MPO density in primary cancer biopsies of ovarian carcinoma enhances the indicative value of IL-17 for chemosensitivity

Cancer of the ovary is mostly discovered at a late stage and cannot be removed by surgery alone. Therefore surgery is usually followed by adjuvant chemotherapy. However, few reliable biomarkers exist to predict response to chemotherapy of ovarian cancer. Previously, we could demonstrate that IL-17 density is indicative for chemosensitivity. This study focuses on the predictive value of myeloperoxidase (MPO) concerning response to chemotherapy of ovarian cancer. Biopsies of mostly high-grade primary serous ovarian carcinomas and their matched recurrences were stained with MPO after fixation in formalin and embedding in paraffin. For this staining the technique of tissue-microarray was used. Recurrence within 6 months of the completion of platinum-based chemotherapy was defined as chemoresistance as previously publised. Data for MPO could be analyzed in 92 biopsies. MPO and IL-17 positive immune cells correlated significantly in biopsies of primary and recurrent carcinomas (rs = 0.41; p = 0.004 and rs = 0.40; p = 0.007, respectively). MPO expression alone did not predict response to chemotherapy, but in multivariate cox regression analysis including age, residual disease, number of chemotherapy cycles, FIGO classification and combined categorized MPO and IL-17 cell densities of primary cancer biopsies, the combination of both immune markers was an independent prognostic factor for recurrence-free survival (p = 0.013, HR = .23, 95CI = 0.07–0.73). There was no chemoresistant patient in the subgroup of MPO + IL-17+, neither in primary nor in recurrent cancer biopsies. High MPO positive cell density enhances the indicative value of IL-17 for response to chemotherapy in ovarian carcinoma. Although, these results have to be validated in a larger cohort.


Background
Ovarian cancer has an incidence range of 5-15/100'000 in Europe [1][2][3]. It is only the 5 th most common female cancer, but even though it is the most lethal of all female genital carcinomas. It is mostly discovered at a late stage and cannot be removed by surgery alone due to late and unspecific symptoms. Surgical debulking is usually followed by adjuvant platinum-based chemotherapy. However, most patients recur with chemoresistant disease.
It is known that tumor microenvironment influences tumor biology and that tumor behavior is affected by the immunological environment. According to several previous publications tumor microenvironment seems to have significant impact on survival and tumor growth [4][5][6][7][8]. Tumor-infiltrating lymphocytes (TILs) are frequently thought to mirror tumor immune response to invasive neoplasms [9]. They were discovered in different solid tumors [10,11]. Indeed, in colorectal cancer (CRC) the "immune contexture" [12], especially cells of the adaptive immune response, are more important concerning prognosis than the TNM staging and might help in decision making for personalized treatment [13,14]. There are several previous studies that analyzed different markers predicting response to platinum-based chemotherapy in ovarian cancer with the scope to optimize adjuvant treatment [15][16][17]. However, only few markers were helpful.
On one side it is known that FOXP3 positive regulatory T (Treg) cell infiltration is associated with decreased survival in ovarian cancer [17][18][19]. On the other side, granulocytes have largely been neglected by tumor immunologists [20]. Challenging this, recent studies implied that neutrophil granulocytes might play an important role in the prevention of cancer metastasis [21]. Finally, neutrophil granulocytes are thought to have the capacity to undergo differentiation into N1 and N2 cells with anti-and protumor properties, respectively [22,23]. Therefore tumor infiltrating granulocytes regain attention in research [24][25][26].
In a previous study we could demonstrate that IL-17, but not FOXP3 positive immune cell infiltration in primary and recurrent ovarian carcinoma were indicative of chemosensitivity [27]. Finally it has been shown that IL-17 can be produced by granulocytes [28,29] and other innate immune cells [30] and not only by TILs. However, in ovarian cancer the role of the innate immune system has not been evaluated to the same extent as the role of the adaptive immune system. In early stage lung cancer granulocytes have recently been shown to stimulate T cell responses in humans [31].
Neutrophilic granulocytes (NG) accumulate myeloperoxidase (MPO) in high amounts during their early maturation phase [32]. MPO produces hypochlorous acid from hydrogen peroxide and chloride anion and is responsible for the oxidization of tyrosine to tyrosyl radicals. Both are cytotoxic to a variety of microorganisms. After activation of granulocytes this enzyme is also implicated in the induction of apoptosis [33,34].
There are few studies reporting the prognostic and predictive role of MPO in ovarian cancer. Therefore, we investigated its predictive value for chemosensitivity alone and in combination with IL-17 expression in a well characterized cohort of primary ovarian carcinomas and their matched recurrences which has also been used for previous publications of our group [35][36][37][38].

Patients
Tissues from ovarian serous carcinomas and their recurrences were available at the Pathology Biobank at Pathology of the University Hospital of Basel and the Cantonal Hospitals of Baden, Liestal and St. Gallen, Switzerland.
Mostly high-grade ovarian carcinomas (5.7 % FIGO stage II, 84.3 % FIGO stage III and 4.3 % FIGO stage IV) were included in this study after typing according to previous publications [39,40]. The tissue microarray (TMA) was available from previous studies [35][36][37][38]. All patients had recurrences after initial surgery and had received at least three cycles of platinum-based adjuvant chemotherapy. The collection was divided into two groups according to response to chemotherapy. Recurrence occurring within 6 months after completion of platinum-based chemotherapy was defined as resistance [41]. The TMA allows investigation of tissues from ovarian carcinomas and matched recurrences from the same patients as previously shown [35][36][37][38]. The statement concerning the clinical data collection and ethical considerations can be found in previous publications [27,[35][36][37][38].

Tissue microarray construction
The construction of the tissue microarray has been previously described [27,42].

Immunohistochemistry (IHC) and visual analysis
Standard indirect immunoperoxidase procedures (ABC-Elite, Vectra Laboratories) were used for immunohistochemistry. For MPO staining the following antibody was used: clone 59A5 Novocastra, Newcastle, UK. In each tissue spot positive stained tumor immune cell infiltration (TICI) in the stroma was counted, representing approximately one high-power-field (10×), intravascular cells were excluded from analysis ( Fig. 1a and b). Two independent experienced observers (RM and GS) analyzed the staining for specificity and the amount of TICI as described above. Cut-off was 22 positive cells/punch for MPO. Conclusive data for MPO were available in 47 biopsies of primary and 45 biopsies of matched recurrent carcinomas, respectively.

Statistical analysis section
Cut-off scores used to classify ovarian carcinomas with low or high MPO infiltration were obtained by regression tree analysis, evaluating the best threshold in order to predict patients' survival status, on all tumor samples [43]. Specific scores were set at 22 positive cells/punch. IL-17 data were available from a previous publication [27]. Kruskal Wallis, Chi-Square or Fisher's Exact tests were used for the association of the clinicopathological features with the corresponding four groups of the biomarkers. Univariate recurrence-free and overall survival analysis was carried out by the Kaplan-Meier method and log rank test.
The assumption of proportional hazards was verified for both markers by analyzing the correlation of Schoenfeld residuals and the ranks of individual failure times. Any missing clinicopathological information was assumed to be missing at random. Subsequently, a multivariate Cox regression analysis was performed including MPO and IL-17. The hazard ratios (HR) and the 95 % confidence intervals (CI) were used to determine prognostic effects on survival time. Spearman's rank correlation was used to analyze the correlation between MPO and IL-17. All statistical analyses were made using STATA software version 13 (StataCorp, College Station, TX, USA).

Patient characteristics
The baseline characteristics of the patient cohort have been described previously [27] (Table 1). Briefly recurrence-free (RFS) and overall survival (OS) in the chemoresistant      Table 2.

MPO positive immune cell infiltration in paired primary and recurrent ovarian carcinoma
Mean number of infiltrating MPO positive cells in primary and recurrent cancer biopsies were 16.6 (±21.6) and 19.0 (±34.8), respectively. Neither for dichotomized MPO density in primary, nor in recurrent cancer biopsies a significant association with any clinicopathological feature was found ( Table 2). Twelve out of 47 and 10 out of 45 displayed a high MPO cell density in primary and recurrent cancer biopsies, respectively. Finally, MPO density in primary and recurrent cancer biopsies did not show any significant association with chemosensitivity (p = 0.249 and p = 0.135) or any other clinicopathological feature ( Table 2).

Combined analysis of MPO and IL-17 positive cell density
Based on the correlation analysis results, a combined analysis of MPO and IL-17 cell density was performed. As shown in Fig. 2a, MPO and IL-17 positive TICI frequency was significantly associated with a longer RFS in biopsies of primary cancers (n = 47, p = 0.011), although the combined marker analysis did not show significant association with OS (Fig. 2b, p = 0.283). In addition, there was a significant association with chemosensitivity (p = 0.004) and FIGO classification (p = 0.029) in primary cancer biopsies (Tables 3 and 4).
In biopsies from carcinoma recurrences, no significant association with RFS could be found (p = 0.121).

Multivariate analysis of synergistic effect
In a multivariate cox regression analysis including age, residual disease, FIGO classification, number of chemotherapy cycles and categorized MPO and IL-17 cell Table 4 Dichotomized distribution of MPO and IL-17 according to defined cut-offs (22 cells/punch for MPO and 1 cell/punch for IL-17 [27]   density, the combination of the immune markers was an independent prognostic factor for RFS in primary cancer biopsies (Table 5).

Discussion
As in ovarian carcinoma, surgical tumor debulking is often followed by adjuvant platinum-based chemotherapy it would be helpful to find predictive markers for chemoresponse. Based on such biomarkers it would be possible to carry out extended chemotherapy regimen and repetitive surgical procedures. In our previous studies we identified different potential therapeutical targets and IL-17 as predictive marker for chemosensitivity [27,[35][36][37][38]. As previously mentioned there is a resurgent interest in the scientific community concerning the role of granulocytes in tumor immunology [22,24,31]. Indeed, they have usually been associated with poor prognosis [20]. But experimental models in the past have proposed an anti-tumor role through the activation of T cells [44,45]. Furthermore recently, granulocyte polarization has been described [22] and the ability of granulocytes to promote lymphocyte activation in the tumor environment has been reported [31]. The composition of tumor microenvironment has been shown to significantly impact tumor progression and clinical outcome [12]. Similar to a variety of cancers of different origins, T lymphocyte infiltration in ovarian cancer has a positive prognostic role [46]. Most interestingly, ovarian cancer infiltration by IL-17 producing cells has been found to be associated with either longer RFS or good prognosis by us and others [27,47]. However, there is no evidence concerning the potential clinical relevance of granulocyte infiltration in ovarian cancer.
In our previous study we found that IL-17 positive TICI were significantly more frequent in the chemosensitive ovarian carcinoma group [27]. Granulocytes and other immune cells of the innate immune system can produce IL-17 [28][29][30]. On the other hand, TNF-alpha in the tumor microenvironment could attract myeloid cells in an IL-17-dependent manner and contribute to tumor-promotion [48]. Therefore we investigated the predictive role of MPO, an enzyme that is expressed in myeloid cells, alone and together with IL-17. Finally we found that the combination of both markers was an independent prognostic factor for RFS. However, in this study IL-17 was the dominant marker for prediction of RFS. A further limitation is the small sample size. Therefore, our results have to be validated in an independent larger patient cohort.

Conclusion
Based on the results in this study, we conclude that the combination of high MPO positive cell density and IL-17 expression enhances the indicative value for the response of ovarian carcinomas to chemotherapy, as it in addition has prognostic value regarding recurrence-free survival in ovarian carcinoma. Although, these results have to be validated in a larger cohort.