Comparison of clinical and imaging features of cerebral small vessel disease associated with heterozygous HTRA1 and NOTCH3 mutations

Background: Heterozygous HTRA1 mutations are the second most common cause of monogenic dominant cerebral small vessel disease ( HTRA1 -AD-cSVD or CADASIL2), after cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) due to NOTCH3 cysteine-altering mutations. However, there have been few studies of cohorts of HTRA1 -AD-cSVD and whether it can be differentiated clinically and on neuroimaging from CADASIL is unclear. Aims: This retrospective study aims to characterize and compare the clinical and neuroimaging features of HTRA1 -AD-cSVD with those of CADASIL. Methods: We identified 21 unrelated Taiwanese subjects carrying 15 heterozygous HTRA1 variants, all functionally validated as pathogenic through in vitro protease activity assays. HTRA1 -AD-cSVD patients were compared with 406 CADASIL patients, including 44 cases carrying NOTCH3 mutations within the high-risk epidermal growth factor-like repeat domains (EGFr), 358 with moderate-risk EGFr mutations, and 4 with low-risk EGFr mutations. Multivariate regression analyses were conducted with adjustments for age at MRI examination and hypertension. Results: Stroke occurred in 81.0% of HTRA1 -AD-cSVD patients, and 47.6% exhibited cognitive dysfunction. MRI revealed moderate-to-severe white matter hyperintensity (WMH) in the deep white matter and external capsule (modified Scheltens’ scale: 5.3 ± 1.0 and 4.1 ± 1.7), mild WMH in the temporal pole (1.0 ± 1.7), lacunes in 90.5%, ⩾10 cerebral microbleeds (CMBs) in 66.7%, and intracranial hemorrhage (ICH) lesions in 46.7%, indicating susceptibility to both ischemic and hemorrhagic strokes. Patients with HTRA1 loss-of-function mutations or protease domain missense mutations exhibited a higher prevalence of ⩾10 CMBs on SWI/T2* imaging (100% and 83.3%) compared to those with missense mutations outside this domain (20%). Symptom onset occurred earliest in patients with NOTCH3 high-risk EGFr mutations (49.2 ± 10.5 years), followed by those with heterozygous HTRA1 mutations (54.3 ± 10.7 years), and latest in NOTCH3 moderate-risk EGFr mutations carriers (59.7 ± 9.5 years). Temporal pole involvement was most prevalent in NOTCH3 high-risk EGFr mutations (88.6%), followed by NOTCH3 moderate-risk EGFr mutations (32.4%), and least common in heterozygous HTRA1 mutations (28.6%). Even after adjusting for age and hypertension, HTRA1 -AD-cSVD patients exhibited significantly milder temporal pole WMH severity compared to NOTCH3 high-risk EGFr mutation carriers (adjusted p < 0.001). In addition, ICH lesions were more frequently observed in HTRA1 -AD-cSVD patients (46.7%) than in patients with NOTCH3 high-risk or moderate-risk EGFr mutations (18.2% and 21.2%), although the difference was not statistically significant. Conclusion: HTRA1 -AD-cSVD shares overlapping clinical and neuroimaging features with CADASIL. Temporal pole WMH involvement can occur in HTRA1 -AD-cSVD but is more common in CADASIL. The high prevalence of ICH in HTRA1 -AD-cSVD has been under-recognized. Data access statement: Data are available upon reasonable request from third parties.