Background
Amyloid plaque which is composed primarily of amyloid-beta (Aβ) has been reported as an important pathological change of some neurodegenerative diseases including Alzheimer’s disease (AD) [
1]. Aβ is generated from the transmembrane polypeptide called amyloid precursor protein (APP) by β- and γ-secretase enzymes [
1,
2]. β-secretase or β-site APP cleaving enzyme (BACE) is the important rate-limiting enzyme and its increased activity may lead to the elevation of Aβ in the brain [
1,
2]. Previous studies discovered that AD subjects have increased cerebrospinal fluid (CSF) BACE enzymatic activity compared with the controls [
3‐
12]. So researchers hypothesized that BACE enzymatic activities in CSF may become potential biomarkers for AD.
The recent development of a sensitive assay for BACE in CSF makes it feasible to study the association between CSF BACE enzymatic activity and AD [
13]. Moreover, the use of quantitative traits in genome-wide association study (GWAS) has been shown to increase statistical power over case-control designs [
14]. In this study we regard CSF BACE enzymatic activity as an endophenotype for a separate GWAS in the ADNI (Alzheimer’s Disease Neuroimaging Initiative database,
adni.loni.usc.edu) cohort in order to discover genetic factors involved in BACE protein.
Methods
Alzheimer’s Disease Neuroimaging Initiative
Data used in this study were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (
www.loni.ucla.edu/ADNI). The ADNI was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD, VA Medical Center and University of California–San Francisco. ADNI includes more than 800 participants ranging in age from 55 to 90. All these individuals were recruited from over 50 sites across the United States and Canada, including approximately 200 healthy controls (HC), 400 patients diagnosed with mild cognitive impairment (MCI) and 200 patients diagnosed with early AD. The AD patients were followed for 2 years and others were followed for 3 years. Structural 1.5-T magnetic resonance imaging (MRI) collected the full sample. PIB and FDG positron emission tomography (PET) imaging of a subset, some other biological markers, and performance on neuropsychological or clinical assessments were collected at baseline and at follow-up visits in 6- to 12-month intervals. The CSF BACE and genome-wide genotyping in this study were respectively available on approximately half of the cohort and the full ADNI sample. Further information about ADNI can be found in previous publications and at
www.adni-info.org [
15].
Standard protocol approvals, registrations, and patient consents
This study was approved by institutional review boards of all participating institutions and written informed consent was obtained from all participants or authorized representatives.
Participants
Our ADNI cohort included all healthy controls (HC), MCI group and AD group participants with available baseline CSF BACE samples and genotype data. To reduce the likelihood of population stratification effects in the GWAS, all the participants were restricted to non-Hispanic Caucasians. We also tested a multidimensional scaling (MDS) plot and found some genetic outliers (Additional file
1). After quality control (QC) of the CSF BACE data and removal of the outliers, there were 340 participants (AD = 86, MCI = 163, HC = 91) with CSF BACE data left. Detailed QC steps for CSF and genotype data have been previously reported and are briefly described below.
CSF BACE measurement and quality control
Samples were obtained from 382 ADNI subjects, enrolled at 56 participating centers using previously reported methods for CSF measurements [
13]. For most samples, the time from collection to freezing was within 60 min. Samples were processed, aliquoted, and stored at − 80 °C according to the ADNI Biomarker Core Laboratory Standard Operating Procedures [
16].
CSF BACE proteins of all the samples were tested by a solution-based BACE enzymatic assay which has been regarded as best assay format for BACE enzymatic activity in previous studies [
13]. This assay format uses a biotin labeled 15 amino acid peptide biotin-KTEEISEVNFEVEFR (NFEV) as the BACE substrate and uses a baculovirus expressed c-terminally truncated BACE (bBACE) as the BACE enzyme standard. A source of BACE using either purified recombinant truncated BACE, human or rhesus monkey CSF was co-incubated with this BACE substrate. The BACE cleavage product was then detected using an “anti-NF” neo-epitope specific rabbit polyclonal antibody and an indirect anti-rabbit horseradish peroxidase (HRP) or alkaline phosphatase (AP) development of the reaction [
13]. The BACE activity assay includes two steps (1 Enzyme+substrate, 2 ELISA to measure product) [
13]. Luminescence from assay plates was read on EnVision (PerkinElmer, model 2104). The counts from individual CSF samples were converted to BACE enzymatic activity using coefficients determined by a quadratic fit to the bBACE standard curve [
13].
Mean and standard deviations (SD) baseline of CSF BACE measures were calculated by observers blind to diagnostic information and subjects who had a value greater or smaller than 3-fold SD from the mean value were regarded as extreme outliers and removed from the analysis.
Genotyping and quality control
Single nucleotide polymorphism (SNP) genotyping for more than 620,000 target SNPs was completed on all ADNI participants using the following protocol. A total of 7 mL blood of each participant was taken in EDTA containing Vacutainer tubes and genomic DNA was extracted using the QIAamp DNA Blood Maxi Kit (Qiagen, Inc., Valencia, CA) following the manufacturer’s protocol. EBV-transformed B lymphoblastoid cell lines were established. Genomic DNA samples were analyzed using the Human 610-Quad BeadChip (Illumina, Inc., San Diego, CA) according to the manufacturer’s protocols (Infinium HD Assay; Super Protocol Guide; rev. A, May 2008).
Stringent QC assessment was performed using the PLINK software package (
http://pngu.mgh.harvard.edu/purcell/plink/), release v 1.07,15 as described previously. Stringent QC assessment followed these criteria: minimum call rate for SNPs and individuals > 95%, minimum minor allele frequencies (MAF) > 0.05, Hardy-Weinberg equilibrium test
P > 0.001. The restriction to SNPs with a MAF greater than 5% served to reduce the likelihood of false-positive results in the context of modest sample size to enhance statistical power [
14]. What’s more, elimination of relatively rare markers reduced the severity of the multiple comparison correction which in turn enhanced statistical power. After the QC procedure, all 340 participants remained in the analysis and only 519,442 out of 620,901 SNPs remained in the analysis. The overall genotyping rate for the remaining dataset was 99.5%.
Statistical analyses
To find the association of CSF BACE with the genetic polymorphism, a separate GWAS was performed using PLINK software under an additive genetic model. The thresholds of
P < 10
− 5 and
P < 10
− 8 were used for suggestive and genome-wide significant associations respectively [
17]. The analysis included a total of 519,442 genotyped variants. Age, gender and APOEε4 status were included as covariates. One-way analysis of variance (ANOVA) and Tukey’s multiple comparisons test were used to determine the difference of CSF BACE enzymatic activity in different diagnostic groups. The effects of genotypes on CSF BACE were examined with a multiple linear regression model. The relations between top SNP and AD-related phenotypes were also performed using PLINK software. Genome-wide associations were visualized by a software program (R, version 3.4.0; The R Foundation). Regional associations were visualized with the LocusZoom web tool (
http://locuszoom.org/).
Discussion
In this multicenter study, we identified a genome-wide significant association of a SNP (rs1481950) in the gene ATP6V1H region with CSF BACE activity and found seven additional suggestive association loci. Interestingly, this study is the first to show that the rs1481950 risk variant in ATP6V1H significantly affects CSF BACE activity. We found statistically significant differences in CSF BACE activity between the two genotype groups (GG + GT and TT). It is worth noting that this statistical differences between gene polymorphism of ATP6V1H and CSF BACE activity existed in both AD (P = 0.0027) and MCI groups (P = 0.0003). These results indicated that the rs1481950 risk variant (G) in ATP6V1H might significantly increase the CSF BACE activity especially in the AD and MCI individuals.
Rs1481950 is in an intronic region. Current studies haven’t found the exact mechanisms of impairing the expression of
ATP6V1H. However, the roles of intronic regions which have been ignored all the time have attracted the attention of scientists. And more and more studies have proved that intronic region may play some important roles in controling the initiation and termination of transcription. Moreover, common molecular mechanisms for an intronic SNP to alter mRNA levels are to affect transcription, RNA elongation, splicing, or maturation [
18‐
22].
The ATPase, H
+ transporting, lysosomal 50/57 kDa, V1 subunit H gene (
ATP6V1H) at Chr8q11.2 encodes for the V1H subunit of vacuolar ATPase (V-ATPase) [
23,
24].
ATP6V1H gene was mainly studied and discussed about its roles in diabetes in previous researches and the data showed that the down-regulation of its gene expression correlates with the presence of type-2 diabetes [
25]. Though there is still no research point out a direct relation between
ATP6V1H gene and AD, some studies about encoded protein and metabolic process of BACE indicate that mutations of
ATP6V1H gene may contribute to the increased BACE activity. V-ATPase belongs to the rotary ATPase family and is a multiprotein membrane complex. The most important function of V-ATPase is to acidify intracellular compartments by using the energy gathered from ATP hydrolysis to pump protons [
23]. One of the most important influence factors of BACE activity is PH. Acidic Intracellular environment (PH = 4.5) is optimal for BACE activity. So the dysfunction of V-ATPase may lead to the change of the acidic Intracellular environment and then influence the BACE activity. BACE activity is also associated with mature processing. The mature process of BACE requires the formation of disulfide bonds, glycosylation and some other modification processes [
26]. These steps occurs in the endoplasmic reticulum and golgi body and may be influenced by the changes of the intracellular environment. Moreover, V-ATPase also plays an important role in lysosomal acidification. Lysosomal pathway is an important degradation pathway of BACE protein [
27,
28]. To sum up, V-ATPase may play some important roles in both BACE protein levels and activity.
The results also showed that mutations of rs1481950 had significant correlations with the atrophy of AD-related encephalic regions including middle temporal gyrus, parahippocampal gyrus, hippocampus and entorhinal cortex. Moreover, the results showed that mutations of rs1481950 had significant correlations with the CSF p-tau. These results indicated that mutations of rs1481950 may relate to AD and may influence the volume of AD-related encephalic regions by changing the metabolism of some AD-related proteins such as tau.
In this study, we did not find any significant differences in CSF BACE activity among the three diagnostic groups. This aspect of our research is in line with two previous studies including an ADNI cohort study [
29,
30]. Some previous studies show different results including increased activity in MCI and AD [
31], and increased BACE1 activity in MCI but not in AD [
13,
32]. These inconsistent results may be explained by the characteristics of the study samples, the wide range of BACE1 activity measurements, and the large overlap between the groups.
There are several potential limitations of this study. First, the ADNI-1 sample was limited in sample size when CSF BACE data, different genotypes and diagnosis subgroups were taken into consideration, which makes effective sample size small for some tests. Thus, it will still be necessary to replicate these findings in a larger dataset. Second, recent studies also indicated that except CSF BACE, plasmatic BACE was also associated with AD [
33]. The associations among gene polymorphism, CSF BACE, plasmatic BACE and AD still need further study. Third, our sample was restricted to Caucasians to avoid genetics stratification across ethnicities, while all the genes may show different frequencies and polymorphisms in different populations. The relationships between these genes and AD need to be tested in more populations. Fourth, the ADNI database did not cover the detailed classification of the AD patients (sporadic and familial; early onset and late onset). So further studies with a better profile of AD patients should be tested for comparative purposes with other studies.
Conclusions
In summary, after a separate GWAS of CSF BACE, we found a top SNP (rs1481950) in ATP6V1H gene with the P value reaching genome-wide significance and seven suggestive SNPs with the P value lower than 10− 5. Rs1481950 risk variant (G) in ATP6V1H may increase the CSF BACE activity. Seven genes (SNX31, RORA, CDH23, RGS20, LRRC4C, MAPK6PS1, LOC105378355) were regarded as candidate genes. These results provide clues to some novel pathogenic genes associated with some BACE related diseases, such as AD. The in-depth discussion and study of these associations can help us to find the exact mechanisms of AD, which may indicate some new diagnostic methods and therapeutic directions.
Acknowledgements
Data collection and sharing for this project was funded by the Alzheimer’s Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). ADNI is funded by the National Institute on Aging, the National Institute of Biomedical Imaging and Bioengineering, and through generous contributions from the following: AbbVie, Alzheimer’s Association; Alzheimer’s Drug Discovery Foundation; Araclon Biotech; BioClinica, Inc.; Biogen; Bristol-Myers Squibb Company; CereSpir, Inc.; Cogstate; Eisai Inc.; Elan Pharmaceuticals, Inc.; Eli Lilly and Company; EuroImmun; F. Hoffmann-La Roche Ltd. and its affiliated company Genentech, Inc.; Fujirebio; GE Healthcare; IXICO Ltd.; Janssen Alzheimer Immunotherapy Research & Development, LLC.; Johnson & Johnson Pharmaceutical Research & Development LLC.; Lumosity; Lundbeck; Merck & Co., Inc.; Meso Scale Diagnostics, LLC.; NeuroRx Research; Neurotrack Technologies; Novartis Pharmaceuticals Corporation; Pfizer Inc.; Piramal Imaging; Servier; Takeda Pharmaceutical Company; and Transition Therapeutics. The Canadian Institutes of Health Research is providing funds to support ADNI clinical sites in Canada. Private sector contributions are facilitated by the Foundation for the National Institutes of Health (
www.fnih.org). The grantee organization is the Northern California Institute for Research and Education, and the study is coordinated by the Alzheimer’s Therapeutic Research Institute at the University of Southern California. ADNI data are disseminated by the Laboratory for Neuro Imaging at the University of Southern California.