Endometrial carcinoma. Improving diagnosis, deconstructing carcinoma invasion, pursuing new therapeutic strategies

Resumen

Endometrial cancer (EC) is the second most frequent cancer of the gynecological female tract with over 287,100 new cases diagnosed per year. It accounts for 9.4% of the worldwide incidence of cancer in women. EC is primarily considered a cancer of the developed world, and its age-standardized incidence rates have been continuously rising in the past decade, due to overall population aging and growth, as well as the adoption of cancer-associated lifestyle choices. Recent data have focused on developed countries, estimating EC as the most frequent cancer of the gynecological female tract and increasing its crude incidence up to 13 new cases per 100,000 women per year. In spite of its high incidence, EC is associated with a relatively low mortality rate, since 70% of the newly reported cases are usually detected in its initial stages, when the cancer is highly curable and the overall survival rates at 1- and 5- years are 92% and 83%, respectively. EC is usually divided into two clinicopathological types: Type I (EEC) and Type II (NEEC). Molecular genetic evidence indicates that ECs, as described in other malignancies, likely develop as the result of a stepwise accumulation of alterations in the cellular regulatory pathways, such as oncogene activation and tumor suppressor gene inactivation, which lead to dysfunctional cell growth. These molecular alterations appear to be specific in Type I and Type II cancers. Among these categories, endometrioid carcinoma (Type I) corresponds to the most common type with an incidence of 80%. At this time, there is an unfortunate lack of accurate screening methods, which possess both high sensitivity and specificity, making the early detection of this disease a challenge. To date, multiple studies have been conducted to profile EC. Risinger et al. reported distinct gene expression profiles among different subtypes of EC using microarray analysis. In our group, we have identified a number of genes associated with endometrial carcinogenesis. Serum markers for the detection of uterine cancer have also been reported in the literature. Yurkovetsky et al. identified prolactin as a serum biomarker possessing both sensitivity and specificity for EC. They found that serum CA 125, CA 15-3 and CEA were higher in patients with Stage III disease when compared to those with Stage I. According to that study, a five-biomarker panel of prolactin, GH, eotaxin, E-selectin and TSH was able to discriminate between EC and ovarian and breast cancer. Nevertheless, none of these potential biomarkers have been validated, nor have they reached clinical practice for diagnosis. According to the current guidelines of the International FIGO and the American Cancer Society (ACS), the standardized protocol for EC diagnosis focuses first on adequately informing at risk women of the importance of reporting to their physicians any unexpected bleeding or spots, since in 90% of all cases women with EC suffer these types of symptoms. When faced with clinical suspicion, a final diagnosis is accurately reached through methods routinely used in the clinics, such as biopsy by aspiration or biopsy guided by hysteroscopy and transvaginal ultrasound. Unfortunately, these methods are usually uncomfortable for the patient and sometimes rely on the subjective interpretation of visual images. Moreover, the actual protocol of diagnosis is not fully efficient, since only 30% of all patients screened are diagnosed with advanced stages of the disease, while 85% of the symptomatic women screened are healthy, but have had to undergo unnecessary and invasive procedures. In conclusion, there is plenty of room for improvement in the area of EC diagnosis. In order to overcome these diagnostic shortcomings, we aimed to identify new molecular markers for EC and to evaluate their validity in endometrial aspirates, in order to set up the basis for the development of a viable diagnostic test. Such a test could be used to screen patient risk groups, to ameliorate the sensitivity and specificity of the endometrial biopsy and to preclude the performance of unnecessary hysteroscopies. In order to accomplish this goal, we conducted a large cDNA microarray study comparing 52 EC to 10 normal hysterectomy tissue samples. This led us to identify more than 100 potential biomarkers and validate 20 of them at RNA level in an independent series of 17 paired hysterectomy and two biopsy tissue samples. Some of these were also validated at the protein level by immunoblotting and tissue microarrays on tissue samples. The candidate genes included ACAA1, AP1M2, CGN, DDR1, EPS8L2, FASTKD1, GMIP, IKBKE, P2RX4, P4HB, PHKG2, PPFIBP2, PPP1R16A, RASSF7, RNF183, SIRT6, TJP3, EFEMP2, SOCS2 and DCN. These specific genes were selected based on their fold expression and p-values, as well as on information that could be obtained on the public database, such as their subcellular localizations, their types of proteins and their associations with the cancer and/or other biological factors, i.e., basically any information that could altogether improve their utility as diagnostic biomarkers. Next, we confirmed whether or not the gene expression profiles of uterine aspirates efficiently mirrored those of the paired tissue samples by analyzing those markers in an independent series of nine paired samples of uterine aspirates and primary tumors from the same patients using RTqPCR. We then confirmed the validity of these uterine aspirate biomarkers at the RNA level, in order to differentially classify the carcinoma cases and the control samples. We screened 26 aspirates from patients with different types and grades of EC and 24 aspirates from healthy donors using RTqPCR. The samples themselves included atrophic endometrium, normal endometrium with polyps from postmenopausal women and samples from premenopausal women, both in secretory phases and in proliferative phases of the cycle. After obtaining successful results in this first validation set, we were able to obtain results from a larger, clinical validation, which included 519 patients from 14 different clinical institutions around Spain. At this present time, the validity and clinical application of a new EC diagnostic test, based on these endometrial aspirate markers, is currently being evaluated. Despite the obvious importance of establishing a valid and effective diagnostic protocol to improve patient survival, it must be remembered that once a definitive diagnosis has been achieved, it is probable that advanced stages of EC will have already been encountered. This, of course, represents a poor prognosis and a decrease in patient survival rates. Nowadays, endometrial carcinoma is the most frequent infiltrating tumor of the female genital tract. Its direct extension to the myometrium is considered its most common route of spread and one of the most important prognostic factors for Type I and Type II cases of EC. The former represents a positive outcome, since optimal surgery, the best therapeutic alternative currently available, is possible. Type II cases are usually correlated to lymphatic spread, risk of recurrence, and overall survival rate. These tumors are generally associated with poor prognosis and a dramatically decreased rate of survival. Therapeutic options are usually based on radiotherapy and chemotherapy, which show limited efficacy when dissemination and metastasis are already present. It follows that an important key to improving patient survival would rely on the characterization of the molecular basis responsible for the initiation of and/or a means to control the invasive processes of endometrial carcinoma. The mechanisms involved in metastasis and tumor invasion are still poorly understood. For this reason, our group began its extensive research focus on the identification of new genes involved in EC invasion, specifically for the most frequent type of EC, the EEC. Tumor invasion defines the frontier between the tissue-restricted carcinoma and the disseminated tumor cells. The data presented in the PhD thesis of Planaguma and Monge showed that an up-regulation of RUNX1 and ETV5 genes specifically and significantly increased in those tumor stages associated with myometrial infiltration. The data proved that ETV5 plays a role in orchestrating the downstream effects that confer onto the tumor the invasive abilities needed for dissemination. Specifically, ETV5 was found to activate MMP2 gelatinase activity, which remodels the ECM surrounding the tumor, and Hep27 expression, which protects tumor cells facing oxidative stress from apoptosis. As a consequence of these findings, our group began investigations in the role of ETV5 in the crucial process of EMT and its relation to the tumor microenvironment. Many authors have stated that this transformation represents an early step in the invasive process of many different epithelial carcinomas. The tumor microenvironment has also become a growing area of research that has already proved to be important in the elucidation of tumor progression and invasion. The findings presented in this thesis depict a new role for ETV5 as a direct transcriptional inductor of the EMT process, able to cross-talk with the environment through its cooperation with LPP. The alterations associated with the acquisition of a mesenchymal phenotype were first characterized at the molecular level, depicting ETV5 as having a main regulatory effect by modulating E-Cadherin repression. It is widely agreed by the scientific community that a loss of E-Cadherin is a principal hallmark of EMT. Consequently, our group demonstrated that ETV5 transcriptionally activated the zinc finger E-box binding transcription factor, ZEB1, an important E-Cadherin repressor. We concluded that ETV5 was capable of inducing EMT. Relating our in vitro findings with the clinics, high levels of ZEB1 have also been described in human samples, encouraging progression in gynecological carcinomas 130. In addition, using high throughput screening, interplay between the transcriptional repressors of E-Cadherin and miRNAs has been linked to EMT-activation in human EC cases. EMT not only endows cells with migratory and invasive properties, it also induces stem cell properties, prevents apoptosis and senescence, and contributes to immunosuppression. The mesenchymal state is associated with the capacity of cells to migrate to distant organs and maintain their stemness, allowing their subsequent differentiation into multiple cell types during development and the initiation of metastasis. In accordance with previous results, we found that the transition to a mesenchymal phenotype upon an increase in ETV5 expression resulted in the acquisition of migratory capabilities. Importantly, the invasive properties of cells presenting increased levels of ETV5 were significantly increased, resulting in an effective and collective directional invasion in a three-dimensional (3D) in vitro assay that mimicked tumor invasion. In addition, ETV5 overexpressing cells became more adhesive, in agreement with the premise that cells need to adhere before further invasion, in order to improve the efficacy of this initial step in the metastatic process. However, as previously mentioned, it is not only the increased adhesion and invasive capabilities of transformed cells that are needed for successful tumor progression, but also the ability of those cells to communicate and understand the signals of the surrounding microenvironment. Hence, in order to better understand the role of ETV5 within the context of actively invading tumors, we used a proteomic approach to identify new agents in endometrial carcinoma invasion. Interestingly, we found that LPP significantly increased in invasive ECs. Our data in the 3D in vitro invasion assay confirmed an independent role for LPP in EC invasion. Moreover, our analysis of the pattern of expression of LPP and ETV5 in human EC cases showed a correlation to nuclear LPP expression and a concomitant specific expression at the invasive area of the tumors. This signaled the cooperation between both of these agents in enhancing the invasive capabilities of ECs. Interestingly, the presence of increased levels of ETV5 resulted in the redistribution of LPP from cell-cell contacts to focal adhesions. Proteins within focal adhesion may function as a cell surface sensor to regulate cellular signaling and dynamic responses. In our model, the translocation of LPP was translated into an increase in ETV5 transcription activity that finally resulted in a feedback loop mechanism, where ETV5 promoted further invasion. In accordance with these data, we also demonstrated a link between ETV5 and LPP in response to extracellular signals in the promotion of carcinoma invasion. As many authors have suggested and we were able to prove, the orchestrated modulation of cell adhesion is essential for appropriate cell responses and intercellular communication. This dynamic control of adhesion molecules is important for progression through the metastatic cascade. It also allows cell release from the primary tumor, invasion of the surrounding matrix and the intravasation and adhesion to vascular endothelial cells, in order to facilitate extravasation. Our model of the early steps of EC invasion postulates that increased levels of ETV5 are translated into a transcriptional promotion of EMT. As a consequence, LPP localizes at the focal adhesions, where extracellular signals are amplified by its translocation to the nucleus. It then promotes a further activation of ETV5 transcription activity. This amplification loop results in a deep and persistent EC invasion. From a clinical point of view, our molecular approach for understanding myometrial infiltration could serve as decisive input towards the design of new therapeutic approaches in the fight against the high-risk phenotype of EEC. However, the definition of a high-risk patient, i.e. a patient possessing an increased risk of relapse, correlates perfectly to the depth of myometrial invasion encountered in EEC patients, though not for NEEC patients. Cases of NEEC are characterized by aggressive histological subtypes with poor prognoses, even when they are diagnosed in their early stages and the tumors confined within the endometrium. Therefore, our group decided to focus on profiling high risk and low risk EC, in order to unveil more molecular events related to the high-risk phenotypes of both subtypes. The goal was to provide targets for developing new therapeutic approaches that would inhibit tumor spread and improve the outcome for high-risk EC patients. First, the molecular phenotype of high-risk of recurrence EC was characterized by high throughput transcriptomic analysis. Focusing on the core of the network that supports the molecular determinants of EC recurrence, we found molecules that were already associated with an aggressive phenotype, such as HIF1A and NR3C1. Interestingly, the majority of the genes at the core of the molecular network relating to a high risk of recurrence in EC referred to the process of trophoblast invasion. Genes, such as FOS, MMP9, MAPK1, RHOA, and TGFß1, have been described as the main directors of the embryo implantation process. And, in fact, the molecular events characterizing trophoblast implantation as a physiological process of invasion into the uterus, strictly controlled at both the time and space levels, have been paralleled to the processes of endometrial tumor invasion in an uncontrolled manner. In particular, TGFß1 was located at a central position in the molecular network, suggesting a critical role for this cytokine in the highly invasive ability of endometrial cancer cells. TGFß1 signals are widely recognized as tumor promoters of cellular responses, such as proliferation, survival, migration and invasion, and its overexpression has been associated with metastatic phenotypes and poorer patient outcomes. With regards to the uterus, the TGFß1 pathway has been associated with decidualization, whereas its disruption in both endometrial hyperplasias and carcinomas has been found to result in the loss of growth inhibition, acting as a key factor in the early steps of endometrial carcinogenesis. On the other hand, TGFß1 receptor type II and other components of the signaling pathway have been associated with myometrial infiltration, local spread and distant metastasis in ECs. Our group also linked the TGFß1 pathway to the increased invasive ability promoted by the ETV5 transcription factor during the initial steps of EC dissemination. In this new study, we showed that TGFß1 induced changes in the Hec1a and RL95-2 cell lines, which were compatible with EMT, both at the morphological and the molecular levels. In addition, an increase in Hec1a and RL95-2 invasion capacity was found when TGFß1 was used as the chemoattractant, though it was completely abolished in the presence of its specific inhibitor, SB-431542. This would indicate that TGFß1 plays a role in endometrial tumor cell invasiveness. The 3D cell invasion assay, stimulated by EGF as the chemoattractant, was used to further evaluate the real impact and dependence on TGFß1 signaling in the acquisition of an aggressive tumor phenotype. Our results showed that when EGF was used as the chemoattractant, EMT promoted by TGFß1 was indispensable, though not sufficient, since TGFß1 represented only a limited step in deep tumor invasion. Moreover, EGF attraction promoted the generation of Hec1a cell spheroids that deeply invaded the matrix, while under TGFß1 stimulated conditions, the cells spread at the lower fields of the Matrigel®, rather than deeply invading the matrix. It has been described that the spheroid morphology of endometrial epithelial cells can develop polarized glandular structures of single-lumens, resembling those observed in endometrial tissue. These spheres seemed to form a cooperating unit with a highly persistent and directional migration into the 3D matrix. In vivo, this type of cluster migration would minimize cell loss, favor highly localized MMPs expression and protect cancer cells from immunological assault. We hypothesized a step-by-step process, where TGFß1 would stimulate EMT in the Hec1a cells, enabling them to migrate through the porous membrane prior to the development of the EGF-dependent invasive glandular structures. This cooperation was recently reported when TGFß1 stimulation of EMT enabled an EGFR-driven breast cancer model to abandon its inherent branching architecture and form large, undifferentiated masses that were hyperinvasive to EGF. Interestingly, when we applied both EGF and TGFß1 as chemoattractants, we observed the same result as when we applied TGFß1 alone, with no spheroid formation. This would suggest that once TGFß1 has initiated tumor infiltration through the promotion of EMT, its contribution must be counteracted for further persistent invasion. This could possibly account for the extensively discussed evasive character of EMT in carcinomas. All of the above suggest a role for TGFß1 in early endometrial carcinoma invasion, which is associated with the EMT process and the acquisition of a migratory/invasive phenotype during myometrial infiltration. Translated to the appearance of secondary metastatic lesions, we hypothesized that TGFß1 would promote the transition of epithelial circulating tumor cells into an invasive phenotype during implantation of the metastatic niche, prior to post-EMT persistent invasion and micrometastasis formation. For this reason, a therapeutic approach that could target TGFß1-promoted migration and invasion of both primary and secondary tumors would represent an attractive strategy for the prevention and treatment of local and distant recurrence in EC. Interestingly, since the TGFß1 signals seem to play a critical role in EGF-promoted invasion, the combination of both TGF-ß1 and EGFR inhibitors could be promising as a therapeutic strategy. Along these lines, it is important to mention that the clinical trial of GOG 181-B failed to show any significant activity for Trastuzumab in EC, even though both Type I and Type II tumors frequently overexpressed EGFR. This expression was correlated to tumor grade, deep myometrial invasion and poor survival. Finally, the translation of these in vitro results into EC animal models is guaranteed to confirm the speculation that supposes a critical role for TGFß1 in the acquisition of an aggressive phenotype in human endometrial carcinomas. Likewise, other molecular pathways associated with a high-risk of recurrence in EC must be validated, in order to obtain a better understanding of the complex mechanisms driving tumor spread. Integration of the oncological parameters associated with an aggressive phenotype (i.e., the TGFß1 pathway), together with the histopathological parameters currently used in the clinics, would help to more accurately stratify an increased risk of recurrence. In conclusion, in this final study we managed to (i) characterize the molecular changes associated with the acquisition of an aggressive phenotype in human EC, (ii) identify and characterize TGFß1 as a main director of endometrial cancer cell invasion, and (iii) demonstrate that our approach represents a realistic strategy for the rational identification and characterization of new, potentially therapeutic targets. All together, these studies represent a concerted effort to improve the clinical management of and the outcome for high-risk endometrial cancer patients.