Type II Endometrial Cancer Overexpresses NILCO: A Preliminary Evaluation

Objective The expression of NILCO molecules (Notch, IL-1, and leptin crosstalk outcome) and the association with obesity were investigated in types I and II endometrial cancer (EmCa). Additionally, the involvement of NILCO in leptin-induced invasiveness of EmCa cells was investigated. Methods The expression of NILCO mRNAs and proteins were analyzed in EmCa from African-American (n = 29) and Chinese patients (tissue array, n = 120 cases). The role of NILCO in leptin-induced invasion of Ishikawa and An3ca EmCa cells was investigated using Notch, IL-1, and leptin signaling inhibitors. Results NILCO molecules were expressed higher in type II EmCa, regardless of ethnic background or obesity status of patients. NILCO proteins were mainly localized in the cellular membrane and cytoplasm of type II EmCa. Additionally, EmCa from obese African-American patients showed higher levels of NILCO molecules than EmCa from lean patients. Notably, leptin-induced EmCa cell invasion was abrogated by NILCO inhibitors. Conclusion Type II EmCa expressed higher NILCO molecules, which may suggest it is involved in the progression of the more aggressive EmCa phenotype. Obesity was associated with higher expression of NILCO molecules in EmCa. Leptin-induced cell invasion was dependent on NILCO. Hence, NILCO might be involved in tumor progression and could represent a new target/biomarker for type II EmCa.


Background
Endometrial cancer (EmCa) is the most common gynecological malignancy of the female reproductive tract [1]. As of 2015, there were 54,870 new EmCa cases reported and 10,170 deaths reported in the United States [2]. The incidence of EmCa is higher in well-developed countries and countries with high obesity rates [3]. Caucasian patients are at a higher risk of developing endometrial cancer when compared to African-American patients in the US. In 2014, EmCa incidence rate of Caucasian patients was 24.8 per 100,000 women, whereas in African-American women it was 20.9 per 100,000 women [3]. There are still controversial opinions on the categorical classification of types I and II EmCa. However, type I EmCa is estrogen dependent. In contrast, type II EmCa is estrogen independent, more aggressive, shows poor prognosis, and is usually associated with endometrial atrophy [1,4].
Obesity, characterized as having a body mass index (BMI) of thirty or higher, is a major risk factor for EmCa and is a comorbid condition found in approximately 17-46% of all EmCa cases [5]. The heaviest women are at the highest risk of EmCa [6]. Studies have shown that overweight women were twice as likely to develop the disease compared to normal-weight women, while obese women are four times more likely to develop EmCa [7]. Interestingly, African-American women show the highest incidence of obesity and type II EmCa and are more likely to die from this disease. However, whether obesity is a driver for EmCa is not very well understood [1,8]. The reason(s) for this cancer health disparity is unknown [8]. Obese individuals have high serum levels of leptin, an adipokine strongly linked to poor prognosis and higher incidence of several cancer types, including breast, colon, pancreas, stomach, and thyroid cancer among others [9]. High circulating levels of leptin in obese people correlate to the size of adipose tissue and generally to BMI [10]. The main function of leptin is the regulation of appetite and energy balance. Leptin exerts its effects on energy balance through specific signaling pathways in hypothalamic neurons that express the leptin receptor, OB-R [10]. However, obesity is characterized by hypothalamic unresponsiveness to leptin signals, which is known as leptin resistance [10].
Leptin regulates glucose homeostasis, growth, reproduction, and immune response [11]. Leptin's pleiotropic functions also involve angiogenic, inflammatory, and antiapoptotic effects, which are important for cells overexpressing OB-R, including cancer cells [12]. Moreover, several reports have shown a link between leptin signaling and the development of cancer stem cells and resistance to chemotherapeutics. Therefore, leptin is a growth, angiogenic, and survival factor for several types of tumors [12].
Leptin is secreted by adipocytes as well as cancer cells [13]. Therefore, leptin produced by adipocytes and cancer cells could act in an autocrine and paracrine manner to promote proliferation, migration, survival, invasion and proinflammatory processes in tumor cells, and tumor angiogenesis [14]. Accumulating evidence strongly suggest that high levels of leptin and OB-R found in tumor tissues are associated with metastasis and decreased survival rates of breast cancer patients [10,14,15]. OB-R has several isoforms. OB-Rl or OB-Rb is the long isoform with full signaling capabilities. OB-Ra, the short form of the receptor, has limited signaling capabilities and has been found in EmCa cells [10,16].
Notch signaling is a hallmark of several cancers. Aberrant activation of the Notch signaling pathway can be initiated through the abnormal expression of Notch ligands, receptors, and target genes; all of which have been reported in several solid tumors, including breast, prostate, and pancreatic tumors. However, the involvement of specific Notch receptors in cancer is dependent on cell type [17]. Similarly, IL-1 is an inflammatory and proangiogenic cytokine that can promote tumor angiogenesis, growth, and metastasis [18]. IL-1 and leptin induce VEGF in cancer [19]. In turn, VEGF can regulate the expression of Notch signaling components [20]. Providing a feedback mechanism, Notch signaling in turn can alter expression levels of all three VEGF receptors [14]. Leptin also induces VEGFR in cancer cells [21] and promotes angiogenic features in endothelial cells via upregulation/transactivation of VEGFR and downstream expression/activation of Notch. Thus, high levels of leptin found in overweight and obese patients might lead to increased angiogenesis by activating VEGFR-Notch signaling crosstalk in endothelial cells [22].
It was previously reported that leptin signaling induced a complex molecular crosstalk in breast cancer that involved Notch and IL-1 signaling (NILCO). Interestingly, the blockade of the IL-1 system, in breast cancer, inhibited leptin's effects on Notch expression, suggesting that leptin-induced Notch activation is reliant on IL-1 signaling [14]. NILCO is a driver of breast cancer cell proliferation, migration, and survival. Moreover, NILCO induces essential proangiogenic molecules (i.e., VEGF/VEGFR-2) and Notch targeted genes (i.e., Survivin and Hey2, and so forth) [14,15,23]. Data generated from obese mice hosting syngeneic mammary tumors showed leptin signaling induces Notch in vivo, which was linked to tumor growth and angiogenesis. These studies further suggest that obesity accelerates breast cancer progression via leptin induction of NILCO [15].
Likewise, EmCa cells express IL-1, leptin/OB-R, and Notch [24][25][26]. In addition, leptin increased the levels of VEGF in EmCa. However, it is unknown whether the expression of NILCO components is associated with EmCa progression, aggressiveness, and prognosis. The main objectives of the current study were to determine whether NILCO components are differentially expressed in the more aggressive type II EmCa and whether obesity may influence this outcome. Current data suggest that NILCO components are expressed higher in type II EmCa, and their level of expression correlates to obesity status. These data suggest that NILCO may be involved in EmCa progression and prognosis. Moreover, current findings suggest that NILCO might be used as a novel biomarker for type II EmCa.

Patients and Specimens.
The endometrial samples were obtained from Grady Memorial Hospital, Atlanta, GA, under IRB approved protocols (Grady Memorial Hospital and Morehouse School of Medicine, Atlanta, GA) and a patient's written informed consent. Nonmalignant and EmCa samples were obtained from African-American women (n = 29; 17 obese and 12 lean) undergoing uterine surgery. Tumor samples were staged and graded by a pathologist and consist of type I EmCa (n = 12; 7 obese and 5 lean), all endometrioid adenocarcinomas; and type II EmCa (n = 17; 10 obese and 7 lean) were papillary adenocarcinoma (n = 4), papillary serous adenocarcinoma (n = 8), carcinosarcoma (n = 1), and mixed clear cell and serous adenocarcinoma (n = 4). Each EmCa sample had a paired control sample from adjacent nontumor endometrial tissue, as determined by pathologists. Additionally, EmCa tissue arrays from Chinese patients (EmCa, n = 120, and nonmalignant endometria, n = 30) were obtained from US Biomax, Inc. (Rockville, MD). Tissue cores were fixed in formalin. EmCa features from biopsies obtained from Chinese patients included age, TNM, and tumor grading. However, no body weight or body mass index (BMI) information was available for Chinese patients. It was also unknown if some nonmalignant samples were obtained from adjacent normal endometria from the Chinese EmCa patients. Each endometrial tissue array from Chinese donors contained type I EmCa (n = 97) and type II EmCa (n = 23) and nonmalignant endometrial tissues: normal (n = 23) and hyperplasia (n = 7) samples that were classified by a pathologist. A total of 120 cores were scored using the HSCORE method described below.

Western Blot.
Frozen tissue sections from African-American patients were used for Western blot (WB) analyses. All available frozen sections for western blot (n = 8, type I and n = 6, type II) were pooled to eliminate bias during the immunoblot analyses. All samples used in WB were represented in the original IHC cohort. Fifty mg of EmCa and nonmalignant tissues were lysed and extracted using radioimmunoprecipitation assay (RIPA) buffer containing a protease/phosphatase inhibitor cocktail (Sigma). The Bradford Assay was used to determine the protein concentrations of tissue lysates. Fifty μg of protein was used for WB. The same antibodies used in immunohistochemistry analysis were used for WB, and a 1 : 200 dilution of primary antibody was used. Specific positive controls (Santa Cruz Biotech) for each primary antibody were used for more accurate identification of specific antigens. WB results were normalized using β-actin as loading control. The NIH Image program (Image J) was used for quantitative analysis of protein bands. Band density values from type II EmCa were calculated assuming 100% of expression in type I EmCa. Representative data were derived from biological triplicates (mean+standard error).
2.5. ELISA. Type I and type II EmCa tissues from African-American patients were mechanically disrupted with RIPA buffer to produce tumors lysates. Human ELISA kits (R&D Systems) were used to determine leptin and IL-1β concentrations in tumor lysates per the manufacturer's instructions. RIPA buffer was used as control. ELISA sensitivities were 7.8pg/mL for leptin and 1pg/mL for IL-1β.
2.8. Statistical Analysis. Statistical significance was established with the STATA statistical software package using ANOVA, two-tailed Student's t-test, and a Chi-square test or Fisher's exact test. p values were 2-sided, and a p value of less than 0.05 was considered statistically significant. Table 1 summarizes the clinicopathological and histological characteristics of EmCa from African-American and Chinese patients. The cohort of African-American women was recruited from those patients currently attending to the Gynecological Medical services in the Grady Memorial Hospital, but not selection criteria was applied. It is unknown how the cohort of Chinese patients was selected as endometrial biopsies that were supplied by a commercial source (US Biomax).

EmCa Tissue Biopsies and Tissue Arrays.
Endometrial biopsies from African-American patients showing histopathological criteria of type I EmCa (n = 12) and type II EmCa (n = 17) were further classified using FIGO system and grading [29,30]. Grades and stages showed significant differences between type I and type II EmCa (p = 0 0010 and 0.0051, resp.). As expected, type I EmCa was low grade (grades 1 and 2), while type II were high grade (grade 3). Likewise, roughly 75% of type I EmCa were low stage (stage I), but 65% of type II were high stage (stages III and IV). African-American patients with type I and type II EmCa showed significant differences between age (p = 0 0076), with type II EmCa patients showing a higher mean age. Noticeably, the mean BMI between patients having type I or type II EmCa were significantly different, with type II EmCa showing higher BMI (type I: 29.81 and type II: 34.17).
The histopathological characteristics of EmCa from Chinese patients are also shown in Table 1. All endometrial biopsies were evaluated and classified as type I (n = 97) and type II (n = 23) EmCa. Similarly, as the case of EmCa from African-American patients, type I and type II EmCa from Chinese patients showed significant differences in tumor grade and stage (p < 0 001). Most type I EmCa from Chinese patients were grade 1 (45%), whereas type II EmCa were mainly grade 3 (52%). In contrast to the EmCa African-American patients, mean ages of Chinese patients were not significantly different when compared to type I and type II EmCa. Figure 1(a) shows representative images of IHC staining of NILCO antigens examined in type I and type II EmCa from <0.001

Western Blot (WB) Analyses.
Next, WB analysis was used to further determine whether NILCO and targets were expressed differently in type I and type II EmCa from African-American patients (Figure 4(a)). Type II EmCa from African-American patients showed significantly higher Interestingly, OB-R protein levels detected by WB were virtually absent in type I EmCa that confirmed IHC results (OB-R HSCORE = 1.1 that indicates very weak detection of the antigen).

mRNA Levels. mRNA levels of NILCO in EmCa from
African-American patients (Figure 4(b)) exhibit a similar pattern of expression as observed in WB analyses. Figure (4b) shows higher mRNA expression of Notch1, Notch4, DLL4, OB-R, and IL-1R tI in type II EmCa (p < 0 01). Notch1 and JAG1 mRNA showed 2-fold higher expression in type II EmCa. Interestingly, DLL4 and Hey2 mRNA both had a 20-fold change difference in expression between the two tumor types. In addition, OB-R mRNA was 3-fold higher in type II EmCa.
3.5. ELISA Results. No significant differences were found for leptin levels in tumor lysates from African-American patients. However, type II EmCa tends to have higher concentrations of the oncogenic adipokine compared to type I EmCa lysates (Figure 5(a)). Similar results were found for IL-1β levels ( Figure 5(b)).

Effects of NILCO Inhibition on Leptin-Induced EmCa
Cancer Cell Invasion. To examine whether a functional NILCO signaling pathway occurs in EmCa, we determined whether leptin-induced cell invasion is affected by the inhibition of Notch and IL-1 signaling in vitro (DAPT; a γ-secretase inhibitor and anti-IL-1Rt I antibody). Additionally, the specificity of leptin's effects on cell invasion was assessed using a leptin receptor inhibitor (IONP-LPrA2). Figure 6 shows that leptin significantly induced invasion in Ishikawa (Figure 6(a)) and An3ca cells (Figure 6(b)). Remarkably, leptin's effects were prominent in the more aggressive An3CA EmCa cell line that was derived from a type II EmCa [28]. Notably, the effects of leptin were abrogated by the inhibition of Notch and IL-1 signaling.

Discussion
Current data illustrate that NILCO molecules are expressed differently in EmCa types. Notably, NILCO proteins and mRNAs were expressed higher in type II EmCa, which strongly suggests that NILCO expression correlates with the development of the more aggressive form of the disease. Higher expression of NILCO was consistently found in type II EmCa biopsies from African-American and Chinese patients. Type II EmCa showed the highest expression of Notch1 and Notch4. Although, a small number of African-American EmCa patients were evaluated, obese African-American patients (n = 17) showed the highest levels of NILCO expression when compared to lean EmCa patients (n = 12). Unfortunately, lack of data on BMI and body weight made it impossible to elaborate on the potential impact of obesity on NILCO expression in EmCa from Chinese patients. These patient cohorts were different than African-American cohort, showing an unequal proportion of type I and type II EmCa that makes it difficult to establish an adequate comparison. However, it was observed that NILCO mean expression was higher in African-American women than in Chinese women suffering from EmCa. A strong correlation between obesity and EmCa incidence has been consistently reported. EmCa is more than three times as common in obese women when compared to normal healthy weight women [31]. After menopause, progesterone synthesis is drastically reduced. However, in postmenopausal women, the adipose tissue can increase estrogen levels through the aromatization of the androgens. Obesity is characterized by high levels of leptin and estrogen, which increases cancer growth. Leptin was positively associated with EmCa [32]. Then, increased EmCa incidence and progression are related to elevated levels of estrogens (unopposed estrogen stimulus), insulin growth factor-1, adipokines (leptin, resistin), and cytokines [1]. Estrogen may influence leptin synthesis in a tissue-and cell type-specific fashion [33]. Moreover, OB-R and estrogen receptor (ER) are coexpressed in cancer indicating a possible interaction between leptin and estrogen systems to promote carcinogenesis. It was recently reported that OB-R was downregulated during the progression of EmCa, which correlated with ER and progesterone receptor expression patterns [34]. Thus, leptin/OB-R and estrogen/ ER signaling pathways may work together to accelerate the progression of type I EmCa. However, in contrast to a previous report [34], we found that type II EmCa expresses higher levels of OB-R. Remarkably, type II EmCa shares genomic features with basal-like breast cancer [8], which shows low levels or absence of ER, progesterone, and EGFR2 (Her2) receptors. Basal-like breast carcinomas (triple negative) are very aggressive and have poor prognosis. Aberrant Notch signaling has been strongly associated with triple negative breast cancer. Moreover, Notch signaling is a target linked to poor prognosis in breast cancer [14]. Leptin upregulates the IL-1 system in endometrial cancer cells [25] and the Notch pathway in breast cancer [14]. A crosstalk between Notch, IL-1, and leptin signaling (NILCO) has been associated with breast cancer growth [14,15,23]. Type II EmCa are independent of ER signaling, very aggressive, have poor prognosis, and no targeted therapies. Therefore, NILCO may be related to type II EmCa progression [1].
It has been suggested that Notch signaling could be important for endometrial proliferation [24]. Previous studies have shown that Notch2 signaling could play a protective role in breast cancer [35]. However, the role of Notch signaling is poorly understood in EmCa [36]. Moreover, previous reports showed inconsistent endometrial expression of Notch. Increased protein expression of Notch1, Notch3, DLL4, and JAG1 was previously reported in EmCa [37]. Though, a different study reported low Notch4 expression in EmCa [38]. However, there no previous reports comparing the expression of Notch and other NILCO molecules between type I and type II EmCa. Present data show that NILCO molecules (Notch1, Notch2, Notch3, Notch4, DLL4, and JAG1) and targets (Survivin, Hey2, IL-1R tI, and OB-R) were expressed in EmCa regardless of ethnicity and cancer type.
Notch1 and Notch4 were expressed significantly higher in type II EmCa from both African-American and Chinese patients when compared to type I EmCa. Additionally, Notch ligands DLL4 and JAG1 were also expressed higher in all type II EmCa patients. Furthermore, OB-R and IL-1R tI were expressed higher in type II EmCa from African-American patients. qPCR analyses corroborated IHC and WB data on type II EmCa. Although, no significant differences were found, leptin and IL-1β levels within tumor tissues were higher in type II EmCa than in type I EmCa. Moreover, leptin-induced invasion of EmCa cells in vitro was remarkably higher in the more aggressive cell line An3Ca (type II EmCA) [28], which was abrogated via inhibition of Notch and IL-1 signaling. Taken together, these data suggest that functional NILCO signaling might be involved in type II EmCa progression.
Notch ligands are mainly found membrane bound as the signaling cascade is initiated upon binding of ligands to transmembrane Notch receptors expressed in adjacent cells [17]. Interestingly, NILCO molecules showed differential localization patterns in type II versus type I EmCa. Several NILCO molecules (Notch1, Notch2, JAG1, DLL4, OB-R, and IL-1R tI) were found to be expressed higher in the cellular membrane and cytoplasmatic compartment in type II EmCa from African-American patients. Additionally, type II EmCa biopsies showed higher expression of nuclear Notch4 and Survivin in these patients. Meanwhile, type II EmCa from Chinese patients showed higher cytoplasmatic levels of Notch4, DLL4, JAG1, and IL-1R tI. To the best of our knowledge, there are no reports on the cellular localization patterns of NILCO molecules in EmCa. However, the implications of the current findings warrant further investigations.
Biomarkers are promising tools for early detection and monitoring the progression of EmCa [39]. Among these biomarkers, Ki67 proliferation marker is increased in high-grade tumors. Also, abnormal activation of β-catenin has been observed in type I EmCa [40,41]. Serum and plasma biomarkers (i.e., CA125) have been explored in EmCa development and progression, but no specific tumor markers have been investigated for obesity-related type II EmCa [42]. Current data may suggest a potential for NILCO expression as a novel biomarker for type II EmCa. Higher NILCO expression in EmCa might indicate the more aggressive disease and a potential new target for treatment.
Data from the present study should be interpreted considering some limitations: (1) We did not establish selection criteria for the patient cohort to collect EmCa samples from African-American patients. Rather, the EmCa samples were obtained from current patient population attending to the Gynecological Medical services at Grady Hospital; (2) We were only able to collect a small number of EmCa samples from African-American women but not from Caucasian women since the patients attending Grady Hospital are predominantly African-American; and (3) The cohort of Chinese patients suffering from EmCa was obtained from a company that provided commercial tissue microarrays and does not fully specify on the clinicopathological characteristics of patients, that is, BMI and body weight. Despite these study limitations, we could present the first preliminary evaluation of NILCO expression in type I and type II EmCa.

Conclusion
Although, there are some controversial opinions on how to clearly define type I and II EmCa, it is generally recognized that type II EmCa comprises a phenotype of malignancy characterized by the lack of hormonal response, high aggressiveness, and poor prognosis. EmCa incidence and progression are conclusively affected by obesity. Currently, there are no consistent data reported on the expression of Notch in type I and type II EmCa. NILCO components are expressed higher in breast cancer and other cancer types [41]. Present data strongly support the notion that type II EmCa expresses higher NILCO molecules, which suggests that the more aggressive and nonresponsive hormonal disease could be dependent on leptin-Notch-IL-1 signaling. Furthermore, current data indicate that obesity induces higher NILCO expression in EmCa that might be linked to leptin signaling. Additionally, inhibition of Notch and IL-1 signaling in vitro significantly reduced leptin-induced invasion of EmCa cells. Therefore, leptin could induce the more malignant phenotype via promoting the expression and crosstalk between Notch and IL-1 in EmCa. Current observations deserve further investigations including higher number of samples from different ethnic groups of lean and obese patients. These investigations might lead to the identification of NILCO as a novel target and biomarker in EmCa, particularly for type II EmCa.

Conflicts of Interest
The authors declare to have no competing interest.

Authors' Contributions
Ruben Rene Gonzalez-Perez and Danielle Daley-Brown were involved in the study design and manuscript preparation. Danielle Daley-Brown was also involved in cell cultures, invasion assays, HSCORE determinations, and IHC, WB, and PCR analyses of NILCO and targets. Gabriela Oprea-Iles performed HSCORE, IHC, and pathological analyses of tissue samples. Alexander Quarshie performed the statistical analysis and interpretation of data. Kiara T. Vann was involved in IHC and HSCORE determinations. Viola Lanier was involved in real-time PCR analyses. Pierre V. Candelaria performed ELISA analyses and interpreted data. Regina Lee and Roland Pattillo were involved in the collection of patient samples and data interpretation.