Department of Pathology, Microbiology and Immunology, University of South Carolina, School of Medicine, Columbia, SC, USA
Carolina Distinguished Professor, Department of Pathology, Microbiology and Immunology and Associate Dean for Basic Sciences, University of South Carolina, School of Medicine, 6439 Garners Ferry Road, Columbia, SC 29209, USA Fax: +1-803-733-3335
Cannabinoid receptor activation by agents such as Δ9-tetrahydrocannabinol (THC) is known to trigger immune suppression. Here, we show that administration of THC in mice leads to rapid and massive expansion of CD11b+Gr-1+ myeloid-derived suppressor cells (MDSC) expressing functional arginase and exhibiting potent immunosuppressive properties both in vitro and in vivo. The induction of MDSC by THC was associated with a significant increase in granulocyte CSF. Moreover, administration of anti-granulocyte CSF Ab inhibited the induction of MDSC by THC. THC was able to induce MDSC in TLR4 mutant C3H and C57BL10/ScN mice and hence acted independently of TLR4. Accumulation of MDSC in the periphery with a corresponding decrease in the proportion of CD11b+Gr-1+ cells in the bone marrow, as well as in vivo BrdU labeling and cell-cycle analysis, showed that THC induced mobilization of these cells from bone marrow and their expansion in the periphery. Use of selective antagonists SR141716A and SR144528 against cannabinoid receptors 1 and 2, respectively, as well as receptor-deficient mice showed that induction of MDSC was mediated through activation of both cannabinoid receptors 1 and 2. These studies demonstrate that cannabinoid receptor signaling may play a crucial role in immune regulation via the induction of MDSC.
Recently, a suppressor cell population of myeloid lineage capable of reducing anti-tumor as well as inflammatory immune responses has been described 1–5. These cells express CD11b and Gr-1 and have been named myeloid suppressor cells or myeloid-derived suppressor cells (MDSC) 2, 3, 6. CD11b+Gr-1+ MDSC are a heterogeneous cell population including immature macrophages, granulocytes, DC and other myeloid cells, and possess potent immunosuppressive properties 2, 6. They have been shown to suppress T-cell proliferation 7, inhibit the cytotoxicity of CD8 T cells or NK cells both in vitro and in vivo3, 6, 8, 9, downregulate L-selectin expression on CD4/CD8 T cells 10 and induce antigen-specific CD8 T-cell tolerance in tumor-bearing hosts 11. L-Arginine metabolism is the major mechanism by which MDSC cause T-cell suppression 1, 2, 12. CD11b+Gr-1+ suppressive cells have also been identified during several inflammatory conditions 13–16.
Endogenous and exogenous cannabinoids signal through two major G-protein coupled cannabinoid receptors, CB1 and CB2 17. Delta-9-tetrahydrocannabinol (THC) is a natural cannabinoid compound from the plant Cannabis sativa18. THC and other cannabinoids have been extensively studied with respect to their immunomodulatory and anti-inflammatory properties 17. Cannabinoids exert their immunomodulatory effects by various mechanisms. We and others have previously shown that the immunosuppressive property of THC can be attributed in part to its ability to induce apoptosis in T lymphocytes, DC and macrophages 19–22. THC has also been shown to trigger Treg 23 as well as induce anti-inflammatory cytokine production 18, 23, 24.
Although both CB1 and CB2 are expressed on cells of the immune system, the precise nature of these receptors and the role of endogenous and exogenous cannabinoids in the regulation of the immune functions are not clear. Also, although many studies have demonstrated that cannabinoids exhibit anti-inflammatory properties, the precise mechanisms are unclear. In this study, we demonstrate for the first time that activation of CB1 and CB2 through administration of cannabinoids such as THC, into mice, triggers massive induction of arginase 1 expressing CD11b+Gr-1+ MDSC, with immunosuppressive properties.
THC administration triggers massive induction of CD11b+Gr-1+ cells
C57BL/6 (B6) WT mice were injected with vehicle or THC i.p. The exudate cells in the peritoneal cavity were harvested after 16 h and analyzed. To our surprise, the administration of THC induced massive local accumulation of cells in the peritoneal cavity when compared with vehicle control (Fig. 1A). Flow cytometric analysis using forward and side scatter revealed that majority of cells induced by THC were granular and larger in size (Fig. 1B). We further phenotyped these cells using mAb to cell-surface markers, CD3, Gr-1, CD11b and F4/80 (Fig. 1C). Cells from vehicle-injected mice showed phenotypic characteristics expected from a normal peritoneum. CD11bhigh cells in the peritoneum typically represent mature macrophages, which also express high F4/80. Interestingly, majority of cells (>90%) from THC-injected mice were positive for CD11b and Gr-1 markers. Close to one third proportion of the population also showed low or intermediate expression of F4/80. We performed a dose–response study by injecting (i.p.) WT mice with different doses of THC and analyzed the peritoneal exudate cells for the coexpression of CD11b and Gr-1 by flow cytometry. THC triggered a dose-dependent increase in the percentage (Fig. 1D) as well as absolute numbers (Fig. 1E) of CD11b+Gr-1+ cells in the peritoneum. In control mice, the CD11bhigh (Gr-1−) cell population represents mature macrophages. This is typical of naïve peritoneum where mature macrophages form a significant proportion, whereas, CD11b+Gr-1+ cells induced by THC expressed intermediate levels of CD11b. The CD11bhigh (Gr-1−) macrophages disappear in the dot plots with increasing doses of THC (Fig. 1D) likely because of the dilution effect resulting from several fold induction of CD11b+Gr-1+ cells or due to THC-induced killing of mature macrophages as reported previously 20. The robust induction of CD11b+Gr-1+ cells seen here in response to THC also shows that unlike lymphocytes 19–21 or macrophages 20, these cells are resistant to THC-induced killing.
THC-induced CD11b+Gr-1+ cells are MDSC and mediate immunosuppression in vitro and in vivo
Next, we investigated if THC-induced CD11b+Gr-1+ cells were MDSC by studying their ability to mediate inhibition of T-cell proliferation, which is one of the most important characteristics of MDSC. To this end, we performed a series of in vitro and in vivo studies. For in vitro assays, we determined the suppressive activity of purified peritoneal CD11b+Gr-1+ cells induced by THC on the proliferation of T cells stimulated with polyclonal as well as antigen-specific stimuli. T-cell proliferation was assessed by [3H]-thymidine uptake. THC-induced CD11b+Gr-1+ cells caused a dose-dependent decrease in proliferation of T cells stimulated with ConA with almost complete inhibition at a 1:2 ratio of CD11b+Gr-1+ cells to T cells (Fig. 2A). THC-induced CD11b+Gr-1+ cells also significantly decreased proliferation of OT-II ova-transgenic T cells stimulated using agonist ova peptide (Fig. 2B). As a control, we purified CD11b+Gr-1+ cells from peritoneal exudates induced by thioglycollate broth (TGB), which at early time points (4 h) triggers primarily neutrophils, and compared their ability to suppress T-cell proliferation. Side-by-side comparison of suppressive activity showed that THC-induced CD11b+Gr-1+ cells were highly immunosuppressive compared with the CD11b+Gr-1+ neutrophils (Fig. 2C and D).
Next, we used a well-established model of T-cell-mediated liver inflammation induced by ConA to test the immunosuppressive properties of THC-induced CD11b+Gr-1+ cells in vivo. We adoptively transferred CD11b+Gr-1+ cells purified from the peritoneum of THC-injected mice into naïve mice before inducing hepatitis using ConA. The transferred CD11b+Gr-1+ cells significantly suppressed liver inflammation as indicated by decreased liver enzyme, alanine transaminase (ALT) levels, a commonly used marker for liver inflammation (Fig. 2E). These data together suggested that the CD11b+Gr-1+ cells induced by THC exhibited the functional characteristics of MDSC.
Kinetics of THC-induced MDSC accumulation
The kinetics of MDSC induction by THC in peritoneum and spleen was determined by injecting mice with a single dose of THC, and analyzing recruitment of MDSC at different time points (0–24 h). A significant increase in frequency (Fig. 3A) and absolute number of MDSC (Fig. 3B) in peritoneum and spleen was detected as early as 8 h after THC injection, which peaked around 16 h and started to decline by 24 h.
Histological analysis of cellular infiltrates in peritoneum
We studied the morphology of THC-induced MDSC by Wright Giemsa staining. Cytospin preparations of peritoneal exudate cells from mice injected with vehicle or THC were stained and analyzed by light microscopy (Fig. 3C). Vehicle-injected mice showed the characteristics mainly of lymphocytes and macrophages. On the contrary, cells from THC-injected mice consisted of PMN-like cells with characteristic circular (or donut shaped) nucleus as well as monocyte-like cells. We also looked at cellular infiltrates in parietal peritoneum sections. H&E stained cross-sections of peritoneal tissue showed significant infiltration of cells in the THC group, whereas vehicle-injected mice showed no infiltration of cells (Fig. 3D).
THC induces MDSC in TLR4-deficient mice
Previous studies have shown that TLR ligands can trigger induction MDSC 25–27. The pure THC preparation used in this study had undetectable levels of endotoxins. However, to further rule out the effect of endotoxins, as well as to investigate the possibility if THC acted via TLR4 in this case, we tested the same batch of THC in TLR4-deficient C57BL10/ScN mice and TLR4-mutant C3H mice. Both C57BL10/ScN and C3H mice showed significant increase in frequency as well as absolute number of CD11b+Gr-1+ MDSC (Fig. 4A and B) thereby suggesting that THC-induced recruitment of MDSC was independent of TLR4.
Cytokine/chemokine response to THC
THC is known to suppress inflammatory cytokines induced by infectious agents or during inflammatory conditions in vivo23, 24. However, we sought to determine if THC can induce any cytokine/chemokines in vivo resulting in the recruitment of MDSC. To this end, we injected WT mice with a single dose of THC or vehicle and collected sera at 6, 12 and 24 h. We analyzed sera for IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12 (p40), IL-12 (p70), IL-13, IL-17, IFN-γ, TNF-α, MIP-1α, MIP-1β, RANTES, MCP-1, GM-CSF, granulocyte CSF (G-CSF), keratinocyte chemoattractant (KC) and eotaxin by Bioplex assay. THC had little or no effect on majority of these cytokines and chemokines at 6, 12 and 24 h as compared with basal levels induced by vehicle (data not shown), except for KC (CXCL1) and G-CSF. THC induced significant levels of chemokine, KC at 12 h and very high and persistent levels of G-CSF (∼1500 pg/mL) at 12 and 24 h (Fig. 5A). It is noteworthy that THC induced G-CSF levels significantly at an early time point, 6 h, and to a very high level at 12–24 h. It is possible that the initial upregulation of G-CSF by THC leads to migration and activation of MDSC which in turn secrete more G-CSF causing increased mobilization and expansion of these cells.
GM-CSF has been shown to play an important role in induction MDSC associated with tumors 28. However, we could not detect increase in GM-CSF in sera in the current model. To further investigate the role of GM-CSF and G-CSF in the induction of MDSC by THC, we determined the levels of these cytokines directly in the peritoneal exudates of vehicle and THC-injected mice. Although there was a robust increase in the levels of G-CSF in the peritoneum of THC-injected mice as compared with vehicle controls, the GM-CSF levels were low and there was no significant difference between the two groups (Fig. 5B).
Role of G-CSF in THC-induced accumulation of MDSC
To further confirm the role of G-CSF in the induction of MDSC by THC in vivo, we pretreated mice with anti-mouse G-CSF Ab before injecting THC. We observed that as low as 10 μg/mouse of anti-G-CSF was able to significantly block the accumulation of MDSC in the peritoneum in response to THC (Fig. 5C), indicating that G-CSF plays a crucial role in this process. These data suggest that THC may trigger migration and expansion of MDSC primarily by inducing G-CSF.
THC-induced MDSC express functional arginase
L-Arginine metabolism is an important pathway used by MDSC to blunt anti-tumor immunity 1 or reduce inflammatory responses 13–16. Expression of arginase, the enzyme which metabolizes and depletes essential amino acid L-arginine, is the hallmark characteristic of MDSC. In order to test if THC-induced MDSC express functional arginase, we injected mice with vehicle or THC and determined arginase enzyme activity by spectrophotometric assay based on the principle of conversion of substrate L-arginine to L-ornithine. THC-induced cells from the peritoneum accounted for several fold higher levels of arginase activity when compared with vehicle controls thereby suggesting that THC-induced MDSC expressed functionally active arginase (Fig. 6A).
THC induces migration and expansion of MDSC from bone marrow
In naïve 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 6, 29. As THC induces mobilization of MDSC massively and rapidly, we speculated that the source of these cells could be bone marrow. To investigate this, we injected WT mice with THC and determined the percentage of CD11b+Gr-1+ cells at 0 and 16 h in bone marrow and peritoneum (Fig. 6B). We observed a significant decrease in the percentage of CD11b+Gr-1+ cells in the bone marrow 16 h after THC administration with a corresponding increase in periphery (peritoneum). This indicated that CD11b+Gr-1+ cells were migrating from bone marrow in response to THC. Further, we wanted to know if the MDSC were also dividing in the periphery. Mice were given vehicle or THC, 2 h after injecting with BrdU. Peritoneal cells harvested after 24 h were triple stained for CD11b, Gr-1 and intranuclear BrdU incorporation (Fig. 6C). Close to 37% of the MDSC in the peritoneum of the THC-injected mice were BrdU+. As cells incorporating BrdU in the bone marrow and then migrating to peritoneum in response THC would have contributed to this percentage, we determined active proliferation in the peritoneum by cell-cycle analysis using PI staining. We found that a significant proportion (22.5%) of THC-induced MDSC and 23.3% of all cells induced by THC in the peritoneum were in S-phase, indicative of active cycling when compared with 11.9% of peritoneal cells found in vehicle-treated mice. These data suggested that MDSC accumulating in the periphery in response to THC were undergoing active proliferation.
Analysis of MDSC subsets induced by THC
Although MDSC are defined by their coexpression of CD11b and Gr-1, these cells have been known to contain heterogeneous mixture of myeloid cells with suppressive function. However, recently two major subsets of MDSC have been identified based on the expression of CD11b, Ly6G and Ly6C antigens. Granulocytic subsets (PMN-like MDSC subset) express both Ly6G and Ly6C along with CD11b (CD11b+Ly6-GhighLy6-Clow/int), whereas monocyte like MDSC subsets express only Ly6-C and CD11b (CD11b+Ly6GnegLy6Chigh) 25, 30, 31. We sought to identify these subsets among MDSC induced by THC. For this, we used Ab specific to Ly6G (clone: 1A8) and Ly6C (clone: HK1.4) along with anti-CD11b. THC administration resulted in significant dose-dependent increase in the frequency (Fig. 7A) as well as absolute numbers of CD11b+Ly6G+Ly6C+(int) granulocytic (PMN) and CD11b+Ly6GnegLy6C+(high) monocytic MDSC in the peritoneum (Fig. 7B and C), suggesting that THC significantly induces both major MDSC subsets.
Role of cannabinoid receptors
To investigate whether induction of MDSC resulted from activation of cannabinoid receptors, we first used CB1 and CB2 antagonists, SR 141716A (SR1) and SR144528 (SR2), respectively. WT mice were injected with SR1 or SR2 or both, 2 h prior to administration of vehicle or THC. Mice pretreated with either SR1 or SR2 showed significantly less frequency (Fig. 8A) and absolute number (Fig. 8B) of MDSC when compared with mice injected with THC alone. This indicated that THC acts through both CB1 and CB2 receptors to induce the mobilization of MDSC. We did not observe any significant synergistic blocking with both the antagonists injected together.
To further establish the role of cannabinoid receptors, we used CB1 and CB2 receptor KO mice. We injected WT, CB1KO (CB1−/−) or CB2KO (CB2−/− or Cnr2−/−) mice with vehicle or THC. In CB1 and CB2 KO mice, the percentages (Fig. 8C) as well as the absolute numbers (Fig. 8D) of MDSC in the peritoneum were significantly reduced after THC challenge when compared with WT mice, confirming the involvement of both the receptors. Furthermore, when we injected CB1 and CB2 KO mice with CB2 or CB1 antagonist (SR2 or SR1), respectively, and then administered THC, we observed that blocking CB2 receptors in CB1 KO mice and CB1 receptors in CB2KO mice had a synergistic effect and almost completely inhibited the induction of absolute numbers of MDSC in response to THC (Fig. 8C and D). Cumulatively, these data suggested that THC induces MDSC by signaling primarily through both CB1 and CB2 receptors.
Immunosuppressive and anti-inflammatory properties of cannabinoids, including THC, have been well established 17. THC is known to cause immune suppression by various mechanisms including induction of apoptosis in effector T cells, suppressing the inflammatory cytokine production under certain activating conditions, as well as arresting cell proliferation 18. However, the effect of cannabinoids on important suppressor cell population, namely, MDSC is not known. MDSC were initially identified in large numbers in tumor-bearing mice and were found to effectively suppress anti-tumor immune responses 4–6. There is strong evidence showing that these cells are immunosuppressive and inactivate T cells 1, 2. Several recent studies have shown their role in modulation of inflammatory diseases 14–16. Induction of MDSC in cancer has also been attributed to be a consequence of chronic inflammation 32, 33.
In this study, we injected THC i.p. in naïve mice in the absence of any activating conditions. Initially, we were perplexed to see the dramatic accumulation of cellular infiltrates in the peritoneal cavity in response to THC, which were found to be large and granular cells by preliminary assessment. Given the potent immunosuppressive properties of THC, we speculated that these may not be inflammatory infiltrates, but some type of suppressor cells. Further phenotypic and functional characterization revealed that THC-induced cells coexpressed CD11b and Gr-1 markers. We further observed that THC-induced CD11b+Gr-1+ cells expressed functional arginase and caused significant inhibition in the proliferation of T cells stimulated with polyclonal (ConA) as well as antigen-specific (ova) stimuli in vitro, and conA-induced T-cell-mediated liver inflammation in vivo, all hallmark characteristics of MDSC. It should be noted that unlike THC-induced CD11b+Gr-1+ cells, purified CD11b+Gr-1+ neutrophils isolated 4 h after thioglycollate administration failed to exhibit similar immunosuppressive properties.
We and others have shown that THC suppresses inflammatory cytokines elevated during an inflammatory response 23, 34. However, the effect of THC on various cytokines in naïve mice has not been thoroughly investigated. Among an array of 23 different cytokines and chemokines tested, we made a striking observation that THC caused a dramatic induction in the levels of G-CSF, and to some extent KC, in sera of naïve mice. Very high induction of G-CSF was also observed directly in the peritoneum. We noted that blocking of G-CSF by anti-G-CSF Ab was able to significantly decrease THC-induced accumulation of MDSC in vivo. G-CSF is a crucial cytokine required for the differentiation of granulocytes such as neutrophils from bone marrow precursors, and which also regulates their migration, proliferation as well as function 35. Anti-inflammatory properties of G-CSF have also been known 36. A very recent study using a spontaneous metastatic memory carcinoma model showed that the number of MDSC in the spleen directly correlated with G-CSF transcript levels 37. Differential accumulation of tumor infiltrating monocytic MDSC over granulocytic MDSC in tumor-bearing hosts mediated by chemokines, particularly G-CSF has also been demonstrated 38.
We further showed that the effect of THC was independent of TLR4 and mediated directly through activation of both CB1 and CB2. This observation was supported by the findings that CB1 or CB2 select antagonists could block MDSC induction by THC partially but significantly and was confirmed using CB1- and CB2-deficient mice. It has been shown that bone marrow is the source of large numbers of CD11b+Gr-1+ precursors 6, 29. In this study, THC seemed to cause the migration of CD11b+Gr-1+ MDSC from bone marrow. THC also induced the proliferation of migrated MDSC in the periphery as evidenced by significant proportion of actively proliferating MDSC in the peritoneum of THC injected mice. Various cell types including macrophages, endothelial cells and fibroblasts are known to express CB1 and CB2 39. THC may activate CB1/CB2 on these cells to induce the secretion of mediators such as G-CSF and KC (CXCL1). Inasmuch as, G-CSF and KC have been shown to induce the development, migration and expansion of granulocytes, particularly neutrophils, it is highly likely that these molecules may play a crucial role in MDSC induction. We have noted that unlike the i.p. administration of THC, i.v. injection does not result in robust induction of MDSC in spleen, lungs and peritoneum although a significant induction in the liver was seen (data not shown). This may result from the fact that i.p. injection of THC may activate resident cells in the peritoneal cavity such as fibroblasts or immature macrophages that are known to produce G-CSF 40. This is also supported by our observation that high levels of G-CSF were found in the peritoneal cavity washings after THC injection by the i.p. route. Although G-CSF is primarily known to promote neutrophil differentiation from bone marrow precursors, it is possible that G-CSF by itself or other factors such as KC induced by THC or a combination of G-CSF and THC may promote mobilization of CD11b+Gr-1+ MDSC and arrest their differentiation into mature neutrophils or macrophages. Currently, we are further testing this possibility.
We observed that a single dose of THC, as low as 5–10 mg/kg, was able to induce significant levels of MDSC in mice, which is clinically relevant. Based on the body surface area normalization method 41, this translates to human equivalent dose of 15–30 mg/m2 (0.40–0.81 mg/kg). In cancer patients, the dose of THC (dronabinol or Marinol®) recommended is usually 2.5–20 mg/m2/day (0.068–0.54 mg/kg), as an anti-emetic treatment during chemotherapy. Chan et al. demonstrated that rats injected with 50 mg/kg body weight of THC had a serum concentration of 10 μM THC within 10 h of administration 42. Azorlosa et al. showed that levels as high as 1 μM could be attained in the plasma of humans after recreational use of marijuana 43, and in a separate report, it was shown that THC can get preferentially concentrated 15- to 20-fold in some tissues 44. Such levels of THC may lead to significant suppression of the immune response, leading to increased susceptibility to opportunistic infections and cancer. In fact, it was shown that treatment of mice with doses of 8 mg/kg significantly suppressed the responses of mice to infection with Legionella pneumophila34. In light of these, our results suggest that THC, at a pharmacological or recreational dose, may induce MDSC and cause immunosuppression.
Studies from our laboratory suggest that cannabinoids are a double-edged sword. On one hand, cannabinoids can suppress malignancies of the immune system by inducing apoptosis of tumors that express cannabinoid receptors 45. On the other hand, because they suppress anti-tumor immune response, they promote growth and metastasis of cannabinoid receptor-deficient tumors such as breast cancer that are resistant to cannabinoid-induced apoptosis 46. Epidemiological studies have shown that marijuana smokers are more susceptible to certain types of cancer 47, 48. It is possible that induction of MDSC by cannabinoids may make an individual more susceptible to certain types of cancer, particularly those which do not express cannabinoid receptors. We previously showed that THC promotes the growth and metastasis of 4T1 mammary tumors that lack cannabinoid receptors by suppressing anti-tumor immunity 46. Recent studies have suggested that 4T1 tumor growth is associated with MDSC induction 49. Thus, it is tempting to speculate that THC further enhances induction of MDSC which may be responsible for the accelerated growth and metastasis of 4T1 tumors and studies are in progress to address this in detail.
This study unravels a new mechanism by which cannabinoids trigger immune suppression. These findings are important in several ways: (i) THC is currently used clinically to ameliorate nausea and vomiting as well as stimulate appetite during chemotherapy in cancer patients, and HIV/AIDS-infected patients and to lower intraocular eye pressure to treat glaucoma 50. Thus, it is important to know if such a treatment would impact the immune system and pose health hazards. As MDSC are known to promote tumor growth, our observations indicate that THC treatment in cancer patients may potentiate adverse effects. (ii) As CB2 is expressed only on immune cells, CB2 select agonists may constitute better anti-inflammatory agents with less nonspecific toxicity than the currently available anti-inflammatory agents. Thus, the demonstration that CB2 activation leads to MDSC which in turn may help suppress inflammation may provide novel avenues to develop anti-inflammatory agents. (iii) Our studies shed light on the potential role of endocannabinoids that activate cannabinoid receptors on MDSC induction and consequent regulation of inflammation.
Materials and methods
THC, SR141716A (CB1 antagonist) and SR144528 (CB2 antagonist) were provided by National Institute on Drug Abuse, National Institutes of Health (Bethesda, MD). The mAb fluorescein isothiocyanate-conjugated anti-CD11b, anti-Ly6C (clone: HK1.4), phycoerythrin-conjugated anti-CD3, anti-Gr-1 (clone: RB6-8C5), anti-F4/80, anti-Ly6G, Alexa-647-labeled CD11b, unconjugated anti-BrdU and mouse Fc-block (anti-CD16/CD32) were purchased from eBioscience. Phycoerythrin-conjugated anti-Ly6G mAb (clone: 1A8) was from BD Biosciences. Purified anti-mouse-G-CSF Ab was from R&D Systems. Anti-PE microbeads, magnetic sorting columns and equipment were from Miltenyi Biotech. PI and RNAase were from Invitrogen. Concanavalin A, L-arginine, L-ornithine standard, Ninhydrin reagent, BrdU, Triton X-100, RBC lysis buffer and all other chemicals and reagents were from Sigma-Aldrich.
Female B6 (WT), C3H and OT-II mice (8–10 wk) were purchased from National Cancer Institute, National Institutes of Health (Frederick, MD). CB2 KO (CB2−/− or Cnr2−/−) mice on B6 background and C57BL/10ScN mice were obtained from The Jackson labs (Bar Harbor, ME, USA). CB1 KO (CB1−/−) mice on B6 background were a gift from Dr. James Pickel, National Institute of Mental Health Transgenic Core Facility. Mice were housed under specific pathogen-free conditions in the Animal Resource Facility of University of South Carolina and all experiments were preapproved by the Institutional Animal Care and Use Committee.
Administration of compounds and preparation of cells
Mice were injected with THC i.p. Exudate cells in the peritoneal cavity were harvested by performing lavage with sterile, ice-cold PBS (5 mL×3). Cells were spun down and reconstituted in FACS buffer (PBS/2% FBS) for flow cytometry staining or in complete RPMI medium for in vitro experiments. Splenocytes were obtained after preparing the single-cell suspension followed by RBC lysis. Bone marrow cells were obtained by flushing tibia with ice-cold PBS followed by RBC lysis. For blocking experiments, SR1, SR2 compounds or anti-G-CSF Ab were injected (i.p.) at the indicated doses 2 h prior to injecting THC.
Immunofluorescence staining and flow cytometry
For FACS analysis, cells were blocked using mouse Fc-block (anti-CD16/CD32) and stained for various cell surface markers using fluorescently labeled mAb (10 μg/mL, in PBS containing 2% FBS). After washing, stained cells were analyzed in a flow cytometer (FC500; Beckman Coulter, Fullerton, CA). Only live cells were counted by setting gates on forward and side scatters to exclude debris and dead cells. Isotype Ab-treated cells served as staining controls. Intracellular staining was performed after fixing and permeabilizing the surface stained cells using Fix-perm reagents (eBioscience). Data obtained were analyzed in Cytomics CXP software (Beckman Coulter).
Harvested parietal peritoneum was fixed in 4% buffered paraformaldehyde and embedded in paraffin. Sections (5 μm) were stained with H&E for histological analysis under light microscope.
Cells were collected by cytospin onto glass slides and dried completely for 30 min. The slides were stained with Wright–Giemsa stain (Fisher Sci) according to the manufacturer's instructions and analyzed by light microscopy.
T-cell proliferation assay
THC-induced MDSC were sorted by magnetic sorting using PE-conjugated Gr-1 Ab and anti-PE microbeads (Miltenyi Biotech). The sorted cells were >95% positive for CD11b and Gr-1 coexpression. For some experiments, CD11b+Gr-1+ MDSC were sorted to >96% purity using FACS Aria (Becton and Dickinson) after labeling with fluorescently conjugated Ab. Purified MDSC were irradiated and cultured at different ratios with purified syngenic T cells (2×105) stimulated with Con A (4 μg/mL) in a 96-well round bottom plates, in complete RPMI 1640 medium (Invitrogen) supplemented with 10% FBS, 10 mM HEPES, 1 mM penicillin–streptomycin and 50 μM β-mercaptoethanol. For antigen-specific stimulation, purified T cells (2×105) derived from OT-II mice were stimulated with agonist ova peptide (OVA323–339, 1 μg/mL) in the presence of irradiated splenocytes (5×105) as APC, and were cocultured with different ratios of purified, irradiated MDSC. T-cell proliferation was determined after 72 h culture by pulsing with [3H]thymidine (1 μCi/well) during the final 12 h of culture. Cultures were harvested using a cell harvester and thymidine incorporation was measured in a β counter (Perkin Elmer).
WT mice were injected with 1 mL of 3% TGB and peritoneal exudates were harvested after 4 h. This mixture of cells predominantly contains CD11b+(high)Gr-1+(high) neutrophils (approximately, 65–70%). The CD11b+Gr-1+ cells were enriched by sorting on FACS Aria to >96% purity and used as neutrophil controls.
Induction of hepatitis and adoptive transfer with MDSC
ConA was dissolved in pyrogen-free PBS at a concentration of 2.5 mg/mL and injected in to B6 mice intravenously at a dose of 12.5 mg/kg body weight to induce hepatitis. MDSC harvested from the peritoneum of THC-injected mice were purified by magnetic sorting using PE-conjugated Gr-1 Ab and anti-PE microbeads (Miltenyi Biotech). For adoptive transfer, mice received 5 million purified MDSC i.v., 12 h before injecting ConA. Sixteen hours after ConA administration, blood was collected by retro-orbital bleeding. Hepatitis was assessed by measuring liver enzyme alanine aminotransferase (ALT) in sera.
Sera from individual mice were obtained 6, 12 and 24 h after injecting with vehicle or THC and stored below −20°C until analysis. IL-1α, IL-1β, IL-2, IL-3, IL-4, IL-5, IL-6, IL-9, IL-10, IL-12 (p40), IL-12 (p70), IL-13, IL-17, IFN-γ, TNF-α, MIP-1α, MIP-1β, RANTES, MCP-1, GM-CSF, G-CSF, KC and eotaxin concentrations in serum samples were determined using Bio-Plex chemiluminescence assay system (Bio-Rad).
Peritoneal exudates were collected by lavage from vehicle or THC injected (i.p.) mice using 1 mL of sterile PBS. Cells were spun down to obtain supernatants. Sandwich ELISA for GM-CSF (BioLegend) and G-CSF (Peprotetch) were performed, and the concentrations of cytokines in the samples were determined from standard curves after correcting for blank.
Arginase activity assay
Arginase activity in cell lysates was measured from the conversion of L-arginine to L-ornithine as described elsewhere 13. Lysates of 2×106 peritoneal exudate cells from vehicle or THC-injected mice were prepared using cell lysis buffer containing protease inhibitor cocktail, and stored below −80°C until use. Arginase was activated by the addition of 10 mmol/L MnCl2 (25 μL) to cell lysate (25 μL) and incubating at 55°C for 20 min. Carbonate buffer (150 μL, 100 mmol/L, pH 10) was then added along with 100 mmol/L L-arginine (50 μL) to initiate the reaction and incubated at 37°C. Arg1 activity was stopped after 10 min by adding glacial acetic acid (750 μL). Ninhydrin reagent, 250 μL (2.5 gm of Ninhydrin+40 mL of 6 M phosphoric acid+60 mL of glacial acetic acid) was added and tubes were covered and boiled at 90–100°C for 1 h. Standards were created by using known amounts of L-ornithine and processed as above. Tubes were cooled and colored end point was measured using a spectrophotometer at 515 nm.
BrdU labeling in vivo
To study MDSC proliferation in vivo, mice were injected with BrdU (2 mg/mouse) 2 h before injecting THC. Mice were maintained with drinking water containing 0.8 mg/mL BrdU After 24 h. Cells from the peritoneal cavity were harvested by gavage and stained for surface CD11b and Gr-1. After fixing and permeabilization, incorporated BrdU was stained by intranuclear staining protocol using DNAase I and FITC-conjugated anti-BrdU Ab (BD Biosciences). Triple-stained cells were analyzed by flow cytometry on FC 500 (Beckman Coulter).
PI stating for cell-cycle analysis was done by standard procedure. Briefly, cells were stained using anti-Gr-1 FITC to mark MDSC (typically >95% Gr-1+ cells were Gr1+CD11b+ among vehicle and THC-induced cells in the peritoneum). Cells were fixed and resuspended in PI reagent containing 0.1% Triton X-100, 0.2 mg/mL RNAase and 20 μg/mL PI in PBS and incubated for 37°C for 15 min. Labeled cells were acquired on FC 500 flow cytometer (Beckman Coulter) using DNA analysis settings to gate out doublets and aggregates. DNA-PI signal on gated Gr-1+ cells representing MDSC was analyzed using Multicycle software (Pheonix Systems).
In most experiments, we used groups of three to six mice, unless otherwise mentioned, that were analyzed individually and data depicted as mean±SEM. The statistical comparisons between different treatment groups were done using Student's t-test and differences with p-value of <0.05 were considered to be significant. All experiments were repeated at least two times to test the reproducibility of results.
The authors thank Shweta Hegde for technical assistance and Dr. James Pickel, National Institute of Mental Health, NIH, for providing them with CB1 knockout mice. The research was funded by National Institutes of Health, USA grant awards, R01-DA016545, R01-ES09098, R01-AI053703, R01-AI058300, R01-HL058641, and P01-AT003961 to M. N. and P. S. N.
Conflict of interest: The authors declare no financial or commercial conflict of interest.