Association of Variants in Alpha Synuclein and Related Genes with Time to Dementia and Bone Outcomes using Framingham Heart Study
Document Type
Oral Presentation
Department
Biological Sciences
Abstract
Background. Alzheimer’s disease and related forms of dementia (ADRD) is prevalent in the aging population. Early diagnosis is essential, as delayed diagnosis leads to the loss of early intervention. Accurate biomarkers are limited, but a novel potential biomarker for ADRD is bone outcome assessment. Bone loss and neurocognitive decline are two common effects of aging and there is a strong association between ADRD, osteoporosis, and greater risk fracture, with negative bone outcomes often occurring prior to onset of neurological symptoms. Alpha Synuclein (SNCA), which can misfold, resulting in aggregates or Lewy Bodies, has been linked to Lewy Body Dementia (LBD), Parkinson’s Disease (PD), and Alzheimer’s Disease (AD) through its presence in the non-amyloid beta component of the AD plaque. F-box only protein 7 (FBXO7), similarly, has also been linked to PD and has been found to cause progressive neurodegeneration in altered forms. PTEN-induced kinase 1 (PINK1), which has also been linked to PD, has been found to be associated with the death of nerve cells. These losses of cells weaken communication between the brain and the muscles, causing the brain to become unable to control muscle movement. We propose that SNCA may contribute to the mechanism underlying bone loss preceding ADRD. To better understand the roles of SNCA and its related genes, FBXO7 and PINK1, on bone loss and presymptomatic dementia, we propose to perform a candidate gene analysis on single nucleotide polymorphisms (SNPs) using the Framingham Heart Study (FHS) Original and Offspring cohorts. Study Design. The Framingham Heart Study is an ongoing three-generation community-based study. At each examination age, height, body mass index (BMI), and extensive questionnaires were obtained according to standardized protocols. The FHS Original and Offspring cohorts were used for the current analysis as they had at least two bone mineral density (BMD) assessments. For our analysis we use these cohorts to create two new distinct cohorts: one for bone loss, and another for dementia. For the bone loss cohort, we used exams 20, 22, and 24 for the Original cohort and exam 6, 7, and 8 for the Offspring cohort. There was a total of 483 participants for the Original cohort and 809 participants for the Offspring cohort, leading to a combined cohort size of 1292. For the dementia cohort, exams 20 and 24 were used for the Original cohort and exams 6 and 8 were used for the Offspring. There were a total of 265 participants in the Original cohort and 1534 participants in the Offspring cohort, leading to a combined cohort size of 1799. For the dementia cohort the baseline visit is the second visit. BMD of the hip at the femoral neck (FN) was measured in grams per centimeter squared. Baseline BMD (FN BMD) was measured at the second visit (dementia baseline) and categorized by quartile. Bone loss is calculated as the percentage of change in FN BMD between first and second visits divided by the time difference between visits in years and then categorized by quartile. We include an age group by decade (e.g. <50, 50-59, 60-69, 70-69, ≥80). We also include sex, current smoking, current estrogen usage, and the ability to transfer (assessed by the Katz activities of daily living scale including the non-missing levels of “no help”, “with device”, or “with human assistance”). For the dementia cohort, ApoE4 status, given by one or two copies of the ε4 allele, is also included. Analysis. To analyze the association of baseline bone mineral density (BMD) and percent decline/year we used a linear mixed model adjusting for interrelatedness between individuals by modeling a kinship matrix as a random effect. We model time to dementia (and time to AD in a secondary analysis) using a Cox Proportional Hazards model. We correct for age, gender, and body mass index (BMI) when modeling. From the results of our analysis, we select all significant SNPs using a significance threshold of α=0.05. We also explore a variety of functional annotation tools to annotate the significant SNPs. Some of these tools include VarGen, ShAn, tmVar, and SNPnexus. We hope using these tools will lead to finding clues regarding potential mechanisms by which these genes impact bone outcomes in preclinical dementia. Results. When running the analysis on SNCA for bone loss we get a total of 40 significant SNPs. The top 5 are 90,764,310 (p = 0.01, coefficient = 0.45), 90,761,357 (p = 0.01, coefficient = 0.6487), 90,765,944, 90,744,993, and 90,704,011. For dementia, there are no significant SNPs for significance level α=0.05. We instead used α=0.1 and got two significant SNPs. These are 90,718,390 (p = 0.05, hazard ratio = 0.22), 90,673,770 (p = 0.07, hazard ratio = 1.38). There was no overlapping SNPs found between bone loss and dementia. All positions for SNCA (found on chromosome 4) uses build hg37. When running the analysis of FBXO7. for bone loss we get a total of 16 significant SNPs. The top 5 SNPs are at positions: 32,867,528 (p = 0.01, coefficient = -0.32), 32869011 (p=0.01, coefficient= -0.31), 32,869,495, 32,865,533, and 32,864,879. For dementia, we get a total of 9 significant positions. The top 5 are: 32,864,294 (p = 0.01, hazard ratio = 2.92), 32,870,769 (p = 0.03, hazard ratio = 0.73), 32,870,661, 32,864,879 and 32,866,934. When looking for overlapping SNPs for both bone loss and dementia we find that there are six overlapping SNPs: 32,869,495, 32,869,011, 32,864,879, 32,865,533, 32,867,528, and 32,870,661. All positions for FBXO7 (found on chromosome 22) uses build hg37. Conclusions. The results of SNCA show that the top significant SNPs lead to a higher decline in bone loss over time, and overall is not protective against dementia. For FBXO7, the top significant SNPs show an increase in bone density over time and is protective against dementia. Future Work. Moving forward, analysis of PINK1 will be done. Along with this, annotations for the significant SNPs for SNCA, FBXO7, and PINK1 will be added to further our analysis.
Association of Variants in Alpha Synuclein and Related Genes with Time to Dementia and Bone Outcomes using Framingham Heart Study
Background. Alzheimer’s disease and related forms of dementia (ADRD) is prevalent in the aging population. Early diagnosis is essential, as delayed diagnosis leads to the loss of early intervention. Accurate biomarkers are limited, but a novel potential biomarker for ADRD is bone outcome assessment. Bone loss and neurocognitive decline are two common effects of aging and there is a strong association between ADRD, osteoporosis, and greater risk fracture, with negative bone outcomes often occurring prior to onset of neurological symptoms. Alpha Synuclein (SNCA), which can misfold, resulting in aggregates or Lewy Bodies, has been linked to Lewy Body Dementia (LBD), Parkinson’s Disease (PD), and Alzheimer’s Disease (AD) through its presence in the non-amyloid beta component of the AD plaque. F-box only protein 7 (FBXO7), similarly, has also been linked to PD and has been found to cause progressive neurodegeneration in altered forms. PTEN-induced kinase 1 (PINK1), which has also been linked to PD, has been found to be associated with the death of nerve cells. These losses of cells weaken communication between the brain and the muscles, causing the brain to become unable to control muscle movement. We propose that SNCA may contribute to the mechanism underlying bone loss preceding ADRD. To better understand the roles of SNCA and its related genes, FBXO7 and PINK1, on bone loss and presymptomatic dementia, we propose to perform a candidate gene analysis on single nucleotide polymorphisms (SNPs) using the Framingham Heart Study (FHS) Original and Offspring cohorts. Study Design. The Framingham Heart Study is an ongoing three-generation community-based study. At each examination age, height, body mass index (BMI), and extensive questionnaires were obtained according to standardized protocols. The FHS Original and Offspring cohorts were used for the current analysis as they had at least two bone mineral density (BMD) assessments. For our analysis we use these cohorts to create two new distinct cohorts: one for bone loss, and another for dementia. For the bone loss cohort, we used exams 20, 22, and 24 for the Original cohort and exam 6, 7, and 8 for the Offspring cohort. There was a total of 483 participants for the Original cohort and 809 participants for the Offspring cohort, leading to a combined cohort size of 1292. For the dementia cohort, exams 20 and 24 were used for the Original cohort and exams 6 and 8 were used for the Offspring. There were a total of 265 participants in the Original cohort and 1534 participants in the Offspring cohort, leading to a combined cohort size of 1799. For the dementia cohort the baseline visit is the second visit. BMD of the hip at the femoral neck (FN) was measured in grams per centimeter squared. Baseline BMD (FN BMD) was measured at the second visit (dementia baseline) and categorized by quartile. Bone loss is calculated as the percentage of change in FN BMD between first and second visits divided by the time difference between visits in years and then categorized by quartile. We include an age group by decade (e.g. <50, 50-59, 60-69, 70-69, ≥80). We also include sex, current smoking, current estrogen usage, and the ability to transfer (assessed by the Katz activities of daily living scale including the non-missing levels of “no help”, “with device”, or “with human assistance”). For the dementia cohort, ApoE4 status, given by one or two copies of the ε4 allele, is also included. Analysis. To analyze the association of baseline bone mineral density (BMD) and percent decline/year we used a linear mixed model adjusting for interrelatedness between individuals by modeling a kinship matrix as a random effect. We model time to dementia (and time to AD in a secondary analysis) using a Cox Proportional Hazards model. We correct for age, gender, and body mass index (BMI) when modeling. From the results of our analysis, we select all significant SNPs using a significance threshold of α=0.05. We also explore a variety of functional annotation tools to annotate the significant SNPs. Some of these tools include VarGen, ShAn, tmVar, and SNPnexus. We hope using these tools will lead to finding clues regarding potential mechanisms by which these genes impact bone outcomes in preclinical dementia. Results. When running the analysis on SNCA for bone loss we get a total of 40 significant SNPs. The top 5 are 90,764,310 (p = 0.01, coefficient = 0.45), 90,761,357 (p = 0.01, coefficient = 0.6487), 90,765,944, 90,744,993, and 90,704,011. For dementia, there are no significant SNPs for significance level α=0.05. We instead used α=0.1 and got two significant SNPs. These are 90,718,390 (p = 0.05, hazard ratio = 0.22), 90,673,770 (p = 0.07, hazard ratio = 1.38). There was no overlapping SNPs found between bone loss and dementia. All positions for SNCA (found on chromosome 4) uses build hg37. When running the analysis of FBXO7. for bone loss we get a total of 16 significant SNPs. The top 5 SNPs are at positions: 32,867,528 (p = 0.01, coefficient = -0.32), 32869011 (p=0.01, coefficient= -0.31), 32,869,495, 32,865,533, and 32,864,879. For dementia, we get a total of 9 significant positions. The top 5 are: 32,864,294 (p = 0.01, hazard ratio = 2.92), 32,870,769 (p = 0.03, hazard ratio = 0.73), 32,870,661, 32,864,879 and 32,866,934. When looking for overlapping SNPs for both bone loss and dementia we find that there are six overlapping SNPs: 32,869,495, 32,869,011, 32,864,879, 32,865,533, 32,867,528, and 32,870,661. All positions for FBXO7 (found on chromosome 22) uses build hg37. Conclusions. The results of SNCA show that the top significant SNPs lead to a higher decline in bone loss over time, and overall is not protective against dementia. For FBXO7, the top significant SNPs show an increase in bone density over time and is protective against dementia. Future Work. Moving forward, analysis of PINK1 will be done. Along with this, annotations for the significant SNPs for SNCA, FBXO7, and PINK1 will be added to further our analysis.
