In this issue, O'Mahony et al. (1) targeted CTL–associated antigen 4 (CTLA-4) in previously vaccinated solid tumor patients using the human anti–CTLA-4 antibody ipilimumab (MDX-010). Ipilimumab, together with a second human anti–CTLA-4 antibody, ticilimumab, are currently in phase 3 clinical trials in a number of cancers. The study was a single center, open-label pilot phase II trial consisting of 11 patients, 4 with non–Hodgkin's lymphoma immunized to tumor-specific idiotype, 4 with prostate cancer immunized using a poxvirus vector encoding prostate-specific antigen, and 3 with colon cancer immunized using a mutated ras peptide–pulsed dendritic cell vaccine, all of whom failed tumor-specific vaccination therapy. Patients received four cycles of ipilimumab, an initial dose of 3 mg/kg i.v. followed by three monthly doses of 1.5 mg/kg. The primary end point of the study was to assess ipilimumab toxicity at the specified dosing schedule. Overall, ipilimumab was well tolerated with a number of grade 1 and 2 toxicities possibly related to treatment. Patients with more advanced metastatic disease manifested grade 3 and 4 toxicities likely attributable to their advanced disease rather than ipilimumab treatment. One patient with advanced metastatic colon cancer died within 30 days of treatment due to disease progression. Tumor responses were limited to two patients with non–Hodgkin's lymphoma who experienced limited tumor regression at selected metastatic sites. No objective tumor responses were seen in the patients with prostate or colon cancer.
In addition to evaluating ipilimumab toxicity and tumor response, the secondary end points of the study were aimed at elucidating the effects of anti–CTLA-4 on relevant immune variables. These included the effects of ipilimumab on the number of peripheral CD4+CD25+ regulatory T cells (Tregs) and the activation of peripheral vaccine-specific CD8+ effector T cells.
CTLA-4 Regulates Immune Responses
The adaptive immune response requires two signals between the antigen-presenting cells (APC) and the effector T cell as shown in Fig. 1. The primary signal is mediated by the T cell receptor and the specific antigenic peptide presented in the context of MHC class I or class II molecules expressed on the APC surface, whereas the secondary signal is mediated through constitutively expressed costimulatory molecules on the T cell (CD28) and the APC (B7.1/CD80 or B7.2/CD86). The presence of both signals triggers intracellular events resulting in the activation and interleukin (IL)-2–dependent clonal proliferation of T cells expressing T cell receptor specific for the presented antigen. Unchecked, T cell proliferation can lead to autoimmunity and even death. The immune system has evolved homeostatic mechanisms that down-regulate this clonal expansion, which include the up-regulation of CTLA-4 (2). Activated T cells up-regulate surface CTLA-4, which has a higher affinity for B7.1 and B7.2 than CD28, effectively competing for B7.1 and B7.2 binding and inducing inhibitory signals in effector T cells. This leads to the dampening of the effector T cell response.
In addition to well-documented tumor escape mechanisms such as the down-regulation of surface MHC class I molecules and cytokine-mediated reduced responsiveness to tumor-associated antigens (TAA), the findings in both preclinical and clinical studies that treatment with anti–CTLA-4 is associated with enhanced T cell responses and antitumor activity have validated it as a target for immunotherapy (3–6). In addition to activity seen when used as a single agent, preclinical studies combining anti–CTLA-4 with vaccine (7), and reports of enhanced immune and tumor responses in patients previously treated with vaccine support its use following vaccine treatment (8).
O'Mahony et al. discuss two general mechanisms by which CTLA-4 blockade may generate antitumor responses. The first is mediated through CTLA-4 blockade on tumor-specific effector T cells (CTL) discussed above, resulting in an increased propensity for clonal activation and expansion. Consistent with this mechanism, Maker et al. recently provided evidence that treatment with ipilimumab in patients with metastatic melanoma or metastatic renal cancer increased the numbers of activated effector cells among peripheral blood mononuclear cells (9). The second possible mechanism is mediated through CTLA-4 targeted depletion of tumor-induced Tregs, which act to inhibit immune responses against TAA. Depletion of Tregs is proposed to overcome nonresponsiveness to TAA and allow the generation of tumor-specific responses. This is the hypothesis favored by O'Mahony et al. and forms the basis for the discussion of CTLA-4 in the context of tumor-associated Tregs.
CTLA-4 and Tregs in Tumor Immunity
Tregs are a suppressive CD4+ T cell population that express high levels of surface CD25 (the high affinity IL-2α receptor subunit), CTLA-4, and the glucocorticoid-induced tumor necrosis factor-α receptor (GITR), which when activated, abrogates Treg function (10). Tregs have also been shown to produce immunosuppressive cytokines such as IL-10 and transforming growth factor-β (11). Tregs are primarily identified by intracellular expression of the forkhead/winged-helix transcriptional factor Foxp3, critical for Treg development (12). Tregs can be further classified into natural Tregs and induced Tregs, which differ in their origin/development, requirements for activation with regards to costimulation (natural Tregs may not require costimulation unlike induced Tregs), and potential mechanisms of action (13). Whereas activation of Tregs is considered antigen-specific, their immunosuppressive function is nonspecific, inhibiting multiple phases and events in the immune response from antigen presentation to effector functions through a variety of identified mechanisms (14). Among these mechanisms is up-regulation of CTLA-4 on the surface of Tregs (see Fig. 1), which through cell-to-cell contact, can suppress the activation and expansion of effector cells specific for both normal self and tumor antigens (15, 16). Therefore, whereas CTLA-4–expressing Tregs may play a critical role in maintaining self tolerance, they may also facilitate a level of nonresponsiveness to tumor antigens.
Tregs have been shown to be present in tumor and tumor-draining lymph nodes in animal models (17), and coexist with primed effector T cells in these compartments (18). This presence of Tregs in tumor-draining lymph nodes and tumor provides a potential inhibitory population blocking or “balancing” effector cell function. Thus, depletion of Tregs or blockade of Treg function using targeted antibodies has the potential to remove Treg suppression and enhance antitumor immunity. O'Mahony et al. provide evidence that treatment of patients with ipilimumab has the potential to result in antitumor responses via Treg depletion or blockade, as well as through the direct activation of tumor-specific CD8+ CTL.
Although the limited data set provided failed to show the effects of ipilimumab on tumor antigen–specific CD8 responses, ipilimumab was shown to modulate numbers of Tregs in the periphery. Peripheral blood mononuclear cells were stained for CD4, CD25, and CD62L (or surface L-selectin, which has been identified as a marker of more potent Treg suppression activity) before and after ipilimumab treatment. The number of CD4+CD25+CD62+ Tregs showed a bimodal pattern following treatment. Tregs initially decreased after ipilimumab dosing (within 3 days) by approximately 1/5 to 1/3 of pretreatment levels. Following the third day posttreatment, the number of Tregs approached pretreatment levels, and within 4 weeks (or immediately prior to the next ipilimumab cycle), had returned to or exceeded pretreatment levels. Further analysis was done on one of the non–Hodgkin's lymphoma patient responders using reverse transcription-PCR for Foxp3 mRNA expression among peripheral blood mononuclear cells. Foxp3 mRNA expression showed a similar bimodal pattern to Treg surface marker expression. Previous data from animal models showed similar results of rebounding Tregs within several days following anti-CD25 antibody (19). Kohm et al. recently provided evidence that rather than resulting in Treg depletion, the effects of systemically administered anti-CD25 may be the result of a functional blockade of Tregs mediated by antibody-induced shedding of surface CD25 as opposed to Treg depletion, which has yet to be definitively established (20). Whether anti–CTLA-4 acts through functional blockade of CTLA-4 on effector T cells as opposed to or in addition to depletion of CTLA-4–expressing Tregs remains unresolved. In any case, the data presented by O'Mahony et al. suggests that ipilimumab is capable of blocking Treg function in the peripheral blood and supports the hypothesis that anti–CTLA-4 therapy might affect antitumor immunity through multiple mechanisms including Treg depletion or blockade. The authors propose that the short-lived early effects on Treg function posttreatment may have been sufficient to have allowed the generation of antitumor immune responses, even though their limited data failed to document such an enhancement.
The Clinical Potential for Anti–CTLA-4–Targeted Therapy
O'Mahony et al. provided evidence that ipilimumab decreases the number of Tregs in peripheral blood mononuclear cells, which may correlate with tumor regression. Although the authors did not observe differences in peripheral CD8+ CTL responders as measured by tumor-specific IFN-γ production, this does not eliminate the possibility that tumor-specific effectors were present in the local tumor environment or tumor-draining lymph node. Yang et al. and others have shown in animal models that tumor-specific CTL could be generated in the tumor-draining lymph node of untreated tumor-bearing mice, but that responders were sequestered to the local tumor environment and were not detected in the spleen or peripheral blood (21). Furthermore, Quezada et al. showed in tumor-bearing mice that anti–CTLA-4 antibody in the presence of a recombinant granulocyte macrophage colony-stimulating factor producing vaccine (Gvax) could increase the number of tumor-specific responders in the tumor microenvironment, altering the balance of primed effectors to Tregs, which allowed for an effective antitumor response (22). In this study, O'Mahony et al. reported the presence of increased infiltrates of CD3+CD4+ and CD3+CD8+ T cells into the lymph nodes of one non–Hodgkin's lymphoma patient following ipilimumab treatment. Although this may be consistent with the hypothesized mechanism of action of ipilimumab, without a larger number of patients, this remains speculative.
It should be noted that patients enrolled on the O'Mahony et al. study all had advanced disease and that each patient had failed tumor-specific vaccination likely due both to the extent or bulk of disease as well as the potential activity of one or a number of immune escape mechanisms. Recent preclinical studies from the Levitsky group have also shown that vaccines encoding tumor antigens, given in the context of Treg cell activation, may actually enhance the activity and/or numbers of Tregs and thus further inhibit tumor-specific effector function (23). These findings support the view that tumor vaccines are likely to be more effective when given in the minimal residual disease state in which both tumor burden and tumor-associated immune dysfunction may be at a minimum.
In summary, there is increasing evidence that anti–CTLA-4 immunotherapy has potential clinical benefit. Preclinical studies, and increasingly, clinical trials have shown antitumor activity. Given as a single agent or in combination with vaccine, preclinical studies clearly show the efficacy of anti–CTLA-4 treatment at enhancing antitumor T cell responses and tumor regression. Clinical studies have shown similar single agent responses with enticing data showing possible enhanced activity in patients previously treated with vaccine. There is substantially less data available on the effects of modulating Treg populations. Although the preclinical support for such manipulation is clear, clinical studies are at the preliminary stage of associating Treg numbers with tumor responses. Whereas there have been a small number of studies reporting the use of the agent Ontak, a fusion molecule between the ligand of the CD25 receptor and a toxin, to diminish Treg function in patients, they are extremely limited and show conflicting results (24, 25). Until there is validation of an agent or regimen that functionally inhibits Tregs in man, the role of Tregs in clinical cancer remains unclear. The findings of O'Mahony et al., which show an apparent reduction in cells positive for Treg markers using ipilimumab, provides support that such modulation may be possible. It should be stressed, however, that given the small number of patients studied and the limited immune data, that although the O'Mahony hypothesis is intriguing, it remains to be proven until appropriately powered clinical trials are carried out.
We apologize to our colleagues for not citing many important publications. Due to the limitation in space and the numbers of references, this was not possible.