Abstract
Over the past decade, advances in drug delivery technology have substantially increased patient compliance, improved drug efficacy, and minimized toxic side effects. With tailorable controlled release mechanisms and developments in nanotechnology, drug delivery systems can now be designed to meet specific goals. In particular, directing a drug away from healthy tissue and delivering it only to the diseased site through the use of targeted systems is paramount in battling toxicity issues. Yet obtaining the next level of clinical relevance will require significant improvements in the methods to generating targeting moieties. Similarly, molecular imaging has also rapidly progressed as a field, but is limited by a lack of quality targeting molecules. In this dissertation work, we sought to create a means of rapidly generating high affinity targeted ligands for use in therapy and imaging. We chose to apply our new approach in response to cutting edge research on the nature of pancreatic adenocarcinoma that implicates the stroma as crucial to the aggressive disease processes. The powerful combination of our new approach and the enhanced biological insight could stand to make a difference for a disease that has seen a significant lack of improvement in patient prognosis for decades.
We undertook this goal to make targeted ligands in a timely, costly, and efficient way by combining the well-established phage display technology with new quantitative analysis that we developed so that rather than relying on extensive validation and being subject to a high failure rate, we could predict which peptide sequences would be specific and selective. Our in silico quantitative analysis was made possible because of advances in sequencing technology that enable the reading of millions of sequences from a given library. We started by building a program to process, translate, and organize the pertinent data from deep sequencing. We then capitalized on the well-characterized results of phage screening on immobilized streptavidin protein to develop a strategy to predict specificity and selectively. We created a normalization strategy that allowed us to compare as many screens as we wanted, and we further developed our NGSanalyze program to include generation of a matrix of normalized data. We present this program as part of this dissertation and hope to enable other researchers to easily apply our process to their work.
Satisfied that our analytic methods were robust and objective, we transitioned to a more challenging task: discovering novel targeted peptides that selectively and specifically bound the cancer-associated fibroblasts (CAFs) in the stroma of pancreatic cancer. We were able to successfully implement our technique to this end and streamline the process of using such peptides to target liposomes. An impressive two of three peptide sequences chosen had great success in vivo in a mouse model. In fact, our targeted liposomes increased nanoparticle accumulation in mouse tumors 2.25- and 1.75-fold over non- targeted nanoparticles.
We took our system one step further by loading the targeted liposomes with a CXCR2 inhibitor, which is known to act as a paracrine signal between CAFs and other tumor cells. Delivery of the inhibitor to subcutaneous pancreatic tumors using our targeted liposome displayed similar growth inhibition as intraperitoneal injection of the free drug. Additionally, immunofluorescence suggested that both the targeted liposome and free drug reduced levels of a downstream protein and decreased angiogenesis as compared to mice who received no treatment. A statistically significant decrease in cell proliferation was also detected for both the targeted liposome and free drug versus the no treatment group. In contrast to the free drug, however, the targeted liposomes have a very different biodistribution that enables circumvention of the toxic effects known to accompany systemic injection of the inhibitor. This proof-of-concept study shows that our targeted liposomes are capable of drug delivery and highlights their potential to improve current PDAC therapy.
