Faculty Bibliography

Methylation quantitative trait loci (meQTLs) quantify the effects of genetic variants on DNA methylation levels. However, most published studies utilize bulk methylation datasets composed of different cell types and limit our understanding of cell-type-specific methylation regulation. We propose a hierarchical Bayesian interaction (HBI) model to infer cell-type-specific meQTLs, which integrates a large-scale bulk methylation data and a small-scale cell-type-specific methylation data. Through simulations, we show that HBI enhances the estimation of cell-type-specific meQTLs. In real data analyses, we demonstrate that HBI can further improve the functional annotation of genetic variants and identify biologically relevant cell types for complex traits.

Oral squamous cell carcinoma (OSCC) is a highly unpredictable disease with devastating mortality rates that have not changed over the past decades, in the face of advancements in treatments and biomarkers, which have improved survival for other cancers. Delays in diagnosis are frequent, leading to more disfiguring treatments and poor outcomes for patients. The clinical challenge lies in identifying those patients at the highest risk of developing OSCC. Oral epithelial dysplasia (OED) is a precursor of OSCC with highly variable behavior across patients. There is no reliable clinical, pathological, histological, or molecular biomarker to determine individual risk in OED patients. Similarly, there are no robust biomarkers to predict treatment outcomes or mortality in OSCC patients. This review aims to highlight advancements in artificial intelligence (AI)-based methods to develop predictive biomarkers of OED transformation to OSCC or predictive biomarkers of OSCC mortality and treatment response. Biomarkers such as S100A7 demonstrate promising appraisal for the risk of malignant transformation of OED. Machine learning-enhanced multiplex immunohistochemistry workflows examine immune cell patterns and organization within the tumor immune microenvironment to generate outcome predictions in immunotherapy. Deep learning (DL) is an AI-based method using an extended neural network or related architecture with multiple "hidden" layers of simulated neurons to combine simple visual features into complex patterns. DL-based digital pathology is currently being developed to assess OED and OSCC outcomes. The integration of machine learning in epigenomics aims to examine the epigenetic modification of diseases and improve our ability to detect, classify, and predict outcomes associated with epigenetic marks. Collectively, these tools showcase promising advancements in discovery and technology, which may provide a potential solution to addressing the current limitations in predicting OED transformation and OSCC behavior, both of which are clinical challenges that must be addressed in order to improve OSCC survival.

Head and neck squamous cell carcinoma (HNSCC) is painful, and perineural invasion (PNI) has been associated with the worst pain. Pain due to HNSCC is diverse and may vary based on clinicopathological factors. This study aims to characterize different pain patterns linked with PNI, its influence on daily functioning, and gain insights into molecular changes and pathways associated with PNI-related pain in HNSCC patients. We conducted a cross-sectional study across 3 medical centers (n = 114), assessing pain phenotypes and their impact on daily functioning using 2 self-reported pain questionnaires, given to patients prior to their cancer surgery. Furthermore, we conducted RNA-seq analysis utilizing the The Cancer Genome Atlas dataset of HNSCC tumor from patients (n = 192) to identify genes relevant to both PNI and pain. Upon adjusting for demographic and clinicopathological variables using linear regression models, we found that PNI independently predicted function-evoked pain according to the University of Calfornia San Francisco Oral Cancer Pain Questionnaire, as well as the worst pain intensity reported in the Brief Pain Inventory. Distinct pain patterns were observed to be associated with daily activities in varying manners. Our molecular analyses revealed significant disruptions in pathways associated with the extracellular matrix structure and organization. The top differentially expressed genes linked to the extracellular matrix are implicated in cancer development, pain, and neurodegenerative diseases. Our data underscore the importance of properly categorizing pain phenotypes in future studies aiming to uncover mechanistic underpinnings of pain. Additionally, we have compiled a list of genes of interest that could serve as targets for both cancer and cancer pain management. PERSPECTIVE: PNI independently predicts function-evoked pain. Different pain phenotypes affect daily activities differently. We identified a list of candidate genes involved in the extracellular matrix structure and function that can be targeted for both cancer and cancer pain control.

Sexual and gender minority (SGM) communities are underrepresented in biomedical studies, highlighting the importance of developing biospecimen collection protocols aimed at engaging SGM participants. We aimed to learn more about SGM participants' experiences with a remote (i.e., not performed at a central location) biospecimen collection study pioneered by The PRIDE Study, a cohort study of SGM adults residing in the United States and its territories. Feedback was collected from 112 SGM participants following blood donation for a parent study investigating the relationship between minority stress, substance use, and epigenetic markers of substance use and minority stress. We used an inductive and collaborative approach to qualitative analysis and identified major themes and areas for protocol improvement. Major themes among participant feedback were: (1) communication with the research team, (2) convenience of donation, (3) interactions with clinical laboratory staff, and (4) anonymity and privacy. Most participants indicated that they experienced little to no problems during the donation process and expressed approval for the clarity and transparency of the informed consent process, ease of communication with the research team, and measures taken to protect participant confidentiality during their appointment. The most common challenges encountered by participants related to the inconvenience of handling and transporting study materials to the clinical laboratory site and clinical laboratory staff's unfamiliarity with the study protocol. Some participants indicated a preference for more elements of the study protocol (e.g., transporting collection materials) to be left to the responsibility of the research team. Future studies should carefully consider the delegation of responsibility between participants and the research team to balance both study reach and participant accessibility. Alternative formats, such as at-home collection or collaboration with community health workers, may further enhance participant satisfaction and convenience.

OBJECTIVE:This epigenomics sub-study embedded within a randomized controlled trial examined whether an evidenced-based behavioral intervention model that decreased stimulant use altered leukocyte DNA methylation (DNAm). METHODS:Sexual minority men with HIV who use methamphetamine were randomized to a five-session positive affect intervention (n = 32) or an attention-control condition (n = 21), both delivered during three months of contingency management for stimulant abstinence. All participants exhibited sustained HIV virologic control - an HIV viral load less than 40 copies/mL at baseline and six months post-randomization. The Illumina EPIC BeadChip measured leukocyte methylation of cytosine-phosphate-guanosine (CpG) sites mapping onto five a priori candidate genes of interest (i.e., ADRB2, BDNF, FKBP5, NR3C1, OXTR). Functional DNAm pathways and soluble markers of immune dysfunction were secondary outcomes. RESULTS: < 0.05) revealed significant intervention-related alterations in DNAm of Reactome pathways corresponding to neural function as well as dopamine, glutamate, and serotonin release. Positive affect intervention effects on DNAm were accompanied by significant reductions in the self-reported frequency of stimulant use. CONCLUSIONS:There is an epigenetic signature of an evidence-based behavioral intervention model that reduced stimulant use, which will guide the identification of biomarkers for treatment responses.

AIMS/OBJECTIVE:The purpose of this study was to assess the levels of cardiovascular health (CVH) of Black and Latino adults with type 2 diabetes (T2D) and examine the association of individual and microsystem level factors with their CVH score. METHODS:This was a cross-sectional design in 60 Black and Latino Adults aged 18-40 with T2D. Data were collected on sociodemographic, individual (sociodemographic, diabetes self-management, sleep disturbance, depressive symptoms, quality of life, and the inflammatory biomarkers IL-6 and hs-CRP) and microsystem factors (family functioning), and American Heart Association's Life's Simple 7 metrics of CVH. Factors significantly associated with the CVH score in the bivariate analyses were entered into a linear regression model. RESULTS:= .0013). CONCLUSIONS:This sample had an intermediate level of CVH, with inflammation and general health associated with overall CVH score.

OBJECTIVE:Inflammation is implicated in the pathophysiology of depression and type 2 diabetes (T2D) and is linked to social determinants of health (SDoH) associated with socioeconomic disadvantage. The objective of this review is to identify and map the range of SDoHs associated with inflammation in depression, T2D, or their co-occurrence among women. METHODS:PubMed, CINAHL, PsychINFO, and Web of Science were searched March-July 2023 to identify studies where 1) an SDoH was a predictor or independent variable, 2) depression or T2D was a clinical focus, 3) inflammatory markers were collected, and 4) analysis was specific to women. We used the National Institute on Minority Health and Health Disparities research framework to guide searching SDoHs, organize findings, and identify gaps. RESULTS:Of the 1135 studies retrieved, 46 met criteria. Within the reviewed studies, the most used inflammatory measures were C-reactive protein, interleukin-6, and tumor necrosis factor-α, and the most studied SDoHs were early life stress and socioeconomic status. Individual and interpersonal-level variables comprised the bulk of SDoHs in the included studies, while few to no studies examined built environment (n = 6) or health system level (n = 0) factors. Disadvantageous SDoHs were associated with higher levels of inflammation across the included studies. CONCLUSION/CONCLUSIONS:The scope and intersection of depression and T2D represent a syndemic that contributes to and results from socioeconomic inequities and disproportionately affects women. Simultaneous inclusion of social and inflammatory measures, particularly understudied SDoHs, is needed to clarify potent targets aimed at advancing health and equity.

Oral squamous cell carcinoma (OSCC) biomarker studies rarely employ multi-omic biomarker strategies and pertinent clinicopathologic characteristics to predict mortality. In this study we determine for the first time a combined epigenetic, gene expression, and histology signature that differentiates between patients with different tobacco use history (heavy tobacco use with ≥10 pack years vs. no tobacco use). Using The Cancer Genome Atlas (TCGA) cohort (n = 257) and an internal cohort (n = 40), we identify 3 epigenetic markers (GPR15, GNG12, GDNF) and 13 expression markers (IGHA2, SCG5, RPL3L, NTRK1, CD96, BMP6, TFPI2, EFEMP2, RYR3, DMTN, GPD2, BAALC, and FMO3), which are dysregulated in OSCC patients who were never smokers vs. those who have a ≥ 10 pack year history. While mortality risk prediction based on smoking status and clinicopathologic covariates alone is inaccurate (c-statistic = 0.57), the combined epigenetic/expression and histologic signature has a c-statistic = 0.9409 in predicting 5-year mortality in OSCC patients.