Engineered T Cells in Synovial Sarcoma: Persistence Pays Off!

Synovial sarcoma (SS) is an aggressive malignancy that disproportionately affects younger patients. SS represents 10% to 20% of all soft-tissue sarcomas in the adolescent and young adult population, with median age at diagnosis between 20 and 30 years (1). In more than 90% of cases, SS is defined by the presence of the t(X;18)(p11.2;q11.2) translocation, which results in fusion of the SS18 (formerly SYT) gene on chromosome 18 and one of several synovial sarcoma X (SSX) genes on chromosome X. The standard treatment for patients with localized SS is complete surgical resection with or without radiotherapy. Metastatic SS is considered more responsive to chemotherapy and targeted agents than other soft-tissue sarcomas, and yet the median overall survival is 15 months and the 5-year survival is less than 5% (2).

Despite major advances made in a growing number of malignancies (including melanoma, urothelial carcinomas, non–small cell lung carcinoma, and microsatellite instability–high malignancies) with the introduction of immune checkpoint blockade, these agents have been ineffective in SS, in which there is generally a lack of tumor-associated inflammatory infiltrate (3).The basis of immune evasion in SS is unknown, although mechanisms besides those currently targeted by immune checkpoint inhibitors (anti-CTLA4 and anti–PD-1) are likely. Specifically, an early study of ipilimumab in SS was closed early for toxicity and futility (4). SARC028 had a dedicated cohort for SS that included 10 patients, of whom only 1 had a short-lived partial response and others progressed rapidly (5). The combination of ipilimumab and nivolumab was also ineffective, although only 2 patients with SS were included (6). As with other translocation-associated tumors, SS has low genetic complexity, including low burden of nonsynonymous somatic mutations (1).

Although there have been limited attempts to directly target the SS18–SSX oncoprotein in SS (7), recent efforts have focused on targeting NY-ESO-1, a cancer testis antigen that is strongly and diffusely expressed in 70% to 80% of SS (8). Past and ongoing clinical trials of adoptive T-cell therapies (ACT) targeting NY-ESO-1 in metastatic SS have shown promise (Table 1). In 2011, Robbins and colleagues reported the first experience with adoptive transfer of T cells engineered to express NY-ESO-1^c259, an affinity-enhanced T-cell receptor (TCR) recognizing an HLA-A2–restricted NY-ESO-1/LAGE-1a–derived peptide, in patients with metastatic melanoma or metastatic SS whose tumors stained strongly for NY-ESO-1 (9). Patients received autologous TCR-transduced T cells plus IL2 following lymphodepletion with fludarabine (25 mg/m²/day for 5 days) and cyclophosphamide (60 mg/kg/day for 2 days). Objective responses were observed in 4 of the 6 patients with SS. Neither in this study nor in a subsequent report that included results for an additional 12 patients treated with SS did persistence of engineered T cells 1 month after T-cell infusion correlate with response to therapy (10).

In this issue of Cancer Discovery, D'Angelo and colleagues (11) evaluated the safety and feasibility of treating patients with metastatic SS expressing NY-ESO-1 with autologous T cells engineered to express NY-ESO-1^c259 without exogenous IL2 following lymphodepletion with fludarabine (30 mg/m²/day for 5 days) and cyclophosphamide (60 mg/m²/day for 2 days). The overall response rate in this study was 50% (6 of 12 patients), with 1 complete response and 5 partial responses. The median time to initial response was 6.2 weeks (range, 4–9 weeks), and median duration of response was 30.9 weeks (range, 13–72 weeks). Treatment-related ≥ grade 3 adverse events (AE) were reported in 11 of 12 patients. AEs were most commonly cytopenias and hypophosphatemia and likely largely attributable to the lymphodepleting chemotherapy regimen. However, despite eliminating IL2, 5 patients did experience cytokine release syndrome.

D'Angelo and colleagues show that in all patients the infused NY-ESO-1^c259 T cells not only expand but persist in the circulation, as measured by circulating NY-ESO-1^c259 vector copy levels. Interestingly, higher peak NY-ESO-1^c259 vector copy levels were observed in responders compared with nonresponders (median 106,174 vs. 30,601 vector copies/μg). Circulating NY-ESO-1^c259 TCR-expressing T cells were detectable in all 7 patients for whom monitoring was continued beyond 200 days (6 responders and 1 nonresponder). Additionally, these engineered NY-ESO-1^c259 T cells remained functional by intracellular cytokine staining and did not express exhaustion markers such as PD-1 and LAG3.

Next, D'Angelo and colleagues characterized the memory cell phenotype of NY-ESO-1^c259 T cells both in the manufactured product prior to infusion and in the circulation at longitudinal time points. NY-ESO-1^c259 T cells in the manufactured product were comprised predominantly of effector memory (CCR7⁻CD45RA⁻) or effector memory RA⁺ (CCR7⁻CD45RA⁺) phenotype. Although limited by small patient numbers, there were no distinguishable differences between manufactured products received by responders versus nonresponders. Upon adoptive transfer, the authors observed an increase in central memory (CCR7⁺CD45RA⁻) and stem cell memory (CCR7⁺CD45RA⁺) cells among the CD4⁺ and CD8⁺ NY-ESO-1^c259 T cells in circulation.

The authors then deeply profiled circulating NY-ESO-1^c259 T cells and tumor biopsy available from 1 patient who achieved a partial response following ACT but later relapsed despite persistence of functional NY-ESO-1^c259 T cells in circulation. Circulating NY-ESO-1^c259 T cells from this patient at time of relapse that were exposed to antigen ex vivo were capable of inducing robust killing in the absence of exogenous cytokines. There was maintenance of a subset of CCR7⁺ and CD45RA⁺ cells that the authors postulate are capable of self-renewal and differentiation and underlie the continued presence of NY-ESO-1^c259–expressing T cells even beyond 6 months after initial engineered T-cell infusion. Interestingly, however, although this patient's tumor at the time of relapse remained NY-ESO-1–positive, there was no evidence of tumor infiltration by NY-ESO-1^c259 T cells.

Although this study was limited with respect to availability of baseline, on-treatment, and at-relapse tumor and blood biopsies, the exciting findings reported here by D'Angelo and colleagues highlight the importance and impact of longitudinal biospecimen collection. Future trial designs in the ACT space should incorporate longitudinal tumor and blood biopsies, as analyses of these precious biospecimens will be critically important toward improving our understanding of cellular and molecular processes occurring in patients who respond to therapy as well as elucidating mechanisms of resistance in those who fail to respond.

The remarkable responses observed in this study are in stark contrast to the lack of therapeutic efficacy of checkpoint blockade in this malignancy, highlighting the fact that the defect in the immune response appears less dependent on T-cell exhaustion or the immunosuppressive microenvironment and more likely to be at the level of immune recognition and evasion. The overall lack of antigenic stimulation in SS perhaps points to strategies to enhance antigen release and presentation either through epigenetic agents, radiotherapy, agonist antibodies, vaccines, or even intratumoral injectables. In those combinations, there could still be a role for anti-CTLA4 antibodies to augment T-cell priming and activation. It is unclear if PD-1 blockade would still be an attractive approach, although the recent success of combinations involving chemotherapy and anti–PD-1 antibodies suggests a potential role for such combinations in SS.

An important aspect of this study was the elimination of high-dose IL2 from the regimen with arguably improved outcomes as manifested by the clonal diversity and persistence of T cells, as well as the predominance of central memory and stem cell memory cells. This tackles one of the main limitations for the generalizability and wide adoption of ACT beyond a very select population that is capable of tolerating the toxicity of a lymphodepleting regimen and high-dose IL2. This modification to the regimen may expand the pool of eligible candidates for this and other ACT approaches. The risk–benefit profile of ACT must still be considered by clinicians, who weigh the potential for improved outcomes against the possibility of treatment-associated toxic effects, when selecting patients for novel therapies. Correlative studies will also allow identification of prognostic and predictive biomarkers that could improve patient selection for ACT.

In summary, D'Angelo and colleagues demonstrate the safety and feasibility of ACT in patients with autologous T cells engineered to express NY-ESO-1^c259 in patients with metastatic SS and observed a 50% clinical response of variable durability. The authors demonstrate expansion of the infused NY-ESO-1^c259 T cells in the absence of exogenous IL2 in all 12 patients who received therapy as well as persistence of NY-ESO-1^c259 TCR-expressing T cells in circulation correlating with 6 responders of 7 patients tested. These results suggest that adoptive T-cell therapy can induce a robust antitumor response and give rise to a persistent pool of NY-ESO-1^c259 T cells that effect meaningful and durable responses in patients with metastatic SS and can form a backbone for future therapeutic strategies.

H.A. Tawbi has received honoraria from the speakers bureaus of BMS, Novartis, Merck, and Genentech/Roche. No potential conflicts of interest were disclosed by the other author.

E.Z. Keung is supported by National Institutes of Health (NIH) grant T32 CA009599.