Evaluation of PTEN and CD4+FOXP3+ T cell expressions as diagnostic and predictive factors in endometrial cancer: A case control study.

To evaluate the potential role of Pten and CD4FOXP3 T cells in prognosis from endometrial cancer.Tissue samples and clinical data were collected from 200 patients with endometrial cancer and 100 control patients with benign uterine diseases. The expressions of Pten and CD4FOXP3 T cells were quantified by immunohistochemistry and immunofluorescence. After surgery, all patients were followed up for an average of 56.3 months. Surgical effects were evaluated based on the patients' symptoms and signs. A two-sided P value < .05 was considered significant.Pten diminished and CD4FOXP3 T cells significantly accumulated with the progression of endometial cancer, in comparison to the controls. Moreover, Pten expression was negatively correlated with the count of CD4FOXP3 T cells. Pten and CD4FOXP3 T cells were correlated with clinical characteristics, including tumor stage, differentiation and associated with patients' disease-free survival.Limited data were available between the expressions of Pten and CD4FOXP3 T cells in patients with endometrial cancer. Our study findings suggested that the expressions of Pten and CD4FOXP3 T cells might become possible biomarkers for the diagnosis and prediction in endometrial cancer.

Introduction

Endometrial cancer (EC) is the second most common cancer among females worldwide. It is more frequent in the developed countries than that in the developing countries, with an incidence of 1.6% and of 0.6%, respectively,[ with more than 61,380 new cases diagnosed in America in the year of 2017.[ However, there is no effective way to early detection.

Nowadays, the early diagnosis of EC is generally based on the clinical manifestation such as postmenopausal bleeding or abnormal serum levels of certain tumor markers. While, about 15% of ECs occur in women without vaginal bleeding.[ Previous literatures reported the role of different serum markers in the diagnosis of endometrial cancer such as the carcinoembryonic antigen (CEA), the carbohydrate antigen-125 (CA-125), and the carbohydrate antigen 19-9 (CA19-9), but these markers were elevated in only 20% to 30% of patients.[

In recent years, with the development of gene sequencing, the roles of gene regulation and the microenvironment in the development of carcinogenesis have been focused and better understood. It is highly potential that some gene markers might also be related to the development of ECs that could be promising in the diagnosis especially early diagnosis of ECs.

Pten, also known as the late tumor mutated gene, locating at 10q23.3, is 200 kb in length, spans 9 exons and 8 introns, and was discovered by researchers from the Columbia University in 1997.[ Pten maintains normal cellular activities by negatively regulating multiple signal transduction pathways, inhibiting the proliferation and migration of tumor cells, and inducing apoptosis of tumor cells, and play a key role in the development of immune response. Accordingly, Pten mutations have been found in various cancers.[ Indeed, Pten tends to become gradually silenced as lung, prostate, colon, and bladder cancers progress.[ In addition, even a slight reduction in Pten expression has a major impact on cancer susceptibility in animal models.[ Previous studies also found that about 20% of patients had endometrial atypical hyperplasia harbor Pten mutations, indicating that the loss of Pten might be an early event in the development of endometrial cancer. Animal experiment also found that Pten+/− mice develop endometrial atypical hyperplasia, of which approximately 20% progress to highly differentiated cancer.[

Immune tolerance is an important phenomenon as a specific non-response or low-response state when immunocompetent cells encounter antigenic substances. Accordingly, this state enables tumor cells escape from the immune surveillance, and limits the effectiveness of tumor immunotherapy. Regulatory T cells, which was firstly described by Gershon and Kondo in the early 1970s,[ are now known to induce and maintain immune tolerance. These cells are sub-divided into CD4+ and CD8+ subtypes based on the surface markers, secreted cytokines, and mechanisms of action. CD4+CD25+FOXP3+ T cells (Treg cells) derived from the thymus are among the most widely studied T cell populations.[

CD4+FOXP3+ T cells in the peripheral blood were reported to be more abundant in patients with breast cancer, colorectal cancer, esophageal cancer, gastric cancer, hepatocellular carcinoma, leukemia, lung cancer, lymphoma, melanoma, ovarian cancer, and pancreatic cancer. These peripheral CD4+FOXP3+ T cells strongly inhibit T cells in vitro, and hence may suppress immune activity against tumor-associated antigens. Indeed, tumors repress immune activity against through transfer, differentiation, and proliferation of Treg cells, enabling escape from immune surveillance and thus continued growth and development.[ For example, the accumulation of CD4+FOXP3+ T cells in the tumor microenvironment and the abundant expression of FOXP3 promote ovarian cancer development in chimeric animal models with severe combined immunodeficiency, presumably by inhibiting cytotoxic CD8+ T cells that target tumor-associated antigens. The accumulation of CD4+FOXP3+ T cells in tumor tissues is also significantly associated with poor prognosis.[ Conversely, conditional knockout of CD4+FOXP3+ T cells was found to impede the progress of melanoma and colon cancer in mice, as well as to elicit massive apoptosis in metastatic breast cancer cells.[

Recently, it has been revealed that PTEN can increase Treg stability and that the loss of PTEN can lead to spontaneous inflammatory disorder.[ The PI3K/AKT/mammalian target of rapamycin (mTOR) signaling pathway is pivotal for Treg cell development and homeostasis.[ And PTEN is one of the pivotal controler in PI3K/AKT/mTOR signaling pathway, which can present in the microclusters of Treg cells and modulate the AKT phosphorylation status in immune synapses in a coordinated manner.[ This pathway is activated downstream of the T-cell receptor (TCR), CD28, and IL-2 signaling. It is critically involved in Treg cell thymic development, peripheral expansion, and suppressive activity.[ These observations highlight PTEN's importance of a stringent negative regulation of PI3K pathway activity in Treg cells.[

Based on the above pathway, Pten mutations have also been detected in the T cell-associated malignancies. For example, mice with specific Pten mutations exhibit increased B cell proliferation and serum autoantibodies as well as enhanced cell migration, proliferation, and resistance to apoptosis.[ In addition, an enhanced proliferation of CD4+FOXP3+ T cells, a phenotype rescued by Pten re-expression, was also shown in the Pten-deficient mice.[ Similarly, functional Pten was found to be essential to maintain the homeostasis of CD4+FOXP3+ T cells and thus of the immune system.[

However, the relationship between the Pten expression and the CD4+FOXP3+ T cells in patients with endometrial cancer has been poorly characterized. A hospital based case control study was conducted to explore the possible correlation.

Materials and methods

Patients and controls

Approval from the ethics committee of West China Second University Hospital, Sichuan University, China, was obtained at the beginning of the project. Two hundred patients with endometrial cancer were recruited on admission to the hospital from January 2014 to January 2015. These patients were diagnosed by curettage of uterus, and each diagnosis was confirmed by two pathologists. Tumors were staged according to the 2009 criteria for adenocarcinoma of the endometrium, as defined by the International Federation of Gynecology and Obstetrics.

For the case group, inclusion criteria were:

pre-operative pathological diagnosis of endometrial cancer;

no chemotherapy or radiotherapy;

no hormone therapy;

good compliance.

The exclusion criteria were:

women with a previous history of tumors;

with chemotherapy or radiotherapy history;

with hormone therapy history;

patients with severe internal and surgical diseases.

Another 100 age-matched females, who underwent hysterectomy at the hospital due to benign uterine diseases, and without history of cancers during the same time period, were recruited as controls. Written informed consent was obtained from all patients, in compliance with the Declaration of Helsinki.

Sectioning of tissues

Endometrium tissues were collected during surgery, and sectioned at 4 μm. Clinical and histological data were collected.

Immunohistochemistry for Pten

As part of routine clinical practice, three samples of primary tissue per patient were randomly selected and assayed for Pten by conducting immunohistochemistry (Abcam, USA). Staining pattern, intensity, and distribution were evaluated by using light microscopy. Pten was considered present if distinct brownish nuclear staining was observed. Stained mononuclear cells were reported as percentage of all immune cells in the field.

Immunofluorescence for CD4+FOXP3+T cells

Three samples of primary tissue per patient were randomly selected. Paraffin-embedded slides were deparaffinized, rehydrated, microwaved in the EDTA pH 9.1 to retrieve antigens, and were then stained for CD4+FOXP3+ T cells by using the fluorescently labeled rabbit anti-human CD4 (Abcam, USA) and the mouse anti-human FOXP3 (Abcam, USA). Slides were imaged under a fluorescence microscope. Images obtained at different wavelengths were superimposed in Image Processing and Analysis in Image J (National Institute of Health, USA) to identify cells with fluorescently stained cell membranes, nuclei, or cytoplasm. CD4+FOXP3+ T cells were quantified in each sample as (cumulative optical density after correction of CD4 fluorescence index + cumulative optical density after correction of FOXP3 fluorescence index)/2.[

Statistical analysis

Data were analyzed by using SPSS version 22.0 (IBM Company, USA) and are reported as mean ± standard deviation and range. Categorical variables were compared between patients and controls by using Chi-squared test. After testing for normality of the raw data, the correlation between clinical factors and levels of Pten and Treg cells was analyzed by using student t-test. A two-sided P value < .05 was considered significant. Figures were generated in GraphPad Prism 5 (GraphPad Software Inc., USA).

Results

As listed in Table 1, cases and controls have comparable clinical characteristics, with P values > .05.

Table 1

Clinical characteristics of patients and controls.

Pten expression in endometrial tissues

Abnormal proliferation of glandular tissues, structural disorder, heterotypic cells, large nuclei, irregular morphology, enhanced mitosis, less differentiation, smaller glandular structures, and obvious lymphocyte infiltration were observed in endometrial tissues collected from patients, as assessed by immunohistochemistry and microscopy (Fig. 1).

Figure 1

Immunohistochemical staining expression of Pten in endometrial normal tissues and adenocarcinoma of all grades. (A) Normal, ×20, ARROW: normal nuclei and glandular tissue structure; (B) highly differentiated, ×20, ARROW: approximately normal glandular structure but with enhanced mitosis and large nuclei; (C) moderately differentiated, ×20, ARROW: large nuclei, structural disorder of glandular tissues; (D) Poorly differentiated, ×20, ARROW: structural disorder and irregular morphology of glandular tissues, heterotypic cells.

Pten expression was the highest in normal endometrial cells, and was the lowest in G3 endometrial cancers, with an average staining score of 6.6 points for the control tissues, however, of 4.4 points for the tumor tissues. The average staining scores for highly differentiated (Grade 1, G1), moderately differentiated (Grade 2, G2), and poorly differentiated (Grade 3, G3) tissues were 5.9, 3.5, and 1.7 points, respectively. Tumors with staining score < 4.4 were considered as weakly expressing Pten (98/200 cases, 49%), while those with staining score ≥ 4.4 were considered as strongly expressing (102/200 cases, 51%). Remarkably, we found that Pten staining intensity was associated with diabetes in cancer patients, and with the clinical stage and differentiation grade of the tumor (Table 2).

Table 2

Association between Pten expression and clinical characteristics in patients.

Similarly, control tissues with staining score < 6.6 were considered as weakly expressing (44/100 cases, 44%), while those with staining score ≥ 6.6 were considered as strongly expressing (56/100 cases, 56%). However, we found that Pten staining intensity was not significantly associated with menopause, obesity, diabetes, and hypertension (Table 3).

Table 3

Association between Pten expression and clinical characteristics in control volunteers.

Relationship between Pten expression and diagnosis

Furthermore, we analyzed the ROC of Pten expressions and the AUC was 0.839, 95% CI:0.795–0.884. The best cut-off value was 4.75 (Fig. 2). Based on the value, the specificity was 88%, and sensitivity was 65.5%.

Figure 2

The ROC curves of Pten expressions among the subjects.

Relationship between Pten expression and prognosis

Excluding 11 patients (3 with weak and 8 with strong expression) who were lost during follow-up, the average disease-free survival (DFS) was 54.53 months and 58.6 months (P = .048) for patients with tumors weakly (97 cases) and strongly (92 cases) expressing Pten.

For patients with weakly infiltrated tumors, 23 cases were recurrenced. The earliest and the latest recurrence times were 5 and 55 months after surgery, respectively, with the median of 29.9 months. Detailed, for patients with strongly infiltrated tissues, 10 cases relapse occurred between 17 and 55 months after surgery with the median time to recurrence of 36.5 months (Fig. 3). Univariate statistical analysis also showed that Pten expression was statistically associated with postoperative recurrence (Table 4).

Figure 3

Relationship between Pten expression and prognosis into tumors.

Table 4

Univariate analysis of Pten staining intensity and disease recurrence.

Accumulation of CD4+FOXP3+ T cells in endometrial tissues

Immunofluorescence staining showed that CD4+FOXP3+ T cells were less abundant in normal endometrial tissues than that in endometrial carcinomas, with average staining scores of 0.7 and 3.3, respectively. Accordingly, tumors with staining score < 3.3 were considered as weakly infiltrated with CD4+FOXP3+ T cells (117 cases, 58.5%), while those with staining score ≥ 3.3 were considered as strongly infiltrated (93 cases, 46.5%). Staining scores for highly, moderately, and poorly differentiated were 1.2, 2.8, and 4.7 points, respectively (P = .015) (Fig. 4). We found that infiltration of CD4+FOXP3+ T cells was related to the presence or absence of BMI, obesity, as well as to tumor differentiation (Table 5). However, there was no significant relationship between CD4+FOXP3+ T cells and the clinical data in the controls (Table 6).

Figure 4

Immunohistochemical staining expression of CD4+FOXP3+ T cell in endometrial normal tissues and adenocarcinoma of all grades. (A-1) CD4+ T cells; (A-2) FOXP3+ T cells; (A-3) cell nuclei in highly differentiated tissue. (B-1) CD4+ T cells; (B-2) FOXP3+ T cells; (B-3) cell nuclei in moderately differentiated tissue. (C-1) CD4+ T cells; (C-2) FOXP3+ T cells; (C-3) cell nuclei in poorly differentiated tissue. (D-1) CD4+ T cells; (D-2) FOXP3+ T cells; (D-3) cell nuclei in normal tissue.

Table 5

Relationship between CD4+FOXP3+ T infiltration and clinicopathological features in patients.

Table 6

Relationship between immunofluorescence staining for CD4+FOXP3+ T cells and clinical characteristics of control subjects.

Relationship between CD4+FOXP3+ T cells expression and diagnosis

The ROC of Pten expressions and the AUC was 0.759, 95% CI: 0.687–0.831. The best cut-off value was 1.15 (Fig. 5). Based on the value, the specificity was 78%, and sensitivity was 83.5%.

Figure 5

The ROC curves of CD4+FOXP3+ T cell expressions among the subjects.

Relationship between prognosis and abundance of CD4+FOXP3+ T cells

Eleven patients (10 in week and 1 in strong) were excluded for final analysis due to the lost of follow up.

The average follow-up time for 97 patients with weakly infiltrated tumors was 59.6 months. Among them, 10 cases suffered recurrence, and the earliest and latest recurrence times were 17 and 55 months after surgery, respectively, with a median of 37.6 months. The mean DFS was 58.42 months (Fig. 6).

Figure 6

Relationship between prognosis and CD4+FOXP3+ T cell expression into tumors.

The average follow-up time for 92 patients with strongly infiltrated tissues was 52.78 months. Twenty-two relapse cases occurred between 5 and 55 months after surgery. The median time to recurrence was 27.1 months. Mean DFS was 54.53 months (P = .044, Fig. 4). Univariate statistical analysis showed that CD4+FOXP3+ T cell infiltration was statistically associated with postoperative recurrence (Table 7).

Table 7

Univariate analysis of CD4+FOXP3+T cell infiltration and disease recurrence.

Correlation between Pten and CD4+FOXP3+ T cell expression

Stainings for Pten and CD4+FOXP3+ T cell were normally distributed in the controls, highly, moderately, and poorly differentiated tissues. Pearson correlation analysis showed that Pten expression was negatively correlated with CD4+FOXP3+ T cell expression in all groups (Fig. 7).

Figure 7

Pearson correlation analysis between Pten expression and CD4+FOXP3+ T cells in (A) highly differentiated tissues, r = −0.6869, R2 = 0.4718, P < .001, 95% confidence interval (−0.7944, −0.5376); (B) moderately differentiated tissues, r = −0.6688, R2 = 0.4474, P = .032, 95% confidence interval (−0.7765, −0.5232); (C) poorly differentiated tissues, r = −0.4818, R2 = 0.2341, P = .02, 95% confidence interval (−0.6639, −0.2457); (D) control tissues, r = −0.4157, R2 = 0.1728, P < .001, 95% confidence interval (−0.5659, −0.2388).

Discussion

In this study, we analyzed the expression of Pten and CD4+FOXP3+ T cell among 200 patients with endometrial cancer who were treated in our hospital, and another 100 controls who underwent hysterectomy for benign gynecological diseases. There was no significant difference between the two groups in terms of age of disease onset, body mass index, family history of cancers, menopausal status, and gestational age.

The Pten mutations and deletions are unremarkable in esophageal cancers, although Pten expression at mRNA or protein level is exceptionally low, implying that defects may happen during the transcription and the translation.[ Gabriel et al also reported that Pten deletion in prostate cancers was predictive to the tumor stage and grade.[ However, among the endometrial diseases, Mutter et al found that Pten mutations were present in about 20% of patients with endometrial atypical hyperplasia, and in about 83% of patients with endometrial cancer. As Pten mutations are normally absent in the healthy population, suggesting that loss of Pten may be an early event in the process of endometrial carcinogenesis.[ In addition, Konopka et al reported that the severity and nature of Pten mutations in endometrial cancer were closely related to the tissue localization and their clinical features.[ Janiec et al also found that Pten mutations in endometrial cancer is positively related to its differentiation level, the more Pten mutations, the higher differentiation level. Also, the Pten mutation was found more common in patients < 55 years than that in the older patients.[ Our study had consistent findings with the previous studies that the average Pten expression was the highest in the normal endometrial tissues, and negatively correlated with tumor grade in endometrial adenocarcinoma, association with the clinical stage, FIGO stage and differentiation grade of the tumor.

CD4+FOXP3+ T cells are important mediators for tumor immune escape, and are currently considered the most representative of immunoregulatory CD4+ T cells. Importantly, immune escape is associated with prognosis from cancer. For example, levels of CD4+FOXP3+ T cells in peripheral blood were significantly enhanced in patients with breast cancer, colorectal cancer, esophageal cancer, gastric cancer, hepatocellular carcinoma, leukemia, lung cancer, lymphoma, melanoma, ovarian cancer, and pancreatic cancer. These cells may inhibit immune responses to tumor-associated antigens.[ In addition, mice with lymphoma and specific Pten mutations in T cells showed increased CD4+ T cell proliferation, elevated serum auto-antibodies, and enhanced lymphoma cell migration and proliferation.[ In this present, we also found that levels of CD4+FOXP3+ T cells in endometrial adenocarcinomas were significantly increased. The increased levels of these cells suggested later tumor stage and more severe grade, which indicates worse prognosis. Correlation analysis confirmed that Pten expression was negatively correlated with CD4+FOXP3+ T cell expression. Indeed, Pten expression decreased with tumor grade, normally accompanied with CD4+FOXP3+ T cells accumulated.

Diabetes is a recognized risk factor that promotes the development of endometrial cancer.[ Our studies also have shown a negative correlation between Pten level and diabetes, with higher Pten levels in diabetes patients, which confirmed the link between diabetes and endometrial cancer. It also may suggest an intrinsic relationship between serum Pten and endometrial cancer.

Conclusion

In comparison with normal endometrial tissues, Pten expression in endometrial adenocarcinomas was weaker, and the degree of attenuation was roughly correlated with tumor grade. In contrast, accumulation of CD4+FOXP3+ T cells was positively correlated with tumor grade. Loss of Pten expression and accumulation of CD4+FOXP3+ T cells in lesions were inversely related to prognosis. The data suggested a possible link between serum Pten and CD4+FOXP3+ T cells in patients with endometrial cancer and may also be useful biomarkers for endometrial cancer prognosis. We also speculated that there may be a potential regulatory pathway between Pten and CD4+FOXP3+ T cells. Such pathway is worthy further investigations as a new avenue for targeted immunotherapy of endometrial adenocarcinoma.

Acknowledgments

Pro. Yang Fan, Department of Pathology, The West China Second University Hospital, Sichuan University, contributed to analyzing the ICH results.

Author contributions

Conceptualization: Zeng Xi, Liao Guang-Dong.

Data curation: Zeng Xi, Lan Zhu, Liao Guang-Dong.

Formal analysis: Zeng Xi, Li Jing, Liao Guang-Dong.

Funding acquisition: Li Jing, Xi Ming-Rong.

Investigation: Zeng Xi, Li Jing, Kang Le-Ni, Deng Ze-Wen, Ye Hui, Xi Ming-Rong.

Methodology: Zeng Xi, Li Jing, Kang Le-Ni, Deng Ze-Wen, Ye Hui, Xi Ming-Rong, Liao Guang-Dong.

Project administration: Zeng Xi, Lan Zhu, Ye Hui, Liao Guang-Dong.

Resources: Zeng Xi, Lan Zhu, Deng Ze-Wen, Ye Hui.

Software: Zeng Xi, Li Jing, Kang Le-Ni, Deng Ze-Wen, Ye Hui, Liao Guang-Dong.

Supervision: Zeng Xi, Lan Zhu, Xi Ming-Rong, Liao Guang-Dong.

Writing – original draft: Zeng Xi, Deng Ze-Wen, Liao Guang-Dong.

Writing – review & editing: Zeng Xi, Xi Ming-Rong, Liao Guang-Dong.