Neurobiology of Aging 28 (2007) 1637–1643 Genetic risk profiles for Alzheimer’s disease: Integration of APOE genotype and variants that up-regulate inflammation Federico Licastro a , Elisa Porcellini a , Calogero Caruso b , Domenico Lio b , Elizabeth H. Corder c,∗ a Department of Experimental Pathology, School of Medicine, University of Bologna, Via S. Giacomo 14, 40126 Bologna, Italy b Department of Biopathology and Biomedical Methodology, University of Palermo, Corso Tukory 211, 90134 Palermo, Italy c Center for Demographic Studies, 2117 Campus Drive, Duke University, Durham, NC 27708-0408, USA Received 11 May 2006; received in revised form 29 June 2006; accepted 10 July 2006 Available online 22 August 2006 Abstract Background: A number of studies associate Alzheimer’s disease with APOE polymorphism and alleles which favor the increased expression of immunological mediators such as cytokines or acute phase proteins. We integrated this information to better define risk and determine the relative importance of APOE and immunological mediators. Methods: We investigated functional gene variants for APOE, IL-10 (3 loci), ACT (2 loci), HMGCR, IL-1α, IL-1β, TNF-α, IFN-γ , and IL-6 found for 260 AD patients and 190 controls enrolled in Northern Italy. A fuzzy latent classification approach, namely grade-of-membership analysis (GoM), was taken to identify extreme pure type risk sets, or profiles. This approach automatically relates individuals to each profile via graded membership scores. Findings: Four extreme pure type risk sets were identified. Set I defined low intrinsic risk and had a low probability of carrying pro- inflammatory alleles or APOE ε4. Three sufficient risk sets were identified: early onset AD (set II) was characterized by a high density of pro-inflammatory alleles, a rapid cognitive decline and independent of APOE ε4. Late onset AD had a lower density (ages 65–74, set III), or a subset homozygous (ages 75+, set IV), for these alleles and a high probability of one or two APOE ε4 alleles. A total of 97% of the subjects who were cases strongly resembled, i.e. had at least 50% membership in, the sufficient risk sets, as did 25% of middle aged control subjects. IL-10, HMGCR, ACT, and IL-1β gene variants were each more informative in identifying the risk sets than was APOE. Interpretation: AD likely has many determinants including APOE polymorphism and gene variants that modulate innate immunity. Iden- tification of these factors, risk prediction for individuals, and successful prevention and treatment trials require integration of relevant information. © 2006 Elsevier Inc. All rights reserved. Keywords: Alzheimer’s disease; GoM analysis; Gene polymorphism; Inflammation; Cognitive deterioration 1. Introduction Inflammation is accepted to be a feature of Alzheimer’s disease (AD) pathology [10,13,15–19,21–23,28,30,33,35, 39]. Inflammatory processes can be observed in the AD brain; circulating acute phase reactant levels in middle age predict AD risk in old age; certain functional promoter polymor- ∗ Corresponding author. Tel.: +1 919 641 3994; fax: +1 270 675 8521. E-mail address: elizabethcorder@hotmail.com (E.H. Corder). phisms in cognate genes that modulate inflammation are often found at elevated frequency among AD cases. However, alle- les that up-regulate inflammation are not, taken individually, found to be elevated in every sample of cases, nor is any one polymorphism sufficient to predict risk for individuals, nor has the relative importance of APOE [3] polymorphism and immunologic mediators been established. No attempt has been made to integrate this information to better describe risk for AD or to gauge whether these factors taken together define sufficient risk sets for AD. 0197-4580/$ – see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.neurobiolaging.2006.07.007