Characterization of endocannabinoid-mediated induction of myeloid-derived suppressor cells involving mast cells and MCP-1

Endocannabinoids induce suppressive myeloid derived suppressor cells, by mast cell mediated MCP-1 secretion.

Abstract

Endocannabinoids are lipid-signaling molecules found in the nervous system; however, their precise role in the periphery is unclear. In the current study, we observed that a single i.p. administration of AEA caused rapid induction of MDSCs. The MDSCs contained a mixture of granulocytic and monocytic subtypes and expressed Arg-1 and iNOS. The MDSCs suppressed T cell proliferation in vitro and used iNOS to mediate their effect. Moreover, adoptive transfer of MDSCs led to suppression of mBSA-induced DTH. Through the use of pharmacological inhibition, as well as genetic knockout mice, we found that the induction of MDSCs by AEA was CB1-dependent. The induction of MDSCs by AEA was reduced significantly in mast cell-deficient mice, while maintained in LPS-insensitive mice, showing that the induction of MDSCs by AEA was dependent, at least in part, on mast cells and independent of TLR4. Chemokine analysis of AEA- treated WT mice showed an early spike of MCP-1, which was decreased in Kit^W/W−sh mice, showing a role of mast cells in the secretion of MCP-1 in response to AEA. Also, use of antibodies against MCP-1 or mice deficient in MCP-1 confirmed the role played by MCP-1. Interestingly, MCP-1 played a significant role in the induction of monocytic but not granulocytic MDSCs. Our studies demonstrate for the first time that endocannaboinids activate CB1 on mast cells to induce MCP-1, which facilitates recruitment of monocytic MDSCs.

Introduction

MDSCs have been characterized as key immunoregulatory cells because of their potent ability to suppress the immune system [1–6]. MDSCs are believed to be bone marrow-derived precursors to monocytes and granulocytes that are CD11b⁺ and Gr-1⁺ and may include immature macrophages, granulocytes, DCs, and other myeloid cells [7, 8]. MDSCs exert many different mechanisms to suppress T cell proliferation, including cysteine sequestration [9], regulation of L-selectin expression on naive T cells [9], direct contact via NKp30 receptor [10],cross-talk with Tregs [11], and metabolic regulation by iNOS, Arg, and HO-1 [12, 13]. MDSCs have been identified to play a critical regulatory role in several clinical disorders, including cancer [11, 14–16], autoimmunity [2, 3, 6], and infection [4].

Cannabinoids are compounds derived from the Cannabis sativa plant and exert many effects on the body, ranging from deleterious to therapeutic [6, 17, 18]. Cannabinoids have been suggested to have potential as therapeutic agents in several different disease conditions [6, 17–20]. In addition, there is the endocannabinoid system, a set of natural cannabimimetic lipid signaling molecules regulating many processes in the CNS. The major members of this family of compounds include AEA and 2-AG. These compounds act by activating specific receptors called CBs, of which there are two: CB1 and CB2. The CBs are members of the GPCR family and mediate their effects through a series of G proteins and adaptors [21].

Immune cells have been shown to express CB1 and CB2, thereby suggesting that cannabinoids play an important role in the regulation of the immune system. Studies in our laboratory and elsewhere demonstrated that cannabinoids, such as THC, the major psychoactive principle in marijuana, mediate immunosuppression through multiple pathways, including induction of apoptosis in T cells and DCs, down-regulation of cytokine and chemokine production, switch from Th1 to Th2, up-regulation of Tregs, and induction of MDSCs [6, 22–26]. These studies also indicated that the endocannabinoid system may play a critical role in the regulation of immune functions. For example, administration of endocannabinoids or use of inhibitors of enzymes that break down the endocannabinoids led to immunosuppression and recovery from immune-mediated injury to organs, such as the liver [23]. Manipulation of endocannabinoids and/or use of exogenous cannabinoids in vivo can constitute a potent treatment modality against inflammatory disorders [23, 24, 27, 28]. Thus, additional studies are necessary to identify the specific molecular and cellular pathways that endocannabinoids use to modulate immune-cell differentiation and functions.

In the current investigation, we examined the mechanisms through which AEA, an endocannabinoid, suppressed T cell activation. Our data suggested that AEA activates mast cells to produce MCP-1, a chemokine that triggers massive mobilization of CD11b⁺Gr-1⁺-immunosuppressive MDSCs at the site of injection.

MATERIALS AND METHODS

Reagents

AEA, SR1, and SR2 were provided by the National Institute on Drug Abuse, NIH (Bethesda, MD, USA). The NOS2 inhibitor L-NMMA and the Arg-1 inhibitor L-NOHA, Con A, and URB597 were purchased from Sigma-Aldrich (St. Louis, MO, USA). FITC-labeled anti-CD11b (M1/70), anti-Gr-1 (RB6-8C5), and anti-Ly-6C (HK1.4); PE-labeled anti-Gr-1 (RB6-8C5), anti-CD11c (N418), anti-Ly-6G (1A8), anti-F4/80 (BM8), and anti-CD115 (CSF-1R); and Alexa-Fluor 647 anti-CD11b (M1/70) mAb were purchased from BioLegend (San Diego, CA, USA). MCP-1 neutralizing antibody (Clone 2H5) was purchased from BioLegend. Complete RPMI was made by addition of 1% penicilin/streptomycin, 10% FBS, 20 mM glutamine, 50 μM β-ME, and 10 mM HEPES.

Mice

Female B6 mice (6–12 weeks old) were purchased from the National Cancer Institute, NIH (Frederick, MD, USA). CB1^−/− mice were a kind gift from Dr. James Pickel (National Institute of Mental Health, NIH, Bethesda, MD, USA). B6.129P2-Cnr2tm1Dgen/J (CB2^−/−), B6.129S4-Ccl2^tm1Rol/J (MCP-1^−/−), B6.Cg-Kit^W−sh/HNihrJaeBsmJ (mast cell-deficient), and C57BL/10ScNJ (LPS-insensitive) mice were purchased from The Jackson Laboratory (Bar Harbor, ME, USA) and bred under appropriate conditions. All protocols were approved by the Institutional Animal Care and Use Committee, and mice were housed under specific pathogen-free conditions.

Administration of compounds and cell preparation

Mice were treated with endocannabinoids (AEA and 2-AG) at increasing doses i.p. After 12, 24, or 96 h, the peritoneal exudate cells were collected by performing lavage of the peritoneal cavity with PBS (5 ml×3). Cells were centrifuged and resuspended in FACS buffer (PBS containing 2% FBS).

Analysis of cells by flow cytometry

For analysis by flow cytometry, cells were first blocked with Fc block (purified α-CD16/CD32) and then stained with various fluorescently labeled mAb (10 μg/ml in FACS buffer). After washing, the stained cells were analyzed using the Beckman Coulter FC500 flow cytometer. Samples were gated on live cells using the forward-/side-scatter plot to exclude dead cells and debris. Unstained and isotype antibody-stained cells were used as negative controls to set gates.

MACS separation of MDSCs

MDSCs were sorted using a PE-positive selection kit, Microbead Magnetic Sorting Kit, from Miltenyi Biotec (Auburn, CA, USA), per the manufacturer's specification. In short, MDSCs were extracted from mice by lavage of the peritoneal cavity, as described previously [24]. The cells were pelleted in PBS (pH 7.2, 0.5% BSA, and 2 mM EDTA) at a concentration of 1 × 10⁸ cells/ml. The cells are stained with 1.5 μg/ml PE-conjugated Gr-1 mAb. A PE selection cocktail (100 μl/ml) was then added, and cells were run on a magentic sorting column. After washing, positive cells are isolated by plunging buffer through the magnetic column (2 ml×2).

Western blot

Proteins were isolated from 5 × 10⁶ MDSCs, induced by AEA, by three cycles of freeze/thaw in RIPA buffer containing protease inhibitors, PMSF, and sodium orthovanadate. Protein (15 μg) was loaded on to a polyacrylamide gel and run at 50 V for 2.5 h. Proteins were then transferred to nitrocellulose using the Trans-Blot SD Electrophoretic Transfer Cell at 15 V for 20 min. The blots are then blocked with 5% nonfat dry milk in TBS. The blots were then probed with antibodies to NOS2 (sc-7271), Arg-1 (SC-18354), and CCR2 (sc-31563) at a 1:200 dilution. The proper secondary antibody conjugated to HRP was used at a 1:1000 dilution. Substrate was added for 5 min, and blots were exposed for 30 s.

PCR

RNA was extracted from bone marrow progenitors, splenocytes, and MDSCs isolated from the peritoneum of AEA-exposed animals using the Qiagen RNAeasy kit, per the manufacturer's specifications. cDNA was made by performing RT-PCR using the Bio-Rad iScript cDNA synthesis kit. Samples were assayed for expression of CBs using the following primers: CB1 (forward 5′ TGGCCTAATCAAAGACTGAGGTTA 3′, reverse 5′ ATTGGGACTATCTTTGCGGTG 3′); CB2 (forward 5′ CATCTTCCACGTCTTCCACG 3′, reverse 5′ CAGCCCAGTAGGTAGTCGTT 3′); and GPR55 (forward 5′ GCATCTGGACCATTGCTACC 3′, reverse 5′ TGGTACTCAGCGAGTGTCAA 3′) with annealing at 58°C.

T cell proliferation assay

B6 T cells were isolated using nylon wool separation, as described [29]. The T cells were then allowed to flow through, and the column was washed with complete RPMI (3 ml, twice). The T cells were then plated at 5.0 × 10⁵ cells/well in 96-well plates. Purified MDSCs were treated with mitomycin C (10 μg/ml) for 4 h and then plated with these T cells at a MDSC:responder ratio of 1:2, 1:5, 1:10, and 1:20. T cells were stimulated with 2.5 μg/ml Con A for 48 h. At 16 h prior to harvest, 1 μCi was added to each well. Cells were harvested and read using a MicroBeta liquid scintillation counter (Perkin Elmer, Waltham, MA, USA).

Induction of mBSA-induced type IV hypersensivitiy and MDSC adoptive transfer

Mice were sensitized to mBSA by giving a s.c. injection of 150 μg mBSA emulsified in CFA. Two injections were given to each mouse, one in each hind flank. After 6 days, mice were rechallenged in the right footpad with 200 μg mBSA in PBS and in the left footpad with PBS only. Percent increase in swelling was determined by the following equation: thickness of mBSA footpad − thickness of PBS footpad/thickness of PBS footpad × 100%. For adoptive transfer, 12 h before rechallenge, PBS, 5 × 10⁶ splenocytes, or 5 × 10⁶ MDSCs were injected into mBSA-sensitized i.v., as described before [24].

Histology

Footpads from PBS-treated, splenocyte-transferred, and MDSC-transferred mice were isolated for histological examination. The feet were cut distal to the ankle and decalcified/fixed in Cal-Rite (Thermo Fisher Scientific, Waltham, MA, USA). The tissues were embedded in paraffin, cut into 6-μm sections, and stained with H&E for generic architecture and infiltration.

ELISA

Peritoneal lavages were taken from mice exposed to AEA at the following time-points after exposure: 0, 3, 6, 9, and 12 h. Briefly, mice were killed at the appropriate time-point, and 1 ml cold PBS was used to lavage the peritoneal cavity of mice. The lavages were analyzed at no dilution and assayed for the expression of MCP-1 using the BioLegend ELISA Max kit, according to the manufacturer's specifications.

MC/9 cells and induction of MCP-1

MC/9 cells, a murine mast cell line, were cultured at 1 × 10⁵ cells/well in a 96-well plate. These cells were treated with increasing doses of AEA (5, 10, and 20 μM) for 24 h. Supernatants were isolated and assayed for MCP-1 by ELISA as before. To test the role of CB1, we pretreated MC/9 cells with SR1 for 1 h and proceeded with the culture as described before.

MCP-1 neutralization

Mice were injected with 200 μg α-MCP-1 (2H5) antibodies or isotype control antibodies, 30 min before treatment with AEA. Mice were then sacrificed, and peritoneal exudates were analyzed for the induction of MDSCs at 12 h after AEA treatment, as described before.

RESULTS

Endocannabinoids induce significant accumulation of Cd11b⁺ Gr-1⁺ cells

B6 WT mice were injected with vehicle or one of either AEA or 2-AG i.p. The exudate cells were then isolated from the peritoneal cavity at 16 h postinjection, and cell numbers were analyzed. We saw a dramatic increase in total cell number when AEA-treated animals were compared with vehicle treated animals (Fig. 1A). When these cells were analyzed using flow cytometry, forward-/side-scatter analysis revealed that cells in AEA-treated animals were larger and more granular (Fig. 1A). We next examined the phenotype of these cells using mAb for Cd11b and Gr-1. We found that administration of 2AG or AEA triggered a large number of Cd11b/Gr-1 double-positive cells in the peritoneum that were not seen in vehicle-treated mice (Fig. 1B and C). AEA or 2AG caused a dose-dependent increase in proportion of CD11b/Gr-1 double-positive cells, as well as a dose-dependent increase in the absolute cell numbers (Fig. 1B and C). These data suggested that endocannabinoids induce MDSCs, and we pursued additional studies using AEA. To rule out the possibility that any LPS contamination of endocannabinoids may have triggered the MDSCs, we investigated the effect of AEA on BL10SCn/J mice, which are LPS-insensitive. We found that there was no effect on MDSC accumulation induced by AEA in LPS-insensitive mice, which excluded the possibility of LPS contamination causing MDSC induction (Fig. 1D). Furthermore, the level of LPS in the AEA solution was measured using the Limulus amebocyte lysate assay, and the results were negative (data not shown).

Figure 1. AEA induces CD11b⁺/Gr-1⁺ cells in vivo.

(A) B6 (WT) mice were treated with vehicle (Veh) or 20 mg/kg body weight AEA i.p. Peritoneal cells were harvested by lavage at 16 h postinjections and counted using trypan blue exclusion. *P < 0.05 versus vehicle. Data are represented as mean ± sem (n=4). Cells were analyzed by flow cytometry for granularity (side-scatter) and size [forward-scatter (FS)]. Endocannabinoids, 2-AG (B) and AEA (C), were injected into WT mice at increasing doses (5–20 mg/kg). Total peritoneal cells were isolated as before and stained using fluorescently labeled mAb for CD11b and Gr-1. Absolute cell numbers were calculated by applying percentages obtained by flow cytometry to the total cell count. Dots plots are representative based on three independent experiments with n = 4. Square gates represent double-positive cells with frequency as indicated. Absolute cell numbers of double-positive cells were calculated by applying the proportion obtained by flow cytometry, as shown in dot plots, to the total viable cell counts. (D) To rule out the role of any LPS contamination, AEA was injected in WT and C57BL/10ScNJ (BL10) LPS-insensitive mice. Cells were isolated and analyzed as before. Data represented as mean ± sem. Statistics are one-way ANOVA with Tukey post hoc test (B–D) or t- test (A). *P < 0.05 was determined to be significant.

Changes in cellular profiles in the peritoneum of AEA-treated mice

To determine the effect AEA had on the peritoneal cellular profiles, we assayed AEA or vehicle-treated mice for the presence of cells that typically reside in the peritoneum, including macrophages (CD11b/F480 double-positive), DCs (CD11c), and granulocytes (Gr-1) at 16 h postinjection. We noted significant changes in the cellular profile of mice treated with AEA, whereas vehicle-treated mice were showing a normal cell profile. A vehicle-treated profile (Fig. 2B) included a high frequency of CD11b hi/F480 hi macrophages, as well as Cd11b lo/F480-negative monocytes, very few granulocytes, as shown by the lack of Gr-1-expressing cells, as well as very few CD11c-expressing DCs. When we treated mice with AEA, we saw a decrease in the percentage of CD11b hi/F480 hi macrophages. This could be explained by a dilution effect caused by an increase in CD11b/Gr-1 double-positive MDSCs (Fig. 2B) coming into the peritoneal cavity in response to treatment with AEA. Also, these MDSCs expressed an intermediate level of CD11b compared with macrophages in the vehicle-treated peritoneum. When we assayed for the level of macrophage marker (F4/80), we saw an intermediate level of expression of F4/80 in AEA-treated mice compared with vehicle controls (Fig. 2B). These cells in AEA-treated mice were indicative of immature monocytes but also expressed Gr-1 (Fig. 2B). We also found that the induction of CD11c was more on AEA-exposed cells (Fig. 2B). These changes in cellular profiles were indicative of significant induction of monocyte/granulocyte precursors in the peritoneum of mice treated with AEA. These changes also translated into significant absolute cell number differences (Fig. 2A).

Figure 2. Characterization of AEA-induced changes in mouse peritoneal cell populations.

Mice were treated with vehicle or 20 mg/kg AEA. Peritoneal cells were assayed for expression of CD11b⁺/Gr-1⁺ cells, F4/80 macrophages, and CD11c⁺ cells at 16 h after injection (A). Statistical significance was determined by t-test, and *P < 0.05 was determined as significant. Dot plots are representative of two independent experiments; n = 3 are shown (B). (C) Cells were isolated as before, and double-positive CD11b⁺/Gr-1⁺ cells were assayed by flow cytometry for the expression of CD124. Total CD11b⁺ cells were also analyzed for the expression of LY6C and Ly6G. Plots are representative of three independent experiments with n = 3. (D) Western blot analysis of purified MDSCs from AEA-treated animals and assayed for expression of Arg-1, NOS2, and CCR2.

Phenotypic analysis of MDSCs induced by AEA

To determine further the phenotype of these MDSCs induced by AEA, we assayed for the expression of several reported markers for MDSCs by gating on CD11b/Gr-1 double-positive cells and found that 56.6% of these MDSCs expressed IL-4R (CD124; Fig. 2C). MDSCs have been classified in literature as one of two subtypes: granulocytic or monocytic, based on their expression of the Ly-6C or Ly-6G isotypes of the Gr-1 molecule. Granulocytic MDSCs express the Ly-6C and -6G isotypes, whereas monocytic MDSCs express only the Ly-6C isotype. By gating on CD11b, we examined only those cells that express CD11b and a Gr-1 isotype. We found that AEA induced both granulocytic and monocytic MDSCs; however, there was shift toward induction of granulocytic MDSCs (CD11b⁺/Ly-6G⁺/Ly-6C⁺) compared with monocytic 54.6 and 26.0%, respectively (Fig. 2C). This 2:1 ratio of granulocytic:monocytic MDSCs stayed true through all of our studies.

Several mechanisms for the immunosuppression mediated by MDSCs have been reported, including Arg-1 and NOS2. We therefore assayed for the expression of Arg-1 and NOS2, as well as CCR2 (MCP-1R), in MDSCs purified from mice treated with AEA by Western blot (Fig. 2D). We found that the AEA-induced MDSCs expressed NOS2 and Arg-1. There was also expression of CCR2, a common receptor found on macrophages and monocytes, that is important for chemotaxis in response to MCP-1.

AEA-induced MDSCs effectively limit inflammation in vivo and in vitro

Next, we investigated the ability of MDSCs to inhibit T cell proliferation, a hallmark of their function. To that end, we first isolated MDSCs from AEA-treated mice by magnetic bead separation, added them to cultures of T cells activated with Con A, and measured T cell proliferation. The data suggested that T cells showed significant proliferation in response to Con A stimulation. However, addition of increasing numbers of MDSCs to coculture with the T cells led to significant inhibition of T cell proliferation (Fig. 3A). Next, to determine if NOS2 or Arg-1 was necessary for suppressing T cell proliferation by MDSCs, L-NMMA and L-NOHA were used, respectively. Increasing concentrations of L-NOHA did not affect the ability of MDSCs to suppress T cell proliferation (Fig. 3B). However, with increasing doses of L-NMMA, the ability for MDSCs to suppress T cell proliferation was ablated (Fig. 3C). Together, these data suggested that MDSCs induced by AEA suppress T cell proliferation using NOS2.

Figure 3. MDSCs induced by AEA mitigate inflammation.

(A) Purified MDSCs from AEA-administered animals were first treated with 10 μm Mitomycin C and then cocultured with T cells in increasing ratios. T cell proliferation was measured by thymidine incorporation. Data are representative of two separate experiments. The importance of NOS2 (B) and Arg-1 (C) was assayed using pharmacological inhibitors to the respective enzymes, L-NOHA and L-NMMA, at increasing doses, as indicated. (D) Purified MDSCs from AEA-administered animals were adoptively transferred into mBSA-sensitized mice, 12 h before rechallenge with mBSA. Footpad swelling was measured using calipers and calculated using this formula: thickness of mBSA-challenged footpad − thickness of PBS-challenged footpad/thickness of PBS-challenged footpad × 100%. (E and F) The footpad of each mouse was excised and decalcified using Cal-Rite. Samples were embedded individually, cut into 5-μm sections, and stained with H&E to assess histopathological damage. Pictures were taken at 4× (E) and 10× (F). Spl, Splenocytes. Black arrows were added to point out the swelling and edema. (A–D) Significance was determined by one-way ANOVA with Tukey post hoc analysis, with *P < 0.05 considered significant.

MDSCs have been shown to mitigate in vivo inflammation upon adoptive transfer [6]. We isolated AEA-induced MDSCs or normal splenocytes as a control and adoptively transferred them through the i.v. route to test whether the MDSCs would suppress the DTH response against mBSA. The group of mice adoptively transferred with MDSCs exhibited decreased levels of the DTH response when compared with the PBS- or splenocyte-treated group (Fig. 3D). This was confirmed by histopathological analysis, where we noted decreased edema and cellular infiltration in MDSC-transferred mice (Fig. 3E and F).

Involvement of CBs in AEA-induced MDSC accumulation

Next, we investigated whether AEA induced MDSCs through activation of CBs. To that end, we used CB1^−/− and CB2^−/− mice. Administration of AEA into CB1^−/− mice resulted in a decrease in proportion (Fig. 4A), as well as absolute cell numbers of MDSCs (Fig. 4B). Interestingly, AEA was able to induce similar levels of MDSCs in CB2^−/− mice, as in WT mice, thereby suggesting that CB2 did not play a role in AEA-mediated MDSC induction (Fig. 4C and D). When WT mice were given a pharmacological inhibitor of CB2, SR2, there was no significant effect on MDSC induction in WT or CB1^−/− mice in response to AEA. Together, these data indicated that AEA was inducing MDSCs through activation of CB1 but not CB2 in vivo.

Figure 4. CB1 is important for MDSC induction by AEA.

(A) B6 WT or CB1^−/− mice were treated with vehicle or SR2, 1 h before treatment with vehicle or 20 mg/kg body weight AEA. Twelve hours post-treatment with AEA, peritoneal exudates were harvested. Cells were stained with fluorescently labeled antibodies to CD11b and Gr-1 and assayed for MDSC induction. Dot plots are representative of three separate experiments, with n = 4. (B) Absolute cell numbers were obtained by applying proportions to total viable cells, as measured by trypan blue exclusion. (C) CB2 is not involved in the induction of MDSCs by AEA. Vehicle or AEA was given to WT or CB2^−/− mice and analyzed as before. Dot plots are representative of three separate experiments, with n = 4. (D) Absolute cell numbers were determined as before. (E) Total RNA was isolated from bone marrow cells (lane 1), purified MDSCs (lane 2), or splenocytes (lane 3), from AEA-treated animals. Each cell type was assayed for expression of CB1, CB2, or GPR55. Gels are representative of three separate reactions, with n = 2. (F) Splenocytes (lane 1) or peritoneal exudate cells (lane 2) were analyzed for expression of CB1, CB2, or GPR55 by the same procedure as in E. (B and D) Significance was determined by two-way ANOVA and Sidak's multiple comparison analysis with P < 0.05 considered as significant. *P < 0.05 compared with vehicle; **P < 0.05 compared with WT; and #P < 0.01 compared with WT.

One possibility was that AEA was acting directly on MDSCs and inducing their proliferation or regulating their functions. To test this, we assayed for the expression of CBs on MDSCs, splenocytes, and bone marrow cells (Fig. 4E). In the spleen and bone marrow, there was high expression of CB1, CB2, and GPR55 orphan receptor, as expected. However, when we investigated expression profiles of MDSCs purified from the peritoneum of AEA-treated animals, we noted that purified MDSCs did not express CB1 and expressed very low levels of CB2. This indicated that AEA may not act directly on MDSCs and that their induction was resulting from the activation of some other responder cells. In search of this responder, we first isolated peritoneal cells from naive mice and noted that such cells exhibited high levels CB1 and CB2 (Fig. 4F). This suggested a role for naive peritoneal cells in the potential induction of MDSCs in response to AEA.

Role of MCP-1 in MDSC induction

We investigated the possible role of chemokines in the accumulation of MDSCs in the peritoneal cavity by performing ELISAs on peritoneal lavages. Pilot studies showed the presence of MCP-1, and therefore, we performed a time-course experiment to elucidate the temporal role of MCP-1 in MDSC induction. There were no detectable levels of MCP-1 before treatment with AEA. However, as early as 1 h postinjection, there was a significant release of MCP-1 in the peritoneum of mice treated with AEA. This MCP-1 release peaked at 3 h postinjection. After 3 h, the levels of MCP-1 declined sharply until ∼12 h postinjection, at which point, the levels of MCP-1 were undetectable (Fig. 5A). We took cells concurrently from the same mice to examine the induction of MDSCs. We found that the MDSCs begin accumulating at 6 h and exponentially increased at 9 and 12 h (Fig. 5B). This can be explained by the fact that it may take some time for the MDSCs to migrate to the site of AEA injection, where they continue accumulate even after MCP-1 production ceases. To confirm the role of MCP-1, we investigated the effect of MCP-1 neutralization on the induction of MDSCs in AEA-treated animals. We found a significant reduction in the number of MDSCs found in the peritoneum of mice treated with the neutralizing antibody 2H5 when compared with isotype control (Fig. 5C).

Figure 5. Role of MCP-1 released by mast cells in response to AEA on MDSC induction.

(A) Time course of MCP-1 release in response to AEA was determined by treating WT mice with 20 mg/kg body weight of AEA and killing at 1, 3, 6, and 12 h post injection. Peritoneal cytokines were obtained by lavage of the peritoneum with 1 ml cold PBS. Cells were excluded from analysis by centrifugation. MCP-1 levels were determined by ELISA. Figure is representative of two separate experiments, with n = 3. (B) Peritoneal lavages were taken from the same mice and measured for MDSCs at 16 h as before. (C) Importance of MCP-1 in the induction of MDSCs was assessed by using a neutralizing antibody to MCP-1. Anti-MCP-1 antibody (Clone 2H5) or isotype control was injected 30 min before injecting 20 mg/kg body weight AEA. Peritoneal exudate cells were harvested in cold PBS, 12 h after injection with AEA, stained with CD11b and Gr-1 fluorescently labeled mAb, and assayed by flow cytometry for MDSCs. Absolute cell numbers were obtained by applying percentage to total cell number. (D) WT or cKit^−sh were injected with vehicle or 20 mg/kg body weight AEA. At 3 h, peritoneal lavage fluid was obtained with 1 ml cold PBS, and MCP-1 levels were assayed by ELISA. Figure is representative of two separate experiments, with n = 3. (E) MC/9 cells were cultured in the presence of AEA at increasing doses for 24 h. Supernatants were measured for MCP-1 by ELISA. (F) MC/9 cells were precultured with increasing doses of SR1 and measured for MCP-1. *P < 0.05 and **P < 0.01 versus isotype IgG.

We then examined the source of MCP-1, which is secreted by many different cell types [26]. We first examined mast cells as a putative source for MCP-1 in AEA-treated animals. We found that when AEA was administered to mast cell-deficient (cKit^w/w−sh) mice, there was a significant decrease in the induction of MCP-1 when compared with WT animals treated with AEA (Fig. 5D). To confirm further that mast cells secrete MCP-1 directly, we used MC/9 cells, a mast cell line that is IL-3-dependent. When cultured in the presence of AEA, MCP-1 was detected in a dose-dependent response (Fig. 5E). To determine the role of CB in MC/9 cells, we used a pharmacological inhibitor for CB1. We found that the AEA-mediated secretion of MCP-1 was CB1-dependent (Fig. 5F). These data together demonstrated that MCP-1, at least in part, plays a role in AEA-mediated MDSC induction and that mast cells secrete MCP-1 in a CB1-mediated fashion in response to AEA.

Deficiency of mast cells or MCP-1 results in a loss of monocytic MDSCs

After we found the link between mast cells and MCP-1, we used genetic knockout mice for these two factors to determine their role in MDSC induction. We treated cKit^w/w−sh or MCP-1^−/− mice with AEA and looked for induction of MDSCs. We found that both of these mice had reduced numbers of AEA-mediated MDSC accumulation (Fig. 6A). We found this interesting and further characterized the MDSCs induced by AEA in these knockout mice. We found that the proportion of monocytic MDSCs (Ly6C⁺, Ly6G⁻) decreased significantly in MCP-1^−/− and cKit^−sh mice (Fig. 6B). Interestingly, the proportion of granulocytic MDSCs (Ly6C⁺, Ly6G⁺) was not altered significantly. We found a significant decrease in the numbers of monocytic MDSCs, whereas granulocytic MDSCs were statistically unaffected. Together, these data suggested that mast cells and MCP-1 may play a role in the induction of monocytic but not granulocytic MDSCs.

Figure 6. Role of mast cells and MCP-1 in the induction of monocytic MDSC accumulation.

(A) B6 WT, cKit^−sh, and MCP-1^−/− mice were treated with 20 mg/kg body weight AEA. Peritoneal cells were harvested in cold PBS, 16 h after treatment. Total viable cell counts were assayed by trypan blue exclusion. (B) Mice were treated as before, and peritoneal exudate cells were stained for CD11b, Ly6G, and Ly6C to determine differences in MDSC subtype. Absolute cell counts were obtained by applying percentage to absolute cell counts. Data are representative of two separate experiments, with n = 3. Significance was determined by one-way ANOVA with Tukey Post-Hoc test. *P < 0.05 compared with WT.

DISCUSSION

In the current study, we demonstrated for the first time that a peripheral exposure to endocannabinoids, such as AEA and 2-AG, resulted in significant induction of MDSCs. The MDSCs constituted a mixture of granulocytic and monocytic subtypes and expressed Arg-1 and iNOS, while primarily using iNOS to mediate their effect. We demonstrated that these MDSCs were functional in vivo using adoptive transfer, which led to suppression of mBSA-induced DTH. MDSC induction by AEA, at least in part, was CB1- and mast cell-dependent. We found that MCP-1 produced by mast cells played a key role in MDSC induction inasmuch as Kit^W/W−sh mice showed decreased levels of MCP-1 in response to AEA, and antibodies against MCP-1 or mice deficient in MCP-1 displayed significantly ablated MDSC induction. Together, our studies demonstrate for the first time that endocannaboinids activate CB1 on mast cells to induce MCP-1, which mobilizes MDSCs at the site.

Whereas exogenous cannabinoids, such as THC, have been shown to suppress the immune response [6, 17, 18, 24], very little literature exists delineating the role of endocannabinoids in the regulation of the immune response [30, 31]. In the same vein, whereas activation of CB2, which is expressed predominantly in the periphery, specifically on immune cells, has been studied extensively and implicated in the regulation of the immune response [19, 32–34], literature is limited with respect to studies on the role of CB1 in the periphery. We show here that activation of CB1 by endocannabinoids results in a significant effect on immune cells. Previous work from our laboratory suggested that activation of CB1 and CB2 led to massive influx of MDSCs into the peritoneal cavity [24]. Also, we noted that THC-mediated MDSC induction was dependent on G-CSF. In contrast, we found in the current study that AEA-induced MDSCs were CB1- but not CB2-dependent, involving the activation of mast cells and release of MCP-1. In our preliminary studies, 2-AG, the other species of endocannabinoid, was shown not to have a clear, receptor-mediated mechanism. Further studies are necessary to parse out the mechanism of 2-AG-mediated MDSC induction. The precise reason as to why THC and AEA mediate differences is not clear. It should be noted that endocannabinoids are derivatives of arachidonic acid; thus, they have a different chemical structure when compared with phytocannabinoids of the Cannabis plant. Moreover, AEA has greater affinity to the CB1 than CB2. Thus, AEA acts as a full agonist at the CB1 and partial agonist at the CB2 [35]. Also, there is clear evidence for the role of endocannabinoids and CB1 in regulating the immune system [35–43]. Such findings may explain the differences in the mechanism of MDSC induction by AEA and THC.

We show a very potent induction of MDSCs in response to AEA in naive mice. This is the first indication of a direct action of endocannabinoids in the induction of a specific suppressor-cell population. Whereas it should be noted that 2-AG also induced MDSCs at a high level, we chose to use AEA in our further experiments, as its effect was mediated by CBs, whereas preliminary data suggested that 2-AG may use alternative mechanisms, which need to be explored further. The CD11b⁺Gr-1⁺ cells induced by AEA had major hallmarks of typical MDSCs, including expression of IL-4R (CD124) [44], low expression of F4/80, and importantly, potent, functional capacity to suppress T cell proliferation. Even though AEA-induced MDSCs expressed Arg-1 and NOS2, we noted that their suppressive activity was mediated by NOS2 signaling, as a specific NOS2 inhibitor was able to block the suppressive functions of MDSCs in a dose-dependent manner. This may be a unique feature of AEA-induced MDSCs. Whereas granulocytic and monocytic subtypes of MDSCs are known to express Arg-1, NOS2 has been shown to be unique to monocytic MDSCs [7]. This is supported further by the fact that only monocytic MDSCs were decreased significantly in MCP-1^−/− and mast cell-defective mice, and granulocytic MDSC numbers were unaltered in response to AEA.

Another interesting observation was the absence of CB expression on purified MDSCs compared with splenocytes and bone marrow precursors. Expression of the CB on specific cell types has been linked to the apoptosis induced by cannabinoids [17, 19, 22, 23]. CB2 activation has also been shown to be directly chemotactic in response to 2-AG [45]. Showing that MDSCs were CB-negative suggests that the effect of AEA was indirect. This also showed that MDSCs may be able to withstand high concentrations of AEA, such as at the primary tumor site, and continue to perform their function, considering that there is a well-defined link between cancer and immunosuppression.

MDSCs have been implicated as one of the major suppressors of the anti-tumor immune response, responsible for immune evasion in many cancers, including breast [46], lung [44], and renal cell carcinoma [47]. However, the precise mechanism for MDSC induction is not well understood. Some studies have shown that mast cells can function to assist in MDSC migration to tumor-bearing sites [48]. We show here that mast cells are important for chemokine-mediated migration of MDSCs to the site of endocannabinoid administration. With the use of cKit^−sh transgenic mice that lack functional tissue mast cells, we noted that these mice produced decreased levels of MCP-1, as well as showed decreased induction of MDSCs. MDSC induction by AEA was not wiped out completely in mast cell-deficient mice, which is not surprising, as MCP-1 is also produced by other cells, such as monocytes and endothelial cells [49, 50] This is a unique observation and the first study demonstrating a direct activation effect of endocannabinoids on mast cells to produce chemokines, such as MCP-1, leading to induction regulatory cells with the ability to suppress the T cell response.

Mast cells have been known to produce MCP-1 in response to various stimuli [51], although its effect on MDSC induction has not been characterized. Previous studies did identify mast cells as producers of MCP-1 in response to various stimuli [51]. Also, other groups have shown that endocannabinoids do have an effect on the ability for mast cells to mature [52], as well as release factors important for regeneration [53]. Also, endocannabinoid-like compounds, such as palmitoylethanolamide, have been shown to act on canine mast cells, causing them to degranulate at a low rate [54]. There have been reports that AEA does not affect the release of histamine in rat peritoneal mast cells [55], which would suggest that the release of MCP-1 by mast cells in response to AEA, as shown in this report, may be a secretory response, independent from regular degranulation. In naive mice, CD11b⁺Gr-1⁺ cells have been shown to be present in small numbers in peripheral tissues, such as spleen, and up to 18–50% in bone marrow, depending on the mouse strain. In our previous studies, we noted that THC induces mobilization of MDSCs from the bone marrow [24], and we propose that AEA also promotes the migration of MDSCs from the bone marrow to the periphery by inducing the production of MCP-1 by the mast cells.

It is interesting to note that mice deficient in MCP-1 showed a significant decrease in the total numbers of MDSCs induced; however, the response was not eliminated completely. Whereas this suggested the involvement of additional cytokines/chemokines, our attempts to detect G-CSF, which was shown to play a critical role in MDSC induction by THC [24], did not yield positive results (data not shown). However, when we studied the induction of granulocytic and monocytic MDSCs, we noted that MCP-1 and mast cell-deficient mice showed a decrease in the induction of only monocytic but not granulocytic MDSCs. This suggested that mast cell and MCP-1 involvement may relate primarily to monocytic MDSCs.

Chemokines have been shown to be used by tumors to escape immune surveillance [56]. It has been suggested that MCP-1 is used to mediate T cell exclusion from the site of tumor growth [57]. The data presented in this report offer a new perspective on the role of MCP-1 in immune evasion and cancer progression, likely with a direct role in the induction of MDSCs in the tumor microenvironment. The endocannabinoid system has been shown to be dysregulated in many different cancers [36], and the reasons are unknown. The novel connection among endocannabinoids, MCP-1, and MDSCs identified in this study opens the door for a new conversation into cancer immunology. Clearly, additional studies are necessary to determine whether endocannabinoids produced by the tumor cells trigger MDSCs. The potential role of endocannabinoids as regulators of immune surveillance in cancer offers new targets for emerging cancer therapies.

It should also be noted that endocannabinoids may act as a double-edged sword with respect to their ability to regulate immune response and cause immunosuppression. Whereas on one hand, their ability to trigger MDSCs and cause immunosuppression may lead to immune evasion by cancer cells, on the other hand, such a mechanism may play a role in maintaining immune-system homeostasis, especially to down-regulate exacerbated inflammation and autoimmunity. In the context of autoimmune disease, endocannabinoids could have evolved as a natural regulatory system, inducing MDSCs to dampen the unchecked inflammation. Overall, our studies suggest an important role for a natural cannabinoid system in immune regulation.

AUTHORSHIP

A.R.J. designed, performed, and analyzed all experiments and wrote the manuscript. V.L.H. provided significant assistance in experimental design, editing, and scientific reasoning. P.N. and M.N. provided the funds and resources to pursue these studies. Also, they provided mentorship in experimental design and analysis, as well as significant assistance in writing and editing the manuscript.

DISCLOSURES

The authors declare no conflicts of interest.