Discovery and validation of autosomal dominant Alzheimer’s disease mutations

Background Alzheimer’s disease (AD) is a neurodegenerative disease that is clinically characterized by progressive cognitive decline. Mutations in amyloid-β precursor protein (APP), presenilin 1 (PSEN1), and presenilin 2 (PSEN2) are the pathogenic cause of autosomal dominant AD (ADAD). However, polymorphisms also exist within these genes. Methods In order to distinguish polymorphisms from pathogenic mutations, the DIAN Expanded Registry has implemented an algorithm for determining ADAD pathogenicity using available information from multiple domains, including genetic, bioinformatic, clinical, imaging, and biofluid measures and in vitro analyses. Results We propose that PSEN1 M84V, PSEN1 A396T, PSEN2 R284G, and APP T719N are likely pathogenic mutations, whereas PSEN1 c.379_382delXXXXinsG and PSEN2 L238F have uncertain pathogenicity. Conclusions In defining a subset of these variants as pathogenic, individuals from these families can now be enrolled in observational and clinical trials. This study outlines a critical approach for translating genetic data into meaningful clinical outcomes.


Background
Alzheimer's disease (AD) is characterized clinically by progressive cognitive decline and neuropathologically by progressive neuronal loss and the accumulation of amyloid plaques and neurofibrillary tangles. Mutations in amyloid-β precursor protein (APP), presenilin 1 (PSEN1) and presenilin 2 (PSEN2) are the pathogenic cause of autosomal dominant AD (ADAD). More than 200 pathogenic mutations have been identified in APP, PSEN1, and PSEN2 (reviewed in [1,2]). PSEN1 and PSEN2 form the catalytic domain of the γ-secretase complex, which is involved in sequential cleavage of APP into amyloid-β (Aβ) peptides.
The Dominantly Inherited Alzheimer Network (DIAN) is an observational study designed to follow families with mutations in APP, PSEN1, and PSEN2 that cause ADAD [3]. The DIAN Expanded Registry (DIAN EXR; www.dianexr.org) is a web-based registry with global outreach to ADAD families and investigators. It functions to identify families with ADAD and to determine the causative mutations through a genetic discovery program. This program includes genetic counseling and testing for known and unknown causes of ADAD. During a period of 6 years (2011 to 2017), the DIAN Clinical/Genetics committee reviewed 150 pedigrees, 76 of which were approved for genetic counseling and testing. Forty-six probands were positive for a coding variant in APP, PSEN1, or PSEN2, including 39 individuals with known pathogenic mutations, 6 with variants of unknown pathogenic significance, and 1 variant that was determined to be an AD risk factor.
In some cases, families are identified with several generations of early-onset AD; however, at the time of enrollment, whether a pathogenic mutation is the cause of disease in these families remains unknown. We performed genetic analyses in research participants from six densely affected early-onset AD pedigrees and identified variants in APP, PSEN1, and PSEN2 with unknown pathogenicity. In all families examined, we lacked sufficient genetic material to perform segregation analyses (e.g., DNA was available from only one or two family members). To assess the pathogenicity of novel variants in APP, PSEN1, and PSEN2 when pedigree and clinical data are limited or incomplete, Guerreiro and colleagues [4] proposed a pathogenicity algorithm. In the present study, we modified and expanded this algorithm to evaluate the pathogenicity of these six variants using genetic, biochemical, biomarker, and clinical data.

Genetic screening
Genomic DNA was extracted from peripheral blood lymphocytes using standard protocols. All coding exons in APP, PSEN1, and/or PSEN2 were amplified and sequenced using the BigDye Terminator version 3.1 cycle sequencing kit (Life Technologies, Carlsbad, CA, USA) and analyzed on an ABI 3500 genetic analyzer (Life Technologies). Sequence analysis was performed using Sequencher software (Gene Codes, Ann Arbor, MI, USA).

Bioinformatics
To determine whether APP, PSEN1, and PSEN2 variants represented rare or common polymorphisms, we investigated two population-based exome sequencing databases: the Exome Variant Server (EVS) and Exome Aggregation Consortium (ExAC) browser. Polymorphism phenotype v2 (PolyPhen-2; [5]) and Sorting Intolerant From Tolerant (SIFT) were used to predict whether the amino acid change would be disruptive to the encoded protein.

PiB imaging
Structural magnetic resonance imaging (MRI) was performed using 3-T scanners and following the Alzheimer's Disease Neuroimaging Initiative (ADNI) protocol [7,8]. T1-weighted scans were processed through FreeSurfer. Positron emission tomography with Pittsburgh compound B (PiB) was coregistered with MRI images. The standardized uptake value ratio (SUVR) was calculated for each region [9,10]. Global burden was characterized using a summary measure [9]. These data were collected for a PSEN1 M84V carrier and noncarrier.

Biochemical analysis Plasmids and mutagenesis
The full-length PSEN1 complementary DNA (cDNA) was cloned into pcDNA3.1 Myc/His vector [11]. The M84V, c.379_382delXXXXinsG, or A396T variant was introduced into the PSEN1 cDNA using the QuikChange II XL Site-Directed Mutagenesis Kit (Agilent Technologies, Santa Clara, CA, USA). Clones were sequenced to confirm the presence of the variant and the absence of additional modifications. PSEN1 wild type (WT) and the pathogenic PSEN1 A79V mutation were included as controls.
The full-length PSEN2 cDNA was cloned into pcDNA3.1 vector [12]. The L238F or R284G variant was introduced into the PSEN2 cDNA and screened as described above. PSEN2 WT and the pathogenic PSEN2 N141I mutation were included as controls [13].
The full-length APP cDNA (isoform 695) was cloned into pcDNA3.1 [14]. The T719N variant was introduced into the APP cDNA and screened as described above. APP WT and the pathogenic APP KM670/671NL(Swe) mutation were included as controls.

Transient transfection
To assess novel PSEN1 and PSEN2 variants, we used neuroblastoma cells (N2A) stably expressing human APP WT (695 isoform; N2A695). To assess novel APP variants, we used N2A. N2A cells were maintained in equal amounts of DMEM and Opti-MEM, supplemented with 5% FBS, 2 mM L-glutamine, 100 μg/ml penicillin/streptomycin, and, for stable cells, 200 μg/ ml G418 (Life Technologies). Upon reaching confluence, cells were transiently transfected with Lipofectamine 2000 reagent (Life Technologies). Culture media were replaced after 24 hours, and cells were incubated for another 24 hours prior to analysis. Four independent transfections were performed for each construct and used for subsequent analyses.

Aβ enzyme-linked immunosorbent assay
Conditioned medium was collected and centrifuged at 3000 × g at 4°C for 10 minutes to remove cell debris. The levels of Aβ 40 and Aβ 42 were measured in cell culture media by sandwich enzyme-linked immunosorbent assay (ELISA) as described by the manufacturer (Life Technologies). To account for variability in transfection efficiency between experiments, ELISA values were obtained (pg/ml) and corrected for total intracellular protein (μg/ml). Statistical difference was measured using an unpaired Student's t test.

PSEN1 M84V
The proband was identified in a family with three generations of early-onset AD and with a mean age at onset of 59 years ( Fig. 1; pedigree not shown to avoid disclosing mutation status to asymptomatic proband and other family members). Sequencing of the proband revealed a single base substitution (ATG to GTG) at codon 84 in exon 4 of PSEN1, resulting in a methionine-to-valine change (M84V). PSEN1 M84V was also identified in a family-based sequencing study of AD (National Institute of Mental Health Alzheimer's Disease Genetics Initiative Study) [15]. Two PSEN1 M84V carriers were confirmed to have had AD at autopsy with ages at symptomatic onset of 70 and 72 years. The third carrier was cognitively normal at age 67 years [15].
To determine whether PSEN1 M84V represents a rare polymorphism, we examined two population-based exome sequencing databases (Table 1). PSEN1 M84V was absent from both EVS and ExAC. PSEN1 M84L was identified in one individual in the ExAC browser. PSEN1 M84V was also absent in more than 1700 AD and control samples with whole-exome sequencing.
CSF Aβ, total tau, and phosphorylated tau were measured in the PSEN1 M84V carrier, who was cognitively normal at the time of lumbar puncture. CSF Aβ (361.5 pg/ml), total tau (84.93 pg/ml), and p-tau 181 (22.02 pg/ml) at 11 years prior to the parental age at symptomatic onset was consistent with other presymptomatic pathogenic mutation carriers but not completely distinct from noncarriers [3,6]. The PSEN1 M84V carrier produced a mean cortical PiB SUVR value of 1.132 at 15 years prior to the parental age at symptomatic onset and 1.209 at 11 years prior to parent age of disease onset (Fig. 2). A related noncarrier was imaged at 16 years prior to the parent age of disease onset and produced mean cortical PiB SUVR values of 1.086 and 1.039 at 13 years prior to age at disease onset in the parent (Fig. 2). Thus, fluid and imaging biomarkers were consistent with those observed in presymptomatic ADAD mutation carriers.
We next sought to determine whether the PSEN1 M84V variant alters Aβ isoform levels in a manner consistent with previously reported pathogenic PSEN1 mutations. We expressed vectors containing PSEN1 WT, A79V (a known pathogenic mutation), and M84V in N2A695 cells. We found that cells expressing PSEN1 M84V produced significantly more Aβ 42 than cells expressing PSEN1 WT (Fig. 3a-c). We also observed a significant increase in the Aβ 42 /Aβ 40 ratio in cells expressing PSEN1 M84V or the pathogenic PSEN1 A79Vcompared with PSEN1 WT-expressing cells (Fig. 3c).
Thus, we can apply a pathogenicity algorithm, modified from Guerreiro and colleagues [4], to assess the pathogenicity of PSEN1 M84V (Fig. 4). PSEN1 M84V occurs within exon 4 and the first transmembrane domain. This residue is highly conserved between PSEN1 and PSEN2 [16]. Additionally, a pathogenic mutation has been reported at this site: ΔI83/M84 [17,18]. In a cell model, Aβ levels were consistent with pathogenic mutations (Fig. 3a-c, g). Thus, PSEN1 M84V satisfies all criteria for pathogenicity based on residue and Aβ levels [4]. Taken together, we propose that PSEN1 M84V represents a pathogenic mutation.

PSEN1 A396T
The proband was identified in a family with three generations of early-onset AD (Fig. 1a). The proband had an age at symptomatic onset of 50 years. The parent of the proband had an age at symptomatic onset of 57 years, with AD confirmed at autopsy at age 67 years. Sequencing of the proband (Fig. 1a) revealed a single base substitution (GCG to ACG) at codon 396 in exon 11 of PSEN1, resulting in an alanine-to-threonine change (A396T). This variant was reported previously in one individual with sporadic AD [19]. PSEN1 A396T was absent in the EVS and ExAC databases (Table 1). We found that cells expressing PSEN1 A396T produced significantly more Aβ 42 than cells expressing PSEN1 WT (Fig. 3a-c). Thus, applying the algorithm for assessing pathogenicity (Fig. 4), we propose that the PSEN1 A396T represents a probable pathogenic mutation.

PSEN1 c.379_382delXXXXinsG
The proband had an age at symptomatic onset of 50 years, with an average age at symptomatic onset in family members aged 43 years (Fig. 1b)   the proband (Fig. 1b) revealed a 4-bp deletion and insertion of G at codon 127 in exon 5 of PSEN1, resulting in the deletion of glutamine and arginine and insertion of glycine (c.379_382delXXXXinsG). This is a novel variant and was absent from the EVS and ExAC databases ( Table 1). N2A695 expressing PSEN1 c.379_382delXXX-XinsG produced Aβ 42 and Aβ 40 levels similar to PSEN1 WT (Fig. 3a-c). Thus, we propose that PSEN1 c.379_382delXXXXinsG is an AD risk factor or benign polymorphism.

PSEN2 R284G
The proband had an age at symptomatic onset of 58 years, with average age at symptomatic onset in the family of 56 years (Fig. 1c). Sequencing of the proband (Fig. 1c) revealed a single base pair substitution (CGG to GGG) at codon 284 in exon 8 of PSEN2, resulting in an arginine-to-glycine change (R284G). This is a novel variant and was absent from the EVS and ExAC databases (Table 1).
To assess the effects of PSEN2 variants on Aβ isoform levels, we expressed vectors containing PSEN2 WT, N141I, and R284G in N2A695 cells. Cells expressing PSEN2 N141I and R284G produced significantly more Aβ 42 than cells expressing PSEN2 WT (Fig. 3d and e). The Aβ 42 /Aβ 40 ratio was also significantly higher in cells expressing PSEN2 N141I and R284G (Fig. 3f ). Thus, applying the algorithm for assessing pathogenicity (Fig. 4), we propose that the PSEN2 R284G represents a probable pathogenic mutation.

PSEN2 L238F
The proband was identified in a family with three generations of early-onset AD (Fig. 1d). The proband had an age at symptomatic onset of 49. The parent of the proband was diagnosed with AD at 57 years of age and died at age 77. Sequencing of the proband (Fig. 1d) revealed a single base pair substitution (CTT to TTT) at codon 238 in exon 7 of PSEN2, resulting in a leucine-to-phenylalanine change (L238F). PSEN2 L238F was also identified in two alleles in EVS, in two alleles in ExAC, and in one individual with sporadic AD (Table 1; [20]). CSF Aβ, total tau, and phosphorylated tau were measured in a PSEN2 L238F carrier related to the proband, who was cognitively normal at the time of lumbar puncture. CSF Aβ (506.54 pg/ml), total tau (38.75 pg/ml), and p-tau 181 (21.33 pg/ml) at 17 years prior to the parental age at symptomatic onset was consistent with biomarker levels in normal controls [3,6]. Cells expressing PSEN2 L238F produced Aβ 42 and Aβ 40 levels similar to PSEN2 WT (Fig. 3d-f). Thus, we propose that PSEN2 L238F represents an AD risk factor or benign polymorphism.

App T719N
The proband was identified in a family with two generations of early-onset AD (Fig. 1e). The proband had an age at symptomatic onset of 45 years. The parent of the proband was diagnosed at 45 years of age. Sequencing of the proband (Fig. 1e) revealed a single base pair substitution (ACC to AAC) at codon 719 in exon 17 of APP, resulting in a threonine-to-asparagine change (T719N). APP T719N was absent in the EVS and ExAC databases and was detected in one individual with early-onset AD (Table 1) [21]. Cells expressing APP T719N produced significantly elevated levels of Aβ 42 and Aβ 42 /Aβ 40 relative to APP WT (Fig. 3g-i ). Thus, we propose that APP T719N is a probable pathogenic mutation.

Conclusions
By applying genetic, bioinformatic, and functional data to an algorithm to assess pathogenicity, we propose that the PSEN1 M84V, PSEN1 A396T, PSEN2 R284G, and APP T719N are likely pathogenic mutations resulting in ADAD, whereas PSEN1 c.379_382delXXXXinsG and PSEN2 L238F are likely benign polymorphisms. This algorithm was adapted and modified from a pathogenicity algorithm originally proposed by Guerreiro and colleagues [4]. We have expanded upon this algorithm in several important ways: (1) expanding the number of controls in the association analyses from 100 to 65,000 by exploiting the EVS and ExAC databases; (2) evaluating the bioinformatic functional findings (e.g., conservation between PSEN1 and PSEN2 and the presence of other mutations at the same residue) independent of the cell-based functional findings; and (3) incorporating cell-based assays to evaluate the impact of novel variants on Aβ levels. We propose that this modified approach to assessing pathogenicity provides an important pipeline for incorporating mutation data at several levels and that this algorithm may be adapted to impute pathogenicity when extensive genetic data are missing for affected families. Designation of a variant as pathogenic will allow individuals to enroll in observational and clinical trials for AD, with clear applications in clinical and research settings.