The wide genetic landscape of clinical frontotemporal dementia: systematic combined sequencing of 121 consecutive subjects

Abstract

PurposeTo define the genetic spectrum and relative gene frequencies underlying clinical frontotemporal dementia (FTD).MethodsWe investigated the frequencies and mutations in neurodegenerative disease genes in 121 consecutive FTD subjects using an unbiased, combined sequencing approach, complemented by cerebrospinal fluid Aβ_1-42 and serum progranulin measurements. Subjects were screened for C9orf72 repeat expansions, GRN and MAPT mutations, and, if negative, mutations in other neurodegenerative disease genes, by whole-exome sequencing (WES) (n = 108), including WES-based copy-number variant (CNV) analysis.ResultsPathogenic and likely pathogenic mutations were identified in 19% of the subjects, including mutations in C9orf72 (n = 8), GRN (n = 7, one 11-exon macro-deletion) and, more rarely, CHCHD10, TARDBP, SQSTM1 and UBQLN2 (each n = 1), but not in MAPT or TBK1. WES also unraveled pathogenic mutations in genes not commonly linked to FTD, including mutations in Alzheimer (PSEN1, PSEN2), lysosomal (CTSF, 7-exon macro-deletion) and cholesterol homeostasis pathways (CYP27A1).ConclusionOur unbiased approach reveals a wide genetic spectrum underlying clinical FTD, including 11% of seemingly sporadic FTD. It unravels several mutations and CNVs in genes and pathways hitherto not linked to FTD. This suggests that clinical FTD might be the converging downstream result of a delicate susceptibility of frontotemporal brain networks to insults in various pathways.

Figure 1

Relative frequencies of mutations in neurodegenerative disease (NDD) genes in a consecutive series of 121 subjects with clinical frontotemporal dementia (FTD). Twenty-three subjects carried mutations, which were distributed across common FTD genes (C9orf72 repeat expansion, GRN, but surprisingly not MAPT or TBK1), less common FTD genes (CHCHD10, SQSTM1, TARDBP, UBQLN2), and also NDD genes not commonly linked to FTD (PSEN1, PSEN2, CTSF, CYP27A1) (a). Mutations were found not only in 34% of familial subjects, but also in 11% of sporadic subjects (b).

Figure 2

Copy-number variants in GRNand CTSF detected by whole-exome sequencing. Two deletions were identified, affecting exons 2–12 of GRN in subject 18167 (a) and exons 1–6 of CTSF (plus exons 8–21 of ACTN3) in subject 19566 (b), respectively. The start and end points of both deletions were located in regions captured by the exome and the exact start and end points could be determined by visualizing the sequence data in the integrative genomics viewer (GRN deletion chr17:42,426,438–42,430,018; CTSF deletion chr11:66,323,324–66,333,606 (hg19)).

Figure 3

Examples of rare genetic causes of clinical frontotemporal dementia (FTD): brain imaging and pedigrees of CTSF,CHCHD10 and CYP27A1 mutation carriers. (a–c) CTSF subject. MRI of the CTSF subject (19566) showed frontotemporal atrophy (a) and thinning of the corpus callosum (b), but no definite white matter hyperintensity, demonstrating that CTSF mutations can present even with only unspecific FTD brain imaging changes. Family history revealed adult-onset behavioral change and cognitive decline in the deceased brother who was diagnosed with “Huntington disease” (c). (d–f) CHCHD10 subject. MRI of the CHCHD10 subject (21854) revealed bilateral frontal atrophy (d), mild cerebellar atrophy (e), and thinning of the corpus callosum (e). This subject appeared to be sporadic, but family history was incomplete owing to early death of the father (f). (g–j) CYP27A1 subject. The MRI of the CYP27A1 subject (23660) also showed predominantly temporal and frontal atrophy (g), and only unspecific, mild periventricular white matter changes (h), but no characteristic signal alterations of the dentate nucleus (i). This demonstrates that CYP27A mutations can also present with only unspecific FTD brain imaging changes, and can thus easily be overlooked in clinical practice. Clinical workup was at first misdirected by the presumed autosomal–dominant pattern of inheritance of a neuropsychiatric disease (j), before next-generation sequencing unraveled clearly pathogenic autosomal–recessive CYP27A1 mutations in the index subject, indicating that there must be other causes for the neuropsychiatric diseases in the other family members (for a more detailed discussion of the subject’s family history, see Supplementary Material S7).