What Is the Role of Cytotoxic T Lymphocyte–Associated Antigen 4 Blockade in Patients with Metastatic Melanoma?

Abstract

With increasing knowledge of the molecular basis of the immune system and mechanisms of tumor tolerance, novel approaches to treating malignant diseases refractory to standard therapies are being investigated. Monoclonal antibodies (mAbs) that bind cytotoxic T lymphocyte–associated antigen (CTLA)-4 can block inhibitory signals normally generated through this receptor, thus prolonging and sustaining T-cell activation and proliferation. These antibodies are being developed and tested in patients with metastatic melanoma. This article reviews data published or presented at scientific congresses describing the clinical safety and antitumor activity of two different anti–CTLA-4 mAbs: tremelimumab (CP-675,206) and ipilimumab (MDX-010). Overall, although the response rate has not been consistently higher than the response rates associated with other treatments, the induction of durable responses and the favorable safety profile observed with anti–CTLA-4 mAbs are encouraging. However, the true advantage of these new drugs may depend largely on the characterization of predictive biomarkers of activity and subsequent targeting of responsive patients.

Cytotoxic T lymphocyte–associated antigen 4, Melanoma, Metastasis, Biomarkers

Introduction

Although the immune system can detect and destroy cancer cells, tumors develop when malignant cells actively evade the immune system or induce host tolerance [1]. One approach to cancer therapy is to increase the activity of a patient’s own immune system and thus increase the possibility of tumor detection and destruction. However, tumor-induced tolerance can be a barrier to successful immunotherapy. The roles of several critical immunoregulatory elements have recently been elucidated, providing new insight and direction for overcoming tumor tolerance. One approach that has evolved from these insights is to promote T-cell activation by blocking inhibitory signals from receptors such as cytotoxic T lymphocyte–associated antigen (CTLA)-4 [2].

T cells are activated when the T-cell receptor (TCR) complex binds antigen presented by major histocompatibility complex (MHC) I and MHC II on antigen-presenting cells (APCs) (Fig. 1A). T cells also express cosignaling receptors that direct and modulate TCR signals. When the principal costimulatory receptor on the T cell (CD28) interacts with B7 (CD80/86) on the APCs, the costimulatory signal induces T-cell proliferation, cytokine secretion, and changes in gene expression and metabolism (Fig. 1B). CTLA-4 is a key element that induces immune tolerance and the main negative regulator of T-cell–mediated antitumor immune responses [3, 4]. CTLA-4 is a homologue of CD28, with a higher affinity for B7. Binding of CTLA-4 and B7 sends an inhibitory signal that downregulates T-cell activation (Fig. 2A). CTLA-4 is upregulated on activated T cells, where it successfully competes with CD28 for binding to B7 [5, 6]. This sends an inhibitory signal that downregulates T-cell activation and serves as a natural brake on the immune system, altering both downstream cytokine production and the cell-cycle machinery required for proliferation [7–10]. One of the mechanisms by which CTLA-4 signaling induces T-cell inhibition may be through CTLA-4 “back signaling” via B7 into APCs by upregulating the indoleamine 2,3 deoxygenase (IDO) enzyme [11]. This enzyme is responsible for degradation of tryptophan, which is essential for T-cell proliferation. Therefore, stimulation of IDO activity exerts an immunosuppressive stimulus by reducing T-cell proliferation. CTLA-4 is also expressed constitutively on the surface of regulatory T cells (Tregs) and is detectable on approximately 50% of Tregs, compared with <1% on naive helper T cells [12]. It was demonstrated recently that, in a murine model, CTLA-4 ligation on Tregs results in a significant decrease in APC presentation capacity and effector T-cell downregulation [13] (Fig. 2B).

Figure 1

Two signals are required for full T-cell activation. T-cell activation involves binding of T-cell receptor (TCR) to antigen (Ag)-bound major histocompatibility complex (MHC) on the antigen-presenting cell (APC) (A). Full activation also requires binding of costimulatory receptors (e.g., B7 [CD80/86]) on the APC to receptors (e.g., CD28) on the T cell (B).

Figure 2

CTLA-4 blockade prevents downregulation of T cells. After T-cell activation, CTLA-4 receptors upregulate from intracellular stores, compete for binding to B7 (CD80/86), and downregulate T-cell activation. Reverse signaling via CTLA-4 induces an increase in indoleamine 2,3 deoxygenase (IDO) in the antigen-presenting cell (APC) that decreases APC function (A). Ligation of CTLA-4 on regulatory T cells (Tregs) to an APC induces a decrease in APC function and in T-cell proliferation (B).

Abbreviations: Ag, antigen; CTLA-4, cytotoxic T lymphocyte–associated antigen 4; MHC, major histocompatibility complex; TCR, T cell receptor.

The significance of the role of CTLA-4 was revealed with murine knockout models. Mice lacking CTLA-4 exhibit massive T-cell proliferation, have an enlarged spleen and lymph nodes, and develop lymphoproliferative disease [14–16]. Furthermore, blockade of CTLA-4 has been shown to enhance T-cell functions, such as cytokine production, both in vitro and in vivo. In a murine model of cancer, anti–CTLA-4 monoclonal antibody (mAb) treatment inhibited the growth of new or established tumors and resulted in greater immunity to secondary exposure with tumor cells [17]. Based on preclinical data, it was hypothesized that blocking the interaction of CTLA-4 and B7 in humans would sustain T-cell activation and proliferation, thereby preventing tumor tolerance and potentially increasing antitumor activity.

CTLA-4 Blockade

mAbs directed against CTLA-4 can have a higher affinity for CTLA-4 than B7 [6]. These mAbs successfully compete for binding to CTLA-4, blocking the inhibitory signal that would have arisen from binding to B7 [18]. With no inhibitory signal present, the brake on T-cell activation is released and T-cell proliferation is sustained, thereby potentially promoting an antitumor response (Fig. 3A, 3B) [19]. Two fully human anti–CTLA-4 mAbs are in clinical development for the treatment of patients with metastatic melanoma as well as other tumor types and have demonstrated antitumor activity as single agents and in combination with other agents: tremelimumab (CP-675,206; Pfizer, Inc., New London, CT), a fully human IgG₂ mAb, and ipilimumab (MDX-010; Medarex, Princeton, NJ), a fully human IgG_1κ mAb.

Figure 3

Antitumor effects of CTLA-4 blockade. Anti–CTLA4 monoclonal antibodies (mAbs) inhibit interaction of B7 (CD80/86) and CTLA-4, thus prolonging T-cell activation (A). Inhibition of CTLA-4 on Tregs induces T-cell proliferation and has antitumor effects (B) [13].

Abbreviations: Ag, antigen; APC, antigen-presenting cell; CTLA-4, cytotoxic T lymphocyte–associated antigen 4; MHC, major histocompatibility complex; TCR, T cell receptor; Treg, regulatory T cell.

Clinical Development of Anti–CTLA-4 mAbs in Patients with Metastatic Melanoma

Both tremelimumab and ipilimumab have been investigated in clinical trials for patients with metastatic melanoma (Tables 1 [2, 20–27] and 2 [28–39]). These studies have evaluated a variety of doses and schedules of anti–CTLA-4 mAbs as single agents or in combination with other cancer therapies [25, 27, 28, 40].

Table 1

Clinical trials with tremelimumab in patients with metastatic melanoma

^aFor the objective response rate, the denominator for A3671009 is all randomized patients, whereas the denominator for all other studies is response-evaluable patients.

^bSurvival data calculations were based on number of enrolled patients.

Abbreviations: AE, adverse event; CR, complete response; DC, dendritic cell; IFN, interferon; MART-1, melanoma antigen recognized by T cells 1; NR, not reported; PR, partial response; Q1M, every month; Q3M, every 3 months.

Table 2

Clinical trials with ipilimumab in patients with metastatic melanoma

^aThese trials only reported immune-related adverse events.

^bThis trial reported grade 3 or 4 treatment-related AEs.

^cNumbers represent the median overall survival duration in treatment-refractory patients; the median overall survival time in treatment-naive patients was not reached.

Abbreviations: AE, adverse event; CR, complete response; DTIC, dacarbazine; gp, glycoprotein; IL, interleukin; NR, not reported; Q12W, every 12 weeks; Q3W, every 3 weeks; Q28D, every 28 days.

Safety of CTLA-4 Blockade in Patients with Metastatic Melanoma

The most commonly reported adverse events (AEs) after anti–CTLA-4 mAb therapy in patients with advanced melanoma include diarrhea/colitis, dermatitis/rash, pruritus, nausea, vomiting, and fatigue [2, 30, 41–43] (Table 3 [23, 24, 35, 38]). Grade 3 or 4 AEs associated with treatment include diarrhea, dermatitis/rash, fatigue, fever, abdominal pain, nausea, hypophysitis, and hepatitis [2, 30, 34, 42, 44, 45]. When agents are used that enhance patient immunoreactivity, it is expected that some level of reactivity to self-antigens will occur, and most toxicities commonly observed with CTLA-4 blockade are considered immune related AEs (IRAEs) [22, 34, 40, 46, 47]. IRAEs appear to correlate with clinical benefit [30, 48]; however, at this time, it is not known whether the appearance of IRAEs in patients receiving anti–CTLA-4 mAbs is associated with longer survival. Most IRAEs experienced by patients treated with CTLA-4 blockade are reversible following cessation of treatment [19].

Table 3

Safety of tremelimumab and ipilimumab in large trials (n ≥ 100)

Table 3

(Continued)

^aAdapted from Kirkwood JM, Lorigan P, Hersey P et al. A phase II trial of tremelimumab (CP-675,206) in patients with advanced refractory or relapsed melanoma [poster and discussion, abstract 9023]. Presented at the 44th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 30 to June 3, 2008.

^bThese included one sudden death and one diverticular perforation.

^cAdapted from Ribas A, Hauschild A, Kefford R et al. Phase III, open-label, randomized, comparative study of tremelimumab (CP-675,206) and chemotherapy (temozolomide [TMZ] or dacarbazine [DTIC]) in patients with advanced melanoma [oral presentation, abstract LBA9011]. Presented at the 44th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL. May 30 to June 3, 2008.

^dAdapted from Hamid O, Chin K, Li J et al. Dose effect of ipilimumab in patients with advanced melanoma: Results from a phase II, randomized, dose-ranging study [poster and discussion, abstract 9025]. Presented at the 44th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 30 to June 3, 2008.

^eDeath caused by grade 3 respiratory infection possibly related to treatment; patient was a tobacco user with lung metastases.

^fAdapted from Weber JS, Berman D, Siegel J et al. Safety and efficacy of ipilimumab with or without prophylactic budesonide in treatment-naive and previously treated patients with advanced melanoma [oral presentation, abstract 9010]. Presented at the 44th Annual Meeting of the American Society of Clinical Oncology, Chicago, IL, May 30 to June 3, 2008.

Abbreviations: AE, adverse event; IRAE, immune-related adverse event; Q90D, every 90 days.

Gastrointestinal toxicities (e.g., diarrhea and colitis) are commonly associated with CTLA-4 blockade [2, 30, 43, 48, 49]. Evidence of neutrophilic and lymphocytic inflammation in gastrointestinal biopsies suggests that these toxicities are IRAEs [48]. In a phase II study (n = 155) of patients treated with 10 mg/kg ipilimumab every 3 weeks for four cycles followed by maintenance dosing every 12 weeks, grade 3 or 4 AEs were reported by 21.9% of patients, and the majority of these were gastrointestinal AEs (8.4% of total patients) [36]. In addition, in an analysis of safety data from eight completed and ongoing trials of patients (n = 786) who received tremelimumab (15 mg/kg), diarrhea was the most common treatment-related AE (occurring in 40% of patients), and 11% of patients had grade ≥3 diarrhea [43]. Although cases of colitis have developed into colon perforation in approximately 1%–2% of patients treated with CTLA-4 blockade [34, 48, 49], the majority of cases of diarrhea/colitis resolve fully. Most events are mild to moderate in severity.

Severe enterocolitis is usually alleviated by systemic steroid treatment and resolves within approximately 1–2 weeks [19, 45, 49]. In cases refractory to steroids, tumor necrosis factor α blockade with infliximab has been effective [48]. Although some studies have reported that systemic steroid administration does not affect the antitumor activity of CTLA-4 blockade [48, 49], a recent study indicated a trend toward a shorter median duration of response in responders receiving high-dose systemic steroids (19.3 months versus 30.6 months for all responders) [50]. Furthermore, in a phase II trial investigating prophylactic budesonide concurrent with ipilimumab in 115 patients with advanced melanoma, the rate of grade ≥2 diarrhea was not lower [38]. Analysis of ongoing trials with larger sample sizes will provide more definitive guidelines for the management of AEs.

In studies of tremelimumab and ipilimumab, there have been deaths that appeared to be related to treatment. Two patients in a study of ipilimumab (n = 198) in patients with melanoma or renal cell carcinoma developed enterocolitis that led to colonic perforation [48]. One of these patients died from sepsis, and the other chose comfort care because of cancer progression. In a phase II study of patients with melanoma who were treated with tremelimumab (n = 246), there were two treatment-related deaths [23]. One was considered sudden death, and one was the result of diverticular perforation. The deaths resulting from colonic perforation emphasize the importance of careful management of gastrointestinal AEs.

Patients treated with anti–CTLA-4 mAbs may experience rash with or without pruritus, and a small percentage of patients have reported grade ≥3 rash [30, 43, 44]. The cutaneous presentation is often a nonspecific, diffuse maculopapular rash in which pathologic examination may reveal infiltration of CD4⁺ and CD8⁺ T cells [30, 51]. Symptomatic treatment with hydroxyzine or diphenhydramine seems to provide relief for some patients [30], but the rash can also resolve spontaneously [2].

Infrequently occurring AEs such as uveitis and hypophysitis may require clinical intervention. Among 163 patients treated with ipilimumab in one study, 5% of patients developed autoimmune hypophysitis [52]. In the cross-study safety analysis of tremelimumab (n = 786), the incidence of hypophysitis was <1% [43]. Autoimmune hypophysitis may require steroids and/or hormone therapy [52], and this may not be a reversible condition. Ocular AEs can be resolved with topical corticosteroids [30, 41, 51–53], whereas the symptoms of hypophysitis have resolved with discontinuation of therapy and physiologic hormone replacements [52]. CTLA-4 blockade has also occasionally caused autoimmune thyroiditis, and the outcome of these toxicities has varied. Some cases are manageable and self-limiting, whereas other cases have led to partial or total loss of function requiring long-term hormonal supplementation [2, 30]. Hepatitis [22, 24, 52, 54, 55], which may be an IRAE [47], and nephritis [48] have also been reported in some patients with melanoma treated with anti–CTLA-4 mAbs.

Overall, these toxicities are usually manageable, treatment with anti–CTLA-4 mAbs is tolerated by patients with metastatic melanoma, and AEs associated with CTLA-4 blockade are generally less severe than those observed with interleukin (IL)-2 or interferon (IFN)-α. Furthermore, unlike administration of IL-2, which requires intensive hemodynamic monitoring of patients, administration of anti–CTLA-4 mAbs is possible in a community hospital setting.

Efficacy of CTLA-4 Blockade in Patients with Metastatic Melanoma

Standard therapy for melanoma with single-agent dacarbazine yields objective responses in approximately 7.5% of patients, and few patients have long-term benefit (durable response rate, 3.6%) [56]. High-dose immunotherapy with the cytokine IL-2 can induce long-term, durable, complete responses (CRs) in a small percentage (6%) of patients [57–59], but it is very poorly tolerated. High-dose IFN-α2a is no more effective than single-agent dacarbazine in patients with malignant melanoma [60]. In phase II trials, the objective response rates of patients with advanced melanoma to IFN-α2a were 8%–22%, but responses were rarely durable [60, 61]. Therefore, treatment options such as anti–CTLA-4 mAbs that have the potential for improving the prognosis of patients with metastatic melanoma are needed.

CTLA-4 blockade is associated with objective responses in patients with metastatic melanoma (Tables 1 and 2). In a phase II trial of tremelimumab (15 mg/kg administered once every 3 months) in patients (n = 251) with previously treated, stage III or IV melanoma (Study A3671008), 7% of patients had an objective response per independent review (16 partial responses [PRs]), and responses lasted 91–540+ days. The median overall survival time was 10.1 months, and the 1-year survival rate was 41% [23]. Several phase II studies of ipilimumab in patients with advanced melanoma have also demonstrated antitumor activity of anti–CTLA-4 mAbs. In a single-arm, phase II study of ipilimumab in patients (n = 155) who had failed IL-2 or chemotherapy, the best overall objective response rate was 6% (9 PRs) [36]. The median overall survival time was 10.22 months, and the estimated 1-year survival rate was 47% [23, 36].

Anti–CTLA-4 mAbs are also being investigated in ongoing and completed phase III studies. The results of a phase III randomized study of 15 mg/kg tremelimumab every 3 months (n = 328) and the physician’s choice of chemotherapy (n = 327) with either dacarbazine or temozolomide in treatment-naive patients with advanced melanoma were recently reported (A3671009) [24]. The primary endpoint was overall survival. At the second interim analysis, the trial was halted based on recommendation from the data safety monitoring board because the log-rank test statistic (p = .729) crossed the prespecified O’Brien–Fleming boundary for futility (p > .473). Although the median survival time was longer (11.76 months) in patients treated with tremelimumab than in patients given chemotherapy (10.71 months), the difference was not statistically significant (hazard ratio for chemotherapy/tremelimumab, 1.04; p = .729) [24]. Patients with clinical benefit from tremelimumab treatment are continuing on study, and more mature survival and response data are anticipated.

Overall, ∼7% of patients achieve objective responses following single-agent treatment with anti–CTLA-4 mAbs, but the critical characteristic of this treatment is that objective responses and disease stabilization with anti–CTLA-4 mAbs seem to be durable (≥6 months in most patients) [2, 20, 22, 28, 50, 62]. Responses may take as many as 12 weeks or more to develop [63]. Indeed, late-onset objective responses are sometimes preceded by months of stable disease (SD) and have even occurred after disease progression [63]. Such delayed responses may relate to the time required for the immune system to mount an antitumor response. A number of different patterns of responses to CTLA-4 blockade have been observed, including SD with a slow decline in lesion burden, responses in some lesions with progression in others, response in baseline lesions, response after initial increase in tumor burden, and responses in index and new lesions after appearance of new lesions [23, 28, 37, 64]. It is possible that standard response criteria may not adequately capture clinical benefit, and caution must be taken to avoid premature treatment termination of patients who may benefit from continued therapy [37, 64]. Therefore, barring rapid disease progression, it has been proposed that patients should continue to receive anti–CTLA-4 mAb therapy for at least 12 weeks [28]. This modified response profile implies that the evaluation criteria for treatment efficacy should also be adapted to the quality and delay of responses observed. Nevertheless, it is encouraging that, as seen in Tables 1 and 2, the median survival duration in studies of patients with metastatic melanoma is often longer than expected (median, 6.2 months in a recent meta-analysis) [65] and may exceed 12 months [34].

Combination Therapy with Anti–CTLA-4 mAbs

Anti–CTLA-4 mAbs have also been investigated in combination with peptide vaccines [30, 41, 55], dacarbazine [31], or cytokines [25, 33] in patients with metastatic melanoma (Tables 1 and 2). In a phase II study, chemotherapy-naive patients with advanced melanoma were treated every 28 days with 3 mg/kg ipilimumab alone (n = 37) or in combination with dacarbazine (n = 35) [28, 31, 32]. The overall survival time was longer in patients treated with ipilimumab and dacarbazine than in those treated with ipilimumab alone (15.0 months versus 11.7 months, respectively) [32]. This combination is currently being tested in a large randomized phase III study. Ipilimumab has been tested in combination with immunotherapies such as IL-2 or glycoprotein (gp)100 peptide vaccines [28, 30]. In a phase II trial for patients with metastatic melanoma (n = 56) receiving ipilimumab and vaccination with gp100 peptides, seven (13%) patients had an objective response [30]. There was no significant difference in the response rate or toxicity between the two regimens tested (3 mg/kg ipilimumab plus gp100 peptides administered once every 3 weeks compared with a single administration of the combination at these doses followed by maintenance dosing with 1 mg/kg ipilimumab plus gp100 peptides) [30]. Ipilimumab has also been tested in combination with IL-2 therapy [33]. That phase I/II study tested several dose levels of ipilimumab (0.1–3.0 mg/kg) administered every 3 weeks in combination with 720,000 IU/kg IL-2 administered every 8 hours. Although eight (22%) patients had an objective response (three CRs, five PRs), this response rate suggested an additive, but not synergistic, antitumor effect of CTLA-4 blockade plus IL-2 [19, 33].

Tremelimumab has also been studied in combination with other immunotherapeutic agents. In a phase I trial, patients (n = 16) with metastatic melanoma received 3–10 mg/kg tremelimumab monthly or 10–15 mg/kg tremelimumab quarterly in combination with an intradermal administration of 1 × 10⁷ autologous cells pulsed with melanoma antigen recognized by T cells (MART)-1_26–35 peptide. There were four objective responses (one CR, three PRs), and all responses were durable [27]. In a phase II trial in patients (n = 16 to date) with recurrent, inoperable stage III or IV melanoma, tremelimumab (15 mg/kg quarterly) plus high-dose IFN-α2b was associated with a 19% objective response rate (three PRs lasting 5.0+, 8.0+, and 9.0 months), and six additional patients achieved SD (lasting 1.5–9.0+ months) [25]. Ongoing clinical trials are evaluating the combination of other agents (e.g., Toll-like receptor 9 agonists) with anti–CTLA-4 mAbs.

Mechanism of Action: T Cells

One of the major issues that has hindered our understanding of CD28–CTLA-4 signaling is understanding the exact nature of its underlying mechanism of action. CTLA-4 may compete with CD28 for B7 ligation [12, 66]. Thus, in situations in which CTLA-4 is highly expressed, it is likely that B7.1 or B7.2 will bind preferentially to CTLA-4 and thereby deprive T cells of CD28 costimulatory signaling (Fig. 3A). CTLA-4 may also transduce an inhibitory signal in T cells by removing TCR signaling components, such as the ζ chain of the TCR, from lipid rafts [67]. Recent studies have indicated that CTLA-4 blockade is associated with decreases in Treg-mediated immune suppression [12, 13], suggesting that the T-cell activation observed with anti–CTLA-4 mAbs may occur, in part, because Tregs no longer modulate dendritic cell (DC) inactivation and T-cell downregulation [13] (Fig. 3B). Tregs expressing CTLA-4 might inhibit T-cell activation by altering the physical interactions between T cells and DCs [68]. Moreover, CTLA-4 can be expressed on the surface of melanoma cells [69], and it is not yet known whether there is a direct effect of the antibodies on the melanoma cells or whether CTLA-4 expression on the target cells could alter drug effects.

Most studies of immune responses after cancer immunotherapy have been performed on patient peripheral blood samples. Immunoassays allow quantitative assessment of the phenotype of immune effector cells in peripheral blood, their ability to bind to defined major histocompatibility complex-peptide determinants, and their functional responses on exposure to antigen [70–72]. The memory T-cell pool is increased after therapy with either ipilimumab [30, 55] or tremelimumab [73], as demonstrated by the increase in number of CD45RO⁺ T cells. Both antibodies were also shown to increase expression of the activation marker human leukocyte antigen (HLA)-DR on the surface of T cells, and an increase in the number of activated T cells has been observed [30, 55, 73]. However, in several studies of peripheral blood samples from patients treated with anti–CTLA-4 mAbs, expansion of tumor antigen-specific T cells was not observed [26, 33, 55, 74–76]. Tregs are more difficult to study at the cellular level because they have a surface phenotype that is indistinguishable from chronically activated helper T cells. Positive identification is possible by detection of intracellular expression of the Treg-specific transcription factor FoxP3 [77, 78]. No major modifications in the number or function of Tregs were noted [26, 75]. In a cohort of 10 patients treated with tremelimumab, there was no decrease in the number or function of Tregs, but T cells became resistant to the immunosuppressive effect of Tregs after treatment [73]. The mechanism underlying this acquired Treg unresponsiveness remains to be elucidated. However, studies performed on peripheral blood might not reflect the biologic effect of CTLA-4 blockade on the tumor-associated immune system, and analysis of tumor biopsies may provide a better indication of the role of CTLA-4 blockade on T cells and of their action on metastases.

Analyses of tumor biopsies have been undertaken in a few patients treated with CTLA-4–blocking antibodies, and these studies have revealed marked intratumoral changes and the presence of inflammatory infiltrates [2, 51, 76, 79]. CD4⁺ and CD8⁺ T-cell infiltrates have been identified in tumor biopsies [2, 51, 76, 79]. In a study of peripheral blood and tumor tissues from patients with bladder cancer who were treated with CTLA-4 blockade, consistently higher IFN-secreting CD4⁺ICOS^hi T cells and a greater ratio of effector T cells to Treg cells in treated patients than in untreated healthy volunteers were identified [80]. Although consistent association of infiltration of a particular immune cell type and clinical benefits from CTLA-4 blockade has been a challenging goal, a study in a subset of patients (n = 7) with melanoma treated with tremelimumab in a phase I/II study found an association between antitumor effects and the presence of infiltrating CD8⁺ T cells in tumor biopsies [22, 79].

Remaining Questions

Biomarkers of Efficacy

One concern for clinicians when treating their patients with an anti–CTLA-4 mAb is the lack of a commonly used biomarker, other than tumor size, to indicate potential efficacy. A number of cytokines and chemokines have been assessed in treated patient sera, but none has been found to be consistently correlated (from patient to patient and from trial to trial) with objective response. There have also been conflicting reports regarding the association of IRAEs and clinical responses [30, 81]. The small number of objective responses do not currently allow for definitive conclusions.

Administration of CTLA-4–blocking antibodies to patients with melanoma does not result in sustained expansion of circulating melanoma antigen–specific CD8⁺ T cells [30, 55, 75, 76]. Markers of T-cell activation (HLA-DR) and memory phenotype (CD45RO) after dosing with tremelimumab have been shown to segregate patients with clinical benefit (n = 3) from nonresponders (n = 9) [26]. In that study, tremelimumab did not increase the number or function of antigen-specific CD8⁺ T cells in peripheral blood nor decrease FoxP3 transcripts but appeared to generally enhance T-cell activation and differentiation. In our experience, observed increases in HLA-DR and amplification of the T-cell memory pool did not correlate with clinical benefit [73].Therefore, it is unlikely that detection of T-cell activation and memory markers may be useful as a readout of efficacy. Further translational immunological studies on larger cohorts of patients are necessary to determine whether resistance to the immunosuppressive effects of T cells is associated with clinical response [73] and to identify other, more convenient biological markers of efficacy.

In a recent report, immunologic monitoring was performed on patients with melanoma who had been treated with ipilimumab (n = 15) [82]. Eight of those patients had evidence of clinical benefit, and seven did not respond. Five of the eight patients who responded to ipilimumab were seropositive for NY-ESO-1 (anti-cancer testis antigen 1B) and had CD4⁺ and CD8⁺ T-cell responses specific to NY-ESO-1 following treatment, whereas the nonresponders were seronegative (one nonresponder did have a CD4⁺ T-cell response to NY-ESO-1) [82]. The authors suggested that these data might provide a rationale for combination therapy with CTLA-4 blockade and NY-ESO-1 vaccination; however, further confirmation in larger patient cohorts is needed.

Side Effects: Association with Tumor Response, Prognostic Factors

Several studies with CTLA-4 blockade have suggested a correlation between antitumor effects and autoimmune phenomena such as thyroiditis, hypophysitis, enteritis, hepatitis, diarrhea/colitis, and rash [30, 41, 55, 83]. The incidence of IRAEs has been associated with antitumor activity in several clinical trials. For example, Attia and colleagues found that five of 14 (36%) patients with metastatic melanoma who experienced grade 3 or 4 autoimmune toxicity experienced an objective response, compared with two (5%) responses among the 42 patients with no IRAE (p = .008) after treatment with ipilimumab and gp100 peptides [30]. Beck and colleagues also reported a significant difference in the objective response rate between patients with enterocolitis and patients without enterocolitis among 137 melanoma patients treated with ipilimumab and gp100 peptides [48]. However, these reports may have introduced bias because responding patients may remain on treatment with a longer period of exposure and have a higher possibility of developing an IRAE than patients who progress or die early, and this association between side effects and antitumor efficacy needs to be confirmed in larger prospective studies.

There is presently no known means to predict the incidence of an AE or to exclude a population at high risk of developing a serious AE; however, patients with inflammatory bowel disease are thought to be at a higher risk for gastrointestinal problems and should be excluded from treatment. Improving our knowledge of the mechanisms underlying these AEs may help to achieve this objective, and several ongoing clinical studies of tremelimumab and ipilimumab are further investigating mechanisms, prevention, and management of drug-induced toxicities.

Dose Level and Administration Schedule

Another aspect of clinical testing of anti–CTLA-4 mAbs is the determination of an optimized dose and administration schedule. Although the study was not powered for a direct comparison, single-agent tremelimumab at a dose of 15 mg/kg every 3 months was associated with a lower incidence of grade 3 or 4 treatment-related AEs than 10 mg/kg tremelimumab administered monthly (13% versus 27%) in the phase I/II trial; this difference was not statistically significant, but the 15-mg/kg dose administered every 3 months regimen was used in subsequent clinical studies [22–24]. Conserved efficacy with a longer interval between doses may be related to the relatively long plasma half-life of tremelimumab (22 days) [2]. Ipilimumab, which has a plasma half-life of 12–15 days [34, 84], has been administered at concentrations of 0.1–20 mg/kg with a variety of dosing schedules (Table 2). The presently accepted treatment regimen is 10 mg/kg administered once every 3–4 weeks. A recent report on a phase II study of ipilimumab confirmed that, among patients treated with 0.3, 3, and 10 mg/kg, there was a statistically significant trend indicating a higher best objective response rate with a higher dose (p = .0015). Patients in the 10-mg/kg ipilimumab dose cohort also had a longer median overall survival time than patients in the other cohorts; however, the rates of IRAEs were also highest in the 10-mg/kg cohort [35].

Both anti–CTLA-4 mAbs are being tested in combination with other anticancer therapies, and these trials use roughly the same schedules and dose levels as the single-agent trials. As new combinations of agents with anti–CTLA-4 mAbs are tested, more consideration must be placed on the administered dose and schedule. For example, sequential combinations may be more appropriate with chemotherapy, which could reduce the tumor burden and increase the probability of antitumor activity with subsequent CTLA-4 blockade immunotherapy.

However, the classical dose-escalation designs used for conventional chemotherapies may not be adaptable for anti–CTLA-4 immunotherapy, and because we do not have reliable biologic markers of activity, we must admit that the choice of dose level and administration schedule remains somehow empirical. Until reliable biologic or clinical markers of activity are identified, our ongoing awareness of the most recent clinical experience and success of various therapies will aid in the selection of rational combinations with the highest likelihood of antitumor activity.

Conclusions

CTLA-4 blockade has been associated with rare, but impressive and durable, objective responses in patients with metastatic melanoma. Even if the immediate objective response rates obtained with anti–CTLA-4 mAbs do not seem to be significantly higher than those observed with conventional chemotherapy or high-dose IL-2, the high frequency of durable responses is promising. This novel immunotherapeutic approach may also generate a T-cell memory antitumor response, potentially increasing long-term survival. Furthermore, CTLA-4 blockade is associated with a more favorable toxicity profile than high-dose IL-2. Anti–CTLA-4 mAbs are also being investigated as adjuvant therapy for patients with melanoma (Table 4). These studies will determine if this early approach to overcome tumor-induced immune tolerance can improve clinical outcomes for patients with melanoma.

Table 4

Proposed studies of anti–CTLA-4 mAbs in the adjuvant setting

Abbreviations: CTLA-4, cytotoxic T lymphocyte–associated antigen 4; EORTC, European Organization for Research and Treatment of Cancer; mAb, monoclonal antibody; SWOG, Southwest Oncology Group.

Altogether, anti–CTLA-4 mAbs appear to be very promising new therapeutic agents for melanoma. The relatively low objective response rate should not be regarded as a failure, but rather as an encouragement to develop clinical and translational research to understand the exact mechanisms underlying both the biologic antitumor activity of anti–CTLA-4 mAbs and the IRAEs associated with these drugs, to identify reliable markers of activity, and to potentially allow selection of a population of patients with a high probability of response and a low risk for severe AEs.

Author Contributions

Conception/Design: Caroline Robert

Data analysis: Caroline Robert

Manuscript writing: Caroline Robert, Francois Ghiringhelli

Final approval of manuscript: Caroline Robert, Francois Ghiringhelli

The authors take full responsibility for the content of the paper, but thank Tamara Fink, Ph.D., of ProEd Communications, Inc. for her assistance in collating the comments of the authors and organizing the published literature.

Author notes

Disclosures: Caroline Robert:  Consultant/advisory role: Bayer, Pfizer, Bristol-Myers Squibb; Honoraria: Bayer, Pfizer, Bristol-Myers Squibb; Francois Ghiringhelli: None.

The article discusses unlabeled, investigational, or alternative use of a product, device or technique: tremelimumab (CP-675,206, Bristol-Myers Squibb) and ipilimumab (MDX-010, Pfizer) for metastatic melanoma.

The content of this article has been reviewed by independent peer reviewers to ensure that it is balanced, objective, and free from commercial bias. No financial relationships relevant to the content of this article have been disclosed by the independent peer reviewers.