Faculty Bibliography

In recent years, there has been a surge of interest in using machine learning algorithms (MLAs) in oncology, particularly for biomedical applications such as drug discovery, drug repurposing, diagnostics, clinical trial design, and pharmaceutical production. MLAs have the potential to provide valuable insights and predictions in these areas by representing both the disease state and the therapeutic agents used to treat it. To fully utilize the capabilities of MLAs in oncology, it is important to understand the fundamental concepts underlying these algorithms and how they can be applied to assess the efficacy and toxicity of therapeutics. In this perspective, we lay out approaches to represent both the disease state and the therapeutic agents used by MLAs to derive novel insights and make relevant predictions.

Lung cancer treatment has benefited greatly through advancements in immunotherapies. However, immunotherapy often fails in patients with specific mutations like KEAP1, which are frequently found in lung adenocarcinoma. We established an antigenic lung cancer model and used it to explore how Keap1 mutations remodel the tumor immune microenvironment. Using single-cell technology and depletion studies, we demonstrate that Keap1-mutant tumors diminish dendritic cell and T cell responses driving immunotherapy resistance. This observation was corroborated in patient samples. CRISPR-Cas9-mediated gene targeting revealed that hyperactivation of the NRF2 antioxidant pathway is responsible for diminished immune responses in Keap1-mutant tumors. Importantly, we demonstrate that combining glutaminase inhibition with immune checkpoint blockade can reverse immunosuppression, making Keap1-mutant tumors susceptible to immunotherapy. Our study provides new insight into the role of KEAP1 mutations in immune evasion, paving the way for novel immune-based therapeutic strategies for KEAP1-mutant cancers.

We introduce a pioneering approach that integrates pathology imaging with transcriptomics and proteomics to identify predictive histology features associated with critical clinical outcomes in cancer. We utilize 2,755 H&E-stained histopathological slides from 657 patients across 6 cancer types from CPTAC. Our models effectively recapitulate distinctions readily made by human pathologists: tumor vs. normal (AUROC = 0.995) and tissue-of-origin (AUROC = 0.979). We further investigate predictive power on tasks not normally performed from H&E alone, including TP53 prediction and pathologic stage. Importantly, we describe predictive morphologies not previously utilized in a clinical setting. The incorporation of transcriptomics and proteomics identifies pathway-level signatures and cellular processes driving predictive histology features. Model generalizability and interpretability is confirmed using TCGA. We propose a classification system for these tasks, and suggest potential clinical applications for this integrated human and machine learning approach. A publicly available web-based platform implements these models.

Skeletal stem and progenitor cells (SSPCs) perform bone maintenance and repair. With age, they produce fewer osteoblasts and more adipocytes leading to a loss of skeletal integrity. The molecular mechanisms that underlie this detrimental transformation are largely unknown. Single-cell RNA sequencing revealed that Notch signaling becomes elevated in SSPCs during aging. To examine the role of increased Notch activity, we deleted Nicastrin, an essential Notch pathway component, in SSPCs in vivo. Middle-aged conditional knockout mice displayed elevated SSPC osteo-lineage gene expression, increased trabecular bone mass, reduced bone marrow adiposity, and enhanced bone repair. Thus, Notch regulates SSPC cell fate decisions, and moderating Notch signaling ameliorates the skeletal aging phenotype, increasing bone mass even beyond that of young mice. Finally, we identified the transcription factor Ebf3 as a downstream mediator of Notch signaling in SSPCs that is dysregulated with aging, highlighting it as a promising therapeutic target to rejuvenate the aged skeleton.

Approximately 30% of early-stage lung adenocarcinoma patients present with disease progression after successful surgical resection. Despite efforts of mapping the genetic landscape, there has been limited success in discovering predictive biomarkers of disease outcomes. Here we performed a systematic multi-omic assessment of 143 tumors and matched tumor-adjacent, histologically-normal lung tissue with long-term patient follow-up. Through histologic, mutational, and transcriptomic profiling of tumor and adjacent-normal tissue, we identified an inflammatory gene signature in tumor-adjacent tissue as the strongest clinical predictor of disease progression. Single-cell transcriptomic analysis demonstrated the progression-associated inflammatory signature was expressed in both immune and non-immune cells, and cell type-specific profiling in monocytes further improved outcome predictions. Additional analyses of tumor-adjacent transcriptomic data from The Cancer Genome Atlas validated the association of the inflammatory signature with worse outcomes across cancers. Collectively, our study suggests that molecular profiling of tumor-adjacent tissue can identify patients at high risk for disease progression.

BACKGROUND/UNASSIGNED:Conotruncal defects due to developmental abnormalities of the outflow tract (OFT) are an important cause of cyanotic congenital heart disease. Dysregulation of transcriptional programs tuned by NKX2-5 (NK2 homeobox 5), GATA6 (GATA binding protein 6), and TBX1 (T-box transcription factor 1) have been implicated in abnormal OFT morphogenesis. However, there remains no consensus on how these transcriptional programs function in a unified gene regulatory network within the OFT. METHODS/UNASSIGNED: RESULTS/UNASSIGNED: CONCLUSIONS/UNASSIGNED:Our results using human and mouse models reveal an essential gene regulatory network of the OFT that requires an anterior second heart field enhancer to link GATA6 with NKX2-5-dependent rotation and septation gene programs.

PURPOSE/OBJECTIVE:Chromosome 1 (chr1) copy number abnormalities (CNAs) and structural variants (SV) are frequent in newly diagnosed multiple myeloma (NDMM) and associate with a heterogeneous impact on outcome the drivers of which are largely unknown. EXPERIMENTAL DESIGN/METHODS:A multiomic approach comprising CRISPR, gene mapping of CNA and SV, methylation, expression, and mutational analysis was used to document the extent of chr1 molecular variants and their impact on pathway utilisation. RESULTS:We identified two distinct groups of gain(1q): focal gains associated with limited gene expression changes and a neutral prognosis, and whole-arm gains, which associate with substantial gene expression changes, complex genetics and an adverse prognosis. CRISPR identified a number of dependencies on chr1 but only limited variants associated with acquired CNAs. We identified seven regions of deletion, nine of gain, three of chromothripsis (CT) and two of templated-insertion (TI), which contain a number of potential drivers. An additional mechanism involving hypomethylation of genes at 1q may contribute to the aberrant gene expression of a number of genes. Expression changes associated with whole-arm gains were substantial and gene set enrichment analysis identified metabolic processes, apoptotic resistance, signaling via the MAPK pathway, and upregulation of transcription factors as being key drivers of the adverse prognosis associated with these variants. CONCLUSIONS:Multiple layers of genetic complexity impact the phenotype associated with CNAs on chr1 to generate its associated clinical phenotype. Whole-arm gains of 1q are the critically important prognostic group that deregulate multiple pathways, which may offer therapeutic vulnerabilities.

BACKGROUND/UNASSIGNED:Central nervous system (CNS) cancer is the 10th leading cause of cancer-associated deaths for adults, but the leading cause in pediatric patients and young adults. The variety and complexity of histologic subtypes can lead to diagnostic errors. DNA methylation is an epigenetic modification that provides a tumor type-specific signature that can be used for diagnosis. METHODS/UNASSIGNED:We performed a prospective study using DNA methylation analysis as a primary diagnostic method for 1921 brain tumors. All tumors received a pathology diagnosis and profiling by whole genome DNA methylation, followed by next-generation DNA and RNA sequencing. Results were stratified by concordance between DNA methylation and histopathology, establishing diagnostic utility. RESULTS/UNASSIGNED:Of the 1602 cases with a World Health Organization histologic diagnosis, DNA methylation identified a diagnostic mismatch in 225 cases (14%), 78 cases (5%) did not classify with any class, and in an additional 110 (7%) cases DNA methylation confirmed the diagnosis and provided prognostic information. Of 319 cases carrying 195 different descriptive histologic diagnoses, DNA methylation provided a definitive diagnosis in 273 (86%) cases, separated them into 55 methylation classes, and changed the grading in 58 (18%) cases. CONCLUSIONS/UNASSIGNED:DNA methylation analysis is a robust method to diagnose primary CNS tumors, improving diagnostic accuracy, decreasing diagnostic errors and inconclusive diagnoses, and providing prognostic subclassification. This study provides a framework for inclusion of DNA methylation profiling as a primary molecular diagnostic test into professional guidelines for CNS tumors. The benefits include increased diagnostic accuracy, improved patient management, and refinements in clinical trial design.

The chromosome 9p21 locus comprises several tumor suppressor genes including MTAP, CDKN2A and CDKN2B, and its homo- or heterozygous deletion is associated with reduced survival in multiple cancer types. We report that mice with germline monoallelic deletion or induced biallelic deletion of the 9p21-syntenic locus (9p21s) developed a fatal myelodysplastic syndrome/myeloproliferative neoplasm (MDS/MPN)-like disease associated with aberrant trabecular bone formation and/or fibrosis in the bone marrow (BM). Reciprocal BM transfers and conditional targeting of 9p21s suggested that the disease originates in the BM stroma. Single-cell analysis of 9p21s-deficient BM stroma revealed the expansion of chondrocyte and osteogenic precursors, reflected in increased osteogenic differentiation in vitro. It also showed reduced expression of factors maintaining hematopoietic stem/progenitor cells, including Cxcl12. Accordingly, 9p21s-deficient mice showed reduced levels of circulating Cxcl12 and concomitant upregulation of the pro-fibrotic chemokine Cxcl13 and osteogenesis- and fibrosis-related multifunctional glycoprotein Osteopontin (OPN)/Spp1. Our study highlights the potential of mutations in the BM microenvironment to drive MDS/MPN-like disease.