Regulatory T Cells Move in When Gliomas Say “I DO”

Commentary on Wainwright et al., p. 6110

In this issue of Clinical Cancer Research, Wainwright and colleagues present data characterizing the role of IDO-induced immunosuppression in glioma, and specifically address the impact of IDO on the recruitment of regulatory T cells (Treg) to tumors in the brain (1).

Glioblastoma is the most common and aggressive tumor of the central nervous system (CNS). It carries an exceedingly poor prognosis despite aggressive combinations of surgical resection, radiation, and temozolomide chemotherapy, as well as salvage therapy with bevacizumab. These conventional therapies also cause incapacitating nonspecific toxicity to healthy tissues. In contrast, immunotherapy offers an exquisitely precise approach; substantial evidence suggests that T cells have the ability to eliminate large, well-established tumors, even within the “immunologically privileged” CNS and without concomitant autoimmunity (reviewed in ref. 2).

Despite this promise, glioblastoma immunotherapy has been limited to date, in part, due to proclivities that the tumor possesses for instigating profound host immunosuppression. While the source of this immune impairment is clearly multifactorial, Tregs are perhaps most frequently implicated as major contributors to cellular immune defects and are present at increased proportions in both the tumors and peripheral blood of patients with glioblastoma (3, 4). In tumors, Tregs curtail the generation and expansion of cellular immune responses, primarily through inhibited production of interleukin (IL)-2 and IFN-γ and the induction of T-cell anergy. Recent studies have suggested that the accumulation of Tregs in brain tumors may be a result of factors associated with the anatomical niche as intracerebral gliomas in mice were found to contain significantly elevated numbers of Tregs compared with identical tumors implanted subcutaneously (5). Although the CCR4/CCL22 axis has certainly been implicated, the data presented here by Wainwright and colleagues suggest an additional mechanism for Treg-to-glioblastoma recruitment, centered about the enzyme indoleamine 2,3-dioxygenase (IDO).

IDO is an intracellular oxidoreductase that catalyzes the first step of tryptophan degradation. In a variety of tumor models, the expression of IDO by dendritic cells in tumor-draining lymph nodes has been shown to potently facilitate antitumor T-cell tolerance, partly via the induction and recruitment of new and existing Tregs (Fig. 1). Currently, it is believed that IDO also inhibits effector T-cell responses through the depletion of essential tryptophan stores. In the absence of tryptophan, effector T cells have been shown to activate the general control nonrepressed-2 (GCN2) stress kinase pathway, which leads to cell-cycle arrest and abortive T-cell activation. Interestingly, IDO expression can be induced in dendritic cells through counter-regulatory mechanisms triggered by proinflammatory mediators (e.g., IFN-γ), or alternatively through interactions with CTLA-4 on the surface of activated Tregs, thereby creating an immunosuppressive feedback “loop.”

Wainwright and colleagues show that IDO expression by glioma leads to the recruitment of Tregs to tumors resulting in impaired immune-mediated rejection of brain tumors in both orthotopic and transgenic models of glioblastoma (1). The effects of IDO proved most impactful when expressed by the tumor itself, with peripheral levels being seemingly less relevant. This was clinically correlated insomuch as a retrospectively reviewed database of tumor samples from patients with glioblastoma revealed an inverse relationship between IDO expression and survival. Given that an IDO inhibitor, 1-methyl-tryptophan (1MT), is now in clinical trials, these data may be leveraged to support the application of IDO inhibition as a novel therapeutic strategy for glioblastoma. The caveat though will be that such inhibition will need to be shown at the level of the tumor and not systemically. This is especially pertinent given that much of what is known about IDO and its inhibition to date has been limited to its peripheral activity in dendritic cells, which the authors advance as less influential in the setting of glioblastoma.

Of further import, the authors' data also suggest that it is the constitution of tumor-infiltrating lymphocytes (TIL), and not their absolute number per se, that impacts survival. Previously, a number of efforts have been made to evaluate the prognostic significance of TILs in glioblastomas; however, these studies have not yielded definitive results to date. In contrast, when considered as a subset of glioma-infiltrating lymphocytes, intratumoral Tregs have been shown to correlate with tumor grade and malignant features. This notion is similarly supported by Wainwright and colleagues, as increased levels of tumor IDO, despite being associated with increased numbers of TILs in general, were simultaneously associated with an increased ratio of Tregs-to-CD8⁺ T cells, and accordingly poorer survival.

Numerous efforts have been made to selectively modulate Treg activity with hopes of improving immunotherapeutic responses against glioblastoma. Importantly, upon removal of Tregs from the peripheral blood of patients with glioblastoma, normal levels of effector T-cell proliferation and cytokine secretion can be restored in vitro, and in vivo depletion of Tregs in tumor-bearing mice has been shown to significantly prolong survival (3, 6, 7). Given the high levels of CD25 on the surface of Tregs, attempts have been made to selectively eliminate Tregs with denileukin diftitox (a fusion protein of diphtheria toxin and IL-2) and LMB-2 (a fusion protein of an anti-IL-2Rα monoclonal antibody and exotoxin). Both have been used to deplete Tregs in humans but with mixed success. A significant limitation is the collateral depletion of effector T cells that transiently upregulate IL-2Rα upon activation. Moreover, due to its IL-2 targeting moiety, denileukin diftitox has the added drawback of unintentionally eliminating an even broader subset of cells that constitutively express lower affinity IL-2βγ receptors.

Although Tregs rely heavily on IL-2 for survival and function, activated effector T cells and tumor-specific memory T cells are not necessarily IL-2–dependent, especially when undergoing homeostatic proliferation after lymphodepleting chemotherapy. This has prompted hope that systemic blockade with unarmed anti-IL-2Rα antibodies (e.g., daclizumab and basiliximab) may selectively impair Tregs based on their differential requirement for IL-2. Importantly, in preclinical models and early clinical trials, monoclonal antibodies that block IL-2Rα have been shown to significantly reduce Treg activity without hampering vaccine-induced immunity (8, 9). Although promising, current strategies to deplete Tregs in vivo have not been shown to efficiently eliminate intratumoral Treg populations, a limitation that IDO inhibition may avoid.

Going forward, it will be important to assess whether IDO inhibition may prove more effective or even synergistic with like-minded therapies. For instance, as noted, one component of IDO activity may be interaction with CTLA-4 on Tregs. CTLA-4 has an established role in facilitating Treg activity as well as in limiting activated T-cell responses, and its blockade with ipilimumab has potent therapeutic value in metastatic melanoma, for which it is FDA-approved (10). Likewise, CTLA-4 blockade has proven effective in experimental models of glioblastoma (11) and in compassionate use protocols for patients with brain metastases (12). Given their likely activity along 2 steps in the same pathway, combining IDO inhibition and CTLA-4 blockade may be one example of a rational therapeutic platform to assess. Clearly, the authors show IDO inhibition as a promising means of countering Treg recruitment and activity. Clarifying its mechanisms, appropriate immune context and interaction with other modalities should be emerging priorities.

No potential conflicts of interest were disclosed.

Conception and design: B.D. Choi, P.E. Fecci, J.H. Sampson

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): P.E. Fecci

Writing, review, and/or revision of the manuscript: B.D. Choi, P.E. Fecci, J.H. Sampson

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.E. Fecci

Study supervision: J.H. Sampson

This work was supported in part by the NIH 5R01-CA135272-04 (to J.H. Sampson), 5P50-NS020023-29 (to D.D. Bigner and J.H. Sampson), 3R25-NS065731-03S1 (to J.H. Sampson), as well as grants from the Pediatric Brain Tumor Foundation (to D.D. Bigner and J.H. Sampson), Ben and Catherine Ivy Foundation (to J.H. Sampson), and Cancer Research Institute (to B.D. Choi).