APRDX1 mutant allele causes a MMACHC secondary epimutation in cblC patients

Abstract

To date, epimutations reported in man have been somatic and erased in germlines. Here, we identify a cause of the autosomal recessive cblC class of inborn errors of vitamin B₁₂ metabolism that we name "epi-cblC". The subjects are compound heterozygotes for a genetic mutation and for a promoter epimutation, detected in blood, fibroblasts, and sperm, at the MMACHC locus; 5-azacytidine restores the expression of MMACHC in fibroblasts. MMACHC is flanked by CCDC163P and PRDX1, which are in the opposite orientation. The epimutation is present in three generations and results from PRDX1 mutations that force antisense transcription of MMACHC thereby possibly generating a H3K36me3 mark. The silencing of PRDX1 transcription leads to partial hypomethylation of the epiallele and restores the expression of MMACHC. This example of epi-cblC demonstrates the need to search for compound epigenetic-genetic heterozygosity in patients with typical disease manifestation and genetic heterozygosity in disease-causing genes located in other gene trios.

Fig. 1

The CHU-12122 epi-cblC case is compound heterozygous for a coding mutation and an epimutation of MMACHC detected in three generations. a Pedigree of the family of case CHU-12122; red bar, mutation (heterozygous c.270_271insA, p.Arg91LysfsTer14) found in the proband, her mother, and her maternal grandfather; green circle, epimutation encompassing the MMACHC promotor/exon 1 found in the proband. b Map of the MMACHC gene and expanded view of the MMACHC CpG island. CpG sites are numbered according to their position upstream and downstream of ATG. Asterisks indicate the CpG sites of the CpG island which were probed in the HM450K array. c Epigrams of MMACHC methylation analyzed by PCR amplification/cloning/Sanger sequencing of bisulfite-treated DNA. Epimutations were detected in the proband, her father, and her paternal grandfather. d RT-PCR of fibroblasts from case CHU-12122 before and after the 3-day treatment with 10 µM 5-azacytidine (5-AZA). Silencing of the wild-type allele (mono-allelic expression in upper panel, red arrow) was reversed after treatment (bi-allelic expression after treatment, lower panel, red arrow)

Fig. 2

The MMACHC epimutation is found in a second case and is present in sperm of both cases. a Pedigree of the family of case WG3838; red bar, mutation (heterozygous c.81G>A splice variant) found in the proband, her mother, and a deceased brother; green circle, epimutation encompassing the MMACHC promotor/exon 1 found in the proband and her mother. No material from the deceased brother was available to study the presence of the epimutation. b MMACHC methylation epigrams after PCR amplification/cloning/Sanger sequencing of the bisulphite-treated DNA from blood obtained from WG-3838, her father CHD867, WG-4152, and from sperm obtained from CHD867. The epimutation was identified in all samples. c HM450K array methylome profiling of the epimutation in the CHU-12122 case and her relatives. The data confirmed the results obtained in Fig. 1c. d Methylome profiling confirms the epimutation in blood DNA from case WG-3838, her father CDH-867, and case WG-4152; the dotted line corresponds to a β value threshold of 0.2, below which the CpG probe was considered fully unmethylated. e Methylome profiling of sperm DNA from the father of case CHU-12122, the father of case WG-3838, and a control population of Utah. The data confirm the epimutation’s presence in the sperm; the dotted line corresponds to a β value threshold of 0.2, below which the CpG probe was considered fully unmethylated. The absence of sperm DNA contamination by DNA from somatic cells was proven by methylation analysis of SNRPN imprinted gene (see Supplementary Fig. 1)

Fig. 3

Epigenome-wide analyses confirm the presence of the epimutation in genomic DNA and sperm. a Epi-Manhattan plot for the differences between β values in the epi-MMACHC subjects and controls (β_Epi-MMACHC–β_Controls). The horizontal line indicates a difference of 0.3; the volcano plot shows the relationship between (β_Epi-MMACHC–β_Controls) and −log₁₀ (P-value) for the t-test. b Enlarged view of the MMACHC locus; β values of the CpG probes in the MMACHC locus and the epi-linkage disequilibrium plot reporting the matrix relationship between the CpG probes based on the blood samples. The dotted line corresponds to a β value threshold of 0.2, below which the CpG probe was considered to be fully unmethylated. c Genomic position of the MMACHC locus according to RefSeq Genes 105v2, NCBI (GRCh37)

Fig. 4

The MMACHC epimutation is associated with a H3K36me3 chromatin mark in the promoter and a mutation in the PRDX1 adjacent gene. a Results of ChIP-Seq analyses at the genomic region encompassing the promoters of MMACHC/CCDC163P and PRDX1. Genomic panels show normalized coverage for histone H3 trimethylated lysine 36 (H3K36me3) mark in patients (tracks 1–4 from top) and controls (tracks 5–7). H3K4me3 track along with the peak calls, for one of the controls, has also been shown. The rectangles indicate the promoter regions. The same scale has been set in all panels. See Methods for more details on samples and analysis. b Identification of PRDX1 splice acceptor variants in cases with MMACHC epimutation by whole-genome sequencing. Top: the vertical black line is positioned on the locus of the two PRDX1 variants within the splice acceptor site (AG sequence) on intron 5. Bottom: zoomed view centered on the splice acceptor site of PRDX1 intron 5. Genomic positions are reported according to the reference sequence GRCh37. The red semicircle denotes the presence of MMACHC genetic mutation. The green semicircle denotes the presence of MMACHC epimutation. The white semicircle denotes the absence of both

Fig. 5

Pedigrees derived from whole-genome sequencing and epigenome analyses and RNA-Seq in case fibroblasts. a Pedigrees derived from WGS of DNA by HiSeq X Ten Illumina System in cases CHU-12122 and WG-3838 and their relatives. WGS evidences c.515-1G>T and c.515-2A>T mutations of PRDX1, in cases CHU-12122 and WG-3838 and the relatives who bear the secondary epimutation, respectively. The mutations are absent in relatives, who bear the MMACHC heterozygous mutation. b NGS sequencing of RNA from case CHU-12122 and HDF control fibroblast line. Overlapping antisense (in blue) transcription of PRDX1/MMACHC/CCDC163P trio of genes is predominant in CHU-12122 fibroblasts while MMACHC sense transcription (in red) is predominant in HDF control fibroblasts

Fig. 6

The PRDX1 mutation causes aberrant antisense transcription through the MMACHC and CCDC163P genes. a qRT-PCR analyses of antisense transcripts encompassing exons 5 and 6 of PRDX1, PRDX1 exon 5—MMACHC exon 1 and MMACHC exon 1-CCDC163P exon 1 showing the aberrant extension of PRDX1 antisense transcription through sense MMACHC and antisense CCDC163P and no aberrant transcripts in HDF control fibroblasts. b RT-PCR detection of an aberrant 1 kb transcript produced with PRDX1 sense and CCDC163P antisense primers. Lanes 2, 4, and 6 correspond to RT-PCR of RNA from fibroblasts CHU-12122, WG-3838, and HDF (control fibroblasts). Lanes 3 and 5 correspond to control experiments without fibroblast RNA in the reaction mixture. They show no amplification artifacts. c, d Sanger sequencing of cDNA shows that the antisense aberrant transcript encompasses PRDX1 exons 4 (red) and 5 (blue), part of exon 1 and promoter of MMACHC, and part of exon 1 of CCDC163P (green). e Proposed mechanisms of epigenetic silencing of the non-mutated F2 allele in patients with a heterozygous mutation of a causal gene that belong to a trio of reverse (R1)–forward (F2)–reverse (R3) genes. f Expression of PRDX1 and MMACHC in control fibroblasts, fibroblasts of patient WG3838, and MeWo-L1 cell line transfected with control and PRDX1 siRNA. g MMACHC methylation epigrams of MeWo-LC1 cells transfected with control and PRDX1 siRNA