Is there any relationship between PSA and increased peripheral CD4 + CD25 high FOX3 + Treg in prostate cancer patients?

Introduction : The aims of this study were ﬁ rst, to determine whether peripheral levels of CD4 + CD25 high Foxp3 + regulatory T cells (Treg) are elevated in Prostate Cancer (PCa) patients, and second, to determine the direct correlation between peripheral Treg and total serum Prostate Speci ﬁ c Antigen (PSA) levels in these patients. Methods : Peripheral Blood Mononuclear Cells from 56 subjects undergoing diagnostic prostate biopsies (PSA ≥ 2.5 ng/ml) were analyzed for Treg numbers. The association between the peripheral Treg and serum PSA values was ﬁ rst determined in the entire population, including people with no prostate pathology and PCa and Benign Prostate Hyperplasia (BPH) patients, and second, in nine PCa patients before and after curative prostatectomy. Results : This project was performed in Akdeniz University immunology laboratory and urology out patient clinic from 2008 to 2010. Peripheral Treg frequencies were signi ﬁ cantly increased in PCa patients (n = 19, 3.23 ± 1.59) compared with BPH patients (n = 27, 1.66 ± 0.80) and healthy subjects (n = 10, 1.08 ± 0.43) ( p < 0.01). The percentage of Treg in BPH patients was also signi ﬁ cantly higher than that of healthy subjects ( p < 0.01). Importantly, the increase in BPH and PCa patients paralleled the elevation in total serum PSA levels, demonstrating a strong positive correlation (r = 0.75; p < 0.01). Conclusion : These results demonstrate that peripheral Treg densities are correlated with PSA in BPH and PCa patients, suggesting that PSA may have a role in Treg induction and/or maintenance in Treg in these people.


Introduction
Prostate cancer (PCa) is the most commonly diagnosed cancer among men in the world (1). Approximately two thirds of PCa cases are confi ned to the prostate and can be treated by radical prostate removal or radiotherapy (2). In addition, approximately 25 to 55% of treated, locally confi ned tumors reappear within 10 years and may progress as either a local recurrence or distant metastases (3). In the quest for eff ective prevention and treatment modalities for metastatic PCa (4), immunotherapy attempts using several diff erent methods have shown very limited success (5)(6)(7)(8)(9). Various immune evasion mechanisms, such as defects in antigen presentation, secretion of immunosuppressive agents by the tumor cells, and T cell receptor defects, are thought to limit the success of these trials. Recently, natural regulatory T cells (CD4 + CD25 high ; Treg), one of the key regulators of self tolerance, have also been implicated in immune evasion by PCa (10,11). CD4 + CD25 high Treg cells, which are known to control self tolerance in the periphery, derive from the thymus (12,13) and constitute 1-2% of peripheral lymphocytes in adult humans (14). Th e importance of these cells in tumor immunity was fi rst demonstrated three decades ago (15). Since then, especially aft er the introduction of Foxp3  (17)(18)(19)(20). Evidence for Treg in PCa patients, however, has been limited to a few recent studies with signifi cant controversy (21). Studies from experimental models have demonstrated that Tregs may either be induced or activated within the tumor draining lymph nodes by tumor-derived factors, including tumorassociated antigens (TAA). Th en, these cells prevent tumor-specifi c immune responses either at the induction phase in the tumor draining lymph nodes or at the eff ector phase within the tumor milieu (22,23). Th e fi rst evidence that TAA-specifi c Tregs may be involved in suppression of tumor-specifi c immune responses in humans has recently been reported (24). Prostate-Specifi c Antigen (PSA), a serine protease produced by the prostate gland, is the best known TAA of the prostate. Large amounts of PSA are produced and released into circulation in PCa and BPH patients, as well as in people with prostate infl ammation. We demonstrate that the prevalence of Tregs is increased not only in PCa patients, but also in patients with BPH, and that this enhancement is strongly correlated with PSA, suggesting that PSA may have a role in Treg induction and/or maintenance in these people. Th e varying results concerning peripheral Treg densities in PCa patients and the lack of information regarding a correlation between PSA levels and Treg cell frequency led us to investigate the number of peripheral CD4 + CD25 high Foxp3 + Tregs and their association with total serum PSA levels in PCa and BPH patients. Our results support that the frequency of Tregs increase not only in PCa patients, but also in BPH patients. Th is enhancement strongly correlates with PSA levels, suggesting a role for PSA in the Treg induction and/or maintenance in these subjects.

Patients and Samples
Th is project was performed in Akdeniz University immunology laboratory and urology out patient clinic from 2008 to 2010. A total of 56 patients with total serum PSA values of >2.5 ng/ ml, who were referred to transrectal ultrasound (TRUS)-guided sextant prostate biopsy at Akdeniz University Urology Outpatient Clinic, were recruited. Th e mean age of the patients was 62 years (range 41 to 78 years). For Treg screening, 8 ml peripheral blood samples were drawn from patients into heparinized tubes just before the biopsy. None of the patients had any known current infections, and none were taking any immunomodulatory medications and none of them had any cancer before. Blood and biopsy samples were obtained with written consent under an institutional review board-approved protocol.
Of the total number of 56 patients that were analyzed, 19 were diagnosed with PCa, 27 with BPH and 10 with no apparent prostate pathology (Table 1). Of Of the 19 PCa patients, 17 had initial stage, locally confi ned cancer, while 2 had radiologically detected bony metastasis. Th e Gleason score was used to grade prostate cancer; where a high score is associated with advanced disease and poorer prognosis. Of the 17 locally confi ned cancer patients, 1 patient scored 5, 5 patients scored 6, 3 patients scored 7 and 10 patients scored 9. Of the 17 with locally confi ned PCa, 9 were treated by laparoscopic radical prostatectomy at our hospital and were re-analyzed for both total serum PSA and peripheral Treg levels one month following the surgery. 10 who were negative for BPH, PCa and prostatitis through pathological assessment were used as healthy controls. For Treg screening, 8 ml peripheral blood samples were drawn from patients into heparinized tubes just before the biopsy. Blood and biopsy samples  were obtained with written consent under an institutional review board-approved protocol.

Phenotypic and Quantitative Analysis of Lymphocytes:
Peripheral blood mononuclear cells (PBMCs) were isolated using a Fycoll-Hypaque density gradient. Surface staining with anti-human CD4-FITC (BD; 555346), CD25-PE (BD; 555432) and intracellular staining with Foxp3-APC (e-Bioscience; 17-4776-73) antibodies was performed as previously described (22,25). Briefl y, PBMCs were washed three times with D-PBS and stained for surface CD4 and CD25 markers for 30 minutes at room temperature. Finally, the cells were washed twice with saponin buff er and once with washing buff er and analyzed using a BD FACSCalibur Flow Cytometer. Flow-Jo soft ware (Tree Star Inc., San Carlos, CA) was used to analyze the samples and determine the frequencies of Treg cells. Absolute lymphocyte counts were determined by using an automated hematological analyzer (Sysmex XT-2000iV).

Serum PSA Quantifi cation:
Total PSA (free + complexed) from serum samples was measured using an electrochemiluminescence immunoassay (ECLIA) with a Roche Elecsys Modular Analytics E170 immunoassay analyzer according to manufacturer's instructions.

Statistical analysis
All of the statistical analyses were performed using SPSS (version 16; SPSS Inc.). Statistical diff erences between groups were evaluated by using Kruskas-Wallis analysis. Th e diff erences between two groups were determined by the Mann-Whitney test with Benferroni Correction. Correlation was tested by the non-parametric Spearman method. Th e statistical signifi cance (p value) was set at <0.05.

Results
In this study, in order to avoid bias, we recruited patients with serum PSA values of >2.5 ng/ml (range 2.5-51) at the time of digital rectal examination and biopsy. Of the total number of 56 patients that were analyzed, 19 were diagnosed with PCa, 27 with BPH and 10 with no apparent prostate pathology. Of the 19 cancer patients, 1 patient had a total Gleason score of 5.5 had a total Gleason score of 6.3 had a total Gleason score of 7 and 10 had a total Gleason score of 9. Seventeen of the cancer patients had initial stage, locally confi ned cancer, while two had radiological detected bone metastasis. Ten people who were negative for BPH, PCa and prostatitis through pathological assessment were used as healthy controls. Peripheral blood samples of all the patients at the time of biopsy were analyzed by fl ow cytometry using CD4, CD25, Foxp3. Th e gating strategy used for selecting CD25high cells was very stringent (Figure 1a). Th e mean frequencies of CD4 + CD25 high Foxp3 + Treg cells as percentages of peripheral lymphocytes were determined as 3.23% ± 1.59% (n = 19), 1.66% ± 0.80% (n = 27) and 1.08% ± 0.43%, (n = 10) for PCa patients, BPH patients and healthy people, respectively (Figure 1b). Mean frequency of Tregs in PCa patients was signifi cantly higher than that of the BPH patients (p < 0.01) and healthy donors (p < 0.01). In addition, the mean frequency of Tregs in BPH patients was also signifi cantly higher than that of the healthy donors (p < 0.01).
Finally, immunohistochemical staining of prostate tissue sections from 19 PCa patients showed a substantial increase in the number of Foxp3 expressing Treg in malignant tissue ( Figure  2b) compared with benign tissue (Figure 2a). Th ere was a strong positive correlation between the two parameters with a correlation coeffi cient of 0.75 (p < 0.01). Nine PCa patients, who were treated by laparoscopic radical prostatectomy, were re-analyzed one-month post surgery for peripheral PSA and Treg levels ( Figure 3). Th e mean level of total serum PSA was 10.00 ± 5.97 and the mean frequency of peripheral Treg was 3.20±1.61 before the surgery. Aft er curative prostatectomy, serum PSA levels of all the patients were reduced to very low/undetectable levels as expected [(0.20±0.49; paired t-test p < 0.01) (Figure 4a, b)]. Strikingly, the mean Treg frequencies in these patients also decreased signifi cantly [(1.09 ± 0.32; paired t-test p < 0.01) (Figure 4c, d)]. Th is result suggests that PSA alone or in combination with other tumor derived factors may be required for the increased presence of Treg in the periphery. In order to rule out a possible post-surgery stress eff ect on Treg frequencies, we recruited six more patients, in addition to our study group described above, that had previously planned to undergo non-PCa related prostate surgeries at our hospital. Th ese people were screened for Treg frequencies before and one month aft er the surgery. In these patients, the frequencies of Treg did not change signifi cantly aft er the surgery (1.11±0.20 to 1.13±0.22; paired t-test p > 0.05), demonstrating that surgery by itself does not cause a decrease in Treg frequency (Figure 4e, f).

Discussion
Th e purpose of this study was two-fold. First, we sought to verify whether the peripheral frequencies of CD4 + CD25 high Foxp3 + Treg cells are elevated  in PCa patients; second, we aimed to determine the direct correlation between the peripheral Treg and total serum PSA levels in these patients. We fi rst demonstrated that CD4 + CD25 high Foxp3 + Treg densities are increased in PCa patients compared with BPH patients and healthy controls. In addition to this observation, which is consistent with the fi ndings of Miller et al., (10) we also recorded a signifi cant enhancement in Treg frequen-   We also do not know whether other factors, either tumor-derived or tumor-induced, are also required in these processes. Previous in vitro studies demonstrated that PSA is able to induce TGF-β production and impair dendritic cell maturation. Both of these pathways are known to induce Tregs, in vivo in experimental models and in vitro in human PBMC cultures and thus, it is reasonable to assume that excessive amounts of PSA in both BPH and PCa patients may invoke one or both of the above pathways to either induce or expand PSA-specifi c Tregs. It is also likely that other mechanisms, such as IDO (Indoleamin 2-3 deoxygenase) or PGE-2 production by the cells within the tumor milieu might contribute to the process. Further studies addressing these questions will be important for our understanding of the biology of the Treg in human cancer.
Restoring peripheral Treg levels upon removal of cancerous prostates also has important implications for immunotherapy for PCa. Many prostate-specific vaccine trials using PSA peptides to stimulate tumor-specifi c immune responses before the surgical removal of tumors have not yielded desired results.

Conclusions
In this study we present, to our knowledge, the fi rst evidence in the literature that increased frequency of circulating CD4 + CD25 high Foxp3 + Tregs in PCa and BPH patients is correlated with PSA levels and also the fi rst to demonstrate a strong dynamic association between a TAA and Tregs in cancer bearing humans. Our results demonstrate that peripheral Treg densities are correlated with PSA in BPH and PCa patients, suggesting that PSA may have a role in Treg induction and/or maintenance in Treg in these people. In the light of these fi ndings, we may expect better response of treatment of tumour specifi c immunotherapy aft er removal of prostate gland which includes adenocarcinoma cells. More comprehensive studies on this issue by holding the light of fi ndings in immunotherapy for PCa may provide new forms of treatments.