Estrogen as Jekyll and Hyde: regulation of cell death

Sustained estrogenic exposure increases the risk and/or the progression of various cancers, including those of the breast, endometrium and ovary. Unexpectedly, physiological level of estrogen together with a novel IKKα inhibitor BAY11-7082 could effectively induce cell apoptosis in ER-positive breast cancer cells, suggesting combining estrogen with IKKα inhibition may be beneficial for breast cancer patients. This opinion article touches upon the dual role estrogen played in inducing cancer cell death and asks whether use of estrogen in combination with IKKα-targeted therapy would be possible reconsider the newly identified crosstalk between ER and NFκB pathway which can be utilized to switch the effects of estrogen on cell death.

Our research projects currently focus on the understanding of the interplay between different signaling cascades and estrogen receptor alpha (ER) dependent transcription activation in breast cancer. ER is a master transcription factor frequently elevated in breast cancer patients. ER-positive breast cancers usually occur in up to two thirds of newly diagnosed breast cancer patients. Drugs (such as the antiestrogens tamoxifen and raloxifene) have been very effective in treating ER-positive breast cancers and have saved millions of lives. However, this standard treatment does not work for all ER-positive breast cancer patients. Therefore, we enthusiastically review recent research breakthroughs in the context of our ever-developing understanding on ER function. These new mechanistic insights on the crosstalk between ER and NFκB pathway will lay the ground work for improved clinical approaches to the treatment of this type of breast cancer.
For many years, those of us working in the field were used to looking at estrogen as a mitogen through its genomic function mediated by ER-dependent transcription program. Now we realize that estrogen through ER, in addition to regulating gene expression, crosstalks with many non-genomic signaling pathways involved in cell growth, differentiation and apoptosis. These newly identified interplays may give a different flavor to long conceived mitogenic role of estrogen. In this opinion article, we generally reviewed the changing attitude toward the use of estrogen in the clinic, with a focus on newly discovered pro-apoptotic roles of estrogen when combined with IKKα inhibitors. We also discussed the possibility of estrogen and IKKα inhibitor dual-therapy in ER-positive breast cancer treatment.

Historical perspectives: changing attitudes toward the use of estrogen in the clinic
Sir George Thomas Beatson (1896) use oophorectomy to reduce estrogen level in premenopausal women in order to prevent breast cancer occurrence. It was for the first time to reveal the mitogenic role of estrogen at physiological concentration 1  ,808 generally healthy women aged 50-79 to evaluate the beneficial effects of postmenopausal hormone replacement therapy on heart diseases, bone fractures, and cancers 4 . Due to the increased incidence of breast cancer, stroke, and cardiovascular complications in women treated with estrogen alone or with a combination of estrogen and progesterone, the study was terminated prematurely in 2002. Though extensively studied, the definite understanding of the mechanism of estrogen action always challenges our mind.

The challenge: paradoxical role of estrogen in cell death
Estrogen regulates the proliferation and development of tissues expressing estrogen receptors and ER is mainly expressed in breast epithelium, ovary and endometrium. Thus, estrogen is mitogenic for cultured ER positive breast cancer lines. The mitogenic effects of estrogen at the G1-to-S transition are mediated by the key effectors of estrogen action, c-Myc, cyclin D1 and E2F-1 5-7 . c-Myc expression occurs within 15 min of estrogen stimulation, among the earliest responses to estrogen. Estrogen also rapidly induces cyclin D1 expression. In the G1 phase, estrogen drives E2F-1 expression. Nongenomically, Estrogen binding to the ER stimulates rapid activation of Src and signaling pathways MAPK and PI3K/Akt pathways that affect cell survival 8,9 . Based on these understandings of estrogen action, ER protein is assayed in newly diagnosed breast cancers because it is a clinically useful prognostic factor and predicts responsiveness to ER blocking drugs such as Tamoxifen.
Paradoxically, estrogen induces apoptosis under certain circumstances. As aforementioned, high-dose estrogen was used to induce tumor regression of hormone-dependent breast cancer in postmenopausal women before the introduction of tamoxifen 2 . This regimen is of clinical interest, given that long-term treatment of breast cancer with the antiestrogen tamoxifen often leads to drug resistance and that sustained tamoxifen exposure may sensitize breast cancer cells to estrogen therapy 10 . The field of the mysterious dual effects of estrogen on apoptosis has not much progress until recently.

Research breakthrough
Perillo Group from the Second University of Naples identified a key player, IKKα in the switch of estrogen action in apoptosis 11 . They found ER agonist 17β-estradiol (E2) and IKKα kinase specific inhibitor BAY11-7082 (BAY) in combination can induce apoptosis in ERα positive breast cancer cell line, and dual-therapy now receives more and more attention.
In the journal Cell Death & Differentiation, Perillo et al. recently reported that the inhibition of IKKα by BAY switched the effect of estrogens on breast cancer cells from anti-to pro-apoptotic, which leads the exploration of therapeutic benefits of estrogen into a new era 11 . IKKα is the kinase responsible for histone H3 Ser 10 phosphorylation (H3pS10) 12 . H3pS10 is known to inhibit H3 Lys 9 dimethylation (H3K9me2) in a space repulsion model 13 . Thus, inhibiting H3pS10 by targeting IKKα facilitates estrogen-triggered

Amendments from Version 1
The authors would like to thank the referees for the helpful and constructive critiques of the manuscript. The observations made by the referees were very helpful and we have adapted the manuscript according to their suggestions. We have rewritten part of the introduction, including the part Dr Young pointed out, and added more discussion on Perillo's article in the body of the text.

REVISED
ER-dependent recruitment of histone methyltransferase Suv39H1. Histone demethylase LSD1 demethylating the Suv39H1 target sites H3K9me2 was increased concomitantly. LSD1-mediated demethylation process is known to produce reactive oxygen species (ROS) and cause ROS-mediated DNA damaging effects 14 . The net results after IKKα knowndown is causing more DNA damages to cancer cells through estrogen triggered ER-dependent Suv39H1 and LSD1 binding to ER target gene promoter ( Figure 1).
In short, Perillo group identifies a novel crosstalk between IKKα and estrogen signaling and shows that inhibition of IKKα-mediated histone phosphorylation switch ER-mediated anti-apoptotic effects to ER-dependent ROS-mediated breast cell death, which implicates potential dual-therapy of ER agonist (E2) together with IKKα inhibitor (BAY) in a variety of hormone-regulated cancers. Questions remain as to whether this study using tissue-culture adapted cell lines such as MCF-7 is relevant to clinic practice. The consideration here is mainly based on the well documented limitations of using established cell lines to predict a clinical response, with several compounds that show dramatic tumor killing in cell models failing to yield clinical benefits. Given that established cell lines do not fully reflect the genomic heterogeneity we now appreciate existing in different breast tumors, it will be important to test the same drug combination in a more clinical-relevant sophisticated tumorstromal microenvironment. This merits further investigation using co-culture between malignant epithelial cells, fat and vascular and inflammatory cells as this likely to influence response to therapy. Additionally, in vivo studies employing patient-derived xenograft model would be very informative in promoting these novel ideas toward clinical trials. Recognizing the known complication of translating scientific discovery to clinical benefit, the question of whether inhibition of IKKα is a valuable therapeutic strategy in ER-positive breast cancer will have to await more positive results from other model systems and also the further development of additional IKKα inhibitors with minimal potential toxicity and highly bioavailable.

Future perspectives
In the last few years, there have been significant shifts in the attitudes toward the use of estrogen in clinic. Estrogen exhibits a broad range of functions that regulates cell proliferation and homeostasis in many tissues. Despite beneficial estrogen functions, sustained estrogenic exposure increases the risk and/or the progression of various cancers, including those of the breast, endometrium and ovary 15 . The International Agency for Research on Cancer (IARC) has listed estrogen as known human carcinogen 16 . Now, the success of the combination of E2 and BAY will certainly become an "accelerator" to the alternative use of estrogen in treating cancers and we expect to see more positive pre-clinical and/or clinical results on BAY or other IKKα inhibitors in ER-positive hormone responsive cancers in the near future.

Author contributions
Conceived and designed the Figure: WZ XZ. Wrote the paper: WZ XZ.

Competing interests
No competing interests were disclosed.

Grant information
This work was partially supported by US Department of Defense pre-doctoral grant W81XWH-11-1-0097 (W.Z.).
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.  :

Suggestions for minor modifications
Section entitled 'Challenges' The authors state that estrogen regulates the proliferation and development of tissues expressing estrogen receptors and that 'ER' is expressed in steroid--responsive female tissues such as the ovary, endometrium and breast. The authors further state 'Thus, estrogen is mitogenic for cultured ER positive breast cancer cell lines." The authors need to amend this paragraph to make it clear that there are two estrogen receptor genes both of which are expressed in breast cancers . Studies on knockout mice have highlighted that expression of the ERα isoform has been identified as critically important to cell proliferation in the uterus and breast . Studies in breast cancer cell lines have highlighted the importance of ERα in both up and down-regulation of cell proliferation and this remains a very active area of research . The authors are encouraged to amend the text of this section changing 'ER' to ERα to make better sense of comments made in the section entitled 'Research Breakthrough'. New ideas about the way oestrogens might affect the growth, differentiation and functional development of cancers, including those of the breast, are always welcome. The authors' Opinion Article questions whether giving oestrogens (rather than the more common anti-oestrogens) may merit exploration as an alternative treatment for breast cancers when provided in combination with agents that alter inflammatory signalling pathways. A limitation of the commentary is that it has focused on data from studies using breast cancer cell lines such as the MCF-7 cells cited in the studies by the Perillo group . A limitation of all studies using tissue-culture adapted cell lines such as MCF-7 is that they do not reflect the genomic heterogeneity we now appreciate exists in different breast tumours , only a proportion of which are 1-3 4 1.
heterogeneity we now appreciate exists in different breast tumours , only a proportion of which are considered classically activated by oestrogens because they are ER alpha positive. In spite of these reservations the authors do raise some interesting points.
Firstly, they highlight the potential for inflammatory mediators such as those that activate the NFκB signalling pathway as targets for therapies in breast cancer treatment. One example provided is the data from the Perillo group, reporting that inhibition of IKKα kinase specific inhibitor BAY11-7082 might be beneficial in switching the effects of oestrogens from anti-to pro-apoptotic. These studies are in agreement with those from a number of other authors claiming that combination therapies are likely to be beneficial in treatment of this malignancy. Notably, there is an interesting paper by Biswas published et al. in 2004 which examines how NFκB activation in breast cancer specimens might have a role in cell proliferation and apoptosis. Notably, in the Biswas paper, they report results suggesting that activated NFκB is predominantly detected in ER rather than ER positive breast tumours. This paper, as negative well as the recent genomic proflling of breast cancers from 2000 patients highlights the importance of considering the both the immediate microenvironment of the tumour as well as the adjacent non-tumour tissue. Notably, in considering how oestrogen therapy might be utilised, it is important to take into account the hormonal status of the woman (pre or post menopausal) as well as the compelling data that has demonstrated intra-tissue oestrogen biosynthesis in fat and other cells close to the malignant epithelial cells . A paper that has just appeared in the Journal of Clinical Endocrinology and Metabolism by in situ Savolainen-Peltonen and colleagues is particularly relevant. In this paper, the authors compared the metabolic pathways producing active oestradiol in breast subcutaneous adipose tissue of postmenopausal women with and without cancer, and these showed there were differences depending on whether the adipose tissue was proximal or distal to a tumour. If the tumour is already bathed in high concentrations of oestrogen either from the periphery or local biosynthesis then addition of more ligand is unlikely to be effective.
The local inflammatory environment of the tumour is an important factor in tumour progression and it is influenced by activation of NFκB activity leading to production of inflammatory mediators such as TNFα, IL-1β and prostaglandin E2 from adipocyteassociated macrophages and COX-2 inhibitors have been cited as effective in reducing recurrence of cancer . Local increases in inflammatory mediators in turn stimulate adipocyte aromatase expression/activity and hence oestradiol production that may be particularly relevant to obese women . Thus, when considering suppression of NFκB activity in combination therapy there may be many context-dependent impacts on the tumour microenvironment to consider.
The role of NFκB activation as a regulator of inflammatory signalling is particularly important to consider in the context of endocrine resistant breast cancer which cannot be addressed using MCF-7 cells. A new paper in Molecular Cellular Endocrinology by Litchfield notes that regulation of et al. endocrine-resistant breast cancer cells may also be mediated by other transcription factors such as COUPTFII. So the keys questions to be addressed in considering the value of the future prospective advanced within this Opinion piece are the following: Are studies in isolated cell lines such as the one described really relevant when it comes to translating new therapies into practice?
Is inhibition of inflammatory mediators such as that using the BAY compound really most effective in cells that are not actually estrogen responsive?This merits further investigation.
Finally, how can we move forward with the field? In our opinion, an excellent way to advance these studies would be to conduct further investigations using methods which better reflect the microenvironment of an intact tumour , this should include studies in situ on the interplay between malignant epithelial cells, fat and vascular and inflammatory cells as this likely to influence response to therapy and it cannot be under-estimated that the individual responses to oestrogens by these different cell types may vary considerably. Examples of such studies currently include 3 being explored in the context of breast cancer, including organoids, tissue slices or xenografts. These approaches are more likely than studies in MCF-7 cells to inform choice of new combination therapies for testing in clinical trials; oestrogen combined with NFκB inhibition should be tested in these systems . The observations made by Dr. Saunders were very pertinent and allowed us to alter the manuscript in several key points. Below we address each one of the reviewer comments. We hope to have provided all the information needed regarding the work submitted and that the manuscript is now suitable for publishing and archiving. Alterations were signalized in the text.

Responses:
"A limitation of the commentary is that it has focused on data from studies using breast cancer cell lines such as the MCF-7 cells cited in the studies by the Perillo group. A limitation of all studies using tissue-culture adapted cell lines such as MCF-7 is that they do not reflect the genomic heterogeneity we now appreciate exists in different breast tumours, only a proportion of which are considered classically activated by oestrogens because they are ER alpha positive … Are studies in isolated cell lines such as the one described really relevant when it comes to translating new therapies into practice?" We thank Dr. Saunders for her comment. Perillo used an experimental approach to evaluate the molecular interaction of both ER and other signaling pathways in MCF-7 cells, as described in many journal publications ( ; ; Garcia ., 1992 et al Planas-Silva ., 1999 et al Sun et ). In these works an evaluation of the biological effect of the molecular interaction ., 2012 al was assessed as Perillo group performed in their work. Thus, we believe Perillo's cell line based studies are meaningful and relevant to find novel targets in hormone responsive cancers. Still, we have altered the manuscript, avoiding the reference to other proportion of breast tumors which are not so responsive to hormone therapy. Although it has been previously reported that ER positive cancers make up to two thirds of newly diagnosed breast cancer cases, we agree with Dr. Saunders that it is important to appreciate the highly existed heterogeneity under the name of breast cancer, to take into account the hormone status of women, and to better understanding the tumor-stromal microenvironment.
"Notably, there is an interesting paper by Biswas published in 2004 which examines et al. 3.
"Notably, there is an interesting paper by Biswas published in 2004 which examines et al. how NFκB activation in breast cancer specimens might have a role in cell proliferation and apoptosis. Notably, in the Biswas paper, they report results suggesting that activated NFκB is predominantly detected in ER negative rather than ER positive breast tumours… Is inhibition of inflammatory mediators such as that using the BAY compound really most effective in cells that are not actually estrogen responsive? " We agree with Dr. Saunders that activated NFκB is predominantly detected in ER negative rather than ER positive breast tumors. Perillo's study we referred in our opinion article also support this point, as "BAY (IKKα inhibitor) only" induced more apoptosis in MDA-MB-231 (ER-negative) cells compared with MCF-7 in their Figure S1C. However, the point to make here, as shown in original Perillo paper, is that estradiol combined with BAY has been more potent in inducing apoptosis in ER positive cells compared with each agent alone in either cell lines. Perillo's study revealed the previously hidden pro-apoptotic property of estrogen in physiological concentration range. On this stand, it inspired us to write this opinion article, and we think it was an interesting observation and worth attracting more attention as well as the warm discussions we have thus far achieved.
"Finally, how can we move forward with the field? In our opinion, an excellent way to advance these studies would be to conduct further investigations using methods which better reflect the microenvironment of an intact tumour in situ, this should include studies on the interplay between malignant epithelial cells, fat and vascular and inflammatory cells as this likely to influence response to therapy and it cannot be under-estimated that the individual responses to oestrogens by these different cell types may vary considerably. Examples of such studies currently include 3 being explored in the context of breast cancer, including organoids, tissue slices or xenografts. These approaches are more likely than studies in MCF-7 cells to inform choice of new combination therapies for testing in clinical trials; oestrogen combined with NFκB inhibition should be tested in these systems." We thank Dr. Saunders for her comments and have changed part of our content according to her suggestion.
No competing interests were disclosed. In this opinion manuscript, Zhou and Zhu highlight recent research that suggests estrogen can act on opposing signaling pathways to either promote proliferation or cell death in breast cancer cells and, further, that the signaling can be biased toward cell death by inhibition of IKKa action. The manuscript is thought provoking and timely and the argument is logical and generally well-constructed. The title is eye-catching and appropriate. The figure is attractive and highlights the key point of the manuscript. If the areas of concern (below) are addressed, the conclusions will be justified by the argument presented.
The concentrations of estradiol required when used together with BAY for the induction of cell death is 10 nM, which is within the normal premenopausal range. We agree with Dr. Young and also another reviewer that it is important to take into Dr. Philippa Sauders account the different hormonal effect on pre-or postmenopausal women, and the dose-related thrombogenic side effect of estrogen.
No competing interests were disclosed. Competing Interests: