Background: Numerous immune genes are epigenetically silenced in tumor cells and agents such as histone deacetylase inhibitors (HDACi), which reverse these effects, could potentially be used to develop therapeutic vaccines. The conversion of cancer cells to antigen presenting cells (APCs) by HDACi treatment could potentially provide an additional pathway, together with cross-presentation of tumor antigens by host APCs, to establish tumor immunity. Methods: HDACi-treated B16 melanoma cells were used in a murine vaccine model, lymphocyte subset depletion, ELISpot and Cytotoxicity assays were employed to evaluate immunity. Antigen presentation assays, vaccination with isolated apoptotic preparations and tumorigenesis in MHC-deficient mice and radiation chimeras were performed to elucidate the mechanisms of vaccine-induced immunity. Results: HDACi treatment enhanced the expression of MHC class II, CD40 and B7-1/2 on B16 cells and vaccination with HDACi-treated melanoma cells elicited tumor specific immunity in both prevention and treatment models. Cytotoxic and IFN-γ-producing cells were identified in splenocytes and CD4⁺, CD8⁺ T cells and NK cells were all involved in the induction of immunity. Apoptotic cells derived from HDACi treatments, but not H₂O₂, significantly enhanced the effectiveness of the vaccine. HDACi-treated B16 cells become APCs in vitro and studies in chimeras defective in cross presentation demonstrate direct presentation in vivo and short-term but not memory responses and long-term immunity. Conclusion: The efficacy of this vaccine derives mainly from cross-presentation which is enhanced by HDACi-induced apoptosis. Additionally, epigenetic activation of immune genes may contribute to direct antigen presentation by tumor cells. Epigenetically altered cancer cells should be further explored as a vaccine strategy.

The immune cell niche associated with oral dysplastic lesion progression to carcinoma is poorly understood. We identified T regulatory cells (Treg), CD8⁺ effector T cells (Teff) and immune checkpoint molecules across oral dysplastic stages of oral potentially malignant disorders (OPMD). OPMD and oral squamous cell carcinoma (OSCC) tissue sections (N = 270) were analyzed by immunohistochemistry for Treg (CD4, CD25 and FoxP3), Teff (CD8) and immune checkpoint molecules (PD-1 and PD-L1). The Treg marker staining intensity correlated significantly (p < 0.01) with presence of higher dysplasia grade and invasive cancer. These data suggest that Treg infiltration is relatively early in dysplasia and may be associated with disease progression. The presence of CD8⁺ effector T cells and the immune checkpoint markers PD-1 and PD-L1 were also associated with oral cancer progression (p < 0.01). These observations indicate the induction of an adaptive immune response with similar Treg and Teff recruitment timing and, potentially, the early induction of exhaustion. FoxP3 and PD-L1 levels were closely correlated with CD8 levels (p < 0.01). These data indicate the presence of reinforcing mechanisms contributing to the immune suppressive niche in high-risk OPMD and in OSCC. The presence of an adaptive immune response and T-cell exhaustion suggest that an effective immune response may be reactivated with targeted interventions coupled with immune checkpoint inhibition.

Oral potentially malignant disorders (OPMDs) are precursors to over 80% of oral cancers. Hematoxylin and eosin (H&E) staining, followed by pathologist interpretation of tissue and cellular morphology, is the current gold standard for diagnosis. However, this method is qualitative, can result in errors during the multi-step diagnostic process, and results may have significant inter-observer variability. Chemical imaging (CI) offers a promising alternative, wherein label-free imaging is used to record both the morphology and the composition of tissue and artificial intelligence (AI) is used to objectively assign histologic information. Here, we employ quantum cascade laser (QCL)-based discrete frequency infrared (DFIR) chemical imaging to record data from oral tissues. In this proof-of-concept study, we focused on achieving tissue segmentation into three classes (connective tissue, dysplastic epithelium, and normal epithelium) using a convolutional neural network (CNN) applied to three bands of label-free DFIR data with paired darkfield visible imaging. Using pathologist-annotated H&E images as the ground truth, we demonstrate results that are 94.5% accurate with the ground truth using combined information from IR and darkfield microscopy in a deep learning framework. This chemical-imaging-based workflow for OPMD classification has the potential to enhance the efficiency and accuracy of clinical oral precancer diagnosis.