An AluYa5 Insertion in the 3'UTR of COL4A1 and Cerebral Small Vessel Disease

Abstract

Importance: Cerebral small vessel diseases (CSVDs) account for one-fifth of stroke cases. Numerous familial cases remain unresolved after routine screening of known CSVD genes.

Objective: To identify novel genes and mechanisms associated with familial CSVD.

Design, setting, and participants: This 2-stage study involved linkage analysis and a case-control study; linkage analysis and whole exome and genome sequencing were used to identify candidate gene variants in 2 large families with CSVD (9 patients with CSVD). Then, a case-control analysis was conducted on 246 unrelated probands, including probands from these 2 families and 244 additional probands. All probands (clinical onset <age 55 years and ≥1 first-degree relative with CSVD) were referred to the French cerebrovascular referral center between 2013 and 2023. The large-scale gnomAD structural variant database and 467 healthy individuals of French ancestry were used as a control group.

Main outcomes and measures: A pathogenic AluYa5 insertion was identified within the COL4A1 3'UTR in the 2 large families with CSVD. Reverse transcriptase-quantitative polymerase chain reaction (RT-qPCR), Western blot, and long-read RNA sequencing were used to investigate outcomes associated with the insertion using patient fibroblasts. Clinical and magnetic resonance imaging features of probands with variants and available relatives were assessed.

Results: Among 246 probands (141 females [57.3%]; median [IQR] age at referral, 56 [49-64] years), 7 patients of French ancestry carried the insertion. This insertion was absent in 467 healthy French individuals in a control group (odds ratio, ∞; 95% CI, 2.78 to ∞; P = 5 × 10-4) and 10 847 individuals from the gnomAD structural variant database (odds ratio, ∞; 95% CI, 64.77 to ∞; P = 2.42 × 10-12). In these 7 patients' families, 19 family members with CSVD carried the insertion. RT-qPCR and Western blot showed an upregulation of COL4A1 mRNA (10.6-fold increase; 95% CI, 1.4-fold to 17.1-fold increase) and protein levels (2.8-fold increase; 95% CI, 2.1-fold to 3.5-fold increase) in patient vs control group fibroblasts. Long-read RNA sequencing data showed that the insertion was associated with perturbation in the use of canonical COL4A1 polyadenylation signals (approximately 87% of isoforms transcribed from the wild type allele vs 5% of isoforms transcribed from the allele with the insertion used the 2 distal canonical polyadenylation signals). The main clinical feature of individuals with CSVD was the recurrence of pontine ischemic lesions starting at an early age (17 of 19 patients [89.5%]).

Conclusions and relevance: This study found a novel mechanism associated with COL4A1 upregulation and a highly penetrant adult-onset CSVD. These findings suggest that quantitative alterations of the cerebrovascular matrisome are associated with CSVD pathogenesis, with diagnostic and therapeutic implications.

Figure 1.. Genealogical Trees of Families F1 to F7

CSVD indicates cerebral small vessel disease; MRI, magnetic resonance imaging.

Figure 2.. Magnetic Resonance Imaging Data of Patients From Family F1

The first column presents sagittal T1–weighted images. Fluid-attenuated inversion recovery images are presented in columns 2 to 4. All patients present pontine infarcts and a vascular leukoencephalopathy associated with hemispheric lacunes.

Figure 3.. Magnetic Resonance Imaging Data of Patients From Family F2

The first column presents sagittal T1–weighted images except for C (patient F2-12), which presents diffusion-weighted images. Fluid-attenuated inversion recovery images are presented in columns 2 to 4, except for A (patient F2-9) and B (patient F2-10) in column 4, which present gradient echo sequences, and D (patient F2-11) in column 3, which presents sagittal T1 images. All patients except patient F2-9 (A) present pontine infarcts and vascular leukoencephalopathy associated with hemispheric lacunes.

Figure 4.. Western Blot Analysis of Wild-Type and Variant COL4A1 Expressed in Transfected Cells and Endogenous Fibroblasts

A-B, A Western blot of conditioned medium and cell lysates from HEK293T cells transfected with wild type and variant COL4A1 and human skin fibroblasts from 4 healthy individuals in the control group (C1-C4), 2 patients carrying the AluYa5 insertion (P1-P2), and 1 patient duplicated at the COL4A1/2 locus. A, In the top panel, blot n°1 was probed with an anticollagen IV antibody, and in the middle and bottom panels, blot n°2 was probed with antifibronectin and anti-MMP2 antibodies used as loading controls (see antibody references in the eMethods in Supplement 1). C-D, Visualizations of relative protein expression levels are presented from A and B, respectively. Circles indicate COL4-dup, the patient duplicated at the COL4A1/2 locus; HEK, HEK293T cells; horizontal lines, means; MW, molecular weight; NT, nontransfected plasmid; pCMV-Col4a1, plasmid-expressing COL4A1; squares, patients with AluYa5 insertion; triangles, control group.

Figure 5.. AluYa5 Insertion and Polyadenylation Signal (PAS) Usage

A, The diagram shows the last exon and 3′UTR of COL4A1 with the 6 PAS annotated as c.*distance to the stop codon. B, The diagram of the 3′UTR shows the 6 possible isoforms depending on PAS usage. The 6 isoforms differ only by the length of their 3′UTR. Cleavage and polyadenylation start approximately 15 to 30 bp downstream of the PAS. C, Proportions of expressed isoforms depending on PAS usage are shown in 3 patients carrying the insertion and 7 individuals in the control group. D, Proportions of each of expressed isoform by the wild type and variant allele depending on PAS usage in patients F2-12 and F8-16 are presented. T_AluYa5+ represents isoforms transcribed from the allele that carries the AluYa5 insertion. G_AluYa5− represents isoforms transcribed from the wild type allele.