Pancreatic cancer : investigation of the prognostic significance of tumor immunophenotype and establishment of a novel ex-vivo tissue slice culture system for drug sensitivity testing

Abstract

Pancreatic ductal adenocarcinoma (PDAC), commonly referred to as pancreatic cancer, is associated with a dismal 5-year overall survival of 9 to 6 %. A major reason for the high mortality is the pronounced resistance to treatment. There is urgent need for the development of novel, effective therapies and for a better understanding as to why some tumors respond better to treatment than others. The process for conventional drug testing relies significantly on cell lines and genetically engineered mouse models. Even if these models recapitulate some of the features of human pancreatic cancer, there are significant limitations, mainly the lack of the native tumor microenvironment and the use of a host that belongs to another species and may be immunocompromised.

In Paper I, a large cohort (n = 409) of adenocarcinomas from the main anatomical locations in the pancreatobiliary system is analyzed by immunohistochemistry with a panel of up to 27 antibodies. Hierarchical clustering and differential expression analysis reveal three immunohistochemical tumor types (extrahepatic pancreatobiliary, intestinal and intrahepatic cholangiocarcinoma) and their distinguishing markers. Among patients who underwent surgical resection of the primary tumor, the intestinal type showed an adjusted hazard ratio of 0.19 for overall survival (p value = 0.014) as compared to the extrahepatic pancreatobiliary type. Furthermore, the characteristic immunohistochemical profile supports the positive diagnosis of intrahepatic cholangiocarcinoma, which is commonly regarded as a diagnosis of exclusion. The presented integrative immunohistochemical classification may contribute to improve diagnosis and prognostic stratification for patients with adenocarcinomas of the pancreatobiliary system.

Paper II introduces a novel ex-vivo model of precision-cut tissue slices that allows culturing of human PDAC including the native tumor microenvironment. Fresh tumor tissue samples (n = 12) were obtained from surgical resection specimens, cut into 350 μm thick tissue slices, and cultured for at least 96 h. Every 24 h, tissue slices were harvested and processed for analysis, including histomorphology, transmission electron microscopy and immunohistochemistry. These revealed good morphological preservation and ultrastructural integrity of the tissue slices, including cancerous, stromal, and immune cell populations. Each tumor retained its characteristic histological and cytological features and grade of differentiation throughout the entire culture period. A cancerous and, to a lesser extent, nonneoplastic cell outgrowth covered gradually the surface of the tissue slices, while tumor cells retained their proliferative (Ki67) and metabolic (pS6 mTOR pathway) activity. This culture system is a close surrogate of the parent tumor and harbors promising potential for drug sensitivity testing and personalized treatment of PDAC.

Paper III is a continuation study in which five tissue slice cultures of human PDAC (from Paper II) are analyzed at the whole transcriptome level (RNASeq). Findings at 24 h intervals (24-72 h of culture) are compared to baseline (0 h, non-cultured tissue). On differential expression analysis, only a limited number of genes (median range 10-25) were up- or downregulated during culture. One culture with morphologically visible regions of necrosis/apoptosis (0-18 % total slice area) showed upregulation of VEGFA and PTGS2. Pathway analysis suggested for this culture a highly significant probability of activation of apoptosis via HIF-1/quercetin/NF-ĸB/AP-1 interaction pathways. These results support that the transcriptome of the parent tumor is largely preserved in ex-vivo cultured tissue slices of PDAC, and that transcriptomic analysis is a valuable complement to morphology for evaluation of the tissue slices.

Finally, in Paper IV, the previously established (in Papers II and III) ex-vivo model of precision-cut tissue slices is used for drug sensitivity testing in nine samples of surgically resected human PDAC. PDAC tissue slices were exposed to selenium compounds (selenite and methylselenocysteine) at various concentrations and gemcitabine 1 μM during 48 h of culture. Selenium compounds administered at concentrations below the maximum tolerated dose in humans significantly reduced PDAC cell viability (p < 0.02) and decreased outgrowth of viable tumor cells onto the tissue surface (p < 0.05), while non-neoplastic tissues remained preserved. Transcriptomic analysis revealed downregulation of CEMIP, PLOD2, DDR2 and P4HA1, genes that are involved in extracellular matrix modulation, cancer growth and metastatic potential, while the cell death-inducing genes ATF3 and ACHE were significantly upregulated (p < 0.0001).

In conclusion, the results of this thesis support the value of immunohistochemical profiling for the diagnosis of prognostically significant subtypes of periampullary adenocarcinomas. The findings confirm the relevance of the precision-cut tissue slice culture model for anticancer drug sensitivity testing and the potential of selenium compounds as candidate drugs for the treatment of PDAC.