Adoptive Transfer of MART-1 T-Cell Receptor Transgenic Lymphocytes and Dendritic Cell Vaccination in Patients with Metastatic Melanoma

The genetic transfer of α and β chains of the T-cell receptor (TCR) endows recipient T cells with the specificity of donor T cells (1), allowing the generation of large numbers of T cells with uniform antigen. Initial clinical experiences testing adoptive cell transfer (ACT) of TCR-engineered T cells in humans have provided clear evidence of antitumor efficacy in patients with metastatic melanoma and sarcoma (2–4). Pioneering TCR-engineering clinical trials by Steven A. Rosenberg and colleagues at the Surgery Branch, National Cancer Institute (NCI, Bethesda, MD; refs. 2, 3, 5) included a 2- to 4-week ex vivo T-cell culture. Preclinical models suggest that extended ex vivo expansion of lymphocytes before ACT results in more terminally differentiated cells with limited proliferation ability in vivo and lower antitumor activity (6, 7). Provision of antigen in the form of a vaccine is required in some animal models to support the antitumor activity of adoptively transferred T cells (8–10). This may be because exposure to antigen while undergoing homeostatic proliferation can stimulate further T-cell expansion (11, 12).

To test this combined cell therapy approach in the clinic, the University of California, Los Angeles (UCLA)/Caltech F5 clinical trial was designed with a short, 1-week, cell manipulation that included initial lymphocyte activation followed by retroviral transduction and limited further ex vivo cell expansion. We also provided autologous MART-1_26–35 peptide-loaded DC, a vaccine that in our prior experience had resulted per se in two complete responders out of 25 patients with metastatic melanoma. These responses are durable over 10 years later (13, 14).

A Simon optimal two-stage phase II clinical trial design (15) was used to allow for the simultaneous testing of three coprimary endpoints, safety, feasibility, and objective tumor response. Patients were enrolled in the clinical trial after signing a written informed consent approved by the UCLA Institutional Review Board (#08-02-020 and #10-001212) under an investigational new drug (IND) filed with the U.S. Food and Drug Administration (IND# 13859). The study was conducted in accordance with local regulations, the guidelines for Good Clinical Practice, and the principles of the current version of the Declaration of Helsinki. The study had the clinical trial registration number NCT00910650.

Eligible patients were HLA-A*0201 by molecular subtyping, had progressive locally advanced (stage IIIc) or metastatic melanoma (stage IV) with either no available standard therapeutic options with a curative intent, or who had progressed on standard options like chemotherapy, high-dose interleukin (IL)-2, interferon, and experimental therapies as listed in Table 1; the melanoma was MART-1–positive by immunohistochemistry, age greater than or equal to 18, Eastern Cooperative Oncology Group performance status 0 or 1, life expectancy greater than 3 months, adequate organ function as routinely required to receive high-dose IL-2 (16), and seronegative for HIV, Hepatitis B and C. Patients with clinically active brain metastases were excluded. Baseline radiological documentation of absence of active brain metastases was required for all patients, but previously treated brain metastases were acceptable. All patients underwent formal ophthalmologic and otologic exams at baseline and periodically after TCR-engineered ACT.

The study outline is included in Fig. 1. Patients underwent baseline leukapheresis for the manufacture of the two cell therapies, the F5 TCR gene–modified T cells, and the MART-1 peptide-pulsed DC (see Supplementary Online Methods). Patients received conditioning chemotherapy as inpatients consisting of cyclophosphamide 60 mg/kg/day × 2 days i.v. and fludarabine 25 mg/m²/day i.v. daily for 4 or 5 days as previously described (2, 3, 5, 17). On day 0, patients received the ACT of TCR-engineered lymphocytes as an i.v. infusion with the cells thawed at bedside under the initial study protocol. Patients were transferred for monitoring to the intensive care unit to start high-dose IL-2 on the next morning, and then they received the first intradermal administration of MART-1_26–35 peptide-pulsed DC. IL-2 (Proleukin; Prometheus) was given at 600,000 IU/kg i.v. every 8 hours for up to 14 (amendments 1–7) or 9 (amendments 8–10) doses, as tolerated, following the standard high-dose IL-2 UCLA protocol (16). Standard supportive care was provided including filgrastim (Neupogen; Amgen), antibiotics for neutropenic fever, and red blood cell and platelet transfusions. When patients had recovered peripheral blood cell counts and were transfusion-free, they were discharged from the hospital. Patients received two more doses of DC vaccines at 2 weeks intervals. Patients were eligible to receive 3 more booster DC vaccines after study day 90 if the level of MART-1 TCR transgenic cells was below 5% of total lymphocytes by MHC tetramer or dextramer assay. A research [¹⁸F]FDG positron emission tomography (PET) scan and biopsy was planned at around day 30, and formal restaging PET computed tomography (CT) scans were performed on day 90.

Safety was determined in stage I, and if 3 of 8 patients have MART-1 F5 TCR-induced dose-limiting toxicities (DLT), then further accrual would not be warranted. Adverse events were analyzed following NCI CTCAE v3.0. The known toxicities and side effects of the chemotherapy preparative regimen, or from the administration of IL-2 or G-CSF, as listed in the protocol or package insert, were not considered for the assessment of DLTs. Analysis of patient blood samples for potential replication competent retrovirus and cytokine storm is described in the Supplementary Online Methods.

Feasibility was also determined in the first stage, and if 3 of 8 patients could not receive the intended cellular therapies, or if they resulted in suboptimal TCR transgenic cell in vivo persistence, further accrual would not be warranted to the protocol as originally designed. Feasibility was assessed after the first 8 patients were followed up for a minimum of 3 months after the last subject had received the infusion of the MART-1 F5 TCR transgenic cells.

Quantification of changes in PET imaging for the intratumoral accumulation of [¹⁸F]FDG was performed by counting the total number of FDG avid lesions as well as the maximum standardized uptake value (SUVmax) averaged over up to 5 hottest lesions at baseline, at day 30, and day 90. Objective clinical response rate was assessed on study day 90 and recorded following a modified Response Evaluation Criteria in Solid Tumors (RECIST; ref. 18).

MHC tetramer analysis in cryopreserved peripheral blood mononuclear cell (PBMC) collected at different time points was performed as previously described (19, 20). Our previous definitions for a positive or negative immunologic response using standardized MHC tetramer assays were used, which are based on the assay performance specifications by defining changes that were beyond the assay variability with a 95% confidence level (19).

The Simon optimal two-stage design (15) was used to determine the sample size, using the coprimary endpoint of response rate as the criterion for the clinical trial statistical design. This clinical trial was set up to rule out the null hypothesis that P₀ ≤ 0.10 (i.e., to rule out that this combined therapy has a response beyond 10%, since several current treatment approaches achieve response rates of 10% in patients with advanced melanoma) versus the alternative that the effect size = P₁–P₀ > 0.25 (α ≤ 0.05, β ≤ 0.20). The alternative hypothesis of a response rate of 35% was chosen because it was felt to represent a clinically meaningful difference and results in a study sample size that is feasible to be conducted within a pilot single-institution study. Using this statistical design, if 2 or more of 8 patients in stage I had an objective response at day 90, the study would proceed to stage II and accrue a total of 22 patients. If 5 or more patients in the overall study had a complete response or a partial response, defined as the objective response rate at day 90, the study would be declared positive. Because of the small sample size, the statistical analyses are mostly descriptive. Descriptive statistics such as mean, SD, median, minimum, maximum, and frequency of variables of interest are calculated. Responses of individual patients are plotted over time to present the data.

Between April 2009 and September 2011, 14 HLA-A*0201–positive patients with MART-1–positive metastatic melanoma were accrued. Patients had a median age of 50 years. Nine patients had M1c metastatic melanoma with visceral and/or bone metastases, 4 had lung metastases (M1b), and 1 had skin, nodal, and subcutaneous-only metastases (M1a). Half of the patients had received prior therapy for metastatic disease, including high-dose IL-2 in 4 patients and ipilimumab in 3 patients (Table 1).

This study underwent nine protocol amendments during its conduct, with some significantly changing the delivered cell therapies; therefore, we analyzed patients in subgroups based on the protocol amendments. The first 9 patients (F5-1 to F5-9) were treated under the original protocol (amendments 1–5) administering up to 1 × 10⁹ previously cryopreserved TCR transgenic lymphocytes following the full cyclophosphamide–fludarabine conditioning regimen and up to 14 infusions of high-dose IL-2. One patient (F5-5) had brain metastases at the baseline MRI screening exam and did not receive the TCR transgenic infusion. Therefore, this patient is not accounted in the safety, feasibility, or efficacy evaluations. After the planned study endpoint assessment based on the first 8 patients receiving the full protocol, it was deemed that the study was safe and feasible, but the antitumor activity assessed on day +90 was suboptimal. Therefore, patients F5-10 and F5-11 were treated under amendments 6–7 that allowed an increase in the cell number of previously cryopreserved TCR transgenic lymphocytes up to 1 × 10¹⁰. Because of an event of delayed pancytopenia in patient F5-10 who had continuing evidence of durable antitumor activity at that time, and fludarabine being the most likely cause of marrow toxicity, the study was further amended (amendments 8–9, patients F5-12 to F5-14) to decrease the conditioning regimen (one fewer day of fludarabine). Because this patient also had received 14 doses of IL-2, a potential adverse contributor to the marrow aplasia, which was more than what the other patients tolerated except for F5-4, the number of potential IL-2 doses was also limited to a maximum of 9. Studies in murine model of ACT showed that cryopreservation has a profound detrimental effect on the in vivo long-term survival of the adoptively transferred T cells and their ability to have a secondary response to antigen exposure. Therefore, we amended to administer noncryopreserved, freshly produced TCR transgenic lymphocytes (Fig. 1B).

TCR transgenic cell preparations were manufactured for all patients that met the lot release criteria of viability >70%, negative for mycoplasma, Gram and fungal stain, endotoxin level of ≤ 5 EU/kg body weight, and >10% MART-1 tetramer-positive CD3 lymphocytes (Supplementary Fig. S1). On average, transduction efficiency was 64% (range, 42%–84%; Supplementary Figs. S2 and S3). The first 10 patients received cryopreserved TCR-engineered cells, with the first 8 receiving up to 1 × 10⁹ cells and the next 2 patients receiving up to 1 × 10¹⁰ cells. However, there was not much difference in the absolute number of delivered MART-1 TCR transgenic cells (Supplementary Fig. S3). Despite administering a similar number of cells to these patients, peripheral cell reconstitution varied widely (Fig. 2). In this group of patients, the mean and median peak blood levels of TCR transgenic cells were 31%, with a maximum of 63% and a minimum of 0.67% of total CD3⁺ T lymphocytes. The peak of TCR transgenic cell frequency was early, within the first 2 weeks after ACT, and the percentage and absolute number of TCR transgenic cells in peripheral blood decreased over time in all patients. There was no clear evidence of enhancing number or function of peripheral blood TCR transgenic cells with the delivery of MART-1 peptide-pulsed DC vaccines on study days 14 and 28.

There was evidence of initial antitumor activity in 8 of 10 patients with metastatic melanoma regression detected at day 30 PET scans (examples in Figs. 3 and 4, and additional explanation in Supplementary Online Methods) or physical examination. However, the initial antitumor activity was incomplete and transient. At the formal restaging PET CT scans on day 90, none of the patients had evidence of a sustained tumor response by RECIST criteria. Therefore, after the first 8 patients had been enrolled, it was deemed that improvements in the protocol were needed. This led to the testing of a higher TCR transgenic cell dose, and then the testing of the infusion of freshly manufactured cells, without cryopreservation.

There were no unanticipated serious toxicities in the first 8 patients in the protocol beyond the ones that would be expected with the conditioning chemotherapy and high-dose IL-2 therapy. Of note, there were no findings of otologic, vestibular, or ophthalmologic toxicities after repeated specialist visits. Therefore, the protocol was deemed to meet the coprimary endpoint of safety when administered as 1 × 10⁹ cryopreserved cells. Because the antitumor activity did not meet the prespecified criteria to proceed to the second stage of study, the next 2 patients were treated with up to 1 × 10¹⁰ cryopreserved cells. However, the outcomes did not improve because one patient had a transient tumor response and the other had stable disease (Supplementary Figs. S4–S6).

A new protocol amendment was introduced to administer 1 × 10¹⁰ TCR-engineered cells to be infused fresh after harvesting from the ex vivo culture, without a cryopreservation step. The baseline leukapheresis was performed on the day of hospital admission, and cells were manufactured concurrent to patients receiving conditioning chemotherapy. The conditioning chemotherapy was shortened by one day and the number of potential IL-2 doses was limited at nine. Three patients were enrolled and had evidence of transient tumor responses by serial X-rays and PET scans, but also had a more pronounced whole body erythematous skin rash compared with the majority of prior patients in this protocol. Furthermore, 2 of the patients (F5-12 and F5-14) had serious adverse events (SAE) of acute respiratory distress requiring intubation associated with patchy pulmonary infiltrates within 1 week of cell infusion (Supplementary Fig. S7), resulting in the discontinuation of this cohort due to increased toxicities. Plasma from peripheral blood was analyzed for the production of multiple cytokines to study the potential development of a cytokine storm (Supplementary Fig. S8). Both patients received corticosteroid therapy and recovered their baseline respiratory function within 2 weeks.

Despite this increased toxicities with the infusion of fresh TCR transgenic cells and the administration of corticosteroids in these 2 patients and lack of administration of further MART-1/DC vaccinations, this group of 3 patients all had evidence of antitumor activity. The serial chest X-rays shown in Supplementary Fig. S7 document a time-course decrease in size of lung metastases in patients F5-12 and F5-14, and the CT scan images from patient F5-13 in Fig. 4 demonstrate the regression of large subcutaneous/muscle metastases. However, none of these responses were durable with all 3 patients having disease progression within 6 months from study start.

The infusion of fresh TCR transgenic lymphocytes resulted in a more prolonged persistence of circulating TCR transgenic cells in blood in the 3 patients from whom we had samples for analysis compared with the previous experience with cryopreserved cells (Fig. 5). It is remarkable that both CD4⁺ and CD8⁺ TCR transgenic cells were approximately 20% of the peripheral T lymphocytes at 3 months after ACT in F5-12 and F5-13. Furthermore, patient F5-13, who did not have a SAE and received the three scheduled MART-1/DC administrations, had evidence of recall whole body rash and reexpansion of the TCR transgenic cells in peripheral blood, demonstrating in vivo activation impact of MART-1/DC vaccination (Supplementary Fig. S9).

The MART-1 TCR ACT protocol described herein resulted in a high rate of transient tumor responses. These results are not too different from the studies at the NCI Surgery Branch despite multiple differences in the cell manufacture protocol (see Supplementary Online Discussion), primarily our short 1-week ex vivo culture using basic equipment available in any facility, resulting in younger cells and the use of DC vaccination. The generation of transgenic T cells, TILs, and endogenous antigen-specific T cells that have been used in the trials so far involved complex processes and at least 4 to 8 weeks of culture, thereby feasible only in a few specialized centers, and many of these patients progressing rapidly do not have this long a time to wait. Tumor progression after an initial response may be due to loss of antitumor activity of the TCR transgenic cells given to patients or changes in tumor cells resulting in acquired resistance to this mode of immunotherapy. The decrease in frequency of TCR transgenic T cells after the initial brisk expansion is a logical evolution of an effective immune response, as is commonly noted with T-cell responses to viral infections (21). In other work (22), we analyzed TCR transgenic T cells administered and recovered from 3 of the patients in this series using new-generation microfluidics-based miniaturized assays able to simultaneously study multiple functional responses of T cells selected based on defined antigen specificities (23, 24). These studies showed that the initial polyfunctionality resulting in high antitumor activity of the administered TCR transgenic T cells is gradually lost over time in vivo, which is temporarily associated with the clinical course of initial tumor response followed by progression. Therefore, there is a need to better maintain the TCR transgenic cell polyfunctionality upon ACT. This could be achieved pharmacologically using monoclonal antibodies blocking negative immune checkpoints (CTLA4; PD-1; ref. 25), with small molecules fostering T memory cell function (26, 27), or by endogenously generating fully active TCR transgenic cells from stem cell precursors (28–31).

Patients who experienced respiratory distress with the administration of noncryopreserved TCR transgenic cells demonstrated an increase in circulating cytokines and chemokines, nevertheless lower than the levels noted in two clinical reported cases of life-threatening cytokine storm (32, 33). In fact, these cytokine levels are comparable with those observed in acute pneumonia (34). Despite these toxicities, this group of patients had favorable antitumor activity with evidence of longer persistence of circulating TCR transgenic cells and in vivo expansion following DC vaccination.

In conclusion, a short ex vivo manufacture protocol was able to generate large numbers of tumor-specific TCR transgenic T cells. Administration of these in combination with DC vaccination is feasible and has high initial antitumor activity. These two cell therapies can be concurrently manufactured while patients undergo conditioning with lymphodepleting chemotherapy over a period of 1 week. The administration of freshly manufactured cells without cryopreservation results in higher in vivo persistence, but also in a higher incidence of side effects when targeting the self-melanosomal antigen MART-1. However, improvements in the conditions to maintain TCR transgenic cell functionality, tumor responses, and to lower toxic side effects are needed.

No potential conflicts of interest were disclosed.

Conception and design: T. Chodon, B. Chmielowski, R.C. Koya, D.W. Gjertson, B. Mukherji, L. Yang, J.A. Zack, O.N. Witte, D. Baltimore, J.S. Economou, A. Ribas

Development of methodology: T. Chodon, B. Chmielowski, R.C. Koya, C.G. Radu, D.W. Gjertson, A.J. Cochran, K. Cornetta, B. Mukherji, L. Yang, D. Baltimore, A. Ribas

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): T. Chodon, B. Comin-Anduix, B. Chmielowski, R.C. Koya, Z. Wu, M. Auerbach, C. Ng, E. Avramis, T.A. McCannel, A. Ishiyama, J. Czernin, K. Cornetta, D.J.L. Wong, P. Kaplan-lefko, O. Hamid, W. Samlowski, P.A. Cohen, G.A. Daniels, J.R. Heath, J.A. Glaspy, A. Ribas

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): T. Chodon, B. Comin-Anduix, B. Chmielowski, R.C. Koya, M. Auerbach, E. Avramis, J. Czernin, X. Wang, D.W. Gjertson, A.J. Cochran, J.R. Heath, J.A. Glaspy, J.S. Economou, A. Ribas

Writing, review, and/or revision of the manuscript: T. Chodon, B. Chmielowski, R.C. Koya, M. Auerbach, E. Avramis, T.A. McCannel, J. Czernin, X. Wang, A.J. Cochran, K. Cornetta, W. Samlowski, G.A. Daniels, J.A. Zack, D.B. Kohn, D. Baltimore, A. Ribas

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): T. Chodon, Z. Wu, E. Avramis, E. Seja, A. Villanueva, T.A. McCannel, P. Kaplan-lefko, O.N. Witte, D. Baltimore, A. Ribas

Study supervision: T. Chodon, B. Chmielowski, R.C. Koya, A.J. Cochran, J.S. Economou, A. Ribas

The authors thank Steven A. Rosenberg, Richard Morgan, Laura Johnson, and Mark Dudley (all from the NCI Surgery Branch, Bethesda, MD) for their guidance in establishing the TCR-engineered ACT protocol at UCLA and allowing access to their clinical grade retroviral vector master cell bank. They also acknowledge the contributions of Erika von Euw, Joanne Cox, and Narsis Attar in some of the cell therapy preparations.

This work was funded by the NCI grants P01 CA132681 (to D. Baltimore, J.S. Economou, O.N. Witte, A. Ribas, D.B. Kohn, and J.A. Glaspy), U54 CA119347 (to J.R. Heath and A. Ribas), P50 CA086306 (to A. Ribas), and RO1 CA129816 (to J.S. Economou), the California Institute for Regenerative Medicine New Faculty Award RN2-00902-1 (to A. Ribas), the Eli & Edythe Broad Center of Regenerative Medicine and Stem Cell Research at UCLA (to O.N. Witte and A. Ribas), the Samuel Waxman Foundation (to J.S. Economou, D. Baltimore, and O.N. Witte), the Keck Foundation (to J.S. Economou, D. Baltimore, and O.N. Witte), The Seaver Institute (to A. Ribas), the PhaseOne Foundation (to A. Ribas), the Louise Belley and Richard Schnarr Fund (to A. Ribas), the Wesley Coyle Memorial Fund (to A. Ribas), the Garcia-Corsini Family Fund (to A. Ribas), the Bila Alon Hacker Memorial Fund (to A. Ribas), the Fred L. Hartley Family Foundation (to A. Ribas), the Ruby Family Foundation (to A. Ribas), the Joy and Jerry Monkarsh Fund (to J.S. Economou), the Caltech/UCLA Joint Center for Translational Medicine (to D. Baltimore, O.N. Witte, and A. Ribas), and the Melanoma Research Alliance (to A. Ribas, D. Baltimore, and J.R. Heath). O.N. Witte is an investigator of the Howard Hughes Medical Institute. R.C. Koya was supported in part by the V Foundation-Gil Nickel Family Endowed Fellowship in Melanoma Research. The UCLA Jonsson Comprehensive Cancer Center (JCCC) Flow Cytometry Core Facility is supported by NIH awards CA-16042 and AI-28697.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.