Pooled-DNA sequencing identifies novel causative variants in PSEN1, GRN and MAPT in a clinical early-onset and familial Alzheimer's disease Ibero-American cohort

Abstract

Introduction: Some familial Alzheimer's disease (AD) cases are caused by rare and highly-penetrant mutations in APP, PSEN1, and PSEN2. Mutations in GRN and MAPT, two genes associated with frontotemporal dementia (FTD), have been found in clinically diagnosed AD cases. Due to the dramatic developments in next-generation sequencing (NGS), high-throughput sequencing of targeted genomic regions of the human genome in many individuals in a single run is now cheap and feasible. Recent findings favor the rare variant-common disease hypothesis by which the combination effects of rare variants could explain a large proportion of the heritability. We utilized NGS to identify rare and pathogenic variants in APP, PSEN1, PSEN2, GRN, and MAPT in an Ibero-American cohort.

Methods: We performed pooled-DNA sequencing of each exon and flanking sequences in APP, PSEN1, PSEN2, MAPT and GRN in 167 clinical and 5 autopsy-confirmed AD cases (15 familial early-onset, 136 sporadic early-onset and 16 familial late-onset) from Spain and Uruguay using NGS. Follow-up genotyping was used to validate variants. After genotyping additional controls, we performed segregation and functional analyses to determine the pathogenicity of validated variants.

Results: We identified a novel G to T transition (g.38816G>T) in exon 6 of PSEN1 in a sporadic early-onset AD case, resulting in a previously described pathogenic p.L173F mutation. A pathogenic p.L392V mutation in exon 11 was found in one familial early-onset AD case. We also identified a novel CC insertion (g.10974_10975insCC) in exon 8 of GRN, which introduced a premature stop codon, resulting in nonsense-mediated mRNA decay. This GRN mutation was associated with lower GRN plasma levels, as previously reported for other GRN pathogenic mutations. We found two variants in MAPT (p.A152T, p.S318L) present only in three AD cases but not controls, suggesting that these variants could be risk factors for the disease.

Conclusions: We found pathogenic mutations in PSEN1, GRN and MAPT in 2.33% of the screened cases. This study suggests that pathogenic mutations or risk variants in MAPT and in GRN are as frequent in clinical AD cases as mutations in APP, PSEN1 and PSEN2, highlighting that pleiotropy of MAPT or GRN mutations can influence both FTD and AD phenotypic traits.

Figure 1

Schematic of the study design. Pooled-DNA sequencing was performed in a single DNA pool of 172 individuals to identify known pathogenic or novel functional variants by using Illumina HiSeq 2000. The SPLINTER software was used to call the variants. High-confident variants were selected for Sequenom genotyping. For those validated functional variants, follow-up genotyping was performed in large case-control series to infer and compare the frequencies. Segregation analysis was then performed to determine whether disease status segregates with risk alleles. Enzyme-linked immunosorbent assays (ELISAs) were used to evaluate the impact of novel GRN splice-site mutation on the changes of GRN plasma levels. GRN, progranulin; SPLINTER, short indel prediction by large deviation inference and nonlinear true frequency estimation by recursion.

Figure 2

The transcript of the novel GRN mutation. A splice-site mutation with a CC-insertion (g.10974_10975insCC) in GRN exon 8 results in a premature stop codon and nonsense-mediated decay of the resultant mRNA. GRN, progranulin.

Figure 3

Enzyme-linked immunosorbent assays comparing GRN plasma levels between GRN mutation carriers and controls. Measurement of GRN plasma levels for a known GRN mutation (left) and the novel GRN mutation (right), Ile256IlefsX27, observed in our study compared with Alzheimer's disease cases and non-demented older controls. GRN, progranulin; mut, mutant; wt, wild-type.