Background: Among the common malignancies, pancreatic cancer has the shortest long-term survival. The aggressive, rapid, and infiltrative growth pattern of pancreatic cancer, together with the lack of specific symptoms, often leads to late diagnosis. Metastases are frequently found at the time of diagnosis, which prevents curative surgical treatment. Good tumor markers would enable early detection, thus improving the prognosis. Unfortunately, no such markers are available in the clinic. The tumor stroma is defined as the non-malignant cells and the extracellular matrix (ECM) of a cancer. Pancreatic cancer is characterized by an abundant tumor stroma, rich in ECM proteins such as collagens, which have been shown to play important roles in tumor progression. Furthermore, pancreatic cancer cells produce large quantities of ECM proteins, especially the basement membrane (BM) protein type IV collagen. All epithelial cells are anchored to a BM, which must be degraded in order for an in situ cancer to become invasive. Matrix metalloproteinases (MMPs) are enzymes involved in BM degradation. In this thesis, the tumor stroma of pancreatic cancer is studied, focusing on the BM proteins type IV and type XVIII collagen, with the aim to clarify if the stroma could be a source of novel tumor markers for this form of cancer. Additionally, the role of type IV collagen produced by the cancer cells is studied. Methods: Expression patterns of type IV and type XVIII collagen, MMPs involved in collagen degradation, and collagen receptors (integrins) were studied by immunoflourescence in both normal and pancreatic cancer tissue, and in pancreatic cancer cell lines. Circulating plasma levels of type IV and type XVIII collagen and conventional tumor markers (TPS, Ca 19-9, CEA and Ca 125) were measured in controls and pancreatic cancer patients at the time of diagnosis and after treatment. The role of cancer cell produced type IV collagen was studied in human pancreatic cancer cell lines by functional blocking of integrin receptors (integrin a1, a2 and b1) and integrin-binding sites on type IV collagen, and by siRNA-induced down-regulation of type IV collagen synthesis. Proliferation was analyzed by a luminescence based cell viability assay, migration by time-lapse microscopy, and apoptosis by M30-neoepitope detection. Results: MMPs involved in BM degradation were upregulated in pancreatic cancer tissue. The expression of type XVIII collagen shifted from a general BM expression pattern in normal tissue, to mainly being found in the tumor vasculature in pancreatic cancer. Type IV collagen, on the other hand, remained highly expressed in the vicinity of the cancer cells. The a1, a2, and b1 integrin receptors were highly expressed at the cancer cell surface. Both down-regulation of type IV collagen synthesis and blocking the integrin/type IV collagen interaction decreased cell proliferation and migration. The proliferative capacity was rescued by the addition of exogenous type IV collagen. Furthermore, the circulating levels of both type IV and type XVIII collagen were increased in pancreatic cancer patients at the time of diagnosis compared to controls. After treatment, the levels were normalized for type XVIII collagen, whereas the levels of type IV collagen remained high after surgery. High postoperative levels of type IV collagen were associated with short overall survival. A similar association to short survival was found for preoperative type XVIII collagen levels. No such associations to survival could be detected for the conventional markers.   Conclusion: The results of this thesis show that type IV and type XVIII collagens can serve as tumor markers for pancreatic cancer with advantages compared to conventionally used markers. Additionally, evidence is provided of an autocrine loop, involving type IV collagen and its integrin receptors, with importance for retaining a proliferative and migratory phenotype in pancreatic cancer cells.

Vascular endothelial growth factor (VEGF-A) was first described in 1989 for its angiogenic and mitogenic properties. Early studies indicated that VEGF-A acts primarily in a paracrine pathway which is limited to vascular endothelium. Further investigation showed that VEGF-A and VEGF receptor-2 (VEGFR-2) are expressed by many solid tumors and improve cell growth and survival. Therefore, VEGF-A may act via an autocrine pathway that effects tumor cellular proliferation by binding VEGFR-2 at the cell surface. This study utilizes small interfering RNA (siRNA) technology to investigate the presence of an autocrine loop in human RKO colorectal cancer cells. RT-PCR demonstrated the expression of VEGF-A, VEGF-B, VEGF-D, placental growth factor (PlGF), VEGFR-2, neuropilin-1 (NP-1) and neuropilin-2 (NP-2) in vitro by RKO cells. Transfection with siRNA against VEGF-A resulted in a 94% knockdown of VEGF-A expression by ELISA. Northern blot, quantitative real time PCR and semiquantitative RT-PCR confirmed the knockdown data. In addition, transfected RKO cells showed a 67% decrease in cellular proliferation by WST-1 assay. This data correlated to the ELISA results. In summary, the presence of VEGF-A and VEGFR-2 argues in favor of an autocrine loop in human colorectal cancer cells. siRNA targeting of VEGF-A remains a promising anti-tumor therapeutic strategy.