Research ArticleCancer

Thrombin contributes to cancer immune evasion via proteolysis of platelet-bound GARP to activate LTGF-β

See allHide authors and affiliations

Science Translational Medicine  08 Jan 2020:
Vol. 12, Issue 525, eaay4860
DOI: 10.1126/scitranslmed.aay4860
  • Fig. 1 N-terminal fragment of mouse GARP is shed from the cell surface.

    (A) Western blot analysis of GARP in the total cell lysate of WT and gp96 KO pre–B cells. Cell lysates were analyzed by using antibodies against mouse GARP and mouse gp96. Reducing condition gel was used. (B) Parallel Western blot analysis of cell lysates and conditioned medium of WT pre–B cells expressing GARP or control vector. Reducing condition gel was used. EV, empty vector. (A) and (B) are representative of more than three independent biological experiments. (C) Mass spectrometry analysis of the fragment present in the conditioned medium of pre–B cells expressing GARP. #PSMS, number of identified sequences (peptide spectrum matches). G3XA59 is UniProtKB for GARP.

  • Fig. 2 Thrombin is responsible for cleavage of human GARP.

    (A) Western blot analysis of cell lysate and conditioned medium of HEK293 cells expressing human GARP (hGARP) or a control (Ctrl) empty vector. Reducing condition gel was used. β-Actin was used as a loading control. (B) Western blot analysis of human platelet lysate (25, 50, and 100 μg) and the releasate obtained from increasing concentrations of platelets (107, 108, and 109 per ml). Reducing condition gel was used. (C) Western blot analysis of HEK293 cells expressing wild-type (WT) human GARP or GARP with single (Mut 1 or Mut 2) or both thrombin binding sites mutated (2xMut). Reducing condition gel was used. (D) Western blot analysis of 1 μg of recombinant human GARP digested with increasing concentration of human thrombin with active (left) and blocked (right) activity. Reducing condition gel was used. (E) Representative molecular model of a thrombin, GARP, and LTGF-β1 complex showing that thrombin can bind GARP on the surface opposite of LTGF-β1. Western blot analysis for TGF-β1 in cell lysate (F) and the conditioned medium (G) of HEK293 cells expressing TGF-β1 only or TGF-β1 in combination with WT or mutant GARP. (H) Same samples as in (G) were immunoblotted for GARP. Nonreducing condition gel was used for (F) to (H). At least three independent experiments were performed with similar findings.

  • Fig. 3 Molecular and biochemical mapping demonstrates mouse cell surface GARP as a thrombin substrate.

    (A) GARP-expressing pre–B cells were digested with increasing concentrations of thrombin (0, 1, 2, and 4 μg in 25 μl), followed by Western blot analysis. Reducing condition gel was used. (B) Sequence of GARP fragments containing the two PR thrombin binding sites, T250. (C) Western blotting analysis of GARP-expressing pre–B cells digested with 4 μg of thrombin in the presence or absence of 4 μg of T250. Reducing condition gel was used. (D) Structural model of the mouse GARP/LTGF-β1 interaction. (E) Flow cytometry analysis of cell surface GARP on WT pre–B cells or pre–B cells expressing various forms of mutant GARP or empty vector. (F) Western blot analysis of GARP in whole-cell lysates and conditioned medium of pre–B cells expressing WT or mutant GARP. Reducing condition gel electrophoresis was performed.

  • Fig. 4 Thrombin cleavage of mouse GARP liberates mature TGF-β1.

    (A) Mouse WT and GARP KO platelets were collected and stimulated with thrombin or not. Western blot analysis was performed on platelet lysate using anti-GARP antibody. Reducing condition gel was used. (B) Platelet releasate from WT and GARP KO platelets was obtained with and without thrombin stimulation and analyzed by Western blot under nonreducing and nondenaturing conditions using anti–TGF-β1 antibody. Nonreducing condition gel was used. (C) Platelet releasate from WT and GARP KO platelets was obtained with and without thrombin stimulation, and active TGF-β1 was quantified by ELISA. Data shown as means ± SD (n = 3). (D) Mouse splenocytes were labeled with carboxyfluorescein diacetate succinimidyl ester (CFSE), followed by activation with anti-CD3ε/CD28 and interleukin-2 for 3 days in the presence of different concentrations of platelet releasate. CD4+ lymphocytes were then analyzed by flow cytometry. (E) WT platelets were untreated or stimulated with the indicated conditions. Platelet releasate was collected, and active TGF-β1 was quantified by ELISA. Data shown as means ± SD (n = 4). (F) Flow cytometry analysis of P-selectin expression in mouse platelets stimulated in the presence of thrombin and/or T250 peptide.

  • Fig. 5 Cleavage of platelet GARP mediates serum LTGF-β1 activation and tumor progression.

    Female C57BL/6 mice were treated for 1 week with dabigatran etexilate, and blood was drawn and analyzed. (A) Platelet number and activation status. (B) Serum active TGF-β1 (as a measurement of active TGF-β1 release from blood clots) and total TGF-β1 quantified by ELISA. ns, not significant. (C) C57BL/6 female mice were treated for 1 week with dabigatran etexilate or clopidogrel and aspirin or left untreated. Serum active TGF-β1 was evaluated by ELISA. Mice with WT and GARP KO platelets were challenged with MC38 cells and treated with dabigatran etexilate or left untreated. (D) Tumor area. (E) Serum active TGF-β1 from MC38-bearing mice was evaluated by ELISA. (F and G) Dabigatran treatment enhances GARP and LAP expression on circulating platelets. WT mice were treated with dabigatran (DAB; 3 mg per mouse) or sham control (UT) daily for 1 week. Platelets were purified and analyzed by flow cytometry for surface expression of (F) GARP or (G) LAP. MFI, mean fluorescent intensity. Data presented as means ± SEM, n = 8 to 10 mice per group. Two-tailed, paired Student’s t test was used in (B), (F), and (G); two-tailed, independent Student’s t test was used in (C) and (E); and repeated measurement two-way ANOVA was performed in (D).

  • Fig. 6 Cleavage of platelet GARP supports TGF-β1–mediated immune evasion in the tumor microenvironment.

    (A) Representative images with relative expression intensity of TGF-β1, collagen, SMA1, and CD61 in MC38 tumors collected from WT mice treated with dabigatran etexilate for 3 weeks or left untreated. Each symbol represents an individual mouse. Scale bars, 100 μm (TGF-β1 and SMA1) and 50 μm (collagen and CD61). (B and C) Mass cytometry analysis of intratumoral cell populations isolated from C57BL/6 mice (n = 10 per group) subcutaneously injected with MC38 and treated daily with dabigatran etexilate for 3 weeks. (B) The combined dataset was down-sampled to 2500 cells per sample for a total of 50,000 cells and subjected to dimensionality reduction using the t-distributed stochastic neighbor embedding (t-SNE) algorithm with the sample-specific composition in tumors of dabigatran etexilate–treated or untreated mice. Relative distribution of immune and nonimmune cells is indicated in the right. CD8+ T cells, NK cells, and neutrophils are indicated by arrows. (C) Statistical comparison of the frequencies of immune cell populations. DC, dendritic cells; TILs, tumor-infiltrating lymphocytes. (D) Representative image of GARP staining in regulatory T cells (Tregs). (E) Flow cytometry analysis of the MFI of GARP expression on Tregs isolated from MC38 tumors in mice treated with dabigatran and control. (F) Tumor-infiltrating CD8+ T cells were treated with pyridine-2-aldoxime methiodide/ionomycin for 4 hours, followed by intracellular staining of IFNγ. Two-tailed, independent Student’s t test was used in (A) and (E); adjusted P value based on computational framework diffcyt (67) was used in (C).

  • Fig. 7 Blocking platelet GARP cleavage enhances immune checkpoint blockade therapy of cancer.

    WT C57BL/6 mice were injected with MC38 colon cancer cells and left untreated or treated with dabigatran etexilate alone, PD-1 blocking antibody (Ab) alone, or combination of both. (A) Tumor growth curves. (B) Mouse survival. (C) Serum active TGF-β1 measured by ELISA. BALB/c female mice were injected into the fourth left mammary fat pad with EMT-6 triple-negative breast cancer cells and treated with dabigatran etexilate alone, PD-1 blocking antibody alone, or both. (D) Tumor growth curves. (E) Mouse survival. (F) Metastatic burden in the lungs. (G) Serum active TGF-β1 in EMT-6–bearing BALB/c female mice treated with indicated conditions. Repeated measurement two-way ANOVA was performed in (A), log-rank (Mantel-Cox) was performed in (B) and (E), and two-tailed, independent Student’s t test was used in (F). The experiment with MC38 tumor models was performed three times, and EMT-6 was performed two times. For each experiment, six to eight animals per group were used. All P values reflect comparison to the combination group (dabigatran plus anti–PD-1).

  • Fig. 8 Both sGARP and sGARP–LTGF-β1 complex are present in patients with cancer.

    (A) sGARP was detected by ELISA in plasma samples collected from patients with prostate cancer and healthy controls. (B) Correlation analysis between GARP-positive or GARP-negative plasma samples and PSA concentration greater or less than 10 ng/ml (left) or the presence of metastasis (right). (C) GARP–LTGF-β1 sandwich ELISA performed on plasma samples collected in heparin-coated tubes from patients with prostate cancer and healthy controls. OD, optical density. Two-tailed, independent Student’s t test was used in (A) and (C). χ2 test was used in (B).

Supplementary Materials

  • stm.sciencemag.org/cgi/content/full/12/525/eaay4860/DC1

    Fig. S1. Both mouse and human GARP contain consensus PR thrombin cleavage sites.

    Fig. S2. Mutation of one thrombin binding site (Mut 1 or Mut 2) does not alter surface expression of GARP-LAP complex.

    Data file S1. Original data.

  • The PDF file includes:

    • Fig. S1. Both mouse and human GARP contain consensus PR thrombin cleavage sites.
    • Fig. S2. Mutation of one thrombin binding site (Mut 1 or Mut 2) does not alter surface expression of GARP-LAP complex.

    [Download PDF]

    Other Supplementary Material for this manuscript includes the following:

Stay Connected to Science Translational Medicine

Navigate This Article