Unraveling the Molecular Underpinnings of Premalignant Oral Lesion Transformation into Head and Neck Cancer

Division Of Basic Sciences - Nci

Investigators

Abstract

In fiscal year 2025, our research has made significant strides in advancing the understanding of head and neck cancer (HNC) through several innovative approaches. We have achieved notable progress in several key areas: 1) Understanding Rare Premalignant Oral Lesions through Spatial Transcriptomics: We have generated one of the largest sets of Spatial Transcriptomics datasets (n=11) for premalignant oral lesions, for which we have already drafted manuscripts ready to be submitted. These new datasets allowed for our team (Nguyen Lab, Molloy Lab, and Sultan Lab) to pioneer new batch correction and data integration techniques allowing for us to gain novel insight into this disease process. More importantly, we have identified that even if tissues looked histologically "normal", through cluster analyses we have determined that normal epithelium, dysplastic epithelium, and squamous cell carcinoma have unique expression profiles. More interestingly is that we have preliminary evidence that immune populations surround early dysplastic cells, suggesting a common mechanism prior to cancer transformation. We have drafted a manuscript for which we intend to submit to a top tier journal. 2) Analyzing Patient-Paired Premalignant and Malignant Biopsies for Whole Exome Sequencing: We have generated the largest dataset of paired biopsy and tumor resection materials for patient matched premalignant and malignant samples. One of the key questions in our field is how the rate of tumor mutagenesis occurs and what are the key mutations that generate neoantigens. Better understanding the rate of neoantigen formation may give us better insight predicting the rate of malignant transformation for patients and possible patient response rates. The comprehensive analysis of patient-paired premalignant and malignant biopsies using Whole Exome Sequencing represents an important and transformative approach in the field of head and neck oncology. This methodology provides insights into the early molecular events of cancer initiation, sheds light on the intricate dynamics of clonal evolution, and unravels the complex mechanisms driving malignant transformation. Better understanding these findings will facilitate the rational design of personalized targeted therapies and immunotherapies, particularly through the identification of tumor-specific neoantigens. Furthermore, this approach empowers future clinicians with the ability to dynamically monitor tumor evolution and adapt treatment strategies in real-time, effectively overcoming the challenges posed by intratumoral heterogeneity and the emergence of resistance mechanisms. 3) With the support of Maggie Cam and Richard Finney, and our UMB collaborators we have developed and refined Artificial Intelligence models that are capable of over 90% accuracy of predicting malignant transformation in a curated dataset of histological samples of premalignant and malignant head and neck cancer. These novel AI analyses enhance diagnostic precision and predict the progression of premalignant oral lesions (PMOLs) to HNC. Our DL models, trained on comprehensive datasets that include clinical, histopathological, and molecular data, have shown promising results in identifying key histopathological patterns indicative of transformation, thus enabling earlier and more accurate intervention strategies. Despite challenges such as data variability and the complexity of integrating multiomic data, our efforts have moved us closer to developing an AI-driven diagnostic tool for HNC, with the potential to greatly improve patient outcomes and reduce HNC mortality rates. 4) Murine Models and mTORC3 Signaling: we have utilized 4NQO-induced murine models to simulate tobacco-related carcinogenesis, closely mirroring human HNC conditions. These models were genetically engineered to delete meak7, a pivotal component of the mTORC3 signaling complex. This experimental setup has allowed us to investigate how the absence of mTORC3 affects cancer progression, tumor burden, survival, and the tumor-immune microenvironment. By employing single-cell RNA sequencing, we are gaining valuable insights into the role of mTORC3 in HNC and identifying potential biomarkers and therapeutic targets that could be pivotal for future treatments. 5) Generation of "mCas9-OSCC" cell lines, Murine Cas9+ 4NQO-generated Oral Squamous Cell Carcinoma Lines: we have developed 2x oral squamous cell carcinoma (OSCC) cell lines from the base of the tongue in 4NQO-treated mice, engineered to express Cas9 for CRISPR-mediated gene editing, mCas9-OSCC-1/2. We have conducted Whole Genome Sequencing, Methylation Sequencing, and validated that these cells have a high rate of in vitro cell growth and in vivo (immunocompetent and immunocompromised mice) growth. Additionally, our mice have been shown to metastasize in vivo. For future studies, we are performing a focused CRISPR screens to uncover gene networks and regulatory pathways crucial for evading the immune checkpoint inhibitor response. Through these methodologies, the Nguyen Lab aims to significantly enhance our understanding of HNC's molecular mechanisms and contribute to the development of more effective, personalized diagnostic and treatment strategies. These advancements are poised to offer breakthroughs in early detection, targeted therapy, and ultimately, improvements in patient outcomes.