Safety and Tolerability of Letetresgene Autoleucel (GSK3377794): Pilot Studies in Patients with Advanced Non–Small Cell Lung Cancer Open Access

The use of immunotherapy with immune checkpoint inhibitors (ICI) targeting the PD-1/PD-L1 axis alone or in combination with anticytotoxic T lymphocyte–associated protein 4, with or without chemotherapy, has been associated with improved outcomes compared with chemotherapy alone in non–small cell lung cancer (NSCLC; refs. 1–6). Despite these improvements, primary or acquired resistance to ICIs remains a clinical challenge (7). Targeted therapies specific to oncogenic driver mutations, such as EGFR, anaplastic lymphoma kinase (ALK), and ROS1, have substantially enhanced outcomes compared with chemotherapy in a subset of patients (8, 9), but the development of resistance to targeted therapy represents another challenge. ICIs are not effective in patients with certain tumor genomic aberrations such as activating EGFR mutations, ALK, and ROS1 translocations, and patients with these tumors are commonly excluded from immunotherapy trials owing to their lack of responsiveness to checkpoint blockade (10).

Strategies aimed at establishing de novo antitumor immunity or rescuing native antitumor immunity via delivery of engineered T cell–based cellular therapies have demonstrated promise in multiple solid tumors, including in lung cancer (11–16). T-cell receptor (TCR)–based therapy uses genetically engineered lymphocytes to target antigens derived from tumor-specific proteins [e.g., cancer testis antigens (CTA) and mutated tumor antigens, such as KRAS] presented by HLA molecules on the tumor cell surface as “non-self neoantigens” and can generate a T-cell immune response (17). The treatment process with TCR-based therapies involves patient screening for HLA typing and target tumor-specific protein, leukapheresis, generation of TCR-transduced T cells, lymphodepletion, and infusion of the TCR-transduced T cells (18). The CTA New York esophageal squamous cell carcinoma 1 (NY-ESO-1)/LAGE-1a are appealing targets because of their known immunogenicity and their upregulated expression on certain tumors while maintaining restricted expression on normal tissue. The NY-ESO-1 protein has been shown to be generally homologous to LAGE-1a (approximately 84% shared protein sequence identity; refs. 19–22). Efficacy of T-cell therapy targeting NY-ESO-1 has been demonstrated in previous clinical trials in hematologic malignancies (e.g., multiple myeloma), as well as in patients with synovial sarcoma and melanoma (23–30).

Letetresgene autoleucel (lete-cel; GSK3377794) is an affinity-enhanced T-cell therapy that consists of autologous CD4⁺ and CD8⁺ T cells genetically modified to express a TCR that recognizes the NY-ESO-1/LAGE-1a shared epitope, SLLMWITQC, upon presentation on HLA-A*02:01, HLA-A*02:05, or HLA-A*02:06 (24, 26). Lete-cel has shown significant activity in synovial sarcomas, an immunologically “cold” tumor in which NY-ESO-1 is upregulated at a high level (28). NY-ESO-1 and LAGE-1a are also upregulated in both driver oncogene-positive and negative NSCLC, making them potential targets for lete-cel in patients with NSCLC (31).

Based on the efficacy of adoptive T-cell transfer with lete-cel demonstrated in other tumor types, we conducted two pilot studies to evaluate the safety, tolerability, and efficacy of lete-cel alone (study 208749, NCT02588612) and in combination with pembrolizumab (study 208471, NCT03709706) in patients with advanced NSCLC. To date, these are the largest clinical studies of a TCR therapy in NSCLC.

Study 208749 is a single-arm, open-label pilot study evaluating the safety and clinical activity of lete-cel in patients with stage IIIB or IV, or recurrent, NSCLC, whereas study 208471 is a multiarm, open-label, phase Ib/IIa study evaluating the safety and clinical activity of lete-cel alone (arm A) or in combination with pembrolizumab (arms B and C) in patients with stage IIIB or IV, or recurrent, NSCLC (32). The original sponsor for the single-arm study was Adaptimmune (ADP-0011-004), with sponsorship transferring to GSK prior to study closure. Both studies complied with the ethical principles of the Declaration of Helsinki. An overview of each study design is shown in Supplementary Fig. S1; a high-level comparison between the two studies is shown in Table 1. Key inclusion and exclusion criteria for the single-arm and multiarm studies are shown in Supplementary Table S1A and S1B, respectively. Eligible patients were ≥18 years of age, diagnosed with advanced NSCLC (stage IIIB or IV), were HLA-A*02–positive (HLA-A*02:01, HLA-A*02:05, and/or HLA-A*02:06), and had tumors expressing NY-ESO-1 or LAGE-1a evaluated on tumor tissue from a formalin-fixed, paraffin-embedded archival sample or fresh biopsy.

The primary objective of the single-arm study was to evaluate the safety and tolerability of lete-cel in the selected patient population. For the multiarm study, the primary objectives were to evaluate the safety, tolerability, and efficacy of lete-cel in patients who received treatment alone or in combination with pembrolizumab (Table 1).

Study protocols and patient informed consent documentation were approved by institutional review boards. Written informed consent was required from all study patients prior to study initiation.

The initial target enrollment for the single-arm monotherapy study was 10 patients; patients were pre-screened for the relevant HLA alleles (HLA-A*2:01, HLA-A*2:05, and/or HLA-A*02:06) and target antigen expression (NY-ESO-1/LAGE-1a) through a separate prescreening protocol, ADP-0000-001 (NCT02636855) before entering the screening phase of the protocol to determine eligibility for enrollment (Supplementary Fig. S2). Eligibility criteria were grouped into two parts and eligibility screening involved the following two steps: leukapheresis eligibility screening and lymphodepletion eligibility screening. Patients were planned to receive lete-cel monotherapy as a single intravenous infusion. Patients who had a confirmed response and documented disease progression, and whose tumors continued to express NY-ESO-1 and/or LAGE-1a, were considered for a second lete-cel infusion.

The initial target enrollment for the multiarm immunotherapy study was 45 dosed patients, or 15 per arm. Eligibility criteria were grouped into four parts and eligibility screening involved the following four steps: target expression screening, leukapheresis eligibility screening, lymphodepletion eligibility screening, and pembrolizumab eligibility screening. Patients who did not have actionable genetic aberrations per National Comprehensive Cancer Network (NCCN) guidelines (33) were assigned to receive either lete-cel monotherapy as a single intravenous infusion (arm A) or a single intravenous infusion of lete-cel on day 1 followed by pembrolizumab to be initiated on day 22 (arm B; Supplementary Fig. S3), with patients being assigned to arm A first and then to arm B. Any patients in arm A who progressed within 25 weeks following lete-cel infusion had the option of receiving pembrolizumab following a benefit–risk evaluation. Patients with actionable genetic aberrations (e.g., sensitizing EGFR mutations or ALK translocations) per NCCN guidelines were assigned to arm C and received the same treatment as patients in arm B. There was no re-treatment with lete-cel in this study.

For target expression screening, HLA type was evaluated using DNA extracted from a blood sample, and archival or fresh tumor tissue was tested for NY-ESO-1 and LAGE-1a expression.

In the multiarm study, the consented patient population was screened using an investigational-use only IHC clinical trial assay for NY-ESO-1 protein expression using the E978 anti-NY-ESO-1 mAb, and an investigational-use only RT-PCR clinical trial assay for LAGE-1a RNA expression detection.

NY-ESO-1 IHC–positive tumor expression was defined as P score ≥ 10%, 1+, 2+, and 3+, and LAGE-1a RNA–positive tumor expression was defined as cut-off = 4 dCT and below. LAGE-1a testing was done in reflex to NY-ESO-1 IHC–negative results (<10%, 1+, 2+, and 3+ IHC; P score). Tumor samples submitted for testing included archival specimens or fresh biopsies from either a primary or metastatic lesion. Calculations for NY-ESO-1 and LAGE-1a prevalence excluded repeat testing or nonunique specimens.

For both studies, patients who met all eligibility criteria following screening were enrolled and underwent leukapheresis to obtain autologous cells for the manufacture of lete-cel (NY-ESO-1^c259 TCR T-cell investigational product; ref. 29). In the single-arm study, lete-cel was manufactured via pilot manufacturing supply, whereas in the multiarm study, lete-cel was manufactured via either pilot or intended commercial manufacturing supply. The pilot lete-cel manufacturing process has been previously described (29); the intended commercial supply followed a similar manufacturing process.

Lete-cel was administered to patients following lymphodepleting chemotherapy with cyclophosphamide and fludarabine (27, 34). A range of 1 to 8 × 10⁹ transduced T cells for the single-arm study and 1 to 15 × 10⁹ transduced T cells for the multiarm study was administered by a single intravenous infusion on day 1 for both studies. The cell dose range was selected based on manufacturing capabilities (35–37). For both studies, planned lymphodepletion consisted of fludarabine 30 mg/m²/day at days −8, −7, −6, and −5 and cyclophosphamide 900 mg/m²/day intravenously at days −7, −6, and −5; adjusted for age ≥60 years, renal impairment, increased body weight, or risk of prolonged cytopenia, among others. Refer to study protocols for more information. Lymphodepletion could be given as outpatient treatment at the discretion of the investigator. Patients received mesna (sodium 2-mercaptoethane sulfonate) per institutional guidelines and G-CSF starting 24 hours following the last dose of chemotherapy until neutrophil count recovery.

Patients in arms B and C of the multiarm study received pembrolizumab at the approved dose of 200 mg every 3 weeks. The first administration was on day 22 after administration of lete-cel and then every 3 weeks for up to 35 cycles or until disease progression.

In both studies, the primary safety endpoints included assessment of adverse events (AE). In the multiarm study, the primary efficacy endpoint was investigator-assessed objective response rate (ORR), per RECIST v1.1.

Secondary efficacy endpoints in both studies included time to response, duration of response (DOR), disease control rate, and progression-free survival (PFS). In addition, the single-arm study also evaluated investigator-assessed ORR (per RECIST v1.1) as a secondary endpoint. Exploratory endpoints for both studies included the correlation between lete-cel persistence and response. Both studies also evaluated overall survival (OS) as an exploratory endpoint.

Following lete-cel infusion, patients were monitored closely and received appropriate supportive care measures. Patients routinely underwent physical exams, safety evaluations, medical imaging, and biological sample collection until the end of interventional phase criteria were met or until patient withdrawal. AEs of special interest (AESI) reported in both studies included cytokine release syndrome (CRS), immune effector cell–associated neurotoxicity syndrome (ICANS), pancytopenia/aplastic anemia, graft-versus-host disease, Guillain–Barre syndrome, and pneumonitis. In the multiarm study, AESIs of grade 4 neutropenia lasting ≥28 days and treatment-related inflammatory response at tumor sites were also reported; in addition, all single-cell cytopenias were listed based on reported laboratory results. Refer to study protocols for more information.

The end of the interventional phase was defined as completion of the follow-up period in the interventional phase, disease progression per RECIST v1.1 in the single-arm study or confirmed disease progression per immune-related RECIST (iRECIST) in the multiarm study, or death after lete-cel infusion, whichever comes first.

On completion of the interventional phase or early withdrawal, patients were transitioned to a long-term follow-up (LTFU) study for observation of delayed AEs for 15 years after lete-cel infusion, in accordance with the FDA and the European Medicines Agency requirements for gene therapy clinical trials.

Patient peripheral blood mononuclear cells (PBMC) were collected for monitoring the persistence of lete-cel–modified cells. DNA PCR-based methods were used to detect the Psi gene contained in the lentiviral vector used to transduce the T cells. Persistence, defined by the duration of maximum serum concentration (C_max; vector copies/µg of genomic DNA), was derived as time from T-cell infusion until C_max was no longer detectable.

Serum samples for determination of antibodies against lete-cel were taken from all participants for antidrug antibody testing.

Replication-competent lentivirus (RCL) was monitored using a PCR-based assay that detects and measures copies of the gene coding for the vector’s envelope protein, vesicular stomatitis virus G protein, which is necessary for the assembly of pseudotyped infectious lentiviral particles but is absent from the vector’s backbone. Vesicular stomatitis virus G protein quantitative PCR testing was carried out on lentiviral vector, T-cell product, and PBMCs at timepoints indicated in the schedule of assessments for each study protocol.

At the first instance of >1% of PBMCs testing positive for the Psi gene at or after 1-year post-infusion, the participant’s PBMCs were evaluated for integration site analysis by assessing clonality and the possibility of insertional oncogenesis following regulatory recommendations (38, 39).

For the single-arm study, the sample size was based on clinical judgment; enrollment of 10 participants was planned. The small study was not powered to conduct any statistical hypothesis testing on either primary or secondary endpoints. For the multiarm study, the sample size was determined using a Bayesian predictive adaptive design that allowed the study to be monitored more frequently while maintaining the desired type I error and power. The sample size was based on testing the hypothesis for each arm separately and not for a comparison between arms. Enrollment of approximately 18 participants was planned for each arm to ensure approximately 15 evaluable participants would receive lete-cel infusion within the target dose range.

The intention-to-treat (ITT) population included all patients who enrolled in the trial and met all eligibility criteria, and the modified ITT (mITT) population comprised patients who received lete-cel infusions. The mITT2 population included all patients who received a second lete-cel infusion.

Safety analyses were performed using the ITT and mITT populations for treatment-dependent safety analysis. For both studies, AEs (coded by Medical Dictionary for Regulatory Activity terms) were tabulated by System Organ Class and Preferred Term and were further classified by toxicity (using the NCI Common Terminology Criteria for Adverse Events, v4.03).

In both studies, the population used for the efficacy analysis was the mITT population. Tumor assessments for response and progression were evaluated according to RECIST v1.1 (40) and iRECIST (41). Any ORR endpoint was summarized by two-sided 95% confidence intervals (CI) using exact Clopper–Pearson methods. DOR (in the subset of patients with a confirmed response), PFS, and OS were summarized and displayed graphically using Kaplan–Meier methodology to estimate the median, 25th and 75th percentiles, and two-sided 95% CIs. Descriptive statistics were provided for disposition, demographics, and safety.

Final analyses for each study were conducted when all patients who received lete-cel infusions (mITT) moved to the separate LTFU protocol, declined LTFU, were lost to follow-up, withdrew early, or died. No interim analyses were planned for the single-arm study. Separate interim analyses were planned for each arm in the multiarm study after at least 10 patients were available in that arm. No formal statistical hypotheses were tested in either study.

The results summary for this study is available on clinicaltrials.gov, the default register for GSK Human Subject Research. All datasets analyzed in the article will be available upon request to the editors and peer reviewers, and to the community at the time of publication. For interventional studies that evaluate our medicines, patient-level data will be made available to independent researchers, subject to review by an independent panel, at www.clinicalstudydatarequest.com. To protect the privacy of patients and individuals involved in our studies, GSK does not publicly disclose patient-level data.

For the single-arm study, patients were confirmed as positive for HLA-A*02 and antigen expression under a separate prescreening protocol from Adaptimmune (ADP-0000-001, NCT02636855). More than 1,000 patients were pre-screened; these data are not included in the current article. Forty-one of the patients identified as positive for HLA-A*02 and antigen expression under ADP-0000-001 consented to the single-arm study to pursue screening (Supplementary Fig. S2).

A total of 1,698 patients in the multiarm study signed the screening informed consent. Of those, 1,638 patients were tested for HLA-A*02 and 738 (45%) patients were HLA-A*02–positive. Among the 525 patients who provided tissue for NY-ESO-1 testing, 62 (12%) were positive [samples from 38 (7%) patients were not evaluable]. Three hundred forty-eight patients were tested for LAGE-1a; 15 (4%) were positive [samples from six (2%) patients were not evaluable; Fig. 1A; Supplementary Fig. S3].

Individual patient demographics and characteristics of both mITT populations, including tumor stage and histology, antigen expression, lymphodepletion regimen, and lete-cel treatment, are shown in Table 2. Patient representativeness by sex, age, race/ethnicity, and geography is reported in Supplementary Table S2.

In the single-arm study, most of the patients in the mITT population were White (80%) and male (60%), with a median age of 59.0 years (range, 29–70 years). Four (80%) patients were HLA-A*02:01–positive, whereas one (20%) was found to be HLA-A*02:06–positive (Table 2). All had NY-ESO-1–positive tumors. In the multiarm study, most patients in the mITT population were White (92%) and female (54%), with a median age of 62.0 years (range, 49–77 years). Eleven (85%) patients were HLA-A*02:01–positive, two (15%) patients were HLA-A*02:06–positive (frequencies include one patient who was positive for both HLA-A*02:01 and HLA-A*02:06), and one (8%) patient was HLA-A*02:05–positive (Table 2). Eleven (85%) patients had NY-ESO-1–positive tumors, and two (15%) patients had LAGE-1a–positive tumors.

Dosed patients across studies received standard-of-care anticancer therapy prior to receiving lete-cel. This included targeted therapy for patients with actionable genetic aberrations, following NCCN or equivalent country-level guidelines (Table 2). In arm C of the multiarm study, four patients had actionable EGFR mutations and two had actionable ALK mutations (Table 2).

In the single-arm study, nine patients underwent leukapheresis, of whom five patients underwent lymphodepletion and T-cell infusion (Supplementary Fig. S2). Reasons for withdrawal prior to lymphodepletion were stable disease at the time of T-cell expiry, withdrawal at investigator’s decision, and low cell yield at apheresis. The median number of transduced T cells was 3.02 × 10⁹ (range, 0.84–4.98). One patient received a second lymphodepleting treatment and T-cell infusion following progression after the first lete-cel treatment. In the multiarm study, a total of 34 patients underwent leukapheresis, of whom 13 underwent lymphodepletion and lete-cel infusion (Fig. 1A; Supplementary Fig. S3). Seven and six patients were assigned to arms A and C, respectively, according to their genetic aberration status (Table 1). Arm B of the trial was never opened. Contributing to attrition rates among patients who underwent leukapheresis in the multiarm study were failed eligibility, lack of fitness, death due to underlying disease, withdrawal at patient/investigator discretion, and study closure. The median number of transduced T cells was 5.2 × 10⁹ (range, 1.0–8.1). Seven (54%) patients (three in arm A; four in arm C) received at least one infusion of pembrolizumab. Two patients in arm C did not receive pembrolizumab (one due to ongoing AEs and one due to disease progression and death). Administration of pembrolizumab was delayed in two patients in arm C due to AEs, including one who did not receive any doses at the time of disease progression). T-cell products meeting the protocol-defined minimum requirement of 1 × 10⁹ transduced T cells were successfully manufactured for 33 out of 34 patients who underwent leukapheresis in the multiarm study; one patient died in between apheresis and manufacturing, and thus, manufacturing never started. Of the 34 patients who underwent leukapheresis, four underwent repeat leukapheresis and nine required a second T-cell manufacturing process. Manufacturing data are not available for the single-arm study.

All dosed patients across studies were followed until completion of the interventional phase or withdrawal. The multiarm study was terminated early at the discretion of the sponsor without reaching criteria for interim analyses or the target of 15 dosed patients per treatment arm.

The NY-ESO-1 IHC assay in the multiarm study produced predominantly cytoplasmic and nuclear staining across the dynamic range of staining intensities (0+, 1+, 2+, and 3+; ref. 42). The expression profile distribution spanned from being fully negative (100%, 0+) to 100% of tumor cells stained, revealing a wide range of expression in a continuous variable fashion, as determined by P score (maximum of 100%; Fig. 1B). The NY-ESO-1 expression profile for the mITT population in the multiarm study is depicted in Fig. 1C and shows the distribution of both the P score and the relative contribution of staining intensity. Antigen analysis of tumor samples that were reflexed to LAGE-1a testing also revealed a wide distribution of LAGE-1a scores (Fig. 1D). The prevalence of NY-ESO-1 and LAGE-1a positivity was assessed by including unique specimens for all patients and excluding multiple tests of the same tissue block, resulting in 12% and 4% positivity, respectively (Fig. 1E). Nonevaluable samples were also included.

There were no fatal treatment-emergent AEs (TEAE) reported in both studies. In the single-arm study, cytopenias were the most common total and grade ≥3 TEAEs (Supplementary Table S3A). In the multiarm study, cytopenias and CRS were the most common TEAEs, and cytopenias were the most common grade ≥3 TEAEs (Supplementary Table S3B). In the single-arm study, four patients (80%) reported at least one serious TEAE [two patients (40%) experienced serious TEAEs of grade ≥3; Table 3]. In the multiarm study, eight patients across arms (62%) reported at least one serious TEAE (all were grade ≥3; Table 4). There were no serious TEAEs attributed to pembrolizumab.

AESIs (defined in footnotes of Tables 3 and 4) were reported in three (60%) patients in the single-arm study: two serious AEs (SAE) of CRS, one SAE of ICANS, and one non-SAE of pancytopenia. The two CRS events from the single-arm study were grades 1 and 3 and lasted from days 9 to 17 and 4 to 11 after T-cell infusion, respectively; tocilizumab was administered for the grade 3 event (Table 3). AESIs were reported in 12 (92%) patients in the multiarm study: nine AEs of CRS (two serious), two AEs of ICANS (one serious), and two AEs of pancytopenia (one serious; Table 4). The events of CRS in the multiarm study had an onset of 1 to 8 days following administration of lete-cel and a median duration of 5 days. The two serious events of CRS were both grade 2 and lasted from days 8 to 12 and 7 to 12 after T-cell infusion, respectively. The latter of the two serious events of CRS was treated with tocilizumab and corticosteroids. There were no reports of graft-versus-host disease or Guillain–Barre syndrome in either study. In both studies, all AESIs were considered related to study treatment, and all were reported as resolved.

In the single-arm study, one of the five patients (ORR 20%; 95% CI, 0.5–71.6) had a confirmed partial response (PR) following the first T-cell infusion. The remaining patients had stable disease (one patient) and progressive disease (three patients) as best response (Fig. 2A). In the multiarm study, six (86%) and two (33%) patients had stable disease as best response in arms A and C, respectively, and no patients had a PR or complete response (Fig. 2A and B).

In the single-arm study, four patients (80%) experienced disease progression per RECIST 1.1 and one (20%) had death from progression not confirmed by RECIST 1.1. In the multiarm study, 10 patients (77%) across arms [six (86%) patients in arm A and four patients (67%) in arm C] experienced disease progression per RECIST 1.1. One death was reported in arm C due to complications of underlying disease. Two patients withdrew from the study. The median PFS in the single-arm study was 1.81 months (95% CI, 0.46–18.23), whereas median PFS was 5.32 months (95% CI, 1.45–5.52) in arm A and 1.48 months (95% CI, 0.62–2.83) in arm C of the multiarm study. Median OS was 8.71 months (95% CI, 0.72–NA) in the single-arm study and 9.33 months (95% CI, 3.15–9.33) in arm A and not reached (95% CI, 0.62–NA) in arm C of the multiarm study (Fig. 2C).

The patient who responded in the single-arm study (Table 2 patient S5) was heterozygous for HLA-A*02:01 and had a poorly differentiated adenocarcinoma without EGFR, ALK, or ROS1 pretreatment alterations, or any other targetable oncogenic mutations. Patient S5 was previously treated with three lines of chemotherapy and immunotherapy (nivolumab) and received pemetrexed as a bridging therapy. The patient had a small reduction in lymphadenopathy that lasted 3 months following prior immunotherapy treatment. Radiographic images illustrating PR per RECIST 1.1 following lete-cel are shown in Fig. 2D. Response was durable until 18 months after initial infusion (Fig. 2E). Following disease progression, the patient was considered to have completed the interventional portion of the study and underwent tumor resection as part of standard of care. The patient received a second lymphodepleting regimen and T-cell infusion approximately 28 weeks after documented progressive disease after first infusion (and 22 weeks after tumor resection). Tumor assessment on initial 8-week scan after second infusion showed stable disease per RECIST 1.1 (with complete response in non-target lesions). The patient then withdrew consent prior to further disease evaluation. Therefore, no DOR data were collected following the second T-cell infusion.

All five patients who completed lymphodepletion and lete-cel infusion in the single-arm study completed the interventional phase, according to protocol definition (RECIST 1.1; Supplementary Fig. S2). Of the 13 patients who completed lymphodepletion and lete-cel infusion in the multiarm study, seven patients completed the interventional phase according to protocol definition (iRECIST; Supplementary Fig. S3). Six patients discontinued the interventional phase early, including four who withdrew prior to scan confirming disease progression per iRECIST (Supplementary Fig. S3).

In the single-arm study, peak cell persistence (defined as median C_max) was 6.8 × 10⁴ copies per µg in the responder, with shorter median time to C_max (T_max) in the responder versus non-responders (Fig. 3A). In the multiarm study, mean peak cell persistence was 1.7 × 10⁵ copies per µg in arm A (lete-cel monotherapy), with a median T_max of 3.0 days. Mean peak persistence was 1.3 × 10⁵ copies per µg in arm C (lete-cel and pembrolizumab combination therapy), with median T_max of 7.9 days. Mean AUC from 0 to 28 days was 2.2 × 10⁶ copies per µg × days in arm A and 1.7 × 10⁶ copies per µg × days in arm C (Fig. 3B).

Two patients in the single-arm study were tested for RCL; both were negative. No integration site analysis was performed because there were no Psi DNA copies of PBMCs >1% at 1-year post-treatment. In the multiarm study, all patients were tested for anti-GSK3377794 antibodies at least once and were negative at all timepoints during the study. All patients were tested for RCL at least once and were negative at all timepoints during the study. No patients remained in the study 1-year post-treatment.

We have reported the findings from the 208749 single-arm, open-label pilot study evaluating lete-cel alone and the 208471 multiarm, open-label, phase Ib/IIa study evaluating lete-cel, either alone or in combination with pembrolizumab, in advanced NSCLC. The rationale behind the combination of lete-cel and pembrolizumab in arms B and C of the 208471 multiarm study was to overcome the hostile immunosuppressive microenvironment observed in solid tumors and to augment responses following lete-cel infusion. This rationale is supported by studies demonstrating the synergistic effects of combining NY-ESO-1–specific T cells with anti–PD-1 therapy (43). To date, these are the largest clinical studies of a TCR therapy in NSCLC. Overall, lete-cel was generally well tolerated and had a manageable safety profile consistent with that described in prior and ongoing studies across the lete-cel program (27). The combination of lete-cel administration and a PD-1 inhibitor did not seem to increase toxicity in NSCLC. Antitumor activity was observed for lete-cel in a subset of patients with only one durable response observed among 18 patients treated across studies. Pharmacokinetic data showed similar T-cell expansion between all patients and a comparable T-cell expansion profile to what has been described in prior lete-cel trials (28, 29).

In these studies, large-scale HLA typing and target expression testing of more than 2,500 were required to ultimately treat 18 patients. The extensive screening needed and the high attrition rate experienced during the screening process are important to consider, especially in the context of the rapid growth of HLA-restricted vaccines and T cell–based therapeutics in solid tumors. Stepwise HLA blood and NY-ESO-1/LAGE-1a tumor testing and apheresis can lead to long time periods between initial screening and treatment. Efforts to reduce screening barriers, such as incorporating HLA typing into standard next-generation sequencing platforms, may be critical to screening efficiency and may enable more equitable access to HLA-restricted therapies in the future. Differential HLA-A*02 allele prevalence by race or ethnicity could affect patient access to T cell–based therapies in certain populations (44, 45). Furthermore, timing apheresis can be challenging in patients heavily pretreated with therapies and can typically impair T-cell activity (46, 47). Planning and consideration of cellular therapy earlier in the course of treatment may be beneficial to both patient and T-cell fitness.

Trafficking of T cells and effective infiltration of the tumor are essential for the direct antitumor activity of adoptive T-cell therapy. Poor infiltration of transduced T cells into NY-ESO-1–expressing tissue could have contributed to the limited antitumor activity observed. Locoregional delivery of cell therapies, including direct intrapleural/tumoral injection, is an active area of investigation to overcome this obstacle (48, 49). There are several additional potential mechanisms of resistance to T-cell therapy, including the existence of inhibitory ligands in the tumor microenvironment and endogenous TCRs that can exhaust chimeric antigen receptor T cells (50). Other potential factors contributing to resistance are defects in antigen presentation and HLA loss with prior exposure to PD-1 blockade. Screening for HLA loss of heterozygosity refines neoantigen prediction and may have implications for our understanding of resistance mechanisms and immunotherapeutic approaches targeting neoantigens (51).

Unlike chimeric antigen receptor T cells, TCR-based approaches enable the recognition of targets beyond the cell surface to include targets located within any subcellular compartment. The use of TCRs also enables the recognition of low concentrations of intracellular cognate antigens and mimics the natural function of the T cell by recruiting the endogenous signaling molecules and adhering to correct spatial orientation between the T cell and its target (52–54). However, TCR-based adoptive T-cell therapies still rely on appropriate target expression and recognition, and ultimately, the NY-ESO-1/LAGE-1a epitope–HLA complex may not be an optimal therapeutic target in lung cancer. The affinity and stability of the engineered TCR–HLA interaction also affect the T-cell response, and the utilization of novel approaches, including machine-learning models and structure-based designs, may improve the selection of TCRs and potential outcomes in future studies (55, 56).

Unlike some soft tissue sarcomas, NY-ESO-1/LAGE-1a is not commonly expressed in lung cancer. In addition, NY-ESO-1/LAGE-1a expression in NSCLC is relatively heterogeneous across the patient population and within a given tumor sample (42). Target expression screening in the multiarm study suggests NY-ESO-1 prevalence in NSCLC to be approximately 12% (because LAGE-1a testing in the multiarm study was done in reflex, it may not accurately reflect its prevalence in NSCLC). It should be noted the NY-ESO-1 IHC CTA test and LAGE-1a RT-PCR CTA tests are also investigational and the set cut-offs for positive expression can include low and/or heterogeneous antigen levels. Low and/or heterogeneous antigen levels may help explain the often limited and transient clinical benefits. Owing to the small population in this trial, it is difficult to have enough data to modify the clinical testing cut-offs or to adjust assay performance.

However, consistent with the potential importance of NY-ESO-1/LAGE-1a antigen levels, the only patient in the single-arm study who had a PR to lete-cel treatment had very high (90%) NY-ESO-1 expression. Lete-cel may be more effective in tumors in which NY-ESO-1 is highly and uniformly expressed, such as in sarcomas. High levels of NY-ESO-1 expression with homogeneous distribution have been reported in myxoid/round cell liposarcomas and synovial sarcomas, which might be related to the promising results obtained in adoptive cellular immunotherapy trials (26, 57–59).

T-cell expansion and cell persistence were not meaningfully different in these NSCLC studies compared with previous studies in other indications in which efficacy has been observed (e.g., synovial sarcoma and myxoid/round cell liposarcoma; refs. 23–26). This suggests that expansion and persistence were not likely to explain the limited efficacy in these studies.

In addition to NY-ESO-1, other tumor neoantigens, such as mutant KRAS G12D and KRAS G12V, are being explored as targets for engineered TCR/adoptive T-cell therapy in solid tumors (60, 61). Other antitumor immunotherapy technologies under development include bispecific T-cell engager antibodies comprising two binding domains that link endogenous T cells to engage specific antigen-expressing tumors (62).

In summary, we demonstrated that lete-cel was generally safe in advanced NSCLC with or without pembrolizumab, but antitumor activity was limited and there were various design challenges associated with both studies that necessitated larger-scale screening efforts and limited study feasibility. Throughout these screening efforts, we gained valuable insights into the landscape of HLA typing and CTA expression in lung cancer, which constitute key learnings for designing future trials of HLA-based therapies. Furthermore, the enrollment challenges outlined above can provide critical insights for design and implementation of future HLA-restricted TCR and vaccine-based therapies, which are increasingly being investigated across solid tumors.

M. Altan reports research funding (to institution) from GSK, Genentech, Nektar Therapeutics, Merck, Novartis, Jounce Therapeutics, Bristol Myers Squibb, Eli Lilly and Company, Adaptimmune, Shattuck Labs, and Gilead; serving on the advisory board of GSK, Shattuck Labs, Bristol Myers Squibb, AstraZeneca, and Insightec; receiving speaker fees from AstraZeneca, Nektar Therapeutics, and Society for Immunotherapy of Cancer; and being a member of the safety review committee for Nanobiotix-MDA alliance and Hengenix. G. Lopes reports grants from GSK during the conduct of the study; stock and other ownership interests in Lucence Diagnostics, Xilis, Biomab, MorphoMetriX and CDR-Life; honoraria from Boehringer Ingelheim, Blueprint Medicines, AstraZeneca, Merck, and Janssen; consulting for Pfizer and AstraZeneca; research funding from AstraZeneca, Lucence, Xilis, E.R. Squibb & Sons, LLC, Merck Sharp & Dohme, EMD Serono, Blueprint Medicines, TESARO, Bavarian Nordic, Novartis, G1 Therapeutics, Adaptimmune, Bristol Myers Squibb, GSK, AbbVie, Rgenix, Pfizer, Roche, Genentech, Eli Lilly and Company, and Janssen; and fees for travel, accommodations, and expenses from Boehringer Ingelheim, Pfizer, E.R. Squibb Sons, & LLC, Janssen, Seattle Genetics, Celgene, Ipsen, Pharmacyclics, Merck, and AstraZeneca. T.J.N. Hiltermann reports grants from Roche, AstraZeneca, and Bristol Myers Squibb outside the submitted work. L.C. Villaruz reports personal fees from AstraZeneca, Johnson & Johnson, EMD Serono, Sanofi, Gilead Sciences, and Daiichi Sankyo outside the submitted work. E. Calvo reports employment with START and HM Hospitales Group; leadership roles in START, PharmaMar, EORTC, Sanofi, BeiGene, Novartis, and Merus N.V.; being a stockholder in START and Oncoart Associated; receiving honoraria from HM Hospitales Group; acting as a consultant/advisor for Nanobiotix, Janssen-Cilag, Roche/Genentech, TargImmune Therapeutics, Servier, Bristol Myers Squibb, Amunix, Adcendo, Anaveon, AstraZeneca/MedImmune, Chugai Pharmaceutical, MonTa, MSD Oncology, Nouscom, Novartis, OncoDNA, T-knife, Elevation Oncology, PharmaMar, Ellipses Pharma, Syneos Health, Genmab, and Diaccurate; and research funding from START. M.J. Edelman reports grants from Stand Up To Cancer, Department of Defense, and NCI; consulting for BioAlta, GE Healthcare, InterVenn, Novocure, touchONCOLOGY, Sanofi, Omega Therapeutics, and Regeneron; meetings/travel for International Association for the Study of Lung Cancer; advisory board membership for AstraZeneca, AnHeart Therapeutics, Takeda, Duke/Hoosier Oncology Group, and GSK; leadership roles, including co-chair, lung committee, NRG, and deputy editor, lung cancer chair, scientific and advisory board; and being a stockholder in Creatv BIO and BioMarker Strategies. M. Furqan reports institutional research funding from AbbVie, Amgen, Aprea, AstraZeneca, Astellas, BeiGene, Bristol Myers Squibb, Checkmate, Elicio, Genmab, Genentech, Gilead, GSK, Immunocore, Incyte, Jacobio, Eli Lilly and Company, Merck, Mirati, Novartis, and Pfizer; a consultant role with Omega Therapeutics and Novartis; and advisory board membership for AbbVie, AstraZeneca, Jazz Pharmaceuticals, BeiGene, and Mirati. J. Neal reports honoraria from CME Matters, Clinical Care Options, Research to Practice, Medscape, Biomedical Learning Institute, MLI PeerView, prIME Oncology, Projects In Knowledge, Rockpointe, MJH Life Sciences, The Medical Educator Consortium, and HMP Education; a consulting role with AstraZeneca, Genentech/Roche, Exelixis, Takeda, Eli Lilly and Company, Amgen, Iovance, Blueprint, Regeneron, Natera, Sanofi/Regeneron, D2G Oncology, Surface Oncology, Turning Point Therapeutics, Mirati Therapeutics, Gilead Sciences, AbbVie, Summit Therapeutics, Novartis, Novocure, Janssen Oncology, and AnHeart Therapeutics; research funding from Genentech/Roche, Merck, Novartis, Boehringer Ingelheim, Exelixis, Nektar Therapeutics, Takeda, Adaptimmune, GSK, Janssen, AbbVie, and Novocure; and royalties from UpToDate. E. Felip reports honoraria from AbbVie, Amgen, AstraZeneca, Bayer, BeiGene, Boehringer Ingelheim, Bristol Myers Squibb, Daiichi Sankyo, Eli Lilly, F. Hoffmann-La Roche, Genentech, Gilead Sciences, Janssen, Medical Trends, Medscape, Merck Serono, Merck Sharp & Dohme, Novartis, PeerVoice, Pfizer, Regeneron, Sanofi, Takeda, and touchONCOLOGY. J.W. Carlisle reports consulting for Amgen, Novocure, and Catalyst, as well as research funding (to institution) from Amgen, AstraZeneca, Daiichi Sankyo, HUTCHMED, and Parexel. J.V. Heymach reports consulting for GSK, Merck, AstraZeneca, Genentech, Regeneron, and EMD Serono, as well as advisory board membership for GSK, Merck, AstraZeneca, Genentech, Regeneron, and EMD Serono. R.E. O'Cearbhail reports being supported in part by the NCI/NIH P30 CA008748 institutional grant; personal fees from Bayer, Curio, ImmunoGen, MJH/OncLive, Regeneron, Seattle Genetics/Seagen, TESARO/GSK, Miltenyi Biotec, 2seventy bio, and Loxo; fees for travel from Gathering Around Cancer, Ireland, outside the submitted work; a board advisory role for GSK; committee membership for PRIMA, Moonstone (TESARO/GSK), DUO-O (AstraZeneca), and ARTISTRY-7 (Mural Oncology) studies; an advisor for Carina Biotech; and funding from AbbVie/Stemcentrx, Atara Biotherapeutics, Bayer/Celgene/Juno, Context Therapeutics, Genentech, Genmab/Seagen Therapeutics, Gynecologic Oncology Foundation, Kite Pharma, Ludwig Cancer Institute, Marker Therapeutics, Merck, Regeneron, SELLAS Therapeutics, Syndax Pharmaceuticals, TCR2 Therapeutics, Lyell Therapeutics, Arsenal Bio, and TESARO/GSK. M. Zauderer reports being supported in part by the NCI/NIH P30 CA008748 institutional grant; grants from Vivace, MedImmune, Precog, GSK, Epizyme, Polaris, SELLAS Life Sciences, Bristol Myers Squibb, Millenium, Curis, and Atara; consulting for Curis and Ikena; honoraria from PER and Medscape; and a leadership role in the Mesothelioma Applied Research Foundation, as chair, board of the directors. M. Chisamore reports other support from Merck & Co. Inc. outside the submitted work, as well as reports employment with and ownership of stock in Merck & Co. Inc. E. Corigliano reports employment with GSK and holding financial equities in GSK. I. Eleftheriadou reports employment with GSK and holding financial equities in GSK. S. Zajic reports employment with GSK and holding financial equities in GSK. B. Jenkins reports employment with GSK and holding financial equities in GSK. S. Goodison reports employment with GSK and holding financial equities in GSK. S. Suchindran reports employment with GSK and holding financial equities in GSK. N. Ramos-Hernandez reports employment with GSK and holding financial equities in GSK. N. Tarek reports employment with GSK and holding financial equities in GSK. A.J. Schoenfeld reports personal fees from Johnson & Johnson, KSQ Therapeutics, Bristol Myers Squibb, Merck, AstraZeneca, Enara Bio, Perceptive Advisors, Oppenheimer and Co., Umoja Biopharma, Legend Biotech, Iovance Biotherapeutics, Obsidian Therapeutics, Prelude Therapeutics, Immunocore, Lyell Immunopharma, Amgen, and Heat Biologics, as well as grants (to institution) from GSK, PACT Pharma, Iovance Biotherapeutics, Achilles Therapeutics, Merck, Bristol Myers Squibb, Harpoon Therapeutics, Affini-T Therapeutics, Legend Therapeutics, and Amgen outside the submitted work. No disclosures were reported by the other authors.

M. Altan: Conceptualization, data curation, investigation, methodology, writing–original draft, writing–review and editing. G. Lopes: Data curation, investigation, writing–review and editing. T.J.N. Hiltermann: Data curation, investigation, writing–review and editing. R. Govindan: Data curation, investigation, writing–review and editing. L.C. Villaruz: Data curation, investigation, writing–review and editing. E. Calvo: Data curation, investigation, writing–review and editing. M.J. Edelman: Data curation, investigation, writing–review and editing. M. Furqan: Data curation, investigation, writing–review and editing. J. Neal: Data curation, investigation, writing–review and editing. E. Felip: Data curation, investigation, writing–review and editing. J.W. Carlisle: Data curation, investigation, writing–review and editing. J.V. Heymach: Data curation, investigation, writing–review and editing. R.E. O’Cearbhaill: Formal analysis, supervision, investigation, writing–review and editing. M. Zauderer: Data curation, investigation, methodology, writing–review and editing. M. Chisamore: Investigation, writing–review and editing. E. Corigliano: Conceptualization, resources, data curation, formal analysis, supervision, investigation, methodology, project administration, writing–review and editing. I. Eleftheriadou: Conceptualization, resources, data curation, formal analysis, supervision, investigation, methodology, project administration, writing–review and editing. S. Zajic: Data curation, software, supervision, validation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. B. Jenkins: Data curation, software, formal analysis, supervision, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. S. Goodison: Data curation, formal analysis, supervision, investigation, methodology, writing–original draft, project administration, writing–review and editing. S. Suchindran: Data curation, software, formal analysis, validation, investigation, visualization, writing–original draft, writing–review and editing. N. Ramos-Hernandez: Conceptualization, data curation, formal analysis, supervision, validation, investigation, visualization, methodology, writing–original draft, writing–review and editing. N. Tarek: Conceptualization, data curation, formal analysis, supervision, validation, investigation, methodology, writing–original draft, writing–review and editing. A.J. Schoenfeld: Conceptualization, resources, data curation, supervision, validation, investigation, methodology, writing–review and editing.

The authors would like to thank the patients, their families, and all investigators involved in this study. Special thanks to others involved in the design and conduct of the study at GSK, including Athanasia (Nancy) Skoura, Benedetto Farsaci, Jimson D’Souza, Thomas Faitg, Yi Hsing, Bryan Barnes, Colleen Valenzuela, Ale Bozac, Sheila Walker, Allison Hainline. Medical writing support was provided by Jessica Men, PharmD, and editorial support was provided by Travis Taylor, BA, both of Scion (a division of Prime), supported by GSK according to Good Publication Practice guidelines (Link). These studies were sponsored by GSK (study IDs: 208749/208471).