Adoptive Cell Transfer for Patients With Metastatic Melanoma: The Potential and Promise of Cancer Immunotherapy

Abstract

Background:

Current FDA-approved therapeutic options for patients with metastatic melanoma include dacarbazine, interleukin 2, ipilimumab, vemurafenib, dabrafenib, and trametinib, but long-term tumor regression using available agents remains out of reach for most patients. Adoptive cell transfer (ACT) with autologous tumor-infiltrating lymphocytes (TILs) has shown encouraging results in clinical trials, with evidence of durable ongoing complete responses in patients with advanced melanoma. Emerging techniques to engineer T-cell receptors (TCRs) or chimeric antigen receptors (CARs) using lymphocytes from peripheral blood may offer new tactics in ACT.

Methods:

We reviewed the literature to provide a synopsis on the development and clinical trial results of ACT, as well as the future outlook for using ACT in patients with metastatic melanoma.

Results:

ACT with TILs as part of a lymphodepleting regimen has been shown in clinical trials to cause objective clinical responses in approximately 40% to 72% of patients with metastatic melanoma, with up to 40% of those patients experiencing complete responses lasting up to 7 years ongoing. Pilot trials using TCR-engineered cells against melanoma-associated antigens MART-1 and gp100 and the cancer-testis antigen NY-ESO-1 have shown clinical responses in patients with melanoma. CAR cells directed against melanoma have been tested only in preclinical models; however, CAR cells targeting other histologies such as lymphoma have elicited antitumor responses in patients.

Conclusions:

An example of state-of-the-art personalized medicine, ACT is a potentially curative therapy for patients with metastatic melanoma. Ongoing trials aiming to simplify the regimens may allow a broader range of patients to be treated and enable ACT to be offered by academic cancer centers.

Graphical Abstract

Adoptive cell transfer can lead to durable tumor response in patients with metastatic melanoma.

Introduction

Cancer immunotherapy can be separated into three broad categories: active immunization, nonspecific immune stimulation, and adoptive cell transfer (ACT). For patients with metastatic melanoma, active immunization with agents such as peptides or whole tumor cell vaccines, recombinant viruses encoding tumor-associated antigens, or dendritic cells has not been shown to produce consistent and clinically relevant rates of tumor regression, which generally have been no more than 5%.^1–4 In the adjuvant setting, even patients with melanoma who have a strong in vitro response to vaccinations with melanoma-associated antigens such as the gp100_{209–217(210M)} peptide (evidenced by the generation of high frequency of antigen-specific T cells) have experienced tumor recurrence.⁵ Non-specific immune stimulation with interleukin 2 (IL-2) and ipilimumab can lead to durable cancer regression, although the overall tumor response rates for each agent have been small (16% for high-dose IL-2⁶ and 11% for ipilimumab¹), with complete response (CR) rates of less than 10%.^1,6,7 A pilot trial of 36 patients with melanoma treated with ipilimumab combined with high-dose IL-2 had overall response (OR) rates of 25%, with 17% achieving CRs lasting more than 8 years ongoing⁸; however, this IL-2 plus ipilimunab combination has not been further tested to confirm these results. Anti-PD1 and anti-PD-L1 antibodies have been recently reported to have OR rates of up to 38%^9,10 and 17%,¹¹ respectively, in patients with melanoma, and OR rates of up to 40% when combined with ipilimumab,¹² although the long-term durability of the responses is not yet known.

ACT entails the ex vivo identification (or production) of antitumor lymphocytes that are then expanded to large numbers and reinfused (“transferred”) back into the patient. ACT has theoretical and practical advantages over active immunization and nonspecific immune stimulation, including the ability to identify the exact population of T cells capable of in vitro tumor killing and select them for expansion. These cells can be activated ex vivo, free from the potentially suppressive tumor microenvironment that may prevent them from fully living up to their antitumor potential. Preparation of the host patient with lymphodepletion immediately before the transfer of the antitumor cells can eliminate potentially suppressive influences (such as regulatory T cells) to provide an optimal milieu for the cells to proliferate and become activated in vivo. When combined with a preparative lymphodepleting regimen pretransfer, ACT has consistently higher OR rates, from 40% to 72%, with long-term durable and potentially curative CR rates of up to 40%.¹³ This review briefly discusses the historical development of ACT, key clinical trials demonstrating its efficacy and potential, and ongoing and future developments that may lead to a wider use of ACT for patients with metastatic melanoma.

Historical Milestones in the Development of ACT

As early as 1922, it was suggested that the presence of a lymphocytic infiltration within resected tumor specimens was associated with longer postoperative survival compared to tumors that lacked lymphocytic infiltration.¹⁴ In 1954, Billingham et al¹⁵ reported that allograft immunity could be transferred by using regional draining lymph node cells and termed it “adoptively acquired immunity, in which a normal subject becomes immune as a result of the transference, not of preformed antibody, but of immunologically activated tissue,” demonstrating the role of the cellular arm of the immune system in tissue rejection. Limited by the ability to grow T lymphocytes in vitro, the evaluation and manipulation of this “adoptively acquired immunity” in the next several decades utilized cells obtained fresh from immunized animals.

In 1976, Morgan et al¹⁶ reported that nontransformed, bone marrow–derived T cells could be cultured in vitro using conditioned medium from stimulated T cells. In 1980, Smith et al^17,18 identified IL-2 as the soluble “T-cell growth factor” responsible for the initiation and proliferation of T lymphocytes. The development of recombinant IL-2 in 1984 led to the mass manufacture of IL-2 for use in humans.¹⁹ A report in 1985 showing that IL-2 (given along with lymphokine-activated natural killer [LAK] cells which are nonspecific non-T and non-B lymphocytes) could cause tumor regression in humans was the first to show the feasibility and potential efficacy of manipulating the immune system in human cancer therapy.²⁰ A follow-up randomized trial showed that the tumor response was due to IL-2, not due to the nonspecific LAK cells.²¹ The search for a more specific cause of immune-mediated tumor rejection led to the identification of tumor-infiltrating lymphocytes (TILs) which are capable of killing established tumors in murine models at much greater efficacy than LAK cells.²² Unlike LAK cells, TILs are classical T cells that become activated only by recognizing, with its T-cell receptor, a specific peptide presented by a human leukocyte antigen (HLA) complex on an antigen-presenting cell or tumor cell. The discovery that TILs grown from human melanoma tumors can lyse fresh autologous tumor cells but not autologous normal cells²³ led to a phase I clinical trial published in 1988 involving 12 patients with advanced cancer of varying histologies²⁴ who had metastases that could be resected to grow TILs while still having residual evaluable tumors. A single dose of cyclophosphamide was used as a preparative regimen (mainly as an “immunomodulatory agent to inhibit suppressor cell function”), and patients were infused with varying numbers of TILs (10⁹ to 10¹¹ cells) and varying doses of IL-2. Predictable toxicities attributable to IL-2 were seen, but none were directly attributable to TILs. Tumor response was seen in 1 patient with melanoma and 1 with renal cell carcinoma. These results showing the potential of treating patients with TILs led to further trials that assessed the exact preparative lymphodepleting regimens, the role of IL-2, and the nature of the infused lymphocytes needed to attain the full power of ACT.

Tumor-Infiltrating Lymphocytes

Generating TILs involves resecting a tumor deposit (generally > 1 cm, preferably ≥ 2 cm in diameter) and establishing multiple individual microcultures grown in vitro from either single-cell suspensions or 1 to 2 mm³ tumor fragments in media containing IL-2 (Figure). Appropriately expanded TIL cultures should reach several million cells (combined) in 2 to 3 weeks and have the capacity to kill autologous tumor cells present within the individual cultures. These T cells then undergo a rapid expansion protocol (REP) using the T-cell–stimulating antibody muromonab-CD3, resulting in billions of cells for patient infusion. In a retrospective study evaluating surgical resections for TILs in 402 patients from 2002 to 2007 at the Surgery Branch of the National Cancer Institute, TILs were successfully generated in 677 (86%) of the 787 specimens from all tumor sites, although tumors from the gastrointestinal tract had a decreased rate of TIL growth (70%; P = .008).²⁵

Figure. —

Adoptive cell transfer therapy using either autologous tumor-infiltrating lymphocytes (TILs; upper panel) from excised tumors or autologous peripheral blood lymphocytes gene-modified to express engineered T-cell receptors or chimeric antigen receptors (lower panel). Patients undergo a preparative lymphodepleting regimen prior to cell infusion.

The first phase II TIL-based ACT trial²⁶ for patients with metastatic melanoma was reported in 1988. The trial involved 20 patients treated with up to 2 × 10¹¹ TILs and high-dose IL-2 (720,000 IU/kg) given every 8 hours as tolerated as previously described.^27,28 Patients also received a single infusion of 25 mg/kg of cyclophosphamide 36 hours before they were given TILs. Eleven patients experienced objective tumor regression occurring in multiple metastatic sites (from subcutaneous tissue to liver and lung). A follow-up of this trial involving additional patients for a total of 86 revealed an OR rate of 34%, which was similar among those who had previously received (but did not respond to) IL-2 and those who were IL–2-naive, suggesting that the tumor responses were due to the combined regimen involving TILs and not necessarily due to the IL-2 component only.²⁹

Concurrent studies in murine models demonstrated that more aggressive lymphodepletion prior to cell transfer resulted in higher response rates suggesting that beyond the immunosuppressive effects of regulatory T cells, endogenous immune cells may compete with the adoptively transferred cells for homeostatic cytokines.^30,31 Lymphodepletion was also associated with increasing levels of homeostatic cytokines IL-7 and IL-15, which may improve the expansion and activation of the transferred cells.³² Thus, pilot trials were conducted to explore the extent of lymphodepletion needed to increase the efficacy of TIL ACT.^13,32 A nonmyeloablative, lymphodepleting chemotherapy regimen used for allogeneic peripheral blood stem cell transplant that consisted of 2 days of cyclophosphamide (60 mg/kg per day) followed by 5 days of fludarabine (25 mg/m² per day) was used.³³ The day after receiving the last dose of fludarabine, patients received TIL cell infusion followed by high-dose IL-2 as tolerated (Table 1).¹³ In two additional sequential trials, in addition to the lymphodepleting regimen described above, patients also received either a total of 2 Gy or 12 Gy total body irradiation (TBI) to further deplete endogenous lymphocytes prior to TIL cell infusion.¹³ The day following TIL infusion, patients receiving TBI also received at least 2 × 10⁶/kg of autologous CD34+ hematopoietic stem cells harvested from a granulocyte colony-stimulating factor–mobilized apheresis performed at least 1 week prior to starting cyclophosphamide. With a median follow-up time of 90 months (43 patients with no TBI), 58 months (25 patients with 2 Gy TBI), and 41 months (25 patients with 12 Gy TBI; Table 2), objective tumor responses were seen in 56% of the 93 patients enrolled.¹³ Responses were seen in all affected organs, including the lungs, liver, and brain, and could affect large bulky tumor burden.^13,35,36 Complete tumor disappearance was seen in 12% of patients without TBI, in 20% of those treated with 2 Gy, and in 40% of those treated with 12 Gy. Nineteen of the 20 patients with CR remain free of disease with some response durations lasting up to 82 months ongoing (82+, 81+, 79+, 78+, 64+, 68+, 64+, 60+, 57+, 54+, 48+, 45+, 44+, 44+, 39+, 38+, 38+, 38+, 37+, and 19 months).¹³

Table 1. —

Protocols for ACT Trials at the Surgery Branch of the National Cancer Institute

	Day −7	Day −6	Day −5	Day −4	Day −3	Day −2	Day −1	Day 0	Day +1	Day +2	Day +3
ACT Without TBI
Chemotherapy	Cy	Cy	Flu	Flu	Flu	Flu	Flu
Cells								Cells
Cytokinesa								IL-2	IL-2	IL-2	IL-2
ACT With 2 Gy TBI
Chemotherapy		Cy + Flu	Cy + Flu	Flu	Flu	Flu
Radiation							TBI
Cells								Cells
Cytokines								IL-2	IL-2	IL-2	IL-2
Stem cells									CD34+
ACT With 12 Gy TBI
Chemotherapy	Cy + Flu	Cy + Flu	Flu	Flu	Flu
Radiation					TBI	TBI	TBI
Cells								Cells
Cytokines								IL-2	IL-2	IL-2	IL-2
Stem cells									CD34+

Data from Rosenberg SA, Yang JC, Sherry RM, et al. Durable complete responses in heavily pretreated patients with metastatic melanoma using T-cell transfer immunotherapy. Clin Cancer Res. 2011;17(13):4550–4557.

^aFor protocols in which cytokines such as IL-2 are given.

ACT = adoptive cell transfer, Cy = cyclophosphamide (60 mg/kg per day), Flu = fludarabine (25 mg/m² daily), IL-2 = interleukin 2 (760,000 IU/kg per dose, given every 8 hours as tolerated up to maximum of 12 doses), TBI = total body irradiation.

Table 2. —

Objective Tumor Regressions by RECIST^a in Selected, Sequential TIL Trials Using Differing Lymphodepleting Regimens

Regimenb	No. of Patients	Partial Response, n (%)	Complete Response, n (%)	Overall Response, n (%)
No TBI	43	16 (37)	5 (12)	21 (49)
2 Gy TBI	25	8 (32)	5 (20)	13 (52)
12 Gy TBI	25	8 (32)	10 (40)	18 (72)
Total	93	32 (34)	20 (22)	52 (56)

Data from Rosenberg SA, Yang JC, Sherry RM, et al. Durable complete responses in heavily pretreated patients with metastatic melanoma using T-cell transfer immunotherapy. Clin Cancer Res. 2011;17(13): 4550–4557.

^aPatient response to treatment was assessed utilizing Response Evaluation Criteria in Solid Tumors (RECIST) guidelines.³⁴

^bSpecifics of the regimens are detailed in the text and in Table 1.

TBI = total body irradiation, TIL = tumor-infiltrating lymphocyte.

Although the data suggest that the patients who received the most aggressive lymphodepleting regimen with the addition of 12 Gy TBI experienced higher OR rates and potentially curative CR rates, these three trials were sequential and not randomized; thus, comparisons between the trials are meant to be hypothesis-generating only. The role of TBI lymphodepletion in affecting the efficacy and toxicity of TIL ACT is currently under investigation: an ongoing clinical trial is randomizing patients with meta-static melanoma to receive preparative lymphodepletion with cyclophosphamide and fludarabine or with cyclophosphamide, fludarabine, and 12 Gy TBI (NCT01319565).

In the trials listed in Table 2,¹³ the TILs used were “selected” TILs. In other words, the individual startup microcultures were separately tested for antitumor recognition by co-culture assays against either autologous tumor or melanoma cell lines, and only microcultures showing expected antitumor reactivity were selected to undergo REP expansion leading to clinical use. This step of selecting for reactivity required additional time and prevented some patients from undergoing therapy if their TIL cultures did not pass the test (approximately one-third of patients did not have adequate in vitro TIL reactivity²⁵) or if rapid disease progression causing a significant decline in performance status occurred during the growth of the TILs (4 to 6 weeks). Furthermore, data emerging from both mouse models and human clinical trials have indicated that TILs grown for a shorter time in culture have characteristics (eg, longer telomeres³⁷ and CD27+^38,39) associated with higher proliferative potential and higher rates of tumor regression. Techniques to grow “unselected” TILs without screening for tumor reactivity were developed and shortened the time TILs spent in culture to 10 to 18 days before undergoing REP expansion.⁴⁰ When comparing “selected” and “unselected” TILs, the efficacy rate was similar by in vitro testing⁴⁰ and was subsequently tested in a pilot trial involving 33 patients who experienced similar tumor response rates (58%) as those seen in prior “selected” TIL trials.⁴¹

This and other simplifications in techniques led other centers to adapt methods to perform TIL trials. Besser et al^42,43 at Sheba Medical Center (Israel) treated 32 patients with metastatic melanoma using (unselected) TIL and IL-2 with the same lymphodepleting cyclophosphamide and fludarabine regimen. Fifteen (48%) of 31 evaluable patients experienced ORs, with 4 (13%) achieving CRs. Pilon-Thomas et al⁴⁴ at Moffitt Cancer Center treated 13 patients with melanoma using selected TIL and IL-2 with the same lymphodepleting regimen. Five patients achieved ORs (38% response among those treated) with 2 achieving CRs. Ullenhag et al⁴⁵ from Uppsala University (Sweden) grew TILs from core biopsies rather than surgical excisions and used low-dose subcutaneous IL-2. They treated 24 patients with metastatic melanoma, with 5 (21%) achieving ORs and 1 achieving a CR. Other groups currently investigating TIL therapy in patients with melanoma include the MD Anderson Cancer Center (NCT00338377)⁴⁶ and Copenhagen University Hospital (Denmark)⁴⁷ (NCT00937625), which reported 2 patients achieving CRs out of 6 treated with TILs and low-dose subcutaneous IL-2.

Genetically Modified T-Cell Receptors

Several important limitations of TIL therapy exist: (1) the need to perform an invasive procedure to obtain tumor tissue to grow TILs, (2) tumor sites that are not easily accessible can increase risks of postoperative morbidity; some tumor sites [eg, lung hilum, head of the pancreas] exclude the option for TIL resection given the potential high morbidity of surgery, and (3) the inability to grow TILs in a small but real number of patients (~ 10% to 15% of patients with melanoma²⁵). The ability to modify genes of any lymphocyte to induce expression of the desired T-cell receptor (TCR) can help bypass these limitations. As an alternative to TILs, genes encoding TCRs that recognize cancer antigens can be introduced into a patient’s peripheral blood lymphocytes (obtained from an apheresis or blood draw) using retroviral or lentiviral vectors.^48,49 High-avidity TCRs can be identified either by isolating highly reactive T-cell clones from patients (usually after extensive in vitro sensitization of peripheral blood lymphocytes with the target tumor antigen) or by immunizing transgenic mice (with human HLA) using the target human antigen. The TCR α and β chains from the reactive T-cell clones can then be isolated and cloned into a gene expression viral vector subsequently used to transduce any lymphocyte into becoming antigen specific similar to the parental clone.

The first ACT clinical trial using genetically engineered TCRs involved 15 patients with metastatic melanoma treated with anti–MART-1 TCR-engineered-lymphocytes and high-dose IL-2 (after receiving a cyclophosphamide and fludarabine lymphodepleting regimen).⁵⁰ Two patients exhibited partial tumor responses, suggesting that these gene-engineered T cells may be active in vivo. Subsequent ACT clinical trials involving gene-modified TCRs used a different clone of an anti–MART-1 T cell with higher avidity (which induced a 30% OR rate in 20 treated patients) and anti-gp100 TCR-transduced lymphocytes that induced a 19% OR rate in 16 treated patients.⁵¹ However, these T cells also caused destruction of normal melanocytes in the skin, eye, and ear in some patients, leading to severe skin rashes, uveitis, or hearing loss (which resolved with local steroid administration). This “on-target/off-tumor” toxicity was also seen in an ACT trial testing anti-carcinoembryonic antigen (CEA) TCR-engineered T cells aimed at treating patients with CEA-positive metastatic colon cancer refractory to standard chemotherapy.⁵² The modified T cells induced a partial response in 1 patient out of the 3 treated, but they also caused severe transient colitis in the normal colon. These studies suggest that the selection of the target antigen for TCR development is critical to prevent potentially life-threatening cross-reactivity.

Cancer-testis antigens (CTAs) are expressed by some tumors but are expressed in normal tissues only during fetal development. NY-ESO-1 is a CTA expressed in a wide range of tumor types. In a trial using anti-NY-ESO-1 TCR-modified peripheral blood lymphocytes, along with high-dose IL-2 following lymphodepletion with cyclophosphamide and fludarabine, in patients with metastatic melanoma and synovial cell sarcoma expressing NY-ESO-1, tumor response was seen in 4 out of 6 patients with synovial cell sarcoma and in 5 out of 11 patients with melanoma.⁵³ No TCR-specific toxicity was seen. This study was the first to show that TCR-transduced lymphocytes could cause objective cancer regression in a nonmelanoma tumor, suggesting the possibility of expanding ACT to nonmelanoma cancers traditionally thought to be less immunogenic.

Chimeric Antigen Receptors

Although the use of TCR-transduced T cells eliminates the need for tumor excision for TIL ACT, the specific nature of the TCR, which is HLA-restricted, limits its applicability in some patients. For example, the NYESO-1 TCR used in the study by Robbins et al⁵³ can recognize and interact only with HLA-A*0201; lymphocytes transduced with that NY-ESO-1 TCR would be functional only in patients with HLA-A*0201, and thus, patients who are not HLA-A*0201 are not eligible for treatments using that particular TCR. Because they rely on an antibody-antigen interaction and not TCR-HLA-tumor peptide interaction, chimeric antigen receptors (CARs) provide an alternative to TCR-transduced cells because they do not require HLA interaction and are not HLA-restricted. A CAR results from the fusion of the intracellular signaling domain of a TCR (eg, CD3ζ and CD28) with the extracellular antigen-binding domain from a single-chain variable fragment of an antibody that recognizes the intended tumor antigen. These CAR-transduced cells have the specificity of an antibody coupled with the cytotoxic effector mechanisms of a T cell. Antigens recognized by CAR cells need to be expressed on the surface of the tumor; thus, intracellular melanoma-differentiation antigens such as gp100 and MART-1 cannot be targeted by CARs. This aspect of CAR cells is an important disadvantage when compared with TCR-transduced cells which can recognize both extracellular and intracellular antigens that are then processed and expressed extracellularly by the HLA molecule.

CAR cells targeting the ganglioside antigens GD2 and GD3 expressed by the majority of melanoma tumors have shown antitumor efficacy in murine melanoma models.^54,55 CAR cells directed against high-molecular-weight melanoma-associated antigen (HMW-MAA; from the CSPG4 gene, a cell-surface proteoglycan expressed on > 90% of melanomas) have been shown in vitro to be cytolytic in response to HMW-MAA–expressing melanoma cell lines but have not yet been clinically tested.⁵⁶ Ongoing clinical trials assessing the efficacy of CARs in melanoma include an anti-vascular endothelial growth factor receptor 2 CAR, which is being tested on patients with any solid tumor histology, including melanoma (NCT01218867). The evidence for the potential power of CAR-transduced cells currently lies in studies for nonmelanoma tumors.

The first study to show clinical responses using CAR-transduced lymphocytes treated 11 pediatric patients with neuroblastoma with CAR cells (2 × 10⁷ to 2 × 10⁸ cells/m²) directed against the GD2 diasialo-ganglioside antigen expressed by most neuroblastoma cells.⁵⁷ Of the 8 patients with evaluable tumors, 4 had evidence of tumor necrosis or regression, including 1 patient with a CR more than 1 year ongoing. CAR cells targeting CD19 have been shown to be efficacious in causing tumor remission in patients with lymphoma or chronic lymphocytic leukemia (CLL).⁵⁸ CD19 is a cell-surface protein expressed only in B-lineage cells (both normal and malignant) such as B cells, B-cell precursors, and plasma cells. Kochenderfer et al⁵⁹ published the first anti-CD19 CAR study showing clinical response. The study participants underwent a preparative lymphodepleting regimen of cyclophosphamide and fludarabine and received 0.3 × 10⁷ to 3.0 × 10⁷ anti-CD19 cells/kg along with high-dose IL-2. Three out of 4 patients with lymphoma experienced partial responses lasting up to 18 months, and 3 out of 4 patients with CLL experienced CRs or partial responses lasting up to 15 months ongoing.⁶⁰ Porter et al⁶¹ also noted complete remission in a patient with CLL treated with a different anti-CD19 CAR after undergoing lymphodepletion with pentostatin and cyclophosphamide; no cytokine was administered after the cell infusion in this study. Because IL-2 was not needed for these CAR cells to cause tumor regression, the anti-CD19 CAR trial at the Surgery Branch of the National Cancer Institute no longer includes IL-2 (NCT00924326).

Although toxicities such as fever, tumor lysis syndrome, transient hypotension, and transient renal and hepatic insufficiency were reported in some CAR studies, many adverse events could be attributable to the administration of IL-2, and the patients recovered from these toxicities. However, unexpected deaths have also been reported in 1 patient treated with anti-CD19 CAR cells⁶² and 1 patient treated with anti-HER2 CAR cells.⁶³ Soon after cell infusion, both patients developed profound hypotension, respiratory distress, and multiorgan failure associated with markedly elevated levels of inflammatory and homeostatic cytokines. Although severe sepsis could cause this clinical scenario given that these patients were neutropenic around the time of cell infusion, this lethal “cytokine storm” could have been due to the overwhelming activation and proliferation of the transferred CAR cells upon target recognition; such events had not been observed in other ACT trials using TILs. Modification of the CD19 trial to decrease the number of CAR cells transferred has allowed additional patients to be safely treated.⁶²

Discussion

Metastatic melanoma remains a formidable disease with 1-year survival rates for patients with M1a, M1b, and M1c being only 62%, 53%, and 33%, respectively.⁶⁴ In 2011, ipilimumab and vemurafenib were approved by the US Food and Drug Administration for use in patients with metastatic melanoma, adding to the previously limited therapeutic armamentarium that consisted of IL-2 (approved in 1998) and dacarbazine (approved in 1975). Dabrafenib and trametinib were approved in May 2013 for patients with mutated BRAF. Anti-PD1 and anti-PD-L1 antibodies may be added to the list in the near future. The most beneficial sequence of therapy for patients is not yet determined, and we and others are evaluating whether these agents may provide synergism with ACT. Clinical trials are ongoing to evaluate TILs with ipilimumab (NCT01701674), TILs with vemurafenib (NCT01585415, NCT01659151), and TILs with dendritic cell immunization (NCT00338377).

Compared with immunomodulatory drugs such as IL-2 or ipilimumab alone, ACT can theoretically overcome the immunosuppressive tumor microenvironment that prevents naturally existing antitumor lymphocytes from proliferating, becoming activated, and killing tumor cells. ACT may be able to sidestep this limitation because activation and expansion of the antitumor lymphocytes occur ex vivo. Furthermore, the manipulation of the host patient precell transfer with lymphodepletion may create an environment conducive for further cell expansion and activation and may also prevent immunosuppressive regulatory T cells from interfering with the full capacity of the transferred cells. Consistent results of multiple clinical trials confirm the potential power of ACT to induce durable clinical responses. Although the ACT regimens with the longest follow-up (and thus having stronger evidence of durable tumor regression) use TILs, TCR– and CAR–engineered cells have elicited notable antitumor responses in clinical trials and may provide options to patients who do not qualify for TIL therapy. Given the simplicity of obtaining lymphocytes for transduction by just blood draws, T-cell engineering technology may in the future lead to “off-the-shelf” reagents that are personalized to an individual’s HLA and tumor antigen status.

The ACT schema we have used most often, which consists of preconditioning with cyclophosphamide and fludarabine precell transfer and high-dose IL-2 postcell transfer, is rigorous. Patients need to be medically fit to safely tolerate the regimen. Because a significant contributor to the adverse events seen in these protocols is the high-dose IL-2 component, ACT trials with low-dose^45,47 or no IL-2 are being evaluated. A TIL trial currently enrolling patients with melanoma is using the preparative cyclophosphamide and fludarabine regimen but does not administer IL-2 or any cytokine following cell transfer; preliminary evaluations have shown some clinical responses, although whether the results are similar to TIL trials using high-dose IL-2 remains to be seen (NCT01468818). Based on encouraging results from murine models,^65,66 another trial is using TILs transduced to express IL-12, an immunostimulatory cytokine, upon antigenic stimulation of the native T-cell receptor (NCT01236573). IL-2 is not given in this trial, and preliminary evaluations have shown some objective responses in patients with metastatic melanoma, including complete responses. Beyond decreasing the incidence of toxicities attributable to high-dose IL-2 and expanding the patient population eligible to receive ACT, these trials may ultimately help simplify the ACT regimens and allow more medical centers to offer these potentially curative therapies.

Conclusions

The cure of patients with solid tumors such as meta-static melanoma requires the induction of durable complete tumor responses. Adoptive cell transfer has the potential to provide lifelong antitumor immunity and the promise of a cure for some patients. Active research is ongoing to bring this to fruition for all patients.