Impact of mouse model tumor implantation site on acquired resistance to anti-PD-1 immune checkpoint therapy

Download pdf:

Introduction

The discovery of immune checkpoint inhibitors (ICI) has deeply modified treatment in several cancer indications. These therapies enhance the activity of immune cells against tumors by impeding the immunoparesis induced by tumor cells. ICI targeting the PD-1/PDL1 axis have shown significant antitumor activities in several tumor types (1–9). However, a majority of patients do not respond to therapy and a majority of those who are initially  sensitive to ICI will eventually relapse. Understanding the mechanisms of resistance to ICI  thus represents a major issue. In the case of primary resistance to ICI, access to patient  samples is fairly straightforward. Conversely obtaining longitudinal samples of primarily sensitive and secondarily resistant patients is far more challenging, explaining the scarcity of data regarding secondary resistance mechanisms in the clinic. The best described acquired resistance mechanisms are the overexpression of alternative ICI such as TIM3 on immune cells or PD-L1 on tumor cells, the dysfunction of the presentation of the antigen by MHC I or the mutations of genes  such as JAK1/2 (10–13).  The majority of murine syngeneic models are resistant to  anti PD1/PDL1 therapies. However, there are very few models of secondary resistance to these compounds (14). To address this issue we have developed syngeneic models of resistance to ICI and found that the development of the resistant phenotype is associated with strong molecular and immunological heterogeneity (15). Another methodological difficulty is the fact that most tumor implantations in mice are performed subcutaneously rather than orthotopically. Since the tumor immune microenvironment (TIME) is critical in the case of ICI therapy, it is likely that the site of tumor implantation will have an impact on the nature and functionality of the tumor immune infiltrate. The development of well characterized orthotopic murine models of sensitivity and resistance to ICI may thus be expected to be better correlated with the situation encountered in patients than subcutaneous (SC) models. Colorectal cancer (CRC), which is the third most prevalent type of neoplasia, was initially found to be poorly sensitive to ICI therapy, with a response rate of 5% (16, 17). However, response rates were found to be much higher when patients with microsatelliteinstability-high (MSI-H) or mismatched repair-deficient phenotypes were considered. Additionally the Immunoscore has been suggested to help select patients with a higher probability of response to ICI therapies (18). There remains a unmet need to better modelize the impact of orthotopic implantation which is expected to be associated with a specific immune infiltrate, exposure to microbiota and ability to disseminate to liver, which is the most common site of metastases in the clinic (19). To explore the impact of the implantation site in mouse colorectal tumor models we chose to compare subcutaneously and orthotopically implanted MC38 colorectal tumors, which were analyzed for their immune microenvironment and sensitivity to anti-PD1 therapy. We also developed resistant variants for both implantations and compared the alterations of the tumor immune microenvironment associated with acquisition of resistance. As presented in this manuscript, the major differences observed according to implantation site may be a major confounding factor in the preclinical modelization of ICI therapy.   Frontiersinimmu-13-1011943