Regulation of arginase I activity and expression by both PD-1 and CTLA-4 on the myeloid-derived suppressor cells

Abstract

An elevated number of Gr-1⁺CD11b⁺ myeloid-derived suppression cells (MDSCs) has been described in mice and human bearing tumor and associated with immune suppression. Arginase I production by MDSCs in the tumor environment may be a central mechanism for immunosuppression and tumor evasion. In this study and before, we found that Gr-1⁺CD11b⁺ MDSCs from ascites and spleen of mice bearing ovarian 18D carcinoma express a high level of PD-1, CTLA-4, B7-H1 and CD80 while other co-stimulatory molecules, namely CD40, B7-DC and CD86 are not detected. Further studies showed that PD-1 and CTLA-4 on the Gr-1⁺CD11b⁺ MDSCs regulated the activity and expression of arginase I. The blockage and silencing of PD-1, CTLA-4 or both PD-1 and CTLA4 molecules could significantly reduce arginase I activity and expression induced with tumor-associated factor. Similar results were also observed while their ligands B7-H1 and/or CD80 were blocked or silenced. Furthermore, CD80 deficiency also decreased the arginase I expression and activity. Antibody blockade or silencing of PD-1, CTLA-4 or both reduced the suppressive potential of PD-1+CTLA-4+MDSCs. Blockade of PD-1, CTLA-4 or both also slowed tumor growth and improved the survival rate of tumor-bearing mice. Thus, there may exist a coinhibitory and costimulatory molecules-based immuno-regulating wet among MDSCs.

Introduction

Tumor progression is associated with accumulation of myeloid-derived suppressor cells (MDSCs), which include immature macrophages, granulocytes, dendritic cells (DCs) and myeloid cells [1–4]. These MDSCs share common features such as myeloid origin, macrophage-like morphology, phenotype of surface receptors, and ability to suppress T cells after culture in vitro [1–5]. The Gr-1⁺CD11b⁺ surface markers are most often associated with a population of immature myeloid cells in the spleen of tumor-bearing mice [5–7]. Such immature myeloid cells are present in the bone marrow and spleen of healthy mice, and differentiate into mature myeloid cells under normal conditions [7]. These cells are also associated with immune suppression during viral infection, transplantation, UV irradiation or cyclophosphamide treatment [8]. Purified MDSCs inhibit both CD4+ and CD8+ T cell responses in vitro and downregulate T cell functions in vivo [9]. Deletion of MDSCs in tumor-bearing mice using an antibody to the granulocyte marker Gr1/Ly6G results in enhanced immune responses to tumors [10].

Several mechanisms have been described for the induction of tolerance. These include release of soluble mediators like NO, lack of costimulation, Fas/FasL interaction, induction of several types of regulatory T cells including Tr1, Th3, NKT cells and CD4⁺CD25⁺ T regulatory (T_reg) cells [11, 12] and generation of MDSCs [9, 13]. MSDCs can express many agents including arginase I, reactive nitrogen and oxygen species, cytokines, proteases and enzymes important in amino acid metabolism to suppress the immune responses. Recently, the importance of arginase I in MDSCs has been addressed in the inhibition and possibly deletion of tumor-specific CTL [5]. CD3ζ is lost in T cells exposed to arginase I [14, 15]. Thus, arginase I production by MDSCs in the tumor environment may play a critical role in tumor evasion and may represent a target for new therapies [15].

The activity and expression of arginase I could be regulated via multiple factors, including tumor-derived factors such as TGFβ [16] and cytokines such as IL-4 and IL-13 [15]. Our previous studies showed that MDSCs from spleen and ascites of tumor-bearing mice expressed an array of coinhibitory and costimulatory molecules such as CD80, B7-H1, PD-1 and CTLA-4. These molecules play a critical role in regulating T and B cell immune responses. While a number of studies show that costimulatory or coinhibitory molecules between antigen-presenting cells or MDSCs and T, B cells are required for the regulation of T cell activity, the function of these molecules among tumor-associated Gr-1⁺CD11b⁺ MDSCs cells is not clear. Herein, we found that these costimulatory or coinhibitory molecules on the surface of MDSCs are involved in regulating the expression and activity of arginase I, implying an immunoregulation wet among the MDSCs in inducing immunosuppression and tumor evasion.

Materials and methods

Mice

Six to eight-week-old male mice, C57BL/6 (NCI) and CD80−/− (B6, 129S4-CD80^TM1SHR, The Jackson Laboratory), were maintained in a pathogen-free animal facility for at least 1 week before use. Experiments were performed in accordance with Institutional Guidelines. For in vivo blockade studies, mice were injected intraperitoneally with 250 μg anti-CTLA-4 mAb (UC10-4F10-11, PharMingen), anti-PD-1 (RMP1-14, eBioscience, San Diego, CA) and normal isotypic IgG as control on days −6, −4, −2 and +1 as that was previously reported [20].

MOSEC ovarian carcinoma model

A syngeneic mouse model for epithelial ovarian cancer [17] based on a spontaneously transformed mouse ovarian surface epithelial cell line 1D8 has been previously described [18]. Ovarian 1D8 carcinoma cells did not express the costimulatory molecules CD40, CD80, CD86, CD22.2, CD72, CTLA-4 or other markers CD5, B220, CD11b and CD11c (data not shown). However, 1D8 tumor cells expressed CD44 and very low levels of costimulatory molecules B7-H1, PD-1 and MHC class II (not shown). Multiple tumors and ascites formed within 60 days after the i.p. injection of 1 × 10⁷ 18D tumor cells in syngeneic C57BL/6 mice. Tissue samples were collected 2 months after the i.p. 1D8 tumor injection. To monitor tumor growth rate, 1 × 10⁷ 18D tumor cells were injected s.c. in the right leg.

Flow cytometric analysis and isolation of PD-1+CTLA-4+Gr-1⁺CD11b⁺ myeloid cells from ascites of mice with ovarian 18D carcinoma

FITC or PE-conjugated anti-mouse CD11c (N418), anti-mouse PD-1 (J43), anti-mouse CD86 (GL1), anti-mouse CD80 (16-10A1), anti-mouse CD40 (3/23) and PE-labeled anti-CD4 (L3T4), anti-mouse B7-H1 (MIH1), anti-mouse CD8α (53–6.6), anti-mouse CD45 R/B220 (RA36B2), anti-mouse CD11b (M1/70), as well as purified anti-mouse I-A^b (25–9-17), anti-mouse Ly-6G (RB6-8C5), anti-CD25 (7D4), CTLA-4 (UC10-4F10-11) and allophycocyanin (APC)-labeled anti-mouse Ly-6G (RB6-8C5) were purchased from PharMingen (San Diego, CA). Single or multiple staining was performed using different monoclonal antibodies. For each analysis, isotype matched control mAb was used as a negative control. Cells were then washed twice, resuspended in PBS containing 1% paraformaldehyde and 1% FCS and kept at 4°C prior to flow cytometric analysis (FAScan, Becton Dickson).

For isolation of Gr-1⁺CD11b⁺ myeloid cells, ascites or splenic cells from mice with ovarian carcinoma or healthy mice were directly stained with FITC-labeled anti-Gr-1 and PE-labeled anti-CD11b antibodies. After washing, cells were gated for myeloid cell characteristics, namely high forward and side scatter and bright staining for CD11b. These selected cells were then sorted using FAScan based on the staining for CD11b and for Gr-1 or absence of staining for both. The purity of Gr-1⁺CD11b⁺ cells was >95%. The sorted Gr1⁺CD11b⁺ cells from the ascites of mice with ovarian carcinoma express high level of PD-1 and CTLA-4; whereas expression of PD-1 and CTLA-4 on the sorted Gr1⁺CD11b⁺ cells from splenic cells of disease-free mice was not detected. The sorted Gr1⁺CD11b⁺ cells from ascites of mice with 1D8 ovarian carcinoma were used in the experiments.

siRNA design, synthesis and transfection

siRNA sequences were designed according to protocol on website http://www.ambion.com. Two target sequences, as 1# and 2#, for each gene were designed. For CTLA-4, target sequence1# AAGTGATTCAGTTTGTGGCAG, siRNA sense 5′GUGAUUCAGUUUGUGGCAGtt and siRNA antisense 5′CUGCCACAAACUGAAUCACtt; target sequence 2# AAGGTGGAACTCATGTACCCA, siRNA sense 5′GGUGGAACUCAUGUACCCAtt and siRNA antisense 5′UGGGUACAUGAGUUCCACCtt; for PD-1, target sequence 1# AACAGAGAGAATCCTGGAGAC, siRNA sense 5′CAGAGAGAAUCCUGGAGACtt and siRNA antisense5′GUCUCCAGGAUUCUCUCUGtt; target sequence 2# AAGGCATGGTCATTGGTATCA, siRNA sense 5′GGCAUGGUCAUUGGUAUCAtt and siRNA antisense 5′UGAUACCAAUGACCAUGCCtt; for B7-H1, target sequence 1# AACAGGCATGACTTCCACATG; siRNA sense 5′CAGGCAUGACUUCCACAUGtt and siRNA antisense 5′CAUGUGGAAGUCAUGCCUGtt; target sequence 2# AATCACGCTGAAAGTCAATGC, siRNA sense 5′UCACGCUGAAAGUCAAUGCtt and siRNA antisense 5′GCAUUGACUUUCAGCGUGAtt; for CD80, target sequence 1# AACTCTCCTCATGAAGATGAG, siRNA sense CUCUCCUCAUGAAGAUGAGtt and siRNA antisense CUCAUCUUCAUGAGGAGAGtt; target sequence 2# AAGACCGAATCTACTGGCAAA, siRNA sense GAUGAGUCUGAAGACCGAAtt and siRNA antisense UUCGGUCUUCAGACUCAUCtt. Non-silencing control siRNA was an irrelevant siRNA with random nucleotides (5′-ACUATCUAAGUUACTACCCCtt). Sequences were synthesized and annealed by the Johns Hopkins University Genetics Core Facility. Searches of the mouse genome database (BLAST) were carried out to ensure that the sequence would not target other gene transcripts. Cells were transfected with the siRNAs using nucleofector device (Amaxa, Germany) according to the manufacturer’s protocol.

Arginase activity assay

Arginase activity was measured as described elsewhere [19]. Briefly, 100 μl of cell lysate was mixed with 100 μl of 10 mM MnCl₂ and the enzyme was activated by heating for 10 min at 56°C. l-arginine hydrolysis was conducted by incubating 25 μl of the lysate with 25 μl of 0.5 M l-arginine, pH 9.7, at 37°C for 60 min. The reaction was stopped with 400 μl H₂SO₄ (96%)/H₃PO₄ (85%)/H₂O (1/3/7 V/V/V). The urea concentration was measured at 540 nm after the addition of 40 μl of α-isonitrosopropiophenone in 100% ethanol followed by heating at 95°C for 30 min. A calibration curve was prepared with increasing amounts of urea between 1.5 and 30 μg. Data (derived from three replica wells) were given as milli units of arginase/10⁶ cells, where 1U of arginase was defined as the amount of enzyme that catalyzes the formation of 1 μg of Urea/min. To determine effect of PD-1 and CTLA-4 on the arginase I activity induced by tumor supernatant (25% V/V) as conditional medium or 10 ng/ml mouse recombinant IL-13 (mrIL-13, eBioscience SanDiego CA) as positive control. Blocking experiments were used to determine the role of costimulatory molecules in the arginase I activity in which various neutralizing antibodies were added into the culture to 10 μg/ml, including anti-PD-1 (RMP1-14, eBioscience, San Diego, CA), anti-mouse CTLA-4 (UC10-4F10-11, Hamster IgG1) or both and isotypic control antibody. To further decide the effect of PD-1 and CTLA-4 on the MDSCs, PD-1 or/and CTLA-4 targeted siRNA, transfected MDSCs were cultured in conditioned medium containing (25% V/V) tumor supernatants or control 10 ng/ml IL-13, the cells were lysed and arginase activity was determined after 24 h.

RT-PCR analysis

Total cellular RNA was prepared using TRIzol reagent (Invitrogen). RT-PCR was performed by SuperScript™ one-step RT-PCR with Platinum Taq according to the protocol provided (Invitrogen). For cDNA synthesis and pre-denaturation, we utilized 1 cycle (50°C for 15 min, 94°C for 2 min); and for PCR amplification we utilized 40 cycles (denature, 94°C for 15 s, anneal, 55°C for 30 s, extend 72°C for 1 min/kb) and for final extension 1 cycle (72°C for 10 min). The primer used included: for murine CTLA-4, sense 5′ ATGGCTTGTCTTGGACTCCGGAG and antisense 5′GTTGATGGGAATAAAATAAGGCTG; for murine PD-1, sense 5′ ATGTGGGTCCGGCAGGTACCCTG and antisense 5′ GAGGCCAAGAACAATGTCCATC; for GAPDH, sense 5′-ATGGTGAAGGTCGGTGTGAACGGATTTGGC and antisense 5′-CATCGAAGGTGGAAGAGTGGGAGTTGCTGT.

Inhibition of antigen-specific responses by ovarian carcinoma-associated Gr-1⁺CD11b⁺ cells

To investigate the suppressive role of ovarian carcinoma-associated Gr-1⁺CD11b⁺ cells in the antigen-specific response, 2 × 10⁶ splenocytes were co-cultured with VLP (final concentration 25 μg/ml) in triplicate in 24 well flat-bottom plates (Falcon, BD Biosciences) in 200 μl of RPMI-1640 medium according to our prior methods [20]. Gr-1⁺CD11b⁺ cells obtained from mice with or without ovarian 18D tumor cells, splenocytes from mice immunized with human papillomavirus type 16 virus-like particles (VLPs) and VLPs were, respectively, used as suppressors, responders and stimulators. Gr-1⁺CD11b⁺ MDSCs (5 × 10⁴ per well unless otherwise stated) were added into culture of splenocytes during VLP stimulation. The supernatants were harvested separately at 4 days. The IFN-γ in the supernatant was assayed by capture ELISA.

Blocking experiments were used to determine the role of costimulatory molecules. The ovarian carcinoma-associated PD-1+CTLA-4+Gr-1⁺CD11b⁺ cells were respectively incubated with neutralizing antibodies including anti-PD-1 (RMP1-14, eBioscience, San Diego, CA), anti-mouse CTLA-4 (UC10-4F10-11, Hamster IgG1) or both and isotypic control antibody and then added into the culture to observe the effect on the antigen-specific T cells to release IFN-γ. The supernatants were collected after 3 days and release of IFNγ was analyzed. All mAbs had an endotoxin level of <1 U/ml.

To further decide the effect of PD-1 and CTLA-4 on the Gr-1⁺CD11b⁺ MDSCs, Gr-1⁺CD11b⁺ MDSCs transfected with the PD-1 targeted, CTLA-4 targeted siRNAs or both PD-1 and CTLA-4 targeted siRNA or control siRNAs (Mock siRNA) (5 × 10⁴ per well unless otherwise stated) were added into culture of splenocytes during VLP stimulation, the supernatants were collected after 3 days and release of IFNγ was analyzed.

ELISA

Commercial sandwich ELISA kits were used for the quantitation of IFNγ (Pierce Endogen). The OD of each of the samples was measured at 450 nm using a SpectraMax 190 ELISA plate reader. Cytokine levels were quantified from two to three titrations using standard curves, and expressed in pg/ml.

Statistical analysis

For statistical analysis, we used Student’s test, and a 95% confidence limit was taken to be significant, defined as P < 0.05. Results were presented as mean ± standard errors. Survival was compared by Kaplan–Meyer analysis.

Results

PD-1+CTLA-4+myeloid-derived suppressor cells express arginase I

Our previous findings suggest that Gr-1⁺CD11b⁺ MDSCs from mice bearing 1D8 ovarian tumor are functionally different from those Gr1⁺CD11b⁺ cells of disease-free mice [20]. These Gr-1⁺CD11b⁺ MDSCs not only expressed high level of PD-1 and CTLA-4 but also expressed B7-H1 (PD-1 ligand) and CD80 (CTLA-4 ligand) (Fig. 1a; [20]); whereas these molecules were not detected on the surface of Gr-1⁺CD11b⁺ myeloid cells from disease-free mice (Fig. 1a; [20]). The Gr-1⁺CD11b⁺ MDSCs from mice bearing ovarian 1D8 tumor were composed of the cells with polymorphonuclear and mononuclear morphology (Fig. 1b). Consistent with other tumor-associated MDSCs, these murine ovarian 1D8 tumor-associated PD-1 and CTLA-4 double positive MDSCs (PD-1+CTLA-4+MDSCs) also expressed higher level of arginase I (Fig. 1c) when compared with the sorted Gr1⁺CD11b⁺ cells from tumor-free mice. The arginase I activity and expression in the sorted PD-1+CTLA-4+MDSCs could be regulated with conditioned medium containing ovarian 1D8 tumor supernatants. As illustrated in Fig. 1, while these sorted PD-1+CTLA-4+MDSCs were exposed to the conditional medium containing tumor supernatants, the activity and transcription of arginase I were remarkably upregulated (Fig. 1c). Furthermore, this upregulation was dependent on the concentration of tumor supernatants. The activity and transcriptional level of arginase I in PD-1+CTLA-4+MDSCs exposed to the medium containing 25% (V/V) or 50% (V/V) tumor supernatants was much higher than those exposed to the medium containing 5% (V/V) tumor supernatants (Fig. 1c). As a positive control, arginase I activity and expression in PD-1+CTLA-4+MDSCs was also regulated by mrIL-13. Thus, PD-1+CTLA-4+MDSCs did have potential to express arginase I. The PD-1+CTLA-4+MDSCs from the ascites of mice bearing ovarian 18D carcinoma were sorted according to our previous methods [20].

Fig. 1.

PD1+CTLA-4+MDSCs derived from ascites of mice bearing ovarian 1D8 carcinoma express arginase I. a Expression of PD1 and CTLA-4 on the surface of murine carcinoma-associated Gr1+MDSCs cells. A1, A2 MDSCs from ascites of mice bearing ovarian 1D8 carcinoma. A3, A4 Gr1+cells from naive mouse splenic cells. b Isolated PD1+CTLA-4+Gr1⁺CD11b⁺MDSCs from ascites of mice bearing ovarian carcinoma (Giemsa staining). PD-1+CTLA-4+MDSCs were sorted from the ascites of mice with ovarian 1D8 carcinoma using PE-labeled anti-Gr1 and FITC-labeled anti-CD11b using flow cytometry and the sorted cells were stained using Giemsa stain kit (Thermo Fisher Scientific, USA) according to the manufacture’s procedure. c Murine ovarian carcinoma supernatants promoted the arginase I activity and expression by PD-1+CTLA-4+MDSCs. Arginase activity (C1) and transcription level (C2) of arginase I in PD-1+CTLA-4+MDSCs upon exposing to different concentration of murine ovarian carcinoma supernatant and mrIL-13 (positive control). Cell contr. CD11b+Gr1+ cells from naive mouse splenic cells, MDSCs PD1+CTLA-4+MDSCs from the ascites of mice bearing ovarian carcinoma, Tu. Sup. tumor supernatants from ovarian 1D8 carcinoma after culturing for 24 h

Antibody blockade of PD-1 and CTLA-4 decreases the activity and expression of arginase I

Both PD-1 and CTLA-4 expressed on the activated T lymphocytes can transmit inhibitory signals to T lymphocytes and regulate an overlapping set of signaling protein [21, 22], implying possible role of PD-1 and CTLA-4 on the Gr1+CD11b+MDSCs. Because arginase I plays a critical role in inducing T cell immunotolerance and tumor evasion, we investigated the effect of PD-1 and CTLA-4 on the expression of arginase I. Ovarian carcinoma-associated PD-1+CTLA-4+Gr-1⁺CD11b⁺ MDSCs were blocked by anti-PD-1, anti-CTLA-4 and both antibodies to demonstrate the possible role of PD-1 and CTLA-4 in regulating arginase I activity and then to determine the effect of blockade on the activity and expression of arginase I induced by tumor supernatants. As expected, PD-1, CTLA-4 or both antibodies significantly decreased the activity and transcriptional level of arginase I while these cells were exposed to tumor supernatants (Fig. 2a), whereas Ab control did not. B7-H1 and CD80 are acted as PD-1 and CTLA-4 ligands, respectively, which are also expressed by ovarian carcinoma-associated PD-1+CTLA-4+MDSCs [20]. While anti-PD-1 ligand B7-H1, anti-CTLA-4 ligand CD80 or both antibodies were used to block PD-1+CTLA-4+Gr-1⁺CD11b⁺ MDSCs, they also showed similar inhibition on the activity and expression of arginase I. Similarly, CD80 and B7-H1 Ab control did not do this also (Fig. 2b). Thus, these results suggest that PD-1 and CTLA-4 may regulate the expression and activity of arginase I induced by tumor derived factors.

Fig. 2.

Blockade of PD-1, CTLA-4 or their ligands decreases arginase I activity and expression by ovarian carcinoma-associated MDSCs. a Blockade of PD-1, CTLA-4 or both reduced arginase I activity and expression. Anti-PD-1, CTLA-4 or both antibodies were, respectively, added into the culture of PD-1+CTLA4+MDSCs in the conditioned medium and then their arginase I activity and transcription level were analyzed after 24 h. b Blockade of B7-H1 and CD80 decreased the arginase I activity and transcription. Similarly, Anti-CD80, B7-H1 or both antibodies were added into the culture of PD-1+CTLA4+MDSCs in the conditioned medium and then their arginase I activity and transcription level were analyzed after 24 h. The isolated PD-1+CTLA4+MDSCs were plated at 1 × 10⁶ cells/well in 24-well tissue culture plates and stimulated with equal amount of tumor supernatants (25% V/V). Following stimulation, cells were washed with PBS and lysed in lysis buffer described in “Materials and methods”. Tu. Sup. Tumor supernatants from ovarian 1D8 carcinoma after culturing for 24 h

PD-1 and/or CTLA-4 knockdown by siRNA decreases the activity and expression of arginase I

To further address the role of PD-1 and CTLA-4 in regulating the expression and activity of arginase I, PD-1+CTLA-4+MDSCs were transfected with PD-1, CTLA-4, both PD-1 and CTLA-4 -targeted siRNA or mock siRNA as control to investigate the effect on arginase I activity and transcriptional level. Upon transfection, PD-1, CTLA-4 or both-targeted siRNA reduced transcript level of PD-1 or CTLA-4 (Fig. 3a) and expression of surface protein (Fig. 3b) in ovarian carcinoma-associated PD-1+CTLA-4+MDSCs, whereas the control siRNA did not. Notably, neither PD-1 nor CTLA-4-targetted nor non-specific control siRNA altered the level of other costimulatory molecules such as CD40 and CD86 in these cells (not shown). While equal amount of tumor supernatants were, respectively, added into PD-1, CTLA-4, both -targeted siRNA or control siRNA transfected Gr-1⁺CD11b⁺ MDSCs, PD-1, CTLA-4 or both -targeted siRNA transfected MDSCs showed the decreased activity and expression of arginase I, whereas control siRNA did alter the activity and transcriptional level of arginase I induced by tumor supernatants as compared to untransfected Gr1+CD11b+MDSCs (Fig. 3c). To further address the function of PD-1 and CTLA-4 on the surface of MDSCs, CD80 (CTLA-4 ligand), B7-H1 (PD-1 ligand) or both-targeted siRNA or control siRNA transfected Gr-1⁺CD11b⁺ MDSCs were exposed to tumor supernatants and then arginase I activity and expression was analyzed. Similarly, CD80, B7-H1 or both -targeted siRNA also decreased arginase I activity and expression (not shown).

Fig. 3.

Knockdown of PD-1, CTLA-4 or both decreases arginase I activity and expression. a PD-1 and CTLA4 siRNA reduced the level of PD-1 and CTLA-4, respectively, but not that of GAPDH transcript expression in the transfected ovarian carcinoma-associated PD-1+CTLA-4+MDSCs, as assessed by RT-PCR. b PD-1, CTLA4 or both siRNA but not non-specific siRNA, reduced the level of surface PD-1 and CTLA4 expression in transfected ovarian carcinoma-associated PD-1+CTLA-4+MDSCs, as assessed by flow cytometric analysis, respectively. B1, B2 Expression of PD-1 and CTLA-4 on the siRNA transfected ovarian carcinoma-associated PD-1+CTLA-4+MDSCs, respectively. Black line Isotype control, dark black line expression of PD-1 (B1) and CTLA-4 (B2) on ovarian carcinoma-associated PD-1+CTLA-4+MDSCs transfected with PD-1 and CTLA-4 specific siRNA, respectively, dotted line expression of PD-1 (B1) and CTLA-4 (B2) on ovarian carcinoma-associated PD-1+CTLA-4+MDSCs transfected by control mock siRNA. c Knockdown of PD-1, CTLA-4 or both reduced arginase I activity (C1) and level of transcription (C2). PD-1, CTLA-4 or both siRNA transfected PD1+CTLA-4+ MDSCs were cultured in the conditioned medium for 24 h and then lysed for analysis of arginase I activity. The total RNA was extracted for transcription assay according to the described protocol in “Materials and methods”. T. Sup. Conditioned medium containing 25% murine ovarian carcinoma supernatant (V/V), PD1siRNA, CTLA4siRNA and PD1+CTLA4siRNA, respectively, were PD1 siRNA, CTLA4 siRNA or both transfected PD1+CTLA4+MDSCs, Mock siRNA control siRNA transfected PD1+CTLA4+MDSCs

CD80 deficiency decreases the arginase I activity and expression in ovarian carcinoma-associated MDSCs

We previously reported that growth of ovarian 1D8 tumor is retarded in CD80 deficient mice [20]. CD80 and CD80 ligand CTLA4-targeted siRNA decrease the expression of arginase I induced by tumor supernatant, suggesting the role of CTLA-4 and its ligand CD80. Thus, we isolated the Gr-1⁺CD11b⁺ MDSCs derived from these CD80−/− mice and wt mice bearing tumor, respectively, and then examined activity and expression of arginase I. As predicted, fresh Gr-1⁺CD11b⁺ cells derived from CD80−/− mice bearing ovarian 1D8 tumor indeed exhibited lower level of activity and expression of arginase I as compared to those isolated from wt mice with the equivalent tumor load (Fig. 4a). While isolated Gr-1⁺CD11b⁺ cells from both CD80−/− and wt mice were cultured in the conditional medium containing the 25% tumor supernatant or in the presence of IL-13, CD80−/− Gr-1⁺CD11b⁺ cells also exhibited reduced arginase I activity and expression in response to equal amount of tumor supernatants as compared to wild type CD80+Gr-1⁺CD11b⁺ cells, further strengthening the findings that CTLA-4 and CD80 expressed by MDSCs regulate the expression of arginase I (Fig. 4b). Since it is difficult to get PD-1 or its ligand knockout mice, the data from these mice is absent.

Fig. 4.

CD80 deficiency decreases arginase I activity and expression by ovarian carcinoma-associated PD1+CTLA-4+ MDSCs. a The activity and expression of fresh CD11b+Gr1+ cells from wild type (WT) or CD80 deficiency (CD80−/−) spleen of mouse with (Tumor) or without ovarian carcinoma (Contr.). Activity (A1) and transcriptional level (A2) of arginase I were analyzed in WT mice or CD80−/− mice by the amount of urea released from cell extract and RT-PCR. b CD80−/− reduced the arginase activity (B1) and transcription (B2) induced using tumor supernatant and mrIL13. Contr. PD1+CTLA-4+MDSCs cells alone, Tu.Sup. PD1+CTLA-4+ MDSCs were cultured in conditioned medium containing 25% tumor supernatant, and mrIL13 PD1+CTLA-4+ MDSCs were cultured in 10 ng/ml IL13 medium

Antibody blockade or siRNA silencing of PD-1 and CTLA-4 reduces the suppressive potential of PD-1+CTLA-4+MDSCs

To address the role of PD-1 and CTLA-4 in the suppressive effects of MDSCs upon antigen-specific T cell responses, we again used splenocytes from mice vaccinated with the potent heterologous antigen, HPV16 L1 VLPs. The pretreatment of MDSCs by PD-1 or CTLA-4 -specific neutralizing antibody, but not by control antibody reduced the suppressive effect of ovarian carcinoma-associated Gr-1⁺CD11b⁺ MDSCs upon VLP-specific and T cell-dependent IFNγ release (Fig. 5a). PD-1 and CTLA-4-specific blocking antibody need not be directly added into culture medium, suggesting that the effect was mediated by binding of PD-1 antibody and CTLA-4 antibody to PD-1+CTLA4+MDSCs. Since antibody blockade suggests that PD-1 and CTLA-4 signaling contributes to suppression of antigen-specific immune responses by the PD-1+CTLA-4+MDSCs, we reasoned that blockade of the PD1 ligand B7-H1 or CTLA-4 ligand CD80 might produce a similar effect. Indeed, as shown previously, antibody pretreatment of B7-H1 and CD80 also significantly decreased the suppression of VLP-specific and T cell-dependent IFNγ release [20]. Combination of PD-1 and CTLA-4 Abs produced a stronger inhibition than PD-1 or CTLA-4 Abs alone (P < 0.05) and their affects appeared to be additive (Fig. 5a). Ovarian carcinoma-associated MDSCs express both PD-1 and CTLA-4. Not only PD-1 but also CTLA-4 could activate the intracellular inhibitory signals. This might account for stronger inhibition of MDSCs suppression by antibody blockade of MDSCs by both PD-1 and CTLA-4 Abs as compared to PD-1 or CTLA-4 Abs alone.

Fig. 5.

Antibody blockade or siRNA silencing of PD-1 and CTLA-4 reduces the suppressive potential of PD-1+CTLA-4+ MDSCs. a Effect of pretreatment of PD-1+CTLA-4+ MDSCs using neutralizing antibody PD-1, CTLA-4 or both on the generation of IFNγ by VLP-specific splenocytes (VLP.SC). Supernatants of cultures were collected after 24 h and IFNγ was assayed by capture ELISA. Isotypic contr. Control antibody. b PD-1, CTLA-4 or both siRNA transfected MDSCs had the reduced suppression on VLP-induced generation of IFN-γ in VLP-specific splenocyte culture. PD-1+CTLA-4+ MDSCs were sorted from the ascites of mice with ovarian 1D8 carcinoma and then transfected using PD-1, CTLA-4 or both siRNA or mock siRNA and then added into VLP-specific splenocyte culture. PD-1+, CTLA-4+ or CTLA-4 and PD1 siRNA represented PD-1+CTLA-4+MDSCs, which were transfected with PD-1, CTLA-4 or both siRNA, respectively, mock siRNA PD-1+CTLA-4+ MDSCs transfected with control siRNA

We next sought to further confirm the role of PD-1 and CTLA-4 in suppression of antigen-specific T cell responses by PD-1+CTLA-4+MDSCs using PD-1, CTLA-4 or both-targeted siRNA. Ovarian carcinoma-associated PD-1+CTLA-4+MDSCs transfected with PD-1, CTLA-4 or both -targeted siRNA or control siRNA were added into culture of VLP-specific splenocytes with VLPs. Ovarian carcinoma-associated PD-1+CTLA-4+MDSCs transfected with PD-1, CTLA-4 or both-targeted siRNA exhibited significantly reduced suppressive potential for VLP-specific and T cell-dependent induction of IFNγ release as compared to control siRNA transfected or untransfected MDSCs (Fig. 5b). Thus, conjugation of PD-1, CD80 and their ligands expressed on the surface of MDSCs possibly plays a role in inhibitory function of MDSCs to antigen-specific T cells by maintaining and promoting the inhibitory potential of MDSCs.

Antibody blockade of PD-1 or CTLA-4 slows tumor growth and improves the survival rate of tumor-bearing mice

To further address the role of PD-1 and CTLA-4-dependent signaling on tumor tolerance, we examined the effect of both PD-1 and CTLA-4-specific neutralizing antibodies upon the growth of the ovarian 1D8 ovarian carcinoma model in C57BL/6 mice. Administration of either PD-1 or CTLA-4-specific neutralizing antibody significantly delayed the growth of tumor (P < 0.05), whereas control antibody failed to retard tumor growth (Fig. 6a). Meanwhile, antibody blockade of combination of PD-1 and CTLA-4 was much stronger than PD-1 or CTLA-4 alone (Fig. 6). Furthermore, PD-1, CTLA-4 or both specific neutralizing antibodies also significantly improved the survival rate of tumor-bearing mice as compared to control mice (P < 0.05) (Fig. 6b). Thus, blockade of PD-1 or CTLA-4 not only slows tumor growth but also improves the survival rate of tumor-bearing mice.

Fig. 6.

Blockade of PD-1 or CTLA-4 slows tumor growth and improves the survival rate of mice bearing tumor. a In vivo injection of neutralizing antibody to PD-1, CTLA-4 or both, significantly slowed and retarded tumor growth. b In vivo injection of neutralizing antibody to PD-1, CTLA-4 or both, significantly prolonged the survival rate. Mice (6/group) were injected using 1D8 cells (s.c. with 1 × 10⁷) and i.p. with 250 μg of control antibody or antibody specific for PD-1 and CTLA-4 or both on days −6, −4, −2 and +1 as previously described [20]

Discussion

In this study, we further confirm our previous findings that a high level of PD-1 and CTLA-4 in mouse ovarian carcinoma-associated Gr-1⁺CD11b⁺ MDSCs are detected, while other co-stimulatory molecules, namely CD40, B7-DC and CD86, are not detected. As a new finding, both PD-1 and CTLA-4 expressed by Gr-1⁺CD11b⁺ MDSCs may regulate the activity and expression of arginase I that plays a critical role in inducing the tolerance of T cells and antigen-specific T cells. Thus, our results suggest an important mechanism of PD-1 and CTLA-4 on the MDSCs in induction and maintenance of tumor-immunotolerance (Fig. 7). MDSCs induced by tumor-derived factors such as VEGF, TGF-β, IL-6 and PEG2 etc. can express high level of PD-1, CTLA4 and also B7-H1 and CD80, which are ligands of PD-1 and CTLA-4, respectively. Both PD-1 and CTLA4 on the surface of MDSCs upregulate the activity and expression of arginase I via binding with their ligands expressed on the MDSCs. These upregulated arginase I can inhibit activation of CD4+ and CD8+ T cells [14, 15]. Other studies also show that binding of CTLA-Ig to dendritic cells (DCs) induces the expression of indoleamine 2, 3-dioxygenase (IDO), an enzyme that degrades the essential amino acid tryptophan [23]. The caused tryptophan metabolites can suppress T cell response in vitro and in vivo [24] as well as T cell clonal expansion [23]. Since PD-1 ligands are constitutively expressed on fresh isolated splenic T cells, CD4+CD25+Treg cells, B cells, macrophages and DCs and get upregulated on T cells, macrophages, and DCs after activation and also on MDSCs, there may exist a costimualtory molecules-based immuno-regulating wet, not only including between antigen-presenting cells and T cells, but also, including among all of these MDSCs, as described by regulation among MDSCs in this paper.

Fig. 7.

Schematic illustration of a mechanism by which PD-1 and B7-H1 on MDSCs subvert antitumor immunity via regulating the activity and expression of arginase I. MDSCs were induced by tumor-derived factors such as VEGF, TGF-β, IL-6 and PEG2 etc. These tumor-associated MDSCs express higher level of PD-1, CTLA4 and also B7-H1 and CD80, which are ligands of PD-1 and CTLA-4, respectively. Both PD-1 and CTLA4 on the surface of MDSCs may upregulate the activity and expression of arginase I via binding with their ligands expressed on the MDSCs. These upregulated arginase I can inhibit the activation of CD4+ and CD8+ T cells to regulate immune responses. BMC Bone marrow cells

Studies in the field of costimulatory molecules focused previously on the effect of costimulatory molecules on the activity and function of effective immune cells, namely T cells, and B cells. The importance of CTLA-4 as a negative T cell costimulatory molecule in physiologic termination of T cell responses is highlighted by the observation that CTLA-4 gene knockout mice develop massive lymphoproliferation and early death [25, 26]. Recent studies strongly suggested that CTLA-4 may function as a master switch, playing a key role in peripheral tolerance [27, 28]. Indeed, we and others showed that blockage of CTLA-4 signaling results in exacerbation of autoimmune disease and decrease of tumor growth [29, 30]. CTLA-4 on the MDSCs can regulate the arginase activity and expression of arginase I, implying the CTLA-4 may inhibit the activating signaling, which can promote the expression and upregulate the activity of arginase I. This may be another mechanism of CTLA-4 to exert immune inhibitory role. Similar to CTLA-4, PD-1 posses only a single V-like domain and an immunoreceptor tyrosine-based inhibition motif (ITIM) within its cytoplasmatic tail. The inhibitory activity of PD-1 is dependent on association of the cytoplasmic tyrosine phosphatase SHP-2 with ITSM rather than with ITIM motif that is more typically associated with inhibitory signals [31]. This molecule is expressed by myeloid cells and B cells [32]. Indeed, PD-1 molecules expressed on the surface of MDSCs can be involved in the regulation of MDSC function by regulating the activation and expression of arginase I.

The regulation of arginase I activity and expression by co-stimulatory or coinhibitory molecules on the MDSCs might be important in inducing and maintaining immunosuppression. Recent studies show that arginase I plays a critical role in tumor growth and escape from the immune-surveillance. Arginase I could be found in the serum and tumor microenvironment of patients with colon-rectal carcinoma [33], breast [34], skin [35] and prostate cancer [36]. Arginase I can also be released by tumor-associated MDSCs [5, 15]. Tumor progression showed a marked increase in arginase I expressing MDSCs among the splenocytes and impaired antitumor cytotoxic lymphocytes that could be recovered by in vitro CD11b+ cell depletion [37]. Thus, our finding may produce the important role in understating the immune regulation of immunosuppression and design of antitumor drugs.