- Email: [email protected]
Whole transcriptome changes in the prefrontal cortex of late-onset Alzheimer's disease patients
S94
Oral O1-01: Genetics I
whether there is an independent signal from a nearby gene, such as TOMM40. In this study we tried to replicate a recently reported association of a PolyT polymorphism in TOMM40 with age at onset among APOE 33 homozygotes. Methods: Risk for disease and age at onset analyses were performed in a total of 1594 LOAD cases (474 APOE 33) and 1190 cognitively normal controls (701 APOE 33). Rs7412 and rs429358, which define the APOE e2/e3/e4 isoforms, rs1160985, and rs4420638 (TOMM40) were genotyped using Kaspar or TaqMan genotyping technology. The polyT repeat inintron 6 of TOMM40 (rs10524523) was genotyped using fluorescence-based fragment size analysis. Expression studies were carried out using cDNA obtained from the parietal lobes of 82 AD cases and 39 cognitively normal individuals (CDR ¼ 0). Association with CSF tau, tau phosphorylated at threonine 181 (ptau181), Ab42, and Ab40 and APOE levels was tested in an independent series of 474 samples from the WU-ADRC and 259 samples from the Alzheimer’s Disease Neuroimaging Initiative. Results: We failed to find an association between any genotyped SNP and age at onset after including APOE isoform genotype in the model. We did find a significant association between the PolyT polymorphism (rs10524523) and risk for LOAD among APOE 33 homozygotes. In the APOE 33 substratum the frequency of thers 10524523 very-long allele was decreased in cases compared to controls (p ¼ 0.004; OR ¼ 0.78; 95%CI ¼ 0.65-0.95), which is the opposite direction to that reportedin the initial study. We found no association between rs10524523 and CSF tau orAb42levels or TOMM40 or APOE gene expression. Conclusions: Although we were unable to replicate the earlier association between the PolyT polymorphism and age at onset, we did observe that the PolyT polymorphism is associated with risk for LOAD among APOE 33 homozygotes in a large case-control series. The very-long allele of the polyT variant, which had previously been reported to be associated with earlier age at onset, was associated with a lower risk for AD. The polyT variant or the SNP, rs1160985, which is in high LD (r2¼0.93) among the APOE 33 carriers, could be used to identify subgroups of APOE 33 individuals with different risk for LOAD. O1-01-05
GENOME-WIDE ASSESSMENT OF COPY NUMBER VARIATIONS IN EARLY-ONSET ALZHEIMER’S DISEASE
D
Basavaraj Hooli1, Lars Bertram2, Antonio Parrado3, Rudolph Tanzi4, Kristina Mullin1, Lucille Gotta1, 1Massachusetts General Hospital, Charlestown, Mass., United States; 2Max Planck Institute for Molecular Genetics, Berlin; 3Massachusetts General Hopsital & Harvard Medical School, Charlestown, Massachusetts; 4Harvard Medical School & Massachusetts General Hospital, Charlestown, Massachusetts.
R
ET R
AC
TE
Background: Alzheimer’s disease (AD) is a genetically complex and heterogeneous disorder. Over 200 rare and highly penetrant mutations in APP, PSEN1 and PSEN2 are known to cause early-onset familial form of AD (EOAD), while APOE- e4 is the only established risk factor in late-onset sporadic form of AD (LOAD). Currently, there are no reported novel EOAD genes, while numerous studies have identified genes that show modest increase in LOAD risk (10-15%). Most of the AD genetic studies are based on the effects of sequence variants (mutations and single nucleotide polymorphisms, SNPs) on AD susceptibility. However, recent studies estimate that copy number variation (CNVs) make significant contributions to genetic and phenotypic variation. Several pathogenic CNVs have been reported to underlie risk for diseases such as cancer, autism, schizophrenia and HIV. To date, duplication in APP is the only CNV known to cause EOAD. In this study we set out to discover novel pathogenic CNVs causing EOAD. Methods: Family sample-sets: Two large family-based AD sample-sets were used in the study: the National Cell Repository for Alzheimer’s Disease (NCRAD) sample and the National Institute of Mental Health Alzheimer’s Disease Genetics Initiative Study (NIMH) sample (263 earlymixed families, age of onset < 65). Genotyping and CNV analyses: DNA samples were processed on Affymetrix Human Genome Wide SNP Array 6.0 panel using the standard protocol. After performing extensive quality checks, CNVs were inferred using two different algorithms: Hidden Markov Model and CopyNumber Analysis Module. CNVs were confirmed on Fluid-
D E T C A R T E
igm Digital 48.776 array using TaqMan copy number probes followed by fluorescence in-situ hybridization (FISH) using lymphoblast cells derived from the subjects. Results: CNV analysis revealed two EOAD families carrying APP locus duplication. In addition, we identified several rare and novel deletions and duplications: six CNVs in NIMH families, and seven CNVs in the NCRAD families. Conclusions: APP duplication remains the only established CNV in early-onset AD. Our comprehensive and systematic genome-wide CNV analyses in EOAD families yielded multiple rare and unique CNVs spanning candidate genes associated with AD, and showing segregation with the disease. This study represents one of the first in-depth genome-wide searches for CNVs in early-onset FAD.
R
O1-01-06
RISK FOR ALZHEIMER’S ASSOCIATED WITH A COPY NUMBER VARIATION IN THE COMPLEMENT RECEPTOR 1 INCREASING C3B/ C4B BINDING SITES
Nathalie Brouwers1, Caroline Van Cauwenberghe1, Sebastiaan Engelborghs1, Jean-Charles Lambert2, Karolien Bettens1, Nathalie Le Bastard1, Florence Pasquier2, Rik Vandenberghe3, Peter De Deyn1, Marc Cruts1, Philippe Amouyel2, Kristel Sleegers1, Christine Van Broeckhoven1, 1University of Antwerp, Antwerpen, Belgium; 2Institut Pasteur de Lille, Lille, France; 3University Hospital Leuven, Belgium. Background: Two multicentre genome-wide association (GWA) studies provided substantial evidence implicating the complement receptor 1 gene (CR1) in Alzheimer’s disease (AD) genetic etiology. CR1 encodes a large transmembrane receptor with a crucial role in the immune complement cascade. Methods: We performed a genetic follow-up of the GWA CR1 association in a Flanders-Belgian cohort (n¼1883), and investigated the effect of single nucleotide polymorphisms (SNPs) located in the CR1 locus on AD risk and cerebrospinal fluid (CSF) biomarker levels. Moreover, we quantified a low copy repeat (LCR) associated copy number variation (CNV) in CR1 producing different CR1 isoforms, CR1-F and CR1-S. To further strengthen our study we replicated our findings in an additional cohort of French individuals (n¼2003). Results: We obtained significant association (padj<0.03; OR¼1.24 (95%, CI: 1.02-1.51)) for one CR1 risk haplotype, which was strongest in individuals carrying apolipoprotein E (APOE) e4 alleles (padj<0.006; OR¼1.50 (95% CI, 1.08-2.09)). Also, four SNPs correlated with increased CSF amyloid Ab1-42 levels suggesting a role for the CR1 protein in Ab metabolism. Quantification of the CR1 CNV obtained significant association in carriers of CR1-S. We replicated the CR1 CNV association finding in a French cohort and calculated in the combined cohorts an OR of 1.32; 95% CI: 1.10-1.59 (p¼0.0025). Conclusions: Our data showed that the common AD risk association at het CR1 locus may well be explained by the presence of CR1-S increasing the number of C3b/C4b and CA sites and AD risk with 30% in CR1-S carriers. How precisely the different functional role of CR1-S in the immune complement cascade contributes to AD pathogenesis will need additional functional studies.
O1-01-07
WHOLE TRANSCRIPTOME CHANGES IN THE PREFRONTAL CORTEX OF LATE-ONSET ALZHEIMER’S DISEASE PATIENTS
Crystal Humphries1, John Gilbert2, Martin Kohli2, Margaret Pericak-Vance2, Deborah Mash3, 1Hussman Institute for Human Genomics, Miami, Florida United States; 2John P Hussman Institute for Human Genomics, Miami, Florida; 3University of Miami, Miami, Florida. Background: To investigate the functional pathways involved in Late-Onset Alzheimer’s disease (LOAD), RNA-seq was performed on total RNA from autopsy confirmed human frontal cortex samples using next-generation sequencing (NGS). NGS permit the identification and quantisation of known genes as well as novel isoforms, novel splicing events, coding RNAs, and non-coding RNAs (ncRNA), allowing us to see more complete differential expression patterns of genes and transcripts and to elucidate the pathways specifically affected between AD and control brain. Methods: Neuropathological specimens were sampled from age, sex and race-matched temporal
Oral O1-02: Prevention Interventions cortex (BA 38) from cases of Alzheimer’s disease, diffuse Lewy-body disease and non-neurologic normal controls. Libraries were prepared from extracted RNA using Illumina’s Directional RNA-seq and DSN normalization protocols. Each sample was run on one lane of Illumina’s HiSeq2000, generating 40-65 million reads. Using the alignment program TopHat, approximately 80% of the reads aligned to the reference genome. The Cufflinks program assembled transcripts and determined expression values. Results: An average of 85,000 unique transcripts were found in each sample with average coverage of 600 reads per transcript. Each sample expressed approximately 16,000 annotated reference sequence (RefSeq) genes. To date, 2,579 novel isoforms of RefSeq genes were discovered with several of them belonging to genes previously associated in AD (e.g. BIN1, PICALM, etc). In total, 76,000 novel transcripts (un-annotated transcripts) were observed, most mapping to intronic regions of genes (41%), the antisense strand of a known gene (46%), or the intergenic regions of annotated genes (6%). Our results also reveal hundreds of anti sense transcripts. Initial results have identified ten novel antisense transcripts that overlap genes or regions that have shown genome wide association to AD. Four of these have been verified, with the analysis of the other six currently underway. Conclusions: A significant amount of the AD transcriptome consists of novel transcripts, suggesting considerable uncharacterized regulation. In addition we have detected over 2500 previously uncharacterized alternative splicing events in AD and novel antisense RNA transcripts to known AD candidate genes and/ or regions. Studies of these antisense transcripts and novel isoforms may yield valuable insight into the etiology of AD.
O1-01-08
DEVELOPMENT OF A GENETIC EDUCATION AND GENETIC RISK DISCLOSURE PROTOCOL FOR INDIVIDUALS WITH MILD COGNITIVE IMPAIRMENT
Denise Lautenbach1, Jason Karlawish2, Robert Green1, J. Scott Roberts3, Margaret Bradbury4, Wendy Uhlmann5, Thomas Obisesan6, L. Adrienne Cupples1, Ronald Petersen7, Tolulope Fafowora6, Kristin Harkins2, Elisabeth Wood2, Lindsay Zausmer3, 1Boston University, Boston, Mass., United States; 2University of Pennsylvania, Philadelphia, Pennsylvania; 3University of Michigan, Ann Arbor, Michigan; 4Genetic and Public Policy Center, Washington, District of Columbia; 5University of Michigan, Ann Arbor, Michigan; 6Howard University, Washington, District of Columbia; 7Mayo Clinic, Rochester, Minnesota. Background: The Risk Evaluation and Education for Alzheimer’s Disease (REVEAL) Study is a series of randomized controlled trials evaluating the impact of disclosing genetic risk for Alzheimer’s disease (AD). During the fourth REVEAL trial, we developed a risk communication protocol to disclose APOE genotype and three-year risk of progressing to AD to individuals with amnestic-Mild Cognitive Impairment (MCI). Methods: We generated risk estimates using data obtained from a clinical trial involving 769 amnestic-MCI patients, which provided three-year conversion data stratified by APOE genotype (Petersen et al., 2005). We used an evidence-based approach in risk communication to develop graphics and language to communicate APOE genotype and a numerical risk estimate. Consensus was reached on specific issues including how to discuss the protective nature of the APOEe2 allele in an APOEe2/e4 genotype and the presence of two copies (versus one) of APOEe4. Results: Three-year risks for each age-group were: 8.4% for APOEe4 negative and 42.0% for APOE e4 positive individuals (ages 55-70), 20.5% for APOEe4 negative and 47.4% for APOE e4 positive (ages 71-77), and 30.7% for APOEe4 negative and 57.1% for APOE e4 positive (age 78 or older). Estimates based on MCI diagnosis and age alone (excluding genotype information) were 25.2% (ages 55-70), 34.0% (ages 71-77) and 43.9% (ages 78 or older). Educational materials were created to describe the possible APOE genotypes, an individual’s APOE genotype result and three-year AD conversion risk. For all risk estimates, pictographs and risk curves were created. Pictographs had 100 total boxes with colored-in boxes depicting a numerical risk percentage for AD. Comparison pictographs were created to display general population
S95
three-year risk for AD (5%) and three-year AD conversion risk among individuals with MCI of the same age (excluding genotype). Language was incorporated for risk disclosure of the “protective” e2 allele and the presence of two e4 alleles which may decrease or increase risk, respectively, and may not be accurately reflected in our risk estimates. Conclusions: We developed a tool for potential use in a clinical setting when individuals with MCI inquire about APOE genotype and risk of progressing to dementia.
SUNDAY, JULY 17, 2011 ORAL O1-02 PREVENTION INTERVENTIONS O1-02-01
ASSOCIATION OF HDL CHOLESTEROL LEVEL WITH THE SEVERITY OF ALZHEIMER’S DISEASE AND IMPROVEMENT FOLLOWING EXERCISE
Anthea Vreugdenhil1, George Razay2, John Cannell3, 1School of Sociology and Social Work, Launceston, Australia; 2Launceston General Hospital, Launceston, Australia; 3Launceston General Hospital, Launceston, Australia. Background: Low plasma HDL cholesterol level has been associated with increased risk of Alzheimer’s disease (AD). However, there have been no studies on the association between HDL cholesterol levels and the severity of AD and improvement following exercise. We have therefore investigated the association of HDL cholesterol levels and the severity of AD and looked at the impact of exercise on HDL levels, cognitive and physical functioning in a 4-month randomised controlled trial of patients with AD. Methods: 85 consecutive patients with AD were selected from the Memory Disorders Clinic in Launceston, Australia, of whom 40 patients were randomly allocated to either the treatment (exercise) group or the control group. Patients in the treatment group participated in a four month at-home exercise program and controls received their usual treatment only. Patients were assessed at baseline and four months followup on measures of cognitive and physical functioning, and fasting plasma lipids were measured using standard assessment tools. Results: Fifty women and 35 men participated in the study, mean age of 74.4 years (range 51-90 years) and a mean Mini Mental State Examination (MMSE) score of 18.6 (range 4-28). Using multiple regression analysis, HDL cholesterol level was independently positively correlated with MMSE score (ß coefficient 6.7, p ¼ 0.007). Patient with sever AD (MMSE score of 13 or less), compared with mild to moderate disease (MMSE score > 13), had lower HDL level by 0.23 mmol/L or 20%, p ¼ 0.008). At 4 months follow-up, patients who exercised, compared with controls, were 2.9 seconds faster on Timed Up and Go (p ¼ 0.004), had increased MMSE scores by 2.6 points (p < 0.001) and had increased plasma HDL by 0.16 mmol/L (p ¼ 0.016). These analyses were adjusted for confounding factors using general linear modelling. Conclusions: This study suggests that a low plasma HDL cholesterol level was associated with the severity of AD. Exercise may help to improve cognitive and physical functioning and this may be partly explained by the increase in plasma HDL level.
O1-02-02
PREVENTING COGNITIVE DECLINE IN THE ELDERLY THROUGH PHYSICAL ACTIVITY IN MIDLIFE
Angela Clifford, Eef Hogervorst, Stephan Bandelow, Loughborough University, Loughborough, United Kingdom. Background: Despite advances in healthcare for those with Alzheimer’s disease and other forms of dementia, the absence of an effective longterm treatment means that the prevalence of age-related cognitive decline
Oral O1-01: Genetics I
whether there is an independent signal from a nearby gene, such as TOMM40. In this study we tried to replicate a recently reported association of a PolyT polymorphism in TOMM40 with age at onset among APOE 33 homozygotes. Methods: Risk for disease and age at onset analyses were performed in a total of 1594 LOAD cases (474 APOE 33) and 1190 cognitively normal controls (701 APOE 33). Rs7412 and rs429358, which define the APOE e2/e3/e4 isoforms, rs1160985, and rs4420638 (TOMM40) were genotyped using Kaspar or TaqMan genotyping technology. The polyT repeat inintron 6 of TOMM40 (rs10524523) was genotyped using fluorescence-based fragment size analysis. Expression studies were carried out using cDNA obtained from the parietal lobes of 82 AD cases and 39 cognitively normal individuals (CDR ¼ 0). Association with CSF tau, tau phosphorylated at threonine 181 (ptau181), Ab42, and Ab40 and APOE levels was tested in an independent series of 474 samples from the WU-ADRC and 259 samples from the Alzheimer’s Disease Neuroimaging Initiative. Results: We failed to find an association between any genotyped SNP and age at onset after including APOE isoform genotype in the model. We did find a significant association between the PolyT polymorphism (rs10524523) and risk for LOAD among APOE 33 homozygotes. In the APOE 33 substratum the frequency of thers 10524523 very-long allele was decreased in cases compared to controls (p ¼ 0.004; OR ¼ 0.78; 95%CI ¼ 0.65-0.95), which is the opposite direction to that reportedin the initial study. We found no association between rs10524523 and CSF tau orAb42levels or TOMM40 or APOE gene expression. Conclusions: Although we were unable to replicate the earlier association between the PolyT polymorphism and age at onset, we did observe that the PolyT polymorphism is associated with risk for LOAD among APOE 33 homozygotes in a large case-control series. The very-long allele of the polyT variant, which had previously been reported to be associated with earlier age at onset, was associated with a lower risk for AD. The polyT variant or the SNP, rs1160985, which is in high LD (r2¼0.93) among the APOE 33 carriers, could be used to identify subgroups of APOE 33 individuals with different risk for LOAD. O1-01-05
GENOME-WIDE ASSESSMENT OF COPY NUMBER VARIATIONS IN EARLY-ONSET ALZHEIMER’S DISEASE
D
Basavaraj Hooli1, Lars Bertram2, Antonio Parrado3, Rudolph Tanzi4, Kristina Mullin1, Lucille Gotta1, 1Massachusetts General Hospital, Charlestown, Mass., United States; 2Max Planck Institute for Molecular Genetics, Berlin; 3Massachusetts General Hopsital & Harvard Medical School, Charlestown, Massachusetts; 4Harvard Medical School & Massachusetts General Hospital, Charlestown, Massachusetts.
R
ET R
AC
TE
Background: Alzheimer’s disease (AD) is a genetically complex and heterogeneous disorder. Over 200 rare and highly penetrant mutations in APP, PSEN1 and PSEN2 are known to cause early-onset familial form of AD (EOAD), while APOE- e4 is the only established risk factor in late-onset sporadic form of AD (LOAD). Currently, there are no reported novel EOAD genes, while numerous studies have identified genes that show modest increase in LOAD risk (10-15%). Most of the AD genetic studies are based on the effects of sequence variants (mutations and single nucleotide polymorphisms, SNPs) on AD susceptibility. However, recent studies estimate that copy number variation (CNVs) make significant contributions to genetic and phenotypic variation. Several pathogenic CNVs have been reported to underlie risk for diseases such as cancer, autism, schizophrenia and HIV. To date, duplication in APP is the only CNV known to cause EOAD. In this study we set out to discover novel pathogenic CNVs causing EOAD. Methods: Family sample-sets: Two large family-based AD sample-sets were used in the study: the National Cell Repository for Alzheimer’s Disease (NCRAD) sample and the National Institute of Mental Health Alzheimer’s Disease Genetics Initiative Study (NIMH) sample (263 earlymixed families, age of onset < 65). Genotyping and CNV analyses: DNA samples were processed on Affymetrix Human Genome Wide SNP Array 6.0 panel using the standard protocol. After performing extensive quality checks, CNVs were inferred using two different algorithms: Hidden Markov Model and CopyNumber Analysis Module. CNVs were confirmed on Fluid-
D E T C A R T E
igm Digital 48.776 array using TaqMan copy number probes followed by fluorescence in-situ hybridization (FISH) using lymphoblast cells derived from the subjects. Results: CNV analysis revealed two EOAD families carrying APP locus duplication. In addition, we identified several rare and novel deletions and duplications: six CNVs in NIMH families, and seven CNVs in the NCRAD families. Conclusions: APP duplication remains the only established CNV in early-onset AD. Our comprehensive and systematic genome-wide CNV analyses in EOAD families yielded multiple rare and unique CNVs spanning candidate genes associated with AD, and showing segregation with the disease. This study represents one of the first in-depth genome-wide searches for CNVs in early-onset FAD.
R
O1-01-06
RISK FOR ALZHEIMER’S ASSOCIATED WITH A COPY NUMBER VARIATION IN THE COMPLEMENT RECEPTOR 1 INCREASING C3B/ C4B BINDING SITES
Nathalie Brouwers1, Caroline Van Cauwenberghe1, Sebastiaan Engelborghs1, Jean-Charles Lambert2, Karolien Bettens1, Nathalie Le Bastard1, Florence Pasquier2, Rik Vandenberghe3, Peter De Deyn1, Marc Cruts1, Philippe Amouyel2, Kristel Sleegers1, Christine Van Broeckhoven1, 1University of Antwerp, Antwerpen, Belgium; 2Institut Pasteur de Lille, Lille, France; 3University Hospital Leuven, Belgium. Background: Two multicentre genome-wide association (GWA) studies provided substantial evidence implicating the complement receptor 1 gene (CR1) in Alzheimer’s disease (AD) genetic etiology. CR1 encodes a large transmembrane receptor with a crucial role in the immune complement cascade. Methods: We performed a genetic follow-up of the GWA CR1 association in a Flanders-Belgian cohort (n¼1883), and investigated the effect of single nucleotide polymorphisms (SNPs) located in the CR1 locus on AD risk and cerebrospinal fluid (CSF) biomarker levels. Moreover, we quantified a low copy repeat (LCR) associated copy number variation (CNV) in CR1 producing different CR1 isoforms, CR1-F and CR1-S. To further strengthen our study we replicated our findings in an additional cohort of French individuals (n¼2003). Results: We obtained significant association (padj<0.03; OR¼1.24 (95%, CI: 1.02-1.51)) for one CR1 risk haplotype, which was strongest in individuals carrying apolipoprotein E (APOE) e4 alleles (padj<0.006; OR¼1.50 (95% CI, 1.08-2.09)). Also, four SNPs correlated with increased CSF amyloid Ab1-42 levels suggesting a role for the CR1 protein in Ab metabolism. Quantification of the CR1 CNV obtained significant association in carriers of CR1-S. We replicated the CR1 CNV association finding in a French cohort and calculated in the combined cohorts an OR of 1.32; 95% CI: 1.10-1.59 (p¼0.0025). Conclusions: Our data showed that the common AD risk association at het CR1 locus may well be explained by the presence of CR1-S increasing the number of C3b/C4b and CA sites and AD risk with 30% in CR1-S carriers. How precisely the different functional role of CR1-S in the immune complement cascade contributes to AD pathogenesis will need additional functional studies.
O1-01-07
WHOLE TRANSCRIPTOME CHANGES IN THE PREFRONTAL CORTEX OF LATE-ONSET ALZHEIMER’S DISEASE PATIENTS
Crystal Humphries1, John Gilbert2, Martin Kohli2, Margaret Pericak-Vance2, Deborah Mash3, 1Hussman Institute for Human Genomics, Miami, Florida United States; 2John P Hussman Institute for Human Genomics, Miami, Florida; 3University of Miami, Miami, Florida. Background: To investigate the functional pathways involved in Late-Onset Alzheimer’s disease (LOAD), RNA-seq was performed on total RNA from autopsy confirmed human frontal cortex samples using next-generation sequencing (NGS). NGS permit the identification and quantisation of known genes as well as novel isoforms, novel splicing events, coding RNAs, and non-coding RNAs (ncRNA), allowing us to see more complete differential expression patterns of genes and transcripts and to elucidate the pathways specifically affected between AD and control brain. Methods: Neuropathological specimens were sampled from age, sex and race-matched temporal
Oral O1-02: Prevention Interventions cortex (BA 38) from cases of Alzheimer’s disease, diffuse Lewy-body disease and non-neurologic normal controls. Libraries were prepared from extracted RNA using Illumina’s Directional RNA-seq and DSN normalization protocols. Each sample was run on one lane of Illumina’s HiSeq2000, generating 40-65 million reads. Using the alignment program TopHat, approximately 80% of the reads aligned to the reference genome. The Cufflinks program assembled transcripts and determined expression values. Results: An average of 85,000 unique transcripts were found in each sample with average coverage of 600 reads per transcript. Each sample expressed approximately 16,000 annotated reference sequence (RefSeq) genes. To date, 2,579 novel isoforms of RefSeq genes were discovered with several of them belonging to genes previously associated in AD (e.g. BIN1, PICALM, etc). In total, 76,000 novel transcripts (un-annotated transcripts) were observed, most mapping to intronic regions of genes (41%), the antisense strand of a known gene (46%), or the intergenic regions of annotated genes (6%). Our results also reveal hundreds of anti sense transcripts. Initial results have identified ten novel antisense transcripts that overlap genes or regions that have shown genome wide association to AD. Four of these have been verified, with the analysis of the other six currently underway. Conclusions: A significant amount of the AD transcriptome consists of novel transcripts, suggesting considerable uncharacterized regulation. In addition we have detected over 2500 previously uncharacterized alternative splicing events in AD and novel antisense RNA transcripts to known AD candidate genes and/ or regions. Studies of these antisense transcripts and novel isoforms may yield valuable insight into the etiology of AD.
O1-01-08
DEVELOPMENT OF A GENETIC EDUCATION AND GENETIC RISK DISCLOSURE PROTOCOL FOR INDIVIDUALS WITH MILD COGNITIVE IMPAIRMENT
Denise Lautenbach1, Jason Karlawish2, Robert Green1, J. Scott Roberts3, Margaret Bradbury4, Wendy Uhlmann5, Thomas Obisesan6, L. Adrienne Cupples1, Ronald Petersen7, Tolulope Fafowora6, Kristin Harkins2, Elisabeth Wood2, Lindsay Zausmer3, 1Boston University, Boston, Mass., United States; 2University of Pennsylvania, Philadelphia, Pennsylvania; 3University of Michigan, Ann Arbor, Michigan; 4Genetic and Public Policy Center, Washington, District of Columbia; 5University of Michigan, Ann Arbor, Michigan; 6Howard University, Washington, District of Columbia; 7Mayo Clinic, Rochester, Minnesota. Background: The Risk Evaluation and Education for Alzheimer’s Disease (REVEAL) Study is a series of randomized controlled trials evaluating the impact of disclosing genetic risk for Alzheimer’s disease (AD). During the fourth REVEAL trial, we developed a risk communication protocol to disclose APOE genotype and three-year risk of progressing to AD to individuals with amnestic-Mild Cognitive Impairment (MCI). Methods: We generated risk estimates using data obtained from a clinical trial involving 769 amnestic-MCI patients, which provided three-year conversion data stratified by APOE genotype (Petersen et al., 2005). We used an evidence-based approach in risk communication to develop graphics and language to communicate APOE genotype and a numerical risk estimate. Consensus was reached on specific issues including how to discuss the protective nature of the APOEe2 allele in an APOEe2/e4 genotype and the presence of two copies (versus one) of APOEe4. Results: Three-year risks for each age-group were: 8.4% for APOEe4 negative and 42.0% for APOE e4 positive individuals (ages 55-70), 20.5% for APOEe4 negative and 47.4% for APOE e4 positive (ages 71-77), and 30.7% for APOEe4 negative and 57.1% for APOE e4 positive (age 78 or older). Estimates based on MCI diagnosis and age alone (excluding genotype information) were 25.2% (ages 55-70), 34.0% (ages 71-77) and 43.9% (ages 78 or older). Educational materials were created to describe the possible APOE genotypes, an individual’s APOE genotype result and three-year AD conversion risk. For all risk estimates, pictographs and risk curves were created. Pictographs had 100 total boxes with colored-in boxes depicting a numerical risk percentage for AD. Comparison pictographs were created to display general population
S95
three-year risk for AD (5%) and three-year AD conversion risk among individuals with MCI of the same age (excluding genotype). Language was incorporated for risk disclosure of the “protective” e2 allele and the presence of two e4 alleles which may decrease or increase risk, respectively, and may not be accurately reflected in our risk estimates. Conclusions: We developed a tool for potential use in a clinical setting when individuals with MCI inquire about APOE genotype and risk of progressing to dementia.
SUNDAY, JULY 17, 2011 ORAL O1-02 PREVENTION INTERVENTIONS O1-02-01
ASSOCIATION OF HDL CHOLESTEROL LEVEL WITH THE SEVERITY OF ALZHEIMER’S DISEASE AND IMPROVEMENT FOLLOWING EXERCISE
Anthea Vreugdenhil1, George Razay2, John Cannell3, 1School of Sociology and Social Work, Launceston, Australia; 2Launceston General Hospital, Launceston, Australia; 3Launceston General Hospital, Launceston, Australia. Background: Low plasma HDL cholesterol level has been associated with increased risk of Alzheimer’s disease (AD). However, there have been no studies on the association between HDL cholesterol levels and the severity of AD and improvement following exercise. We have therefore investigated the association of HDL cholesterol levels and the severity of AD and looked at the impact of exercise on HDL levels, cognitive and physical functioning in a 4-month randomised controlled trial of patients with AD. Methods: 85 consecutive patients with AD were selected from the Memory Disorders Clinic in Launceston, Australia, of whom 40 patients were randomly allocated to either the treatment (exercise) group or the control group. Patients in the treatment group participated in a four month at-home exercise program and controls received their usual treatment only. Patients were assessed at baseline and four months followup on measures of cognitive and physical functioning, and fasting plasma lipids were measured using standard assessment tools. Results: Fifty women and 35 men participated in the study, mean age of 74.4 years (range 51-90 years) and a mean Mini Mental State Examination (MMSE) score of 18.6 (range 4-28). Using multiple regression analysis, HDL cholesterol level was independently positively correlated with MMSE score (ß coefficient 6.7, p ¼ 0.007). Patient with sever AD (MMSE score of 13 or less), compared with mild to moderate disease (MMSE score > 13), had lower HDL level by 0.23 mmol/L or 20%, p ¼ 0.008). At 4 months follow-up, patients who exercised, compared with controls, were 2.9 seconds faster on Timed Up and Go (p ¼ 0.004), had increased MMSE scores by 2.6 points (p < 0.001) and had increased plasma HDL by 0.16 mmol/L (p ¼ 0.016). These analyses were adjusted for confounding factors using general linear modelling. Conclusions: This study suggests that a low plasma HDL cholesterol level was associated with the severity of AD. Exercise may help to improve cognitive and physical functioning and this may be partly explained by the increase in plasma HDL level.
O1-02-02
PREVENTING COGNITIVE DECLINE IN THE ELDERLY THROUGH PHYSICAL ACTIVITY IN MIDLIFE
Angela Clifford, Eef Hogervorst, Stephan Bandelow, Loughborough University, Loughborough, United Kingdom. Background: Despite advances in healthcare for those with Alzheimer’s disease and other forms of dementia, the absence of an effective longterm treatment means that the prevalence of age-related cognitive decline