Published online before print June 19, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1,1,2
,
* Departments of Surgical Oncology,
Imunopathology, and
Oral and Maxillofacial Sciences, Gifu University Graduate School of Medicine, Gifu, Japan;
Department of Cell Biology, New York University School of Medicine, New York, New York, USA; and
|| Department of Preventive Medicine, University of Debrecen, Debrecen, Hungary
2 Correspondence: Department of Immunopathology, Gifu University Graduate School of Medicine, 1-1 Yanagido, Gifu, 501-1194, Japan. E-mail: saio{at}gifu-u.ac.jp
ABSTRACT
Mononuclear phagocytes (MPCs) at the tumor site can be divided into subclasses, including monocyte-lineage myeloid-derived suppressor cells (MDSCs) and the immunosuppressive tumor-infiltrating macrophages (TIMs). Cancer growth coincides with the expansion of MDSCs found in the blood, secondary lymphoid organs, and tumor tissue. These MDSCs are thought to mature into macrophages and to promote tumor development by a combination of growth-enhancing properties and suppression of local antitumor immunoresponses. As little is known about either subset of MPCs, we investigated MPCs infiltrating into murine adenocarcinoma MCA38 tumors. We found that these MPCs displayed immunosuppressive characteristics and a MDSC cell-surface phenotype. Over 70% of the MPCs were mature (F4/80+Ly6C–) macrophages, and the rest were immature (F480+ Ly6C+) monocytes. MPC maturation was inhibited when the cells infiltrated a tumor variant expressing IL-2 and soluble TNF type II receptor (sTNFRII). In addition, the IL-2/sTNFRII MCA38 tumor microenvironment altered the MPC phenotype; these cells did not survive culturing in vitro as a result of Fas-mediated apoptosis and negligible M-CSFR expression. Furthermore, CD4+ tumor-infiltrating lymphocytes (TILs) in wild-type tumors robustly expressed IL-13, IFN-
, and GM-CSF, and CD4+ TILs in IL-2/sTNFRII-expressing tumors expressed little IL-13. These data suggest that immunotherapy-altered Th cell balance in the tumor microenvironment can affect the differentiation and maturation of MPCs in vivo. Furthermore, as neither the designation MDSC nor TIM can sufficiently describe the status of monocytes/macrophages in this tumor microenvironment, we believe these cells are best designated as MPCs.
Key Words: cancer • tumor necrosis factor • interleukin-2 • interleukin-13 • granulocyte macrophage-colony stimulating factor
INTRODUCTION
Myeloid-derived suppressor cells (MDSCs) are myelomonocytic cells that can suppress the proliferation and function of T cells [1
, 2
]. MDSCs were initially examined using the bone marrow of tumor-bearing animals [3
4
5
6
7
]. These studies showed that a functional subset of MDSCs originates from myeloid cells in the bone marrow, and more recent research further revealed that MDSCs involved in immunosuppression are derived from a monocytic, rather than a granulocytic, population. The cell-surface phenotype of MDSCs was identified as CD11b+CD11c+Gr-1+IL-4R
+, indicating that they belong to the "inflamed monocyte" class of cells [8
]. In addition, a subset of the bone marrow-developed MDSC population that matures at the tumor site [1
] differentiates into tumor-infiltrating macrophages (TIMs) [9
].
Before eliciting effector functions, MDSCs must first be activated by cytokines such as IL-4, IL-13, or IFN-, which are generated by tumor-infiltrating lymphocytes (TILs) or by MDSCs themselves (in an autocrine manner) [10
]. In addition, the activation of monocyte/macrophage lineage cells depends on Th cell products [11
], and monocyte/macrophage activation is classified according to the Th1/Th2 paradigm [12
, 13
]. It is thus important to consider how T cells affect MDSC differentiation in situ, particularly, as there are currently no reports directly showing the role of tumor-infiltrating T cells in MDSC maturation.
We previously analyzed tumor-infiltrating mononuclear phagocytes (MPCs) comprised of monocytes and macrophages in animal models [14 15 16 ]; however, these previous studies did not thoroughly address whether tumor-infiltrating MPCs could be classified as tumor-infiltrating MDSCs [9 ]. Furthermore, we have not yet examined how tumor-site T cells affect the differentiation of tumor-infiltrating MPCs. In the present study, we address these questions using an IL-2-based immunotherapy model. We found that tumor-infiltrating MPCs could be classified phenotypically as MDSCs and that the maturation of tumor-infiltrating MPCs was affected by the Th balance of tumor-infiltrating CD4+ TILs in situ.
MATERIALS AND METHODS
Mice
Male C57BL/6 (B6) and gld mice, 6–8 weeks old, were purchased from SLC (Wilmington, MA, USA). TNF receptor type I (TNFRI)–/– (Tnfrsf1atm1Mak) and TNFRII–/– (Tnfrsf1btm1Mwm) mice with B6 backgrounds were purchased from The Jackson Laboratory (Bar Harbor, MA, USA). Mice were maintained in accordance with the guidelines of the Committee on Animals of the Gifu University School of Medicine (Japan).
Cell lines
A murine colon carcinoma cell line (MCA38) and its derivatives were cultured in RPMI-1640 medium containing 10% FCS, L-glutamate, and penicillin-streptomycin (all from Invitrogen Life Technologies, Tokyo, Japan). Cells were maintained at 37°C in a humidified 5% CO2 atmosphere. Murine IL-2 and soluble (s)TNFRII cDNA preparation, expression plasmid construction, and MCA38 transfectant isolation were described in detail previously [16
]. In the experiments presented here, we used one clone each of IL-2-expressing MCA38 cells (#75, designated MCAIL-2), sTNFRII-expressing MCA38 cells (#3, MCAsTNFR), and MCA38 cells expressing IL-2 and sTNFRII (#15, MCAIL-2/sTNFR). A clone of MCA38 expressing an "empty" expression plasmid was used as a control (MCAmock). Each sub-cell line used in this study displayed the same growth kinetics in vitro as nonmodified (wild-type) cells (data not shown).
Tumor growth
Mice were anesthetized by an injection of pentobarbiturate (2.5 mg/mouse). For intrahepatic implantation, 106 cells were inoculated into the sub-capsule of the liver. Tumor development was observed in the parenchyma, with no tumor growth observed outside of the liver. For s.c. implantation, 3 x 106 cells were inoculated into the s.c. space of the abdominal wall. Tumor size (longitudinalxlateral diameter) was measured by caliper.
Isolation of tumor-infiltrating cells
Mice were killed 14 days after tumor implantation, and cells were isolated by collagenase digestion of the tumor tissue and magnetic immunobeading, as described previously [17
] (Miltenyi Biotec, Berdish-Gladbach, Germany). According to flow cytometric analysis, the purity of the purified tumor-infiltrating CD4+ cells (CD4+CD3+ cells) was more than 95%.
Preparation of peritoneal exudate cells (PECs)
PECs were collected from B6 mice 4 days after i.p. administration of 4 ml 4% thioglycolate.
Flow cytometry
Cells for flow cytometric analysis were preincubated with 0.010 mg/ml anti-CD16/32 (clone 2.4G2, BD Biosciences, Rockville, MD, USA) at 4°C for 30 min prior to staining. For all stainings, Via-Probe (BD Biosciences) was used for dead cell exclusion following the manufacturer’s instructions. The following antibody reagents were used, each at 0.010 mg/ml: allophycocyanin (APC)-conjugated anti-CD11b (clone M1/70, BD Biosciences), FITC- or PE-conjugated anti-F4/80 (clone F4/80, Serotec Ltd., Oxford, UK), FITC-conjugated anti-Ly6C (clone ER-MP20, BMA Biomedicals, Augst, Switzerland), PE-conjugated anti-Gr-1 (clone RB6-8C5, BD Biosciences), PE-conjugated anti-CD119 (clone 2E2, BD Biosciences), and PE-conjugated anti-CD40 (clone 3-23, BD Biosciences). Control hamster IgG and rat Ig (BD Biosciences) were used as FITC or PE conjugates. Biotin-conjugated goat polyclonal anti-CD124 (R&D Systems, Minneapolis, MN, USA) followed by PE-conjugated avidin (BD Biosciences)- and biotin-conjugated control goat IgG (R&D Systems) were used for the analysis of CD124 expression in MDSCs.
To determine the percentages of CD4+ and CD8+ TILs and MDSCs in the total tumor digest, APC-conjugated anti-CD3 (clone 145C11, BD Biosciences) and anti-CD11b (M1/70, BD Biosciences) and PE-conjugated anti-CD4 (RM4-5, BD Biosciences), anti-CD8b (53-5.8, BD Biosciences), anti-CD24 (M1/69, BD Biosciences), and anti-F4/80 (clone F4/80, Serotec Ltd.) antibodies were used. Cells were incubated with antibodies for 30 min at 4°C, washed with PBS, fixed with 1% paraformaldehyde/PBS, and analyzed using a FACSCalibur flow cytometer and CellQuest software (Becton Dickinson, Tokyo, Japan).
Cell viability assays
CD11b+ cells isolated from tumors were cultured in vitro (24 h), and their viability was evaluated by a lactate dehydrogenase (LDH) cytotoxicity assay (CytoTox-OneTM, Promega Corp., Madison, WI, USA), according to the manufacturer’s instructions. In brief, dead cell-released LDH was quantified from the cellular supernatants, and the total LDH in cultured cells was determined after cell lysis.
Apoptosis analysis
Cells were washed with annexin-binding buffer (0.01 M HEPES, 0.14 M NaCl, 2.5 mM CaCl2, pH 7.4) before staining with PE-conjugated annexin V and Via-Probe (BD Biosciences) for 15 min at 4°C and flow cytometric analysis. For microscopic observation of apoptosis, cells were washed with PBS and stained with 1 nM Hoechst 33258 in PBS.
In vitro coculture experiment
Activated splenic CD8+ T cells and tumor-infiltrating MDSCs were cocultured as described previously [18
]. Briefly, Day 12 tumor-infiltrating MDSCs (0.75x106 cells/ml) were purified as described above and cultured for 4 h. Twenty-four hours before coculture, CD8+ T cells were purified from normal mouse spleens using CD8+ magnetic immunobeads, according to the manufacturer’s instructions (Miltenyi Biotec). CD8+ T cells were then stimulated with plate-bound anti-TCRβ antibody (H57-597, BD Biosciences). Activated T cells (0.5x106 cells/cm2) were then cocultured with MDSCs.
To quantify suppression of T cell proliferation, the activated, normal T cells were labeled with CFSE (Invitrogen Life Technologies) prior to cocultivation. After 24 h of coculture, the cells were stained with APC-anti-CD3, PE-anti-CD8, and Via-Probe to identify the T cells, and their CFSE levels were analyzed by FACS.
NO analysis
Cells (106) were plated onto 24-well plates in 2 ml complete medium and incubated with or without 100 U/ml recombinant IFN- (BD Biosciences) for 24 or 48 h. Supernatants (0.1 ml) were collected and mixed with 0.1 ml Greiss reagent (Sigma Chemical Co., St. Louis, MO, USA) in 96-well plates for 15 min. Absorbance at 540 nm was then determined with a Benchmark® Microplate reader and analyzed using Microplate Manager III software (BioRad Laboratories, Hercules, CA, USA) with sodium nitrite as a standard. Unstimulated PECs, cultured for 24 or 48 h, were used as a negative control to confirm that PECs did not produce detectable levels of NO (
0.391 ng/ml) during the culture period.
Multiple PCR and real-time PCR analysis
Total RNA was purified using Trizol (Invitrogen Life Technologies), and 600 ng total RNA was used for RT using Superscript III (Invitrogen Life Technologies). For multiple PCR analysis, we used a cytokine Th1/Th2 Multiplex PCR kit (Mouse Th1/Th2 Set-2, Maxim Biotech, Inc., South San Francisco, CA, USA). Briefly, we mixed 100 ng cDNA or the supplied positive-control DNA with buffer, Taq polymerase, and a multiple primer pair mixture. PCR was then performed according to the conditions provided by the manufacturer. Amplified DNA was analyzed by 5% acrylamide gel electrophoresis followed by ethidium bromide staining.
For real-time PCR analysis, 5 µL 20x diluted cDNA samples, 1 µL each 10 mM primers, 3 µL PCR-grade water (Roche, Indianapolis, IN, USA), and 10 µL 2x concentrated Syber Green and Taq enzyme-premixed reaction mixture (Takara Chemical Co., Japan) were used. The sequences of the primer pairs used in this analysis were: GM-CSF, 5'-AAGGGCGCCTTGAACATGAC-3', 5'-AAATCCGCATAGGTGGTAACTTGTG-3'; GM-CSFR, 5'-GAGAACCTGACCTGCGAGATCC-3', 5'-TCGCAATGCACTGTGTGATGA-3'; GM-CSFRβ, 5'-CACATTTCTGGCTCTGTGTGAAAG-3', 5'-TAGCGACAAGAGCCTGACCAA-3'; M-CSF, 5'-GAACAGCCTGTCCCATCCATC-3', 5'-TGAGGCCAGCTCAGTGCAA-3'; M-CSFR-variant1, 5'-CTTTGGACTGGCTAGGGACATCA-3', 5'-ATGCCGTAGGACCACACATCAC-3'; M-CSFR-variant2, 5'-CTTTGGACTGGCTAGGGACATCA-3'. The reaction conditions were: one cycle of 95°C for 5 min, 45 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 10 s, followed by a melting curve analysis. The reaction and data analysis were performed using a LightCycler (Roche).
mAb purification
For antibody purification, hybridoma cells were cultured in CD Hybridoma serum-free culture medium supplemented with L-glutamate and penicillin-streptomycin (all from Gibco, Grand Island, NY, USA). The supernatant was concentrated with an Amicon® Model 8200 ultra-concentration module and YM-30 membrane (Millipore, Bedford, MA, USA) and then dialyzed in 0.02 M phosphate buffer (pH 8.0). The concentrated supernatant was passed through a Hi-Trap protein G affinity column (Amersham Biosciences, Piscataway, NJ, USA) and washed with 0.02 M phosphate buffer (pH 8.0). The antibody was eluted from the column by 0.02 M citrate buffer (pH 2.0), and the eluted fraction was dialyzed into PBS. Antibody concentration was determined with a DC Protein Assay (Bio-Rad Laboratories).
In vivo depletion of NK cells and T cells
Purified antibody (100 µL 500 µg/ml; anti-NK1.1, clone PK136; anti-CD4, clone GK1.5; and anti-CD8a, clone 53.6.72, American Type Culture Collection, Manassas, VA, USA) was administered i.p. every 2 days for the 5 days before tumor inoculation.
SDS-PAGE and immunoblotting
CD11b+ cells from tumor tissues or PECs were washed three times with PBS before being incubated on ice for 1 h with lysis solution [50 mM Tris-HCl (pH. 7.5), 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, 5 mM iodoacetamide, 1 mM Na3VO4, and protease inhibitor cocktail (P-8340, Sigma Chemical Co.)]. Cell supernatants were then collected and subjected to electrophoresis on 10% PAGE gels. After transferring the proteins onto polyvinylidene difluoride membranes (Bio-Rad Laboratories), the membranes were blocked with skim milk and then reacted for 1 h with 0.002 mg/ml mouse anti-NOS II (Clone 6, BD Biosciences) or anti-rat/mouse Arginase I (Clone 19, BD Biosciences) mAb, diluted in a primary antibody signal-enhancing solution (Can Get SignalTM, Toyobo Co., Ltd., Osaka, Japan). After washing, the membranes were incubated for 1 h with HRP-conjugated goat anti-mouse IgG (Nichirei Biosciences, Inc., Tokyo, Japan), diluted 50 times in a secondary antibody signal-enhancing solution (Toyobo Co., Ltd.). The blots were developed with ECL (Amersham Pharmacia Biotech, Buckinghamshire, UK), and the images were captured by a Cool Saver Lumino-Capture System (Model AE-6955, ATTO, Tokyo, Japan) before analysis using CS-Analyzer software (ATTO). The membrane was then stripped and reprobed with 0.002 mg/ml anti-GAPDH (clone 9.B.88, United States Biological, Swampscott, MA, USA) and HRP-conjugated goat anti-mouse IgG (Nichirei Biosciences, Inc.). Image intensities were then analyzed by CS-Analyzer software (ATTO).
Immunocytochemistry
Mice were killed 14 days after tumor inoculation, and tissue samples were collected and embedded with optimal cutting temperature compound (Tissue Tek, Miles, Elkhart, IN, USA) before freezing under liquid nitrogen. Sections (4 µ) were cut using a cryostat and then fixed with cold acetone for 10 min. The specimens were pretreated with anti-CD16/32 (2.4G2, 10 µg/ml, PharMingen, San Diego, CA, USA) and then reacted with FITC-conjugated rat anti-mouse CD11b (M1/70, 10 µg/ml, PharMingen) and PE-conjugated rat anti-mouse CD4 (RM4-5, 10 µg/ml, PharMingen) or CD8b (53-5.8, 10 µg/ml, PharMingen). After washing three times with PBS, specimens were analyzed with a model DM RA fluorescence microscope with QFISH® software (Leica Microsystems Imaging Solutions Ltd., Cambridge, UK).
Statistics
We used the Student’s t-test for all statistical analyses, and P values <0.05 were considered significant.
RESULTS
IL-2 and TNF affect maturation but not characteristics of tumor-infiltrating MPCs. We observed growth curves for each tumor subtype in vivo to address how tumor modifications affected tumor growth. As shown in Figure 1A , expression of sTNFRII alone did not affect tumor size. However, expression of IL-2 alone or coexpression of IL-2 and sTNFRII strongly affected tumor growth in vivo, and three out of five IL-2-expressing and five out of five IL-2/sTHFRII-expressing tumors demonstrated regression
 View larger version (42K): [in a new window] |
Figure 1. (A) Tumor growth of mock-, IL-2-, sTNFRII-, and IL-2/sTNFRII-expressing tumors in vivo. Tumor cells (3x106) were inoculated into the s.c. space of the abdominal wall of B6 mice, and the tumor size was measured as a function of time. (Left panel) Average tumor size of all tumor subtypes. (Right panel) Tumor size of individual IL-2-expressing (black) and IL-2/sTNFRII-expressing (white) tumors from B6 mice. Note that three out of five IL-2-expressing and five out of five IL-2/sTNFRII-expressing tumors began to regress from Day 9 of tumor inoculation. (B and C) Flow cytometric analysis of MPC markers. On Day 12 of tumor growth, tumors were collected, and CD11b+ cells were isolated and analyzed by flow cytometry. (B) CD11c (left panel) and Gr-1 (right panel) expression by MCA38/mock-infiltrating CD11b+F4/80+ cells is demonstrated. (C) A representative histogram of CD124 (upper) and CD40 (lower) expression on MPCs from each tumor type is shown. (D) Effect of tumor-infiltrating MPCs on the growth of activated CD8+ T cells. Splenic CD8+ T cells were activated with anti-TCR antibodies for 24 h. CD8+ T cells were then labeled with CFSE prior to cocultivation with MPCs. PECs were prepared from B6 mice 4 days after i.p. injection with thioglycolate and served as a control. The MPCs and PECs were cultivated alone for 4 h and cocultured with the activated CD8+ T cells for 24 h; T cells were labeled with anti-CD3 and anti-CD8 antibodies and Via-Probe; and CFSE levels were determined by flow cytometric analysis. Numbers indicate the percentage of M1 (proliferating)- and M2 (not proliferating)-phase cells. (E and F) Expression of the maturation marker Ly6C on MPCs. CD11b+ cells were recovered from each tumor 12 days after tumor inoculation and stained with CD11b, Ly6C, F4/80, and Via-Probe. CD11b+ Via-Probe– cells were analyzed for Ly6C and F4/80 expression. (E) Samples indicated in the histogram are the cells recovered from the MCA38/mock case. M, Mature; IM, immature. (F) We present analyses of CD11b+ MPC maturation for cells derived from each tumor type in immature (left panel) and mature (right panel) areas from E.
|
+ phenotype [2
]. We probed the expression levels of CD11c and Gr-1 on tumor-infiltrating CD11b+ cells and found that all of the tumor-infiltrating CD11b+F4/80+ cells from each immunotherapeutically modified tumor microenvironment expressed CD11cint and Gr-1low. Representative data of CD11c and Gr-1 expression on MCA38 mock-derived, tumor-infiltrating CD11b+F4/80+ cells are shown in Figure 1B
. We further investigated whether the MDSC-phenotypic cell-surface characteristics of the tumor-infiltrating CD11b+F4/80+ cells were affected by immunotherapy. As shown in Figure 1C
, immunotherapy did not alter the MDSC phenotype of CD11b+F4/80+ cells, which displayed the markers CD80+CD86+CD120a+CD120b+ (data not shown). However, in the tumor microenvironment, Gr-1 expression by tumor-infiltrating F4/80+ cells is not necessary to define cells as MDSCs [9 ]. Rather, MDSCs are defined solely as immunosuppressive myeloid-derived cells [2 ]. Therefore, we further analyzed the immunosuppressive abilities of the tumor-infiltrating CD11b+ cells and confirmed that the cells could inhibit cocultured CD8+ T cell proliferation in vitro (Fig. 1D) . As Ly6C is expressed in immature monocyte lineage cells and used for the determination of monocyte/macrophage maturation [19 ], we also investigated whether immunotherapy affected the maturation of tumor-infiltrating MPCs. Wild-type, tumor-infiltrating MPCs were predominantly mature (F4/80+Ly6C– macrophages), whereas MPCs in IL-2/sTNFRII-expressing tumors were significantly less mature (F4/80+Ly6C+ monocytes), implying that modification of the tumor microenvironment prevented MPC maturation, despite maintenance of the MDSC phenotype (Fig. 1E and 1F) .
Tumor-infiltrating MPCs in IL-2/sTNFRII-expressing tumors did not survive in vitro
As MPC maturation was affected by IL-2, particularly for MPCs in IL-2/sTNFRII-expressing tumors (Fig. 1E
and 1F)
, we examined the characteristics of these MPCs in vitro. Interestingly, compared with MPCs from other tumor types, tumor-infiltrating MPCs from IL-2/sTNFRII-expressing tumors had the least production of NO in response to ex vivo IFN- (Fig. 2A
). Furthermore, we observed no difference in the expression of the IFN- receptor (CD119) between the tumor groups (Fig. 2B)
. As tumor-infiltrating MPCs in IL-2/sTNFRII-expressingtumors had a MDSC phenotype (Fig. 1A
and 1B)
, we hypothesized that the reduced NO production was not related to the absence of MPC maturation.
 View larger version (36K): [in a new window] |
Figure 2. (A) NO production in purified MPCs after treatment with different concentrations of IFN- in vitro. Purified MPCs were cultured in the presence or absence of different concentrations of IFN- (untreated, 10 ng/ml or 100 ng/ml) for 24 (left panel) or 48 (right panel) h, and the supernatants were analyzed. *, Significance as compared with the mock group under the same treatment conditions. (B) Freshly isolated CD11b+ MPCs were analyzed for CD119 expression by flow cytometry. (C) Annexin V labeling of MPCs. CD11b+ MPCs were isolated from each tumor subtype after 12 days of growth and analyzed for annexin V binding before and after 24 h of culture (with no cytokine supplements). (D) Cell toxicity assay of MPCs after 24 h of culture. The percentage of dead cells was measured by a LDH release assay of culture supernatants. (E) MPC staining with Hoechst 33258. MPCs isolated from IL-2/sTNFRII-expressing tumors were analyzed by vital dye staining and microscopy (x200) after 5 h in culture.
|
Tumor-infiltrating MPC apoptosis was partially rescued in gld mice
We next investigated whether we could rescue tumor-infiltrating MPCs in IL-2/sTNFRII-expressing tumors from apoptosis by inoculating the tumor in Fas ligand mutant gld mice (Fig. 3A
and 3B
). We found that although MPC death was prevented, and NO production was improved, cell death was not recovered completely (Fig. 3C)
. These findings suggest that there is another mechanism of cell death induction in addition to the involvement of Fas signaling in MPCs recovered from IL-2/sTNFRII-expressing tumors.
 View larger version (15K): [in a new window] |
Figure 3. Characterization of MPCs derived from IL-2/sTNFRII-expressing tumors inoculated in gld mice. Tumors were formed by injection of IL-2/sTNFRII-expressing MCA38 cells into gld mice, and MPCs were isolated. (A) CD11b+ MPCs were isolated after 12 days of growth and analyzed for annexin V binding. (B) NO production by MPCs was determined after 48 h of culture in the presence (+) or absence (–) of 100 ng/ml IFN-. (C) Death of MPCs isolated from IL-2/sTNFRII-expressing tumors from gld or control (normal) mice was determined using a LDH release assay.
|
 View larger version (45K): [in a new window] |
Figure 4. Analysis of GM-CSF and M-CSFR expression. (A) Real-time PCR analyses of GM-CSFR (upper left and right histograms) and M-CSFR variants 1 (v1; lower left histogram) and 2 (v2; lower right histogram) in MPCs recovered at Day 12 of tumor growth in wild-type mice. *, Significance as compared with the mock group under the same treatment conditions. (B) Real-time PCR analysis of M-CSF in MPCs recovered at Day 12 of tumor growth in wild-type mice. *, Significance as compared with the mock group under the same treatment conditions. (C) Flow cytometric analysis of M-CSFR (CD115) for total F4/80+ MPCs (left column), immature MPCs (Ly6C–, center column), and mature MPCs (Ly6C+, right column). CD11b+ MPCs were isolated from the indicated tumors grown in wild-type mice for 12 days and stained with antibodies reactive to F4/80, Ly6C, and M-CFSR. Percentages are the percent CD11b+ cells in each tumor type. (D) Flow cytometric analysis of M-CSFR expression on CD11b+ MPCs isolated from IL-2-expressing tumors (MCAIL-2) grown in TNFRI–/– or TNFRII–/– mice. KO, Knockout.
|
Tumor-infiltrating CD4+ T cells affect the maturation of tumor-infiltrating MPCs
As T cell-derived cytokines such as IL-4, IL-13, and IFN- are involved in the activation of MDSCs [10
], we next analyzed tumor-infiltrating CD4+ T cells for their expression of various cytokines. Tumor-infiltrating CD4+ T cells in IL-2/sTNFRII-expressing tumors were skewed to a Th1-dominant condition (Fig. 5A
). In addition, in the draining lymph node, the percentage of CD4+ T cells expressing CXCR3, a chemokine receptor expressed in Th1 cells [21
, 22
], was increased under the IL-2/sTNFRII condition but not under the IL-2-alone condition (unpublished observation). As we could rescue MPCs in IL-2/sTNFRII-expressing tumors from cell death in vitro by inoculating them into gld mice (Fig. 3)
, we examined which cytokines affected MPC maturation under the rescued condition (Fig. 5B)
. We found that IL-4, IL-13, and GM-SCF, but not IL-2, IL-6, IL-7, IL-10, IL-12, IFN-, or TNF-, induced MPC maturation. Furthermore, the cytokines that did not induce maturation also had histogram patterns similar to the control (no cytokine) condition shown in Figure 5B
(data not shown). GM-CSF expression was not detected in MPCs from any tumor subtype, despite its observed expression in tumor-infiltrating CD4+ T cells from all tumor subtypes (Fig. 5C)
.
 View larger version (53K): [in a new window] |
Figure 5. CD4+ T cells are important for the maturation of MPCs. (A) CD4+ TILs, recovered on Day 12 after tumor inoculation in wild-type mice, were analyzed for cytokine mRNA expression by RT-PCR. (B) Effect of cytokines on MPC maturation in vitro. MPCs were isolated from gld mice bearing IL-2/sTNFRII-expressing tumors, cultured in vitro with IL-4, IL-13, or GM-CSF (100 ng/ml each) for 5 days, and analyzed by flow cytometry. (C) Real-time PCR analysis of MPCs and CD4+ TILs. Day 12 tumors were isolated, and cytokine RNA analysis for purified MPCs and CD4+ TILs was performed. *, Significance as compared with the mock group under the same treatment conditions. (D) CD11b+ MPCs were isolated from CD8+ T cell/NK cell-depleted (upper left panel) or CD4+ T cell-depleted (upper right panel) mice bearing IL-2/sTNFR-expressing tumors. CD11b+ MPCs were isolated from PBS-treated (lower left panel) or CD4+ T cell-depleted (lower right panel) mice bearing MCAmock tumors. MPCs were analyzed by flow cytometry for maturation status (expression of Ly6C and F4/80). (E) Immunohistochemical analysis of tumor tissue from MCAmock or IL-2/sTNFR-expressing tumors. CD11b+ cells were FITC-positive (green), and CD4+ T cells were PE-positive (red). (F) Annexin V labeling of MPCs. CD11b+ MPCs were isolated from in vivo CD4+ T cell-depleted, IL-2/sTNFRII-expressing tumors after 12 days of growth and analyzed for annexin V binding after 24 h of culture. (G) Flow cytometric analysis of M-CSFR (CD115) for total F4/80+ MPCs (left column), mature MPCs (Ly6C–, center column), and immature MPCs (Ly6C+, right column). CD11b+ MPCs in CD4+ T cell-depleted, IL-2/sTNFRII-expressing tumors were stained with antibodies reactive with F4/80, Ly6C, and M-CFSR. (H) NO production in tumor-infiltrating MPCs from IL-2/sTNFRII-expressing tumors from in vivo CD4+ T cell-depleted (CD4del) or CD4+ T cell-not depleted, wild-type (WT) mice. Forty-eight hours after culture with or without IFN- treatment, NO concentration in the supernatant was measured by the same assay as in Figure 2A
. As a control macrophage sample, the supernatant of PECs cultured for 48 h without IFN- (PEC) was measured. n.d., Not detected. *, Significance as compared with the IL-2/sTNFRII (CD4del) group under the same in vitro treatment conditions. (I) Analysis of inducible NO synthase (iNOS) and arginase I levels of freshly isolated, IL-2/sTNFRII-expressing, tumor-derived MPCs (Day 12 tumors). Comparison is between CD4+ T cell-depleted (CD4 del) and nondepleted wild-type conditions. Freshly isolated PECs were used for monocyte/macrophage control. Representative data from duplicated experiments are shown. Enzyme expression levels were standardized by each sample’s GAPDH level by calculating the intensity of the image by Image software.
|
We next examined the effect of CD4+ T cells on cell viability, CD115 expression, and NO production in MPC from IL-2/sTNFRII-expressing tumors (Fig. 5F 5G 5H) . When CD4+ T cells were depleted, most of the MPCs stained with annexin V (Fig. 5F) , in contrast with freshly isolated cells, which did not (data not shown). This cell death phenomenon was also observed in MPCs in the CD4-nondepleted condition (Fig. 2E) . CD4 depletion in vivo did not lead to CD115 expression on MPCs (Fig. 5G) . Furthermore, MPCs had no detectable NO production, in contrast with cells from the nondepleted condition (Fig. 5H) .
As a result of the in vitro cell death of MPCs from IL-2/sTNFRII-expressing tumors, we were unable to directly demonstrate the effect of in vivo depletion of CD4+ T cells. Therefore, we checked iNOS and arginase I expression by MPCs freshly isolated from IL-2/sTNFRII-expressing tumors under the CD4-depleted and nondepleted conditions. As shown in Figure 5I , we observed decreased iNOS and increased arginase I expressions in the CD4-depleted compared with the nondepleted condition. These data suggest that CD4+ T cells are not only essential for the maturation of tumor-infiltrating MPCs but are also important in affecting their function, especially in the immunotherapeutic condition.
DISCUSSION
In the present study, we showed that tumor-infiltrating monocyte/macrophage lineage cells could be classified as MDSCs based on their cell-surface phenotypes. Furthermore, the most important characteristic of MDSCs is their ability to suppress T cell proliferation and function [2 ], and we also demonstrated that tumor-infiltrating CD11b+ cells suppressed T cell proliferation (Fig. 1D) . We previously showed that tumor-infiltrating CD11b+F4/80+cells can induce T cell apoptosis [18 ]. Sica and Bronte [9 ] recently summarized that Gr-1+ MDSCs mature into a Gr-1–F4/80+ phenotype at the tumor site. Therefore, although the tumor-infiltrating MPCs in our model were Gr-1low, these cells could be referred to as MDSCs, as they demonstrated immunosuppressive properties (Fig. 1D) . However, such a designation runs the risk of over-reading the definition of a MDSC.
Until recently, tumor-infiltrating MPCs were typically referred to as TIMs, a MPC subset that displays anti- and protumor characteristics [23 ]. The MPCs studied here displayed immunosuppressive properties (Fig. 1D) , possibly distinguishing them as immunosuppressive TIMs. However, we are hesitant to designate the cell subset identified here as tumor-infiltrating "macrophages." The CD11b+ population of MPCs contained a prominent monocytic population, particularly under the immunotherapeutic condition (Fig. 1D 1E 1F) . In addition, our recent report indicates that tumor-infiltrating MPCs cannot be defined as tumor-associated macrophages (TAMs), which are defined as typical M2-type macrophages, as the cells displayed pleiotropic M1 and M2 characteristics [24 ]. For these reasons, we designated the cells of interest in the present study as MPCs, rather than as MDSCs, TAMs, or TIMs.
In terms of the immunosuppressive abilities of MPCs from IL-2/sTNFRII-expressing tumors, we attempted to perform a MDSC/T cell coculture experiment similar to that for Figure 1D . We found that the percentage of T cells in the proliferating phase in MPCs from mock tumors or IL-2/sTNFRII-expressing tumors was 24.24 ± 3.91% or 40.89 ± 2.51%, respectively, and the percentage of T cells in the proliferating phase when cultured alone was 44.09 ± 1.33% (data not shown). Although this result makes it appear as though MPC immunosuppression was reduced in MPCs derived from immunologically modified tumors compared with those from mock tumors, as we show in Figures 2 3 4 5 , MPCs from immunologically manipulated tumors did not survive ex vivo. Therefore, we cannot conclude that the immunological manipulation of tumors decreased MPC-induced immunosuppression. Rather, the inability of MPCs from immunologically manipulated tumors to survive ex vivo made it difficult for us to evaluate their functional properties. What we can conclude is that MPCs from immunologically modified tumors do not alter their cell-surface characteristics similar to other MPCs.
In the present study, we concluded that the suppressed maturation of tumor-infiltrating MPCs in IL-2/sTNFRII-expressing tumors was a result of an alteration of the Th balance of CD4+ TILs into a Th1-dominant state. To adequately make this assertion, we must first address the possibility that the reduced size of IL-2/sTNFRII-expressing tumors led to less-mature MPCs. CD4+ T cell-depleted, IL-2/sTNFRII-expressing tumors had a 3.6x greater mass per mouse at Day 12 of tumor collection compared with nondepleted tumors (data not shown). This, in addition to evidence that MPC maturation was suppressed in depleted tumors in vivo (Fig. 5E) , allowed us to conclude that tumor size and MPC maturation were not inter-related.
T cell-derived, cytokine-related activation is related to the classical (M1) and alternate (M2) activation types. In the presence of GM-CSF, LPS, or IFN-, monocytes differentiate into M1 macrophages that play a role in tumor suppression. Conversely, in the presence of M-CSF, IL-4, IL-13, IL-10, or corticosteroids, monocytes differentiate into M2 macrophages and assist in tumor promotion [25
]. In this regard, the IL-2/sTNFRII-expressing tumor microenvironment is an M1-type cytokine-dominant condition. That is, in the IL-2/sTNFRII-expressing tumor microenvironment, GM-CSF is supplied from CD4+ T cells but not from MDSCs (Fig. 5C)
and binds to GM-CSFRs expressed by MDSCs (Fig. 4A)
. MDSCs were unresponsive to M-CSF as a result of a lack of M-CSFR expression (Fig. 4A
4B
4C
4D)
, and the reduced generation of IL-13 by CD4+ T cells resulted in the suppressed maturation of MPCs (Fig. 5A)
. Taken together, these results show that the Th1-skewed tumor microenvironment resulted in the prevention of MPC maturation. Moreover, the IL-2-alone tumor microenvironment did not differ from the MCA38/mock condition in terms of Th balance (Fig. 5A)
, and the only observed difference was a reduced expression of M-CSFR by MPCs (Fig. 4C)
. These findings indicate that reduced M-CSFR expression leads to the altered maturation of MPCs, even when the cytokine supply is sufficient. However, we would like to stress that even in a Th1-dominant tumor microenvironment, MPC maturation occurred as a result of CD4+ T cell depletion; neither CD8+ T cells nor NK cells could completely abolish MPC maturation or alter their function (Fig. 5D
5E
5F
5G
5H
5I)
.
Several previous studies have shown that tumor-infiltrating monocytes/macrophages are M2-polarized [12 , 26 , 27 ]. However, recently, Biswas et al. [28 ] showed that M1- and M2-type markers were simultaneously expressed on identical macrophages. As described earlier, we have also shown an instance where tumor-infiltrating monocyte/macrophages were not classified as typical M2-type TAMs [24 ]. Therefore, the findings of Biswas et al. [28 ] and our results indicate that M1- and M2-type stimuli can induce MPC maturation. As Stout et al. [29 , 30 ] recently demonstrated, macrophages have plastic cell characteristics that can be adapted to each stimuli when the same cells are sequentially stimulated with M1- and M2-type cytokines/reagents. Indeed, as we demonstrated in Figure 5B , Th1-skewed, IL-2/sTNFRII-expressing, tumor-derived MPCs can be matured with M2 (IL-4 and IL-13)- and M1-type (GM-CSF) cytokines. Our findings therefore suggest that MPCs are pleiotropic and possess plastic characteristics.
The proliferation and differentiation of monocytes/macrophages usually correlate with M-CSF and GM-CSF [20 ]. Indeed, in our model, we interpreted MPC capture of GM-CSF from tumor-infiltrating CD4+ T cells through the presence of GM-CSFR (Figs. 4 and 5) . Under normal conditions, M-CSF works in an autocrine manner; M-CSF did not function in this study, as its receptor was not found on MPCs from IL-2/sTNFRII-expressing tumors (Fig. 4) . Therefore, we conclude that the lack of M-CSFR on the MPC cell surface and T cell-derived GM-CSF (Fig. 4A 4B 4C) resulted in the death of MPCs from IL-2/sTNFRII-expressing tumors in gld mice (Fig. 3C) . Indeed, M-CSF shortages prevent monocytes/macrophages proliferation in vitro [31 ], and op/op mice, which have a genetic mutation in the M-CSF gene, have dramatically decreased systemic infiltration of macrophage lineage cells in their liver, spleen, and bone marrow [32 ].
We demonstrated that the expression of TNFRI and TNFRII is vital for the induction of M-CSFR expression in IL-2-based immunotherapy (Fig. 4D)
. However, this phenomenon was not observed when TNF alone was blocked (Fig. 4C)
, indicating that loss of the TNF signal was not sufficient for a loss in M-CSFR expression. IL-4 is important for the induction of M-CSFR expression in bone marrow macrophages [33
], and IL-4R is the common receptor for IL-4 and IL-13 [10
]. Our research indicates that the major inducing cytokine for M-CSFR expression is IL-13, as CD4+ TILs rarely expressed IL-4 compared with IL-13 (Fig. 5A)
. Therefore, a Th1-skewed microenvironment in IL-2/sTNFRII-expressing tumors could explain why MPCs did not express M-CSFR.
A schematic summary of the present study is shown as Figure 6
. In wild-type tumor growth, MPCs do not express M-CSFR until approximately Day 5 of tumor growth; by this time, CD4+ T cells have not yet infiltrated the tumor (data not shown). Although the infiltration of CD4+ T cells is small, upon secretion of IL-13 from CD4+ TILs, MPCs express M-CSFR, which permits cell survival/proliferation in an autocrine manner. In addition, CD4+ TIL-derived cytokines IL-4, IL-13, and GM-CSF partially contribute to MPC maturation. GM-CSF could also affect MPC survival, but we did not observe any direct evidence for this in our study. On the other hand, in tumors that expressed IL-2/sTNFRII, which skews the CD4+ TIL phenotype toward Th1 and induces marked T cell infiltration, MPCs received IFN- and GM-CSF, but not IL-4 or IL-13, from CD4+ TILs. Thus, the absence of IL-4 and IL-13 in the Th1-dominant and T cell infiltration-rich environment emphasizes the suppression of MPC maturation compared with the wild-type tumor microenvironment.
 View larger version (42K): [in a new window] |
Figure 6. Schematic summary of the current study. A detailed explanation of this scheme is described in Discussion.
|
ACKNOWLEDGEMENTS
This work was supported, in part, by Grants-in-Aid for Scientific Research to M. S. (Grants 16591437 and 18390393) from the Japan Society for the Promotion of Science.
FOOTNOTES
1 These authors contributed equally to this work. 
Received November 6, 2007; revised May 26, 2008; accepted May 29, 2008.
REFERENCES
plus tumor necrosis factor- synergistically reduce suppressor cell activity Int. J. Cancer 46,245-250[Medline]
B and enhanced IRF-3/STAT1 activation) Blood 107,2112-2122
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
