Article Text

Download PDFPDF
  1. CF Slattery,
  2. J Beck,
  3. L Harper,
  4. G Adamson,
  5. Z Abdi,
  6. J Uphill,
  7. T Campbell,
  8. R Druyeh,
  9. CJ Mahoney,
  10. JD Rohrer,
  11. J Kenny,
  12. J Lowe,
  13. KK Leung,
  14. J Barnes,
  15. SL Clegg,
  16. M Blair,
  17. JM Nicholas,
  18. RJ Guerreiro,
  19. JB Rowe,
  20. C Ponto,
  21. I Zerr,
  22. H Kretzschmar,
  23. P Gambetti,
  24. SJ Crutch,
  25. JD Warren,
  26. MN Rossor,
  27. NC Fox,
  28. J Collinge,
  29. JM Schott,
  30. S Mead


Objective Variants in TREM2, a gene expressed on microglia and involved in CNS innate immunity, have recently been reported as rare but strong risk factors for Alzheimer's disease. Microglial mediated inflammation is implicated in several dementias. It remains unresolved whether TREM2 variant effects are selective for Alzheimer's disease or generic to neurodegeneration, and whether there is a distinct clinical phenotype associated. This study sought to provide the first systematic characterisation and detailed clinical phenotyping of TREM2 associated neurodegeneration across multiple dementias.

Method We used next generation sequencing of the entire coding sequence (n=700), targeted Sanger sequencing of exon 2 (n=2634), p.R47H genotyping (n=3518) and imputation from genome wide association study data (n=3383) to determine TREM2 variants in large independent cohorts of patients with clinical diagnoses of predominantly early-onset Alzheimer's disease (n=1002), frontotemporal dementia (n=358), sporadic Creutzfeldt-Jakob disease (n=2437) and variant Creutzfeldt-Jakob disease (n=115). We describe the demographics, disease duration, neuroimaging, neuropsychological profiles and pathological findings in Alzheimer's disease cases with p.R47H and perform a case-control study comparing the detailed clinical presentations, rates of mini mental state examination decline and volumetric neuroimaging characteristics.

Results We confirm previous reports that p.R47H is a risk factor for Alzheimer's disease vs. unselected controls (OR=2.19; 95% CI=1.04–4.51; P=0.03) and show that it does not significantly alter the risk for frontotemporal dementia (OR=0.81, 95% CI=0.09–3.36, P=1.00), sporadic or variant Creutzfeldt-Jakob disease (ORs=0.97–1.54 in three populations). We found a total of 49 non-synonymous alleles in 47 Alzheimer's patients, including one homozygous p.R47H case. Where age of symptom onset was available, Alzheimer's disease individuals with p.R47H TREM2 variants (n=12) were significantly younger at symptom onset than individuals with no TREM2 variants (n=551) (p.R47H: 55.2 years vs nil TREM2 variants: 61.7 years, P=0.024). 10/12 (83%) of the p.R47H Alzheimer's disease cases had age-at-onset below 65 years, and 4/12 below 50 years. Heterozygous p.R47H Alzheimer's disease typically had a memory led presentation with disease duration, rates of mini mental state examination decline, and neuroimaging and pathological features indistinguishable from ‘typical’ sporadic Alzheimer's. The one notable exception was the homozygous p.R47H individual who had a behavioural presentation. Despite having typical Alzheimer's disease pathology confirmed on histology, they had an unusual degree of frontal atrophy, and relative sparing of the hippocampi.

Conclusion Heterozygous p.R47H variants confer specific risk for Alzheimer's disease, and result in earlier symptom onset in a disease that appears otherwise typical for sporadic Alzheimer's disease.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.