Immune Compartmentalization of T cell Subsets in Chemically-induced Breast Cancer

Authors


G. Esendagli, PhD, Department of Basic Oncology, Institute of Oncology, Hacettepe University, 06100 Sihhiye, Ankara – Turkey. E-mail: gunes.esendagli@lycos.com

Abstract

In cancer, the phenotype and/or the function of T cells may differ according to their distribution through immune-associated tissues, namely immune compartments. Here, in N-methyl-N-nitrosourea (MNU)-induced mammary carcinomas of rat as a relevant model for human breast tumors, the impact of tumor burden on the T cell subsets populating the tumor microenvironment, the tumor-adjacent and -opposite mammary lymph nodes, and the spleen was assessed. In the tumors, ratio of CD8+ cytotoxic and CD4+ helper T cells were not significantly different than other immune compartments. On the other hand, most of these cells were further identified with CD4+ CD25hi or CD4+ Foxp3+, CD8+ Foxp3+ regulatory phenotype. The selective presence of Tregs in the mammary tumors but not in neighboring-mammary tissue was also confirmed by the expression of Treg-associated genes. The percentage of CD161+ NKT cells was also significantly increased especially in the tumors and mammary lymph nodes. In the lymph nodes of tumor-bearing animals, in contrast to the spleen, total amount of CD8+ cells and CD4+ cells were increased but both of these compartments harbored high numbers of CD4+ CD25hi Treg cells. TGF-β was determined as the major suppressive cytokine secreted by the immune cells of tumor-bearing animals, in addition, proliferation capacity of the T cells was diminished. Hence, the differential distribution of T cell subsets through the spleen, the mammary lymph nodes and the tumor mass in MNU-induced mammary tumor-bearing animals may contribute to a tumor-associated immunosuppression.

Introduction

Breast cancer is the most frequently diagnosed malignancy and the leading cause of cancer related deaths in women despite early detection and treatment [1]. One of the most promising approaches for breast cancer therapy is immune intervention. On the other hand, especially in the local environment of tumors, immune system is diverted and cannot eliminate the neoplastic cells, even may contribute to cancer progression.

The immune cells found in tumor microenvironment such as regulatory T cells (Tregs), myeloid-derived suppressor cells, tumor-associated macrophages and plasmacytoid dendritic cells (DC) are important players in tumor’s immune escape [2]. Except the subclasses with regulatory and/or suppressive functions, T cells are critical for anti-tumor immune responses. However, even the activated T cells may become ineffective in the tumor microenvironment. The factors secreted by the tumor(-associated) cells may influence the activation and proliferation of T cells directly; or indirectly via interfering with DC functions and support Treg differentiation [2–4].

There is a strong relationship between Tregs and the development and progression of most solid cancers [5]. On the other hand, in certain types of hematological malignancies, the presence of Tregs found to be associated with good prognosis [6–8]. Naturally occurring CD4+ CD25hi Foxp3+ Tregs cells are able to actively inhibit CD4+ CD25 and cytotoxic T cells, DC, natural killer cells and B cells via cell-to-cell contact, e.g. membrane bound transforming growth factor-beta (TGF-β), cytotoxic T lymphocyte antigen-4 (CTLA-4), programmed death-1 (PD-1), Fas, and via secreted factors such as interleukin (IL)-10, TGF-β and IL-3 [9]. The inhibition of Treg functions may enhance the immune surveillance and the success of immunotherapeutic modalities [5].

In human breast cancers, each immune compartment, i.e. tumor tissue, tumor-draining (sentinel) lymph nodes and peripheral blood, has been individually studied [10–12]. However, the heterogeneity of patient samples may preclude with the significance of biological finding. As an experimental cancer model, the chemical carcinogenesis is a conventional method to study the process of tumor initiation and promotion/progression, in vivo [13]. The chemical carcinogen N-methyl-N-nitrosourea (MNU) has been optimized to specifically target the mammary gland and does not require metabolic activation [13, 14]. Mammary tumors induced with MNU are relevant to human breast cancers with similar histopathological lesions, originating from ductal epithelial cells, and may carry altered TGF-β, erbB2, cyclin D1 expression [13]. Therefore, this experimental model has been widely used in chemoprevention and carcinogenesis studies [13]. Moreover, several studies characterizing the local immune aspects of MNU-induced breast carcinomas showed compatibility with the immune status of human breast tumors [15].

The aim of the current study was to monitor the distribution of T cell subsets (helper, cytotoxic, regulatory, and NKT) in tumors, tumor-adjacent and -opposite mammary lymph nodes, and in spleen of the chemically-induced mammary tumor-bearing animals. Functional assays were performed with the immune cells isolated from these compartments. In addition, tumor-neighboring normal mammary tissue was analyzed to support the selectivity of T cell infiltration. Here, we demonstrated the differential distribution of T cell subsets in MNU-induced mammary tumor-bearing animals and potential contribution of T cell localization to the tumor-associated immunosuppression.

Material and methods

Animals.  Twenty six, 21-day-old female Sprague–Dawely rats were obtained from the Experimental Animals Breeding Unit of Hacettepe University. Twelve rats were used for chemical carcinogenesis experiments, and 14 rats were used as controls. All rats were housed under environmentally controlled standard conditions. The Guiding Principles in the Care and Use of Laboratory Animals together with those described in the declaration of Helsinki were strictly adhered in the conduct of all experimental procedures described within this manuscript. This project was approved by the Institutional Animal Care and Use Committee of Hacettepe University, Ankara, Turkey (Approval No. 2007/68-3) before commencement of the project.

Establishment of mammary tumors and histopathological evaluation. N-methyl-N-nitrosourea (Sigma, St Louis, MO, USA) was dissolved under sterile conditions in physiological saline (0.9%, w/v; pH 5.00). The solution was injected intraperitoneally (i.p.) at a dose of 50 mg/kg, for 4 weeks (one injection per week) [16]. The tumors (largest dimension ≥ 1 cm) were excised not earlier than 10 weeks following the last MNU injection. For histopathological evaluation, mammary tumors were excised and fixed in 10% formalin solution, embedded in paraffin and examined under light microscopy following hematoxylin – eosin staining.

Isolation of immune cells.  Finely chopped mammary tumor samples (n = 8) were passed through 40 μm cell strainers (Becton Dickinson, San Jose, CA, USA) and subjected to discontinuous Percoll gradient separation (Sigma) using the dilutions 75%, 50% and 25%. Cells collected from the respective density fractions were directly prepared for flow cytometric analysis.

Following the isolation of splenocytes lymphocytes were enriched by density gradient separation by Bicoll (Biochrom, Berlin, Germany) procedure. The cells isolated from lymph nodes (from control group, n = 11; from tumor-bearing group – tumor-adjacent lymph node, n = 8, tumor-opposite lymph node, n = 5) and spleens (from control group, n = 11; from tumor-bearing group, n = 11) were filtered through 47 μm nylon mesh.

Flow cytometry.  The following monoclonal antibodies were used for flow cytometry: anti-rat-CD3 (G4.18), -CD25 (OX-39), -Foxp3 (FJK-16s), -CD4 (OX35) (eBioscience, San Deigo, CA, USA), -CD8a (OX-8), -CD161 (10/78) (Biolegend, San Deigo, CA, USA). For intracellular staining, Cytofix/Cytoperm™ Fixation/Permeabilization Solution Kit (Becton Dickinson) was used according to the recommendations of manufacturer. Samples were analyzed on an EPICS XL-MCL flow cytometer (Beckman Coulter, Fullerton, CA, USA). The percentage of positive cells was calculated by comparison with appropriate isotype-matched antibody controls. The number of immunophenotyped T cells in specimen suspensions was calculated with the equation: Number of isolated cells per ml × Total volume (ml) × Ratio of T cells (%).

Enzyme-linked immunosorbent assay (ELISA).  The isolated cells (5 ×106 cells/ml) from tumor, spleen and lymph nodes (n = 10) were incubated for 18 h in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin and 100 μg/ml streptomycin at 37 °C in a humidified 5% CO2 incubator with or without phorbol 12-myristate 13-acetate (PMA, 40 ng/ml) and ionomycin (2 μg/ml). The supernatants and/or plasma samples collected were tested using IL-10, IFN-γ (Thermo scientific, Rockford, IL, USA) and TGF-β (R&D systems, Minneapolis, MN, USA) ELISAs according to the manufacturer’s instructions.

Proliferation assay.  The isolated cells (4 × 106) from tumor, spleen and lymph nodes (n = 6) were labeled with carboxyfluorescein succinimidyl ester (CFSE, 5 μm) (CellTrace™; Invitrogen, Eugene, OR, USA). Labeled cells were incubated for 6 days in RPMI 1640 medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin and 100 μg/ml streptomycin at 37 °C in a humidified 5% CO2 incubator with or without 1 μg/ml phytohemagglutinin (PHA, Sigma). Following the incubation cells were harvested and stained with anti-rat CD3 monoclonal antibody (eBioscience) and percentage of proliferated cells (>one division) were analyzed by flow cytometry.

Reverse transcription-polymerase chain reaction (RT-PCR).  Tumor tissues and tumor-neighboring mammary tissues (in 0.5 cm proximity to a well demarcated site of the tumor mass) were sampled. Total RNA was extracted from homogenized tissues using QIAamp RNA Blood Mini Kit (Qiagen, Maryland, MD, USA). To ensure the absence of contaminating DNA, RNA samples were treated with DNA-free DNase (DNA-free kit; Ambion, Austin, TX, USA). cDNA was synthesized from 0.5 μg of RNA, using oligo(dT) primers and RevertAid™ First Strand cDNA Synthesis Kit (Fermentas, Vilnius, Lithuania) according to manufacturer’s instructions. PCR reactions were performed with primers designed for rat TGF-β, IL-10, Foxp3, CTLA-4, Fas, PD-1, neuropilin-1 (NP-1), IFN-γ, and CD40 ligand (CD40L) are listed in Table 1. The primer sequences for β-actin and tumour necrosis factor-alpha (TNF-α) were previously published [17]. PCR products were separated on 2% agarose gel electrophoresis, stained with ethidium bromide and documented under UV light (Kodak Gel Logic 1500 Imaging System; Carestream Health, Rochester, NY, USA).

Table 1.   Primer sequences designed for RT-PCR.
PrimerSense (5′-3′)Antisense (5′-3′)Product size (bp)Accessiona
  1. aNCBI nucleotide database accession numbers of the cDNAs available online at http://www.ncbi.nlm.nih.gov

TGF-βccagatcctgtccaaactaaggcgcgggtgctgttgtacaaagc102NM_021578
IL-10agaccagcaaaggccattccaacttgactgaaggcagccctca103NM_012854
Foxp3cacagctctgctggagaaagctgacttccaagtctcgtgtg122NM_001108250
CTLA-4agtgacagaggtctgtgccacgaagtacagtccggtgtcagca143NM_031674
Fasgatgagatcgagcacaacagcgcggttagcttttctgagacc130NM_139194
PD-1gccgccttctgcaatggttaactgtcattgcgccgtgcat120NM_001106927
NP-1agtggctcctggaagaggagcaatgcctgaattcgcgag119NM_145098
IFN-γtgaaagacaaccaggccatcagcggatctgtgggttgttcacctcg138NM_138880
CD40Ltcgggaacctttgagtcaacgagcttggaggaactgtgggt111NM_053353

Statistical analysis.  All the values are expressed by arithmetic mean ± standard deviation (SD). Statistical difference between groups was determined using student’s t-test (paired or unpaired). When P ≤ 0.05 the differences were regarded statistically significant.

Results

Presence of T cell subsets in the mammary lymph nodes, spleen and tumor tissue

In the flow cytometry analyses following different isolation procedures for the enrichment of immune cell-containing fractions, lymphocytic cells were gated and presence of CD3+ cells in tumor tissue, tumor-adjacent and -opposite mammary lymph nodes, and in spleen was determined (Fig. 1A). The mean value of freshly isolated T cells that could be obtained from tumor tissue was 28.64 ± 6.9%. There was no significant difference in the T cell ratio of immune compartments of tumor-bearing animals when compared that of the control group. Moreover, the percentage of CD3+ cells was similar in tumor-adjacent and -opposite mammary lymph nodes (Fig. 1B). In addition, there was no relation between the amount of T cells and the multiplicity, size or location of the mammary tumors (data not shown).

Figure 1.

 CD3+ T cell population in tumor tissue, tumor-opposite and -adjacent lymph nodes, and spleen was determined by flow cytometry. (A) Lymphocytic cells were gated according to forward-scatter (FS) and side-scatter (SS) properties. Representative dot plots of the isolated cells and histograms of CD3 staining of a tumor-bearing rat are shown. (B) Percentages of CD3+ cells in control and tumor-bearing group samples can be seen (black, isotype staining; green, control group; red, tumor-bearing group; gr., group; Op., opposite; Ad., adjacent).

Cytotoxic and NKT cell populations are differentially distributed in the immune compartments of tumor-bearing animals

When we went further through the analysis of T cell subsets on CD3-gated cells (Fig. 2A), a significant increase in CD4+ T cells was noted in the spleen of tumor-bearing animals compared to the control animals. On the other hand, the ratio of CD4+ cells was decreased both in tumor-adjacent and -opposite lymph nodes. The 60.51 ± 9.9% of the T cells infiltrating the tumor tissues, not being different from other immune compartments of the tumor-bearing animals, was immunophenotyped as CD4+ (Fig. 2B).

Figure 2.

 Determination of helper, cytotoxic and NKT cell subsets. Immunophenotyping was done by flow cytometry analysis of CD4, CD8, and CD161 surface markers on CD3 gated cells. Representative flow cytometry dot plots of a control and a tumor-bearing group animal can be seen in panel (A). The plots of tumor-adjacent mammary lymph nodes are given. The mean percentage values of (B) helper T cells, (C) cytotoxic T cells and (D) NKT cells are shown (gr., group; Op., opposite; Ad., adjacent, *< 0.05 and **< 0.01).

Regarding cytotoxic T cells, there was a significant decrease in their mean ratio within the splenic T cells, whereas they were at a higher ratio in both mammary lymph nodes of tumor-bearing rats. Although it did not reach to a statistical significance, the number of CD8+ T cells in the tumor tissue was relatively enhanced, 37.33 ± 8.45%, (Fig. 2C).

Through the other immune compartments analyzed, tumors were the most preferred location for CD161+ CD3+ NKT cells (20.48 ± 9.27%). NKT cells were also significantly increased in the lymph nodes of tumor-bearing animals when compared to the control group (Fig. 2D).

When the total number of the cells was calculated, a significant increase in CD3+ CD4+ T cells was observed in spleen and tumor-adjacent lymph nodes whereas the amount of CD3+ CD8+ and CD3+ CD161+ cells were significantly higher only in the lymph nodes of the tumor-bearing animals compared to that of the control group (Table 2). Furthermore, in the tumor-bearing group, the highest number of helper and cytotoxic T cells was found in spleen, where the lowest was in the lymphocytic fractions isolated from the tumors. On the other hand, the number of NKT cells in the tumor-adjacent lymph nodes was lower than the amount infiltrating the tumors (Table 2).

Table 2.   Total numbers (×106) of T cell subsets obtained from the immune compartments of control and tumor-bearing groups.
 SpleenAdjacent lymph nodeTumor
Control gr.Tumor gr.Control gr.Tumor gr.
  1. Control gr., Control group; Tumor gr., Tumor-bearing group.

  2. Statistical significance for a specific immune compartment between control and tumor-bearing groups was designated as *< 0.05, **< 0.01.

  3. Statistical significance between different immune compartments of tumor-bearing animals (between spleen and lymph node, aP<0.05, aaP<0.01; spleen and tumor tissue, bP<0.05, bbP<0.01; lymph node and tumor tissue, ccP<0.01; spleen and lymph node and tumor tissue, dP<0.05, ddP<0.01) were also shown.

CD3 + CD4+1.96 ± 0.035.53 ± 2.7*,a0.37 ± 0.031.88 ± 0.22**0.42 ± 0.21dd
CD3 + CD8+1.04 ± 0.172.41 ± 1.15a0.11 ± 0.020.89 ± 0.2**0.23 ± 0.09d
CD3 + CD161+0.13 ± 0.070.53 ± 0.3a0.006 ± 0.0040.06 ± 0.01**0.15 ± 0.04cc
CD4 + CD25+0.26 ± 0.051.51 ± 0.12**,aa0.030.27 ± 0.03**0.26 ± 0.08bb
CD4 + CD25hi0.13 ± 0.020.75 ± 0.06**,aa0.0150.13 ± 0.02**0.13 ± 0.04bb
CD4 + Foxp3+0.02 ± 0.030.77 ± 0.08**,aa0.009 ± 0.00050.16 ± 0.03**0.12 ± 0.01bb
CD8 + CD25+0.11 ± 0.060.27 ± 0.07aa0.004 ± 0.0020.09 ± 0.03**0.15 ± 0.08
CD8 + Foxp3+0.004 ± 0.0040.22 ± 0.1*,a0.002 ± 0.0010.02 ± 0.005**0.04 ± 0.005b

Tumors are the preferential site for CD4+ and CD8+ T cells to possess regulatory phenotype

CD25 and Foxp3 markers were used to determine the Treg subset. The double positive CD4+ CD25+ and CD8+ CD25+; CD4+ Foxp3+ and CD8+ Foxp3+ populations were analyzed by flow cytometry. Compared to the control group, the ratio of CD25+ within CD4+ cells in splenocytes and in cells isolated from tumor-adjacent mammary lymph nodes of tumor-bearing rats was significantly increased (Fig. 3A, B). Especially in the tumors, the population of CD4+ helper or CD8+ cytotoxic T cells carrying CD25 was very high, 88.13 ± 3.17% and 68.89 ± 8.14%, respectively (Fig. 3A, B and C).

Figure 3.

 CD25 and Foxp3 expression in T cell populations. (A) Representative flow cytometry dot plots showing the presence of CD25 on CD4+ or CD8+ T cells. Only the results of tumor-adjacent mammary lymph nodes are shown. Evaluation of the CD25+ cell ratio was done following the gating (B) CD4+ or (C) CD8+ cells. In panel (D) representative flow cytometry analyses showing the intracellular staining of Foxp3 in CD4+ or CD8+ T cells is shown. The amount of regulatory T cells was determined by analyzing the mean percentage values of (E) CD25hi, (F) Foxp3+ populations within CD4+ cells and (G) Foxp3+ populations within CD8+ cells (gr., group; *< 0.05 and **< 0.01).

The CD4+ CD25hi Treg cells are regarded as major effector cells involved in the suppression of anti-tumor immune responses. Therefore, we determined the ratio of CD4+ CD25+ cells highly expressing CD25. CD4+ CD25hi population in the spleen and mammary lymph nodes of tumor-bearing rats were found to be significantly increased (Fig. 3E). There was no difference between the percentages of CD4+ CD25hi cells in opposite- and adjacent-mammary lymph nodes (data not shown). Intracellular presence of Foxp3 was tested as another important Treg marker (Fig. 3D). There was a significant difference in the Foxp3+ population within CD4+ T cells in spleen of tumor-bearing rats compared to the control group. On the other hand, no difference was observed in CD4+ Foxp3+ cell ratio between the mammary lymph nodes. In the tumor tissue, 36.68 ± 9.65% of helper T cells and 16.55 ± 4.9% of cytotoxic T cells was significantly determined to carry the Treg marker, Foxp3 (Fig. 3F, G).

When the total numbers of CD25 or Foxp3 expressing T lymphocytes was calculated, a significant increase was determined in spleen, except for CD8+ CD25+ cells, and in tumor-adjacent lymph nodes of the tumor-bearing animals compared to that of the control group (Table 2). In the tumor-bearing group, the highest number of regulatory cells was found in spleen, whereas tumor-adjacent lymph nodes were harboring approximate numbers of CD4+ or CD8+ Tregs with the tumor tissue (Table 2).

Immune cells obtained from the compartments of tumor-bearing animals respond to activating stimuli despite suppressive influences, ex vivo

To assess the functional capacity of the isolated immune cells, ex vivo analyses were carried out. In the presence of T cell stimulator, PHA, proliferation of the T cells isolated from spleen and tumor-adjacent lymph nodes of tumor-bearing animals were significantly induced. However, the percentage of proliferated cells was rather low when compared to proliferation response obtained with the control group. Interestingly, the unstimulated cells isolated from the tumor tissue showed a proliferation ratio as high as stimulated lymphocytes from tumor-adjacent lymph nodes and did not respond to PHA-stimulation (Fig. 4A).

Figure 4.

Ex vivo functional capacity of immune cells isolated from tumor tissue, tumor-opposite and -adjacent lymph nodes, and spleen. (A) Proliferation capacity of the T cells upon stimulation with PHA was tested by flow CFSE proliferation assay. The production of cytokines (B) TGF-β, (C) IL-10, and (D) IFN-γ by the cells with or without PMA/ionomycin stimulation were determined by ELISA (gr., group; *< 0.05 and **< 0.01).

We checked the concentrations of the cytokines TGF-β, IL-10, and IFN-γ in the supernatants of the cells incubated for 24 h with or without stimulation. TGF-β secretion by the cells isolated from the compartments of tumor-bearing group was significantly higher. Upon stimulation, TGF-β raised in the supernatants of cells isolated from tumor-adjacent mammary lymph nodes and tumors. On the other hand, this effect was not seen with the spleen of tumor-bearing animals or with the lymph nodes of control animals (Fig. 4B). In case of IL-10, no difference was observed between stimulated or unstimulated cells from either control or tumor-bearing groups. The cells isolated from spleen were found with the highest expression of IL-10 (Fig. 4C). Also, there was no significant difference in the concentration of IL-10 plasma levels between tumor-bearing and control animals (data not shown).

IFN-γ could be detected in the supernatants derived from stimulated cells from the control spleens. The cell supernatants harvested from other compartments did not contain detectable levels of IFN-γ (Fig. 4D). On the other hand, the levels of IFN-γ in the plasma of tumor-bearing animals (0.29 ± 0.7 pg/ml) were significantly lower than the control group (17.1 ± 0.4 pg/ml).

The expression of regulatory T cell-related genes is absent in the tumor-neighboring mammary tissue

The expression of the genes mainly associated with the Tregs, IL-10, Foxp3, CTLA-4, Fas, PD-1, was only found in the tumors whereas TGF-β and NP-1 were also detected at lower levels in the tumor-neighboring mammary tissues. The genes with proinflammatory functions, IFN-γ and TNF-α were expressed in tumors; on the other hand, CD40L could not be amplified either in tumor or tumor-neighboring mammary tissues (Fig. 5).

Figure 5.

 Expression of T cell-associated genes in tumor and tumor-neighboring mammary tissue was determined by RT-PCR (cDNA obtained from splenocytes was used as positive control, C (+); M, DNA size marker; T-NMT, tumor-neighboring mammary tissue; C (−), PCR-negative control).

Discussion

The phenotype or even the function of immune cells may vary according to their localization in specific tissues, namely immune compartments. As a dynamic member of immune system, T cells can be found in circulation, migrated into inflamed sites, or residing in immune-associated tissues [18]. Moreover, the activation state of T cells is also detrimental for a proper adaptive immune response. Under pathological circumstances such as malignancies, the distribution of T cell subsets can be reorganized and/or they may be functionally disabled [19]. Even though the neoplastic tissue itself is the center for immune regulation in cancer, other immune compartments can be pathophysiologically influenced [20]. Using the MNU-induced tumor model in rats, the distribution of CD8+ cytotoxic, CD4+ helper, CD4+ CD25hi or CD4+ Foxp3+, CD8+ Foxp3+ regulatory, and CD161+ NKT, T cell subsets in tumor tissues, tumor-adjacent and -opposite mammary lymph nodes, and in spleen were compared with that of healthy control rats. The systemic injection of MNU may have toxic effects on hematopoietic cells. This effect has been reported as a temporary residual toxicity and colony formation by bone marrow cells can return to normal levels following the third week post-MNU administration [21, 22]. Accordingly, we did not observe any toxicity-related alteration (e.g. number of isolated cells) in the bone marrow, spleen or lymph nodes of tumor-bearing animals.

In the chemically-induced mammary tumor model, the increase in the numbers of T cell subsets may indicate an immune activity. As expected, the highest amount of any specific subset was found in the spleen, then in the tumor-adjacent lymph nodes or the lymphocyte-enriched fraction of the tumor. The highest ratio of T cell subsets populating the tumor tissue was helper, then cytotoxic and the lowest was NKT. In the same experimental model, a similar phenomenon was also previously reported [15]. Since, CD4+ or CD8+ T cells may possess regulatory/suppressor functions, these cells were further stained with CD25 or Foxp3 markers. Almost half of the amount of helper T cells and 15% of cytotoxic T cells were identified with regulatory phenotype. Especially, CD4+ Tregs were increased in all immune compartments of the tumor-bearing group. In accordance with our results, tumor tissue is the site with the highest level of Treg accumulation, while the frequency of Tregs in the peripheral compartments such as tumor-draining lymph nodes, spleen and in peripheral blood may vary depending on the origin of tumor and the stage of disease progression [23].

In cancer, although the number of CD4+ or CD8+ T cells can be decreased, the percentage of CD4+ CD25hi T cells in the tumor-draining and regional lymph nodes may be higher in comparison to those in lymph nodes of healthy controls [24, 25]. While we observed an increase in the total numbers of CD4 or CD8 T lymphocytes, the ratio of Tregs was significantly high in the tumor-adjacent mammary lymph nodes of MNU-induced mammary carcinomas. The conversion of naive CD4+ T cells into antigen-specifically induced Tregs directly takes place within the tumor microenvironment [26]. In our study, the presence of Tregs in the tumors but not in tumor-neighboring mammary tissue was supported by the expression of Treg-associated genes such as Foxp3, IL-10, TGF-β and NP-1. Although CTLA-4, PD-1 and Fas can be associated with both regulatory and activated T cell phenotype [9], they were selectively found in the tumors. Immune suppression in the tumor microenvironment is mediated by a unique subset of Tregs which produces both IL-10 and TGF-β [27]. Furthermore, these cytokines can also be secreted by mammary cancer cells [15, 28]. Previously, IL-10 expression by the MNU-induced tumor cells was reported [15]. Although in the current study, we did not observe high levels of IL-10 in the culture supernatants of immune cells isolated and in the sera of tumor-bearing animals, a significant increase of TGF-β in all immune compartments were evident. Thus, in our model, TGF-β might be the major immunosuppressor cytokine secreted by the immune cells in the tumor microenvironment and in various compartments of tumor-bearing animals.

The presence of CD25 on CD8+ or CD4+ (non-regulatory) T cells may indicate some level of activation especially in tumors [29]. TNF-α and IFN-γ expression in the tumor supports an inflammatory microenvironment. However, the most probable source of IFN-γ is the tumor-infiltrating immune cells and the analysis of these cells shown that the level of IFN-γ was very low and upon short-term stimulation, its level was decreased in contrast to (or as an indirect result of) increased secretion of TGF-β. Furthermore, CD40L, the critical costimulatory molecule mainly expressed by helper T cells, was not detected. This may indicate a deficit in the activation status of T cells [30]. Hill et al. [15] reported no CD25, TNF-α and IFN-γ expression in MNU-induced tumors. However, their study was based on immunohistochemical analyses and the methods employed in the current study are more sensitive for immunological analyses, so, this contradiction may be arisen because of the methodological difference.

Consistent with our findings, an unusual compartmentalization of CD8+ Foxp3+ Tregs was reported in murine cancers [31]. In the presence of TGF-β, increased numbers of NKT cells promote the generation of CD8+ Treg subset which suppresses the anti-tumor responses directly by killing immune cells or indirectly by co-opting the other cells to secrete suppressive molecules such as TGF-β, IL-10 and indolamine 2,3-dioxygenase (IDO) [32, 33]. In addition, CD8+ Tregs do not function unless they produce and respond to IFN-γ [32]. Correspondingly, in the inflammatory scene of chemically-induced tumors used in this study, the ratio of CD4+ or CD8+ Tregs, NKT cells and the levels of TGF-β were selectively increased in the presence of IFN-γ.

CD161+ CD3+ NKT cell population in the tumors was significantly high. Since tumor-infiltrating NKT cells can suppress tumor immunity through TGF-β production [34], this may also constitute for an additional immune evasion mechanism in the breast tumors studied. In addition, Tregs have the potential to repress anti-tumor activity of type I NKT cells by cell-to-cell contact [35].

According to our findings, the proportion of CD4+ to CD8+ T cells was significantly decreased in the mammary lymph nodes of tumor-bearing animals. However, an inverse situation was reported in a study performed with MTAG (MMTV-PyMT/B6) transgenic mice [36]. More specifically in the lymph nodes, the amount of Tregs (CD25hi, Foxp3+) and activated (CD25low) CD4+ T cells was rather low compared to that of in the tumor tissue. Both Foxp3+ regulatory and anti-tumor T cells are primed in the tumor-draining lymph nodes during disease progression [37]. The lack of antigen presentation via inhibition of DC migration into the tumor-draining lymph nodes emerges as novel immune escape mechanism of tumors [38]. In addition, the expression of immune suppressor mediators can be higher in breast cancer sentinel lymph nodes than in non-sentinel lymph nodes while the level of stimulatory cytokines are not affected [39]. Correspondingly, diminished proliferation capacity of the cells isolated from spleen and lymph nodes of tumor-bearing group may indicate the lack of stimulation in vivo or the anergic state of T cells. Also, the tumor-infiltrating T cells were intriguingly unresponsive to PHA-stimulation. The significant decrease in IFN-γ in the sera of tumor group may indicate a systemic immune suppression. Thus, as currently observed in the chemical carcinogen-induced mammary carcinoma model of the rat, the immune regulatory mechanisms applied by breast tumors may hinder immune stimulation both in the local tumor environment and in the tumor-draining lymph nodes.

In the spleen of tumor-bearing animals, a significant increase versus decrease in the ratio of CD4+ and CD8+ T cells was evidenced, respectively. A similar situation in splenic T cells was previously reported in a transgenic model of mammary cancer [36]. In cancer, spleen can be the target organ for migration of CD4+ T cells expressing homing receptor CD62L in where they serve as the suppressor-inducer cells with TGF-β production to induce mainly CD4+ CD25+ regulatory T cells [40]. Our findings on splenocytes demonstrated the increased ratio of Tregs and production of TGF-β in tumor-bearing animals, besides IL-10 was abundantly secreted compared to that of the cells isolated from other compartments. Other than suppressor T cells, myeloid-derived suppressor cells can be harbored by the spleen of mice with breast cancer [41, 42]. Accordingly, the spleen may serve as an important site for immunosuppressive process.

The suppressive features of tumor microenvironment, e.g. the presence of Tregs, have been considered to be one of the major obstacles for an efficient immunotherapy, especially in which the T cell responses are directed against a specific antigen [43]. Here, in a chemically-induced mammary tumor model of rat, in addition to the tumor tissue, spleen and mammary lymph nodes were determined as critical sites of immune modulation. The differential distribution of T cell subsets, especially Tregs and NKT cells, and the engulfing inflammatory network in the immune compartments must be considered in the immune responses against tumors.

Acknowledgment

This study was supported by Eczacibasi Scientific Research and Award Fund, and Hacettepe University Scientific Research Unit. We thank The Turkish Academy of Sciences for additional support. We gratefully acknowledge the administrative help of Dr. Hande Canpinar in the conduct of the project.

Ancillary