Background
Epithelial ovarian cancer is associated with the highest mortality among gynecologic cancers [
1]. Approximately 60–70% of patients are detected in an advanced stage (stages III and IV) at the time of diagnosis, because ovarian cancer rarely disseminates through the vasculature but has a tendency to metastasize to the peritoneum, unlike many other types of solid tumors [
2] [
3]. These patients with peritoneal dissemination at diagnosis have a 5-year survival rate of less than 30% and a very high recurrence rate even with a successful response to surgery or chemotherapy [
1] [
2] [
4]. In recent years, successful improvement of survival with newly developed therapeutic agents, such as anti-angiogenic agents and poly ADP ribose polymerase (PARP) inhibitors, has been reported [
5] [
6] [
7]. However, the mortality rate of ovarian cancer patients is still far from satisfactory, and so a new therapeutic strategy is warranted.
P-cadherin, a member of classical cadherin superfamily, is a cell adhesion factor first reported in 1986 as a new subtype of the cadherin found in mouse embryonic development [
8]. Besides its regulatory role in implantation, embryo morphogenesis, cellular homeostasis, cell differentiation, cell shape, cell polarity, and growth and migration in fetal development [
8] [
9] [
10], aberrant expression in various cancers has been reported. Since it is only faintly expressed in limited organs of normal adult tissue, it has been attracting attention as a therapeutic target.
Some studies on P-cadherin and ovarian cancer have been reported. Decreased E-cadherin and increased P-cadherin expression, so-called cadherin switching, has been observed when a tumor progresses from stage I to II, indicating its functional role at the stage of cancer spreading from the primary lesion to pelvic cavity [
11] [
12]. High expression of P-cadherin was reported to have a negative impact on survival in patients with high-grade serous subtypes of ovarian cancer [
12]. Moreover, using the mouse ovarian cancer xenograft model, inhibition of P-cadherin by RNAi resulted in decreased peritoneal implantation [
3]. These reports suggest the involvement of P-cadherin in ovarian cancer progression and the effectiveness of developing treatment targeting P-cadherin for ovarian cancer; however, the precise profile of P-cadherin expression in terms of histological subtypes, and the proportion and distribution of positive cancer cells in metastatic lesions and association with clinicopathological characteristics, are largely unknown. The aims of this study were to provide the expression profiles of P-cadherin in ovarian cancer, which are mandatory for development of targeted therapy in near future.
Discussion
This is the first study to demonstrate the expression profile of P-cadherin in ovarian cancer patients from the viewpoint of histology and tumor distribution for target therapy. First, we documented that P-cadherin was strongly corelated with unfavorable prognostic factors, with a trend toward poor survival. Second, we showed that P-cadherin is intensely and broadly expressed in specific tissue subtypes. Third, we showed that the expression pattern of P-cadherin in the primary lesion was preserved in the disseminated lesions and probably in recurrent lesions, but was reduced in metastatic lymph nodes. Together, these findings will shed light on the future development of drug-delivery strategies targeting P-cadherin in advanced ovarian cancer.
The most important finding of this study was that we clarified the expression profile of P-cadherin in ovarian cancer patients. In endometrioid and serous subtypes, 21/30 (70%) and 59/75 (78.7%) were classified as “High”, with a total immunohistochemistry score of 4 or more, whereas 18/34 (52.9%) of the clear cell subtype did not express P-cadherin at all and 30/34 (88.2%) were classified as “Low” in primary lesions. These results indicate that patients with specific subtypes of ovarian cancer are candidates for P-cadherin-targeted treatment. Furthermore, when P-cadherin expression in metastatic foci was evaluated, interestingly, corresponding disseminated foci showed similar staining to the primary foci, but it was decreased in metastatic lymph nodes (Fig.
4). Decreased E-cadherin and increased P-cadherin, called cadherin switching, has been observed when cancer progresses from FIGO stages I to II, when cancer cells spread from the primary tumor to pelvic cavity [
11] {Patel, 2003 #489} [
12]. Some of the mechanisms have been reported. Cheung LW et al. reported that promotions of cell migration and invasion by gonadotrophin releasing hormone (GnRH) via activation of p120 catenin signaling are mediated by the P-cadherin/insulin-like growth factor-1 receptor (IGF-1R) complex. Usui et al., using an in vitro and mouse xenograft model of ovarian cancer, demonstrated that inhibition of P-cadherin by RNAi decreased the aggregation and survival of cancer cells floating in ascites and reduced the number of peritoneal implants [
3]. Taken together, these reports suggest that P-cadherin contributes to the establishment of peritoneal dissemination in ovarian cancer. Our results further indicate that P-cadherin is broadly and strongly maintained after the completion of dissemination and that P-cadherin-targeting therapy may have effects on a wide range of lesions including not only primary but also disseminated foci. From another perspective, it is suggested that biopsy of the disseminated foci acts as a substitute when biopsy of the primary lesion is difficult or highly invasive. On the contrary, P-cadherin expression was reduced in metastatic lymph nodes compared with that in concomitant primary lesions. The effect of P-cadherin-targeted therapy may not be predictable based on P-cadherin expression of the primary tumors thus, sampling is mandatory when treating patients with lymph node metastases. In fact, P-cadherin expression of dissected recurrent metastatic lymph nodes demonstrated a similar P-cadherin score to the primary tumor (Table S1), and so the patient may benefit from P-cadherin-targeted therapy. We also examined the P-cadherin expression in recurrent foci in contrast to the primary lesion. Although it is a small-scaled analysis, the P-cadherin expression pattern was maintained in the recurrent site, indicating that the results of immunohistochemistry of a primary lesion may sufficiently predict the expression of a recurrent lesion when sampling is deemed highly invasive.
Another important finding in this study was elucidation of the correlation of prognostic factors with P-cadherin in ovarian cancer patients. P-cadherin expression was significantly correlated with the histological subtype and unfavorable clinicopathologic features of ovarian cancer patients, including a high FIGO stage, positive peritoneal dissemination, and distant site metastases. Although not significant, patients with high P-cadherin expression showed a trend toward shorter DFS and OS on univariate survival analysis (Table
3, Fig.
2). When the survival analysis was limited to P-cadherin high-expressing subtypes, mucinous, endometrioid, and serous subtypes, the Kaplan-Meier survival curve demonstrated significantly decreased survival probability of DFS (
P = 0.0264) and OS (
P = 0.0227) in the P-cadherin High population (Fig. S
1). In multivariate analysis, CA125, stage, peritoneal dissemination, and histological type, but not P-cadherin expression, were confirmed as independent prognostic factors in ovarian cancer patients. Together with clinicopathological analysis, P-cadherin may act as a confounder of other prognostic factors. As Van Marck et al. [
15] reported, the impact of P-cadherin on survival differs in cancer types. In gastric [
16] and oral [
17] cancer, it has reported to be a good prognostic factor whereas in breast [
18] [
19], endometrial [
20], gallbladder [
21], colon [
22], and pancreatic cancer [
23], its overexpression indicates a poor prognosis. The fact that P-cad expression strongly correlates with unfavorable prognostic factors in ovarian cancer patients indicates that P-cadherin-targeted treatment benefits such a population, and the development of effective treatments may contribute to the improvement of survival.
Table 3
Univariate and multivariate analyses of predictive factors for survival of ovarian cancer patients
Age a (years) | | .187 | | | | .176 | | | |
≤ 55 | 79 | | | | | | | | |
> 55 | 83 | | | | | | | | |
CA125a | | 6.3e-4* | 2.29 | 1.08–4.84 | .030* | .0017* | 2.31 | 1.10–4.84 | .027* |
≤136 | 79 | | | | | | | | |
> 136 | 81 | | | | | | | | |
Stageb (FIGO2014) | 1.0e-7* | 3.40 | 1.32–8.80 | .011* | 4.9e-7* | 4.03 | 1.53–10.6 | .0048* |
Stage I/II | 88 | | | | | | | | |
Stage III/IV | 74 | | | | | | | | |
Peritoneal dissemination | 5.0e-8* | 4.06 | 1.42–11.6 | .0090* | 2.5e-6* | 2.94 | 1.03–8.38 | .044* |
Positive | 76 | | | | | | | | |
Negative | 86 | | | | | | | | |
Lymph node metastasis | .091 | | | | .109 | | | |
Positive | 24 | | | | | | | | |
Negative | 138 | | | | | | | | |
Distant site metastasis | .062 | | | | .095 | | | |
Positive | 10 | | | | | | | | |
Negative | 152 | | | | | | | | |
Histologic typeb | | .014* | 0.42 | 0.20–0.92 | .029* | .030* | 0.42 | 0.19–0.93 | .032* |
Serous | 75 | | | | | | | | |
Others | 87 | | | | | | | | |
P-cadherin | | .058 | 1.18 | 0.56–2.49 | .669 | .050 | 1.19 | 0.55–2.58 | .653 |
High | 95 | | | | | | | | |
Low | 67 | | | | | | | | |
In recent years, antibody drug therapy is being intensively researched and developed in the field of cancer treatment. Antibody therapy is largely classified in to two categories: those that act directly on functional molecules, and those that are used as drug carriers. Numerous monoclonal antibody drugs, which act on functional molecules, have been developed and have proven survival benefits. Although still in the experimental stage, the effectiveness of cell adhesion inhibition by an antibody targeting the X-dimer of P-cadherin has been reported [
24]. Further advances in antibody technology have led to the rapid development of an antibody-drug conjugate (ADC) and radioimmunotherapy (RIT) that use antibodies specifically binding to cancer antigens as carriers. The first ADC approved by U.S. Food and Drug Administration (FDA) was Gemtuzumab ozogamicin for acute myelogenous leukemia (AML) in 2001 [
25]. Since then, several ADC, such as Brentuximab vedotin for relapsed Hodgkin’s lymphoma (HL), systemic anaplastic large cell lymphoma (sALCL), Trastuzumab emtansine for HER2-positive metastatic breast cancer, and Inotuzumab ozogamicin for relapsed CD22-positive B-cell precursor acute lymphoblastic leukemia (ALL), have been applied in clinical practice. As for RIT, Ibritumomab tiuxetan and Iodine tositumomab, both of which use anti-CD20 to conjugate Yttrium-90 or Iodine-131, respectively, have been used to treat non-Hodgkin’s lymphoma patients [
26]. Because P-cadherin shows faint expression in normal adult tissues in restricted organs such as hair follicles and the breast [
27], and aberrant expression in numerous cancers [
18] [
19] [
20] [
21] [
22] [
23], drug-delivery systems targeting P-cadherin as a tumor-associated antigen are attractive treatment strategies. In this study, advanced ovarian cancer, especially endometrioid and serous subtypes, demonstrated intense and broad expression of P-cadherin not only in primary foci but also in disseminated lesions. Because the dissemination is the main cause of treatment failure and a poor prognosis in ovarian cancer patients, the development of ADC or RIT targeting P-cadherin may be optimal treatment to improve the prognosis of ovarian cancer patients. Although molecular targeting therapy or ADC targeting P-cadherin has yet not been developed, radionuclide bound to P-cadherin antibody has been developed. In the United States, phase 1 clinical trials have been conducted, in which Yttrium-90 bound to anti-P-cadherin antibody (FF21101) was administered to patients with solid tumors [
28]. Advanced ovarian cancer, especially the endometrioid or serous subtype, may be a good candidate for such treatment.
A limitation of this study is that we could not obtain and analyze P-cadherin expression in distant metastatic lesions in the lung or liver. The small sample number of recurrent lesions is also a limitation.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.