First Application of Whole Genome Sequencing in Myelinated Retinal Nerve Fibers (MRNF)

Summary

Genetic features are currently unknown in myelinated retinal nerve fibers (MRNF). For a 20-year-old asymptomatic female with unilateral MRNF, we performed whole genome sequencing (WGS) by standard workflow protocol to produce contiguous long-read sequences with Illumina DNA PCR-Free Prep. After tagmentation, libraries were sequenced on separate runs via NovaSeq 6000 platform at 2 × 150bp read length. Gene variants included rs2248799, rs2672589, rs7555070, rs247616_T and rs2043085_C all associated with an increased macular degeneration risk, and seven novel variants of uncertain significance. For optic disc enlargement, variants rs9988687_A, rs11079419_T, rs6787363 and rs10862708_A suggested an increased risk for this condition. In contrast, modeling revealed retinal detachment risk was reduced by variants identified at rs9651980_T, rs4373767_T, and rs7940691_T which were among five other previously unreported variants. WGS data placed proband at the 66^th and 64^th percentiles for disc anomaly and retinal detachment risk, respectively. Additionally, risk determined from 16 loci associated with age-related macular degeneration found the patient to be at the 18^th percentile for this diagnosis (i.e., below average genetic predisposition). Fundoscopic findings showed mean RNFL thickness was lower with MRNF (77 OS vs. 96μm OD) and RNFL symmetry was impaired (43 %) but stable between 2020 and 2023. Rim area and cup volume were also substantially different (2.33 OS vs. 1.34mm² OD, and 0.001 OS vs. 0.151mm³ OD, respectively). As the first known evaluation of MRNF via WGS, these data reveal a mixed picture with variants associated with different risks for potentially related ocular pathologies. In addition, we identify multiple new variants of unknown significance. Factors affecting gene expression in MRNF require further study.

Myelinated retinal nerve fibers (MRNF) describe retinal nerve fibers anterior to the lamina cribrosa which retain a myelin sheath, an event noted in about 1 % of patients. It is often diagnosed incidentally with few clinical complaints. First described in 1856 by Virchow, the condition appears as sharply circumscribed gray-white retinal patches [1]. Its etiology remains not known. The finding is often present at birth as a nonprogressive lesion, and familial cases of MRNF have been reported. Here we describe genetic features in the setting of asymptomatic non-progressive unilateral MRNF, believed to be the first application of whole genome sequencing (WGS) for this condition.

Clinical presentation: A 20-year-old female attended for follow-up ophthalmic examination. The patient was a well-adjusted college student with BMI = 17.5. She did not use eye drops or tobacco and her surgical history was negative. Her medical history was notable for being a known carrier of three gene variants at RUNX2, SALL1, and SAMD9, with ophthalmology referral and WGS included in her post-hospitalization care plan after COVID-19. She had been diagnosed with strabismus during early pediatric checks, and corrective lenses at age 18mo were revised at periodic exams. By high school graduation, glasses were no longer needed. At the most recent exam there were no field defects and acuity by Snellen was 20/30, bilaterally. Unilateral (left) MRNF was serially monitored with no changes recorded over 10 years. Her only daily medications were oral contraceptives and 2.5mg enalapril, the latter prescribed for intermittent proteinuria. When increased urine protein and fluctuating glomerular filtration rate emerged in 2023, an unscheduled eye evaluation was recommended before proceeding to renal biopsy.

Assessment included optical coherence tomography (macula & anterior segment), ONA/optic nerve, endothelial cell count, amplitude scan, bright scan ultrasound, fluorescein & ICG angiography, pachymetry, slit-lamp exam, color perception, refraction and visual field mapping. Retinal nerve fiber layer (RNFL) thickness was notably reduced with MRNF (77 OS vs. 96μm OD), and RNFL symmetry was low (43 %) but unchanged between 2020 and 2023. Rim area and cup volume in 2023 were substantially different (2.33 OS vs. 1.34mm² OD, and 0.001 OS vs. 0.151mm³ OD, respectively).

WGS data were collected using standard workflow to produce contiguous long-read sequences on the NovaSeq^TM 6000 System and NovaSeq X Series [2]. Briefly, long DNA fragments were enzymatically indexed with landmarks. Unmarked libraries to produce contiguous long reads for the original single-molecule fragment were prepared using Illumina DNA PCR-Free Prep, Tagmentation (Illumina cat.# 20041794). Libraries were then sequenced on separate runs via NovaSeq 6000 System at 2 × 150bp read length. Because current experience has yet to correlate specific variants with MRNF, this investigation cross-referenced regions with known association with retinal structure or function based on reported variants associated with optic disc morphology, retinal detachment, and age-related macular degeneration.

Reference metrics of optic disc morphology were derived from 115 genetic variants associated with optic disc size and vertical cup-to-disc ratio [3]. From such reports, proband data were calculated at the 66^th percentile (i.e., above average genetic predisposition to disc enlargement). For retinal detachment 11 genetic variants associated with this condition were considered [4], and the proband was placed at the 64^th percentile consistent with above average genetic predisposition to retinal detachment. However, macular degeneration risk based on 16 loci associated with age-related macular degeneration [5] placed the proband at the 18^th percentile (i.e., below average genetic predisposition). Variants sufficiently rare or new were provisionally classified under macular degeneration (n=7), optic disc enlargement (n=59), and retinal detachment (n=8) as shown in Table 1a and 1b. Prior exon sequencing [6] identified variants Q253H in SALL1 and R824Q in SAMD9 plus a previously unreported multiexon 3′ terminal duplication of RUNX2. Since retinal evaluation was reassuring with no vascular changes (Fig. 1), her kidney findings were attributed to a local manifestation of ‘long Covid’ as an isolated renal process.

Table 1a.

Leading variants identified in unilateral MRNF via whole genome sequencing (WGS) with reference to age-related macular degeneration, optic disc size anomalies, and retinal detachment vs. a WGS private library (n=5,000) ranked by effect size. Imputed polygenic scores (+/−) estimated proband’s relative risk for each disorder.

	variant	CR	PG	effect size	VAF ( %)
macular degeneration (−)	rs2248799	10q26	C/T	0.61	49
	rs2672589	10p14	A/A	0.47	61
	rs7555070	1p13	T/T	0.46	68
	rs247616_T	16q13	C/T	0.17	33
	rs2043085_C	15q21	C/C	0.14	61
optic disc (+)	rs9988687_A	10p13	A/A	0.07	77
	rs6787363	3p13	A/A	0.02	38
	rs10862708_A	12p13/12q21	A/A	0.02	56
retinal detachment (+)	rs9651980_T	12p14	T/T	0.17	9
	rs4373767_T	1p14	C/T	0.12	63
	rs7940691_T	11p13	C/T	0.11	36

Notes: CR=cytogenetic region, PG=proband genotype, VAF=variant allele frequency. For all entries, p<0.0001.

Table 1b.

Previously unknown or new variants identified in unilateral MRNF via whole genome sequencing (WGS) with reference to reports associated with age-related macular degeneration (n=69), optic disc size anomalies (n=115), and retinal detachment (n=11).

	variant	PG	effect size	VAF (%)
macular degeneration	rs3825991_A	C/A	0.09	48
	rs1926564_A	A/A	−0.14	90
	rs1378940_A	C/A	0.09	68
	rs259842_C	C/C	−0.08	62
	rs11120691_G	G/G	−0.08	44
	rs1005819_T	C/T	−0.08	42
	rs17421419_G	A/G	0.15	7
optic disc	rs11079419_T	T/C	0.03	20
	rs9905786_T	T/T	0.02	64
	rs4839470_T	C/T	−0.02	24
	rs12436074_A	A/A	−0.02	56
	rs73173591_A	T/A	0.03	9
	rs72759609_T	T/T	0.03	90
	rs76567987_A	A/A	0.02	84
	rs3857971_A	G/A	−0.02	30
	rs74056359_A	A/A	−0.02	83
	rs7744813_A	C/A	−0.02	59
	rs61975075_A	A/A	0.03	94
	rs9330799_A	A/A	−0.02	53
	rs2092524_A	G/A	0.01	34
	rs9967780_T	G/T	0.02	22
	rs906568_T	T/T	−0.01	36
	rs1905014_T	T/T	−0.01	57
	rs1550094_A	G/A	0.01	69
	rs10764494_A	C/A	0.01	68
	rs35285683_A	A/A	−0.02	85
	rs7717697_T	T/T	0.01	59
	rs11564398_T	T/T	0.01	72
	rs7188859_T	T/C	0.01	63
	rs72784719_A	A/A	0.06	99
	rs77877421_A	A/A	−0.03	94
	rs698153_A	A/A	−0.03	95
	rs71296770_A	A/A	−0.02	90
	rs6673575_A	G/A	0.01	32
	rs2033054_T	C/T	−0.01	65
	rs10957177_A	A/G	0.01	75
	rs11584075_A	A/A	0.02	92
	rs10164395_T	C/T	−0.01	33
	rs6999835_T	T/C	0.01	63
	rs11734073_A	A/A	−0.02	87
	rs251526_A	A/G	−0.02	92
	rs2149108_T	C/T	−0.01	40
	rs13417287_T	T/T	−0.01	77
	rs2761882_T	C/T	−0.01	50
	rs12619508_A	C/A	−0.01	47
	rs3754442_T	C/T	0.01	59
	rs59199978_A	A/A	−0.01	82
	rs11627052_A	G/A	0.01	22
	rs13022913_T	T/T	−0.01	57
	rs78977588_A	C/A	0.02	12
	rs10823610_A	C/A	0.01	56
	rs599892_C	C/C	0.01	70
	rs1852148_A	A/G	−0.01	47
	rs10910_T	T/T	0.01	69
	rs6860726_C	G/C	−0.01	49
	rs7620608_T	T/T	−0.01	61
	rs7615960_A	A/T	−0.02	93
	rs2266963_C	C/C	−0.01	81
	rs28840750_T	T/G	−0.03	96
	rs2430356_A	A/T	−0.01	77
	rs11684168_T	T/T	0.01	83
	rs8076249_T	T/T	−0.01	60
	rs1901440_A	C/A	−0.01	66
	rs30371_T	T/T	−0.01	63
	rs28603236_A	A/A	0.01	85
	rs13264644_A	A/G	−0.01	57
retinal detachment	rs74764079_T	T/T	−0.33	97
	rs4373767_T	C/T	0.12	63
	rs11187838_G	G/A	−0.11	57
	rs1248634_G	G/G	−0.12	71
	rs11217712_T	T/T	−0.11	31
	rs4243042_T	A/T	−0.12	47
	rs7940691_T	C/T	0.11	36
	rs9651980_T	T/T	0.17	9

Notes: PG=proband genotype, VAF=variant allele frequency. Negative effect size indicates reduced risk. Retinal detachment variants appear in both tables because all variants were new/unreported. For all entries, p<0.0001.

Fig. 1.

Fundoscopic findings (left) documenting myelinated retinal nerve fibers (M) at the 2023 exam.

For WGS, effect size was defined as contribution of the SNP/mutation to trait genetic variance (gv) = 2β² f (1 – f) where f is allele frequency for either of two alleles and β is coefficient for a SNP when outcome is modeled by regression [7].

Discussion

MRNFs are caused by presence of ectopic oligodendrocyte-like cells in the retina, leading to myelination of retinal ganglion cell fibers. It occurs when the protective lamina cribrosa barrier is underdeveloped, allowing encroachment of oligodendrocytes into retina. Visual pathway myelinization is evident by the 8^th gestational month and reaches the posterior globe near term, with most fibers attaining full myelination by age seven months [8]. Myelination usually stops at this stage, but when it progresses beyond the lamina cribrosa, MRNFs are the result [9,10]. The impact of MRNF on visual functioning is variable and likely depends on degree of myelination or macula involvement.

While this project does contribute original WGS data to the understanding of MRNF, there are important constraints on our research. First, WGS sampling was not performed on relatives to construct an informative pedigree. Additionally, MRNF is joined here by exon variants involving RUNX2, SALL1 and SAMD9, making generalization to a standard background population difficult. Indeed, these variants draw notice to their potential relevance in the emergence of MRNFindependent of information supplied by WGS, particularly as axonal myelination is organized by oligodendrocyte progenitors which migrate via neuroendocrine signaling [8]. Although RUNX2 has not yet been implicated in MRNF, this is a highly-conserved gene known to regulate ocular development [11], and the protein product RUNX2 guides transcriptional programs and eye morphogenesis. RUNX2 also influences extracellular matrix receptor interaction pathways related to corneal dystrophy/keratoconus [12] and promotes Müller glial cell activation and maintenance of the blood-retina barrier [13]. Recently, a proteomic study of vitreous samples was referenced against a WGS dataset and discovered a previously unknown RUNX2 pathway involved in angiogenesis [14]. While SALL1 disruption is not thought to affect microglial colonization of retina and cortex [15], microglial morphology during retinal development changed from ramified to amoeboid in a murine SALL1 knockout model [16].

Regarding SAMD9, no retinal features have been reported among patients with mutations at this locus or with its associated ‘MIRAGE’ syndrome (i.e., myelodysplasia, infection, growth restriction, adrenal hypoplasia, genital phenotypes, and enteropathy). Current guidance for eye care with SAMD9 variants is routine (i.e., artificial tears for hypolacrima) [17].

Conclusion

This report shares the first WGS data on unilateral asymptomatic MRNF. In our case, MRNF remained stable over many years, despite—or perhaps because of—three gene variants not previously associated with MRNF. As a developmental mechanism for MRNF awaits clarification, it is hoped that increased availability and lower cost of WGS will enable broader knowledge of this unusual condition going forward.