We are pleased to present highlights of the latest advances in immunotherapy emerging from the 2026 AACR Annual Meeting, April 17–22, 2026.
(Thanks to Jennifer Anderson, MD, PhD and Terri Holzen, PhD for compiling these scientific highlights)
Phase 1 study of GV20-0251, a human antibody against the novel immune checkpoint IGSF8, for solid tumors
CT001. Evaluation of pharmacodynamic and potential predictive biomarkers for GV20-0251, an anti-IGSF8 antibody, as monotherapy from ongoing Phase 1/2a study
Xingfeng Bao (GV20 Therapeutics LLC, Newton, MA, USA) presented an ongoing phase 1/2a study of GV20-0251, a human antibody targeting the novel checkpoint immunoglobulin superfamily member 8 (IGSF8). IGSF8 is an immune checkpoint expressed on some solid tumors. Preclinical studies indicate that inhibition of IGSF8 by GVS20-0251 or loss of Igsf8 suppresses tumor growth in mice by enhancing NK cell cytotoxicity and dendritic cell antigen presentation. 42 patients with solid tumors enrolled in the phase I dose escalation phase of the study received GVS20-0251 monotherapy at doses ranging from 0.5 mg/kg to 20.0 mg/kg. 98% of patients had received prior systemic therapies, and 71% of patients had received prior PD-(L)1 therapy, including ICI monotherapy or dual ICI therapy. Of the 37 evaluable patients, 16 patients had cutaneous melanoma. Three evaluable patients experienced a confirmed partial response. All three responders had cutaneous melanoma, two of the three had liver metastases, and one responder has remained on GVS20-0251 for almost 2 years. Nine of the 16 evaluable patients with melanoma had primary resistance to anti PD-1 therapy, and of those 9 patients, 3 experienced a partial response (overall response rate 33.3%), and 3 experienced tumor shrinkage (disease control rate 77.8%). GVS20-0251 exhibited a favorable safety profile, with no observed dose-limiting toxicities and only one episode of grade 3pneumonitis that was observed in the context of disease progression. Paired tumor biopsies indicated that most tumors treated with GV20-0251 exhibited increased NK and CD8+ T cell infiltration after treatment as well as increased antigen presentation and NK cell activity. Tumors from all 3 responders exhibited high levels of IGSF8 expression and low levels of PD-L1 at baseline, suggesting that response was independent of PD-L1 expression. GV20-025 exhibited promising clinical activity, a favorable safety profile, and evidence of immunostimulatory effects in treated tumors. Given that several tumor types that are refractory to immunotherapy exhibit high levels of IGSF8 expression, GV20-0251 could potentially treat a variety of high-unmet-need tumor types. Phase 2a trials of GV20-0251 as monotherapy or in combination with PD-1 inhibitors for solid tumors are currently enrolling patients.
PLN-101095, an alpha v beta 8 / alpha v beta 1 integrin inhibitor, in combination with pembrolizumab for patients with advanced solid tumors with secondary resistance to immune checkpoint inhibitors
CT002. First-in-human phase I study of PLN-101095, a first-in-class dual alpha v beta 8 / alpha v beta 1 integrin inhibitor, as monotherapy and in combination with pembrolizumab in patients with advanced solid tumors refractory to immune checkpoint inhibitors (ICI)
Timothy A. Yap (University of Texas MD Anderson Cancer Center, Houston, TX, USA) presented results from a phase 1 open-label, dose-escalation study of PLN-101095 in patients with solid tumors and prior resistance to immune checkpoint inhibitors. PLN-101095 is an orally bioavailable inhibitor of alpha v beta 8 (avb8) and alpha v beta 1 (avb1) integrins. PLN-101095 inhibits avb8/avb1 to block integrin-driven activation of tumor growth factor beta (TGFb) and selectively enhancing T cell interferon gamma (IFN-g) function and reducing fibroblast activation and fibrotic tumor stroma. This study addressed the hypothesis that PLN-101095-mediated inhibition of avb8/avb1 blocks activation of TGFb, reducing immunosuppression and promoting a new or reinvigorated anti-tumor immune response. Prior preclinical studies indicated that PLN-101095 inhibits TGFb, shifting solid tumors to a high-IFN-g signature and resensitizing them to immune checkpoint inhibitors. Following the dose-escalation phase, all patients received PLN-101095 monotherapy for two weeks followed by PLN-101095 combined with pembrolizumab. Tumor response was assessed every 8 weeks, beginning on Week 10 of treatment. 16 patients with solid tumors with primary or secondary resistance to immune checkpoint inhibitors, as defined by SITC, were treated. Tumor types in the study included non-small cell lung cancer (NSCLC; n=3), cholangiocarcinoma (n=3), head and neck squamous cell carcinoma (n=2), and renal cell carcinoma (RCC; n=2). Patients received a median of 3 lines of prior therapy, and 12 patients experienced secondary resistance to prior immune checkpoint inhibitors. PLN-101095 was well-tolerated and safe. Most treatment-related adverse events (TRAEs) were Grade 1 or 2. The most common TRAE was rash, and all rashes were Grade 1 or 2. One rash was reported during the monotherapy period, and most were primarily observed within two days of starting combination treatment. Four responses were observed in the overall study population (1 complete response, 2 partial responses (PR), and 1 unconfirmed PR), with an objective response rate (ORR) of 19% and a disease control rate (DCR) or 56%. All 4 responders had secondary resistance to prior immune checkpoint inhibitors in the subgroup of patients with secondary resistance to prior immune checkpoint inhibitors. In this patient subgroup, ORR was 30% and DCR was 60%. The 3 confirmed responses were clinically meaningful and durable, and occurred in patients receiving doses of 1000 mg PLN-101095 twice daily. The median time on treatment for these 3 responders is 19 months, and the average maximum tumor reduction was 89%. Response to PLN-101095 was associated with increased levels of IFN-g and PD-L1 in plasma after 14 days of monotherapy, suggesting increased IFN-g as a potential biomarker of TGF-b inhibition. PLN-101095 was well-tolerated as monotherapy and in combination with pembrolizumab, and combination treatment was associated with promising anti-tumor activity, especially in patients with secondary resistance to immune checkpoint inhibitors. Dose-expansion cohorts of PLN-101095 with pembrolizumab for NSCLC, clear cell RCC, and tumor mutation burden-high cancers have begun, and these studies will further address the potential of IFN-g as a predictive biomarker
AMG 355, a novel CCR8-targeting antibody, modulates the tumor immune microenvironment but exhibits limited clinical activity
CT003. Initial results of a phase 1 study of AMG 355, a novel anti-CCR8 antibody as a single agent and in combination with pembrolizumab in patients with solid tumors
Marwan Fakih (City of Hope National Medical Center, Duarte, CA, USA) reported initial results of phase 1 study of AMG 355, a novel antibody that targets chemokine receptor 8 (CCR8), which is selectively enriched on intratumoral regulatory T cells (Treg) High levels of intratumoral Tregs are associated with resistance to immune checkpoint inhibitors and poor prognosis. AMG 355 selectively depletes CCR8+ intratumoral Tregs by activating NK cell-mediated ADCC induced killing, resulting in improved safety compared to systemic Treg modulators. Patients with solid tumors received AMG 355 as monotherapy (n=47) or in combination with pembrolizumab (n=30). Patients received monotherapy at doses ranging from 0.28 mg to 700 mg every three weeks. Patients receiving combination therapy received AMG 355 at doses of 70 mg to 700 mg every three weeks. Patients with microsatellite stable colorectal cancer, gastric cancer, non-small cell lung cancer, or melanoma were enrolled in the study. Patients had received a median of 3 prior lines of therapy, and approximately 40% of patients had received prior PD-(L)1 therapy. No dose-limiting toxicities were observed. 4.3% of patients in the monotherapy arm and 16.7% of patients in the combination arm experienced Grade 3 or higher treatment-related adverse events. Serum levels of AMB 355 increased in a dose-proportional manner. One patient responded to AMG 355 monotherapy, and 1 patient responded to AMG 355 in combination with pembrolizumab, with objective response rates of 2.1% and 3.3%, respectively. Patients at all dose levels, regardless of response, exhibited reduction of CCR8 on treatment and reduction of CCR8/CD8 ratio in paired tumor biopsies, suggesting modulation of the tumor immune microenvironment consistent with the mechanism of action of AMG 355. Although AMG 355 was safe and well-tolerated, its clinical activity as monotherapy or in combination with pembrolizumab was limited. Additional tissue-based multi-omic studies are planned to further characterize changes in the tumor microenvironment, including immune cell profiles.
Final results of the NIBIT-M2 trial: Nivolumab and ipilimumab in patients with melanoma and asymptomatic brain metastases
CT008. Ten-year survival and cell-free DNA methylation profiling of melanoma patients with asymptomatic brain metastases treated with nivolumab plus ipilimumab: The multicenter phase III NIBIT-M2 trial
Anna Maria Di Giacomo (University Hospital of Siena, Siena, Italy) presented the final results of the phase 3 Italian Network for Tumor Biotherapy (NIBIT)-M2 Trial, which investigated the long-term efficacy of fotemustine, ipilimumab in combination with fotemustine, and ipilimumab with nivolumab in patients with melanoma and asymptomatic brain metastases. Although this study previously showed nivolumab and ipilimumab to be effective in treating patients with melanoma and asymptomatic brain metastases, with 7-year overall survival (OS) of 41%, no predictive biomarkers of response have been identified. Patients were randomly assigned to Arm A (fotemustine; n=23), Arm B (ipilimumab + fotemustine; n=26), or Arm C (ipilimumab + nivolumab; n=27). At a median follow-up of 125 months, 10-year OS was 13.0% for Arm A, 7.7% for Arm B, and 32.1% for Arm C. Ten-year melanoma-specific survival were similar, with 13.0% for Arm A, 7.7% for Arm B, and 35.7% for Arm C. Global progression-free survival (PFS) was 4.3% for Arm A, 7.7% for Arm B, and 19.8% for Arm C. Intracranial PFS was 4.3% for Arm A, 7.7% for Arm B, and 28.8% for Arm C, further indicating that the combination of ipilimumab and nivolumab reduces the risk of progression of brain metastases. Exploratory analyses of cell-free DNA (cfDNA) were conducted with patient plasma samples at baseline (n=57) and paired patient plasma samples at week 12 (n=29). Tumor fraction (TF) was estimated from low pass whole genome sequencing, and tumor-specific methylation score (T-meth Score) was calculated as the ratio of observed methylated regions analyzed by cf-methylated DNA immunoprecipitation and high-throughput sequencing to melanoma-specific methylated regions previously identified in the TCGA melanoma cohort. OS was significantly improved in patients with TF and T-meth Scores below median baseline values compared to patients with TF and T-meth Scores above the median (p=0.033 and p=0.002, respectively). Patients with TF and T-meth Scores below the median were enriched in Arm C, compared to Arms A and B, and had the longest OS. . Results from this study further reinforce long-term therapeutic efficacy of ipilimumab and nivolumab in patients with melanoma and asymptomatic brain metastases and suggest that plasma TF and T-meth Score may act as predictive biomarkers of long-term survival.
Predictive value of ctDNA status after neoadjuvant therapy for triple-negative breast cancer
CT013. Whole-exome sequencing tumor-informed circulating tumor DNA detection after completion of neoadjuvant treatment predicts non-pCR and distant recurrence in patients with early triple-negative breast cancer (TNBC) — Results from a sub-study of the NSABP B-59/GBG-96-GeparDouze Trial
Marija Balic (University of Pittsburgh School of Medicine, Pittsburgh, PA, USA) addressed the association of circulating tumor DNA (ctDNA) status after neoadjuvant therapy with pathological complete response (pCR) and distant recurrence in patients receiving perioperative therapy for triple negative breast cancer (TNBC). This analysis was performed on samples from 160 patients in the NSABP B-59/GBG-96-GeparDouze Trial, which compared neoadjuvant chemotherapy with atezolizumab followed by adjuvant atezolizumab vs. neoadjuvant chemotherapy with placebo followed by adjuvant placebo for TNBC. A previous analysis from this trial indicated that patients who were ctDNA-positive after surgery were at a 30-fold higher risk of distant recurrence. The Oncodetect tumor-informed ctDNA assay was used for this study. 155 (96%) of 160 patients were ctDNA-positive. 9% of the 155 patients positive baseline ctDNA were ctDNA-positive after neoadjuvant treatment (post-NAT), and 6% of patients were ctDNA positive after surgery (post-surgery). Among patients who were ctDNA positive at baseline who also had a post-NAT blood draw (n=136), 90.4% exhibited ctDNA clearance post-NAT. Among patients who were post-NAT ctDNA negative, 95 achieved a pCR, and 46 had a non-pCR. Among patients who were post-NAT ctDNA-positive, 2 achieved a pCR, and 23 achieved a non-pCR. Post-NAT ctDNA positivity showed 97.9% specificity for non-pCR, and the positive predictive value for non-pCR was 85.7%. 85% of patients who were persistently ctDNA-positive post-NAT did not achieve a pCR, and 69% of patients who exhibited ctDNA clearance post-NAT achieved a pCR. ctDNA levels correlated with tumor burden and lymph node stages. Post-NAT ctDNA positivity was significantly associated with distant recurrence, with a hazard ratio (HR) 10.2 (p<0.0001). Although some sample sizes were small, patients who were ctDNA-negative post-surgery exhibited a lower risk of distant recurrence compared to those who were ctDNA-positive, regardless of pCR status. These analyses indicate that post-NAT ctDNA positivity is associated with a statistically significant 10-fold increased risk of distant recurrence, reinforcing previous observations with post-surgery ctDNA positivity and distant recurrences. Post-NAT ctDNA positivity is also a strong predictor of non-pCR. Taken together, these results suggest post-NAT ctDNA status has high potential as an early surrogate for response or resistance to neoadjuvant chemoimmunotherapy for TNBC and may one day be used for risk stratification and informing future treatment decisions.
Tumor alteration status and ctDNA clearance as early markers of clinical efficacy of perioperative nivolumab for non-small cell lung cancer
CT015. Clinical outcomes by genomic markers and ctDNA dynamics with perioperative nivolumab (NIVO) for resectable NSCLC from CheckMate 77T
Tina Cascone (University of Texas MD Anderson Cancer Center, Houston, TX, USA) presented exploratory biomarker analyses of tumor genomic alteration status and circulating tumor DNA (ctDNA) dynamics from CheckMate 77T, which demonstrated statistically significant and clinically meaningful event-free survival (EFS) benefits associated with perioperative nivolumab (NIVO) compared to placebo (PBO) for patients with stage III resectable non-small cell lung cancer (NSCLC). Perioperative NIVO for resectable NSCLC has been approved by the United States Food and Drug Administration, but there is still an unmet need to identify early markers of clinical efficacy, especially for patients with tumors with driver mutations, patients who are ctDNA-positive after neoadjuvant therapy, and patients with minimum residual disease after surgery. Samples from 190 patients (41%) were biomarker-evaluable. EFS analysis was performed by tumor mutation status for KRAS, KEAP1, and STK11 as single or co-mutations. EFS was similar in the NIVO and PBO arms for patients with KRAS mutations (HR 0.94), while EFS was improved with NIVO compared to PBO in patients with wild-type KRAS (HR 0.60). NIVO improved EFS compared to PBO in patients with KEAP1 mutations (HR 0.60), wild-type KEAP1 (HR 0.64), STK11 mutations (HR 0.63), and wild-type STK11 (HR 0.65). Conclusions could not be drawn from this analysis, as these patient subgroups were small. EFS analyses were also performed in patient subgroups with one or more tumor genomic alterations in KEAP1, STK11, CDKN2A and/or SMARCA4, which are associated with poor prognosis in NSCLC. NIVO improved EFS for patients with at least one of these alteration(s) (n=60) compared to PBO (n=45; HR 0.48), while in patients with no tumor genomic alterations, EFS was similar between the two treatment groups (HR 0.90). EFS was similar in the NIVO groups regardless of tumor genomic alterations, whereas the PBO group with tumor genomic alterations had poorer outcomes than the PBO group without alterations. Of the 46 patients in the NIVO arm and 44 patients in the PBO arm with evaluable ctDNA before and after neoadjuvant treatment, 40/46 (87%) in the NIVO arm and 37/44 (84%) in the PBO arm had detectable ctDNA before neoadjuvant treatment. In the NIVO arm, among the 30 (75%) patients who exhibited ctDNA clearance after neoadjuvant treatment, 21 (70%) achieved a pCR, and 9 (30%) did not. In the PBO arm, 17 patients (46%) achieved ctDNA clearance after neoadjuvant therapy, 3 (18%) achieved a pCR, and 14 (82%) did not. Post-surgery MRD-negativity rates were low in both treatment arms, regardless of pCR status. Among patients who did not exhibit ctDNA clearance after neoadjuvant treatment, almost none achieved a pCR, and MRD-negativity rates were also low, regardless of treatment arm. EFS was improved in the NIVO arm compared to PBO in patients who were ctDNA-positive at baseline (HR 0.58). Patients with ctDNA clearance after neoadjuvant treatment had improved EFS for both treatment arms, although NIVO was associated with improved EFS compared to PBO in patients with ctDNA clearance (HR 0.48) and in patients without ctDNA clearance (HR 0.76). NIVO also improved EFS in MRD-negative patients before adjuvant treatment compared to PBO (HR 0.75). A machine learning model trained with data from biomarker-evaluable patients across both treatment arms indicated that ctDNA clearance, pCR, non-N2 nodal status, squamous tumor histology, and treatment with NIVO were the most predictive factors for longer EFS. SMARCA4 mutation, CDKN2A alteration, and KEAP1 mutation were less predictive of EFS. These results indicate that perioperative NIVO is consistently associated with EFS benefits compared to PBO across patient subgroups and further support the use of perioperative NIVO for resectable NSCLC. Markers such as ctDNA clearance, pCR, non-N2 nodal status, and squamous histology, may have potential use for risk stratification and guiding clinical decisions, but analyses of larger patient cohorts are needed to validate these findings.
Phase 1 trial of mesothelin-targeting CAR T cells for mesothelin-positive esophagogastric cancer with peritoneal carcinomatosis
CT026. A phase I trial of intraperitoneal (IP) mesothelin (MSLN)-targeted CAR T-cell therapy in patients with MSLN-positive esophagogastric cancer (EGC) with peritoneal carcinomatosis
Joan Choo (Memorial Sloan Kettering Cancer Center, New York, NY, USA) presented a phase 1 trial of M28z1XXPD1DNR, a mesothelin (MSLN)-targeting CAR T cell for MSLN-positive esophagogastric adenocarcinoma (EGA) with peritoneal carcinomatosis. Peritoneal carcinomatosis occurs in up to 40% of patients with EGA, and outcomes are poor. MSLN is expressed in 40 to 60% of cases of EGA. M28z1XXPD1DNR expresses a CD28 costimulatory domain, a modified CD3z signaling domain, and a PD-1 dominant negative receptor that provides T cell checkpoint blockade. Previous studies have indicated that regional delivery of MSLN-targeting CAR T cells in combination with systemic anti-PD-1 inhibitors is safe and clinically effective in patients with malignant pleural disease. Seven patients with MSLN-positive esophagogastric cancer with evidence of peritoneal carcinomatosis and at least one prior line of therapy received treatment. After lymphodepletion, CAR T cells were administered intraperitoneally. Patients received doses of 3 x 106 (n=2), 6 x 106 (n=2), and 1 x 107 (n=3) cells/kg. All patients had peritoneal disease, and some had extraperitoneal disease at baseline. All patients had received 3 to 5 prior lines of therapy. Most patients experienced Grade 1 or 2 cytokine release syndrome (CRS) early in treatment, which was resolved quickly with tocilizumab. For all but one patient, this occurred 1 to 3 days after administration. No cases of neurotoxicity, pleuritis, or pericarditis were observed. Low-grade gastrointestinal toxicities were the most common treatment emergent adverse event observed, and all were easily managed. No dose-limiting toxicities or Grade 5 adverse events occurred. Three patients experienced stable disease by RECIST criteria. One patient experienced tumor shrinkage in the target lesions but progressive disease elsewhere, including lymph nodes, suggesting heterogenous MSLN expression limited CAR T cell efficacy. CAR T expansion peaked on Day 25 to 30, and vector copies in peripheral blood remained detectable beyond 100 days. Although M28z1XXPD1DNR exhibited limited clinical activity, intraperitoneal administration of the CAR T cells was safe, treatment was well-tolerated, and the CAR T cells were persistent in peripheral blood. Lack of clinical activity may have been due to heterogenous expression of MSLN in solid tumors, and the study is currently being amended to recruit patients with peritoneal mesothelioma, which is associated with higher levels of MSLN expression. Translational studies of tumor biopsies are also underway to better understand the mechanisms underlying resistance to treatment.
Deletion of GPNMB enhances anti-tumor immunity through reprogramming of tumor-associated macrophages
1349. Deletion of host-derived GPNMB positively regulates anti-tumor response by reprogramming tumor-associated macrophages
Ruxandra Tonea (University of Chicago, Chicago, IL, USA) discussed a translational study of GPNMB, a novel target against anti-inflammatory macrophages, and potential approaches to reprogram immunosuppressive tumor microenvironments. GPNMG, a type I transmembrane glycoprotein, was previously found to be to be upregulated on dysfunctional cells from mouse melanoma tumors, and single-cell RNA sequencing indicated it is upregulated in myeloid cells in patients with breast cancer that does not respond to anti-PD-1 therapy. Studies in mice indicate that GPNMG is upregulated in anti-inflammatory macrophages. Gpnmb knockout (KO Gpnmb) mice were generated, and they exhibited significant tumor control and survival compared to wild type in models of melanoma, breast cancer, and lung metastases. Tumors from KO Gpnmb mice had increased levels of CD8+ T cells, enhanced activation and memory-like phenotypes, and increased numbers of proinflammatory macrophages and monocytes. Single cell RNA sequencing of immune cells from tumors indicated that macrophages and T cells are most affected by Gpnmb knockout. KO Gpnmb induced macrophage subtypes C1qa+ and Spp1+ to shift from anti-inflammatory to pro-inflammatory gene expression, and the gene expression signature from Spp1+ macrophages was found to be upregulated in inflamed breast cancer tumors from humans. Macrophages from tumors of KO Gpnmb mice exhibited reduced lipid uptake, which was consistent with KO Gpnmb shifting macrophages from an anti-inflammatory phenotype to a pro-inflammatory phenotype. KO Gpnmb tumor-associated macrophages exhibited reduced suppression of CD8+ T cells compared to wild type. Tumor control in KO Gpnmb mice was dependent on T cells, and depletion of CD8+ T cells and blockade of IFN gamma led to partial loss of tumor control. Tumor growth was not affected in Vav1-Cre Gpnmb mouse models, which lack Gpnmb in hematopoietic cells. Knockout of Gpnmb in the host mouse led to tumor control, whereas no tumor control was observed in mice lacking Gpnmb in the bone marrow, suggesting that absence of Gpnmb in non-immune compartments or tissue-derived macrophages is important for tumor control. LysM-Cre Gpnmb mouse models, which lack Gpnmb in myeloid cells, exhibited improved tumor control compared to wild type, indicating that Gpnmb knockout targets non-bone marrow derived tissue resident macrophages for tumor control. These results led to the development of a model in which knockout of host-derived Gpnmb leads to a shift to inflammatory gene expression in subsets of anti-inflammatory tumor associated macrophages, enhancing anti-tumor immunity and tumor control.
Vaccines stimulate bystander T cells, promoting anti-tumor immunity
1350. Vaccine stimulation of bystander T cells in the lymph node unexpectedly promotes tumor-specific host immunity to regress solid tumors
Leyuan Ma (Children’s Hospital of Philadelphia, Philadelphia, PA, USA) presented a study of bystander T cells promoting tumor-specific immunity after vaccination. This work is related to reports of SARS-COV-2 mRNA vaccines sensitizing tumors to immune checkpoint blockade and prior observations of spontaneous complete tumor regression after booster doses of vaccines not related to cancer. In at least one case, the SARS-COV-2 mRNA booster vaccine was associated with enrichment of T cells in the tumor-draining lymph node and the tumor bed. A similar phenomenon was observed independently in mice, with an amphiphile-ligand vaccine stimulating “tumor-irrelevant” CAR-T cells (bystander CAR T-vax) and causing tumor regression in multiple tumor models, including pancreatic ductal adenocarcinoma (PDAC), melanoma, and breast cancer. An amphiphile-ligand vaccine carrying the chemical ligand fluorescein (FITC) was used to stimulate bispecific T cells expressing a FITC-CAR and a mesothelin-targeting CAR (meso-CAR). As expected, the FITC vaccine stimulated anti-tumor activity of the bispecific T cell expressing FITC-CAR and meso-CAR, but it also unexpectedly stimulated anti-tumor activity of the control “tumor-irrelevant” T cell expressing only the FITC-CAR. Tumor regression occurred in vivo but not in vitro, and regression occurred quickly, 24 hours after vaccination. Regression persisted up to two weeks and cured 40% of mice in some models of melanoma. The boosting of FITC-CAR-T cells by the FITC vaccine (bystander FITC-CART-Vax) required the duration of CAR T cells for more than 24 hours and up to 3 days. Bystander FITC-CAR T-Vax induced increased levels of CD4+ and CD8+ T cells in the lymph nodes and tumor microenvironment (TME) as well as dendritic cells in the lymph nodes. Bystander FITC-CAR T-Vax also induced a transient release of cytokines IFN gamma and CCL2 24 hours after vaccination that induced remodeling of the PDAC TME, which promoted immune cell infiltration over time. Bystander FITC-CAR T-Vax-mediated tumor regression in PDAC models was dependent on IFN-gamma and the presence of tumor-reactive cytotoxic lymphocytes in the host and combining bystander FITC-CART-Vax with PD-L1 inhibitors enhanced tumor progression. Antigen presentation to CD8+ T cells is also required since inhibition of IFN gamma resulted in downregulation of MHC class I and beta 2 microglobulin knock out resulted in loss of tumor regression. These results have developed a hypothesis that the observed bystander effect is caused by vaccine-stimulated dendritic cell priming of CAR T cells in the lymph nodes, which then activates host immunity and stimulates remodeling of the TME. By activating host immunity, anti-tumor T cells are reinvigorated and infiltrate the remodeled TME, leading to tumor regression. While these findings need to be validated in patients, they may support new approaches to developing off-the-shelf cancer vaccines and adoptive T cell therapies.
Investigating the roles of adaptive and innate immune effector cells in IL18-mediated clearance of heterogenous solid tumors
1351. IL-18 secreting CAR-T cells uncover divergent immune mechanisms of antigen-heterogeneous solid tumor clearance
Audre May (Stanford University School of Medicine, Stanford, CA, USA) presented a study exploring the mechanisms underlying the anti-tumor activity of IL-18 secreting CAR T cells in heterogeneous solid tumors. Heterogeneous antigen expression and antigen escape in solid tumors present a problem for CAR T cell therapy, and endogenous immune responses involving both innate and adaptive effector cells are necessary for regression of solid tumors. The goal of this study was to study the biology of solid tumor clearance using cytokine-secreting CAR T cells that induce an endogenous immune response. This study focused on three previously developed cytokine-producing CAR T cells: IL-18 secreting CAR T cells, which promote lymphoid activation and IFN gamma secretion; IL-36 secreting CAR T cells, which stimulate myeloid cell activation and antigen presentation; and FLT3L secreting CAR T cells, which stimulate dendritic cell expansion and survival. In order to study the contributions of adaptive and innate effector immunity, heterogenous tumors consisting of 80% CD19-positive tumor cells and 20% CD19-negative tumor cells were introduced subcutaneously in mice and treated with CD19-targeting cytokine-producing CAR-T cells. IL-18 secreting CAR T cells exhibited superior clearance of CD19-positive and –negative tumor cells, compared to the IL-36 secreting and the FLT3L secreting CAR T cells. IL-18 secreting CAR T cells exhibited low persistence in tumor draining lymph nodes and the tumor, but large accumulations were found in the spleen. Although there were no differences in the absolute number of T cells in the spleens and tumors of the different treatment groups, the spleens and tumors of mice treated with the IL-18 secreting CAR T cells exhibited higher levels of CD8+ T cells compared to CD4+ T cells, which were enriched for effector memory phenotypes and T cell activation markers. Bulk RNA sequencing of endogenous T cells indicated that IL-18 secreting CAR T cells induced expression of cytotoxic/effector and cycling gene profiles and decreased expression of stemness/persistence markers. Surprisingly, the IL-36 CARs with largest numbers of tumor reactive CD8+ T cells exhibited the lowest levels of tumor control. To determine the relative contribution of adaptive immune responses to tumor clearance by cytokine-secreting CAR T cells, heterogeneous tumors were introduced in wild type mice and Rag1-/- mice lacking endogenous T cells. Tumor control by IL-18 secreting CAR T cells was relatively the same in wild type and Rag1 -/- mice, indicating tumor clearance was independent of adaptive immunity. The tumor draining lymph nodes and spleens of mice treated with IL-18 secreting CAR T cells exhibited increased levels of myeloid cells compared to control mice. Myeloid cells from the spleens of mice treated with IL-18 secreting CAR T cells exhibited upregulated expression of genes related to IFN gamma response, antigen presentation machinery, and complement pathways. Tumor clearance by IL-18 secreting CAR T cells was independent of neutrophil activity. These results indicate that superior heterogenous tumor control of IL-18 secreting CAR T cells was not dependent on CAR T cell persistence and adaptive immune responses, but it was associated with increased myeloid cell expansion and activation. These results also indicate the potential of innate immune cells, like myeloid cells, for developing effective cellular therapies for solid tumors. Studies investigating the role of macrophages in IL18-mediated heterogenous tumor clearance are underway.
Preliminary results from a Phase 1 study of ABO2203, a lipid nanopartical-encapsulated mRNA encoding a T cell engager for relapsed/refractory B cell non-Hodgkin lymphoma
CT032. ABO2203, a messenger RNA encoding a CD3×CD19 T cell engager, for relapsed/refractory B cell non-Hodgkin lymphoma: Preliminary results from first-in-human study
Li Wang (Ruijin Hospital Affiliated to Shanghai Jiao Tong University School of Medicine, Shanghai, China) reported preliminary results from a first-in-human study of ABO2203, a first-in-class lipid nanoparticle (LNP)-encapsulated messenger RNA (mRNA) encoding a CD3 x CD19 T cell engager for relapsed/refractory B cell non-Hodgkin lymphoma. Preclinical studies indicate that ABO2203 exhibited more favorable pharmacokinetics, lower cytokine release, and improved B cell depletion compared to its matched T cell engager protein P4107. The dose escalation and expansion stage of the study consisted of three cohorts. Cohort 1 (n=3) received doses ranging from 3 to 480 micrograms ABO2203, Cohort 2 (n=3) received doses ranging from 120 to 960 micrograms, and Cohort 3 (n=3) received doses ranging from 240 to 1920 micrograms. Doses were administered as weekly subcutaneous injections. Three patients in the study had follicular lymphoma (FL), 2 had diffuse large B cell lymphoma (DLBCL), 2 had marginal zone lymphoma, and 2 had mantle cell lymphoma. Patients had received a median of 4 prior lines of therapy, and all had failed CD20-targeting therapies. ABO2203 T cell engagers exhibited peak concentration in the serum at 5.5 days after each dose, and the mean T cell engager half-life was 7.9 days. This slow pharmacokinetic profile led to a natural priming of T cells, enabling higher starting doses without increased risk of cytokine release syndrome (CRS). Across all 3 cohorts, the overall response rate was 67% (6/9), and the disease control rate was 78% (7/9). 44% (4/9) of patients achieved a complete metabolic response. The overall response rate was 100% at the highest dose. ABO2203 was well-tolerated, with dose-limiting toxicities observed. No CRS, neurotoxicity, or hepatic adverse events higher than Grade 1 were observed. The most common treatment-emergent adverse events were Grade 1 or 2 pyrexia. Grade 3 or 4 TEAEs of decreased neutrophils or white blood cells occurred in 3 patients, and all patients recovered within one week. Analysis of peripheral immune cells demonstrated depletion of B cells within one week. ABO02203 exhibits favorable safety and pharmacokinetic profiles and preliminary clinical efficacy, and these results indicate that it may be a promising treatment option for relapsed/refractory B cell non-Hodgkin lymphoma. Dose expansion with the regimen for Cohort 3 (240 to 1920 micrograms) is ongoing.
CAR-PRISM: BCMA-targeting CAR T cell therapy for high-risk smoldering melanoma
CT103. Ciltacabtagene autoleucel in high-risk smoldering myeloma: Results from the CAR-PRISM trial
Omar Nadeem (Dana-Farber Cancer Institute, Boston, MA, USA) presented initial safety and efficacy results from the phase 2 CAR-PRISM study, which investigated the safety and efficacy of ciltacabtagene autoleucel (citla-cel), a BCMA-targeting CAR T cell therapy, for treating patients with high-risk smoldering myeloma (HR-SMM). HR-SMM is a precursor plasma cell neoplasm, and 50% of patients risk progression to multiple myeloma within 2 years of diagnosis. Previous studies of citla-cel indicate its curative potential for multiple myeloma, especially when it is administered to patients in early disease states. The CAR-PRISM trial hypothesized that early use of cilta-cel in patients with early-stage HR-SMM could potentially lead to disease eradication. In the dose escalation phase of the trial, 20 patients with HR-SMM received cilta-cel at doses of 0.3, 0.5, or > 0.5 x 106 CAR T cells/kg following lymphodepletion, with a target dose of 0.3 - 0.75 x 106 viable CAR+ T cells/kg. Twelve patients received lower doses (0.3 to 0.5 x 106 CAR T cells/kg) of cilta-cel, and 8 patients received doses higher than 0.5 x 106 CAR T cells/kg. Patients did not receive induction or bridging therapy. No dose-limiting toxicities were observed. Cytokine release syndrome (CRS) occurred in all patients, but all cases were grade 1 or 2. Hematologic toxicities were common, with 90% of patients experiencing grade 3 neutropenia, but all cases were resolved. Increase in absolute lymphocyte count was associated with neurotoxicity, which required a protocol amendment. Neurotoxicities that were not immune effector cell-associated neurotoxicity syndrome occurred in 7 patients, and most cases were in patients who received the highest dose level. Four of the 7 cases comprising cranial nerve palsies have been resolved, and the remaining 3 cases are ongoing. All patients received high-dose steroids and have experienced improvement of symptoms, with mild motor symptoms having minimal impact on daily activities. At a median of 15.3 months follow-up, all patients achieved an objective response, with 18 patients (90%) experiencing a complete or stringent complete response at the time of data cut-off. In the remaining 2 patients, follow-up was approximately 3.5 months. All 6 patients with follow-up beyond 18 months have sustained minimum residual disease (MRD)-negative status. No patients developed disease progression and median progression-free survival was not reached. After cilta-cel infusion serum BCMA (sBCMA) levels declined rapidly, independently of dose, and baseline sBCMA did not correlate with increased incidence of neurotoxicities. Patients who developed neurotoxicities exhibited higher post-infusion absolute lymphocyte counts, increased levels of effector memory T cells, and increased CD4/CD8 ratios. This is the first study of CAR T cells as a primary therapy in a precursor cancer setting. Although more time is needed to determine the durability of responses and the risk/benefit ratio of using cilta-cel in a precursor cancer setting, these early results suggest that the use of CAR T cell therapy in an earlier disease state while patients' immune systems are intact improves outcomes and prevents progression to multiple myeloma.
Early safety and efficacy results of STAR-101, a first-in-human study of multi-chain CAR T cells for mesothelin-expressing solid tumors
CT104. Initial results of a first in human dose-escalation study of KIR-CAR in patients with advanced mesothelin-expressing solid tumors
Janos L. Tanyi (Abramson Cancer Center, University of Pennsylvania, Philadelphia, PA) reported results from STAR-101, a first-in-human phase 1 trial of mesothelin-targeting Killer Immunoglobulin-like Receptors (KIR)-CAR T cells for refractory solid tumors (SynKIR-110). While conventional single-chain CAR T cells have been successful in treating hematologic cancers, they have not been successful at treating solid tumors, often demonstrating on- and off-target toxicities. Multi-chain KIR-CARs separate target binding from T cell signaling and activation, more closely resembling endogenous T cell signaling and activation and allowing for T cell rest and recovery. Preclinical studies with KIR-CAR T cells indicated that KIR-CAR T cells exhibit reduced levels of activation and exhaustion markers compared to single-chain CAR T cells, demonstrating ability of T cells to rest. Cell-killing activity of KIR-CAR T cells was faster and more specific than single-chain CAR T cells, and in mouse mesothelial xenografts, KIR-CAR T cells were associated with increased anti-tumor efficacy, decreased duration of IFN gamma secretion, and fewer circulating T cells. Based on this preclinical data demonstrating superiority of multi-chain over single chain CAR T cells, a phase I dose escalation study was designed. Patients with advanced ovarian/fallopian carcinoma, epithelial mesothelioma, or cholangiocarcinoma that was refractory to standard of care treatment participated in the study. This analysis focused on 9 patients in the first three dose expansion cohorts. Patients received single doses of 1 x 107 (n=3), 3 x 107 (n=3), or 10 x 107 (n=3) SynKIR-110 cells/m2 after lymphodepletion. Patients had received an average of 4 prior systemic treatments. No dose-limiting toxicities were observed, and 3 patients (33%) experienced grade 1 or 2 cytokine release syndrome. No immune effector cell-associated neurologic events were observed. Multiple grade 4 hematologic toxicities occurred due to lymphodepletion. Levels of SynKIR-110 cells generally increased with dose level, and CAR T cells could be identified in peripheral blood for up to 28 days. Patient serum levels of inflammatory cytokines such as IFN gamma and TNF alpha were relatively low and did not correlate with CRS, and serum cytokine kinetics demonstrated on-target CAR T cell activation profiles. Tumor reductions were observed with increasing doses of SynKIR-110 cells, and one patient who received the highest dose in this analysis achieved a partial response that has been ongoing for over 3 months. With the favorable safety profile and encouraging early clinical activity, these data support further clinical study of SynKIR-110 for mesothelin-expressing solid tumors. Patient enrollment and dose-escalation for STAR-101 is ongoing to identify the maximum tolerated dose and recommended phase 2 dose.
Tumor-localized 4-1BB agonist in combination with a CEA-directed T cell engager for microsatellite stable metastatic colorectal cancer
CT105. FAP-targeted 4-1BBL trimeric costimulation amplifies T-cell activation and antitumor efficacy of a CEA-directed T-cell engager in MSS colorectal cancer
Ignacio Melero (Clínica Universidad de Navarra/Nuffield Department of Medicine, University of Oxford, Pamplona/Oxford, Spain) reported results of a phase 1b study evaluating the combination of cibisatamab, a CEA-directed CD3 T cell engager, with an FAP-targeted 4-1BBL trimeric costimulator in patients with microsatellite stable/mismatch repair proficient (MMS/pMMR) metastatic colorectal cancer (mCRC). mCRC is the third most common cancer and the second most common cause of cancer-related deaths. Treatment options are limited for patients whose disease progresses after first- and second-line therapy, and PD-(L)1 inhibitors, while effective against microsatellite instability-high/mismatch repair deficient CRC, have limited clinical activity against MMS/pMMR mCRC. Cibisatamab, a CEA-targeting T cell engager, has shown signs of clinical activity against MMS/pMMR mCRC, but full T cell activation requires signaling through the T cell receptor-CD3 complex as well as costimulatory pathways, such as those mediated by the 4-1BB receptor. FAP-4-1BBL is a bispecific fusion protein consisting of a domain that binds to fibroblast activating protein (FAP) and a 4-1BBL agonist domain to provide a co-stimulatory signal to T cells. Preclinical studies indicate that FAP-4-1BBL and cibisatamab work synergistically to inhibit tumor growth in humanized mice with CRC xenografts. Fifty-two patients with MMS/pMMR mCRC received cycles of cibisatamab with sequential FAP-4-1BBL at escalating doses weekly (Part 1, n=30) or every three weeks (Part 2, n=22). 73% of patients had liver metastases, 50% had lung metastases, and 15% had peritoneal involvement. Patients had received a median of 3 prior lines of therapy, and only 2 patients (3.8%) had received prior checkpoint inhibitors. Thirty patients (57.7%) experienced cytokine release syndrome, and all but 2 cases were grade 1 or 2. Two patients experienced Grade 5 treatment-related adverse events (TRAEs) which included general physical health deterioration and CMV-related colitis. Gastrointestinal toxicities, a known on-target, off-tumor effect of cibisatamab, were also observed. Cibisatamab + FAP-4-1BBL was associated with increased levels of IFN gamma and sCD25 and enhanced T cell activation compared to cibisatamab monotherapy. Sequential tumor biopsies indicated that combination therapy was associated with increased levels of intratumoral T cells and a more immune-engaged state. T cell infiltration often occurred in the fibrotic tracks of tumors, which are typically immune deserts, thus indicating remodeling of the tumor immune microenvironment. Confirmed partial responses were observed across all FAP-4-1BBL dose levels and dosing regimens. The confirmed objective response rate of 13% and the confirmed disease control rate of 50% compare favorably to prior reports of cibisatamab monotherapy. This is the first clinical study investigating the coordination of two T cell activation signals in patients with immunologically cold tumors and FAP-4-1BBL in combination with cibisatamab may someday represent a new treatment option for a patient population with unmet clinical needs. These data also support further studies of 4-1BB agonists to enhance the anti-tumor activity of T cell engagers.
Safety and clinical efficacy of an mRNA-based immunotherapy in combination with pembrolizumab for locally advanced/metastatic melanoma
CT106. First-line mRNA-4359 plus pembrolizumab (pembro) in locally advanced or metastatic melanoma: Results from the phase 1/2 mRNA-4359-P101 study
Pavlina Spiliopoulou (University of Glasgow, Glasgow, United Kingdom) presented safety, clinical, and translational data of the mRNA-4359-P101 study, a phase 1/2 trial of mRNA-45359 in combination with pembrolizumab for first-line treatment of locally advanced/metastatic melanoma. mRNA-4359 is an investigational immunotherapy designed to address immune resistance. mRNA-4359 consists of lipid nanoparticle-encapsulated mRNA encoding fragments of PD-L1 and IDO1. mRNA-4359 is administered as an intramuscular injection. It is hypothesized that antigen presenting cells such as dendritic cells take up mRNA-4359, then produce and present the PD-L1 and IDO1 antigens to T cells. T cells are activated against PD-L1 and IDO1 infiltrate the tumor, recognizing and destroying immunosuppressive cells expressing PD-L1 and/or IDO1. Destruction of PD-L1 and/or IDO1-expressing immunosuppressive and tumor cells remodels the tumor microenvironment, converting it to a more immune-permissive state. Tumor cell death also causes the release of epitopes that may stimulate expansion of additional tumor-specific T cells, broadening the immune response. The presentation focused on data from Arm 2a of the study, which was a dose expansion of mRNA-4359 in combination with pembrolizumab for patients with previously untreated locally advanced or metastatic melanoma. Twelve patients were treated with combination therapy for up to 9 cycles. At a median follow-up of 54.2 weeks, 4 patients completed mRNA-4359 and are ongoing with maintenance pembrolizumab, 1 patient is ongoing with combination therapy, 3 patients discontinued pembrolizumab only, and 4 patients discontinued combination therapy. Most patients had stage 4 melanoma, and 8 patients (67%) had PD-L1 TPS score of 1% or higher. The cohort included mucosal melanoma as well as cutaneous melanoma. Three patients (25%) had received prior (neo)adjuvant immunotherapy. All observed mRNA-4359 adverse events (AEs) were grade 1 or 2. No dose limiting toxicities were observed, and the AEs observed in this patient population were consistent with AEs observed in other studies of immunogenic RNA-based therapies. AEs associated with pembrolizumab were mostly low-grade. No dose limiting toxicities were observed, one patient had to discontinue treatment due to AEs, and one patient experienced grade 4 immune-mediated pancreatitis. The overall response rate was 83%, and the disease control rate was 92%. The median time to response was 6 weeks, and responses were observed regardless of PD-L1 status or prior (neo)adjuvant treatment. Most responders experienced deep, durable responses, and 2 patients experienced a complete response. All evaluable patients exhibited increases in PD-L1- and IDO1-specific T cells from baseline in peripheral blood, and all patients exhibited an increase in novel expanded T cell receptor clones, supporting the proposed mechanism of action for mRNA-4359. Although these results need to be replicated and validated in larger patient populations, these findings support further clinical development of mRNA-4359 in combination with immune checkpoint inhibitors to overcome immune resistance in patients with locally advanced or metastatic melanoma.
Targeted lipid nanoparticle-mediated delivery of IL-12 circular RNA remodels the tumor immune microenvironment in mouse models of lung cancer
4042. Tumor-tailored ionizable lipid nanoparticles facilitate IL-12 circular RNA delivery for enhanced lung cancer immunotherapy
Yue Xu (Baylor College of Medicine, Houston, TX, USA) presented a preclinical study of the feasibility and anti-tumor activity of intratumoral delivery of IL-2 as circular RNA (circRNA). mRNA-based immunotherapies have generated interest in the immuno-oncology field, but they are often associated with systemic toxicities and transient expression. CircRNA offers several advantages over mRNA, including more sustained protein expression, easier manufacturing, robust stability, better efficacy and safety, and easier logistics for administration. IL-12 is a potent pro-inflammatory cytokine that acts on multiple arms of the immune system. Despite its potent immunomodulatory activity, clinical development has been limited due to systemic toxicity. Intratumoral delivery provides a mechanism to reduce systemic toxicity. IL-12 circRNA was synthesized and packaged into lipid nanoparticles (LNPs) that were identified in a screen for LNPs suitable for circRNA delivery. IL-12 circRNA exhibited more consistent, longer-lasting IL-12 expression in LLC1 and HKP1 cell lines in vitro, and circRNA exhibited more favorable, significantly longer-lasting kinetics when injected intramuscularly in mice. Intratumor injection of IL-12 circRNA LNP alone had a robust anti-tumor effect in anti-PD-L1 resistant LLC1 lung cancer mouse models. Intratumoral injection of IL-12 circRNA LNP and anti-PD-L1 antibodies had a synergistic effect in inhibiting tumor growth, and combination treatment was associated with increased levels of intratumoral CD45 cells, CD8 T cells, CD4 T cells. Intratracheal administration of IL-12 circRNA LNPs slowed tumor progression in mouse models of Kras-driven non-small cell lung cancer, and L-12 circRNA LNPs were associated with favorable safety profiles and increased infiltration of CD45 cells, CD4 and CD8 T cells, natural killer T cells, granulocytes, and monocytes in the tumor, indicating a complete remodeling of the tumor microenvironment. Based on these data, the LNP-circRNA-based platform provides localized cytokine expression that remodels tumors, making them immunologically “hot,” with minimal side effects. This approach may someday represent a safer strategy for targeted delivery of cytokines, stimulating anti-tumor immunity in patients with lung cancer.
A novel approach using tumor-associated macrophages to deliver payloads to aggressive solid tumors
4043. A novel TAM-targeted therapeutic for rare and aggressive solid tumors
Faith H. Barnett (Resolute Science, Inc., San Diego, CA) discussed a novel approach to target pro-tumorigenic tumor-associated macrophages (TAMs) in solid tumors. TAMs are present in most aggressive cancers, representing 30 to 50% of tumor cells and promoting tumor growth and metastasis. Attempts to eliminate or reprogram TAMs have been unsuccessful, so macrophage-targeted conjugates (MAC-TACs) have been developed to target TAMs for delivery of payloads to tumors, similar in concept to antibody-drug conjugates (ADCs). MAC-TACs consist of a TAM receptor-targeting ligand, a backbone, linker, and payload. The TAM receptor targeted by MAC-TACs is highly and uniformly in a wide range of malignancies, including sarcomas, prostate cancer, gastric cancer, melanoma, pancreatic cancer, and bladder cancer, suggesting that multiple cancer types could be treated with a single drug. MAC-TACs are 18 times smaller than ADCs, improving their delivery and retention in the tumor, and they have potential to treat cancers of the central nervous system. In vivo and in vitro studies indicate that MAC-TAMs are taken up quickly by TAMs, which internalize the constructs and transfer the payload to neighboring cancer cells. The therapeutic MAC-TAC RS-5, which carries the antineoplastic drug MMAE as a payload, exhibits anti-tumor activity in mouse models of glioma and melanoma and in an aggressive sarcoma patient-derived xenograft model. Two doses of RS-5 rapidly reduced tumor size in an intracranial sarcoma model, and in an intercranial glioma model, RS-5 extended median survival by 79%. In the glioma model, anti-tumor activity was retained with reduction of dosing to once every two weeks, suggesting feasibility to translation to clinical application. In all models tested, body weight and all vital signs for mice receiving MAC-TACs have remained within normal ranges, suggesting MAC-TACs are well-tolerated. Overall, RS-5 has demonstrated efficacy in pre-clinical models of drug-resistant tumors and difficult to treat tumors such as sarcoma and glioma. The United States Food and Drug Association has granted Orphan Drug Designation to RS-5 for the treatment of soft tissue sarcomas, and the first in-human dose-escalation study is planned for the summer of 2026
Development and characterization of a novel DLL3-targeting trispecific T cell engager that integrates CD2/CD58 co-stimulation
4046. A novel DLL3 trispecific T cell engager antibody integrating the CD2/CD58 co-stimulatory pathway demonstrates superior antitumor efficacy and balanced safety profile
Yu Liang (Probio, Inc, Pennington, NJ) presented the development and characterization of a novel DLL3-targeting trispecific T cell engager for solid tumors. There is a large gap in the efficacies of T cell engagers for hematologic malignancies compared to T cell engagers immunotherapies for solid tumors. One challenge for T cell engagers in the treatment of solid tumors is T cell exhaustion caused by the immunosuppressive tumor microenvironment (TME). One approach to address this challenge is to engineer a co-stimulatory domain in the T cell engager to enhance T cell activation. CD2-CD58 co-stimulation was examined in this study because CD2 expression is constitutive and robust on human T cells, and the CD2-CD58 axis plays roles in organizing and stabilizing immune synapses and in maintaining T cell fitness. A novel DLL3 x CD3 x CD2 trispecific antibody was developed and compared to the DLL3 x CD3 bispecific antibody tarlatamab. The trispecific antibody exhibited increased tumor cell killing and inflammatory cytokine release compared to the bispecific across multiple peripheral blood mononuclear cell (PBMC) donors. CD2 co-stimulation in the trispecific also enhanced T cell proliferation and survival in vitro, compared to the bispecific tarlatamab. The DLL3 x CD3 x CD2 trispecific antibody also consistently promoted T cell-mediated cytotoxicity after 7 serial tumor challenges due to ability to maintain T cell activation. T cell bridging and T cell fratricide were not observed with the trispecific antibody, and trispecific-mediated T cell activation and proliferation occurred only in the presence of the DLL3 target. The DLL3 x CD3 x CD2 trispecific or tarlatamab were administered to cell line-derived xenograft mouse models of small cell lung cancer with human PBMC transplants, and 5/7 mice that received the trispecific antibody had a complete response, compared to 2/6 mice that received tarlatamab. Arming a DLL3 x CD3 T cell engager with CD2 co-stimulation resulted in improved T cell proliferation and survival in vitro and greater efficacy in tumor killing in vivo. Toxicity and pharmacokinetic studies of this next-generation T cell engager in non-human primates are ongoing, with the potential to initiate clinical studies soon.
Final results of cohort 2 of the INFINITY study: STRIDE regimen of tremelimumab and durvalumab followed by non-operative management of microsatellite instability-high resectable gastric or gastroesophageal junction adenocarcinoma
CT230. STRIDE regimen of tremelimumab and durvalumab as non-operative management strategy of microsatellite instability-high (MSI-H) resectable gastric or gastroesophageal junction adenocarcinoma (GAC/GEJAC): Final results of the cohort 2 of INFINITY study by GONO GI
Alberto Giovanni Leone (Fondazione IRCCS Istituto Nazionale dei Tumori, Milan, Italy) presented the final results of cohort 2 of the INFINITY trial investigating neoadjuvant treatment with the Single Tremelimumab Regular Interval Durvalumab (STRIDE) regimen for microsatellite instability-high (MSI-H) resectable gastric or gastroesophageal junction adenocarcinoma (GAC/GEJAC). This trial aimed to evaluate the efficacy of non-operative management for patients who had a clinical complete response to neoadjuvant therapy. Cohort 1 of the INFINITY trial indicated that the STRIDE regimen - a single high dose of tremelimumab (anti-CTLA-4; 300 mg) combined with durvalumab (anti-PD-L1; 1500 mg every four weeks) in the neoadjuvant setting – followed by surgery and standard follow-up resulted in a pathological complete response (pCR) rate of 60%. Cohort 2 was designed to determine whether surgery was necessary in patients who achieved a pCR with the STRIDE regimen. Patients in Cohort 2 underwent restaging during weeks 12 through 14 of the protocol, after receiving the STRIDE regimen, to assess clinical response. Clinical complete response (cCR) was assessed by bite-on-bite biopsies, 18F-FGD-PET, CT scan, EGD/EUS, and tissue-informed cDNA. Patients who achieved cCR underwent non-operative management (NOM), which involved intense surveillance. Patients who had an incomplete response to STRIDE underwent salvage surgery followed by standard follow-up. Seventeen patients were evaluable for Cohort 2, and at two years, the cCR rate was 71%. Thirteen patients achieved a cCR after the STRIDE regimen and began NOM. One patient experienced locoregional progression after initial cCR and underwent salvage surgery. The four patients who did not achieve a cCR all underwent salvage surgery. Three of the four patients were circulating tumor DNA (ctDNA)-positive at restaging, and one patient was ctDNA-negative. At a median follow-up of 27.1 months, the 24-month progression-free survival rate was 94.1%, the 24-month overall survival rate was 100%, and the 24-month gastrectomy-free survival rate was 70.6%. The STRIDE regimen was well-tolerated. 72% of patients experienced an adverse event (AE) of any grade, 3 patients experienced grade 3 AEs, and no grade 4 or 5 AEs occurred. All grade 3 AEs were treatable. One patient had to delay durvalumab administration due to grade 2 immune-related nephritis, and no cases of treatment discontinuation were reported. Based on patient-reported outcomes, quality of life was preserved throughout treatment. Exploratory analyses of plasma ctDNA levels at baseline and post STRIDE treatment indicated that ctDNA was detectable in 13 of 17 patients at baseline, and of these 13 patients, 11 (84%) achieved ctDNA clearance after treatment. Higher baseline ctDNA levels were associated with a lower possibility of a cCR, and ctDNA status at restaging predicted failure to achieve a cCR, with 50% sensitivity and 100% specificity. Differences observed in the exploratory analyses were not statistically significant due to the small patient population. Baseline transcriptional profiling indicated that transcriptional profiles involving immune recognition and tumor-intrinsic profiles were upregulated in patients who achieved a cCR, and transcriptional profiles related to stromal activation, tissue remodeling, and immune trafficking were upregulated in patients who did not achieve a cCR. Patients who did not achieve a cCR also exhibited increased levels of myeloid infiltration in tumors at baseline, suggesting a myeloid-rich tumor microenvironment may have negatively affected intratumoral T cell activity. These results indicate that a NOM strategy after 3 months of treatment with tremelimumab and durvalumab is safe and viable in carefully selected patients with resectable MSI-H GAC/GEJAC. Early data also suggest that biomarkers such as ctDNA or tumor microenvironment markers may be useful in selecting patients for NOM. Although larger randomized or adaptive studies are needed to validate these results, results from the INFINITY trial indicate that neoadjuvant immunotherapy for resectable MSI-H GAC/GEJAC followed by organ-sparing NOM may become a treatment option with curative potential for some patients while reducing morbidity of treatment.
Interim results of the MATISSE trial: Perioperative blockade of the adenosine pathway in combination with chemoimmunotherapy in patients with early-stage resectable non-small cell lung cancer
CT231. Dual CD39 and PD-L1 inhibition: Interim results from the phase 2 MATISSE trial of IPH5201 plus durvalumab (durva) and platinum-based chemotherapy (CT) in patients (pts) with resectable NSCLC
Fabrice Barlesi (Gustave Roussy / University Paris Saclay, Villejuif, France) reported interim results from the single-arm phase 2 MATISSE trial, which investigated IPH5201, an anti-CD39 monoclonal antibody, in combination with durvalumab and platinum-based chemotherapy in patients with resectable non-small cell lung cancer (NSCLC). CD39 is an enzyme overexpressed in the tumor microenvironment (TME) that degrades immunostimulatory extracellular ATP into AMP, which is then further degraded to immunosuppressive adenosine. Blockade of CD39 reduces adenosine levels to enhance dendritic cell activity and restore T cell stimulation. Preclinical studies of IPH5201 indicate that it enhances the anti-tumor activity of durvalumab and chemotherapy by turning immunotherapy-resistant tumors into more immunotherapy-sensitive environments. Patients with early-stage, previously untreated resectable NSCLC received perioperative PIH5201 combined with durvalumab and chemotherapy. Clinical results were available for 40 patients, all of whom had received at least one cycle of neoadjuvant therapy. Perioperative treatment was associated with a favorable safety profile, with no new safety signals, and 35 of the 40 received surgery. One grade 5 treatment-emergent adverse event (TEAE) occurred due to a postoperative pneumonia, and 72.5% of patients experienced at least one IPH5201-related TEAE. The objective response rate to systemic treatment was 62.5%, with 3 patients (7.5%) achieving a complete response and 22 (55.0%) achieving a partial response. Of the 35 patients who underwent surgery, R0 resection was reached in 31 (88.6%) cases. The overall pathological complete response (pCR) rate was 27.5% and major pathologic response (MPR) rate was 32.5%. Among patients with PD-L1 TPS >= 1% (n=28), the pCR rate was 35.7% and 50.0% in patients with PD-L1 TPS >= 50% (n=14). Patients with an MPR or pCR tended to have higher baseline levels of intratumoral CD39 and CD8 cell density, and there was no correlation between CD39 cell density and PD-L1 expression in the tumor. MATISSE is the first study to show feasibility and early clinical activity of neoadjuvant CD39 combined with PD-L1 inhibition and chemotherapy for patients with resectable early-stage NSCLC. Although more time and larger sample sizes are needed to determine durability of responses and survival, improved pCR rates and high proportion of patients reaching surgery support further investigation of the CD39/adenosine pathway to enhance immunotherapy in patients with NSCLC, especially those with PD-L1-positive tumors. The MATISSE trial is ongoing and currently recruiting patients with PD-L1-positive (TPS >=1%) tumors.
First in-human study of denikitug, a CCR8-targeting antibody, for advanced solid tumors
CT232. Preliminary results of a phase 1, first-in-human, dose-escalation study of the anti-CCR8 antibody denikitug in participants with advanced solid tumors
Bruno B. Bockorny (Beth Israel Deaconess Medical Center, Boston, MA, USA) discussed preliminary results from a first-in-human dose-escalation phase 1 study of denikitug, an anti-CCR8 monoclonal antibody in patients with advanced solid tumors. Regulatory T cells (Tregs) Tregs drive immunosuppression in the tumor microenvironment (TME), and they represent the main mechanism of resistance to immune checkpoint blockade. CCR8 is preferentially expressed on intratumoral Tregs as opposed to peripheral Tregs. Denikitug is engineered to maximize antibody-dependent cellular toxicity, blocking CCR8 signaling and promoting apoptosis of CCR8-expressing Tregs. The phase I study investigated denikitug monotherapy and in combination with zimberelimab; the results presented today included the monotherapy data. Sixteen patients participated in Part A, the dose escalation stage of the study, where patients received doses of denikitug ranging from 1 to 100 mg every three weeks, and 41 patients participated in Part B of the study, which analyzed paired tumor biopsies. All patients in the study had advanced solid tumors with no standard treatment options available. Diverse tumor types were represented in the study, including non-small cell lung cancer, gastric adenocarcinoma, gynecologic cancers, breast cancer, and head and neck squamous cell carcinoma. Patients had received a median of 3 prior lines of therapy, 38 patients (67%) had received prior anti-PD-(L)1 therapy. The overall safety profile of denikitug was manageable. 98% of patients in the study experienced a treatment-emergent adverse event (TEAE), and 74% of patients experienced a grade 3 TEAE. All grade 3 TEAEs were manageable with supportive care. No grade 4 or 5 treatment-related adverse events were observed. No dose-limiting toxicities were observed, and 14 patients discontinued treatment due to AEs. Skin rash and pruritis were the most common treatment-related AEs, consistent with mechanism of action of Treg depletion. Confirmed partial responses were observed across multiple tumor types and seen at doses as low as 10 mg. Responders and non-responders exhibited a wide range of CCR8 expression, and some responders had undetectable levels of CCR8. The objective response rate was 8% (n=4), which 4 confirmed partial responses, and disease control rate was 46% (n=24). The median duration of response was 4.3 months, and some responses lasted as long as 14.6 months. Paired biopsies from Part B of the study indicated that denikitug was associated with a reduction in intratumoral CCR8-expressing Tregs and an increase in intratumoral Ki67+ CD8+ effector T cells. Similar changes were observed in peripheral blood. With its manageable safety profile and promising anti-tumor activity in heavily pretreated patients, these results support further investigation of denikitug as monotherapy and in combination with immune checkpoint inhibitors.
Large-scale multi-omic analyses to identify mechanisms of resistance to immunotherapy in non-small cell lung cancer
CT233. Multi-modal multi-omic analyses reveal mechanisms of immunotherapy resistance in non-small cell lung cancer (NSCLC)
Archana Balan (The Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University School of Medicine, Baltimore, MD, USA) reported results from multi-omic analyses to identify mechanisms of resistance to immunotherapy. Exploratory analyses including tumor genomic profiling, circulating tumor DNA analysis, transcriptomic profiling, and single cell RNA-sequencing were performed on biospecimens from 951 patients from the HUDSON study, the largest immunotherapy resistance cohort to date. HUDSON was a phase 2 biomarker-directed study of combination regimens in patients with non-small cell lung cancer that progressed on PD-(L)1 inhibitors. Post-immunotherapy tumors in the acquired resistance group exhibited a higher tumor mutation burden that tumors in the primary resistance group, but circulating tumor DNA levels at resistance were lower in the acquired resistance group than the primary resistance group. Tumors from both the primary and acquired resistance groups showed increased levels of genomic instability, such as large-scale transitions, telomeric allele imbalances, and homologous recombination deficiencies post-immunotherapy, compared to pre-immunotherapy tumors. Surveying the genomic landscape of resistant tumors revealed that genomic alterations that drive resistance span multiple pathways and cancer hallmarks. Approximately 40% of patients with either primary or acquired resistance exhibited alterations in STK11 and/or KEAP1, known genomic features of resistance to immunotherapy. Post-immunotherapy tumors showed increased frequencies of mutations that deregulated cell cycle, DNA damage repair, chromatin maintenance, proliferative signaling, and de-differentiation. Differential enrichment in driver mutations in acquired and primary resistance demonstrated these processes are driven by distinct molecular programs. Transcriptomic analyses indicated that in post-immunotherapy tumors, tumors in the acquired resistance groups showed upregulation of gene expression programs involving the epithelial-to-mesenchymal-transition (EMT), IFN-gamma, TNF-alpha signaling, and conserved inflammatory responses compared to the primary resistance group. Post-immunotherapy tumors from the acquired resistance group also showed altered gene expression profiles that shortened overall survival, including changes in metabolic programming, the IFN gamma response, and EMT. Single cell transcriptomics of acquired resistance tumors showed enrichment of tissue-resident CD8+ T cells, suggesting a partially functional anti-tumor response, as well as enrichment of stem-like precursor exhausted T cells, indicating immune dysfunction. Acquired resistance was also associated with enrichment of cells in pleuripotent and EMT states, indicative of cellular reprogramming, and downregulation of antigen presentation. These data indicate that primary and acquired resistance to immunotherapy is dynamic and multi-factorial, involving genomic and transcriptomic alterations, lineage plasticity, and metabolic and immunological reprogramming. While further characterizing these changes will likely identify potential targets for therapeutic intervention, interventions or strategies to curb immunotherapy resistance will likely be highly personalized and involve a combination of approaches.
Tertiary lymphoid structure-associated immune niches may identify patients with HER2-enriched early breast cancer who are most likely to respond to chemotherapy-free treatment strategies
6741. Tertiary lymphoid structure (TLS)-associated immune circuits define response to durvalumab, trastuzumab, and pertuzumab (DTP) in HER2-enriched early breast cancer
Mailin Li (Houston Methodist Research Institute, Houston, TX, USA) presented a study to define the cellular mechanisms underlying response and resistance to a chemotherapy-free approach to the treatment of HER2-enriched early breast cancer. In a phase 2 clinical trial, 37 patients with ER-negative, PR-negative, HER2-enriched breast cancer received neoadjuvant treatment with durvalumab, trastuzumab, and pertuzumab (DTP). Response was assessed after 6 cycles of DTP; responders (n=31) underwent surgery, and non-responders (n=6) received standard of care chemotherapy before surgery. 49% of patients (n=18) who received DTP achieved a pathological complete response, indicating chemotherapy was not necessary in a subset of patients. Pre-treatment samples from patients who responded to DTP and did not receive chemotherapy were analyzed by single-nucleus RNA sequencing, Xenium spatial transcriptomics, and multiplex IHC to determine which cell types were associated with response to DTP and to map their spatial organization. Pre-treatment samples from patients who experienced a pCR were found to be significantly enriched for T follicular helper (Tfh)-like CD4+ T cells, mature dendritic cells, and cycling (Ki67+) B cells. Hierarchical clustering indicated these populations were part of an immune module that correlated with pCR to DTP. Because these cells were part of the same immune module, it was hypothesized that they were part of tertiary lymphoid structures (TLSs). Spatial mapping indicated that the TFH-like CD4 T cells, dendritic cells, and B cells colocalized with each other as well as other CD4 T cells, CD8 T cells, plasma cells, reticular fibroblasts, and lymphatic endothelial cells, which reflects the spatial organization of tertiary lymphod structures (TLS). To quantify these response-associated interactions, a niche was defined as a co-occurrence of a CD4+ T cell, a B cell and a dendritic cell within a 20 unit radius. Tumor samples from patients who achieved a pCR had more tumor-infiltrating lymphocytes and immune niches than patients who did not respond to DTP. CellChat analyses indicate that Tfh CD4+ T cells support B cell maturation. Results from this study indicate chemotherapy is not necessary for some patients with HER2-enriched early breast cancer to achieve a complete response. Tertiary lymphoid structure-associated niches, that include Tfh CD4+ T cells, B cells, are enriched in tumors from patients who achieve a pathologic complete response, thus potentially acting as biomarkers to identify patients who are candidates for chemotherapy-free treatment strategies.
Immune profiling of patients receiving high-fiber diets or healthy control diets in parallel with immune checkpoint blockade for melanoma
6743. Peripheral immune profiling from the DIET trial - A randomized double blinded dietary intervention study in melanoma patients receiving immunotherapy
Carrie R. Daniel-McDougall (University of Texas MD Anderson Cancer Center, Houston, TX, USA) presented immune profiling from the phase 2 Diet and Immune Effects Trial (DIET), a randomized double-blind dietary intervention protocol comparing the effects of a high fiber diet compared to a healthy control diet in patients receiving standard-of-care immune checkpoint blockade for melanoma. The DIET trial is based on prior results that dietary fiber influences the gut microbiome and improves progression-free survival for patients whoimmunotherapy for melanoma. Three cohorts of patients receiving immune checkpoint inhibitors as standard-of-care for melanoma were randomized to receive a high fiber diet (n=28; 30 to 50 g/day) or a control healthy diet (n=15; 20 g/day). Patients in the high-fiber arm experienced numerically greater gastrointestinal symptoms over the 10 weeks of the diet compared to the control arm, but symptoms decreased for both arms over time. Shifts in either alpha or beta diversity were not observed between the control or high fiber diet cohorts, however compositional shifts in response-associated taxa were observed. Patients in the high fiber arm exhibited a decrease in secondary-to-primary bile acid ratios, a shift that is associated with anti-tumor immunity in preclinical models. Clinical outcomes were evaluated in unresectable and neoadjuvant cohorts. The response rates in the unresectable and neoadjuvant cohorts were 77% (10/13) for the high fiber arm and 29% (2/7) for the control arm. Updated survival data indicate that event free survival and recurrence-free survival are improved in the high fiber arm compared to the control arm. Both arms had comparable safety profiles, but the high fiber arm exhibited a significantly lower rate of cutaneous toxicities (pruritis and rash) compared to the control arm. Peripheral blood from patients in the control arm exhibited increased levels of immunosuppressive classical monocytes compared to patients in the high fiber arm, and similar results were observed in non-responders vs. responders. Digital spatial profiling indicated that levels of CD8+ T cells and cytotoxic lymphocytes in the tumor microenvironment increased from the initiation of intervention to the end of intervention in complete responders compared to non-responders, but the increases observed in responders were more dramatic in the high fiber arm compared to the control arm. These results indicate that a high fiber diet delivered in parallel with standard of care immune checkpoint blockade is safe and is associated with favorable shifts in gut and host metabolism and in the anti-tumor immune response and comparable safety profiles. More studies are underway with humans and with preclinical models to interrogate the mechanisms underlying these shifts and their effects on response to immunotherapy. Trials of prebiotic food-enriched diets of multiple daily snacks for patients receiving immune checkpoint inhibitors for melanoma in the neoadjuvant or metastatic settings are also underway.
Multi-omics profiling and longitudinal clinical studies identify potential biomarkers and mechanisms underlying immune-related adverse events in the central nervous system
6744. Persistent neuro-antigen-reactive Th17 cells drive IL-17-dependent neurotoxicity in CNS immune-related adverse events
Jingyao Zhang (Danbury Hospital, Danbury, CT, USA) reported results from a study of immune-related adverse events (irAEs) of the central nervous system (CNS). irAEs of the CNS (irAE-CNS) are rare but life-threatening, poorly characterized, and lack effective treatment. The goal of the study was to identify a disease-specific CNS immune program that explains the severity of irAE-CNS and reveals a targetable pathway. The study consisted of sequential sampling of 4 independent longitudinal cohorts. Proteomic analyses of cerebrospinal fluid screened 406 inflammatory proteins and identified 53 proteins that were enriched in patients with irAE-CNS compared to control groups. CD137, IL-17A, MCP1, MCP4, CCL11, and GZMB were among the top hits. Single-cell profiling identified a distinct polyfunctional cell cluster of CD3+ CD4+ PD-1+ helper T cells (Poly6+ CD4+ Th cells) that was enriched in patients with irAE-CNS. Poly6+ CD4+ Th cells secrete IL-17A, MCP1, MCP1, CCL11, GZMB, and CD137. Gene ontology studies indicated that the Poly6+ program was strongly associated with IL-17 biology. Increased levels of Poly6+ CD4+ Th cells in patients correlated with higher neuroinflammation and worse longitudinal cognitive and neurologic outcomes, suggesting that Poly6+ Th cells may act as biomarkers of irAE-CNS severity and prognosis. Single cell analyses indicated that CD137+ Poly6+ cells exhibited phenotypic stability, persistent clonal expansion, and strong neuroantigen reactivity. NMDAR1 was identified as the primary target of reactive Poly6+ CD4+ Th cells, and CD137+ NMDAR1 reactive T cells were associated with increased neuroinflammation and functional decline. To determine whether blockade of downstream signals associated with Poly6+ CD4+ Th cells, patients with irAE-CNS were treated with IL-17A inhibitors (n=17) and compared to patients who received the standard of care for irAE-CNS (n=33). IL-17A blockade significantly reduced neuroinflammation across multiple brain regions, and neurological function improved significantly with IL-17A inhibitors. IL-17A blockade did not increase cancer-related or overall mortality. These results have identified NMDAR1-reactive Poly6+ CD4+ Th cells as potential biomarkers of irAE-CNS. IL-17, which is secreted by Poly6+ CD4+ Th cells, drives irAE-CNS and is targetable for managing and preventing irAE-CNS. Although larger studies are needed to validate antigen specificity and test IL-17A blockade in a randomized trial, this study provides a framework for identifying biomarkers of irAE-CNS and managing them through non-steroidal approaches.