The authors have no financial conflict of interest, except for Dr. John M. McPherson who was a former employee of Genzyme Corporation, and he retired from this company earlier this year. Xin Chen is an employee of Leidos Biomedical Research Inc. (formerly known as SAIC-Frederick), which is a contractor of NCI. This manuscript is not involved in any product and patent from Leidos Biomedical Research Inc. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Conceived and designed the experiments: XC LMW JJO. Performed the experiments: XC YY QZ JMW OMZH. Analyzed the data: XC YY QZ LMW. Contributed reagents/materials/analysis tools: JMM. Wrote the paper: XC LMW JJO.
TGFβ is reportedly responsible for accumulation of CD4+Foxp3+ regulatory T cells (Tregs) in tumor. Thus, we treated mouse 4T1 mammary carcinoma with 1D11, a neutralizing anti-TGFβ (1,2,3) antibody. The treatment delayed tumor growth, but unexpectedly increased the proportion of Tregs in tumor. In vitro, 1D11 enhanced while TGFβ potently inhibited the proliferation of Tregs. To enhance the anti-tumor effects, 1D11 was administered with cyclophosphamide which was reported to eliminate intratumoral Tregs. This combination resulted in long term tumor-free survival of up to 80% of mice, and the tumor-free mice were more resistant to re-challenge with tumor. To examine the phenotype of tumor infiltrating immune cells, 4T1-tumor bearing mice were treated with 1D11 and a lower dose of cyclophosphamide. This treatment markedly inhibited tumor growth, and was accompanied by massive infiltration of IFNγ-producing T cells. Furthermore, this combination markedly decreased the number of splenic CD11b+Gr1+ cells, and increased their expression levels of MHC II and CD80. In a spontaneous 4T1 lung metastasis model with resection of primary tumor, this combination therapy markedly increased the survival of mice, indicating it was effective in reducing lethal metastasis burden. Taken together, our data show that anti-TGFβ antibody and cyclophosphamide represents an effective chemoimmunotherapeutic combination.
It has been proposed that breast cancer is a naturally immunogenic tumor, since tumor antigen specific immunity can be detected in breast cancer patients, and tumor-reactive T cells are known to localize to the breast tumor microenvironment
TGFβ is a potent immunosuppressive cytokine which has the capacity to convert naïve CD4 cells into FoxP3-expressing Tregs
The DNA alkylating agent cyclophosphamide (CY) is a commonly used cytotoxic medicine in the treatment of cancer
The highly tumorigenic and invasive mouse 4T1 mammary carcinoma model shares many of the characteristics of human breast cancer, particularly its ability to spontaneously metastasize to the lungs
Female wild type 8 to 12 wk old Balb/c mice were provided by the Animal Production Area of the NCI (Frederick, MD). Foxp3/gfp KI mice were kindly provided by Dr. Yasmine Belkaid at NIAID, and maintained in the NCI-Frederick. BALB/c IFNγ−/− mice were obtained from Jackson Laboratories. NCI-Frederick is accredited by AAALAC International and follows the Public Health Service Policy for the Care and Use of Laboratory Animals. Animal care was provided in accordance with the procedures outlined in the "Guide for Care and Use of Laboratory Animals" (National Research Council; 1996; National Academy Press; Washington, D.C.). Animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) of National Cancer Institute (Frederick, MD).
4T1 breast cancer cells were obtained from ATCC (11/112003, lot No. 3306022 CRL-2539) and from Dr. Fred Miller (3/262003, Barbara Ann Karmacos Institute, Wayne State University School of Medicine) who firstly described this cell line
Antibodies purchased from BD Biosciences (San Diego, CA) consisted of anti-CD3 (145-2C11), CD4 (GK1.5), CD16/CD32 (2.4G2), INFγ (XMG1.2). Foxp3 Staining Set (FJK-16s), anti-mouse TCRβ Ab (H57-597) and functional grade purified anti-mouse CD3e Ab (eBio500A2) were purchased from eBioscience (San Diego, CA). The anti mouse TGFβ monoclonal antibody, 1D11, which neutralizes all three isoforms of TGFβ, and an isotype-matched mouse IgG1 monoclonal antibody, 13C4, were provided by Genzyme Corp.
4T1 tumor cells were injected into right mammary fat pads (thoracic No. 2 mammary glands) of recipient mice in single cells suspension with 50,000 cells in 0.2 ml PBS per mouse. After indicated times, tumors were excised, minced and digested in RPMI 1640 supplemented with 1 mg/ml collagenase IV and 0.1 mg/ml DNase I. The fragments were pushed through a 70-µm pore size cell strainer to create a single-cell suspension. In some experiments, two weeks after last treatment (60 days after initial tumor inoculation), tumor free mice after CY+1D11 treatment were re-inoculated with 4T1 cells (50,000) into the right mammary fat pads (thoracic No. 2 mammary glands), and the same number of CT26 colon carcinoma cells were s.c. injected to the left flank. Tumor size was calculated by the formula: (
Mice were treated by the following dose schedule: 1D11 or mouse IgG1 (13C4) were administered three times per week i.p at 0.1 mg in 0.2 mL PBS, starting 1 or 3 days after inoculation of 4T1 cancer cells. After 4 weeks, the three times weekly treatment was reduced to one. A single dose of CY was injected i.p. at 4 mg in 0.2 mL PBS by 3 days after cancer cell inoculation. For a reduced dose schedule, 1D11 or Mu IgG1 was i.p. administered at 0.1 mg, starting 3 days after 4T1 cancer cell inoculation. After 3 weeks, the three times weekly treatment was reduced to one. A single dose of CY was i.p. injected at 2 mg 3 days after cancer cell inoculation. For the spontaneous metastasis study with surgical resection of primary tumor, 1D11 or mouse IgG1 were administered three times per week i.p at 5 mg/kg for 2 wks starting 7 days after inoculation of 4T1 cells, followed by once a week treatment for the duration of the experiment. A single dose of 50 mg/kg CY was injected i.p. at day 14 after cancer cell inoculation.
After blocking FcR, cells were incubated with appropriately diluted antibodies. Acquisition was performed using a SLRII (BD Biosciences, Mountain View, CA) and data analysis was conducted using FlowJo software (Tree Star Inc., Ashland, OR). For intracellular cytokines staining, cells were re-stimulated with BD Leukocyte Activation Cocktail for 4 h. FACS analysis was gated on the live cells only by using LIVE/DEAD Fixable Dead Cell Stain Kit (Invitrogen Life Technologies). FACS analysis of TILs was gated on live CD45+TCRβ+(or CD3+) cells.
CD4+Foxp3/gfp+ Tregs were sorted from LNs and spleens of Foxp3/gfp KI mice using Cytomation MoFlo cytometer (Fort Collins, CO), yielding a purity of ∼98% Tregs. T-depleted spleen cells were used as APCs and irradiated with 3,000 R. Tregs were seeded into round-bottom 96-well plate at 2×104 cells/well. The cells were stimulated with 2×105 APCs/well plus 0.5 µg/ml of soluble anti CD3 Ab, with or without murine TNF (10 ng/ml, PeproTech, Rocky Hill, NJ), in the presence of medium alone or increasing concentration of recombinant human TGFβ1 (0.1∼1 ng/ml, R&D Systems, Minneapolis, MN) or 1D11 (1∼20 µg/ml). Cells were pulsed with 1 µCi [3H]thymidine (Perkin Elmer Life Sciences, Boston, MA) per well for the last 6 h of the 72-hour culture period.
All data was compared and analyzed by two-tailed Student’s
To examine if anti-tumor effect of the anti-TGFβ Ab 1D11
(A-D) Mice were treated with 0.1 mg 1D11 or mouse IgG1 (i.p., 3×week), starting at day 1 of tumor inoculation. (A) Kinetics of tumor growth. (B) Weight of tumors isolated at 14 days after inoculation. (C-E) Effect of 1D11 on Tregs. CD4 cells and Tregs was analyzed with FACS at 14 days after tumor inoculation. (C, D) Typical FACS analysis of CD4 cells and Tregs. Number represents the percentage of CD4+ cells in total tumor infiltrating CD45+ leukocytes (C) or Foxp3+ cells in intratumoral CD45+CD4+ cells (D). (E) Summary of proportion of Foxp3+ cells in CD4 cells present in the tumor, spleen, mesenteric LNs and axillary/inguinal LNs (N = 3∼7). (F-G) TGFβ inhibits, while 1D11 promotes, proliferation of Tregs in vitro. (F) CD4+Foxp3/gfp+ Tregs were stimulated in the presence of TNF with increasing concentrations of rhTGFβ1. (G) Tregs were stimulated in the presence of TNF or medium alone with increasing concentration of 1D11. After incubation for 72 hours, the proliferation of Tregs was determined by [3H] thymidine incorporation. By compared with respective control (without rhTGFβ1or 1D11), *p<0.05, ** p<0.01. N = 3. The data are representatives of three separate experiments with same results.
This result suggest that neutralization of TGFβ might promote proliferation of Tregs in the tumor inflammatory environment. To test this, we examined the effect of TGFβ and anti-TGFβ Ab on the proliferation of Tregs in vitro. Previously we showed that the profound hyporesponsiveness of Tregs to TCR stimulation in vitro could be overcome by exogenous TNF
Although 1D11 suppressed the growth of 4T1 tumors, it failed to completely control their growth, which may be attributable to the expansion of Tregs in the tumor. A therapeutic with the capacity to eliminate tumor infiltrating Tregs may enhance the anti-tumor action of 1D11. It was reported that tumor infiltrating TNFR2+ highly suppressive Tregs could be eliminated by CY
Three days after tumor inoculation, the mice were i.p. treated with single dose of CY (4 mg) or 1D11 (0.1 mg, 3×week), or combination of CY and 1D11 or mouse IgG1. (A) Percent tumor-free mice (%). (B) Survival of tumor inoculated mice. (C) Tumor size in groups treated with PBS, or CY alone or 1D11 alone. (D) Tumor size in groups treated with CY+1D11 or CY+Mu IgG1. Two weeks after last 1D11 treatment (60 days after initial tumor inoculation), the tumor-free mice (designated as pre-treated) were re-inoculated with 4T1 cells into the right thoracic mammary fat pad, and CT26 cancer cells were inoculated (s.c.) into the left flank. For comparison, age- and gender-matched normal Balb/c mice (designated as untreated) were inoculated with 4T1 and CT26 tumor cells in the same manner. (E) Incidence of 4T1 and CT26 tumor development on day 18 after tumor inoculation. (F) Growth of 4T1 tumor and (G) growth of CT26 tumor. Data shown in C, D, F and G are means±SEM (N = 5∼10). Comparison of two groups: * p<0.05; **p<0.01. The data are representatives of three separate experiments with similar results.
Three days after tumor inoculation, the mice were i.p. treated with single dose of CY (2 mg) or 1D11 (0.1 mg, 3×week), or combination of CY and 1D11 or mouse IgG1. Mice were sacrificed ∼4 wks after tumor inoculation. (A) Kinetics of tumor growth. The data are representatives of three separate experiments with similar results (N = 10, data shown as means±SEM). (B) Weight of tumors after 4 wks of inoculation (N = 17, pooled from two separate experiments). (C) The number of grossly visible metastatic nodules in the lung (N = 14, pooled from two separate experiments). Comparison of indicated groups, * p<0.05, ** p<0.01, *** p<0.001.
To examine whether the tumor-free mice developed 4T1 tumor-specific immunity, those mice surviving after CY+1D11 treatment were re-inoculated with 4T1 tumor cells and the same number of mouse CT26 colon cancer cells on the contralateral flank. As a control, normal Balb/c mice were also inoculated with 4T1 cells and CT26 cells in the same manner. All mice (100%) in the control group developed measurable 4T1 and CT26 tumor by day 13 after inoculation (
Although our original treatment regimen achieved an optimal anti-tumor effect, it did not allow us to examine Tregs and other TILs, since the majority of mice did not develop tumor at all. Therefore, we administered lower doses of both 1D11 and CY in order to allow tumor growth for analysis. The results show that 90% of mice developed tumor after treatment with reduced doses of CY and 1D11. This treatment regimen also markedly inhibited the growth of primary tumor (p<0.05∼0.001), as shown in
Since tumor-free mice after CY+1D11 treatment developed partial 4T1 tumor-specific resistance, we hypothesized that T cells should be mobilized and activated. Indeed, 1D11 treatment alone, and CY treatment alone to a lesser extent, increased T cell infiltrating the tumor (p<0.01,
Four weeks after 4T1 tumor inoculation, cell suspension was prepared from tumor tissues. (A-B) Proportion of TCRβ+ T cells in total tumor infiltrating CD45+ leukocytes. Typical flow plots are shown in (A), and summary of data from three separate experiments is shown in (B, Means±SEM, N = 14∼20). (C-F) IFNγ expression by CD8 and CD4 TILs. Intracellular expression of IFNγ was analyzed by FACS, gating on live CD45+TCRβ+CD8+ cells (C-D) or CD45+TCRβ+CD4+ cells (E-F). Data shown are typical FACS plots (C, E) and summary of data (D, F) from three separate experiments (Means±SEM, N = 9). Comparison of indicated groups, * p<0.05, ** p<0.01.
To evaluate the role of IFNγ in the anti-tumor effect of the combination treatment of CY+1D11, we examined its effect in 4T1 tumor-bearing IFNγ KO mice. As can be seen in
Since we had shown that 1D11 treatment increases intratumoral Tregs, and CY was reported to reduce Tregs, we predicted that CY would abrogate 1D11-driven expansion of Tregs. However, unexpectedly, we did not find any change in intratumoral Tregs following combination therapy (data now shown). We therefore looked for alternative mechanisms. The mouse 4T1 tumor model is characterized by the accumulation of MDSCs in the spleen which causes splenomegaly
Four weeks after 4T1 tumor inoculation, cell suspensions were prepared from spleen and MDSCs were analyzed by FACS, gating on live CD45+ cells. (A-B) Proportion and number of Gr1+CD11b+ cells in the spleens. Typical FACS plots are shown in (A, gating on total live splenic CD45+ cells) and summary of data pooled from three experiments are shown in (B, percent of PBS control group, Means±SEM, N = 10∼13). (C) Absolute number of Gr1+CD11b+ cells in the spleen (N = 11, pooled from two separate experiments). (D-E) Expression of I-A/I-E and CD80 on Gr1+CD11b+ splenic cells. (D) Typical FACS plots (gating on Gr1+CD11b+ cells) and (D) summary of data pooled from three separate experiments (Means±SEM, N = 9). Comparison of indicated groups, *p<0.05, ** p<0.01, ***p<0.001.
We further utilized the 4T1 model in a format where the primary 4T1 tumor was surgically excised and mouse survival is driven by metastatic lung disease
4T1 cells were inoculated into the mammary fat pad and primary tumors were surgically removed at day 12 after inoculation. 1D11 or Mu IgG1 (13C4) control antibody (5 mg/Kg, i.p.) were administered three times per week for the first two weeks and followed by once a week, starting from 7 days after tumor inoculation. A single subtherapeutic dose of CY (50 mg/Kg, i.p.) or vehicle was given on day 14, two days following surgical resection of the primary tumor. The therapeutics was treated alone or combined as indicated. (A) Survival curves for the different treatment groups. Survival curves are significantly different between the groups (p<0.0007; Log-Rank (Mantel-Cox) Test). (C) Representative images from lungs at gross and (D) in histological cross-section (H&E stained).
TGFβ is a pleiotropic cytokine that plays a key role in the interplay of tumor cells and other cells in the tumor environment
Thus, TGFβ has dual and biphasic effects in tumor development, and this complex nature of TGFβ in cancer biology poses the challenge for the application of TGFβ inhibitor as a sole therapeutic. Combination with other therapeutics has the potential to reinforce the beneficial anti-tumor effects, while minimizing the undesirable effects of a TGFβ inhibitor. Our study clearly shows that CY is one such chemotherapeutic. This combination treatment likely targets multiple cellular and molecular events simultaneously. Nevertheless, the activation of anti-tumor immune responses contributes substantially to the anti-tumor effect , for the reason that: 1) the combination treatment resulted in a massive infiltration of IFNγ-producing cells to the tumor; 2) anti-tumor effect of the combination therapy was reduced in IFNγ KO mice; 3) the number of splenic MDSCs was markedly reduced and the residual MDSCs showed a more mature phenotype; 4) tumor-free mice after CY+1D11 treatment developed long term anti-tumor immunity.
We anticipated that CY treatment would reduce the number of Tregs, as previously reported
Although TGFβ is able to convert naïve CD4 cells into FoxP3-expressing induced Tregs (iTregs)
A previous report showed that the combination of CY with an anti-TGFβ receptor II antibody had an additive effect in the suppression of primary tumor growth and lung metastasis in mice bearing EMT6 mammary cancer
In our studies, Mu IgG1 13C4 by itself or in combination with CY had a consistent inhibitory effect on 4T1 tumor growth (
Taken together, our study showed a commonly used chemotherapeutic CY was able to enhance the anti-tumor effect of TGFβ inhibitor, resulting in the potent inhibition in the development of 4T1 mouse mammary carcinoma. Although multiple mechanisms may underlie the anti-tumor effect of combination therapy of CY and TGFβ inhibitor, this efficacy is in part due to improvement in the quality of anti-tumor immunity in both adaptive and innate arms. This combination regimen thus represents a successful approach to the chemoimmunotherapy of primary and metastatic cancer.