Real-world clinical and molecular management of 50 prospective patients with microphthalmia, anophthalmia and/or ocular coloboma

Background/aims Microphthalmia, anophthalmia and coloboma (MAC) are clinically and genetically heterogenous rare developmental eye conditions, which contribute to a significant proportion of childhood blindness worldwide. Clear understanding of MAC aetiology and comorbidities is essential to providing patients with appropriate care. However, current management is unstandardised and molecular diagnostic rates remain low, particularly in those with unilateral presentation. To further understanding of clinical and genetic management of patients with MAC, we charted their real-world experience to ascertain optimal management pathways and yield from molecular analysis. Methods A prospective cohort study of consecutive patients with MAC referred to the ocular genetics service at Moorfields Eye Hospital between 2017–2020. Results Clinical analysis of 50 MAC patients (15 microphthalmia; 2 anophthalmia; 11 coloboma; and 22 mixed) from 44 unrelated families found 44% had additional ocular features (complex) and 34% had systemic involvement, most frequently intellectual/developmental delay (8/17). Molecular analysis of 39 families using targeted gene panels, whole genome sequencing and microarray comparative genomic hybridisation identified genetic causes in, 28% including novel variants in six known MAC genes (SOX2, KMT2D, MAB21L2, ALDH1A3, BCOR and FOXE3), and a molecular diagnostic rate of 33% for both bilateral and unilateral cohorts. New phenotypic associations were found for FOXE3 (bilateral sensorineural hearing loss) and MAB21L2 (unilateral microphthalmia). Conclusion This study highlights the importance of thorough clinical and molecular phenotyping of MAC patients to provide appropriate multidisciplinary care. Routine genetic testing for both unilateral and bilateral cases in the clinic may increase diagnostic rates in the future, helping elucidate genotype–phenotype correlations and informing genetic counselling.


INTRODUCTION
Microphthalmia, anophthalmia and ocular coloboma (MAC) fall within the same phenotypic spectrum of ocular maldevelopment.Anophthalmia is defined as no visible ocular globe, while microphthalmia consists of a small eye (axial length ≥2 SD below the normal axial length), both of which are caused by disrupted eye development, resulting in failure to initiate or premature arrest of oculogenesis. 1 Ocular coloboma is a persistent inferonasal tissue defect arising from incomplete fusion of the optic fissure, affecting one or more structures including the iris, ciliary body and retina. 2MAC may present in one or both eyes (unilateral/bilateral) and can occur in combination (mixed) within the same eye or contralateral

Clinical science
eyes.MAC can be simplex, with no additional ocular complications, but frequently manifests with complex associated features including cataracts and anterior segment dysgenesis. 1 Furthermore, between 33% and 95% of patients display extraocular involvement (syndromic). 3 4MAC has a combined prevalence of between 1 and 4 live births per 10 000 4 5 and often results in reduced visual acuity, contributing up to 15% of childhood blindness and severe visual impairment worldwide. 6enetic or environmental factors disturbing eye development from 3 weeks of gestation can cause MAC by altering molecular signalling pathways regulating oculogenesis. 1 4 7Over 100 monogenic associations with MAC have been identified as well as large chromosomal anomalies detected in up to 15% of patients through conventional cytogenetics, 5 7 8 although lower rates are reported in non-syndromic cohorts (8%). 9Clinical heterogeneity in MAC cohorts is partially due to the spectrum of causative genes playing diverse roles in controlling multiorgan/ tissue development.Discovering genotype-phenotype correlations can guide patient management and genetic counselling. 3 10hile a genetic cause can be identified in 80% of bilateral severe microphthalmia and anophthalmia patients, only 10% of unilateral patients typically receive a molecular diagnosis. 7Improving molecular diagnostic rates through increased genetic testing and identification of novel variants will improve understanding of genotype-phenotype relationships.
This study reports clinical evaluation and genetic testing results of 50 MAC patients referred to Moorfields Eye Hospital (MEH) NHS Foundation Trust ocular genetics service prospectively over a 29-month period.

MATERIALS AND METHODS Editorial policies and ethical considerations
This study had relevant local and national research ethics committee approvals (MEH and the Northwest London Research Ethics Committee) and adhered to the tenets of the Declaration of Helsinki.Patients and relatives gave written informed consent for participation through either the Genetic Study of Inherited Eye Disease (REC reference 12/LO/0141, 10 October 2016) or Genomics England 100 000 Genomes project (REC reference 14/EE/1112, 20 February 2015).

Clinical evaluation
Patients with microphthalmia (axial length <16 mm at birth, <19 mm at 12 months, <21 mm in adult eyes), anophthalmia (no evidence of a globe or ocular tissue in the orbit on clinical examination/MRI scan) and/or ocular coloboma presenting to the ocular genetics service at MEH between 1 September 2017 and 30 January 2020 were included in this study.Patients were classed as simplex MAC if no other additional ocular phenotype was seen; mixed MAC was a combination of two or more MAC features in the same or contralateral eye; complex MAC was considered the presence any non-MAC ocular features in the MAC-affected or contralateral eye.Patients were classed as syndromic if any non-ocular phenotype was identified or non-syndromic is no systemic/extraocular involvement.Each patient underwent detailed clinical evaluation, including full history, orthoptic assessment, refraction, best-corrected visual acuity (BCVA) measured using LogMAR or Cardiff cards for preverbal children up to 36 months of age where possible; slit lamp examination and fundus examination were recorded with anterior segment imaging using the Haag-Streit slit lamp camera (Haag-Streit Holdings AG, Köniz, Switzerland) and ultrawidefield fundus colour imaging with the Optos California.
Further investigations included orbital ultrasound to measure axial length, electrophysiology and MRI brain and orbits (where applicable).Every patient was evaluated by a paediatrician for systemic associations.

Patients and genetic analysis
Whole genome sequencing (WGS) analysis was undertaken through the 100 000 Genomes Project. 11Genomic DNA was processed using Illumina TruSeq DNA PCR-Free Sample Preparation kits (Illumina) and sequenced using the Illumina HiSeq X Ten high-throughput sequencing platform, generating minimum coverage of 15X for >97% of the callable autosomal genome. 12eadings were aligned to build GRCh37/GRCh38 of the human genome using an Isaac aligner (Illumina Inc).Single-nucleotide variants and indels were identified using Platypus software (V.0.8.1) and annotated using Cellbase (https://github.com/opencb/cellbase).Variant filtering was performed using minor allele frequency <0.001 in publicly available and in-house data sets, predicted protein effect and familial segregation.Surviving variants were prioritised using the 'microphthalmia or anophthalmia' or 'ocular coloboma' virtual gene panels (https:// panelapp.genomicsengland.co.uk/panels/).WGS data were not analysed for copy number variants (CNVs) through an automated pipeline; however, CNVs were investigated on a case-bycase basis.
Targeted gene panel testing of MAC genes was conducted through the Rare & Inherited Disease Genomic Laboratory at Great Ormond Street Hospital for Children (GOSH, London, UK).Variant screening was carried out by library preparation using Agilent-focused clinical exome+1 kit followed by nextgeneration sequencing on Illumina platforms.Data were analysed using in-house pipelines and virtual gene panels, with variants confirmed by Sanger sequencing.Larger insertion/ deletion mutations and CNVs are routinely investigated using ExomeDepth, but the sensitivity to detect these mutation types is unknown.aCGH, which is able to detect pathogenic CNVs, was performed through GOSH, together with segregation of known familial variants as previously described. 12iterature searches along with public databases HGMD (http://www.hgmd.cf.ac.uk), 13 Clinvar (https://www.ncbi.nlm.nih.gov/clinvar/) 14 and gnomAD 15 were examined for prior reports of variants found in this cohort.Likely pathogenicity of novel variants was assessed using the predictive algorithms gnomAD, PolyPhen-2, 16 SIFT 17 and PredictSNP. 18Genetic results were reviewed by a multidisciplinary team (including clinical scientists, specialists in clinical genetics and ophthalmology), to confirm variant pathogenicity, prevalence in publicly available genome databases, the clinical phenotype and mode of inheritance, before the molecular diagnosis was established.Novel variants in this study were submitted to ClinVar.

Clinical science
Of the 44 families, 36 were simplex cases with no known family history (82%), 2 had affected siblings only (5%) and 6 had a multi-generational family history (14%).Two cases were known to be the offspring of consanguineous parents (5%).
Of the eight cataract patients, seven had lensectomy and were left aphakic.The age of surgical intervention ranged from 6 weeks to 28 years.Two eyes developed secondary glaucoma following lensectomy; there were no primary cases of glaucoma in this series.
Retinal detachments (RDs) were observed in seven affected eyes from six patients (7/80, 9%) all with ocular coloboma±microphthalmia. Six patients had unilateral RD; patient 10-1 had bilateral consecutive RDs: patient 10-1 had symptoms in one eye, but the other eye was found to have a recent onset RD in the other better-seeing eye during time of examination under anaesthesia.The detachments were thought to have occurred within several months of each other, but no trauma had been reported.No molecular tests were performed for this patient.Three eyes had only coloboma (8-1 had optic disc; 10-1 had bilateral chorioretinal) while four had coloboma with microphthalmia (12-1 had optic disc; 4-1, 17-1 and 26.1 had maculasparing chorioretinal).RDs occurred before the age of 1 year in three eyes (12-1, 17-1, 26-1).Vitreoretinal surgery was performed on patient 10-1 as both eyes had visual potential, but other patients were treated conservatively due to the extent of the detachment at presentation or poor visual potential.Unoperated eyes became painless pthtisical eyes (5/7).
The majority of patients did not need or were unsuitable for prosthetic rehabilitation because of normal or near-normal axial lengths (31/50, 62%).Of the 19/50 who were candidates for prosthetics, 2/19 (11%) did not undergo prosthetic rehabilitation despite low axial lengths (patient 6-1 decided not to pursue prosthetics and patient 39-1 could not retain a prosthesis due to presence of a developmental cyst).Nineteen microphthalmic eyes underwent prosthetic rehabilitation, along with three anophthalmic eyes.Patients who underwent prosthetic rehabilitation had bespoke treatment plans depending on factors such as age of presentation, size of affected eye and patient/parent request (table 1): 10 customised cosmetic prosthesis (CCP), 7 customised clear shell (CCS).All eyes fitted with CCS were microphthalmic (7/7), while 75% CCP eyes were microphthalmic (9/12) with the remainder being anophthalmic (3/12).
Developmental delay/intellectual disability was found in eight patients, all of whom had an MRI: six bilateral and two unilateral.Patient 9-1 with Kabuki syndrome, 22-1 with learning difficulties and 22-1 with intellectual disability and autism all showed no intracranial abnormalities.Patients 35-1 and 35-2 were diagnosed with global developmental delay and severe intellectual disability, respectively, and both had bilateral sensorineural hearing loss, and 35-1 had cochlear implants fitted, however, neither had abnormal MRI findings.

DISCUSSION
The aetiology of MAC is complicated by a multitude of genetic and environmental factors resulting in highly variable phenotypes.The multiplex role of genes in regulating the development of ocular and extraocular tissues means that MAC is often associated with complex ocular phenotypes and systemic associations in 33%-95% of patients. 3 4Here, we report clinical presentation of 50 patients referred to the ocular genetic service at MEH. Paediatric and adult patients in this cohort were referred to the ocular genetic service at different stages of life, most having experienced various investigations and management by local teams.The care pathway provided by MEH enables a multidisciplinary approach from the outset, which incorporates full phenotyping (with ophthalmic and systemic examination including parents, appropriate imaging and paediatric review for children), visual and aesthetic rehabilitation, developmental support and genetic investigation.Overall, lower BCVA in microphthalmic eyes compared with those with coloboma likely reflects the global effect on eye development, whereas visual impairment can vary depending on size and location of the colobomatous defect.It highlights the importance of determining the axial length through ultrasound B-scan or orbital MRI.Complex ocular features were found in 44% of patients, with cataracts occurring most frequently.An RD rate of 9% was identified in this cohort ranging from the first to the third decade; this varies widely in the literature (2.4%-42%) for patients with chorioretinal coloboma. 28 29Guidance regarding screening intervals and overall screening period is not clear and it is challenging to be prescriptive for rare and heterogeneous groups of patients with variable visual potentials.From this data, patients suitable for vitreoretinal surgery should be screened during their preverbal years and parents are given advice about monitoring for change in visual behaviour between intervals.
Heterogeneous MAC conditions in this study required a customised approach in prosthetic management, but general principles include assessment of ocular visual potential and the requirement for periorbital tissue rehabilitation before considering CCP treatment.Most patients (66%) did not require CCP and 7/50 were advised to have CCS as there was visual potential, but the eyes were deemed too small to support normal periocular development.It must be noted that some patients may still require socket surgery to support the use of a prosthesis or provide more symmetry between the periocular areas.
Seventeen patients (34%) had systemic involvement, in line with previous reports, 3 4 and 56% had brain MRI scans, with intracranial abnormalities were found in 7/28, associated with intellectual disability/developmental delay in 3/7.The overwhelming majority of patients with abnormal intracranial findings had bilateral MAC (6/7, 86%) although previous reports have found no association with frequency/type of systemic abnormality by laterality. 30There is no definitive guidance on neuroimaging for MAC; however, this study reinforces the association with midline neurological defects including pituitary abnormalities and, hence, it is advisable to consider endocrinology work up and neuroimaging, particularly for bilateral cases.
Previous studies report higher molecular diagnostic rates for bilateral MAC patients (26-80%) than unilateral (10-20%). 4 7 31olecular testing of 45/50 patients in the mixed MAC cohort yielded an overall diagnostic rate of 33%, which was consistent for both bilateral and unilateral cases, despite only 79% of unilateral cases having genetic testing compared with 97% of bilateral patients.Hence, both presentations warrant access to genetic testing as the rate of diagnosis is not preferentially higher for bilateral cases as previously considered.The highest proportion of molecular diagnoses were made from WGS (36%).A novel association of known cataract gene EPHA2 with microphthalmia was identified, 19 and the remaining monogenic causes had been previously associated with MAC, although with previously unreported pathogenic variants.Most genetic changes were in transcription factors (SOX2, PAX6 and FOXE3) 10 21 and molecules which regulate gene expression, such as histone modifier KMT2D and corepressor BCOR. 23 26 32 previously unreported association of FOXE3 with bilateral sensorineural hearing loss was identified in two patients from family 35, alongside bilateral enlarged kidneys with multiple cysts.Biallelic pathogenic variants of forkhead transcription factor FOXE3 are associated bilateral microphthalmia and coloboma, typically in non-syndromic patients; a recent study of 102 individuals found only 8% displayed extraocular features, including global developmental delay and autism. 10 27 31Systemic features were more commonly found in patients with truncating DNA changes; however, sensorineural hearing loss and kidney anomalies have not previously been reported.No other pathogenic variants were found from exome panel testing, and segregation of FOXE3 with hearing loss in two patients suggests c.763dup; p.(Ala255Glyfs*30) is likely a recessive disease-causing variant affecting both eye and ear development; all parents were unaffected and found to be carriers.However, while the presence of an additional pathogenic mutation is unlikely it cannot be entirely disregarded.
We also report the first association of MAB21L2 with unilateral MAC in a patient with non-syndromic, simplex microphthalmia and novel heterozygous nonsense variant, c.379A>T; p.(Lys127*).Pathogenic MAB21L2 variants are found in patients with a spectrum of developmental ocular conditions, including bilateral MAC with or without systemic features, yet prior to this study, no unilateral MAC patients have been reported. 2 24 33ntrafamilial phenotypic variation was observed in MAC phenotype of family 20 and both MAC phenotype and additional ocular features of family 40.Patient 20-1 displayed bilateral chorioretinal colobomas with a microphthalmic right eye, while their affected sibling exhibited no microphthalmia, but bilateral iris and chorioretinal colobomas and an optic nerve coloboma in the right eye.Patient 20-2 also had myopia, while patient 20-1 had no non-MAC ocular conditions.Patient 40-1 had unilateral mixed MAC while their affected son exhibited bilateral MAC.Neither of these heterogeneous families obtained a genetic diagnosis, so the underlying cause of the variation in these cases cannot be yet established.However, a multitude of factors can lead to intrafamilial phenotypic variability, such as genetic modifiers, epigenetic variation and environmental factors influencing gene expression.These epistatic/external factors can make genetic diagnosis more difficult, as they can influence segregation patterns, and more work understanding and diagnosing these effects would provide further insight into MAC aetiology.
Patients diagnosed with a molecular cause were directed to appropriate specialists for investigation and management of ocular/systemic features where genotype-phenotype correlations were known.Furthermore, genetically solved patients in this cohort were given access to genetic counselling to provide information on the potential impacts of their test results and guide future choices.However, with less than a third of families (28%) obtaining a genetic diagnosis, unfortunately many Clinical science families could not be given specific clinical care or advice based on their underlying molecular cause, resulting in greater uncertainty regarding prognosis and family planning.
Clear understanding of MAC aetiology is key to providing appropriate clinical care and genetic counselling.However, genetic heterogeneity and complex inheritance patterns make molecular diagnosis challenging, particularly for unilateral cohorts.This work highlights the importance of careful phenotyping, to assemble the appropriate multidisciplinary care team to undertake investigations such as MRI brain imaging and genetic testing.Increased usage of next-generation sequencing to identify novel variants and complex non-mendelian causes of MAC can clarify genotype-phenotype relationships, pointing to potential comorbidities, which ensure that patients are referred to the correct specialists, which can improve prognosis.

F
kidney larger than other (left duplex, right dysplastic), multiple UTI, single palmar crease left hand, low set ears, developmental delay, brachycephaly, severe microcephaly, abnormal intracranial findings (simple gyral pattern, cerebellar hypoplasia, short corpus callosum and post positioned pituitary) intracranial findings (vascular anomaly of the scalp), mild developmental delay/ intellectual disability

Figure 1
Figure 1 Demographics and clinical phenotype of 50 microphthalmia, anophthalmia and ocular coloboma (MAC) patients: (a) Sex; (b) Ethnicity; (c) MAC phenotype of patients divided into exclusively microphthalmia/anophthalmia/ocular coloboma or a combination of MAC (mixed); (d) Ocular and systemic features of MAC cohort: laterality (unilateral/bilateral MAC); presence of associated non-MAC ocular phenotype in affected or contralateral eye (simplex/complex); presence of systemic features (syndromic/non-syndromic). (e) Mean best corrected visual acuities of MAC-affected eyes; (f) Additional ocular phenotypes present in MAC patients; (g) Extraocular associations in MAC patients (abnormal intracranial findings refer to significant findings which would affect the health of the patient, including agenesis of corpus callosum, hydrocephalus and cerebellar hypoplasia; developmental delay includes intellectual/learning difficulties and global developmental delay.ASD, anterior segment dysgenesis; G6P6, glucose-6-phosphate dehydrogenase; RD, retinal detachment; UTI, urinary tract infections. Clinical sciencedeclined testing; and three families did not attend their genetics appointments following referral.Patient 6-1 had previously undergone an aCGH prior to this study.Following genetic diagnosis of patient 35-2 from a targeted gene panel, patient 35-1 underwent a single gene test to confirm genetic cause.Patient 7-1 underwent a single gene test for PAX6 variants, following previous familial diagnosis.Fifteen patients from 11 families obtained a genetic diagnosis (11/39, 28%), while 30 from 28 families had no primary findings.Family solve rate by test was 4/17 WGS (24%), 3/3 by aCGH (100%), 3/18 by targeted panel testing (17%) and 1/1 by familial variant testing only (figure2A).Pathogenic variants were identified in eight genes: SOX2, PAX6, KMT2D, EPHA2, MAB21L2, ALDH1A3, BCOR and FOXE3, in addition to large deletions in chromosome 10, 11 and X.

Table 2
Summary of phenotype data of microphthalmia, anophthalmia and coloboma (MAC) patients with molecular diagnostic testing

Table 3
Variant details for solved families of microphthalmia, anophthalmia and coloboma (MAC) patients