Membrane expression of TRAIL receptors DR4, DR5, DcR1 and DcR2 in the normal endometrium, atypical endometrial hyperplasia and endometrioid adenocarcinoma: a tissue microarray study

Purpose To evaluate the membrane expression of DR4, DR5, DcR1 and DcR2 in the normal endometrium (NE), atypical endometrial hyperplasia (AEH) and endometrioid adenocarcinoma (EAC). Methods The study comprised 197 patients: 20 NE, 18 AEH and 159 EAC. Tissue microarrays were constructed. Membrane expression of DR4, DR5, DcR1 and DcR2 was examined and presented as total score (TS). Results In EAC, the membrane expression of DR4, DR5 and DcR2 was less common compared to NE (p < 0.001; p < 0.001; p = 0.018) and AEH (p < 0.001; p < 0.001; p = 0.004). In EAC the membrane expression of DcR1 did not differ when compared to NE (p = 0.055) and AEH (p = 0.173). A strong correlation was found between the type of endometrial tissue (NE/AEH/EAC) and the TS of DR4 (p < 0.001), DR5 (p < 0.001), DcR1 (p = 0.033) and DcR2 (p < 0.001). In EAC, the TS of DR4, DR5, DcR1 and DcR2 was not related to grading and staging. In EAC, the membrane expression of DR5, but not DR4, DcR1 and DcR2, was related to better disease-free survival (DFS). The overall survival (OS) was not related to membrane TRAIL receptors expression. Conclusions The membrane expression of the receptors for TRAIL DR4, DR5, DcR1 and DcR2 is greater in NE than EAC. The level of membrane staining of the receptors in EAC is not dependent on grading and staging. In EAC patients, membrane expression of DR4, DR5, DcR1 and DcR2 are not independent predictors of survival.


Introduction
Programmed cell death (apoptosis) is a crucial process in the development and homeostasis of multicellular organisms, and the dysfunction of apoptosis is regarded as an important step in the development of cancer and spreading of metastases [1,2]. The cell death program consists of three essential types of elements: activators, inhibitors and effectors. The cytokine TNF-related apoptosis-inducing ligand (TRAIL), also known as Apo-2L, is one of the most important extracellular activators of apoptosis. TRAIL is a type II transmembrane protein located on chromosome 8p21-22, that selectively induces apoptosis in tumor cells while leaving normal cells intact [1][2][3][4][5].
The ligand TRAIL binds to five receptors. Two of them, DR4 (TRAIL-R1) and DR5 (TRAIL-R2) are membranebound and contain a death domain in their intracellular portion, which is able to transmit an apoptotic signal (''death receptors''). In contrast, the soluble receptor osteoprotegerin is incapable of transmitting an apoptotic signal. Similarly, two other membrane-bound receptors-DcR1 (TRAIL-R3), which lacks the complete intracellular portion and DcR2 (TRAIL-R4), which has a truncated cytoplasmic death domain-do not transmit an apoptotic signal (''decoy receptors'') [6,7].
The expression of TRAIL and its receptors has been widely described in numerous normal and cancerous tissues [8][9][10][11][12][13]. Although endometrial cancer is the most common gynecologic malignancy, to date there have been only a few studies attempting to evaluate the expression of TRAIL receptors [14,15].
The aim of our study is to assess the membrane expression of DR4, DR5, DcR1 and DcR2 in the normal endometrium (NE), atypical endometrial hyperplasia (AEH) and endometrioid adenocarcinoma (EAC) using the tissue microarray method.

Tissue collection
The study examined endometrial tissues from surgery in 197 consecutive patients who had undergone hysterectomy in the Pirogow Memorial Hospital of Lodz between 2000 and 2007. Twenty patients were treated due to uterine fibroids presented with a NE, 18 patients were diagnosed with AEH, and in 159 patients EAC was identified. The distribution of EAC patients by grade was: G1, 59 (37.1 %); G2, 82 (51.6 %); G3, 18 (11.3 %). There were 109 EAC patients at stage I (68.6 %), 24 patients at stage II (15.1 %), and 26 patients at stage III (16.3 %). The patients with EAC were subsequently treated and examined in the Regional Cancer Center, Copernicus Memorial Hospital of Lodz. Diseasefree survival (DFS) and overall survival (OS) were analyzed as functions of TRAIL receptor expression. DFS was defined as the period from primary surgery until relapse. OS was defined as the period from primary surgery until the end of the follow-up (60 months) or until the death of the patient. For the study, the approval of the Ethics Committee of the Medical University of Lodz (RNN/82/11/KE) was obtained.
The tissue blocks were fixed in formalin and embedded in paraffin. In all cases hematoxylin-eosin stained slides were available. They were reviewed by a pathologist to confirm the diagnosis of EAC. Additional histological features were recorded: histological grade and FIGO staging (The International Federation of Gynecology and Obstetrics). The patients with EAC were staged according to the FIGO 2009 Staging System [16]. The EAC grading was defined according to the 2003 World Health Organization classification criteria [17]. In each case the NE, the AEH and the EAC was diagnosed in the uterus removed during surgery (not obtained from the curettage).

Results
In EAC the membrane expression of DR4 and DR5 was less common when compared to NE (p \ 0.001; p \ 0.001) and AEH (p \ 0.001; p \ 0.001). The expression of DR4 and DR5 in NE and AEH was similar (p = 0.564; p = 0.385) (Fig. 4a, b). A strong correlation between increased TS of DR4 and DR5 in endometrial tissue from NE through AEH to EAC was present (Table 1). In EAC the TS value of DR4 and DR5 was not related to grading (Table 2) and staging (Table 3). Membrane expression of both death receptors: DR4 and DR5 was present in 11 NE (55.0 %), 9 AEH (50.1 %) and 10 cases of EAC (6.3 %). Only one receptor DR4 or DR5 expressed 5 NE (25.0 %), 8 AEH (44.4 %) and 29 cases of EAC (18.2 %). Lack of membrane expression of both the DR4 and DR5 receptors was found in 4 NE (20.0 %), 1 AEH (5.5 %) and 120 cases of EAC (75.5 %). The DR4/ DR5 receptor status correlated with the type of endometrial tissue from NE through AEH to EAC (r = -0.457; p \ 0.001). In univariate analysis, EAC with positive membrane staining of DR5 presented better DFS when compared to EAC negative for DR5 (p = 0.033; Fig. 5b). The membrane expression of DR4 was not related to the DFS (Fig. 5a). The membrane expression of DR4 and DR5 was not related to the OS (Fig. 6a, b).
The expression of DcR1 in NE was lower than in EAC, but not significantly (p = 0.055). The presence of DcR1 did not differ between AEH/NE (p = 0.735) and AEH/ EAC (p = 0.173) (Fig. 4c). DcR2 was more frequently expressed in NE than in AEH and EAC (p = 0.018; p = 0.004). The expression of DcR2 in AEH and in EAC was similar (p = 0.962) (Fig. 4d). A correlation was found between increased TS of DcR1 and DcR2 in endometrial tissue from NE to AEH and EAC (    (Fig. 5c, d) or the OS (Fig. 6c, d).

Discussion
The TMA is a piece of histological equipment designed to efficiently and economically estimate the expression of proteins across large sets of tissue specimens assembled on a single glass microscope slide [18,19]. The use of TMAs has significant advantages over standard techniques: it allows amplification of limited tissue resources by allowing the production of a large number of small core biopsies, rather than generating a single section saving time, antibodies and costs associated with analyzing multiple specimens at once [18,20]. Our previous studies confirmed that two 2-mm-sized tissue cores from biopsy of the donor block constructed into the TMAs are sufficient to obtain EAC tissue in above 90 % cases [21].
Since Wiley et al. [1] in 1995 reported the isolation of a new novel member of the TNF ligand family, designated TRAIL, many studies have reported on its significant role in cellular homeostasis. Alterations of TRAIL-induced apoptosis are suspected to be important in the development and progression of cancers, but to date, the mechanism of sensitivity and resistance for TRAIL-induced apoptosis has not been clearly explained [22]. TRAIL receptors show both cell membrane and cytoplasmic staining [9,11,12], and in some cases, the receptors can be also expressed in the nucleus [12]. Nowadays it is known that TRAIL and its receptors are present in many normal human tissues, e.g., in peripheral leukocytes, hepatocytes, neurons, renal tubuli contorti, heart myocytes, colonic luminal epithelium and crypt cells, bronchial epithelium and alveolar septa, heart myocytes, germ and Leydig cells [8][9][10]. Our study also showed membrane staining of DR4, DR5, DcR1 and DcR2 to be present in NE and AEH.
High levels of TRAIL receptor expression were described in the majority of cancer cell lines in vitro, and in many human malignant tumors in vivo [23,24]. To  characterize TRAIL receptors in cancers, cytoplasmic staining is usually analyzed rather than membrane or nuclear staining [9,12]. The cytoplasmic expression of DR4 and DR5 was found to be strong in primary and metastatic brain tumors, leukemias, malignant melanomas and cancers of the breast, lung, head and neck, esophagus, colorectum, pancreas, renal, urinary bladder, uterine cervix and ovaries [8, 9, 11-13, 25, 26]. Similarly, the cytoplasmic expression of DcR1 and DcR2 was reported, e.g., in primary and metastatic brain tumors [8] and in prostate cancer [27]. In endometrial cancer, DcR1 cytoplasmic expression was evaluated by Tarragona et al. [15]. The authors studied 62 cases of endometrial cancers by tissue microarray, and found DcR1 to be present in 98.1 % of cases [15]. Several recent studies in cancers have reported the occurrence of a correlation between the level of cytoplasmic expression of both death receptors for TRAIL, tumor grading and survival, but the data is generally unclear [12,25]. For example, Li et al. [25] report a significantly longer postoperative recurrence-free rate in patients with bladder cancer with either high DR4 or DR5 expression, than in those with low expression of both receptors identified during a 10-year follow-up. Koksal et al. [27] report correlation of high cytoplasmic DcR2 expression with high Gleason scores, prostate-specific antigen recurrence and decreased survival. In contrast to this, Koornstra et al. [9] describe a lack of correlation of DR4 and DR5 cytoplasmic expression with grading and staging in cancer of the colon, and Zhuang et al. [28] found DcR1 and DcR2 cytoplasmic expression not to be related to progression of neoplasmic disease. In endometrial cancer patients, Tarragona et al. [15] did not confirm the association of DcR1 expression with histological type of the cancer (EAC vs. non-EAC), its grade and stage, as well.
In contrast to many previous reports, the present study analyzes not cytoplasmic, but also the cell membrane expression of death and decoy receptors for TRAIL in NE, AEH and EAC. It should be noted that although cytoplasmic and nuclear staining of all the receptors was frequently observed during our examination of specimens for DR4, DR5, DcR1 and DcR2, this was not subjected to further analysis. As with other similar studies on breast and esophageal cancers [11,12] The existence of multiple receptors for TRAIL and its different cell distribution suggests that the regulation of signaling by TRAIL cytokine is of a more complex nature [3]. The observation that TRAIL selectively induces cancer cells to undergo apoptosis, while sparing normal cells, has raised great interest in using TRAIL in clinical applications as an anti-tumor weapon [2,4,7,14]. The therapeutic potential of a recombinant soluble human TRAIL has been evaluated in several human tumors, including breast, prostate, glioma and colorectal cancers [29]. Sadarangani et al. [5] describe the positive response of endometrium and endometrial cancer cell lines to TRAIL, and Llobelt et al. [30] demonstrate that endometrial cancer cells, primarily insensitive to TRAIL, can be sensitized to the ligand after administration of chemotherapeutics. These findings are in agreement with other authors using chemotherapeutics in many other forms of cancer, not only gynecologic ones [14,31,32]. However, the possibility of applying TRAIL in clinical cancer therapy should be performed with great caution due to the reported possible toxicity of TRAIL to primary human hepatocytes, especially in patients with a diseased liver [29,33,34].
In conclusion, our study demonstrates that malignant transformation of the endometrial tissue is related to the reduction of membrane expression of DR4, DR5, DcR1 and DcR2. The level of membrane staining for receptors of TRAIL in the EAC is not dependent on grading and staging of the cancer. Analysis of cell membrane expression of the receptors for TRAIL is not a good predictor of survival in EAC patients. However, the small number of patients used for this study demands further prospective studies in larger populations to confirm these results and to assess its value in clinical practice.