Data Report | Open | Published:

Excess of rare coding variants in PLD3 in late- but not early-onset Alzheimer’s disease

Human Genome Variation volume 2, Article number: 14028 (2015) | Download Citation

Abstract

Recently, mutations in phospholipase D3 (PLD3) were reported in late-onset Alzheimer's disease (AD). By screening the coding regions of PLD3 for variants in a European cohort of 1,089 AD cases, 182 individuals with frontotemporal lobar degeneration and 1,456 controls, we identified 32 variants with a minor allele frequency <5% and observed an excess of rare variants in individuals with late- but not early-onset AD (P=0.034, χ 2-test; odds ratio=1.46).

Genome-wide association studies and linkage analyses have identified at least 25 genes associated with sporadic and familial Alzheimer's disease (AD).1 These genes include classical genetic factors contributing to familial, early-onset forms of AD, such as the β-amyloid precursor protein and the presenilins (PSEN1 and PSEN2),2,​3,​4,​5 as well as several more recently discovered genes that harbor common variants associated with increased risk of late-onset AD (LOAD).6,​7,​8,​9 Together, these genes explain ~61% of the population attributable risk of AD,8,10 and novel genetic factors continue to be revealed. Most recently, rare variants in phospholipase D3 (PLD3) were implicated in late-onset familial and sporadic AD both by family-based whole-exome sequencing and by genotyping and gene-based resequencing.11 Here we assessed the role of PLD3 variants in central European AD and frontotemporal lobar degeneration (FTLD) patients, particularly investigating the role of PLD3 variants in early-onset AD (EOAD).

Using Idaho LightScanner high-resolution melting curve analysis (Biofire Diagnostics, Inc., Salt Lake City, UT, USA), we screened the 13 coding exons and exon–intron boundaries (±10 bp) of PLD3 in 1089 German AD case subjects (75.6±18.6 years, 59.3% female, including 139 cases with an age of onset younger than 65 (61.5±5.5 years, 53.2% female)), 138 German FTLD cases (63.7±8.1 years, 42.0% female) and 1,456 general population controls belonging to the KORA general population cohort12 (58.3±12.0 years, 48.2% female) based in southern Germany. When altered melting patterns suggested variants, Sanger sequencing ensued to identify the underlying genetic alteration. Gene-based burden tests (cohort allelic sum test) and single-variant association tests were performed using χ2 analysis.

All subjects were diagnosed according to the NINCDS-ADRDA criteria or the revised Neary et al.13 criteria as appropriate by a senior psychiatrist specializing in dementias. Ethics review board approval and participants’ written informed consent were obtained before the initiation of the study.

We observed a total of 32 variants with minor allele frequency (MAF) <5% within the coding regions ±10 bp, including: 3 near-splice variants, 1 deletion, 9 synonymous and 16 non-synonymous variants and 3 newly introduced stop codons (Figure 1, Table 1). Eight of the coding variants (27.5%) had been observed previously.11 Notably, we found the variant most significantly associated with the AD phenotype in the previous study,11 PLD3 p.Val232Met (rs145999145), more frequently in controls (n=6; MAF=0.20%; 55.6±12.6 years; 33% older than 65 years) than in AD patients (n=1; MAF=0.05%; 66 years). The single AD individual harboring the p.Val232Met variant presented with AD at age 64 and reported that her mother had also suffered from dementia. Moreover, the synonymous variant (PLD3 p.Ala442Ala; rs4819) that had been significantly associated with LOAD in the original publication11 did not show association with the AD phenotype in our sample in either LOAD or EOAD cases (Table 1). However, we did identify significant associations between two different variants in PLD3 and AD in our sample: PLD3 p.Ile364Ile (rs51787324; Pnominal=3.0×10−8; Pcorrected=9.6×10−7; χ2 test; odds ratio (OR)=63.50 (95% confidence interval (CI): 3.85–1,046.15)) in both the LOAD-only and combined AD sample and PLD3 p.Asn284Ser (rs200274020; Pnominal=5.0×10−6; Pcorrected=1.6×10−5; χ2 test; OR=52.66 (95% CI: 2.52–1,099.83)). To date, these variants have only been found in individuals with either LOAD11 or EOAD (this study).

Figure 1
Figure 1

Rare coding variants in PLD3 in AD, FTLD and control subjects. (a) Schematic representation of rare coding variants identified in PLD3 in AD and FTLD case subjects (above the gene) and KORA general population controls (below the gene). Numbers in parentheses indicate variant counts. If no number is given, variants were identified only once. (b) Aggregate minor allele frequencies (MAFs) for variants of different classes in the different subsamples. P-values represent gene-based burden tests for rare variants in PLD3 (MAF<5%) based on variant counts (Supplementary Table S1) and calculated using χ2 test statistics. AA, amino acids; AD, Alzheimer's disease; EOAD, EOAD, early-onset AD; FTLD, frontotemporal lobar degeneration; NS, not significant.

Table 1: Rare variants (MAF<5%) identified by high-resolution melting curve analysis is 1,089 AD case subjects, 138 FTD case subjects and 1,456 KORA general population controls

A priori power calculation based on published variant frequencies and effect sizes11 using the Purcell power calculator14 suggested that a sample size of 1,089 AD cases and an equal number of general population controls would be sufficient to reach 100% power to detect an excess of rare variants with a MAF <5% under an autosomal dominant model and 36% power for recessive effects. Accordingly, we also performed gene-based burden tests for variants with a KORA-derived MAF<5%. Interestingly, although we observed an excess (Figure 1, Supplementary Tables S1–S4) of rare variants of all classes in LOAD (P=0.034, χ2 test; OR=1.46 (95% CI: 1.02–2.07)) and LOAD+EOAD combined (P=0.031, χ2 test; OR=1.45 (95% CI: 1.03–2.05)), the same was not observed for EOAD alone (P=0.54, χ2 test; OR=1.28 (95% CI: 0.60–2.73)). This excess of rare variants included mainly synonymous variants (LOAD: P=0.007, χ2 test, OR=1.74 (95% CI: 1.15–2.62); LOAD+EOAD combined: P=0.009, χ2 test, OR=1.68 (95% CI: 1.13–2.52)). In the EOAD cases, however, non-synonymous variants were encountered at a MAF=1.1%, which was almost twice as frequent as in either controls (MAF=0.7%) or LOAD (MAF=0.5%; Figure 1, Supplementary Table S2). This result fell short of statistical significance, possibly because of the small number of EOAD cases (n=139) in our sample.

Because an overlap in the genetic architecture of different dementia syndromes as well as neurodegenerative conditions15,​16,​17 has been described, we also assessed the contributions of rare variants in PLD3 to the genetic framework of our FTLD samples. Rare variants in PLD3 were found at an equal or lower frequency in FTLD case subjects relative to controls (Figure 1, Supplementary Tables S1–S3), making a large-scale contribution of rare genetic variants in PLD3 to the genetics of FTLD unlikely.

Both our data as well as previously published data11 point to a significant contribution of a number of different rare synonymous variants in PLD3 to the LOAD phenotype. This is especially interesting in light of the fact that although over 50 human diseases associated with synonymous mutations have been reported to date, few examples exist with regard to neuropsychiatric disorders.18 Functional assays have demonstrated that PLD3 can directly alter β-amyloid precursor protein processing and β-amyloid formation,11 and the apparent lack of contribution of rare PLD3 variants in another neurodegenerative phenotype (that is, FTLD) indirectly supports this notion. It is known that PLD3 p.Ala442Ala is associated with lower expression of total PLD3 mRNA as well as lower levels of exon 11-containing transcripts. Whether a similar mechanism could be implicated in the association between PLD3 p.Ile364Ile and, to a much lesser extent, p.Pro17Pro remains to be elucidated. Human Splicing Finder19 predicts that p.Ile364Ile ablates an enhancer, whereas p.Pro17Pro generates a novel enhancer site. However, experimental evidence is lacking to date. In this context, although we did not observe a similar role of rare synonymous variants in EOAD cases, it seems noteworthy that we identified several non-synonymous PLD3 variants in our small EOAD sample. One could hypothesize that non-synonymous variants with possibly larger effects might also contribute to this comparatively more severe phenotype. However, given the dearth of statistical significance to support this assumption, it currently remains a hypothesis at best.

When considering the individual variants identified in our screening and their contribution to AD genetics, an additional caveat would have to be that association P-values and ORs appear inflated possibly due to the small total number of variants, the small sample sizes (for EOAD) or the possible existence of unaccounted population substructure. Conversely, because we used general population controls, we are unable to exclude the possibility that some of the controls have or will develop AD, thus underestimating the calculated effect sizes. ORs between 50 and 60 typically suggest (near) monogenic disease. However, OR estimates for the total rare genetic variation in PLD3 and its contribution to the AD phenotype (that is, ORs of ~1.5–2.0) seem more realistic because it is likely that these variants contribute to AD risk but are not causal by themselves.

In summary, our data corroborate the role of rare variants in PLD3 and further highlight the significant contribution of rare synonymous variants in this gene to the genetic architecture of LOAD. Interestingly, the association between PLD3 and LOAD was largely driven by variants not significantly associated with the phenotype in the original study,11 whereas the individual variants showing significant associations in the original study could be replicated directly. While rare variants overall or synonymous variants alone do not seem to play a large role in bringing about EOAD, the role for non-synonymous PLD3 variants in EOAD remains open to debate.

References

  1. 1.

    , , , , , et al. Pathways to Alzheimer's disease. J Internal Med 2014; 275: 296–303.

  2. 2.

    , , , , , et al. Cloning of a gene bearing missense mutations in early-onset familial Alzheimer's disease. Nature 1995; 375: 754–760.

  3. 3.

    , , , , , et al. Early-onset Alzheimer's disease caused by mutations at codon 717 of the beta-amyloid precursor protein gene. Nature 1991; 353: 844–846.

  4. 4.

    , , , , , et al. Familial Alzheimer's disease in kindreds with missense mutations in a gene on chromosome 1 related to the Alzheimer's disease type 3 gene. Nature 1995; 376: 775–778.

  5. 5.

    , , , , , et al. Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer's disease. Nature 1991; 349: 704–706.

  6. 6.

    , , , , , et al. Genome-wide association study identifies variants at CLU and PICALM associated with Alzheimer's disease. Nature Genet 2009; 41: 1088–1093.

  7. 7.

    , , , , , et al. Common variants at ABCA7, MS4A6A/MS4A4E, EPHA1, CD33 and CD2AP are associated with Alzheimer's disease. Nature Genet 2011; 43: 429–435.

  8. 8.

    , , , , , et al. Common variants at MS4A4/MS4A6E, CD2AP, CD33 and EPHA1 are associated with late-onset Alzheimer's disease. Nature Genet 2011; 43: 436–441.

  9. 9.

    , , , , , et al. Meta-analysis of 74,046 individuals identifies 11 new susceptibility loci for Alzheimer's disease. Nature Genet 2013; 45: 1452–1458.

  10. 10.

    , . Review: the genetics of Alzheimer's disease; putting flesh on the bones. Neuropathol Appl Neurobiol 2014; 40: 97–105.

  11. 11.

    , , , , , et al. Rare coding variants in the phospholipase D3 gene confer risk for Alzheimer's disease. Nature 2014; 505: 550–554.

  12. 12.

    , , Group MKS. KORA-gen—resource for population genetics, controls and a broad spectrum of disease phenotypes. Gesundheitswesen 2005; 67(Suppl 1): S26–S30.

  13. 13.

    , , , , , et al. Frontotemporal lobar degeneration: a consensus on clinical diagnostic criteria. Neurology 1998; 51: 1546–1554.

  14. 14.

    , , . Genetic Power Calculator: design of linkage and association genetic mapping studies of complex traits. Bioinformatics 2003; 19: 149–150.

  15. 15.

    , , , , , et al. C9orf72 hexanucleotide repeat expansions in clinical Alzheimer disease. JAMA Neurol 2013; 70: 736–741.

  16. 16.

    , , , , , et al. Association of missense and 5'-splice-site mutations in tau with the inherited dementia FTDP-17. Nature 1998; 393: 702–705.

  17. 17.

    , , , , , et al. Significant association between the tau gene A0/A0 genotype and Parkinson's disease. Ann Neurol 2000; 47: 242–245.

  18. 18.

    , , , , . Exposing synonymous mutations. Trends Genet 2014; 30: 308–321.

  19. 19.

    , , , , , . Human Splicing Finder: an online bioinformatics tool to predict splicing signals. Nucleic Acids Res 2009; 37: e67.

Download references

Data Citations

  1. 1.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.526 (2014)

  2. 2.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.530 (2014)

  3. 3.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.532 (2014)

  4. 4.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.534 (2014)

  5. 5.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.536 (2014)

  6. 6.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.538 (2014)

  7. 7.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.540 (2014)

  8. 8.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.542 (2014)

  9. 9.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.544 (2014)

  10. 10.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.546 (2014)

  11. 11.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.548 (2014)

  12. 12.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.550 (2014)

  13. 13.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.552 (2014)

  14. 14.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.554 (2014)

  15. 15.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.556 (2014)

  16. 16.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.558 (2014)

  17. 17.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.560 (2014)

  18. 18.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.562 (2014)

  19. 19.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.564 (2014)

  20. 20.

    Schulte, Eva HGV Database http://dx.doi.org/10.6084/m9.figshare.hgv.566 (2014)

Author information

Affiliations

  1. Neurologische Klinik und Poliklinik, Klinikum rechts der Isar, Technische Universität München, Munich, Germany

    • Eva C Schulte
    •  & Juliane Winkelmann
  2. Institut für Humangenetik, Helmholtz Zentrum München, Neuherberg, Germany

    • Eva C Schulte
    •  & Juliane Winkelmann
  3. Klinik und Poliklinik für Psychiatrie und Psychotherapie, Klinikum rechts der Isar, Technische Universität München, Munich, Germany

    • Alexander Kurz
    • , Panagiotis Alexopoulos
    •  & Janine Diehl-Schmid
  4. AXA Research Fund

    • Harald Hampel
  5. Sorbonne Universités, Université Pierre et Marie Curie, Paris 06, Institut de la Mémoire et de la Maladie d’Alzheimer (IM2A) & Institut du Cerveau et de la Moelle épinière (ICM), Département de Neurologie, Hôpital de la Pitié-Salpétrière, Paris, France; Université Pierre et Marie Curie-Paris 6, AP-HP, Hôpital de la Salpêtrière, Paris, France

    • Harald Hampel
  6. Institute of Epidemiology II, Helmholtz Zentrum München, Neuherberg, Germany

    • Annette Peters
  7. Institute of Genetic Epidemiology, Helmholtz Zentrum München, Neuherberg, Germany

    • Christian Gieger
  8. Klinik für Psychiatrie, Universität Halle, Halle, Germany

    • Dan Rujescu
  9. Institut für Humangenetik, Technische Universität München, Munich, Germany

    • Juliane Winkelmann
  10. Munich Cluster for Systems Neurology (SyNergy), Munich, Germany

    • Juliane Winkelmann
  11. Department of Neurology and Neurosciences, Stanford University, Palo Alto, CA, USA

    • Juliane Winkelmann

Authors

  1. Search for Eva C Schulte in:

  2. Search for Alexander Kurz in:

  3. Search for Panagiotis Alexopoulos in:

  4. Search for Harald Hampel in:

  5. Search for Annette Peters in:

  6. Search for Christian Gieger in:

  7. Search for Dan Rujescu in:

  8. Search for Janine Diehl-Schmid in:

  9. Search for Juliane Winkelmann in:

Competing interests

The authors declare no conflict of interest.

Corresponding author

Correspondence to Juliane Winkelmann.

Supplementary information

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/hgv.2014.28

Further reading