Aberrant COL11A1 splicing causes prelingual autosomal dominant nonsyndromic hearing loss in the DFNA37 locus

Alpha‐chain collagen molecules encoded by genes that include COL11A1 are essential for skeletal, ocular, and auditory function. COL11A1 variants have been reported in syndromes involving these organ systems. However, a description of the complete clinical spectrum is lacking, as evidenced by a recent association of autosomal dominant nonsyndromic hearing loss due to a splice‐altering variant in COL11A1, mapping the DFNA37 locus. Here, we describe two German families presenting prelingual autosomal dominant nonsyndromic hearing loss with novel COL11A1 heterozygous splice‐altering variants (c.652‐1G>C and c.4338+2T>C) that were molecularly characterized. Interestingly, the c.652‐1G>C variant affects the same intron 4 canonical splice site originally reported in the DFNA37 family (c.652‐2A>C) but elicits a different splicing outcome. Furthermore, the c.4338+2T>C variant originated de novo. We provide clinical and molecular genetic evidence to unambiguously confirm that COL11A1 splice‐altering variants cause DFNA37 hearing loss and affirm that COL11A1 be included in the genetic testing of patients with nonsyndromic deafness.

2A>C) but elicits a different splicing outcome. Furthermore, the c.4338+2T>C variant originated de novo. We provide clinical and molecular genetic evidence to unambiguously confirm that COL11A1 splice-altering variants cause DFNA37 hearing loss and affirm that COL11A1 be included in the genetic testing of patients with nonsyndromic deafness. is present in roughly 20% of hearing-impaired individuals. More than 20% of genes exhibit pleiotropy, whereby variants in a single gene can be associated with syndromic or nonsyndromic hearing loss and can follow an autosomal dominant or recessive inheritance pattern .
The gene COL11A1 (collagen type XI alpha-1 chain; MIM #120280) is associated with autosomal dominant Marshall syndrome (MRSHS) and autosomal dominant or recessive Stickler syndrome type II (STL2), as well as autosomal recessive fibrochondrogenesis (FBCG1). Each of these syndromes has a phenotypic overlap that includes skeletal abnormalities and dysmorphic features, as well as variable cleft palate, ocular, and auditory phenotypes that can include mild-to-moderate hearing loss and outer ear malformations.
Recently, the gene COL11A1 has been associated with ADNSHL (DFNA37; MIM #618533) through the genetic analysis of a large European-American family presenting a novel splice-site altering variant (Booth et al., 2019).
Written informed consent was obtained from participants (Approval of the University of Tübingen Ethics Commission; Nos.: 016/2014BO1 and 197/2019BO1). Medical history excluded acquired forms of hearing loss. The 37-year-old proband in Family 1 (III:3) (Figure 1a), presents stable, down-sloping, moderate-to-severe, high-frequency sensorineural hearing loss (Figure 1c), as well as hypothyroidism and diabetes. The speech discrimination with regard to monosyllables was 70% and 50% at 65-dB hearing level (HL) on the right side and 80% and 80% at 65 dB(HL) on the left side when evaluated at 34 and 37 years of age, respectively. Regarding the speech recognition threshold (SRT), the proband achieved a score of 35-and 32.5-dB hearing loss ("a1 value") on the right and 27-and 30-dB hearing loss on the left ear at 34 and 37 years of age, respectively. He has worn hearing aids since age 6 years. His daughter (Family 1, IV:1) is currently 6.9 years old and was born after an unremarkable pregnancy and delivery. She failed newborn hearing screening but passed follow-up testing. At the age of 2 years, she was diagnosed with severe hearing loss and has used hearing aids since diagnosis. Serial audiometry has revealed stable, moderate-tosevere, high-frequency sensorineural hearing loss (Figure 1c). Her monosyllable discrimination was 75% and 65% at 65 dB(HL) on the right and 65% and 100% at 65 dB(HL) on the left side at 4.8 and 6.9 years, respectively. SRT was 27 and 30 dB hearing loss on the right, 25 and 27.5 dB hearing loss on the left at 4.8 and 6.9 years, respectively. Syndromic features including myopia, retinal detachment, midface hypoplasia, submucous cleft palate, and arthritis/joint pain have been excluded in both affected individuals in Family 1 (III:3 and IV:1). The 31-year-old proband (III:3) in Family 2 ( Figure 1b) presented moderate sensorineural hearing loss in the mid-and highfrequencies since early childhood ( Figure 1d) and she has worn hearing aids since age 27 years. Her pure-tone audiograms showed nearly identical thresholds taken 5 months apart at the age of Thresholds are shown with circles and crosses/vertical rectangles, for right and left ears, respectively bilateral sensorineural hearing loss and no other abnormalities. He was too young to undergo speech audiometry. The mother (Family 2, III:3) reported a cleft lip and palate, an occurrence not previously reported with STL2 and assumed to be due to other genetic or multifactorial causes. She reported no other abnormalities. Submucous cleft palate in her son (Family 2, IV:1) has been ruled out.
Ophthalmic examination excluded amblyopia, strabismus, and phoria and confirmed normal vision. Affected individuals in both families do not report otorrhea, otalgia, tinnitus, or vertigo. Bone conduction thresholds from individuals III:3 and IV:1 in Family 1 and III:3 in Family 2 confirm sensorineural hearing loss ( Figure S1).
Blood samples were collected from affected (III:3 and IV:1 of Family 1, as well as III:3 and IV:1 of Family 2) and unaffected (III:4 from Family 1 and II:1 and II:2 from Family 2) individuals in two unrelated German families (Figure 1a An in vitro splicing assay was carried out as previously described Tompson & Young, 2017)   with the molecular characterization of splice variants that revealed multiple in-frame mutant transcripts, we hypothesize that these inframe variants may lead to a milder phenotype (DFNA37) due to partial residual function of the protein (i.e., noncomplete loss-offunction alleles), whereas those with more severe phenotypes may be due to frameshift, truncating loss-of-function alleles, and missense variants which induce splice altering effects or substitution of glycine in a repeat Gly-Xaa-Yaa region. Interestingly, one additional variant was recently described in a Czech family (NM_080629.2:c.1560delC) with ADNSHL. The variant was described as a splice-altering variant but a splicing assay was not performed and in silico tools did not predict a splicing effect (Čopíková et al., 2020). Functional characterization of this variant will be important for a refined genotype-phenotype correlation due to variants in COL11A1. Although the precise mechanism of COL11A1-associated hearing impairment has not been elucidated, disruption of COL11A1 is consistently associated with hearing loss, as demonstrated by about 84% of individuals diagnosed with MRSHS having hearing loss, which supports the critical function of this protein in the auditory system (Bacciu et al., 2018). It also demonstrated that the source of COL11A1 mRNA is in the tectorial membrane and suggested its mutation affects normal cochlearfunction (Shpargel et al., 2004). Moreover, the pleiotropy exhibited by this gene, like other genes that are associated with syndromic and nonsyndromic hearing loss, remains to be fully characterized.
In summary, we report on and characterize two novel splicealtering variants associated with DFNA37, providing confirmatory evidence of COL11A1 as a bona fide ADNSHL gene. We recommend COL11A1 be included in the routine diagnostic testing of patients with both syndromic and nonsyndromic forms of deafness.

ACKNOWLEDGMENTS
The authors thank the families for their participation in this study. All authors approved the final version of the manuscript.

DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available upon request from the corresponding author.