Pancreatic cancer (PC) is highly lethal malignancy and consider the fourth most common cause of cancer-related deaths worldwide. The majority of patients, approximately 80%, are diagnosed at an advanced and incurable stage, advanced local (III) or metastatic (IV), of disease, and only around 20% of cancers are suitable for surgical resection. The prognosis for PC is extremely poor, with only 7% five-year survival rate. Few chemotherapy treatments have been shown to slightly improve survival for advanced pancreatic cancer patients, but even with the best treatments currently available, the median survival for these patients is less than one year. For over 40 years, survival rates in pancreatic cancer have not shown any improvement, and there is a clear need for further research of this disease to improve patient outcomes.
Palliative indiscriminate chemotherapy still is the standard of care for patients with advanced PC with no biomarkers currently available to guide the likelihood of an individual’s response to treatment. In recent decades, the treatment of many cancers has been evolved, in which screening for specific molecular changes is applied to patients first, and then targeted therapy is applied accordingly. Examples of this approach include the use of EGFR and ALK inhibitors in non-small cell lung cancers harbouring EGFR mutations or ALK gene rearrangements; BRAF and MEK inhibitors in melanoma; and trastuzumab and pertuzumab in HER2 amplified breast cancer, all of which have led to clinically meaningful improvements in patient outcomes. Given the limitations of current treatment regimens, the prospect of applying precision medicine to the treatment of PDAC holds great appeal.
Pancreatic ductal adenocarcinoma (PDAC) accounts for about 90-95% of all pancreatic cancer cases
Genomic landscape of PDAC
Recent advances in next-generation sequencing technologies have allowed for significant progress in the characterization of the genomic landscape of pancreatic cancer, leading to better understanding of the pathogenesis of this disease and identification of a number of potential therapeutic targets.
An activating somatic mutation of the KRAS gene has long been implicated as a critical event occurring early in the development of the overwhelming majority of human PDAC. Mutation of this gene has been implicated in the progression of pre-malignant pancreatic intra-epithelial neoplasia (Pan-IN) into invasive malignancy, and has also been demonstrated to play a vital role in tumour maintenance. Early mutation in KRAS is typically followed by the loss of multiple tumour suppressor genes, most notably CDKN2A, TP53 and SMAD4. However, beyond these common mutations, recent studies have revealed significant tumour heterogeneity among PDAC patients, highlighting a major challenge to the application of precision medicine to this disease.
In 2014, a large genome wide association study of more than 7000 PDAC patients identified numerous susceptibility loci for PDAC lying in close proximity to a variety of genes, some of which have previously been implicated in oncogenesis (e.g. BCAR1, KLF14, PDX1, CHEK2, TERT). More recently, whole exome sequencing on a smaller cohort of 109 patients identified that 5% of tumours contained 24 significantly mutated genes with potential prognostic significance (e.g. KRAS) as well as potential for therapeutic targeting (e.g. BRAF, PIK3CA). Further, comprehensive analysis of 24 PDAC tumours identified an average of 63 genetic mutations in each tumour and described alterations in 12 core signaling pathways, some of which (e.g. DNA damage repair, alterations in cell cycle regulation) may also be amenable to targeted therapy. Indeed, a number of gene mutations with potential for targeted therapy can be identified using resources such as the COSMIC database (Catalogue of Somatic Mutations in Cancer), but most occur at a low overall frequency in PDAC. Despite this, a recent comprehensive analysis integrating genomic, transcriptomic and proteomic profiling of 150 PDAC specimens identified that 42% of patients harboured at least one alteration which could potentially inform enrolment in a genotype directed clinical trial.
EUS FNA as a source of tissue and genetic material
Overwhelmingly, genomic profiling of PDAC has relied on surgical resection specimens to obtain tumour material. EUS FNA is a well-established, minimally invasive biopsy technique which can be utilized in patients at any stage of disease. It is generally considered a safe procedure, as evidenced by a large systematic review of over 10,000 patients undergoing EUS FNA across multiple institutions which reported reassuringly low morbidity (0.98%) and mortality (0.02%) rates associated with the procedure.
EUS FNA biopsy specimens can be used as a source of tissue for genetic analysis of PDAC, although widespread clinical use has been limited due to concerns regarding small tissue quantities, suboptimal yield of genetic material, and potential contamination of samples with non-malignant cells such as blood, inflammatory cells and stomach or intestinal wall cells. However, despite some inherent challenges, the potential to use this technique to obtain tissue from patients at all disease stages with relative ease is a clear advantage.
There are several approaches which have been demonstrated to improve the sensitivity and yield of EUS FNA-derived tissue for diagnosis and genetic analysis, including using larger needles and increased number of passes, utilizing on-site cytology services, and optimizing sample processing including using techniques such as snap freezing biopsies in liquid nitrogen and using RNA-preserving agents such as RNAlater™. Multiple studies have reported improved diagnostic sensitivity of the procedure using EUS FNA-derived genomic DNA to detect KRAS mutations, showing that this technique can be reliably used as a source of genomic DNA. Several studies have also demonstrated that EUS FNA can be used to reliably extract and sequence RNA for gene-expression profiling and to derive diagnostic gene signatures.
EUS FNA biopsies are also poised to play a role in establishing valuable pre-clinical disease models for precision therapy, such as patient-derived xenografts and organoid cultures. Pre-clinical disease models which accurately reflect the histological and molecular make-up of a tumour allow for rapid in-vivo testing and drug screening, and can support in-vitro findings. PDX involves implanting patient-derived cancer cells into immunodeficient mice to grow tumours, which have been shown to reliably retain the original histological architecture, cellular characteristics and molecular profiles of the original patient tumour over multiple passages. However, the lack of human stromal elements and the absence of a functional immune system is a limitation of these PDX models, particularly when considering in-vivo testing of immunotherapeutic agents. These limitations can be addressed to a degree through “humanizing” PDX models in-vivo, by reconstituting with human immune cells or tissue.
While PDX models have historically been largely reliant on surgical specimens, our group is among those to recently show that EUS FNA biopsies can also be successfully used to create PDX models which maintain their original tumour characteristics. Using defined media and conditions, patient derived tumour cells can also be used to grow organoid cultures, which provide three-dimensional cell cultures that can be used to complement PDXs in molecular analysis of PDAC and for longitudinal therapeutic testing in “real-time”. These organoid cultures can be generated using EUS FNA derived tissue in a short period of time, serially passaged, and used to create PDX models.
Our group has recently optimized a novel protocol for the simultaneous extraction of genomic DNA and RNA from EUS FNA biopsies, and demonstrated that EUS-FNA derived PDAC biopsies can be used to establish patient derived xenografts (PDXs) which can then be utilized for in-vivo testing of targeted therapies on molecular-profiled tumours. This translational pipeline allows for molecular screening and testing of targeted therapy on tumours derived from all clinical stages of disease. The benefits of this approach and contribution to bench-to-bedside care are evidenced by the subsequent initiation of the phase II clinical trial “Panitumumab in KRAS wild-type pancreatic cancer” at the Monash Health Translation Precinct, which is described in detail in the methodology section below. This pipeline leaves us well poised to examine our immune targets of interest in PDAC.
Immune regulation in pancreatic cancer
Dysregulation of the innate immune system can lead to chronic inflammation in the pancreas, which is well established as a risk factor for development and progression of invasive malignancy. Pro-inflammatory factors in the tumour microenvironment including cytokines (e.g. TNF-α, IL-6), chemokines (e.g. CCL-2), reactive oxygen species, angiogenic factors and growth factors, can promote growth of the underlying pancreatic epithelium and facilitate oncogene activation, leading to promotion of tumour cell proliferation, invasion and metastasis.
The tumour microenvironment in PDAC is rich in immune cells, and antigen presenting cells are capable of evoking an anti-tumour response by activating natural killer (NK) and cytotoxic T cells to eradicate tumour cells. However, spontaneous regression of tumours is extremely rare; and it is evident that pancreatic cancer cells are capable of inducing local immune dysfunction, leading to an immunosuppressive tumour microenvironment and protecting the tumour cells from immune attack. Immune cell function has also been demonstrated to be modulated by chemotherapy agents, and via interaction with stromal elements such as pancreatic stellate cells, which contribute to the typical desmoplastic reaction seen in PDAC.
In recent years, we have seen the emergence of immunotherapy as a viable and effective treatment modality which has been demonstrated to improve overall survival rates in several malignancies while maintaining a favourable toxicity profile; the most notable clinical responses are seen in melanoma, renal cell carcinoma and non-small cell lung cancer. These successes, along with the above observations noting an immune cell rich tumour microenvironment capable of eliciting anti-tumour response in PDAC, support the targeting of innate and adaptive immune responses as an appealing therapeutic prospect despite the known heterogeneity of the tumour genetics and local environment. There are a number of broad approaches to immunotherapy, including passive approaches such as the use of antibodies or effector T cells to target tumour-specific antigens or molecules involved in tumour progression; and active approaches such as vaccination with tumour antigens or irradiated tumour cells, or targeting immune checkpoint inhibition. However, to date the application of immunotherapy in unselected PDAC patients has proven disappointing. Applying a precision medicine approach to immunotherapy in PDAC may improve responses, by identifying predictive biomarkers to select and stratify patients towards specific treatments.
Our group has performed RNA sequencing on a number of EUS FNA-derived PDAC biopsies, and identified toll-like receptor 2 (TLR2), bone-morphogenetic protein 4 (BMP4) and interleukin 6 (IL-6) as genes of interest which display dysregulated expression across all clinical stages of PDAC. All three genes regulate a wide range of oncogenic processes affecting both tumour and immune cells, have previously been associated with carcinogenesis, and also represent potentially “actionable” targets with available therapeutic inhibitors. They are therefore attractive candidates to study further for immune-based precision medicine in PDAC. My initial focus is on TLR2, which is discussed in further detail below.