Published online before print February 19, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1,1,2,
,
* Departments of Oral and Maxillofacial Sciences,
Immunopathology,
Surgical Oncology, and
Neurosurgery, Gifu University Graduate School of Medicine, Gifu, Japan; and
|| Department of Preventive Medicine, University of Debrecen, Hungary
2Correspondence: Gifu University Graduate School of Medicine, Dept. of Immunopathology, 2S29 School of Medicine, Main Building, 1-1 Yanagido, Gifu-city, Gifu 501-1194 Japan. E-mail: saio{at}gifu-u.ac.jp
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
+ phenotype. In addition, the cells expressed CX3CR1 and CCR2 simultaneously, which are the markers of an inflammatory monocyte. The MDSCs expressed CD206, CXCL10, IL-1β, and TNF- mRNAs. They also simultaneously expressed CXCL10 and CD206 proteins, which are typical, classical (M1) and alternative (M2) macrophage activation markers, respectively. Peritoneal exudate cells (PECs) strongly expressed CD36, CD206, and TGF-β mRNA, which is characteristic of deactivated monocytes. The MDSCs also secreted TGF-β, and in vitro culture of MDSCs and PECs with anti-TGF-β antibody recovered their ability to secrete NO. However, as a result of secretion of proinflammatory cytokines, MDSCs could not be categorized into deactivated monocyte/macrophages. Thus, tumor-infiltrating MDSCs bear pleiotropic characteristics of M1 and M2 monocytes/macrophages. Furthermore, CD206 expression by tumor-infiltrating MDSCs appears to be regulated by an autocrine mechanism that involves TGF-β.
Key Words: myeloid-derived suppressor cells • TGF-β
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
+ inflammatory monocytes [8
]. MDSCs within tumors also have a potent, immunosuppressive phenotype [10
] and produce reactive nitrogen compounds (peroxynitrites, among others) that have recently been shown to nitrosylate the CD8 and TCR molecules on T cells, thereby inhibiting T cell reactivity [11
].
Tumor-associated macrophages (TAM) are the cells that had been considered to be alternatively activated "M2"-type macrophages [12
, 13
], which are one of the two monocyte/macrophage types in the classification system of Mantovani et al. [14
]. This system evolved after the Th1/Th2 concept was proposed, and it proposes that macrophages can be similarly divided into classically activated "M1" macrophages, which are activated/differentiated by LPS and the Th1 cytokine IFN-
, and alternatively, activated M2-type macrophages, which are activated/differentiated by glucocorticoid, IL-10, and Th2 cytokines IL-4 and IL-13 [15
]. However, Van Ginderachter et al. [16
] suggested that this classification system may be too simplistic. Indeed, a recent review by Gordon [17
] has classified monocytes/macrophages into another three categories in addition to the M1 and M2 categories in terms of the stimuli that induce their activation/differentiation, namely, innate activation via microbial stimulus, humoral activation via Fc and complement receptors, and deactivation by TGF-β, M-CSF, IL-10, IFN-/β, or glucocorticoid receptors.
In the present study, we sought first to determine whether tumor-infiltrating myelomonocytoid cells can be defined as MDSCs. We found that indeed, they bear known markers and functions of MDSCs. Thus, we refer to these cells hereafter as tumor-infiltrating MDSCs. We also examined whether these cells can be defined as alternatively activated monocytes/macrophages such as TAM. However, we found that tumor-infiltrating MDSCs cannot be classified by current monocyte/macrophage classification systems, as they simultaneously bear some but not all of the markers of several monocyte/macrophage categories.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell lines
The murine colon adenocarcinoma (MCA)38 cell line was originally provided by Dr. Yang Liu (Ohio State University, Columbus, OH, USA), and the murine glioma cell line GL261 was originally provided by Dr. Elizabeth Newcomb (Dept. of Pathology, New York University School of Medicine, New York, NY, USA). The cells were cultured at 37°C in a humidified 5% CO2 atmosphere in RPMI 1640 or D-MEM (Invitrogen Life Technologies, Carlsbad, CA, USA) containing 10% FCS, L-glutamate (Invitrogen Life Technologies), and penicillin-streptomycin (Invitrogen Life Technologies).
Implantation of tumor cells
All mice were anesthetized by an intramuscular injection of pentobarbiturate (2.5 mg/mouse, Dainippon-Sumitomo, Osaka, Japan) and inoculated by s.c. injection in the left flank with 3 x 106 MCA or GL261 cells.
Preparation of tumor-infiltrating cells
Mice were killed 14 or 21 days after tumor implantation, and the tumor-infiltrating cells were prepared as described previously [18
]. Briefly, the tumors was collected and minced into small pieces before incubation for 15 min at 37°C with the following enzymes dissolved in HBSS: collagenase Type I (0.05 mg/ml), collagenase Type IV (0.05 mg/ml), hyaluronidase (0.025 mg/ml, all from Sigma Chemical Co., St. Louis, MO, USA), and DNase I (0.01 mg/ml) and soybean trypsin inhibitor (0.2 trypsin inhibitor unit/ml, both from Roche Diagnostics, Nutley, NJ, USA). The digested cells were harvested, and the RBCs were lysed by using hypotonic buffer (0.155 M NH4Cl, 0.1 mM EDTA, 10 mM KHCO3) for 1 min. CD4+ and CD8+ cells were depleted by applying anti-CD4 and anti-CD8a magnetic beads, after which CD11b+ cells were isolated by using anti-CD11b magnetic immunobeads according to the manufacturer’s instructions (MACS, Miltenyi Biotec, Berdish-Gladbach, Germany).
Preparation of peritoneal exudate cells (PECs)
PECs were collected from B6 mice 4 days after i.p. administration of thioglycolate.
Flow cytometric analysis
Prior to flow cytometric analysis, all cells were preincubated with 10 µg/ml anti-CD16/32 antibody (4.2G2, PharMingen, San Diego, CA, USA) at 4°C for 30 min before staining with the following specific antibodies. To analyze CD11b+ cell surface antigen expression, allophycocyanin (APC)-conjugated anti-CD11b (M1/70, PharMingen) and FITC-, PE-, or cychrome-conjugated anti-F4/80 were used. Other antibodies used were FITC-conjugated anti-Ly6C (ER-MP20, BMA Biomedicals, Augst, Switzerland), PE-conjugated anti-Ly6G (Gr-1; RB6-8C5, PharMingen), biotin-conjugated anti-IL-4R (polyclonal goat IgG, R&D Systems, Inc., Minneapolis, MN, USA), FITC-conjugated CXCL10 (clone 134013, R&D Systems, Inc.), PE-conjugated anti-CD206 (clone MR5D3, PharMingen), and PE-conjugated anti-CD11c (clone HL3, PharMingen). FITC- or PE-conjugated hamster IgG (PharMingen), rat IgG1 (PharMingen), rat IgG2a (PharMingen and Serotec Ltd., Oxford, UK), rat IgG2b (PharMingen), and biotin-conjugated control goat IgG (R&D Systems, Inc.) served as control antibodies. Anti-CX3CR1 (polyclonal rabbit antibody, ProSci, Poway, CA, USA) and control rabbit IgG (Abcam Japan, Tokyo) were labeled by using a FITC-conjugation kit (Dojin, Tokyo, Japan), and anti-CCR2 (polyclonal rabbit antibody, Abcam Japan) and control rabbit IgG (Abcam Japan) were labeled by using a PE-conjugation kit (Dojin) and used for analysis. All antibodies were used at 10 µg/ml. The cells were incubated with the antibodies for 30 min at 4°C and washed with PBS. If the biotin-conjugated antibody was used, the samples were subsequently stained with avidin-conjugated PE (PharMingen) for 30 min at 4°C and then washed with PBS. The samples were fixed with 1% paraformaldehyde/PBS and analyzed by using a FACSCalibur flow cytometer and CellQuest software (Becton Dickinson Japan, Tokyo).
Multiple PCR analysis
Total RNA was purified by using Trizol (Invitrogen Life Technologies), and 600 ng total RNA was used for reverse transcription, using Superscript III RT (Invitrogen Life Technologies). For multiple PCR analysis, 100 ng cDNA samples or positive control DNA mixtures were mixed with a multiple primer pair mixture, buffer, and Taq polymerase, which were provided by the Multiplex PCR kits for mouse chemokine receptors (CCR Set 1 and CCR Set 2, Maxim Biotech, Inc., South San Francisco, CA, USA), and then subjected to PCR using the conditions indicated in the manufacturer’s instruction manual. The amplified DNAs were analyzed by 5% acrylamide gel electrophoresis followed by ethidium bromide staining.
Real-time PCR analysis
Total RNA was purified as described above, and 600 ng was subjected to reverse-transcription using Superscript III RT. For real-time PCR analysis, 5 µl 20x diluted cDNA sample, 1 µl each 10 µM upper and lower primer, 3 µl PCR grade water (Roche Diagnostics), and 10 µl 2x concentrated Syber Green and Taq enzyme-premixed reaction mixture (SYBR® Premix Ex TaqTM, Takara Bio, Japan) were used. The sequences of the primer pairs used in this analysis are indicated in Table 1
. The reaction conditions consisted of one 5-min cycle at 95°C, followed by 45 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 10 s, which were concluded by the melting curve analysis process. The reaction and analysis were performed by Light Cycler (Roche Diagnostics).
 View larger version (64K): [in a new window] |
Figure 1. Characterization of tumor-infiltrating MDSCs. B6 mice were inoculated s.c. with 3 x 106 MCA38 tumor cells, and the resulting tumors were excised 14 days later. The cells infiltrating the tumors were collected, and the MDSCs were purified by selecting CD11b+ cells. (A and B) Effect of tumor-infiltrating MDSCs on the growth of activated CD8+ T cells. Splenic CD8+ T cells from a GFP mouse (A) and a normal mouse (B) were activated with anti-TCR antibody for 24 h. The normal mouse-derived CD8+ T cells were then labeled with CFSE prior to cocultivation with the MDSCs. PECs were prepared from B6 mice 4 days after i.p. injection with thioglycolate and served as a control. The MDSCs and PECs were cultivated alone for 4 h and cocultured with the activated CD8+ T cells for 36 (A) or 24 h (B), and the T cells were examined by fluorescence microscopy (A; arrows indicate apoptotic CD8+ T cells) or labeled with anti-CD3 and anti-CD8 antibodies and Via-Probe, after which the CFSE levels were determined by flow cytometric analysis (B). (C) Flow cytometric analysis of the tumor-infiltrating MDSCs. Regions 1- and 2 (R1 and R2)-gated cells (CD11b+ and Via-Probe–) were analyzed for the expression of CD206 and the MDSC markers CD11c, Gr-1, and IL-4R. Representative data of at least three different preparations, all of which showed similar phenotypes, are shown. F4/80 and Gr-1 staining is shown on the upper light part of the panel to indicate size of the monocyte/macrophage (F4/80+Gr-1low) cell population in our model. SSC, Side-scatter; FSC, forward-scatter. (D) May-Giemsa staining of the tumor-infiltrating MDSCs. The arrow indicates the large cytoplasm and vacuoles of the MDSCs. Representative data from at least two independent experiments, which yielded similar data, are shown. (E) Appearance of tumor-infiltrating MDSCs from MCA38 tumors (left) and GL261 tumors (right) after in vitro culture in serum-free medium for 5 h. Representative data of at least two independent experiments, which yielded similar data, are shown.
|
In vitro coculture experiment
Activated, splenic CD8+ T cells and tumor-infiltrating MDSCs were cocultured as described in our previous report [18
]. Briefly, Day 14 tumor-infiltrating MDSCs were purified as described above, and 1.5 x 106 cells/cm2 were cultured alone for 4 h. Twenty-four hours before coculture, CD8+ T cells were purified from a GFP transgenic or normal mouse spleen by using CD8+ magnetic immunobeads, according to the manufacturer’s instructions (Miltenyi Biotec) and then stimulated with plate-bound anti-TCR-β antibody (H57-597, PharMingen). The activated T cells were cocultured with the MDSCs at 1 x 106 cells/cm2. The GFP mouse-derived T cells were cocultured for 36 h and then observed under a fluorescent microscope. As a control, PECs were used instead of tumor-infiltrating MDSCs. To quantitate the 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 (1x106) were plated onto 24-well plates in 2 ml complete medium and incubated with or without 0.5 ng/ml recombinant IFN- (PharMingen) for 24 h with medium alone or chicken control IgY or anti-TGF-β antibody (R&D Systems, Inc.) before collection of the supernatants. Supernatant samples (0.1 ml) were mixed with 0.1 ml Greiss reagent (Sigma Chemical Co.) in 96-well plates for 15 min, and then absorbance at 540 nm was determined (Benchmark® Microplate reader) and analyzed using Microplate Manager III software (both from Bio-Rad Laboratories, Richmond, CA, USA) and with sodium nitrite as a standard.
ELISA
The 24-h culture supernatants of MDSCs and PECs were subjected to an ELISA for IL-1β and TGF-β by using ELISA kits (R&D Systems, Inc.), according to the manufacturer’s instructions.
SDS-PAGE and immunoblotting
CD11b+ cells 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-X100, 1 mM EDTA, 1 mM EGTA, 1 mM PMSF, 5 mM iodoacetamide, 1 mM Na3VO4, and protease inhibitor cocktail (Sigma Chemical Co., P-8340)]. The supernatants were 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 anti-NOS II (rabbit polyclonal, Upstate, Charlottesville, VA) or anti-rat/mouse Arg I (Clone 19, BD Biosciences, Rockville, MD, USA) antibody, diluted in Can Get SignalTM Immunoreaction Enhancer Solution 1 for primary antibody (Toyobo Co., Ltd., Osaka, Japan). After washing, the membrane was stained with 16 ng/ml HRP-conjugated goat anti-rabbit IgG or anti-mouse IgG (both from Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA) diluted in Can Get SignalTM Immunoreaction Enhancer Solution 2 for 1 h. The blots were developed with ECL (Amersham Pharmacia Biotech, Buckinghamshire, UK), and the images were captured by the Cool Saver Lumino-capture system (Model AE-6955, ATTO, Tokyo, Japan) before analysis using CS Analyzer software. 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 16 ng/ml of HRP-conjugated goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc.).
Statistics
For statistical analysis, we used the Student’s-t test. P values less than 0.05 were considered to indicate a significant difference.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Flow cytometric analysis of the tumor-infiltrating MDSCs revealed they had the CD11b+CD11c+Gr-1+IL-4R+ MDSC phenotype (Fig. 1C)
, as defined by Gallina et al. [8
]. In addition, the MDSCs expressed CD206 (Fig. 1C)
, which is a M2-type monocyte/macrophage marker [19
]. As MDSCs belong to the myeloid lineage, which consists of granulocyte and monocyte lineage cells [1
], we examined the morphology of the tumor-infiltrating MDSCs of our model. This revealed they had a large cytoplasm with a vacuole and a kidney-shaped, twisted, or horseshoe-shaped nucleus (Fig. 1D)
. In addition, flow cytometric analysis revealed that of the CD11b+ cells, less than 5% (3.72% in the representative experiment shown in Fig. 1C
) were F4/80intGr-1high, and over 90% (91.3% in the representative experiment shown in Fig. 1C
) were F4/80+Gr-1low. Thus, the majority of the cells were monocyte-macrophage lineage cells. Upon in vitro culture, the MDSCs adhered to the plastic plate and had an oval or spindle shape (Fig. 1E
, left). This confirms that the cells belong to the monocyte-macrophage lineage rather than to the granulocyte lineage. This was also observed for MDSCs harvested from 14-day-old tumors induced in B6 mice by inoculation with the murine glioma cell line GL261 (Fig. 1E
, right). Thus, tumor-infiltrating, myelomonocytoid cells have a MDSC phenotype and monocyte/macrophage-like properties.
Tumor-infiltrating MDSCs bear CCR2+ and CX3CR1+ inflammatory monocyte characteristics
We further characterized the tumor-infiltrating MDSCs by comparing their chemokine receptor expression pattern. According to the recent classification of murine monocytes by Geissmann et al. [20
], monocytes can be classified into two groups, namely, CD11b+Gr-1high CX3CR1+CCR2+ inflammatory (from the blood) monocytes and CD11b+Gr-1low CX3CR1+CCR2– resident monocytes [21
]. Analysis of the chemokine receptors expressed by GL261 tumor-infiltrating MDSCs revealed they strongly expressed CCR2, CCR5, CX3CR1, and CCR7 mRNAs (Fig. 2A
). Flow cytometric analysis also showed that these cells simultaneously expressed CX3CR1 and CCR2 protein (Fig. 2B)
as well as CCR5 protein (data not shown). On the one hand, this confirms that these cells are monocytoid [19
]. On the other hand, the chemokine receptor expression profile of tumor-infiltrating MDSCs suggests that they originated from blood monocytes.
 View larger version (38K): [in a new window] |
Figure 2. Further characterization of tumor-infiltrating MDSCs. Multiple PCR analysis of chemokine receptor expression (A) and flow cytometric analysis of CX3CR1 and CCR2 expression (B) by GL261 tumor-infiltrating MDSCs obtained 14 days after tumor cell inoculation. (B) The CD11b+Via-Probe– population was examined. (C) Flow cytometric analysis of MCA38-derived, tumor-infiltrating MDSCs on Day 21 after tumor inoculation. These data can be compared with the MDSC data in Figure 1C
.
|
+ (Fig. 2C)
. As macrophages were recently, narrowly classified as CD11b+CD11c–Ly-6CintF4/80high cells [22
], this suggests that only few of the tumor-infiltrating MDSCs had matured into macrophages, regardless of when after tumor inoculation they were obtained.
Tumor-infiltrating MDSCs cannot be strictly categorized as M1- or M2-type monocytes/macrophages
Our observations described above (Figs. 1
and 2)
suggest that tumor-infiltrating MDSCs bear inflammatory, monocyte characteristics. The FACS analysis also revealed that they expressed CD206 (Fig. 1C)
, which is a known marker of M2 cells [19
, 23
]. To test whether these cells bear more characteristics of M2 cells, we subjected MDSCs from Day 14 MCA38- or GL261-derived tumors to real-time PCR analysis using primer pairs (Table 1)
for molecules that have been used to classify monocytes/macrophages into the M1 and M2 subtypes [19
, 23
]. Of the M2-type markers CCL22, CCL17, IL-10, and CD206, the MCA38 and GL261 tumor-derived MDSCs expressed CD206 mRNA (Fig. 3A
). The GL261 tumor-derived MDSCs were also weakly positive for CCL17 mRNA (Fig. 3A)
. Of the M1-type markers, both MDSC preparations expressed IL-1β, TNF-, and CXCL10 but not CXCL11 or IL-12 mRNAs (Fig. 3A)
. In addition, flow cytometric analysis revealed that the MDSCs simultaneously expressed CXCL10 and CD206 protein (Fig. 3B)
. Thus, MDSCs bear some M1 markers and some M2 markers, which indicates they are pleiotropic in terms of the M1 and M2 classification system.
|
View this table: [in a new window] |
Table 1. Primer Pairs Used in Real-Time PCR Analysis
|
 View larger version (28K): [in a new window] |
Figure 3. Analysis of M1 and M2 marker expression by Day 14 MDSCs and PECs. (A) Real-time PCR analysis of the expression by MDSCs of molecules used to classify monocytes/macrophages. MCA38 and GL261 tumor-infiltrating MDSCs were subjected to total RNA purification, RT, and real-time PCR analysis for the monocyte/macrophage markers shown in Table 1
. (B) Flow cytometric analysis of GL261 tumor-infiltrating MDSCs. The CD11b+Via-Probe– population was examined for the expression of CXCL10 and CD206. (C) Real-time PCR analysis of the expression by PECs of molecules used to classify monocytes/macrophages, as described in A.
|
MDSCs strongly produce TGF-β1, and tumor cells weakly produce TGF-β1, -2, and -3
TGF-β is well-known to suppress the immune system [24
]. Consequently, we examined whether tumor-infiltrating MDSCs produce any of the three TGF-β subtypes. Relative to GAPDH mRNA expression, MCA38 and GL261 tumor-derived MDSCs expressed TGF-β1 at relatively high levels (Fig. 4A
). The MCA38 and GL261 tumor cells themselves weakly produced TGF-β1, -2, and -3 (Fig. 4B)
. In contrast, PECs expressed TGF-β1 at high levels and TGF-β2 at low levels (Fig. 4C)
. These data suggest that tumor-infiltrating MDSCs are the main producers of TGF-β in the tumor microenvironment.
 View larger version (11K): [in a new window] |
Figure 4. Real-time PCR analysis of TGF-β expression by tumor-infiltrating MDSCs, tumor cells, and PECs. (A–C) Day 14 tumor-infiltrating MDSCs (A), cultured tumor cells (B), and PECs (C) were subjected to total RNA purification, RT, and real-time PCR analysis.
|
(data not shown). These data are supported by Figure 3A
, which shows MDSCs express IL-1β and TNF- mRNAs. In contrast, PECs did not express IL-1β and TNF- mRNAs (Fig. 3C)
.
 View larger version (34K): [in a new window] |
Figure 5. PECs are deactivated monocytes/macrophages, but tumor-infiltrating MDSCs are activated monocytes/macrophages whose TGF-β production affects their expression profiles. (A) ELISA assay for IL-1β and TGF-β. Day 14 MCA38 tumor-infiltrating MDSCs and PECs (1x106 cells/ml) were cultured for 24 h, and the supernatant was subjected to ELISA analysis. (B) Effect of TGF-β neutralization on the nitrite production of tumor-infiltrating MDSCs and PECs. Day 14 MCA38 or GL261 tumor-infiltrating MDSCs were cultured for 24 h in serum-free conditions with or without IFN- in the presence or absence of anti-TGF-β-neutralizing antibody or control chicken IgY. The nitrite concentrations in the medium were then measured. n.d., Not detected. (C) Western blotting analysis of inducible NO synthase (iNOS) and arginase I levels in Day 14 tumor-infiltrating MDSCs and PECs. The cells were cultured for 24 h with medium alone (–) or in the presence or absence of anti-TGF-β-neutralizing antibody or control chicken IgY and then subjected to Western blotting analysis. (D) Effect of TGF-β neutralization on MDSC expression of IL-4R, CXCL10, and CD206. Day 14 tumor-infiltrating MDSCs were cultured with anti-TGF-β antibody or control IgY for 48 h. The cells were then analyzed for IL-4R, CXCL10, and CD206 expression by flow cytometry. To show the effect of anti-TGF-β antibody treatment on CD206 expression levels, the control antibody staining for CD206 and the CD206 histograms were overlaid (lower panel). The percentage of CD206+ cells is indicated on each panel. In the upper panel, the CD206 staining patterns of cells treated with control IgY and anti-TGF-β antibodies were overlaid.
|
stimulation to recover [25
]. Consequently, we examined the effect of anti-TGF-β antibody treatment on the ability of GL261-derived MDSCs to produce nitrite under serum-free conditions. Indeed, the nitrite production of the MDSCs recovered after anti-TGF-β antibody treatment, even in the absence of IFN- stimulation (Fig. 5B)
. The NO production of MCA38 tumor-derived MDSCs and PECs also recovered after anti-TGF-β antibody treatment, although only when they were treated with IFN- (Fig. 5B)
. As the iNOS and arginase I levels are an important feature of M1 and M2 monocytes/macrophages, respectively [14
], we subjected MDSCs to Western blot analysis of iNOS and arginase I levels. The MDSCs expressed iNOS and arginase I at higher levels than PECs, and anti-TGF-β antibody treatment enhanced their iNOS expression while decreasing their arginase I expression (Fig. 5C)
.
We also examined the effect of anti-TGF-β antibody treatment on the cell surface phenotype of tumor-infiltrating MDSCs (Fig. 5D)
. The anti-TGF-β antibody suppressed CD206 expression but had no effect on CXCL10 and IL-4R expression. Moreover, the in vitro treatment of tumor-infiltrating MDSCs with IFN- or IL-4 did not alter the MDSC expression levels of CD206, CXCL10, or IL-4R (data not shown).
In conclusion, it appears that tumor-infiltrating MDSCs are unique monocyte/macrophages that do not fall into any known category. They simultaneously express several (but not all) M1 (iNOS, IL-1β, TNF-, and CXCL10) and M2 (CD206 and arginase I) markers. They resemble the inflammatory monocyte/macrophage population. They also resemble deactivated monocyte/macrophages, as they are CD206+ and CD36+, and their NO production recovered after blocking TGF-β, yet they are not deactivated, as they produce inflammatory cytokines such as IL-1β and TNF-, which indicates an M1-like activation status [17
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
[1
, 8
]. In our study here, we showed that tumor-infiltrating CD11b+ cells indeed express CD11c, Gr-1, and IL-4R. On the basis of the recent definition of macrophages as CD11b+CD11c–Ly-6CintF4/80high cells [22
], these cells are not macrophages. Instead, they resemble the inflammatory monocytes described by Strauss-Ayali et al. [21
], who recently reported that monocytes could be categorized into two populations, namely, CCR2+CX3CR1+ blood-derived, inflamed monocytes and CCR2–CX3CR1high resident monocytes. The tumor-infiltrating MDSCs fell into the inflammatory monocyte category because of the expression pattern of CCR2 and CX3CR1 (Fig. 2A
and 2B)
. IL-4R expression is an important characteristic of MDSCs, as Gallina et al. [8
], who also showed that MDSCs can be categorized as inflammatory monocytes, found that the suppressive phenotype of MDSCs requires the expression of IL-4R. Indeed, in the present study, we could demonstrate that tumor-infiltrating MDSCs are IL-4R+ inflammatory monocytes/macrophages and show immunosuppressive property (Fig. 1A
1B
1C)
.
We then asked whether tumor-infiltrating MDSCs could be classified as alternatively activated monocyte/macrophages but found that these cells expressed some, but not all, classical activation markers along with some, but not all, alternative activation markers. There are two possible explanations for this eclectic phenotypic profile. Either all tumor-infiltrating MDSCs bear the M1 and M2 characteristics simultaneously (i.e., they are pleiotropic), or the infiltrating MDSC populations are heterogenous mixtures of cells with M1 or M2 characteristics. We showed the former is the case, as the tumor-infiltrating MDSC populations only expressed three of the five M1 (IL-1β, TNF-, and CXCL10)- and only three of the five M2-type markers (CCL17, CD206, and CD36). Moreover, CXCL10 and CD206 were shown by flow cytometric analysis to be expressed simultaneously on the same cells (Fig. 3B)
. Thus, it appears that tumor-infiltrating MDSCs are pleiotropic in character, as they simultaneously bear M1 and M2 characteristics. Supporting this is a study by Biswas et al. [26
], who reported that tumor-infiltrating monocyte/macrophages simultaneously express M1 and M2 markers. It remains possible that RT-PCR and sample artifacts may account for our failure to detect the expression by tumor-infiltrating MDSCs of the remaining M1 and M2 markers. However, we found that the tumor-infiltrating MDSCs we isolated were highly pure (Fig. 1C)
. Moreover, in our analysis of the RT-PCR data, we expressed the mRNA levels of the M1 and M2 genes relative to GAPDH expression. Thus, we believe our M1/M2 mRNA expression data are quite reliable.
Figures 3
4
5
confirmed that tumor-infiltrating MDSCs have a unique character, as they not only expressed CD206 and CD36 (which are deactivation markers as well as being part of the M2 marker profile), but also, their NO production was down-regulated by TGF-β. Similarly, the PECs also expressed CD206 and CD36, and their NO production was down-regulated by TGF-β. These observations suggest that MDSCs, like PECs, are deactivated monocytes. However, the fact that MDSCs secrete proinflammatory cytokines such as IL-1β (Fig. 5A)
, which is a monocyte/macrophage activation marker [17
], suggests that tumor-infiltrating MDSCs cannot be categorized as deactivated monocytes/macrophage, unlike PECs (which do not express any M1 markers). Instead, it appears that tumor-infiltrating MDSCs have a unique character, as they simultaneously express the M1 markers iNOS, IL-1β, TNF-, and CXCL10, the M2 and deactivated monocyte markers CD206 and CD36, and the M2 marker arginase I. Indeed, the fact that blocking TGF-β changed the MDSC phenotype to one with high iNOS expression and lower CD206 and arginase I expression (Fig. 5C
and 5D)
suggests that tumor-infiltrating MDSCs can change their phenotype and that this is regulated to some extent by TGF-β in an autocrine manner. In conclusion, we showed that tumor-infiltrating MDSCs have a unique, inflammatory monocyte/macrophage character and cannot be classified into any known monocyte/macrophage categories.
We observed that although the tumors also expressed low levels of TGF-β, the main source of this cytokine in the tumor microenvironment was the tumor-infiltrating MDSCs. Thus, although Bronte et al. [7 ] have shown in their review that tumors produce immunoregulatory cytokines such as IL-10 and TGF-β, our present study indicates that TGF-β is produced by tumor-infiltrating MDSCs and acts in an autocrine manner on these cells (Fig. 4) . In terms of the factor(s) inducing TGF-β expression, Terabe et al. [27 ] reported that IL-13 produced by NKT cells induces MDSCs to express TGF-β. Gallina et al. [8 ] have also mentioned that splenic MDSCs themselves can secrete IL-13, but we found that in the tumor microenvironment, tumor-infiltrating MDSCs do not produce IL-13 (unpublished observations). Thus, if the stimulatory factor is IL-13, it is likely to be provided by a paracrine manner in our models.
Finally, we will comment on the nomenclature of tumor-infiltrating monocyte/macrophage lineage cells. Sica and Bronte [12 ] recently described MDSCs as cells that coexpress the myeloid cell markers CD11b and Gr-1 (Ly-6C/6G) and inhibit T cell activation. However, they also said that some tumor-infiltrating and splenic MDSCs, although retaining their suppressive properties, lose their Gr-1 expression, and the cell surface phenotype is CD11b+F4/80+. Thus, even if the CD11b+F4/80+ cells in the tumor microenvironment are Gr-1–, they can still be referred to as MDSCs. This means that TAM cannot be distinguished from the tumor-infiltrating Gr-1– MDSC population on the basis of cell surface phenotype, as F4/80 antigen is a well-known macrophage marker as well [28 ]. Instead, their immunosuppressive properties must be demonstrated before tumor-infiltrating F4/80+Gr-1low cells can be referred to as MDSCs. Although, if you would call tumor-infiltrating F4/80+Gr-1low cells TAM, you should demonstrate M2-type polarization of the cells, which is defined as typical property of TAM [23 ]. However, the concept of MDSC originally contains monocyte lineage and granulocyte lineage [1 ]. Therefore, if we describe the issue of a tumor-infiltrating monocyte/macrophage population, MDSCs may not be the best term to use. Instead, we might be able to refer to the tumor-infiltrating F4/80+Gr-1low population as "tumor-infiltrating (or -associated) mononuclear phagocytic cells (MPC)" as the cell lineage-based but not function-based terminology. That is because the definition of TAM is too narrow, and the definition of tumor-infiltrating MDSC is too broad to be used referring to tumor-infiltrating MPC.
Received September 6, 2007; revised January 7, 2008; accepted January 30, 2008.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
stimulation through a mechanism dependent on endogenous TNF- and IL-1 Eur. J. Immunol. 30,1263-1271[CrossRef][Medline] plus tumor necrosis factor- synergistically reduce suppressor cell activity Int. J. Cancer 46,245-250[Medline]
B and enhanced IRF-3/STAT1 activation) Blood 107,2112-2122This article has been cited by other articles:
 |
|
 |
|
  K. Lolmede, L. Campana, M. Vezzoli, L. Bosurgi, R. Tonlorenzi, E. Clementi, M. E. Bianchi, G. Cossu, A. A. Manfredi, S. Brunelli, et al. Inflammatory and alternatively activated human macrophages attract vessel-associated stem cells, relying on separate HMGB1- and MMP-9-dependent pathways J. Leukoc. Biol., May 1, 2009; 85(5): 779 - 787. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Ostrand-Rosenberg and P. Sinha Myeloid-Derived Suppressor Cells: Linking Inflammation and Cancer J. Immunol., April 15, 2009; 182(8): 4499 - 4506. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. C. Gensel, S. Nakamura, Z. Guan, N. van Rooijen, D. P. Ankeny, and P. G. Popovich Macrophages Promote Axon Regeneration with Concurrent Neurotoxicity J. Neurosci., March 25, 2009; 29(12): 3956 - 3968. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Sag, D. Carling, R. D. Stout, and J. Suttles Adenosine 5'-Monophosphate-Activated Protein Kinase Promotes Macrophage Polarization to an Anti-Inflammatory Functional Phenotype J. Immunol., December 15, 2008; 181(12): 8633 - 8641. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. Sinha, C. Okoro, D. Foell, H. H. Freeze, S. Ostrand-Rosenberg, and G. Srikrishna Proinflammatory S100 Proteins Regulate the Accumulation of Myeloid-Derived Suppressor Cells J. Immunol., October 1, 2008; 181(7): 4666 - 4675. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Nonaka, M. Saio, T. Suwa, A. B. Frey, N. Umemura, H. Imai, G.-F. Ouyang, S. Osada, M. Balazs, R. Adany, et al. Skewing the Th cell phenotype toward Th1 alters the maturation of tumor-infiltrating mononuclear phagocytes J. Leukoc. Biol., September 1, 2008; 84(3): 679 - 688. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
