High-Dose Interleukin-2: Is It Still Indicated for Melanoma and RCC in an Era of Targeted Therapies?
High-Dose Interleukin-2: Is It Still Indicated for Melanoma and RCC in an Era of Targeted Therapies?
ABSTRACT: Immunotherapy with interleukin-2 (IL-2) has been the mainstay of systemic therapy for advanced kidney cancer and melanoma. Although IL-2 treatment is limited to healthy patients, a select group of these patients have derived substantial, durable benefit from it—in some translating into cures with no ongoing therapy or chronic toxicity. Over the past 10 years, insights into the biology of renal cell carcinoma and into key signaling mechanisms in melanoma, and growth in our understanding of immune checkpoints, have led to the development and approval of targeted and immune-modulatory therapeutic options with clinically relevant benefit. Our improved understanding of the relationship between the host environment, immune system, and malignancy has helped identify compounds and therapies that are changing the way we think about cancer and our approach to cancer therapeutics. While the newer options may be applicable to most patients, durable responses measured in years are rare. In this review, we examine the currently approved options available for these disease processes, including the newer agents and selected combinatorial approaches under investigation, and we attempt to identify the role of high-dose IL-2 in the context of current clinical practice.
The generation of antitumor immune response by T lymphocytes is a complex multi-step process that is modulated by several signals. Primary signal requires that antigen be presented by antigen-presenting cells (APCs) in the context of self–human leukocyte antigen (HLA) molecules—for recognition by T cells. HLA class I molecules present antigens to CD8+ T lymphocytes, and class II molecules present antigens to CD4+ T lymphocytes. CD4+ T helper cells subsequently produce a number of cytokines, including interleukin-2 (IL-2), that can propagate the immune response (Figure A). IL-2, a nonspecific T-cell growth factor, can result in the expansion of all T-cell subsets, including T-cytotoxic, helper, and regulatory cells. After initial recognition of the antigen, the T-cell function is modified by complex interactions between co-stimulatory or co-inhibitory proteins and their ligands expressed on the APCs (or in some cases tumor cells) and on the T cells; these interactions constitute the secondary regulatory signal that fine-tunes the immune response. A number of immune-modulatory proteins, termed checkpoints, have been identified; the list continues to expand as we get better insights into the anatomy of the immune response; some of these checkpoints are illustrated in Figure 1. Most of these proteins belong to the immunoglobulin superfamily or to the tumor necrosis factor (TNF)/TNF receptor (TNFR) family. While not inherently active, these regulatory molecules exert their effect via the recruitment of TNFR-associated factor (TRAF) adapter proteins. Cytotoxic T lymphocyte antigen 4 (CTLA-4) and programmed death protein 1 (PD-1) have been targeted successfully in cancer immunotherapy; several other checkpoints are under active investigation.
Immunotherapy With IL-2–Based Approaches
Malignant melanoma and renal cell carcinoma are immunologically modulated malignancies, as demonstrated by observations of spontaneous regression without systemic intervention.[1,2] These malignancies show relative resistance to cytotoxic chemotherapy, and therefore, until recently, patients have had limited treatment options.
IL-2 was initially described as a growth factor that was necessary and sufficient for T-cell growth and activation. Identification of IL-2 as a critical T-cell growth factor and its subsequent availability for clinical use facilitated the development of immunotherapeutic approaches against diseases amenable to immune modulation.[4-6] The Surgery Branch of the National Cancer Institute (NCI) demonstrated in 1985 that administration of exogenous IL-2 could result in modulation of the immune response with generation of durable regression of established tumors in a murine model.
Based on the animal model, IL-2, with or without the addition of lymphokine-activated killer (LAK) cells, was administered to patients with advanced renal cell carcinoma and metastatic melanoma, eliciting responses, with a small proportion of these measured in years.
Seven phase II studies, including 255 patients with metastatic renal cell carcinoma who were treated with high-dose (HD) IL-2 (600,000–720,000 IU/kg) yielded an overall response rate of 15%, with 7% complete responders. Median duration of complete responses could not be calculated at that time because > 50% of the patients had not experienced recurrence of their disease. Based on the duration of response, HD IL-2 was approved by the US Food and Drug Administration (FDA) in 1992 for the treatment of advanced renal cell carcinoma. These findings have been confirmed by the Cytokine Working Group, as well as by our group in 104 patients (Figure 2). Updated response data from patients with advanced renal cell carcinoma who were treated at the Surgery Branch of the NCI between 1986 and 2006 include some responses that have lasted over 20 years. The above studies are summarized in Table 1.
The data set that led to FDA approval of HD IL-2 for melanoma consisted of a series of eight phase II studies that included a total of 270 patients who were treated with HD IL-2 at doses of 600,000–720,000 IU/kg. The overall response rate was 16%; 6% had a complete response. The median duration of response had not been reached at the time of reporting. Thus, HD IL-2 was approved for advanced melanoma in 1998. The Cytokine Working Group conducted three phase II studies of HD IL-2 (at a dose of 600,000 IU/kg) along with gp100 peptide that showed an overall response in 16%; complete response was observed in 9%. In 684 consecutive patients with metastatic melanoma treated at the NCI with HD IL-2 alone or in combination with a vaccine, the overall response rate was 13% for patients who received IL-2 alone and 16% for patients who received IL-2 with the vaccine. A phase III study comparing HD IL-2 alone to HD IL-2 with gp100 vaccine showed the overall response rate to be 6% for IL-2 alone and 16% for the combination. Our own experience with HD IL-2 in 117 patients has been consistent with the above outcomes (see Figure 2). These data are summarized in Table 2.
Despite these responses and the likely cures for some patients, adoption of HD IL-2 treatment has been hampered by its toxicity profile—primarily capillary leak syndrome, which manifests as oliguria, mild hypoxemia, generalized edema, tachyarrhythmias, and hypotension. Toxicities also include fever, nausea, diarrhea, catheter-related sepsis, and death. These toxicities require clinical expertise and hospitalization for monitoring but can be managed with limited mortality.[18,19] In our 17-year experience treating 221 patients with 744 cycles of HD IL-2, only one treatment-related death has occurred.
In order to mitigate the toxicities of HD IL-2, several studies have been performed with low doses of IL-2 in both malignant melanoma and renal cell carcinoma. However, these have consistently shown the apparent superiority of HD IL-2 as the optimal regimen for appropriately selected patients.[11,20,21]
CTLA-4 Checkpoint Inhibition
Cytotoxic T lymphocyte antigen 4 (CTLA-4) is an inhibitory checkpoint that is expressed on activated T cells. T-cell activation requires presentation of antigen in the appropriate context by APCs; this constitutes the initial signal in the activation process. Once the antigen is recognized as non-self, a secondary regulatory interaction occurs between CD28 expressed on the T cell and molecules in the B7 family (CD80, CD86) on the APC. This constitutes a stimulatory signal that results in the activation of the T cell. Subsequent down-regulation of the T-cell activation ensues when the co-inhibitory regulator CTLA-4 is expressed on activated T cells. CTLA-4 binds to CD80 and CD86 with much higher affinity than does CD28, thus displacing CD28 and leading to inhibition of the T cell (Figure B). Blocking the CTLA-4 checkpoint results in unrestrained activation of the T cell, which, when appropriately harnessed, translates into enhanced antitumor activity.
Ipilimumab (Yervoy) is a fully human immunoglobulin G1 (IgG1) antibody that binds to the CTLA-4 molecule and blocks it (Figure 3). Demonstration that CTLA-4 blockade could result in clinically meaningful tumor regression in patients with advanced melanoma led to its clinical development in this setting. Monotherapy with ipilimumab in the phase II setting has shown overall responses in the range of 6% to 16%, and the disease control rate (defined as the percentage of complete responses, partial responses, and stable disease) has ranged between 27% and 32% in patients with advanced melanoma, including initially surprising response rates in visceral disease, including in the liver. The overall 5-year survival rate for treatment-naive patients has reached as high as 49%.[24,25] In a phase III study, ipilimumab alone was compared with the glycoprotein 100 peptide (gp100) vaccine and with a combination of both. The median overall survival in the ipilimumab-containing arms of the trial was 10 months, compared with 6.4 months for the vaccine-alone arm. The overall survival rates for ipilimumab alone vs gp100 alone were 45.6% vs 25.2% at 12 months, and 23.5% vs 13.7% at 24 months. The disease control rate was 28.5% in the ipilimumab-alone arm and 11% in the gp100 vaccine−alone arm. This was the first phase III study to show a survival benefit for immunotherapy in the setting of advanced melanoma, and it led to FDA approval of ipilimumab for this indication in 2011. This study allowed individuals to receive reinduction with the same regimen if they demonstrated clinical benefit and then subsequent progression. Thirty-one individuals met the criteria and were reinduced. Those reinduced patients who had previously received ipilimumab alone showed a response rate of 37.5% and a disease control rate of 75%.
In a second phase III study, the combination of ipilimumab and dacarbazine was compared to dacarbazine alone. The median overall survival in the combination arm was 11.2 months compared with 9.1 months in the dacarbazine-alone arm. The estimated survival rate for the combination was 47.3% at 1 year, compared with 36.3% for dacarbazine alone; the rates were 28.5% and 17.9% at 2 years, and 20.8% and 12.2% at 3 years, respectively (hazard ratio [HR] = 0.72; P < .001). While no difference was noted in the disease control rate (32.2% for the combination vs 30.2% for dacarbazine alone), the median duration of response in complete and partial responders was 19.3 months in the combination arm vs 8.1 months in the dacarbazine-alone arm.
Ipilimumab is now an established therapeutic option for patients with advanced melanoma, yet it comes with attendant expected toxicity on account of the associated breakdown in self-tolerance. The most common autoimmune toxicities include dermatologic reactions; gastrointestinal complications, such as colitis; endocrinopathies; and hepatitis. Administration of ipilimumab therefore requires substantial care, with an emphasis on patient education, ongoing close clinical vigilance, and timely appropriate response to control severe autoimmune manifestations.
PD-1 Checkpoint Inhibition
PD-1 is an inhibitory molecule that is expressed on T cells and that interacts with molecules in the B7 family—B7-H1 (also called PD ligand 1 [PD-L1]) and B7-DC (or PD ligand 2 [PD-L2])—to down-regulate the peripheral T-cell response to infection and to limit autoimmunity. PD-1 is also involved in controlling T-cell exhaustion that may result with continued antigenic stimulation. PD-L1 can be expressed on hematopoietic cells, in normal tissues (pancreatic islets, heart, endothelium, small intestine, and placenta) and on several tumor cell types, thereby providing the tumor cells with a protective mechanism against tumor cell–specific T-cell responses, a concept known as adaptive immune resistance (Figure C).[31,32] Blocking the interaction of the PD-1 receptor or its ligand PD-L1 using an anti–PD-1 or an anti–PD-L1 antibody prevents down-regulation of T cells (Figure 4). Several antibodies exploiting this approach are in clinical testing (MDX-1106, MK-3475, Amp224, CT-011, MDX-1105).
Initial evidence of clinical activity with PD-1 blockade using the anti–PD-1 antibody BMS-936558 (MDX-1106 [nivolumab]) was noted in a phase I study that included 39 patients with various malignancies. One durable complete response was noted in a patient with colorectal cancer, and two partial responses were seen in patients with melanoma and renal cell carcinoma; two additional patients with diagnoses of melanoma and non–small-cell lung cancer had significant tumor regression. A subsequent phase I study included 296 patients with advanced solid malignancies (melanoma, non–small-cell lung cancer, renal cell carcinoma, castration-resistant prostate cancer, and colorectal cancer) who were treated with escalating doses of the anti–PD-1 antibody BMS-936558. The incidence of grade 3 or 4 adverse events attributed to the drug was 14%; there were 3 deaths due to pulmonary toxicity. The autoimmune breakthrough toxicities were different from those observed with the anti–CTLA-4 antibody: the incidence of diarrhea was lower, but the incidence of pneumonitis was higher. Responses occurred in 28% of the patients with melanoma, in 18% of those with non–small-cell lung cancer, and in 27% of those with renal cell carcinoma. The responses were durable: 20 out of 31 lasted more than 1 year. In 42 patients, pretreatment tumor samples were obtained and tested for expression of PD-L1. None of the 17 patients with PD-L1–negative tumors had a response; 9 of the 25 patients with PD-L1–positive tumors had an objective response, suggesting PD-L1 expression as a potential tumor marker for response prediction, although this needs validation in larger studies. Building on the concept of preventing interaction between PD-1 and PD-L1, 207 patients with advanced solid malignancies (non–small-cell lung cancer, melanoma, colorectal cancer, renal cell carcinoma, ovarian cancer, pancreatic cancer, gastric cancer, and breast cancer) were treated with the anti–PD-L1 antibody BMS-936559 (MDX-1105) in a phase I study. Objective responses occurred in 17% of patients with melanoma, in 10% of those with non–small-cell lung cancer, in 12% of those with renal cell carcinoma, and in 6% of those with ovarian cancer. Responses were durable and lasted more than 1 year in 50% of patients. Very impressive early results in melanoma have been reported by Hamid et al with another anti–PD-1 antibody, lambrolizumab (previously known as MK-3475). In a phase I study that included 135 patients with advanced melanoma, the overall response rate was 38%. At the highest dose level (10 mg/kg every 2 weeks), the overall response was 52%; responses were observed both in patients who had previously received ipilimumab monotherapy and in those who were ipilimumab-naive. Both anti–PD-1 and anti–PD-L1 antibodies are in advanced stages of clinical development and are expected to gain FDA approval.