pituitary and umbilical abnormalities (Idrees et al., 2006). ARS has 

varying severity can be from delicate findings (which can be detected using 

gonioscopy) to a more severe malformation (Stahl, 2014). 

1.6.4.2 Peters anomaly and Peters-plus syndrome 

Peters’ anomaly is a developmental abnormality in the anterior section of the eye. In 

1906, Peters described the syndrome as it was characterised by cornea-iris adhesion, 

corneal leukoma and shallow anterior chamber (Harissi-Dagher and Colby, 2008). 

The severity of the disorder can vary from mild opacification of the cornea to severe 

glaucoma, cataract and microphthalmia (Stahl, 2014). Peters-plus syndrome results 

when abnormality of the system such as developmental delay and skeletal changes 

also occur (Yang and Lambert, 2001). The two modes of inheritance (autosomal 

recessive and dominant) have been reported for peters anomaly, which involves 

mutations in PAX6, FOXC1, PITX2 and CYP1B1 genes (Doward et al. 1999; 

Nishimura et al. 2001; Vincent et al. 2001). 

1.6.4.3 Aniridia 

Aniridia (lack of iris), is a rare autosomal dominant disease which is described by 

varying eye abnormalities such as corneal opacity, lens dislocation, ciliary body 

hypoplasia, and iris hypoplasia (Lee et al., 2008). The following can also lead to 

impaired or complete loss of vision, foveal hypoplasia, cataract, glaucoma early 

onset, retinal detachment and aniridia-associated keratopathy. Systemic 

abnormalities do not usually present with aniridia but deletions in genes such as 

FOXC1 and PAX6 enables the presentation of systemic anomalies in aniridia 

patients (Ito et al., 2009). 

 



31 
 

1.6.4.4 Sclerocornea 

Sclerocornea is a congenital anomaly that is characterised by noninflammatory, 

nonprogressive asymmetric scleralisation of the cornea (Kenyon, 1975). In 

Sclerocornea the scleral tissue is extended anteriorly which in turn replaces the 

peripheral cornea leading to the vessels of both the conjunctiva and episclera 

crossing the cornea, this disorder may associate with cataracts and coloboma 

(Harissi-Dagher and Colby, 2008). 

1.6.4.5 Primary congenital glaucoma 

Primary congenital glaucoma (PCG) is part of infantile glaucoma which can lead to 

eye defects and can be associated with ARS (Sarfarazi et al., 2003, Micheal et al., 

2016). At the age of 6 months about 60% of the patients presenting with such 

disorders can be diagnosed and 80% within the first year. PCG accounts for 

approximately 0.04% of blindness and can cause permanent visual impairment 

(Gould and John, 2002). Across various populations the incidence rate of PCG varies 

such that in western population it affects 1 in 10000, 1 in 3300 in Indians, 1 in 2500 

in Saudi Arabians and 1 in 1250 Slovakian gypsies (Bejjani et al., 1998; Plásilová et 

al., 1999; Tanwar et al., 2009). 

Elevated intraocular pressure (IOP), coupled with optic nerve damage and elevated 

corneal diameter are the characteristics of PCG (Tanwar et al., 2009; Kim et al., 

2011). Some of the clinical features of PCG include descemet membrane 

opacification coupled with rupture, buphthalmos, edema of the cornea, anterior 

sclera thinning, iris atrophy, and deep anterior chamber (Sarfarazi et al., 2003; 

Micheal et al., 2016). Some of the less common features include photophobia, 

epiphora, and blepharospasm. GLC3A, GLC3B, GLC3C at 2p21, 1p36.2-1p36.1, 

and 14q24.3 respectively have been associated with PCG pathogenesis. CYP1B1 

along with LTBP2 are the only target genes that have be located thus far, while 

CYP1B1 gene was found to be frequently mutated, in about 20% to 100% cases of 

PCG from Saudi Arabia, Slovakian gypsies and Japan (Kim et al., 2011; Plásilová et 

al., 1999; Tanwar et al., 2009; Bejjani et al., 1998). 

 



32 
 

1.6.4.6 Microphthalmia 

Microphthalmia, anophthalmia and coloboma disorders, usually referred to as MAC, 

present as a syndrome rather than separate malformations (FitzPatrick and Van 

Heyningen, 2005; Ragge et al., 2007). The disruption of normal formation of anterior 

segment and congenital cataracts may be seen in patients with microphthalmia, this 

will lead to opacification of the cornea, abnormal irides and small abnormal lenses 

(Verma and FitzPatrick, 2012). A study of congenital eye disease at a teaching 

hospital in Sagamu, Nigeria, found that corneal opacities is the fourth most common 

congenital eye disorder causing blindness in 5.7% of infants with congenital eye 

diseases (Bodunde and Ajibode, 2006). 

Microphthalmia (small eyes) refers to an eye with reduced volume and may be 

associated with Anophthalmia (complete absence of the eye) and Coloboma (a defect 

in the iris of the eye) (Skalicky et al., 2013). MAC contributes an important 

percentage of congenital ocular disorders worldwide (Hornby et al., 2000).  The 

presence of any of the three disorders always increases the risk of development of 

cataracts and glaucoma if left untreated. The aetiology of the spectrum of eye disease 

is yet to be fully understood. The evidence gathered from experimental and clinical 

investigations suggests a heterogenous genetic basis which also includes a disruption 

(Local) in development of the eye (Verma and Fitzpatrick, 2007; Bardakjian and 

Schneider, 2011). There are various factors involved in the early stages of eye 

development such as transcription factors, complex signalling molecules, and 

structural proteins, adhesion factors, cell cycle regulators which are downstream 

targets (Yun et al., 2009; Fuhrmann, 2010; Eiraku et al., 2011). The regulation of 

eye development pathways interacts in a specific manner which may be time and 

tissue specific. This may help explain the overlap of the underlying genetic 

heterogeneity of the disorders with the clinical phenotypes (Bardakjian and 

Schneider, 2011). 

 

 

 



33 
 

1.6.5 The genetics of anterior segment dysgenesis 

Mutations in genes that initiate and regulate the complex pathways involved in eye 

development can cause a spectrum of disorders such as ASD which can present with 

MAC. Mutations in genes such as PXDN, PAX6, SOX2, ALDH1A3, STRA6, OTX2, 

RAX, BMP4, VSX2, GDF6, VAX1, SMOC1 and SIX6 to name a few have been found 

in patients presenting with ASD and MAC disorders (Schneider et al., 2009; Khan 

et al., 2011; Slavotinek et al., 2011; Aldahmesh et al., 2013; Yan et al., 2014; Choi 

et al., 2015). Genes such as FOXE3, PITX3, PAX6, B3GALTL, COL4A1 and 

SH3PXD2B have been screened for mutations and deletions in patients with ASD, 

and there is substantial heterogeneity in both MAC and ASD disorders which causes 

overlap thus (Figure 1.10) leading to no clear-cut molecular diagnosis (Semina et 

al., 2001; Mao et al., 2012; Alzuhairy et al., 2013; Prokudin et al., 2013).  

 
Figure 1.10: Genes that overlap in clinical features of congenital eye disorders 

due to genetic heterogeneity. Most common congenital eye disorders are shown 

along with the genes that can cause them. Eye disorders that manifest without 

abnormality of the system are represented by genes in normal font. The bolded genes 

are mostly found in eye disorders with syndromic features (Choi et al., 2015). 



34 
 

Whilst there are many unidentified genes that are involved in the development of 

congenital eye disorders, Figure 1.11 shows some of the genes known to be active 

at different stages of eye development (Sinn and Wittbrodt, 2013) and the 

chromosomal location along with the function of the corresponding proteins is listed 

in Table 1.1. It is imperative to recognise ASD because it is commonly linked with 

cataracts and influences congenital glaucoma development in 50% of diagnosed 

patients as well as an increase in ocular pressure (Reis and Semina; 2011; Choi et 

al., 2015). 

 
Figure 1.11: Various genes activated at different stages in the development of 

the eye in vertebrates. Various key transcription factors such as PAX6, SIX3, 

OTX2, WNT as well as LHX2 are required for normal eye development starting 

from the forebrain development and eye field specification stages in the presence 

FGF. Expression of FGF as well as TGF-β leads to upregulation of SHH, SIX3 and 

RX3 which drive eye field split with RX3 further expressed for optic vesicle 

evagination. NUMB and OPO which regulated cell differentiation are expressed to 

enable optic cup formation, with FGF 3 and 8 expression enabling SHH to initiate 

retina differentiation. (Sinn and Wittbrodt, 2013).  

 



35 
 

Investigations have been carried out with regards to some of the above-mentioned 

genes in cohorts of various ethnic groups. Several other genes other than PXDN were 

also investigated in other cohorts for their association with ocular congenital 

disorders (Table 1.2). Mutations in genes such as PNPT1, COL4A1, OTX2, GDF6, 

RARB, and STRA6 were detected in 32 patients diagnosed with microphthalmia and 

other defects of the eye (Slavotinek et al., 2015). A Pakistani family of seven and an 

Indian family were screened for mutations via sequencing, and three mutations 

(novel) were detected in the gene ALDH1A3 [c.964G>A (p.V322M), 

c.1310_1311delAT (p.Tyr437Trpfs*44)] as well as c.289A>G (p.Ile97Val) 

missense mutation in FOXE3. Families harbouring these mutations were presenting 

with microphthalmia (Ullah et al., 2016). In a study by (Prokudin et al., 2014) five 

causal variants in PAX6, CRYGC, CYP1B1, and GJA8 were identified from 11 Indian 

families (unrelated) with Peter’s anomaly, cataract, coloboma and microphthalmia.   

Recently, mutations in PXDN have been associated with ASD. In a study by Khan 

et al. (2011), it was found that two Pakistan families (consanguineous) and one 

Cambodian family (consanguineous) harboured homozygous mutations in PXDN 

gene (Khan et al., 2011). The two Pakistani families presented with congenital 

cataract and corneal opacity which was mild and moderate, while the Cambodian 

family presented with severe corneal opacification and congenital glaucoma. An 

exon 17 missense mutation c.2638C>T (p.Arg880Cys) was detected in PXDN via 

direct sequencing. Having scored 0.98 with PolyPhen the mutation was likely to be 

deleterious and the amino acid change was within the conserved peroxidase domain, 

which could affect both protein structure and enzymatic ability of the domain (Khan 

et al., 2011). Another variant in exon 10 of PXDN c.1021C>T (p.Arg341X) was 

found to likely cause protei