Search for a command to run...
To the Editor: PSEN1, PSEN2, and APP have been verified as causative genes for familial Alzheimer’s disease (AD). Meta-analysis suggests that patients with AD causative mutations may develop symptoms at a younger age (average age of onset at 46.2 years).[1] However, some cases are late-onset or sporadic. Therefore, studies confined to early-onset or familial AD populations are incomprehensive, leading to an incomplete description of the genetic AD spectrum. In the pathogenesis of AD, amyloid beta (Aβ) is derived from amyloid precursor protein (APP) through proteolysis via the β-secretase and γ-secretase complex, whereas presenilin-1 (PS1) and presenilin-2 (PS2) are critical components of γ-secretase complex.[2] Previous research on transgenic animal models or transfected cells indicated that mutations could increase the level of Aβ (especially Aβ42) or Aβ42/Aβ40 ratio. This study aimed to screen for pathogenic mutations in a large sample of Chinese population with dementia (Peking Union Medical College Hospital [PUMCH] dementia cohort) and determine the functional effect of the identified mutation by in vitro study. Participants were recruited from the PUMCH dementia cohort in Beijing, China. This study was approved by the Ethics Committee of PUMCH (No. JS2810) and registered at ClinicalTrials.gov (NCT05023564). Written informed consent was obtained from all participants and their legal guardians. A total of 1277 participants diagnosed with dementia according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) were enrolled, including those with AD, frontotemporal dementia, vascular dementia, dementia with Lewy bodies, mixed dementia, and dementia with leukoencephalopathy. Data, including detailed clinical information, physical examination, and family history, were collected. All patients underwent standard blood tests, neuropsychological assessments, and magnetic resonance imaging (MRI), with detailed protocols provided in the Supplementary Materials, https://links.lww.com/CM9/C822. DNA was extracted from the peripheral blood of participants in the Clinical Biobank (ISO 20387) of PUMCH, which has standard quality control. Next-generation sequencing was performed on 1277 patients recruited between 2007 and 2019, among which, 267 underwnt whole exome sequencing (WGS), and the remaining 1010 underwent targeted panel sequencing (targeting on 278 dementia-related genes, including APP, PSEN1, and PSEN2). Next-generation sequencing technology was performed on Illumina HiSeq (Illumina Inc., San Diego, USA). All detected variants were verified by Sanger sequencing and classified as pathogenic, likely pathogenic, or of uncertain significance according to the American College of Medical Genetics and Genomics (ACMG) standards and searched in the database of Alzforum and PubMed to determine whether they were novel or reported. The inclusion criteria for analysis were based on bioinformatics analysis: filtering was restricted to causative AD genetic mutations; variants previously reported to be causative AD mutations; non-synonymous variants causing missense, frameshift, or splicing mutations; the frequency of variants in the general population (1000 Genomes project, Exome Aggregation Consortium [ExAC] database, 1000 healthy Chinese controls) must be <1%; the variants were predicted to be functional protein damage by three function prediction software analysis (Sorting Intolerant From Tolerant [SIFT], PolyPhen, and Likelihood Ratio Test [LRT]); benign and likely benign variants on ACMG classification were excluded. Methods for in vitro functional validation of variants of uncertain significance and statistical analysis methods are provided in the Supplementary Materials, https://links.lww.com/CM9/C822. This study enrolled 1277 participants, of which 331 had a family history of dementia, 754 were early-onset (onset age <65 years), and 154 have early-onset dementia and a family history of dementia. Forty-six patients (46/1277, 3.6%) harbored 36 variants. The demographic and clinical information of patients with AD causative mutations is summarized in Supplementary Table 1, https://links.lww.com/CM9/C822. Among the 46 patients with causative mutations, 23 patients (50.0%, 23/46) had family history of dementia, 32 patients (69.6%, 32/46) had early-onset dementia, and 17 (37.0%, 17/46) had early-onset dementia and a family history of dementia. The prevalence of causative AD mutations in those who have dementia family history, early-onset dementia, and early-onset dementia with a family history of dementia was 7.0% (23/331), 4.2% (32/754), and 11.0% (17/154), respectively. Among the 36 identified variants, 33 variants are missense mutations (10 APP, 11 PSEN1, and 12 PSEN2 mutations), two PSEN2 mutations are splice mutations, and one APP mutation is a non-frameshifting insertion mutation. According to the ACMG classification, five variants were pathogenic (one APP, three PSEN1, and one PSEN2), five variants were likely pathogenic (three APP and two PSEN1), and 26 variants were of uncertain significance (seven APP, eight PSEN1, and eleven PSEN2). Among the detected variants, 26 variants were novel. Sixteen patients (16/1277, 1.3%) carried 11 causative APP variants, including seven novel variants; 16 patients (16/1277, 1.3%) carried 13 causative PSEN1 variants, including 9 novel variants and 14 patients (14/1277, 1.1%) carried 12 causative PSEN2 variants, including 10 novel variants. The profiles of the identified mutations are summarized in Supplementary Table 2, https://links.lww.com/CM9/C822. The mutation sites in APP and PSEN1 identified in this study are indicated in the gene schematic diagrams [Figure 1A, B]. Functional validation was performed for six novel APP variants (APP E232K, 280insT, R397K, S614N, A344T, and L545V). The Human Embryonic Kidney 293 cells (HEK293) transfected with these six APP mutations show no influence on the secretion of Aβ42, Aβ40, and Aβ42/Aβ40 ratio relative to wild-type APP. Functional validation was performed for PSEN1 W165C and eight novel PSEN1 variants (PSEN1 Q223K, H214R, F447S, D458N, R27L, R42P, E72K and E363Q). The HEK293 cells transfected with PSEN1 E72K showed increased Aβ42 and Aβ40 levels, PSEN1 Q223K showed increased Aβ42 level, PSEN1 E363Q showed increased Aβ42 level and ratio of Aβ42/Aβ40 relative to wild-type PSEN1 [Figure 1C, D; Figure 1B [yellow]]. All the negative variants validated are illustrated in the Supplementary Material, https://links.lww.com/CM9/C822. Functional validation was performed for nine novel PSEN2 variants (PSEN2 N141S, M298T, I368F, S147N, M239T, L396I, G117X, H220Y, and R62C). These results were reported in our previous article.[3]Figure 1: Illustration of APP and PSEN1 variants. (A) Detected APP variants. (B) Detected PSEN1 variants. Variants with positive functional validation were highlighted in yellow. (C) PSEN1 Q223K elevated Aβ42 in HEK293 cells. (D) PSEN1 E72K and E363Q elevated Aβ42 and Aβ42/Aβ40 ratio in HEK293 cells. Aβ: Amyloid beta; APP: Amyloid precursor protein; PSEN1: Presenilin-1; Wt: Wild type. * P <0.05, † P <0.001, ‡ P <0.0001.The results of investigations on the frequency of AD causative mutations are inconsistent. The previously reported frequency of mutations in early-onset Alzheimer’s disease (EOAD) patients with family history of dementia ranges from 11–68%. This huge variation might be due to the different inclusion criteria, such as the different strategies of variant interpretation, definition of dementia family history of dementia, and included population. The detected mutation frequency in the Japanese population with early-onset dementia and a family history of dementia was 11.0% (5/45), which is similar to our results. In this study, among the 46 patients carrying causative mutations, 23 (50.0%, 23/46) had a family history of dementia, whereas 32 (69.6%, 32/46) had early-onset dementia. Notably, many patients have late-onset or sporadic disease. Possible reasons for this negative family history is that some patients might carry the de novo mutations that spontaneously arose in the proband, family members might be too young to manifest relative symptoms, or family members may show atypical symptoms rather than memory impairment. Thus, even late-onset or sporadic cases can carry causative mutations. An important study finding is the discovery of novel mutations and their associated functional validations. We identified seven novel APP, nine novel PSEN1, and ten novel PSEN2 mutations, of which three PSEN1 mutations (PSEN1 E72K, Q223K, and E363Q) and three PSEN2 mutations (PSEN2 N141S, I368F, and M239T, reported in another article) were validated by the in vitro functional examination to increase the secretion of Aβ or Aβ42/Aβ40 ratio, suggesting their pathogenicity. Research indicated that the mutation hotspots in APP are the region encoding Aβ and splice sites, whereas hotspots in PSEN1 and PSEN2 are the region encoding transmembrane domains that may impact the function of PS1 and PS2 as the active center of γ-secretase complex. PSEN1 E72K is not located in the transmembrane-encoding region, but has several near variants, such as E69D and A79V. PSEN1 E69D was functionally validated to increase levels of Aβ42 and Aβ40, whereas PSEN1 A79V variant was reported in patients with autopsy-confirmed AD. These two variants near PSEN1 E72K may be pathogenic, enhancing the possibility of PSEN1 E72K AD pathogenicity. PSEN1 E363Q is located in the encoding region of a large hydrophilic loop between the sixth and seventh transmembrane regions and is an important PS1 component. Importantly, the 363 glutamate showed high evolutionary conservation across species, and several variants have been reported nearby PSEN1 E363Q, such as L364P and S365A, indicating its possible important role in AD pathogenesis. PSEN1 Q223K is located in the fifth transmembrane-encoding region and is a mutation hotspots. The PSEN1 Q223R variant has been previously reported at the same site. Brain biopsies from patients carrying this variant revealed cortical cotton-wool plaques and neurofibrillary tangles.[4] Pathogenic mutation from the same site indicated the possible AD pathogenicity of PSEN1 Q223K. This study had some limitations. First, the PUCMH dementia cohort was a single-center study; thus, it may not reflect the profile of the Chinese population. Second, further functional validation of some reported mutations is required. Finally, not all the patients underwent WGS. Although the panel included common pathogenic mutations, some important undiscovered potential genes may have been missed. Therefore, WGS should be further optimized in subsequent studies. Overall, among the 1277 participants from the PUMCH dementia cohort, 46 patients with causative AD mutations (36 variants) were identified, with a frequency of 3.6% (1.3% of APP, 1.3% of PSEN1, and 1.1% of PSEN2). Three variants (PSEN1 E72K, E363Q, and Q223K) were identified to significantly increased the level of Aβ42, Aβ40, and Aβ42/Aβ40 ratio compared with wild-type variant by in vitro study. Acknowledgment We are grateful to the Clinical Biobank (ISO 20387) of PUMCH for their support in processing and storing the clinical specimens. Funding This research was supported by grants from the National Key Research and Development Program of China (Nos. 2020YFA0804500 and 2020YFA0804501), Chinese Academy of Medical Sciences (CAMS) Innovation Fund for Medical Sciences (Nos. 2020-I2M-C&T-B-010 and 2021-I2M-1-020), and National Natural Science Foundation of China (Nos. 81550021 and 30470618). Conflicts of Interest None.