Abstract
Background: High-grade epithelial ovarian cancer (HGEOC) is a heterogeneous disease and among the deadliest types of cancer. It often acquires resistance to conventional chemotherapy and its prognosis remains highly poor. The tissue protein nestin, implicated in the assembly and disassembly of intermediate filaments, has been reported to be an unfavourable prognostic factor in several cancer types. We hypothesized that HGEOC progression is regulated by the proliferation of chemoresistant cancer stem cells, in which nestin might be implicated. This preliminary study aimed to evaluate nestin as a prognostic biomarker in HGEOC treated by neoadjuvant chemotherapy (NACT) followed by cytoreductive surgery. Patients and Methods: A retrospective study (2009-2019) was conducted on 92 patients with primary ovarian, fallopian tube or peritoneal HGEOC who underwent NACT followed by cytoreductive surgery. Nestin expression in tissue samples was semi-quantitatively evaluated defining nestin positivity for those with histochemical score ≥30. We then evaluated the prognostic value of nestin expression. Results: The median progression-free survival was similar between nestin-positive (22 months) and nestin-negative (19 months) groups (p=0.57). Interestingly, the median overall survival was shorter for the nestin-positive group (48 vs. 67 months, respectively), however the difference did not reach statistical significance (p=0.43). Conclusion: Tissue nestin expression does not appear to be a relevant prognostic biomarker in HGEOC treated by NACT. However, we believe that prospective studies in larger cohorts should be conducted and evaluation of nestin in pre-NACT HGEOC samples needs to be explored.
Ovarian cancer remains the most lethal gynaecological cancer and the seventh most commonly diagnosed cancer of females in the world (1). Worldwide, 295,000 women are diagnosed each year with ovarian cancer and 184,800 will die from the disease (2). The number of new cases of malignant ovarian tumour in France was estimated at 5,193 in 2018 and the median age at diagnosis was 68 years (3). Epithelial ovarian cancer (EOC), also known as ovarian carcinoma, represents 95% of ovarian malignant tumours and is a highly heterogeneous disease. The revised histomolecular model of EOC distinguishes two tumour types, with different origins and histological and genomic features, as well as clinical behaviour. Type I tumours (30% of all EOCs) have a generally low proliferative index and a better prognosis, with the exception of clear-cell carcinoma, considered a high-grade neoplasm. They are divided into three groups: i) Endometriosis-related tumours that include endometrioid, clear-cell and seromucinous carcinomas; ii) low-grade serous carcinomas; and iii) mucinous carcinomas and malignant Brenner tumours. Type II tumours (70% of all EOCs) develop rapidly, have a high proliferative index, and respond relatively well to chemotherapy but have a high relapse rate and a poor prognosis. They account for more than 90% of ovarian cancer-related deaths and mostly consist of high-grade serous carcinomas (4, 5) .
Due to the late appearance of clinical symptoms, the great majority of EOCs (especially type II) are diagnosed at an advanced stage, with up to 85% of patients having locally advanced or distant metastatic disease [International Federation of Obstetrics and Gynecology (FIGO) stages III and IV] (6). Patients with late-stage disease have a 5-year relative survival rate of only 29%, in contrast to 92% for those with early-stage disease (7).
Recent discoveries in ovarian tumour molecular biology and histopathology have shown that ovarian cancer is a general term for several distinct diseases that only share an anatomical location. For example, fallopian tube malignancies were considered rare among ovarian neoplasms. However, histological, molecular and genetic evidence has shown that approximately 80% of tumours that were classified as highgrade serous carcinomas of the ovary or peritoneum may have originated from the fimbrial end of the fallopian tube and that serous tubal intraepithelial carcinoma represents the precursor of these neoplasms (8).
Serous borderline ovarian tumours, precursors of low-grade serous carcinomas, have a high frequency of activating mutations in B-Raf Proto-Oncogene serine/threonine kinase and Kirsten rat sarcoma virus genes, whereas clear-cell carcinomas have a high frequency of activating phosphatidylinositol 3-kinase mutations. Immunological markers and genomic studies have shown that many tumours previously designated as highgrade endometrioid carcinomas should be classified as serous carcinomas [reviewed in (9)].
From a clinical perspective, inter-individual and intra-tumoural heterogeneity (10) suggests that treatment of type II EOC should not only be based on the disease stage. Detection of specific biomarkers is crucial for therapeutic decision-making. A well-known example is the presence of a deleterious BRCA DNA repair-associated gene (BRCA) mutation which confers treatment efficacy based solely on that information (11). It is now well established that highly penetrant mutations of BRCA1 and BRCA2 genes increase the lifetime risk for EOC. EOCs carrying BRCA1/2 mutated genes account for the majority of hereditary EOC cases and 10-15% of all EOC (12-14).
Discovery of molecular pathways crucial in ovarian cancer development has led to approval of the first targeted therapies and their companion molecular tests for this type of cancer. The development of poly(ADP-ribose) polymerase (PARP) inhibitors and angiogenesis inhibitors has shifted the unique treatment approach to EOC towards a stratification based on molecular drivers. After more than 20 years of stable poor survival rates (15), these novel treatments have significantly improved prognosis for patients with EOC (11, 16, 17).
Besides mutated BRCA1/2 genes, more than 33% (18) of EOCs exhibit defective DNA-repair mechanisms via homologous recombination due to alterations of other homologous recombination pathway genes (19). myChoice® test (20, 21) (Myriad Genetics, Salt Lake City, UT, USA) is a test which detects homologous recombination deficiency in multiple tumour types [ovarian (17), breast (22), prostatic (23), and pancreatic (24)] and has been approved in the United States as being predictive for EOC maintenance therapy by PARP inhibitors.
While the standard therapeutic strategy in advanced EOC stages (25) is based on surgical staging and complete or optimal [residual disease ≤1 cm (26)] cytoreductive surgery preceded or followed by a platinum- and taxane-based chemotherapy, most patients will experience relapse and the response rate diminishes with each recurrence.
Complete resection of the entire macroscopic disease has been shown to be the most important independent prognostic factor in advanced high-grade disease (HGEOC) (27).
If complete resection of the entire macroscopic disease can be obtained with acceptable operative morbidity, primary cytoreductive surgery followed by carboplatin/paclitaxel-based adjuvant therapy is the standard of care for EOC (28). Non-candidate patients for primary cytoreductive surgery should be offered neoadjuvant chemotherapy (NACT), based on European Society of Gynaecological Oncology 2017 recommendations (25). However, in spite of major advances in the understanding of HGEOC at the molecular, clinical and surgical levels, biomarkers to predict disease recurrence are still lacking. Cytotoxic chemotherapy regimens are in absolute need of complementary treatment. Even though anti-PARP agents reduce the risk of disease progression and death, only up to 50% of patients with HGEOC (10) are eligible for their use. Lessons from the discovery of markers for lung cancer, such as epidermal growth factor receptor (29), ALK receptor tyrosine kinase (30) and their respective treatments, have taught us that identifying and targeting new specific ovarian cancer biomarkers may be the way to drastically improve patients, survival.
Nestin is a protein belonging to class VI intermediate filaments, first identified in the central nervous system as a component of neuroepithelial stem cells which generate neurons and glia (31, 32). The name ‘nestin, is derived from its location: Neural stem cells. Nestin is mainly expressed by immature cells (33); upon cell differentiation, it is down-regulated and replaced by the tissue-specific intermediate filaments (34, 35). In adults, expression of nestin is induced in regeneration and development of tissues such as muscle (36), intestine (37), endothelium (38), and bone marrow (39), when activation of the stem cell reserve is required (40-42). In malignant tumours, nestin expression has been described in cancer stem cells and in newly generated blood vessels. Isolated nestin-positive glioma stem cells exhibited increased sphere-forming capacity and size (43), indicating involvement of nestin in glioma development and progression. Studies in glioma (44) and prostate cancer (45) have demonstrated nestin expression in proliferating intra-tumoural blood vessels, indicating the role of nestin in (neo)angiogenesis (46-48). Although the mechanisms of action of nestin in cancer development require further investigation, recent biological studies (49, 50) suggest that nestin might play a major role in the regulation of cytoskeleton homeostasis. Cyclin-dependent kinase 5 phosphorylates numerous nestin sites and in that way regulates important aspects of cell differentiation (51). In neuronal cells, hyperactivation of cyclin-dependent kinase 5 in oxidative stress situations, such as ischaemia, is associated with proteasome-mediated degradation of nestin followed by apoptosis (52). The existence of a cancer stem cell population in ovarian cancer was confirmed by experimental findings showing that a small and distinguishable group of quiescent malignant cells, capable of organizing themselves into spheroid colonies, were able to generate ovarian cancer when transferred into immune-deficient mice (53). Nestin has been reported to be a negative prognostic factor in several malignancies. In breast cancer, Piras et al. found nestin to be predominantly expressed by aggressive subtypes (54), and basal-like aggressive subtypes such as triplenegative carcinomas (47), but was not an independent factor prognostic of poor survival. In melanoma, a retrospective study of 150 tissue samples reported a shorter overall survival (OS) in patients with nestin-positive disease (54).
Nestin has rarely been evaluated as a biomarker for ovarian cancer and not at all in a neoadjuvant setting. Based on the hypothesis that HGEOC recurrence is provoked by proliferation of chemoresistant CSCs, in which nestin might play a major role (34, 55), this preliminary study aimed to evaluate nestin as a prognostic biomarker in patients with HGEOC treated by NACT, as this therapeutic approach is currently used for a great number of patients (56).
Patients and Methods
Patients. Patients with primary ovarian, fallopian tube or peritoneal high-grade epithelial cancer who underwent NACT followed by cytoreductive surgery between January 2009 and December 2019 at the Jean Perrin Centre in Clermont-Ferrand, France, were retrospectively included.
Histological type, grade and disease stage were determined according to the World Health Organization criteria (57) and the FIGO (58) system. Exclusion criteria were: Absence of cytoreductive surgery following NACT, clear-cell or mucinous histotype, and insufficient samples for evaluation of nestin by immunohistochemistry.
Patient demographic and clinical characteristics, namely age, histopathological type and grade, presence of a deleterious germline or somatic BRCA mutation, FIGO stage, chemotherapy protocols and number of cycles, cytoreductive surgery peritoneal carcinomatosis index (PCI) (59), surgical classification of post-NACT residual disease, pathologically determined number of invaded locoregional tissues at surgery, chemotherapy response score (60) and the date of first relapse were collected by reviewing patients, electronic files.
Primary endpoints were OS, defined as the time from diagnosis to death, and progression-free survival (PFS), defined as the time from diagnosis to clinically or radiologically confirmed relapse, progression, death, or the date of the last follow-up visit. Relapse was confirmed according to the Response Evaluation Criteria for Solid Tumours. Isolated cancer antigen-125 increase was not considered to indicate relapse.
Tissue samples and immunohistochemistry. To assess nestin expression, a discovery cohort was assembled from the formalin-fixed, paraffin-embedded HGEOC tissue samples obtained by biopsy before NACT, available at the archives of Pathology and Molecular Pathology Department of the Jean Perrin Centre, Clermont-Ferrand, France.
Tumour tissue sections 4 μm-thick were first treated by Cell Conditioning Buffer 1 (Ventana/Roche Diagnostics, Meylan, France) for 1 h at 95°C, for antigen retrieval. Then they were incubated for 1 h at 37°C with a rabbit monoclonal antibody to human nestin (clone EP287; Epitomics, MM France, Brignais, France) at 1:200 dilution. The antigen-antibody reaction was revealed by ultraView Universal DAB Detection Kit (Ventana/Roche Diagnostics). The entire immunohistochemical procedure was performed in a Benchmark Ultra automate (Ventana/Roche Diagnostics).
Nestin expression was then semi-quantitatively evaluated by the H-score (61, 62), which was calculated as the sum of products obtained by multiplication of the percentage of stained tumour cells with the staining intensity graded as 1 (weak), 2 (moderate) or 3 (strong). The scoring was performed by a pathologist unaware of the patients, outcomes.
Statistical analysis. Primary analyses were performed on the entire study population. Associations of clinicopathological features with disease progression and survival were investigated by Cox proportional hazards model.
Multivariate analysis was conducted to identify independent prognostic factors as well as to estimate their effect on PFS and OS when adjusted for covariates. To perform model selection in multivariate models, the least absolute shrinkage and selection operator method was used.
Kaplan-Meier PFS and OS survival curves related to clinical characteristics were prepared and compared using the log-rank test. Time-dependent receiver operating characteristics curves and maximally selected rank statistics were used to investigate the possibility of defining nestin-positive and negative groups. Fisher’s exact test and Wilcoxon-Mann-Whitney test were used to compare subgroups. All statistical tests were two-tailed, and p-values of less than 0.05 were considered to indicate statistical significance. All statistical analyses were performed using R software (version 3.6.1) for statistical computing and graphics.
Ethical considerations. The study was initiated after obtaining Institutional Board approval no. 2022/CE 32 (Ethics Committee Sud-Est 6 of Clermont-Ferrand). All patients were informed about the purpose and nature of the study; no objection was received. All deceased patients had signed an agreement allowing the use of their tissue samples for research purposes. The study was designed, conducted, and analysed independently of any funding or sponsorship.
Results
Patient characteristics. A total of 92 patients were included in the study. Their median age was 65 years (range=42 to 83 years). Fifty-four patients (58.7%) had FIGO stage III and 36 (39.1%) had stage IV disease; 85 patients (92.4%) had serous histological type and 79 (85.9%) had high MD Anderson grade; 82 patients (89.1%) had no macroscopic residual disease after surgery. Mutation profiling of BRCA was performed in 69 patients (75%): eight patients (8.7%) had BRCA1 mutation, eight (8.7%) had BRCA2 mutation and 53 (57.6%) presented no BRCA mutation.
Study treatments: All patients received a carboplatin-paclitaxel chemotherapy regimen with a median of 4 neoadjuvant and 3 adjuvant cycles. Fifty-eight patients (63%) received between 4 and 6 NACT cycles; only two patients (2.2%) had disease progression during NACT; 80 patients (87%) received 3 or fewer adjuvant cycles; 27 patients (29.3%) received adjuvant bevacizumab.
Outcomes. The median follow-up period was 43 months. Sixty-four (69.6%) relapses and 38 (41.3%) deaths were documented. Forty-eight patients who experienced recurrence had peritoneal carcinomatosis (75%). Median PFS and OS were 20 [95% confidence interval (CI)=17-23] and 49 (95% CI=44-78) months, respectively. Table I summarizes the baseline clinico-pathological characteristics of the study population.
Patient characteristics.
Univariate and multivariate analysis of known clinico-pathological prognostic factors. We performed an evaluation of known prognostic factors in HGEOC for the total study population (n=92). Clinico-pathological variables were included in a Cox proportional hazards regression model to identify independent prognostic factors.
In univariate analysis, higher PCI, FIGO stage and number of involved locoregional post-NACT tissues removed at surgery were all associated with worse PFS (Table II), without statistically significant impact on OS (Table III). Absence of a platinum-based treatment of the first relapse was not associated with reduced OS and PFS. In multivariate regression analysis, higher initial FIGO stage was an independent prognostic factor for reduced PFS (HR=2.14, 95% CI=1.37-3.34, p=0.001) and OS (HR=1.73, 95% CI 0.98-3.05, p=0.057). Increasing PCI was associated with reduced PFS (HR=1.11, 95% CI=1.04-1.19, p=0.001) and OS (HR=1.06, 95% CI=0.99-1.14, p=0.063), with each PCI unit increasing the risk of disease progression by 12%. The number of locoregional tissues involved, chemotherapy response score and cytoreduction score were not significantly associated with PFS or OS.
Progression-free survival according to prognostic factors.
Overall survival according to prognostic factors.
Evaluation of nestin prognostic value. The cohort included 86 patients for whom the tumours were high-grade serous carcinomas. Nestin was predominantly located on the cell membrane. The median percentage of stained tumour cells was 15% (range=0-100%). The median H-score was 20 (range=0-290 and 95% CI=15-40). There were 72, 59, 51 and 40 nestin-positive tumours, according to cut-offs using 5% and 10% stained cells, and H-score of 20 and 30 (Table IV).
Number of patients in each nestin group according to H-score threshold (n=92).
Median PFS and OS for the whole cohort of 92 patients were 20 (95% CI=17-23) and 49 (95% CI=44-78) months, respectively (Figure 1). A time-dependent receiver operating characteristic approach was unable to determine a cut-off with sufficient capacity to separate the cohort into groups with statistically significant differences in PFS or OS. Although statistically non-significant, nestin positivity defined by an H-score ≥30 was shown to be the most reliable among several tested cut-offs (5% and 10% of nestinexpressing tumour cells, as well as H-scores of 10, 20 and 30). The results obtained with this cut-off are reported in Table V. A median PFS of 22 months was recorded for the nestin-positive subpopulation (26 events) and 19 months for nestin-negative subpopulation (38 events) (p=0.57) (Figure 2). The median OS was shorter for the nestin-positive group than for the nestin-negative group (48 vs. 67 months, respectively) but the difference did not reach statistical significance (p=0.43) (Figure 2).
Kaplan-Meier plot of the progression-free (A) and overall (B) survival of patients with primary ovarian, fallopian tube or peritoneal highgrade epithelial ovarian cancer who underwent neoadjuvant chemotherapy followed by cytoreductive surgery (n=92).
Kaplan-Meier plot of the progression-free (A) and overall (B) survival of patients with high-grade epithelial ovarian cancer who underwent neoadjuvant chemotherapy followed by cytoreductive surgery according to nestin assessment (nestin positive when H-score ≥30).
Cohort characteristics according to nestin H-score (positive: H-score≥30).
Discussion
This study was undertaken to explore the relationship between nestin protein expression assessed by Immunohistochemistry and the prognosis of HGEOC. To the best of our knowledge, this is the first evaluation of the prognostic value of nestin in primary ovarian or peritoneal high-grade carcinomas treated by NACT followed by cytoreductive surgery.
No single parameter or signature has been confirmed to be a universal prognostic biomarker of HGEOC(61). In such a context, Qin et al. (62) (n=123) and He et al. (63) (n=113) reported a statistically significant positive influence of nestin expression on chemotherapy resistance and survival in HGEOC but those results need further confirmation. In addition, a recently published meta-analysis including six retrospective studies reported that nestin expression was significantly associated with advanced FIGO stage of HGEOC and larger size of residual tumour after first-line surgery (64).
Our study has several limitations, among which the small size of the discovery cohort is the most important. Interestingly, despite having more patients with a low chemotherapy response score (i.e., more viable cells after NACT), the cohort was not able to provide enough power for statistical confirmation of a putative prognostic value for nestin. We conclude that the cohort likely did not contain enough high expressors to confirm their shorter OS. Indeed, we did not include patients with primary resistance to NACT, whose tumours might have had higher nestin expression. Literature data showed a positive association between nestin and chemoresistance (65, 66). Finally, the inclusion of patients whose NACT was initiated before presentation at our Institution may have increased the selection bias.
Our evaluation of clinical, surgical and pathological prognostic factors results for our cohort (n=92) are in agreement with those previously reported (67, 68). Complete resection of all macroscopic disease is a well-established negative independent prognostic factor for ovarian cancer (27). However, the very low proportion of patients with residual macroscopic disease (5.4%) did not allow for confirmation of the unfavourable impact of this factor as reported in larger studies (69). This discrepancy might be explained by the high quality of surgical resection of ovarian cancer deposits in a comprehensive cancer centre like ours, as well as our smaller sample size.
As a cancer stem cell marker, nestin is most frequently reported to be expressed by a relatively small tumour cell population, which might have been missed in our pre-NACT biopsy sampling of HGEOC. In that light, evaluation of nestin as a prognostic biomarker might be better for post-NACT HGEOC residues, in larger samples with a possibly increased number of chemotherapy-resistant cancer stem cells.
Nestin has also been evaluated as a neovascularization marker (46-48). In this study, we did not measure its expression by intratumoural vessels due to the small biopsy samples used. However, it would be of interest to evaluate nestin expression in post-NACT HGEOC residues in blood vessels. In residual disease post-NACT, nestin expression in tumour cells or vessels, as well as both scored together, might provide prognostic information.
As with any biomarker detected by immunohistochemistry, the best procedure for detection and measurement of nestin expression remains to be defined. The authors of previous studies used mouse monoclonal and rabbit polyclonal antibodies to nestin at 1/50-200 dilution (62, 63). We used a highly specific rabbit monoclonal antibody, validated for immunohistochemistry by the manufacturer, which resulted in intense staining of nestin expressed by the tumour vessels (internal control) and no background staining in the negative control (external tissue control). Because in the majority of samples, tumour cell expression of nestin was weak, the immunohistochemistry method we exploited needs to be verified in terms of sensitivity using specifically constructed tissue controls containing tissues with low nestin expression.
Since prospective biomarker trials are very costly, independent confirmation of findings for specific biomarkers in multiple prospective-retrospective studies remains highly important (70). Assessable by a low-cost and highly accessible method, immunohistochemistry, nestin remains an interesting putative prognostic biomarker, potentially transposable to real-life cancer-care strategies.
Although we were unable to demonstrate the prognostic value of nestin in HGEOC treated by NACT, we believe that the methodology employed in this preliminary study is a valid approach towards further exploration of nestin as a biomarker for this disease. Therefore, we are intending to pursue evaluation of nestin expression in a larger cohort of pre-NACT HGEOC samples, as well as in post-NACT residues, in order to better position this putative biomarker.
Footnotes
Authors’ Contributions
All Authors actively contributed to the conception, design, execution, or interpretation of the current study.
Conflicts of Interest
No potential conflicts of interest were disclosed by the Authors.
- Received March 14, 2022.
- Revision received May 5, 2022.
- Accepted June 1, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
