Molecular Psychiatry (2006) 11, 182–186 & 2006 Nature Publishing Group All rights reserved 1359-4184/06 $30.00 www.nature.com/mp ORIGINAL ARTICLE Genome scan on Swedish Alzheimer’s disease families A Sille´n1, C Forsell1, L Lilius1, K Axelman1, BF Bjo¨rk2, P Onkamo3, J Kere3, B Winblad1 and C Graff1 1Department Neurotec, Karolinska Institutet Sumitomo Pharmaceuticals Alzheimer Center (KASPAC), Karolinska Institutet, Novum, Huddinge, Sweden; 2Department Neurotec, Experimental Geriatrics, Novum, Karolinska University Hospital Huddinge, Huddinge, Sweden and 3Department of Biosciences at Novum, Centre for Biotechnology, Karolinska University Hospital Huddinge, Huddinge, Sweden

Alzheimer’s disease (AD) is an age-related disease, which affects approximately 40% of the population at an age above 90 years. The heritability is estimated to be greater than 60% and there are rare autosomal dominant forms indicating a significant genetic influence on the disease process. Despite the successes in the early 1990s when four were identified, which directly cause the disease (APP, PSEN1 and PSEN2) or greatly increase the risk of disease development (APOE), it has proved exceedingly difficult to identify additional genes involved in the pathogenesis. However, several linkage and association studies have repeatedly supported the presence of susceptibility genes on (chrms) 9, 10 and 12. The study populations have, however, mostly been of great genetic heterogeneity, and this may have contributed to the meagre successes in identifying the disease associated genetic variants. In this study, we have performed a genome wide linkage study on 71 AD families from the relatively genetically homogeneous Swedish population where it is also possible to study the genetic ancestry in public databases. We have performed nonparametric linkage analyses in the total family material as well as stratified the families with respect to the presence or absence of APOE e4. Our results suggest that the families included in this study are tightly linked to the APOE region, but do not show evidence of linkage to the previously reported linkages on chrms 9, 10 and 12. Instead, we observed the next highest LOD score on 5q35 in the total material. Further, the data suggest that the major fraction of families linked to this region is APOE e4 positive. Molecular Psychiatry (2006) 11, 182–186. doi:10.1038/sj.mp.4001772; published online 15 November 2005 Keywords: linkage analysis; familial dementia; APOE; Alzheimer’s disease; genome scan

Introduction development of AD. To find additional AD genes, association studies in various candidate genes have Alzheimer’s disease (AD) is a severe neurodegenera- been performed; however, without any conclusive tive disorder characterized by progressive cognitive results.7 Recent whole-genome scans have shown decline. Although the genetic aetiology of the com- linkage to chrms 9, 10 and 12,8–10 which have been mon forms of AD is complex, there are rare families confirmed by fine mapping studies.10–17 Several asso- with autosmal dominant inheritance patterns. Muta- ciation studies, both genome-wide and in linkage tions in three genes amyloid precursor (APP), regions, have also been performed in case–control presenilin 1 (PSEN1), presenilin 2 (PSEN2) are known samples, which have resulted in numerous regions to cause early-onset (before the age of 65 years) AD.1–4 harbouring potential candidate genes.18–24 Linkage- These genes have been identified by linkage analysis based covariate analysis has also been made to identify and are located on chromosomes (chrms) 21, 14 and 1, loci linked to AD, by including age at onset, rate of respectively. However, these genes contribute to less decline or presence of the e4 allele as a covariate.17,25 than 1% of all AD cases. In contrast, the only known We have performed a whole-genome scan on susceptibility apolipoprotein E (APOE) has an Swedish AD families with the aim of identifying attributable risk of 10–50%.5,6 Further, the e4 allele of novel genes involved in AD pathogenesis. There are the APOE gene gives a three- to five-fold increased several strengths of studying the Swedish population. risk for AD, but is not sufficient, nor necessary for the One is the publicly available genealogic information on all Swedish citizens dating back to the mid-17th Correspondence: Dr C Graff, Department Neurotec, Karolinska century thanks to a law founded in 1686, which Institutet Sumitomo Pharmaceuticals Alzheimer Center (KAS- imposed parishes to register the Swedish population PAC), Karolinska Institutet, Novum 5th floor, Huddinge SE-141 in order to find tax payers. In separate books, birth, 57, Sweden. marriage, children, death and sometimes cause of E-mail: [email protected] 26 Received 14 January 2005; revised 5 October 2005; accepted 11 death were registered. The families included in this October 2005; published online 15 November 2005 study were traced back 3–7 generations using these Genome scan on Swedish AD families A Sille´n et al 183 parish registries. In addition, the genetic heterogene- labelled microsatellite markers (Applied Biosystems ity is likely lower in Sweden compared to many other Linkage Mapping Set version 2) evenly distributed countries such as North America used in the majority with an average spacing of 10 cM across the genome. of previous linkage studies. Here we present our The analyses were performed on a Megabase capillary linkage results on 188 individuals affected by AD DNA sequencer (Amersham), at the Finnish Genome from 71 Swedish families. Center (FGC) in Helsinki. The technicians were blind to family structure. Further genotyping of eight markers to fill in gaps greater than 20 cM on Materials and methods chrms 2, 5, 8, 10, 14, 15 and 17 was performed by Samples us. PCR assays were set up in 10 ml volumes using DNA was prepared according to standard procedures 0.4 mM primers, 0.2 mM dNTPs, 1.5 mM MgCl2, 0.25 U from whole blood. Samples from 188 individuals with AmpliTaq Goldt and 30 ng DNA. The fragments AD from 71 multiplex and sib-pair families were were separated and genotypes were scored using the genotyped. The average age at onset was 6976.8 years ABI PRISMs 3100 Genetic Analyser (Applied Bio- and 41% were heterozygous for the APOE e4 allele, systems) and GeneMappert version 2.0.1 programme, 35% were homozygous and 24% completely lacked respectively. The CEPH DNAs: 133101, 133102 the e4 allele. The AD cases (men n = 65, women and 134702 were genotyped to ensure correct scoring n = 121) were from 71 families with a history of of alleles. dementia in two or more generations and with DNA samples from at least two or more living individuals Statistical analysis affected with AD. Table 1 illustrates the number of affected individuals available for genotyping as well as Linkage analysis. We used the programme Allegro the average number of known affected in each family. (version 1.2)27 to calculate nonparametric allele- Families with mutations in the APP and the PSEN1 sharing LOD scores for multipoint (mpt) and single and PSEN2 genes were excluded. The AD cases were point (spt) analysis as well as LOD exact P-values clinically evaluated and diagnosed according to (pointwise P-values). It was performed on a Unix Sun NINCDS-ADRDA criteria as possible or probable AD. Solaris platform using the exponential allele-sharing In 23 cases representing 18 families, the clinical model28 and scoring function ‘all’29,30 with family diagnosis was confirmed by neuropathological exam- weight = 0.5, as recommended in the Allegro manual ination at the Karolinska University Hospital, Hud- to adjust differences in pedigree size. Due to compu- dinge, Sweden. For all affected individuals, detailed tational limitations (which occurred because some data on family history of dementia, age at onset of the of the families studied were > 26 bits), some ‘un- disease (defined as the age when the first symptoms informative’ siblings in the largest families were appeared) and its course were collected by interview- removed from this analysis. Model-free Allegro ing their next of kin and by studying medical records. calculations were first made on the whole set of 71 An additional 19 affected, 32 unaffected siblings families and secondly after stratification according to and 70 healthy spouses were genotyped for the APOE e4 status. In total, 46 families were stratified markers used to fill in gaps in the genetic map, see into the APOE e4 positive group ‘e4 pos families’, below. All samples were collected after informed where all affected had at least one e4 allele, and 25 consent of the participating party or next of kin. The families into the APOE e4 negative group ‘e4 neg study was approved by the local Ethical committee at families’, where at least one affected completely Karolinska Institutet, Huddinge, Sweden. lacked an APOE e4 allele. In the e4 neg families, affection status for individuals carrying an e4 allele Genotyping was set as unknown. Thereby, these individuals The genome scan was performed by PCR and do not contribute to the LOD score but they add subsequent fragment analysis of 399 fluorescently information about phase. Microsatellite allele frequencies were calculated from the genotypic data of the analysed AD families Table 1 Distribution of the 71 AD families in the genome together with genotypes from nondemented Swedish scan individuals, who were investigated earlier at the FGC in Helsinki. The nondemented individuals were No. of No. of Average No. of sib anonymous to us and we had no clinical information genotyped families total no. of pairs on these samples. Thus, for most markers 1000 to affected in affected in genotyped 1200 alleles were available. Genetic maps and inter- each family family marker distances were obtained from the deCODE map31 or the Marshfield Center for Medical Genetics 2 43 4.5 43 map (http://research.marshfieldclinic.org/genetics/). 317651 4 7 7.3 38 The genotypes obtained from filling gaps on chrms 5 2 7.5 14 2, 5, 8, 10, 14, 15 and 17 were included in the Allegro 6 2 8.5 30 calculations, including the genotypes from the addi- tional 121 samples (see above in Samples section).

Molecular Psychiatry Genome scan on Swedish AD families A Sille´n et al 184 Simulations. Simulation analysis with 1000 itera- Linkage calculations. When the whole set of families tions was performed under the null hypothesis of no was analysed, a significant LOD score of 4.84 linkage anywhere in the genome. The same marker (pointwise P = 1.1 Â 10À6, simulated global P-value maps, allele frequencies and family structures as in 0.01) was found close to marker D19S902 in the the linkage analyses were used. Three different region of the APOE gene in 19q13.33, Table 2. This simulation sets were made: the total set of 71 was the only LOD score that reached a global families, the APOE e4 negative and the APOE e4 significant linkage value according to Lander and positive families. Genotypes were simulated for all Kruglyak’s criteria.32 The e4 pos group of families also chromosomes except the X chromosome. We used the obtained its highest linkage value in the same region. simulated LOD scores to estimate the global signi- Thus, the linkage peaks most probably reflect the ficance of our three highest observed LOD scores. genetic effect of the APOE gene. The reported values are uncorrected for multiple testing. The second Genealogy highest LOD score for the 71 families was 2.40 Sweden offers good opportunities for genealogical (pointwise P = 4.2 Â 10À4) obtained in 5q35.2 with research thanks to a law from 1686, which imposed marker D5S498. This LOD score reached the level of parishes to register the Swedish population in order suggestive linkage as estimated by our simulation to find taxpayers. In separate books, birth, marriage, analysis. The 5q35 region has previously been children, death and sometimes cause of death were reported to be linked to LOAD sib-pairs.8,10 The registered. Church records at Arninge Riksarkiv, major contribution of the peak comes from the e4 Sweden were investigated for the families included pos substrata, which obtained its third highest LOD in this study. score of 1.91 (pointwise P = 0.0014) for the mpt analysis, and a LOD of 1.91 in the spt analysis at Results marker D5S498. When the sample was stratified according to APOE Successful genotyping was performed on 365 markers e4 status, the e4 pos families (n = 46) showed sig- with an average inter marker distance of 8.97 cM. nificant linkage to 19q13.32 (LOD = 4.34, pointwise When failed markers were excluded, the success P = 3.70 Â 10À6, simulated global P-value = 0.006) at rate for the genotyping at FGC was 95.5%, after position 71.5 cM, which is between marker D19S420 Mendel and haplotype checking and 87% including and D19S902. The second highest LOD score in the failed markers. After we filled in the gaps on chrms APOE e4 subgroup was suggestive of linkage with 2, 5, 8, 10, 14, 15 and 17 there remained gaps marker D4S406 in 4q25 (mpt LOD = 2.00, spt greater than 20 cM on chrms 6, 13, 18, 21 and the X LOD = 2.35). chromosome. The highest multipoint LOD score observed in the e4 neg subgroup approached the level of global suggestive linkage (LOD = 1.36, pointwise P = 0.0063 in 19q13). Simulations. Simulations were used to estimate global P-values. We have used the threshold levels suggested by Lander and Kruglyak32 to define Genealogy significant and suggestive linkage in this study. As Genealogic studies of church records identified more shown in Table 2, the highest LOD score on than 3000 ancestors dating back 3–7 generations for chromosome 19q13 reached a significant global P- the families. Research in families descending from the value for the total material as well as the APOE e4 same geographical areas made it possible to make positive families. The second highest LOD score five large families out of 12 smaller families (data (2.40) located in 5q35.2 is suggestive of linkage. not shown). However, the disease did not always

Table 2 The two highest multipoint and single-point nonparametric LOD scores in the complete family material, the APOE e4 positive and APOE e4 negative families, respectively, are listed with the global significance values estimated by simulation

Locus Stratum (no. of fam) Two-point analysis Multipoint analysis Global significance value Mpt

Max LOD Peak marker Max LOD Peak marker

19q13 Genome scan (71) 2.99 D19S902 4.84 73 cMa 0.014 19q13 e4 pos (46) 2.34 D19S902 4.34 D19S420 0.006 19q13 e4 neg (25) 1.25 D19S902 1.36 D19S902 0.77 5q35.2 Genome scan (71) 2.05 D5S498 2.40 D5S498 0.26 4q25 e4 pos (46) 2.35 D4S406 2.00 113 cMb 0.35

a73 cM, nearest genotyped marker is D19S902. b113 cM, nearest genotyped marker is D4S406.

Molecular Psychiatry Genome scan on Swedish AD families A Sille´n et al 185 segregate through the common ancestor, which is not No evidence of linkage was found in the present unexpected for a complex disease. In some families, study to any of the earlier reported AD linkage regions there was evidence of consanguinity and the inheri- on chrms 9, 10 and 12. Nor did we find linkage to tance pattern of AD in some of these and other APP, PSEN1 or PSEN2 on chrms 21, 14 and 1, families where the disease is present in both parental respectively, which is not unexpected since families lineages may be explained by an autosomal recessive with identified mutations in these genes had been inheritance (data not shown). excluded from the study. In conclusion, our results differ from previous Discussion performed genome scans on familial AD, which clearly illustrates the heterogeneity of the disease. We present the results on the first whole-genome scan This could reflect genetic differences between popu- on Swedish familial AD sib-pairs, trios and multiplex lations and can be supported by the fact that previous families. We reach statistically significant linkage reported genome scans have mostly been performed (LOD = 4.84) in chromosome 19q13 close to the APOE on American Caucasians. Concerning genetic differ- gene. The whole-sample set, including 71 families, ences between populations, we have reported that generated its highest multipoint linkage peak at PSEN1 mutations are less frequent (3%) in our location 73.06 cM, close to D19S902, located approxi- Swedish early onset AD cases38 compared to other mately 3 Mb downstream of APOE, whereas the APOE studies, which have a mutation detection rate of e4 positive substratum reached its maximum close to 20–50%.39 D19S420, approximately 1.5 Mb upstream of the This is the first genome scan performed on Swedish APOE locus. Long haplotype blocks without recom- AD families. It is possible that the Swedish popula- bination were seen in all families linked to this tion is homogeneous enough to identify genes for region, which makes it difficult to determine if APOE complex diseases such as AD in parallel to the alone contributes to the linkage peak or if there are successes in Finland and Iceland. One of the great additional AD loci in the neighbouring region. The advantages is the genealogic resources, which allow possibility of two different genes involved in the AD for the identification of common ancestries between disease process on chrm 19 has previously been families. In addition, we expect to find founder effects proposed33 and shown.34 With a denser map of in the northern parts of Sweden, as many of the markers, we hope to be able to determine if other families have lived in the same region for several genes in the region are involved in the pathogenesis of generations. Future plans are to continue the genea- AD in our families. For the APOE e4 negative families, logical investigations to try to find common ancestors the strongest signal in the multipoint linkage analysis in our AD families. Additional analysis is now was also obtained in the APOE region in 19q13; ongoing in an extended family material, using addi- however, it did not reach the level of suggestive tional families and by adding more affected indivi- linkage. This is not surprising, given the small duals to existing pedigrees, as well as using a denser number of families in this substratum. In order to map of microsatellite markers to identify new AD achieve significant LOD scores, it would be necessary genes and thus the present results should be con- to increase the number of families and/or family sidered to be preliminary. members in this group. Besides the strong linkage to the known suscept- ibility gene APOE in chr 19, we also obtained Acknowledgments suggestive linkage (LOD = 2.40) to in the whole material with the peak at D5S498. This We thank Sumitomo Pharmaceuticals for generous marker located in 5q35.2 showed signs of linkage support throughout the study. We thank Dr Ingrid also in the APOE e4 positive substrata, however, Kockum at Department of Molecular Medicine, nonsignificant. Earlier reports of linkage in 5q35 CMM, Karolinska Institutet, Sweden for guidance in have been made with a peak at marker D5S211, using the Allegro Programme, Associate Professors only 2 cM centromeric to our finding.8 Candi- Nenad Bogdanovic and Inger Nennesmo for the date genes in 5q35 are among others HMP19, the neuropathology and Drs Anders Sandstro¨mat dopamine receptor (DRD1), histamine receptor Sunderby Hospital, Lulea˚,Bo¨rje Lind at Ga¨vle (HRH), complexin-2 (CPLX2) and the beta-synuclein Hospital, Ga¨vle and Stellan Ba˚tsman at Kalix Primary gene (SCNB). Care Health, Kalix for referring new families to The suggestive linkage in the APOE e4 positive the study, Kurt Johansson and Lars Lannfelt for substrata at 4q24–25 colocalizes with the recently clinical examinations at Karolinska University identified amyloid associated protein CLAC.35,36 The Hospital, Huddinge, Sweden. We also wish to thank alpha-synuclein gene, which is associated with all the participating families without whom the study Parkinson’s disease and furthermore has been re- could not be performed. ported to be associated with AD is located in 4q22, which is proximal to our linkage peak.37 Two groups Electronic-database information have previously reported linkage to the more distal Marshfield Center for Medical Genetics maps, http:// region 4q32 on the chromosome.8,13 research.marshfieldclinic.org/genetics/.

Molecular Psychiatry Genome scan on Swedish AD families A Sille´n et al 186 References 20 Luedecking-Zimmer E, DeKosky ST, Chen Q, Barmada MM, Kamboh MI. Investigation of oxidized LDL-receptor 1 (OLR1) as 1 Goate A, Chartier-Harlin M-C, Mullan M, Brown J, Crawford F, the candidate gene for Alzheimer’s disease on chromosome 12. Fidani L et al. Segregation of a missense mutation in the amyloid Hum Genet 2002; 111: 443–451. precursor protein gene with familial Alzheimer’s disease. Nature 21 Zubenko GS, Hughes HB, Stiffler JS, Hurtt MR, Kaplan BB. A 1991; 349: 704–706. genome survey for novel Alzheimer disease risk loci: results at 10- 2 Sherrington R, Rogaev EI, Liang Y, Rogaeva EA, Levesque G, Ikeda cM resolution. Genomics 1998; 50: 121–128. M et al. Cloning of a gene bearing missense mutations in early- 22 Hiltunen M, Mannermaa A, Thompson D, Easton D, Pirskanen M, onset familial Alzheimer’s disease. Nature 1995; 375: 754–760. Helisalmi S et al. Genome-wide linkage disequilibrium mapping 3 Levy-Lahad E, Wijsman EM, Nemens E, Anderson L, Goddard of late-onset Alzheimer’s disease in Finland. Neurology 2001; 57: KAB, Weber JL et al. A familial Alzheimer’s disease locus on 1663–1668. chromosome 1. Science 1995; 269: 970–973. 23 Prince JA, Feuk L, Sawyer SL, Gottfries J, Ricksten A, Nagga K et 4 Rogaev EI, Sherrington R, Rogaeva EA, Levesque G, Ikeda M, Liang al. Lack of replication of association findings in complex disease: Y et al. Familial Alzheimer’s disease in kindreds with missense an analysis of 15 polymorphisms in prior candidate genes for mutations in a gene on chromosome 1 related to the Alzheimer’s sporadic Alzheimer’s disease. Eur J Hum Genet 2001; 9: 437–444. disease type 3 gene. Nature 1995; 376: 775–778. 24 Emahazion T, Feuk L, Jobs M, Sawyer SL, Fredman D, St Clair D et 5 Slooter AJ, Cruts M, Kalmijn S, Hofman A, Breteler MM, Van al. SNP association studies in Alzheimer’s disease highlight Broeckhoven C et al. Risk estimates of dementia by apolipoprotein problems for complex disease analysis. Trends Genet 2001; 17: E genotypes from a population-based incidence study: the 407–413. Rotterdam Study. Arch Neurol 1998; 55: 964–968. 25 Holmans P, Hamshere M, Hollingworth P, Rice F, Tunstall N, Jones 6 Ashford JW. APOE genotype effects on Alzheimer’s disease onset S et al. Genome screen for loci influencing age at onset and rate of and epidemiology. J Mol Neurosci 2004; 23: 157–165. decline in late onset Alzheimer’s disease. Am J Med Genet B 7 Kamboh MI. Molecular genetics of late-onset Alzheimer’s disease. Neuropsychiatr Genet 2005; 135: 24–32. Ann Hum Genet 2004; 68: 381–404. 26 Sjo¨gren T, Sjo¨gren H, Lindgren A˚ GH. Morbus Alzheimer and 8 Blacker D, Bertram L, Saunders AJ, Moscarillo TJ, Albert MS, Morbus Pick: a genetic, clinical and patho-anatomical study. Acta Wiener H et al. Results of a high-resolution genome screen of 437 Psychiatr Neurol Scand 1952; S82: 1–66. Alzheimer’s disease families. Hum Mol Genet 2003; 12: 23–32. 27 Gudbjartsson DF, Jonasson K, Frigge ML, Kong A. Allegro, a new 9 Kehoe P, Wavrant-De Vrieze F, Crook R, Wu WS, Holmans P, computer program for multipoint linkage analysis. Nat Genet Fenton I et al. A full genome scan for late onset Alzheimer’s 2000; 25: 12–13. disease. Hum Mol Genet 1999; 8: 237–245. 28 Kong A, Cox NJ. Allele-sharing models: LOD scores and accurate 10 Myers A, Wavrant De-Vrieze F, Holmans P, Hamshere M, Crook R, linkage tests. Am J Hum Genet 1997; 61: 1179–1188. Compton D et al. Full genome screen for Alzheimer disease: stage 29 McPeek MS. Optimal allele-sharing statistics for genetic mapping II analysis. Am J Med Genet 2002; 114: 235–244. using affected relatives. Genet Epidemiol 1999; 16: 225–249. 11 Bassett SS, Avramopoulos D, Fallin D. Evidence for parent of 30 Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES. Parametric and origin effect in late-onset Alzheimer disease. Am J Med Genet nonparametric linkage analysis: a unified multipoint approach. 2002; 114: 679–686. Am J Hum Genet 1996; 58: 1347–1363. 12 Bertram L, Blacker D, Mullin K, Keeney D, Jones J, Basu S et al. 31 Kong A, Gudbjartsson D, Sainz J, Jonsdottir G, Gudjonsson S, Evidence for genetic linkage of Alzheimer’s disease to chromo- Richardsson B et al. A high-resolution recombination map of the some 10q. Science 2000; 290: 2302–2303. . Nat Genet 2002; 31: 241–247. 13 Pericak-Vance MA, Grubber J, Bailey LR, Hedges D, West S, 32 Lander E, Kruglyak L. Genetic dissection of complex traits: Santoro L et al. Identification of novel genes in late-onset guidelines for interpreting and reporting linkage results. Nat Alzheimer’s disease. Exp Gerontol 2000; 35: 1343–1352. Genet 1995; 11: 241–247. 14 Scott WK, Grubber JM, Conneally PM, Small GW, Hulette CM, 33 Poduslo SE, Yin X. A new locus on chromosome 19 linked with Rosenberg CK et al. Fine mapping of the chromosome 12 late-onset late-onset Alzheimer’s disease. Neuroreport 2001; 4: 3759–3761. Alzheimer disease locus: potential genetic and phenotypic 34 Wijsman EM, Daw EW, Yu CE, Payami H, Steinbart EJ, Nochlin D heterogeneity. Am J Hum Genet 2000; 66: 922–932. et al. Evidence for a novel late-onset Alzheimer disease locus on 15 Scott WK, Hauser ER, Schmechel DE, Welsh-Bohmer KA, Small chromosome 19p13.2. Am J Hum Genet 2004; 75: 398–409. GW, Roses AD et al. Ordered-subsets linkage analysis detects 35 Hashimoto T, Wakabayashi T, Watanabe A, Kowa H, Hosoda R, novel Alzheimer disease loci on chromosomes 2q34 and 15q22. Nakamura A et al. CLAC: a novel Alzheimer amyloid plaque Am J Hum Genet 2003; 73: 1041–1051. component derived from a transmembrane precursor, CLAC-P/ 16 Myers A, Holmans P, Marshall H, Kwon J, Meyer D, Ramic D et al. collagen type XXV. EMBO J 2002; 21: 1524–1534. Susceptibility locus for Alzheimer’s disease on chromosome 10. 36 So¨derberg L, Zhukareva V, Bogdanovic N, Hashimoto T, Winblad Science 2000; 290: 2304–2305. B, Iwatsubo T et al. Molecular identification of AMY, an 17 Olson JM, Goddard KA, Dudek DM. A second locus for very-late- Alzheimer disease amyloid-associated protein. J Neuropathol onset Alzheimer disease: a genome scan reveals linkage to 20p and Exp Neurol 2003; 62: 1108–1117. epistasis between 20p and the amyloid precursor protein region. 37 Matsubara M, Yamagata H, Kamino K, Nomura T, Kohara K, Kondo Am J Hum Genet 2002; 71: 154–161. I et al. Genetic association between Alzheimer disease and 18 Bertram L, Hiltunen M, Parkinson M, Ingelsson M, Lange C, the alpha-synuclein gene. Dement Geriatr Cogn Disord 2001; 12: Ramasamy K et al. Family-based association between Alzheimer’s 106–109. disease and variants in UBQLN1. N Engl J Med 2005; 352: 38 Forsell C, Froelich S, Axelman K, Vestling M, Cowburn R, Lilius L 884–894. et al. A novel pathogenic mutation (Leu262Phe) found in the 19 Li Y, Hollingworth P, Moore P, Foy C, Archer N, Powell J presenilin 1 gene in early-onset Alzheimer’s disease. Neurosci Lett et al. Genetic association of the APP binding protein 2 gene 1997; 234: 3–6. (APBB2) with late onset Alzheimer disease. Hum Mutat 2005; 25: 39 Tandon A, Fraser P. The presenilins. Genome Biol 2002; 3: 3014. 270–277. 1–3014.9.

Molecular Psychiatry