Search

Pancreatic ductal adenocarcinoma (PDAC) is an aggressive, painful disease with a 5-year survival rate of only 9%. Recent evidence indicates that distinct epigenomic landscapes underlie PDAC progression, identifying the H3K9me pathway as important to its pathobiology. Here, we delineate the role of Euchromatic Histone-lysine N-Methyltransferase 2 (EHMT2), the enzyme that generates H3K9me, as a downstream effector of oncogenic KRAS during PDAC initiation and pancreatitis-associated promotion. EHMT2 inactivation in pancreatic cells reduces H3K9me2 and antagonizes Kras(G12D)-mediated acinar-to-ductal metaplasia (ADM) and Pancreatic Intraepithelial Neoplasia (PanIN) formation in both the Pdx1-Cre and P48(Cre/+) Kras(G12D) mouse models. Ex vivo acinar explants also show impaired EGFR-KRAS-MAPK pathway-mediated ADM upon EHMT2 deletion. Notably, Kras(G12D) increases EHMT2 protein levels and EHMT2-EHMT1-WIZ complex formation. Transcriptome analysis reveals that EHMT2 inactivation upregulates a cell cycle inhibitory gene expression network that converges on the Cdkn1a/p21-Chek2 pathway. Congruently, pancreas tissue from Kras(G12D) animals with EHMT2 inactivation have increased P21 protein levels and enhanced senescence. Furthermore, loss of EHMT2 reduces inflammatory cell infiltration typically induced during Kras(G12D)-mediated initiation. The inhibitory effect on Kras(G12D)-induced growth is maintained in the pancreatitis-accelerated model, while simultaneously modifying immunoregulatory gene networks that also contribute to carcinogenesis. This study outlines the existence of a novel KRAS-EHMT2 pathway that is critical for mediating the growth-promoting and immunoregulatory effects of this oncogene in vivo, extending human observations to support a pathophysiological role for the H3K9me pathway in PDAC.

This dissertation explores the effects of heterogeneity across different biological scales in cancer as well as normal cells. At the tissue scale, we investigated the variability present in the tumor microenvironment and its effect on patient chemotherapeutic outcomes using a mathematical model of drug transport. We found that parameters such as tumor blood perfusion and radius of blood vessel had an impact on the tumor cytotoxicity. This indicated that the physical microenvironment of the tumor is an important regulator of the tumor response to chemotherapy. At the cellular scale, we investigated the heterogeneity present on the membrane landscape of ErbB2 and ErbB3, two receptors that are upregulated in cancer, using a spatial stochastic model of receptor dimerization and phosphorylation. We found that membrane domains played an important role in regulating signaling emanating from this receptor dimer. In our next study, we developed a 3-D spatial stochastic model of pre-BCR, a receptor which is crucial in the development of B lymphocytes and also upregulated in a subset of patients with B-Cell Precursor Acute Lymphoblastic Leukemia, to investigate the effects of ligand independent (tonic signaling) originating from this receptor. We populated our model with single particle tracking data from two different leukemic cell lines which had different dimer off rates and diffusion coefficients, along with experimental measurements. Other important signaling molecules such as Lyn and Syk, which are active in this pathway, were also included in the model. We found that the variability in characteristics between the two cell lines led to differences in downstream signaling events from the receptor. The cell line with the lower dimer off rate formed higher order oligomers and had more overall molecule phosphorylation compared to the other. Thus, this spatial stochastic model was able to shed light on threshold signaling events which take place during tonic signaling.