Abstract
Background: Cancer stem cells (CSCs) are tumour-initiating cells with self-renewal properties and chemo/radio-resistance. Regulatory T-cells (Tregs) influence CSCs through several mechanisms. In different solid tumours, the presence of both cell populations correlated with poor survival. In vulvar cancer, little is known regarding biological markers able to predict patient prognosis. We investigated the presence and clinical impact of CSCs and infiltrating Treg in primary vulvar cancer. Materials and Methods: Paraffin-embedded tissue specimens derived from 43 patients with vulvar cancer were analyzed by immunohistochemistry for the expression of prominin-1 (CD133), CD24, ATP-binding cassette sub-family G member 2 (ABCG2) (CSC markers) and forkhead box protein P3 (FOXP3) (Treg marker). Results: CD133 expression correlated with younger age at diagnosis (p<0.01), lymph-node metastasis (p<0.05) and larger tumour diameter (p<0.05). CD133+ tumours showed a high FOXP3+ T-cell infiltration. Overall survival and progression-free survival were not influenced by the expression of the analyzed biomarkers. Conclusion: In vulvar cancer, CSCs were more frequently expressed in younger aged patients and those with aggressive disease. Their presence was also associated with high Treg infiltration, which contributes to the generation of an immunosuppressive milieu.
- Vulvar cancer
- cancer stem cells
- CD133
- CD24
- ABCG2
Vulvar cancer is an uncommon gynaecological malignancy; approximately 4,850 new cases of vulvar cancer and 1,030 deaths from this disease were projected for the United States in 2014 (1). Most patients are elderly, with a peak of incidence in the eighth decade; nearly 30% are diagnosed at international Federation of Gynaecology and Obstetrics (FIGO) stage III or IV with a 5-year overall survival of 43% and 13%, respectively (2). Although representing a rare disease of elderly women with a current incidence of 2-3 per 100,000 women and a median age of 65-70 years old, vulvar cancer has shown an increasing incidence with concurrently decreasing median age of 55-60 years at onset over the past few decades (1, 3).
The standard treatment of vulvar cancer includes radical surgery and adjuvant radiotherapy in selected cases. Considering the median age of these patients, clinical and pathological prognostic factors are constantly being explored in order to minimize unnecessary treatments. Furthermore, new molecules are being investigated to propose target therapies and increase patient survival.
In the past decade, cancer stem cells (CSCs) have been identified as tumour-initiating cells. This subpopulation is resistant to cancer therapy such as chemo- and radiotherapies and successful treatments are dependent on the elimination of these cells (4). In addition, several reports highlight the interaction between CSCs and the immune system. Regulatory T-cells (Tregs) influence the stemness, progress and control of CSCs through regulating angiogenesis influenced by vascular endothelial growth factor (VEGF) (5, 6). CSCs impact on Tregs mainly through their recruitment and induction (7).
Prominin-1 (CD133) and forkhead box protein P3 (FOXP3) represent well-established markers of CSC and Tregs, respectively. Prominin-1 (CD133) is associated with tumour progression. Its up-regulation in tumours correlated with poor prognosis, although its function is unknown (8). FOXP3 is the master regulator of Tregs and it is essential for Treg development and function (9). In recent years, the CD24 and ATP-binding cassette transporter G2 (ABCG2) molecules have also been proposed as additional CSC markers for several solid tumour types (10). CD24 is a glycoprotein preferentially expressed by B-lymphocytes, that positively regulates the proliferation of activated T-cells, and it is also described in the central nervous system. Several studies conducted on colorectal cancer demonstrated that this molecule is associated with a poor prognosis and more aggressive phenotype of tumour cells (11). ABCG2 causes decreased drug concentration within the cells by pumping chemotherapeutic drugs out of the cell (12). In addition, it is implicated in the survival of stem cells and tumour cells in an hypoxic environment (13). In cancer, ABCG2 is associated with a poor prognosis and is often highly up-regulated in the small stem-like cell subpopulation (14).
In this work, we analyzed the expression of CD133, FOXP3, ABCG2 and CD24 in women affected by vulvar cancer, correlating these with common clinical prognostic factors.
Materials and Methods
Patient characteristics. This retrospective multi-institutional study included patients affected by vulvar cancer and treated between 1999 and 2010 in the following Italian Gynaecologic Oncology Units: Sapienza University of Rome, University of Chieti, Hospital of Bergamo. The Internal Review Boards of the involved Institutions granted approval for this study (protocol number 703/08). Data, including age, histopathology, tumour grading, lymph node involvement and treatment protocol, were obtained from clinical charts and pathological records. Follow-up was closed in March 2014.
Paraffin-embedded samples were obtained from patients affected by primary vulvar cancer and treated with radical surgery. Surgical procedures were carried out with the triple-incision technique. Patients with lesions involving only the vulva were subjected to radical vulvectomy or wide local excision in the attempt to obtain at least a 1 cm tumour-free resection margin. In cases of lower urethral involvement, a partial distal urethrectomy was performed. Superficial and deep groin lymph nodes were removed in cases of suspected metastatic involvement for optimal surgical staging. Patients with multiple lymph node metastases were treated with adjuvant radiotherapy.
Immunohistochemistry. Serial formalin-fixed, paraffin-embedded vulvar tumour samples were deparaffinised in xylene, followed by absolute ethanol, 95% ethanol and distilled water. Before immunostaining procedures, sections were incubated in citrate buffer (pH 6.0) or in EDTA buffer (pH 8.0) in a pressure cooker at 112°C to enhance immunoreactivity of samples. Endogenous peroxidase activity was quenched by treatment with 3% H2O2. After blocking of the nonspecific sites, the sections were incubated with the following mouse monoclonal antibodies: anti-CD133/1 (AC133 clone, dilution 1:50, overnight at 4°C; Miltenyi Biotech, Bergisch Gladbach, Germany), anti-CD24 (ready to use, 1 h at room temperature; Biocare Medical, Concord, CA, USA), anti-ABCG2 (6D171 clone, dilution 1:100 overnight at 4°C; Santa Cruz, Heidelberg, Germany), anti-CD3 (PS1 clone, ready to use, 2 h at room temperature; UCS Diagnostic, Morlupo, Italy) and anti-FOXP3 (MCA2376 clone, 1:200, 1.5 h at room temperature; AbDSerotec, Kidlington, UK). Markers were visualized using the DakoEnVision + System Horseradish peroxidise (HRP)-labelled polymer kit (Dako, Santa Clara, CA, USA) following the manufacturer's instructions and the tissue sections were counterstained with haematoxylin (ScyTek Laboratories, West Logan, UT, USA).
The images were acquired with Olympus BX51 microscope (Olympus Europe SE & Co. KG, Hamburg, Germany) and analyzed with IAS software (Rehlingen-Siersburg, Germany). Staining was graded according to the number of positive tumour cells as follows: <5%, negative; >5-20%, weak; >20-50%, moderate; >50%, strong. Negative control slides were incubated with MOC21 monoclonal antibody (AbCam, Cambridge, UK) as isotype control. Three independent investigators blinded to the patient clinical information evaluated all specimens.
The expression of each marker was evaluated using a x20 objective selecting 10 independent areas. In particular for the quantification of the CD3 and Treg markers, 10 areas with the most abundant tumour-infiltrating lymphocytes were selected, digitally photographed at a size of 0.0625 mm2, and counted manually. The count was performed three times for each photograph by the same investigator (E.S.) without knowledge of earlier results.
Statistical analysis. Statistical analyses were performed with the Fisher's exact test. Survival curves were plotted by means of Kaplan–Meier method and compared by using the log-rank test. A p-value (two-tailed) lower than 0.05 was considered significant.
Results
Patient characteristics. Forty-three patients were included. Patients' characteristics are listed in Table I. Briefly, the median age at diagnosis was 71 years. Most of the patients were affected by FIGO II (49%) or III (42%) stage tumours. Monolateral or bilateral inguinal lymphadenectomy for optimal surgical staging was performed in 93% (40/43) of cases.
CD133, ABCG2 and CD24 expression. Forty-three patients were evaluated for the expression of CD133, concurrently tumours from 31 women were also analyzed for the expression of ABCG2 and CD24.
In 43 patients examined, 11 samples were positive for the expression of CD133. This marker was weakly and moderately expressed in eight (18.6%) and three (7%) samples, respectively, while no case showed strong CD133 expression (Table II).
In the 31 patients analyzed, ABCG2+ and CD24+ expressing cells were found in 90.3% (28/31) and 54.8% (17/31) of patients, respectively. ABCG2 was weakly expressed in 64.5% of samples, while 26% of women had a moderate or strong expression of this marker. CD24 presented a weak expression in 12 women (39%), while four (13%) and one (3.2%) case showed moderate and strong expression, respectively. Three (9.6%) and 14 (45%) samples out of 31 were completely negative for the presence of ABCG2 and CD24, respectively. Figure 1 presents an example of the antigen distribution observed.
The expression of the three markers examined on 31 patients revealed that 16% (5/31) of the tumours expressed all three antigens, while only one sample (3.2%) was completely negative.
No samples were CD133+ABCG2−CD24+, while 3/31 patients (9.6%) were CD133+ABCG2+CD24−. Eleven out of 31 patients (35.5%) were CD133−ABCG2+CD24+. Results are summarized in Figure 2.
Clinicopathological parameters and CD133, ABCG2 and CD24 expression. Correlation between clinical parameters and tumour stem cell markers are reported in Table III. CD133+ tumours were more frequently encountered in patients with lymph nodal metastasis (p=0.03), larger tumour diameter (p=0.03) and younger age at diagnosis (p=0.001) respect to CD133− tumours. ABCG2 and CD24 were not significant correlated with commonly identified prognostic factors.
Tumour lymphocyte infiltration. All patients were also evaluated for the presence of CD3 and FOXP3 markers in tumour (Figure 3). The immunosuppressive tumour pattern was defined as tumour samples presenting ≥20% FOXP3+ T-lymphocytes among all infiltrating CD3+ T-cells and this cut-off of 20% represents the lowest quintile (Treg/CD3 >0.195). Figure 3A shows the expression of CD3 and FOXP3 in patients with a low and a high Treg/CD3 ratio. The analyses performed on all patients showed that 34 out of 43 woman had a Treg/CD3 ratio >0.195 corresponding to 79% of cases, while 9/43 patients had a low proportion of immunosuppressive infiltrating T-cells (Figure 3B).
In addition, the expression of CD133 was evaluated in those patients with high immunosuppressive tumour pattern. Among the CD133+ patients (11/43), 10 of these woman (91%) had a higher Treg/CD3 ratio and had a significant Treg infiltration compared to those with CD133- samples (91% CD133+ vs. 50% CD133-; p<0.05). ABCG2 and CD24 did not exhibit any significant correlation with tumour immunosuppression (data not shown).
Survival analyses. The median follow-up of patients was 87.5 months. Overall and progression-free survival were not statistically correlated to the expression of any single marker analyzed. Figure 4 shows the results obtained analysing survival by CD133 and FOXP3 markers.
Discussion
CSCs have been identified in different solid tumours, including, glioma, melanoma, and prostate, ovarian, lung, colorectal and pancreatic cancer (15-18). This cell population appears to have self-renewal capacity and differentiation potential and inside the tumour it can maintain the stem cell pool, thus sustaining the heterogeneous growth of tumour mass.
In addition, CSCs appear dormant or with slow-cycling states (19), increase resistance to chemo-/radiotherapy (20) and furthermore, CSCs induce angiogenesis and lymphangiogenesis (21, 22). Currently, the prognostic implications of CSC presence within tumours are still debated. While different recent studies support the theory that the presence of CSCs within the tumour mass correlates with poor clinical outcome, there are studies that show no such correlation (23, 24).
CSCs show phenotypic heterogeneity in the same type of cancer (25), but they can be isolated utilizing specific patterns of surface biomarkers, which allows their identification from non-tumorigenic cells. CSCs can be recognized through the expression of well-characterized cell surface markers, including CD molecules (CD133, CD44, CD24, CD166, etc.), ATP-binding cassette transporters (ABCG2, ABCB5), aldehyde dehydrogenase 1, epithelial cell adhesion molecule, C-X-C chemochine receptor type 4, and nestin (26).
CD133 is the most investigated CSC marker (27). A very recent meta-analysis carried out in 470 patients affected by different solid tumours treated with chemoradiotherapy assessed the prognostic impact of CD133 expression (28). The analysis pointed out that CD133 expression positively correlated with poorer progression-free and overall survival. In addition, we demonstrated that CD133+ CSCs isolated from endometrial cancer express the tumour-associated antigen mucin-1 (MUC1). MUC1 make these cells potentially susceptible to immune system attack and contributes to maintaining resistance to cisplatin and paclitaxel (29). Again, CD133 expression was found to be the only independent risk factor in 1,366 patients affected by ovarian cancer with central nervous system metastasis (30).
In the present study, we performed a preliminary evaluation of the immunological and clinical impact of CD133 expressing CSCs in patients with vulvar cancer. We have shown that CD133 expression is present in approximately a quarter of patients suffering from vulvar cancer and that over 90% of tumour samples expressing this marker exhibited a strong tumour infiltration of Tregs, indicating the presence of a significant immunosuppressive microenvironment. Furthermore, CSCs were more frequently encountered in patients with unfavourable clinical factors such as lymph nodal metastasis, larger tumour diameter and younger age at diagnosis, although no correlation was found with patient survival. To our knowledge, this study is the first experience, in the field of vulvar cancer setting, in which a link between CSCs and immune infiltrate has been found. Despite the limited sample size of 43 patients, a significant correlation between CD133 CSC marker and immunosuppression, evaluated by examining the percentage of Tregs expressing FOXP3+, was found. Several hypotheses have been drawn to define this complex interaction highlighting the mutual interaction between Tregs and CSCs. One of these suggests angiogenesis and VEGF levels in tumour microenvironment as an intermediate connection between these two cell populations. VEGF signaling is in fact involved in the regulation of CSC stemness and expansion and in the recruitment of Tregs in the tumour microenvironment (5, 6). Recently, several reports have demonstrated that the administration of a monoclonal antibody to VEGF (bevacizumab) as an anti-angiogenesis factor, significantly inhibited the proliferation of CSCs and concurrently the level of Tregs in the tumour microenvironment (31-32).
Tregs release VEGF in response to hypoxia and their elimination depletes VEGF production in ovarian cancer (33). In addition, tumour growth factor (TGF)-β expressed by Tregs is also involved in this process of increasing the level of VEGF. However, an increase of Tregs is also attributed to the expression of VEGF (32-33). Intriguingly, recent reports demonstrate that interleukin 17-producing Tregs drive cells to become cancer-initiating cells in colorectal cancer (34). Finally, Tregs indirectly interact with CSCs partially driving the differentiation of tumour-associated macrophages that release TGFβ (35). On the other hand, CSCs also impact on the recruitment, function and status of Tregs. Glioma-associated CSCs induce Tregs, inhibit T-cell proliferation and activation and trigger T-cell apoptosis by programmed death-ligand 1 and galectin-3 (36). CSCs also produce TGFβ that contribute to the induction of Tregs and release the chemokine (C-C motif) ligand 2, which is a potent chemoattractant for Tregs (37).
Vulvar cancer frequently occurs in extremely frail patients. Standard treatment includes radical vulvectomy and bilateral lymphadenectomy. Furthermore, patients with negative prognostic factors are treated with adjuvant radiotherapy. These combined therapeutic strategies frequently determine an impairment of patient mobility due to lymphoedema and nervous and vascular lower limb damage. In the era of personalized medicine, in which new cellular pathways have already been identified as key markers for new successful target therapies (38), increasing evidence is emerging regarding the identification of new biomarkers and their prognostic and predictive role in vulvar cancer (39). If confirmed on larger series, the direct association between CD133, the immunosuppressive pattern and tumour size and lymph node metastases could have several therapeutic implications that could include tailoring lymphadenectomy and introduce new immunological drugs.
Acknowledgements
This work was supported by Sapienza Ateneo 2014 (C26A14WW3A, C.N.), Sapienza Avvio alla Ricerca 2014 (C26N14AHKZ, I.G.Z.; C26N14X3H8, S.C.) and Sapienza Avvio alla Ricerca 2015 (C26N15M7K2, I.R.).
Footnotes
* These Authors contributed equally to this study.
This article is freely accessible online.
Conflicts of Interest
The Authors declare no conflict of interest in regard to this study.
- Received May 9, 2016.
- Revision received August 31, 2016.
- Accepted September 6, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved