Sage Journals: Discover world-class research

Abstract

Immune cell-based therapies can induce potent antitumor effects but are also often associated with severe toxicities. We previously developed a PD-1-based small molecule-regulated reversible T cell activation switch to control the activity of cellular immunotherapy products. This chemically regulated and SH2-delivered-inhibitory tail (CRASH-IT) switch relies on the noncovalent interaction of switch SH2 domains with phosphorylated ITAM motifs in either chimeric antigen receptors or T cell receptors. After this interaction, the immunoreceptor tyrosine-based inhibition motif/switch motif (ITIM/ITSM) containing PD-1 domain present in the CRASH-IT switch induces robust inhibition of T cell signaling, and CRASH-IT-mediated suppression of T cell activity can be reversed by small molecule-induced switch proteolysis. With the aim to develop improved second-generation switch systems, we here analyze the possibility space of both the immune cell receptor docking and inhibitory signaling domains that allow control over T cell activity. Importantly, these analyses demonstrate that the inhibitory domains that most potently suppress antigen receptor signaling in primary human T cells are not derived from inhibitory receptors, such as PD-1 and BTLA, that are endogenously expressed in T cells, but include ITIM/ITSM containing inhibitory domains derived from receptors present in myeloid cells. In addition, we demonstrate that physical proximity of the inhibitory domain to the antigen receptor is crucial to efficiently suppress T cell activation, as only switch designs that employ SH2 domains directly interacting with ITAM motifs in antigen receptors efficiently and reversibly inhibit T cell functionality. These data demonstrate the flexible and interchangeable nature of immune cell signaling domains, and inform the design of a synthetic proximity-based switch system with a superior dynamic range.

Introduction

Adoptive T cell therapies, employing T cell products modified with either T cell receptor (TCR) or chimeric antigen receptor (CAR) genes, but also unmodified allogeneic T cell products in case of donor lymphocyte infusion (DLI), are potent therapeutic strategies used in cancer treatment.^1,2 In addition to their capacity to directly kill antigen expressing cells, and unlike small molecule- or antibody-based drugs, transferred T cells can engraft long-term and expand in vivo, and produce cytokines to induce a broadening of immune responses. The adverse effects that have been associated with T cell-based therapies reflect a combination of the functionalities of this drug class: Excessive activation of CAR-T cells can induce a self-reinforcing cascade of cytokine production involving elevated levels of IL-6, interferon gamma or granulocyte macrophage colony-stimulating factor, and the resulting cytokine release syndrome (CRS) is, among others, characterized by hypoxia, disseminated intravascular coagulation, end-organ dysfunction, and capillary leak.³ In patients with severe CRS, cytokine-driven activation of endothelial cells increases blood–brain-barrier permeability, thereby promoting neurotoxicity (NT).^4,5 While CRS and NT can be considered indirect toxicities associated with T cell therapies, transferred T cells can also directly attack healthy tissues, if such tissues show expression of the relevant antigen(s).^6
–8 For instance, in case of DLI after HLA-matched hematopoietic stem cell transplantation, such on-target off-tumor effects are due to recognition of minor histocompatibility antigens (MiHA) on host cells, frequently resulting in graft-versus-host-disease due to the broad expression pattern of most MiHAs.^9,10 The side effects of T cell-based therapies can range from mild to fatal, and management of toxicities is costly and sometimes interferes with therapeutic activity of the transferred cells.¹¹

To abrogate T cell activity, or to transiently halt T cell activity, a number of irreversible and reversible switch designs have been developed.^9,12
–14 Those strategies that allow reversible control of T cell activity all rely on modification of the antigen receptor, for instance by splitting the receptor into multiple parts, or by its fusion with regulatory domains.¹⁵ However, this design is not compatible with control of endogenous TCRs, and such regulatory systems cannot be “retrofitted” into existing CAR T cell therapies. To address these issues, we designed a chemically regulated and SH2-delivered-inhibitory tail (CRASH-IT) switch that does not require hard modification of the antigen receptor.¹⁶ This design allows CRASH-IT switches to be combined with T cells employing their endogenous TCRs, or T cells modified with transduced TCRs or CARs.

Inhibition of T cell activity by the first-generation CRASH-IT switch was achieved by the immunoreceptor tyrosine-based inhibition motif/switch motif (ITIM/ITSM) containing intracellular domain of the PD-1 receptor. Recruitment of this inhibitory domain to activated antigen receptors by the Zap70 SH2 domains present in the switch induces proximity of the SHP2 phosphatase, thereby suppressing T cell activation through dephosphorylation of the CD28 costimulatory receptor¹⁷ or direct inhibition of TCR signaling.¹⁸ Finally, the ability to control the level of T cell activity is provided by a small molecule-regulated protein stability control domain, such as the small molecule-assisted shutoff (SMASh) domain¹⁹ that can be regulated by asunaprevir, or the FKBP12^F36V [FKBP(V)] domain²⁰ that can be regulated by dTAG-13. Thus, CRASH-IT switches can be considered tripartite regulators of T cell activity, in which one domain controls vicinity to activated antigen receptors, one domain controls recruitment of phosphatase activity, and the final domain is used to influence the levels of this designed regulatory protein. This modular nature of the system makes it an attractive platform to dissect the rules for optimal control of T cell activity, and to test whether second-generation switch designs with improved properties can be developed.

Immune cell signaling pathways are frequently based on a common blueprint, in which the information of a ligand binding event is propagated by a cascade of phosphorylation and dephosphorylation events. Short recognition motifs (ITAM/ITIM/ITSM) are phosphorylated, followed by the highly specific recognition of these phospho-motifs by SH2 domains.²¹ ITAM containing activating receptors associate with SH2 domain containing tyrosine kinases such as Zap70, whereas ITIM/ITSM containing inhibitory receptors generally recruit SH2 domain containing phosphatases such as SHP1/2 to counteract ITAM signaling.

To design synthetic T cell activity regulators, the use of intracellular domains that are derived from proteins that are physiological inhibitors of TCR signaling forms a logical first choice. However, the fact that different immune receptors and immune receptor signaling domains can influence cell activity when introduced into other cell types^22
–24 argues for a level of interchangeability. Based on this potential interchangeability, and the abundance of inhibitory signaling and SH2 domain containing proteins in diverse immune cell types, we set out to investigate the possibility space of CRASH-IT-based T cell switch designs, by replacing both the ITAM docking domain and inhibitory signaling domain with structurally or functionally homologous domains.

Comparison of switch activity of designs employing different docking domains underscores the importance of physical proximity of the inhibitory fusion protein to the antigen receptor complex. Furthermore, by exploring the activity of 21 inhibitory domains, we reveal that ITIM/ITSM containing inhibitory domains that are derived from myeloid cell inhibitory receptors, such as SIGLEC11 and SIRPA, yield synthetic regulators of T cell activity with a significantly improved dynamic range, as compared with regulators based on T cell inhibitory receptors. Specifically, both the improved T cell inhibition in the absence of asunaprevir or dTAG-13 and the near full restoration of T cell function in the presence of these drugs identify SIGLEC11- and SIRPA-based switches as optimal systems to control T cell activity. Finally, the observation that some of the most efficient regulators of TCR signaling capacity are found in proteins expressed in different cell types underscores the interchangeable nature of protein modules in phospho-ITIM/ITSM/ITAM-based signaling networks.

Methods

Retroviral DNA constructs

Codon optimized DNA sequences were synthesized by IDT as gene fragments and cloned into the MP71 vector by Gibson assembly. MP71-Zap70 2xSH2-PD1tail-SMASh-IRES-EGFP, MP71-Zap70 2xSH2-SMASh-IRES-EGFP (no signaling domain), MP71-PD1tail-SMASh-IRES-EGFP (no docking domain), MP71-Zap70-PD1-FKBP12^F36V-IRES-EGFP, MP71 CD19 ScFv-CD28-CD3 ζ CAR-IRES-huEGFRt vectors,¹⁶ and the HLA class I-restricted CDK4 TCR (TCR 17)²⁵ have been described previously. To generate CRASH-IT variants encoding alternative docking domains, the Zap70 (2xSH2 domains) in MP71-Zap70 2xSH2-PD1tail-SMASh-IRES-EGFP were replaced with the codon optimized sequences listed in Supplementary Table S1. CRASH-IT variants encoding alternative inhibitory domains were created by replacing the PD-1 signaling domain in MP71-Zap70 2xSH2-PD1tail-SMASh-IRES-EGFP and MP71-Zap70-PD1-FKBP12^F36V-IRES-EGFP vectors with the codon optimized sequences listed in Supplementary Tables S2 and S3, respectively, or the variant sequences described in Figs. 1D and 4A. A CAR panel encoding various ITAM containing signaling domains was created using the signaling and structural domains described in Supplementary Table S4, using the same vector backbone and CD19 ScFv antibody sequence as used in the MP71 CD19 ScFv-CD28-CD3 ζ CAR-IRES-huEGFRt vector.

Figure 1.

CRASH-IT components required for efficient and reversible inhibition of T cell functions. (A, B) Schematic overview of T cell signaling and mechanism of action of CRASH-IT rheostat. After peptide-MHC ligation to the TCR complex, Lck phosphorylates ITAM motifs on the CD3zeta chain. Phospho-ITAMs on CD3zeta act as a docking platform for the SH2 domains of the Zap70 kinase. CD3zeta bound Zap70 phosphorylates downstream targets, such as the LAT adaptor protein that propagates TCR signaling by recruitment of other signaling molecules. (A) The Zap70 SH2 domains of the CRASH-IT switch compete with endogenous Zap70 for CD3zeta binding in switch equipped T cells. ITIM/ITSM motifs within the intracellular domain of PD-1 that forms part of the switch are phosphorylated by Lck. The phosphorylated PD-1 domain enables SHP-2 SH2 domains to dock in the vicinity of TCR complex. Phosphatase activity of PD-1-bound activated SHP-2 results in inhibition of T cell signaling. (B) In the presence of a small molecule regulator of the protein stability control domain [asunaprevir or dTAG-13 in case of SMASh- and FKBP(V)-based switch designs, respectively], the CRASH-IT switch is targeted for proteasomal degradation. As a result, T cell effector functions such as cytokine release and cytotoxicity are restored. (C) Schematic representation of CRASH-IT variants. (D–G) Primary human T cells modified with the HLA class I restricted CDK4 TCR plus vector control or either intact Zap70-PD-1-SMASh switch, CRASH-IT variants with indicated ITIM and/or ITSM mutations, or CRASH-IT variants lacking one or both Zap70 SH2 domains or the inhibitory PD-1 domain, were pretreated with 10 μM asunaprevir or DMSO control. Data depict intracellular IFNγ, IL2, TNFα, and cell surface LAMP1 expression of CDK4 TCR⁺, EGFP high, CD8⁺ T cells upon coculture with NKIRTIL006 melanoma cells in the continued presence or absence of asunaprevir. Error bars represent standard deviation from three biological replicates (n = 3). Data are representative of two independent experiments. CRASH-IT, chemically regulated and SH2-delivered-inhibitory tail; SMASh, small molecule-assisted shutoff; TCR, T cell receptor. Color images are available online.

Cell lines and cell culture

FLYRD18, NKIRTIL006²⁶ and Daudi cells were cultured in IMDM (Thermo Fisher Scientific), supplemented with 8% FCS (Thermo Fisher Scientific) and penicillin–streptomycin (100 IU/mL penicillin, 100 μg/mL streptomycin; Sigma-Aldrich). FLYRD18 and NKIRTIL006 cells were passaged every 2–3 days with trypsin-EDTA (Thermo Fisher Scientific). The human NK cell line KHYG-1 (DSMZ) was cultured in RPMI supplemented with 8% FBS and penicillin–streptomycin (100 IU/mL penicillin, 100 μg/mL streptomycin) containing 500 IU/mL IL-2 (Novartis). All cell lines were tested for mycoplasma using PCR-based screening and found negative.

Retrovirus production

Retroviral particles were produced in FLYRD18 packaging cells. In brief, 700,000 FLYRD18 packaging cells were plated per 10 cm dish 1 day before transfection. The next day, cell culture medium was refreshed with IMDM supplemented with 8% FCS without antibiotics. Twenty-five microliters of X-tremeGENE 9 (Roche) was mixed with 800 μL Opti-MEM (Thermo Fisher Scientific) and incubated for 5 min. Subsequently, the Optimem-X-tremeGENE 9 mixture was added on top of 10 μg retroviral plasmid DNA dissolved in water and incubated for 15 min, and the resulting transfection mixture was added dropwise onto the packaging cells. Retrovirus containing supernatant was harvested 48 h after transfection and immediately used or snap-frozen in liquid nitrogen.

T cell isolation and activation

Peripheral blood mononuclear cells (PBMC) were isolated from buffy coats from healthy donors in accordance with local regulations (Sanquin, Amsterdam, NL) by Ficoll-Isopaque density centrifugation²⁷ and were stored frozen until further use. To generate activated T cell populations, PBMCs were thawed in PBS containing 5% FCS, counted, and mixed with CD3/CD28 Dynabeads (CTS) at a 1:1 cell to bead ratio, at a density of 10⁷ cells/mL. After a 30-min incubation at room temperature on a tumbler, the mixture was put on a magnet and unbound cells were removed. Bead bound T cells were subsequently resuspended in RPMI supplemented with 10% human serum (Sigma-Aldrich) and penicillin–streptomycin containing 100 IU/mL IL-2 and 5 ng/mL IL-15 (Peprotech), and plated at a density of 0.75 × 10⁶ cells/mL.

Spin-transduction of T cells

Twenty-four-well nontreated cell culture plates were covered with 10 μg/mL RetroNectin (Takara) overnight at 4°C. The next day, the RetroNectin solution was removed and wells were blocked with 2% BSA (Sigma-Aldrich) in PBS for 30 min. Activated T cells (0.25 × 10⁶ cells/mL in RPMI/10% human serum/penicillin–streptomycin/200 IU/mL IL-2 and 10 ng/mL IL-15) were then mixed with retroviral supernatant at a 1:1 ratio (volume/volume) in RetroNectin-coated 24-well plates and centrifuged at 2,000 RPM for 90 min at room temperature. Transduction efficiencies and estimated multiplicity of infection (MOI) values are listed in Supplementary Tables S5–16. MOI values were calculated using Poisson distribution equation P(0) = e ^−λ, where P(0) is the fraction of untransduced cells, and λ is the MOI.

Small molecule compounds

Asunaprevir, an orally available drug for the treatment of hepatitis C virus (HCV) infection,²⁸ was obtained from MedChemExpress (No. HY-14434, supplied as 10 mM stock solution in DMSO). dTAG-13 was obtained from Tocris (No. 6605) as powder, and a 10 mM stock solution in DMSO was prepared (No. D4540–500 ML; Sigma).

Cytokine release assay and antibody staining

T cells were pretreated with 10 μM asunaprevir, 0.5 μM dTAG-13 or DMSO control in T cell medium (RPMI/10% human serum/penicillin–streptomycin, 100 IU/mL IL2 and 5 ng/mL IL15) for 24 h before coculture experiments. 100,000 T cells and 100,000 of the indicated tumor cells were mixed in T cell medium supplemented with golgi-plug (1:1,000 dilution; BD) and anti-LAMP1-APC (No. 328620; 1:100 dilution; Biolegend), in the presence of indicated drugs or DMSO control in 96-well plates and incubated for 5 h at 37°C. In all coculture experiments, transduced cell populations from the same donor were for each experimental condition (i.e., DMSO control and either asunaprevir or dTAG-13 treatment) split into three wells before coculture with target cells. After incubation, cells were washed once with PBS and stained with IR dye (No. L34976; Invitrogen,) at 1:400 dilution for 5 min at 4°C. Subsequently, cells were washed once with FACS buffer (PBS plus 0.5% BSA) and stained with anti-CD8-PerCP Cy5.5 (No. 341050, 1:20 dilution; BD), anti-CD4 BV711 (No. 317440, 1:50 dilution; Biolegend), anti-murine constant TCR-PE (in experiments using TCR transduced T cells, 1:200 dilution, No. 553172; BD) or cetuximab-PE (in experiments using CAR transduced T cells that express the truncated human EGFR reporter, 1:200 dilution, No. FAB9577P; R&D Systems) for 20 min at 4°C. Cells were then washed once with FACS buffer and fixed with BD fixation and permeabilization solution for 20 min at 4°C. After permeabilization, cells were washed twice with BD perm/wash buffer and stained with anti-IFNγ-BV421 (1:100 dilution, No. 564791; BD), anti-IL-2-PE-Cy7 (1:100 dilution, #560707; BD), and anti-TNFα-BV650 (1:100 dilution, #502938; Biolegend) diluted in perm/wash buffer for 20 min at 4°C. Cells were then washed twice, resuspended in 100 μL FACS buffer and analyzed directly on a Fortessa Special Order analyzer. Data were analyzed using FlowJo and Prism 7 software. Flow gating strategy and representative flow data are provided in Supplementary Fig. S7. Absolute cytokine expression and degranulation profiles are listed in Supplementary Tables S17–34.

Results

Characterization of CRASH-IT components required for efficient and reversible inhibition of T cell function

Upon pMHC ligation to the TCR, Lck binds to CD4/CD8 costimulatory receptors and phosphorylates ITAM motifs present in the CD3zeta chain. The ITAM-phosphorylated CD3zeta chain subsequently forms a docking platform for the tandem SH2 domains of the Zap70 tyrosine kinase. To characterize components of the CRASH-IT switch (Fig. 1A, B) that are required for efficient and reversible inhibition of T cell signaling, we first investigated whether CRASH-IT switches require bivalent binding to the CD3zeta chain, or whether either of the two individual Zap70 SH2 domains suffices to bring the PD-1 inhibitory tail domain to the TCR complex. Primary human T cells were modified with a high affinity CDK4 neoantigen-specific class I-restricted TCR together with CRASH-IT variants that either contain one of the individual or both SH2 domains of Zap70.

Analysis of cytokine production and degranulation of modified T cells cocultured with NKIRTIL006 melanoma cells endogenously expressing the CDK4 neoantigen revealed efficient control over T cell activity in T cells equipped with the full-length first-generation CRASH-IT switch (a 15-, 33-, 20-, and 14-fold difference in intracellular IFNγ, IL2, TNFα, and surface LAMP-1 levels in absence or presence of 10 μM of the HCV NS3/4A protease inhibitor asunaprevir that is used as to induce degradation of SMASh containing CRASH-IT switches). In contrast, suppression of cytokine production and degranulation in T cells equipped with CRASH-IT variants containing a single SH2 domain was <1.2-fold (Fig. 1C–G).

To subsequently understand which domains within the PD-1 tail are critical for T cell inhibition by CRASH-IT switches, we focused on the PD-1 ITIM and ITSM motifs that are phosphorylated after PD-L1 ligation, and thereby form a docking platform for the SH2 domains of the inhibitory SHP-2 phosphatase.²⁹ Mutation of Y248 to phenylalanine identified a critical role for the PD-1 ITSM, with very limited control over all T cell activities in the absence of asunaprevir (Fig. 1C–G). In contrast, mutation of Y223 to phenylalanine resulted in only a mild impairment in switch activity, indicating a modest but significant contribution of the PD-1 ITIM, with a 1.8-, 1.5-, 2.3-, and 1.7-fold increase in intracellular IFNγ, IL2, TNFα, and surface LAMP-1 levels in T cells equipped with the ITIM deficient switch, relative to cells expressing the parental switch in the absence of asunaprevir (p = 0.0097, 0.0206, 0.0077, and 0.0336, respectively). In line with expectations, asunaprevir addition did not influence the activity of non-switch modified (i.e. EGFP-negative) CDK4 TCR+ T cells present in the same cultures, indicating that control over T cell function is fully cell-intrinsic (Supplementary Fig. S1).

Docking domain requirements to achieve efficient inhibition of T cell signaling

Having established the importance of both individual Zap70 SH2 domains, the PD-1 ITSM and, to a lesser extent, the PD-1 ITIM to allow efficient inhibition of T cell function, we explored the possibility space for domains facilitating efficient docking of inhibitory domains to the TCR complex by replacing the Zap70 SH2 domains of the CRASH-IT switch with interaction partners of the CD3zeta chain, with direct fusion to the CD3zeta chain, or with signaling/adaptor proteins involved in proximal or more distal TCR signaling. Using protein interaction information from the Uniprot database and literature,³⁰ a total of 27 putative switch constructs was designed employing the docking domains or full-length sequences from 19 signaling or adaptor proteins (Supplementary Table S1).

Analysis of switch function in primary human T cells modified to express the CDK4 TCR plus individual CRASH-IT variants indicated that robust inhibition of antigen-induced T cell activation was highly dependent on the docking domain used (Fig. 2 and Supplementary Fig. S2). Specifically, although Nck2-, TRAT1-, Grb2-, and CD3z-based switch designs showed a low-level inhibitory effect on T cell function, particularly IL2 production, the magnitude of these effects was minimal as compared with Zap70 SH2-based CRASH-IT switches.

Figure 2.

Docking domain requirements to achieve efficient inhibition of T cell signaling. Primary human T cells modified with the CDK4 TCR plus vector control, plus a switch without docking domain, or plus a docking domain-PD1-SMASh switch in which the docking domain was chosen from Zap70 (2xSH2), Syk (2xSH2), Lck (SH4-Unique-SH3-SH2), Nck1 (SH3.1 and SH2), Nck1 (SH2), Nck2 (SH3.1 and SH2), Nck2 (SH2), SLA (SH3-SH2), SLA2 (SH3-SH2), SHB (SH2), SHC1 (SH2), TRAT1, DOCK2 (SH3), SH2D1A (SH2), SH2D2A (SH2), SH2D1B (SH2), CRK (SH2), Grb2 (full-length), Grb2 (SH2), Numb (PTB), SLP76 (SH2), truncated CD3zeta (1-51aa, 1-56aa, 1-60aa), or full-length CD3zeta, were pretreated with 10 μM asunaprevir or DMSO control. Data depict percentage inhibition and percentage restoration of intracellular IFNγ (A), IL-2 (B), TNFα (C), and surface LAMP1 (D) expression of CDK4 TCR⁺, EGFP high, CD8⁺ T cells compared with vector control upon coculture with NKIRTIL006 cells in the continued presence (y-axis) or absence (x-axis) of asunaprevir. Data are representative of two independent experiments. Color images are available online.

Notably, a CRASH-IT switch employing the SH2 domains from the Syk protein, a close homolog of Zap70, was shown to allow reversible inhibition T cell activation. Similar to Zap70, the Syk protein contains 2x SH2 domains for interaction with phosphorylated ITAM motifs.³¹ Despite the conserved functions of the SH2 domains in both proteins, T cell reactivation levels in cells expressing the Syk SH2-based switch was reduced, relative to cells expressing the Zap70-based switch. A switch design that contained the SH4-Unique-SH3-SH2 domains of Lck likewise allowed reversible inhibition of T cell functions. However, transduction efficiencies with the Lck-based CRASH-IT switch were substantially lowered relative to other switch designs, suggesting a detrimental effect in producer cells, and thereby limiting its clinical utility. The observation that a large majority of switch designs employing binding domains from different levels in the TCR signaling pathway provide little control over T cell activity, whereas Zap70, Syk, and Lck-based CRASH-IT switches that contain SH2 domains that directly interact with phosphorylated ITAM motifs present in CD3zeta chain do allow such control, suggests a strict spatial or temporal requirement to efficiently delivery inhibitory signals to the T cell signaling pathway.

Improved CRASH-IT switch designs with ITIM/ITSM containing signaling domains derived from non-T cell inhibitory receptors

We subsequently explored the possibility space of the inhibitory module in proximity-based T cell switches by replacing the PD-1 tail with signaling domains derived from a diverse set of inhibitory receptors. In this analysis, we did not restrict the inhibitory receptor set to proteins that are endogenously expressed in T cells or proteins containing ITIM/ITSM, but also sampled signaling domains from inhibitory receptors with dominant expression in either T cells, B cells, or myeloid cells. Side-by-side comparison with the PD-1-based CRASH-IT switch revealed that a CRASH-IT switch based on the ITIM/ITSM containing BTLA signaling domain inhibited cytokine production and degranulation with similar efficiency in the absence of asunaprevir, and this suppression was reversed in the presence of 10 μM asunaprevir (Supplementary Fig. S3). In contrast, CD22-, FCGR2B-, CTLA4-, TIM3-, LAG3-, and TIGIT-based switches all failed to efficiently suppress T cell function. Notably, use of a switch system based on the 3xITIM 1xITSM containing SIRPA signaling domain resulted in a suppression of T cell activity that exceeded that observed with the PD-1-based switch. Considering that substantial inhibitory activity, in one case exceeding that achieved with the PD-1 inhibitory domain, was observed with three different ITSM containing inhibitory receptor domains, we created an additional panel of switch constructs with a focus on ITSM containing inhibitory domains. To this purpose, we assembled a library of all ITSM containing inhibitory domains within the human proteome²¹ with a total length of <200 amino acids. To ensure utility also in CAR-modified T cells, this new panel of CRASH-IT variants, plus the variants from the first screen, were first evaluated with respect to their capacity to control signaling through a second-generation CD19-specific CAR in primary human T cells (Supplementary Fig. S4), and the activity of well-performing switch designs was subsequently verified in TCR-T cells (Supplementary Fig. S5). T cell activation upon culture with CD19⁺ Daudi cells in the absence or presence of 10 μM asunaprevir, was used to quantify and rank the degree of T cell inhibition provided by the various signaling domains (Supplementary Fig. S4B). Furthermore, to be able to identify those switches that already mediate efficient suppression at low expression levels, we focused on T cell inhibition as a function of switch expression level. Notably, analysis of suppression of T cell activity in cells with intermediate switch expression levels revealed that switch systems that were based on the myeloid cell inhibitory receptors SIGLEC5, SIGLEC9, and SIGLEC11 all yielded significantly stronger inhibition of T cell activity, as compared with the PD-1-based switch (p ≤ 0.000001, p = 0.000003, p ≤ 0.000001, respectively), and a SIRPA-based switch trended toward higher activity, whereas there was no statistically significant difference between inhibition provided by PECAM1 and PD-1-based switches. Substantial, but lower, control over T cell activity was obtained with BTLA, SIGLEC6, and LY9 containing CRASH-IT variants, whereas SLAMF1, SLAMF6, SLAMF7, LAG3, and TIM3 containing switches inhibited T cell activation only weakly. CD22, FCGR2B, CTLA4, TIGIT, CD84, CXADR, and CD244 containing switches inhibited T cell activation even less than the switch without inhibitory tail, possibly because of impaired folding or stability of these fusion proteins (Supplementary Fig. S4B). Notably, in cells with high-level switch expression, asunaprevir-induced degradation of SIRPA and SIGLEC5, 9, or 11 containing switches failed to efficiently restore T cell effector functions, further suggesting the strong inhibitory activity of these signaling domains (Supplementary Fig. S6), and indicating a need for improved control over switch protein levels for such second-generation switch systems. Importantly, efficient suppression of T cell activity by SIGLEC5, SIGLEC9, SIGLEC11, and SIRPA-based CRASH-IT switches was also observed in primary human T cells modified with a tumor-reactive TCR, indicating that strong T cell signaling inhibitory domains efficiently downregulate T cell functions independent of the type of antigen receptor (Supplementary Fig. S5).

Efficient control over TCR-modified T cell function with optimized FKBP(V)-based CRASH-IT switches

In T cells modified with SIRPA/SIGLEC5/SIGLEC9/SIGLEC11 inhibitory domain-based CRASH-IT switches that employ the SMASh domain to regulate switch protein levels, asunaprevir addition only results in a partial restoration of T cell activity, suggesting that residual switch protein levels still result in substantial inhibition of TCR signaling (Supplementary Figs. S3–S6). To determine whether induced protein degradation systems with a higher dynamic range could increase the utility of such highly active designs, we created a panel of dTAG-13 inducible FKBP(V)-based CRASH-IT variants. Importantly, despite the very potent inhibitory potential of these four myeloid inhibitory receptor domains in the absence of the dTAG-13 PROTAC, cytokine production, and surface LAMP-1 levels were largely restored in the presence of 0.5 μM dTAG-13, both in cells with high and intermediate levels of switch expression (Fig. 3). Particularly tight regulation of T cell output signals was observed with the SIGLEC11-based switch system, with PROTAC addition boosting the fraction of intracellular IFNγ, IL2, TNFα, and surface LAMP-1-positive cells in the presence antigen-positive tumor cells by 61-, 422-, 235-, and 29-fold in cells with high level switch expression. Furthermore, relative to PD-1 switch expressing cells, a 2.3-, 4.5-, 4.2-, and 2.5-times greater fold induction of intracellular IFNγ, IL2, TNFα, and surface LAMP-1-positive cells was observed in SIGLEC11-based switch expressing cells, with the activity of T cells in the presence of 0.5 μM dTAG-13 closely approaching the activity of control vector modified cells. These data identify second-generation CRASH-IT-based switch designs with improved properties over PD-1-based switches and indicate a particular utility of a number of members of the SIGLEC family. Site-directed mutagenesis of the ITIM and/or ITSM motifs in the SIGLEC11-based CRASH-IT switch revealed that both the SIGLEC11-ITIM and SIGLEC11-ITSM contributed substantially to the high degree with T cell inhibition, but with a larger contribution of the former (Fig. 4). Together with the prior data revealing a dominant role for the PD-1 ITSM in T cell inhibition (Fig. 1C–G), these data suggest that the relative contribution of ITIM and ITSM motifs to T cell inhibition is context dependent.

Figure 3.

Improved dynamic range of PROTAC-based highly active CRASH-IT switch designs. Primary human T cells modified with the CDK4 TCR plus vector control, plus a switch without inhibitory domain, or plus a Zap70 (2xSH2)-inhibitory domain-FKBP(V) switch in which the inhibitory domain was chosen from PD-1, SIRPA, SIGLEC5, SIGLEC9, or SIGLEC11 signaling domains, were pretreated with 0.5 μM dTAG-13 or DMSO control. Data depict intracellular IFNγ, IL2, TNFα, and cell surface LAMP1 expression of CDK4 TCR⁺, CD8⁺, EGFP high (A) and EGFP intermediate (B) T cells upon coculture with NKIRTIL006 cells in the continued presence or absence of dTAG-13. Red horizontal lines indicate cytokine production and degranulation levels of (A) PD-1 tail-based switch expressing cells in the presence of dTAG-13 and (B) PD-1 tail-based switch expressing cells in the absence of dTAG-13. Error bars represent standard deviation from three biological replicates (n = 3). Data are representative of two independent experiments. Color images are available online.

Figure 4.

Context-dependent contribution of ITIM/ITSM inhibitory motifs to T cell inhibition. (A) Schematic representation of SIGLEC11-based CRASH-IT variants. (B) Primary human T cells modified with the CDK4 TCR plus vector control, plus a switch without inhibitory domain, or plus a Zap70 (2xSH2)-inhibitory domain-FKBP(V) switch in which the inhibitory domain was chosen from wild-type PD-1 or SIGLEC11 signaling domains, or the SIGLEC11 variants described in (A), were pretreated with 0.5 μM dTAG-13 or DMSO control. Data depict intracellular IFNγ, IL2, TNFα, and cell surface LAMP1 expression of CDK4 TCR⁺, EGFP intermediate, CD8⁺ T cells upon coculture with NKIRTIL006 cells in the continued presence or absence of dTAG-13. Error bars represent standard deviation from three biological replicates (n = 3). Data are representative of two independent experiments. (C) Same data as in (B) depicted as the percentage inhibition and percentage restoration of intracellular IFNγ, IL-2, TNFα, and surface LAMP1 expression of CDK4 TCR⁺, EGFP intermediate, CD8⁺ T cells compared with vector control-modified cells upon coculture with NKIRTIL006 cells in the continued presence (y-axis) or absence (x-axis) of dTAG-13. Data are representative of two independent experiments. ITIM/ITSM, immunoreceptor tyrosine-based inhibition motif/switch motif. Color images are available online.

Efficient regulation of a diverse panel of ITAM receptors with a SIGLEC11-based CRASH-IT switch

The aforementioned data indicate a substantial flexibility in the capacity of different inhibitory domains to regulate signaling through the CD3 ITAM motifs in TCR-modified cells and the CD3zeta ITAM motifs in CAR-modified T cells. To explore whether a given inhibitory domain that can be used to control T cell signaling through the CD3 complex can also be used to control T cell activation through other ITAM containing domains, we generated primary human T cells modified to express CD19-specific CARs that contained the ITAM containing domains of either CD3zeta, the gamma chain of the immunoglobulin E receptor (FCER1G), CD3epsilon, or DAP12. Exposure of resulting CAR T cells to CD19⁺ Daudi cells in the absence or presence of 0.5 μM dTAG-13 revealed tight control over T cell activation in all four cases (Fig. 5).

Figure 5.

Efficient regulation of different CAR designs by second-generation CRASH-IT switches. (A) Schematic representation of CD19-specific CARs containing the CD3zeta, FCERG1, CD3epsilon, and DAP12-ITAM signaling domains and a control CAR without an ITAM containing domain. Primary human T cells modified with the CD19-specific CAR constructs depicted in (A) plus Zap70-PD1-FKBP(V) switch (B), Zap70-SIGLEC11-FKBP(V) switch (C), or vector control (D) were pretreated with 0.5 μM dTAG-13 or DMSO control. Data depict intracellular IFNγ, IL2, TNFα, and cell surface LAMP1 expression of CAR⁺, EGFP high, CD8⁺ T cells upon coculture with Daudi cells in the continued presence or absence of dTAG-13. Error bars represent standard deviation from three biological replicates (n = 3). Data are representative of two independent experiments. CAR, chimeric antigen receptor. Color images are available online.

To explore the capacity of CRASH-IT-based systems to also control activity of ITAM containing NK cell receptor complexes in their physiological context, we modified the KHYG-1 NK cell line with first- or second-generation CRASH-IT switches. Coculture of resulting cells with HLA class I deficient K562 cells showed efficient inhibition of cytokine production in the absence of small molecule for both PD-1 and SIGLEC11-based designs, but with superior restoration of cytokine production in the presence of 0.5 μM dTAG-13 for the second-generation switch (Fig. 6). Thus, whether antigen recognition is provided by the TCR, CD3 zeta chain-containing classical CAR designs, a range of other CAR designs with alternative ITAM containing domains, or by NK cell receptors, SIGLEC11-based switch systems provide superior control over immune cell activity.

Figure 6.

Control of NK cell function by second-generation CRASH-IT switches. The human NK cell line KHYG-1 modified with Zap70-PD1-FKBP(V) switch (A), Zap70-SIGLEC11-FKBP(V) switch (B), or vector control (C) were pretreated with 0.5 μM dTAG-13 or DMSO control. Data depict intracellular IFNγ, IL2 and TNFα expression of EGFP intermediate NK cells upon coculture in the presence or absence of K562 tumor cells in the continued presence or absence of dTAG-13. (D) Representative plot showing EGFP intermediate gate used in (A–C). Error bars represent standard deviation from three biological replicates (n = 3). Data are representative of two independent experiments. Color images are available online.

Conclusion

The efforts to optimize the different structural domains of CRASH-IT switches described in this study yielded second-generation switch designs with improved capacity to control T cell activity. In particular, SIGLEC11-based switches allow an extremely tight control over T cell activity, with an on average 189-fold induction in the fraction responding cells by dTAG-13 addition over the four T cell functions tested. As demonstrated using the dTAG-13 regulated FKBP(V)-based designs, such switch systems can be assembled fully from human protein domains, thereby significantly reducing the probability of immune-mediated rejection. Furthermore, the fact that control over T cell activity is achieved in cells with both low and high switch expression also significantly facilitates clinical implementation and may, for instance, enable engineering of compact vector designs that combine antigen receptor and CRASH-IT in a single vector.

CRASH-IT-based switches may potentially be used for the clinical testing of gene-modified T cells in settings in which on-target off-tumor toxicity forms a concern, or to reduce the likelihood of cell therapy-induced CRS or NT. In addition, a recent study has suggested that CAR T cells may benefit from pulsatile—rather than constitutive—antigen receptor signaling,³² and CRASH-IT triggering could provide a means to achieve this. On a more general note, the ability of ITIM/ITSM containing inhibitory signaling domains derived from different lineages of the blood system to control signaling through a variety of ITAM containing receptors provides clear evidence of the modular nature of immune receptor signaling. This malleability of immune receptor systems makes the use of components of this system in other synthetic biology approaches that aim to develop novel signaling networks particularly attractive.

Footnotes

Authors' Contributions

A.C.S. and T.N.S. conceived the project, analyzed the data, and wrote the article. A.C.S. performed experiments and visualized the data. T.N.S. supervised the project.

Acknowledgments

We would like to thank R.N. Ozdeslik (UMC Utrecht, The Netherlands) for illustrations.

Author Disclosure

A.C.S and T.N.S. are inventors on the patent application PCT/NL2020/050655 that covers the CRASH-IT switch system. T.N.S. is advisor for Adaptive Biotechnologies, Asher Biotherapeutics, Allogene Therapeutics, Merus, and Neogene Therapeutics; is a recipient of grant or research support from Merck KgaA; is a stockholder in Allogene Therapeutics, Asher Biotherapeutics, Merus, and Neogene Therapeutics; and is venture partner at Third Rock Ventures, all outside of this study.

Funding Information

This study was supported by Oncode Institute core funding.

Supplementary Material

Supplementary Figure S1

Supplementary Figure S2

Supplementary Figure S3

Supplementary Figure S4

Supplementary Figure S5

Supplementary Figure S6

Supplementary Table S1

Supplementary Table S2

Supplementary Table S3

Supplementary Table S4

Supplementary Table S5

Supplementary Table S6

Supplementary Table S7

Supplementary Table S8

Supplementary Table S9

Supplementary Table S10

Supplementary Table S11

Supplementary Table S12

Supplementary Table S13

Supplementary Table S14

Supplementary Table S15

Supplementary Table S16

Supplementary Table S17

Supplementary Table S18

Supplementary Table S19

Supplementary Table S20

Supplementary Table S21

Supplementary Table S22

Supplementary Table S23

Supplementary Table S24

Supplementary Table S25

Supplementary Table S26

Supplementary Table S27

Supplementary Table S28

Supplementary Table S29

Supplementary Table S30

Supplementary Table S31

Supplementary Table S32

Supplementary Table S33

Supplementary Table S34

References

Brudno

, Kochenderfer

. Chimeric antigen receptor T-cell therapies for lymphoma. Nat Rev Clin Oncol, 2018; 15:31–46.

Bollard

, Rooney

, Heslop

. T-cell therapy in the treatment of post-transplant lymphoproliferative disease. Nat Rev Clin Oncol, 2012; 9:510–519.

Belin

, Devic

, Ayrignac

, et al. Description of neurotoxicity in a series of patients treated with CAR T-cell therapy. Sci Rep-uk, 2020; 10:18997.

Norelli

, Camisa

, Barbiera

, et al. Monocyte-derived IL-1 and IL-6 are differentially required for cytokine-release syndrome and neurotoxicity due to CAR T cells. Nat Med, 2018; 24:739–748.

Brudno

, Kochenderfer

. Recent advances in CAR T-cell toxicity: Mechanisms, manifestations and management. Blood Rev, 2018; 34:45–55.

Berg

JH van den

, Gomez-Eerland

, Wiel

B van de

, et al. Case report of a fatal serious adverse event upon administration of T cells transduced with a MART-1-specific T-cell receptor. Mol Ther, 2015; 23:1541–1550.

Parkhurst

, Yang

, Langan

, et al. T cells targeting carcinoembryonic antigen can mediate regression of metastatic colorectal cancer but induce severe transient colitis. Mol Ther, 2011; 19:620–626.

Lamers

, Sleijfer

, Steenbergen

S van

, et al. Treatment of metastatic renal cell carcinoma With CAIX CAR-engineered T cells: Clinical Evaluation and Management of On-target Toxicity. Mol Ther, 2013; 21:904–912.

Bergen

CAM van

, Luxemburg-Heijs

SAP van

, Wreede

LC de

, et al. Selective graft-versus-leukemia depends on magnitude and diversity of the alloreactive T cell response. J Clin Invest, 2017; 127:517–529.

10.

Negrin

. Graft-versus-host disease versus graft-versus-leukemia. Hematology, 2015; 2015:225–230.

11.

Schubert

M-L

, Schmitt

, Wang

, et al. Side-effect management of chimeric antigen receptor (CAR) T-cell therapy. Ann Oncol, 2020; 32:34–48.

12.

Giordano-Attianese

, Gainza

, Gray-Gaillard

, et al. A computationally designed chimeric antigen receptor provides a small-molecule safety switch for T-cell therapy. Nat Biotechnol, 2020; 38:426–432.

13.

Jones

, Lamb

, Goldman

, et al. Improving the safety of cell therapy products by suicide gene transfer. Front Pharmacol, 2014; 5:254.

14.

Bachmann

. The UniCAR system: A modular CAR T cell approach to improve the safety of CAR T cells. Immunol Lett, 2019; 211:13–22.

15.

Sahillioglu

, Schumacher

. Safety switches for adoptive cell therapy. Curr Opin Immunol. Epub ahead of print 2021. DOI: 10.1016/j.coi.2021.07.002.

16.

Sahillioglu

, Toebes

, Apriamashvili

, et al. CRASH-IT switch enables reversible and dose-dependent control of TCR and CAR T-cell Function. Cancer Immunol Res, 2021; 9:999–1007.

17.

Marasco

, Berteotti

, Weyershaeuser

, et al. Molecular mechanism of SHP2 activation by PD-1 stimulation. Sci Adv, 2020; 6:eaay4458.

18.

Mizuno

, Sugiura

, Shimizu

, et al. PD-1 primarily targets TCR signal in the inhibition of functional T cell activation. Front Immunol, 2019; 10:630.

19.

Chung

, Jacobs

, Huo

, et al. Tunable and reversible drug control of protein production via a self-excising degron. Nat Chem Biol, 2015; 11:713–720.

20.

Nabet

, Roberts

, Buckley

, et al. The dTAG system for immediate and target-specific protein degradation. Nat Chem Biol, 2018; 14:431–441.

21.

Liu

, Li

, Voss

, et al. A comprehensive immunoreceptor phosphotyrosine-based signaling network revealed by reciprocal protein–peptide array screening. Mol Cell Proteomics, 2015; 14:1846–1858.

22.

Klichinsky

, Ruella

, Shestova

, et al. Human chimeric antigen receptor macrophages for cancer immunotherapy. Nat Biotechnol, 2020; 38:947–953.

23.

Xie

, Dong

, Liang

, et al. CAR-NK cells: A promising cellular immunotherapy for cancer. Ebiomedicine, 2020; 59:102975.

24.

Ren-Heidenreich

, Mordini

, Hayman

, et al. Comparison of the TCR ζ-chain with the FcR γ-chain in chimeric TCR constructs for T cell activation and apoptosis. Cancer Immunol Immunother, 2002; 51:417–423.

25.

Strønen

, Toebes

, Kelderman

, et al. Targeting of cancer neoantigens with donor-derived T cell receptor repertoires. Science, 2016; 352:1337–1341.

26.

Kvistborg

, Shu

, Heemskerk

, et al. TIL therapy broadens the tumor-reactive CD8 + T cell compartment in melanoma patients. Oncoimmunology, 2012; 1:409–418.

27.

Linnemann

, Heemskerk

, Kvistborg

, et al. High-throughput identification of antigen-specific TCRs by TCR gene capture. Nat Med, 2013; 19:1534–1541.

28.

Eley

, Garimella

, Li

, et al. Asunaprevir: A review of preclinical and clinical pharmacokinetics and drug–drug interactions. Clin Pharmacokinet, 2015; 54:1205–1222.

29.

Hui

, Cheung

, Zhu

, et al. T cell costimulatory receptor CD28 is a primary target for PD-1–mediated inhibition. Science, 2017; 355:1428–1433.

30.

Ngoenkam

, Schamel

, Pongcharoen

. Selected signalling proteins recruited to the T-cell receptor–CD3–complex. Immunology, 2018; 153:42–50.

31.

Mócsai

, Ruland

, Tybulewicz

VLJ

. The SYK tyrosine kinase: a crucial player in diverse biological functions. Nat Rev Immunol, 2010; 10:387–402.

32.

Weber

, Parker

, Sotillo

, et al. Transient rest restores functionality in exhausted CAR-T cells through epigenetic remodeling. Science, 2021; 372:eaba1786.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.01 MB

0.08 MB

0.36 MB

0.30 MB

0.43 MB

0.31 MB

0.20 MB

0.46 MB

0.07 MB

Multimodular Optimization of Chemically Regulated T Cell Switches Demonstrates Flexible and Interchangeable Nature of Immune Cell Signaling Domains

Abstract

Introduction

Methods

Retroviral DNA constructs

Cell lines and cell culture

Retrovirus production

T cell isolation and activation

Spin-transduction of T cells

Small molecule compounds

Cytokine release assay and antibody staining

Results

Characterization of CRASH-IT components required for efficient and reversible inhibition of T cell function

Docking domain requirements to achieve efficient inhibition of T cell signaling

Improved CRASH-IT switch designs with ITIM/ITSM containing signaling domains derived from non-T cell inhibitory receptors

Efficient control over TCR-modified T cell function with optimized FKBP(V)-based CRASH-IT switches

Efficient regulation of a diverse panel of ITAM receptors with a SIGLEC11-based CRASH-IT switch

Conclusion

Footnotes

Authors' Contributions

Acknowledgments

Author Disclosure

Funding Information

Supplementary Material

References

Supplementary Material