GITR Antibodies in Cancer: Not Ready for Prime Time

In this issue of Clinical Cancer Research, Davar and colleagues (1) present the results of the phase Ib clinical study for TRX-518, a fully humanized Fc-dysfunctional aglycosylated IgG1 mAb directed to the TNF receptor (TNFR) family member glucocorticoid-induced TNFR-related protein (GITR).

GITR is a type I transmembrane glycoprotein mainly expressed in naive and activated T- and natural killer (NK) cells, and more specifically in CD4⁺CD25⁺ regulatory T cells (Tregs; refs. 2, 3). GITR agonistic antibodies increase T-cell activation allowing for secondary Tregs expansion. Once T-cell receptor-mediated activation occurs, GITR is upregulated by responder T cells (CD4⁺CD25^– T cells or CD8⁺CD25^– T cells). GITR interaction with its ligand (GITRL) on dendritic cells (DC) leads to an increase in IL2 production, activating T-cell differentiation into effector T cells (Fig. 1). Once the response is established, CD4⁺CD25⁺ Treg cells are expanded by their binding to GITRL in DCs, achieving a final suppression of T-cell activity (4–6).

In the past years, GITR has become an interesting target as potential enhancer of classical immunotherapeutic agents (5). The first mAb developed for GITR targeting was DTA-1, a mouse IgG2b mAb with agonistic effect over GITR (7). Several targeted agents have followed, some antibodies with specific structural variations for improvement and directing of human GITR, and others based on recombinant fusion proteins that mimic GITRL.

TRX-518 is an IgG1 aglycosylated mAb that showed promise in preclinical testing, achieving effective targeting of the GITR receptor and enhancing NK cell activity and cytotoxicity. In vivo, TRX-518 was able to bind to GITR on CD4⁺ and CD8⁺ naïve and memory T cells, B cells, NK cells (leading to their activation), invariant NK T cells, monocytes, and macrophages, without depleting them (8). Recently, the IgG form has proven to be required for GITR activation, and not the Fab form of the compound (9).

In the first-in-human (FIH) trial (NCT-01239134) TRX518 monotherapy effectively depleted Tregs in the periphery, with an apparent direct correlation with Treg depletion in tumor tissue. Peripheral CD4⁺ and CD8⁺ effector T cells and NK cells were not activated, and no clinical responses were seen (8).

The present paper by Davar and colleagues, shows the results of TRX-518 in combination with chemotherapy and anti-PD1 therapy. The study enrolled 109 patients in parts A (monotherapy dose escalation) and B (monotherapy expansion); 30 patients in part C (combination with gemcitabine); 26 in part D (combination with pembrolizumab); and 10 in part E (combination with nivolumab). Parts A and B were composed of mainly patients with microsatellite stable colorectal carcinoma (28%), melanoma (12%), and breast cancer (9%), among others. Toxicity profile was favorable and no dose-limiting toxicities (DLT) were observed. Clinical activity was low. The median duration on treatment was 11.4 weeks (range: 0.1–97.9 weeks), with one partial response (PR) in a patient with hepatocellular carcinoma that had progressed on ipilimumab/nivolumab, and 2 patients (appendiceal carcinoma and prostate adenocarcinoma) had long lasting disease stabilization (>6 months). The overall response rate was 3.2% in total for the monotherapy arms (1/31 evaluable patients).

Chemotherapy combination, which mainly included patients with pancreatic adenocarcinoma and biliary tract carcinoma (half of which had progressed on gemcitabine) did not exhibit better clinical activity. The median duration on treatment was 7.0 weeks (range: 0.1–61.1 weeks). Objective response rate was 3.8% (1/26 evaluable): 1 patient with pancreatic adenocarcinoma had a partial response, and 6 patients achieved stable disease (SD) as best response lasting more than 6 months (including one appendiceal carcinoma, one biliary tract carcinoma, and 2 patients with pancreatic adenocarcinoma).

In the anti–PD-1 combination arms, clinical activity was also discouraging, with only one PR of 25 evaluable patients in the pembrolizumab combination cohort, an esophageal squamous cell carcinoma PD1-naïve; and another PR in 1 of 8 evaluable patients in the nivolumab cohort, a PD1-refractory urothelial carcinoma who achieved a PR lasting more than 162 days. In cohort D with pembrolizumab, a long-lasting disease stabilization was observed on a patient with choroidal melanoma progressing to prior anti–cytotoxic T-lymphocyte antigen-4⁺ tebentafusp.

From the pharmacodynamic perspective, as previously observed in the single agent phase I trial where a dose-dependent depletion of Tregs peripherally was achieved with little or no translation on GITR CD8⁺ T cells and GITR CD4⁺ Foxp3⁻ T effector cells, the expansion phase for monotherapy arms (parts A + B on present study) confirmed similar results. The patient achieving PR showed a significant and more profound increase in CD8⁺ tumor-infiltrating lymphocytes (TIL) than those patients with SD, and with progressive disease. Chemotherapy combination showed entirely different kinetics. Gemcitabine administration on days 1 and 8 allowed for depletion of peripheral Tregs and effector Tregs, whilst TRX-518 administered on day 2 would increase Treg expansion, resulting in a counterintuitive immune response, with no signs of clinical activity.

The combination arms with anti-PD1, showed that a significant depletion of Tregs together with a high CD8⁺ TIL infiltration is required for clinical activity. 2 patients achieving response had a significant depletion of tumor Tregs, in comparison with nonresponders, all of which had Treg increase in tumor. Question arises on whether the latter would have been previously PD-1–treated patients. Unfortunately, little clinical activity is seen in any of these cohorts to draw meaningful conclusions, and no improvement on prior experience with other GITR agonists has been demonstrated in this paper.

Similar results were seen for another GITR agonist, MK-4166 (10). The FIH trial was not able to reach a maximum tolerated dose, and no responses were seen in monotherapy. Others such as BMS-986156 and AMG228, both agonistic human IgG1 GITR-mAbs showed no signs of clinical activity and were not developed further. INCAGN1876 (ragifilimab), also showed preclinically to effectively activate the GITR pathway in recently primed T cells, but no clinical activity was seen in the phase I dose escalation trial (11). Further variations on GITR-targeted mAbs focused on developing structural changes like tetravalent properties such as ASP-1951 (PTZ-522), with no success. GWN-323, an anti- GITR IgG1 mAb also showed little activity as single agent (12). Currently, a trial with another GITR agonistic antibody, REGN6569 is undergoing and pending results (NCT-04465487).

MEDI-1873 is a hexameric GITR-ligand/IgG1 agonistic fusion protein (GITRL FP). The FIH trial (NCT02583165) resulted to be apparently more toxic than mAbs. Grade 3 treatment-related adverse events were reported in 22.5% of patients. MEDI-1873 increased CD4⁺Ki67⁺ T cells and induced a significant decrease in GITR⁺/FoxP3⁺ T cells in the evaluable patients peripherally, as well as GITR⁺ and FoxP3⁺ TILs in tumor samples, confirming the proof of concept of the compound, but again, very little clinical activity was seen. The best response achieved was disease stabilization (42.5%; ref. 13).

In summary, the current paper by Davar and colleagues (1) provides more insight into the value of GITR agonism to modify the immune response. There is clear translational evidence that GITR agonists induce co-stimulation of T cells and dampen Tregs suppressive functions. However, to date, the clinical results of GITR targeting agents seem to be insufficient to induce meaningful clinical responses, and little is known about further potential combinatorial approaches that could enhance the efficacy of these compounds. Additional translational work will be needed to translate the potential benefit of GITR agonists into the clinic.

T. Hernandez-Guerrero reports institutional funding from AbbVie, AceaBio, Adaptimmune, ADC Therapeutics, Aduro, Agenus, Amcure, Amgen, Astellas, AstraZeneca, Bayer HealthCare, BeiGene, BioInvent International AB, BMS, Boehringer Ingelheim, Boston, Celgene, Daichii Sankyo, DEBIOPHARM, Eisai, e-therapeutics, Exelisis, Forma Therapeutics, Genmab, GSK, Harpoon, Hutchison, Immutep, Incyte, Inovio, Iovance, Janssen, Kyowa Kirin, Eli Lilly and Company, Loxo, MedSir, Menarini, Merck, Merus, Millennium, MSD, Nanobiotix, Nektar, Novartis, Odonate Therapeutics, Pfizer, PharmaMar, Principia, PsiOxus, Puma, Regeneron, Rigontec, Roche, Sanofi-Aventis, Sierra Oncology, Synthon, Taiho, Takeda, Tesaro, Transgene, Turning Point Therapeutics, and Upshersmith. V. Moreno reports personal fees from BMS, Bayer HealthCare, Roche, Basilea, and Janssen outside the submitted work; in addition, V. Moreno has been a principal investigator/received institutional funding from AbbVie, AceaBio, Adaptimmune, ADC Therapeutics, Aduro, Agenus, Amcure, Amgen, Astellas, AstraZeneca, Bayer HealthCare, BeiGene, BioInvent International AB, BMS, Boehringer Ingelheim, Boston, Celgene, Daichii Sankyo, DEBIOPHARM, Eisai, e-therapeutics, Exelisis, Forma Therapeutics, Genmab, GSK, Harpoon, Hutchison, Immutep, Incyte, Inovio, Iovance, Janssen, Kyowa Kirin, Eli Lilly and Company, Loxo, MedSir, Menarini, Merck, Merus, Millennium, MSD, Nanobiotix, Nektar, Novartis, Odonate Therapeutics, Pfizer, PharmaMar, Principia, PsiOxus, Puma, Regeneron, Rigontec, Roche, Sanofi-Aventis, Sierra Oncology, Synthon, Taiho, Takeda, Tesaro, Transgene, Turning Point Therapeutics, and Upshersmith.