The Genetic Determinants of Axial Length: From Microphthalmia to High Myopia in Childhood

The axial length of the eye is critical for normal visual function by enabling light to precisely focus on the retina. The mean axial length of the adult human eye is 23.5 mm, but the molecular mechanisms regulating ocular axial length remain poorly understood. Underdevelopment can lead to microph-thalmia (defined as a small eye with an axial length of less than 19 mm at 1 year of age or less than 21 mm in adulthood) within the first trimester of pregnancy. However, continued overgrowth can lead to axial high myopia (an enlarged eye with an axial length of 26.5 mm or more) at any age. Both conditions show high genetic and phenotypic heterogeneity associated with significant visual morbidity worldwide. More than 90 genes can contribute to microphthalmia, and several hundred genes are associated with myopia, yet diagnostic yields are low. Crucially, the genetic pathways underpinning the specification of eye size are only now being discovered, with evidence suggesting that shared molecular pathways regulate under-or overgrowth of the eye. Improving


INTRODUCTION Axial Length Specification: From Disrupted Early Development in Microphthalmia to Abnormal Postnatal Growth in High Axial Myopia
Early eye growth is under strict spatiotemporal control coordinated by a self-regulatory network of genes beginning from 3 weeks' gestation in humans and is specified in the anterior neuroectoderm (117).Upregulation of genes encoding eye field transcription factors, including RAX, PAX6, and SIX3, by OTX2 allows for eye field specification (16,117,148,152).Anterior migration of cells splits the eye field in two, evaginating toward the overlying surface ectoderm and forming two optic vesicles (76).This induces the specification of the lens placode, as well as concurrently inducing invagination of the optic vesicle to form a bilayered optic cup (16,21,117,148,152).Retinal pigment epithelium (RPE) is derived from the outer layer of the cup, while the inner layer forms the neural retina (Figure 1a-c).The optic fissure forms from an opening on the ventral surface of the optic cup, allowing periocular mesenchyme to form the hyaloid vasculature to support ocular development, and closes by gestational week 7 in humans (36,44).The human sclera-the dense fibrous tissue forming the outer coat of the eye-differentiates from both the neural crest (surrounding the optic cup) and the mesoderm (contributing to the vascular endothelium and extraocular muscles) in week 6 of human embryonic development (98).Differentiation occurs anterior to posterior, reaching the equator by week 8 and the posterior pole by week 12.The scleral spur can be identified by month 4, and the lamina cribosa is formed by month 5 (109).Microscopically, the sclera is a fibrous connective tissue consisting of lamellae of collagen fibrils interspersed with proteoglycans and noncollagenous glycoproteins.Type I collagen forms approximately 90% of the sclera, along with types III, IV, V, VI, VIII, XII, and XIII.Scleral connective tissue consists of extracellular matrix (ECM) and fibroblasts.Scleral fibroblasts dynamically alter the composition and biomechanical properties of the sclera by remodeling the ECM through the expression of matrix metalloproteinases (MMP-1, MMP-2, MMP-3, MMP-9, and TIMP-1) (98).In addition to the collagen, proteoglycans consisting of a protein core and glycosaminoglycan chain form another key part of the ECM, increasing from childhood throughout life (98).
At birth, the average human eye has an axial length of approximately 16.5-17 mm, growing to 21-23.5 mm in adulthood (83).The newborn eye is hypermetropic, with changes in the cornea, lens, and axial length resulting in adjustment of the refractive state of the eye in a process called emmetropization, which occurs in childhood, with the cornea and lens undergoing the largest transition (118).The incidence of myopia, usually attributed to increased axial length, increases approximately sevenfold during puberty, peaking between ages 9 and 14, with axial length increasing by 0.5 mm (25,29).Myopia developed during childhood or early adolescence worsens throughout adolescence, stabilizing by the early 20s.Adult-onset myopia is typically less severe compared with myopia developing in childhood.Postnatal eye growth is influenced by visual stimuli that trigger a signaling cascade initiated within the retina, passing through the RPE and choroid and resulting in scleral remodeling (91,98,105,122).Evidence has been obtained largely from form-deprivation and optical-defocus animal models, which suggest that the neural retina is the source of growthregulating signals (105).High myopia is morphologically characterized by scleral thinning and posterior scleral ectasia (98) (Figure 1d).

Disorders of Axial Length
Disorders of axial length give rise to a spectrum of ocular conditions, ranging from reduced eye size arising in early development, such as microphthalmia and nanophthalmia, to pathological overgrowth resulting in axial myopia.Microphthalmia has a prevalence of 1 in 7,000 births and is defined as the presence of a small eye with an axial length of less than 19 mm at 1 year of age or less than 21 mm in an adult measured on a B-scan ultrasound, representing two or more standard deviations below normal (44, 100) (Figure 2a-c).Nanophthalmia and posterior microphthalmia are rare subsets, where the eye is structurally normal overall but has a reduced axial length of less than 20 mm with high hypermetropia (more than +8.00 diopters) (58,102).
Microphthalmia is reported in up to 11.2% of blind children and contributes up to 15% of severe visual impairment in children worldwide, with the effect on vision dependent on the severity of the abnormality, the size of the eye, and the associated ocular malformations (47,136).Microphthalmia exhibits high phenotypic heterogeneity that is often complex and associated with other ocular abnormalities, such as anterior segment dysgenesis, ocular coloboma, cataract, or vitreoretinal dysplasia (103).Systemic malformations may be present, with up to 45% of individuals diagnosed with a recognized syndrome and up to 95% of patients having extraocular features (111,119).There are no treatments to encourage axial growth in microphthalmic patients, and current management focuses on maximizing existing vision and enhancing cosmetic appearance.Depending on methodology, phenotype, and severity, diagnostic rates vary between 33% and 80% for bilateral severe microphthalmia patients, but typically only up to 33% of unilateral patients receive a molecular diagnosis (43,96).Improving molecular diagnostic rates through increased genetic testing and identification of novel variants will improve understanding of genotype-phenotype relationships and guide patient management and genetic counseling (136).
Refractive error is a function of corneal curvature, lens power, lens position, and axial length.Myopia is a refractive state that occurs when light focuses in front of the retina and can be either refractive or axial.The most common form of myopia, termed axial myopia, is usually attributable to an abnormally long eye.Myopia is classified as low (or common) myopia if the refractive error is less than −6.00 diopters and as high myopia if the refractive error is −6.00 diopters or greater and/or the eye has an axial length of 26.5 mm or more.Pathological myopia is defined as high myopia with any posterior myopia-specific pathology from axial elongation, such as myopic macular degeneration, which reportedly affects 3% of the world's population and can lead to severe vision loss (30,87,139).Myopia can be simple or complex depending on whether the eye exhibits other abnormalities, as well as syndromic or nonsyndromic depending on whether there are extraocular features or systemic disease.
Myopia is a highly complex trait of multifactorial etiology with environmental, behavioral, and genetic influences (68,147).The global prevalence of myopia has increased drastically over the last two decades, with 30-50% of the adult population now myopic in Europe, America, Australia, and Asia.It is predicted that by 2050, one-fifth of patients with myopia will develop high myopia (46).Refractive symptoms can be treated with glasses, contact lenses, or refractive surgery, and atropine and orthokeratology may slow or prevent axial elongation in common myopia; however, the risk of complications such as retinal detachment, glaucoma, and myopic macular degeneration increases with longer axial length (135) (Figure 2d,e).These complications can potentially lead to visual impairment and blindness (32,106).The development and progression of myopia occur in childhood and adolescence, with up to 16.4% of teenagers in the United Kingdom and more than 95% of those in the Far East affected (32).Emmetropization, a developmental process whereby the refractive systems of the eye balance with the axial length so that light is focused on the retina when viewing at distance, is heavily influenced by environmental factors, resulting in the development of low or moderate myopia during adolescence.Light exposure is a key component of emmetropization: Children with myopia spend less time outdoors, and a lack of sunlight linked to an increasingly urban lifestyle is a key environmental risk factor (127).Randomized controlled trials with children in East Asia demonstrated that increasing time outdoors delayed the onset of myopia and slowed down progression (143).
There is a strong genetic component to the onset and progression of myopia.Mendelian inheritance has been reported (37,138); however, many common genetic variants (>100) have been associated with risk of refractive error and common myopia via large-scale studies, including studies performed by the Consortium for Refractive Error and Myopia in conjunction with the UK Biobank Eye and Vision Consortium and 23andMe (27,48,125,126).A significant number of patients who present in early childhood (preschool) with high myopia (−6.00 diopters or more) then continue to progress, with increasing axial length and potentially pathological myopia developing over time.These patients also have a strong family history with the same presentation, highly suggestive of a genetic cause, with heritability estimated at over 90% in large twin studies (41,74).This review focuses on the genetic basis of preschool children with high myopia as a genetic model of axial length determination.
There are an increasing number of examples of how shared genetic pathways between microphthalmia and myopia may influence axial growth.This article discusses the current understanding of the underlying genetic determinants of axial length and potential molecular mechanisms to improve our understanding of the pathology of these two conditions, with a view to developing future therapeutic interventions.

REDUCED AXIAL LENGTH: THE GENETICS OF MICROPHTHALMIA AND NANOPHTHALMIA
Microphthalmia exists on a phenotypic spectrum of ocular maldevelopment with anophthalmia, defined as aborted eye development during optic vesicle formation, leaving a cystic remnant, and ocular coloboma, a structural malformation resulting from incomplete optic fissure fusion.There is high genotypic heterogeneity, with more than 90 monogenic causes having been identified (Table 1), as well as large chromosomal abnormalities in up to 8-15% of patients (55,92,96,107).Environmental causes, such as maternal vitamin A deficiency and alcohol consumption, contribute to approximately 2% of microphthalmia cases (8,13,111).All forms of inheritance have been noted (de novo sporadic, autosomal dominant, autosomal recessive, X-linked dominant, and X-linked recessive), with most pathogenic mutations associated with nonsyndromic cases that sporadically arise de novo, including missense, nonsense, frameshift, and splice-site variants.Variable expressivity and nonpenetrance have been observed, as well as germline mosaicism, creating significant challenges in counseling of patients (6,96).
Much of our understanding of microphthalmia derives from human cellular and animal models, including mice, zebrafish, and Xenopus, given the challenges of studying the molecular development of the human eye at such an early gestational stage (42).The two main groups of disease-causing genes in microphthalmia are (a) genes encoding eye field-initiating transcription factors, including SOX2, OTX2, RAX, VSX2, and PAX6, which regulate further downstream signaling pathways central to tissue specialization (e.g., WNT, BMP, TGFβ, and SHH), and (b) the retinoic acid signaling component-encoding genes STRA6, ALDH1A3, and RARβ, which are key for early eye morphogenesis (44).
Heterozygous mutations of SOX2 account for up to 40% of bilateral microphthalmia (35,136).More than 70 disease-causing variants have been identified, 60% of which arise de novo, with autosomal dominant inheritance also being reported, leading to microphthalmia from haploinsufficiency (13,119).OTX2 mutations account for up to 8% of microphthalmia cases; combined with SOX2 mutations, these account for up to 60% of bilateral microphthalmia cases (35,136,141).Animal modeling in Xenopus has shown that Sox2 and Otx2 play key roles in early eye specification by coordinating Rax signaling, positively regulating the expression of eye field transcription factors, which subsequently downregulates OTX2 in a negative feedback loop (22,79).Homozygous or compound heterozygous mutations in RAX are responsible for approximately 2% of microphthalmia cases, usually presenting with a bilateral severe phenotype, with rx3 mutant zebrafish failing to produce optic vesicles (13,72).OTX2 is vital for eye development at both early and later stages, including regionalization of the optic vesicle following TGFβ signaling and RPE differentiation following Wnt signaling and regulation of MITF (148).Nonpenetrance in OTX2 mutations is high, with 35% of mutations inherited from unaffected parents with no ocular phenotype and thought to be due to gonosomal mosaicism (35,99,141).OTX2 mutations are also associated with highly variable ocular and systemic phenotypes, including anterior segment developmental abnormalities, retinal dystrophy, optic chiasm aplasia, hypopituitarism, and developmental delay, thus presenting a challenge for diagnosis and patient management (108,124).
PAX6 is a highly conserved regulator of ocular development, mutations in which can cause numerous phenotypes, including aniridia; microphthalmia, anophthalmia, and coloboma; and optic nerve hypoplasia, with patients often having a complex phenotype (112).Most mutations in PAX6 that cause aniridia are thought to act through a loss of function (haploinsufficiency). Mutations affecting PAX6 enhancers can also cause aniridia (5).Missense mutations, occurring in the pairedbox domain that binds to SOX2 and to DNA, are more frequently associated with microphthalmia (56,136), which has been attributed to residual DNA-binding activity of the mutant PAX6 leading to a worse-than-null phenotype (137).
Pax6 mutant mouse models exhibit small eyes; however, phenotypic severity varies in zebrafish, where missense mutations in the sunrise pax6b homozygous line replicate the milder microphthalmia phenotype observed in patients with some missense PAX6 mutations, while morpholino-induced knockdown of pax6a results in more extreme phenotypes, including reduced  body size and abnormal brain development.Caution is required in interpreting zebrafish models in these circumstances, as the human phenotype may not be accurately recapitulated due to gene duplication leading to role sharing among multiple orthologous genes (18,24,63,104).
In addition to the eye field-initiating transcription factors, retinoic acid signaling is vital for early optic cup morphogenesis, and mutations in several stages of the pathway lead to microphthalmia.Retinoic acid is released from the developing lens placode and surrounding mesenchyme, which bind to the optic vesicle, stimulating invagination of the optic vesicle from 31 days' gestation (21).Mutations in STRA6-encoding a transmembrane receptor for a retinol-binding protein responsible for mediating vitamin A uptake into cells-cause variable phenotypes, ranging from anophthalmia/microphthalmia to coloboma, highlighting the importance of the retinoic acid signaling pathway in eye development (12,14).RARβ, encoding one of three retinoic acid receptors that transduce retinoic acid signals, is implicated in microphthalmia and anophthalmia through missense, nonsense, and frameshift mutations (121).Mutations in ALDH1A3-encoding one of three human retinaldehyde dehydrogenases that are crucial to the retinoic acid synthesis pathway via oxidation of retinaldehyde acid (21)-may cause up to 10% of microphthalmia cases (119).
Abnormalities in the correct assembly of the ECM may also form an important underlying mechanism in the etiology of microphthalmia.The ECM plays a critical role in early optic vesicle development and in retinal progenitor cell differentiation (110).A notable example is COL4A1, encoding a subunit of type IV collagen that regulates RPE growth, mutations of which lead to a range of ocular maldevelopment issues, including microphthalmia (78).Further characterization of the role of the ECM in the developing eye is required.
Nanophthalmia can be defined as a subset of microphthalmia.Diagnostic criteria vary widely, including not only a high hypermetropic refractive error (more than +8.00 diopters) but also axial length (less than 20 mm), posterior wall thickness, and lens/eye volume ratio metrics, and there is overlap with posterior microphthalmia, in which only the posterior segment is shortened (140,144).Clear distinctions are required in phenotyping patients, which will subsequently allow identification of genotype-phenotype correlations and improve understanding of the mechanism of genetic control of axial length.There is consensus that a nanophthalmic eye is small but structurally macroscopically normal other than thickened sclera and choroid (11,123,132).However, significant microscopic morphological changes underlie the pathology, with abnormal collagen fibers that are split and frayed in all scleral layers (145).Because many nanophthalmia genes are expressed in the retina and RPE, it is unclear whether the primary pathology seen is of the sclera or a secondary abnormality within a yet-to-be-identified retina-RPE-sclera signaling cascade.Despite having a grossly normal macroscopic eye structure, patients with nanophthalmia are still at high risk of sight-threatening sequalae, including amblyopia, uveal effusion syndrome, and angle closure glaucoma (11,90).
Nanophthalmia can be sporadic or inherited in autosomal recessive and autosomal dominant forms (3,77,80,90) (Table 1).Six genes have now been identified in familial forms of nanophthalmia, with PRSS56 and MFRP mutations being the most common (discussed further below, along with TMEM98) (2,85,97,115).Interestingly, variants in genes usually associated with retinal dystrophy have also been associated with nanophthalmia.Mutations in CRB1, encoding a transmembrane protein on the photoreceptor inner segments, typically result in either Leber congenital amaurosis or retinitis pigmentosa (73).Novel homozygous missense variants have been identified in exon 5 (c.1125C>G) and exon 7 (c.2498G>A) of CRB1 in two separate families with retinitis pigmentosa and nanophthalmia (93,149).Additionally, mutations in BEST1, encoding a calcium-activated chloride channel localized to the RPE, are associated with vitelliform macular dystrophy and (rarely) autosomal dominant vitreoretinochoroidopathy; the latter occurs from mutations in splicing regulators of BEST1 and is associated with nanophthalmia and chorioretinal atrophy (65,146).Finally, a C-terminal mutation in the penultimate MYRF exon has recently been identified to cause a rare form of autosomal dominant nanophthalmia (116).One further chromosomal locus, NNO3 (OMIM 611897), reported to be linked to chromosome 2q11-14 (66), has been associated with autosomal dominant nanophthalmia, but the gene is not yet known.

THE GENETICS OF AXIAL MYOPIA
More than 25 chromosomal loci and 100 genes are known to be associated with nonsyndromic axial high myopia, identified through linkage analysis, twin studies, candidate gene analysis, genome-wide association studies (GWASs), pathway analysis, and next-generation sequencing (9,133).GWASs have identified many common variants and have been conducted among numerous ethnic groups with phenotypic features of high myopia, including refractive error, axial length, and decreased macular thickness.Several GWASs and meta-analyses have found that myopia or refractive error is associated with single-nucleotide polymorphisms in the vicinity of GJD2, RASGRF1, GRIA4, KCNQ5, RDH, LAMA2, BMP25, SIX6, PRSS56, CTNND2, ZC3H11B, SNTB1, VIPR2, and ZFHX1B (26,60,69,113), with functional analyses producing useful subdivisions and giving insight into potential mechanisms, which include ECM remodeling, retinoic acid signaling and photoreceptor development, neurotransmission, ion channel activity, and ocular and central nervous system development (48,49,61,131).The understanding of how these genes relate to axial growth and ocular development is limited, however.
There is increasing and reproducible evidence that the TGFβ signaling pathway is disrupted in high myopia.LRPAP1 encodes a chaperone, LRP1, that regulates TGFβ activity, with homozygous nonsense (p.R68X) and frameshift (p.Q67Sfs * 8) mutations having been identified by whole-exome sequencing in high myopia (1,50).The regulation of TGFβ activity by LRPAP1 is supported by the Lrpap1 (SM22-Cre + ;LRP flox/flox ;LDLR −/− ) knockout mouse model, which revealed activation of TGFβ target genes (TSP1 and PDGFRβ) upon loss of Lrpap1 (7).Further functional evidence that the TGFβ/BMP pathway is involved in myopia pathogenesis comes from pathogenic variants in SLC39A5; a total of 13 missense and 2 nonsense mutations cause high myopia (10,28).In a lymphocyte cell line carrying the SLC39A5 p.Y47X nonsense mutation, there was upregulation of SMAD1, which encodes a downstream transcription factor of the TGFβ/BMP pathway.SLC39A5 encodes a zinc transport complex situated in the Golgi apparatus and is expressed in mouse sclera, inner retina, and retinal outer plexiform and ganglion cell layers.Immunofluorescence demonstrates that SLC39A5 is detectable in both embryonic and postnatal mouse eye cryosections (38).
In addition to being central to early eye development and retinal morphogenesis, TGFβ signaling is thought to modulate the postnatal scleral ECM, dysregulation of which is associated with high myopia in both tree shrew animal models and GWAS data (53,54,131).
ZNF644 encodes a transcription factor that is expressed in all tissues, including human neural retina and RPE, and is postulated to regulate protein domain function.Whole-exome sequencing showed that a missense mutation in exon 3 (p.S672G) was responsible for autosomal dominant myopia in a Chinese family (114).Further heterozygous mutations in ZNF644 responsible for high myopia following screening across different populations were also identified, implicating this gene in high myopia (128).A novel missense variant in CCDC111 (p.Y89D) has been identified in a family with high myopia (151); CCDC111 is highly conserved across species and ubiquitously expressed in numerous tissues, including scleral fibroblasts, RPE, and Müller cells, but its precise functions remain unknown.
BSG mutations predispose to early-onset high myopia; six missense mutations (p.P221S, p.G297S, p.G219R, p.G37R, p.G24S, and p.R129C), one nonsense mutation (p.Q69X), and one splicing defect (c.415+1G>A) in this gene have been identified (10,52).BSG plays a role in developmental cell signaling, encoding a photoreceptor-specific transmembrane protein, which interacts with rod-derived cone viability factor in retinal maturation (15).The role of BSG in influencing axial length has been supported by a Bsg knock-in mouse model with a c.901G>A mutation, corresponding to the human c.889G>A p.G297S mutation, which led to an elongated axial length and myopic phenotype (50).
Retinal signal transduction has also been implicated in myopia.Three heterozygous variants in ARR3 (p.A298D, p.R200X, and p.L80P), located on Xq13.1, have been associated with high myopia in female patients in a highly unusual pattern of X-linked female limited inheritance (142).ARR3 encodes a cone arrestin that is enriched in the retina.Arr4 knockout mice develop a conelike dystrophy; however, the mechanisms and role of human ARR3 remain poorly understood (23).OPN1LW, which is located on Xq28 and encodes one of three cone light-absorbing opsins, has been identified as a cause of X-linked syndromic and nonsyndromic high myopia with a frameshift p.Phe208Argfs * 51 mutation (67).Investigation of a multigenerational family with X-linked high myopia and cone dystrophy showed rare exon 3 interchange haplotypes of the OPN1LW and OPN1MW genes, causing apparently nonsyndromic high myopia in young patients but leading to progressive cone-rod dystrophy with loss of color vision and visual acuity in middle-aged patients (88).
SCO2 and NDUFAF7, which are involved in mitochondrial function, have been reported to be responsible for pathological myopia.SCO2 encodes a cytochrome c oxidase assembly protein, dysfunction in which may alter copper homeostasis in ocular tissues, resulting in axial elongation.To date, 10 mutations (including nonsense and missense variants) have been associated with autosomal dominant high myopia (129).This finding has been contested, however, as mice with p.E129K knock-in mutations (corresponding to p.E140K in humans) that result in Sco2 deficiency do not exhibit any axial elongation; therefore, it remains uncertain whether SCO2 is related to axial or other forms of myopia (95).
CTSH encodes one of the cysteine proteases with a role in the degradation of lysosomal proteases.Whole-exome sequencing has identified a 4-bp deletion in CTSH causing high myopia (1).The role of CTSH in axial length is further supported by Ctsh-knockout mice that develop a markedly abnormal globe with an unusual <-shaped morphology that is thought to be due to lengthening of the posterior chamber, although the association has not been formally measured (1).Monogenic causes of high myopia are typically complex and syndromic, caused by a highly penetrant gene.High myopia is typically a feature of Marfan syndrome (FBN1), Stickler syndrome (COL2A1, COL11A1, COL11A2, COL9A1, COL9A2, and COL9A3), and several inherited retinal dystrophies (71,75).The most common inherited retinal dystrophies include congenital stationary night blindness (CACNA1F and NYX) and X-linked retinitis pigmentosa caused by hemizygous variants in RPGR (45,68).Syndromic forms of myopia are typically rare and form the exception.The association with inherited retinal dystrophies and myopia provides evidence that retinal genes are influencing scleral remodeling and eye growth.These results are further supported in form-deprivation myopia, whereby alterations of the visual stimulus trigger a signaling cascade originating in the retina.Interestingly, Stickler and Marfan syndromes are also associated with high myopia, suggesting that scleral weakness results in a mechanical stretch that causes axial elongation as part of the primary pathology.Ultimately, the role of the retina-RPE-sclera pathway and its modulators in specifying axial length requires detailed characterization.

SHARED MOLECULAR PATHWAYS SPECIFYING AXIAL LENGTH
Shared genes and molecular pathways between microphthalmia/nanophthalmia and myopia are now being identified; candidate genes are summarized in Figure 3. Ocular growth has been divided into two distinct phases, prenatal and postnatal.Prenatal growth, in the absence of visual stimuli, is controlled by genetic influences, perturbations of which lead to microphthalmia.Conventionally, postnatal growth is guided by visual stimuli as part of the process of emmetropization, ensuring that the axial length of the eye is appropriate for the optical power, errors in which lead to axial myopia.The presence of shared genes involved in both phases of ocular growth would suggest that there are underlying molecular pathways specifying eye size, which act both dependently and independently of visual stimuli.PRSS56 and MFRP mutations have been identified as a cause of familial autosomal recessive nanophthalmia (2,85,89,97,115).GWAS meta-analyses have also associated common singlenucleotide polymorphisms in PRSS56 and MFRP (rs1550094 and rs10892353, respectively) with myopia (48,125).Human MFRP, expressed in the RPE and ciliary body, encodes a transmembrane protein with an extracellular frizzled-related domain (123).Such proteins act through the Wnt signaling pathway to regulate ocular growth, differentiation, and cellular polarity (4,57).There are at least 17 identified missense and nonsense mutations in MFRP in different populations leading to reduced axial length (82,86,134).In addition to reduced axial length, patients with MFRP mutations can exhibit foveoschisis, optic nerve head drusen, and retinitis pigmentosa (86,123,150).The phenotypic spectrum is quite varied, with no genotype-phenotype correlations yet identified (134).Both zebrafish and mouse models recapitulate the human phenotype of reduced axial length and retinal abnormalities (17,31,130).Interestingly, adenoviral-based gene therapy may reverse some of these pathogenic changes in Mfrp rd6 /Mfrp rd6 mice, with resulting rescue of axial length in adult mice.
It is postulated that MFRP may act as part of a regulatory network specifying ocular axial length in conjunction with the serine protease-encoding genes PRSS56 and ADAMTS19 (64,120).PRSS56 is expressed in retinal ganglion cells and the internal limiting membrane of adult animal and embryonic eyes, at the boundary of the vitreous and retina, where a basement membrane containing laminins and type IV collagen resides (34).PRSS56 might remodel ECM via the Müller glia, which span the retina, allowing transduction to RPE.Prss56 −/− mutant mice display a phenotype of reduced axial length and, like Mfrp rd6 /Mfrp rd6 mice, demonstrate significant upregulation of Adamts19 in the retina (64,85).Interestingly, loss of PRSS56 or MFRP function prevents excessive ocular axial growth in a mouse model of early-onset myopia caused by a null mutation in Irbp, further supporting the evidence of their role in axial elongation (64).
TMEM98 encodes a transmembrane protein expressed throughout the body and in the RPE, ciliary body, and iris of the adult eye (3).Two missense variants, p.A193P and p.H196P, and a small exon 4 deletion that also involves the adjacent intron are associated with autosomal dominant nanophthalmia; however, GWASs have identified 5 variants associated with myopia (3,61,94,125).One of the 5 variants, rs10512441, located 15 kb upstream of the TMEM98 transcription start site (70), has strong evidence of regulatory potential, locating within a DNase I hypersensitive region reported in many different cell types (70).Such potential regulatory intragenic variants may explain the divergent ocular phenotypes seen from the same gene.No dominant phenotype is seen in mice when the human disease-causing missense variants are introduced; however, mice with homozygous or compound heterozygous introduction of these mutations displayed retinal folding but otherwise normal eye size (20).Complete loss-of-function mutations in mouse Tmem98 are lethal, but selective loss of Tmem98 in mouse RPE produces a greatly enlarged eye phenotype, mimicking axial myopia, with expanded and thin retina and sclera, although the sclera remains microscopically normal (19).Furthermore, TMEM98 inhibits self-cleavage of MYRF, mutations in which cause nanophthalmia, as part of a postulated regulatory mechanism in eye size specification (19).
Finally, BMP4, part of the TGFβ family, is pivotal in the development of the optic vesicle and lens placode via LHX2-regulated expression, as demonstrated in mouse models (33,62).Mutations in BMP4 are typically associated with microphthalmia, syndromic 6 (MCOPS6; OMIM 112262) (101).Novel heterozygous BMP4 truncation mutations (c.43delC, c.97A>T, c.419delT, and c.766C>T) have, however, recently been identified to cause a phenotype characterized by pathological myopia in eight patients from four Chinese families (51).The bidirectionality regarding BMP4 and its role in axial length requires further study.

FUTURE PERSPECTIVES AND CONCLUSION
Molecular determinants of axial length remain poorly understood, although there is growing evidence for a wealth of genetic influencers.Both myopia and microphthalmia are highly genetically heterogeneous groups of diseases that display significant phenotypic variability.GWASs have provided clues to the genetic basis of myopia, but single-nucleotide polymorphisms explain only part of the heritability (i.e., the missing heritability problem), with a large proportion of variants with small effects yet to be discovered.Whole-exome sequencing has predominantly been used to investigate rare variants to date, with subsequent low diagnostic yields.Given that exomes account for only 1-2% of the human genome, patients need to be recruited for whole-genome sequencing as a priority to allow interrogation of noncoding DNA and regulatory elements.Furthermore, the complex interactions of numerous ocular tissues, along with questions about the tissue origin of the primary pathology, with much of the process occurring embryonically, make studying disorders of axial length particularly challenging.This also has implications for any potential therapy, particularly in identifying the tissue targets and age of intervention.
Much progress has been made over the last decade regarding gene discovery; however, there remains a large disconnect between these gene effectors and the actual molecular pathways involved, particularly with regard to myopia signaling cascades.Given the increasing prevalence of myopia and associated disease burden, understanding the pathways that determine eye size must be a priority.
Research on disorders of axial length has promise in elucidating the mechanisms of preschool childhood myopia, which will also better inform gene-environment interactions that underpin later-onset common myopia.In microphthalmia, many monogenic cases have been identified, but diagnostic yields remain low.Further modes of inheritance need to be considered, as well as the influence of genetic modifiers in susceptible individuals.If shared pathways are identified, then it is possible that eye growth could be manipulated using a therapy that could encourage axial growth in those with microphthalmia in the very early postnatal period; conversely, a molecular inhibitor or activator could be used to retard excessive elongation, thus reducing the morbidity of high or pathological myopia.More extensive studies identifying novel genes, expanding mutations, genotype-phenotype correlations, and biological mechanisms utilizing human and animal models are required.Such studies will ultimately provide better targeted patient management, including prevention and potential treatment of visual loss.

Figure 1
Figure 1 The development of the eye and features of the myopic eye.(a) Early development of the optic vesicle at gestational day 22.(b) Invagination of the optic vesicle, which occurs by gestational week 5, and scleral development from the embryonic mesenchyme and neural crest from inside to outside (small arrows) and anterior to posterior (large arrows), which occurs from gestational week 6 onward.(c) Postnatal eye with all intact structures.The bilayered optic cup with defined neural retina and retinal pigment epithelium formed at gestational week 7.The lens formed from the hollow lens vesicle, and the cornea developed from the overlying surface ectoderm.(d) Macroscopic morphological features of the myopic eye (purple text) and complications of high myopia (green text).

Figure 2 (
Figure 2 (a) External eye photograph of right-sided microphthalmia.(b) Slit lamp image of a microphthalmic right eye with an associated anterior segment developmental anomaly.(c) B-scan ultrasonography in microphthalmia, showing an axial length of 14.6 mm.(d) Pseudocolor fundus image from the right eye of a patient with high myopia, showing posterior pole and peripapillary retinal atrophy and vitreous opacity.(e) Optical coherence tomography of myopic macular degeneration with active choroidal neovascular membrane in high myopia.

TMEM98Prenatal:Figure 3
Figure 3Shared genes involved in both microphthalmia and myopia and their expression in prenatal eye development (red text) and the postnatal eye (black text and leader lines).Further studies are required to determine the role of these genes in human early eye development and postnatal growth.