| dc.description.abstract | Seminal research conducted over the last two decades has confirmed that the immune system plays a crucial role in the processes that regulate tumor growth and development. So-called immune surveillance is an idea that has persisted for decades, and evidence supporting the hypothesis that the immune system detects and destroys neoplastic growth has accumulated over time. More recently the understanding of the role of immune checkpoints in cancer has led to breakthrough immunotherapeutic drugs such as ipilimumab (anti-CTLA-4 antibody), and pembrolizumab and nivolumab (anti-PD-1 antibodies). These drugs, also known as checkpoint inhibitors, act by blocking inhibitory immune receptors on T cells thereby enhancing the ability of the immune system to mount an anti-tumoral response. While immunotherapies can lead to durable, robust responses, the rate of response is quite variable across and within tumor types (Topalian, Taube, Anders, & Pardoll, 2016). There is currently an intense effort to characterize what differentiates responders from non-responders that will lead to the development of predictive criteria capable of identifying these populations prior to the initiation of treatment. For instance, the FDA has approved companion diagnostics that quantify the amount of programmed cell death ligand 1 (PD-L1) there is within a tumor for prediction of response to nivolumab and pembrolizumab (Topalian, Taube, Anders, & Pardoll, 2016). However, a recent analysis found that the response rate to anti-PD-1 antibody in PD-L1+ tumors was 48% while PD-L1- tumors still maintained a 15% response rate (Sunshine & Taube, 2015), demonstrating the difficulty that exists in predicting absolute response to these antibodies.
Ongoing work at Merck Research Laboratories (MRL) in Boston is focused on developing next generation immuno-oncology therapeutics. Understanding the molecular mechanisms that regulate response and non-response is central to this work. Traditional mouse tumor models relied on implantation of cells derived from human tumors, called xenografts, that required immunodeficient hosts. These models were suitable for the study of cytotoxic drugs designed to kill tumor cells directly. However, they would not be suitable for the study of tumor-immune interaction. The preclinical research at Merck currently relies heavily on mouse syngeneic tumor models that, crucially, use mice with an intact immune system capable of mounting immune responses directed towards the tumor. Here, tumor bearing mice are generated by exogenous implantation of cultured tumor cell lines derived from primary mouse tumors obtained from either naturally occurring (spontaneous) or carcinogen-induced tumors raised in mice of the same genetic background. Upon subcutaneous inoculation implanted tumor cells grow into solid tumors that can be easily measured for response to therapeutic intervention. The most commonly used murine syngeneic model at MRL uses the MC38 colon adenocarcinoma cell line. Tumors generated from these cells demonstrate robust responsiveness to murine anti-PD-1 antibody (mDX400). Typical response rates in mice bearing 80-120mm3 tumors are between 60-80% complete regressions (with additional partial responses) following 5 milligram/kilogram (mg/kg) dosing of mDX400, administered intraperitoneally (IP) every 5 days (Q5D) for approximately 15-20 days (data not shown). As MC38 tumors grow larger than this 100mm3 average size they have markedly lower response rates. It is possible that immune cell populations contribute to this observation. Phenotyping of explanted dissociated 80-120mm3 tumors via mass cytometry time of flight (Helios, Fluidigm) revealed that as much as 50% of cells within these tumors are CD45+ immune cell infiltrate. Furthermore, the PD-1/PD-L1 axis is abundant: 60+% of T cells express PD-1, ~90% of myeloid cells express PD-L1, and among CD45- cells, which consist of tumor and stromal cells, PD-L1 expression is nearly 50% (data not shown). Moreover, in tumors of this size the CD4+ T cell compartment is typically 10-12% CD25+Foxp3+ regulatory T cells (Treg). Opportunistic phenotyping of large tumors (600mm3 and 1000mm3) revealed drastic increases in Treg cells within the tumor infiltrating lymphocytes (TIL): ~30% and ~60% of total CD4+, respectively. Several lines of research have demonstrated the importance of IL-10 in the activation and anti-tumor activity of tumor resident CD8+ cytotoxic T cells, which is type I IFN dependent (Emmerich, et al., 2012) (Mumm, et al., 2011) (Stewart, et al., 2013). Additionally, it has been shown that the majority of intratumoral IL-10 is generated by Treg, which plays an important role in restraining tumorigenic Th17 inflammation (Stewart, et al., 2013). Therefore, one possible explanation for the loss of responsiveness to anti-PD-1 treatment as tumors progress is that while Treg numbers increase they become dysfunctional and lose the ability to both suppress Th17 inflammation and activate intratumoral CD8+ T cells. Elucidating the role of the type I IFN/IL-10/Treg axis in this phenomenon could provide valuable insight into strategies for treating patients that respond poorly to checkpoint inhibitor therapy.
The current study interrogates whether changes in the type I IFN/IL-10/Treg axis within the tumor microenvironment contributes to poor therapeutic response in syngeneic tumors across tumor progression through a comprehensive profiling of surface markers and cytokine production. | |
