Lenalidomide boosts CD1d-Vδ2 bispecific antibody-engaged Vγ9Vδ2-T cell effector functions via CD28, Notch signaling and IL-2 release Authors: Milon de Jong, Myrthe Veth, Tereza Brachtlová, Lisa A King, José Saura-Esteller, Tamara M Bechler, Pauline M van Helden, Canan Alhan, Roeland Lameris, Tanja D de Gruijl, Hans J van der Vliet Published: 2026 ABSTRACT AbstractBackgroundVγ9Vδ2-T cells form a conserved T-cell subset known for its potent intrinsic antitumor activity and versatility in recognizing diverse cancer types independently of the major histocompatibility complex. Previously, we reported the preclinical activity of a bispecific T-cell engager (bsTCE) specific for both CD1d and the Vδ2-TCR that engaged both Vγ9Vδ2-T and type 1 natural killer T cells to CD1d-expressing hematological malignancies, including multiple myeloma (MM) and acute myeloid leukemia (AML). Here, we evaluated whether various standard-of-care drugs for patients with MM and AML/myelodysplastic syndromes (MDS) affected the in vitro antitumor activity of CD1d-Vδ2 bsTCE-activated Vγ9Vδ2-T cells.MethodsMM and AML/MDS standard-of-care drugs were tested for antagonistic, additive, or synergistic effects on CD1d-Vδ2 bsTCE-induced and Vγ9Vδ2-T cell-mediated tumor cell lysis. Based on observed synergy, the immunomodulatory drug (IMiD) lenalidomide was studied in more detail, exploring effects on Vγ9Vδ2-T cell proliferation, phenotype, cytokine profile, and cytolytic activity, as well as its mechanism of action, using healthy donor peripheral blood mononuclear cells (PBMC) and MDS patient PBMC and bone marrow samples.ResultsLenalidomide exerted synergistic activity on CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T cell antitumor activity and was found to substantially enhance CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell activation, degranulation, Th1-type cytokine secretion, and expansion. Stimulatory effects of lenalidomide on Vγ9Vδ2-T cell effector functions were mediated via enhanced intracellular CD28 phosphorylation, activation of Notch signaling, and interleukin-2 release. Importantly, lenalidomide facilitated otherwise absent proliferation of Vγ9Vδ2-T cells with oncolytic potential in response to CD1d-Vδ2 bsTCE exposure in MDS patient-derived bone marrow samples.ConclusionOur findings provide a rationale to explore the combination of CD1d-Vδ2 bsTCE and the IMiD lenalidomide in patients with CD1d-expressing hematological malignancies. WHAT IS ALREADY KNOWN ON THIS TOPIC A CD1d-Vδ2 bispecific T-cell engager (bsTCE) was shown to selectively engage both Vγ9Vδ2-T and type 1 natural killer T cells to trigger a potent antitumor response to CD1d-expressing hematologic malignancies and proceeded to a first-in-human phase 1 dose escalation trial in patients with chronic lymphocytic leukemia, multiple myeloma, or acute myeloid leukemia refractory to prior therapy. A CD1d-Vδ2 bispecific T-cell engager (bsTCE) was shown to selectively engage both Vγ9Vδ2-T and type 1 natural killer T cells to trigger a potent antitumor response to CD1d-expressing hematologic malignancies and proceeded to a first-in-human phase 1 dose escalation trial in patients with chronic lymphocytic leukemia, multiple myeloma, or acute myeloid leukemia refractory to prior therapy. WHAT THIS STUDY ADDS This study demonstrates the immunomodulatory drug lenalidomide to promote CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell activation, proliferation, Th1 type cytokine secretion, and cytolytic activity through CD28 phosphorylation, activation of Notch signaling, and interleukin-2 release. Importantly, lenalidomide promotes Vγ9Vδ2-T cell responsiveness to CD1d-Vδ2 bsTCE exposure in bone marrow samples of patients with myelodysplastic syndrome. This study demonstrates the immunomodulatory drug lenalidomide to promote CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell activation, proliferation, Th1 type cytokine secretion, and cytolytic activity through CD28 phosphorylation, activation of Notch signaling, and interleukin-2 release. Importantly, lenalidomide promotes Vγ9Vδ2-T cell responsiveness to CD1d-Vδ2 bsTCE exposure in bone marrow samples of patients with myelodysplastic syndrome. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY This study enhances our understanding of the mechanism of action of lenalidomide and provides a rationale for combining Vγ9Vδ2-T cell-directed therapy with the immunomodulatory drug lenalidomide. This study enhances our understanding of the mechanism of action of lenalidomide and provides a rationale for combining Vγ9Vδ2-T cell-directed therapy with the immunomodulatory drug lenalidomide. Introduction Bispecific T-cell engagers (bsTCE) offer an “off-the-shelf” immunotherapeutic modality by triggering T cell-mediated tumor lysis via crosslinking of tumor antigens and, most commonly, the CD3-T cell receptor (TCR) complex.1 CD3-directed bsTCE showed impressive clinical results, particularly in B-cell malignancies, but toxicities, including cytokine release syndrome (CRS), neurotoxicity, and on-target off-tumor toxicity often occur.2 Moreover, collateral engagement of regulatory T cells (Tregs) negatively impacts clinical outcome.3 In preclinical and early clinical settings, we demonstrated that selective engagement of innate-like T-cell subsets with inherent antitumor activity, such as Vγ9Vδ2-T and type 1 natural killer T (NKT) cells, can overcome these challenges and may allow for a favorable balance between efficacy and toxicity.49 Vγ9Vδ2-T cells represent a sizeable immune-effector cell population (1–10% of peripheral blood (PB) T cells) that responds to intracellular accumulation of phosphoantigens (pAgs) by sensing conformational changes in the butyrophilin 2A1/3A1 complex.6 As metabolites of the mevalonate pathway, pAgs frequently accumulate in malignant cells due to metabolic dysregulation, which sensitizes these cells to Vγ9Vδ2-T cell recognition. On activation, Vγ9Vδ2-T cells rapidly expand, secrete pro-inflammatory cytokines, exert a direct cytotoxic effect via perforin/granzyme B and Fas ligand, and can (cross-)present antigens in a co-stimulatory context.6 Notably, Vγ9Vδ2-T cell frequencies in PB and the tumor microenvironment (TME) positively correlate with clinical outcome in multiple malignancies.1012 While therapies using pAg-based approaches and/or adoptive cell transfer of ex vivo expanded Vγ9Vδ2-T cells were proven safe, overall clinical efficacy was limited.6 We and others have shown that tumor-targeted activation of Vγ9Vδ2-T cells via a bsTCE that crosslinks the TCR Vγ9 or Vδ2 chain with a tumor-associated membrane antigen could significantly improve the consistency and robustness of Vγ9Vδ2-T cell-directed therapies.4 5 8 13 14 Indeed, by using a CD1d-γδ bsTCE we demonstrated strong activity in in vivo and ex vivo tumor models, and good tolerability in a non-human primate model.8 CD1d is a major histocompatibility complex class I-like molecule expressed by healthy and malignant cells of myelomonocytic and B-cell lineage origin5 8 and some solid malignancies,1518 and presents both self and foreign (glyco)lipid antigens to type 1 NKT cells, a relatively rare innate T-cell subset (~0.1% of PB T cells) with potent inherent antitumor properties.19 By directly crosslinking CD1d and the Vδ2-TCR chain of Vγ9Vδ2-T cells, and concurrent stabilization of the interaction between CD1d and the type 1 NKT TCR, the CD1d-Vδ2 bsTCE selectively engages both Vγ9Vδ2-T and type 1 NKT cells to trigger a potent antitumor response to CD1d-expressing tumors in chronic lymphocytic leukemia (CLL), multiple myeloma (MM) and acute myeloid leukemia (AML) models.8 Accordingly, early evidence of CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell engagement was reported in a first-in-human phase 1 dose escalation trial.7 Rational combination therapies are common in the treatment of hematologic malignancies in an attempt to enhance treatment efficacy. As bsTCE activity intrinsically depends on the endogenous T-cell population, combinations with therapies that boost T-cell activity and expansion may enhance clinical efficacy. In the case of Vγ9Vδ2-T and type 1 NKT cells in particular, combinations that increase the frequency of these cells can be expected to enhance their therapeutic potential, as numerical defects have been described by us and others with increasing age, in the context of disease, and due to previous therapy.2023 In this study, we set out to evaluate the impact of several standard-of-care drugs used in the treatment of MM, AML, and myelodysplastic syndromes (MDS) on the antitumor activity of CD1d-Vδ2 bsTCE engaged Vγ9Vδ2-T cells. Among these drugs, the immunomodulatory drug (IMiD) lenalidomide (LEN) stood out as it significantly boosted Vγ9Vδ2-T cell antitumor effector functions through increased activation, degranulation, and Th1 cytokine production. Additionally, LEN promoted the CD1d-Vδ2 bsTCE-induced expansion of both type 1 NKT and Vγ9Vδ2-T cells, through enhanced CD28 phosphorylation, Notch signaling, and interleukin (IL)-2 secretion. As LEN potentiated CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell proliferation in bone marrow samples of patients with MDS, and we observed potent lysis of CD1d-expressing blast cells by CD1d-Vδ2 bsTCE-activated Vγ9Vδ2-T cells, our study underscores the relevance of exploring this combination in future clinical studies in patients with MDS. Materials and methods Generation and production of CD1d-Vδ2 bsTCE, generation and culture conditions of tumor cell lines and Vγ9Vδ2-T and type 1 NKT cells, human healthy donor and patient sample processing, and flow cytometry are described in the online supplemental methods; see online supplemental material. Activation, degranulation, cytokine, and cytotoxicity assays In cytotoxicity assays, 5×104 target cells (MM.1s.CD1d or MOLM-13) or 1×105 bone marrow mononuclear cells (BMMC) were co-cultured with or without 5×104 Vγ9Vδ2-T or type 1 NKT cells±concentration range or 10 nM of the CD1d-Vδ2 bsTCE and/or concentration range of the indicated standard-of-care drugs; LEN (Sigma-Aldrich), dexamethasone, bortezomib, daratumumab, azacytidine, venetoclax (MedChemExpress). To assess Vγ9Vδ2-T cell degranulation, PE-labeled CD107a antibody was added during incubation, if indicated. After 48 hours incubation, cells were stained with fluorescently labeled antibodies for flow cytometry and living tumor (7AAD−CD3−TRGV9−) or blast cells (7AAD−(SSClow)(CD7±)(CD13±)(CD33±)CD34+CD45dimCD117+) were quantified by 123 counting eBeads (Thermo Fisher Scientific (TFS)). Specific tumor/blast lysis was calculated as follows: 100 – ((absolute cell number (condition) / absolute cell number (target cells only)) × 100). Synergy was assessed using the Bliss Independence model.24 To assess the effect of LEN on Vγ9Vδ2-T cell effector functions and tumor cytotoxicity, 1×105 peripheral blood mononuclear cell (PBMC) or 5×104 γδ-T cells (isolated using the untouched human TCR γ/δ+-T cell isolation kit, Miltenyi; >85% Vγ9Vδ2-T of CD3+ cells after isolation) were primed for 72 hours with LEN (1 µM) or directly co-cultured with LEN for 18 hours or 48 hours with MM.1s.CD1d cells (10:1 PBMC:tumor or 1:1 Vγ9Vδ2-T:tumor ratio)±concentration range of the CD1d-Vδ2 bsTCE, in the presence of PE-labeled CD107a antibody. Subsequently, cells were stained with fluorescently labeled antibodies and specific tumor lysis (living tumor cells; 7AAD−CD3−(CD45−/dim) TRGV9−TRDV2−) was calculated as described above. Supernatants were collected, stored at −20°C until analysis of granzyme B secretion (granzyme B CBA flex set (BD)) and/or cytokine secretion (human Th1/Th2/Th17 CBA kit (BD)). For intracellular cytokine analysis, 1×105 PBMC were primed 72 hours with or without LEN (1 µM), cultured overnight with 1×104 MM.1s.CD1d cells plus GolgiPlug (BD) and GolgiStop (BD), ±CD1d-Vδ2 bsTCE (10 nM) or phorbol myristate acetate (PMA)/ionomycin (eBioscience Cell Stimulation Cocktail, TFS), and subsequently stained using the Cytofix/Cytoperm Fixation/Permeabilization Kit (BD) according to manufacturer’s instructions, and analyzed by flow cytometry. Vγ9Vδ2-T and type 1 NKT cell proliferation, expansion and phenotype assays To assess Vγ9Vδ2-T and type 1 NKT cell expansion, 1×105 healthy donor or MDS patient-derived PBMC/BMMC were cultured 7 days±1×104 MM.1s.CD1d cells±LEN (1 µM)±CD1d-Vδ2 bsTCE (10 nM). At baseline and day 7, cells were stained with fluorescently-labeled antibodies and Vγ9Vδ2-T and type 1 NKT cells were quantified by counting beads and flow cytometry. Expansion was calculated by dividing the cell count after co-culture by baseline cell count. Vγ9Vδ2-T cell proliferation was assessed by staining PBMC with carboxyfluorescein succinimidyl ester (CFSE) at baseline (CellTrace CFSE Cell Proliferation Kit, TFS), according to manufacturer’s instructions, or by intracellular staining with fluorescently labeled Ki-67 antibody at the end of co-culture. Using intracellular FoxP3 staining and flow cytometric counting beads, Tregs (7AAD−CD3+CD4+CD25highCD127−FoxP3+)25 were quantified at day 7 of the co-culture. To assess Vγ9Vδ2-T cell functionality following expansion, 7-day cultured PBMC were treated for 4 hours with PMA/ionomycin and analyzed for intracellular cytokines as described above, or were exposed overnight to C1R.CD1d cells in the presence of PE-labeled CD107a antibody after which Vγ9Vδ2-T cell degranulation and tumor cytotoxicity were determined. Where indicated, CD4+ cells were removed from PBMC before start of co-cultures by magnetic bead isolation using BV786-conjugated CD4 antibody and anti-mouse IgG microbeads (post-depletion, ≥96% of T cells and ≥99% of Vγ9Vδ2-T cells were CD4−). Where indicated, tissue culture plates were overnight coated with human CD1d (10 µg/mL, SinoBiological). Where indicated, cells were cultured in the presence of rhIL-2 (100 IU/mL, Proleukin), neutralizing human IL-2 antibody (10 µg/mL, R&D Systems), abatacept (CTLA-4 Ig, 10 µg/mL, MedChemExpress), CB-103 (10 µg/mL, MedChemExpress), rapamycin (1 µM, MedChemExpress), or RO429097 (γ-secretase inhibitor (GSI), 10 µM, MedChemExpress). Abatacept, CB-103, rapamycin, and RO429092 were added again to the co-cultures after 72 hours. Phospho-flow cytometry To assess CD28 phosphorylation of Vγ9+-T cells, PBMC were exposed to medium or CD1d-Vδ2 bsTCE (10 nM, 45 min at 4°C), washed on ice and subsequently incubated ±LEN (1 µM) at 37°C for 5, 15, or 60 min. Directly following incubation, cells were fixed with Phosflow Fix Buffer I (BD), incubated with FcR blocking reagent (Miltenyi), and stained using the Phosflow Perm Buffer III/Pharmingen Stain Buffer (BD), according to the manufacturer’s instructions. Specific binding of the phospho-CD28 antibody (Tyr218, TFS) was detected by an FITC-labeled anti-rabbit-antibody (online supplemental table S1). QuantitativePCR For analysis of Hes1 expression, expanded Vγ9Vδ2-T cells (purity >96%) were exposed to medium or LEN (1 µM) for 24 hours. RNA was extracted using RNeasy mini kit (Qiagen) and cDNA was synthesized with FireScript RT cDNA Synthesis Kit (Solis BioDyne) according to the manufacturer’s instructions. Quantitative PCR (qPCR) was run on 200 ng cDNA with Roche LightCycler 480 qPCR system using 5×HOT FIREPol EvaGreen qPCR Mix Plus (no ROX) (Solis BioDyne) in combination with Hs_HES1_1_SG QuantiTect Primer Assay (Qiagen) for the detection of Hes1. Hs_GAPDH_2_SG QuantiTect Primer Assay (Qiagen) was used for GAPDH expression as a housekeeping gene. The reactions were performed in technical duplicates and the 2–ΔΔCT method was applied to calculate the relative expression. Statistical analysis Data were analyzed using paired t-test, one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test, or two-way ANOVA followed by Šídák’s multiple comparisons test as indicated. Dose-response curves and half maximal effective concentrations (EC50) were calculated using non-linear regression analysis. Statistical analyses were performed using GraphPad Prism (V.10.2.0). P value ≤0.05 was considered significant and showed with asterisks: *p≤0.05, **p≤0.01, ***p≤0.001, ****p≤0.0001. Evaluation of the Bliss synergy score for drug interactions is based on the Bliss Independence model and performed with SynergyFinder R package26 using R (V.4.3.2). Results Lenalidomide boosts CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T cell effector functions resulting in enhanced antitumor activity In an effort to identify drugs that may enhance the efficacy of CD1d-Vδ2 bsTCE therapy in patients with MM, AML, or MDS, a panel of standard-of-care drugs approved for these indications was selected. To assess potential synergy, the CD1d-Vδ2 bsTCE was combined with each standard-of-care drug in 48 hours co-cultures of expanded Vγ9Vδ2-T cells and CD1d+ MM or AML cell lines, using a checkerboard format with concentration ranges based on the EC50 values of each drug as a single agent including clinically relevant concentrations27 (cytotoxicity dose-response curves (including EC50s) for each drug are provided in online supplemental figure S1A–C). Due to the scarcity of well-characterized and relevant MDS cell lines,28 we here focused on CD1d+ MM or AML cell lines. By calculating Bliss synergy scores over the range of concentrations,24 we assessed whether the drug combinations yielded cytotoxicity that was either synergistic (Bliss scores >0) or antagonistic (Bliss scores <0). Combinations of the CD1d-Vδ2 bsTCE with azacytidine, venetoclax and dexamethasone showed no clear synergistic or antagonistic effects with scores predominantly around 0, indicating additive effects (online supplemental figure S1D). However, co-treatment with bortezomib resulted in a predominant antagonistic effect across all concentrations (figure 1A), while daratumumab showed antagonistic effects specifically when concentrations exceeded EC50 (online supplemental figure S1D), likely due to fratricide of CD38-expressing Vγ9Vδ2-T cells (online supplemental figure S1E). Interestingly, the combination of the CD1d-Vδ2 bsTCE with the IMiD LEN showed an overall synergistic effect on tumor cell lysis, particularly when suboptimal (below EC50) concentrations of the CD1d-Vδ2 bsTCE were used (figure 1B). As the CD1d-Vδ2 bsTCE selectively engages both Vγ9Vδ2-T and type 1 NKT cells (EC50s in low-pM and low-nM range, respectively, online supplemental figure S1A,F), we additionally assessed the combination of LEN and the CD1d-Vδ2 bsTCE in co-cultures with type 1 NKT and CD1d+ tumor cells. In line with the results obtained with Vγ9Vδ2-T cells, co-treatment with LEN showed an overall synergistic effect on tumor cell lysis (online supplemental figure S1F). Lenalidomide boosts CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T cell activation, cytokine production, and target cell lysis. (A–B) Bliss synergy scores (synergistic, scores >0; antagonistic, scores <0) for specific lysis of MM.1s.CD1d cells after 48 hours co-cultures with Vγ9Vδ2-T cells (ratio 1:1)±concentration range of CD1d-Vδ2 bsTCE and concentration range of bortezomib (A, n=3) or lenalidomide (B, n=3); right panel shows specific lysis of MM1s.CD1d cells±1.5 pM CD1d-Vδ2 bsTCE±1.25 µM lenalidomide (corresponding to Bliss synergy score of 19.3)). (C–F) CD25 expression on Vγ9Vδ2-T cells (C, representative flow cytometry plots), CD107a expression on Vγ9Vδ2-T cells (D, representative histograms, gMFI), secreted granzyme B (E, n=3), and specific lysis of MM1s.CD1d cells (F, n=7) after 18 hours co-culture of LEN pretreated (1 µM for 72 hours) PBMC and MM.1s.CD1d cells (ratio 10:1)±100 pM ((C–E) or concentration range (F) of CD1d-Vδ2 bsTCE. (G) IL-2 expression in CD4+ αβ-T cells, CD8+ αβ-T cells and Vγ9Vδ2-T cells after 4 hours co-cultures of LEN pretreated (1 µM for 72 hours) PBMC and MM.1s.CD1d cells (ratio 10:1)±PMA/ionomycin or 1 nM CD1d-Vδ2 bsTCE (n=4). (H) Cytokine secretion by freshly isolated Vγ9Vδ2-T cells (>85% purity) after 48 hours co-culture with MM.1s.CD1d cells (ratio 1:1)±1 µM LEN and/or 100 pM CD1d-Vδ2 bsTCE (n=4). Box-and-whisker plots indicate median, 25th–75th percentiles, and minimum-maximum (F), bar graphs indicate mean±SD. (G) Bliss independence model with SynergyFinder R package (A), paired two-tailed t-test (D), non-linear regression with 95% confidence bands (dotted lines) and EC50s depicted (E), two-way ANOVA with Sidák’s multiple comparisons test (E, F), one-way ANOVA with Tukey’s multiple comparisons test (B left panel, G). *p<0.05; **p<0.01. See also online supplemental figures S1 and S2. ANOVA, analysis of variance; bsTCE, bispecific T-cell engagers; EC50, half maximal effective concentrations; gMFI, geometric mean fluorescence intensity; IFN, interferon; IL, interleukin; LEN, lenalidomide; MM, multiple myeloma; PBMC, peripheral blood mononuclear cells; PMA, phorbol myristate acetate; TNF, tumor necrosis factor. The reported pleiotropic antitumor effect of IMiDs, ranging from direct tumor cytotoxicity to enhanced antitumor immunity,29 30 prompted us to further explore the largely unknown effects of LEN on Vγ9Vδ2-T cell effector functions. While analyses thus far focused on cultured (expanded) Vγ9Vδ2-T cells, we next evaluated the effect of LEN using healthy donor-derived PBMC (ie, non-enriched Vγ9Vδ2-T cells). PBMC were pretreated for 72 hours with a clinically relevant concentration of LEN (1 µM; achieved at 25 mg oral daily dose),27 and subsequently co-cultured 18 hours with CD1d+ MM cells and the CD1d-Vδ2 bsTCE. LEN treatment resulted in a significant enhancement of CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell activation (CD25 upregulation, figure 1C), degranulation (CD107a upregulation; figure 1D, granzyme B secretion; figure 1E), and subsequent CD1d+ MM cell lysis (figure 1F, online supplemental figure S1G for gating strategy). In the absence of the CD1d-Vδ2 bsTCE, LEN pretreatment did not affect Vγ9Vδ2-T cell activation, degranulation, or tumor cell lysis. LEN pretreatment of PBMC enhanced IL-2 expression in CD4+ and CD8+-T cells on stimulation with PMA/ionomycin (figure 1G, online supplemental figure S2A for gating strategy). Interestingly, on CD1d-Vδ2 bsTCE exposure, also IL-2 production in Vγ9Vδ2-T cells was selectively and significantly potentiated when PBMC had been pretreated with LEN (figure 1G), indicating that the known LEN-induced potentiation of IL-2 production is not restricted to αβ-T cells.30 31 Although LEN did not significantly affect the cytokine profile in supernatants from 48 hours co-cultures of CD1d-Vδ2 bsTCE-exposed PBMC and CD1d+ MM cells (online supplemental figure S2B), it clearly affected the cytokine release profile in 48 hours co-cultures of Vγ9Vδ2-T cells freshly isolated from PBMC and CD1d+ MM cells. As shown in figure 1H and online supplemental figure S2C, LEN itself did not trigger cytokine release, but it significantly enhanced the CD1d-Vδ2 bsTCE-induced release of IL-2, tumor necrosis factor (TNF), and interferon (IFN)-γ by Vγ9Vδ2-T cells in a (CD1d-Vδ2 bsTCE) dose-dependent manner. Of note, secretion of IL-10 was significantly reduced in the presence of LEN, while IL-4 secretion was low and did not significantly change. No secretion of IL-6 or IL-17 was detected. Lenalidomide enhances CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell expansion. (A–D) Ki-67 expression (A left panel, n=19) and representative histogram of CFSE-dilution (A right panel, n=10), fold expansion (B, representative flow cytometry plots, n=39; CD1d-Vδ2 bsTCE; 8.3±41.6 vs CD1d-Vδ2 bsTCE+lenalidomide; 30.8±73.6; median±IQR), activating and inhibitory marker/receptor expression (C, n=10–27), and Vγ9Vδ2-T cell differentiation state (D, naïve (N; CD27+ CD45RA+), central memory (CM; CD27+ CD45RA−), effector memory (EM; CD27− CD45RA−), or terminally differentiated (TEMRA; CD27− CD45RA+), n=6) after 7-day co-culture of PBMC with MM.1s.CD1d cells (ratio 10:1)±1 µM LEN and/or 10 nM CD1d-Vδ2 bsTCE (D, asterisks are shown compared with negative control). (E) CD107a expression of Vγ9Vδ2-T cells and specific lysis of C1R.CD1d cells after 18 hours co-culture with PBMC (ratio 1:1) that were co-cultured for 7 days with MM.1s.CD1d cells (ratio 10:1)±1 µM LEN and/or 10 nM CD1d-Vδ2 bsTCE (n=4). (F–G) CD4 expression on Vγ9Vδ2-T cells (F, percentage of CD4+ cells within total Vγ9Vδ2-T cell population, n=17) and CD4 and CD8 expression on Vγ9Vδ2-T cells (G, n=17) after 7-day co-culture of PBMC with MM.1s.CD1d cells (ratio 10:1)±1 µM LEN and/or 10 nM CD1d-Vδ2 bsTCE. (G, asterisks are shown compared with negative control). (H) Cytokine expression of CD4−CD8−, CD4+, CD8+ Vγ9Vδ2-T cells after 4 hours stimulation with PMA/ionomycin of PBMC that were co-cultured for 7 days with MM.1s.CD1d cells (ratio 10:1) in the presence of 1 µM LEN and 10 nM CD1d-Vδ2 bsTCE (n=7). Box-and-whisker plots indicate median, 25th–75th percentiles, and minimum−maximum. One-way ANOVA with Tukey’s multiple comparisons test (A–C, E, F, H), two-way ANOVA with Šídák’s multiple comparisons test (D, G). *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. See also online supplemental figure S4. ANOVA, analysis of variance; bsTCE, bispecific T-cell engagers; CFSE, carboxyfluorescein succinimidyl ester; CM, central memory; EM, effector memory; IFN, interferon; IL, interleukin; LEN, lenalidomide; MM, multiple myeloma; PBMC, peripheral blood mononuclear cells; PMA, phorbol myristate acetate; TEMRA, terminally differentiated; TNF, tumor necrosis factor. As secreted IL-2 may stimulate Tregs in the TME,32 we assessed Treg activation in 7-day co-cultures of PBMC and CD1d+ MM cells. No activation and/or proliferation of activated (FoxP3++ CD45RA−) or resting (FoxP3+CD45RA+) Tregs25 was observed on treatment with LEN and/or the CD1d-Vδ2 bsTCE (online supplemental figure S3A–C for gating strategy). Collectively, these results demonstrate that LEN potentiates CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell effector functions, which results in enhanced degranulation, skewed Th1-type cytokine secretion, and increased antitumor activity. Lenalidomide enhances CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell proliferation and expansion Since we previously demonstrated CD1d-Vδ2 bsTCE induced expansion of Vγ9Vδ2-T cells,8 we next investigated the effect of LEN on CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell proliferation. To this end, PBMC were co-cultured with CD1d+ MM cells for 7 days in the presence or absence of LEN and/or the CD1d-Vδ2 bsTCE. While the CD1d-Vδ2 bsTCE induced a strong proliferative response in Vγ9Vδ2-T cells, as reflected by increased Ki-67 expression, CFSE dilution, and fold expansion (8.3±41.6; median±IQR), proliferation of Vγ9Vδ2-T cells was significantly boosted by LEN, resulting in a mean fold expansion of 30.8 (±73.6; median±IQR, p<0.0001; figure 2A,B). Consistent with a previous report,33 LEN treatment alone did not trigger Vγ9Vδ2-T cell proliferation. Of interest, LEN also enhanced the CD1d-Vδ2 bsTCE-induced expansion of type 1 NKT cells in these co-cultures (online supplemental figure S4A). Vγ9Vδ2-T cells expanded using the CD1d-Vδ2 bsTCE expressed significantly higher levels of the activation markers CD25 and HLA-DR, and this was further potentiated by LEN (figure 2C). Moreover, while expression of the activating/co-stimulatory receptors DNAM-1 and NKG2D was generally high on Vγ9Vδ2-T cells, expression of DNAM-1 was further increased in PBMC cultured with the combination of CD1d-Vδ2 bsTCE and LEN. In parallel, expression of the co-stimulatory molecules CD80 and CD86—typically expressed on professional antigen presenting cells (APCs) and linked to the ability of Vγ9Vδ2-T cells to present antigens and prime αβ-T cells9 34—was also upregulated on LEN co-treatment (online supplemental figure S4B). An increase in the expression of inhibitory receptors PD-1 and NKG2A was observed on Vγ9Vδ2-T cells cultured in the presence of the CD1d-Vδ2 bsTCE, though this did not increase further on LEN co-treatment (figure 2C). While this upregulation of PD-1 and NKG2A may be reflective of activation, it could also be a sign of T-cell exhaustion. Of note, expression of the terminal differentiation marker CD57 remained low and was even significantly downregulated on LEN co-treatment (figure 2C). CD1d-Vδ2 bsTCE and LEN combination treatment significantly increased the proportion of central memory Vγ9Vδ2-T cells, whereas almost no Vγ9Vδ2-T cells with a terminally differentiated phenotype were observed, supporting an activated, more stem-like phenotype (figure 2D, online supplemental figure S4C). Finally, on subsequent rechallenge with CD1d+ tumor cells, expanded Vγ9Vδ2-T cells present in these co-cultures readily degranulated and triggered robust tumor cell lysis (figure 2E), thus negating the induction of exhaustion. The majority of Vγ9Vδ2-T cells in PB are negative for the classical T-cell markers CD4 and CD8, although small fractions of Vγ9Vδ2-T cells express these co-receptors.35 Interestingly, after the 7-day co-culture of PBMC with CD1d+ MM cells, CD1d-Vδ2 bsTCE treatment induced a significant increase in relatively dim CD4-expressing Vγ9Vδ2-T cells and this was substantially further enhanced in the presence of LEN (figure 2F,G, online supplemental figure S4D,E for representative plots). As Vγ9Vδ2-T cells can trogocytose membrane fragments on immune synapse formation,36 this could possibly account for the acquisition of CD4 by Vγ9Vδ2-T cells following CD1d-Vδ2 bsTCE-induced cross-linking with CD4 co-expressing CD1d+ PBMC (eg, monocytes) or CD1d+ MM cells. As Vγ9Vδ2-T cells did not acquire relevant levels of the monocyte marker CD14 during the co-culture (online supplemental figure S4F), the MM cells did not express CD4, and CD4 expression was still acquired by Vγ9Vδ2-T cells when PBMC were cultured in the absence of tumor cells (online supplemental figure S4G,H), trogocytosis was considered an unlikely cause of the induced expression of CD4 on Vγ9Vδ2-T cells. As pre-culture depletion of CD4+ cells did not affect the induction of CD4dim-expressing Vγ9Vδ2-T cells (online supplemental figure S4I), this suggested that Vγ9Vδ2-T cells that had expanded under these conditions might have acquired membrane CD4dim expression de novo, a phenomenon previously reported in CD8+ αβ-T cells on strong activation.37 38 Alternatively, the emergence of a CD4dim population of Vγ9Vδ2-T cells may reflect preferential expansion of the minor (≤1%) CD4+ Vγ9Vδ2-T cell subset that escaped magnetic bead depletion. We next evaluated potential differences in phenotype or cytokine profiles between CD4+, CD8+, and CD4−CD8− double negative Vγ9Vδ2-T cells. When PBMC were stimulated with PMA/ionomycin after the 7-day co-culture with CD1d-Vδ2 bsTCE and LEN, the proportion of Vγ9Vδ2-T cells positive for TNF or IFN-γ was similar between CD4+, CD8+, and CD4−CD8− Vγ9Vδ2-T cells (figure 2H), consistent with a previous report.39 However, single positive CD4+ and CD8+ Vγ9Vδ2-T cells contained a larger proportion of IL-2 producing Vγ9Vδ2-T cells. Individual Vγ9Vδ2-T cells frequently expressed multiple cytokines, most commonly the combination of TNF and IFN-γ (online supplemental figure S4J). No significant differences in expression of CD25, HLA-DR, CD80, CD86, and CTLA-4, the proliferative marker Ki-67, or in CFSE dilution were noted between CD4+, CD8+, and CD4−CD8− Vγ9Vδ2-T cells after the 7-day cultures (online supplemental figure S4K,L). Overall, these findings demonstrate that LEN potently boosts CD1d-Vδ2 bsTCE-mediated expansion of polyfunctional Vγ9Vδ2-T cells with a highly activated phenotype, which are enriched for Vγ9Vδ2-T cells that express membrane CD4. Lenalidomide boosts Vγ9Vδ2-T cell immunity via CD28 phosphorylation, Notch signaling, and IL-2 release IMiDs, including LEN, have a reported immunostimulatory effect on αβ-T cells through activation of the CD28 axis40 and induction of cereblon (CRBN)-mediated degradation of Ikaros (IKZF1) and Aiolos (IKZF3),30 which relieves the suppression of Notch signaling and IL-2 expression, thereby promoting T-cell function and expansion.31 To refine our understanding of the observed immune stimulatory effect of LEN on bsTCE-activated Vγ9Vδ2-T cells, we first aimed to explore the role of IL-2, as we found that LEN co-treatment enhanced secretion of IL-2 (figure 1G,H, online supplemental figure S2C), a key driver of T-cell proliferation. In 7-day co-cultures of PBMC and CD1d+ MM cells, neutralization of IL-2 completely abrogated the CD1d-Vδ2 bsTCE-mediated expansion of Vγ9Vδ2-T cells regardless of the presence of LEN (figure 3A). While these data underscored the pivotal role of IL-2 in supporting Vγ9Vδ2-T cell expansion in this setting, we did note that LEN still had an additional stimulatory effect on CD1d-Vδ2 bsTCE-mediated Vγ9Vδ2-T cell expansion when exogenous IL-2 was added in surplus (figure 3B), suggesting an effect beyond autocrine (or paracrine) IL-2 signaling. Lenalidomide boosts CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T cell functions via CD28 phosphorylation and Notch signaling. (A, B) Vγ9Vδ2-T cell expansion (A, n=3; B, n=10) after 7-day co-culture of PBMC with MM.1s.CD1d cells (ratio 10:1)±1 µM LEN, 10 nM CD1d-Vδ2 bsTCE, 10 µg/mL neutralizing human IL-2 antibody (αIL-2), and/or 100 IU/mL rhIL-2. (C) Representative histograms (n=3 independent experiments) depicting Tyr218 CD28 phosphorylation of Vγ9+ T cells in PBMC following 15-minute co-culture±10 nM CD1d-Vδ2 bsTCE and/or 1 µM LEN. (D–E) Vγ9Vδ2-T cell percentage (of total T cells) after 7-day co-culture (D, n=14) and IL-2 secretion (fold change relative to CD1d-Vδ2 bsTCE condition) after 24 hours co-culture (E, n=12) of PBMC with MM.1s.CD1d cells (ratio 10:1)±1 µM LEN, 10 nM CD1d-Vδ2 bsTCE, 10 µg/mL CTLA-4 Ig (abatacept). (F) qPCR analysis showing relative mRNA expression levels of Hes1 after 24 hours co-culture of expanded Vγ9Vδ2-T cells (purity >96%)±1 µM LEN. The 2–ΔΔCT method was applied to calculate the relative expression levels and normalized to GAPDH. n=3 in duplicate. (G–J) Vγ9Vδ2-T cell percentage (of total T cells) after 7-day co-culture (G, n=12; I, n=12) and IL-2 secretion (fold change relative to CD1d-Vδ2 bsTCE condition) after 24 hours co-culture (H, n=8; J, n=6) of PBMC with MM.1s.CD1d cells (ratio 10:1)±1 µM LEN, 10 nM CD1d-Vδ2 bsTCE, 10 µg/mL CTLA-4 Ig (abatacept), and/or 10 µM GSI (RO429097). Box-and-whisker plots indicate median, 25th–75th percentiles, and minimum−maximum. One-way ANOVA with Tukey’s multiple comparisons test (B, D–I), paired t-test (F). *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. See also online supplemental figure S5. ANOVA, analysis of variance; bsTCE, bispecific T-cell engagers; CTLA-4, cytotoxic T-lymphocyte associated protein 4; GSI, γ-secretase inhibitor; IL, interleukin; LEN, lenalidomide; MM, multiple myeloma; mRNA, messenger RNA; PBMC, peripheral blood mononuclear cells; qPCR, quantitative PCR. Since LEN showed no effect on Vγ9Vδ2-T cell activation and expansion without CD1d-Vδ2 bsTCE stimulation, and is known to trigger CD28 phosphorylation in αβ-T cells,40 we hypothesized a co-stimulatory role for LEN in bsTCE-mediated activation of Vγ9Vδ2-T cells. Indeed, we found that LEN initiated CD28 phosphorylation in Vγ9+-T cells within 15 min (figure 3C, online supplemental figure S5A), which was substantially potentiated when combined with the CD1d-Vδ2 bsTCE, possibly as a result of CD1d-Vδ2 bsTCE-induced crosslinking between Vδ2+ and CD1d+ cells in PBMC (ie, monocytes, B cells and circulating dendritic cells expressing CD28 ligands). Vγ9Vδ2-T cell membrane expression of CD28 was not different between the culture conditions (online supplemental figure S5B). Of interest, while blocking CD28 ligation using recombinant CTLA-4-Ig (abatacept) impaired the enrichment of Vγ9Vδ2-T cells and IL-2 secretion in PBMC cultured with CD1d-Vδ2 bsTCE alone, it did not significantly impact Vγ9Vδ2-T cell enrichment nor IL-2 secretion when PBMC were cultured with both LEN and the CD1d-Vδ2 bsTCE (figure 3D,E). This indicates that LEN bypassed CTLA-4 Ig inhibition, likely through direct phosphorylation of CD28, thereby bypassing, at least in part, the need for CD28 ligation. As LEN has been reported to enhance intracellular Notch signaling,31 we next explored the role of Notch activation in LEN-enhanced Vγ9Vδ2-T cell responses on bsTCE engagement. Using qPCR analysis, we first showed that LEN directly activates Notch signaling in Vγ9Vδ2-T cells as shown by increased expression of Hes1, a primary downstream target gene of the canonical Notch signaling pathway41 42 (figure 3F). In line, inhibition of Notch signaling using a GSI, which prevents proteolytic release of the Notch intracellular domain,31 and CB-103, which blocks downstream Notch transcriptional complex activation,43 not only attenuated the enrichment of Vγ9Vδ2-T cells and IL-2 secretion in cultures of PBMC and the CD1d-Vδ2 bsTCE alone, it also did this in LEN-containing PBMC cultures (figure 3G,H and online supplemental figure S5CS5C). Notably, when CTLA-4 Ig-mediated inhibition was combined with Notch blockade, LEN was no longer able to overcome the suppressive effect of CTLA-4 Ig (figure 3I,J), further supporting that Notch signaling, at least in part, contributes to LEN’s mode of action. In the presence of rapamycin, an inhibitor of the mTOR kinase that integrates signaling downstream of the TCR, CD28, and IL-2 receptor,44 the CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell enrichment was completely abrogated, irrespective of LEN co-treatment (online supplemental figure S5D). Collectively, these data support that CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T cell effector functions are boosted by LEN via enhanced intracellular CD28 phosphorylation, activation of Notch signaling, and IL-2 release. Lenalidomide potentiates CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T cell activation and expansion in MDS patient-derived bone marrow samples LEN is of therapeutic importance both in MM and MDS with 5q deletion (del(5q)) by having an immunomodulatory effect and by inducing apoptosis of malignant cells via induction of IKZF1/3 and casein kinase 1A1 (CK1α) degradation, respectively.30 45 In MM, combination treatment with LEN and CD3-directed bsTCE has been shown to boost bsTCE efficacy and is currently being investigated in multiple phase 3 clinical trials.46 Although early clinical reports in MDS suggest that bsTCE treatment could be a valuable therapeutic option,47 combination therapy with LEN to overcome the severely dysregulated immune system has not yet been studied. Therefore, we investigated the effect of LEN on CD1d-Vδ2 bsTCE-activity in MDS patient samples. First, we assessed frequency and phenotype of Vγ9Vδ2-T cells in PB of patients with MDS that were LEN therapy naïve (LEN-naïve, n=5) or on LEN-maintenance therapy (LEN-exposed, n=3; 5–10 mg once daily) (patient characteristics in online supplemental table S1). The Vγ9Vδ2-T cell frequency was within the range reported in healthy donor PBMC23 for both LEN-naïve and LEN-exposed patients (LEN-naïve: 1.04%±0.52 vs LEN-exposed: 4.71%±6.08, mean±SD; figure 4A). The proportion of Vγ9Vδ2-T cells expressing the activation markers CD69 and CD25 was overall low in both patient groups and immune checkpoint expression (programmed cell death protein 1 (PD-1), T cell immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domain (TIGIT), and T cell immunoglobulin and mucin domain-3 (TIM-3), but not cytotoxic T-lymphocyte associated protein 4 (CTLA-4)) was variable and not statistically significantly different between LEN-exposed and LEN-naïve patients (figure 4B). A similar pattern of expression was observed in αβ-T cells (online supplemental figure S6A). The majority of Vγ9Vδ2-T and αβ-T cells in these PBMC samples displayed a central memory phenotype, with a generally low fraction of terminally differentiated Vγ9Vδ2-T cells (figure 4C, online supplemental figure S6B). We next explored the ability of the CD1d-Vδ2 bsTCE to trigger Vγ9Vδ2-T cell activation and expansion in these MDS PB samples. In the majority of samples, the CD1d-Vδ2 bsTCE triggered Vγ9Vδ2-T cell specific activation and expansion to a comparable level as observed in healthy donors in 7-day cultures (figure 4D,E online supplemental figure S6C), which was in most cases (and consistently in LEN-exposed patients) potentiated by LEN co-treatment. Lenalidomide potentiates CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell expansion and activation in MDS patient samples. (A–C) Vγ9Vδ2-T cell frequency (A, % within total T-cell population), CD69, CD25, PD-1, TIGIT, TIM-3, and CTLA-4 expression (B), and differentiation state (C, naïve (N; CD27+ CD45RA+), central memory (CM; CD27+ CD45RA−), effector memory (EM; CD27− CD45RA−), or terminally differentiated (TEMRA; CD27− CD45RA+)) of Vγ9Vδ2-T cells in PBMC samples of MDS patients that were either LEN-naïve (LEN-naïve, n=5) or receiving LEN therapy (LEN-exposed, n=3) (horizontal line depicts the mean). (D, E) Vγ9Vδ2-T cell CD25 expression (D) and fold expansion (E) after 7-day co-culture of LEN-naïve (n=5) versus LEN-exposed (n=3) MDS patient-derived PBMC with±1 µM LEN and/or 10 nM CD1d-Vδ2 bsTCE. (F) Vγ9Vδ2-T cell frequency (% within total T-cell population) in bone marrow samples of patients with MDS (horizontal line depicts the mean). (G, H) CD25 expression on Vγ9Vδ2-T cells (G), and representative flow cytometry plots (H, left panel) and Vγ9Vδ2-T cell fold expansion (H, right panel) after 7-day co-culture of MDS patient-derived BMMC (n=5) with±1 µM LEN and/or 10 nM CD1d-Vδ2 bsTCE. (I) CD1d expression on blast cells in MDS patient-derived BMMC (n=5, geometric mean fluorescence intensity index). (J, K) CD107a on Vγ9Vδ2-T cells (J, %) and specific blast cell lysis (K) after 18 hours co-culture of MDS patient-derived BMMC (n=4) with Vγ9Vδ2-T cells (1:2 effector:BMMC ratio) ±10 nM CD1d-Vδ2 bsTCE. Box-and-whisker plots indicate median, 25th–75th percentiles, and minimum−maximum (G, H). Two-way ANOVA with Sidák’s multiple comparisons test (D, E), one-way ANOVA with Tukey’s multiple comparisons test (G, H), and paired t-test (J, K). *p<0.05; **p<0.01. See also online supplemental figure S6. ANOVA, analysis of variance; APC, antigen presenting cell; BMMC, bone marrow mononuclear cell; bsTCE, bispecific T-cell engager; CM, central memory; CTLA-4, cytotoxic T-lymphocyte associated protein 4; EM, effector memory; LEN, lenalidomide; MDS, myelodysplastic syndromes; PBMC, peripheral blood mononuclear cells; PD-1, programmed cell death protein 1; TEMRA, terminally differentiated; TIGIT, T cell immunoglobulin and immunoreceptor tyrosine-based inhibitory motif domain; TIM-3, T cell immunoglobulin and mucin domain-3. Next, we assessed Vγ9Vδ2-T cell functionality in MDS patient-derived bone marrow samples. Although Vγ9Vδ2-T cells were present in MDS BMMC (0.49%±0.44, % of total T cells; mean±SD; figure 4F, patient characteristics in online supplemental table S1) and CD1d-Vδ2 bsTCE treatment triggered Vγ9Vδ2-T cell activation, no expansion was observed in 7-day cultures (figure 4G,H). However, co-treatment with LEN did allow for CD1d-Vδ2 bsTCE-mediated Vγ9Vδ2-T cell expansion and further boosted activation. In co-cultures with healthy donor-derived BMMC, the CD1d-Vδ2 bsTCE alone induced robust Vγ9Vδ2-T cell activation (online supplemental figure S6E) and consistent Vγ9Vδ2-T cell expansion when combined with LEN (online supplemental figure S6F). Of note, MDS blast viability was too low in the 7-day MDS BMMC cultures to reliably determine specific tumor lysis. However, in overnight co-cultures of BMMC from patients with MDS and healthy donor-derived Vγ9Vδ2-T cells, the CD1d-Vδ2 bsTCE induced robust Vγ9Vδ2-T cell activation and lysis of CD1d+ MDS blast cells (figure 4I–K). Overall, these data demonstrate that LEN potentiates CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell activation and expansion in bone marrow samples of both healthy donors and patients with MDS. Discussion Due to its ability to induce durable remissions in patients with various tumor types, bsTCE treatment has generated considerable excitement. Unfortunately, primary and secondary resistance are common and bsTCE use is often associated with toxicities that can be severe.2 Selective engagement of innate-like cells, such as Vγ9Vδ2-T and type 1 NKT cells, may allow for a favorable balance between efficacy and safety.48 We previously demonstrated a bsTCE, specific for both CD1d and the Vδ2-TCR, to have trispecific properties by engaging not only Vγ9Vδ2-T but also type 1 NKT cells to CD1d+ tumors in AML, MM and CLL models.8 In the present study, we explored how different standard-of-care drugs used for the treatment of MM and AML/MDS could shape CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T-cell effector functions. Intriguingly, we found that the IMiD LEN—approved for the treatment of patients with MM, del(5q) MDS, and several lymphomas, and known to exert pleiotropic antitumor effects ranging from direct tumor cytotoxicity to enhanced antitumor immunity29 30—synergized with the CD1d-Vδ2 bsTCE to enhance Vγ9Vδ2-T cell activation, expansion, proinflammatory cytokine secretion, degranulation, and subsequent cytotoxicity. Although we focused primarily on Vγ9Vδ2-T cells in this study, we also found LEN to enhance CD1d-Vδ2 bsTCE-induced type 1 NKT cell proliferation and antitumor activity, in accordance with work demonstrating LEN to enhance α-galactosylceramide induced type 1 NKT cell stimulation.48 LEN was shown to enhance IL-2 production by αβ-T cells via CRBN-mediated proteasomal degradation of IKZF1 and IKZF3.30 31 We here report that the enhanced IL-2 production induced by LEN is not restricted to αβ-T cells but also extends to Vγ9Vδ2-T cells. In parallel, LEN co-treatment enhanced IFN-γ and TNF release on CD1d-Vδ2 bsTCE exposure, whereas the already low levels of IL-10 were further reduced, highlighting that LEN can favorably modulate the immune microenvironment. Moreover, we observed no secretion of IL-6 on CD1d-Vδ2 bsTCE and/or LEN exposure, which may have clinical relevance as IL-6 has been identified as a key driver of CRS in patients treated with CD3-bsTCE and chimeric antigen receptor (CAR)-T cell therapy.49 Importantly, despite the observed enhancement of IL-2 secretion, Tregs, known to suppress bsTCE-mediated antitumor activity,3 were not stimulated following CD1d-Vδ2 bsTCE and LEN co-treatment. Phenotypic analysis of Vγ9Vδ2-T cells cultured in the presence of LEN and the CD1d-Vδ2 bsTCE revealed enhanced expression of multiple activation markers and the agonistic receptor DNAM-1. In addition, we observed enhanced expression of inhibitory receptors PD-1 and NKG2A on LEN and CD1d-Vδ2 bsTCE exposure. Although expression of PD-1 and NKG2A is often linked to exhaustion,50 51 it here likely represented an activated state of the Vγ9Vδ2-T cells, as CD1d-Vδ2 bsTCE and LEN-treated Vγ9Vδ2-T cells remained highly functional as evidenced by their ability to readily degranulate and induce robust tumor cell lysis. In support, the expression of NKG2A has previously been associated with highly functional Vγ9Vδ2-T cells in response to tumors.52 This aligns with our observation that most Vγ9Vδ2-T cells exhibited a predominant central memory and not a terminally differentiated phenotype. Moreover, LEN co-treatment downregulated expression of CD57, which is associated with T-cell senescence and functional deficiency.53 The combination of CD1d-Vδ2 bsTCE and LEN also resulted in a Vγ9Vδ2-T cell population significantly enriched for (relatively dim) membrane expression of CD4. As de novo CD4dim expression was reported on highly activated and cytolytic CD8+ αβ-T cells,37 38 the observed CD4 expression may similarly reflect an activated Vγ9Vδ2-T-cell phenotype induced by the combination of CD1d-Vδ2 bsTCE and LEN. Of particular interest was our observation that the expression of B7 ligands CD80 and CD86, typically expressed on professional APCs, was upregulated on Vγ9Vδ2-T cells exposed to CD1d-Vδ2 bsTCE and most pronounced when this was combined with LEN. In CD8+ αβ-T cells, CD80/CD86 were shown to evoke cell-autonomous CD28 co-stimulation in cis,54 translating into enhanced T-cell function and antitumor immunity. It is plausible that B7 ligands could have a similar effect on Vγ9Vδ2-T cells. Moreover, as Vγ9Vδ2-T cells can effectively present antigens and prime αβ-T cells,9 34 the LEN-induced upregulation of CD80 and CD86 could potentially enhance their APC functionality. To elucidate the mechanism by which LEN potentiated the immune-stimulatory effects of bsTCE-activated Vγ9Vδ2-T cells, we first investigated the critical role of IL-2 in Vγ9Vδ2-T cell activity. Neutralization of IL-2 completely abrogated Vγ9Vδ2-T cell expansion and LEN was unable to rescue this. However, in the presence of exogenous IL-2, we still found LEN to significantly enhance CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell expansion, suggesting an effect of LEN beyond IL-2 signaling. Indeed, LEN induced intracellular CD28 phosphorylation, consistent with previous findings in conventional αβ-T cells.40 In conjunction with TCR stimulation, CD28 promotes cytokine expression with specifically high levels of IL-2, induced resistance to apoptosis, and expansion of T cells.55 In line with this, we demonstrated that blockade of CD28 ligation using CTLA-4 Ig impaired CD1d-Vδ2 bsTCE-triggered Vγ9Vδ2-T cell proliferation and IL-2 secretion, confirming the critical role of CD28 co-stimulation in Vγ9Vδ2-T cell functionality. Importantly, however, in the presence of LEN, CTLA-4 Ig-mediated CD28 blockade no longer suppressed Vγ9Vδ2-T cell activity, suggesting that LEN effectively bypassed the need for extracellular CD28 ligation. This could be of particular relevance in the immune suppressed TME where inadequate co-stimulation is a key factor limiting bsTCE activity.56 LEN has been reported to enhance T-cell immunity through potentiation of Notch signaling,31 an effect now confirmed by us in Vγ9Vδ2-T cells. As Vγ9Vδ2-T cell enrichment and IL-2 secretion induced by the CD1d-Vδ2 bsTCE and LEN were abrogated by inhibition of Notch signaling, these findings support a contributory role for Notch signaling in the immunostimulatory effects of LEN on Vγ9Vδ2-T cells. Finally, the potential clinical relevance of our findings was validated in MDS patient samples where we found LEN to promote CD1d-Vδ2 bsTCE-induced Vγ9Vδ2-T cell expansion in PBMC and BMMC. LEN is thought to be of therapeutic importance in MDS in part by inducing CK1α degradation, which is particularly relevant in MDS cells with del(5q) as these cells express CK1α at haploinsufficient levels.45 LEN further reduces CK1α levels, resulting in selective TP53-dependent apoptosis of dysplastic blast cells. Yet, also a proportion of non-del(5q) patients with MDS responds to LEN treatment which has been attributed to its immunomodulatory effects.57 This aligns with our observation that circulating Vγ9Vδ2-T and αβ-T cells in patients with MDS that received LEN-maintenance therapy had a general low expression of activation markers and inhibitory immune checkpoint receptors (PD-1, TIGIT, and TIM-3). Although our results should be interpreted with caution given the low patient number, they are supported by reports demonstrating reduced immune checkpoint receptor expression following LEN-maintenance therapy in patients with AML, MM and CLL.5862 In healthy donor and MDS patient bone marrow samples, Vγ9Vδ2-T cells could be activated but failed to expand in response to CD1d-Vδ2 bsTCE exposure. Importantly, LEN not only promoted Vγ9Vδ2-T cell activation but also allowed for their proliferation in response to the CD1d-Vδ2 bsTCE in these bone marrow samples. As dysplastic blast cells in the MDS bone marrow samples expressed CD1d and the CD1d-Vδ2 bsTCE triggered allogeneic Vγ9Vδ2-T cell-mediated blast lysis, it is conceivable that combination treatment with LEN and the CD1d-Vδ2 bsTCE could elicit immunostimulatory effects while also promoting direct lysis of dysplastic blast cells, independent of del(5q) status. Future (clinical) studies are needed to determine the precise role of combination treatment with IMiDs and CD1d-Vδ2 bsTCE in patients with MDS. Collectively, we present preclinical data demonstrating that LEN co-stimulates Vγ9Vδ2-T cells through a cascade involving CD28, Notch signaling, and IL-2 release, leading to highly functional Vγ9Vδ2-T cells upon bispecific engagement which translates to enhanced activity in MDS patient samples. Our findings enhance our understanding of the mechanism of action of LEN and provide a new rationale for combining Vγ9Vδ2-T cell-directed therapy with the IMiD LEN. Supplementary material