Something in the Air: Hyperoxic Conditioning of the Tumor Microenvironment for Enhanced Immunotherapy

Abstract

Recent clinical trials in cancer therapy have demonstrated unprecedented responses through blockade of CTLA-4 and PD-1 immune checkpoint pathways. In a provocative recent paper in Science Translational Medicine, Hatfield and colleagues demonstrate the ability of supplemental oxygen to act as a novel immune checkpoint inhibitor by disrupting the hypoxia-adenosine-A2aR pathway.

Recent clinical successes of immunotherapy for cancer support a model whereby T cell responses are generated against tumor antigens only to be thwarted by negative regulatory mechanisms that comprise the tumor microenvironment (Postow et al., 2015). Indeed, it is clear that, for some patients, checkpoint blockade directed against PD-1 and CTLA-4 is able to unleash T cells capable of promoting durable tumor regression. Recently, Hatfield et al. (2015) presented data accentuating the role of the hypoxia-adenosine-A2aR axis as a checkpoint in thwarting anti-tumor immune responses. For checkpoint blockade, they employ oxygen.

Over the past two decades, a multitude of studies have demonstrated the critical role of extracellular adenosine as an immunoregulatory agent to prevent the development of over-exuberant immune responses (Young et al., 2014). As an immune response develops, against pathogenic invasion for example, increased cellular turnover and hypoxia promote the production and release of large amounts of adenosine into the local microenvironment (Figure 1). Signaling through the A2a and A2b adenosine receptors (receptors that are broadly expressed on immune cells), adenosine triggers a pleiotropic cascade of immunosuppressive effects, dampening the inflammatory response and protecting surrounding tissues from immune-mediated damage. Increasing hypoxia in inflamed tissues triggers the accumulation of adenosine by depleting intracellular ATP, increasing intracellular AMP, increasing expression of 5′-nucleotidase, and, through hypoxia-inducible factor-1α (HIF-1α) activity, inhibiting adenosine kinase activity (Decking et al., 1997; Sitkovsky, 2009). The increased activity of HIF-1α in hypoxic conditions also drives the expression of the CD73 and CD39 transmembrane ectonucleotidases that work in tandem to degrade extracellular ATP to adenosine (Young et al., 2014). As the adenosine concentration increases in the extracellular environment, increased signaling through G_s-protein-coupled receptors (A2a and A2b) produces an elevation in intracellular cAMP, ultimately suppressing the release of inflammatory mediators and enhancing the expression of immunosuppressive factors.

Figure 1. The Hypoxia-Adenosinergic Pathway in Immune Regulation.

Partially through stabilization of HIF-1α, hypoxia in inflamed and cancerous tissues drives the intracellular accumulation of adenosine by causing decreased ATP and adenosine kinase (AK) activity as well as increasing intracellular AMP and 5′-nucleotidase activity (5′N). Additionally, HIF-1α activity drives increased expression of the 5′-ectonucleotidases CD39 and CD73, which are transmembrane proteins that catalyze the enzymatic breakdown of ATP to adenosine. Both of these mechanisms lead to significant elevation of extracellular adenosine concentrations in the microenvironment. Signaling through two G_s-protein-coupled adenosine receptors (A2aR and A2bR), adenosine triggers an increase in the intracellular signaling molecule cyclic AMP (cAMP). Within immune effector cells such as cytotoxic T cells and NK cells, cAMP causes a buildup of a range of immunosuppressive factors (TGF-β, IL-10, CTLA-4, PD-1, Galectin-1, FoxP3), as well as the down-regulation of key effector molecules (IL-2, INF-γ, TNF-α, perforin, IL-12, MIP1α) involved in an active immune response. A2aR, adenosine A2a receptor; A2bR, adenosine A2b receptor; CTLA-4, cytotoxic T-lymphocyte-associated protein 4; FoxP3, forkhead box P3; HIF1-α, hypoxia inducible factor-1α; INFγ, interferon-γ; IL-2, interleukin-2; IL-10, interleukin-10; IL-12, interleukin-12; MIP1α, macrophage inflammatory protein 1α; PD-1, programmed cell death protein 1; TGF-β, transforming growth factor-β; TNFα, tumor necrosis factor-α.

In their article recently published in Science Translational Medicine, Hatfield et al. (2015) not only hypothesized that the hypoxia-adenosine-A2aR axis was acting as a “checkpoint” to inhibit anti-tumor immunity, but, more provocatively, that they could overcome this inhibition by using supplemental oxygen to produce hyperoxia in mice. To this end, multiple pulmonary tumors were established by the intravenous injection of B16 melanoma or MCA205 fibrosarcoma tumor cells. By simply allowing mice to breath 60% oxygen, they were able to demonstrate a marked decrease in tumor size as well as a marked increase in overall survival when compared to mice breathing ambient air. Notably, the anti-tumor responses were completely abrogated in γc/Rag-2^−/− mice that are deficient in T and NK cells, thus demonstrating the requirement of immune effector cells to enact the antitumor response. With the use of monoclonal antibodies to deplete CD4⁺, CD8⁺, and NK cells, they demonstrated further that, while the effect of hyperoxic breathing on tumor regression was greatly attenuated in the absence of CD4⁺ and CD8 T⁺ cells, the effect was completely abrogated after the depletion of the NK cell population. Importantly, the authors also showed that, in contrast to wild-type mice, hyperoxia had no detectable effect on enhancing anti-tumor immunity further in A2aR null mice. This latter finding implicates a role for the A2aR in mediating the hypoxia-induced immunosuppression. Lastly, administration of a reactive oxygen species (ROS) scavenger had no effect on the tumor response, indicating that ROS and related cytotoxicity did not play a meaningful role in the observed tumor response.

Interestingly, altering the oxygen tension of the tumor microenvironment also had an effect on T cell trafficking. First, through immunohistochemical staining of tumor tissue sections, the authors were able to demonstrate that CD8⁺ T cells preferentially localized to non-hypoxic regions within a particular tumor. Second, by using an in vivo marker of hypoxia, they demonstrated that hyperoxic respiration decreases CD8⁺ and CD4⁺ T cell exposure to hypoxic microenvironments. Third, the authors observed a dramatic increase in recruitment of CD8⁺ T cells to established MCA205 tumors in the setting of hyperoxic breathing. Importantly, when evaluated for CD69 and CD44 expression by flow cytometry, these CD8⁺ T cells were significantly more highly activated than those isolated from mice exposed only to ambient O₂. The increased CD8⁺ T cell trafficking was supported by the change in expression of over 50 cytokines, chemokines, and chemokine receptors. These findings were extended by showing that the number of adoptively transferred, CFSE-labeled T cells that traffic to established tumors was 3-fold higher per unit area in mice undergoing hyperoxic respiration than control mice. Furthermore, hyperoxia resulted in a decrease in the percentage of T regulatory cells in the tumor microenvironment, decrease in Foxp3 expression, decrease in CD39 and CD73 expression on the T regulatory cells, as well as a decrease in the expression of CTLA-4 on the T regulatory cells. While it is thought that tumor infiltrating T regulatory cells easily adapt to the tumor microenvironment, it appeared as though altering the oxygen tension in the tumor disrupted this adaptive process.

Overall, these studies have potentially important implications for the further development of immunotherapy for cancer. First, they confirm and provide insightful details concerning the role of hypoxia, adenosine, and the A2aR in inhibiting anti-tumor immune responses. To this end, these findings support the development of A2aR antagonists and CD73 inhibitors as novel agents to enhance immunotherapy. Second, these observations immediately lend themselves to devising novel and synergistic clinical approaches. The simplicity of oxygen therapy as a means of enhancing immunotherapy is appealing. Yes, persistent hyperoxia can lead to tissue damage both in the lungs and liver (Kallet and Matthay, 2013; Zangl et al., 2014). Thus, reproducing these findings in humans in a way that limits hyperoxia-induced injury will be challenging. However, one can imagine sequencing oxygen therapy with anti-PD-1, anti-CTLA-4, or even A2aR blockade as a means of making the tumor microenvironment more hospitable for immune-mediated tumor destruction. As higher populations of intratumoral CD8⁺ cells have been correlated to better clinical responses to PD-1 checkpoint blockade (Tumeh et al., 2014), the observation that hyperoxia can increase CD8⁺ T cells and decrease Foxp3⁺ T regulatory T cells within the tumor portends well for combining oxygen and other modalities. Along these lines, given the pivotal role of oxygen in regulating both tumor and immune cell metabolism and function (Palazon et al., 2014; Zeng et al., 2015), increasing the oxygen tension within the tumor microenvironment might prove to be a simple means of simultaneously dampening multiple immune regulatory checkpoints.