Animal Model Contributions to Primary Congenital Glaucoma

Primary congenital glaucoma (PCG) is an ocular disease characterized by congenital anterior segmental maldevelopment with progressive optic nerve degeneration. Certain genes, such as cytochrome P450 family 1 subfamily B member 1 and latent TGF-β-binding protein 2, are involved in the pathogenesis of PCG, but the exact pathogenic mechanism has not yet been fully elucidated. There is an urgent need to determine the etiology and pathophysiology of PCG and develop new therapeutic methods to stop disease progression. Animal models can simulate PCG and are essential to study the pathogenesis and treatment of PCG. Various animal species have been used in the study of PCG, including rabbits, rats, mice, cats, zebrafish, and quails. These models are formed spontaneously or by combining with genetic engineering technology. The focus of the present study is to review the characteristics and potential applications of animal models in PCG and provide new approaches to understand the mechanism and develop new treatment strategies for patients with PCG.


Introduction
Primary congenital glaucoma (PCG) is one of the leading causes of childhood blindness and is a characteristic of incomplete development of the trabecular meshwork (TM) and anterior chamber angle with optic nerve degeneration [1,2]. Worldwide, the overall incidence of PCG is 1-80 patients per 100,000 live births [3]. In the highly consanguineous subpopulation of Slovakia, the incidence of 1 : 1250 individuals has been reported [4]. e incidence of PCG in offspring may be 5 to 10 times higher if there is consanguinity in the parents [2]. PCG cases are unilateral or bilateral, showing variable penetrance (40-100%) and various severities [5]. Elevation in intraocular pressure (IOP) is a common risk factor that threatens the survival of retinal ganglion cells (RGCs) and favors disease progression. e mechanism to maintain IOP is a normal circulation of aqueous humor (AH) through the drainage pathway ( Figure 1).
Various animal models of glaucoma have been described in other review articles [20,21]. At present, animal models used in PCG include rabbits [22], rats [23], mice [24], cats [25], zebrafish [26], and quails [27]. e study of these models provides valuable information on the pathophysiology of PCG, as it is associated with changes in the angle of the anterior chamber, the optic nerve, and the retina, although the mechanisms that caused these changes are unclear. e purpose of the present study is to review the pathological features of animal models and the possible genetic mechanisms associated with PCG, which would provide references for future modeling and pave the way for the elucidation of the pathogenic mechanism of PCG.

Genes Related to PCG
Although exact causative genes and pathways have yet to be confirmed, several related genes, including CYP1B1, seem to be involved in the development of the anterior segment of the eye, and they may participate in the progression of PCG (see Table 1).

CYP1B1.
e biallelic variants of CYP1B1 represented approximately 64.8% of PCG cases in the Middle East and 54.4% of PCG cases in the Maghreb [41]. Similarly, the rate of CYP1B1-mutated alleles was 34.7% in Europe, 21.3% in Asia, and 14.9% in the United States [41]. ere exists a strong clinical genotype/phenotype correlation with some variants causing PCG with histopathological features, while A330F caused by c.988G>T&c.989C>T and R390H caused by c.1169G>A are associated with the severe glaucomatous phenotype [42].
Gene product loss-of-function (LoF) caused by CYP1B1 mutations has been reported to be involved in the development of TM of PCG. e CYP1B1 enzyme metabolizes a signaling molecule required in the development of ocular tissues (possibly endogenous steroid metabolism) and eliminates toxic metabolites [10,43,44] involved in the cAMP/protein kinase A pathway together with MYOC to affect TM formation [45][46][47]. However, the endogenous target in vivo is still unknown. In zebrafish, CYP1B1 was expressed in the dorsal and ventral retina in an ocular fissure, overlapped with retinoic acid (RA) synthesis enzymes (RALDH2 and RALDH3) [28,48]. CYP1B1 is a component that mediates the conversion of vitamin A (Vit A) into the retina and then into RA by the dehydrogenase-independent pathway [48]. e overexpression of CYP1B1 inhibited ocular fissure closure, causing colobomatous defects [28]. By comparison, CYP1B1 knockdown led to the premature closure of inferior optic fissure and changes neural crest (NC) later migration via the RA-independent pathway [28]. Furthermore, the formation of iris stroma derived from the crest and the recovery of retinal development were delayed in the larval stage in cyp1b1-knockdown zebrafish [49]. Null CYP1B1 activity altered the expression of ECM, lipid metabolism genes, and inflammation [50].
Other findings reported that CYP1B1 is an important regulator of redox homeostasis, which is related to increased oxidative stress in trabecular meshwork cells (TMCs) and retinal vascular cells in PCG [51]. And the increased oxidative stress alters the production of ECM protein periostin (Postn) that influences the cellular integrity and function of TM [51,52]. Postn is an ECM protein that interacts with collagen fibers and is involved in TM's function and morphology [53]. Moreover, CYP1B1 metabolite, arachidonic acid, regulates the transparency of the cornea and the function of the Na + -K + ATPase in corneal microsomes [54,55]. us, the later migration of NC mediated by CYP1B1 in proper spatialtemporal scope is essential for ocular tissues to form. Also, CYP1B1 contributes to the maintenance of the homeostasis of structure-function of TM tissue through regulating the expression of Postn [29].

LTBP2.
Ali et al. reported that p.R299X in LTBP2 was the main variant of the founder of PCG in the Gypsy population, which accounted for more than 50% of CYP1B1negative cases and ∼40% of all PCG cases [15]. LTBP2 is identified as a PCG-associated causative gene and is expressed in the TM and ciliary body, particularly in the ECM protein of the ciliary body [15,31]. e homonymous protein encoded by LTBP2 is an ECM protein and is thought to be a member of latent TGF-β complex and a structural component of microfibrils. LTBP2 plays an important role in cell adhesion and supporting ciliary muscle tone, binding to fibrillin-1 through the carboxyl-terminal (C-terminal) and fibulin-5 through the amino-terminal (N-terminal) [56,57]. Ali et al. demonstrated that LTBP2 mutations increased the elasticity of ciliary body tissue and altered the structural support of the context tissues [15].
Furthermore, LTBP2 may indirectly negatively regulate the concentration of the large latent TGF-β complex on microfibrils [56]. A hypothesis is that LTBP2 competes with LTBP1 to bind to the same binding site in fibrillin-containing microfibrils, leading to the release of LTBP1 from microfibrils [56,58]. By strong noncovalent bonds, TGF-β binding to latency-related protein (LAP) with LTBP1 forms the large latent complex (LLC) [58]. When LAP is integrated by integrating the cell surface, LTBP1 promotes the activation of TGF-β via fixing to the ECM to produce traction [58]. is traction leads to deformation of the LAP to release active growth factor [58]. Active TGF-β is released and BMP binds to the domain of their receptors, and phosphorylation of the serine/threonine kinase of receptors results in activation of the SAMD pathway to transmit the signal (Figure 2) [58,59]. Suri et al. confirmed that LTBP2 had a putative effect on TGF-β signaling and the ECM of TM [32]. LTBP2 knockdown increased samd2 phosphorylation and decreased phosphorylation of samd1/5/8, suggesting LTBP2 knockdown promotes TGF-β signaling and inhibits the bone rs35717904, c. * 734A>T; rs185790394, c.-244C>T; and rs79691946, c. * 454C>T) in 10 patients without a family history of glaucoma, and this indicated that FOXC1 variants may contribute to the formation of goniodysgenesis in PCG [62]. e expression of FOXC1 is in periocular mesenchyme (POM) derived from the NC and appears to be related to the development of iridocorneal angle. Seo et al. showed that NC-FOXC1 −/− reduced the expression of paired like homeodomain 2 (PITX2) and downstream effector dickkopf WNT signaling pathway inhibitor 2 (DKK2), suggesting that FOXC1 as a regulator in WNT signaling regulates and maintains ocular development [33]. In addition to that the expression of PITX2 and FOXC1 depends on RA signaling in the POM [34,63,64].
2.4. ANGPT1 and TEK. ANGPT1/TEK plays an essential role in vascular development and is also implicated in the development of Schlemm's canal (SC). Dysplasia of SC and TM in AH drainage was frequently observed in PCG patients with elevated IOP [65]. is indicates that the ANGPT1/TEK pathway is involved in the development of AH drainage structures in PCG. Souma et al. identified 10 rare variants of TEK in 189 pedigrees with PCG, and each variant resulted in haploinsufficiency due to the LoF protein [17]. Similarly, omson et al. reported that 3 variants (p.Q236 * , p.R494 * , and p.K249R) of ANGPT1 were identified in the cohort of 284 PCG patients, and ANGPT1 mutations reduced the ability to activate TEK signaling [66].
TEK, a receptor tyrosine kinase, regulates vascular homeostasis by its autophosphorylation and transphosphorylation, which is highly expressed in the endothelium of SC [17,67,68]. ANGPT1, which is a primary ligand expressed in pericytes and other vascular supporting cells, plays a significant role in proangiogenic and vascular stabilization by activating TEK [66]. en, vacuoles formed Figure 2: LTBP2 indirectly mediates the TGF-β/BMP signal pathway to regulate elastogenesis or anchors to components of the matrix of the trabecular meshwork. Mutant LTBP2 loses competition with LTBP1 for fibrillin-containing microfibrils, which increases the concentration of activated TGF-β and upregulates TGF-β signal to activate target gene. BMP signal negatively regulates the TGF-β signal to downregulate TGF-β gene expression. LTBP2: latent transforming growth factor-beta binding protein 2; TGF-β: transforming growth factor-beta; BMP: bone morphogenetic protein; LAP: latency-related protein; TGF-βR: transforming growth factor-beta receptor; BMPR: bone morphogenetic protein receptor. from the endothelium of SC, which subsequently resulted in SC endothelial thinning and formation of pores to induce the normal flow of AH [69,70]. It has been reported that disruption in the ANGPT1/TEK pathway possibly caused PCG-like phenotypes, such as elevated IOP, RGC degeneration, and buphthalmos [71]. Several studies have shown that the ANGPT1/TEK pathway supports and maintains the stability of the function of SC and TM, which is one of the pathological mechanisms of PCG [17,66].

GPATCH3.
Ferre-Fernández et al. identified two heterozygous variants (c.701A>G, p.Asn234Ser and c.1424G>A, p.Gly475Glu) of GPATCH3 in a PCG family [18]. e disease phenotype in the PCG family caused by GPATCH3 variants was autosomal recessive inheritance [18]. It was predicted that c.701A>G, p.Asn234Ser and c.1424G>A, p.Gly475Glu affected highly conserved amino acid residues and were disruptive [18]. e variants of GPATCH3 showed that nuclear localization was similar to FOXC2. Embryonically, the pattern of GPATCH3 expressed is similar to that of PITX2 and overlaps with FOXC1 presented in the POM [18]. GPATCH3 encoded a protein with the G-patch domain, which was present in both RNAbinding and DNA-binding and participated in proteinnucleic acid interactions [72]. It is assumed that GPATCH3 regulates iridocorneal angle development by activating the GPATCH3 protein that activates the promoter of the C-X-C motif chemokine receptor4 (CXCR4) to regulate DNA sequences or indirectly mediates the mechanisms involved in PCG [18]. Compared with the wild-type protein (WT), the affected protein increased ∼17% of the transactivation activity of CXCR4, suggesting that changes in amino acids caused hypermorphic variants and led to functional alterations of the affected protein [18]. CXCR4 is linked to NC migration through encoding chemokine receptors [38]. is provides evidence that the function of GPATCH3 may interact with PITX2 and FOXC1 and that GPATCH3 is implicated in PCG [18].

GUCA1C.
Morales-Cámara et al. identified one homozygous variant of GUCA1C (c.52G>T; p. Glu18Ter) in two affected siblings with PCG by an autosomal recessive pattern [19]. It is predicted that c.52G>T in GUCA1C resulted in complete LoF of the gene product [19]. e GUCA1C-encoded guanylate cyclase 3 activating protein (GCAP3) belongs to the member of the guanylate cyclaseactivating protein family, which is related to phototransduction and regulates IOP [19,39,73]. GCAP3 was expressed in the iris tissues, corneal tissues, retina, photoreceptors, and RGCs, suggesting that GCAP3 regulated the homeostasis of IOP and that GCAP3 LoF may contribute to congenital glaucoma [19]. Furthermore, GUCA1C reduces IOP by regulating the volume of SC and TMCs in a cGMP-dependent manner to increase the outflow of AH [19,39]. However, to prove whether the GUCA1C variant causes PCG, more experimental data are required [40].

Treatment Methods for PCG
Currently, clinical treatments for PCG include surgery and drug-assisted treatment. Carbonic anhydrase inhibitors (CAIs) are aimed at inhibiting the secretion of bicarbonate ions and reducing the flow of liquid to reduce the production of AH in the ciliary body [74,75]. Compared with β-receptor blockers, CAIs effectively reduce IOP by ∼25% with fewer side effects but will cause anorexia, thirst, fatigue, renal acidosis, and growth retardation in the long term [76,77]. Furthermore, certain medications, for example, latanoprost, the prostaglandin F2α (FP) receptor agonist, increase the amount of AH in the uveoscleral outflow pathway [78,79]. Uva et al. pointed out that the mean reduction in IOP controlled by latanoprost in children was 35.6% and lasted 18 ± 22 months, but adverse events included conjunctival hyperemia, increased pigmentation of the iris, changes in eyelashes, and irritation of the upper airways [80].
So far, almost all cases of PCG require initial surgery because medications are not always effective or feasible and are poorly tolerated long term in pediatric patients [2,81]. Goniotomy is the first choice in PCG because the conjunctiva is preserved and therefore does not jeopardize the success of procedures in the future [82]. e advantage of trabeculotomy is that it possibly is used in eyes with corneal opacity and may have a higher success rate than goniotomy [81,83]. Trabeculectomy is performed after other procedures failed, with the overall goal of allowing AH to outflow from the anterior chamber into the sub-Tenon space [84,85]. e success rate of trabeculectomy varies from 55.3% to 92.3% and has a high rate of complications, such as endophthalmitis and other herpes-related infections (6.7%) [81]. More procedures also are performed to increase AH outflow for PCG, for example, filtration surgery, cyclodestructive procedure, and combined trabeculotomy and trabeculectomy [86][87][88].
e advantages and success rates of these procedures are varied, and their application should be weighed against the effectiveness of each method and the complexity of the disease [81]. Preoperative and postoperative medications are used to help control IOP and prevent refractive error, amblyopia, and other secondary complications [86,89,90]. Importantly, long-term follow-up and lifelong control of IOP are necessary. e independent anatomical structure of the eyes, easily transduced cell population, and the unique anatomical barrier and physiological environment of the eyes constitute suitable conditions for gene therapy [91]. e preferred target tissue to reduce IOP is TM while for neuroprotection is RGCs [92]. e progress made in gene therapy brings hope to the treatment of ophthalmic diseases, especially in PCG. Recently, advances and progress made in gene therapy of glaucoma bring hope to the treatment of PCG [93]. Connective tissue growth factor was reported to be inhibited by the AAV-CRISPR-Cas9 system in rabbit glaucoma, which improved the outcomes of glaucoma filtration surgery [94]. Yang et al. reported that delivery of exogenous p27 to the rabbit model would inhibit the proliferation of Tenon's capsule fibroblasts at glaucoma surgery sites and the formation of scars, thus improving the efficacy of surgery [95].

Journal of Ophthalmology
Silencing the P2Y 2 receptor by siRNAs in rabbits reduced IOP by 67%, and the treatment effect lasted for 5 days [96]. Similarly, silencing of GAS5 relieved symptoms of rat glaucoma by rescuing RGC apoptosis [97]. Jiang et al. pointed out that treatment of glaucoma in a rat model with superoxide dismutase (SOD2) mediated by recombinant adeno-associated virus (AAV) may protect RGCs from chronic IOP elevation damage [98]. Zhou et al. reported that AAV-mediated Map2k1 gene transfer into the RGCs of rats significantly increased neurons' survival from 680 ± 86 RGCs/mm 2 to 1366 ± 70 RGCs/mm 2 [99]. Similarly, Luna et al. indicated that delivery of miR-146a provided a longterm reduction of IOP without inflammation in a rat model [100].
ese findings suggest that gene therapy may be promising for patients with PCG. However, gene therapy for PCG has a long way to go, and identifying the causative gene is a prerequisite for it.

Animal Models Related to PCG
Currently, animal models are divided into two categories, the spontaneous model (see Table 2) and the genetic model (see Table 3), which also elucidate the clinical phenotype and pathological mechanisms of PCG from different perspectives.

Spontaneous Models
e rabbit model with PCG is mainly spontaneously inherited. Congenital abnormalities in anterior chamber development in rabbits with PCG were similar to those in humans. Abnormalities included corneal epithelial (sometimes stromal) edema, larger corneal diameter, hyperemia of the ciliary body, compression or loss of the iris pillars, increased cell cornification, and hypoplasia or posterior displacement of the aqueous plexus [101,102]. Additional findings included loss of association of trabecular endothelial cells to cells, disorganization of trabecular lamellae, compression or dilation of intertrabecular spaces, and a decrease in the number of collagen fibers [102]. Others reported that TM was replaced by abundant ECM, unidentified round cells below the aqueous plexus, and basement membrane-like material [114].
In the rabbit model, except for the typical manifests mentioned, variable elevation of IOP (up to 48 mmHg), retina nerve degeneration, and outflow facility defect were observed [101,120]. An important finding was that there was no marked blockage of axonal transport in rabbits when IOP was elevated acutely because rabbits lack a true lamina cribrosa (LC) [120]. is was consistent with the hypothesis that elevated IOP mechanically damages the optic nerve. PCG in rabbits is likely to be inherited through an autosomal recessive pattern with incomplete phenotypes and penetrance [121]. Gene mutation in rabbits inactivated genes of TM development and eventually led to maldevelopment of TM tissue [22,101]. In addition, a lack of vitamin A may be associated with the severity of PCG, as it participates in the stability of the cell's membrane structure and maintenance of the optic nerve [101]. Furthermore, in rabbit models, abnormal expression of AH proteins in PCG led to a decrease in the number of RGCs, thickening of Descemet's membrane and anterior lens capsule [103].
However, interest in the rabbit model gradually decreased as genetic models of PCG became available. Ishida et al. performed trabeculectomy and Ex-PRESS filtering to treat PCG in rabbits, respectively, and demonstrated that the overall effect of trabeculectomy on blood pressure reduction was relatively stable [122]. e inflammation of Ex-PRESS filtering was milder than that of trabeculectomy. Furthermore, nanostructured glaucoma drainage implants have also been successful in rabbit models, showing good tolerance, minimal leakage of implants, good compatibility with cells and tissues, and IOP reduction by 33-44% [123]. ese studies suggest that the rabbit model is ideal for simulating surgical treatment.
e rabbit model of PCG should avoid the interference of conjunctivitis, which is a common disease in rabbits caused by infectious and/or noninfectious diseases during the experiment [124]. And the intermittent elevation of IOP may lead to unstable experimental results [114]. Furthermore, distinctions in the anatomical structure of TM and the drainage route of AH make it difficult to establish an anterior chamber angle association between rabbits and humans [125]. e SC drainage function replaced by the aqueous plexus in rabbit models restricts the ability to study the genetic mechanism of PCG [103]. In addition, genetic testing in rabbits is restricted due to the limited rabbit genomic resource available for testing [103]. e CYP1B1 mutation in rabbits could not be identified [103].

Rat Model.
In spontaneous rat PCG models, like rabbits, it showed symptoms similar to humans [23,105,106]. Additionally, the rat model showed unilateral or bilateral enlarged globes with the IOP ranging from 29 to 42.5 mmHg [105,107]. After 18 months, the rat model showed that the remaining RGC was 92 ± 26 RGC/ mm 2 , while the WT group was 1887 ± 117 RGC/mm 2 [107]. Furthermore, the condition worsened with age, showed mitochondrial dysfunction, lack of nutritional proteins, accumulation of toxic substances, and oxidative stress [107]. Others reported upregulation of c-Myc in the RGCs and neurons of the inner nuclear layer promoted cell differentiation and induced cell apoptosis [106]. Elevated levels of heat shock protein 27 and VEGF in RGCs exerted a protective effect on RGCs, while the upregulation of glial fibrillary acidic protein, growth-associated protein 43, and endothelin-1 in glial cells promoted the apoptosis of RGCs [106,107]. Additional findings included large anomalous tissue deposition accumulated in the suprachoroidal space, and the uveoscleral outflow pathway was severely impaired [108]. ese rat models used in PCG offer valuable information on both RGC loss and uveoscleral outflow.
Although the rat is a common tool for studying disease in the laboratory, some questions need to be considered. First, the TM of rats is less intensive than that of humans and cannot generate enough flow resistance because it is a few layers thick [109,126]. Second, differences in LC and vascular supply to the optic nerve head (ONH) between rats and humans need to be taken into account compared to the human situation [109]. In addition, rodents, such as rats and mice, do not have a fovea, their LC is faintly present and also both choroid and pial vessels possibly exist [104,109]. Hence, the rat is not suitable to study LCspecific function.

Cat Model.
Cats with PCG showed IOP > 30 mmHg (normal IOP: 20 ± 5 mmHg), and those with nonspecific conjunctivitis were likely to suffer from glaucoma [25]. e red eye is an early sign of PCG in cats and indicates the optimal time for treatment [25]. In the spontaneous cat model, arrested vascular development, posterior  displacement of intrascleral vessels, and loss of the RGCs resembled human PCG [25,110,111,127]. e IOP in cats continued to increase until at least 6 months of age [127]. Elevated IOP ranged from 30 to 40 mmHg with intermittent IOP peaks of 50 to 70 mmHg [111]. e mean average retinal nerve fiber layer (RNFL) thickness of cats with PCG ranged from 8 to 20 μm (normal RNFL is 40 μm) [110].
Kuehn et al. reported the first case of hereditary glaucoma with complete penetrance in a domestic cat, which was caused by LTBP2 mutation [112]. e sequence homology between cat and human orthologs reaches 87% over 100% length [112]. e LTBP2 mutation in the cat model was shown to be autosomal recessive inheritance, and the postnatal development of the AH drainage system of the anterior segment was arrested [112]. Except for elevated IOP, cats were reported to have disrupted microstructural integrity [128]. e clinical appearance in cats with PCG included iris hypoplasia and ectopia lentis [112]. An open, slightly narrowed iridocorneal angle and mild maldevelopment of the pectinate ligament were observed by goniophotography [112]. e ONH cupping and degeneration in cats with advanced PCG were observed by fundus photographs [112]. Compared ∼83398 optic nerve axons in normal cats, PCG-affected cats had a moderate loss of axons, which remained ∼30365 optic nerve axons [112]. e AH outflow pathway composed of the uveal, corneoscleral TM, and angular aqueous plexus appeared collapsed [112]. Genetic analysis of the B3 chromosome LTBP2 in domestic cats showed a 4 bp insertion and deletion in exon 8, resulting in a frame shift mutation that produces a truncated protein; the protein containing an abnormal C-terminal may accumulate in cells or become degraded and participate in the development of PCG [112].
However, in cats, there is no structure similar to human SC, and the uveoscleral route accounts for less than 3% of the AH outflow [129]. PCG also progresses slowly and degeneration of the ciliary body may cause inaccurate IOP measurements in cats [113]. Furthermore, creating cat disease models for PCG research is complicated by the need to remove various mixed phenotypes to guarantee a mono-disease reproductive population [111,130].

Quail Model.
e pattern of inheritance in albino mutant quails is sex-linked recessive (al mutant) [27]. Albino mutant quails exhibited clinical features similar to other spontaneous glaucoma animals, such as enlarged eyes, RGC degeneration, and optic disc cupping [27]. e mean IOP in the al mutant quail was significantly higher than that in the normal quail at six months of age (25.2 ± 2.2 vs. 17.3 ± 1.7 mmHg), and some of them showed markedly elevated IOP that was over 35 mmHg at one year [27]. e opened iridocorneal angle, the shallowed anterior chamber, and the flat corneal curvature were observed in al mutant quails at 3 months [27]. In addition, in the ganglion cell layer, 11%-55% of the total number of RGCs decreased, resulting in a reduction of the iso-density map in the retina [115]. At the advanced stage, almost all RGCs had disappeared and the eye of the quail model showed anterior synechia [27,115].
us, the quail model seems suitable for studying spontaneous glaucoma because of easy maintenance and handling in the laboratory [27]. However, the small size of the quail cornea makes IOP measurement a challenge [114,115]. Furthermore, the al mutant has not been described in detail [114].

Genetic Models.
No further reports have been published describing the clinical, pathological, and genetic characteristics of PCG in spontaneous models. Genetic models have validated some genes with altered expression and potential pathogenic signaling pathways in human PCG.

Mouse Model.
Mice are becoming the most common experimental animals in the laboratory because of their advantages, such as rapid reproduction, easy breeding, the highly conserved genome between mice and humans, and the availability of genetic manipulation [131]. e morphology of the iridocorneal angle and the development sequence of ocular structures in mice are similar to those of humans [109].
(1) Cyp1b1-Deficient Model. e Cyp1b1 sequence homology between mice and human orthologs reaches 83% over 100% length (https://blast.ncbi.nlm.nih.gov/). A Cyp1b1-deficient mouse model has been developed to simulate PCG to further understand the molecular mechanisms. Cyp1b1 −/− mice also showed a genotype/phenotype correlation that was similar to that in PCG patients [30]. During the early stages of development of mice, spatiotemporal expression patterns showed that Cyp1b1 was involved in the establishment and (or) maintenance of polarity along the axis of embryonic development [132]. e expression of Cyp1b1 was in the retina, in the tissues around the optic nerve, and the developing and mature ciliary body of the mice eyes [24].
In Cyp1b1-deficient mice (Cyp1b1 −/− ), both the elevation of IOP and the ultrastructural abnormalities of ECM in TM were similar to those of human PCG [29,30]. e mean IOP in Cyp1b1 −/− mice was 11.45 ± 0.20 mmHg, which was higher than that in Cyp1b1 +/+ mice (10.14 ± 0.07 mmHg). At 3 weeks, Cyp1b1 −/− mice showed obvious disruption of the anterior and posterior TM, which were characterized by multifocal atrophy of the trabecular beams with the expansion of the intertrabecular spaces [30]. Marked fragmentation and irregular distribution of collagen fibers on the collagenous layer of the trabeculae were observed [30]. TMCs showed irregular cytoplasmic processes, a lack of attachment to the basement membrane, cytoplasmic vacuolation, and accumulation of organelle debris within the cytoplasm [30]. With increasing age, Cyp1b1 −/− mice exhibited progressively severely disarranged trabecular collagen. At 8 months, Cyp1b1 −/− mice exhibited complete collapse of the TM and formed an irregular atrophic trabecular beam across the anterior and posterior TM [30]. In addition, the relative amount of TM collagen was less and progressively decreased with age in Cyp1b1 −/− mice [30].
Zhao et al. also found severe morphological disruption in collagen fibers and increased intracellular oxidative stress in Cyp1b1 −/− TM tissues [29]. In response to stress, Cyp1b1 −/− TMC trended to apoptosis and the secretion of Postn was decreased [29].
Falero-Perez et al. determined the role of Cyp1b1 −/− in retinal endothelial cells (EC), pericytes (PCs), or astrocytes (ACs) on retinal neovascularization by developing cell-specific targeted Cyp1b1-transgenic mice [133]. During oxygeninduced ischemic retinopathy (OIR), attenuation of the pathological retinal neovascularization of PC-Cyp1b1 −/− mice was similar to the phenotype previously observed in Cyp1b1 −/ − mice worldwide [133]. Also, Cyp1b1-transgenic mice exhibited destruction and loss of ECM and the degenerative cellular alterations in TM tissues [133]. is indicated that the expression of CYP1B1 in perivascular supporting cells and EC affected the development of ischemia-mediated retinal neovascularization and maintains the integrity of TM by regulating cellular redox homeostasis. However, this work did not report the abnormal IOP in these mice.
(2) Tyr-Deficient Model. Two mutant genes (Cyp1b1 and tyrosinase (Tyr)) in the mouse model showed that Tyr served as a modifier in maintaining the phenotype of the drainage structure and was possibly associated with the severity of PCG [116]. is model showed that Tyr deficiency increased the magnitude of the dysgenesis of the structure of the ocular angle, altered the penetrance linked to the inheritance of mutant orthologs of two known human PCG pathogenic genes (CYP1B1 and FOXC1) [116]. Furthermore, oral administration of L-dihydroxyphenylalanine (L-dopa) to pregnant mice improved partial angle dysgenesis of both Cyp1b1 −/− mice and Foxc1 −/− mice [116]. erefore, the pathway involving L-dopa and tyrosinase participates in angle formation or the function of AH drainage structures or functions in the time of prenatal development in mice [116].
TYR as a modifier of the phenotype of PCG caused by Cyp1b1 mutations has not been confirmed in humans [117]. A study in PCG patients found that 16 of 19 nonpermeable families had the same homozygous CYP1B1 mutations (G61E), resulting in changes in conserved amino acids but did not interrupt protein synthesis as in Cyp1b1 −/− mice [116,117]. By analysis of single nucleotide polymorphisms (SNPs), individuals from these nonpenetrant families did not show a significant link to the TYR locus [117]. Most of the subjects in this experiment carried recessive missense mutations in CYP1B1, and some mutations had been shown to have residual enzymatic activity [117,135]. erefore, the limitation of the Tyr −/− mouse model is that the role of TYR in the eye of humans is different from that of the mouse model, leading to inconsistencies in the role of Tyr in CYP1B1-induced PCG between mice and humans.
(3) Angpt1-Deficient or Tek-Deficient Model. Mouse models harboring a condition by inversion alleles (Tek COIN mice) have been developed [17]. In transgenic mice (LoF angiopoietin-Tek signaling), PCG-like phenotypes were observed [17]. In Tek hemizygous and conditional knockout (KO) mice, a 25% elevation in IOP was found in haploinsufficient mice (Tek +/− ), and IOP levels were moderately related to the degree of SC hypomorphism [17]. Tek +/− mice exhibited severe convolutions and focal narrowing of the SC and hypoplastic TM [17]. is evidence supports that the dose of the TEK gene plays a significant role in the proper development of AH outflow, and loss of TEK signaling underlies PCG with variable expressivity.
ose data demonstrated that Angpt1 mutation in the ANGPT/TEK signaling axis is one of the PCG pathogenic mechanisms.
Mice with PCG may provide evidence that abnormal development in anterior chamber structures and glaucomatous neuro-changes play a key role in revealing the genetic mechanism for PCG. However, abnormal IOP was not reported in some Cyp1b1 −/− or Tyr −/− mouse models. However, compared with humans, the LC is not present, and blood supply for the ONH region is not abundant (absence of choroid and pial vessels), and the number of RGC is less in mice [109]. ereby, the mouse model is not suitable for the study of specific structures and functions [136]. Moreover, most mouse models are single-gene models, which pose certain limitations regarding the study of multigene interaction diseases in PCG [118].

Zebrafish Model.
Ortholog genes are shared among the zebrafish, humans, and mice, more than 70% of human genes have at least one ortholog in zebrafish, and 47% of orthologous genes show a one-to-one relationship in humans and zebrafish [137]. e biologically conserved nature of anterior segment development between zebrafish and humans provides an opportunity for studying inherited Journal of Ophthalmology eye diseases in recent years [26]. In addition to maintaining significant evolutionary proximity to humans, zebrafish possess unique advantages such as high fecundity, rapid extrauterine development, and transparency during organogenesis, making them ideal for high-throughput drug screening, mechanistic studies, and behavioral genetics [138]. e available genetic manipulation techniques, such as Crispr/Cas9, easily generate zebrafish models with mutation of target genes [139].
(1) Cyp1b1-Deficient Model. Cyp1b1 LoF has been studied in zebrafish mainly through morpholino (MO)-mediated knockdown [28,140]. By this approach, protein expression is inhibited in the early stages of development, and the development of tissues derived from NC cells is affected, further exploring the function of cyp1b1 in early embryonic development [49].
Williams et al. reported that overexpression of cyp1b1 inhibited ocular fissure closure and caused craniofacial and ocular defects through the RA-independent pathway [28]. Importantly, administration of human CYP1B1 mRNA to embryos resulted in large colobomas in 71.8 ± 17.9% of embryos and disrupted NC-derived tissue formation, supporting evolutionary conservation of cyp1b1 function between zebrafish and humans [28]. Additionally, at 4 hours postfertilization (hpf), the volume of embryos from cyp1b1-KO zebrafish was 60% of that of the WT group [50]. However, at 168 hpf, no significant histological differences in glaucoma-related structures were observed between cyp1b1-KO zebrafish and the WT group [50]. Similar to previous zebrafish models, cyp1b1-KO zebrafish exhibited variable craniofacial defects, which may be caused by dysregulation of ECM gene expression induced by cyp1b1 disruption [50]. e limitation of cyp1b1-KO zebrafish is that they do not show a PCG-like phenotype but provide information that tyrosinase does not modify the effect of cyp1b1 on eye development [28,50]. It further illustrates that the relationship between TYR and CYP1B1 may vary from species to species.
(2) Foxc1-Deficient Model. Foxc1a and foxc1b of zebrafish are orthologs of human FOXC1 and FOXC2, respectively; the null mutation of foxc1 in zebrafish models with PCG was generated by Crispr/Cas9 [141]. In zebrafish, foxc1a was expressed primarily in NC cells and POM, and foxc1b was partially overlappingly expressed with foxc1a in the POM [141]. Compared with WT embryos, at 5 days postfertilization (dpf ), foxc1b −/− alone neither affected the cells counts in the RGC layer nor the thickness of the optic nerve nor inhibited foxc1a on its own [141]. en, after administration of foxc1a-MO to foxc1 −/− embryos, the number of cells in the RGC layer reduced significantly (146.1 ± 3.6/section vs. WT 163.5 ± 3.2/section) [141]. Similarly, in foxc1b −/− zebrafish with inhibited foxc1a, it showed that the thickness of the optic nerve was less than that of the WT group (9.9 ± 0.55 vs. WT12.2 ± 0.42 μm) [141]. Moreover, the loss of foxc1 homologs in zebrafish negatively reduced the expression of atoh7. After the loss of atoh7, pou4f2 expression was reduced, which determined RGC fate [64,141]. However, loss of foxc1 homologs did not reduce the level of the marker of amacrine cell fate scrt1a [141]. Zebrafish with foxc1 LoF provides valuable information associated with endophenotypes of glaucoma, suggesting the role of RGC differentiation defects in PCG.
(3) Guca1c-Deficient Model. e guca1c-KO zebrafish model was developed through Crispr/Cas9 genome editing to study the role of GUCA1C LoF in congenital glaucoma and retinal physiology [19]. Guca1c-KO zebrafish did not show significant gross external macroscopic changes and PCG-like phenotypes [19]. e thickness of the corneal epithelium in guca1c-KO zebrafish was less than that in WT zebrafish (5.1 ± 0.2 μm vs. 8.6 ± 0.2 μm) [19]. GCAP3-immunopositive signals were observed in the nonpigmented epithelial cells of the ciliary zone in both guca1c-KO and WT zebrafish [19]. In addition, GCAP3 immunoreactivity was not shown in the nonpigmented epithelium, keratocytes, photoreceptors, inner and outer plexiform layers, and the RGC layer in guca1c-KO zebrafish [19]. Glial fibrillary acidic protein was upregulated, and RGC apoptosis was observed in Müller cells of guca1c-KO zebrafish [19]. e mechanism of guca1c LoF may be through nonsense-mediated decay, leading to mRNA degradation and causing the LoF gene product, consequently disrupting the structures of SC and TM through a cGMPdependent pathway [19,142]. However, predictive data and functional evidence insufficiently support a gene-disease relationship; further studies are required to demonstrate the pathogenicity of guca1c variants in PCG [40].
(4) Gpatch3-Deficient Model. In the gpatch3 −/− zebrafish model, Gpatch3 was expressed in the corneal endothelium and POM-like cells, head cartilages, and skeletal muscles [18]. Gpatch3 −/− zebrafish exhibited variable degrees of abnormality in ocular structures and dose-dependent phenotypes [18]. At 24 hpf, 53% of gpatch3 −/− embryos showed a proportion of lethal phenotypes [18]. At 96 hpf, gpatch3 −/− zebrafish exhibited defects in ocular development relevant to glaucoma in the anterior angle and periocular tissue [18]. Dorsal angle alterations showed decreased silver and foamy cells corresponding to iridophores and xanthophores, respectively [18]. Also, undifferentiated mesenchymal-like cells were accumulated in the anterior segment angle, and hypoplastic development of the ventral angle was observed in the gpatch3 −/− zebrafish model [18]. Other features include increased periocular space and dysplasia of the pharynx and cartilage, indicating a defective ocular structure [18]. Some features resembled that observed in pitx2-KO or foxc1-KO embryos, which indicated a functional relationship between gpatch3 and pitx2 and foxc1 and provided evidence that gpatch3 was implicated in PCG [18]. However, because of the different components of the cytoskeleton and different proteoglycans in the TM between zebrafish and humans, zebrafish are not suitable for studying the ultrastructure changes of human PCG [143]. IOP elevation and loss of RGCs were not reported in zebrafish. Furthermore, the ECM was not found in the TMCs of zebrafish, and the annular ligament of zebrafish differs from the TM of humans [143].

Conclusions
Animal models with various traits play a pivotal role in exploring the pathophysiology and new treatment methods for PCG patients. Animal models are simple representations of complex physiological changes in humans. Each animal model that simulates human PCG has its strengths and weaknesses. Spontaneous animal models are mainly focused on the study of clinical manifestations, pathological changes, and treatment of PCG, providing valuable information on the external phenotype and treatment. e limitation of spontaneous animals is that no more human causative gene is given, except for LTBP2. is requires additional animal models to study the molecular mechanisms of human PCG and gene therapy approaches, as well as new molecular techniques to selectively specifically alter the expression of specific genes in specific tissues. Recently, advances in genome editing have specifically modified the expression of individual genes, which has generated new PCG models with specific gene mutations and may be introduced into therapies in the future. Genetic models (mouse and zebrafish) are studied from the causative gene to the cellular and tissue level, providing information on the pathology of PCG from embryonic development to adulthood.
But there are still some problems to be solved. First, some animal models present atypical clinical manifestations, such as a slight or no elevation of IOP. Second, not all genomes of animals are available.
ird, the relevance and "translatability" from PCG animal models to humans is limited. Furthermore, topical ocular drug administration usually causes more systemic exposure. Because of the grooming behavior and the relatively large volume of delivery versus the smaller ocular surface area of rodents, rodents rub the remaining dose of the drug from the eyes and then lick the forelimb or remaining hair [21]. Of course, minimal interference factors, reproducibility, and cost-effectiveness should also be taken into account. erefore, an excellent animal model with PCG should have the following characteristics: (i) markedly elevated IOP; (ii) similarity to the human eye in terms of anatomical structures and physiological function; (iii) optic nerve degeneration similar to human PCG; (iv) animals have available orthologs of human PCG pathogenic genes; and (v) the results of animal models specifically and credibly reflect the pathological process of human PCG.
Due to the anatomical, physiological, and genetic differences between humans and animals, the role of animal models in PCG is to reproduce one or several specific aspects of the disease to provide valuable information to the researchers.
erefore, ongoing efforts focus on the development of a more optimized animal model to simulate the etiology and effects of the treatment of human PCG. e usage of genome editing and available animal genome resources in mice and zebrafish gives a hint that optionally altering the expression of target genes in a given tissue will be an upcoming opportunity to identify new disease-causing genes and mutational loci of PCG. e future development of animal models may provide fascinating information on understanding pathological and genetic characteristics and developing new treatment strategies for PCG.

Data Availability
e data used to support the findings of this study are included within the article.

Conflicts of Interest
e authors declare that they have no conflicts of interest.

Authors' Contributions
QX reviewed the literature and wrote the manuscript. DZ designed and revised the manuscript. YZ, YD, HJ, QD, TW, and YJ collected related articles. DZ gave final approval for publication. All the authors read and approved the final manuscript. QX, DZ, YZ, and YD contributed equally to this study.