Research ArticleClinical Studies

Surface-bound TGF-β1 on Effusion-derived Exosomes Participates in Maintenance of Number and Suppressive Function of Regulatory T-Cells in Malignant Effusions

JUNJI WADA, HIDEYA ONISHI, HIROYUKI SUZUKI, AKIO YAMASAKI, SHUNTARO NAGAI, TAKASHI MORISAKI and MITSUO KATANO
Anticancer Research September 2010, 30 (9) 3747-3757;

Abstract

Background: This study analysed the contribution of malignant-effusion derived exosomes (Eff-Ex) to the number and function of regulatory T-cells (Treg) in malignant effusions. Patients and Methods: Eff-Ex were collected from the malignant effusions of 24 cancer patients. Peripheral blood mononuclear cells (PBMCs) were co-cultured with different concentrations of Eff-Ex. FOXP3+ CD4+ T-cells were defined as Treg. Expression of molecules on Eff-Ex was determined by flow cytometric analysis. Results: The number of Treg decreased daily in parallel with the FOXP3 expression level. Purified Eff-Ex prevented the decreases in both Treg number and FOXP3 expression levels in a dose-dependent manner. Pre-treatment of Eff-Ex with a neutralizing mAb against TGF-β1 significantly reduced these effects and the suppressive function of Treg. Conclusion: Elimination of Eff-Ex or control of Eff-Ex expressing TGF-β1 may be new therapeutic strategies in immunotherapy for advanced cancer patients with malignant effusions.

A unique CD4+ T-cell population, designated as regulatory T-cells (Treg), plays an important role in maintaining immunological tolerance to self-antigens (1, 2). Recent studies have shown that the transcription factor forkhead box P3 (FOXP3) is not only a key intracellular marker but also a crucial developmental and functional factor for CD4+ Treg (3, 4). Thus, it is now generally considered that FOXP3+ CD4+ T-cells are Treg (3). Importantly, it has been shown that increases in Treg numbers at tumour sites, including malignant effusions, indicates that tumour-infiltrating Treg may negatively control the antitumour immune response (5-7). Most studies into the mechanisms involved in increasing Treg numbers at tumour sites have investigated the possible conversion of FOXP3−CD4+ T-cells to FOXP3+ Treg in situ. For example, it has been indicated that TGF-β, which is frequently secreted at tumour sites, can induce this conversion in vitro (8) and in vivo (9). In mice, TGF-β has been shown to be essential for maintaining FOXP3 expression and Treg function (10). Importantly, the suppressive function of Treg is partially mediated by surface-bound TGF-β on Treg (11). However, the mechanisms by which surface-bound TGF-β suppresses the function of effector cells are not fully understood.

Exosomes are endosome-derived organelles of 50-100 nm size, which are actively secreted through an exocytosis pathway by many cell types (12, 13). Exosomes are very rigid and resistant to enzymatic degradation in blood, ascites or effusion. (14). These biophysical properties allow exosomes to play an important role in cell to cell communication, in particular the communication between immune cells (15-17). In fact, exosomes express many cell-cell communication-related molecules, including MHC class I and II, CD86, tetraspanins, and heat-shock proteins (13, 14). It has been previously reported that exosomes secreted by monocyte-derived dendritic cells (Mo-DCs) support naïve CD4+ T-cell survival via NF-κB activation, induced by the interaction of HLA-DR with the TCR (18). Tumour cells also secrete exosomes, known as tumour-derived exosomes (TuEx) (19-22) and as a result, TuEx accumulate in malignant effusions from where it is relatively easy to obtain intact exosomes (23-25). Biophysical analysis of TuEx revealed that TuEx express several functional molecules such as MHC class I, tumour antigens and FasL, which can have effects on the biological properties of immune cells (23, 26-29). Interestingly, the possibility that surface-bound TGF-β on TuEx plays a role in tumour immune evasion has been indicated (22, 26).

As described above, both Treg and TuEx accumulate in malignant effusions. This study analysed the possible interaction between Treg and TuEx in malignant effusions, providing evidence that surface-bound TGF-β1 on TuEx plays a role in increasing Treg numbers in malignant effusions by prolonging survival and/or conversion of FOXP3−CD4+ T-cells to Treg. Elimination of TuEx from the tumour site might be a valuable therapeutic strategy for tumour immunotherapy.

Materials and Methods

Patients and patient-derived materials. Twenty four advanced cancer patients from the Fukuoka-Cancer General Clinic with malignant pleural effusions, malignant ascites and distant metastases were enrolled in this study. Written informed consent was obtained from all patients and healthy volunteers. Malignant pleural effusions, ascites and peripheral blood were obtained from these patients. Peripheral blood mononuclear cells (PBMCs) were isolated from heparinized peripheral blood by Ficoll–Hypaque (Life Technologies, Gaithersburg, MD, USA) and were maintained at 37°C under a humidified atmosphere of 5% CO2 in RPMI-1640 medium (Nacalai tesque, Kyoto, Japan) supplemented with 2% human albumin (Mitsubishi Pharma, Osaka, Japan) and antibiotics (100 units/ml penicillin and 100 μg/ml streptomycin; Meiji Seika, Tokyo, Japan). PBMCs from 13 healthy volunteers were used as controls. In several experiments, CD4+ T lymphocytes were selected from PBMCs using the Dynabeads FlowComp Human CD4 kit (Invitrogen, Tokyo, Japan).

Exosome isolation and purification. To purify Eff-Ex from malignant effusions, a generally accepted purification procedure was used (30). Briefly, malignant effusions were centrifuged at 300×g for 30 minutes to remove floating cells and pellets were collected after 3 rounds of centrifugation. Briefly, supernatants were centrifuged at 800×g for 30 minutes. Then the supernatants were centrifuged at 10,000×g for 30 minutes using an Optima LE-80k Ultracentrifuge (Beckman Coulter, Fullerton, CA, USA). Finally, the supernatants were underlayed with 30% sucrose/D2O cushion, and were subjected to ultracentrifugation at 100,000×g for 1 hour. The collected pellets were resuspended in PBS and centrifuged at 90,000×g for 85 minutes to allow irrelevant cell debris to form pellets. The exosome concentration was determined using a Nanodrop system (Thermo Scientific, Wilmington, DE, USA) before storage at −80°C.

Semiquantitative reverse transcription-polymerase chain reaction (RT-PCR). Total RNA was extracted by the guanidinium isothiocyanate-phenol-chloroform method and quantified by spectrophotometry (Ultrospec 2100 Pro; Amersham Pharmacia Biotech, Cambridge, UK). For semi-quantitative RT-PCR, pd (N) 6 random hexamers (GE healthcare UK Ltd, Buckinghamshire, UK) were used for priming. The sequences of the primers used were: FOXP3 forward, 5′-CAAGTTCCACAACATGCGAC-3′, reverse, 5′-ATTGAGTGTCCGCTGCTTCT-3′ and glyceraldehydes-3-phosphate dehydrogenase (GAPDH) forward, 5′-CCACCCATG GCAAATTCCATGGCA-3′, reverse, 5′-TCTAGACGGCAGGT CAGGTCCACC-3′. Amplification conditions comprised an initial denaturation step for 2 minutes at 95°C followed by 30 cycles of 94°C for 1 minute, 58°C for 1 minute and 72°C for 2 minutes. RT-PCR products were separated on ethidium bromide 2% agarose gels and visualised using a Molecular Imager FX (Bio-Rad, Hercules, CA, USA). The band intensities were calculated using Image J software (NIH, Bethesda, MD, USA).

Figure 1.
Figure 1.

Ratio of Treg to CD4+ T-cells (%Treg) is higher in malignant effusions than in peripheral blood from healthy volunteers or patients with malignant effusions. A: %Treg were determined by FACS. Bars show the means±SD. *p<0.001. B: FOXP3 mRNA expressions were examined by RT-PCR (upper panel). Band intensities were calculated and were normalized to GAPDH (lower panel).

FACS analysis of malignant effusion-derived mononuclear cells (EDMCs) and PBMCs. EDMCs and PBMCs were stained with FITC-conjugated mouse IgG2b-κ (isotype control), anti CD25, PE-conjugated mouse IgG2b-κ (isotype control) and PE-cy5-conjugated anti CD4 (all BD Pharmingen, San Jose, CA, USA) for 30 minutes at 4°C. In some experiments, CD4+ T lymphocytes were selected using the Dynabeads FlowComp Human CD4 kit (Invitrogen) and were labeled by CFSE (Invitrogen) for the evaluation of proliferation. Intracellular staining of FOXP3 was performed using the PE-conjugated anti-human FOXP3 Staining Set (clone PCH101; eBioscience, San Diego, CA, USA) according to the manufacturer's instructions. EDMCs and PBMCs were washed twice by PBS and were resuspended in 200 μl of PBS. Expression levels were examined using a FACSCalibur™ (Becton Dickinson Immunocytometry Systems, San Jose, CA, USA) using CELLQuest software (Becton Dickinson Immunocytometry Systems, San Jose, CA, USA).

FACS analysis of exosomes. Purified exosome preparations were stained directly by FITC-conjugated mouse IgG2b-κ (isotype control, BD Pharmingen), anti carcino embryonic antigen (CEA, Affinity BioReagents, Rockford, IL, USA), anti human HLA-ABC (BD Pharmingen), and cytokeratin (Beckman Coulter), PE-conjugated mouse IgG2b-κ (isotype control, BD Pharmingen), anti human LAP (TGF-β, R&D system, Minneapolis, ML, USA), and anti human HLA-DR (BD Pharmingen) for 30 minutes at 4°C. In some experiments, 1 μg/ml of anti-TGF-β1 neutralizing mAb (R&D Systems) or control mouse IgG2B (R&D System) was treated on malignant effusion-derived exosomes (Eff-Ex) for 1 hour at 22°C. Exosomes were washed twice by PBS and centrifuged at 100,000×g for 60 minutes. Exosomes were resuspended in 200 μl of PBS and the expression of surface molecules examined using a FACSCalibur™.

Electron microscopy. Eff-Ex were fixed in 3% glutaraldehyde in 0.1 M cacodylate buffer (CB) at pH 7.3 for 3 hours at 4°C and washed in 0.1 M CB. Eff-Ex were resuspended and embedded in 4% agar. The agar was cut into 1-mm3 pieces, and the pieces were fixed in 1% osmium tetroxide in 0.1 M CB overnight and then washed in distilled water. The specimens were dehydrated in a graded series of ethanol and embedded in Epon 812. Ultrathin sections were treated with uranyl acetate followed by lead citrate and were examined with an electron microscope (JEM-1200EX, JEOL, Tokyo, Japan).

Immunoblotting. Eff-Ex extract (50 μg) was separated by electrophoresis on 12.5% SDS-polyacrylamide gels and transferred to Protran nitrocellulose membranes (Schleicher and Schnell BioScience, Dassel, Germany). Blots were incubated with anti-MHC class I pAb (1:200, Santa Cruz) or anti-HSP70 mAb (1:200, Santa Cruz) overnight at 4°C. Blots were then incubated with secondary antibody at room temperature for 1 h. Immunocomplexes were detected with an enhanced chemiluminescence reagent (Amersham Biosciences) and visualised with a Molecular Imager FX (Bio-Rad).

Enzyme-linked immunosorbent assay (ELISA). Cell free effusions were collected by centrifuging at 2,500×g for 30 minutes. The concentrations of IL-10, TGF-β1, and VEGF in effusions were measured by ELISA (Biosource International, Inc, Camarillo, CA, USA) according to the manufacturer's protocol. The detection limits of the IL-10, TGF-β1, and VEGF assays were 1 pg/ml, 15.6 pg/ml, and 5 pg/ml, respectively.

Mixed lymphocyte reaction. CD4+ T lymphocytes from patients were cultured with or without IgG-treated Eff-Ex or anti-TGF-β1 mAb-treated Eff-Ex for 1 day. The 3 types of CD4+ T lymphocytes were then co-cultured with healthy volunteer derived PBMCs at a 1:1 cell ratio for one day using polycarbonate 6-well plates (Becton Dickinson Labware, NJ, USA). IFN-γ in the supernatants was measured by ELISA.

Statistical analysis. Fisher's exact probability test was used for the statistical analyses. Data was analysed with an SAS statistical software package (Abacus Concepts, Berkeley, CA, USA). A p-value of <0.05 was considered statistically significant.

Results

%Treg is higher in malignant effusions than in healthy volunteer peripheral blood. In the present study, FOXP3+CD4+ T-cells were defined as Treg. The ratio of Treg to CD4+ T-cells (%Treg) in malignant effusion-derived cells obtained from 24 untreated patients with various types of tumours was measured. In non-malignant exudative effusions collected from 8 patients with liver cirrhosis, %Treg was less than 1% (data not shown). Concerning %Treg in PBMCs of the 24 patients examined in the present study, data was obtained from only 3 patients. Therefore, PBMCs collected from 13 new patients with malignant effusions were analysed. The %Treg in malignant effusions was significantly higher than that in PBMCs collected from healthy volunteers or patients with malignant effusions (Figure 1A). The level of FOXP3 mRNA expression was examined by RT-PCR using CD4+ T-cells isolated from effusion-derived cells or healthy volunteer-derived PBMCs. FOXP3 expression in effusion-derived CD4+ T-cells was clearly higher than that in healthy volunteer-derived CD4+ T-cells (Figure 1B).

Effusions reduce the decrease of PBMC-derived Treg in an ex vivo culture system. To investigate the mechanisms involved in mediating the increase in %Treg in malignant effusions, the concentrations of TGF-β1, IL-10, and VEGF in 18 malignant effusions were measured (Figure 2A), where %Treg represents the ratio of FOXP3+CD4+ T-cells to total CD4+ T-cells in malignant effusions. These cytokines have previously been indicated to play a role in Treg induction or maintenance (8, 9, 31, 32). To exclude the influence of these cytokines, strictly target effusions were selected at the following three steps. Step 1: 9 effusions were first selected (case 7, 8, 9, 10, 11, 13, 14, 15, and 17) which did not contain these cytokines. Step2: then one effusion (case 14) in which %Tregs was similar to that of PBMCs of healthy volunteers' (4%) was excluded. Step 3: finally, the effusions of 4 patients (case 8, 9, 10, and 11) in whom collection of PBMCs was possible throughout experimental period were selected (Figure 2A). When PBMCs collected from these 4 patients were cultured in conventional RPMI medium in the absence of effusions, the %Treg and Treg number decreased daily (data not shown). In case 9, addition of the effusions to the culture medium significantly reduced this decrease in %Treg (Figure 2B; left panel) and Treg number (Figure 2B; middle panel) in PBMCs in a dose-dependent manner. However, the effusions did not affect FOXP3−CD4+ T-cell numbers in PBMCs (Figure 2B; right panel). Another remaining 3 cases showed a similar pattern (data not shown). On the other hand, 5 non-malignant effusions collected from patients with liver cirrhosis did not reduce the decrease in Tregs (data not shown). Next, effusions were filtered to allow components smaller than 100 kDa to pass through. Both passed (cytokine-rich) and non-passed (cytokine-poor) fractions were re-adjusted to the original volume with RPMI medium before addition to PBMCs. Neither fraction contained detectable cytokines such as TGF-β, VEGF, and IL-10. Only the non-passed fraction reduced the decrease in %Treg in PBMCs from cases 8, 9, 10 and 11 (Figure 2C).

Eff-Ex reduce the decrease in number and FOXP3 expression levels of PBMC-derived Treg. It was speculated that insoluble substances, such as exosomes, may contribute to this effect because it has previously been demonstrated that Mo-DC-derived exosomes support naïve CD4+ T-cell survival by interaction of HLA-DR with the TCR (18). Exosomes were collected and purified from these malignant effusions as described in the Materials and Methods section. Purified Eff-Ex used in the present study satisfied three major criteria as exosomes: they had a size of 60-100 nm in diameter (Figure 3A; left panel), a density of 1.13 to 1.21 g/dl in a sucrose gradient and they were enriched with HSP 70 and MHC molecules (Figure 3A; right panel) indicating an endocytic origin (13, 30, 33). Exosomes have been commonly used between 1 μg/ml and 100 μg/ml in the literature (34-37). Therefore, in the present study, Eff-Ex at 0-50 μg/ml was used. Eff-Ex significantly reduced the decrease in both %Treg (Figure 3B; left panels) and Treg number (Figure 3B; middle panels) in PBMCs from cases 8 and 9 in a dose-dependent manner. Interestingly, Eff-Ex also reduced the decrease in Treg FOXP3 expression levels in ex vivo culture (Figure 3B; right panels). The remaining two cases showed a similar pattern (data not shown). To determine the mechanisms by which Eff-Ex prevented the Treg decrease, we investigated the possibility of Eff-Ex-dependent Treg proliferation. CD4+ T-cells purified from PBMCs were pre-labeled with CFSE and cell division was assessed by CFSE dilution. The values of upper right corner show mean fluorescence intensity of CFSE (Figure 3C; lower panels). Although Eff-Ex reduced the decrease in %Treg (Figure 3C; upper panels) they did not induce proliferation of Treg or FOXP3−CD4+ T-cells (Figure 3C; lower panels). These data suggest that Eff-Ex did not induce Treg proliferation.

Figure 2.
Figure 2.

Effusions reduce the decrease of PBMC-derived Treg in an ex vivo culture system. A: Cytokine concentrations of malignant effusions were measured by ELISA. B: After PBMCs from case 9 were cultured in the absence or presence of malignant effusion at the indicated concentrations for 5 days, %Treg (left panel), Treg (middle panel) and FOXP3−CD4+ T-cell numbers (right panel) in the PBMCs were evaluated by FACS. Bars show the means±SD. *, p<0.001. C: After PBMCs from cases 8, 9, 10, and 11 were cultured in the presence of whole effusion, passed fraction (cytokine filtrate rich fraction) and non-passed fraction (cytokine non-filtrate poor fraction), %Treg in the PBMCs was evaluated by FACS. Bars show means±SD. *p<0.001, **p<0.05.

Figure 3.
Figure 3.

Malignant-effusion-derived exosomes (Eff-Ex) reduce the decrease in PBMCs-derived Treg numbers and FOXP3 expression levels. A: Left panel: representative electron microscopic image of Eff-Ex. Right panel: Western blot analysis of protein extracted from Eff-Ex. B: After PBMCs from case 8 and 9 were cultured in the absence or presence of Eff-Ex at the indicated concentrations for 5 days, %Treg (left panels), Treg number (middle panels) and FOXP3 expression (right panels) in the PBMCs were examined by FACS. Bars show means±SD. *p<0.001, **p<0.05. C: CD4+T-cells purified from PBMCs of case 8 were pre-labeled with CFSE and cultured in the absence or presence of 50 μg/ml Eff-Ex for 5 days. CFSE expression in CD4+T-cells was then evaluated by FACS.

Figure 4.
Figure 4.

Eff-Ex are of tumour cell origin and express surface-bound TGF-β1. A: The expression of MHC class I, HLA-DR, CEA and TGF-β1 on Eff-Ex from cases 8, 9, and 11 were evaluated by FACS. Light line: isotype controls, heavy line: the expressions in class I, HLA-DR, CEA and TGF-β1 in each case. B: The expressions of CEA and TGF-β1 on Eff-Ex (upper panels in each case) and CEA, TGF-β1 and cytokeratin on malignant effusion derived cells (lower panels in each case) in case 9 and 11 were evaluated by FACS. The value shows the percentage of positive cells.

Figure 5.
Figure 5.

Surface-bound TGF-β1 on Eff-Ex reduces the decrease in PBMC-derived Treg numbers and FOXP3 expression. A: The expression of TGF-β on Eff-Ex cultured with or without 1 μg/ml of anti-TGF-β1 neutralizing mAb was evaluated by FACS. Dotted line: isotype controls, thick line: TGF-β1 expression in Eff-Ex cultured with anti-TGF-β1 neutralizing mAb, thin line; TGF-β1 expression in Eff-Ex cultured with control IgG. B: PBMCs from cases 8 and 9 were cultured with autologous Eff-Ex treated with anti TGF-β1 neutralizing mAb or control IgG for 1 day and the %Treg (left panels), Treg number (middle panels) and the expression of FOXP3 (right panels) were examined by FACS. Bars show the means±SD. *p<0.001.

Eff-Ex are mainly of tumour cell origin and express surface-bound TGF-β1. The expression of several molecules was investigated, including MHC class I, HLA-DR, tumour antigen CEA, and TGF-β1, on Eff-Ex by FACS analysis in 3 of the 4 effusions (Figure 4A). In cases 8 and 9, most Eff-Ex expressed MHC class I, CEA, and TGF-β1, while in case 11, CEA and TGF-β1 were only slightly expressed. However, HLA-DR expression was not detected on Eff-Ex of these 3 cases. In cases 9 and 11, the expression of CEA, TGF-β1 and cytokeratin on effusion-derived cells could also be detected (Figure 4B; lower panels in each case). Eighty-seven percent and almost 100% of effusion-derived cells were cytokeratin positive, respectively (Figure 4B), suggesting that most T-cells existing in these effusions are epithelial cells, that is, tumour cells. In addition, most TGF-β1-positive effusion-derived cells are cytokeratin positive (Figure 4B; lower right panels in each case). Taken together, these data indicate that the TGF-β1 on Eff-Ex cells examined here is mainly of tumour cell origin (TuEx).

Surface-bound TGF-β1 on Eff-Ex reduces the decrease in PBMC-derived Treg numbers and FOXP3 expression levels. To directly confirm whether surface-bound TGF-β1 on Eff-Ex can prevent the decrease in Treg, TGF-β1 neutralization experiments were performed using autologous PBMC-derived CD4+ T-cells as target T-cells. Briefly, Eff-Ex were incubated with neutralising mAb against TGF-β1 or control IgG for 1 hour and then excessive antibodies were eliminated by ultracentrifugation. TGF-β1 neutralization significantly reduced the expression of surface-bound TGF-β1 on Eff-Ex (Figure 5A). Eff-Ex again reduced the decrease in both %Treg (Figure 5B; left panels) and Treg number (Figure 5B; middle panels). Pretreatment of Eff-Ex with TGF-β1 neutralizing mAb (but not irrelevant antibody, not shown) significantly reduced the effect of Eff-Ex on the Treg decrease (Figure 5B; left and middle panels). Furthermore, the FOXP3 expression levels in Treg also decreased daily under the current culture conditions (data not shown). Eff-Ex inhibited the decrease in FOXP3 expression levels in Treg, and anti-TGF-β1 mAb treatment significantly reduced this effect (Figure 5B; right panels).

Surface-bound TGF-β1 on Eff-Ex participates in the suppressive function of Treg. Finally, whether surface-bound TGF-β1 on Eff-Ex participates in the suppressive functions of Treg was examined. Before the experiments, two types of Eff-Ex were prepared: anti-TGF-β1 neutralizing mAb-treated Eff-Ex and control IgG-treated Eff-Ex. Patient CD4+ T-cells were cultured in the presence or absence of the 2 types of Eff-Ex for 5 days. Cells were then washed intensively to eliminate the Eff-Ex and the cell number was readjusted. Each of the three different types of CD4+ T-cells, including non-stimulated CD4+ T-cells, were then mixed with healthy volunteer-derived PBMCs at a 1:1cell ratio and cultured for a further 24 hours. Cell-free supernatants were then collected and concentrations of IFN-γ were measured by ELISA (Figure 6). The IFN-γ concentration was the highest in the non-stimulated CD4+ T cell group, the lowest in the IgG-treated Eff-Ex-treated CD4+ T cell group, while the anti-TGFβ1 mAb-pretreated Eff-Ex-treated CD4+ T cell group was in the middle. These data indicates that surface-bound TGF-β1 on Eff-Ex contributes to the suppressive function of Treg.

Figure 6.
Figure 6.

Surface-bound TGF-β1 on Eff-Ex participates in the suppressive function of Treg. Two types of Eff-Ex were prepared; one is anti TGF-β1 neutralizing mAb-treated Eff-Ex and the other is control IgG-treated Eff-Ex. Patient-derived CD4+ T-cells were cultured in the presence or absence of the 2 types of Eff-Ex for 5 days. CD4+ T-cells were washed intensively and were co-cultured with healthy volunteer derived PBMCs at a 1:1 cell ratio for 1 day. IFN-γ secretion in the cell supernatants was measured by ELISA. NS, Non-stimulated CD4+ T-cells; IgG, IgG-treated Ex-stimulated CD4+ T-cells; TGF-β1, anti-TGF-β1 neutralizing mAb-treated Ex–stimulated CD4+ T-cells. Bars show the means±SD. *p<0.001.

Discussion

This study demonstrated that TuEx contribute to the maintenance of Treg numbers and suppressive function in malignant effusions. It was also shown that surface-bound TGF-β1 on the TuEx may mediate these abilities through the control of FOXP3 expression.

It has been shown that Treg numbers increase in malignant effusions (5). On the other hand, it has been indicated that they are susceptible to apoptosis and have critically short telomeres and low telomerase activity (5, 38), suggesting a short survival potential. In fact, Treg numbers more rapidly decreased compared with FOXP3−CD4+ T-cells in the current study (Figure 2B). Importantly, the addition of malignant effusions to the in vitro culture system selectively reduced the decrease in Treg numbers, compared with FOXP3−CD4+ T-cells (Figure 2B). To determine the mechanisms by which effusions selectively reduce the rapid decrease in Treg numbers, the concentrations of Treg-related cytokines, including TGF-β1, IL-10 and VEGF, in the malignant effusions were first examined (8, 9, 31, 32). Interestingly, in 4 of the 18 effusions examined concentrations of these cytokines were below the detection limit, while the inhibitory effect of the effusions on the rapid decrease in Treg numbers mainly existed in the non-passed fraction, over 100 kDa. This suggests that insoluble and large molecules or substances confer this ability (Figure 2C). Since it has previously been shown that Mo-DC-derived exosomes can prolong naïve CD4+ T-cell survival through the interactions of HLA-DR on exosomes and the TCR on CD4+ T-cells (18), this study focused on exosomes existing in malignant effusions. Exosomes are very rigid, stable at pH 7, and resistant to enzymatic degradation (13, 14). Therefore, it is possible to purify Eff-Ex from malignant effusions relatively easily (23-25). The first question was whether Eff-Ex could prevent the rapid decrease in Treg numbers. In this experiment, autologous PBMC-derived CD4+ T-cells were used as target T-cells. As expected, Eff-Ex reduced this rapid decrease in Treg numbers in a dose-dependent manner (Figure 3B). Therefore, the current data suggest a contribution of Eff-Ex to the increase of Treg in malignant effusions. The effect of Eff-Ex on the survival of Treg using frozen-preserved malignant effusions has been examined previously. Four out of the seven effusions reduced the decrease of PBMC-derived Treg (data not shown). Thus, we believe that this effect of Eff-Ex on Treg is not universal but also is not rare.

The second question concerned the molecular mechanisms by which Eff-Ex selectively prevented the decrease in Treg numbers in the current culture system. Recent evidence indicates that surface-bound TGF-β on exosomes or several types of cells, including tumour cells, is functional (11, 22, 39, 40). In particular, Clayton et al. (22) demonstrated that the suppressive function of CD25+CD4+ T-cells was enhanced by TuEx, and suggested that surface-bound TGF-β1 on TuEx may play an important role in this phenomenon. Based on these findings, this study showed, for the first time, direct evidence that Eff-Ex express surface-bound TGF-β1 using FACS analysis (Figure 4). In addition, it was shown that a TGF-β1-neutralizing mAb is able to block the expression of surface-bound TGF-β1 on Eff-Ex (Figure 5A). Pre-treatment of Eff-Ex with TGF-β1 mAb significantly abrogated the ability of Eff-Ex to prevent the rapid decrease in Treg numbers (Figure 5B; left and middle panels). Although this strongly indicates a contribution of surface-bound TGF-β1 on this property of Eff-Ex, it is still unclear what signaling pathways of CD4+ T-cells are stimulated by surface-bound TGF-β1. Nevertheless, the observation that TGF-β1 on TuEx may affect Treg FOXP3 expression levels is noteworthy (Figure 5B; right panels). This indirectly indicates that TGF-β1 on Eff-Ex is functional, because FOXP3 is a master control gene for the development and function of naturally-occurring Treg (4). The Treg FOXP3 expression level decreased daily in the current culture system. Eff-Ex reduced this decrease in a dose-dependent manner, and neutralizing TGF-β1 mAb significantly reduced this ability of Eff-Ex (Figure 5B right panels). These observations suggest several possibilities concerning the relationship between Eff-Ex and Treg. The first possibility is that Treg spontaneously lose FOXP3 expression during in vitro culture, and consequently a conversion from Treg to FOXP3−CD4+ T-cells occurs. If so, TGF-β1 on Eff-Ex may inhibit this conversion. The second possibility is that Treg showing high-level expression of FOXP3 are a subpopulation of cells that survive for a short time compared with cells showing low-level expression of FOXP3. If so, TGF-β1 on Eff-Ex may selectively suppress the decrease in Treg numbers that highly express FOXP3. The third possibility is that TGF-β1 on Eff-Ex induces FOXP3 expression in FOXP3−CD4+ T-cells. Several investigators have indicated this possibility (8, 9). This study attempted to confirm this using CD25-CD4+ T-cells, but results were inconclusive. The fourth possibility is that TGF-β1 on Eff-Ex induces proliferation of Treg. This is unlikely as Eff-Ex did not induce cell division in Treg or FOXP3−CD4+ T-cells (Figure 3C). It has been indicated that FOXP3 expression levels are closely associated with the suppressive function of Treg (2, 41, 42). Since there was a possibility that TGF-β1 on Eff-Ex controls FOXP3 expression levels, it was examined whether TGF-β1 on Eff-Ex affect Treg suppressive function using a mixed lymphocyte culture (MLC) system. When CD4+ T-cells stimulated with Eff-Ex instead of unstimulated CD4+ T-cells were added to the MLC, concentrations of IFN-γ in the MLC medium were lower (Figure 6). This suggested that the suppressive effect of Eff-Ex-stimulated CD4+ T-cells may be higher than that of unstimulated CD4+ T-cells (Figure 6). Importantly, pre-treatment of Eff-Ex with neutralizing TGF-β1 mAb reduced this suppressive effect of Eff-Ex-stimulated CD4+ T-cells, suggesting that TGF-β1 on Eff-Ex may be controlling Treg suppressive function. In conclusion, the current data strongly indicate that surface-bound TGF-β1 on Eff-Ex is functional and contributes to the maintenance of both Treg number and function in malignant effusions.

The final question was what kind of cells produce Eff-Ex. Several types of cells, including tumour cells, mesothelial cells, and leukocytes, exist in malignant effusions, and most of these cells secrete exosomes (13). In addition, it has been reported that tumour cells, immature DCs, thymic cells, and Treg express surface-bound TGF-β1 (11, 22, 39, 40, 43). In the present study, FACS analysis of molecules expressed on Eff-Ex collected from 3 malignant effusions was performed (Figure 4). Expression of CEA, indicating tumour cell-origin, was confirmed in most of the Eff-Ex from 2 effusions. In one Eff-Ex, in which only weak CEA expression was detected, most of the effusion-derived cells were cytokeratin-positive (Figure 4B). On the other hand, expression of HLA-DR, which indicates DCs- or activated B cell-origin, was not detected in any of the Eff-Ex. Taken together with these data, it seems that the Eff-Ex used here are derived from tumour cells, that is, TuEx.

Evidence that exosomes are the fourth information transmitter in addition to the cytokine, hormone, and neurotransmitters is rapidly increasing (15-17). In the field of immunotherapy, TuEx were at first utilised as a source of tumour antigens to induce CTLs against tumours (19). Lately, however, the use of TuEx as a source of immunosuppressive molecules has gained more attention (21, 22, 28). Such a complex property of TuEx teaches us the necessity to fully understand the function of an individual molecule expressed on TuEx. The present study provides evidence that surface-bound TGF-β1 on TuEx plays a role in the creation of an immunosuppressive microenvironment in malignant effusions by maintaining Treg number and function. If this is indeed so, elimination of TuEx from malignant effusions may be a valuable strategy for immunotherapy against patients with malignant effusions. However, it is still unclear if blocking of surface-bound TGF-β on Eff-Ex can induce a type I antitumour response.

Acknowledgements

This study was supported by General Scientific Research Grants (18591440 and 21591670) from the Ministry of Education, Culture, Sports and Technology of Japan. The authors thank Kaori Nomiyama for her skillful technical assistance.

Footnotes

  • ↵* These Authors contributed equally to this study.

  • Conflict of interest

    The authors declare no financial or commercial conflict of interest.

  • Received May 14, 2010.
  • Revision received June 2, 2010.
  • Accepted June 8, 2010.

References

  1. ↵
    1. Sakaguchi S
    : Naturally arising CD4+ regulatory T-cells for immunologic self-tolerance and negative control of immune responses. Annu Rev Immunol 22: 531-562, 2004.
    OpenUrlCrossRefPubMed
  2. ↵
    1. Sakaguchi S,
    2. Yamaguchi T,
    3. Nomura T,
    4. Ono M
    : Regulatory T-cells and immune tolerance. Cell 133: 775-787, 2008.
    OpenUrlCrossRefPubMed
  3. ↵
    1. Hori S,
    2. Nomura T,
    3. Sakaguchi S
    : Control of regulatory T-cell development by the transcription factor Foxp3. Science 299: 1057-1061, 2003.
    OpenUrlAbstract/FREE Full Text
  4. ↵
    1. Fontenot JD,
    2. Gavin MA,
    3. Rudensky AY
    : Foxp3 programs the development and function of CD4+CD25+ regulatory T-cells. Nat Immunol 4: 330-336, 2003.
    OpenUrlCrossRefPubMed
  5. ↵
    1. Curiel TJ,
    2. Coukos G,
    3. Zou L,
    4. Alvarez X,
    5. Cheng P,
    6. Mottram P,
    7. Evdemon-Hogan M,
    8. Conejo-Garcia JR,
    9. Zhang L,
    10. Burow M,
    11. Zhu Y,
    12. Wei S,
    13. Kryczek I,
    14. Daniel B,
    15. Gordon A,
    16. Myers L,
    17. Lackner A,
    18. Disis ML,
    19. Knutson KL,
    20. Chen L,
    21. Zou W
    : Specific recruitment of regulatory T-cells in ovarian carcinoma fosters immune privilege and predicts reduced survival. Nat Med 10: 942-949, 2004.
    OpenUrlCrossRefPubMed
    1. Gao Q,
    2. Qiu SJ,
    3. Fan J,
    4. Zhou J,
    5. Wang XY,
    6. Xiao YS,
    7. Xu Y,
    8. Li YW,
    9. Tang ZY
    : Intratumoral balance of regulatory and cytotoxic T-cells is associated with prognosis of hepatocellular carcinoma after resection. J Clin Oncol 25: 2586-2593, 2007.
    OpenUrlAbstract/FREE Full Text
  7. ↵
    1. Bates GJ,
    2. Fox SB,
    3. Han C,
    4. Leek RD,
    5. Garcia JF,
    6. Harris AL,
    7. Banham AH
    : Quantification of regulatory T-cells enables the identification of high-risk breast cancer patients and those at risk of late relapse. J Clin Oncol 24: 5373-5380, 2006.
    OpenUrlAbstract/FREE Full Text
  8. ↵
    1. Chen W,
    2. Jin W,
    3. Hardegen N,
    4. Lei KJ,
    5. Li L,
    6. Marinos N,
    7. McGrady G,
    8. Wahl SM
    : Conversion of peripheral CD4+CD25-naïve T-cells to CD4+CD25+ regulatory T-cells by TGF-β induction of transcription factor Foxp3. J Exp Med 198: 1875-1886, 2003.
    OpenUrlAbstract/FREE Full Text
  9. ↵
    1. Huber S,
    2. Schramm C,
    3. Lehr HA,
    4. Mann A,
    5. Schmitt S,
    6. Becker C,
    7. Protschka M,
    8. Galle PR,
    9. Neurath MF,
    10. Blessing M
    : Cutting edge: TGF-β signaling is required for the in vivo expansion and immunosuppressive capacity of regulatory CD4+CD25+ T-cells. J Immunol 173: 6526-6531, 2004.
    OpenUrlAbstract/FREE Full Text
  10. ↵
    1. Marie JC,
    2. Letterio JJ,
    3. Gavin M,
    4. Rudensky AY
    : TGF-β1 maintains suppressor function and Foxp3 expression in CD4+CD25+ regulatory T-cells. J Exp Med 201: 1061-1067, 2005.
    OpenUrlAbstract/FREE Full Text
  11. ↵
    1. Nakamura K,
    2. Kitani A,
    3. Strober W
    : Cell contact-dependent immunosuppression by CD4+CD25+ regulatory T-cells is mediated by cell surface-bound transforming growth factor β. J Exp Med 194: 629-644, 2001.
    OpenUrlAbstract/FREE Full Text
  12. ↵
    1. Schartz NE,
    2. Chaput N,
    3. André F,
    4. Zitvogel L
    : From the antigen-presenting cell to the antigen-presenting vesicle: the exosomes. Curr Opin Mol Ther 4: 372-381, 2002.
    OpenUrlPubMed
  13. ↵
    1. Mignot G,
    2. Roux S,
    3. Thery C,
    4. Segura E,
    5. Zitvogel L
    : Prospects for exosomes in immunotherapy of cancer. J Cell Mol Med 10: 376-388, 2006.
    OpenUrlCrossRefPubMed
  14. ↵
    1. Laulagnier K,
    2. Motta C,
    3. Hamdi S,
    4. Roy S,
    5. Fauvelle F,
    6. Pageaux JF,
    7. Kobayashi T,
    8. Salles JP,
    9. Perret B,
    10. Bonnerot C,
    11. Record M
    : Mast cell-and dendritic cell-derived exosomes display a specific lipid composition and an unusual membrane organization. Biochem J 380: 161-171, 2004.
    OpenUrlCrossRefPubMed
  15. ↵
    1. Johnstone RM
    : Exosomes biological significance: a concise review. Blood Cells Mol Dis 36: 315-321, 2006.
    OpenUrlCrossRefPubMed
    1. Li XB,
    2. Zhang ZR,
    3. Schluesener HJ,
    4. Xu SQ
    : Role of exosomes in immune regulation. J Cell Mol Med 10: 364-375, 2006.
    OpenUrlCrossRefPubMed
  17. ↵
    1. Ohno H
    : Overview: membrane traffic in multicellular systems: more than just a housekeeper. J Biochem 139: 941-942, 2006.
    OpenUrlAbstract/FREE Full Text
  18. ↵
    1. Matsumoto K,
    2. Morisaki T,
    3. Kuroki H,
    4. Kubo M,
    5. Onishi H,
    6. Nakamura K,
    7. Nakahara C,
    8. Kuga H,
    9. Baba E,
    10. Nakamura M,
    11. Hirata K,
    12. Tanaka M,
    13. Katano M
    : Exosomes secreted from monocyte-derived dendritic cells support in vitro naive CD4+ T cell survival through NF-κB activation. Cell Immunol 231: 20-29, 2004.
    OpenUrlCrossRefPubMed
  19. ↵
    1. Wolfers J,
    2. Lozier A,
    3. Raposo G,
    4. Regnault A,
    5. Théry C,
    6. Masurier C,
    7. Flament C,
    8. Pouzieux S,
    9. Faure F,
    10. Tursz T,
    11. Angevin E,
    12. Amigorena S,
    13. Zitvogel L
    : Tumor-derived exosomes are a source of shared tumor rejection antigens for CTL cross-priming. Nat Med 7: 297-303, 2001.
    OpenUrlCrossRefPubMed
    1. Gastpar R,
    2. Gehrmann M,
    3. Bausero MA,
    4. Asea A,
    5. Gross C,
    6. Schroeder JA,
    7. Multhoff G
    : Heat shock protein 70 surface-positive tumor exosomes stimulate migratory and cytolytic activity of natural killer cells. Cancer Res 65: 5238-5247, 2005.
    OpenUrlAbstract/FREE Full Text
  21. ↵
    1. Liu C,
    2. Yu S,
    3. Zinn K,
    4. Wang J,
    5. Grizzle WE,
    6. Zinn KR,
    7. Zhang HG
    : Murine mammary carcinoma exosomes promote tumor growth by suppression of NK cell function. J Immunol 176: 1375-1385, 2006.
    OpenUrlAbstract/FREE Full Text
  22. ↵
    1. Clayton A,
    2. Mitchell JP,
    3. Court J,
    4. Mason MD,
    5. Tabi Z
    : Human tumor-derived exosomes selectively impair lymphocyte responses to interleukin-2. Cancer Res 67: 7458-7466, 2007.
    OpenUrlAbstract/FREE Full Text
  23. ↵
    1. Andre F,
    2. Schartz NE,
    3. Movassagh M,
    4. Flament C,
    5. Pautier P,
    6. Morice P,
    7. Pomel C,
    8. Lhomme C,
    9. Escudier B,
    10. Le Chevalier T,
    11. Tursz T,
    12. Amigorena S,
    13. Raposo G,
    14. Angevin E,
    15. Zitvogel L
    : Malignant effusions and immunogenic tumor-derived exosomes. Lancet 360: 295-305, 2002.
    OpenUrlCrossRefPubMed
    1. Bard MP,
    2. Hegmans JP,
    3. Hemmes A,
    4. Luider TM,
    5. Willemsen R,
    6. Severijnen LA,
    7. van Meerbeeck JP
    : Proteomic analysis of exosomes isolated from human malignant pleural effusions. Am J Respir Cell Mol Biol 31: 114-121, 2004.
    OpenUrlCrossRefPubMed
  25. ↵
    1. Navabi H,
    2. Croston D,
    3. Hobot J,
    4. Clayton A,
    5. Zitvogel L,
    6. Jasani B,
    7. Bailey-Wood R,
    8. Wilson K,
    9. Tabi Z,
    10. Mason MD,
    11. Adams M
    : Preparation of human ovarian cancer ascites-derived exosomes for a clinical trial. Blood Cells Mol Dis 35: 149-152, 2005.
    OpenUrlCrossRefPubMed
  26. ↵
    1. Clayton A,
    2. Mitchell JP,
    3. Court J,
    4. Linnane S,
    5. Mason MD,
    6. Tabi Z
    : Human tumor-derived exosomes down-modulate NKG2D expression. J Immunol 180: 7249-7258, 2008.
    OpenUrlAbstract/FREE Full Text
    1. Abusamra AJ,
    2. Zhong Z,
    3. Zheng X,
    4. Li M,
    5. Ichim TE,
    6. Chin JL,
    7. Min WP
    : Tumor exosomes expressing Fas ligand mediate CD8+ T-cell apoptosis. Blood Cells Mol Dis 35: 169-173, 2005.
    OpenUrlCrossRefPubMed
  28. ↵
    1. Huber V,
    2. Fais S,
    3. Iero M,
    4. Lugini L,
    5. Canese P,
    6. Squarcina P,
    7. Zaccheddu A,
    8. Colone M,
    9. Arancia G,
    10. Gentile M,
    11. Seregni E,
    12. Valenti R,
    13. Ballabio G,
    14. Belli F,
    15. Leo E,
    16. Parmiani G,
    17. Rivoltini L
    : Human colorectal cancer cells induce T-cell death through release of proapoptotic microvesicles: role in immune escape. Gastroenterology 128: 1796-1804, 2005.
    OpenUrlCrossRefPubMed
  29. ↵
    1. Iero M,
    2. Valenti R,
    3. Huber V,
    4. Filipazzi P,
    5. Parmiani G,
    6. Fais S,
    7. Rivoltini L
    : Tumour-released exosomes and their implications in cancer immunity. Cell Death Differ 15: 80-88, 2008.
    OpenUrlCrossRefPubMed
  30. ↵
    1. Lamparski HG,
    2. Metha-Damani A,
    3. Yao JY,
    4. Patel S,
    5. Hsu DH,
    6. Ruegg C,
    7. Le Pecq JB
    : Production and characterization of clinical grade exosomes derived from dendritic cells. J Immunol Methods 270: 211-226, 2002
    OpenUrlCrossRefPubMed
  31. ↵
    1. O'Garra A,
    2. Vieira PL,
    3. Vieira P,
    4. Goldfeld AE
    : IL-10-producing and naturally occurring CD4+ Tregs: limiting collateral damage. J Clin Invest 114: 1372-1378, 2004.
    OpenUrlCrossRefPubMed
  32. ↵
    1. Wada J,
    2. Suzuki H,
    3. Fuchino R,
    4. Yamasaki A,
    5. Nagai S,
    6. Yanai K,
    7. Koga K,
    8. Nakamura M,
    9. Tanaka M,
    10. Morisaki T,
    11. Katano M
    : The contribution of vascular endothelial growth factor to the induction of regulatory T-cells in malignant effusions. Anticancer Res 29: 881-888, 2009.
    OpenUrlAbstract/FREE Full Text
  33. ↵
    1. Raposo G,
    2. Nijman HW,
    3. Stoorvogel W,
    4. Liejendekker R,
    5. Harding CV,
    6. Melief CJ,
    7. Geuze HJ
    : B lymphocytes secrete antigen-presenting vesicles. J Exp Med 183: 1161-1172, 1996.
    OpenUrlAbstract/FREE Full Text
  34. ↵
    1. Abusamra AJ,
    2. Zhong Z,
    3. Zheng X,
    4. Li M,
    5. Ichim TE,
    6. Chin JL,
    7. Min WP
    : Tumor exosomes expressing Fas ligand mediate CD8+ T-cell apoptosis. Blood Cell Mol Dis 35: 169-173, 2005.
    OpenUrlCrossRefPubMed
    1. Hao S,
    2. Yuan J,
    3. Xiang J
    : Nonspecific CD4+ T-cells with uptake of antigen-specific dendritic cell-released exosomes stimulate antigen-specific CD8+ CTL responses and long-term T cell memory. J Leukocyte Biol 82: 829-838, 2007.
    OpenUrlAbstract/FREE Full Text
    1. Yu S,
    2. Liu C,
    3. Su K,
    4. Wang J,
    5. Liu Y,
    6. Zhang L,
    7. Li C,
    8. Cong Y,
    9. Kimberly R,
    10. Grizzle WE,
    11. Falkson C,
    12. Zhang HG
    : Tumor exosomes inhibit differentiation of bone marrow dendritic cells. J Immunol 178: 6867-6875, 2007.
    OpenUrlAbstract/FREE Full Text
  37. ↵
    1. Mitchell JP,
    2. Court J,
    3. Mason MD,
    4. Tabi Z,
    5. Clayton A
    : Increased exosome production from tumour cell cultures using the Integra CELLine culture system. J Immunol Methods 335: 98-105, 2008.
    OpenUrlCrossRefPubMed
  38. ↵
    1. Vukmanovic-Stejic M,
    2. Zhang Y,
    3. Cook JE,
    4. Fletcher JM,
    5. McQuaid A,
    6. Masters JE,
    7. Rustin MH,
    8. Taams LS,
    9. Beverley PC,
    10. Macallan DC,
    11. Akbar AN
    : Human CD4+ CD25hi FOXP3+ regulatory T-cells are derived by rapid turnover of memory populations in vivo. J Clin Invest 116: 2423-2433, 2006.
    OpenUrlCrossRefPubMed
  39. ↵
    1. Gandhi R,
    2. Anderson DE,
    3. Weiner HL
    : Cutting Edge: Immature human dendritic cells express latency-associated peptide and inhibit T cell activation in a TGF-β-dependent manner. J Immunol 178: 4017-4021, 2007.
    OpenUrlAbstract/FREE Full Text
  40. ↵
    1. Wang GJ,
    2. Liu Y,
    3. Qin A,
    4. Shah SV,
    5. Deng ZB,
    6. Xiang X,
    7. Cheng Z,
    8. Liu C,
    9. Wang J,
    10. Zhang L,
    11. Grizzle WE,
    12. Zhang HG
    : Thymus exosomes-like particles induce regulatory T-cells. J Immunol 181: 5242-5248, 2008.
    OpenUrlAbstract/FREE Full Text
  41. ↵
    1. Longhi MS,
    2. Hussain MJ,
    3. Mitry RR,
    4. Arora SK,
    5. Mieli-Vergani G,
    6. Vergani D,
    7. Ma Y
    : Functional study of CD4+CD25+ regulatory T-cells in health and autoimmune hepatitis. J Immunol 176: 4484-4491, 2006.
    OpenUrlAbstract/FREE Full Text
  42. ↵
    1. Yagi H,
    2. Nomura T,
    3. Nakamura K,
    4. Yamazaki S,
    5. Kitawaki T,
    6. Hori S,
    7. Maeda M,
    8. Onodera M,
    9. Uchiyama T,
    10. Fujii S,
    11. Sakaguchi S
    : Crucial role of FOXP3 in the development and function of human CD25+CD4+ regulatory T-cells. Int Immunol 16: 1643-1656, 2004.
    OpenUrlAbstract/FREE Full Text
  43. ↵
    1. Annunziato F,
    2. Cosmi L,
    3. Liotta F,
    4. Lazzeri E,
    5. Manetti R,
    6. Vanini V,
    7. Romagnani P,
    8. Maggi E,
    9. Romagnani S
    : Phenotype, localization, and mechanism of suppression of CD4+CD25+ human thymocytes. J Exp Med 196: 379-387. 2002.
    OpenUrlAbstract/FREE Full Text
Back to top

In this issue

Print
Download PDF
Article Alerts
Email Article
Citation Tools
Reprints and Permissions
Surface-bound TGF-β1 on Effusion-derived Exosomes Participates in Maintenance of Number and Suppressive Function of Regulatory T-Cells in Malignant Effusions
JUNJI WADA, HIDEYA ONISHI, HIROYUKI SUZUKI, AKIO YAMASAKI, SHUNTARO NAGAI, TAKASHI MORISAKI, MITSUO KATANO
Anticancer Research Sep 2010, 30 (9) 3747-3757;
Twitter logo Facebook logo Mendeley logo

Jump to section