NPC1 silent variant induces skipping of exon 11 (p.V562V) and unfolded protein response was found in a specific Niemann-Pick type C patient

Abstract

Background: Niemann-Pick type C (NPC, MIM #257220) is a neuro-visceral disease, caused predominantly by pathogenic variants in the NPC1 gene. Here we studied patients with clinical diagnosis of NPC but inconclusive results regarding the molecular analysis.

Methods: We used a Next-Generation Sequencing (NGS)-panel followed by cDNA analysis. Latter, we used massively parallel single-cell RNA-seq (MARS-Seq) to address gene profiling changes and finally the effect of different variants on the protein and cellular levels.

Results: We identified novel variants and cDNA analysis allowed us to establish the functional effect of a silent variant, previously reported as a polymorphism. We demonstrated that this variant induces the skipping of exon 11 leading to a premature stop codon and identified it in NPC patients from two unrelated families. MARS-Seq analysis showed that a number of upregulated genes were related to the unfolded protein response (UPR) and endoplasmic reticulum (ER) stress in one specific patient. Also, for all analyzed variants, the NPC1 protein was partially retained in the ER.

Conclusion: We showed that the NPC1 silent polymorphism (p.V562V) is a disease-causing variant in NPC and that the UPR is upregulated in an NPC patient.

Keywords: NPC1; Niemann-Pick type C; RNA-seq; exon skipping; silent variant; unfolded protein response.

FIGURE 1

The NGS‐targeted gene panel allowed the screening of two NPC patients. One harbors known disease‐causing variants and for the second patient, a functional study was conducted. (a) Workflow of the study of the patients. (b) Clinical manifestation timeline for patient P1, with the first symptoms at 6 years old (juvenile form), cognitive impairment, ataxia, dystonia, and vertical supranuclear gaze palsy (VSGP) at 21 years old, compatible with an NPC diagnosis. (c) Unesterified cholesterol was labeled with filipin and staining intensity calculated from three independent experiments for patient P1, compared with a classical (patient P5) and variant (siblings P3/P3′). Data are mean ± SEM, ***p < 0.001, **p < 0.01 by ANOVA with Tukey's post hoc test. ns = not significant

FIGURE 2

The silent variant p.V562V leads to the skipping of exon 11 and a premature stop codon (PTC). (a) The discrepancies between patient P1 gDNA (heterozygous for two nucleotide substitutions: c.1514T=G; p.V505G and c.1686G=A; p.V562V) and the cDNA (one allele more expressed than the other) led to the hypothesis of an unstable transcript. After segregation studies (chromatograms not shown) we observed that the allele inherited from the mother, that carries the variant c.1686G=A; p.V562V (b), is expressed in low abundance. Treatment with cycloheximide (CHX) partially stabilized the transcript (a, right and lower panel), suggesting NMD. (c) Agarose gel (ethidium‐bromide 1%) electrophoresis and sequencing chromatograms of cDNA‐derived PCR products (using PAXgene blood RNA) comprising exons 9‐13. The 697 bp amplified fragment represents the normal splice product and the 594 bp fragment represents the aberrant transcript, lacking exon 11. However, the upper band of 697 bp (corresponding to the normal length transcript) was sequenced separately (the band excised, purified, and sequenced) and it was shown that both the 1686G and 1686A alleles are present, suggesting that c.1686G=A; p.V562V is a leaky splicing variant. CTRL = control cDNA; negative = water control. At the boundaries of exon 10 and 11 it is possible to observe the second transcript, lacking exon 11 (c, right) and the PTC, marked in red. The same transcript was observed in the mother's cDNA (c, lower panel)

FIGURE 3

The silent variant p.V562 V was found in a second unrelated family (a) gDNA analysis of exons 9 and 11 in siblings P3 and P3′ (Family 2—F2), showing the variants in heterozygosity. (b) Subsequent/additional analysis on the parents and a non‐affected sister shows the segregation of the variants. (c) Sequencing chromatogram of cDNA‐derived PCR product comprising exons 9–13; at the boundaries of exon 10 and 11 it is possible to observe the second transcript, lacking exon 11 and with the premature stop codon. (d) Schematic representation of p.V562V localization on exon 11 and the effect on splicing based on in silico predictions (Human Splicing Finder—HSF and EX‐SKIP tools). EX‐SKIP compares the Exonic Splicing Enhancer (ESE)/Exonic Splicing Silencer (ESS) profile of a wild type (wt) and a mutated allele to determine if a specific exonic variant increases the chance of exon skipping. It calculates the total number of ESSs, ESEs, and their ratio. The V562V mutant is associated with a change in the ESS/ESE ratio which is compatible with a higher chance of exon skipping than in the wt allele. In addition, HSF (a tool to predict the effects of mutations on splicing signals or to identify splicing motifs in any human sequence) predicts that the V562V mutant leads to the creation of an ESS site. It involves the cDNA sequences CTTGTAAT (Zhang & Chasin, 2004) or CTTGTA (Goren et al., 2006), which might be associated with a potential alteration of splicing

FIGURE 4

Patient P1 has differentially expressed genes (upregulated) associated with the UPR, but other NPC1 mutants exhibit delayed trafficking to lysosome. (a) List of genes with altered expression (p < 0.05, fold change =1.5) in P1 patient compared with P4, P5, and controls and P4 and P5 versus controls. Those that are upregulated exclusively in P1 are marked red. (b) After cluster analysis, the results showed that the most part of the genes are associated with ER stress and the UPR. (c) Relative expression levels of mRNA for HSPA1A, HSPA5, HSP90B1, and HMOX1 genes (four of the obtained hits). Error bars represent standard deviation (SD) from three independent experiments. ***p < 0.001 P1 versus CTRL by one‐way ANOVA with Bonferroni's multiple comparison post hoc test (n = 3 in each group); ns = not significant. (d) CTRL and NPC1 primary human fibroblasts were treated with PNGaseF to deglycosylate the NPC1 protein (all N‐linked oligosaccharides) and Endo H, and subjected to Western blotting (WB) using anti‐NPC1 antibody. In CTRL the NPC1 protein is mostly Endo H‐resistant, but in all patients, the Endo H‐sensitive form is more abundant, suggesting that part of the protein was retained in the ER, and did not go for complex sugars formation, a process that only happens in the Golgi. Bands corresponding to complex glycosylated (Endo H‐resistant) and mannose rich forms (Endo H‐sensitive) were quantified and are displayed as Endo H‐sensitive/Endo H‐resistant ratio. (e) Immunostaining followed by confocal laser scanning microscopy of CTRL fibroblasts revealed that endogenous NPC‐1 protein colocalizes with LAMP1 and partially colocalizes with calnexin. In primary human fibroblasts of patient P1, NPC1 colocalizes much less with LAMP1. Instead, it presents a higher colocalization with calnexin. Scale bar 25 µm