In patient II-4, exome sequencing had revealed a homozygous NMNAT1 variant (p.Glu257Lys), which did not fully segregate with the intellectual disability in this family. Individual II-1 also carries this variant in a homozygous state and on ophthalmologic examination (ie, electroretinography and optical coherence tomography) had no ocular abnormalities.
Siemiatkowska AM, Schuurs-Hoeijmakers JHM, Bosch DGM, Boonstra FN, Riemslag FCC, Ruiter M, de Vries BBA, den Hollander AI, Collin RWJ, Cremers FPM. Nonpenetrance of the Most Frequent Autosomal Recessive Leber Congenital Amaurosis Mutation in NMNAT1. JAMA Ophthalmol. 2014;132(8):1002-1004. doi:10.1001/jamaophthalmol.2014.983
The NMNAT1 gene was recently found to be mutated in a subset of patients with Leber congenital amaurosis and macular atrophy. The most prevalent NMNAT1 variant was p.Glu257Lys, which was observed in 38 of 106 alleles (35.8%). On the basis of functional assays, it was deemed a severe variant.
The p.Glu257Lys variant was 80-fold less frequent in a homozygous state in patients with Leber congenital amaurosis than predicted based on its heterozygosity frequency in the European American population. Moreover, we identified this variant in a homozygous state in a patient with no ocular abnormalities.
Conclusions and Relevance
On the basis of these results, the p.Glu257Lys variant is considered not fully penetrant. Homozygotes of the p.Glu257Lys variant in most persons are therefore not associated with ocular disease. Consequently, genetic counselors should exercise great caution in the interpretation of this variant.
In the September 2012 issue of Nature Genetics, 4 research groups reported NMNAT1 (OMIM 608700) mutations in patients with Leber congenital amaurosis (LCA).1- 4 Most patients with these mutations had severe, early-onset vision loss accompanied by macular atrophy. Homozygous Nmnat1 knockout mice displayed early lethality, whereas heterozygous animals remained unaffected.5 In our LCA cohort, we found compound heterozygous NMNAT1 variants in 6 probands. In previous studies, the most prevalent NMNAT1 variant was c.769G>A (p.Glu257Lys), observed in 38 of 106 alleles (35.8%). Collectively, this variant was found in 43 of 957 patients with LCA.1- 4
Eleven of 4289 healthy European Americans are p.Glu257Lys heterozygotes (http://evs.gs.washington.edu/EVS/). When the Hardy-Weinberg principle is applied, homozygous alleles should be present in approximately 1 in 600 000 individuals. Because LCA occurs in 1 in 50 000 individuals, it can be inferred that this homozygous variant, if fully penetrant, should be responsible for approximately 8% of patients with LCA. Only 1 in 957 patients with LCA (0.1%), however, carried the p.Glu257Lys variant in a homozygous state.3 This 80-fold discrepancy between true and expected homozygosity rates in LCA led us to hypothesize that p.Glu257Lys is a hypomorphic variant that is not fully penetrant.
The activity of NMNAT1 was measured in erythrocytes in the 1 homozygous patient and was significantly lower than in his heterozygous mother.3 This variant was also tested in a gene construct in HeLa cells and was suggested to be the most severe allele of 4 NMNAT1 variants.3
In the Department of Human Genetics, Radboudumc, Nijmegen, exome database, the p.Glu257Lys variant was found in 10 of 850 persons. One of these individuals (patient II-4), who had intellectual disability, was carrying this substitution in a homozygous state (Figure). Interestingly, an unaffected sibling (individual II-1) was also homozygous for p.Glu257Lys. After clinical and ophthalmologic examination, individuals II-1 and II-4 did not have any ocular abnormalities.
On the basis of these findings, we assume that in persons of European descent there is a much larger group of healthy individuals who carry the p.Glu257Lys variant in a homozygous state than there are patients with LCA who carry this variant on both alleles. In patients with LCA carrying one p.Glu257Lys variant, the other NMNAT1 allele would then be predicted to be more severe. Additional evidence for this assumption was obtained using haplotype analysis. The patient with LCA who was previously homozygous for the p.Glu257Lys mutation was analyzed for single-nucleotide polymorphisms in NMNAT1 and proven to harbor this variant on 2 different haplotypes.3 In our patients, one of these haplotypes was present in every person with the p.Glu257Lys variant, including the healthy individual (patient II-1) (data not shown). Consequently, we hypothesize that there may be differential expression of the 2 alleles harboring the p.Glu257Lys mutation due to trans- or cis-acting elements. This would render one of them more severe, which might explain the nonpenetrance observed in individuals II-1 and II-4.
Similarly behaving variants were previously found in ABCA4 (OMIM 601691) and BBS1 (OMIM 209901). A frequent ABCA4 variant, p.Gly863Ala, was identified in a compound heterozygous state in one-third of Dutch persons with Stargardt disease and in a heterozygous state in 3% of healthy controls.6 In several thousands of patients with Stargardt disease, this variant was never found in a homozygous state (F.P.M.C., unpublished data, December 1, 2013). Consequently, it was deemed a hypomorphic allele causing Stargardt disease only if in combination with a severe ABCA4 variant. The most frequent Bardet-Biedl syndrome–associated variant, p.Met390Arg in BBS1, was observed in a homozygous state in individuals with Bardet-Biedl syndrome with nonsyndromic retinitis pigmentosa, and healthy persons.7,8 The presence of low-penetrant variants has also been observed in more common human hereditary conditions, such as cystic fibrosis.9 Reduced penetrance is frequently observed in both dominant and recessive disorders. The relevant variants can be hypomorphic for several reasons. Besides environmental factors (such as in the case of breast cancer in BRCA1/2 [OMIM 113705/600185] mutation carriers), incomplete penetrance may be due to trans- or cis-acting elements, which would influence its expression, rendering the variant mild and benign. An example of a trans-acting modifier was provided for the PRPF31 gene implicated in autosomal-dominant retinitis pigmentosa with reduced penetrance.10 A haploinsufficiency model was proposed for PRPF31-associated retinitis pigmentosa. CNOT3 (OMIM 604910) was found to act as a transcriptional repressor of PRPF31, thereby influencing the expression level of the wild-type PRPF31 gene.11,12 The identification of this kind of transcriptional regulators is possible if the target gene of interest and its regulators are expressed highly enough in accessible patient cells. The identification of this kind of modifiers through high-throughput sequencing is challenging in view of the enormous genetic variability observed in the human population.
Taken together, we propose that the NMNAT1 p.Glu257Lys variant is a hypomorphic variant that almost without exception causes LCA in combination with more severe NMNAT1 variants. Other variants in NMNAT1 are far less frequent and therefore have not been observed in a homozygous state. However, it may be speculated that they also represent combinations of mild and severe mutations or 2 alterations exerting a moderate effect, which cumulatively result in disease. In a homozygous state, the mild variants would generally not cause LCA, unless one or both gene copies show reduced transcription, possibly due to trans- or cis-acting modifier elements. This conclusion should be taken into consideration when families carrying this variant are genetically counseled, especially when the retinal phenotype lacks the macular atrophy and does not occur early in life. In those cases, the causal variants are most likely situated in other genes.
Submitted for Publication: December 2, 2013; final revision received January 21, 2014; accepted January 31, 2014.
Corresponding Author: Frans P. M. Cremers, PhD, Department of Human Genetics, Radboud University Medical Center, Geert Grooteplein 10, PO Box 9101, 6500 HB Nijmegen, the Netherlands (email@example.com).
Published Online: May 15, 2014. doi:10.1001/jamaophthalmol.2014.983.
Author Contributions: Dr Cremers had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Siemiatkowska, Collin, Cremers.
Acquisition, analysis, or interpretation of data: Siemiatkowska, Schuurs-Hoeijmakers, Bosch, Boonstra, Riemslag, Ruiter, de Vries, Collin.
Drafting of the manuscript: Siemiatkowska, Cremers.
Critical revision of the manuscript for important intellectual content: Bosch, Riemslag, Ruiter, de Vries, den Hollander, Collin, Cremers.
Statistical analysis: Siemiatkowska.
Obtained funding: den Hollander, Cremers.
Administrative, technical, or material support: Bosch, Boonstra.
Study supervision: Collin, Cremers.
Conflict of Interest Disclosures: None reported.
Funding/Support: This research was supported by TOP grant 91209047 from the Netherlands Organization for Scientific Research, the Algemene Nederlandse Vereniging ter Voorkoming van Blindheid, the Gelderse Blinden Stichting, the Landelijke Stichting voor Blinden en Slechtzienden, the Macula Degeneratie Fonds, the Rotterdamse Stichting Blindenbelangen, and the Stichting Blinden-Penning (Drs den Hollander and Cremers).
Role of the Sponsors: The funding sources sponsored the design and conduct of the study; collection, management, analysis, and interpretation of the data; and preparation, review or approval of the manuscript.