Remember me
Research ArticleGeneticsOphthalmology
Open Access | 
10.1172/jci.insight.194681
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Rozumek, G. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Brinkmeier, M. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by
Guan, B.
in:
PubMed
|
Google Scholar
|
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Wang, S. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Tower, C. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Yang, N. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Lim, R. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Kong, D. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Soden, D. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Zhang, Q. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Han, J. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Miller, J. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by
Dong, L.
in:
PubMed
|
Google Scholar
|
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Hannum, D. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Moroi, S. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Richards, J. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Hufnagel, R. in: PubMed | Google Scholar
1Department of Ophthalmology and Visual Sciences,
2Department of Pathology, and
3Department of Human Genetics, University of Michigan, Ann Arbor, Michigan, USA.
4National Eye Institute, NIH, Bethesda, Maryland, USA.
5Cellular and Molecular Biology Program, University of Michigan, Ann Arbor, Michigan, USA.
6Department of Ophthalmology and Visual Sciences, The Ohio State University, Columbus, Ohio, USA.
7Department of Epidemiology, The University of Michigan, Ann Arbor, Michigan, USA.
8Center for Integrated Healthcare Research, Kaiser Permanente Hawaii, Honolulu, Hawaii, USA.
Authorship note: GMR and MLB contributed equally to this work.
Find articles by Prasov, L. in: PubMed | Google Scholar
Authorship note: GMR and MLB contributed equally to this work.
Published February 26, 2026 - More info
Published in Volume 11, Issue 6 on March 23, 2026
AbstractImproper light focus on the retina, refractive error, is primarily caused by eye size differences and is the leading cause of vision loss worldwide. C-terminal variants in the Myelin regulatory factor (MYRF) gene, a retinal pigment epithelium–derived (RPE-derived) transcription factor, lead to isolated nanophthalmos characterized by a small, though structurally sound eye. However, other MYRF loss-of-function variants cause syndromic disease. To address this discrepancy, in vitro and animal studies were performed on a pathogenic C-terminal variant dG-MYRF (p.Gly1126fs30*, c.3376-1G>A). Human RPE cells or primary RPE transduced with dG-MYRF showed reduced target gene expression, with decreased steady-state levels of the C-terminal cleavage product, but normal cleavage and localization. A homozygous humanized MYRF C-terminal mouse model (MyrfhumdG/humdG) was embryonic lethal by E18.5, while WT (MyrfhumWT/humWT) mice were viable. Single-cell RNA-seq from E17.5 MyrfhumdG/humdG and KO RxCre;Myrffl/fl (E15.5 and P0) mice revealed shared differentially expressed genes, with decreased effect size in the MyrfhumdG/humdG eyes. These findings support dG-MYRF as a hypomorphic allele. Additionally, 2 MYRF splicing variants creating nonfunctional isoforms were found in families with isolated nanophthalmos. Overall, hypomorphic MYRF alleles underlie isolated nanophthalmos, supporting a tissue-specific threshold effect and highlighting unique roles for the MYRF C-terminus in the RPE.
Graphical Abstract
Introduction
Myelin regulatory factor (MYRF) is a membrane-associated transcription factor initially identified for its role in oligodendrocyte maturation and myelination (1). It is a type II transmembrane protein that homotrimerizes in the endoplasmic reticulum (ER) lumen and autocleaves to release an N-terminal fragment that translocates to the nucleus and acts as a transcriptional activator (2). More recently, pathogenic variants in MYRF have been associated with an ocular-cardiac-urogenital syndrome featuring congenital heart defects, diaphragmatic hernias, pulmonary hypoplasia, genital abnormalities, and high hyperopia (3–7). MYRF variants have also been described in isolated nanophthalmos, an ocular disease featuring a small but structurally normal eye, with resulting extreme farsightedness (hyperopia). C-terminal frameshift variants have been identified in multiple large pedigrees with predominantly isolated nanophthalmos, suggesting a unique role for the MYRF C-terminus in the eye (8, 9). However, it is not clear why some variants in MYRF result in isolated ocular disease while others produce syndromic phenotypes.
In the eye, MYRF is predominantly and highly expressed in the developing and mature retinal pigment epithelium (RPE) (8, 10). Conditional loss of mouse Myrf in the eye leads to defects in RPE development and retinal degeneration in mice, with perturbations in TGF-β/BMP signaling, pigmentation, cell structure, and cell viability (10). Although the mouse models of Myrf show disparate phenotypes compared with patients, they are invaluable for understanding retinal disease pathogenesis in vivo (8, 11, 12).
Two large nanophthalmos families have been reported with genetic changes that lead to the same C-terminal frameshift mutation in MYRF causing a 31–amino acid extension in the last exon (8, 9). Interestingly, a cluster of patients with pathogenic variants located in the conserved C2 domain manifest with predominantly ocular diseases (Figure 1) (13). While the functional consequences of variants in the DNA binding domain (DBD) and intramolecular chaperone autocleavage (ICA) domain have been well studied (14, 15), there has been little to no progress on understanding the function or importance of the ER-resident, conserved C2 domain.
Figure 1Processing of C-terminal MYRF variant in vitro. (A) Diagram of amino acid changes in WT-MYRF, dG-MYRF, and cleavage deficient variant V679A-MYRF. ProRich, proline rich; DBD, DNA binding domain; ICA, intramolecular chaperone auto-processing; TM, transmembrane; C2, C-terminal. (B) Localization of FLAG-tagged MYRF constructs in ARPE-19 cells show normal nuclear localization of dG-MYRF (n = 3). Scale bar: 50 μm. (C) Western blot of transfected ARPE-19 cells shows no change in cleavage of N-terminal fragment (n = 3). (D) qPCR analysis of RNA from ARPE-19 cells transduced with dG-MYRF, compared with WT-MYRF show decreased levels of transcripts for total Myrf and endogenous Tmem98 mRNA (n = 3) by Student’s t test. *P < 0.05, **P < 0.01.
Toward that end, we generated a humanized mouse model of the C-terminal Myrf variant and performed in vitro functional assays in RPE cell lines and primary human RPE cultures to better understand the molecular mechanism by which C-terminal extension alleles cause disease. Furthermore, we identified 2 additional rare splice site variants in MYRF in families with isolated nanophthalmos, supporting this as a more common mechanism for disease pathogenesis. Together, our studies support that partial loss-of-function alleles in MYRF contribute to isolated nanophthalmos and provide mechanistic insights into the role of the MYRF C2 domain.
ResultsC-terminal MYRF variant is processed normally but shows decreased stability and transcriptional activity. C-terminal MYRF variants and splicing variants are associated with predominantly ocular phenotypes and have familial inheritance, while early truncating and ICA and DBD variants are associated with syndromic phenotypes and predominantly occur de novo (9, 13, 15–17). Given this discrepancy, the C-terminal frameshift alleles could act as dominant negative or hypomorphic alleles. To distinguish these models, we evaluated localization, cleavage, and protein stability of the C-terminal variant in vitro in the ARPE-19 human RPE-like cells. When overexpressed with lentiviral vectors, both WT MYRF and C-terminal frameshift MYRF (dG-MYRF) (p.Gly1126fs30*) showed nuclear localization, while a known cleavage-deficient mutant V679A-MYRF (14) displayed only cytoplasmic localization and was excluded from the nucleus (Figure 1, A and B). Since homotrimerization at the ER lumen is necessary for cleavage of the N-terminal MYRF fragment and its subsequent translocation to the nucleus (1, 2, 18, 19), our results support a model in which dG-MYRF can homotrimerize and cleave.
To further evaluate this, we used Western blotting to systematically define the cleavage dynamics (ratio of full length/cleaved) of the dG-MYRF variant compared with WT-MYRF in ARPE-19 cells. We observed faint bands for full-length FLAG-MYRF (160 kDa) in both WT-MYRF and dG-MYRF and strong transcriptionally active N-MYRF cleavage products (~70 kDa). In contrast, the cleavage-deficient mutant, V679A-MYRF, shows 1 strong full-length MYRF band (Figure 1C). Given normal cleavage and localization, we next evaluated protein stability through computational tools examining degree of disorder Predictor of Natural Disordered Regions (PONDR) and protein folding (AlphaFold2) (20–23). PONDR uses neural networks trained on ordered and disordered regions of short and/or long regions of amino acid sequences from NMR or x-ray crystallography data to predict intrinsic disordered regions. PONDR analysis predicted that dG-MYRF alters the C-terminus from a highly ordered to a highly disordered structure (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.194681DS1). De novo structural modeling using AlphaFold2 predicted the loss of a β-sheet and alpha-helix structure in the conserved C2 domain of the C-terminal frameshift variant of MYRF (Supplemental Figure 1, B and C). To experimentally validate this, cycloheximide pulse-chase experiments were conducted in ARPE-19 cells transduced with WT-MYRF and dG-MYRF. Protein extracts were analyzed by Western blotting using antibodies to the N-terminus or C-terminus of MYRF. The rates of decay of the C-terminal and N-terminal MYRF cleavage product were unchanged in the dG-MYRF variant following a 24-hour chase (Supplemental Figure 2A). However, steady state levels of C-terminal cleavage product for dG-MYRF were reduced compared with WT (Supplemental Figure 2, B and C). This was unlikely due to reduced epitope detection because the antibody was raised to an epitope present in both forms of MYRF (anti-MYRF393-766AA) (1), and the C2 domain is predicted to fold independently.
To evaluate the functional effect of the dG-MYRF variant on transcriptional activity, we tested the ability of dG-MYRF to autoregulate MYRF transcription and its downstream target TMEM98 using qPCR, as compared with an internal transduction GFP control, present in each of the constructs. ARPE-19 cells transduced with dG-MYRF showed decreased levels of endogenous MYRF and TMEM98 mRNA expression (relative to GFP) as compared with WT-MYRF (Figure 1D).
To evaluate whether RPE differentiation state and polarization alter cleavage dynamics of MYRF, we repeated key ARPE-19 experiments in primary human RPE cultures grown on microporous supports, as described previously (24). We again observed normal localization and cleavage of virally transduced WT-MYRF and dG-MYRF primary RPE cultures (Figure 2, A–D). Additionally, transepithelial electrical resistance (TEER) was measured as a proxy for barrier function and overall cell health since we have previously demonstrated the close relationship between assays for cell toxicity and declines in TEER in primary RPE cultures (25). We observed no difference in TEER in primary RPE transduced with Empty Vector, WT-MYRF, or dG-MYRF 1 week after transduction compared with the same well before transduction (Figure 2E). Overall, the results obtained in the ARPE-19 cell line were consistent with results seen in mature and polarized human primary RPE cultures, demonstrating the dG-MYRF variant does not affect localization, processing, barrier function, or overall cell health but seems to reduce transcriptional activation function of MYRF.
Figure 2Processing and functional studies of C-terminal MYRF variant in human primary RPE. (A–C) Localization of FLAG-tagged MYRF constructs in mature human primary RPE show normal nuclear localization of dG-MYRF. Scale bar: 5 μm. (D) Western blot of transduced human primary RPE cells shows no change in cleavage of N-terminal fragment. (E) Functional assessment of barrier function via TEER (normalized to pretransduction TEER value for each replicate) shows no significant difference in TEER 1 week after transduction with Empty Vector, WT-MYRF, or dG-MYRF by 1-way ANOVA (all experiments are 3 reps from n = 2 donors).
Mouse model of human C-terminal MYRF is prenatal lethal. In vitro analysis of the dG-MYRF variant (Figures 1 and 2) provided insight into the potential molecular mechanism of the patient variant MYRF p.Gly1126fs30*. To understand the pathogenesis of nanophthalmos in patients with the MYRF p.Gly1126fs30* variant, we used CRISPR/Cas9 homology–directed repair to generate a humanized mouse model that mimics the frameshift variant detected in patients, as well as a matched humanized WT control (Figure 3A). In this model, we replaced mouse exon 26 with either the human WT or human C-terminal frameshift exon 27 DNA sequence, including the endogenous-3′UTR sequence, and fused this with mouse exon 25 (Supplemental Figure 3). We previously showed that Myrf is highly expressed in RPE via qPCR of RNA from optic cups and RNAscope in situ hybridization (8, 10). To confirm proper cell specificity and expression of the humanized alleles of Myrf, we used a validated Myrf RNAscope probe to evaluate Myrf expression in MyrfhumdG/humdG and MyrfhumWT/humWT eyes at E15.5. These studies showed consistent Myrf expression pattern and levels among MyrfhumWT/humWT and MyrfhumdG/humdG eyes within the RPE (Figure 3B). To further define the function of the dG allele, we first evaluated survival of MyrfhumdG/humdG embryos. We observed that homozygous MyrfhumdG/humdG embryos were never detected after birth, while MyrfhumWT/humWT mice were viable and fertile. MyrfhumdG/humdG embryos were present in normal Mendelian ratios at E12.5–E17.5 but underrepresented at E18.5 (χ2 prob = 0.015) and at birth (P0) (χ2 prob = 0.042) (Figure 3C).
Figure 3Homozygous Myrf humdG mice are embryonic lethal. (A) Schematic showing the (humdG) fusion allele amino acid sequence changes in the humanized mouse model. (B) In situ hybridization using RNAscope detected Myrf mRNA in humanized humWT and humdG mouse RPE (n = 4–5 per genotype). Scale bar: 100 μm. (C) Progeny from intercrossing MyrfhumWT/humdG mice exhibit skewed Mendelian ratios at E18.5 (χ2P = 0.015) and P0 (χ2P = 0.002). E12.5–E14.5 (n = 35 mice from 4 litters), E15.5 (n = 24 mice from 3 litters), E16.5 (n = 26 mice from 3 litters), E17.5 (n = 44 mice from 5 litters), E18.5 (n = 24 mice from 3 litters), and P0 (n = 56 mice from 5 litters).
Analysis of ocular histology in E16.5 homozygous MyrfhumdG/humdG mice revealed normal pigmentation, which differed from the severe and early-onset depigmentation phenotype we observed in our RxCre;Myrffl/fl mice (Supplemental Figure 4) (8). Tmem98, a direct target of MYRF, is expressed at similar levels in control eyes (MyrfhumWT/humWT) and MyrfhumdG/humdG eyes at the RNA and protein level (Supplemental Figure 5) (8, 26). We cannot rule out the possibility that there are small expression differences not detected by the mRNA and immunostaining assays. However, the RxCre;Myrffl/fl eyes had profoundly reduced expression of Tmem98 (10). The lethality of MyrfhumdG/humdG mice confirm the pathogenicity of the C-terminal Myrf mutant allele, and the lack of obvious ocular defects suggest that it likely functions as a hypomorphic (partial loss-of-function allele) rather than null allele.
MyrfhumdG/+ have normal eye size, retinal and RPE morphology, and retinal function. The RxCre;Myrffl/+ mice had white spots in the retina and RxCre;Myrffl/fl mice had severe retinal degeneration, though no measurable eye size phenotype was observed in either genotype into early adulthood (8). We systematically profiled MyrfhumWT/+ and MyrfhumdG/+ mice over the course of 1 year to determine whether changes ocular dimensions, retinal structure, or function would emerge over time. Noninvasive imaging (SD-OCT and fundus photography) and electroretinography (ERG) were carried out on the same mice over the course of 1-year at 3-month intervals. After 1 year, MyrfhumdG/+ show no significant difference in eye size, axial length, corneal diameter, anterior chamber depth, or retinal thickness and only a modest decline in vitreous chamber depth compared with controls (0.53 ± 0.03 mm versus 0.50 ± 0.01 mm, P = 0.0342) (Figure 4, A–G). Additionally, scotopic and photopic ERGs showed no decrease in visual function, consistent with the SD-OCT findings of normal retinal thickness and morphology (Figure 4, H–J). Fundus photography showed little to no signs of atrophy after 1 year (Figure 4, B and C). To investigate for more subtle changes to RPE morphology similar to those that were evident in RxCre;Myrffl/fl eyes (8), RPE flatmounts were stained with phalloidin, segmented using REShAPE software (27), and evaluated for morphometric features (Supplemental Figure 6). These analyses reveal no differences in the median cell area (P = 0.899), aspect ratio (P = 0.847), or hexagonality (P = 0.7568) of 1-year-old MyrfhumdG/+ eyes compared with MyrfhumWT/+ controls. There was also no change in the median number of neighboring RPE cells between genotypes.
Figure 4Heterozygous MyrfhumdG/+ mice show no gross morphological or retinal eye phenotype after 1 year. (A) Images of enucleated eyes showing similar eye size. Scale bar: 1 mm. (B) Spectral domain optical coherence tomography (SD-OCT) images showing no change in overall thickness or individual retinal layers with white lines representing measurements for vitreous chamber depth (VCD) and retinal thickness. (C) Fundus photographs showing no signs of retinal white spots in MyrfhumdG/+ (pink) compared with controls (black). (D–G) Quantitative assessment of eye size (D), VCD (E), anterior chamber depth (F), and retinal thickness (G) by Student’s t test. (H–J) Retinal function measured by scotopic electroretinogram (ERG) (H and I) and photopic ERG (J) showed no differences (n = 6–8 per genotype). Scale bar: 1 mm.
To evaluate for more subtle transcriptomic changes in MyrfhumdG/+ compared with WT mice, we performed qPCR in P21 adult mouse optic cups including the retina and RPE (n = 4–9 pairs of eyes per genotype). In agreement with the lack of phenotypic changes in these mice, we found no significant differences in gene expression between Myrf+/+ and MyrfhumdG/+ mice in key RPE genes and previously defined MYRF target genes (10) including Mitf, Ermn, Myrf, Sox10, or Tmem98 (Supplemental Figure 7). Together, these results suggest that, at both histologic and the molecular level, MyrfhumdG/+ mice are similar to their matched controls.
To determine the contribution to disease pathogenesis of the Myrf humdG allele, we generated RxCre;MyrfhumWT/humWT, RxCre;MyrfflhumWT, and RxCre;Myrffl/humdG mice for observational studies over a 6-month period to see the effect of the humdG allele on a sensitized background with 1 allele of Myrf missing (Supplemental Figure 8), but the numbers were limited due to poor breeding and survival (n = 4 eyes, 2 mice per genotype). Despite limited sample size, the effect was so severe that it was clear that RxCre;Myrffl/humdG showed signs of RPE degeneration compared with RxCre;MyrfhumWT/humWT and RxCre;Myrffl/humWT controls as early as 2 months and more strikingly by 6 months (Supplemental Figure 8A). In addition, RxCre;Myrffl/humdG mice showed significant reduction in vitreous chamber depth at 4 months (0.4 ± 0.2 mm) versus RxCre;Myrffl/humWT (0.54 ± 0.01 mm, P = 0.0163) and RxCre;MyrfhumWT/humWT (0.525 ± 0.008 mm, P = 0.0106), and 6 months (0.38 ± 0.03 mm) versus RxCre;Myrffl/humWT (0.486 ± 0.005 mm, P = 0.0130) and RxCre;MyrfhumWT/humWT (0.54 ± 0.01 mm, P = 0.0013) (Supplemental Figure 8, B and D). No differences were observed between genotypes in total retinal thickness or overall eye size (Supplemental Figure 8, E and F). In addition to anatomical changes, we observed molecular changes such as mislocalization of a partner protein TMEM98 in 6-month-old RPE flat mounts from RxCre;Myrffl/humdG mice, which was not seen in controls (Supplemental Figure 8C). These data support the Myrf humdG allele as a partial loss-of-function allele.
Single-cell RNA-seq shows altered RPE gene regulatory networks in MyrfhumdG/humdG eyes. To better understand how the C-terminal MYRF variant affects eye development, we performed single-cell RNA-seq (scRNA-seq) on E17.5 optic cups of MyrfhumWT/humWT and MyrfhumdG/humdG mice using the 10X genomics platform (Figure 5A). We chose E17.5, as it is occurs after our RPE pigmentation phenotype (E15.5) is observed in RxCre;Myrffl/fl mice (10) and just before Mendelian ratios become skewed at E18.5. We collected 11,281 and 11,322 cells for MyrfhumWT/humWT and MyrfhumdG/humdG, respectively. The median genes per cell were comparable in WT (2,429) and variant (2,415) samples. Quality control filtering was performed to filter out dead cells, doublets, and poor-quality cells (10). After integration with previously published datasets from conditional KO Myrf eyes (RxCre;Myrffl/fl and Myrffl/fl) at E15.5 and P0, we performed unsupervised clustering and were able to identify 19 clusters, including all the major cell types within the optic cup using previously established cell type–specific markers (Supplemental Figure 9) (10). All clusters were present in both the MyrfhumWT/humWT and MyrfhumdG/humdG mice (Figure 5B), but there was a slight reduction in the RPE cell proportions in MyrfhumdG/humdG mice (5.0%) compared with control (6.4%) (Supplemental Figure 9, C and D).
Figure 5scRNA-seq reveals molecular changes in RPE of mice carrying C-terminal Myrf Allele (MyrfhumdG/humdG). (A) Schematic depicting optic cups and samples used for scRNA-seq including control and variant mouse optic cups (n = 3 pooled per genotype). (B) UMAPs of control versus Myrf C-terminal variant mouse optic cups demonstrating cell type distributions. (C) Feature plot demonstrating expression of Myrf mRNA in both control and variant mice within the RPE cluster. (D) Volcano plot of differentially expressed genes in MyrfhumdG/humdG mice compared with control (MyrfhumWT/humWT). DEGs were defined by a Bonferroni Padj < 0.1. (E) Top 10 Gene Ontology (GO) Biological Pathway terms for upregulated and downregulated genes.
Myrf expression was detected exclusively in the RPE cluster (Figure 5C), consistent with previous results, leading us to focus our downstream analysis on the RPE cluster. We performed differential gene expression (DEG) analysis between MyrfhumWT/humWT and MyrfhumdG/humdG RPE clusters. We found 735 upregulated and 170 downregulated differentially DEGs with avg_log2FC > ± 0.25 and Padj ≤ 0.1 (Figure 5D), MyrfhumdG/humdG relative to MyrfhumWT/humWT controls. Downregulated genes included extracellular matrix–related (ECM-related) genes Col3a1, Col2a1, Col1a1, and Col1a2 produced by the RPE that play an important role in structure and barrier maintenance (28). Interestingly, some of the most downregulated genes in the variant mice are related to eye size disorders including Serpine3 (avg_log2FC = –1.11, Padj = 2.76E-14), Ankfn1 (avg_log2FC = –1.138, P
Comments (0)