Post-transcriptional microRNA repression of PMP22 dose in severe Charcot-Marie-Tooth disease type 1

Abstract

Copy number variation (CNV) may lead to pathological traits, and Charcot-Marie-Tooth disease type 1A (CMT1A), the commonest inherited peripheral neuropathy, is due to a genomic duplication encompassing the dosage-sensitive PMP22 gene. MicroRNAs act as repressors on post-transcriptional regulation of gene expression and in rodent models of CMT1A, overexpression of one such microRNA (miR-29a) has been shown to reduce the PMP22 transcript and protein level. Here we present genomic and functional evidence, for the first time in a human CNV-associated phenotype, of the 3' untranslated region (3'-UTR)-mediated role of microRNA repression on gene expression. The proband of the family presented with an early-onset, severe sensorimotor demyelinating neuropathy and harboured a novel de novo deletion in the PMP22 3'-UTR. The deletion is predicted to include the miR-29a seed binding site and transcript analysis of dermal myelinated nerve fibres using a novel platform, revealed a marked increase in PMP22 transcript levels. Functional evidence from Schwann cell lines harbouring the wild-type and mutant 3'-UTR showed significantly increased reporter assay activity in the latter, which was not ameliorated by overexpression of a miR-29a mimic. This shows the importance of miR-29a in regulating PMP22 expression and opens an avenue for therapeutic drug development.

Keywords: Charcot-Marie-Tooth disease type 1A; dosage-sensitive genes; microRNAs; peripheral myelin protein 22-kD; post-transcriptional regulation.

Figure 1

Family pedigree and clinical photography of affected family members. Family pedigree (A), with the arrow indicating the proband and filled symbols denoting affected individuals; squares are males and circles are females. In cases where the genotype was available, it is indicated under the corresponding symbol as WT/WT (wild-type/wild-type) or WT/M (wild-type/mutant). Still photographs showing forearm and intrinsic hand muscle amyotrophy in the proband (B) and a milder degree of forearm, hand and distal leg amyotrophy in the proband’s daughter (C).

Figure 2

Aligned sequence reads at the PMP22 locus visualized in the IGV browser. Sequence reads from the trio WGS study (A) and the quad WES study (B). In the WGS study, the panels correspond to the proband (top), proband’s daughter (middle) and the proband’s mother (bottom) and the black arrows indicate the 50% reduction in the coverage plot. The approximate primer annealing sites and direction are indicated. In the WES study (panel order from top to bottom is proband, proband’s daughter, proband’s mother, proband’s father), there is a hint of reduced coverage at the PMP22 3′-UTR in the proband and the proband’s daughter, but is not as obvious and well delineated as in the WGS study, and hence could be overlooked. IGV = Integrative Genomics Viewer; WES = whole exome sequencing; WGS = whole genome sequencing.

Figure 3

Delineation of the breakpoint junctions of the structural variant. Gel electrophoresis of PCR amplicons (A): primers A + B correspond to Lanes 1, 4, 7, 10, primers A + C correspond to Lanes 2, 5, 8, 11, and primers D + B correspond to Lanes 3, 6, 9 and 12. Lanes 1 and 4 illustrate the presence of both the wild-type PCR amplicon (779 bp) and mutated (129 bp). Lanes on either side of Lanes 1–12 were used for the mid-range ladder. Breakpoint junction sequencing was performed on the amplicons excised from Lanes 4 and 7 and the corresponding electropherograms are shown in B (midline of the electropherogram is centred on the SNP position indicated by an asterisk). The electropherogram from the small amplicon in Lane 4 containing the 3′-UTR deletion shows the interruption and skipping of the reference nucleotide sequence whereas the electropherogram from the amplicon in Lane 7 shows an uninterrupted reference sequence. The asterisk at genomic position chr17:15,230,858 corresponds to the common SNP rs13422 (MAF G = 0.32), which was present homozygously in both the proband’s parents as well as the proband and her daughter. The predicted miR-29a target site is also illustrated in respect to the PMP22 3′-UTR deletion.

Figure 4

The 3′-UTR deletion elevates PMP22 transcript levels. (A) For skin biopsy samples from controls (n = 16), HNPP (n = 1) and CMT1A patients (n = 15) and the proband (from two duplicate skin biopsies) in this study, purified RNA was submitted for Nanostring transcript quantitation analysis. The box and whisker plot shows PMP22 transcript values for the samples after normalization to Schwann cell-specific genes. (B) The graph shows normalized reporter activity from luciferase plasmids containing the human PMP22 3′-UTR or a version in which the neuropathy-associated deletion was introduced (3′-UTR del). The transfections were performed in the RT4 Schwann cell and Oli-neu oligodendrocyte cell lines. Error bars indicate standard deviations. (C) The two reporter plasmids were transfected with a control siRNA or a miR-29a mimic in RT4 cells, and resulting luciferase activities indicate that deletion of the 3′-UTR renders the luciferase reporter resistant to miR-29a repression. Reduction of luciferase activity by miR-29a co-transfection was significant (P < 0.02, t-test, n = 3) for the wild-type construct. HNNP = hereditary neuropathy with liability to pressure palsies.