IL7 and depression: The importance of gender and blood fraction

Poster Presentations: P1 modeled using longitudinal mixed effects models with random intercepts and slopes, time was modeled with both linear and quadratic terms. For each plasma marker, two types of models were fit: 1) to assess the effect of baseline marker value on 3MS trajectory over time, and 2) to assess the time-varying effect of marker value on 3MS score at each time point. Models were fit for the full sample, and then for the subsamples of individuals who developed AD or MCI during the trial. Results: In the full sample, there were no statistically significant effects for any plasma markers on 3MS. As baseline predictors of 3MS trajectory, among 57 who converted to AD, a 1-unit difference in total cholesterol was associated with an additional decrease of 0.07 points/year (p¼0.004). A 1-unit difference in ldl was associated with an additional decrease of 0.06 points/year (p¼0.021). There were no significant baseline predictors of trajectory among 64 who converted to MCI. As time-varying predictors of 3MS, among AD converters, ldl (beta ¼ -0.05, p¼0.01) was associated with lower 3MS scores. Among MCI converters, a 1-unit difference in glucose was associated with higher scores (beta ¼ 0.02, p¼0.02), and a 1-unit difference in urea nitrogen (beta ¼ -0.19, p¼0.004) was associated with lower scores. Conclusions: Consistent with existing reports, cholesterol was associated with slightly greater decline among AD converters. Though higher uric acids levels have been previously associated with reduced dementia risk, and creatinine kinase and bilirubin have been identified as possible protective agents, we found no associations between these markers and decline. The negative effect of urea nitrogen may be reflective of a relationship between poor general health and cognitive decline.

P1-223

PENTRAXIN INFLAMMATORY BIOMARKERS AND ASSOCIATIONS WITH RISK OF DEMENTIA AND ALZHEIMER’S DISEASE IN OLDER ADULTS

Annette Fitzpatrick1, Gloria Chi1, Elizabeth Hom1, Nancy Jenny2, Oscar Lopez3, Steven DeKosky4, 1University of Washington, Seattle, Washington, United States; 2University of Vermont, Burlington, Vermont, United States; 3University of Pittsburgh, Pittsburgh, Pennsylvania, United States; 4University of Virginia, Charlottesville, Virginia, United States. Contact e-mail: [email protected] Background: Recent studies suggest that the development of Alzheimer’s disease (AD) may involve vascular as well as neurodegenerative components and that inflammation may be an early marker of this progression. Methods: We investigated two vascular endothelial biomarkers from the pentraxin family, pentraxin-3 (PTX-3) and serum amyloid protein (SAP), and subsequent development of dementia, AD and mixed dementia (vascular and AD) among 1,319 participants of the Ginkgo Evaluation of Memory Study (GEMS) over an average follow-up of 5.2 years. Cox proportional hazards regression was used to evaluate risk of dementia and AD in models adjusted for demographics, cardiovascular risk factors, physical function, and ApoE genotype. Results: Participants were a mean age of 79.0 years (SD 3.4), primarily Caucasian (95%) and 45% were female. Mean baseline levels of serum PTX-3 were 1.00 ng/mL (SD: 0.67) and mean SAP was 44,936 ng/mL (SD: 12,678). Using Cox proportional hazards regression, a 20% increased risk of dementia was found for each standard deviation increase in PTX-3 adjusted for age, gender, race, education and clinic (HR 1.20, 95% CI: 1.12 to 1.29, p < 0.001). The association remained statistically significant after adjustment for cardiovascular risk factors, physical function and ApoE-4 genotype (HR: 1.16, 95% CI: 1.06 to 1.26, p¼0.001). Associations between PTX-3 and both AD and mixed dementia were similar in the fully adjusted model (For AD -HR: 1.12, 95%CI: 1.00 -1.26, p¼0.06; mixed dementia - HR:1.16 , 95% CI: 0.99 to 1.35, p¼0.07). While no associations were found between SAP and dementia in adjusted models, evaluation of AD alone found that an increase of one SD of SAP decreased the risk of AD by 16% (HR: 0.84, 95% CI: 0.75 0.95, p¼0.004) adjusted for demographics. The association remained significant after adjustment for CVD risk factors and physical function, but was attenuated after adjustment for ApoE genotype (HR: 0.92, 95% CI: 0.79 -1.07, p¼0.28). Conclusions: As PTX3 may represent damaged endothelium while SAP may act to inhibit fibril formation of amyloid deposits as-

P235

sociated with Alzheimer’s disease, use of these biomarkers may be important for early prediction of dementia and AD in older adults. P1-224

INCREASED MYELOPEROXIDASE (MPO) PLASMA LEVELS IN PATIENTS WITH ALZHEIMER’S DISEASE

Brigitte Schreitm€uller1, Christoph Laske1, Elke Stransky2, Konstantinos Stellos3, 1Department of Psychiatry, T€ubingen, Germany; 2 Department of Psychiatry, T€ubingen, Germany; 3Department of Cardiology, Frankfurt, Germany. Contact e-mail: christoph.laske@med. uni-tuebingen.de Background: Increasing evidence supports the role of cerebrovascular risk factors in the development of Alzheimer’s disease (AD). However little is known about the molecular mechanisms involved. In the present pilot-study we investigated plasma concentrations of the pro-atherogenic marker myeloperoxidase (MPO) and its possible association with plasma amyloid-beta (Ab) 1-42/1-40 ratio - a well established predictor for the development of AD - in AD patients and healthy elderly controls. Methods: The study sample included 28 AD patients and 27 gender-matched elderly individuals with a normal cognitive status as a control group. The Mini-Mental Status Examination (MMSE) was used to determine the severity of dementia. MPO, Ab1-40 and Ab 1-42 plasma concentrations were measured by enzyme linked immunoabsorbent assay. Results: AD patients showed significantly higher plasma concentrations of MPO in comparison to healthy elderly controls (AD vs. healthy elderly controls (mean 6 SD): 132.86114.8 ng/mL vs. 55.06 42.6 ng/mL; p¼0.002). MPO plasma concentrations were negatively associated with the baseline score of MMSE (r¼-0.274; p¼0.043) and positively with plasma Ab1-42/1-40 ratio (r¼0.400; p¼0.048). In a binary logistic regression model plasma MPO concentrations were independently associated with the presence of AD (p¼0.010). Conclusions: AD patients showed significantly increased plasma levels of pro-atherogenic MPO, which could be an important molecular link between atherosclerosis and AD. Its close association with plasma Ab 1-42/Ab 1-40 ratio indicates that MPO might be a useful biomarker for the detection and risk stratification of AD patients and a potential new treatment target in AD. P1-225

IL7 AND DEPRESSION: THE IMPORTANCE OF GENDER AND BLOOD FRACTION

James Hall1, Leigh Johnson2, Melissa Edwards2, Tori Como2, Robert Barber3, Neill Graff-Radford4, Michael Devous5, Sid O’Bryant3, 1 University of North Texas Health Science Center at Fort Worth, Fort Worth, Texas, United States; 2UNTHSC, Fort Worth, Texas, United States; 3 University of North Texas Health Science Center, Fort Worth, Texas, United States; 4Mayo Clinic Jacksonville, Jacksonville, Florida, United States; 5 University of Texas Southwestern Medical Center, Dallas, Texas, United States. Contact e-mail: [email protected] Background: Interleukin 7 (IL7) is involved in B and T cell development and differentiation. Recent work suggests serum IL7 may be altered in depression. We have shown depression, and biomarkers related to depression are impacted by gender and disease status. The current study examined the link between IL7 (serum and plasma) and depression scores by gender. Methods: Serum (150 AD cases150 controls) and plasma (100 AD cases, 100 controls) levels of IL7 were assayed from the Texas Alzheimer’s Research & Care Consortium (TARCC). Non-fasting samples were collected according to standardized protocols using serum-separating and EDTA plasma tubes. IL7 was assayed via electochemiluminescense (ECL). Results: Box-Cox transformation was used. The correlation between serum and plasma for IL-7 was 0.34. In the total sample, serum (r 2 ¼ 0.16, p¼0.006) and plasma (r 2 ¼ -0.20, p¼0.007) IL7 levels were significantly, but inversely, correlated with GDS-30 scores. When split by gender, serum IL7 levels were significantly positively associated GDS scores among men (r 2 ¼0.34, p¼0.001) whereas plasma IL7 levels (r 2 ¼-0.23, p¼0.008) were significantly negatively associated with GDS scores among women. A logistic regression model predicting depression status (GDS30 >¼10) included age, gender, education, plasma and serum IL7 levels, found both significantly associated with depression status but in opposite directions. Plasma

P236

Poster Presentations: P1

IL7 was significantly negatively associated with depression (OR¼0.01, 95% CI .00-0.47). Serum IL7 was significantly positively associated with depression (OR¼3.01, 95% CI 1.02-9.31). Preliminary analyses suggest that diagnosis (AD versus control) also impacted the relation between GDS scores and IL7. Plasma IL7 levels were significantly negatively associated with GDS scores among women diagnosed with AD whereas serum IL7 was significantly negatively associated with GDS scores among female normal controls. Conclusions: Our results confirm prior work that IL7 is significantly associated with depression scores and shows that the IL7 - depression link varies by gender and blood fraction. IL7 levels are weakly correlated across blood fractions. Future work examining blood-based markers of depression must consider gender as well as blood fraction in analyses. If IL-7 is considered in as a therapeutic agent for depression or AD, the current findings suggest that gender should be considered in trial design. P1-226

EARLY DETECTION OF ALZHEIMER’S DISEASE BY ANALYSIS OF BRAIN-ENRICHED MIRNA BIOMARKERS IN PLASMA

Kira Sheinerman1, Vladimir Tsivinsky1, Fiona Crawford2, Michael Mullan2, Laila Abdullah2, Samuil Umansky1, 1DiamiR, LLC, Princeton, New Jersey, United States; 2Roskamp Institute, Sarasota, Florida, United States. Contact e-mail: [email protected] Background: Development of Alzheimer’s disease (AD) and other neurodegenerative disorders begins long before their clinical manifestation. Detailed stratified analyses of recent Phase III clinical trials of anti-AD therapies revealed promising results for the treatment of AD patients with earlier, mild stages of the disease, highlighting a high need for a screening test capable of detecting AD at early, preferably asymptomatic, stage. Methods: Recently we have demonstrated that Mild Cognitive Impairment (MCI), a heterogeneous condition characteristic of early stages of many neurodegenerative diseases, can be detected with high accuracy by analysis of brain, synapse, and neurite enriched miRNA biomarker pairs in patients’ plasma (Sheinerman et al., Aging, 2012; 4:590-605). The identified miRNA biomarker pairs successfully detected MCI in the majority of patients included in the study at asymptomatic stage 1-5 years prior to clinical diagnosis. In the follow-on studies we aim to identify effective miRNA biomarkers for differentiation of AD from other neurodegenerative diseases and monitoring of MCI to dementia progression. The approach is based on: (i) analysis of circulating miRNA, which are enriched in particular brain areas affected in various pathologies, e.g. hippocampus in AD, midbrain in Parkinson’s disease, frontal lobe in frontotemporal dementia; and (ii) analysis of changes in levels of circulating miRNA caused by processes underlying AD progression, such as destruction of synapses and neurites in earlier stages of the disease, neuronal death and spreading of pathology to previously unaffected brain areas and glial cells in later stages of the disease, and inflammation. Results: Several miRNA pairs that appear highly promising for monitoring the transition from MCI to AD dementia have been identified. We will present the most recent data and discuss the implications of our findings for the development of screening and companion diagnostic tests. Conclusions: Analysis of brain-enriched miRNA circulating in plasma represents a new promising approach for early detection of Alzheimer’s disease and its differentiation from other neurodegenerative pathologies. P1-227

THYMOSIN b4 AND RETINOL-BINDING PROTEIN 4 AS POTENTIAL SERUM BIOMARKERS FOR DIAGNOSIS OF DEMENTIA WITH LEWY BODIES

Satoko Nakashita, Kenji Wada, Nakashima Kenji, Tottori University, Department of Neurology, Yonago City, Japan. Contact e-mail: satoko-n@ med.tottori-u.ac.jp Background: Dementia of Lewy bodies (DLB) is the second most common form of neurodegenerativedementia following Alzheimer’s disease (AD). Early detection of DLB is important because its clinical features are different from those of AD. Biomarkers using cerebrospinal fluids and neuro-imaging have been studied in AD. Development of possible biomarkers for DLB is also desired. In this study, we explored serum biomarkers for DLB, which might be easily applied to routine clinical practice and epide-

miological research. Methods: Serum samples from the subjects were subjected to surface-enhanced laser desorption/ionization time-of-flight mass spectrometry (SELDI-TOF MS) analysis in order to identify a diagnostic marker for DLB. Four putative protein peaks were differentially expressed in the DLB patients compared to AD patients and the control (CTL) subjects. Tandem mass spectrometry revealed those protein peaks to be thymosin b4 (THB4) and retinol binding protein 4(RBP4). Then, serum samples from 78 subjects with Parkinson’s disease (PD), 52 subjects with AD, 22 subjects with DLB, and of 46 subjects with CTL, were used to measure the serum level of each protein by ELISA kit. Results: The mean serum THB4 level of the DLB patients was significantly higher than those of other disease patients and CTL subjects. The mean serum RBP4 level of the DLB patients was significantly higher than those of AD patients. Receiver operating characteristics (ROC) analysis of a multivariate logistics model of combination of THB4 and RBP4 exhibited the highest discriminatory ability of DLB subjects from non-DLB subjects with a sensitivity of 64%, a specificity of 87% (AUC¼0.823). Conclusions: Potential biomarkers for the diagnosis of DLB have been shown in the measurement of serum levels THB4 and RBP4. A combination of multiple biomarkers might improve a higher diagnostic accuracy. P1-228

WITHDRAWN

P1-229

STNF-R1 AS A BIOMARKER OF ALZHEIMER’S DISEASE VARIES BY APOE-ε4 STATUS

Melissa Edwards1, Tori Como2, Leigh Johnson3, James Hall4, Robert Barber5, Christoph Laske6, Neill Graff-Radford7, Michael Devous8, Sid O’Bryant5, 1University of North Texas, Fort Worth, Texas, United States; 2 University of North Texas Health Sciences Center, Fort Worth, Texas, United States; 3UNTHSC, Fort Worth, Texas, United States; 4University of North Texas Health Science Center at Fort Worth, Fort Worth, Texas, United States; 5University of North Texas Health Science Center, Fort Worth, Texas, United States; 6Department of Psychiatry, University of T€ubingen,, T€ubingen, Germany; 7Mayo Clinic Jacksonville, Jacksonville, Florida, United States; 8University of Texas Southwestern Medical Center, Dallas, Texas, United States. Contact e-mail: [email protected] Background: sTNF-R1 is one of the receptors for TNFa, which is a central immune mediator. Prior work has linked sTNF-R1 to Ab plaque formation as well as Ab neurotoxicity. TNF-R1 has also been found elevated in AD brains. sTNF-R1 was recently found to be a key marker in a serum-based algorithm that yielded excellent discrimination between AD cases and normal controls. The current study was undertaken to examine the performance of sTNF-R1 as a putative biomarker of AD presence. Methods: Serum and plasma levels of sTNF-R1 were assayed from 300 participants (150 AD cases and 150 controls) from the Texas Alzheimer’s Research & Care Consortium (TARCC) via electochemiluminescense (ECL). Results: Serum sTNF-R1 levels were significantly higher among AD cases as compared to NCs (t¼0.04) though plasma levels were not significantly different. In an adjusted model, neither serum nor plasma sTNF-R1 was a significant predictor of disease presence. However, when split by APOE4 genotype, serum sTNF-R1 was a significant negative predictor of AD status (B¼0.02; 95%CI¼0.001-0.386, p¼0.01). Plasma sTNF-R1 was not significantly associated with AD status for either APOE4 genotype. Conclusions: Our results confirm that sTNF-R1 levels are related to AD status; however, they suggest a sample medium (serum/ plasma) by genotype (APOE) interaction. This experiment highlights the need to consider blood fraction as well as APOE4 genotype when examining putative blood-based biomarkers of Alzheimer’s disease. P1-230

DESMOSTEROL: A NEW PLASMA BIOMARKER FOR ALZHEIMER’S DISEASE

Yoshiaki Sato1, Yasukazu Yamanaka2, Ikumi Suzuki3, Mamoru Yanagimachi2, Francois Bernier2, Ken Aoshima2, Yoshiya Oda4, Akinori Miyashita5, Takeshi Ikeuchi6, Ryozo Kuwano6, 1Eisai Product Creation Systems, Ibaraki, Japan; 2Eisai Product Creation Systems, Tsukuba, Japan; 3Eisai Product Creation Systems, Tsukuba, Japan; 4Eisai Product Creation Systems, Andover, Massachusetts, United States; 5Brain