R47H TREM2 variant increases risk of typical early-onset Alzheimer's disease but not of prion or frontotemporal dementia

Alzheimer’s & Dementia 10 (2014) 602-608

R47H TREM2 variant increases risk of typical early-onset Alzheimer’s disease but not of prion or frontotemporal dementia Catherine F. Slatterya, Jonathan A. Beckb, Lorna Harpera, Gary Adamsonb, Zeinab Abdia, James Uphillb, Tracy Campbellb, Ron Druyehb, Colin J. Mahoneya, Jonathan D. Rohrera, Janna Kennyb, Jessica Loweb, Kelvin K. Leunga, Josephine Barnesa, Shona L. Clegga, Melanie Blaira, Jennifer M. Nicholasc, Rita J. Guerreirod, James B. Rowee, Claudia Pontof, Inga Zerrf, Hans Kretzschmarg, Pierluigi Gambettih, Sebastian J. Crutcha, Jason D. Warrena, Martin N. Rossora, Nick C. Foxa, John Collingeb, Jonathan M. Schotta,1, Simon Meadb,*,1 a

Department of Neurodegenerative Disease, Dementia Research Centre, UCL Institute of Neurology, London, UK b Department of Neurodegenerative Disease, MRC Prion Unit, UCL Institute of Neurology, London, UK c Department of Medical Statistics, London School of Hygiene and Tropical Medicine, London, UK d Department of Molecular Neuroscience, UCL Institute of Neurology, London, UK e Department of Clinical Neurosciences, Cambridge University, UK f Department of Neurology, Clinical Dementia Center, Georg-August University G€ottingen, G€ottingen, Germany g Center for Neuropathology and Prion Research, Ludwig-Maximilians-University Munich, Munich, Germany h Department of Pathology, Case Western Reserve University, Cleveland, OH, USA

Abstract

Background: Rare TREM2 variants are significant risk factors for Alzheimer’s disease (AD). Methods: We used next generation sequencing of the whole gene (n 5 700), exon 2 Sanger sequencing (n 5 2634), p.R47H genotyping (n 5 3518), and genome wide association study imputation (n 5 13,048) to determine whether TREM2 variants are risk factors or phenotypic modifiers in patients with AD (n 5 1002), frontotemporal dementia (n 5 358), sporadic (n 5 2500), and variant (n 5 115) Creutzfeldt-Jakob disease (CJD). Results: We confirm only p.R47H as a risk factor for AD (odds ratio or OR 5 2.19; 95% confidence interval or CI 5 1.04-4.51; P 5 .03). p.R47H does not significantly alter risk for frontotemporal dementia (OR 5 0.81), variant or sporadic CJD (OR 5 1.06 95%CI 5 0.66-1.69) in our cohorts. Individuals with p.R47H associated AD (n 5 12) had significantly earlier symptom onset than individuals with no TREM2 variants (n 5 551) (55.2 years vs. 61.7 years, P 5 .02). We note that heterozygous p.R47H AD is memory led and otherwise indistinguishable from “typical” sporadic AD. Conclusion: We find p.R47H is a risk factor for AD, but not frontotemporal dementia or prion disease. Ó 2014 The Alzheimer’s Association. All rights reserved.

Keywords:

TREM2; Alzheimer’s; Frontotemporal dementia; Prion; Phenotype

1. Introduction Polycystic lipomembranous osteodysplasia with sclerosing leukoencephalopathy (PLOSL) is a recessively inherited early onset frontal dementia with bone cysts and 1

Joint last authors. *Corresponding author. Tel.: 144-(0)2076762164; Fax: 144-(0) 2076922406. E-mail address: [email protected]

basal ganglia calcification [1,2] due to variants in either TREM2 (triggering receptor expressed on myeloid cells 2, chr6:41,126,244-41,130,924, hg19) or TYROBP (TYRO protein tyrosine kinase binding protein) [3–5]. This phenotype has since been expanded, with PLOSL associated homozygote TREM2 variants (p.Q33X, p.T66 M, and p.Y38 C) described in three individuals with typical cognitive impairment, white matter change, and frontal atrophy, but without bone cysts [6]. Association study

1552-5260/$ - see front matter Ó 2014 The Alzheimer’s Association. All rights reserved. http://dx.doi.org/10.1016/j.jalz.2014.05.1751

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

has also identified p.R47H (rs75932628), a rare variant in TREM2, that confers risk for late onset Alzheimer’s disease (LOAD) in populations of European descent [7,8]. p.R47H associations have subsequently been reported with frontotemporal dementia (FTD) [9] and Parkinson’s disease [9,10]. Enrichment of rare TREM2 variants has been observed in both AD and FTD patients compared with controls [11,12]. Furthermore, microglial proliferation and CSF1R (colony stimulating factor 1 receptor) activation, implicated in the same inflammatory pathway as the TREM2/TYROBP complex, are thought to be a major component of prion related neurodegeneration [13] and a partial loss of function variant in CSF1R causes hereditary diffuse leukoencephalopathy with spheroids [14]. Given that microglial mediated inflammation is implicated in several dementias, this raises the question as to whether p.R47H TREM2 effects are selective for AD or generic to neurodegeneration, and it remains unclear whether TREM2 variants in AD have a distinctive phenotype, of importance clinically. This study sought to provide detailed clinical phenotyping of TREM2 associated neurodegeneration across multiple dementias; AD, FTD, and Creutzfeldt-Jakob disease (CJD). 2. Methods 2.1. Cohorts DNA samples from individuals with clinical diagnoses of AD (n 5 1002), FTD (n 5 358), variant CJD (vCJD, n 5 115), and sporadic CJD (sCJD, n 5 2500) were identified from the Medical Research Council Prion Unit research sample database (1990 onwards). For further details see supplementary methods. Ethical approval was obtained from the National Hospital for Neurology and Neurosurgery Research Ethics Committee. Informed consent for genetic studies was obtained from all participants. 2.2. Genetics Exon 2 of the TREM2 gene was Sanger sequenced for individuals with AD (n 5 971), FTD (n 5 358), vCJD (n 5 115), sCJD (UK n 5 721, German n 5 49), and UK controls (n 5 534). A total of 2381 UK controls from the 1958 birth cohort (University of Leicester) and 2437 sCJD samples (UK n 5 722, German n 5 824 and US n 5 891) were directly genotyped for p.R47H using a Taqman minor groove binding probe. CJD study cases (UK sCJD n 5 731, German sCJD n 5 818, US sCJD n 5 951) and controls (UK n 5 5020, German n 5 2691 and US n 5 2837) were genotyped on several different Illumina Genome Wide Bead Chip arrays (550 Bead Chip arrays or larger platforms), see supplementary methods.

603

Exon capture and next generation sequencing were undertaken on genomic DNA from patients with AD (exome sequencing n 5 33, dementia gene sequencing n 5 504), FTD (n 5 29), vCJD (n 5 86), and UK sCJD (n 5 48), see supplementary methods. APOE status p.R47H variant associated AD (n 5 13) was ascertained by Minor Groove Binding probe and fluorescent PCR. 2.3. Clinical phenotyping We established age at symptom onset (AAO), where available, and reviewed medical records for p.R47H Alzheimer’s cases (n 5 14) and a group of nil TREM2 variant cases (n 5 33), matched for sex and age at symptom onset (Table S1) to determine sex, annualized rates of decline on the Mini-Mental State Examination (MMSE) based on the longest available interval, presenting clinical features (visual, language, behavioral/dysexecutive, memory), neurological, and psychiatric signs and symptoms. We extracted the following neuropsychometric data, where available: general intellectual function—Wechsler Adult Intelligence Scale-Revised or the Wechsler Abbreviated Scale of Intelligence [15,16]; verbal and visual memory—Recognition Memory Test for words and faces, respectively [17]; and visuospatial and visuoperceptual skills—Visual Object and Spatial Perception battery [18]. Raw scores were converted into percentiles for reporting. Post mortem data reports for individuals with p.R47H variants held in the Queen Square brain bank (University College London) and Institute of Psychiatry brain bank (Kings College London) were also reviewed where available. 2.4. Brain magnetic resonance imaging T1 weighted volumetric brain magnetic resonance imaging (MRI) scans were reviewed retrospectively and volumetric region of interest comparisons performed for p.R47H variant individuals (3T n 5 3 and 1.5 T n 5 1) and 22 AAO and disease duration matched (Table S1) AD individuals with no TREM2 variants (3T n 5 17, 1.5 T n 5 5). All analyses were performed blind to genetic status. 2.5. Statistics The association between TREM2 variants and each neurodegenerative condition was examined using odds ratios and Fisher exact test based on allelic frequencies. The characteristics of the p.R47H variant and nil TREM2 variant AD cases were compared using two tailed t-tests or Wilcoxon-Mann-Whitney test where appropriate. Brain and total hippocampal volumes (expressed as ratio of TIV to correct for head size) between p.R47H positive and TREM2 negative AD subjects were compared using the Wilcoxon-Mann-Whitney test. All statistical tests were two-tailed, with significance set at P , .05 without any correction for multiple hypothesis

604

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

testing. Statistical analyses were performed using Stata (version 12).

3. Results 3.1. p.R47H TREM2 variants in AD, FTD, and CJD Minor allele frequencies for p.R47H in AD (n 5 1002), FTD (n 5 358), vCJD (n 5 115), and sCJD (genotyped n 5 2437, combined genome wide association study analysis n 5 2500) cases are shown in Table 1. Fifteen nonreference alleles causing p.R47H variants (13 heterozygote individuals and 1 homozygote) were found in the AD population, giving an OR of increased risk verses UK controls based on allelic frequencies of 2.19 (95%CI 5 1.04–4.51, P 5 .03), confirming the significant association demonstrated in previous studies [7,8,19–21]. We found no significant association between p.R47H variants and FTD or sCJD. Outside exon 2 p.S162 R and p.L211p may be associated with AD, but this must be considered very provisional given the small sample numbers and the fact that both variants are thought to be nondamaging; further studies will be required (see Table S2 for other variants in exon 2, and Table S3 for variants in exons 1 and 3–5).

3.2. TREM2 variants are associated with earlier disease onset in AD Patients with p.R47H TREM2 variants had significantly younger ages at onset than individuals with no TREM2 variants (AAO 5 55.2 6 8.5 years, n 5 12 vs. AAO 5 61.7 6 13.1 years, n 5 551, P 5 .024). 10/12 (83%) of these p.R47H variant individuals met criteria for young onset AD (YOAD, defined as symptom onset less than 65 years), with 4/12 (33%) of individuals having an age at onset ,50 years, indicating very early onset disease.

3.3. R47H variants in AD: clinical features and neuropsychological profiles Disease duration (age from first reported symptom to death) was known for 6/14 individuals, for whom the mean was 11.3 years (range 7–15 years). 6/12 (50%) of the individuals with detailed clinical information had at least one first or second degree relative with a diagnosis of AD. Case 1 (p.R47H homozygote) had a mother who developed AD in her 70s who died in her 80s, and a brother with disease onset at 52 years, who died in his 60s. Their TREM2 statuses are unknown. Most p.R47H individuals (10/12, 83%) for whom clinical information was available had an amnestic presentation. This was supported by neuropsychology data, available on seven individuals (Table 2). The disease duration at the time of testing varied from 1 to 7 years. All cases had impairment (,5th percentile) on at least one recognition memory test at the time of testing. Three also had some evidence of visuospatial and/or visuoperceptual disturbance, but in no case was this disproportionate to the degree of amnesia. Anecdotally, none of the AD p.R47H patients were reported to have had bone cysts or pathological fractures, but this was not examined or investigated for systematically. There was no significant difference in the annual rate of MMSE decline between the p.R47H variant (n 5 5) and nil TREM2 variants (n 5 21) groups, albeit with small numbers (4.3 points/year 6 3.8 vs. 3.2 points/year 6 2.6, respectively, P 5 .43). Although most individuals in both TREM2 positive and nil TREM2 variants groups had an amnestic presentation, there was a trend for cognitive deficits referable to parietal lobe dysfunction as the presenting feature to be less common than in the p.R47H variant cases; P 5 .09 (Table 3). There were no significant differences in the frequency of neurological signs between the p.R47H variant positive (n 5 12) and nil TREM2 variant (n 5 33) cases (Table 3).

Table 1 p.R47H TREM2 variants in AD, FTD, and CJD cohorts

AD FTD UK vCJD UK sCJD UK sCJD German sCJD US sCJD Combined sCJD Total:

n

Nonreference alleles

MAF cases

MAF controls

OR

95% CI

P

1002 358 115 722

15* 2* 0* 7*

0.0075* 0.0028* 0* 0.0049* 0.0057y 0.0020y 0.0055y 0.0044y

0.0034* 0.0034* 0.0034* 0.0034* 0.0041y 0.0026y 0.0058y 0.0042y

2.19 0.81 NA 1.41 1.39 0.80 0.94 1.06

1.04–4.51 0.09–3.36 NA 0.50–3.49 0.66–2.93 0.24–2.63 0.46–1.89 0.66–1.69

.03z 1.00z 1.00z .47z

824 891 2500 3975

2y 7y 22.1z

.60y

Abbreviations: AD, Alzheimer’s disease; FTD, frontotemporal dementia; CJD, Creutzfeldt-Jakob disease; sCJD; sporadic CJD; MAF; minor allele frequency, OR; odds ratio, CI; confidence interval. *Directly genotyped by Sanger sequencing or MGB R47H probe. Both genotyped and imputed data are shown for UK sCJD and controls. y Allele counts, frequencies and P values calculated using IMPUTEv2 and SNPTEST from genome wide array data (including 731 UK sCJD, 818 German sCJD, and 951 US sCJD compared against 5020 UK, 2691 German, and 2837 US controls). z P values were calculated using two-sided Fisher’s exact test.

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

605

Table 2 Clinical features of AD individuals with p.R47H variant Neuropsychology TREM2 Case M/F R47H 1 2 3 4 5 6 7 8 9 10 11 12

F F M M M M F F M F F F

13 14

F F

Other Age of gene APOE ε onset variants genotype (yrs)

Homozygote Heterozygote Heterozygote Heterozygote Heterozygote Heterozygote Heterozygote Heterozygote Heterozygote Heterozygote Heterozygote Heterozygote

nil nil nil nil nil nil nil nil nil nil nil E318 G PS1 Heterozygote nil Heterozygote nil

Disease duration Family (yrs) history

Years VisuoVisuoLeading since Perceptua spatial symptoms onset VIQ PIQ RMT-W RMT-F (VOSP) (VOSP)

14 – 15 – – – 8 – – 15 7 9

Frontal Memory Memory Memory Memory Memory Memory Memory Language Memory Memory Memory

– 3 3 3 4 1 – – – 7 4 5

– 86 67 80 55 87 – – – 78 – 73

– 77 64 65 63 72 – – – 83 62

,5 ,5 ,5 ,5 ,5 ,5 ,5

,5 ,5 ,5 10–25 ,5 25–75

10–25 25–75 ,5 ,5 25–75 25–75 25–75

25-75 ,5 25-75 25-75 10-25 25-75

Unknown Unknown Unknown Unknown

– –

– –

– –

-

-

-

-

34 33 34 33 33 33 34 33 44 44 33 34

64 44 49 49 54 60 71 46 50 51 59 65

– 33

Unknown – Unknown –

Yes No Yes No Yes Yes Yes No No No Yes No

Abbreviations: AD, Alzheimer’s disease; F, female; M, male; positive family history denotes those with at least one secondary case of AD diagnosed in a first or second degree relative; VIQ, verbal intelligence quotient; PIQ, performance intelligence quotient; RMT-W, Recognition Memory Test for words; RMT-F, Recognition Memory Test for faces; VOSP, Visual object and spatial perception battery. RMT and VOSP scores given as percentiles. “–” indicates no data available.

3.4. Neuroimaging in p.R47H variants MRI in the p.R47H cases (n 5 4) was typical for AD, revealing generalized cerebral and symmetrical hippocampal atrophy (Figure 1). Other than case 10, who had participated in the AN1792 vaccination trial [22], none of the other individuals with neuroimaging had any white matter disease Table 3 Leading symptoms and neurological features in AD patients by p.R47H genotype

Leading symptom Memory Language Frontal Parietal Neurological features Myoclonus Seizures Cerebellar signs Extrapyramidal motor features Pyramidal motor features Dystonia Hallucinations Other psychiatric symptoms Sleep disturbance Dyspraxia

AD p.R47H variant (n 5 12)

AD nil TREM2 variant (n 5 33)

n

%

n

%

P value*

10 1 1 0

83.3 8.3 8.3 0.0

21 1 2 9

63.6 3.0 6.1 27.3

.29 .47 1.00 .09

3 2 2 2 1 1 1 1 1 2

25.0 16.7 16.7 16.7 8.3 8.3 8.3 8.3 8.3 16.7

10 2 0 3 3 0 3 12 2 5

30.3 6.1 0.0 9.1 9.1 0.0 9.1 36.4 6.1 15.2

1.00 .29 .07 .60 1.00 .27 1.00 .13 1.00 1.00

Abbreviation: AD, Alzheimer’s disease. *P values calculated using Fisher’s exact test.

greater than would have been expected for age. No basal ganglia calcification was evident on any of the T1 sequences. Quantitative analysis of cross-sectional brain volumes revealed no difference in brain volume/TIV (median, interquartile range [IQR] 5 0.69, 0.66–0.70 vs. 0.70, 0.66– 0.73, P 5 .40), or total hippocampal volume*1000/TIV (median, IQR 5 3.5, 3.2–3.6 vs. 3.3, 2.9–3.6, P 5 .43), between p.R47H positive (n 5 4) and nil TREM2 variant (n 5 22) AD cases. 3.5. Pathology in p.R47H cases AD pathology was confirmed in all 4 p.R47H AD individuals who had post mortem examination, and two had at least moderate cerebral amyloid angiopathy. Pathological slides for case 10, previously published elsewhere [7], showed mature diffuse amyloid plaques. Case 1 (p.R47H homozygote, behavioral/dysexecutive presentation) showed marked frontal atrophy macroscopically, as is typical in PLOSL cases [23], however, the associated typical white matter lesions were absent. There was extensive formation of senile plaques, neurofibrillary tangles, and neuropil threads throughout the grey matter, but relative preservation of the hippocampus histologically. In the sCJD cohort, all five p.R47H patients having had autopsy demonstrated the classical neuropathology of CJD, however, there was evidence of concurrent Abeta or tau pathology in three individuals. 4. Discussion We report a sequencing and genotyping survey of TREM2 variants in three major neurodegenerative dementias. We

606

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

Fig. 1. Coronal magnetic resonance imaging images in four individuals with Alzheimer’s disease and p.R47H variant. Disease duration denotes time from first symptom to time of scan.

confirm the significant association between p.R47H and AD and describe the related clinical phenotype. We find that the possession of a p.R47H TREM2 variant in AD is associated with a significantly younger age at symptom onset than nil TREM2 variant cases, and p.R47H associated AD is otherwise usually indistinguishable from typical, amnestic AD on clinical, imaging, and neuropsychometric grounds. It remains unclear whether p.R47H is a risk factor for neurodegeneration in general, or is specific to AD. We did not find any association between p.R47H and CJD, and no other groups have yet published data on TREM2 variants in prion disease. We did not find any evidence that p.R47H variants are associated with FTD. Although there were no p.R47H variants identified in either French (n 5 175) [24] or Spanish (n 5 628) [25] FTD populations, a North American cohort (n 5 609) found a significant association (OR 5 5.06, P 5 .001) [9]. Our UK data were not consistent with an association as large as an OR 5 5 but do not rule out a more modest association between p.R47H and frontotemporal dementia (95% CI 5 0.09–3.36). Whether these differences reflect the underlying pathological heterogeneity of patients presenting with behavioral problems or population differences in risk remains to be determined and further studies are warranted, ideally with post mortem confirmation of the underlying pathology, or using other biomarkers (e.g. CSF tau and beta, amyloid imaging) to improve the diagnostic certainty in life. Most individuals with p.R47H had YOAD, with 4/12 of our cases identified having exceptionally young AAO (,50 years) in the absence of other known genetic variants, consistent with results from a French YOAD population [21]. p.R47H has previously been found to correspond with earlier age of onset in both Icelandic (3.18 years, P 5 .20) and Dutch populations (3.65 years, P 5 .13) [8]. In our UK YOAD enriched AD population we find that the mean AAO for p.R47H variant AD patients was significantly earlier than nil TREM2 variant cases (6.5 years, P 5 .024). In most cases, heterozygous p.R47H variants were associated with typical “amnestic” AD presentations. Atypical presentations are more commonly seen in YOAD than LOAD, accounting for between 30% and 40% of cases [26,27]. Our data suggest that, if anything, individuals with TREM2 variants were less likely to have a

nonmemory presentation compared with other YOAD individuals. Our p.R47H cohort had a mean disease duration fairly typical for sporadic AD, rates of MMSE decline were within the published range [28,29] and there were no specific neurological features that could reliably be useful in identifying p.R47H TREM2 variants. Our brain volume analysis revealed no differences between p.R47H carriers and noncarriers, which may reflect the small numbers in this study. Interestingly, data from Alzheimer’s Disease Neuroimaging Initiative has shown that individuals with TREM2 variants lose 1.4% to 3.3% more brain tissue per year than non carriers, and p.R47H is significantly associated with smaller hippocampal volumes [30]. All p.R47H AD individuals with post mortem data available had Alzheimer’s pathology confirmed, although the topography of atrophy seen in case 1 (p.R47H homozygote) was similar to the “generalized cerebral gyral atrophy with frontal accentuation” pathologically observed in a case series of eight patients with PLOSL [23]. Due to the relative rarity of TREM2 variants the numbers of individuals identified were small, and not all had post mortem diagnostic confirmation. The retrospective nature meant clinical information was limited and collected in a nonstandardized manner, hence limiting direct comparisons and inferences with respect to the whole cohort. Although very large multicentre prospective studies will be needed to establish the true spectrum of clinical features, neuroimaging and pathological signatures of TREM2 variants in AD, our findings suggest that p.R47H confers specific risk for typical, amnestic, and often very YOAD. Acknowledgments The authors thank all the patients and their families who have taken part in research leading to this publication. Funding: This work was funded by the Medical Research Council, UK and Alzheimer’s Research UK (ARUKNetwork2012-6-ICE). The authors acknowledge the support of the NIHR Queen Square Dementia Biomedical Research Unit, Leonard Wolfson Experimental Neurology Centre, and the University College London Hospitals NHS Trust

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

Biomedical Research Centre. The authors acknowledge the use of genotype data from the British 1958 Birth Cohort DNA collection, funded by MRC grant G0000934 and Wellcome Trust grant 068545/Z/02. The Dementia Research Centre is an Alzheimer’s Research UK coordinating centre and has also received equipment funded by Alzheimer’s Research UK and Brain Research Trust. NCF is a NIHR senior investigator and MRC Senior Clinical Fellow. JDW is supported by a Wellcome Trust Senior Clinical Fellowship (091673/Z/10/Z). MNR is a NIHR Senior Investigator. JMS is a HEFCE Senior Clinical Lecturer. J Barnes is supported by an Alzheimer’s Research UK Senior Research Fellowship. JBR is supported by the Wellcome Trust (Senior Clinical Fellowship 088324) and NIHR Cambridge Biomedical Research Centre.

[3]

[4]

[5]

[6]

[7]

RESEARCH IN CONTEXT [8]

1. Systematic review: We searched PubMed up to December 2013 with the terms “Alzheimer’s disease”, “young onset Alzheimer’s disease”, “frontotemporal dementia”, “prion”, “Creutzfeldt-Jakob disease”, and “TREM2”. p.R47H is a risk factor for Alzheimer’s disease (AD) and can be associated with young onset. Associations have also been reported with frontotemporal dementia. The clinical phenotype of p.R47H associated AD and p.R47H association with prion disease is unknown. 2. Interpretation: We show that individuals with p.R47H associated AD are significantly younger at symptom onset than individuals with no TREM2 variants. Heterozygous p.R47H AD is memory led with disease duration, neuroimaging and pathological features otherwise indistinguishable from “typical” sporadic AD. p.R47H does not significantly alter risk for frontotemporal dementia or Creutzfeldt-Jakob disease in our cohorts. 3. Future directions: The mechanism by which p.R47H leads to earlier age at Alzheimer’s symptom onset remains to be elucidated, but may lead to the identification of novel therapeutic targets.

[9]

[10] [11]

[12]

[13]

[14]

[15] [16] [17] [18] [19]

[20]

References [1] Hakola HP. Neuropsychiatric and genetic aspects of a new hereditary disease characterized by progressive dementia and lipomembranous polycystic osteodysplasia. Acta Psychiatr Scand Suppl 1972; 232:1–173. [2] Nasu T, Tsukahara Y, Terayama K. A lipid metabolic disease—“membranous lipodystrophy”—an autopsy case demonstrating numerous

[21]

[22]

607

peculiar membrane-structures composed of compound lipid in bone and bone marrow and various adipose tissues. Acta Pathol Jpn 1973; 23:539–58. Paloneva J, Kestila M, Wu J, Salminen A, Bohling T, Ruotsalainen V, et al. Loss-of-function mutations in TYROBP (DAP12) result in a presenile dementia with bone cysts. Nat Genet 2000;25:357–61. Paloneva J, Manninen T, Christman G, Hovanes K, Mandelin J, Adolfsson R, et al. Mutations in two genes encoding different subunits of a receptor signaling complex result in an identical disease phenotype. The American Journal of Human Genetics 2002; 71:656–62. Klunemann HH, Ridha BH, Magy L, Wherrett JR, Hemelsoet DM, Keen RW, et al. The genetic causes of basal ganglia calcification, dementia, and bone cysts: DAP12 and TREM2. Neurology 2005; 64:1502–7. Guerreiro RJ, Lohmann E, Bras JM, Gibbs JR, Rohrer JD, Gurunlian N, et al. Using exome sequencing to reveal mutations in TREM2 presenting as a frontotemporal dementia-like syndrome without bone involvement. JAMA Neurol 2013;70:78–84. Guerreiro R, Wojtas A, Bras J, Carrasquillo M, Rogaeva E, Majounie E, et al. TREM2 variants in Alzheimer’s disease. N Engl J Med 2013;368:117–27. Jonsson T, Stefansson H, Steinberg S, Jonsdottir I, Jonsson PV, Snaedal J, et al. Variant of TREM2 associated with the risk of Alzheimer’s disease. N Engl J Med 2013;368:107–16. Rayaprolu S, Mullen B, Baker M, Lynch T, Finger E, Seeley WW, et al. TREM2 in neurodegeneration: evidence for association of the p.R47H variant with frontotemporal dementia and Parkinson’s disease. Mol Neurodegener 2013;8:19. Benitez BA, Cruchaga C. TREM2 and neurodegenerative disease. N Engl J Med 2013;369:1567–8. Borroni B, Ferrari F, Galimberti D, Nacmias B, Barone C, Bagnoli S, et al. Heterozygous TREM2 mutations in frontotemporal dementia. Neurobiol Aging 2014;35:934.e7–93410. Cuyvers E, Bettens K, Philtjens S, Van Langenhove T, Gijselinck I, van der Zee J, et al. Investigating the role of rare heterozygous TREM2 variants in Alzheimer’s disease and frontotemporal dementia. Neurobiol Aging 2014;35:726.e11–9. Gomez-Nicola D, Fransen NL, Suzzi S, Perry VH. Regulation of microglial proliferation during chronic neurodegeneration. J Neurosci 2013;33:2481–93. Rademakers R, Baker M, Nicholson AM, Rutherford NJ, Finch N, Soto-Ortolaza A, et al. Mutations in the colony stimulating factor 1 receptor (CSF1R) gene cause hereditary diffuse leukoencephalopathy with spheroids. Nat Genet 2011;44:200–5. Wechsler D. Manual for the Wechsler Adult Intelligence Scale. New York: Psychological Corporation; 1981. Wechsler D. Abbreviated Scale of Intelligence. San Antonio, Texas: Psychological Corporation; 1999. Warrington EK. Manual for the recognition memory test for words and faces. Windsor, UK: NFER-Nelson; 1984. Warrington E, James M. The visual object and space perception battery. Bury St Edmunds, UK: Thames Valley Test Company; 1991. Benitez BA, Cooper B, Pastor P, Jin SC, Lorenzo E, Cervantes S, et al. TREM2 is associated with the risk of Alzheimer’s disease in Spanish population. Neurobiol Aging 2013;34:5. Giraldo M, Lopera F, Siniard AL, Corneveaux JJ, Schrauwen I, Carvajal J, et al. Variants in triggering receptor expressed on myeloid cells 2 are associated with both behavioral variant frontotemporal lobar degeneration and Alzheimer’s disease. Neurobiol Aging 2013; 9:016. Pottier C, Wallon D, Rousseau S, Rovelet-Lecrux A, Richard AC, Rollin-Sillaire A, et al. TREM2 R47H variant as a risk factor for early-onset Alzheimer’s Disease. J Alzheimers Dis 2013;4:4. Gilman S, Koller M, Black RS, Jenkins L, Griffith SG, Fox NC, et al. Clinical effects of Abeta immunization (AN1792) in patients with AD in an interrupted trial. Neurology 2005;64:1553–62.

608

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

[23] Paloneva J, Autti T, Raininko R, Partanen J, Salonen O, Puranen M, et al. CNS manifestations of Nasu-Hakola disease: a frontal dementia with bone cysts. Neurology 2001;56:1552–8. [24] Lattante S, Le Ber I, Camuzat A, Dayan S, Godard C, Van Bortel I, et al. TREM2 mutations are rare in a French cohort of patients with frontotemporal dementia. Neurobiol Aging 2013;4:030. [25] Ruiz A, Dols-Icardo O, Bullido MJ, Pastor P, Rodrıguez-Rodrıguez E, L opez de Munain A, et al. Assessing the role of the TREM2 p.R47H variant as a risk factor for Alzheimer’s disease and frontotemporal dementia. Neurobiol Aging 2014;35:444.e1–4. [26] Koedam EL, Lauffer V, van der Vlies AE, van der Flier WM, Scheltens P, Pijnenburg YA. Early-versus late-onset Alzheimer’s disease: more than age alone. J Alzheimers Dis 2010;19:1401–8.

[27] Balasa M, Gelpi E, Antonell A, Rey MJ, Sanchez-Valle R, Molinuevo JL, et al. Clinical features and APOE genotype of pathologically proven early-onset Alzheimer disease. Neurology 2011;76:1720–5. [28] Tan KS, Libon DJ, Rascovsky K, Grossman M, Xie SX. Differential longitudinal decline on the Mini-Mental State Examination in frontotemporal lobar degeneration and Alzheimer disease. Alzheimer Dis Assoc Disord 2013;14:14. [29] van der Vlies AE, Koedam EL, Pijnenburg YA, Twisk JW, Scheltens P, van der Flier WM. Most rapid cognitive decline in APOE epsilon4 negative Alzheimer’s disease with early onset. Psychol Med 2009; 39:1907–11. [30] Rajagopalan P, Hibar DP, Thompson PM. TREM2 and neurodegenerative disease. N Engl J Med 2013;369:1565–7.

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

Supplementary methods Cohorts Alzheimer’s disease (AD) (n 5 1002) and frontotemporal dementia (FTD) (n 5 358) cases were recruited via tertiary specialist clinics at the National Hospital for Neurology and Neurosurgery, London. Clinical diagnoses of AD and FTD were supported by participation in longitudinal research studies at University College London and the University of Cambridge; however, as these sample collections were acquired over two decades the comprehensive use of research diagnostic criteria cannot be confirmed and some samples were assigned to these diagnostic categories based on clinical diagnoses only. Probable or definite variant Creutzfeldt-Jakob disease (vCJD) patients (n 5 115), diagnosed according to established World Health Organization criteria, were recruited by the National Prion Clinic, London, or the National CJD Research and Surveillance Unit, Edinburgh, from 1995 to 2010. Probable or definite sporadic Creutzfeldt-Jakob disease (sCJD) patients (according to World Health Organization criteria) were recruited by the National Prion Clinic, London, the National CJD Research and Surveillance Unit, Edinburgh (n 5 722 UK samples), the National Reference Centre for Transmissible Spongiform Encephalopathy, Georg-August University Goettingen, the Institut f€ ur Neuropathologie, Ludwig-Maximilians Universit€at, Munich, Germany (n 5 824); and the National Prion Disease Pathology Surveillance Centre, Case Western University, Cleveland, Ohio, USA (n 5 891). This sample collection was used for Sanger sequencing, a similar sample collection was used in the imputation analysis using genome wide association study data (see section below for n). All patients of known non-white ethnicity were excluded. Individuals known to have pathogenic disease causing variants in the amyloid precursor protein, APP; Presenillin 1, PSEN1; Presenillin 2, PSEN2; prion protein, PRNP, chromosome 9 open reading frame 72, C9orf72; microtubule associated protein tau, MAPT; or progranulin, PGRN genes were excluded although the entire sample collection was only partially screened for mutations in these genes. Control samples were obtained from the Human Random Control panel (European Collection of Cell Cultures n 5 534), the UK 1958 Birth cohort (n 5 2381) (University of Leicester), and genome-wide control data were used that was generated by the Wellcome Trust Case Control Consortium 2 (UK, n 5 5020), KORA-gen (German, n 5 2691), Framingham and Geisinger control cohorts (US, n 5 2837). Sanger sequencing Polymerase chain reaction amplification used 20 ng of genomic DNA, 2! PCR MegaMix-RoyalÒ (Microzone) and forward (50 -gaccatacgatgggttttcc-30 ) and reverse (50 ccgctcccaacttgtataagaa-30 ) primers. PCR products were cleaned, and 5–20 ng was sequenced in a reaction containing 5! Reaction Buffer, BigDyeÒ (Applied Biosystems) and

608.e1

the forward primer. Sequencing reaction products were cleaned, and sequencing was performed on a 3730 DNA Analyzer (Applied Biosystems). Sequence traces were analyzed using Seqscape software (version 2.7). Imputation of TREM2 R47H using genome-wide array data Case and control samples were genotyped genome-wide using several different Illumina Bead Arrays (at least 550K content). Samples with call rates ,98% were excluded; SNPs with call rates ,99%, minor allele frequency ,1% and Hardy–Weinberg equilibrium test P , .001 in controls were excluded before phasing and imputation. Standard methods were implemented with PLINK to remove related samples, cryptic duplicates and ethnic outliers. In the combined analysis after QC 731 UK sCJD, 818 German sCJD, and 951 US sCJD cases remained which were compared against 5020 UK, 2691 German and 2837 US controls with the same QC filters. Shapeit_v2 was used to align samples to the positive strand using reference files from the March 2012 release of the 1000 Genomes Phase 1 Integrated Variant Set before imputation of chromosome 6 including rs75932628 using IMPUTE2 (v.2.3.0). The info score quality estimates for rs75932628 were 0.635 for samples imputed on the 660 platform, 0.68 on the OmniExpress platform and 0.758 for the 12.5 M platform. These metrics are well above standard thresholds used to exclude poorly imputed SNPs. Association testing was done using the SNPTESTv2.5 Beta frequentist additive score method, which takes into account imputation uncertainty. This test also incorporated the first four covariates from a multidimensional scaling plot (generated in PLINK) to correct for population stratification, resulting in an acceptable genomic inflation factor of 1.05. The info score quality estimate for rs75932628 (R47H) derived from SNPtestv2 was 0.82. Exome Sequencing The solution-phase Agilent SureSelect 38 Mb exome capture (SureSelect Human All Exon Kit, Agilent, CA, USA) and the Illumina HiSeq2000 sequencer (Illumina, CA, USA) were used and methods carried out according to standard manufacturer protocols. Paired-end 100 bp reads were aligned to the hg19 human reference sequence using the Burrows-Wheeler Aligner [1]; average sequencing depth on target was 27, and 90% of the targeted region was covered with a minimum read depth of 8 (minimum depth used for data analysis). Data handling was performed using Picard (www.picard.sourceforge.net/) and a combination of SAMTools [2] and ANNOVAR [3] were used for calling and annotation of sequence variants. Dementia Gene Sequencing An Ion Torrent Personnel Genome Machine sequencer (Life Technologies Corporation, CA, USA) was used with

608.e2

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

Ampliseq PCR amplicon-based library preparation (Life Technologies) to sequence approximately 17.7 Kb across 16 dementia related genes [4] including TREM2 in patients with Alzheimer’s disease (n 5 504). All five exons of TREM2 were targeted for capture using eight amplicons which equaled a total of 878 base pairs of target of which 693 bp were open reading frame. Average read depth for all cases screened was 581!. Variant Call Format and associated Mutation Reports files were generated using NextGENe software version 2.3.3 (SoftGenetics, PA, USA) following alignment of 200 bp reads to the hg19 reference using an adapted version of Burrows-Wheeler Aligner. Filters were relaxed to allow mutation to reference allele frequency of 0.1 through with an F/R balance ratio of 0.01, and a minimum read depth of 10!. Results were then filtered further to exclude, all variants with mutation allele frequency between 0.1 and 0.3 after visual inspection, overall mutation (OM) score less than 12 and synonymous variants. OM scores were generated by the NextGENe software algorithm where quality scores are logarithmically linked to error probabilities using factors including coverage, read balance, allele balance, wrong allele score, homopolymer score, and mismatch score. Brain magnetic resonance imaging T1-weighted volumetric brain magnetic resonance imagings were acquired using a Magnetization Prepared Rapid Gradient Echo sequence: 1.5 T GE Signa scanner (General Electric Milwaukee, WI, USA) (256 ! 256 matrix; 1.5 mm slice thickness) and 3.0 T Siemens Trio scanner (Siemens, Germany) (256 ! 256 matrix; 1.1 mm slice thickness). MR images were corrected for intensity inhomogeneity using N3 [5,6]. Whole brain volumes were generated using an automated segmentation technique (brain-multi-atlas

propagation and segmentation [MAPS]) [7]. Hippocampi were delineated using the automated similarity and truth estimation for propagated segmentations (STEPS) algorithm [8]. Both the brain regions and hippocampal regions were checked by experienced raters. Total intracranial volumes were calculated by summing grey matter, white matter and cerebrospinal fluid volumes acquired using the new segmentation toolbox within Statistical Parametric Mapping– version 8 [9].

References [1] Li H, Durbin R. Fast and accurate short read alignment with BurrowsWheeler transform. Bioinformatics 2009;25:1754–60. [2] Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 2009; 25:2078–9. [3] Wang K, Li M, Hakonarson H. ANNOVAR: functional annotation of genetic variants from high-throughput sequencing data. Nucleic Acids Res 2010;38:3. [4] Beck J, Pittman A, Adamson G, Campbell T, Kenny J, Houlden H, et al. Validation of next generation sequencing technologies in genetic diagnosis of dementia. Neurobiol Aging 2014;35:261–5. [5] Sled JG, Zijdenbos AP, Evans AC. A nonparametric method for automatic correction of intensity nonuniformity in MRI data. IEEE Trans Med Imaging 1998;17:87–97. [6] Boyes RG, Gunter JL, Frost C, Janke AL, Yeatman T, Hill DL, et al. Intensity non-uniformity correction using N3 on 3-T scanners with multichannel phased array coils. Neuroimage 2008;39:1752–62. [7] Leung KK, Barnes J, Modat M, Ridgway GR, Bartlett JW, Fox NC, et al. Brain MAPS: an automated, accurate and robust brain extraction technique using a template library. Neuroimage 2011;55:1091–108. [8] Jorge Cardoso M, Leung K, Modat M, Keihaninejad S, Cash D, Barnes J, et al. STEPS: Similarity and Truth Estimation for Propagated Segmentations and its application to hippocampal segmentation and brain parcelation. Med Image Anal 2013;17:671–84. [9] Interactive algorithms for the Statistical Parametric Mapping. Available at: http://www.fil.ion.ucl.ac.uk/spm. Accessed July 31, 2014.

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

608.e3

Table S1 Demographics of age and sex matched p.R47H and nil TREM2 variant AD cases AD p.R47H variant Leading symptom and neurological features n 12 M:F 5:7 AAO (mean yrs 6 SD yrs) 55.2 6 8.5 Rate MMSE decline n 5 M:F 2:3 AAO (mean yrs 6 SD yrs) 53.6 6 6.5 Volumetric imaging analysis, n/a n 4 M:F 3:1 AAO (mean yrs 6 SD yrs) 53.5 6 4.8 Disease duration at time of scan 5.0 6 3.4 (mean yrs 6 SD yrs)

AD nil TREM2 variants

P value

33 9:24 56.1 6 7.2

n/a .47* .72y

21 9:12 56.6 6 7.4

n/a 1.0* .42y

22 7:15 56.0 6 7.0 3.5 6 2.3

.26* .45z .33z

Abbreviations: AD, Alzheimer’s disease; MMSE, Mini-Mental State Examination; M, male; F, female; AAO, age at onset; yrs, years; SD, standard deviation. *P values calculated using Fisher’s exact test. y P value calculated using t test for equal variance. z P values calculated using the Wilcoxon-Mann-Whitney test.

Table S2 TREM2 coding variants identified using Sanger sequencing of exon 2 Sanger sequencing of exon 2

Variant p.Q33 R p.D39 E p.R47H p.C51Y p.R62H p.R62 C p.D87 N p.T96 K p.R98 W p.R98Q p.A105 V Total

SNP number n/a rs200392967 rs75932628 n/a rs143332484 n/a rs142232675 rs2234253 rs147564421 n/a rs145080901

Position*

Reference allele

Minor allele

PolyPhen-2y? damaging

41129296 41129275 41129252 41129264 41129207 41129208 41129133 41129105 41129100 41129099 41129078

A G C G C G C G G C G

G C T A T A T T A T A

Benign (0) Possibly (0.89) Probably (1.00) Probably (1.00) Benign (0.02) Possibly (0.99) Probably (1.00) Probably (1.00) Probably (1.00) Benign (0.02) Probably (1.00)

AD (n 5 971)

FTD (n 5 358)

sCJD (n 5 778)

vCJD (n 5 115)

Controls (n 5 534)

No. nonref alleles

No. nonref alleles

No. nonref alleles

No. non ref alleles

No. non ref alleles

0 1 15 1 22 1 0 8 1 0 0 49

MAF 0.0000 0.0005 0.0077 0.0005 0.0113 0.0005 0.0000 0.0041 0.0005 0.0000 0.0000

1 0 2 0 8 0 0 1 0 0 0 12

MAF 0.0014 0.0000 0.0028 0.0000 0.0112 0.0000 0.0000 0.0014 0.0000 0.0000 0.0000

0 0 7 0 19 1 1 1 0 1 0 30

MAF 0.0000 0.0000 0.0045 0.0000 0.0122 0.0006 0.0006 0.0006 0.0000 0.0006 0.0000

0 0 0 0 2 0 1 0 0 0 0 3

MAF 0.0000 0.0000 0.0000 0.0000 0.0096 0.0000 0.0048 0.0000 0.0000 0.0000 0.0000

0 1 5 0 9 0 1 1 1 0 1 19

MAF 0.0000 0.0009 0.0047 0.0000 0.0084 0.0000 0.0009 0.0009 0.0009 0.0000 0.0009

Abbreviations: AD, Alzheimer’s disease; FTD, frontotemporal dementia; sCJD, sporadic Creutzfeldt-Jakob disease; MAF, minor allele frequency; SNP, single nucleotide polymorphism. The 15 nonreference alleles at 41129252 (R47H variant) in the AD cohort represent 13 heterozygote individuals and 1 homozygote individual. One further individual had two different variants (R62H and T96 K). *Position denotes the location of the variant in base pairs in chromosome 6 (hg19). y PolyPhen-2 refers to the pathogenicity prediction on polymorphism phenotyping, version 2. The numbers in brackets represent prediction scores ranging from 0 (benign) to 1 (damaging).

608.e4

C.F. Slattery et al. / Alzheimer’s & Dementia 10 (2014) 602-608

Table S3 TREM2 coding variants in exons 1 and 3–5, identified using next generation sequencing (dementia chip n 5 504, exome sequencing n 5 196) UK sCJD FTD (n 5 86) (n 5 48)

AD (n 5 537) Reference Minor allele allele Poly Phen-2y

Variant

SNP position

Position*

p.E151 K

rs79011726

41127561 C

T

41126801 G 41126655 A 41126613 G 41126524 A

p.S162 R

TMP_ESP_6 _41126801 p.L211P rs2234256 p.P225 L n/a c.677-6T . C n/a Total

Nonref alleles MAF

OR (95%CI) z

vCJD (n 5 86)

non non non P ref ref ref valuez alleles MAF alleles MAF alleles MAF

1

0.0009 2.67 (0.05–33.29)

0.38

0

0.000 1

0.042 0

0.000

C

Possibly damaging (0.95) Benign (0.42)

2

0.0019 16.04 (0.83–946.48) 0.03

0

0.000 0

0.000 0

0.000

T C T

Benign (0.00) Benign (0.01) n/a

7 1 1 12

0.0065 4.70 (1.56–13.00) 0.0009 – 0.0009 8.01 (0.10–629.11)

0 0 0 0

0.000 0 0.000 0 0.000 0 1

0.000 0 0.000 0 0.000 0 0

0.000 0.000 0.000

0.003 – 0.21

Abbreviations: OR, odds ratio; CI, confidence interval. *Position denotes the location of the variant in base pairs in chromosome 6 (hg19). y PolyPhen-2 refers to the pathogenicity prediction on polymorphism phenotyping, version 2. The seven nonreference alleles at 41126655 (p.L211P variant) in the AD cohort represent five heterozygote individuals and one homozygote individual. Three of these individuals also had p.T96 K variants. z OR and P values calculated using control data from http://evs.gs.washington.edu/EVS/ (European American controls n 5 4300).