Disulfiram bolsters T‐cell anti‐tumor immunity through direct activation of LCK‐mediated TCR signaling

Abstract

Activation of the T‐cell antigen receptor (TCR)–CD3 complex is critical to induce the anti‐tumor response of CD8⁺ T cells. Here, we found that disulfiram (DSF), an FDA‐approved drug previously used to treat alcohol dependency, directly activates TCR signaling. Mechanistically, DSF covalently binds to Cys20/Cys23 residues of lymphocyte‐specific protein tyrosine kinase (LCK) and enhances its tyrosine 394 phosphorylation, thereby promoting LCK kinase activity and boosting effector T cell function, interleukin‐2 production, metabolic reprogramming, and proliferation. Furthermore, our in vivo data revealed that DSF promotes anti‐tumor immunity against both melanoma and colon cancer in mice by activating CD8⁺ T cells, and this effect was enhanced by anti‐PD‐1 co‐treatment. We conclude that DSF directly activates LCK‐mediated TCR signaling to induce strong anti‐tumor immunity, providing novel molecular insights into the therapeutic effect of DSF on cancer.

The FDA‐approved drug disulfiram, previously used to treat alcohol dependency, directly activates T‐cell receptor signaling to activate CD8⁺ T cells and exerts potent anti‐tumor effects in vivo and in vitro.

Introduction

The adaptive immune system is composed of highly specialized, systemic cells and processes that eliminate pathogens or prevent their growth (Cao, 2016). During adaptive immune responses, CD8⁺ T cells are the major type of lymphocytes, which play important roles in autoimmunity and tumor immunity (Brownlie & Zamoyska, 2013). Recognition of tumor antigen by TCR‐CD3 complex has to be correctly translated into signal transduction events necessary for the induction of an effective immune response. However, the activation of CD8⁺ T cells is suppressed in the tumor microenvironment (TME) to facilitate the immune escape of tumors (Mellman et al, 2011; Joyce & Fearon, 2015), therefore, it is of great clinical interest to find a way to efficiently promote the anti‐tumor response of CD8⁺ T cells (Roberts et al, 2016; Yang et al, 2016).

Based on the events in generating and regulating antitumor immunity, enhancing the TCR signaling of CD8⁺ T cell is one of the most attractive approaches to achieve therapeutic efficacy (Mellman et al, 2011). Tyrosine phosphorylation of the CD3 immunoreceptor tyrosine‐based activation motifs (ITAMs) is an early and essential step in TCR‐mediated T cell activation (Smith‐Garvin et al, 2009). Recent studies have focused on the mechanism that initiates and propagates the TCR signaling upon stimulation with antibody or antigenic ligands, which is a prerequisite for T cell response (Gillis & Watson, 1980; Weiss et al, 1984; Abraham & Weiss, 2004; Brownlie & Zamoyska, 2013; Shi et al, 2013). While it is still lacking a safe compound that can directly trigger TCR activation and promote T cell‐mediated anti‐tumor immunity for clinical application.

The SRC family protein tyrosine kinase LCK is the first kinase to phosphorylate TCR signaling in T cells, leading to the phosphorylation of CD3 ITAMs. Amounting evidence has demonstrated that LCK can function either be loaded on coreceptor CD4/CD8 through a “zinc‐clasp” structure or be free (Ehrlich et al, 2002; Irvine et al, 2002; Kim et al, 2003; Salmond et al, 2009). LCK is regulated by trans‐autophosphorylation, and by dephosphorylation of phosphatases, or conformational changes (Davis & van der Merwe, 2011). It is critical for controlling the activity of LCK in T cell‐mediated anti‐tumor immune responses (Bommhardt et al, 2019). Free LCK molecules are critical for initial TCR activation, and co‐receptor‐bound Lck [N‐terminal 20/23 Cysteine (Cys) (C20/23)] is recruited to the TCR complex at later time points (Terashima et al, 2020). A recent study showed that asparagine promotes the CD8⁺ T cell activation through the tyrosine 394 (Tyr394) site in the C‐terminal catalytic domain (Wu et al, 2021). However, the regulatory mechanism of the LCK N‐terminal domain in the TCR activation remains elusive.

The discovery of DSF (disulfiram, tetraethylthiuram disulfide) as a promising anti‐cancer drug raises new hope for drug repurposing. With well‐established and excellent safety profiles of DSF, it has been used for the treatment of alcohol dependence for over six decades. A number of in vitro and in vivo studies have indicated the anti‐tumor activity of DSF‐derived metabolites in different cancers, such as urinary bladder cancer, intestinal cancer, acute and chronic lymphocytic/myelocytic leukemia, breast cancer, colorectal cancer, non‐small cell lung cancer, ovarian cancer, and renal cancer (Jiao et al, 2016). Accumulating clinical trials were ongoing or completed that evaluate DSF’s anti‐cancer efficacy. These trials have tested the use of DSF for advanced solid malignancies and newly diagnosed non‐small cell lung cancer (Jiao et al, 2016), supporting the notion that DSF is a promising cancer‐therapeutic agent. A previous study focused on the direct tumor‐killing function of ditiocarb–copper complex [diethyldithiocarbamate (DTC)‐copper, CuET], which is a metabolite of DSF. However, the role of DSF in T cell immune response remains largely unknown. In the present study, we performed a high throughput screening assay, and identified the FDA‐approved alcohol‐abuse drug, DSF, directly activates TCR signaling through binding to LCK and enhancing its kinase activity, consequently triggering the immune response of CD8⁺ T cells and bolsters in vivo anti‐tumor immunity.

Results

DSF activates TCR signaling in CD8⁺ T cells

To discover whether any compounds could enhance TCR signaling, we performed a screen of 1,835 compounds to detect their effect on TCR‐mediated phosphorylation of the S6 ribosomal subunit (p‐S6), a critical downstream component of the TCR signaling that controls protein synthesis (Fig 1A) (Wang et al, 2016). By evaluating S6 phosphorylation in wild‐type (WT) mouse splenic cells treated with the compounds, we found that DSF significantly promoted the level of p‐S6 in both CD4⁺ and CD8⁺ T cells (Figs 1B and EV1A). Since CD8⁺ T cells play a central role in anti‐tumor immunity; therefore, we focused on exploring the function of DSF in CD8⁺ T cells. DSF pre‐treatment resulted in a dose‐dependent elevation of TCR‐triggered S6 phosphorylation (Fig 1C). Consistently, DTC‐sodium complex (NaDTC) as an active metabolic product of DSF, can also promote p‐S6 in T cells upon α‐CD3ε stimulation (Fig EV1B). To further investigate the function of DSF in p‐S6, immunoblot analysis was performed to detect the phosphorylation of S6 and p70S6 kinase, a kinase of the S6 protein (Wu et al, 2021). We examined DSF‐treated primary mouse CD8⁺ T cells and found that DSF enhanced the phosphorylation of both S6 and p70S6K (Fig 1D). Surprisingly, DSF could activate the phosphorylation of S6 and p70S6K without the stimulation of TCR crosslinking (time point 0; Fig 1D), indicating that DSF may directly activate a critical functional protein in the proximal signaling network of the TCR pathway.

Figure 1. DSF activates TCR signaling in CD8⁺ T cells.

• A
  FACS screening assay in WT mouse splenic cells.
• B
  FACS analysis of p‐S6 in WT mouse splenic CD8⁺ T cells stimulated for 8 min with α‐mouse CD3ε(145–2C11) after pretreatment with DSF (5 µM) for 2 h (n = 3).
• C
  Representative FACS of p‐S6 in WT mouse splenic CD8⁺ T cells stimulated for 8 min with α‐mouse CD3ε(145–2C11) after pretreatment with the indicated concentrations of DSF for 2 h (n = 2).
• D
  Western blot analysis of the indicated proteins in WT naïve CD8⁺ T cells stimulated by α‐mouse CD3ε(145–2C11) (1 µg ml⁻¹) in the presence of α‐mouse CD28 (1 µg ml⁻¹) for the indicated times after pretreatment with DSF (5 µM) for 2 h. The immunoblot shows a normalized expression of p‐S6(S235/236) to S6, p‐p70S6K to p70S6K, p‐Zap70 to Zap70, as well as that of p‐Lck(Y394) and p‐Lck(Y505) to Lck.
• E
  FACS analysis of p‐CD3ζ(Tyr142) in mouse naïve CD8⁺ T cells treated with DSF or stimulated with α‐mouse CD3ε(145–2C11) in the presence of α‐mouse CD28 (1 µg ml⁻¹) for 2 h (n = 3).
• F
  Western blot analysis of the indicated proteins in Jurkat T cells treated with the indicated concentrations of DSF for 2 h. The immunoblot analysis shows a normalized expression of p‐CD3ζ to CD3ζ as well as that of p‐Zap70(Tyr319)/Syk(Tyr352), p‐LCK(Y394), and p‐LAT(Tyr191) to β‐actin.
• G, H
  Luciferin assay of NFAT activity (G) and real‐time quantitative PCR (q‐PCR) analysis of Nfatc1 mRNA expression (H) in J76‐NFATRE‐luc and TCR052 cells stimulated with α‐human CD3(OKT3) (1 µg ml⁻¹) in the presence of α‐human CD28 (1 µg ml⁻¹) for 4 h or treated with DSF (5 µM) for 2 h. The levels of NFAT activity were normalized with TCR052 cells treated with DMSO (n = 3).

Data information: In (B–H), data are representative of 3 independent experiments, mean ± SD. **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant [unpaired t‐tests for measurements between the two groups in (B), (C), and (E), Image J analysis in (D) and (F), or two‐way analysis of variance (ANOVA) in (G, H)].

Source data are available online for this figure.

Figure EV1. DSF activates TCR signaling in CD8⁺ T cells.

• A
  FACS analysis of p‐S6 expression in WT mouse splenic CD4⁺ T cells stimulated for 8 min with α‐mouse CD3ε(145‐2C11) after pretreated with DSF (5 μM) for 2 h (n = 3).
• B
  FACS analysis of p‐S6 expression in WT mouse splenic CD4⁺ or CD8⁺ T cells stimulated for 8 min with α‐mouse CD3ε(145‐2C11) after pretreated with NaDTC (5 μM) for 2 h.
• C
  Western blot analysis of the phosphorylation level of CD3ζ in Jurkat T cells stimulated by α‐CD3(OKT3) (1 μg ml⁻¹) in the presence of α‐CD28 (1 μg ml⁻¹) for indicated times after pretreated with DSF (5 μM) for 2 h. The immunoblot shows a normalized expression of p‐CD3ζ to CD3ζ, p‐LCK(Y394), and p‐LCK(Y505) to β‐actin.
• D
  Western blot analysis of the phosphorylation level of Zap70 and Lck in primary OT‐I T cells stimulated by α‐CD3ε(145‐2C11) (1 μg ml⁻¹) in the presence of α‐CD28 (1 μg ml⁻¹) for indicated times after pretreated with DSF (5 μM) for 2 h. The immunoblot shows a normalized expression of p‐Zap70 to Zap70 as well as p‐Lck(Y505) to Lck.
• E
  Western blot analysis of the phosphorylation level of Zap70, Lck, and S6 in primary OT‐I T cells stimulated by α‐CD3ε(145‐2C11) (1 μg ml⁻¹) in the presence of α‐CD28 (1 μg ml⁻¹) for indicated times after pretreated with CuET (1 μM) for 2 h. The immunoblot shows normalized expression of p‐Zap70 to Zap70, p‐Lck(Y394), and p‐Lck(Y505) to Lck as well as p‐S6 to S6.
• F
  Identification of the expression of TCRα/β and CD3 of Jurkat, J76‐NFATRE‐luc, and TCR052 cells. Cells were gated from live single cells.
• G
  The luciferin assay of NFAT activity in TCR052 cells stimulated by α‐CD3(OKT3) (1 μg ml⁻¹) in the presence of α‐CD28 (1 μg ml⁻¹) for 4 h or followed by treating with DSF for 2 h. The levels of NFAT activity were normalized with DMSO (n = 3).
• H, I
  FACS analysis of ROS (H) and CD69 (I) in WT naïve CD8⁺ T cells treated with NAC (10 mM) for 24 h followed by with or without treatment with DSF (5 µM) or CuET (1 µM) for 2 h (n = 3).
• J
  CCK8 analysis of the viability of primary OT‐I T cells treated with DSF or CuET for 72 h (n = 5).

Data information: In (A–J), data are representative of 3 independent experiments, mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001. ns, not significant [multiple t‐tests between the two groups in (A) and (H), or two‐way ANOVA in (G)].

Source data are available online for this figure.

Upon TCR engagement by antigenic ligands, LCK is recruited to the TCR‐CD3 complex to phosphorylate the ITAMs of CD3 subunits, consequently phosphorylates ZAP‐70 (Chan et al, 1994; Ehrlich et al, 2002; Abraham & Weiss, 2004). To confirm whether DSF directly activates TCR signals, we examined the effect of DSF on TCR signals in mouse naïve CD8⁺ T cells. We found DSF robustly activated TCR signals, evidenced by the activation of Zap70 and Lck by increasing p‐Lck(Y394) and decreasing p‐Lck(Y505) (time point 0; Fig 1D). When naïve CD8⁺ T cells were treated with DSF, we observed that DSF induced the phosphorylation of CD3ζ (p‐CD3ζ), the initial step of TCR activation in the absence of stimulating antibodies (Fig 1E). To further investigate the direct effect of DSF on TCR signaling, Jurkat T cells were treated with DSF. The data showed that DSF directly facilitated proximal TCR signaling, marked by elevated p‐CD3ζ, and activation of LCK by increasing p‐LCK(Y394) and decreasing p‐LCK(Y505) (time point 0; Fig EV1C). DSF directly facilitated TCR signals p‐CD3ζ, p‐Zap70, p‐LAT, and p‐LCK(Y394) in a dose‐dependent manner (Fig 1F). Similar results were observed in primary OT‐I T cells from TCR‐transgenic OT‐I mice [CD8⁺ T cells that specifically recognize ovalbumin peptide residues 257–264 (OVA_257–264)] (Fig EV1D). As a derivative of DSF, CuET can also activate the TCR signal (time point 0; Fig EV1E). All these data support the function of DSF in triggering TCR signaling.

We next quantified the DSF‐induced transcription of NFAT, which is a downstream transcriptional factor for T cell activation. Reporter Jurkat T cells were engineered to express luciferase reporters under the NFAT promoter (Glanville et al, 2017). DSF directly activates NFAT transcription in TCR052 cells that expressed tuberculosis (TB) antigen‐specific TCR (Figs 1G right, and Fig EV1F). The level of DSF‐induced NFAT transcription was comparable with the effect induced by α‐CD3 and α‐CD28 antibodies (Figs 1G and H, and EV1G). In contrast, the DSF‐induced NFAT expression was abolished in the control J76‐NFATRE‐Luc cells, which are TCR‐deficient cells (Fig 1G and H left). These data indicated TCR itself is required for triggering the downstream signaling pathway by DSF, that is, DSF could directly activate T cell signaling in a TCR‐dependent manner.

We next measured whether the ROS level was affected during activation. We found that the ROS level after being treated with DSF or CuET was comparable with the control group. While, both DSF and CuET significantly increased the expression of CD69 in T cells (Fig EV1H and I). NAC treatment decreased the level of ROS in T cells (with slightly decreased expression of activation marker CD69) (Fig EV1H and I). Since CuET elicits cell death in a number of cancer cell lines, we performed CCK8 analysis to test the T cell viability. We found DSF did not elicit cell death by 5 μM, while CuET elicits cell death by 1 μM or above in primary OT‐I T cells (Fig EV1J).

DSF directly binds to and activates LCK

We next explored the underlying mechanism of DSF in TCR activation. DSF is suggested as a Zn‐finger active compound (Nash & Rice, 1998). LCK is an attractive candidate for DSF binding because LCK and the CD8 coreceptor form a zinc(Zn)‐finger domain, CD8 C‐terminal 225/227 Cys and LCK N‐terminal 20/23 Cys, but also free LCK is responsible for the initial phosphorylation of TCR (Kim et al, 2003; Yang et al, 2016; Terashima et al, 2020). First, we used microscale thermophoresis (MST) to quantify the direct interaction between DSF and LCK (Hu et al, 2020). The MST binding assay revealed that DSF binds to wild‐type LCK [LCK(WT)] with a dissociation constant of 86.812 nM (Fig 2A). DSF has been shown to bind Cys residues by covalent disulfide bonds (Rice et al, 1993). Consistently, our MST assay demonstrated that the interaction between DSF and LCK was reduced in the presence of dithiothreitol (DTT) (Fig EV2A). We then mutated Cys20 and Cys23 to Ser20 and Ser23 [LCK(C20/23S)], respectively, at the N‐terminal of LCK (Fig EV2B). DSF bound to LCK(C20/23S) at a dissociation constant of 22.797 µM, several hundred‐fold weaker than its binding to the WT counterpart (Fig 2B). Furthermore, the drug affinity responsive target stability (DARTS) assay was used to confirm the direct binding of DSF to LCK, which is based on the altered protease susceptibility of target proteins upon drug binding (Pai et al, 2015). Exposure to DSF led to a differential pronase‐dependent proteolysis pattern of LCK(WT) but not LCK(C20/23S) (Fig EV2C). To further confirm the covalent binding site of DSF–LCK, we used nano‐liquid chromatography‐tandem mass spectrometry (nano‐LC–MS/MS) to analyze DSF‐treated human LCK protein. Tryptic fragments indicated a dithiodiethylcarbamoyl adduct of Cys23, in which half of the symmetrical DSF molecule was attached to the thiol (Fig 2C and D). Together, these data demonstrated that LCK was covalently modified by DSF and the critical Cys residues of LCK were required for this modification.

Figure 2. DSF directly binds to and activates LCK.

• A, B
  MST analysis determined the K _d of DSF towards His‐LCK(WT) (86.812 nm) (A) or His‐LCK(C20/23S) (22.797 µM) (B) labeled with RED‐tris‐NTA 2nd Generation dye. Concentration is reported in nanomolar.
• C, D
  MS/MS spectra of the Cys20/23‐containing human LCK peptide WMENIDVCENCHYPIVPLDGK (aa 13–33; 2532.87 Da) (LC retention time was 49.16 min; a triplet‐charged precursor ion m/z 844.7163 (mass, 2531.1284 Da; ΔM of −0.5 ppm; 10logP = 200) was observed) (C), or of the corresponding LCK peptide after LCK incubation with DSF, which was modified on Cys23 (red) by the diethyldithiocarbamate moiety of DSF (an increase of 147.0176 Da; LC retention time was 56.31 min; a triplet charged precursor ion m/z 893.7295 (mass, 2678.1460 Da; ΔM of 7.7 ppm; −10logP = 116.99) was observed (D).
• E
  Recombinant human LCK proteins left untreated (Ctrl) or incubated with ATP and increasing concentrations of DSF for 15 min. Enzyme‐linked immunosorbent assay (ELISA) of LCK kinase activity was analyzed using α‐p‐LCK(Y394) antibody. The results show the average values of the optical density at 450 nm (OD₄₅₀) from five replicates (n = 5).
• F
  Confocal analysis of p‐Lck(Y394) (red) in naïve CD8⁺ T cells treated with DSF (5 µM) or stimulated with α‐mouse CD3 and α‐mouse CD28 for 2 h (left). The total fluorescence intensity (IntDen) was quantified (right; n = 20 cells). Scale bars, 3 µm.
• G
  Western blot analysis of the active LCK enrichment [assessed by p‐Lck (Y394)] in Triton X‐100‐insoluble fractions after DSF (5 µM) treatment in naïve CD8⁺ T cells.
• H
  IVP assay of His‐tagged CD3ε_CD peptides was phosphorylated by LCK in the presence of DSF. CD3ε phosphorylation was analyzed by immunoblotting using α‐p‐Tyr‐100.
• I
  Western blot analysis of p‐CD3ζ, LCK, and p‐Zap70 in LCK‐deficient cells were infected with the lentivirus collected from human HEK293T cells transfected with the control plasmid (Control) (1 µg), LCK(WT) (1 µg), or LCK(C20/23S) (1 µg) with psPAX2 (1 µg) and pMD2.G (1 µg) for 48 h followed by treatment with DSF (5 µM) for 2 h stimulated with α‐human CD3 and α‐human CD28 for 5 min.

Data information: In (A–I), data are representative of 2 (C–E, H) or 3 (A, B, F, G, and I) independent experiments, mean ± SD. **P < 0. 01; ***P < 0.001; ns, not significant [one‐way ANOVA in (E) and (F)].

Source data are available online for this figure.

Figure EV2. DSF directly binds to Cys20/23 of LCK.

• A
  MST analysis determined the K _d of DSF toward His‐LCK(WT) in the presence of DTT (5 mM). Concentration is reported in nanomolar.
• B
  Sequencing verification of the codon replacement by specific locus mutation resulting in LCK(C20/23S).
• C
  DARTS analysis of recombinant human LCK proteins shows that differential pronase‐mediated proteolysis after DSF (50 μM) addition is apparent for LCK(WT) but not for LCK(C20/23S), detected by SDS–PAGE and blotting with an α‐LCK monoclonal antibody.
• D
  Confocal analysis of p‐LCK(Y394) (red) in Jurkat T cells treated with DSF (5 µM) or stimulated with α‐human CD3 and α‐human CD28 for 2 h (left). The total fluorescence intensity (IntDen) was quantified (right; n = 20 cells). Scale bars, 3 µm.
• E, F
  Western blot analysis of LCK expression (E) and FACS analysis of CD69 [(F), (n = 3)] in Jurkat T cells and LCK‐deficient cells stimulated with α‐human CD3 and α‐human CD28 for 5 min.
• G
  Q‐PCR analysis of Nfatc1 mRNA expression in LCK‐deficient cells was infected with the lentivirus collected from human HEK293T cells transfected with the control plasmid (Control) (1 µg), LCK(WT) (1 µg), or LCK(C20/23S) (1 µg) with psPAX2 (1 µg) and pMD2.G (1 µg) for 48 h followed by treatment with DSF (5 µM) for 2 h stimulated with α‐human CD3 and α‐human CD28 for 5 min (n = 3).

Data information: In (A–G), data are representative of 3 independent experiments, mean ± SD. **P < 0.01; ***P < 0.001; ****P < 0.0001; ns, not significant [multiple t‐tests between the two groups in (D), or two‐way ANOVA in (F) and (G)].

Source data are available online for this figure.

LCK is regulated by trans‐autophosphorylation (Tyr394 phosphorylation) (Davis & van der Merwe, 2011). To further investigate whether DSF can activate LCK kinase activity, the purified recombinant human LCK protein was untreated (Ctrl) or incubated with ATP and DSF, DSF robustly enhanced LCK Tyr394 phosphorylation (Fig 2E). Through high‐resolution imaging of DSF‐treated naïve CD8⁺ T cells, we found that DSF increased microclusters of active LCK (Fig 2F). Consistent results were observed in Jurkat T cells (Fig EV2D). In addition, DSF‐induced immobilization of the active form of LCK was observed by the accumulation of p‐LCK(Tyr394) in detergent‐insoluble fractions from naïve CD8⁺ T cells (Fig 2G). Phosphorylation of the ITAMs of CD3 by LCK is essential for TCR activation in T cells (Salmond et al, 2009). We next set up an in vitro phosphorylation (IVP) assay to test the effects of DSF on the activity of LCK and its substrate, His‐tagged CD3ε_CD peptides (Xu et al, 2008). In this assay, we incubated CD3ε_CD peptides and LCK protein solution with or without DSF, which showed that p‐CD3ε was robustly enhanced after treatment with DSF (Fig 2H). To further confirm the effect of DSF on TCR phosphorylation by binding to LCK, we knocked out LCK in Jurkat T cells (LCK‐deficient cells) and verified the efficiency of the knockout by immunoblotting (Fig EV2E), and the abolish of activation by both immunoblotting and FACS analysis (Fig EV2E and F). We found DSF could not induce CD3ζ and Zap70 phosphorylation in these LCK‐deficient cells, supporting that the DSF‐induced activation of TCR signaling requiring LCK functionality (Fig 2I). DSF enhanced both CD3ζ and Zap70 phosphorylation in LCK (WT)‐transducing cells, but not the LCK(C20/23S)‐transducing cells (Fig 2I), supporting that DSF enhanced TCR activation via LCK Cys20/23. Consistent with the TCR activation results, DSF did not influence the mRNA expression of Nfatc1 in the LCK‐deficient cells. The mRNA expression of Nfatc1 was upregulated by DSF in LCK(WT)‐transducing cells. While the mRNA expression of Nfatc1 was not changed after DSF treatment in the LCK(C20/23S)‐transducing cells (Fig EV2G). Taken together, these data demonstrated DSF directly binds to LCK and in‐turn immobilizes and activates its kinase activity.

DSF activates T cell function

Phosphorylation of TCR signaling leads to T cell activation, marked by IL‐2 production, proliferation, and generation of the effector cytokines. To determine the effect of DSF on T cell activation, we treated T cells with DSF and analyzed their functions with different assays (Fig 3A). We performed RNA‐seq analysis, and functional pathway enrichment analysis showed a significant upregulation of the TCR signaling pathway in DSF pre‐treated primary OT‐I T cells (Fig 3B). Compared with the cells of the control group, DSF pre‐treated primary OT‐I T cells upregulated 164 genes and downregulated 25 T cell response‐related genes (Fig 3C). The expression of T cell activation‐related genes, including Cd69, Cd44, Dual Specificity Phosphatase 18 (Dusp18), Serpin Family E Member 1 (Serpine1), and Enolase 2 (Eno2), was upregulated in these DSF pre‐treated T cells (Fig 3D), indicating DSF induced strong T cell activation. To verify the corresponding gene expression patterns, we performed a q‐PCR analysis in cells treated as mentioned in Fig 3D, the mRNA expression of Serpine1, Crip2, Dusp18, Pik3ap1, Cd44, Cd69, Prkcg, Atf3, and Eno2 was comparable after treated with DSF or OVA in naïve CD8⁺ T cells (Fig EV3A). Although the mRNA expression of Tagap was relatively higher after being treated with OVA than with DSF, the general pattern of gene expression is similar between DSF and OVA stimulated naïve CD8⁺ T cells (Fig EV3A). Consistent with these gene expression data, the elevated protein level of the activation marker CD69 in DSF‐treated T cells could also be detected by flow cytometry (Fig EV3B). The secretion of IL‐2 is a critical and early landmark in the activation program of CD8⁺ T cells in vitro. Thus, we wondered whether DSF could influence the production of IL‐2. ELISA and q‐PCR data showed that DSF increased the level of secreted IL‐2 in primary OT‐I T cells treated with DSF (Figs 3E and EV3C left). Given that IL‐2 is a critical factor to promote T cell proliferation, we next tested the influence of DSF on the proliferation of T cells. In CFSE‐labeled primary OT‐I T cells following treatment with DSF, we found that the CFSE dilution was increased after treatment with DSF (Fig 3F). To further investigate the effect of DSF on the production of other effector cytokines, we primed primary OT‐I T cells with two peptides (N4 and A4) of different affinities followed by DSF treatment, and found that DSF enhanced the production of both interferon‐γ (IFNγ) and tumor necrosis factor α (TNFα) (Figs 3G and EV3C right). Together, these results demonstrated that DSF promoted T cell proliferation and induced the production of functional cytokines.

Figure 3. DSF directly activates T cell effector function.

• A
  Schematic depicting the treatment of naïve primary OT‐I T cells for assessment of T cell activation.
• B
  Functional enrichment analysis of Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways upregulated by DSF compared with DMSO (n = 3).
• C
  Scatter plot comparing global gene‐expression profiles of DMSO and DSF (n = 3).
• D
  A heatmap of upregulated genes associated with TCR and T cell activation‐related genes in DSF relative to that in DMSO. One column indicates one sample in each condition (n = 3).
• E
  ELISA analysis of the production of IL‐2 in primary OT‐I T cell supernatants treated with DMSO, DSF (5 µM), or OVA peptide (5 µg ml⁻¹) for 4 h (n = 3).
• F
  CFSE dilution of the proliferation of primary OT‐I T cells treated with DMSO or DSF (5 µM) for 2 h (n = 3).
• G
  FACS analysis of IFNγ and TNFα production in primary OT‐I T cells primed by N4 (SIINFEKL, 0.01 μg ml⁻¹) and A4 (SAINFEKL, 0.01 μgml⁻¹) for 2 h followed by treatment with DSF (5 µM) for 2 h (n = 3).
• H, I
  Extracellular acidification rate (ECAR) (H) and oxygen consumption rate (OCR) (I) in primary OT‐I T cells treated with DMSO, DSF (5 µM), or stimulated with α‐mouse CD3 and α‐mouse CD28 for 2 h (n = 3).

Data information: In (B–I), data are representative of 3 independent experiments, mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. ns, not significant [unpaired t‐tests for the measurements between the two groups in (E), (F), and (H), two‐way ANOVA in (G) and (I)].

Source data are available online for this figure.

Figure EV3. DSF enhances T cell activation and proliferation.

1. Q‐PCR analysis of the indicated gene expression in Figure 3D (n = 3).
2. FACS analysis of CD69 expression in primary OT‐I T cells stimulated treated with DSF (5 μM) for indicated times (n = 3).
3. Q‐PCR analysis of the Il‐2 and Ifng mRNA expression in naïve CD8⁺ T cells treated with DMSO or DSF (5 μM) for 2 and 4 h (n = 3).

Data information: In (A–C), data are representative of 3 independent experiments, mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001 [two‐way ANOVA in (A–C)].

Source data are available online for this figure.

Following early TCR signaling, T cells upregulate the metabolic machinery necessary to proliferate and sustain effector function (Chapman et al, 2020). A previous study showed that LCK activation can induce TCR‐dependent glycolysis in T cells (Menk et al, 2018). Our RNA‐seq analysis also revealed the upregulation of metabolic pathways following DSF treatment (Fig 3B). We next examined the metabolic function of T cells under DSF treatment conditions by Seahorse assay. Consistent with the profiling of gene expression, primary OT‐I T cells pre‐treated with DSF showed a much higher ECAR and OCR than the control cells (Fig 3H and I). Furthermore, the DSF effect on metabolism was comparable to the strong T cell activation induced by treatment with α‐CD3 and α‐CD28 antibodies (Fig 3H and I). These data demonstrated that DSF directly enhanced the metabolic program in CD8⁺ T cells.

DSF bolsters the anti‐tumor immunity of CD8⁺ T cells

To explore the in vivo anti‐tumor effect of DSF in mice, we compared groups of mice inoculated with murine B16F10 melanoma cells and fed with control (Vehicle) or DSF. Then tumor volume and survival rates were measured over time. Compared with those in the vehicle group, tumor volumes in the DSF‐treated group were significantly decreased (Fig 4A). One tumor disappeared on day 14 after oral administration of DSF (Fig 4B). Furthermore, the mortality in the DSF‐treated group was dramatically reduced (Fig 4C). Parallel studies in the MC‐38 tumor model revealed that DSF‐treated group mice displayed similarly reduced tumor growth (Fig EV4A–C). We analyzed the tumor‐infiltrating CD8⁺ T cells by FACS and immunofluorescence (IF) analysis, CD8⁺ T cells in the DSF‐treated group mice were increased (Fig 4D left, and 4E), as well as the CD8⁺/CD4⁺ T cell ratio (Fig 4D right), alongside enhanced activity marked as the frequency and number of IFNγ⁺CD8⁺ and Granzyme B (GzmB)⁺CD8⁺ T cells (Fig 4F), demonstrating that DSF enhanced the CD8⁺ T cell anti‐tumor immunity in TME. Next, we wondered whether DSF activates T cells in the draining lymph nodes (dLN) (Buchtova et al, 2021). We isolated the dLN from the melanoma‐bearing mice, the IF data showed that the number of CD8⁺ T cells in dLN does not show a difference between DSF and Vehicle group (Fig EV4D). While FACS analysis demonstrated that the frequency of IFNγ⁺CD8⁺ and TNFα⁺CD8⁺ T cells was elevated in the DSF‐treated group in the samples from dLN (Fig 4G), indicating that DSF not only enhanced the effector T cell function in TME but also in dLN.

Figure 4. DSF enhances the anti‐tumor immunity of CD8⁺ T cells.

• A, C
  Tumor growth [(A), (Vehicle, n = 5; DSF, n = 4, one tumor had disappeared)] and survival curves [(C), (n = 10 mice per group)] in DSF (suspended in 0.5% NaCMC)‐ and Vehicle (equal volume of 0.5% NaCMC)‐treated mice after B16F10 melanoma inoculation.
• B
  Images and tumor weight of subcutaneous mouse B16F10 tumors extracted from mice on day 21 (Vehicle, n = 5; DSF, n = 4, one tumor had disappeared).
• D–F
  Phenotypic (D, F) and IF (E) analysis of tumor‐infiltrating T cells on day 21 after melanoma inoculation (Vehicle, n = 5; DSF, n = 4, one tumor had disappeared). Scale bars, 20 µm. Statistics data are from a 40× field of view.
• G
  Phenotypic analysis of IFNγ and TNFα production in T cells isolated from dLN on day 19 after melanoma inoculation (n = 3).
• H, I
  Tumor growth [(H), (n = 5 mice per group)] and survival curves [(I), (n = 10 mice per group)] in Vehicle‐ and DSF‐treated mice injected s.c. with B16F10‐melanoma cells followed by intraperitoneal (i.p.) injection with PD‐1 antibody on days 0 and 4 and then every third day.

Data information: In (A–I), data are representative of 3 independent experiments, mean ± SD. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. ns, not significant [multiple t‐tests between the two groups in (B), (D) to (G), or two‐way ANOVA in (A) and (H), log‐rank (Mantel‐Cox) test in (C) and (I)].

Source data are available online for this figure.

Figure EV4. DSF inhibits MC38‐tumor growth in mice and did not influence memory CD8⁺ T cell subsets in vivo .

• A
  Diagram of MC38 tumor‐bearing mice model construction.
• B
  Tumor growth in DSF (suspended in 0.5% NaCMC)‐ and Vehicle (equal volume of 0.5% NaCMC)‐treated mice after MC38 colon cancer cells inoculation (n = 5 mice per group).
• C
  Photographs and tumor weight of subcutaneously growing mouse MC38 tumors extracted from mice at day 17 (n = 5 mice per group).
• D
  IF analysis of CD8⁺ T cells in dLN at day 21 after melanoma inoculation. Scale bars, 20 µm.
• E–H
  Phenotypic analysis of memory T cell (CD44, CD62L) in tumor (E) or dLN (F) and stem‐cell‐like memory marker (CD95, TCF1) in tumor (G) or dLN (H) at day 21 after melanoma inoculation (n = 5 mice per group).

Data information: In (B–H), data are representative of 3 independent experiments, mean ± SD. *P < 0.05; ***P < 0.001; ****P < 0.0001. ns, not significant [multiple t‐tests between the two groups in (C), (G) and (H), or two‐way ANOVA in (B), (E), and (F)].

Source data are available online for this figure.

We next detected memory CD8⁺ T cell subsets in the tumor‐infiltrating lymphocytes and lymph nodes. The percentage of CD62L⁺ central memory and CD62L⁻ effector memory CD8⁺ T cells in the DSF group was comparable with the control group (Fig EV4E and F). Furthermore, the expression of memory markers (CD95 and TCF1) in the control and DSF group showed no significance (Fig EV4G and H). To test the function of DSF in immune checkpoint blockade conditions, we treated B16F10 tumor‐bearing mice with α‐PD‐1, which could induce the expansion and boost the function of tumor‐infiltrating CD8⁺ T cells (Hu et al, 2019). The DSF‐treated group mice treated with α‐PD‐1 showed a lower rate of tumor growth and a higher survival rate than the α‐PD‐1 monotherapy group (Fig 4H and I), indicating that DSF can further enhance anti‐tumor immunity under PD‐1 blockade conditions. These data suggested that DSF has the potential to be combined with immune checkpoint blockade antibodies in the clinic.

Our data showed DSF directly activated T cells and induced a strong anti‐tumor effect in vivo, indicating that the DSF‐mediated anti‐tumor response directly involved CD8⁺ T cells. Using an in vivo α‐CD8 depleting antibody, we found that DSF‐mediated anti‐tumor function was largely abolished in the CD8⁺ T cell‐depleted group (Fig 5A), demonstrating that CD8⁺ T cells were necessary for the DSF‐mediated anti‐tumor immunity. To confirm the direct effect of DSF on CD8⁺ T cells, we co‐cultured B16‐OVA cells and DSF‐pre‐treated primary OT‐I T cells to examine their killing efficiency. Data showed that DSF‐pre‐treatment significantly promoted the in vitro cytotoxicity of CD8⁺ T cells (Fig 5B). Furthermore, we adoptively transferred DSF‐pre‐treated primary OT‐I T cells into B16‐OVA‐tumor‐bearing mice, and observed that mice receiving DSF‐treated T cells displayed a lower tumor growth rate and a prolonged survival time compared with those receiving the untreated control T cells (Fig 5C and D). Taken together, our data demonstrated DSF that directly enhances CD8⁺ T cell‐mediated anti‐tumor immunity.

Figure 5. DSF bolsters anti‐tumor immunity via activating CD8⁺ T cells.

• A
  Tumor growth in Vehicle‐ and DSF‐treated mice injected s.c. with B16F10 melanoma cells followed by intraperitoneal (i.p.) injection with CD8(2A3) antibody on days −1, 0, and 2 and then every second day (n = 5).
• B
  FACS analysis of cytotoxicity of B16F10‐OVA cells after co‐culture with DMSO‐ or DSF‐treated primary OT‐I T cells (n = 3).
• C, D
  Tumor growth [(C), (n = 5 mice per group)] and survival curves [(D), (n = 10 mice per group)] of B16F10‐OVA tumor‐bearing mice after adoptive transfer of Vehicle‐ or DSF‐treated primary OT‐I T cells.

Data information: In (A–D), data are representative of 3 independent experiments, mean ± SD. *P < 0.05; **P < 0.01; ****P < 0.0001. ns, not significant [two‐way ANOVA in (A) to (C), log‐rank (Mantel‐Cox) test in (D), mean ± SD].

Source data are available online for this figure.

Discussion

Conceptually, our findings provide an example of applying “compound‐mediated T cell receptor activation” to directly promote T cell anti‐tumor immunity. We found that the FDA‐approved alcohol drug DSF directly binds to LCK via Cys20/23 to promote its kinase activity, thereby boosting T cell effector response and enhancing the anti‐tumor immunity of CD8⁺ T cells (Nash & Rice, 1998; Nika et al, 2010). LCK, which belongs to the SRC family of kinases and is the first molecule to be recruited to the TCR complex, can either be free or associated with CD8 or CD4 coreceptors (Kim et al, 2003; Salmond et al, 2009; Wang et al, 2018). Free LCK, which is associated with the TCR‐CD3 complex, phosphorylates CD3ζ ITAMs and consequently activates TCR signaling pathways (Terashima et al, 2020; Wu et al, 2021). Thus, it is rational that DSF directly activates TCR signaling by enhancing the LCK kinase activity.

DSF can efficiently activate Zn‐finger proteins or complexes (Nash & Rice, 1998), therefore, we chose LCK as an attractive candidate because LCK and CD4/CD8 coreceptor form a Zn‐finger complex via Cys residues. Our results firstly proposed a novel mechanism of TCR activation, and identified a previously unknown target of DSF in T cells, evidenced by the fact that DSF can covalently bind to LCK via Cys20/Cys23, in which half of the symmetrical DSF molecule is attached to the thiol of Cys, resulting in CD3ζ ITAMs phosphorylation. Previous studies demonstrated that DSF inhibits human PHGDH through binding to Cys116 residue (Wang et al, 2018); DSF covalently modifies human/mouse Cys191/Cys192 in GSDMD to block pore formation in THP‐1 cells (Hu et al, 2020). Together, these studies indicated that the thiol‐linked covalent modification of DSF to specific target proteins might be a common mechanism to regulate physiological effects in various cell types.

However, it remains unclear whether DSF can regulate T cell function in anti‐tumor responses. In this study, our in vitro results showed that DSF robustly enhanced T cell‐mediated cytotoxicity directly. Moreover, our in vivo study demonstrated that the anti‐tumor function was abolished after CD8⁺ T cell depletion. These data support that the anti‐tumor function of DSF is dependent on the direct activation of CD8⁺ T cells. In addition, our adoptive transfer data showed that DSF could boost the in vivo anti‐tumor function of CD8⁺ T cells through in vitro pre‐treatment, broadly highlighting its potential application in TCR‐T or CAR‐T cell therapy, because LCK activation was the critical event in both cases.

Previous studies suggested that CuET (chemically bis‐DTC–copper complex) accumulates in tumor cells, binds and aggregates NPL4, to inhibit protein processing and degradation, thus leading to cancer cell death (Skrott et al, 2017, 2019). Our data confirmed that CuET, but not DSF, kills the cancer cell in vitro (Fig EV5). It has been identified that cannabidiol, a component of the marijuana plant used by cancer patients to mitigate side effects of chemotherapy, triggers the expression of metallothioneins providing protective effects by binding heavy metal‐based substances including CuET. Whereas, at the same time, CuET’s anticancer toxicity is neutralized by metallothioneins (Buchtova et al, 2021). The latest study showed that copper can directly induce cell death in a mitochondrial respiration‐dependent manner (Tsvetkov et al, 2022), adding additional complexity to applying the copper‐related metabolite in clinics. In this study, we demonstrated that DSF directly activates T cell by binding and activating LCK kinase activity through half of the DSF (DTC, which is independent of copper ion). Thus, we hypothesize that cannabidiol may not influence DSF‐induced T cell activation and anti‐tumor effect. However, it needs to be addressed by further experiments. Together, our work provides evidence for repurposing the FDA‐approved drug, DSF, in cancer immunotherapy.

Figure EV5. CuET suppresses the viability of tumor cells.

• A–C
  CCK8 analysis of the viability of mouse breast cancer 4T1 (A), mouse melanoma B16F10 (B), and mouse colon adenocarcinoma MC38 (C) cell lines treated with DSF or CuET for 24 h (n = 5).

Data information: In (A–C), data are representative of 3 independent experiments, mean ± SD.

Source data are available online for this figure.

Materials and Methods

Cell culture

CD8⁺ T cells isolated from OT‐I or WT mice were cultured in complete RPMI 1640 media (10% FBS, 0.05 mM 2‐mercaptoethanol, 1 mM NEAA). J76‐NFATRE‐Luc, TCR052, and 4T1 cells were maintained in complete RPMI 1640 media containing 10% FBS. Human HEK293T and B16F10 cells were cultured in complete DMEM media containing 10% FBS. MC38 cells were cultured in complete DMEM media (10% FBS, 20 mM HEPES).

Murine models

C57BL/6 mice were purchased from SLAC. OT‐I TCR transgenic mice were from The Jackson Laboratory. All mice were maintained in pathogen‐free facilities in a specific pathogen‐free facility at Shanghai Jiao Tong University. All animal experiments used mice with matched age and sex. Animals were randomly allocated to experimental groups. All animal experiments were approved by the Institutional Animal Care and Use Committee of Shanghai Jiao Tong University School of Medicine. The maximal tumor measurements/volumes are in accordance with the IACUC.

Compound screening assay

Peripheral cells were isolated from the C57BL/6 mouse spleen. Cells were treated with different compounds (Topscience Co., Ltd.) for 2 h at 37°C, and then stimulated with 1 μg ml⁻¹ α‐mouse CD3ε(145‐2C11, BD) for 8 min at 37°C. Cells were fixed with 4% paraformaldehyde (PFA) and permeabilized with 70% methanol for 4–24 h. Next, cells were stained with anti‐CD4(RM4‐5, BV421), anti‐CD8(53‐6.7, APC), and anti‐p‐S6(D57.2.2E, PE). BD LSRFortessa X‐20 was used for data acquisition and FlowJo (Tree Star) was used for data analysis.

Western blot analysis of TCR signal

To generate primary OT‐I T cells, lymph nodes isolated from OT‐I mice were stimulated with OVA_257–264 peptide for 3 days in the presence of 1 ng ml⁻¹ recombinant mouse IL‐2 (rmIL‐2) (R&D, MX2816041). Cells were centrifuged and cultured in a fresh medium containing 1 ng ml⁻¹ rmIL‐2 for 2 more days, after which most of the cells in the culture were primary OT‐I T cells. Primary OT‐I T cells were treated with control (DMSO) or DSF (5 μM) or CuET (1 μM) for 2 h at 37°C, and then stimulated with 1 μg ml⁻¹ α‐mouse CD3ε(clone: 145‐2C11, BD, 553057) plus α‐mouse CD28(clone: 37.51, BD, 553294) for 0, 15, and 30 min.

Jurkat T cells were treated with control (DMSO) or DSF (5 μM) for 2 h at 37°C, and then stimulated with 1 μg ml⁻¹ α‐human CD3(clone: OKT3, eBioscience, 16‐0037‐85) plus α‐human CD28(clone: CD28.2, BD, 555725) for 0, 2, and 5 min.

Cells were lysed in cell lysis buffer (CLB) (Cell Signaling Technology, #9803) containing Phosphatase Inhibitor Cocktail (Roche, 04906837001) and Protease Inhibitor Cocktail (Roche, 04693116001). Protein concentrations were quantified with the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, A53225). For western blot, equal amounts of protein were heat denatured in the presence of a reducing agent and separated on 10% SDS‐PAGE, and transferred to PVDF membranes (Millipore, IPVH00010). Antibodies used for western blot were as followed: anti‐p‐S6 (CST, 5364S), anti‐S6 (CST, 2217S), anti‐p‐p70S6K (CST, 9204S), anti‐p70S6K (CST, 2708S), anti‐p‐CD3ζ(Tyr141) (CST, 67748S), anti‐CD3ζ (Santa Cruz Biotechnology, 6B10.2), anti‐p‐Zap70(Tyr319)/Syk(Tyr352) (CST, 2717S), anti‐p‐Zap70(Tyr319)/Syk(Tyr352) (CST, 2701S), anti‐p‐LAT(Tyr171) (CST, 3581S), anti‐p‐LCK(Y394) (R&D, MAB7500), anti‐p‐LCK(Y505) (CST, 2751S), anti‐Lck(3A5) (Santa Cruz Biotechnology, sc‐433), and anti‐β‐actin (CST, 4970S). Proteins were detected using ECL Plus (Tanon, 180‐5001) using the ChemiDoc Imaging System (AI600). Data were analyzed by Image J and normalized with control (as mentioned in figure legend).

Luciferase assay of NFAT transcription activity

The 2 × 10⁵ J76‐NFATRE‐luc and TCR052 cells were stimulated with 1 μg ml⁻¹ α‐human CD3(OKT3) plus α‐human CD28(CD28.2) for 4 h, or treated with control (DMSO) or DSF (5 μM) for 2 h. Cells were collected and luciferase activity was measured using Nano‐Glo Luciferase Assay (Promega, N1110). The absorbance was calculated using the SpectraMax i3 (Molecular Devices) and SoftMax Pro 6.3 software. Fold induction of luciferase activity was calculated by referring to unstimulated samples.

Q‐PCR analysis

Q‐PCR analysis was performed as previously described (Wang et al, 2018). Briefly, cells were collected after being treated as legend mentioned, extraction of RNA and operation of q‐PCR were performed by Fastagen RNA extraction and Bimake RT‐PCR kits according to the protocols. Sequences of the primers for targeting of indicated gene expression are in Table EV1. Data were normalized by the level of indicated β‐actin expression in each sample.

FACS analysis of T cell ROS level and activation

Wildtype naïve CD8⁺ T cells were treated with NAC (10 mM) for 24 h or DMSO, DSF (5 μM), CuET (1 μM) for 2 h, washed cells 2 times with pre‐warmed PBS, stained with LIVE/DEAD(APC‐Cy7) (Invitrogen, L34976A) and anti‐CD8(53‐6.7, BV785), anti‐CD69(H1.2F3, Percp‐cy5.5) for 15 min at 4°C, then washed cells with pre‐warmed PBS for 2 times, stained with CM‐H2DCFDA (Thermo Fisher Scientific, C6827) for 30 min at 37°C, washed cells with cold PBS for 2 times. BD LSRFortessa X‐20 was used for data acquisition and FlowJo (Tree Star) was used for data analysis.

ELISA

For IL‐2 production: primary OT‐I T cells (1 × 10⁶ cells) were plated in 6‐well plates with a complete growth medium. Cells were treated with control (DMSO), DSF (5 μM) or OVA_257–264 peptide (5 μg ml⁻¹) for 4 h. IL‐2 in supernatants were measured by ELISA as follows: coat 96 well plates with 50 μl/well of purified rat anti‐mouse IL‐2 (Pharmingen) at 2 μg ml⁻¹ in PBS. Incubate the plate at 4°C overnight. Block the plates by blocking buffer (PBS containing 5% BSA) for 2 h at room temperature (RT). Wash the plate six times with wash buffer (PBS containing 0.05% Tween 20). Add 100 μl supernatants or IL‐2 standard (Pharmingen) for 2 h at RT, wash six times with wash buffer, and incubate with 100 μl/well of purified Biotin rat anti‐mouse IL‐2 (Pharmingen) at 1 μg ml⁻¹ in blocking buffer for 1 h at RT, wash six times with wash buffer, add 100 μl/well of Streptavidin‐HRP (Biolegend, 405210) for 30 min at RT, wash six times with wash buffer, add 100 μl/well of 1× TMB solution (Biolegend, 421501) at RT, stop the reaction with 100 μl/well of 2 M HCl. The data were collected at 450 nM and normalized to DMSO control.

For LCK kinase activity assays: LCK kinase activity was described as previously (Wu et al, 2021). Briefly, to measure the phosphorylation of Tyr394 on LCK, recombinant human LCK proteins pre‐incubated with a buffer containing 60 mM HEPES (pH 7.5), 5 mM MgCl₂, 5 mM MnCl₂, 1 mM DTT, Phosphatase Inhibitor Cocktail, Protease Inhibitor Cocktail, and with or without DSF (5 µM), were incubated at 37°C for 15 min. ATP was then added to the mixtures to a final concentration of 500 µM and incubated for 15 min at 37°C. The reaction mixtures were transferred into a 96‐well ELISA plate (100 µl per well) to detect the phosphorylation of Tyr394 on LCK.

For the ELISA: add 100 μl supernatants or IL‐2 standard (Pharmingen) or the mixtures for 2 h at RT, wash six times with wash buffer, incubate with 100 μl/well of purified Biotin rat anti‐mouse IL‐2 (Pharmingen) at 1 μg ml⁻¹ or 100 µl mouse anti‐LCK p‐Tyr394 antibody (1:2,000) for 1 h at RT, wash six times with wash buffer, add 100 μl/well of Streptavidin‐HRP (Biolegend, 405210) for 30 min or goat anti‐mouse IgG‐HRP was added (100 µl per well) for 1 h, respectively, at RT, wash six times with wash buffer, add 100 μl/well of 1× TMB solution (Biolegend, 421501) at RT, stop the reaction with 100 μl/well of 2 M HCl. The data were collected at 450 nM and normalized to DMSO control.

IVP assay

Large uniflagellar vesicles with 100 nm diameter were described as previously (Xu et al, 2008). POPC (90%) and DGS‐NTA(Ni) (10%) were used for LUV preparation. His‐tagged CD3ε_CD peptides and His‐tagged LCK (Millipore) were incubated with LUVs for 1 h at room temperature for membrane coating. ATP was then added to initiate the reaction. The reaction volume was 50 μL, and the final concentrations of ingredients were: 0.5 μM His‐CD3ε_CD, 7.1 nM His‐tagged LCK, 1 mM phospholipids, 65 mM HEPES (pH7.0), 5 mM MgCl₂, 3 μM Na₃VO₄, 1.25 mM DTT, 150 mM KCl, and 0.2 mM ATP. The TCR phosphorylation reactions were proceeded at 37°C for 0, 30, and 60 min and terminated with 2× SDS loading buffer and then boiled for 10 min. Then, the samples were subjected to SDS‐PAGE gel for electrophoresis. After that, the proteins on the gel were transferred onto the PVDF membrane for subsequent immunoblot analysis. The tyrosine‐phosphorylated and total protein levels were detected by anti‐p‐Tyr‐100 and anti‐CD3ε primary antibodies, respectively, followed by the corresponding HRP‐conjugated secondary antibodies. Proteins were detected using ECL Plus (Tanon, 180‐5001) using the ChemiDoc Imaging System (AI600).

Lentivirus preparation and LCK‐deficient cells transfection

Human HEK293T cells were cultured and transfected with 3 μg DNA (1ug constructs, 1 µg psPAX2 and 1ug pMD2.G) using PEI transfection reagent according to the protocols. After that, collected the medium contains the target lentivirus and is directly apply to a 0.22 μm filter after 48 h. LCK‐deficient cells were infected with the lentivirus by the ratio of 1:3 (virus vs. fresh complete medium) for 48 h, the cells were collected for immunoblot analysis. The data were analyzed by Image J.

Confocal immunofluorescence imaging

Imaging was performed on a custom modified Olympus FV3000 Laser Scanning Microscope equipped with a × 60 oil immersion lens. Briefly, naive CD8⁺ T cells or Jurkat T cells were seeded in chambers and stimulated with control (DMSO) or DSF (5 μM) or 1 μg ml⁻¹ α‐CD3 plus 1 μg ml⁻¹ α‐CD28 antibodies for 2 h at 37°C. The cells were then fixed with 4% paraformaldehyde (PFA) and 0.1% Triton X‐100 for 1 h at 4°C, followed by 5% BSA blocking for 30 min at 37°C. The cells were gently washed three times with PBS and stained overnight with the indicated antibody at 4°C. After three washes with PBS, the cells were incubated with the secondary antibody for 1h at 4°C and gently washed three times with PBS before imaging. The images were analyzed using Image J.

Cell fractionation for Triton‐X100 insoluble pellets

Cell fractionation assay was described as previously (Skrott et al, 2017). Briefly, naïve CD8⁺ T cells were treated as indicated, washed in cold PBS, and lysed in lysis buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 2 mM MgCl₂, 10% glycerol, 0.5% Triton X‐100, and protease inhibitor cocktail) for 10 min gently agitating at 4°C. Then, cells were kept for another 10 min on ice with intermittent vortexing. After that, the lysate was centrifuged at 15,000 g for 10 min at 4°C. The insoluble fraction (pellets) and soluble fraction were separately resuspended in 1× LSB buffer and terminated with 5× SDS loading buffer and then boiled for 10 min. Then, the samples were subjected to SDS–PAGE gel for electrophoresis. After that, the proteins on the gel were transferred onto the PVDF membrane for subsequent immunoblot analysis. The active accumulation was detected by anti‐p‐Lck(Y394), followed by the corresponding HRP‐conjugated secondary antibodies. Proteins were detected using ECL Plus (Tanon, 180‐5001) using the ChemiDoc Imaging System (AI600).

Protein labeling and MST analysis

The binding affinity of LCK and DSF was measured at 25°C in a binding buffer (PBS containing 0.05% Tween 20) by Microscale Thermophoresis using MONOLITH NT.115. 100 μl His‐LCK(WT) or His‐LCK(C20/23S) at a concentration of 160 nM was incubated with 100 μl 100 nM RED‐tris‐NTA 2nd Generation dye in the dark for 30 min. After incubation, the labeled LCK was mixed 1:1 with DSF in a two‐fold dilution series from 49 nM‐150 μM for the measurement or in the presence of 5 mM DTT. The samples were loaded into NanoTemper Monolith NT.115 glass capillaries and MST was carried out using 40% MST power. K_d values were calculated using the mass action equation and NanoTemper software, the data were plotted by the GraphPad Software.

Nano‐LC–MS/MS assay

The sequence analysis of LCK binding by DSF was performed on an Orbitrap Fusion LUMOS mass spectrometer (Thermo Fisher Scientific) connected to an Easy‐nLC 1200 via an Easy Spray (Thermo Fisher Scientific). The peptides were loaded onto a self‐packed analytical PicoFrit column with an integrated spray tip (New Objective, Woburn, MA, USA) (75 μm × 15 cm length) packed with ReproSil‐Pur 120A C18‐AQ 1.9 μm (Dr. Maisch GmbH, Ammerbuch, Germany) and separated within a 60 min linear gradient from 95% solvent A (0.1% formic acid/2% acetonitrile/98% water) to 28% solvent B (0.1% formic acid/80% acetonitrile) at a flow rate of 300 nl/min at 50°C. The mass spectrometer was operated in positive ion mode and employed in the data‐dependent mode within the specialized cycle time (2S) to automatically switch between MS and MS/MS. One full MS scan from 350 to 1500 m/z was acquired at high‐resolution R = 120,000 (defined at m/z = 400); MS/MS scans were performed at a resolution of 30,000 with an isolation window of 1.6 Da and higher‐energy collisional dissociation (HCD) fragmentation with a collision energy of 30% ± 5. Dynamic exclusion was set to 30 s.

Database search and data analysis

All MS/MS ion spectra were analyzed using PEAKS 10.6 (Bioinformatics Solutions) for processing, de novo sequencing, and database searching. The resulting sequences were searched against the known peptide (WMENIDVCENCHYPIVPLDGKG) with mass error tolerances of 10 ppm and 0.02 Da for parent and fragment, respectively, the digest mode specified as No digestion, and Oxidation of methionine (M + 15.99), and DSF of Cys (C + 147.02) specified as variable modifications. FDR estimation was enabled. Peptides were filtered for −10log P ≥ 20 (P < 0.01).

RNA‐seq analysis

Total RNA was extracted from primary OT‐I T cells treated with DSF (5 μM) or control (DMSO) for 2 h. RNA extraction was performed using TRIzol (Invitrogen, 15596026) following the manufacturer’s protocol. Scatter plot comparing global gene‐expression profiles, functional enrichment, and heatmap analysis were performed using the free online platform of Majorbio Cloud Platform (www.majorbio.com).

CellTrace CFSE analysis of T cell proliferation

Primary OT‐I T cells were treated with control (DMSO) or DSF (5 μM) for 2 h, washed cells 2 times with pre‐warmed PBS, and labeled cells with 5 μM CFSE (Thermo Fisher, C34554) staining solution for 20 min at 37°C water bath, incubate cells with 5 ml pre‐warmed complete medium for 5 min, centrifuge cells for 5 min at 300 g and washed cells two times with pre‐warmed PBS. Cells were stimulated with OVA_257–264 peptide (5 μg ml⁻¹) for 1 day, and then added 1 ng ml⁻¹ rmIL‐2 for 2 more days. Collect cells and stained with LIVE/DEAD(APC‐Cy7) (Invitrogen, L34976A) and anti‐CD8(53‐6.7, BV785). BD LSRFortessa X‐20 was used for data acquisition and FlowJo (Tree Star) was used for data analysis.

FACS analysis with T cell mediated‐cytotoxicity

Primary OT‐I T cells were treated with control (DMSO) or DSF (5 μM) for 2 h, washed cells 2 times with pre‐warmed PBS, and cocultured with B16F10‐OVA cells with indicated effector versus target ratio, both supernatants and cells were collected after co‐culture for 16–24 h. Cells were stained with LIVE/DEAD(BV510) (Invitrogen, L34957) and anti‐CD8(53‐6.7, APC). BD LSRFortessa X‐20 was used for data acquisition and FlowJo (Tree Star) was used for data analysis.

FACS analysis with T cell effector cytokines

Primary OT‐I T cells were primed with 0.01 μg ml^‐1 N4 or A4 peptide for 2 h in the presence of brefeldin A (BFA), followed by treated with control (DMSO) or DSF (5 μM) for 2 h. Cells were collected and stained with LIVE/DEAD(APC‐Cy7) (Invitrogen, L34976A), anti‐CD8(53‐6.7, APC), anti‐IFNγ(XMG1.2, PE), and anti‐TNFα(MP6‐XT22, PE‐Cy7) by Foxp3/Transcription Factor Staining Buffer Set (eBioscience, 2159394) according to manufacturer’s protocol.

For T cell activation validation, primary OT‐I T cells were treated with control (DMSO) or DSF (5 μM) for 2 h. Cells were stained with LIVE/DEAD(APC‐Cy7) (Invitrogen, L34976A), anti‐CD8(53‐6.7, APC), anti‐CD69(H1.2F3, Percp‐Cy5.5).

BD LSRFortessa X‐20 was used for data acquisition and FlowJo (Tree Star) was used for data analysis.

Glycolytic and mitochondrial respiration rate measurement

Seahorse XFe96 Extracellular Flux Analyzer (Agilent Technologies) was used for metabolic experiments. Primary OT‐I T cells were treated with control (DMSO), DSF (5 μM) or stimulated with 1 μg ml⁻¹ α‐mouse CD3ε(145‐2C11) plus α‐mouse CD28 for 2 h. Cells were seeded at a density of 2 × 10⁵ per well. The ECAR and OCR for each well were calculated, while the cells were subjected to the XF Cell Mito or the XF Glycolytic stress test using the following concentrations of injected compounds: 10 mM glucose, 2 mM oligomycin, 50 mM 2‐DG, 1 mM FCCP, 0.5 mM rotenone/antimycin A. The XF Cell Mito and the XF Glycolytic stress test kits were purchased from Agilent Technologies, Inc.

In vivo mouse models

To test the effect of DSF on the skin melanoma mouse model, B16F10 cells were washed three times with PBS, and B16F10 cells (2 × 10⁵) were subcutaneously injected into the dorsal part of the mouse (aged 8–10 weeks). After 7 days, mice were randomly divided into three groups, each of 5 mice, and treated as follows: (i) normal diet plus oral administration of 0.5% carboxymethyl cellulose sodium (NaCMC) as control (Vehicle) every day; (ii) normal diet plus oral administration of 10 mg kg⁻¹ DSF every day; and (iii) normal diet plus oral administration of 50 mg kg⁻¹ DSF every day. From day 7, tumor volume was measured every 2–3 days, and animal survival rate was recorded every day. Tumor volume was calculated as length × width × width × 0.5. Mice with tumor sizes larger than 20 mm at the longest axis were euthanized for ethical considerations. To analyze the effector function of tumor‐infiltrating T cells, mice were euthanized on day 21. T cells were isolated and analyzed as mentioned above.

To test the effect of DSF on mouse colon cancer mouse model, MC38 cells were washed three times with PBS, MC38 cells (1 × 10⁶) were subcutaneously injected into the dorsal part of the mouse (aged 8–10 weeks). After 7 days, mice were randomly divided into two groups, each of 5 mice, and treated as follows: (i) normal diet plus oral administration of Vehicle every day; (ii) normal diet plus oral administration of 50 mg kg⁻¹ DSF every day. Tumor growth and survival rates were detected as mentioned above.

In vivo experiments with CD8 depletion or DSF combine anti‐PD‐1 therapy

B16F10 cells (2 × 10⁵) were subcutaneously injected into 7–8 weeks old male C57BL/6 mice. For CD8 depletion assay, on day 7, melanoma‐bearing mouse with similar tumor size were randomly divided into four groups (n = 5–10): Vehicle + α‐Isotype IgG2a (clone: 2A3, Bio X cell, BE0089), Vehicle + α‐CD8, DSF + α‐Isotype IgG2a, DSF + α‐CD8. The α‐CD8 antibody (clone: 53‐6.7, Bio X cell, 100 μg per mouse) was injected intraperitoneally on day −1, day 0, day 2 and then every 2^nd day. For DSF combined α‐PD‐1 therapy, a melanoma‐bearing mouse with a similar tumor size were randomly divided into four groups (n = 5–10) and received Vehicle + α‐Isotype IgG2a, Vehicle + α‐PD‐1, DSF + α‐Isotype IgG2a, DSF + α‐PD‐1. The α‐PD‐1 antibody (clone: RMP1‐14, Bio X cell, 200 μg per mouse) was injected intraperitoneally on day 0, day 4 and then every 3^rd day. Tumor growth and survival rates were detected as mentioned above.

In vivo experiments with primary OT‐I T cell transfer

B16F10‐OVA cells (2 × 10⁵) were subcutaneously injected into 7–8 weeks old male C57BL/6 mice. On day 6, primary OT‐I T cells were treated with control (DMSO) or DSF (5 μM) for 2 h, then adoptive transfer of primary OT‐I T cells (2 × 10⁶) to donor mice (n = 5–10). Tumor growth and survival rates were detected as mentioned above.

FACS analysis with tumor‐infiltrating T cells

After implantation as mentioned above, tumors were dissected from the surrounding fascia, weighed, mechanically minced, and treated with DNase I (0.1 mg ml⁻¹) and collagenase D (1 mg ml⁻¹) for 30 min at 37°C. Cells were passed through a 70‐μm filter to remove clumps, diluted in medium, and single‐cell suspensions were stained with the following antibodies: LIVE/DEAD(BV510) (Invitrogen, L34957), anti‐CD45(30‐F11, APC‐Cy7), anti‐CD4(RM4‐5, BV421), anti‐CD8(53‐6.7, APC), anti‐CD44(IM7, PE‐Cy7), anti‐CD62L(MEL‐14, BV711), anti‐CD95(J02, PE‐Cy7), anti‐TCF1(S33‐966, BV421), anti‐IFNγ(XMG1.2, PE), anti‐GzmB(NGZB, FITC), and anti‐TNFα(MP6‐XT22, PE‐Cy7) by eBioscience™ Foxp3/Transcription Factor Staining Buffer Set (eBioscience, 2159394) according to manufacturer’s protocol.

For dLN evaluation in the tumor mice model, draining lymph nodes were isolated from B16F10‐melanoma mice after implantation on day 19, single‐cell suspensions were stimulated with Cell Stimulation Cocktail (plus protein transport inhibitors) (Thermo Fisher, 00‐4975‐93) for 5 h, then cells were stained with following antibodies: LIVE/DEAD(BV510) (Invitrogen, L34957), anti‐CD45(30‐F11, APC‐Cy7), anti‐CD8(53‐6.7, APC), anti‐CD44(IM7, PE‐Cy7), anti‐CD62L(MEL‐14, BV711), anti‐CD95(J02, PE‐Cy7), anti‐TCF1(S33‐966, BV421), anti‐IFNγ(XMG1.2, PE), and anti‐TNFα(MP6‐XT22, PE‐Cy7), by eBioscience^TM Foxp3/Transcription Factor Staining Buffer Set (eBioscience, 2159394) according to manufacturer’s protocol. BD LSRFortessa X‐20 was used for data acquisition and FlowJo (Tree Star) was used for data analysis.

IF analysis with tumor‐infiltrating T cells

After implantation as mentioned above, tumors and lymph nodes are embedded in paraffin sections. Paraffin sections were deparaffinized to water, xylene for 5 min for 2 times, gradient alcohol (anhydrous, 95, 85, and 75%) for 5 min per gradient, and washed three times with TBS for 3 min. After antigen retrieval in a pressure cooker, it was blocked with 10% donkey serum for 30 min at 37°C. Discard serum and incubate overnight at 4°C with CD8 antibody working solution. After 20 min at room temperature, wash three times with TBST for 5 min each time, add Alexa Fluor^®488 donkey anti‐rabbit lgG(H + L) for 45 min at 37°C, wash three times with TBST for 5 min each time, discard TBST and stained with DAPI working solution in the dark for 5 min. Mount slides with fluorescent mounting medium. Data were scanned and viewed by CaseViewer software.

DARTS

DARTS was performed according to a modified published protocol. Purified GST‐LCK(WT) and GST‐LCK(C20/23S) proteins were diluted by 100 mM phosphate buffer, pH 7.4 to a final concentration of 0.03 μg μl⁻¹. The proteins were treated with DSF (50 μM, dissolved in 10% DMSO) for 1h and equal amounts of DMSO were added to the solutions, which served as control samples. Pronase (Roche, 10165921001) was dissolved in TNC buffer (50 mM Tris–Cl, 50 mM NaCl, 10 mM CaCl₂, pH 7.5). The 0.25 μg of pronase was added to 50 μl of protein solution and incubated for 1 h at 37°C. Samples without pronase served as the input controls. The reaction mix was stopped by the addition of 5 × SDS loading buffer; the samples were boiled at 95°C for 15 min and loaded on SDS–PAGE gels, then transferred to PVDF membranes for western blot analysis.

CCK8 analysis of cell viability

4T1, B16‐F10, and MC38 cells were seeded in 96 platform plates with 1 × 10⁴ cells/well, after 24 h, treated with mentioned concentrations of DSF or CuET for 24 h. Replaced the medium with fresh medium, add CCK8 solution, and detect the OD₄₅₀ value with a microplate reader after 2–4 h.

For evaluation of DSF and CuET on the T cells, primary OT‐I T cells (4 × 10⁵ cells/well) were seeded in 96 platform plates, treated with mentioned concentrations of DSF or CuET in the presence of OVA peptide, added IL‐2 after 24 h. After 72 h, added CCK8 solution for 2–4 h to detect the OD₄₅₀ value by a microplate reader.

Data were analyzed as follows: Viability (%) = [OD₄₅₀ (Cells treated with the indicated concentration of drug in the presence of CCK8) − OD₄₅₀ (Only medium)]/[OD₄₅₀ (Cells treated with DMSO in the presence of CCK8) − OD₄₅₀ (Only medium)] ×100.

Author contributions

Feng Wang: Conceptualization; Resources; Supervision; Funding acquisition; Investigation; Writing—original draft; Writing—review & editing. Qinlan Wang: Investigation; Formal analysis; Writing—original draft; Writing—review & editing. Ting Zhu: Investigation. Naijun Miao: Investigation; Data curation. Yingying Qu: Investigation. Zhuning Wang: Investigation. Yinong Chao: Investigation; Data curation. Jing Wang: Supervision; Software; Data curation. Wei Wu: Investigation. Xinyi Xu: Investigation; Methodology. Chenqi Xu: Methodology. Li Xia: Methodology.

Disclosure and competing interests statement

The authors declare that they have no competing interests.

Supporting information

Expanded View Figures PDF

Table EV1

Source Data for Expanded View

Source Data for Figure 1

Source Data for Figure 2

Source Data for Figure 3

Source Data for Figure 4

Source Data for Figure 5

The EMBO Journal (2022) 41: e110636.

Data availability

The RNA sequencing data have been deposited in NCBI's Gene Expression Omnibus (GEO) and are accessible through accession number GSE202131 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE202131).