Novel Alzheimer's disease (AD) risk factors have been identified by genome-wide association studies. Elucidating the mechanism underlying these factors is critical to the validation process and, by identifying rate-limiting steps in AD risk, may yield novel therapeutic targets. Here, we evaluated the association between the AD-associated polymorphism rs3851179 near PICALM, which encodes a clathrin-coated pit accessory protein. Immunostaining established that PICALM is expressed predominately in microvessels in human brain. Consistent with this finding, PICALM mRNA expression correlated with expression of the endothelial genes vWF and CD31. Additionally, we found that PICALM expression was modestly increased with the rs3851179A AD-protective allele. Analysis of PICALM isoforms found several isoforms lacking exons encoding elements previously identified as critical to PICALM function. Increased expression of the common isoform lacking exon 13 was also associated with the rs3851179A protective allele; this association was not apparent when this isoform was compared with total PICALM expression, indicating that the SNP is associated with total PICALM expression and not this isoform per se. Interestingly, PICALM lacking exons 2–4 was not associated with rs3851179 but was associated with rs592297, which is located in exon 5. Thus, our primary findings are that multiple PICALM isoforms are expressed in the human brain, that PICALM is robustly expressed in microvessels, and that expression of total PICALM is modestly correlated with the AD-associated SNP rs3851179. We interpret these results as suggesting that increased PICALM expression in the microvasculature may reduce AD risk.
Citation: Parikh I, Fardo DW, Estus S (2014) Genetics of PICALM Expression and Alzheimer's Disease. PLoS ONE 9(3): e91242. https://doi.org/10.1371/journal.pone.0091242
Editor: David R. Borchelt, University of Florida, United States of America
Received: September 16, 2013; Accepted: February 10, 2014; Published: March 11, 2014
Copyright: © 2014 Parikh et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The authors acknowledge the National Institute on Aging for funding (P01-AG030128 (Estus) and K25-AG043546 (Fardo)), the University of Kentucky Alzheimer's Center for tissue (P30-AG028383) and James Simpson for technical assistance. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Alzheimer's disease (AD) is a devastating disease marked by cognition and memory decline, affecting the elderly population. Twin and family-based studies suggest that sporadic late onset AD risk is genetically linked , . Recent genome wide association studies (GWAS) have identified loci of genetic variance, single nucleotide polymorphisms (SNP)s, that are associated with AD risk –. Elucidating the mechanism of action of these SNPs validates the SNP as an AD risk factor and may identify novel AD pathways. Additionally, since steps in AD pathways that are modulated by genetics may be susceptible to pharmacologic manipulation, identifying the actions of AD-associated SNPs may lead to robust new pharmacologic targets.
One of these SNPs is near the gene PICALM (phosphatidylinositol binding clathrin assembly protein) which is involved in endocytosis. The primary AD-associated SNP is rs3851179 –, which lies approximately 80 kb 5′ of PICALM. PICALM itself is encoded by 21 exons, several of which are variably spliced . Here, we sought to elucidate how rs3851179 alters PICALM expression or splicing to modulate AD risk. We report that PICALM is expressed robustly in microvessels and moderately in other cell types. Rs3851179 was modestly associated with total PICALM expression as well as the major PICALM isoform lacking exon 13. In contrast, the expression of rare PICALM isoforms lacking exons 2, 2–4, or 18–19 was not associated with rs3851179. We interpret our results as suggesting that the PICALM is robustly expressed in microvessels and that the protective rs3851179A allele is associated with modestly increased PICALM expression. We speculate that increased microvessel PICALM reduces AD risk, perhaps by facilitating Aβ clearance from the brain through enhanced translocation across the blood brain barrier.
Materials and Methods
The work described here was performed with approval from the University of Kentucky Institutional Review Board.
The RNA and DNA samples for this study were from de-identified AD and non-AD autopsy samples. Anterior cingulate specimens were provided by the University of Kentucky AD Center Neuropathology Core and have been described previously , . A total of 52 brain samples, 28 male and 24 female, were used for this study. All of the non-AD individuals were cognitively intact at their last visit (MMSE of 27.9±3.4 (mean ± SD)). AD individuals were demented (MMSE of 12.2±8.3). For the AD autopsy samples, the average age at death and postmortem interval was 82.9±6.4 years (mean ± SD) and 3.4±0.7 hrs, respectively. For the non-AD samples, the average age at death and postmortem interval was 82.3±8.7 years (mean ± SD) and 2.8±0.8 hrs, respectively. By NIARI neuropathology criteria, the non-AD individuals included 21 samples with a score of no-low likelihood of AD, and 6 with intermediate. The AD samples were uniformly high-likelihood of AD. RNA was prepared by the method of Chomczynski and Sacchi  and converted to cDNA with random hexamers and Superscript II, as described previously . Although RNA integrity analyses were not performed prior to reverse transcription, others have demonstrated that for qPCR with short amplicons, normalized expression differences are comparable between samples with moderate RNA degradation and those with high integrity RNA .
Paraffin-embedded anterior cingulate tissue sections (5 μm thick) were rehydrated, underwent heat-induced antigen retrieval in citrate buffer (pH 6.0) and were quenched in 0.3% H2O2. Sections were immersed in 10% rabbit serum in Tris-buffered saline followed by an overnight incubation in anti-CALM (sc-6433, Santa-Cruz; 1∶400 dilution). After thorough rinsing in Tris-buffered saline, sections were incubated in biotinylated secondary antibody for 1 h, rinsed, incubated in ABC reagent (Vector) for 1 h, developed in Nova Red chromagen (Vector) and counterstained with Hematoxylin.
Identification of PICALM Splice Variants in Human Brain
Screening for PICALM splice variants was performed on a pool of cDNA samples from five AD and five non-AD individuals. This cDNA pool was amplified by PCR by using forward and reverse primers designed to produce overlapping products; this enabled evaluation of splicing efficiency of each internal exon (Table 1). The identity of splice variants was determined by sequencing. To estimate the distribution of these splice variants, exon 12–20 PCR products from three rs3851179 homozygous minor (A/A) and three homozygous major (G/G) individuals were TA-cloned (Invitrogen) and 847 random clones were sequenced. For this work, thirty cycles of PCR (Platinum Taq, Invitrogen) were performed by using primers corresponding to exons 12 and 20 (Table 1). PCR conditions were 94° for 15 seconds, 60° for 15 seconds, and 72° for 60 seconds (Veriti 96-Well Thermal Cycler, Life Tech). PCR was conducted using approximately 30 ng of cDNA template. After PCR, samples were cloned into pcDNA2.1 according to the manufacturer's instructions (TA-Cloning Kit, Invitrogen) and sequenced.
Quantitation of PICALM Expression
Total PICALM expression was quantified by qPCR using primers corresponding to sequences within the constitutively present exons 9 and 10 (Table 1); PICALM isoforms lacking exon 2, exons 2–4, 13 or 18–19 were quantified similarly (Table 1). As no single Ensembl transcript incorporates each of the exons that we identify here, note that our exon designations are derived from ENST00000393346 for exons 1–16. Exons 17–21 correspond to the final five exons within ENST00000532317. PCR was conducted using an initial 2-minute incubation at 95°, followed by cycles of 10 seconds at 95°, 20 seconds at 60°, and 20 seconds at 72°. The 20 μL reactions contained 1 μM of each primer, 1x PerfeCTa SYBR Green Super Mix (Quanta Biosciences), and 30 ng cDNA. Experimental samples were amplified in parallel with serially diluted standards that were generated by PCR of cDNA using the indicated primers followed by purification and quantitation by UV absorbance. Results from samples were compared relative to the standard curve to calculate copy number in each sample. Real time assays were performed at least twice and the average copy number used for data analyses. Since PICALM was expressed in microvessels, neurons and astrocytes, we wished to compare PICALM expression to that of genes specific to these cell types. Hence, we also quantified two microvessel-specific mRNAs, CD31 and von Willebrand Factor (vWF), neuron-specific mRNA SYP and astrocyte-specific GFAP. , . The copy number for each mRNA was then normalized to the geometric mean of reference genes RPL32 and EIF4H, previously quantified in this sample set , . The linear regression statistical model used to analyze the data included the geometric mean of CD31 and vWF (microvessel mRNA), GFAP, SYP, AD status and the number of rs3851179 minor alleles (SPSS version 21).
To begin to evaluate the role of PICALM in AD, we localized PICALM expression in human brain by performing immunohistochemistry. We used an antibody that recognizes an epitope at the extreme PICALM carboxyl terminus that is found in all PICALM isoforms (see below). Robust PICALM expression was observed in microvessels in both non-AD and AD brain sections (Figure 1). Consistent with other reports, we also observed less robust PICALM immunostaining in other cell types that have been identified as neurons and glia , .
Human anterior cingulate was immunostained with anti-CALM antibody, revealing robust microvessel labeling (bar = 100 μm).
To elucidate the impact of the primary AD-associated SNP rs3851179 on PICALM, we chose a three-tiered approach. First, we evaluated whether a non-synonymous PICALM SNP was in linkage disequilibrium with rs3851179. The rs3851179 minor allele frequency in European Americans is 35%. According to the Exome Variant Server (http://evs.gs.washington.edu/EVS/), there are no non-synonymous PICALM SNPs with a minor allele frequency above 0.2% . Hence, the rs3851179 association with AD is not likely to be explained by a non-synonymous PICALM SNP.
The second tier of our approach to elucidate SNP action was to evaluate the extent that PICALM expression correlated with rs3851179 genotype and/or AD status. To this end, total PICALM expression was quantified in 52 brain samples by using qPCR and primers corresponding to sequences within exons 9 and 10 which are constitutively present (see below). PICALM copy number was normalized to the geometric mean of two housekeeping genes, RPL32 and eIF4H , . Inspection of the results supports that total PICALM expression correlated positively with microvessel mRNA expression (Figure 2A). To evaluate the statistical correlation between PICALM expression and relevant indices, we analyzed PICALM expression relative to AD status, rs3851179 genotype, and several cell-type specific mRNAs. Linear regression analysis found an overall significant model (adjusted R2 = 0.46) with a significant correlation between PICALM and rs3851179 as well as cell type markers but not AD (Table 2). Rs3851179, GFAP and microvessel mRNA correlated positively with total PICALM expression, whereas SYP showed negative correlation. The AD-protective, minor rs3851179A allele was associated with increased total PICALM expression.
The indicated mRNAs or isoforms was quantified by qPCR and compared relative to the AD-associated SNP rs3851179 (A–E) or rs592297, an exon 5 SNP (F).
The third tier of our approach to determine possible SNP function was to evaluate the extent that a PICALM splice variant was associated with rs3851179 genotype and/or AD status. We began by identifying PICALM splice variants present in human brain. PCR was performed by using a series of primer pairs that flank PICALM internal exons, e.g., primers corresponding to exons 1 and 5 were used to evaluate whether exons 2, 3 or 4 were variably spliced. This study found that multiple PICALM exons were inefficiently spliced (Figure 3). Sequencing of the exon 1–5 amplicons found that most PICALM isoforms contained exons 2, 3 and 4 while apparently rare isoforms lacked exon 2 or exons 2–4. Amplifying from exon 3 to exon 9, and exon 7 to exon 12 showed that exons 5–11 were consistently present (Figure 3). This supports the use of primers corresponding to exons 9 and 10 for qPCR for total PICALM. Amplification reactions between exons 10–21 overall found multiple PICALM isoforms. These isoforms were not sufficiently resolved by polyacrylamide gel electrophoresis to allow sequencing of individual gel-purified products. To overcome this issue, PICALM from exon 12 to exon 20 was PCR-amplified, and the PCR products cloned and sequenced. To gain an initial evaluation of whether rs3851179 may be associated with PICALM splice variants, we analyzed RNA from three rs3851179 G/G and three rs3851179 A/A homozygous individuals. This effort revealed that exons 13, 14, 18 and 19 were inefficiently spliced. The most common PICALM variant lacked exon 13 and contained each of the other exons from 12 to 20 (Table 3). Other common variants contained each exon from 12–20, or lacked exon 13 and the initial 15 bp of exon 15, or lacked both exon 13 and 18. A comparison of the abundance of each isoform in rs3851179G/G versus rs3851179A/A individuals did not reveal striking differences (Table 3). Overall, we interpret these data as indicating that multiple PICALM exons are variably spliced. Although these isoforms were not associated with rs3851179 in this semi-quantitative assay, their abundance warranted a more quantitative evaluation.
PCR amplification across the indicated exons was performed on cDNA pooled from AD and non-AD brain samples. The products were separated by polyacrylamide gel electrophoresis and visualized by SYBR-Gold fluorescence. Single PCR products from amplifications between exon 3–9 and 7–12 indicate that individual exons between 5–11 are included with high efficiency. The presence of multiple products in other lanes represents inefficiently spliced exons as confirmed in Table 3.
For quantitation, we initially focused on exon 13 because (i) this exon is commonly skipped and (ii) this exon encodes the DPF peptide motif that contributes to PICALM binding to AP2 . We quantified PICALM lacking exon 13 (D13-PICALM) by using qPCR primers corresponding to sequences within exon 11 and the exon 12- exon 14 junction (Table 1). D13-PICALM correlated well with total PICALM expression and constituted about 40% of total transcript (Figure 2B). D13-PICALM expression was analyzed as a function of rs3851179, AD status, and several cell-type specific mRNAs. The expression of D13-PICALM correlated with rs3851179, AD status, as well as the cell-type specific mRNAs (adjusted R2 = 0.54, Table 2). The minor rs38555179A allele and the absence of AD correlated with increased D13-PICALM expression.
To evaluate whether rs3851179 was associated with D13-PICALM independently of the SNP association with total PICALM expression, we analyzed D13-PICALM expression as a function of rs3851179, AD status and total PICALM expression. With this analysis, we found that D13-PICALM was associated with AD status and total PICALM, but not rs3851179. Hence, D13-PICALM expression is associated with rs3851179 only because total PICALM expression is associated with rs3851179.
We next analyzed PICALM splice variants that lacked exons 18 and 19 (D18–19 PICALM), noting that the PICALM carboxyl region that includes amino acids encoded by exon 18 and 19 is critical for PICALM function . This qPCR assay used forward and reverse primers that recognized exon 17 and the exon 17 - exon 20 junction, respectively (Table 1). We found that D18–19 PICALM represented 1–2% of total PICALM expression (Figure 2C) and correlated with neuronal and astrocyte content but not rs3851179 (Table 2).
We next quantified isoforms that lack exon 2 (D2-PICALM). This isoform is expected to not encode a functional protein because the loss of exon 2 introduces a codon frameshift with a premature stop codon in exon 3. Exon 2 encodes a portion of the ANTH domain that binds PIP2 on the plasma membrane during the initial stage of clathrin-coated pit formation , We found that D2-PICALM was typically rare, representing less than 1% of total PICALM expression (Figure 2D). However, two samples showed increased D2-PICALM expression, ranging as high as 3.6%. The reason underlying the higher D2-PICALM in these individuals was unclear; these individuals both had AD, they differ in sex (one female and one male), and had a post-mortem interval similar to the other samples (2.4–4.0 hours). When these outlier samples were excluded from analysis, D2-PICALM was significantly associated with microvessel and neuronal content, as well as AD status but not rs3851179 genotype (adjusted R2 = 0.41, Table 2). D2-PICALM was increased in AD individuals.
We also identified a PICALM isoform lacking exons 2–4 (D2-4 PICALM). The D2-4 PICALM isoform was also present at low levels, with an average of 0.28±0.15% (mean ± S.E.) of total PICALM expression (Figure 2E). Expression of D2-4 PICALM was associated with AD but not with rs3851179 or microvessel content, suggesting that this variant is not expressed in microvessels (Table 2). Interestingly, Schnetz-Boutaud et al have reported that an exon 5 SNP rs592297 is in linkage disequilibrium with rs3851179 (D′ = 1, r2 = 0.34) and proposed that rs592297 modulates the activity of an exon splicing enhancer . Therefore we evaluated whether rs592297 was associated with D2-4 PICALM expression. We found that rs592297 associated with the D2-4 PICALM (Figure 2F, Table 2). Hence, higher D2-4 PICALM expression is associated with the rs592297C minor allele. The percentage of PICALM expressed as D2-4 PICALM was quite low but was increased from 0.23±0.11% in rs592297 major allele homozygous samples to 0.36±0.17% in samples with the rs592297 minor allele (Figure 2F, p = 0.004). Although we and others have not examined the association of this SNP with AD directly, based on the linkage between rs592297 and rs3851179, the minor rs592297C allele is likely to associated with increased AD risk .
The primary findings of this paper are (i) multiple PICALM isoforms are expressed in human brain, (ii) consistent with immunohistochemistry results that PICALM is commonly found in microvessels, expression of total PICALM and the abundant D13-PICALM is positively correlated with the expression of microvessel mRNAs, (iii) total PICALM expression correlates modestly with the AD-associated SNP rs3851179, (iv) D2-4 PICALM was associated with AD status and an exon five SNP, rs592297, which is in linkage disequilibrium with rs3851179 (r2 = 0.34). However, D2-4 PICALM was a rare isoform, suggesting that this association is not responsible for the SNP association with AD, and (v) two additional rare PICALM isoforms, D18-19 PICALM and D2-PICALM were variably associated with AD and cell-specific mRNAs. Overall, we interpret our results as suggesting that multiple PICALM isoforms are expressed in the brain, and that correcting for cell-specific mRNAs allows the discernment that the AD-protective allele of rs3851179 is associated with increased PICALM expression.
Immunostaining showed abundant PICALM expression in microvessels. Consistent with this observation, total PICALM expression correlated with CD31 and vWF expression, genes highly expressed in endothelial cells , . Hence, our statistical model for PICALM expression included the geometric mean of these microvessel mRNAs, well as SYP and GFAP. When we analyzed PICALM expression in this fashion, PICALM expression correlated with rs3851179 genotype. Indeed, inclusion of the expression of these cell-type specific mRNA is the primary difference between our study which detected an association between PICALM expression and rs3851179 and prior studies that did not discern this association , . The modest association that we observed may reflect that rs3851179 is not a functional SNP but rather is in linkage disequilibrium with SNP(s) that directly modulate PICALM expression. Rs3851179 is unlikely to be a directly functional SNP since its well removed from PICALM at 80 kbp upstream and does not alter a transcription factor binding site as predicted by ENCODE . Hence, we speculate that another SNP, more proximal to PICALM, is the functional SNP and is in moderate linkage disequilibrium with rs3851179.
Variation in PICALM expression associated with rs3851179 genotype may have several effects. At the cellular level, PICALM mediates clathrin-coated-pit endocytosis; the amino-terminus of PICALM binds phosphatidyl-inositol 4,5 bisphosphate (PIP2), while the central portion binds adaptor protein-2 (AP-2) and the carboxyl terminus binds clathrin , . Reducing PICALM expression by siRNA leads to altered size and shape of the clathrin-coated pit , . Since the AD-protective allele of rs3851179 correlates with increased PICALM expression, we considered several mechanisms whereby PICALM may modulate AD risk. First, PICALM expression modulates APP metabolism in vitro . Decreased PICALM expression leads to increased APP at the cell surface while increased PICALM expression leads to increased APP internalization. Since APP is metabolized in a non-amyloidogenic pathway at the cell surface but in an amyloidogenic pathway in endosomes, the effects of PICALM on APP localization lead to altered Aβ levels: PICALM knockdown reduces Aβ while PICALM overexpression increases Aβ . This pathway is not consistent with our finding that the protective rs3851179 allele increases PICALM expression. A second pathway whereby altered PICALM may alter AD risk recognizes that altered PICALM expression modulates cell surface proteins in a protein-specific fashion. For example, decreased PICALM leads to increased GluR2 which may promote excitotoxicity ; the protective rs3851179 allele that increases PICALM expression may reduce AD risk by reducing excitotoxicity. Increased PICALM also leads to increased cell surface transferrin and EGFR , , , . Consistent with a critical role for PICALM in iron homeostasis, PICALM-deficient mice suffer from severe anemia and poor erythroid development and, at the cellular level, show reduced transferrin uptake; iron supplementation ameliorates some aspects of PICALM deletion . Recognizing that PICALM was robustly expressed in microvessels and that PICALM expression correlated positively with microvessel mRNAs, we speculate that increased PICALM may be AD-protective by facilitating Aβ clearance across the blood brain barrier . Overall, altered PICALM levels may modulate AD risk by multiple mechanisms and is the subject of ongoing investigation.
Multiple PICALM exons were spliced inefficiently in human brain. Isoforms lacking many of these exons are likely to encode PICALM with altered function. Isoforms lacking exon 13 were especially common. Since a critical AP-2 binding DPF peptide motif is encoded by exon 13, the loss of exon 13 is expected to reduce AP-2 binding . Loss of this DPF motif may be compensated by the DIF and/or FESVF motifs encoded within exons 12 and 14, respectively . Isoforms lacking exons 13 and 14 were also detected that would lack both the DPF and FESVF motifs and would be expected to have particularly low AP-2 binding. Rare isoforms also showed an absence of exons 2 or 2–4. Since exon 2 is 143 bp, isoforms lacking exon 2 undergo a codon frameshift such that D2-PICALM and D2-4-PICALM are predicted to encode only an amino terminal PICALM fragment. Since the exon 5 SNP, rs592297, was associated with exon 2 splicing, we sought to evaluate whether this SNP was associated with AD. Although rs592297 was not available in data from Naj et al, rs1237230 is highly linked with rs592297 (r2 = 0.95 in Europeans, (http://www.broadinstitute.org/mpg/snap/ldsearch.php) and is present in this dataset. Rs1237230 was modestly associated with AD (p = 0.018), relative to rs3851179 (p = 0.00015) . Hence, rs592297 does not appear to be robustly associated with AD risk relative to the primary PICALM SNP. Although rs592297 may be a functional SNP in modulating exon 2–4 splicing, the modest proportion of PICALM present in this isoform may mitigate the SNP effects on overall PICALM function.
In summary, our primary findings are that multiple PICALM isoforms are expressed in human brain, with prominent presence in microvessels, and that overall PICALM expression is correlated with the AD SNP rs3851179. Rare PICALM isoforms are associated with AD status and/or rs592297, a SNP that is in moderate linkage disequilibrium with rs3851179. The rarity of these isoforms and their lack of association with rs3851179 suggest they are unlikely to contribute to AD risk. Since D13-PICALM is the most abundant PICALM isoform, future studies of PICALM function may wish to evaluate this isoform.
Conceived and designed the experiments: IP SE. Performed the experiments: IP. Analyzed the data: IP DF SE. Wrote the paper: IP DF SE.
- 1. Gatz M, Reynolds CA, Fratiglioni L, Johansson B, Mortimer JA, et al. (2006) Role of genes and environments for explaining Alzheimer disease. Arch Gen Psychiatry 63: 168–174.
- 2. Avramopoulos D (2009) Genetics of Alzheimer's disease: recent advances. Genome Med 1: 34.
- 3. Harold D, Abraham R, Hollingworth P, Sims R, Gerrish A, et al. (2009) Genome-wide association study identifies variants at CLU and PICALM associated with Alzheimer's disease. Nat Genet 41: 1088–1093.
- 4. Lambert JC, Zelenika D, Hiltunen M, Chouraki V, Combarros O, et al.. (2011) Evidence of the association of BIN1 and PICALM with the AD risk in contrasting European populations. Neurobiol Aging 32: : 756 e711–755.
- 5. Naj AC, Jun G, Beecham GW, Wang LS, Vardarajan BN, et al. (2011) Common variants at MS4A4/MS4A6E, CD2AP, CD33 and EPHA1 are associated with late-onset Alzheimer's disease. Nat Genet 43: 436–441.
- 6. Harold D, Abraham R, Hollingworth P, Sims R, Gerrish A, et al. (2009) Genome-wide association study identifies variants at CLU and PICALM associated with Alzheimer's disease. Nature genetics 41: 1088–1093.
- 7. Lambert JC, Heath S, Even G, Campion D, Sleegers K, et al. (2009) Genome-wide association study identifies variants at CLU and CR1 associated with Alzheimer's disease. Nat Genet 41: 1094–1099.
- 8. Jun G, Naj AC, Beecham GW, Wang LS, Buros J, et al. (2010) Meta-analysis confirms CR1, CLU, and PICALM as alzheimer disease risk loci and reveals interactions with APOE genotypes. Archives of Neurology 67: 1473–1484.
- 9. Naj AC, Jun G, Beecham GW, Wang LS, Vardarajan BN, et al. (2011) Common variants at MS4A4/MS4A6E, CD2AP, CD33 and EPHA1 are associated with late-onset Alzheimer's disease. Nature genetics 43: 436–441.
- 10. Hollingworth P, Harold D, Sims R, Gerrish A, Lambert JC, et al. (2011) Common variants at ABCA7, MS4A6A/MS4A4E, EPHA1, CD33 and CD2AP are associated with Alzheimer's disease. Nature genetics 43: 429–435.
- 11. Morgan K (2011) The three new pathways leading to Alzheimer's disease. Neuropathology and applied neurobiology 37: 353–357.
- 12. Paul Flicek IA, Ridwan Amode M, Barrell D, Beal K, Brent S, et al.. Ensembl 2013. 41 ed. Nucleic Acids Research 2013 41 Database issue:D48–D55.
- 13. Ling IF, Bhongsatiern J, Simpson JF, Fardo DW, Estus S (2012) Genetics of clusterin isoform expression and Alzheimer's disease risk. PLoS One 7: e33923.
- 14. Malik M, Simpson JF, Parikh I, Wilfred BR, Fardo DW, et al. (2013) CD33 Alzheimer's Risk-Altering Polymorphism, CD33 Expression, and Exon 2 Splicing. J Neurosci 33: 13320–13325.
- 15. Chomczynski P, Sacchi N (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162: 156–159.
- 16. Fleige S, Pfaffl MW (2006) RNA integrity and the effect on the real-time qRT-PCR performance. Mol Aspects Med 27: 126–139.
- 17. Jackson DE (2003) The unfolding tale of PECAM-1. FEBS Lett 540: 7–14.
- 18. Sadler JE (1998) Biochemistry and genetics of von Willebrand factor. Annu Rev Biochem 67: 395–424.
- 19. Ando K, Brion JP, Stygelbout V, Suain V, Authelet M, et al. (2013) Clathrin adaptor CALM/PICALM is associated with neurofibrillary tangles and is cleaved in Alzheimer's brains. Acta Neuropathol 125: 861–878.
- 20. Baig S, Joseph SA, Tayler H, Abraham R, Owen MJ, et al. (2010) Distribution and expression of picalm in Alzheimer disease. J Neuropathol Exp Neurol 69: 1071–1077.
- 21. Exome Variant Server, NHLBI Exome Sequencing Project (ESP), Seattle, WA (URL:http://evs.gs.washington.edu/EVS/) [December 15, 2011].
- 22. Meyerholz A, Hinrichsen L, Groos S, Esk PC, Brandes G, et al. (2005) Effect of clathrin assembly lymphoid myeloid leukemia protein depletion on clathrin coat formation. Traffic 6: 1225–1234.
- 23. Scotland PB, Heath JL, Conway AE, Porter NB, Armstrong MB, et al. (2012) The PICALM protein plays a key role in iron homeostasis and cell proliferation. PLoS One 7: e44252.
- 24. Dreyling MH, Martinez-Climent JA, Zheng M, Mao J, Rowley JD, et al. (1996) The t(10;11)(p13;q14) in the U937 cell line results in the fusion of the AF10 gene and CALM, encoding a new member of the AP-3 clathrin assembly protein family. Proc Natl Acad Sci U S A 93: 4804–4809.
- 25. Schnetz-Boutaud NC, Hoffman J, Coe JE, Murdock DG, Pericak-Vance MA, et al. (2012) Identification and confirmation of an exonic splicing enhancer variation in exon 5 of the Alzheimer disease associated PICALM gene. Ann Hum Genet 76: 448–453.
- 26. Karch CM, Jeng AT, Nowotny P, Cady J, Cruchaga C, et al. (2012) Expression of novel Alzheimer's disease risk genes in control and Alzheimer's disease brains. PLoS One 7: e50976.
- 27. Allen M, Zou F, Chai HS, Younkin CS, Crook J, et al. (2012) Novel late-onset Alzheimer disease loci variants associate with brain gene expression. Neurology 79: 221–228.
- 28. Rosenbloom KR, Sloan CA, Malladi VS, Dreszer TR, Learned K, et al.. (2012) ENCODE Data in the UCSC Genome Browser: year 5 update. Nucleic Acids Res.
- 29. Ramanan V, Agrawal NJ, Liu J, Engles S, Toy R, et al. (2011) Systems biology and physical biology of clathrin-mediated endocytosis. Integr Biol (Camb) 3: 803–815.
- 30. Tebar F, Bohlander SK, Sorkin A (1999) Clathrin assembly lymphoid myeloid leukemia (CALM) protein: localization in endocytic-coated pits, interactions with clathrin, and the impact of overexpression on clathrin-mediated traffic. Mol Biol Cell 10: 2687–2702.
- 31. Xiao Q, Gil SC, Yan P, Wang Y, Han S, et al. (2012) Role of phosphatidylinositol clathrin assembly lymphoid-myeloid leukemia (PICALM) in intracellular amyloid precursor protein (APP) processing and amyloid plaque pathogenesis. J Biol Chem 287: 21279–21289.
- 32. Harel A, Mattson MP, Yao PJ (2011) CALM, a clathrin assembly protein, influences cell surface GluR2 abundance. Neuromolecular Med 13: 88–90.
- 33. Huang F, Khvorova A, Marshall W, Sorkin A (2004) Analysis of clathrin-mediated endocytosis of epidermal growth factor receptor by RNA interference. J Biol Chem 279: 16657–16661.
- 34. Sagare AP, Bell RD, Zlokovic BV (2012) Neurovascular dysfunction and faulty amyloid beta-peptide clearance in Alzheimer disease. Cold Spring Harb Perspect Med 2..