Combining targeted therapy and immune checkpoint inhibitors in the treatment of metastatic melanoma

Epidemiology of melanoma

Skin cancer is among the most common cancer types globally¹. Melanoma is the deadliest skin cancer². Its burden on public health continues to rise, with its incidence increasing faster than any other cancer in recent years^1,2. Early stage melanoma is treatable with surgery, but the late stage of this disease is often fatal³. This review discusses the treatment options for patients with advanced melanoma and the rationale for combining targeted therapy and immune checkpoint inhibitors to treat these patients.

Chemotherapy

Treatment options for advanced melanoma were limited in the past decade, and prognosis was universally poor. Cytotoxic chemotherapy was the main treatment strategy but was marginally effective only in the treatment of locally advanced or metastatic disease. Dacarbazine [5(3,3-dimethyl-1-triazeno)-imidazole-4-carboxamide] was the primary agent used, and this drug remains the only FDA-approved chemotherapy for metastatic melanoma. However, therapy is characterized with low overall response rates (approximately 10%-15%), and the drug offers no survival benefit^4,5.

In addition to dacarbazine treatment, biochemotherapy regimens and combined chemotherapeutic agents (e.g., dacarbazine or temozolomide with vinblastine and cisplatin) with cytokine-based therapy [e.g., interleukin-2 (IL-2) and interferon-alpha] had been administered to improve response rates by inducing immunogenic cell death. Two large meta-analyses that evaluated the standard chemotherapy versus biochemotherapy in 18 randomized controlled trials (RCTs) involving more than 2,600 patients with metastatic melanoma show that biochemotherapy regimens can improve overall response rates, but with greater systemic toxicity and without a statistically significant survival benefit^6,7.

Immunotherapy

Another modality in melanoma treatment involves the use of immunotherapy. The first immune-based therapy with demonstrated clinical benefit in melanoma patients was IL-2, an immune stimulating cytokine integral to T cell activation and proliferation. Atkins et al.⁸ examined 270 patients with metastatic melanoma treated with high dose IL-2, and 16% of patients who achieved complete response (CR) or partial response (PR) showed long-term responses, with a median progression free survival (PFS) of 13.1 months. Longer follow-up time of the patients demonstrated an approximately 6% CR rate⁹. A follow-up phase III RCT by Schwartzentruber et al.¹⁰ demonstrated a small but statistically significant improvement in objective response rate (ORR) associated with the addition of the gp100 peptide vaccine to high dose IL-2, although, again, in only a small percentage of patients (16% vs. 6%) treated with vaccine plus IL-2 versus IL-2 alone.

Immune checkpoint inhibitors have also been successfully used to treat melanoma. This therapy is based on the fact that T lymphocytes are critical to antitumor immunity, and activation by an antigen-specific T cell receptor in the context of costimulatory activation is required. However, a naturally occurring feedback mechanism that prevents excess immune activation through the expression of negative costimulatory molecules exists¹¹. These negative costimulatory molecules, also known as “checkpoints”, such as cytotoxic T-lymphocyte antigen 4 (CTLA-4), programmed death 1 (PD-1), T cell immunoglobulin 3, and lymphocyte-activation gene 3, act as “brakes” on T cell activation and serve as negative feedback mechanism¹¹. Interestingly, tumor-infiltrating T lymphocytes (TIL) in many tumor types express high levels of negative costimulatory markers, suggesting a tumor-derived mechanism of suppressing antitumor immunity and providing rationale for T cell checkpoint blockade¹².

In a milestone phase III RCT of 676 patients with unresectable stage III or IV melanoma treated with either anti-CTLA-4 antibody ipilimumab, gp100 peptide vaccine, or combined ipilimumab plus vaccine, patients treated with either ipilimumab arm had improved overall survival (OS) compared with those treated with vaccine alone (10.0 vs. 6.4 months)¹³. Ipilimumab alone achieved the best overall response rate in 10.9% of patients, and 60% of these patients benefitted from long-term responses lasting greater than 2 years. However, ipilimumab therapy was also associated with higher toxicity rate, with 10%-15% of patients suffering from grade 3 or 4 immune-related adverse events (AEs) such as diarrhea or colitis, dermatitis, and pruritis¹³. Similar results were reported in a second RCT, which compared ipilimumab plus dacarbazine versus dacarbazine alone in 502 patients with advanced melanoma, but this study utilized a higher dose (10 mg/kg) of ipilimumab¹⁴. Response rates were 15% in the ipilimumab with dacarbazine-treated group but with higher toxicities. Grade 3 or 4 AEs occurred in 56% of patients¹⁴.

Topalian et al.¹⁵ recently reported the results of the phase I trial of 296 patients with either advanced melanoma or other solid tumors, which included non-small cell lung cancer, prostate cancer, renal cell carcinoma, and colorectal cancer, in which the checkpoint blocking antibody anti-PD-1 (BMS-936558, nivolumab) achieved a 28% response rate in melanoma patients, with long-term responses longer than 1 year in 50% of responding patients. Anti-PD-1 therapy was associated with a lower rate of grade 3 or 4 AEs compared with ipilimumab. Interestingly, Topalian et al.¹⁵ suggested a possible association between tumor expression of the PD-1 ligand PD-L1 and response to anti-PD-1 therapy. However, further studies are necessary to confirm this finding. In a pooled analysis of 411 melanoma patients treated with the anti-PD-1 antibody MK-3475 (pembrolizumab, Merck Sharpe & Dohme) with over 6 months of follow-up data, the ORR was 40% in ipilimumab naïve patients and 28% in ipilimumab refractory patients¹⁶. Median PFS was 24 weeks, but median OS had not been reached at the time of analysis. Pembrolizumab was well-tolerated with 12% of patients experiencing a grade 3 or 4 AE attributed to the drug, but only 4% of patients discontinued treatment because of AE¹⁶. This agent was FDA-approved for metastatic melanoma in early September 2014. The antibody blockade of PD-L1 in the phase I trial of 207 patients using BMS-936559, including 55 with advanced melanoma, achieved an objective response in 17% of melanoma patients, with more than half of patients achieving long-term responses lasting longer than 1 year and a comparable rate of grade 3 or 4 AEs¹⁷. Although this agent is currently not being tested, two anti-PD-L1 antibodies, namely, MPDL3280A (Genentech) and MEDI4736 (MedImmune), are being tested in solid tumors under early phase clinical trials.

In addition to their use for monotherapy, different immune checkpoint inhibitors are now being combined in clinical trials, which showed impressive response rates. In the phase I trial of 86 patients with unresectable stage III or IV melanoma treated with either concurrent or sequential ipilimumab and nivolumab, concurrent CTLA-4 and PD-1 blockade achieved a higher ORR of 40%, with 53% of patients achieving CR or PR at the maximum doses tested, whereas 31% of responders demonstrated tumor regression of 80% or more even with bulky disease¹⁸. In 53% of patients, grade 3 or 4 AEs occurred at a higher frequency in combined therapy than in monotherapy, the majority of which were reversible with appropriate supportive management¹⁸. In a recent update, the 1-year OS rate for patients treated with combined immune checkpoint inhibitors was 85%, and their 2-year survival rate was 79%¹⁹.

Additional strategies targeting checkpoint inhibitors and other immunomodulatory molecules are currently being studied. However, a thorough discussion of these strategies (as well as other strategies such as treatment with tumor-infiltrating lymphocytes) is beyond the scope of this review.

Targeted therapy

Concurrent advances in targeted molecular therapy have also improved the treatment and prognosis of a subset of advanced melanoma patients. Approximately 50% of cutaneous melanomas harbor an activating mutation in the BRAF oncogene, leading to constitutive activation of the mitogen-activated protein kinase (MAPK) signaling pathway involved in cellular proliferation and survival²⁰. Preclinical studies of vemurafenib (PLX4032), a potent oral small molecule inhibitor of mutated BRAF, have culminated in a phase III RCT of vemurafenib versus dacarbazine in 675 patients with BRAF-mutated metastatic melanoma²¹. The vemurafenib arm resulted in improved OS (84% vs. 64% at 6 months) and higher response rate (48% vs. 5%) than standard chemotherapy, representing the only treatment other than anti-CTLA-4 to improve survival in metastatic melanoma²¹. Similar results were demonstrated in a phase III RCT, in which another potent BRAF^V600E inhibitor, dabrafenib, was compared with dacarbazine²². Despite improvements in response rate and survival, BRAF inhibition achieved only a median PFS of 6 months, implying rapid development of tumor resistance^21,22. Further investigation of resistance mechanisms has suggested that BRAF-mutated melanoma cells can maintain MAPK signaling through RAF-independent activation of MEK, a kinase downstream of RAF in the MAPK cascade²³. Translating these findings to clinical studies, Flaherty et al.²⁴ demonstrated that combined BRAF and MEK inhibition (dabrafenib plus trametinib vs. dabrafenib alone)achieved a higher overall response rate of 76% versus 54%, as well as a longer median PFS of 9.4 vs. 5.8 months. Similar to immunotherapy, combination molecular therapy, which targets multiple levels of an oncogenic signaling cascade or multiple different cell survival pathways, will likely enhance tumor response. In fact, dual immune and molecular therapy together may offer the best possibility of both long-term and frequent response²⁵.

Rationale for combination therapy

Newly approved targeted and immune-modulating agents have provided numerous treatment options. However, the optimal sequencing of these agents remains controversial. Even though BRAF inhibition through selective BRAF inhibitors produces excellent early disease control for patients with V600E/K mutations, the response duration of this approach is limited to less than a year because of the development of multiple molecular mechanisms of resistance^23,26–32. Checkpoint blockade with the CTLA4 inhibitor, ipilimumab, and anti-PD-1 antibodies produces responses in a minority of patients, but with long-term responses^13,15. Thus, the combination of targeted therapy and immunotherapy may lead to early and robust antitumor responses from targeted therapy with long-term benefit from the influence of immunotherapy.

In vitro studies

To date, numerous studies have investigated combined targeted therapy and immunotherapy in melanoma. The first report suggesting that oncogenic BRAF^V600E can lead to tumoral immune escape was published in 2006³³. Further in vitro studies have been performed after the development of specific BRAF inhibitors, and BRAF inhibition in BRAF mutant melanoma cell lines and fresh tumor digests has been demonstrated to result in up regulation (up to 100-fold) of melanoma differentiation antigens³⁴. Additionally, inhibition with BRAF and MEK inhibitors increased the recognition of these melanoma antigens by antigen-specific T lymphocytes. However, MEK inhibitors adversely affect the T cell function whereas those treated with BRAF inhibitors maintained functionality³⁴. Further independent studies on the effects of dabrafenib (BRAF inhibitor), trametinib (MEK inhibitor), or their combination on T lymphocytes have also shown that trametinib alone or in combination suppressed T-lymphocyte proliferation, cytokine production, and antigen-specific expansion, whereas treatment with dabrafenib had no effect³⁵. Callahan et al.³⁶ suggested that BRAF inhibitors can modulate T cell function by potentiating T cell activation through ERK signaling in vitro and in vivo.

In vivo studies

Importantly, the effect of BRAF inhibition has also been studied in patients with metastatic melanoma. Results showed a similar increase in melanoma differentiation antigens and a significant increase in intratumoral CD8⁺ T cells, which were more clonal 10-14 days after initiation of BRAF inhibition^37–39. These findings were also associated with down regulated IL-6, IL-8, IL-1α and vascular endothelial growth factor (VEGF)^38,40,41. The increased immunomodulatory molecules, PD-1 and PD-L1, 10-14 days after BRAF inhibition initiation are also important, and this condition suggests a potential immune-based mechanism of resistance³⁸. The up regulated PD-L1 expression may have been caused by infiltrating IFN-γ-secreting T cells⁴², although stromal components may also be involved⁴³. Jiang et al.⁴⁴ also suggested PD-L1 expression as a mechanism of resistance to BRAF inhibitors because BRAF-resistant cell lines expressed high PD-L1, and the addition of MEK inhibitors has a suppressive effect on PD-L1 expression. These data indicate that addition of immunotherapy and specifically immune checkpoint blockade may enhance antitumoral response when combined with a BRAF inhibitor.

Mouse models have provided important insights into cancer development, treatment, and therapeutic resistance. Several preclinical mouse models have been used to examine in detail the potential of combining immunotherapy with BRAF-targeted therapy, and most studies have indicated an additive or synergistic effect. In the syngeneic SM1 mouse model of BRAF^V600E melanoma, an improved antitumor activity was observed after combining BRAF inhibition with adoptively transferred T cells, leading to increased in vivo cytotoxic activity and intratumoral cytokine secretion by the transferred T cells. Interestingly, BRAF inhibition did not alter adoptively transferred T cell expansion, distribution, or intratumoral density⁴⁵. Liu et al.⁴¹ also studied the effects of BRAF inhibition on adoptively transferred cells by using pmel-1 TCR transgenic mice on a C57BL/6 background and xenografts of melanoma cells transduced with gp100 and H-2Db. They found an increase in T cell infiltration, which was associated with VEGF. Additionally, they showed that melanoma cell VEGF over expression abrogated T cell infiltration, and these findings were validated in patients treated with BRAF-directed therapy considering that down regulated intratumoral VEGF is correlated with a denser intratumoral T cell infiltrate after melanoma patients were treated with BRAF inhibitors⁴¹. Knight et al.⁴⁶ utilized several mouse models, which included SM1, SM1WT, and a transgenic mouse model of melanoma, to support the therapeutic potential of combining BRAF inhibitors and immunotherapy. They observed an increase in CD8:Treg ratio after BRAF inhibition and that depleting CD8⁺ T cells, not NK cells, was required for antitumor activity of BRAF inhibitors. They also showed that CCR2 demonstrates an antitumoral role after BRAF inhibition and that combination of BRAF-targeted therapy and anti-CCL2 or anti-CD137 led to a significant increase in antitumoral activity in these mouse models⁴⁶.

The authors recently demonstrated a potential synergy when immune checkpoint blockade was added to BRAF inhibition⁴⁷. This process was performed using a BRAF^V600E/PTEN^–/–syngeneic subcutaneous mouse model, which showed an increase in intratumoral CD8 T cells after BRAF inhibitor initiation, similar to melanoma patients. Either PD-1 or PD-L1 blockade addition to BRAF inhibition resulted in enhanced response, which slowed tumor growth and enhanced survival. Additionally, increased number and activity of infiltrating TIL was observed⁴⁷.

However, a previous study has shown the absence of synergy in combined BRAF-targeted and immunotherapy⁴⁸. This study conducted experiments using a conditional melanocyte-specific expressed BRAF^V600E and PTEN gene that led to 100% penetrance, short latency, and lymph node and lung metastases. These induced tumors were similar to human melanoma tumors from a histologic standpoint, but the immune response to BRAF inhibition was distinct from that observed in BRAF-inhibitor-treated patients with metastatic melanoma^38,39,48. In this model, treatment with anti-CTLA4 blockade and BRAF inhibition was not associated with improved survival or decreased tumor outgrowth⁴⁸. These results are contrary to those observed in several other models. Thus, understanding the translational relevance of individual models and their utility in guiding the development of human clinical trials is important.

Therefore, combined BRAF-targeted therapy with immunotherapy based on preclinical in vitro and in vivo work is advantageous. Aggregate data suggest that BRAF inhibitor treatment is associated with increased melanoma antigens, increased CD8 T cell infiltrate, and decreased immunosuppressive cytokines and VEGF early in the course of therapy (within 2 weeks of initiating treatment in patients)^38,40,41. However, a simultaneous increase in immunomodulatory molecules was also found, which may contribute to therapy resistance. Adding BRAF-targeted therapy to a number of different treatment modalities could improve responses (Figure 1), and these combinations are currently being tested in murine models and clinical trials.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Putative effects of adding BRAF targeted therapy to immune-based therapies. Without treatment, melanomas demonstrate an immunosuppressive environment with generally low levels of melanoma antigens, low levels of infiltrating cytotoxic T lymphocytes, and high levels of immunosuppressive cytokines and VEGF (A). Treatment with a BRAF inhibitor results in a favorable tumor microenvironment with increased antigens and CD8+ T cells and decreased immunosuppressive cytokines and VEGF, but with concurrent increase in immunomodulatory molecules, such as PD-1 on T cells and PD-L1, in the tumor microenvironment (B). BRAF-targeted therapy may synergize with different treatment modalities, and this phenomenon is being tested in murine models and in clinical trials. Evidence for synergy exists with immune checkpoint blockade (C), considering that the BRAF inhibitor addition has positive effects on the tumor microenvironment (with increased antigens, CD8+ T cells, and decreased immunosuppressive cytokines/VEGF) and that the simultaneous increase in immunomodulatory molecules is tempered by immune checkpoint blockade (D). The potential synergy includes combined treatment with IL2 (E), in which the addition of BRAF-targeted therapy may augment the antitumor immune response by its favorable effects on the tumor microenvironment (F). Nevertheless, regulatory T cells in this setting may be controversial. Adoptive cell therapy works through ex vivo activation of autologous antigen-reactive T cells (G), and this behavior may be enhanced by the favorable effects of BRAF-targeted therapy on the tumor microenvironment (H). Radiation therapy has clear effects on the tumor microenvironment with effects on both tumor cells and antitumor immunity (I), which may be augmented by the addition of BRAF-targeted therapy (J).