A Preclinical Tale for Melanoma: The Aspirin and the Antibody

A preclinical study using mouse and human melanoma cell lines published in a September, 2015, issue ofCell, outlines mechanisms by which melanoma can evade the immune system. They found that cyclooxygenase (COX)-driven prostaglandin E2 enables malignant growth of melanoma through evasion of immune mechanisms, and suggest that COX inhibitors could be useful adjuvants for immune-based therapies.¹

The team used a transplantable melanoma tumor cell line from an immune competent BRAF^+LSL-V600E; Tyr::CreERT2^+/o;p16^INK4a-/-mouse. These were chosen because the mice were genetically prone to develop tumors which evidently had mechanisms in place to avoid attack by the mouse immune system. Evidence of this was gained by transplanting these BRAF^V600Emelanoma cells into wild-type mice. The result was progressively growing tumors, and this growth rate was barely increased when the cells were transplanted into T- and B-cell deficient mice (RAG^-/-strain). Again, these melanoma cells had evolved mechanisms to avoid attack by the immune system.¹

Low Level of Immunogenicity

Next, the team investigated whether or not the low level of immunogenicity of the melanoma cell line resulted from some defect they were able to induce in antigen-presenting cells, dendritic cells, and monocyte-derived cells. They took culture medium (CM) in which the BRAF^V600Ecells were grown, and used it to culture mouse bone marrow-derived mononuclear cells (BMMCs) comprising dendritic cells and monocyte-derived macrophages, with or without lipopolysaccharide (LPS). LPS was used to evoke a vigorous immune response.¹

The CM from the BRAF^V600Ecells powerfully lowered the production of TNF, IL-12/23p40 and MIP1a (an antitumor immune response) by the BMCCs. In contrast, the addition of the CM alone resulted in an increase in proinflammatory agents, such as IL-6, CXCL1, and G-CSF. These proinflammatory agents, in addition to IL-b, IL-10, and RANTES, were likewise induced by LPS but CM increased their buildup. The authors stated that, “Thus, tumor-derived secreted factors subvert the normal pattern of myeloid cell-driven inflammation.”¹

The Factors

Treatment of the CM with heat and benzonase did not alter the effect of the medium on BMMCs, ruling out a protein, and suggesting it may be a lipid. Analyzing the CM in which the melanoma cells had been cultured revealed that it contained high levels of the prostanoid PGE2. The cell line expressed two enzymes required for prostanoid synthesis, cyclo-oxygenase-1 and -2 (COX -1, COX-2), and COX-2 expression was dependent on activity in the RAF/MEK pathway. Addition of PGE2, to BMMCs had the same effect as the CM.¹

To confirm the role of COX-2, they generated Ptgs2^-1-(COX-2 deficient) BRAF^V600Emelanoma cells and found that CM from these cells no longer inhibited LPS-induced TNF production or IL-6 secretion in BMMCs. This established that COX-2 expression and activity in the CM from the BRAF^V600melanoma cells largely accounted for modulation of the properties of the BMMCs. To completely eliminate the production of PGE2, the team created a BRAF^V600Ecell line that was deficient in COX-1 and COX-2, and found no modulation of BMCC activity in vitro using this CM.¹

The Aspirin and The Antibody

Aspirin blocks COX-1 and COX-2; however, the addition of aspirin to the drinking water of mice implanted with COX-competent BRAF^V600melanoma cells exerted no effect on the progression of the tumors. The authors explain that this was probably because of inadequate inhibition of COX-2. However they reasoned that, “Even a modest degree of COX inhibition might help the efficacy of immunotherapies, including those based on immune checkpoint blockade.”¹

When an anti—PD-1 antibody was injected into the mice, with continued addition of aspirin in the drinking water, there was a much more rapid regression of the tumor and eradication of BRAF^V600Emelanoma cells than treatment with anti—PD-1 alone. The effect was absent in Rag1^-1-mice, demonstrating a dependence on adaptive immunity. Furthermore, once the melanoma had been cleared from mice using these two treatments, the mice were immune to further challenges with the BRAF^V600cells, in the absence of treatment.¹

Of Mice and Men

The team looked for evidence of a similar Cox-dependent modulation of the immune system in human melanomas. PTGS2, which encodes COX-2 mRNA expression levels in human melanoma biopsies, showed a strong positive correlation with mRNA levels for IL-6, G-CSF, CXCL1, and other tumor-promoting inflammatory factors in human cutaneous melanoma. Summarizing with additional findings, the authors concluded that, “Together, these data indicate a qualitative change in immune infiltrate composition that is driven by COX expression and shows remarkable parallels between mice and men.”¹

Implications for Treatment

The authors end the paper by stating that COX inhibitors, “might help unleash anticancer immunity and thereby constitute useful additions to the arsenal of conventional and immune-based cancer therapies, most notably those based on immune checkpoint blockade.”¹

Reference

1. Zelenay S, van der Veen AG, Böttcher JP, et al. Cyclooxygenase-dependent tumor growth through evasion of immunity.Cell. 2015;162:1257-1270.