Neoantigen-Reactive TILs With CISH Knockout Are Safe and Effective in Metastatic GI Tumors

Gastric Cancer | Image Credit: ©

Sebastian Kaulitzki – stock.adobe.com

Tumor-infiltrating lymphocyte (TIL) therapy with CRISPR-Cas9–mediated CISH gene knockout was well tolerated and demonstrated early efficacy signals in patients with metastatic gastrointestinal (GI) epithelial cancer, according to findings from a first-in-human, single-center phase 1 trial (NCT04426669), which were published in Lancet Oncology.1

At a median follow-up of 129 days (IQR, 15-283), all 12 evaluable patients experienced severe treatment-related adverse effects (AEs), none of which were attributed to the TIL therapy used in the study. The most common grade 3/4 AEs included hematologic toxicities (100%), which were attributed to the pre-infusion lymphodepleting chemotherapy regimen and included decreased white blood cell counts (100%), decreased platelet counts (100%), decreased lymphocyte counts (100%), anemia (100%), febrile neutropenia (83%), increased alkaline phosphate levels (25%), and increased bilirubin levels (25%); AEs following TIL infusion and interleukin-2 (IL-2) administration (fever, chills, hypoalbuminemia, fatigue, vascular disorders, and respiratory disorders); fatigue (33%), and anorexia (25%). On-study deaths of any cause were attributed to the underlying disease (metastatic GI cancer) and related complications (n = 10), and infection (grade 5 septicemia; n = 1). No grade 3 or higher cytokine release or neurotoxicity events were reported. Notably, the maximum tolerated dose (MTD) was not reached, as all doses of the TIL therapy were tolerated by all patients.

Regarding efficacy, 50% of patients achieved stable disease (SD) by day 28, and 33% of patients had ongoing SD at day 56.

“We demonstrate proof of principle that CISH function can safely be disrupted to potentially improve the efficacy of TILs, a recently approved cell therapy for solid cancers,” lead study author Emil Lou, MD, PhD, FACP, and coauthors wrote in the paper.

Lou is a professor of medicine in the Division of Hematology, Oncology, and Transplantation in the Department of Medicine at the University of Minnesota Medical School in Minneapolis.

Impetus for Continued TIL Investigationin Oncology

In February 2024, the FDA granted accelerated approval to lifileucel (Amtagvi) for the treatment of adult patients with unresectable or metastatic melanoma who have previously received treatment with a PD-1–directed antibody, and if BRAF V600 positive, a BRAF inhibitor with or without a MEK inhibitor.2 This regulatory decision represented the first FDA approval of a TIL therapy for the treatment of patients with cancer.

Moreover, preclinical studies have shown that CISH disruption enhances antitumor responses against several cancer types, and that these responses occur independently of the presence of canonical cell surface immune checkpoints like PD-1.1

“Given that CISH inhibition is not yet tractable through small molecule–directed therapeutic approaches, we designed and executed a, to our knowledge, first-in-human clinical trial, in which we used CRISPR-Cas9 gene editing to knock out CISH in TILs derived from patients with drug-refractory forms of metastatic gastrointestinal cancer,” the authors wrote.

Phase 1 Trial Design

This phase 1 study included patients 18 to 70 years of age with metastatic GI epithelial cancer who had progressive disease following treatment with 1 or more standard first-line therapies, received their last dose of prior systemic therapy at least 4 weeks before receiving on-study lymphodepleting chemotherapy, acute toxicities resolved to grade 1 or less, measurable disease with 1 or more lesion that was resectable for TIL generation, at least 1 other measurable lesion per RECIST 1.1 criteria in which to study disease response, and an ECOG performance status of 0 or 1. Patients were excluded if they had comorbid conditions including coagulation disorders, any form of primary immunodeficiency or concurrent opportunistic infection, active systemic infection requiring anti-infective treatment; any other active major medical illnesses; history of severe immediate hypersensitivity reaction to fludarabine, aldesleukin, cyclophosphamide, or dimethylsulfoxide; history of coronary revascularization or ischemic symptoms; inadequate pulmonary or cardiac function; or prior treatment with any cell therapy product.

TILs obtained from tumor biopsies were expanded based on neoantigen reactivity and underwent CRISPR-Cas9–mediated CISH knockout. The edited TILs were expanded and cryopreserved before being infused intravenously into patients following receipt of non-myeloablative lymphocyte-depleting chemotherapy (consisting of cyclophosphamide at 60 mg/kg per dose on days –6 and –5, as well as fludarabine at 25 mg/m2 per dose on days –7 to –3) and a maximum of 6 doses of high-dose IL-2 (consisting of aldesleukin at 720,000 IU/kg per dose).

This study had a phase 1 dose-escalation design with an initial accelerated dose-escalation stage in which 1 patient was enrolled per dose level until a dose-limiting toxicity (DLT) was encountered or the highest dose level was received without reaching a DLT. The doses for total cells infused ranged from less than 3.0 x 109 cells to 5.1 x 1010 to 1.0 x 1011 cells.

The primary end point was the safety and MTD of the investigational therapy. Key secondary end points included progression-free survival (PFS), overall survival (OS), and toxicity profiles.

Baseline Characteristics and TIL Manufacturing Details

Between May 12, 2020, and September 16, 2022, the trial enrolled 22 patients; notably, 1 patient was enrolled twice due to a lack of TIL outgrowth on the first expansion attempt. CISH-knockout TIL products were successfully manufactured for 86% of patients, and 63% of patients received autologous CISH knockout TIL infusion.

Patients had received a median of 5 prior lines of therapy (IQR, 3-8). In total, 11 patients were male, and 10 patients were female. One patient was Asian, and the rest were White.

The time from initial diagnosis to trial enrollment ranged from 7 to 106 months for all treated patients; this time ranged from 7 to 50 months when excluding the patient who had microsatellite instability–high (MSI-H) disease and had received prior immunotherapy-based treatment before disease progression. Patients had 1 to 4 metastatic sites, and 67% of the 12 treated patients had liver tumor involvement. Notably, liver metastases were a preferred biopsy site for TIL derivation due to their relative ease of access. Among the 12 treated patients, most (83%) had a baseline ECOG performance status of 0.

The investigators developed a clinical-scale, cGMP-compliant manufacturing process for CRISPR-Cas9–mediated CISH knockout in TILs. The authors noted that this process resulted in efficient CISH knockout with concurrent high viability and expansion, explaining that the results of this study therefore “have implications beyond the scope of this specific trial by demonstrating the feasibility of efficient gene editing of polyclonal TIL products.”

Four patients received the TIL therapy at the 2 highest dose levels, including 1 patient who was dosed at 1.0 x 1011 cells and did not experience stopping rules.

Additional Efficacy and Safety Findings

The Luminex assay revealed that most cytokines were expressed at higher levels by CISH-knockout TILs vs control TILs. Correlation analysis showed that higher cytokine expression levels generally correlated with a lower number of TILs produced, although higher expression was also positively associated with PFS.

The authors explained that persistence is a common challenge with T-cell therapies, including TILs. To evaluate the persistence of CISH-knockout TILs in patients following infusion, the investigators collected peripheral blood mononuclear cells (PBMCs) at various time points and conducted targeted CISH PCR amplicon sequencing. This analysis showed that peripheral CISH-knockout TIL proliferation peaked at day 7 and declined by day 21. In 2 patients, a small increase in CISH-knockout TIL abundance was noted at day 21 in CD8-positive, PD-1–negative PBMC T cells. Moreover, although the relative frequencies of CD4- and CD8-positive T cells and their PD-1–positive or –negative phenotype varied, a trend toward increased frequency of CD8-positive, PD-1–positive T cells was observed.

The median PFS was 57 days (95% CI, 25-156), and the median OS was 129 days (95% CI, 46-290). At the time of the PFS analysis, 11 of 12 evaluable patients had progressive disease (PD), and 1 patient had clinical complete response (CR). By the time of the OS analysis, all patients had died except for the patient who had achieved clinical CR.

Notable Patient Cases

One of the patients with ongoing SD at day 56 experienced an initial 17% tumor volume reduction at day 28 and had eventual progression per RECIST 1.1 criteria by day 84, measured by a 23% increase in tumor volume compared with baseline. Notably, that patient’s response occurred following tumor progression that occurred after 2.17 months of treatment with fluorouracil.

Another patient, who had microsatellite-stable, tumor mutational burden–low, chemotherapy-refractory metastatic colorectal cancer (CRC) with liver, lung, and ovary metastases had PD before trial enrollment; this patient had received oral capecitabine and bevacizumab for 6.7 months prior to disease progression, and following TIL infusion, she had SD for more than 8 months and lived an additional 2 years and 11 months. In total, 4 patients received any subsequent form of anticancer therapy.

The investigators emphasized that the most notable response was observed in a young adult patient with MSI-H CRC that was refractory to PD-1– and CTLA-4–directed therapies. This patient achieved a CR with TIL therapy that was ongoing for over 21 months. An analysis of this patients PBMCs showed increased levels of circulating CD8-positive, PD-1–negative T cells, cells around 21 days, as well as a low but continuous presence of CD4-positive, PD-1–positive T cells.

This patient’s frequency of CISH-knockout indels peaked at approximately 4% in the CD4-positive, PD-1–negative fractions and at approximately 1% in the CD8-positive, PD-1–negative fractions at day 10, which represented the initial wave of TIL efficacy after infusion. Secondary increases in the CD4-positive, PD-1–positive and CD4-positive, PD-1–negative fractions were reported at days 14 and 21, respectively, which may have been due to a tumor-specific recognition and expansion event. Furthermore, in an analysis of the T-cell receptor (TCR) clone expression changes at different time points, the investigators saw 3 significant disruptions in the patient’s immune repertoire at 4, 10, and 56 days post-infusion. Although the immune repertoire remained relatively stable between these disruptions, each disruption was followed by significant increases in the expression levels of different clone sets, suggesting that the active immune repertoire was completely altered following disruption. Multiplexed ion beam imaging also showed that this patient’s tumor harbored regions of low tumor cellularity with robust immune infiltration.

“The finding that CISH-knockout, neoantigen-selected TILs can be effective in a patient with MSI-H CRC that was refractory to dual inhibition of PD-L1 and CTLA-4 provides evidence of a therapeutic strategy for circumventing resistance to currently available immune checkpoint inhibitors,” the authors wrote. “More broadly, these observations provide a strong motivation to disable CISH activity using other therapeutic approaches, including small molecule inhibitors and targeted protein degraders.”

Pharmacokinetics Findings and Next Steps

A post hoc pharmacokinetics investigation of infused TIL clonotypes with confirmed neoantigen reactivity that had undergone paired TCR single-cell RNA sequencing after co-culture with neoantigen peptides showed that TIL clusters had a unique transcriptional profile of activation in response to neoantigen response vs pan-CD3 and non-peptide stimulation. Additionally, within the subset of neoantigen-responsive TILs, approximately 21 unique TCR clonotypes were identified; a subset of these clonotypes persisted and expanded after 10 days in 1 patient after infusion. This effect was temporally consistent with the observed increases in CISH-edited allele frequencies, as detected by next-generation sequencing. Furthermore, the persisting clonotypes in this patient’s blood stemmed from a unique cluster of neoantigen-specific activation defined by high expression of several genes. Moreover, 4 of the clonotypes that showed prolonged persistence for longer than 1 year following infusion also had significantly reduced or undetectable CISH expression relative to the total infused population, indicating that these clonotypes consisted of TILs with CRISPR-Cas9–induced CISH knockout.

“Although we used genome editing as a means to a greater end to disrupt CISH, our findings suggest that alternative next-generation small molecule drugging strategies that inhibit this intracellular immune checkpoint might also have utility, and the virtue of more democratized access, in patients with malignancies refractory to current forms of immunotherapy, and this work is actively underway,” the authors concluded.

References

1. Lou E, Choudhry MS, Starr TK, et al. Targeting the intracellular immune checkpoint CISH with CRISPR-Cas9-edited T cells in patients with metastatic colorectal cancer: a first-in-human, single-centre, phase 1 trial. Lancet Oncol. 2025 May;26(5):559-570. doi:10.1016/S1470-2045(25)00083-X
2. FDA approves first cellular therapy to treat patients with unresectable or metastatic melanoma. FDA. February 16, 2024. Accessed May 28, 2025. https://www.fda.gov/news-events/press-announcements/fda-approves-first-cellular-therapy-treat-patients-unresectable-or-metastatic-melanoma