The Pretreatment Tumor Infiltrating T Lymphocytes (CD8+, CD4+, FOXP3+) and Systemic Neutrophil-Lymphocytes Ratio in Definitively Treated Cervical Cancer Patients: The Correlation to Clinicopathological Factors and Survival

Purpose: To evaluate the prognostic potential of pre-treatment tumour infiltrating T lymphocytes (TILs) (CD8+, CD4+, FOXP3+) and systemic neutrophil to lymphocyte ratio (NLR) in predicting the outcome of definitively treated cervical cancer patients. Methods: Densities of pre-treatment TILs (CD8+, CD4+, FOXP3+) in cervical biopsies and NLR were correlated with clinicopathological parameters. The prognostic value of pre-treatment TILs and NLR for disease free survival (DFS) and overall survival (OS) were assessed using Log rank and Cox regression. Results: The final analysis included 28 who had radical hysterectomy while 20 had definitive concurrent chemoradiation. Elevated CD8+, CD8+/CD4+ and low FOXP3+ were associated with node negative, early stage disease and radical hysterectomies. Conversely, elevated NLR was associated with advanced stages, nodal involvement and definitive chemoradiation. Cox regression multivariate revealed that elevated NLR along with nodal involvement were independently correlated with poor prognosis with hazard ratio (HR 3.06 (95% confidence interval [CI], 3.45-9.24),5.63 (95% CI, 2.61-9.32) for OS and (HR 8.21 (95% CI, 4.21-16.53) and 5.32 (95% CI, 2.37-10.24) for DFS respectively. Additionally, FOXP3+ ≥ 19 and CD8+/CD4+ < 2 were significantly associated with decreased OS (HR 4.37 (95% CI, 2.48-12.37), 2.31 (95% CI, 2.34-9.32) and poorer DFS (HR 3.61 (95% CI, 1.389.32), 4.32 (95% CI,3.12-8.34) respectively. Conclusion: The Pre-treatment NLR, CD8+, FOXP3+ and C8/CD4+ showed a significant association with different clinicopathological prognostic factors in definitively treated cervical cancer patients. Additionally, they may be considered as potential independent prognostic indicators of clinical outcomes.


Introduction
Cervical cancer is the second most common female malignancy, accounting for 500,000 new cases and over 200,000 deaths worldwide annually [1,2]. Although many independent prognostic factors were devised in cervical cancer as tumor size, lymph node status, International Federation of Gynecology and Obstetrics (FIGO) staging criteria as well as pretreatment hemoglobin level, they are considered to be insufficient because of the inconsistent specificity and sensitivity among patients [3][4][5][6]. Consequently, more prognostic parameters are still warranted, to further increase predictive accuracy in cervical cancer patients.
An increasingly prominent role has recently been given to the immune system in cancer pathogenesis. There is a link between inflammation and outcome in a number of solid tumors. Accumulating evidence demonstrates that Neutrophil to lymphocyte ratio (NLR) has prognostic significance in patients with various types of cancers [7][8][9][10][11][12][13]. The peripheral (NLR), is a marker for evaluating the systemic potential balance between neutrophil-dependent pro-tumor inflammation and lymphocyte-associated anti-tumor immune response [14,15]. A higher level of NLR could represent a trend towards increased pro-tumor inflammation and decreased antitumor immune capacity.
Moreover, the immune response of the host against the tumor, defined as cancer "immunoediting" is reflected by the presence of tumor-infiltrating lymphocytes (TILs) either within the tumor stroma or tumor epithelium reflects [16]. Many TILs have been incorporated in immunoediting. Of those, fork head box P3-positive (Foxp3 +) regulatory T cells (Tregs), CD8 + T cells, and CD4 + T cells are known to be the main keys for immune surveillance and tolerance, respectively [17]. CD8 + T cells are mediators of antitumor immunity and can lyse tumor cells directly [18]. The clinical importance of CD8 + T cells has been suggested by many recent studies that reported a survival benefit in correlation with an increase in CD8 + T cells in large cohorts of various human cancer patients [19][20][21][22][23][24][25].
as priming tumor-specific cytotoxic T cells or macrophages that are involved in clearance of tumor cells [18]. In contrast, Tregs are known to have a very important role in escape of antitumor T-cell response in cancer cells, due to their ability to potently suppress immune reaction against tumors in vivo resulting in immune tolerance. These Treg cells can be identified by their expression of cluster of differentiation CD4, the interleukin (IL)-2 receptor alpha chain CD25, and the fork head family transcription factor FOXP3 [26][27][28]. FOXP3 is considered as a T-cell marker and is more specific than other markers, such as glucocorticoidinduced tumor necrosis factor receptor (GITR) and Cytotoxic T Lymphocyte Antigen-4 or the co-expression of CD4 (CTLA-4) and CD25 [29]. Transfection of FOXP3 gene will convert naïve CD4 + CD25 + T cells toward a regulatory phenotype in both mice and humans, and therefore, this molecule represents a functionally important marker of this Treg-cell population. Highly suppressive FoxP3 + Tregs expressing CTLA-4, GITR and T-cell immunoglobulin and mucin-domain containing-3 (TIM-3) were isolated from immune infiltrates of HCC, CRC, cervical and ovarian carcinomas [30][31][32]. These observations suggest that the FOXP3 is presently the most definitive marker of the Treg cells, and its detection could allow a better understanding of the mechanisms through which Treg cells are involved in autoimmune and infectious diseases, as well as transplantation and tumor immunity. Moreover, high density of infiltrating regulatory T cells (CD4 + CD25 + FOXP3 +) , which can suppress other effector T cell activities, is associated with adverse prognosis in some cancer types [33,34]. However, it has been a challenge to delineate the role that each TIL type plays as an independent prognostic and/or predictive biomarker as their clinical relevance in cervical cancer remains unclear.
Consequently, our study rationale was to evaluate the correlation of the pretreatment NLR and TILs (CD8 +, CD4 + and FOXP3 +) to clinicopathological parameters in cervical cancer treated definitively in addition to their potential in predicting the outcome in this group of patients.

Material and Methods
Retrospective review of cervical cancer patients treated or referred to Radiation Oncology Department King Fahad Specialist Hospital Al Dammam and Clinical Oncology department Alexandria University between January 2013 and December 2016 after obtaining Institutional Board Approval (IRB) approval. All patients signed informed consent. Medical records were reviewed to select non-metastatic cervical cancer patients who had cervical biopsies prior to surgery or definitive radiation therapy. The prognostic variables including: age, histology, grade, stage, response to definitive chemoradiation, in addition to the tested pretreatment neutrophil count to lymphocyte count (NLR), was calculated from peripheral blood cell count which were calculated from peripheral blood count.
All cases were histopathologically classified according to the International Federation of Obstetrics and Gynecology (FIGO) criteria. The tissue samples were fixed in formalin and embedded in paraffin.

Inclusion criteria
• Non-metastatic, localized, histologically proven squamous cell or adenocarcinoma of uterine cervix.
• Clinically FIGO Stages IA1, IA2, IB1and IIA1 who had radical hysterectomy and pelvic lymph nodes dissection. Adjuvant radiation ± chemotherapy was delivered based on risk factors: A. Post-operative radiotherapy after radical hysterectomy for patients who had at least 2 of the following risk factors:

Exclusion criteria
1) Patients who had evidence of distant metastasis during initial work up.
2) Patients who develop distant metastasis during the course of treatment.
3) Patients with small cell or clear cell carcinoma of the cervix.

Immunohistochemistry
Immunohistochemistry staining was conducted on 4-to 5-mm paraffin sections using horseradish peroxidase from the Envision Kit or the Catalyzed Signal Amplification Kit (Dako Corporation,Copenhagen, Denmark) for single staining, and alkaline phosphatase was used for double staining, as recommended by the manufacturer.
The primary antibodies were mouse monoclonal anti-human anti-CD8 (DakoCytomation, Glostrup, Denmark; 1: 100 dilution), anti-CD4 (Novocastra; 1: 30), anti-FOXP3 (Abcam, Cambridge, UK; 1: 50). Briefly, the sections were dewaxed in xylene and rehydrated using graded concentrations of ethanol and distilled water. Endogenous peroxidase activity was blocked by submersion of the sections in a 0.5% H 2 O 2 / methanol solution for 10 min at room temperature (RT). Antigen was retrieved by heating the sections under high temperature and pressure in a stainless-steel pressure cooker for 1 min and 30 s in unmasking solution (0.01 M citrate buffer, pH 6.0). After cooling the cooker to RT in running tap water, the sections were placed in TBS-Tween buffer for 5 min, blocked with 3% goat serum for 10 min at 37°C, and incubated with primary antibody (at RT: CD8, 1 h; and CD4 overnight at 4°C). Tonsil tissue was used as a positive control for all the antibodies. Immunohistochemical staining for CD4/CD8 was considered positive when lymphocytes showed evident membranous immunoreactions. Immunohistochemical staining for FOXP3 was considered positive when lymphocytes showed evident nuclear immunostaining. The numbers of labeled TILs were counted by the classical counting method with the light microscope using an ocular grid. adjacent stroma (within the same high-powered field; HPF) was counted manually in 3 HPFs (400x) then the mean of each marker positive cells was calculated. Counting was performed twice for each slide.

Patient follow-up
All patients were followed up regularly until December 2016 or until death (every 3 months for the first 2 years and then every 6 months up to 5 th year). Physical examination, laboratory tests and imaging were conducted. The follow up periods varied from 3 to 48 months, with a median of 26 months. Overall survival (OS) was calculated from diagnosis to death. For drop-out patients, the date of the last follow-up was applied. Disease-free survival (DFS) was calculated from surgery to disease relapse or until the date of last follow-up.

Statistical analysis
The correlations between subsets of TILs (CD8 +, CD4 +, FOXP3 + Tregs, and CD8 +/ CD4 + ratio), NLR and other clinicopathological characteristics listed before were analyzed by Spearman test. The impact of the pretreatment CD8 +, CD4 + and FOXP3 + tumor infiltrating T cells and the pretreatment NLR on different clinicopathological parameters in cervical cancer treated either with upfront surgery (radical hysterectomy) or definitively with concurrent chemoradiation were evaluated by Mann-Whitney U test (between 2 groups) or Kruskal-Wallis test (≥3 groups). Receiver operator characteristic (ROC) curves were used to identify pretreatment CD8 +, CD4 + and FOXP3 + tumor

Results
The final analysis included 48 cervical cancer patients, 20 (41.7%) patients had definitive concurrent chemoradiation and 28 (58.3%) had radical hysterectomy that was followed by adjuvant radiation in 12 (25%) or chemoradiation in 16 (33.3%) based on postoperative risk factors ( Table 1). The patients were aged from 35 to 72, while the median age was 52 years. With regards to histological type, 38 (79.2%) patients had squamous cell carcinoma, whereas 10 (20.8%) patients had adenocarcinoma of the uterine cervix.
The median value of baseline inflammatory response biomarker NLR was 1.95. Additionally, the mean scorings of the number of CD8 +, CD4 + and FOXP3 + CD4 positive T lymphocytes present within the malignant epithelium and just adjacent stroma (within the same high-powered field; HPF) were 58, 30 and 18 Cells/HPF respectively. Moreover, CD8 +/ CD4 + TILs ratio was 1.9. Correspondingly, all other patients' baseline characteristics are presented in Table 1.

The optimal thresholds for NLR and Tumor Infiltrating Lymphocytes (TILS) (CD8 +, CD4 + and FOXP3 + CD4 + T lymphocytes)
The receiver operating characteristic (ROC) curve, used DFS as the end-point for NLR and tumor infiltrating lymphocytes (CD8 +, CD4 +, FOXP3 + CD4 and CD8 +/ CD4 + TILs ratio) (Figures 1 and 2). The receiver operating curve (ROC) demonstrated a baseline NLR of 2 cut off value area under the curve (AUC: 0.873) for predicting DFS with a sensitivity 97.4% and specificity of 82.9%. Additionally, CD8 + cut off 64 (AUC: 0.794) yielded a sensitivity of 93.3% and specificity of 84.3% in predicting the DFS. Moreover, the ROC curve illustrated the ability of baseline FOXP3 + and CD8 +/ CD4 + TILS ratio to predict DFS (AUC were 0.854 and 0.859 and the best cut-off value were 19, 2 respectively). The base line FOXP3 + cutoff value 19 yielded a sensitivity of 94% and specificity of 84.5% in predicting DFS respectively. Moreover, CD8 +/ CD4 + TILS ratio cut off 2 values resulted in a sensitivity of 92.3% and specificity of 81.5% in predicting DFS respectively. Moreover, the baseline CD4 + cutoff value of 32 (AUC of 0.721) yielded the lowest sensitivity of 85.1% and specificity 75.3% in predicting DFS. Patients were subsequently divided into two groups according to the optimal cut-off levels, with the high group ≥ the optimal cut-off levels and the low group that < the optimal cutoff levels. Our results revealed that patients with high tumor grade, advanced stage III-IV, lymph node involvement and definitive concurrent chemoradiation had a higher NLR than the others (Table 2). Additionally, higher median values of pretreatment CD8 + and CD8 +/ CD4 + were associated with node negative disease, earlier disease stages and radical hysterectomies. While CD4 + T lymphocytes were not associated with any of the clinicopathological parameters (Table 2). On the hand, low pretreatment FOXP3 + T lymphocytes were associated with node negative patients, early stage disease and radical hysterectomies ( Table 2). It is worth mentioning that on multiple regression analysis pretreatment NLR is found to be significantly related to pretreatment FOXP3 + and CD8 +/ CD4 + ratio, as low median levels of NLR were associated with low levels of FOXP3 + and high levels CD8 +/ CD4 + ratio (OR2.4, 95% CI 1.6-6.8, OR 2.1, 95% CI 1.1-7.8 respectively) ( Table 3).

Clinical factors associated with path CR with definitive concurrent chemoradiation
Twenty patients with locally advanced disease were offered   (Table 3) and any pathologic response. Moreover, high cut off levels of pretreatment CD8 + (P=0.002) and CD8 +/ CD4 + ratio (0.002) while low cut off levels of FOXP3 + (P=0.003) were found to be significantly associated with path CR (Table 4) and any pathologic response. While pretreatment CD4 + was not associated with any pathological response (Table 4).

Clinicopathological parameters of patients treated with radical hysterectomy with or without adjuvant radiation or concurrent chemoradiation
The clinicopathological parameters of the twenty-eight patients treated with radical hysterectomy were displayed in (Table 5). On multiple regression analysis, lymph nodal involvement was found to be the single parameter significantly associated with high levels of pretreatment NLR (P=0.002), FOXP3 + (P=0.001) and low levels of pretreatment CD8 + (P=0.01), CD8 +/ CD4 + ratio (P=0.02) respectively. While CD4 + was not associated with any clinicopathological parameters.

The association between baseline characteristics and clinical prognosis
The median follow-up period was 26 months. During the follow-up period, 13 (27.1%) patients developed either local recurrence or distant metastasis. Among them, 10 (20.8%) patients were dead from cancer related disease. The median of DFS and OS were not reached. Moreover, the 4-year OS and DFS were 81.2% and 73.7% respectively (Figures  3 and 4). To evaluate the association of baseline characteristics with clinical prognosis, Kaplan-Meier survival analysis and log-rank tests were conducted. Our results confirmed that advanced disease stage, nodal involvement, cervical adenocarcinoma, poorly differentiated tumors, definitive concurrent chemoradiation and partial response to chemoradiation were significantly associated with decreased both OS and DFS (Table 6). Moreover, patients who had low pretreatment NLR < 2 (Figures 5 and 6)

Discussion
Accumulating evidence demonstrates that in cervical carcinoma, tumor infiltrating lymphocytes (TILs) have been associated with an  improved clinical outcome [35]. In the present study, we did a detailed analysis of the number and functional status of the pretreatment intraepithelial TILs (CD8 +, CD4 + and FOXP3 +) in relation to prognosis of 48 patients with cervical cancer in FIGO stages IA-IVA treated definitively either with radical hysterectomies ± adjuvant chemo/ radiation or definitive concurrent chemoradiation.
When we consider the 48 studied cancer cervix patients, our results demonstrated that higher levels of pretreatment CD8 + and CD8 +/ CD4 + ratio were associated with node negative disease (P=0.0126, 0.0132), earlier disease stages (P=0.0161, 0.0167) and radical hysterectomies (P=0.003, 0.003) respectively. While CD4 + T lymphocytes were not associated with any of the clinicopathological parameters. Our results are concordant with Piersma who confirmed that evaluation of cervical cancer patients with respect to their lymph node status revealed that LN negative patients, who in general had a good prognosis, displayed a significantly higher number of intraepithelial CD8 + T cells (P < 0.01), a higher CD8 +/ CD4 + T-cell ratio (P=0.01), and a higher CD8 +/ regulatory T-cell ratio than the group of patients with lymph node metastases [25]. On the other hand, low pretreatment FOXP3 + T lymphocytes were associated is node negative patients (P=0.0112), early stage disease (P=0.011) and radical hysterectomies (P=0.002) respectively. Similarly, Wu reported that cervical cancer patients with lymph node metastasis had significantly higher levels of FOXP3 + T cells than in patients without lymph node metastasis (P=0.045) [36]. Shah also reported that FOXP3 + was significantly higher in the advanced FIGO stage IIIB group when     compared to earlier stages (P=0.023) [37]. Our results also revealed that patients with high tumor grade (P=0.0311), advanced stage III-IV (P=0.0145), lymph node involvement (P=0.0124) and definitive concurrent chemoradiation (P=0.0132) had a higher NLR respectively.
Similarly, Huang reported in their meta-analysis that increased NLR was significantly correlated with tumor size (OR 2.05, 95% CI 1.14-3.65), advanced FIGO stage (OR 2.12, 95% CI 1.28-3.49) and lymph node involvement (OR 2.24, 95% CI 1.65-3.04) [38]. Moreover, on multiple regression analysis pretreatment NLR was found to be significantly related to pretreatment FOXP3 + and CD8 +/ CD4 + ratio, as low median levels of NLR were associated with low levels of FOXP3 + and high levels CD8 +/ CD4 + ratio (OR 2.4, 95% CI 1.6-6.8, OR 2.1, 95% CI 1.1-7.8 respectively). It is worth mentioning, that our study is the first to illustrate the correlation between the pretreatment inflammatory response biomarker NLR and pretreatment TILs (CD8 +, CD4 + and FOXP3 +) in tumor milieu. Furthermore, the (ROC) demonstrated the ability of a baseline NLR and TILs (CD8 +, CD4 +, FOXP3 + and CD8 +/     The 48 cervical cancer patients were further classified based on the definitive treatment delivered. Twenty patients with locally advanced disease were offered definitive concurrent chemoradiation. Many clinical factors were found to be associated with pathological response as tumor grade, stage and lymph nodal involvement. Moreover, the high cut off levels of pretreatment CD8 + (P=0.002) and CD8 +/ CD4 + Ratio (P=0.002) while the low cut off levels of FOXP3 + (P=0.003) were found to be significantly associated with path CR and any pathologic response whereas pretreatment CD4 + was not associated with any pathological response. Draghiciu also reported that in locally advanced cervical cancer treated with definitive concurrent chemoradiation, high CD8 + staining was less frequently observed in patients that had a recurrence of disease (OR=0.562; 95% CI=0.319-0.991; P=0.046) [39]. The remaining 28 patients with earlier disease stage were treated with radical hysterectomy followed by adjuvant radiation ± chemotherapy based on postoperative pathological risk factors. Notably, in surgically treated patients, lymph nodal involvement was found on regression analysis to be the single parameter significantly associated with high levels of pretreatment NLR (P=0.002), FOXP3 + (P=0.001) and low levels of pretreatment CD8 + (P=0.01), CD8 +/ CD4 + ratio (P=0.02), respectively. Interestingly, the present study scrutinized on the correlation of pretreatment NLR, TILs (CD8 +, CD4 +, FOXP3 + and CD8 +/ CD4 +) and other clinicopatholgical parameters in patients treated definitively with either radical hysterectomies or definitive concurrent chemoradiation, a concept which was not tested in contemporary studies.
Furthermore, Jordanova revealed that high number of intraepithelial Treg (FoxP3 +) and a low CD8 +/ regulatory T-cell ratio was associated with worse survival (P=0.034 and 0.025) respectively. On subsequent Cox multivariate analysis, low CD8 +/ Treg ratio (P=0.047), low CD8 +/ Treg ratio (P=0.002) were found to be independent unfavorable prognostic predictors in cervical carcinoma [40]. Similarly, Piersma found a significantly stronger CD8 + T-cell tumor infiltration, a higher CD8 +/ CD4 + T-cell ratio, and a higher CD8 +/ Tregs ratio in patients with tumors that failed to metastasize to the tumor-draining lymph node, therefore correlated indirectly with good prognosis [25]. With regards to the impact of FOXP3 + Treg cells on prognosis, Shah stated a significantly lower survival rate in patients with high levels of CD4 + FOXP3 + Tregs as compared with those with low levels of these cells (35.3% versus 88.9%, P=0.001) [37].
The current study has several strengths 1. To the best of our knowledge, it is considered to be the first study to illustrate the association of the pretreatment NLR to TILs (CD8 +, FOXP3 + CD8 +/ CD4 +) in the tumor milieu of cervical cancer patients.

2.
It delineated the role that each TIL type plays as an independent prognostic and/or predictive biomarker in definitively treated cervical cancer patients.