Mutation of FOXC1 and PITX2 induces cerebral small-vessel disease

Abstract

Patients with cerebral small-vessel disease (CSVD) exhibit perturbed end-artery function and have an increased risk for stroke and age-related cognitive decline. Here, we used targeted genome-wide association (GWA) analysis and defined a CSVD locus adjacent to the forkhead transcription factor FOXC1. Moreover, we determined that the linked SNPs influence FOXC1 transcript levels and demonstrated that patients as young as 1 year of age with altered FOXC1 function exhibit CSVD. MRI analysis of patients with missense and nonsense mutations as well as FOXC1-encompassing segmental duplication and deletion revealed white matter hyperintensities, dilated perivascular spaces, and lacunar infarction. In a zebrafish model, overexpression or morpholino-induced suppression of foxc1 induced cerebral hemorrhage. Inhibition of foxc1 perturbed platelet-derived growth factor (Pdgf) signaling, impairing neural crest migration and the recruitment of mural cells, which are essential for vascular stability. GWA analysis also linked the FOXC1-interacting transcription factor PITX2 to CSVD, and both patients with PITX2 mutations and murine Pitx2-/- mutants displayed brain vascular phenotypes. Together, these results extend the genetic etiology of stroke and demonstrate an increasing developmental basis for human cerebrovascular disease.

Figure 3. foxc1 and PITX2 regulate vascular smooth muscle cell numbers.

Number of sox10-positive neural crest cells that associated with the cerebral vasculature in WT embryos at 32 hpf (A) was significantly reduced in foxc1 morphants (B and C). Cerebral vascular mural cells expressed smooth muscle actin by 4 dpf (D), with fewer smooth muscle actin–positive cells observed in foxc1 morphants (E and F). Patients with PITX2-attributable ARS exhibited CSVD: (G) WMH, (arrows) and (H) dilated perivascular spaces (arrowheads). Compared with WT embryos (I), murine Pitx2^–/– mutants demonstrated reduced and discontinuous smooth muscle actin staining of large and small cerebral vessels (J). Original magnification, ×200 (A and B), ×100 (D and E), and ×400 (I and J).

Figure 2. Foxc1 regulates Pdgf signaling.

Transcripts for pdgfra (Pdgfrα) and pdgfrb (Pdgfrβ) were expressed in neural crest cells at 20 hpf (arrows) in WT embryos (A and C) and were highly downregulated in foxc1 morphants (B and D). At 48 hpf, expression of pdgfra was observed in major cerebral blood vessels and ventral head mesenchyme (E, arrow), with the latter being reduced in foxc1 morphants (F), whereas pdgfrb expression was unaltered (G and H). A genetic interaction was observed between subthreshold inhibition of foxc1 and pdgfra (I) or with treatments that affected both Pdgfrα and Pdgfrβ signaling, including dual-morpholino suppression (K) and the pan-Pdgfr inhibitor crenolanib (L). No interaction was observed using subthreshold inhibition of foxc1 and pdgfrb alone (J). Scale bar: 200 μm.

Figure 1. Aberrant FOXC1 function causes CSVD in patients and cerebral hemorrhage in zebrafish.

(A) Regional association plot (meta-analysis, CHARGE consortium) displaying linkage disequilibrium between the 10 SNPs with WMH. (B) Diagrammatic representation of the missense or nonsense FOXC1 mutations or CNV extent for the 18 patients imaged. (C–H) MRIs of patients with FOXC1 mutation (C–F), 6p25 deletion (G), and 6p25 duplication (H) displaying WMH (arrow), dilated perivascular spaces (arrowhead), and lacunar infarct (asterisk). (I and J) Injection of 2 ng of foxc1a morpholino (MO) and 2 ng of foxc1b morpholino resulted in cerebral hemorrhage (arrow) at 48 hpf compared with injection of equimolar amounts of 5-bp mismatched morpholinos. Overexpression of foxc1a mRNA (75 pg) caused cerebral hemorrhage (K). Scale bars: 200 μm.