Combinations of immunostimulatory antibodies with synergistic effects against spontaneous cancer.

Immunostimulatory monoclonal antibodies can be given in combinations, hence modulating the activity of 2 or more receptors of the immune system. Some of these combinations have been shown to synergize at the elicitation of therapeutically relevant immune responses in transgenic mice developing spontaneous, oncogene-driven tumors, including multifocal hepatocellular carcinomas expressing ovalbumin as a surrogate tumor-associated antigen.

Making the most of the immune system to tackle cancer needs the simultaneous or sequential modulation of multiple immunological mechanisms. The concept of combinatorial immunotherapy originates from the fact that it is virtually impossible to develop a single molecule that would impact all such mechanisms. Hence, efforts should be (and are being) refocused on the use of a palette of interventions to identify the most suitable synergistic approaches. Immunotherapeutic approaches can be combined not only in pairs, but also in triplets and quadruplets., In this setting, improved efficacy can certainly be attained, but the possibility that the combinatorial approach may result in synergistic adverse effects should be carefully monitored.

Transplanted tumors are no longer the workbench to test the efficacy of combinatorial immunotherapy, mainly because in this setting treatments are often curative and the predictability of using each agent in suboptimal conditions is not likely to render solid preclinical evidence. Rather, investigation should be based on transgenic mice that accurately mimic human oncogenesis, tumor progression and response to therapy. Many such models are available, though often they involve multifocal carcinogenesis with weak evidence of immunoediting, mainly because of the limited accumulation of mutations that may give rise to tumor-associated antigens (TAAs) and short disease latency.

We have used a very aggressive model of multifocal hepatocellular carcinoma (HCC) based on the inducible, liver-specific expression of human MYC. If deprived of doxycycline at birth, animals develop multifocal HCCs in 3–4 wk, and usually die before 8–10 wk of age. Importantly, in this model, a bidirectional transgene ensures that cancer cells express MYC together with chicken ovalbumin (OVA) as a surrogate TAA. Previous work had unveiled the strong tolerogenic properties of OVA as a result of its expression on the surface of cancer cells. In fact, upon adoptive transfer to these mice, OT-1 T lymphocytes (which express an OVA-targeting TCR) become anergic along with TCR downregulation.

In this dire scenario, we tested a triple combination of immunostimulatory monoclonal antibodies, encompassing molecules targeting CD274 (best known as PD-L1 or B7-H1), tumor necrosis factor receptor superfamily member 9 (TNFRSF9, best known as CD137 or 4–1BB), and TNFRSF4 (best known as OX40 or CD134). PD-L1 and its receptor, programmed cell death 1 (PDCD1, best known as PD-1), were expressed within neoplastic lesions and conceivably were repressing antitumor immune responses. CD137 and OX40 were also available for antibody ligation and these interactions are known to act as agonists for T-cell and NK-cell co-stimulation.

To our surprise, our combinatorial immunotherapeutic regimen prolonged the survival of MYC- and OVA-expressing mice. Such an effect was totally dependent on CD8+ T cells as well as on NK1.1+ lymphocytes. Conversely, CD4+ T cells were dispensable, despite expressing the targets for all antibodies. The combinatorial administration of anti-PD-L1, anti-CD137 and anti-OX40 mAbs mediated long-term antineoplastic effects in 20% of the animals. In this setting, the adoptive transfer of activated OT-1 and OT-2 cells (which also express an OVA-specific TCR) is totally ineffective even when supported by intraperitoneal interleukin-2 (IL-2). However, adoptive T-cell therapy with activated OT-1 and OT-2 T lymphocytes combined with anti-PD-L1, anti-CD137 and anti-OX40 mAbs achieved an impressive efficacy, resulting in the survival in more than 70% of mice at day 250. Of note, a mAb specific for cytotoxic T lymphocyte-associated protein 4 (CTLA4) did not improve further the efficacy of our immunotherapeutic intervention in this model.

Experiments addressing the mechanisms underlying our observations revealed a dramatic infiltrate of tumor lesions by CD4+ and CD8+ T cells. Both such lymphocyte subsets expressed the targets of our immunostimulatory mAbs, namely, PD-L1, CD137, and OX40. Interestingly, the co-administration of anti-PD-L1, anti-CD137 and anti-OX40 mAbs promoted the expression of their targets on T cells.

The cellular response orchestrated around interferon γ (IFNγ), perforin, and granzyme B is the ultimate effector mechanism of anticancer immunity and is significantly inhibited by the profound tolerogenic nature of the most tumors. The spontaneous immune reactivity against OVA is completely suppressed in our model, implying that the effector immune response elicited by the combinatorial administration of anti-PD-L1, anti-CD137, and anti-OX40 mAbs must be directed to other TAAs. We also ruled out the possibility that such response may be directed against the bacterial protein that regulates the tetracycline-repressible expression cassette employed or human MYC. Hence, an immune response targeting other, weak antigens must be involved. The nature of such antigens, however, remains elusive.

As a whole, our work demonstrates the efficacy of a triple immunostimulatory mAb-based approach acting on interrelated target molecules (Fig. 1). The efficacy of such an approach mainly relies on the release of PD-1-dependent immunosuppression coupled to dual, CD137- and OX40-driven co-stimulation.

Figure 1. Molecular mechanisms underlying the therapeutic efficacy of anti-PD-L1, anti-CD137, and anti-OX40 monoclonal antibodies. Interactions between a T lymphocyte and a cancer cell. In this setting, monoclonal antibodies (mAbs) targeting PD-L1 (B7-H1) are de-repressing T-cell activation while anti-CD137 (4–1BB) and anti-OX40 (CD134) mAbs provide robust co-stimulatory signals.

The field of HCC therapy is potentially very fertile for immunotherapy. For instance, an anti-CTLA4 mAb has recently been associated with signs of clinical activity in HCC patients, and an anti-PD-1 mAb is currently being tested in a multicenter clinical trial (NCT01658878). Our proof of concept study based on murine models supports the therapeutic potential of triple immunostimulatory mAb combinations. The addition of adoptive T-cell therapy further enhanced the efficacy of our approach, suggesting the value of testing immunotherapeutic regimens to the limit. Combinations are nowadays perceived as the most suitable strategy when seeking superior efficacy.