The timing of TGF-β inhibition affects the generation of antigen-specific CD8+ T Cells
© Quatromoni et al.; licensee BioMed Central Ltd. 2013
Received: 23 April 2013
Accepted: 2 July 2013
Published: 17 July 2013
Transforming growth factor (TGF)-β is a potent immunosuppressive cytokine necessary for cancer growth. Animal and human studies have shown that pharmacologic inhibition of TGF-β slows the growth rate of established tumors and occasionally eradicates them altogether. We observed, paradoxically, that inhibiting TGF-β before exposing animals to tumor cells increases tumor growth kinetics. We hypothesized that TGF-β is necessary for the anti-tumor effects of cytotoxic CD8+ T lymphocytes (CTLs) during the early stages of tumor initiation.
BALB/c mice were pretreated with a blocking soluble TGF-β receptor (sTGF-βR, TGF-β-blockade group, n=20) or IgG2a (Control group, n=20) before tumor inoculation. Tumor size was followed for 6 weeks. In vivo lymphocyte assays and depletion experiments were then performed to investigate the immunological basis of our results. Lastly, animals were pretreated with either sTGF-βR (n=6) or IgG2a (n=6) prior to immunization with an adenoviral vector encoding the human papillomavirus E7 gene (Ad.E7). One week later, flow cytometry was utilized to measure the number of splenic E7-specific CD8+ T cells.
Inhibition of TGF-β before the injection of tumor cells resulted in significantly larger average tumor volumes on days 11, 17, 22, 26 and 32 post tumor-inoculation (p < 0.05). This effect was due to the inhibition of CTLs, as it was not present in mice with severe combined immunodeficiency (SCID) or those depleted of CD8+ T cells. Furthermore, pretreatment with sTGF-βR inhibited tumor-specific CTL activity in a Winn Assay. Tumors grew to a much larger size when mixed with CD8+ T cells from mice pretreated with sTGF-βR than when mixed with CD8+ T cells from mice in the control group: 96 mm3 vs. 22.5 mm3, respectively (p < 0.05). In addition, fewer CD8+ T cells were generated in Ad.E7-immunized mice pretreated with sTGF-βR than in mice from the control group: 0.6% total CD8+ T cells vs. 1.9%, respectively (p < 0.05).
These studies provide the first in vivo evidence that TGF-β may be necessary for anti-tumor immune responses in certain cancers. This finding has important implications for our understanding of anti-tumor immune responses, the role of TGF-β in the immune system, and the future development of TGF-β inhibiting drugs.
KeywordsMalignant mesothelioma Tumor immunology Immune suppression TGF-β CD8+ Cytotoxic T cell
Transforming growth factor (TGF)-β is a multifunctional cytokine that is capable of either stimulating or inhibiting growth and differentiation of a wide range of cell types, including many of those in the immune system[1–3]. Understanding the role of TGF-β in tumor biology is important to both basic science and translational medicine[4–6].
TGF-β functions primarily as an immunosuppressive cytokine in the tumor microenvironment due to its ability to interfere with the generation, expansion, and function of anti-tumor immune cells[2–4, 7]. In a number of in vitro and ex vivo studies, TGF-β has been associated with the suppression of growth and/or activity of T cells[8–12], NK cells[6, 13], and dendritic cells[14–16]. The current in vivo evidence further supports this hypothesis; using a number of approaches that include anti-TGF-β antibodies, soluble receptors, or TGF-β-binding proteins[6, 17], translational investigators have consistently reported that the blockade of TGF-β is therapeutically useful in a number of murine tumor systems, including renal cell cancer, melanoma, hepatocellular carcinoma, and glioma.
The purpose of this study is to further characterize the role of TGF-β-inhibition in tumorigenesis. The findings of these studies have important implications for our overall understanding of the generation of anti-tumor immune responses, the role of TGF-β in the immune system, and the future use and development of drugs that inhibit TGF-β.
Pathogen-free female BALB/c and C57BL/6 mice (6–8 weeks old; weight ~20-25 g) were purchased from Taconic Labs (Germantown, NY). CB-17 SCID mice (6–8 weeks old; weight ~20-25 g) were bred at the Wistar Institute (Philadelphia, PA). All mice were maintained in a pathogen-free animal facility for at least 1 week before each experiment. The animal use committees of the Wistar Institute and University of Pennsylvania approved all protocols in compliance with the Guide for the Care and Use of Laboratory Animals.
Four murine tumor cell lines were investigated in this study: the AB12 and AB-1 mesothelioma cell lines, the TC-1 non-small-cell lung carcinoma cell line, and the L1C2 bronchoalveolar carcinoma cell line. The non-malignant mink lung epithelial cells (MLECs) were also investigated. The AB12 and AB-1 cell lines were obtained from Dr. Bruce Robinson. These lines were derived in BALB/c mice and grow well as flank tumors in this model. The ability of these lines to secrete TGF-β spontaneously in culture has been studied in detail. AB12 cells secrete large amounts of TGF-β (462 pmol/106 cells/24 hours), mostly in its latent form. AB-1 cells, on the other hand, do not secrete significant quantities of TGF-β. The TC-1 cell line was generated by transduction of C57BL/6 primary lung epithelial cells with a retroviral vector expressing HPV16 E6/E7 plus a retrovirus expressing activated c-Ha-ras. This line is highly tumorigenic in C57BL/6 mice and grows well as flank tumors in this model. The L1C2 cell line, obtained from the American Type Culture Collection (Rockville, MD), is highly tumorigenic in BALB/c mice and grows well as flank tumors in this model. MLECs, previously transfected with a plasminogen activator inhibitor-1 (PAI-1) promoter-luciferase construct, were obtained from Dr. Daniel Rifkin.
AB12, AB-1, L1C2, and MLECs were cultured and maintained in high glucose Dulbecco’s modified Eagle’s medium (DMEM; Mediatech, Washington DC) supplemented with 10% fetal bovine serum (FBS; Georgia Biotechnology, Atlanta GA), 100 units/mL penicillin, 100 μg/mL streptomycin, and 2 mM glutamine. TC-1 was cultured in in RPMI 1640 (Life Technologies, Inc.) supplemented with 10% FBS, 100 units/mL penicillin, 100 μg/mL streptomycin, and 2 mM glutamine. All cell lines were regularly tested and maintained negative for Mycoplasma species.
Quantitative TGF-β bioassay
TGF-β production by the tumor cell lines was quantified using a highly sensitive and specific, nonradioactive, bioassay. This bioassay is based on the ability of TGF-β to induce PAI-1 expression. Briefly, MLECs stably transfected with a construct containing the human PAI-1 promoter fused to the firefly luciferase reporter gene were suspended in DMEM containing 10% FBS and seeded in 96-well plates at a density of 1.6×104 cells per well. Samples and standards (human platelet-derived TGF-β1 protein; BD Biosciences PharMingen, San Diego, CA) were added in triplicate to the plate of MLECs and incubated for 16 hours at 37°C in a 5% CO2 incubator. Cells were then lysed with 1x cell-lysis buffer (Analytical Luminescence, Promega, Madison, WI) and the lysates were transferred to a 96-well plate. Both substrate A and substrate B (Analytical Luminescence, Promega) were then added to the samples. Luciferase activity (a proxy for quantity of TGF-β in the sample) was measured using an ML1000 luminometer (Dynatech Laboratories Inc., Alexandria, VA) and reported as relative light units (RLU).
Soluble TGF-β inhibitor
The soluble recombinant murine TGF-β type II-murine Fc: IgG2a chimeric protein (sTGF-βR) has previously been described[22, 28]. This chimeric protein binds and inhibits TGF-β1 and TGF-β3 in the 1 nM range and has a half-life in mouse plasma of 14 days. Previous studies have shown biological effects at 1 mg/kg, 2 mg/kg, and 5 mg/kg. Based on these reports, we injected sTGF-βR at a concentration of 1.0 mg/kg in all of our experiments. Murine IgG2a antibody was used as a control and injected at the same concentration. The use of murine IgG2a as a control has been described in previous studies[29, 31].
Animal tumor models
To confirm the effect of sTGF-βR on established tumors, we injected BALB/c mice in 1 flank with 1×106 AB12 tumor cells and then initiated treatment with sTGF-βR (TGF-β-blockade group, n=5) or mouse IgG2a (Control group, n=5) when the tumors reached a minimal volume of 100 mm3 (approximately 10 days after tumor cell-inoculation). Animals in the TGF-β-blockade group received 1 intraperitoneal (IP) injection of sTGF-βR, once every 3 days, for a total of 6 doses (a 16-day course). Control animals received murine IgG2a according to the same schedule. We then followed tumor burden with serial estimates of tumor volume.
To test the efficacy of pretreatment with sTGF-βR, we administered sTGF-βR or IgG2a 2 days before inoculation of 1×106 AB12, AB-1, L1C2, or TC-1 tumor cells into the flank of each animal. The TGF-β-blockade group received 1 IP injection of sTGF-βR, once every 3 days, for a total of 3 doses (a 7-day course). The control group received murine IgG2a according to the same schedule. We then followed tumor burden with serial estimates of tumor volume. As part of our investigation into the basis of our results, this protocol was subsequently implemented in SCID animals using AB12 cells.
Lastly, we created a reproducible animal model of metastatic disease to study sTGF-βR in this context. First, we injected 1×106 AB12 tumor cells into the right flank of animals. When the tumors reached a minimal volume of 100 mm3, we initiated treatment with sTGF-βR or IgG2a: animals received 1 injection, once every 3 days. After 3 doses of either sTGF-βR or IgG2a, 1×106 AB12 cells were inoculated into the opposite flank, thus modeling a “metastatic focus”. After tumor re-challenge, 3 additional doses of sTGF-βR or IgG2a were administered. We then followed tumor burden in the primary and secondary inoculation sites with serial estimates of tumor volume.
In all instances, tumor volume was calculated according to the formula (π*long axis*short axis*short axis)/6, as described previously. We measured tumor volume at least twice weekly. Unless otherwise mentioned, each control or experimental group had a minimum of 5 mice. Each experiment was repeated at least once.
Flow cytometry on tumor-infiltrating lymphocytes and lymphocytes in the tumor-draining lymph nodes
To study tumor-infiltrating lymphocytes (TILs) and lymphocytes in the tumor-draining lymph nodes (TDLNs), we compared 3 groups: 1 non tumor-bearing group (Naïve group, n=9) and 2 groups of tumor-bearing animals (Control group, n=9 and TGF-β-blockade group, n=9). The naïve group consisted of BALB/c mice that received a one-time IP injection of BD Matrigel™ matrix (BD Biosciences, Bedford, MA) without tumor cells into both flanks. The control group consisted of BALB/c mice that were injected with 1x106 AB12 cells in 250 μL of serum-free DMEM media mixed with 250 μL of BD Matrigel™ matrix into both flanks. Two days before tumor cell-inoculation and once every 3 days thereafter, for a total of 3 doses, these mice received IP injections of IgG2a (Day −2, Day 1, and Day 4). The TGF-β-blockade group consisted of BALB/c mice that were injected with 1×106 AB12 cells in 250 μL of serum free DMEM media mixed with 250 μL of BD Matrigel™ matrix into both flanks. Two days before tumor cell-inoculation and once every 3 days thereafter, for a total of 3 doses, these mice received IP injections of sTGF-βR (Day −2, Day 1, and Day 4).
Two, 4, and 7 days after tumor cell-inoculation, tumors and bilateral inguinal lymph nodes from both the control (n=3) and TGF-β-blockade (n=3) groups were harvested. Single-cell suspensions were generated by mincing these tissues on ice and subsequently filtering them through a 70μm BD Falcon™ cell strainer (BD Biosciences Discovery Labware, Bedford, MA). These populations were then stained with the following antibodies: allophycocyanin (APC) conjugated to rat anti-mouse CD45 or CD8a (BD Biosciences PharMingen, San Diego, CA), fluorescein isothiocyanate (FITC) conjugated to rat anti-mouse CD4, CD11c (BD Biosciences PharMingen, San Diego, CA), or MHC class I (34-1-2s; eBioscience, San Diego, CA), and phycoerythrin (PE) conjugated to rat anti-mouse CD8a, CD11c, CD86 (BD Biosciences PharMingen, San Diego, CA), or MHC class II (M5/114.15.2; eBioscience). We then used flow cytometry to analyze these populations (FACScalibur, Becton-Dickinson, Mountain View, CA).
Of note, the rationale for inoculation of AB12 tumor cells in a Matrigel™ matrix for this experiment was based on the difficulty of generating single cells suspensions from 2-day old tumors.
Animal vaccine models
To determine if TGF-β-inhibition affects the ability of mice to generate antigen-specific CD8+ T cells, we studied the effect of pretreatment with sTGF-βR in animals immunized against the human papillomavirus E7 protein using an adenoviral vaccine (Ad.E7). First, 6-to-8 week-old female C57BL/6 animals were treated with either sTGF-βR or IgG2a. Two days later, these animals were immunized with Ad.E7 via subcutaneous injection of 1×109 plaque-forming units (pfu), as previously described[32, 33]. Seven days after immunization, splenocytes were isolated from each group and analyzed by flow cytometry to establish the percentage of E7-specific CD8+ T cells (as determined by tetramer binding).
To determine if TGF-β-inhibition affects the period of viability of established antigen-specific CD8+ T cells, 6-to-8 week-old female C57BL/6 mice were immunized with 1×109 pfu of Ad.E7 and treated 7 days later with either sTGF-βR or IgG2a. Then, 7 days after treatment (14 days after immunization), splenocytes from each group (n=3) were analyzed by flow cytometry to establish the percentage of E7-specific CD8+ T cells (as determined by tetramer binding).
Unless otherwise mentioned, each control group or experimental group had a minimum of 3 mice. Each experiment was repeated at least once.
Analysis of E7-specific CD8+ T cells by flow cytometry
Tetramer staining of spleen cells was performed as previously described. Single cell suspensions were generated by filtering spleens through a 70 μm BD Falcon™ cell strainer and then incubating the isolated cells for 15 minutes with BD PharM Lyse™, an ammonium chloride-based red blood cell-lysing reagent (BD Biosciences PharMingen, San Diego, CA). The remaining viable cells were incubated with anti-CD16 mAb (eBioscience, San Diego, CA) for 30 minutes to block non-specific binding of spleen cells to the Fc portion of test-antibody. Then, the spleen cells were stained FITC-conjugated anti-CD8a antibody (BD Biosciences PharMingen) and APC-conjugated E7 tetramer for 30 minutes and 1.5 hours, respectively.
APC-labeled H-2Db tetramers loaded with E7 peptide (RAHYNIVTF) were obtained from the National Institute of Allergy and Infectious Diseases (eNIAID, Bethesda, MD) tetramer core. Flow cytometry was performed using a DakoCytomation CyAn (DakoCytomation, Fort Collins, CO).
In Vivo depletion of CD8+ T cells
To deplete CD8+ T cells prior to, and during, treatments with sTGF-βR or IgG2a in our AB12 tumor model, mice received 200 μg IP injections of monoclonal antibody purified from the anti-CD8 hybridoma 53–6.7 (obtained from the American Type Culture Collection). Mice received injections both 1 and 3 days prior to inoculation with AB12 tumor cells. Thereafter, a maintenance dose was administered once every 7 days throughout the experimental period to ensure continued depletion. CD8+ T cell depletion was confirmed by flow cytometric analysis of spleen cells (preparation described above) at the time of tumor injection and weekly thereafter.
Evaluation of effector function (Winn Assay)
We performed Winn Assays as previously described. This assay allows for assessment of anti-tumor activity of immune effector cells in vivo without the need for ex vivo stimulation. We first prepared a single-cell suspension of splenocytes as described above. Then, CD8+ T cells were isolated from this suspension using the MACs system (Miltenyi Biotec, Auburn, CA). This cell population contained greater than 90% CD8+ T cells as determined by flow cytometry (data not shown). The CD8+ T cell-enriched populations from non tumor-bearing (naïve), IgG2a-pretreated animals (control), or sTGF-βR-pretreated animals (TGF-β-blockade) were admixed with viable AB12 tumor cells at a ratio of 3 purified CD8+ T cells per 1 tumor cell. This ratio has previously been determined to be optimal for detecting positive and negative effects. This mixture was then inoculated subcutaneously (SC) into the flanks of naïve BALB/c mice. Each mouse thus received a total of 0.5×106 tumor cells and 1.5×106 CD8+ T cells. Tumor growth was measured after 1 week and expressed as the mean ± standard error of the mean. Each group contained at least 5 mice unless otherwise stated.
We implemented unpaired Student’s t-tests to compare differences in continuous variables between control and experimental groups. Analysis of variance (ANOVA) with post-hoc testing was used for multiple comparisons. We considered differences statistically significant when the p-value was less than 0.05. Statistical analysis was conducted using the StatView 5.0 for Windows program (Abacus Concepts Inc., Berkeley, CA).
AB12 and TC-1 cells produce a large amount of TGF-β
To determine the level of TGF-β production by the murine cancer cell lines under investigation, we measured soluble TGF-β by the quantitative bioassay described above. AB12 and TC-1 cell lines produced more TGF-β (277.3 pg/mL and 263.3 pg/mL, respectively) than AB-1 and L1C2 (88.9 pg/mL and 92.3 pg/mL, respectively).
The administration of sTGF-βR to animals with established AB12 tumors inhibits tumor growth, while treatment before AB12 inoculation stimulates tumor growth
The increased rate of AB12 tumor growth after pretreatment with sTGF-βR is abolished in the SCID animal model
Previous reports have suggested that TGF-β acts as a direct growth-inhibitor of certain cancer cell lines[37, 38]. Neutralization of TGF-β might therefore induce more rapid growth. However, our lab has shown that TGF-β-inhibition results in neither direct stimulation nor inhibition of AB12 cell proliferation in vitro. To assess the possibility of indirect immunologically-mediated effects of TGF-β on tumor cell growth, we repeated our pretreatment studies using the AB12 cell line in the immunodeficient CB-17 SCID animal model (which lacks B and T lymphocytes). The pretreatment of SCID mice with sTGF-βR before AB12 inoculation abolished the augmentation of growth seen in BALB/c mice (Figure 2F), as tumor growth rates did not differ between mice pretreated with sTGF-βR and control mice pretreated with IgG2a. These experiments show that the increased rate of tumor growth resulting from pretreatment with sTGF-βR in the BALB/c tumor model is not the result of neutralizing direct growth-inhibiting effects of TGF-β; rather, these results support an immunologically-mediated mechanism that is dependent on the presence of B and/or T cells.
The increased rate of AB12 tumor growth after pretreatment with sTGF-βR is abolished in CD8+ T cell-depleted animals
We then designed a lymphocyte-depletion experiment to further probe the immunologic basis of our findings and determine which cells were responsible for this effect. We depleted CD8+ T cells after finding small numbers of CD4+ T cells in AB12 tumors by flow cytometry (data not shown). The pretreatment of naïve (non CD8+ T cell-depleted) BALB/c animals with sTGF-βR resulted in larger tumors compared to control animals pretreated with IgG2a (Figure 3A). At day 17, tumors in control mice were 260 mm3 compared to 350 mm3 in animals pretreated with sTGF-βR, a 34% augmentation of size (p < 0.05). However, when BALB/c mice were depleted of their CD8+ T cells, this significant difference in tumor growth rates between animals pretreated with sTGF-βR or IgG2a disappeared (Figure 3B). Mean tumor volume at day 17 in the animals pretreated with sTGF-βR was 550 mm3 compared to 520 mm3 in the control animals. This 5% difference in tumor growth was not statistically significant (p>0.05). These results, in combination with the SCID animal experiments, demonstrate that the stimulatory effect on tumor growth resulting from pretreatment with sTGF-βR relies on the presence of CD8+ T lymphocytes.
Pretreatment with sTGF-βR before AB12 tumor challenge abolished tumor-specific CTL activity
The more rapid absolute growth of AB12 tumors in SCID and CD8+ T cell-depleted mice regardless of treatment (Figure 2) suggests that the wild-type BALB/c animals mount a tumor-specific, although ultimately ineffective, CD8+ T cell response against the tumor at early time points. We have previously documented the presence of anti-tumor CTLs that arise early in the course of tumor growth and then disappear as the tumors grow to larger sizes (>500 mm3) using an in vivo tumor neutralization assay (Winn Assay)[22, 39]. In order to determine if the increased rate of AB12 tumor growth associated with sTGF-βR pretreatment was dependent on the inhibition of naturally-occurring endogenous anti-tumor CTL, we conducted a Winn Assay as outlined above.
Pretreatment with sTGF-βR before tumor challenge affects neither the migration of DCs nor their expression of CD86, MHC class I, or MHC class II
We have shown that anti-tumor CTLs develop spontaneously in small AB12 tumor-bearing mice and that these endogenous CTLs are not active when sTGF-βR is given before AB12 tumor cell-inoculation. Anti-tumor CTLs develop from naïve CD8+ T cells that are sensitized to tumor antigen when it is presented by antigen-presenting cells (mainly dendritic cells (DCs)) in TDLNs. Initial sensitization of CD8+ T cells typically requires 4 steps: migration of DCs into tumor nodules, ingestion and subsequent internal-processing of apoptotic cancer cell debris, presentation of processed peptide fragments in both MHC class I and class II complex clefts (by activated DCs), and migration of the activated DCs into TDLNs where T cell-sensitization occurs. In order to determine if pretreatment with sTGF-βR affects anti-tumor CTLs indirectly through interruption of these 4 steps, we used flow cytometry to study the effect of pretreatment with sTGF-βR on both the number of DCs and the expression of DC-activation markers (CD86, MHC class I and II) in the tumor and TDLNs.
When we compared tumors between groups, as expected, the average AB12 tumor weight at day 7 post tumor cell-inoculation in mice pretreated with sTGF-βR was significantly greater than the average tumor size in mice pretreated with IgG2a (data not shown). However, no significant differences were found in the total numbers of tumor-infiltrating CD45+ cells, DCs, or CD8+ T cells between tumor-bearing mice pretreated with sTGF-βR and tumor-bearing mice pretreated with IgG2a (Figure 5B). These findings demonstrate that the increased rate of AB12 tumor growth resulting from pretreatment with sTGF-βR is not due to an effect on the migration or activation (assessed by expression levels of CD86 and MHC class I and II) of DCs.
Administration of sTGF-βR to animals with established AB12 tumors does not increase the growth rate of secondary “metastatic” tumors
Pretreatment with sTGF-βR before immunization with Ad.E7 inhibits the generation of E7-specific CD8+ T cells
To determine if TGF-β is required to generate antigen-specific CD8+ T cells, we utilized a previously developed adenoviral vector that expresses the well-studied viral tumor antigen human papilloma virus E7 protein (Ad.E7). In this independent and more quantifiable system, we investigated how the blockade of endogenous TGF-β, at a time point before antigen immunization, affected the generation and maintenance of antigen-specific CD8+ T cells[32, 33].
The administration of sTGF-βR after E7-immunization prevents the spontaneous loss of E7-specific CD8+ T cells
Because of its multiple distinct functions in a variety of experimental models of T cell immunology, it has been difficult to develop a clear model of the in vivo roles of TGF-β[7, 40]. There is ample data to support the hypothesis that TGF-β is an immunosuppressive factor. As summarized previously[3, 8, 15, 16, 40], TGF-β has been reported to (1) inhibit T cell proliferation, CTL generation, and T cell cytokine production; (2) interfere withTH1/TH2 differentiation and the differentiation of naïve T cells towards central memory cells; and (3) inhibit dendritic cell (DC)-mediated antigen presentation by inhibiting DCs’ endocytic and phagocytic activities, preventing DC-maturation, and blocking the up-regulation of critical DC-associated co-stimulatory molecules.
In contrast, there are other studies (although fewer) that have reported that TGF-β exerts stimulatory effects on human T cells and dendritic cells. There is evidence that under some conditions, TGF-β can (1) support the generation of effector cells; (2) augment the development of memory and mature T cell populations; (3) co-stimulate the growth and maturation of CD4+ and CD8+ T cells; (4) inhibit the apoptosis of CD4+ T cells; (5) promote the in vitro development of DCs from hematopoietic progenitors; and (6) regulate the chemotaxis of DCs via regulation of chemokine receptor expression[40–46].
Based on the paradigm that TGF-β is “one of the most potent immunosuppressors described to date”, translational investigators have tried to inhibit tumor growth in animal models by blocking TGF-β production, receptor binding, or function. Using a number of approaches that include anti-TGF-β antibodies, soluble receptors, or TGF-β-binding proteins[6, 17], investigators have consistently reported that blockade of TGF-β is therapeutically useful in a number of murine tumor systems, including renal cell cancer, melanoma, hepatocellular carcinoma, and glioma.
The literature is currently unable to bridge these seemingly contradictory findings regarding TGF-β in cancer biology. The observed results likely depend on the experimental models used, the type of stimulus, the presence of other cytokines, the dose of TGF-β, the distribution of TGF-β in its latent and active form, the duration of the stimulation, and possibly, the genetic background of the cell populations studied. Regardless of the reasons, since TGF-β blocking agents are currently being developed for clinical use, it has become increasingly important to better understand the effects of TGF-β on in vivo anti-tumor immune cell function.
We observed that blockade of TGF-β with sTGF-βR before the inoculation of tumor cells (and for 3 doses afterwards) resulted in significantly enhanced tumor growth of one particular tumor cell line, the AB12 line (but not in others). This response was in marked contrast to the inhibition of tumor growth associated with administration of the same TGF-β blocking agent after the establishment of the same tumor cell line. In this study, we examined the mechanism responsible for the increased rate of AB12 tumor growth resulting from pretreatment with sTGF-βR. We demonstrated that altered anti-tumor immune responses were responsible for this augmentation of tumor growth; specifically, administration of sTGF-βR before tumor cell-inoculation resulted in the failure to generate active anti-tumor CTLs.
The specific characteristics of the relatively immunogenic tumor model used in these studies are important to understand our findings. Mesotheliomas usually result from prior asbestos exposure. They are associated with a high degree of MHC class I expression and TGF-β production. Clinically, they respond to some immune-based therapies. The mouse mesothelioma tumor cells used in this study are very similar to human tumors. When AB12 cells are injected into syngeneic BALB/c mice, their initial growth is quite slow until about 20 days, at which point their size begins to increase rapidly (Figures 2A,2E, and6A). It appears that this initial slow growth phase is due to a partially effective anti-tumor immune response mediated by endogenous, functionally active tumor antigen-specific CTLs. We have observed that AB12 tumors grow much more rapidly in SCID mice (Figure 2F), in CD8+ T cell-depleted mice (Figure 3B), and in IFNγ-knockout or IFNγ-neutralized mice (data not shown). We have also directly examined the ability of AB12 tumors to generate anti-tumor immune responses. Within 4–10 days after subcutaneous injection of AB12 tumor cells, we have detected CD8+ T cells in the spleen that have cytolytic activity. We confirmed the presence of these spontaneously generated anti-tumor CTLs in this study (Figure 4) using a Winn assay that demonstrated markedly inhibited tumor growth when tumor cells were mixed with CD8+ splenocytes from control tumor-bearing animals before inoculation into naïve non tumor-bearing animals. These anti-tumor CTLs persist until the tumor reaches a size of approximately 400 mm3 (usually about 20 days after injection). At this time, CTL activity can no longer be detected and tumor growth rate rapidly increases.
Our experiments indicate that the increased rate of AB12 tumor growth resulting from pretreatment with sTGF-βR was due to a loss of this normal, low-level, and only partially-effective anti-tumor CTL immune response. First, the growth-augmenting effects of sTGF-βR relative to IgG2a were lost in T cell-deficient SCID mice (Figure 2F) and CD8+ T cell-depleted mice (Figure 3B). Second, we showed that the inhibition of TGF-β negatively impacts the functionality of CD8+ CTLs, as the Winn assay (essentially an in vivo test of CD8+ T cell functionality) demonstrated a reduced anti-tumor response with an equivalent number of CD8+ T cells from mice pretreated with sTGF-βR compared to control animals pretreated with IgG2a (Figure 4). Together, these results implicate the inhibition of anti-tumor CD8+ CTLs as central to the augmentation of AB12 tumor growth associated with sTGF-βR pretreatment.
In addition to our tumor study, we also investigated the effect of TGF-β-blockade on the generation of active antigen-specific CTLs against a known viral tumor antigen in an independent and more quantifiable system. Pretreatment with sTGF-βR, at a time point before immunization with an adenovirus encoding the HPV E7 protein (Ad.E7), inhibited the generation of E7-specific CD8+ T cells as compared to control pretreatment with murine IgG2a. These experiments show that TGF-β is required for the generation of active CTLs, at least in models employing AB12 tumor cells or vaccination with Ad.E7.
Unfortunately, despite further investigation, the mechanism by which pretreatment with sTGF-βR inhibits CTL-activity remains unclear. Initial sensitization of CD8+ T cells typically requires 4 steps as described above. We showed that pretreatment with sTGF-βR does not decrease the activation status or the number of DCs, CD4+ T cells, or CD8+ T cells in the TDLNs or tumor beds compared to IgG2a. These data indicate that TGF-β may not be required for the migration or proliferation of DCs, CD4+ T cells, or CD8+ T cells or the activation of DCs. Although studies of expression levels of CD86, MHC class I, and MHC class II are important to evaluate the activation levels of DCs in anti-tumor immune responses, other activation markers for DCs might exist, such as ICAM-1 or B7. It may also be important to test the expression levels of accessory molecules on T lymphocytes, such as LFA-1 or CD28. Thus, the mechanism by which pretreatment with sTGF-βR stimulates the growth of tumors in our AB12 tumor model remains unclear.
Another interesting question relates to the issue of why sTGF-βR did not inhibit the generation of anti-tumor CD8+ CTL activity in other tumor models as it did in the AB12 tumor model. We explored a number of obvious explanations: low amounts of TGF-β produced, lack of tumor immunogenicity, or animal strain differences. With regard to TGF-β production, we know that AB-1 cells make very little TGF β which could explain the lack of effect in this cell line. However, the TC-1 cell line makes sizeable amounts of TGF-β and yet it is still resistant. We have also studied the L1C2 and TC-1 cell lines in the past and have shown them to be moderately or highly immunogenic, similar to the AB12 model, and able to induce anti-tumor CD8+ T cells. To address the issue of strain differences, we also studied L1C2 cells, another tumor line that grows in BALB/c mice (like AB12), and saw no response. We thus have no simple explanation for the selectivity for our observation. The tumor microenvironment is a complex ecosystem which is unique to each tumor model. Given the genetic modifications required for malignant transformation, it is likely that a myriad of factors, including various cytokines, chemokines, other soluble factors, and even cell-bound mediators play significant roles in tumor development and in the interaction with the host’s immune system. The key point is that this stimulation of tumor growth after early TGF-β inhibition can occur in at least one animal model and thus should be carefully looked for in future clinical trials. Additional ongoing research that identifies the key factors responsible for this effect will be needed.
In conclusion, this paper provides the first in vivo evidence, to our knowledge, that the blockade of TGF-β inhibits the initial generation of functionally active anti-tumor CTLs and antigen-specific CD8+ T cells after Ad.E7 vaccination. These findings support the novel hypothesis that, at least under some circumstances, TGF-β is required for the generation of active anti-tumor CTLs. Given the complexity of the in vivo anti-tumor immune response, we have not yet defined the step at which TGF-β-blockade inhibited CTL activation. Although pretreatment with sTGF-βR may not be involved in the migration of immune cells (Figure 5), possible mechanisms include inhibition of either antigen presentation by DCs or other antigen-presenting cells, T cell differentiation, or generation of memory/effector cells. Experiments to differentiate among these potential mechanisms are in progress.
The implications of our findings are significant. From an immunological standpoint, our results support the complex in vivo functions of TGF-β and suggest a potentially new paradigm for its role in the generation of CD8+ memory and/or effector cells. Since it is extremely difficult to model all the variables that factor into an in vivo immune response, it will be very important to study the effects of TGF-β manipulation in a variety of animal models. From a more practical standpoint, these results may help guide the use of TGF-β inhibitors. Given our observation that TGF-β is required for anti-tumor immune responses, along with other data showing that TGF-β-blockade can enhance carcinogenesis through tumor cell-intrinsic mechanisms[48–50], the use of TGF-β inhibitors in a chemopreventive mode should be undertaken with caution. On the other hand, the use of TGF-β inhibitors in patients with established tumors might prove very useful. One encouraging finding from our study was that the blockade of TGF-β did not lead to increased growth rates at secondary (“metastatic”) sites. These data support the hypothesis that blockade of TGF-β does not enhance tumor growth after anti-tumor CTLs have been induced.
We also have evidence from the Ad.E7 model that TGF-β-blockade promotes the persistence of established antigen-specific CD8+ T cells that were induced by immunization at a time point prior to sTGF-βR administration (Figure 8B). While the percentage of E7-specific CD8+ T cells in control animals decreased significantly 1 week after IgG2a administration, the percentage of E7-specific CD8+ T cells in animals treated with sTGF-βR remained stable at the same time point. These results thus support the use of TGF-β-inhibition in patients with established tumors.
In summary, we present an in vivo tumor model demonstrating that the timing of TGF-β blockade can determine whether tumor growth is inhibited or enhanced. These experiments highlight the pleomorphic effects of TGF-β and emphasize the importance of careful patient-selection for novel TGF-β inhibitors.
Transforming growth factor-β
Soluble Type II TGF-β receptor
Fluorescence-activated cell sorting
Severe combined immunodeficiency
Analysis of variance
Cytotoxic CD8+ T cells
Adenoviral vector encoding the human papillomavirus E7 gene
Non-malignant mink lung epithelial cells
Plasminogen activator inhibitor-1
Fetal bovine serum
Dulbecco’s modified Eagle’s medium
Relative light units
Tumor-draining lymph nodes
Major histocompatibility complex
Mean fluorescence intensity.
The authors would like to thank Ellen Pure and Yvonne Paterson for reading our manuscript.
This study was supported by the Lavin Family Foundation (SS), The The OneBreath® Clinical Research Award in Lung Cancer from the CHEST Foundation (SS), and NCI PO1 CA 66726 (ES, SA).
JQ was supported by the National Center for Research Resources, Grant TL1RR024133, and is now at the National Center for Advancing Translational Sciences, Grant TL1R000138. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
- Letterio JJ, Roberts AB: Regulation of immune responses by TGF-beta. Annu Rev Immunol. 1998, 16: 137-161. 10.1146/annurev.immunol.16.1.137.View ArticlePubMed
- Achyut BR, Yang L: Transforming growth factor-beta in the gastrointestinal and hepatic tumor microenvironment. Gastroenterology. 2011, 141: 1167-1178. 10.1053/j.gastro.2011.07.048.View ArticlePubMed
- Kubiczkova L, Sedlarikova L, Hajek R, Sevcikova S: TGF-beta - an excellent servant but a bad master. J Transl Med. 2012, 10: 183-5876. 10.1186/1479-5876-10-183. 10-183PubMed CentralView ArticlePubMed
- de Visser KE, Kast WM: Effects of TGF-beta on the immune system: implications for cancer immunotherapy. Leukemia. 1999, 13: 1188-1199. 10.1038/sj.leu.2401477.View ArticlePubMed
- Yang L: TGFbeta, a potent regulator of tumor microenvironment and host immune response, implication for therapy. Curr Mol Med. 2010, 10: 374-380. 10.2174/156652410791317039.View ArticlePubMed
- Wrzesinski SH, Wan YY, Flavell RA: Transforming growth factor-beta and the immune response: implications for anticancer therapy. Clin Cancer Res. 2007, 13: 5262-5270. 10.1158/1078-0432.CCR-07-1157.View ArticlePubMed
- Yang L, Pang Y, Moses HL: TGF-beta and immune cells: an important regulatory axis in the tumor microenvironment and progression. Trends Immunol. 2010, 31: 220-227. 10.1016/j.it.2010.04.002.PubMed CentralView ArticlePubMed
- Gorelik L, Flavell RA: Transforming growth factor-beta in T-cell biology. Nat Rev Immunol. 2002, 2: 46-53. 10.1038/nri704.View ArticlePubMed
- Somasundaram R, Jacob L, Swoboda R, Caputo L, Song H, Basak S, Monos D, Peritt D, Marincola F, Cai D, Birebent B, Bloome E, Kim J, Berencsi K, Mastrangelo M, Herlyn D: Inhibition of cytolytic T lymphocyte proliferation by autologous CD4+/CD25+ regulatory T cells in a colorectal carcinoma patient is mediated by transforming growth factor-beta. Cancer Res. 2002, 62: 5267-5272.PubMed
- Nakamura K, Kitani A, Strober W: Cell contact-dependent immunosuppression by CD4(+)CD25(+) regulatory T cells is mediated by cell surface-bound transforming growth factor beta. J Exp Med. 2001, 194: 629-644. 10.1084/jem.194.5.629.PubMed CentralView ArticlePubMed
- Carrier Y, Yuan J, Kuchroo VK, Weiner HL: Th3 cells in peripheral tolerance. II. TGF-beta-transgenic Th3 cells rescue IL-2-deficient mice from autoimmunity. J Immunol. 2007, 178: 172-178.View ArticlePubMed
- Fahlen L, Read S, Gorelik L, Hurst SD, Coffman RL, Flavell RA, Powrie F: T cells that cannot respond to TGF-beta escape control by CD4(+)CD25(+) regulatory T cells. J Exp Med. 2005, 201: 737-746. 10.1084/jem.20040685.PubMed CentralView ArticlePubMed
- Rook AH, Kehrl JH, Wakefield LM, Roberts AB, Sporn MB, Burlington DB, Lane HC, Fauci AS: Effects of transforming growth factor beta on the functions of natural killer cells: depressed cytolytic activity and blunting of interferon responsiveness. J Immunol. 1986, 136: 3916-3920.PubMed
- Tian F, Wang L, Qin W, Wang F, Song B, Li Y, Wen W, Zhang Z, Zhu K, Zhang Q, Lee C, Zhong W, Guo Y, Wang H: Vaccination with transforming growth factor-beta insensitive dendritic cells suppresses pulmonary metastases of renal carcinoma in mice. Cancer Lett. 2008, 271: 333-341. 10.1016/j.canlet.2008.06.025.View ArticlePubMed
- Ito M, Minamiya Y, Kawai H, Saito S, Saito H, Nakagawa T, Imai K, Hirokawa M, Ogawa J: Tumor-derived TGFbeta-1 induces dendritic cell apoptosis in the sentinel lymph node. J Immunol. 2006, 176: 5637-5643.View ArticlePubMed
- Kobie JJ, Wu RS, Kurt RA, Lou S, Adelman MK, Whitesell LJ, Ramanathapuram LV, Arteaga CL, Akporiaye ET: Transforming growth factor beta inhibits the antigen-presenting functions and antitumor activity of dendritic cell vaccines. Cancer Res. 2003, 63: 1860-1864.PubMed
- Dumont N, Arteaga CL: Targeting the TGF beta signaling network in human neoplasia. Cancer Cell. 2003, 3: 531-536. 10.1016/S1535-6108(03)00135-1.View ArticlePubMed
- Wang L, Wen W, Yuan J, Helfand B, Li Y, Shi C, Tian F, Zheng J, Wang F, Chen L, Liang L, Zhou L, Lee C, Chen Z, Guo Y, Wang H, Zhang Q, Qin W: Immunotherapy for human renal cell carcinoma by adoptive transfer of autologous transforming growth factor beta-insensitive CD8+ T cells. Clin Cancer Res. 2010, 16: 164-173. 10.1158/1078-0432.CCR-09-1758.View ArticlePubMed
- Quatromoni JG, Wang Y, Vo DD, Morris LF, Jazirehi AR, McBride W, Chatila T, Koya RC, Economou JS: T cell receptor (TCR)-transgenic CD8 lymphocytes rendered insensitive to transforming growth factor beta (TGFbeta) signaling mediate superior tumor regression in an animal model of adoptive cell therapy. J Transl Med. 2012, 10: 127-5876. 10.1186/1479-5876-10-127. 10-127PubMed CentralView ArticlePubMed
- Maggard M, Meng L, Ke B, Allen R, Devgan L, Imagawa DK: Antisense TGF-beta2 immunotherapy for hepatocellular carcinoma: treatment in a rat tumor model. Ann Surg Oncol. 2001, 8: 32-37.PubMed
- Witham TF, Villa L, Yang T, Pollack IF, Okada H, Robbins PD, Chambers WH: Expression of a soluble transforming growth factor-beta (TGFbeta) receptor reduces tumorigenicity by regulating natural killer (NK) cell activity against 9L gliosarcoma in vivo. J Neurooncol. 2003, 64: 63-69.PubMed
- Suzuki E, Kapoor V, Cheung HK, Ling LE, DeLong PA, Kaiser LR, Albelda SM: Soluble type II transforming growth factor-beta receptor inhibits established murine malignant mesothelioma tumor growth by augmenting host antitumor immunity. Clin Cancer Res. 2004, 10: 5907-5918. 10.1158/1078-0432.CCR-03-0611.View ArticlePubMed
- Davis MR, Manning LS, Whitaker D, Garlepp MJ, Robinson BW: Establishment of a murine model of malignant mesothelioma. Int J Cancer. 1992, 52: 881-886. 10.1002/ijc.2910520609.View ArticlePubMed
- Odaka M, Sterman DH, Wiewrodt R, Zhang Y, Kiefer M, Amin KM, Gao GP, Wilson JM, Barsoum J, Kaiser LR, Albelda SM: Eradication of intraperitoneal and distant tumor by adenovirus-mediated interferon-beta gene therapy is attributable to induction of systemic immunity. Cancer Res. 2001, 61: 6201-6212.PubMed
- Fitzpatrick DR, Bielefeldt-Ohmann H, Himbeck RP, Jarnicki AG, Marzo AL, Robinson BW: Transforming growth factor-beta: antisense RNA-mediated inhibition affects anchorage-independent growth, tumorigenicity and tumor-infiltrating T-cells in malignant mesothelioma. Growth Factors. 1994, 11: 29-44. 10.3109/08977199409015049.View ArticlePubMed
- Lin KY, Guarnieri FG, Staveley-O’Carroll KF, Levitsky HI, August JT, Pardoll DM, Wu TC: Treatment of established tumors with a novel vaccine that enhances major histocompatibility class II presentation of tumor antigen. Cancer Res. 1996, 56: 21-26.PubMed
- Abe M, Harpel JG, Metz CN, Nunes I, Loskutoff DJ, Rifkin DB: An assay for transforming growth factor-beta using cells transfected with a plasminogen activator inhibitor-1 promoter-luciferase construct. Anal Biochem. 1994, 216: 276-284. 10.1006/abio.1994.1042.View ArticlePubMed
- Zheng H, Wang J, Koteliansky VE, Gotwals PJ, Hauer-Jensen M: Recombinant soluble transforming growth factor beta type II receptor ameliorates radiation enteropathy in mice. Gastroenterology. 2000, 119: 1286-1296. 10.1053/gast.2000.19282.View ArticlePubMed
- Smith JD, Bryant SR, Couper LL, Vary CP, Gotwals PJ, Koteliansky VE, Lindner V: Soluble transforming growth factor-beta type II receptor inhibits negative remodeling, fibroblast transdifferentiation, and intimal lesion formation but not endothelial growth. Circ Res. 1999, 84: 1212-1222. 10.1161/01.RES.84.10.1212.View ArticlePubMed
- George J, Roulot D, Koteliansky VE, Bissell DM: In vivo inhibition of rat stellate cell activation by soluble transforming growth factor beta type II receptor: a potential new therapy for hepatic fibrosis. Proc Natl Acad Sci USA. 1999, 96: 12719-12724. 10.1073/pnas.96.22.12719.PubMed CentralView ArticlePubMed
- Lammerts E, Roswall P, Sundberg C, Gotwals PJ, Koteliansky VE, Reed RK, Heldin NE, Rubin K: Interference with TGF-beta1 and -beta3 in tumor stroma lowers tumor interstitial fluid pressure independently of growth in experimental carcinoma. Int J Cancer. 2002, 102: 453-462. 10.1002/ijc.10722.View ArticlePubMed
- He Z, Wlazlo AP, Kowalczyk DW, Cheng J, Xiang ZQ, Giles-Davis W, Ertl HC: Viral recombinant vaccines to the E6 and E7 antigens of HPV-16. Virology. 2000, 270: 146-161. 10.1006/viro.2000.0271.View ArticlePubMed
- Kowalczyk DW, Wlazlo AP, Giles-Davis W, Kammer AR, Mukhopadhyay S, Ertl HC: Vaccine-induced CD8+ T cells eliminate tumors by a two-staged attack. Cancer Gene Ther. 2003, 10: 870-878. 10.1038/sj.cgt.7700653.View ArticlePubMed
- Lamikanra A, Pan ZK, Isaacs SN, Wu TC, Paterson Y: Regression of established human papillomavirus type 16 (HPV-16) immortalized tumors in vivo by vaccinia viruses expressing different forms of HPV-16 E7 correlates with enhanced CD8(+) T-cell responses that home to the tumor site. J Virol. 2001, 75: 9654-9664. 10.1128/JVI.75.20.9654-9664.2001.PubMed CentralView ArticlePubMed
- Winn HJ: In vivo methods for the assessment of antibody-mediated tumor immunity. Natl Cancer Inst Monogr. 1972, 35: 13-18.PubMed
- DeLong P, Tanaka T, Kruklitis R, Henry AC, Kapoor V, Kaiser LR, Sterman DH, Albelda SM: Use of cyclooxygenase-2 inhibition to enhance the efficacy of immunotherapy. Cancer Res. 2003, 63: 7845-7852.PubMed
- Wu SP, Sun LZ, Willson JK, Humphrey L, Kerbel R, Brattain MG: Repression of autocrine transforming growth factor beta 1 and beta 2 in quiescent CBS colon carcinoma cells leads to progression of tumorigenic properties. Cell Growth Differ. 1993, 4: 115-123.PubMed
- Wu SP, Theodorescu D, Kerbel RS, Willson JK, Mulder KM, Humphrey LE, Brattain MG: TGF-beta 1 is an autocrine-negative growth regulator of human colon carcinoma FET cells in vivo as revealed by transfection of an antisense expression vector. J Cell Biol. 1992, 116: 187-196. 10.1083/jcb.116.1.187.View ArticlePubMed
- Kruklitis RJ, Singhal S, Delong P, Kapoor V, Sterman DH, Kaiser LR, Albelda SM: Immuno-gene therapy with interferon-beta before surgical debulking delays recurrence and improves survival in a murine model of malignant mesothelioma. J Thorac Cardiovasc Surg. 2004, 127: 123-130. 10.1016/j.jtcvs.2003.08.034.View ArticlePubMed
- Cerwenka A, Swain SL: TGF-beta1: immunosuppressant and viability factor for T lymphocytes. Microbes Infect. 1999, 1: 1291-1296. 10.1016/S1286-4579(99)00255-5.View ArticlePubMed
- Strobl H, Knapp W: TGF-beta1 regulation of dendritic cells. Microbes Infect. 1999, 1: 1283-1290. 10.1016/S1286-4579(99)00256-7.View ArticlePubMed
- Cerwenka A, Kovar H, Majdic O, Holter W: Fas- and activation-induced apoptosis are reduced in human T cells preactivated in the presence of TGF-beta 1. J Immunol. 1996, 156: 459-464.PubMed
- Zhang X, Giangreco L, Broome HE, Dargan CM, Swain SL: Control of CD4 effector fate: transforming growth factor beta 1 and interleukin 2 synergize to prevent apoptosis and promote effector expansion. J Exp Med. 1995, 182: 699-709. 10.1084/jem.182.3.699.View ArticlePubMed
- Ayoub IA, Yang TJ: Growth regulatory effects of transforming growth factor-beta 1 and interleukin-2 on IL-2 dependent CD4+T lymphoblastoid cell line. Immunol Invest. 1996, 25: 129-151. 10.3109/08820139609059297.View ArticlePubMed
- Lee HM, Rich S: Differential activation of CD8+ T cells by transforming growth factor-beta 1. J Immunol. 1993, 151: 668-677.PubMed
- Sato K, Kawasaki H, Nagayama H, Enomoto M, Morimoto C, Tadokoro K, Juji T, Takahashi TA: TGF-beta 1 reciprocally controls chemotaxis of human peripheral blood monocyte-derived dendritic cells via chemokine receptors. J Immunol. 2000, 164: 2285-2295.View ArticlePubMed
- Tanaka T, Delong PA, Amin K, Henry A, Kruklitis R, Kapoor V, Kaiser LR, Albelda SM: Treatment of lung cancer using clinically relevant oral doses of the cyclooxygenase-2 inhibitor rofecoxib: potential value as adjuvant therapy after surgery. Ann Surg. 2005, 241: 168-178.PubMed CentralPubMed
- Bottinger EP, Jakubczak JL, Haines DC, Bagnall K, Wakefield LM: Transgenic mice overexpressing a dominant-negative mutant type II transforming growth factor beta receptor show enhanced tumorigenesis in the mammary gland and lung in response to the carcinogen 7,12-dimethylbenz-[a]-anthracene. Cancer Res. 1997, 57: 5564-5570.PubMed
- Akhurst RJ: TGF-beta antagonists: why suppress a tumor suppressor?. J Clin Invest. 2002, 109: 1533-1536.PubMed CentralView ArticlePubMed
- Amendt C, Schirmacher P, Weber H, Blessing M: Expression of a dominant negative type II TGF-beta receptor in mouse skin results in an increase in carcinoma incidence and an acceleration of carcinoma development. Oncogene. 1998, 17: 25-34. 10.1038/sj.onc.1202161.View ArticlePubMed
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.