Poster Gallery

This poster gallery features research from new and early career Pitt researchers from undergraduate students through junior faculty. Their research stretches across basic, psychosocial/behavioral, clinical, and translational research on a broad range of topics related to Alzheimer’s disease and related disorders (ADRD).

Topic Areas

Posters are labeled with the following color-coded topic areas.

Basic Science & Pathogenesis Clinical Manifestations Drug Development Psychosocial Public Health Translational

Live Interaction on June 17th

To interact with the presenters during the 10 & 11am live sessions on 6/17, join the Poster Session zoom meeting (register via the event page).  Attendees can navigate the posters via breakout rooms which are assigned by the poster numbers below.

Join the Poster Session zoom at 9:45 for further information on navigating the posters.

0Reusable 3D printed enclosure with integrated cutting guides for the alignment of high resolution ex-vivo MRI with ex-vivo gross brain photographs.

Presented by Nadim Farhat, MS
PI: Tamer Ibrahim, PhD

Introduction: High resolution postmortem brain Magnetic Resonance Imaging (PMB MRI) enables submillimeter visualization of neuroanatomy and forms an intermediate space to correlate premortem radiology and histology findings. Indeed, PMB MRI can guide sampling in histology and magnifies high value targets for radiologists.  The goal of this project is to achieve an accurate histology to premortem MRI alignment while using PMB MRI as intermediate step. Towards this goal, we present a reusable 3D printed enclosure that enables the alignment of gross brain photographs to PMB MRI.

Method: Four left hemisphere forebrains (LHB) were acquired from the ADRC at the university of Pittsburgh. On the day of the imaging, he LHB are placed inside a 3D printed enclosure with an integrated cutting guide, are embedded in an agar solution and sent to MRI using our custom build TTT coil 16Tx/32Rx.  Afterwards, the neuropathologist sections the LHB into several coronal slabs coronally with the cutting guides. The knife traces on the surface of agarose are used to mark the knife positions. We matched the knife positions and replicated them with MRI multiplanar reconstruction to extract coronal PMB MRI slabs. By comparing, the cortical and ventricular profiles of both imaging techniques, we matched each gross coronal photograph to a coronal PMB MRI.

Results and Discussion: We have applied this procedure to four LHB (approx. 48 slabs) and were able to match each gross coronal photographs to a coronal PMB MRI. The procedure of matching PMB MRI to gross photographs is highly repeatable, robust, and invariant to experimental conditions.

 1TDP-43 pathology and nucleocytoplasmic transport defects in traumatic injury.

Presented by Eric Anderson, PhD
PI: Udai Pandey, PhD

Abstract
Traumatic brain injury (TBI) is a predisposing factor for many neurodegenerative diseases, including amyotrophic lateral sclerosis (ALS), Alzheimer’s disease (AD), Parkinson’s disease (PD), and chronic traumatic encephalopathy (CTE). Although defects in nucleocytoplasmic transport (NCT) is reported ALS, AD and other neurodegenerative diseases, whether defects in NCT occur in TBI remains unknown. We performed proteomic analysis on Drosophila exposed to repeated trauma and identified resultant alterations in several novel molecular pathways. Trauma upregulated nuclear pore complex (NPC) and nucleocytoplasmic transport (NCT) proteins as well as alter nucleoporin levels. Traumatic injury was further shown to disrupt RanGAP1 and NPC protein distribution in flies and a rat model and led to coaggregation of NPC components and TDP-43. In addition, trauma-mediated NCT defects and lethality are rescued by nuclear export inhibitors. Importantly, genetic upregulation of nucleoporins in vivo and in vitro triggered TDP-43 cytoplasmic mislocalization, aggregation, and altered solubility and reduced motor function and lifespan of animals. We also found NUP62 pathology and elevated NUP62 concentrations in postmortem brain tissues of patients with mild or severe CTE as well as co-localization of NUP62 and TDP-43 in CTE. These findings indicate that TBI leads to NCT defects, which potentially mediate the TDP-43 pathology in CTE.

 2Cumulative stress burden and cognitive function and impairment in African American adults living in low-income neighborhoods.

Presented by Erica Fan, MPH
PI: Andrea Rosso, PhD

Background: Alzheimer’s disease prevalence in African Americans (AA) is reported to be twice as high than in White Americans. Greater exposure to stressors may contribute to disparities in cognition. We investigated associations between stress and cognitive function and impairment in older AAs by building a cumulative stress burden (CSB) index.

Methods: Stressors, cognitive function, and cognitive impairment were assessed in 253 participants aged >50 years from primarily AA neighborhoods in Pittsburgh, PA. We created a CSB index by dichotomizing and summing perceived stress, psychological distress, unfair treatment, and post-traumatic stress disorder scores. Six cognitive domains were Z-scored and adjusted for age, sex, and education. Global cognition scores were created by averaging domain scores. Cognitive impairment (mild cognitive impairment or dementia) was clinically adjudicated using race-specific and local norms. Adjusted generalized linear models assessed relations of the CSB index or specific stressors with cognitive domains and cognitive impairment in separate models. Interactions between age and CSB were assessed.

Results: Greater CSB was associated with greater impairment (b=0.30, p=0.001) and poorer language (b=-0.11, p=0.03) and executive function (b=-0.087, p=0.04) in adjusted models. Specific stressors of the CSB index were not related to any cognitive domain or cognitive impairment. Age interactions indicate that domain findings were generally stronger for younger participants. No interactions were found for cognitive impairment.

Discussion: Greater cumulative stress exposure is associated with poorer domain-specific cognitive function and greater impairment in older AAs. A comprehensive stress assessment is vital in understanding the dimensionality of racialized stress for older adults at risk for cognitive decline.

 3A Cellular Optogenetic Model of Tau Protein Aggregation.

Presented by Bryan Hurtle
PI: Amantha Thathiah, PhD

Abstract
Alzheimer’s disease (AD) and >20 other dementias termed tauopathies are pathologically defined by insoluble aggregates of the microtubule-associated protein tau (MAPT). The onset and progression of clinical AD is strongly associated with both specific tau post-translational modifications found in cerebrospinal fluid and tau aggregation in patient brains. However, the biological basis for the transition of tau from a microtubule-associated protein to an insoluble aggregate, as well as the precise relationship between tau aggregation and neuronal dysfunction, remains unknown. Recent studies have reported the development of photo-inducible systems to control protein association using photo-oligomerizable proteins. We hypothesized that the development of a photo-inducible system, which permits temporal control of tau aggregation, would provide a novel tool to study the cause and consequence of tau aggregation. First, we fused Vivid (VVD), a photoreceptor that self-associates upon exposure to blue light, to full-length (2N4R) tau, termed optoTAU. We then determined that light-treated cells that express optoTAU develop soluble and insoluble high-molecular-weight (HMW) aggregates, thereby allowing for direct investigation of the temporal aggregation of tau. Importantly, light-treated optoTAU aggregates display amyloid positivity, are stable after light treatment, and demonstrate a pattern of insoluble aggregation that is both tau-isoform specific and significantly increased by dementia-associated MAPT mutations. OptoTAU is a novel disease model to directly study HMW tau formation in cells. Furthermore, these exciting results demonstrate the utility of optoTAU as a discovery platform to explore both cellular and pharmacological modifiers of tau aggregation.

 4DNA methylation in the human precuneus changes with age at many genomic sites.

Presented by Alex Iacovelli
PI: Brandon McKinney, MD, PhD

Background: A consistent gene set undergoes age-related changes in expression in the human cerebral cortex. The Age-by-Disease Model posits that these changes contribute to late-life neuropsychiatric diseases by “pushing” the expression of disease-associated genes in disease-promoting directions. DNA methylation (DNAm), the addition of a methyl group to a cytosine in DNA, regulates gene expression and is an attractive candidate for contributing to changes in age-related gene expression. The precuneus, located on the medial surface of the parietal lobe, is one of the nodes in the default mode network and is particularly vulnerable to early deposition of amyloid in individuals with Alzheimer’s disease. The function of the precuneus, generally, and particularly its role in Alzheimer’s disease is poorly understood. We hypothesize that age-related changes in DNAm in the precuneus pushes expression of Alzheimer’s disease-associated genes in disease-promoting directions.

Methods: To begin the test of our hypothesis, we used the Illumina Infinium MethylationEPIC array to characterize geneome-wide DNAm in postmortem precuneus from 98 human subjects without neuropsychiatric disease between the ages of 20 and 96 years.

Results and Conclusion: Though analyses of these data are in early stages, initial results show robust age-related chages in DNAm at many genomic sites in the human precuneus.

Future Directions: To further address our hypothesis, future analyses will integrate these data with transcriptomic and proteomic data from the same subjects and explore the relationship among DNAm, transcription, and protein levels for Alzheimer’s disease-associated genes across the lifespan.

 5Temporal relations among emotional and behavioral factors in late-life depression during the COVID-19 pandemic.

Presented by Joseph Kazan, MD
PI: Sarah Stahl, PhD

Abstract
Late-life depression is associated with social isolation and poor quality of life. The COVID-19 pandemic created unique psychological stressors and disrupted social support systems, which may have uniquely impacted older adults’ depressive symptomatology. We aimed to study the relationship between emotional (depression, anxiety, stress) and behavioral (sleep and physical activity) factors over time in older adults with a history of depression. We conducted weekly assessments over 12 weeks with 20 participants aged 60 years and older with a previous diagnosis of Major Depressive Disorder. Assessments consisted of telephone/zoom interviews and included the following 5 standardized questionnaires: Montgomery –Åsberg Depression Rating Scale (MADRS), Hamilton Anxiety Rating Scale (HAM-A), Perceived Stress Scale (PSS), Insomnia Severity Index (ISI), and Physical Activity Scale for the Elderly (PASE). We performed a series of cross-lagged panel models (CLPM) to examine within-week correlations among the five variables. We also conducted a set of five CLPM models, each focusing on a single measure acting as a predictor for all five variables the subsequent week. Results indicate that all five variables are strong predictors for their own scores the following week. Depression was the only variable found to be a strong predictor for all other variables the following week and was associated with increases in anxiety, stress, insomnia; and decreases in physical activity. Our study highlights the protracted negative effects of depression on the emotional and behavioral wellbeing of older adults and supports the need for longitudinal assessments and targeted interventions for late-life depression.

.

 6Examining the neural basis of worry, stress, anxiety, and rumination in late life.

Presented by Antonija Kolobaric, ScB
PI: Carmen Andreescu, MD

Background: Late life anxiety disorders are associated with worsening cognitive decline, lower gray matter volumes, and accelerated brain aging. Despite high prevalence, various anxiety disorders such as generalized anxiety disorder (GAD) are often undiagnosed and untreated. Symptoms such as severe worry or rumination respond poorly to standard treatment and drive the morbidity associated with GAD by increasing the cognitive load and acting as a risk factor for Alzheimer’s Disease (AD). We explored the association of these distinct anxiety phenotypes (worry/rumination/global anxiety) with functional connectivity markers in networks involved in both emotion regulation and stress.

Methods: We recruited participants with varying levels of worry (N=91) to undergo assessments and resting state fMRI. Following standard preprocessing, we computed seed-based connectivity for the bed nucleus of the stria terminalis (BNST), the paraventricular nucleus (PVN), habenula, and left/right amygdala (AMG). We extracted regions for each network based on their canonical networks in 1000 subjects (neurosynth). Using connectivity and clinical factors, we fit cross-validated elastic net models to predict participants scores on: Penn State Worry Questionnaire, Response Styles Questionnaire, Hamilton Anxiety Rating Scale, and Perceived Stress Scale.

Results: We found that greater worry was associated with greater PVN-subgenual ACC, parahippocampal (PHC), and olfactory and AMG-PHC connectivity. Additionally, anxiety and stress had distinct patterns of connectivity.

Conclusions: Our study demonstrated that different anxiety phenotypes are associated with a unique set of connectivity markers. This may aid in the development of future targeted interventions to decrease the cognitive load of anxiety and lessen its role in accelerated brain aging.

 7Neural asymmetry during working memory and its association with markers of preclinical Alzheimer’s disease.

Presented by Elizabeth Mountz
PI: Helmet Karim, PhD

Introduction: Beta-amyloid (Aβ) has been shown to deposit asymmetrically early in the Alzheimer Disease (AD) trajectory, which has functional consequences (e.g., asymmetric hypometabolism). We aimed to investigate whether AD markers and cognitive function were correlated with neural activation asymmetry during a working memory task.

Methods: We recruited 87 cognitive healthy older adults to undergo functional magnetic resonance imaging (fMRI), positron emission tomography (PET) imaging, and cognitive testing. We identified regions of significant (voxel-wise) asymmetry during the digital symbol substitution task (DSST). We examined the association between hemispheric asymmetry (directional and absolute separately) and age, sex, education, global cerebral amyloid, APOE allele status, global cerebral metabolism, DSST reaction time, white matter hyperintensities, and multiple domains of cognitive function.

Results: We identified seven regions of significant asymmetry. Six regions showed significant left-hemisphere asymmetry: precuneus, mediodorsal thalamus, superior parietal, precentral, supplementary motor area, and rolandic operculum. Angular gyrus alone exhibited significant right-hemisphere asymmetry. Participants who have at least one APOE4 allele tended to show more symmetric activation in the angular gyrus. Greater education was associated with more symmetric activation in mediodorsal thalamus. Functional asymmetry of any region was not associated with cognitive function.

Discussion: Greater genetic risk factor is associated with symmetric activation in the angular gyrus, which may represent compensatory activation. Greater education was associated with symmetric activation in the mediodorsal thalamus, which may be an indicator of greater cognitive reserve. Greater symmetric recruitment of subcortical and cortical regions during working memory may underlie maintained cognitive function in the preclinical stage.

 8G protein-coupled receptor kinases are associated with Alzheimer’s disease pathology.

Presented by Thais Rafael Guimarães, BS
PI: Amantha Thathiah, PhD

Abstract
Alzheimer’s Disease (AD) is characterized by extracellular deposition of amyloid-β (Aβ) in amyloid plaques and intracellular aggregation and accumulation of hyperphosphorylated tau in neurofibrillary tangles (NFTs). Although several kinases have been identified to contribute to the pathological phosphorylation of tau, kinase-targeted therapies for AD have not been successful in clinical trials. Critically, the kinases responsible for numerous identified tau phosphorylation sites remain unknown. G protein-coupled receptor (GPCR) kinases (GRKs) have recently been implicated in phosphorylation of non-GPCR substrates, e.g., tubulin and α-synuclein, and in neurological disorders, including schizophrenia and Parkinson’s disease. Accordingly, we investigated the involvement of GRKs in the pathophysiology of AD. We performed a comprehensive immunohistochemical and biochemical analysis of the ubiquitously expressed GRKs, namely GRK2, 3, 5, and 6, in postmortem human brain tissue of control subjects and AD patients. The GRKs display unique cell type-specific expression patterns in neurons, astrocytes, and microglia. Levels of GRKs 2, 5, and 6 are specifically decreased in the CA1 region of the AD hippocampus. Biochemical evidence indicates that the GRKs differentially associate with total, soluble, and insoluble pools of tau in the AD brain. Complementary immunohistochemical studies indicate that the GRKs differentially co-localize with total tau, phosphorylated tau, and NFTs. Notably, GRKs 3 and 5 also co-localize with amyloid plaques. These studies establish a link between GRKs and the pathological phosphorylation and accumulation of tau and amyloid pathology in AD brains and suggest a novel role for these kinases in regulation of the pathological hallmarks of AD.

 9Prediction of cognitive function based on structural MRI images using a 3D convolutional neural net among cognitively normal older adults.

Presented by Ashti Shah, BS
PI: Howard Aizenstein, MD, PhD

Background: Machine learning algorithms have been successful at predicting whether an individual has Alzheimer’s Disease based off an MRI image. We draw inspiration from 3D convolutional neural nets (CNNs) to predict the current cognitive function of a cognitively normal adult based on a single MRI.

Methods: The neural net included four convolutional layers that used 3x3x3 filters and the weights were adjusted using mean squared error loss (MSEL) (n=159). The inputs to the CNN were MPRAGE MR-T1 weighted images and the outcomes were individual cognitive score and overall subdomain scores (i.e., memory-retrieval, memory-learning, language, visuospatial, and executive attention) calculated from neuropsychological tests. A cross-validated linear regression to predict cognitive function from age served as a comparative model.

Results: Across all cognitive domains, the 3D CNN was up to 2.1 times better at predicting cognitive function than the linear regression model. A backpropagation analysis indicated that the neural net relied on brain regions known to be correlated with certain cognitive domains (i.e., parietal, temporal, and frontal lobes) to make its cognitive function predictions without prior knowledge of the importance of certain brain regions associated with cognition. The cerebellum crus 1 and 2 are also predictive of cognitive function.

Conclusions: Representation learning models (i.e., convolutional neural nets) can be used to predict cognitive function based off MRI images of cognitively normal older adults. The success of this model has been determined by its reliance on well-established correlations between certain brain regions and cognitive function. This algorithm may be a promising methodology for predicting changes in cognitive function in elderly individuals.

10Dual-task gait parameters are associated to cognitive function in older adults.

Presented by Anisha Suri, MS
PI: Andrea Rosso, PhD, MPH

Introduction: Limitations in community mobility are a potential biomarker for dementia risk. Community mobility involves walking with physical and cognitive tasks that challenge gait. We tested gait performance during complex walking conditions that mimic community environments and evaluated associations of gait parameters with cognitive function. 

Method: In community-dwelling older adults, accelerometry signals in Vertical(V), Anterior-Posterior (AP) and Mediolateral (ML) directions were recorded while walking on even and uneven surfaces with or without an alternate alphabeting dual-task (ABC). Gait speed was recorded. Additional gait parameters quantifying variability (stride time CoV), rhythmicity (cadence), smoothness (harmonic-ratio), and complexity (lempel-ziv) were extracted. Spearman correlations were computed between gait in the four conditions of walking (even, even-ABC, uneven, uneven-ABC) and Mini-Mental State Exam (MMSE), a measure of global cognitive function.

Result: Older adults (N=114, Age=74±6 years, 60% females, MMSE=28.5±1.7) participated. Greater gait speed in all four conditions was associated with better cognition (ρ=0.19 to 0.24, p<0.05). We used gait speed as a covariate for other gait parameters and found less variability in even-ABC (ρ_p=-0.33, p<0.001) and uneven-ABC (ρ_p=-0.24, p<0.05) associated with better cognition. Greater rhythmicity (ρ_p=0.21 to 0.29, p<0.05) in all four tasks indicated better cognition. Greater smoothness V and AP (ρ_p=0.19, p<0.05) in even-ABC, greater complexities ML, V and AP in even-ABC and uneven-ABC (ρ_p=0.21 to 0.29, p<0.05) associated to better cognition.

Conclusion: A better gait performance under various conditions is associated to better MMSE scores. Gait parameters, particularly under challenged conditions, may be a biomarker for poor cognitive function and risk for dementia.

11Discrepancy between plasma pTau181 and tau-PET statuses.

Presented by Cécile Tissot, BSc
PI: Tharick Ali Pascoal, MD, PhD

Background: Tau is usually assessed via cerebrospinal fluid or Positron Emission Tomography (PET). However, those methods are expensive and not readily available. Research focuses on cost-effective blood-based biomarkers, leading to a novel immunoassay for plasma pTau181 (plasma181). Even though plasma181 and PET are strongly associated, no study has yet investigated whether the status of plasma181 always correlate with tau-PET status.

Methods: 269 individuals from the TRIAD cohort (164 cognitively unimpaired (CU), 60 MCI and 45 AD) underwent plasma181 assessment, [¹⁸F]MK6240 tau-PET, [¹⁸F]AZD4694 amyloid-PET, MRI and neuropsychological evaluation. We conducted statistical analyses to calculate the threshold values for positivity (plasma181 and tau-PET) and make comparisons between groups. For tau-PET, we used [¹⁸F]MK6240 SUVR in the temporal meta-ROI.

Results: Using Receiver Operating Characteristic (ROC) curves, we established cut-offs for positivity for plasma181 at 11.1060 pg/mL, and tau-PET at 1.1827. 182 individuals had concordant statuses, 87 were discrepant. 64 were plasma+/PET-, (mainly CU) and 23 plasma-/PET+, (mostly MCI or AD) (Figure1). Moreover, concomitant plasma/PET statuses were usually accompanied by similar Ab-status. CU-Ab+ individuals were found in high proportion in the plasma+/PET- group.

Conclusion: Most participants had concordant statuses for plasma181 and tau-PET, which correlate with memory scores and diagnoses. However, there was a significant portion of discordant individuals, predominantly plasma+/PET-. This corroborates findings from CSF pTau181 and tau-PET, in which CSF positivity precedes PET. Plasma181 is also great to assess Ab-status, even before the onset of tau-PET abnormalities. Assessing tau through plasma or PET is thought to suggest different stages of pathological progression.

12[¹⁸F]FDG hypometabolism and blood-brain barrier damage precede cognitive impairment in rats recovered from a systemic inflammatory episode.

Presented by Bruna Bellaver, PhD
PI: Eduardo R. Zimmer, PhD

Methods: Wistar rats (90-days old) were submitted to cecal ligation and perforation (CLP; a model of sepsis) or sham surgery. Short-term memory was assessed 30- and 120- days later through the object recognition test. Brain glucose metabolism was evaluated via micro-PET [¹⁸F]FDG. CSF albumin levels and cellularity were assessed by HPLC and flow cytometry, respectively, to evaluate blood-brain barrier (BBB) integrity.

Results: A whole brain [¹⁸F]FDG hypometabolism was observed in sepsis group 30-days post-surgery. Percentage of change and t-statistical maps revealed a peak of ~17% in the hippocampal area. The evaluation of [¹⁸F]FDG metabolism in rats 120-days after CLP demonstrated a persistent brain hypometabolism. Regional analysis evidenced a significant decrease in [¹⁸F]FDG uptake in cortical and subcortical regions. Despite a brain hypometabolism being observed 30-days after sepsis, no cognitive deficit was evident at this point. An impairment in short-term memory was observed 120-days post- CLP. It was evidenced an increase in CSF albumin levels and cellularity in sepsis animals 30-days following surgery.

Conclusion: These results demonstrate that a systemic inflammatory episode causes long-term brain [¹⁸F]FDG hypometabolism and BBB damage that can be detected earlier than behavioral alterations. It suggests that [¹⁸F]FDG hypometabolism might be a link between peripheral inflammation and cognitive decline/neurodegeneration.

13Interaction between vascular risk and Alzheimer’s disease pathology boosts neurodegeneration and cognitive decline in cognitively unimpaired individuals.

Presented by João Pedro Ferrari-Souza
PI: Tharick Pascoal, MD, PhD

Background: Vascular risk factors (VRFs) are associated with cerebrovascular lesions in Alzheimer’s disease (AD) patients and the presence of these injuries contributes to dementia onset. Nevertheless, studies have reported conflicting results concerning how VRFs are associated with AD pathology to promote cognitive impairment.

Objective: Investigate whether VRF burden interacts with AD pathology to accelerate neurodegeneration and cognitive decline in cognitively unimpaired (CU) individuals and determine if VRF burden is related to changes in AD biomarkers.

Methods: We assessed CU participants from the ADNI database. Individuals were dichotomized as having an elevated VRF burden (V+) and as presenting biological AD (AD+). Linear mixed- effects models (LMMs) were performed to evaluate the longitudinal trajectory of (i) plasma neurofilament light (n=269), (ii) cognition indexed by mPACC (n=503), and (iii) cerebrospinal fluid (CSF) Aβ1-42 and p-tau181 (n=284). Additionally, time-to-event analysis was carried to evaluate the risk of clinical progression to cognitive impairment (n=503).

Results: LMMs demonstrated that an elevated VRF burden interacted with AD pathology to promote higher rates of neurodegeneration (β=5.68; p=.005)(Figure 1A) and cognitive decline (β=-0.43; p=.019)(Figure 1B). Nonetheless, VRF burden was not associated with CSF Aβ1-42 or p-tau181 changes over time. Moreover, only the AD+V+ group presented a greater risk of clinical progression in comparison to the AD-V- group (adjusted HR=3.56; p<.001)(Figure 2).

Discussion: Clinical assessment of VRF burden contributes to determining the prognosis of patients at higher risk for developing cognitive impairment. Our findings present an immediate relevance for public health since most of VRFs are preventable or treatable.

14The combination of plasma p-Tau181 and p-Tau231 enhances Alzheimer’s disease in vivo classification.

Presented by Pâmela Ferreira, PhD
PI: Pascoal Tharick, PhD

Background: Blood-based biomarkers, mainly measuring phosphorylated tau epitopes (p-Tau) have shown promising results in detecting Alzheimer’s disease (AD). However, it is still unknown whether the use of different plasma p-Tau epitopes offer overlap or complementary information. Here, we tested the hypothesis that plasma p-Tau181 and p- Tau231 markers provide complementary information to each other to identify AD pathophysiology.

Methods: Plasma p-Tau181, p-Tau231, [¹⁸F]AZD4694 amyloid-PET, [¹⁸F]MK6240 tau- PET, MRI and cognitive assessment from 284 individuals [30 CU young adults, 155 CU, 54 MCI and 38 AD] were obtained from the McGill TRIAD cohort. The individuals with AD or MCI clinical diagnosis were classified as CI. Cut-points (p-tau181, 11.1 ng/mL; p-tau231, 11.8 ng/mL) were derived using young CU controls as reference. Individuals were group based on these two biomarkers. Then, imaging and cognition measurements were compared across groups.

Results: Plasma p-Tau p-Tau181 and p-Tau231 showed moderate correlation with each other (r=0.64, p <0.0001)(Figure 1). The p-Tau181+/p-Tau231+ group presented higher tau PET and amyloid PET uptake compared to groups negative for both or only one p-tau biomarker (all p<0.01)(Figure 2A, B). Similarly, p-Tau181+/p-Tau231+ group showed worse cognitive performance and hippocampal atrophy than individuals with at least one p-tau biomarker negative(Figure 2C-E).

Conclusion: Individuals positive for both p-Tau231 and p-Tau181 showed worse cognitive performance, lower hippocampal volume, higher amyloid and tau burden than individuals with only one abnormal marker. Our findings highlight that quantifying more than one plasma p-tau epitope could lead to enhancing AD in vivo detection, with potential applications to trial screening and recruitment.

15New directions in understanding Tau-mediated neurodegeneration.

Presented by Patrick Flaherty, PhD
PI: Patrick Flaherty, PhD

Abstract
Alzheimer’s Disease (AD) and other Tau-related neurodegenerative diseases are common and cause serious morbidity in patients and significant societal burden. There are no effective treatments that cure or prevent the progression of AD. Considerable work has been conducted on strategies to reduce or reverse the accumulation of proteins such as hyperphosphorylated Tau. These approaches arise logically from inherited neurodegeneration resulting from mutations in the Tau gene, and correlations between the neuroanatomical location and amount of Tau deposition in the CNS with the type and severity of the cognitive deficits in AD over time. However, these approaches have not delivered an FDA-approved drug that is disease-modifying. A broader approach for designing new therapeutic strategies employs whole-animal models of tauopathy to identify chemicals that rescue the phenotype in a model system. The Burton Lab has developed a transgenic zebrafish model that expresses human Tau, resulting in neurodegeneration, tauopathy, and neurological phenotypes. The kinase medicinal chemistry program in the Flaherty Lab at Duquesne University provides a complimentary skill set that advances this approach. This work simultaneously explores identification of novel biological/pharmacological strategies for mitigating tauopathy and produces new drug-like small molecules that modify new targets. This combination provides insight into the pathology of the disease as well as small molecules with translational potential to prevent disease progression.

16Networks of worry – the complex neural underpinning of a popular phenotype.

Presented by Andrew Gerlach, PhD
PI: Carmen Andreescu, MD

Background: Worry is a transdiagnostic phenotype encountered in multiple mental disorders and independently associated with increased morbidity, including cognitive impairment and cardiovascular diseases. We investigated the neurobiological basis of worry in older adults by analyzing resting fMRI from a systems neuroscience perspective.

Methods: We collected resting fMRI on 77 participants (>50 yo) with varying worry severity. We computed region-wise connectivity across the Default Mode Network (DMN), Anterior Salience Network (ASN), and left Executive Control Network (LECN). All 22,366 correlations were regressed on worry severity and adjusted for age, sex, race, education, disease burden, depression, anxiety, rumination, and neuroticism. We employed higher criticism (HC) thresholding, a second-level method of significance testing for rare/weak features, for correlation selection. Aggregate correlations were used to summarize network-level signatures of worry.

Results: Half the relevant intra-network connections are within DMN. Negative correlations with worry severity dominate throughout the cingulate, temporal lobe, and cuneus, while frontal regions show bidirectional associations with worry. Within ASN, negative correlations with worry severity abound, particularly in the ACC, inferior frontal regions, and thalamus. Positive correlations with worry severity in the left PCC and right temporal lobe and negative correlations in frontal regions are notable within LECN. Inter-network analysis reveals a rich, but complex pattern of connectivity.

Conclusions: Worry severity is associated with complex resting state intra- and inter-network connectivity signatures independent of other clinical and demographic variables. Identifying the most salient and unique connections may be useful for targeted interventions for reducing morbidity associated with severe worry in older adults.

17Biased GPR3 signaling reduces Aβ pathology in vivo.

Presented by Yunhong Huang, PhD
PI: Amantha Thathiah, PhD

Abstract
Selective G protein or β-arrestin signaling, i.e., biased signaling, more precisely regulates the biological functions of G protein-coupled receptors (GPCRs). This discovery has led to development of therapeutic drugs with superior efficacy and reduced side-effects in heart disease, pain management, and neuropsychiatric disorders. Accumulation and aggregation of amyloid-β (Aβ) and tau are pathological hallmarks of Alzheimer’s disease (AD). We previously discovered that the GPCR GPR3 is involved in regulation of Aβ pathology and cognitive deficits in AD mouse models and that GPR3 levels are elevated in a subset of AD patients. Our in vitro evidence showed that β-arrestin (rather than classic G protein) signaling modulates GPR3-mediated Aβ accumulation. Here, we demonstrate that GPR3 is phosphorylated in vitro and that mutagenesis of C-terminal serine phosphorylation residues (GPR3-Ala) abrogates β-arrestin recruitment and Aβ generation in a cell culture model. Importantly, G protein-signaling and downstream mitogen-activated protein kinase activation are unaffected, indicating that the GPR3-Ala mutant specifically alters β-arrestin signaling and Aβ generation. We then generated a G protein-biased GPR3 mouse model (Gpr3^HA-Ala) to investigate the physiological consequences of eliminating β-arrestin signaling in vivo. In contrast to the phenotypes observed in Gpr3-deficient mice, we show that Gpr3^HA-Ala mice do not display memory deficits, anxiety-like behavior, or reduced fertility. Moreover, Aβ accumulation is reduced in a Gpr3^HA-AlaAD mouse model. Collectively, in vivo G protein-biased GPR3 signaling reduces Aβ pathology without side-effects associated with the absence of G protein signaling. We propose that biased GPR3 signaling may be a viable AD therapeutic strategy.

18Norm and raw hippocampal volumes as predictors of memory performance.

Presented by Arabella Johnson
PI: Akiko Mizuno, PhD

Introduction: Alzheimer’s disease (AD) is associated with reduced hippocampal volume and function, resulting in the hallmark symptom of memory impairment. However, whether to account for hippocampal atrophy by raw volume or age- and sex-adjusted normative volume is unclear.

Methods: We analyzed the ADNI2 baseline data of all diagnostic groups [cognitively normal (CN), AD, subjective memory complainers (SMC), mild cognitive impairment (MCI)]. Means and standard deviations were calculated from CN patients’  hippocampus/intracranial-volume for each age (above or below 75) and sex groups. These were used to create normed z-scores for other groups. We ran linear regressions to examine associations between normed hippocampal volume and objective memory [(Rey Auditory Verbal Learning test (RAVLT)] as well as subjective cognition [Everyday Cognition (ECOG)]. Separate models with raw volumes accounted for age and sex as covariates.

Results: Greater normed hippocampal volume was associated with greater RAVLT score in AD (p=.034) and MCI (p<.001) but not in SMC (p=0.96) (Bonferroni corrected). Raw hippocampal volume was  associated with RAVLT score only in MCI (p<.001) For the ECOG, none of the associations, neither norm or raw scores, were predictive of subjective cognitive decline.

Discussion: Normed hippocampal volume successfully predicted objective memory performance in early and late stages of AD, whereas raw volume predicted only at an early stage. Neither normed nor raw hippocampal volume predicted subjective cognitive function. Age- and sex-normed hippocampal volume may be a good index of AD disease stages. Subjective cognitive function may reflect atrophy or neurofunctional pathology of other brain regions.

19Neural asymmetry during memory encoding and its association with markers of preclinical Alzheimer’s disease.

Presented by Jinghang Li
PI: Helmet Karim, PhD

Introduction: Alzheimer’s Disease (AD) causes debilitating cognitive dysfunction and has a preclinical period where beta-amyloid (Aβ), a cytotoxic protein, is present without cognitive impairment. Asymmetric Aβ deposition is associated with asymmetric hypometabolism. Task-induced functional asymmetry and its association with AD risk factors during the preclinical period is still unknown. We investigated the association between AD risk factors, cognitive function, and activation asymmetry during a well-established memory encoding task.

Methods: We recruited cognitively normal participants for functional magnetic resonance imaging (fMRI), positron emission tomography (PET) imaging, and cognitive testing. We conducted voxel-wise analyses to identify regions of significant activation during associative memory encoding and then those with significant asymmetry. We examined the association between asymmetry and absolute asymmetry with age, sex, education, global cerebral amyloid, global cerebral metabolism, memory encoding task performance, white matter hyperintensities, and multiple domains of cognitive function.

Results: We identified expected regions of activation, including hippocampus, and identified regions with left (superior medial frontal gyrus, middle frontal gyrus, supplemental motor area, and medial orbitofrontal gyrus) and right putamen and ventral posterolateral nucleus of the thalamus) asymmetry. Greater left asymmetry in the middle frontal gyrus was correlated with greater global cerebral glucose metabolism. Better performance in memory, learning, and executive attention was associated with greater thalamus absolute symmetry, while better visuospatial performance was associated with greater putamen absolute symmetry. 

Discussion: Functional asymmetry may be an early indicator of functional change in older cognitively normal adults. Longitudinal studies may help us understand the causal impact of neural activation asymmetry.

20Low Thalamic Activity During a Computerized Digit Symbol Substitution Task in Subjective Cognitive Decline.

Presented by Akiko Mizuno, PhD
PI: Beth Snitz, PhD

Abstract
Self-perceived cognitive decline without objective cognitive impairment is known as subjective cognitive decline (SCD) and may represent the earliest detectable preclinical stage of Alzheimer’s Disease (AD). However, the underlying neurobiology of SCD is not completely understood. Since executive function may be affected earlier than memory function in the progression of AD, we aimed to characterize SCD symptoms in terms of fMRI brain activity during the digit-symbol substitution task (DSST), an executive function task. We also examined its association with DSST task performance, brain activation, SCD symptom severity, and amyloid-ß (Aß) load. We analyzed data from 63 cognitively normal older individuals (mean age 73.6 ± 7.2) with varying degree of SCD symptoms. Participants completed a computerized version of DSST in the MR scanner and a Pittsburgh Compound-B (PiB)-PET scan to measure global cerebral Aß load. We conducted regression analyses to investigate voxel-wise association between DSST activation and SCD symptom severity. Greater SCD symptom severity was associated with lower dorsomedial thalamus activation during DSST. Task performance was not associated with brain activation. Slower reaction time was associated with greater SCD symptom severity, but not with Aß load. The observed lower dorsomedial thalamus activation may reflect declining familiarity-based memory function and the trans-thalamic executive function pathway in SCD. SCD symptoms may reflect altered neural function and subtle decline of executive function, while Aß load may have an indirect impact on neural function and performance. Self-perceived cognitive decline (i.e., metacognitive awareness) may serve as a psychological/subjective marker reflecting subtle brain changes.

21Risk of dementia and trajectories of domain-specific cognition differ across lifestyle engagement subgroups of older adults.

Presented by Kyle Moored, PhD
PI: Michelle Carlson, PhD

Background: Characterizing shared lifestyle activities among older adults may both identify subgroups at risk of cognitive impairments and inform group-level interventions. We used latent class analysis to characterize distinct lifestyle engagement groups and examined whether they had differential risk of dementia and domain-specific cognitive declines.

Method: Participants were 3,068 individuals from the Ginkgo Evaluation of Memory Study who completed neuropsychological testing up to 8 years (M=5.0±2.4). We assessed 18 physical, intellectual, and social activities using the Lifestyle Activity Questionnaire. We used discrete-time proportional hazards and linear mixed effects models, adjusted for demographic and health covariates, to examine differences in dementia risk and cognition performance over time.

Result: In the four-class model, the Variety lifestyle group (22%) had high engagement in both intellectual (e.g., classes) and social activities (e.g., concerts, church). The Intellectual (18%) group had high engagement in intellectual/social leisure activities (e.g., concerts). The Social (32%) group had high engagement in social institutional activities (e.g., church). Compared to the Least Active (28%) group, the Variety and Intellectual groups had reduced dementia risk (HR=.62-.65, p’s<.05). The Variety group had higher baseline cognitive performance across domains (B’s=.14-.25) and attenuated declines in memory (B=.02±.01) and attention (B=.02±.01) over time (p’s<.05).

Conclusion: Higher lifestyle engagement, especially in intellectual activities, was associated with better cognition later in life. Yet, only those with the greatest diversity of activities (Variety group) were protected against declines in memory and attention, which likely drove protective associations with dementia. Subgroup differences in activity types may provide novel intervention targets to mitigate aging-related cognitive declines.

22A non-canonical GRK phosphorylation barcode regulates γ-secretase activity and amyloid-β generation.

Presented by Nicholas Todd, BS
PI: Amantha Thathiah, PhD

Abstract
The Aph1a subunit of the γ-secretase complex is required for complex assembly and processing of the amyloid precursor protein (APP) to generate the β-amyloid (Aβ) peptide. Aph1a is a 7-transmembrane domain protein that shares many structural similarities to G protein-coupled receptors (GPCRs). We previously identified a direct interaction between the GPCR scaffolding protein β-arrestin 2 (βarr2) and Aph1a and demonstrated a role of βarr2 in positively regulating γ-secretase activation and Aβ generation. GPCR kinases (GRKs) are responsible for phosphorylating residues on the carboxyl terminus (C-terminus) and intracellular loops (ICLs) of GPCRs to initiate β-arrestin binding and differential intracellular signaling outcomes. Interestingly, Aph1a contains putative phosphorylation sites within its C-terminus and ICL2. As such, we hypothesized that GRKs phosphorylate Aph1a to regulate βarr2 binding and downstream APP processing and Aβ generation. Here, we utilize label-free phosphorylation tandem mass spectrometry (LC-MS/MS) to determine that Aph1a is phosphorylated in cells and that loss of phosphorylation at one, or both, C-terminal serines hinders βarr2 binding and downstream Aβ generation. Further analysis suggests that GRK2, GRK3, and GRK5 are responsible for a phosphorylation barcode on Aph1a, which, in turn, regulate βarr2 binding and downstream Aβ generation. Finally, molecular dynamic simulation studies reveal the βarr2 finger loop region engages primarily with ICL3 and ICL2 of Aph1a to facilitate binding in a similar manner as a GPCR-β-arrestin complex. Overall, we present a novel model of a GRK-mediated phosphorylation barcode on Aph1a that facilitates distinct βarr2 binding conformations and differential downstream effects on γ-secretase activity and Aβ generation.