Introduction
We have engineered a novel adoptive T cell therapy platform by integrating a T cell receptor fusion construct (TRuC
®) into the T cell receptor (TCR). This leverages the full signaling capacity of the TCR in a human leukocyte antigen (HLA)-independent manner and addresses the limitations of chimeric antigen receptor engineered T (CAR-T) cells and TCR-engineered T (TCR-T) approaches. TRuC-T cells consist of a tumor antigen binding domain fused to the CD3ε subunit of the TCR complex, which upon its integration into the TCR redirects T cell killing against tumor cells. This novel design has shown functional advantages over CAR-T cells in preclinical models, including faster tumor regression, lower cytokine production, increased solid tumor infiltration, increased oxidative metabolism, and enhanced persistence [
1]. Based on promising preclinical evidence, a Phase 1/2 clinical trial examining TC-210 (
gavocabtagene autoleucel [gavo-cel]) is ongoing in patients with advanced mesothelin (MSLN)-expressing cancer (NCT03907852).
The tumor microenvironment (TME) of solid tumors presents a major hurdle in realizing the full potential of T cell therapies, as immunoinhibitory molecules are often abundant, and positive costimulatory molecules are lacking [
2]. Overexpression of programmed cell death ligand (PD-L)1 and PD-L2 on tumor cells directly inhibits T cell function by activating the programmed cell death protein 1 (PD-1) [
3].
Furthermore, full T cell activation requires TCR recognition of cognate peptide major histocompatibility complexes (MHCs) (signal 1) in conjunction with costimulation, driven most prominently by activation of CD28 (signal 2) [
4]. A lack of sufficient costimulatory signaling leads TCR-activated T cells to enter a hyporesponsive state known as anergy [
5].
Suppression by the PD-1/PD-L1 axis-mediated suppression within the TME and the lack of intrinsic CD28 signaling afforded by the TRuC construct may present hurdles to optimal TRuC-T cell efficacy. Thus, we engineered and preclinically tested chimeric switch receptors (CSRs) designed to co-opt the immunosuppressive PD-1/PD-L1 axis and, at the same time, deliver a CD28-mediated costimulatory signal. CSRs are designed to convert a normally immunoinhibitory interaction into an immunostimulatory event by genetically linking the extracellular domain of a suppressive receptor (in this case PD-1) to the signaling domain of an activating receptor (in this case CD28).
We benchmarked the activity of anti-MSLN TRuC-T cells coexpressing CSRs to T cells bearing only the anti-MSLN TRuC (TC-210 T cells). As it has been established that the transmembrane (TM) domain can influence the functionality of a chimeric receptor [
6], we compared anti-MSLN TRuC-T cells engineered to coexpress PD1-CD28 CSRs containing either PD1TM or CD28TM. The resulting lead construct, PD1TM CSR, was integrated into a novel anti-MSLN cell therapy product designated TC-510. The efficacy of TC-510 and TC-210 were compared using an in vitro stimulation assay. Durable protection against tumor rechallenge in vivo was also assessed.
Materials and methods
T cell engineering
MSLN-targeting ɛ-TRuC was generated as described [
1]. A PD1TM CSR was generated by isothermal assembly of the ecto- and TM domains of PD-1 (Q15116 amino acids 1–191) to the intracellular domain of CD28 (P10747 amino acids 180–220). Similarly, a CD28TM CSR was generated by isothermal assembly of the ectodomain of PD-1 (Q15116 amino acids 1–170) to the TM and intracellular domains of CD28 (P10747 amino acid 153–220). MSLN-targeting ɛ-TRuC and the CSR were cloned on the same lentivirus expression vector upstream and downstream of a T2A sequence, respectively.
Lentiviruses were prepared by transient transfection of HEK293 suspension cells with packaging plasmids and the TRuC or CAR lentiviral transfer plasmids. Supernatants were collected 48 h post-transfection, centrifuged, filtered, and precipitated. Clarified supernatants were resuspended in TexMACS medium (Miltenyi Biotech, Berisch Gladbacj, Germany) supplemented with 3% human antibody serum (Gemini Bio-Products, West Sacramento, CA) and stored at –80°C until use.
On Day 0, primary human T cells were isolated by magnetic bead separation using anti-CD4 and anti-CD8 microbeads. T cells were activated using Human T Cell TransAct (Miltenyi Biotech) at a 1:1 ratio and cultured in TexMACS medium with 3% human antibody serum (Gemini Bio-Products), 12.5 ng/mL human IL-7, and 12.5 ng/mL human IL-15 (Miltenyi Biotech). T cells were transduced with the respective lentiviral vectors on Day 1, harvested on Day 10, and frozen prior to use in functional assays.
Cell lines
Tumor cell lines were purchased from ATCC (mesothelioma [MSTO]-211H [CRL-2081™]; Mannasas, VA) or Millipore Sigma (A2780 [C30]; St Louis, MO). For the generation of target cell lines, full-length firefly luciferase (Luc) or the PD-L1 ecto- and TM domains were cloned into pCDH-CMV-MCS-EF1a-Neo. Full-length human MSLN was cloned into pCDH- pCDH-EF1a-MCS-T2A-Puro (SBI, Palo Alto, CA), using XbaI and EcoRI restriction sites. Stably transduced cells were selected with neomycin (Millipore Sigma) and/or puromycin (Corning, Bedford, MA).
Flow cytometry analysis
The transduction efficiency, in vitro expansion, activation/exhaustion, and proliferation of engineered T cells were analyzed by flow cytometric analysis. Cells were stained using fluorescently labeled antibody cocktails, and data were acquired on the BD LSR Fortessa™ X-20 cell analyzer. Data analysis was performed using FlowJo software (TreeStar Inc, Ashland, OR). Detailed methods are provided in the supplemental material.
Luciferase activity-based tumor cell cytotoxicity assay
Luciferase-expressing tumor cells were plated in triplicate in a 96-well plate at 1.0 × 104 cells per well, and T cells were added at the desired effector-to-target (E-to-T) ratios. After 24-h coculture, 50% of the culture supernatant was removed for cytokine analysis. Cell viability was determined using the Bright-Glo™ Luciferase Assay System (Promega, Madison, WI) according to the manufacturer’s protocol. Relative luminescence units (RLU) were measured using the SpectraMax M5 plate reader (Molecular Devices, Sunnyvale, CA). The percentage of tumor cell killing was calculated by the following formula: % tumor cell lysis = 100% × [(1 − RLU (tumor cells + T cells)/RLU (tumor cells)].
Coculture assays
For TRuC-T cell coculture assays with target cell lines, TRuC-T cells were first thawed and rested in IL-2 (300 U/mL) for 72 h. At the end of the rest period, TRuC-T cells were then normalized for transduction efficiency and then plated in a 96-well U-bottom plate at a 1:1 ratio with 1.0 × 105 Streck-treated tumor cells (Streck, La Vista, NE) for up to 96 h. Culture supernatants were harvested from replicate plates at 24 or 72 h and stored at –80 °C until sample analysis. Detailed methods, including rechallenge assay conditions, are provided in the supplemental material.
Plate-bound MSLN and PD-L1 assay
TRuC-T cells were recovered from cryopreservation by incubation in IL-2 (300 IU/mL) for 72 h. MSLN- and PD-L1-coated 96-well ELISA microplates were prepared by treatment for 24 h with PD-L1-Fc alone (2 mg/mL) or MSLN (1 mg/mL) with varying concentrations of PD-L1-Fc (0–10 mg/mL), washed, and stored semi-dry prior to use. Recovered TRuC-T cells were normalized for transduction efficiency, and incubated at 1 × 105 TRuC-T cells/well in coated-plates for 72 h. The resulting levels of IL-2, IFN-γ, TNF-α, and GM-CSF were measured using a Meso-Scale Discovery gold kit (Mesoscale Diagnostics, Rockville, MD) per the manufacturer’s instructions.
In vivo efficacy of engineered T cells
For the subcutaneous xenograft model, 1.0 × 106 MSTO-MSLN-PD-L1-Luc cells were resuspended in sterile PBS, mixed 1:1 with ice cold Matrigel® (Corning, Tewksbury, MA), and then injected subcutaneously in the dorsal hind flank of 7–8-week-old female class I/class II negative NOD scid gamma (NSG) mice (NOD.Cg-Prkdcscid H-2K1tm1BpeH2-Ab1em1MvwH2D1tm1BpeIl2rgtm1Wjl/SzJ) from the Jackson Laboratory (Bar Harbor, ME). Mice were randomized into treatment groups by tumor burden prior to injection of human T cells; n = 10 mice per group. Engineered human T cells were administered at a dose of 2.0 × 106 TRuC+ T cells per mouse, via tail vein injection when the tumor size was 150–200 mm3 (Day 0). Tumor growth was monitored as tumor volume by caliper measurement twice weekly. The volume of tumor was calculated as: tumor volume = (length × width2)/2. For tumor rechallenge in mice that had become tumor-free, 1.0 × 106 MSTO-MSLN-PD-L1-Luc cells were prepared as described above and injected subcutaneously in the opposing flank on Day 44. Data represent two independent experiments with two T cell donors.
Discussion
As previously shown, TRuC-T cells differ from CAR-T cells by enabling faster tumor regression, lower cytokine production, increased tumor infiltration, and a shift toward oxidative metabolism [
1]. However, like native T cells, TRuC-T cells remain susceptible to inhibition by the PD-1/PD-L1 axis [
8,
9]. Furthermore, whereas engineered costimulatory signals are not required for the in vivo efficacy of TRuC-T cells, in contrast to CAR-T cells [
10], we hypothesized that the delivery of costimulation in conjunction with TRuC activation would enhance T cell function and persistence. In the present study, we describe a PD1-CD28 CSR that co-opts tumor PD-L1 expression and, at the same time, drives CD28 costimulation. The potential of PD1-CD28 CSRs to improve the function of engineered T cells has previously been established in preclinical models [
11‐
15] and in an early clinical trial [
16].
Both blocking PD-1 and enhancing CD28 signaling are attractive strategies for driving tumor-targeted T cell activity and persistence in the TME. Indeed, the approved anti-PD-1 drugs have significantly changed the standard-of-care treatment in multiple cancers, improving overall survival, progression-free survival and durability of response [
17,
18]. PD-1 inhibition of T cell signaling is mediated by recruitment of SHP phosphatases that deactivate TCR signaling by targeting key kinases of the TCR and CD28 signaling pathways [
19,
20], and CD28 itself is a target of PD-1-activated SHP phosphatases [
21]. Regardless of PD-1-mediated inhibition, activation of CD28 by ligation to B7.1 and B7.2 on antigen-presenting cells provides a costimulatory signal that can regulate and augment endogenous TCR signaling [
22‐
24]. CD28 phosphorylation potentiates T cell activation, leading to enhanced proliferation, effector function, and notably induction of proinflammatory cytokines, such as IL-2 [
25‐
27].
The selection of the TM domain utilized in a chimeric receptor can have a profound effect on its function. To our knowledge, we were first to functionally compare PD1-CD28 CSRs utilizing either a PD1TM or CD28TM. Although both versions of the CSR provided a CD28 signaling enhancement to TRuC-T cells, there were potentially important functional differences between them. The CD28TM failed to discriminate between low and high PD-L1 expression density on target cells and demonstrated an exaggerated response to low levels of PD-L1 in a plate-bound assay. Furthermore, the CD28TM consistently produced more cytokines against MSTO cells that express low levels of both MLSN and PD-L1, suggesting that the CD28TM increases TRuC antigen sensitivity when PD-L1 levels are low, but not when PD-L1 levels are high. A signaling imbalance between the TRuC and CSR in the low antigen/high PD-L1 setting may explain this observation. In contrast, the PD1TM demonstrated a greater dynamic response to PD-L1 density as it exhibited lower activation compared with CD28TM when PD-L1 density was low, but rivaled or surpassed the CD28TM when the PD-L1 density was high. Moreover, the PD1TM also maintained better T cell viability at high PD-L1 expression densities in a plate-bound assay. The stochastic sensitivity of the CD28TM to PD-L1 levels may be explained by CSR homodimerization and/or heterodimerization with endogenous CD28, resulting in amplified signaling [
28]. We selected the PD1TM version of the CSR for its greater sensitivity to regulation by PD-L1 levels, which we believe reduces the risk of cytokine release syndrome and on-target/off-tumor toxicity, and integrated this with the anti-MSLN TRuC utilized in TC-210 to create a second-generation TRuC-T cell product that we call TC-510.
Like TC-210, cytotoxicity and cytokine release in TC-510 cells are dependent upon TRuC engagement by MSLN. This indicates that the costimulatory activity of the PD1-CD28 CSRs adheres to the two-step model of T cell activation [
29], which is an important feature in the context of off-tumor toxicity risk.
The PD1-CD28 CSR contained in TC-510 offers two potential mechanisms for TRuC-T cell enhancement: (1) acting as a PD-1 dominant-negative receptor (DNR); and (2) delivering a costimulatory signal upon PD-L1 engagement. To elucidate the relative contributions of these two mechanisms, we compared TC-510 with T cells bearing the same TRuC but with the PD1TM CSR replaced by either a truncated PD-1 lacking an intracellular signaling domain (PD1
Trunc) or a PD1TM CSR in which CD28 signaling was mutationally inactivated (PD1TM
Mutant). TRuC-T cells coexpressing either of these constructs failed to enhance the activity of TC-510; rather, they performed comparably to TC-210. These findings support the assertion that CD28 costimulatory signaling is the primary mechanism by which the PD1TM CSR enhances TC-510 effector function. The lack of TRuC-T cell functional enhancement by the PD-1 DNR seems unexpected but is consistent with prior observations [
30]; however, others have reported CAR-T cell enhancement by a PD-1 DNR [
31]. This may reflect limitations in the ability of our assays to detect PD-1-mediated suppression or an intrinsic resistance of TRuC-T cells to PD-1-mediated suppression.
Our observation of enhanced expansion and persistence of TC-510 upon serial tumor rechallenge in vitro was further supported by an in vivo study in which TC-510 showed a superior ability to durably protect mice from tumor rechallenge compared with TC-210. This result indicates that integrating a PD1-CD28 CSR into TRuC-T cells enhances their capacity for improved persistence both in vitro and in vivo, which could translate into improved clinical efficacy in cancer patients with solid tumors. Understanding the potential impact of increased activation and persistence mediated by the CSR on safety will be an important focus of planned clinical studies.
Potential limitations of these preclinical studies include the artificial, nonphysiological nature of the in vitro assays used to assess T cell functionality and the use of a xenograft mouse model lacking both an intact immune system and expression of human PD-L1 or MSLN on normal tissues.
Based on these promising preclinical findings, TC-510 is currently being evaluated in a Phase 1/2 clinical trial in patients with advanced MSLN-expressing solid tumors (NCT05451849).
Acknowledgements
The authors would like to thank Seema Shah for technical assistance in the performance of in vivo studies. This study was sponsored by TCR
2 Therapeutics Inc. Medical writing support, under the guidance of the authors, was provided by Macarena Ramos Gonzalez, PhD, CMC Connect, a division of IPG Health Medical Communications, in accordance with Good Publication Practice (GPP) guidelines [
32]. SK is supported by the Marie-Sklodowska-Curie Program Training Network for Optimizing Adoptive T Cell Therapy of Cancer funded by the H2020 Program of the European Union (Grant 955575); by the Hector Foundation; by the International Doctoral Program i-Target: Immunotargeting of Cancer funded by the Elite Network of Bavaria; by Melanoma Research Alliance Grants 409510; by the Else Kröner-Fresenius-Stiftung; by the German Cancer Aid (to SK); by the Ernst-Jung-Stiftung; by the LMU Munich’s Institutional Strategy LMUexcellent within the framework of the German Excellence Initiative by the Go-Bio initiative; by the m4 Award of the Bavarian Ministry of Economical Affairs, by the Bundesministerium für Bildung und Forschung; by the European Research Council Grant 756017, ARMOR-T, and the ERC proof-of-concept Grant 101100460; by the German Research Foundation (DFG) (KO5055-2-1 and 510821390) by the SFB-TRR 338/1 2021-452881907; by the Wilhelm-Sander-Stiftung, by the Fritz-Bender Foundation; and by the Deutsche José-Carreras Leukämie Stiftung.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.