Introduction

Accumulating evidence have clearly shown strong interactions between tumor cells and microenvironment, confirming the influence of the surrounding cells on tumor progression [1]. Platelets can adhere to the surface of tumor cells, and have been identified as a vital external stimuli to tumor cells in the process of hematogenous metastasis. As a matter of fact, platelets have been found to promote tumor growth and metastasis. They can form a physical shield surrounding tumor cells to protect against the lysis by natural killer (NK) cells [2,3,4,5], reduce the damage caused by shear stress [6, 7], and promote the adhesion to endothelium [7, 8]. In addition, we speculate that platelets may also provide additional signals for tumor cells to affect their metastatic potential. To explore the mechanisms involved, we treated murine CT-26 colon cancer cells with platelets and then performed RNA-sequencing (RNA-seq) analysis. Among differentially expressed genes, CD39 is one of the most significantly up-regulated genes.

CD39 has attracted much attention as a newly discovered immune checkpoint molecule in recent years. It can hydrolyze extracellular adenosine triphosphate (ATP) and adenosine diphosphate (ADP) to adenosine monophosphate (AMP) [9, 10], and AMP is further hydrolyzed by CD73 to adenosine [11]. Adenosine is a potent immune regulatory molecule that exerts immunosuppressive effects by binding to adenosine A2A receptor (A2AR) expressed by lymphocytes [12]. CD39 is identified to be the immunological switch shifting the ATP-induced pro-inflammatory status to the adenosine-driven anti-inflammatory status [13]. CD39 shows a high expression level in tumor-infiltrating immunocytes—especially the effector T cells and regulatory T cells (Tregs). The number of tumor-infiltrating CD39highCD8+ T cells increases as tumor grows, which display exhaustion feature [14]. The presence of high levels of CD39 on regulatory immune subsets is directly correlated with immune dysfunction [15, 16]. In CD39-deficient mice, the growth of several syngeneic tumors is inhibited [17, 18]. In addition, the CD39-deficient mice also show metastasis resistance in spontaneous metastasis or disseminated disease models [19, 20]. The promotion of tumor metastasis is observed in transgenic mice with CD39 over-expression [21]. Besides, blocking CD39 activity pharmacologically using the inhibitor or antagonistic antibody remarkably restricts tumor development and improves the anticancer immunity [17, 19, 22, 23].

In the present study we demonstrated that direct contact with platelets induces CD39 expression in tumor cells and the ability of platelets to prime tumor cells for metastasis depends on CD39. Furthermore, different from previous studies, we explored the effect of CD39 expressed by tumor cells on tumor metastasis. The results show that platelet-mediated CD39 induction inhibits anti-tumor immunity and promotes tumor metastasis, indicating that platelets not only serve as the physical shield to tumor cells, but also offer pro-metastatic signals through inducing CD39 expression.

Material and Methods

Human tissue samples

Tumor tissues were collected from colon cancer patients or patients with colon cancer and lung metastasis. All the samples were obtained from the Affiliated Hospital of Jining Medical University (Shandong, China). The present study was approved by the Ethics Committee of the Affiliated Hospital of Jining Medical University. Informed consent was obtained from all individual participants included in the study.

Mice, cell lines and reagents

BALB/c, NCG and CD39 gene-knockout (CD39−/−) mice provided by Nanjing University Model Animal Institute were raised in Jining Medical University at specific pathogen-free (SPF) conditions. NCG mice lacking mature NK cells and B/T lymphocytes were created through sequential clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) editing of the protein kinase, DNA-activated, Catalytic subunit (PRKDC) and interleukin 2 receptor gamma (IL2RG) genes in the NOD/ShiLtJ Gpt mice. The CD39−/− mice were generated through genome engineering mediated by CRISPR/Cas9. Mice were used at 6–8 weeks of age and randomly divided into different treatment groups. The sample size was determined based on published data or preliminary experiments [24]. Investigators were blinded to the group allocation during the procedure and data analysis. All experiments involving animals were approved by the Medical Animal Care & Welfare Committee of Jining Medical University. Human colon cancer HCT116 cells, murine prostate cancer RM-1 cells and murine colon cancer CT-26 cells were obtained from American Type Culture Collection (ATCC, Manassas, VA, USA). Cultured cells were all mycoplasma negative. SB203580, SCH58261, crystal violet and Calcein-AM were purchased from Beyotime (Shanghai, China). PSB-603 was purchased from MedChemExpress (New Jersey, USA). The antibodies used are detailed in Supplementary Table 1.

Preparation of platelets and platelet fractions

Mouse platelets and platelet fractions were prepared as described previously [25]. Platelets at a concentration of 1.5 × 108 platelets/ml were added to the cell medium when used to treat tumor cells in vitro. Human platelets were obtained from the Affiliated Hospital of Jining Medical University (Shandong, China).

Platelet depletion

To deplete platelets, mice received 2 µg/g rat monoclonal antibodies against mouse glycoprotein (GP) Ibα intravenously. CT-26 cells were injected through tail veins 12 h after antibody administration.

RNA extraction and real-time quantitative PCR (RT-qPCR)

The EasyPure1 RNA Kit (Invitrogen, Carlsbad, CA, USA) was applied in isolating the total RNA following instructions. Reverse transcription was performed using Transcriptor First Strand cDNA Synthesis Kit (Roche, Mannheim, Germany). Target gene expression levels were analyzed by SYBR Green I-based RT-qPCR. The sequences of each primer set are shown in Supplementary Table 2.

Western blot analysis

Western blot was performed according to the established protocols described previously [26]. The RIPA buffer was utilized for extracting cellular proteins and BCA protein assay kit (Solarbio, Beijing, China) was utilized for quantification. Proteins were separated through SDS-PAGE, transferred into the PVDF membrane, and incubated with the primary antibody. Afterwards, the membrane was subjected to further incubation using specific horseradish peroxidase (HRP)-conjugated secondary antibodies (Beyotime, Shanghai, China). Quantity One v4.62 was applied in analyzing protein expression after exposure to enhanced chemiluminescence (ECL, Pierce, Rockford, IL).

RNA-Seq

RNA was extracted using the EasyPure1 RNA Kit (Invitrogen, Carlsbad, CA) following specific instructions and RNA quality was assessed by BGI Tech (Shenzhen, China). RNA-Seq was conducted using the BGISEQ-500 sequencing platform in BGI Tech.

Plasmids and transfection

pCMV3-CD39 plasmid was obtained from Sino Biological Inc (Beijing, China). pGPU6/Hygro shRNA expression vector that contained CD39-targeting DNA oligos was provided by Genepharma Inc (Shanghai, China), and oligonucleotides sequences were shown below: 5’-GCATGCGCTTGCTTAGAATGG-3’ (shRNA1), 5’-GGGCAGATTCACTCAGGAACA-3’ (shRNA2), 5’-GCCTGTTCATCTATAGCAAGC-3’ (shRNA3) and 5’-TTCTCCGAACGTGTCACGT-3’ (negative control shRNA, NC-shRNA). Plasmids were transfected with the Effectene Transfection Reagent (Qiagen, Valencia, CA, USA) according to the manufacturer’s instructions. The cells with stable transfection were obtained by hygromycin B selection.

Flow cytometry

Blood was collected into the sodium heparin-coated tubes. Red blood cells (RBCs) lysis buffer was used to lyse RBCs. The cells were then incubated with mouse antibodies against CD3, CD4, CD8, programmed cell death 1 (PD1), T cell immunoglobulin mucin-3 (TIM3) and lymphocyte-activation gene 3 (LAG3) according to the standard procedure. After washing by PBS, cells were analyzed on a flow cytometer (BD BioSciences, Bedford, MA, USA).

ATP hydrolysis assay

After 30 min of cell incubation with 10 μM ATP, concentration of unhydrolyzed extracellular ATP was determined using the ATP detection kit (Beyotime, Beijing, China) according to the instructions.

Cell viability assay

Cell viability was determined using CCK-8 solution (Beyotime, Beijing, China) according to the manufacturer’s instructions.

Cytotoxic T lymphocyte (CTL) cytotoxicity

CTLs were prepared as described previously [27]. Splenocytes from CT-26 tumor-bearing mice (after inoculation for 10 days) were cocultured with 100 U/ml interleukin 2 (IL-2) and irradiated CT-26 tumor cells at a ratio of 10:1. After 10 days, viable lymphocytes were collected as CTLs.

Cell migration and invasion assay

Cell migration capacity was assessed by wound-healing assay. The cell monolayer was scraped using a micropipette tip and cells were then incubated in serum-free medium for 24 h. All assays were carried out in triplicate. For the invasion assays, 24-well Matrigel invasion chambers (Corning, NY, USA) were used. Cells (5 × 104) were resuspended in 100 µl of serum-free medium in the upper chamber, and 600 μl of medium with 10% fetal bovine serum (FBS) was added to the bottom chamber. The cells that migrated through the pores were fixed with methanol and stained with crystal violet. The cell number was calculated under the phase contrast microscope (Nikon, Tokyo, Japan).

Immunohistochemistry (IHC)

IHC assay was conducted using Histostain Plus Kits (Bioss, China) according to the manufacturer’s instructions. The immunohistochemical staining was semi-quantified using the H-score method [28]. The H-score (range, 0–300) is given as the sum of the percent staining multiplied by an ordinal value corresponding to the intensity degree (where 0, 1, 2 and 3 indicate none, weak, moderate and strong, respectively).

Statistical analysis

Results were presented in the manner of mean ± standard deviation. Two-sided Student’s t test was used to compare the two groups. One-way analysis of variance (ANOVA) was used to analyze differences among three or more groups. The survival curves were determined using the Kaplan-Meier method. A difference of P < 0.05 indicated statistical significance.

Results

Platelets induce CD39 expression in tumor cells

To explore the mechanisms involved in the promotion of tumor metastasis induced by platelets, we treated murine CT-26 colon cancer cells with platelets and performed RNA-seq analysis. Platelets have a significant impact on tumor cells and numerous differentially expressed genes were identified. Among the most significantly up-regulated genes, CD39 has aroused our interest. Considering CD39 is a driver of immunosuppression, we speculate that platelet-induced CD39 expression may contribute to the immune escape of tumor cells during hematogenous metastasis. The expression of CD39 in CT-26 cells was verified by RT-qPCR and western blot, and showed consistent with the results of RNA-seq (Fig. 1a). Furthermore, this finding was reproduced in murine RM-1 prostate cancer cells (Fig. 1b). To avoid possible interference with CD39 detection by residual platelets, we removed platelets prior to RNA or protein extraction of tumor cells. The results of immunofluorescence staining showed that there were very few platelets remain attached to tumor cells treated with platelets (Fig. 1c). We also detected the expression of platelet marker CD41 and obtained similar results (Fig. 1d). In addition, we found that white and red blood cells had no effect on the expression of CD39 in tumor cells (Fig. S1). CD39 collaborates with CD73 to produce adenosine, so we examined CD73 expression levels in CT-26 and RM-1 cells. As shown in Figure S2, CD73 was expressed in both CT-26 and RM-1 cells. We next sought to define the molecular mechanisms underlying platelet-mediated induction of CD39. First, we analyzed whether platelets promoted CD39 expression in tumor cells via direct contact. The results showed that separation of platelets from tumor cells using a semipermeable membrane abolished the up-regulation of CD39 in CT-26 (Fig. 1e) and RM-1 cells (Fig. 1f).

Fig. 1: Direct contact with platelets induces CD39 expression in tumor cells.
figure 1

a Upper panel, relative CD39 mRNA levels in CT-26 cells treated with buffer or platelets for 12 h. Values are normalized to GAPDH (n = 3). Lower panel, CD39 protein levels in CT-26 cells treated with buffer or platelets for 24 h. b The experimental protocol in RM-1 cells is the same as that in CT-26 cells (n = 3). c Platelets were dyed with Calcein-AM (2 μM), washed and added to tumor cells. After 12 h, the co-culture was washed to remove non-adherent platelets. Cells were then collected and centrifuged (200 g, 10 min) twice to remove residual platelets. Washed cells were smeared and photographed with fluorescence microscopy (lower panel). Co-culture with platelets not removed is shown in the upper panel. Scale bar = 100 μm. d CD41 protein levels in CT-26 and RM-1cells treated with platelets for 24 h. Protein extracted from mouse platelets was used as a positive control. GAPDH was used as an internal control. e Upper panel, relative CD39 mRNA levels in CT-26 cells treated with buffer or platelets seeded either at the bottom or in the upper chamber of a transwell (0.4 μm pore size) for 12 h. Values are normalized to GAPDH (n = 3). Lower panel, CD39 protein levels in CT-26 cells treated with buffer or platelets seeded either at the bottom or in the upper chamber of a transwell (0.4 μm pore size) for 24 h. f The experimental protocol in RM-1 cells is the same as that in CT-26 cells (n = 3). **p < 0.01, ***p < 0.001.

Full size image

As platelets can be activated by tumor cells and release a variety of cytokines [29], we next investigated whether exposure to the releasates or pellets from activated platelets would be sufficient to promote the expression of CD39 in tumor cells. Platelets were therefore activated with thrombin, and platelet fractions, including the releasates and pellets, were separated by centrifugation. The results showed that cell pellets induced the expression of CD39 in CT-26 (Fig. 2a) and RM-1 cells (Fig. 2b), whereas releasates had no significant effect. We also treated CT-26 and RM-1 cells with cytokines that could be released by platelets, including epidermal growth factor (EGF), fibroblast growth factor (FGF)-2, insulin-like growth factor (IGF)-1, transforming growth factor β (TGFβ)-1 and platelet-derived growth factor (PDGF), and no significant changes in CD39 expression were observed (data not shown). Meanwhile, thrombin also did not affect CD39 expression (Fig. S3). Next, we explored the signaling pathway through which platelets regulated the expression of CD39. As shown in Fig. 2c, d, platelets activated the p38 signaling pathway in CT-26 and RM-1 cells. Moreover, p38 signaling pathway inhibitor SB203580 suppressed platelet-mediated up-regulation of CD39 in CT-26 (Fig. 2e) and RM-1 cells (Fig. 2f), demonstrating that platelets enhanced the expression of CD39 through p38 signaling pathway.

Fig. 2: Platelet up-regulated the expression of CD39 via p38 signal pathway.
figure 2

a Upper panel, relative CD39 mRNA levels in CT-26 cells treated with buffer, platelets, releasate from activated platelets (releasate), or the pellet fraction from activated platelets (pellet) for 12 h. Values are normalized to GAPDH (n = 3). Lower panel, CD39 protein levels in CT-26 cells treated with buffer, platelets, releasate from activated platelets (releasate), or the pellet fraction from activated platelets (pellet) for 24 h. b The experimental protocol in RM-1 cells is the same as that in CT-26 cells (n = 3). c Phosphor or total p38 protein levels in CT-26 cells treated with buffer or platelets for 30 min. d Phosphor or total p38 protein levels in RM-1 cells treated with buffer or platelets for 30 min. e Upper panel, relative CD39 mRNA levels in CT-26 cells treated with buffer, platelets or platelets + SB203580 (10 μM) for 12 h. Values are normalized to GAPDH (n = 3). Lower panel, CD39 protein levels in CT-26 cells treated with buffer, platelets or platelets + SB203580 (10 μM) for 24 h. f The experimental protocol in RM-1 cells is the same as that in CT-26 cells (n = 3). **p < 0.01, ***p < 0.001.

Full size image

The Effects of CD39 on tumor cells

We next evaluated the effects of CD39 on tumor cells. First, we constructed CD39- overexpression and -silencing CT-26 cell lines (Fig. S4). To assess broad effects mediated by CD39, we performed RNA-seq analysis. Differential gene expression analysis revealed significant changes induced by CD39 overexpression (Fig. 3a). We further examined the effects of CD39 on the behaviors of tumor cells. Neither overexpression nor silencing of CD39 affected the proliferation and migration of tumor cells (Fig. 3b–d). However, we found that overexpression of CD39 promoted cell invasion (Fig. 3e), possibly because it up-regulated the expression of matrix metalloproteinase (MMP)-2 and MMP-3, as demonstrated by RNA-seq analysis. Instead, silencing of CD39 had no significant effect on cell invasion (Fig. 3f), which may be due to the low background expression of CD39 in CT-26 cells.

Fig. 3: The effects of CD39 on tumor cells.
figure 3

a Heat map of differentially expressed genes in vector or CD39-expressing CT-26 cells. b Effects of CD39 on proliferation of CT-26 cells. Cells were seeded in 96-well plates. CCK-8 solution was added to the medium 48 hours later, and incubated for another 2 h at 37 °C. Absorbances of each well were quantified at 450 nm (n = 3). c, d Effects of CD39 on cell migration activity. Cell migration activity was evaluated using a wound-healing assay. Cells were plated in 6-well dishes, and the cell monolayer was scraped using a micropipette tip. Subsequently, cells were incubated in serum-free medium for 24 h (n = 3). The data are normalized to the control group. Scale bar = 250 μm. e, f Effects of CD39 on cell invasion activity. Matrigel invasion chambers was used for the invasion assays. Cells that migrated through the pores onto the lower side of the membrane were stained with crystal violet and counted using a phase contrast microscope (n = 3). Scale bar = 50 μm. **p < 0.01.

Full size image

Platelets prime tumor cells for metastasis depends on CD39

We first examined whether the pro-metastatic effects of platelets was related to the induction of CD39 in tumor cells. In the experimental tumor metastasis model by injecting tumor cells intravenously, pretreatment of CT-26 with platelets increased lung metastasis, while silencing CD39 in CT-26 cells reduced this effect (Fig. 4a). This suggests that the ability of platelets to prime tumor cells for metastasis depends on CD39 expressed by tumor cells. Moreover, we examined tumor metastasis in platelet-depleted mice which received platelet depletion antibody against mouse GPIbα. CD39 silencing resulted in reduced tumor metastasis in mice that received isotype control antibody, and this difference was significantly reduced in platelet-depleted mice (Fig. 4b). This result further confirmed the important role of platelets in promoting tumor metastasis by regulating CD39 expression.

Fig. 4: Platelets prime tumor cells for metastasis depends on CD39.
figure 4

a Each group of mice was injected intravenously with 5 × 105 NC- or CD39-shRNA expressing CT-26 cells which were pretreated with buffer or platelets for 24 h. The lungs were removed and photographed after 12 days (n = 5). b Mice received 2 µg/g platelet depletion antibody or isotype control antibody intravenously, and 5 × 105 CT-26 cells were injected through tail veins 12 h after antibody administration. The lungs were removed and photographed after 12 days (n = 5). *p < 0.05, **p < 0.01, ***p < 0.001.

Full size image

The impact of CD39 on tumor growth has mostly been delineated using global CD39 targeted mice; previous studies have shown that tumor growth and metastasis were significantly inhibited in CD39 knockout mice [17, 18, 23, 30]. However, the effect of CD39 expressed by tumor cells on tumor metastasis is unclear. Next, we evaluated the effect of overexpression or silencing of CD39 on tumor metastasis in the experimental tumor metastasis model. The results showed that overexpression of CD39 promoted lung metastasis (Fig. 5a), whereas silencing of CD39 resulted in fewer lung metastases (Fig. 5b) and improved overall survival in mice (Fig. 5c). In addition to tumor cells, CD39 is also expressed in immune cells and vascular endothelial cells. In order to eliminate the effect of CD39 expressed by these cells, we detected the effect of CD39 expressed by tumor cells on tumor metastasis in CD39−/− mice and obtained similar results as in wild type (WT) mice (Fig. 5d–f). These results indicate that platelets prime tumor cells for metastasis depends on CD39 and CD39 expressed by tumor cells has pro-metastatic effects.

Fig. 5: CD39 promoted tumor metastasis.
figure 5

a Each group of WT mice was injected intravenously with 5 × 105 vector or CD39-expressing CT-26 cells. The lungs were removed and photographed after 12 days (n = 5). b Each group of WT mice was injected intravenously with 5 × 105 NC- or CD39-shRNA expressing CT-26 cells. The lungs were removed and photographed after 12 days (n = 5). c The survival period of the CD39 shRNA-expressing tumor-bearing WT mice was significantly prolonged compared with control groups (n = 10, P < 0.001; Kaplan-Meier analysis). d Each group of CD39−/− mice was injected intravenously with 5 × 105 vector or CD39-expressing CT-26 cells. The lungs were removed and photographed after 12 days (n = 5). e Each group of CD39−/− mice was injected intravenously with 5 × 105 NC- or CD39-shRNA expressing CT-26 cells. The lungs were removed and photographed after 12 days (n = 5). f The survival period of the CD39 shRNA-expressing tumor-bearing CD39−/− mice was significantly prolonged compared with control groups (n = 10, P < 0.01; Kaplan-Meier analysis). **p < 0.01, ***p < 0.001.

Full size image

CD39 suppresses antitumor immunity

We first evaluated whether the pro-metastatic effect of CD39 is achieved primarily through the modulation of the immune response. We intravenously injected CD39-overexpressing or -silencing CT-26 cells into immunodeficient NCG mice. There was no significant difference in the number of pulmonary nodules between different groups, although NCG mice showed a significantly higher number of pulmonary nodules than WT mice (Fig. 6a, b). These data suggest that CD39 affects tumor metastasis depending on an active immune system. In addition, we found that A2AR antagonist SCH58261 could significantly inhibit CD39-mediated promotion of tumor metastasis (Fig. 6c), whereas adenosine A2B receptor (A2BR) antagonist PSB-603 showed no significant effect (Fig. S5). These results indicate that the ability of CD39 to prime tumor cells for metastasis depended on adenosine and A2AR. CD39 expressed by immune cells can exert immunosuppressive effects [14, 31]. However, the effect of CD39 expressed by tumor cells on anti-tumor immune response is not clear. First, we examined whether CD39 affects cancer cell killing by CTLs. CT-26-specific CTLs were generated as previously described [27] and added to CT-26 cell cultures. As expected, CD39-overexpressing CT-26 cells resisted killing by CTLs (Fig. 6d). Next, we further examined the effect of CD39 on CD8+ T cell exhaustion. Results showed that overexpression of CD39 in tumor cells enhanced the expression of inhibitory receptors such as PD1, LAG3 and TIM3 in circulating CD8+ T cells (Fig. 6e). In contrast, CD39 overexpression had no effect on CD4+ T cells (Fig. S6). Collectively, the above data shows that CD39 expressed by tumor cells can suppress CD8+ T cell-mediated antitumor immunity, which may be contributing to tumor cell immune escape.

Fig. 6: CD39 suppressed antitumor immunity.
figure 6

a Each group of NCG mice was injected intravenously with 5 × 105 vector or CD39-expressing CT-26 cells. The lungs were removed and photographed after 10 days (n = 5). b Each group of NCG mice was injected intravenously with 5 × 105 NC- or CD39-shRNA expressing CT-26 cells. The lungs were removed and photographed after 10 days (n = 5). c Each group of mice was injected intravenously with 5 × 105 vector or CD39-expressing CT-26 cells. Meanwhile, tumor-bearing mice were treated with SCH58261 (5 mg/kg, i.p.) or vehicle every other day. The lungs were removed and photographed after 12 days (n = 5). d CD39-overexpressing CT-26 cells resisted killing by CTLs. Tumor-specific CTLs were added into the vector or CD39-expressing CT-26 cells, and the number of remaining viable cancer cells was quantified by a CCK8 assay 72 h later (n = 3). e Each group of mice was injected intravenously with 2 × 106 vector or CD39-expressing CT-26 cells. The expression of PD-1, TIM-3 and LAG-3 in CD8+ T cells were measured by flow cytometry 48 h later (n = 3). *p < 0.05, **p < 0.01, ***p < 0.001.

Full size image

Expression of CD39 in human tumor cells and tissues

Based on the animal experiments, we then assessed CD39 expression in human tumor cells and tissues. Results showed that human platelets also could up-regulate the expression of CD39 in human HCT116 colon cancer cells (Fig. 7a). IHC analysis of tumor samples from patients showed that CD39 expression was significantly higher in lung metastases from colon cancer than in colon cancer in situ (Fig. 7b, c). Considering that colon cancer cells metastasize to lung through the blood, these results indicate that the enhanced expression of CD39 in lung metastases may be related to the interaction between tumor cells and platelets during hematogenous metastasis.

Fig. 7: Expression of CD39 in human tumor cells and tissues.
figure 7

a Upper panel, relative CD39 mRNA levels in HCT116 cells treated with buffer or platelets for 12 h. Values are normalized to GAPDH (n = 3). Lower panel, CD39 protein levels in HCT116 cells treated with buffer or platelets for 24 h. b The expression of CD39 in lung metastases (n = 9) and in situ tissues (n = 10) of colon cancer was detected by IHC. Scale bar = 100 μm. c The H-score is given as the sum of the percent staining multiplied by an ordinal value corresponding to the intensity level (0 = none, 1 = weak, 2 = moderate, 3 = strong). d Research hypothesis. Previous studies have shown that platelets can be activated by tumor cells and release endogenous ADP. Here we show that platelets can promote the expression of CD39 in tumor cells. On the basis of these information, we speculate that platelets and tumor cells interact and influence each other. Tumor cells activate platelets and promote the release of ADP; platelets induce the expression of CD39 in tumor cells. Interestingly, it appears that platelets not only induce CD39 expression, but also provide substrates for CD39. ADP can be hydrolyzed to AMP by CD39, and AMP is then hydrolyzed to adenosine by CD73. Adenosine can inhibit CD8+ T cell-mediated anti-tumor immune response, thus promoting metastasis. **p < 0.01.

Full size image

Discussion

Tumor microenvironment (TME) plays a key role in tumorigenesis and development [32]. The primary TME provides signals for the tumor cells to invade and access the vasculature [33]. Nonetheless, the metastatic potential of tumor cells can be further defined responding to signals provided during the intravascular transit [25]. The present study suggests that the interaction between platelets and tumor cells increases tumor cells metastatic ability through inhibiting antitumor immune response. Immune surveillance poses a major threat to tumor cells during hematogenous metastasis, so platelet-mediated immunosuppression may play a critical role in tumor metastasis. Our findings give evidence that platelets not only serve as the physical shield to tumor cells, and reveal a new mechanism of tumor metastasis mediated by platelets.

Platelets can adhere to the surface of tumor cells, and serve as an important external stimulating factor for tumor cells during hematogenous metastasis. As discovered in this work, platelets increased the CD39 level within tumor cells in the manner of direct contact. Notably, tumor cells can activate platelets to release various cytokines, such as PDGF, FGF2, EGF, TGFβ1, and IGF1 [29]. To elucidate whether platelets regulated the expression of CD39 by releasing cytokines, this study adopted thrombin to activate platelets in vitro, and then collected the supernatant and pellet fractions to treat tumor cells. As a result, the activated platelet pellets, rather than supernatants, induced CD39 expression. In addition, EGF, FGF2, IGF1, TGFβ1, and PDGF were unable to modulate CD39 expression (data not shown). Interestingly, platelets can be activated by tumor cells and release endogenous ADP [34, 35]. Therefore, it is reasonable to assume that platelet-released ADP can be subsequently hydrolyzed on the surface of tumor cells by CD39, which is up-regulated by platelets (Fig. 7d). This indicates that platelets and tumor cells are interrelated and interact with each other, leading to the generation of adenosine.

Our data showed that SCH58261 inhibited CD39-mediated pro-metastatic effects. However, SCH58261 could not completely abolish the effect of CD39 on tumor metastasis, potentially because CD39 could also facilitate metastasis through other ways, such as promoting cell invasion. The effects of CD39 expressed by immune cells on antitumor immunity have been evaluated in prior studies [14, 15]. This work demonstrated that, CD39 expression within tumor cells had certain effect on the antitumor immune responses. Notably, overexpression of CD39 led to the exhaustion of circulating CD8+ T cells. The underlying mechanism may be similar to that of Tregs, which can mediate immune suppression via the CD39-CD73-adenosine axis. CD39 and CD73 are coexpressed in Tregs, and Treg-produced adenosine can induce CD8+ T cell exhaustion through the A2AR pathway [36, 37]. These results indicate that CD39 may facilitate tumor metastasis through suppressing antitumor immunity. Interestingly, both CT-26 and RM-1 cells are resistant to anti-PD1 therapy [23], suggesting that CD39 may be a potential target for “cold” anti-PD1 refractory tumors.

Finally, IHC analysis of tumor samples from patients showed that CD39 expression was higher in lung metastases from colon cancer than in colon cancer in situ. Although this may be related to the interactions between colon cancer cells and platelets during hematogenous metastasis, it may also be caused by other factors. Therefore, more investigation is needed to explore the underlying mechanisms.

Evidence that the platelets contribute to cancer progression is substantial and continues to mount. Here, we show that platelets provide pro-metastatic signals to tumor cells by enhancing CD39 expression. The findings identify a new mechanism of tumor immune evasion mediated by platelet-tumor interaction, and highlight a novel cancer immunotherapy strategy focused on targeting the enzymatic activity of tumor CD39.