Using Polycationic Polymers to Interrogate and Mitigate the Effects of Toll-Like Receptor Mediated and Extracellular Microvesicle Induced Inflammation and Metastasis in Pancreatic Cancer

Abstract

Pancreatic cancer is a devastating disease with the worst prognosis of all the major cancers with a 5-year survival of less than 7%. This is due, in part, to the insidious development of pancreatic cancer which is largely asymptomatic until the tumor is too far developed to be efficiently combatted. The poor prognosis of pancreatic cancer is also due to its aggressively metastatic nature. Patients that are diagnosed with localized, surgically resectable tumors are treated with neoadjuvant (preoperative) and adjuvant (post-operative) chemotherapy, with or without radiation therapy, in the hopes of eviscerating micrometastatic disease. However, many of these patients progress to metastatic disease even in the face of aggressive therapeutic intervention. Tumor metastases to other organ sites, primarily the liver, wreak havoc on patients with pancreatic cancer. Unfortunately, most patients with metastatic pancreatic cancer decline quickly. In order to change the tide against pancreatic cancer and other aggressively metastatic tumor types, there need to be new clinical approaches in the arena of anti-metastatic therapeutics. Tumor metastasis is an incredibly complex process involving interactions between tumor cells, the tumor microenvironment, the peripheral stromal tissues, hematologic and lymphatic vessels, and the immune system. In order to reach a metastatic site, the tumor cell has to undergo epithelial to mesenchymal transition, detaching itself from its primary site, traversing the tumor stroma and extracellular matrix to reach hematologic or lymphatic vessels, traveling through the vasculature, and then embedding itself into the distant organ site. Contributing to the complexity of tumor metastasis are the toll-like family of receptors (TLRs). These receptors evolved as a part of the innate immune system to detect pathogen associated molecular patterns (PAMPs) to combat infectious agents such as bacteria and viruses. They are also known, however, to respond to molecules released by innate cells known as damage associated molecular patterns (DAMPs). These receptors are typically expressed by immune cells such as dendritic cells or neutrophils but have more recently been discovered to be expressed by both cancer cells and stromal cells within the tumor. The expression and activation of these receptors by both PAMPs and DAMPs has been implicated in tumor progression and metastasis in a multitude of cancers. Extracellular nucleic acids and nucleic acid-protein complexes have come to the fore as DAMPs that activate nucleic acid sensing TLRs such as TLR9 and TLR4, which sense DNA and nucleosomes, respectively. These innate molecules and their recognition receptors have been strongly implicated in tumor progression and metastasis in lung, breast, skin and pancreatic cancers. Studies have shown that specific activation of TLRs 4 and 9 contribute to invasion in vitro and metastasis in vivo. Recent work has also shown that DNAse treatment in animal models of cancer reduces metastatic burden. Thus, targeting extracellular nucleic acids and nucleic acid-protein complexes seems like a therapeutic strategy that warrants further exploration. In addition to circulating freely, nucleic acids and nucleic acid-protein complexes can also circulate in tumor derived lipid particles such as microparticles and exosomes. Recent research has shown these particles to have a myriad of activities including autocrine signaling amongst tumor cells, paracrine signaling between tumor cells, and communication with tissues of distant metastatic sites. Specifically, studies have shown that tumor derived microparticles can induce invasion in breast cancer cells. Additionally, pancreatic cancer derived exosomes have been shown to pre-condition the liver for metastatic implantation. The discovery of the diverse activities these particles are capable of has uncovered a new layer of the complexity via which tumors communicate and metastasize. These discoveries have also yielded new therapeutic targets that may yield novel clinical approaches to limit tumor progression and metastasis.Cationic polymers were initially developed exclusively as tools for gene delivery. Our group has previously shown that these polymers have the ability to bind to nucleic acid and nucleic-acid protein complexes and neutralize their ability to activate their specific TLRs. This anti-inflammatory ability of cationic polymers has led to beneficial effects in multiple mouse models of diseases in which nucleic acid mediated TLR signaling and inflammation are essential to disease progression. For example, in mouse models of acute liver failure, cutaneous lupus erythematosus, and influenza, cationic polymer therapy decreased the severity of the inflammatory response and increased survival in treated mice. These studies gave light to the new possibilities never before considered of using cationic polymers as multifaceted neutralizing agents of extracellular nucleic acid mediated inflammation. It is after considering all of the previous observations in aggregate that the work described in this thesis was undertaken. Pancreatic cancer was chosen as the disease model due to its aggressively metastatic nature. We first quantified the levels of extracellular nucleic acids and nucleic acid-protein complexes from the sera of patients with known pancreatic cancer. We found that the sera of patients with early stage pancreatic cancer had elevated levels of cell-free DNA (cfDNA), nucleosomes, and HMGB-1 as compared to sera from healthy volunteers. We also found that these factors increased in the sera of pancreatic cancer patients with standard therapies such as combination chemo-radiation therapy and surgical resection of the tumor. After quantifying these circulating factors, the next step was to interrogate the functional significance of elevated cfDNA, nucleosomes, and HMGB-1. Using a panel of TLR reporter cell lines, we tested the ability of pancreatic cancer serum to activate TLRs. Interestingly these receptors were strongly activated by pancreatic cancer patient sera and this activity was neutralized in some TLRs by cationic polymer treatment.TLR expression by cancer cells has been observed in the literature and activation of these receptors on tumor cells has been shown to induce invasive and aggressive phenotypes. We confirmed that a number of pancreatic cancer cell lines express TLRs and numerous studies from the literature show that tumor sections from patients with pancreatic cancer show increased staining for TLRs 4 and 9. Next we investigated the functional significance of TLR expression in pancreatic cancer cell lines using a Transwell-Matrigel invasion assay. We observed that treatment with DNA based TLR agonists induced a significantly invasive phenotype in a multitude of pancreatic cancer cell lines. This pro-invasive effect was abrogated with cationic polymer treatment. Next, we directly tested the pancreatic cancer patient serum in our invasion assay to ascertain whether the TLR9 activity we observed in our TLR reporter cell assays translated to functional effects on tumor cells. Indeed, pancreatic patient sera induced significant invasion in a Transwell-Matrigel invasion assay when compared to sera from healthy volunteers. Moreover, cationic polymer treatment significantly inhibited the ability of cancer patient sera to induce invasion.In addition to extracellular nucleic acids and nucleic acid-protein complexes, microvesicles derived from cultured tumor cells were also tested for their ability to induce invasion in the Transwell-Matrigel invasion assay. Interestingly, both microparticles and exosomes isolated from cultured tumor cells induced invasion in tumor cells and concurrent cationic polymer treatment significantly inhibited the pro-invasive phenotype. This observation provided a basis for the multi-modal use of cationic polymers as potential inhibitors of tumor invasion.In order to elucidate the use of cationic polymers as cancer therapeutics, we established a syngeneic, bioluminescent, immunocompetent murine model of pancreatic cancer metastasis. This model facilitated our ability to track the progress of liver metastases as well as growth of the primary tumor in an immunocompetent setting, thus allowing a more accurate recapitulation of tumor growth and metastasis as it occurs in humans. Cationic polymer therapy significantly reduced liver metastases in our model of pancreatic cancer. Specifically, liver metastases were greatly reduced as measured by total bioluminescent flux (tumor cell specific) and by gross organ weight. In fact, liver weights in polymer treated mice were equivalent to weights from the livers of healthy mice. Liver metastases in polymer treated mice were reduced by >90% overall. Interestingly, polymer treatment had no effect on primary tumor growth as measured by bioluminescent flux and gross organ weight. The lack of any effect on the primary tumor is supported by the in vitro data that we gathered, showing that cationic polymer treatment inhibited tumor invasion but had no effect on tumor cell viability.The second portion of this thesis was focused on characterizing the nucleic acids and nucleic acid-protein complexes in the cancer patient sera that were bound to the cationic polymers. In order to accomplish this goal, we chemically modified an existing cationic polymer to add a biotin tag to the polymer, allowing us to pull the polymer and bound molecules out of solutions using streptavidin coated beads. By performing this modification, we were able to take cancer patient sera and pullout the polymer-nucleic acid/protein complex. By doing so, we were able to show that A) pulling the bound species out of the sera reduced the ability of the sera to activate TLRs and B) The bound species that is pulled out of the sera is indeed responsible for the activation of TLRs. In order to be able to do more in depth analysis of the molecules being bound by the polymers, we had to increase the scale of our source material beyond patient serum due to the limited nature of harvesting blood from patients with pancreatic cancer. To accomplish this, we cultured human pancreatic cell lines to confluence and collected conditioned media from these cells. Cells with media alone represented patients with untreated pancreatic cancer. Cells were also treated with radiation prior to collection of conditioned media to recapitulate the fractionated radiation patients with pancreatic cancer receive. This gave us a resource for biomolecules that would represent the secretions of pancreatic tumors under a variety of conditions.The next step was to characterize the polymer bound molecules, analyzing the nucleic acid and proteinaceous components pulled out of the conditioned media. In order to accomplish this analysis, we displaced the polymer bound molecules using soluble heparin, a negatively charged polymer that would knock off negatively charged DNA and DNA-protein complexes. Using this strategy, we were able to isolate the proteinaceous and nucleic acid components of the polymer bound molecules. The protein components were sent for qualitative mass spectrometry which yielded numerous known markers of pancreatic cancer and agonists of TLR4. The nucleic acid components of the polymer bound molecules were run on agarose gels to analyze their size patterns. The gels yielded a wide diversity of bands, representing the variety of DNA sizes that are released by cancer cells at baseline and in response to radiation therapy. In addition, we adopted a parallel strategy to characterize the proteins the polymer was binding to by analyzing the conditioned media before and after biotinylated-polymer treatment with ELISA’s specific for HMGB-1 and nucleosomes. These proteins are important mediators of TLR based inflammation, particularly due to their ability to bind to and act as carriers for pro-inflammatory nucleic acids that potently activate TLR9. Using this strategy, we discovered that after pulling out the biotinylated polymer, the concentrations of HMGB-1 and nucleosomes decreased in the conditioned media. Additionally, both protein complexes were found in the heparin displaced supernatant from the polymer. The polymer treated and untreated conditioned media were incubated with our TLR reporter cells and we observed decreased TLR activity in the conditioned media that was treated with biotinylated polymer. This strategy also confirmed our previous observation that in addition to nucleic acids, the polymer is binding to nucleic acid-protein complexes and preventing them from activating TLRs. This strategy was also repeated with different cell types, including immortalized, normal pancreatic epithelial cells, human endothelial cells, and lymphoid cells to encompass the cell types that are likely exposed during standard radiation therapy. Interestingly, the most potent release of TLR agonists was seen from tumor cells with other cell types having some TLR activity but significantly less than tumor cells. The use of polycationic polymers as anti-metastatic agents in cancer has never before been considered as a potential use for these incredibly malleable tools. Moreover, the ability of these polymers to be chemically modified allows their use as potential therapeutic agents, tools for liquid biopsy of patient derived and cell culture samples, and a means for understanding the mechanistic underpinnings of the in vivo efficacy of polymers in cancer models. The work described here highlights the diverse capabilities of polycationic polymers in pancreatic cancer. They act as potent anti-inflammatory agents, reducing the levels of TLR activation. This activity translates to an inhibition of in vitro invasion in pancreatic cancer cell lines which is further reflected in the significant inhibition of liver metastasis in vivo. The potential clinical benefits are further explained with the use of a biotin-modified polymer to bind, extract, and identify the pro-inflammatory nucleic acids and nucleic-acid protein complexes. The use of this modified polymer yielded both the identities of the bioactive molecules as well as provide a means to perform a liquid biopsy of cell culture media and patient sera. In totality, the work described here signifies the diverse capability of polycationic polymers in cancer.