Predominant overexpression of CD 25 / FOXP 3 , IFN-γ , and suppressive cytokines in high-grade lesion samples infected with human papillomavirus

Introduction: Human papillomavirus (HPV) persistent infection is the leading cause of cervical cancer and its precursor lesions, and the inappropriate immune response is among the factors that contribute to viral persistence. This may be influenced by regulatory T (Treg) cells and the production of immunosuppressive cytokines, such as transforming growth factor beta (TGF-β) and interleukin-10 (IL-10). Objective: We established the profile of the predominant response, Th1 or immunosuppressive response, in the tissue microenvironment, by detecting interferon-gamma (IFN-γ), TGF-β, and IL-10, as well as the co-expression of IL-2 receptor alpha (CD25) and forkhead box P3 (FOXP3). Methods: Seventy-four samples from uterine cervix biopsies that underwent HPV deoxyribonucleic acid (DNA) detection and histopathology analysis were immunostained to detect CD25/FOXP3, IFN-γ and suppressive cytokines in lymphocytes. Results: The microenvironment of high-grade squamous intraepithelial lesion (HSIL) samples with high numbers of viral particles (≥ 10,000 copies/ml) contained high numbers of CD25/FOXP3+, TGF-β+, IL-10+, and IFN-γ+ cells. Conclusion: The co-expression of CD25/FOXP3 and the expression of TGF-β, and IL-10 in HSIL samples suggest the existence of Treg cells in these locations, although IFN-γ expression was observed in several cells in these samples. Our data suggest that this cytokine could be related to immunosuppressed microenvironment maintenance, favoring the persistent HPV infection and the progression to carcinoma.

Camila M. Bonin; Cacilda T. J. Padovani; Alda Maria T. Ferreira; Leandro S. Ávila; Ana Paula Machado; Thiago T. M. Prata; Carlos Eurico S. Fernandes; Inês Aparecida Tozetti found in HSIL than in low-grade squamous intraepithelial lesion (LSIL) (10) .This decrease may be related to the presence of cells with immunoregulatory properties, capable of interfering with antigen presentation (11) and with the activation and proliferation of antigen-specific T cells (12) .
Among the various cell subsets with immunoregulatory properties, regulatory T (Treg) cells are strongly associated with neoplastic progression.They are considered as a predictive factor for poor prognosis when present in high density in tumor microenvironments (13) .Such cells can be phenotypically characterized by the expression of the alpha chain of the interleukin 2 receptor, IL-2Rα (CD25) and the transcription factor forkhead box protein P3 (FOXP3) (14) , which are essential components in the generation, maintenance, development, and function of Treg cells (15) .One way by which Treg cells perform their function is through the production of immunosuppressive cytokines such as interleukin-10 (IL-10) and transforming growth factor beta (TGF-β).They are then able to trigger the anergy of T cells and allow cancer progression through the local production of these cytokines (16,17) .Thus, the presence of Treg cells can also be associated with the detection of IL-10 and TGF-β in the tissue microenvironment.
Taking into account that HPV infection is restricted to the epithelium (18) , the understanding of the active immune mechanisms in the infected tissue microenvironment is fundamental.Thus, in the present study, we aimed to detect the co-expression of the markers CD25/FOXP3 and the expression of IFN-γ, IL-10, and TGF-β in lymphocytes, by performing immunohistochemistry, to help elucidate the prevailing T helper response profile (Th1 or immunosuppressive response) in the HPVinfected microenvironment, and to provide a better understanding of the pathological processes associated with persistence or regression of the infection, or progression to the malignant form.

Samples
This is a cross-sectional study in which seventy-four biopsy samples of uterine cervix were collected from June 2010 to June 2011 at the Cancer Prevention Center of Campo Grande (MS), Brazil.These samples were selected in a non-probability sampling by convenience.A first fragment of these samples was previously subjected to HPV viral load quantification (VL) and viral typing.Quantification was done by real-time polymerase chain reaction (qPCR) with SYBR Green using LightCycler ® .Copy number quantification was determined in deoxyribonucleic acid (DNA) extraction volume.This, in turn, was classified as follows: VL0 (negative for HPV), VL1 (1,050 to < 10,000 viral copies/ml) and VL2 (≥ 10,000 viral copies/ml) (Table 1).For viral typing, the PCR end point was used followed by restriction fragment length polymorphism (RFLP) analysis (19) .The primers used were those described by Payan et al. (2007)  (20) .
A second fragment of the same samples was embedded in paraffin and sent to histopathology and immunohistochemistry. Based on the histological analysis, the samples were classified as LSIL (CIN I), HSIL (CIN II, III), carcinoma, and negative for intraepithelial lesion and malignancy (NILM) (Table 1).
The Universal LSAB Kit/HRP (Dako, Carpinteria, CA, USA) was used as the detection system for simple marking, and diaminobenzidine (DAB) (Dako, Carpinteria, CA, USA) was used as the chromogen.The Vectastain ABC system was employed for marking multiple antigens, enzyme substrate NovaRED (Vector Laboratories, Burlingame, CA, USA), and plated DAB (Ni-DAB) (Vector Laboratories, Burlingame, CA, USA) were used for the detection of the double staining.Counterstaining was performed using hematoxylin.Human tonsil tissue stained using the primary antibody was used as a positive control.A negative control was obtained by replacing the primary antibody by phosphate buffer (pH 7.4) containing 1% albumin.

Quantitative analysis
The quantification of the immunostained cells to establish cytokines expression (IL-10, TGF-β and IFN-γ) and Treg cells was performed by evaluating 10 random fields including only stromal area of cervical tissue.According to the presence of immunomarked cells, the histological sections were classified in low (when there was up to three stained cells per field) and high (when there were more than three stained cells per field) scores (21) .The biopsy sections were analyzed by two independent observers, previously calibrated (κ = 0.98), and the final result of discordant cases was obtained by common analysis to achieve a consensus.

Statistical analysis
Statistical analysis was performed using SPSS software version 17.0 and BioEstat version 5.0.The frequency analysis of histology and viral load in accordance with the intensity of expression of the markers was compared by Pearson's chi-squared test for contingency tables.When significant, Fisher's exact test was applied.

Ethical considerations
This study was approved by the Ethics Research Committee of Universidade Federal do Mato Grosso do Sul (UFMS), Brazil, protocol number 87527, August 30, 2012.
The distribution of histopathological findings according to the expression of IFN-γ, TGF-β, IL-10, and co-expression of CD25/ FOXP3 is presented in Table 3.
We emphasize that the difference in the number of samples used among the IHC markers in this study was due to the scarcity of material.Some samples did not have enough material for all markers.We also emphasize that the patients with no cytological abnormalities were not subjected to biopsy, what justifies the small quantity of NILM samples (n = 6).

DiSCuSSion
In the present study, high expression of CD25/FOXP3, TGF-β, IL-10 and IFN-γ was observed in the microenvironment of tissue samples with high-grade lesions, as well as in samples with high levels of HPV viral copies.Cytokine expression and Treg cells were not evaluated in the epithelial area, since the IL-10 expression and Treg cells had been evaluated in this location by our group in another study (22) .
Among the analyzed samples, there was a predominance of HPV 16 and 18 in high-grade lesions and carcinoma samples.Our results support the notion that high-risk oncogenic HPV infection, mainly HPV 16 and 18, is the major cause of cervical cancer and its precursor lesions (2) .
In most cases, the production of high levels of IFN-γ by activated Th1 cells is associated with the control on the progression of HPV-associated lesions (9) , because IFN-γ plays an essential role in defending against viruses and intracellular pathogens, and inducing inflammatory responses (6) .Moreover, this cytokine also 16.4% (11/67), 15.3% (11/72), 16.7% (12/72) and 10.1% (7/69) were cells expressing CD25/FOXP3, TGF-β, IL-10 and IFN-γ, respectively.increases expression of major histocompatibility complex (MHC) class I on tumor cells, thereby increasing the immunogenicity of these cells (23) .However, the development of high-grade lesions induced by HPV was shown to be associated with a decrease in the level of IFN-γ (1) .A study assessing mRNA levels of sub-epithelial IFN-γ on precancerous samples infected by HPV16 showed a significant decrease in IFN-γ mRNA levels in precancerous lesions when compared to normal tissue.This study also showed that the level of IFN-γ mRNA was significantly lower in CIN II and CIN III (HSIL) than in CIN I (LSIL), indicating that IFN-γ expression decreased according to the degree of the lesion (10) .

figurE − Immunohistochemical double staining of CD25/FOXP3 and simple staining of TGF-β, IL-10 and IFN-γ A) positive control human tonsil stained with anti-human CD25 and anti-human FOXP3 antibodies; B) negative control human tonsil with omission of the primary antibodies, showed no staining; C) HSIL, with CD25 + FOXP3 + cells in large quantities; D) positive control human tonsil stained with anti-human TGF-β antibodies; E) negative control human tonsil with omission of the primary antibodies, showed no staining; F) HSIL, with TGF-β + cells in large quantities; G) positive control human tonsil stained with antihuman IL-10 antibodies; H) negative control human tonsil with omission of the primary antibodies, showed no staining; I) HSIL, with IL-10 + cells in large quantities; J) positive control human tonsil stained with anti-human IFN-γ antibodies; K) negative control human tonsil with omission of the primary antibodies, showed no staining; L) HSIL, with IFN-γ + cells in large quantities. All figures are presented in the same magnification (400×) CD25: interleukin-2 receptor alpha; FOXP3: forkhead box P3; TGF-β: transforming growth factor beta; IL-10: interleukin-10; IFN-γ: interferon-gamma; HSIL: high-grade squamous intraepithelial lesion.
In contrast with these findings, in our study, the analysis of IFN-γ expression and the histopathology findings (p = 0.096) indicate a large number (high) of cells expressed IFN-γ in 43.5% of the HSIL samples.Although this result was not significant, we suggest that high expression of IFN-γ in the HSIL samples can be linked with immunosuppressive mechanisms that promote progression to carcinoma.Some authors noted the invariant NKT cells (iNKT) are capable of suppressing the local immune environment in persistent HR-HPV infected cervical tissues by producing IFN-γ to induce high-grade CIN (24) .IFN-γ is also involved in the induction (25) and migration of Treg cells, in particular Treg CXCR3 + cells to the site of inflammation (26) .In fact, the CXCR3 ligand production (CXCL9, CXCL10, and CXCL11) involved in the recruitment of CXCR3 + cells is induced by IFN-γ (27,28) .Thus, the high expression of IFN-γ in the samples with high-grade lesions, carcinoma and with high viral load observed in our study, may have induced the recruitment of Treg cells to these sites, hindering the elimination of HPV and the tumor cells.However, evidence to support this suggestion cannot be provided for an experimental model, considering the characteristics of this virus in vitro culture.
The presence of Treg cells has been associated with viral persistence, the development of high-grade lesions, and their progression to cervical cancer (1,8,29) .This association may be justified by the fact that Tregs CD4 + CD25 + FOXP3 + negatively regulate the immune response at the site of HPV infection, hampering the elimination of the viruses and neoplastic cells (1,29) .
The predominance of large numbers (high) of cells coexpressing CD25/FOXP3 in high-grade lesion samples (41.8%, p = 0.022) suggests that the increase of this cell population promotes local immunosuppression, facilitating lesion progression to cervical cancer.These findings are consistent with previous studies, which reported that these cells predominate in high-grade lesions and invasive cancer (8,30) .
The high expression of FOXP3 is associated with the persistent high-risk oncogenic HPV (1) .Kojima et al. (2013) (31) have shown that the high prevalence of Treg FOXP3 in precursor lesions of cervical cancer inversely correlates with the spontaneous regression of the lesions (31) , demonstrating that Treg cells may be responsible for the viral persistence and progression of the lesion.In the present study, we observed a predominance of large numbers (high) of CD25/FOXP3 positive cells in samples with high viral load (VL2) (p = 0.049).This reinforces the negative role played by these cells in viral clearance, i.e., their influence on the maintenance and persistence of high levels of viral copies and subsequent progression to neoplasia.
Other immunosuppressive components, associated with the persistent HPV infection and the development of lesions to cervical cancer, are TGF-β and IL-10 (8,32) .These cytokines are part of one of the main immunosuppressive mechanisms of Treg cells (33) , capable of suppressing the Th1 response (34) .
Besides being one of the cytokines involved in the immunosuppression performed by Tregs (35) , TGF-β is also involved in the maintenance of peripheral Treg cells (adaptive/induced).Once TGF-β is secreted, it induces the expression of FOXP3 in CD4 + CD25 -FOXP3 -naïve peripheral T cells, converting them into FOXP3 + Treg cells (36,37) .In this context, the large number of cells expressing TGF-β in HSIL (56.9%; p = 0.001) observed in our study, added to the large number of cells expressing CD25/FOXP3 in the same lesion, suggesting that the transformation of CD4 + CD25 -FOXP3 -T cells naïve in FOXP3 + occurred, allowing for the maintenance of induced Treg cells in these locations.Furthermore, the association between the expression of TGF-β and FOXP3 and the histopathological findings (p < 0.05) may suggest that, in HSIL and carcinoma samples, Treg cells producing TGF-β are present and involved in the evolution of the lesions to cervical cancer.
Regarding the association between the expression of TGF-β and the viral load (p = 0.034), we observed that 73.6% of the samples with high viral load presented a large amount of cells expressing TGF-β.This is an indication that, besides being involved in the formation of immunoregulatory cells, this cytokine fulfills its suppressive function on the immune response, possibly by inhibiting proliferation and activating effector T lymphocytes and NK cells, thereby contributing to the immune escape of the virus and the neoplastic cells (36)(37)(38)(39) .IL-10, another major cytokine secreted by Treg cells, has been associated with cervical cancer (40) .Its expression is directly associated with the degree of cervical lesions, increasing with the severity of the lesion (41) .In the present study, we observed a predominance of large numbers of IL-10 expressing cells (high) in HSIL samples (54.2%, p = 0.03).Corroborating these findings, the expression of IL-10 has been shown to increase during the different stages of cervical cancer, being more predominant in high-grade lesions and carcinoma (41) .Taken together, our results reflect the existence of a well-established immunosuppressed microenvironment, which greatly contributes to the persistent infection and progression of the lesion.
Because of its various immunosuppressive mechanisms (42,43) , IL-10 has been shown to contribute to the development of cervical neoplasia associated with HPV by favoring viral persistence (44) .In our study, we observed that 72.2% of the samples with high viral load (VL2) presented large numbers (high) of IL-10 expressing cells (p = 0.6).Although this result was not significant, it is worth highlighting its importance for the occurrence of the persistent infection, since high levels of IL-10 contribute greatly to the failure of viral clearance (44) .
In our study, the main limitations were the small number of samples obtained in the collection period and the shortage of tissue obtained in the biopsy, which made it impossible to make all the necessary slides to perform all the markings in the same sample.

ConCLuSion
Our study evaluated the microenvironment immune status of cervical samples representing different histopathological and HPV infected statuses, mostly for high-risk oncogenic HPV.The microenvironment of high-grade lesion presented large numbers of CD25/FOXP3 + , TGF-β + , IL-10 + , and IFN-γ + cells.The co-expression of CD25/FOXP3 and the expression of TGF-β and IL-10 in these samples suggest the existence of Treg cells in the microenvironment.Although a great number of cells expressing IFN-γ have also been observed in these samples, our data suggest that this cytokine could be related to immunosuppressed microenvironment maintenance, favoring the persistent HPV infection and the progression to carcinoma.

taBLE 3 −
Distribution of histopathological findings according to the co-expression of CD25/FOXP3 and the expression of TGF-β, IL-10 and IFN-γ