Does Metformin affect ER, PR, IGF-1R, β-catenin and PAX-2 expression in women with diabetes mellitus and endometrial cancer?

ER

In our material, although we observed decreasing expression of ER parallel with decreasing histological differentiation of cancer cells, we did not find statistically significant difference in ER expression according to the grade of EC (G1 vs G2 p = 0.99; G2 vs G3 p = 0.14; G1 vs G3 p = 0.13). No correlation with FIGO staging was observed. No difference in the positive ER rate was found when comparing diabetic and non-diabetic subgroups of the EC patients (p = 0.18). However, it was shown that DM2 women with EC receiving insulin in monotherapy had a significantly higher ER expression than non-diabetic women with EC (p = 0.0046).

Moreover, patients treated with metformin demonstrated a statistically significant reduction in ER expression in comparison to the group receiving insulin in monotherapy (p = 0.014 for all patients treated with metformin; p = 0.004 for metformin in monotherapy) (Figure 5).

PR

Comparing EC of low and high grade, a reduced PR expression was found in samples with poorly differentiated cells (G1 vs G3 p = 0.029; G1 vs G2 and G2 vs G3 p > 0.05). However, no differences were observed in PR expression in relation to FIGO staging, the presence of diabetes, and the method of its treatment (p > 0.05).

IGF-1R

No difference in the rate of positively stained cells was found in the EC group in relation to grading and FIGO staging.

Cancer patients with diabetes showed a significantly higher IGF-1R expression then the non-diabetic group (p = 0.0012). This trend is present in groups of women treated with metformin (p = 0.0056) or other drugs (p = 0.01). Comparing different types of anti-diabetic treatments in EC patients, we did not find any significant changes in the expression of IGF-1R (p > 0.05).

β-Catenin

All cases under study, excluding one (a case of FIGO IIIA G3 endometrioid cancer in a non-diabetic patient with a strong nuclear reaction) demonstrated a dominant, membranous reaction of β-catenin.

The group of diabetic EC patients did not show a difference in membranous protein expression in comparison to the control groups (p = 0.32). No differences were demonstrated in relation to cancer grading or clinical advancement of the disease according to FIGO (p > 0.05). However, although not statistically significant, patients on metformin monotherapy displayed a lower protein expression than those on insulin monotherapy (p = 0.07).

PAX-2

The intensity of the nuclear reaction for PAX-2 was not statistically varied between DM2 EC patients and control groups (p = 0.53). No differences were observed between FIGO staging and grading, and the type of anti-diabetic drug administered (p > 0.05).

ER, PR, IGF-1R

The increased risk of developing EC in women with DM2 is an indisputable fact [23]. There is also clear evidence showing a positive association between the increased levels of circulating insulin and the incidence of endometrioid adenocarcinoma [24, 25]. But despite numerous studies, there is still no unanimous data explaining the role of IGF-1 on endometrial carcinogenesis. Additionally, there is also a lack of consensus on the prevailing type of receptor (IR or IGF-1R) acting on EC promotion.

Endometrial cancer type I is a hormone dependent malignant disease, in which the balance between estrogen and progesterone is disrupted and an increase of estrogen stimulation leads to excessive cell proliferation.

High ER and PR expression in EC are frequently connected with endometrioid type of neoplasm, better cells differentiation, lower risk of lymph node metastases and better prognosis [26, 27]. Based on studies of endometrial and breast cancer tissue, we can observe a profound and complex crosslink between estrogen, progesterone, insulin or IGF-1 and its receptors.

In the normal endometrium, especially during the proliferative phase, estrogen binding to ER receptor acts as a transcription factor triggering local IGF-1 production [28]. On the other hand, endometrial stromal cells, due to progesterone induction, produce IGFBP-1 (IGF-1 binding protein), which after binding to IGF-1 reduces its bioavailable fraction. Mitogenic function of IGF-1 may occur through different mediators, the dominant pathway being PI3K/Akt, but also MAPK, both involved in regulation of cell proliferation and apoptosis. Furthermore, active MAPK signaling can phosphorylate serine in NH2-terminal region of the ER- AF1, and thereby increase ER activation [29]. Estrogen not only influences IGF-1 production but also takes part in the regulation of IGF-1R, elevating its endometrial expression [30]. The inhibition of IGF-1R promotor may occur via e.g. p53 [31], but its mutation is often seen in EC leading to IGF-1R overexpression.

High IGF-1 plasma concentrations and the presence of IGF-1R on the cell surface are found in many cancers [32, 33], but in the case of EC, the role of circulating IGF-1 is controversial. It was noted that high fasting glucose levels in women not on hormone therapy was correlated with the development of endometrioid adenocarcinoma, whereas the rise in free IGF-1 plasma fraction, has no effect or even decreases the risk of developing EC [24, 34]. These observations may suggest the dominant role of the local production of IGF-1 in the neoplastic endometrium. Most likely, the rise of local free IGF-1 in diabetic patients is due to insulin mediated inhibition of the production of IGFBP-1 [35].

In EC tissue, IGF-1 can influence PR expression. Its high levels are associated with improved prognosis and response to gestagen treatment, especially in advanced or recurrent carcinoma [36]. Xie et al. [37] observed IGF-1 and IGF-2-induced reduction of PR in EC cell lines, which was related to the activation of the PI3K/Akt/mTOR pathway and phosphorylation of the p70S6K effector protein. In an in vitro study, metformin was found to inhibit the growth of ECC-1 and Ishikawa EC cells in a dose dependent manner via activation of AMPK and inhibition of mTOR [9]. It was noted that the administration of this drug can raise PR expression in EC [37]. Similar results were presented by Berstein et al. [38] in 90 breast cancer samples from patients with DM2. In immunohistochemical assessment of ER and PR researchers found no difference in the ER expression of cancer cell in women receiving metformin, insulin, sulfonylurea derivatives and those who were exclusively on a diabetic diet. However, an increased percentage of positive PR in breast cancer specimens was found in patients treated with metformin mono- or polytherapy (p = 0.04).

Our study does not show any difference between PR expression in patients receiving different types of pharmacotherapy in DM2. However, we found a reduction in the rate in which cells displayed a strong ER reaction in EC patients receiving metformin in comparison to those patients on insulin monotherapy. It is believed that metformin may decrease estrogen concentration in neoplastic tissue via local inhibition of aromatase activity suppressing synthesis of the enzyme through interaction with its promoter, PII [39]. But the exact mechanism linking metformin uptake with ER reduction is unknown. We can speculate that the reduction of ER after metformin treatment may decrease the number of cells sensitive to high levels of estrogens, affecting their proliferative abilities and at the same time may influence the prognosis. But further studies are needed to confirm this hypothesis.

Pengchong H and Tao H. [40] showed a greater IGF-1R expression in EC than in normal endometrium and indicated a correlation between IGF-1R overexpression and metastasis to the lymph nodes. Roy et al. [41] in turn found no statistically significant differences between the amount of mRNA IGF-1R in normal cells and EC cells. Interestingly, there was a significant reduction in IGF-1 expression in EC when comparing the endometrium at proliferative (p = 0.0078) and early secretory (p = 0.0002) phases. This disproportion between a relatively high IGF-1R and low IGF-1 in EC cells may reflect intense disruptions in IGF-1R encoding gene expression, leading to protein overproduction (e.g. mutations in suppressor genes i.e. p53 or BRCA1 resulting in the lack of inhibitory effect on the IGF-1R promotor).

In our material women with EC and diabetes demonstrated a significantly higher IGF-1R expression in comparison to the non-diabetic women. This may be associated with e.g. different affinity to specific receptors (IR- α and β, IGF-1R, hybrid receptors) of insulin, and the circulating IGF-1 and IGF-2 or their local production in neoplastic tissue. Metformin administration did not significantly influence the number of IGF-1R in EC. Presumably, the drug inhibits further stages of the PI3K/Akt/mTOR pathway initiated by IGF-1R, without direct influence on IGF-1R membrane expression. In future research it is valid to compare the expression and activity not only of IGF-1R but also of IR isoforms in normal and neoplastic endometrium of diabetic and non-diabetic patients and to assess the influence of metformin on their expression.

Beta-catenin

β-Catenin (CTNNB1) plays a double role in cells –first, it regulates cellular adhesion via interaction with membrane protein E-cadherin – second, in the cell nucleus it plays a key role in the Wnt signaling pathway.

The normal endometrium displays a strong membranous reaction of β-catenin. However, a reduction in β-catenin and E-cadherin expression is noted in different types of neoplastic tissue, including endometrial tissue. Seagusa et al. [42] revealed a gradual decrease in cell membrane β-catenin immunoreactivity from normal to atypical hyperplasia and to grade 3 carcinomas. Disruption in both β-catenin and E-cadherin expression is associated with reduced cell to cell adhesion, the advancement of the clinical stage of the disease and the increased risk of metastasis [15, 16].

The Wnt signaling pathway is a multi-protein complex consisting of axins, β-catenin, APC (adenomatous polyposi coli), kinase CKI (casein kinase I) and GSK3β (glycogen synthase 3β). In the absence of Wnt ligands, the APC/GSK3β/Axin complex is activated which results in phosphorylation of β-catenin and its ubiquitination and proteasomal inactivation. Once the Wnt pathway is activated, APC/GSK3β/Axin does not form a complex with β-catenin and phosphorylation of this protein is inhibited. Stabilized β-catenin is translocated to the nucleus and is assembled into a complex with transcription factors LEF (lymphoid enhancer factor) and TCF (T-cell factor), promoting expression of many genes involving in cell proliferation (Myc or cyclin D1), angiogenesis (VEGF), adhesion and apoptosis (survivin).

Activation of β-catenin/Wnt pathway is a physiological phenomenon during the menstrual cycle, but it differs among cycle phases. Nei et al. [43] showed the presence of nuclear β-catenin in the proliferative phase, whereas during the secretory phase the protein was found mainly in the cytoplasm and the cell membrane.

Estradiol can change the expression of the Wnt ligands including Wnt 4 and 7A [43, 44]. Furthermore, it is also linked to Wnt and PI3k/Akt signaling pathways. It was found that after estrogen stimulation, ER alpha induced PI3K and subsequently activated Akt. This in turn resulted in phosphorylation and inhibition of GSK3 [45]. Consequently, β-catenin cannot be phosphorylated and degraded in proteasomes.

A number of studies indicate disruptions in β-catenin expression in EC, most frequently in the endometrioid type. Mutations in CTNNB1, but also in other genes encoding proteins taking part in Wnt pathway (APC or axin), result in stabilization of β-catenin, its excessive nuclear accumulation and promotion of many genes which lead to neoplastic transformation [46]. Konopka et al. [47] found mutations in the CTNNB1 region in 16.1% of ECs. The mutations detected at the atypical hyperplastic endometrium and the early stages of EC suggest their significant role in early carcinogenesis [44].

Limited literature concerning β-catenin expression in diabetic EC patients makes it impossible to compare our findings to other studies. In our material concerning EC, β-catenin nuclear staining was found in 13.9% of cases which corresponds accordingly to the Nout et al. research [48] of 14%. In our findings the majority of cases concerned non-diabetic patients with EC (21.5% vs 8.3% of patients with diabetes) (Table 3). Due to a small number of diabetic patients with a positive nuclear reaction, it is difficult to assess the influence of anti-diabetic treatment on nuclear accumulation of β-catenin. However, if metformin reduces the expression of ER, which was demonstrated in our study, it is presumed that it may also decrease the activation of PI3K/Akt signaling, increasing the unphosphorylated fraction of GSK3 and reducing the amount of β-catenin. Further studies are necessary to examine the correlation between β-catenin expression, Wnt pathway activation and diabetes in women with EC.

PAX-2

PAX-2 participates in regulating the proper development of the central nervous system, the kidneys and the Müllerian ducts (upper part of the vagina, uterus, fallopian tubes). Even more evidence indicates that it also has a significant role in oncogenesis, including EC.

Monte et al. [20] described the loss of PTEN and PAX-2 expression in normal, hyperplastic, and cancer cells, indicating that independent from PTEN, PAX-2 acts as a suppressor gene undergoing inactivation during cancer transformation. In the normal endometrial tissue, versus precancerous lesions and cancer, the level of the PAX-2 protein loss increases progressively at the rate of 36%, 71%, and 77% respectively. Similar results were obtained by Allison et al. [49], which suggests that PAX-2 protein losses occur at an early stage of carcinogenesis. Unfortunately the mechanism explaining this phenomenon is yet unknown.

Researchers do not know of any study regarding PAX-2 expression in EC in relation to coexisting glucose tolerance disorders.

In our material, no difference in PAX-2 nuclear expression was found in patients with EC in relation to the presence of diabetes or the type of treatment administered. However, these results may be considered as questionable, because of the strong staining of the cytoplasm which may have deterred the assessment of the nuclear reaction. Further research is needed in order to determine if there is a relationship between PAX-2 expression and diabetes in patients with EC, if proven so, then to additionally determine the influence of metformin administration.

Consent

Written informed consent was obtained from the patient for the publication of this report and any accompanying images.