THIRD INTERNATIONAL CONFERENCE ON ALZHEIMER’S DISEASE S67

261 GENETIC TRANSMISSION FOR EARLY-ONSET ALZHEIMER’S DISEASE, L.A. Farrer, C.M. van Duijn, L.A. Cupples and A. Hofman. Dept of Neurology and School of Public Health; Boston University, 02118 Boston, USA and Dept of Epidemiology & Biostatistics; Erasmus University Medical School, 3000 DR Rotterdam, The Netherlands.

It has been suggested that early-onset Alzheimer’s disease (AD) may be explained as an autosomal dominant trait. We have studied the genetic transmission of the disease in 198 patients with early-onset AD. The study comprised all patients with AD living in two areas of The Netherlands with an onset of disease before the age of 85 years. All cases were diagnosed between 1980 and 1987 and met the NINCDS-ADRDA criteria for probable AD. Detailed information on family history was collected by interviewing a next of kin of the patient. These data were always verified by a sibling of the patient. Segregation analysis was carried out using a maximum likelihood method implemented in the computer program POINTER. A variety of single gene models, polygenic models, mixed models (i.e., liability to AD determined by a major locus component, a polygenic background and random environment) and sporadic models were tested. The -2 log likelihood values for the models were compared by a likelihood ratio test in a sequential fashion.

The hypothesis that susceptibility to AD was not transmitted was rejected (p<O.OOOi). Among models postulating a single major locus or a mixed model for AD, recessive inheritance was excluded (p<O.OOOi). Hypotheses suggesting that susceptibility to AD is determined by multifactorial inheritance only (p<O.OOOl) or by a single gene (~~0.02) were rejected in favor of the mixed model. The mixed model suggested non-Mendelian transmission of AD as there was an excess transmission from the heterozygote (~~0.02) and the parent- to-child heritability ratio was significantly greater than one (p<O.OOl). In the best fitted model, the AD susceptibility allele at the major locus had a frequency of 0.1% and behaved in an autosomal dominant manner. In this model, the major locus accounted for 39% of the transmission variance; 61% of the variance was accounted for by a multifactorial component.

Our findings are very similar to a clinic-based study of early- and late-onset AD. The present study suggests that although a dominantly transmitted gene is likely implicated in early-onset AD, other genetic or environmental factors may account for the disease.

262 PRESENILE DEMENTIA AND CEREBRAL HAEMORRHAGE CAUSED BY A MUTATION AT CODON 692 OF THE P-AMYLOID PRECURSOR PROTEIN GENE. L. Hendriks, C.M. van Duijn, P. Cras, M. Cruts, W. Van Hul, F. van Harskamp, J-J. Martin, A. Hofman, and C. Van Broeckhoven. Born Bunge Foundation, University of Antwerp WA), Depts. of Biochemistry, Medicine; Universiteitsplein 1, B-2610 Antwerpen, Belgium; Erasmus University Medical School, Depts. of Epidemiology and Biostatistics; Neurology, 3000 DR Rotterdam, The Netherlands. Single base changes have been reported in exon 17 of the Pamyloid precursor protein gene (APP, transcript 770) in some families with Alzheimer disease (AD) and in hereditary cerebral haemorrhage with amyloidosis Dutch type (HCHWA-D). In AD, three different amho acid substitutions were found at codon 717 while in HCHWA-D an amino acid was changed at codon 693. In an epidemiological study of genetic risk factors in AD, we identified a four generation Dutch family, 1302, with autosomal dominant presenile dementia fulfilling the NINCDS criteria for probable AD. The mean age at onset of the dementia was 47.3 f 4.6 years (n = 6, range 41 to 55 years). In the family, we also identified 4 patients with a cerebral haemorrhage at a mean age of 39.5 f 3.1 years (range 35 to 42 years). Prior to the cerebral haemorrhage these patients did not show signs of dementia. Immunohistochemical analysis of biopsy material showed that the brain pathology was similar to that of HCHWA-D patients. The family is however, not related to the known HCHWA-D families. Direct sequencing of exon 17 revealed a base change (C + G) at position 2074 substituting alanine into glycine (Ala + Gly) at codon 692 of APP. The mutation was found in both patients with dementia or cerebral haemorrhage. Further, single strand conformational analysis (SSCA) showed that the mutation was absent in the unaffected members of the family and in 100 unrelated normal individuals. Our results suggest that in family 1302, both the dementia and the cerebral haemorrhage are caused by the same mutation of codon 692. Following this observation, we analyzed another 103 AD probands

(86 sporadic and I7 familial cases) from the epidemiological study for the mutation at codon 692 and codon 693 by SSCA or restriction enzyme digestion of exon 17. Neither one mutation was found indicating that these mutations are rare events in AD.

263 INHERITANCE OF MULTIPLE LOCI IN FAMILIAL ALZHEIMER DISEASE, J.L. Holnes’, P.H. St George-Hyslop’, J.8 Rlmmler). L. Yamooka’, A. Kazantseti, R.E. Tanzl , J.F. Gusella’, A.D. Roses’, M.A. Pericak-Vance3. ‘Molecular Neurogenetics Lab,, Massachusetts General Hospital, Boston MA 02129 U.S.A.; ?anz Institute for Neurosciences. University of Toronto, Toronto, Ontario, Canada; 3Division of Neurology, Duke University Medical Center, Durham, NC 27710 U.S.A.

Numnrous studies have indicated that Alzheimer disease has an ..- .- --. __ _. inherited component, and that about 15% of cases can be described as familial Alzheimer disease (FAD). Genetic linkage analysis has shown that a locus for FAD exists on chromosome 21 (Ch 21). and suggests another locus on chromosome 19 (Ch 19). The linkage to Ch 21 was confirmed with the description of mutations in a few early onset (mean age-at-onset ~65) dedigrees, but one pedigree without an APP mutation still gives good evidence of Ch 21 linkage. The bulk of the linkage evidence fo;Ch 19 comes from families with a late (>65) mean age-at-onset; no specific gene has yet been identified Using affected-pedigree-member (APM) analysis, the late onset oediarees also aave evidence of linkaae to Ch 21, We have now r--s. --~ w-

examined a combined set of 59 families with both early and late onset from Boston, Durham, and Toronto for linkage to Ch 21 and Ch 19 markers, including several highly informative dlnucleotide repeats. We examined the data usina four comolementarv analvtical methods. We first employed full likelihood analysis, including all available aae-at-onset Information. To examine the effect of including unaf- fected at-risk individuals, we also used likelihood analysis including only affected members of the pedigrees. To eliminate any bias Of assuming a mode of inheritance, we used APM analysis. Finally, to test the hypothesis of multiple disease loci simultaneously influencing FAD, we used a two-locus linkage model. Our results suggest: 1) that the APP gene explains only a tiny portion of FAD; 2) that Ch 21 may harbor a gene besides APP that influences FAD in both early and late onset pedigrees; 3) that Ch 19 harbors a gene influencing primarily late onset FAD; and 4) that a two disease locus model for FAD pro- vides a better explanation of the data than assuming any single locus model. As an example of the latter, affecteds-only lod scores for ATPlA3 and for D21S13 alone were 2.0 and 0.3, respectively. In the simultaneous two-locus analysis, the overall lad score was a substan- tially higher 2.9. The implications of these results will be discussed.

264 GENETIC ANALYSIS OF FAMILIAL ALZHEIMER’S DISEASE: THE APP GENE, CHROMOSOME 21 AND CHROMOSOME 19. G. Schellenberg, K Kamino, E. Wijsman, L. H&on. H. Orr. J. White. H. Payami, M. Ball, J. Kaye, A. Warren, M. McInnis, S. Antonarakis, G. Martin, and T. Bird. Depts of Neurology, Pathology, Psychiatry, and Medical Genetics, University of Washington, Seattle, WA, 98195, USA; Dept.s of Medical Gene& and Neurology Orepn Health Sciences University, Portland, OR, 97201-3098, USA; The Institute of Human Genetics, University of Minnesota, Minneapolis, MN, 55455; Dept of Laboratory Medicine Depts of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, MD 21205. USA;

Familial Alzheimer’s disease is a genetically heterogeneous disease. A sub-set of early onset families have mutations in codon 717 of the APP gene on chromosome 21. There is also evidence for a second chromosome 21 locus centromeric to the APP gene. We evaluated familial Alzheimer’s disease kindreds for linkage to chromosome 21 markers and for the presence of APP gene mutations. The family groups include Volga German families, non-Volga German early-onset kindreds and late-onset families. Linkage to the APP gene was evaluated using a CA-repeat marker D21S210 (GT12). Linkage to the centromeric region was evaluated using both RFLP and CA-repeat markers at DZlSl/Sll and D21S13. To date, screening both familial and sporadic subjects for APP mutations at codon 717 indicates that mutations at this site are rare even in early-onset FAD kindreds. No variant at this site has been identified either in sporadic AD or in normal controls. Direct DNA sequence analysis and single strand conformation polymorphism analysis of APP gene exons 16 and 17 (which encode the region of APP containing the ,9A4-peptide) has revealed 3 additional variants. At present, none of these can be directly related to FAD. Linkage analysis with GT12 strongly indicates that APP gene mutations are not responsible for FAD in both the Volga and non-Volga German early-onset families. Many early-onset kindreds have obligate recombinants behveen FAD and GT12. Likewise, results for late-onset kindreds yield significant negative evidence for linkage to this locus.