Steroids and DNA in Reproductive Tissues: Dangerous Liaisons - Serdar Bulun, MD

Steroids and DNA in Reproductive Tissues: Dangerous Liaisons - Serdar Bulun, MD

Endofound Medical Conference 2017
"Breast, Ovary and Endometriosis"
October 28, 2017 - Lotte New York Palace Hotel

Steroids and DNA in Reproductive Tissues: Dangerous Liaisons Q&A
Serdar Bulun, MD
Chair of Dept. of OBGYN Northwestern University

Thanks so much for this opportunity. When Tamer called me about this meeting and mentioned that we are going to be able to discuss breast and endometriosis and ovarian cancer together, I was really excited. And I asked him, "Really, Tamer? Can I come and talk about some of my even crazy ideas?" And he said, "Absolutely. You can speculate."

I included a few slides in the very beginning. Not necessarily data slides, but some common unifying themes between breast cancer, ovarian cancer, endometriosis and fibroids, the diseases that I have been working on. As you can see, they are quite common problems and worldwide hundreds of millions of women are affected by them. I think it's an epidemiologist's nightmare to be able to try to establish links between these common problems through some time honored epidemiologic studies.

But molecular studies and hormonal studies suggest that there are links. I think the most important link in my mind is the repeated episodes of ovulation and menstruation, which the signals start from the brain with this FSH and LH. And when ovaries getting an egg ready to be fertilized sends these side signals both through the endometrium to get ready for implantation. And breast is not in the picture, but also breast gets some signals by progesterone, "Hey, there could be an embryo coming up soon and you should be ready." At the same time, these hormones being exposed to sequential production of both estradiol and progesterone, I argue and I submit to you, that could be detrimental to a woman's health.

Let's explore that. You all know that breast cancer is a very nasty disease. It does develop in breast fat tissues. It doesn't develop in vacuum. This fat tissue provides both biochemical and structural support for breast cancer. When you look at distribution of breast cancer in the world, you could see that, like in sub-Saharan African and in some of the probably developing countries is less prevalent. One of the, again, themes that I would like to bring up is one difference between these countries and more developed countries with prevalent breast cancer cases is that the women in these countries go through many more pregnancies and lactation. They experience less episodes of ovulation and menstruation.

Again, endometriosis, we all know is a nasty disease. It's very common. And it's the presence of endometrium-like tissue outside of the uterine cavity and also it could involve the ovaries and the entire pelvis like that. It is, again, Tamer showed these pictures. It's quite common when we do laparoscopy at a time of menstruation that we see this huge amounts of blood mixed with endometrial stromal cells and epithelial cells in the abdominal cavity. Again, I think most of us agree that there is enough evidence that 99% of endometriosis possibly have something to do with repeated episodes of ovulation and menstruation. These blood and tissues, they accumulate on pelvic surfaces and we call it peritoneal endometriosis if we see it only peritoneal surfaces. This could lead to, if endometriotic implants probably are involved with hemorrhagic corpus luteum. This eventually turns into a endometrioma. Repeated episodes of their position of these tissue and blood in the cul de sac eventually gives rise to tissue remodeling and complex rectovaginal nodules.

Again, endometriosis is linked to ovarian cancer and when you look at the distribution of ovarian cancer across the world, you see a similar distribution that in probably countries where there are more likely of disruption of ovulation and menstruation in a consistent way, ovarian cancer is less prevalent in these countries.

Uterine fibroids is another nasty disease. It's much more prevalent than all of these conditions I talked about. And it could cause a lot of severe symptoms including excessive uterine bleeding, anemia, recurrent pregnancy loss. It accounts over 200,000 hysterectomies a year in the United States.

I just want to bring up this slide about the history of repeated episodes of ovulation and menstruation, but this is the life expectancy. As you can see, I mean, the life expectancy in most of the continents across the world was not more than 35 even as late as 100 years ago. So, a lot of women probably during these times, they were going through repeated pregnancies, either died of some complication of pregnancy and lactated a lot too. If you look at, again, the number of pregnancies, again, there was a huge change that probably took place after the industrial revolution and after the lifestyle change of women that happened around this time. Again, I think I will like to, maybe even after my talk or later on, have some discussion as to how these changes could be contributing to being more exposed to ovulation and repeated menstruation. And being more exposed to these sequential secretion of endogenous estradiol and progesterone.

Again, I mean, there is some evidence, these circumstantial evidence, that when the most of the treatments that target suppressing ovulation decreases endometriosis, fibroids, ovarian cancer, and again, for example, if you would be using other methodologies other than birth control pills, it would be possible that even breast cancer could be prevented. For example, oophorectomy or even adrenalectomy in the past, it's not done anymore, decreased the incidence of breast cancer.

All right. I'm switching gears now. These are less speculative issues. Endometriosis, breast cancer and leiomyoma, they all respond to these hormones, estradiol and progesterone. But as you could see, I mean, the details are very complex. In endometrium, for example, estradiol is carcinogenic as it is in breast cancer and is tumorogenic in leiomyoma. But the role of progesterone is totally different. Progesterone is protective for endometrial cancer and is carcinogenic or tumorogenic in breast cancer and leiomyomas. We always thought that the issue is the presence of a sophisticated stroma in endometrium that makes a difference with respect to response to progesterone. And endometriosis is totally different and I'll talk about it a little bit more later.

In my early career, we'd done a lot of studies on the role of estrogen in breast cancer and aromatase and how aromatase inhibitors helped. To make a long story short, in breast cancer, breast fat plays a huge role in promoting the growth of carcinogenic epithelial cells, epithelial cancer cells. And estradiol has a direct effect on their growth, but estradiol also induces angiogenesis. And it also induces secretion of some cytokines that would keep the breast fat in undifferentiated state as adipose fibroblasts. These adipose fibroblasts in turn express a lot of aromatase and secretes a lot of estradiol that goes back to the breast cancer and causes growth. And I think this is a very interesting and complex mechanisms between these cell types. It's not that different sometimes when you think of it in some of the diseases that we study, such as endometriosis. This is also, as you all know, a very inflammatory condition, breast cancer.

Just a few words about uterine leiomyomas. Again, during each episode of ovulation, some signals go to the stem cells in myometrium and the signal goes to the myometrium and probably tells the myometrium that, "Look, you need to get ready. There is a fertilized egg coming. You need to get ready for pregnancy." And probably myometrium ... not probably, myometrium doesn't know about the social life of a woman and it gets ready for the pregnancy. And all of these cells get going, but if there is no pregnancy, then it's like there's a menstruation and it gets ready for the next event. Ovulation occurs again.

So, you could imagine a situation in which during these repeated episodes of activation of these myometrial stem cells, one of them can get mutated, such as a MED12 mutation. And this could give rise to a monoclonal growth of a single tumor that we call fibroids. Eventually, it gives rise to these tumors of various size with a large extracellular matrix component. Probably that's how it happens. Most of the time, the signal that turns on the stem cells come from the neighboring mature cells because the stem cells, they do not have sufficient concentrations of estrogen or progesterone receptor.

What happens is estrogen and progesterone work through their receptors in the mature fibroid cells that are terminally differentiated and this in turn sends signals to the stem cells, such as the WNT pathway, such as TGFβ. This stem cell is vulnerable with a mutation in, for example, MED12 is one of them. It regulates a lot of chromatin and nuclear processes. This gives rise to formation of a tumor, but most of the action is in these stem cells in the case of fibroids.

I'll switch to endometriosis. Again, some of my colleagues have shown some of these slides. Again, I think most of us who deal with endometriosis, we do know that ovulation is totally bad and we do target the suppression or interruption of ovulation through various medications. It's important to note that the pain may come from the endometriotic implants on the peritoneum or elsewhere, but I think there's also some evidence that eutopic endometrium is kind of different. In fact, one Australian investigator has discovered neuronal types of cells that could serve as nociceptors in the eutopic endometrium of these patients. The pain could be quite complex. It can arise from the peritoneum. It could arise from the eutopic endometrium. But estradiol is definitely, it promotes inflammation as Rob and Tamer said about this disease. So, the major biologic process, the bad biologic process, in endometriosis is inflammation together with fibrosis.

I heard this concept once from a German gynecologist named Leyendecker. He was hypothesizing that, for example, this is endometrium, which gets renewed every month, right? You can imagine that and the functionis portion is subject to retrograde or forward menstruation. You could imagine a situation in which the majority of the stem cells are in the basalis so that they could build a normal, the next set of endometrial tissue in response to estrogen. If Leyendecker thought that the separation occurs at a more superficial layer, the endometriosis is maybe less likely because lesser number of tissues and stem cells would be carried to the peritoneal cavity. On the other hand, in a more unfortunate woman, endometriosis maybe more likely if this occurs at a deeper layer. In fact, this kind of resonates. Most endometriosis patients will give you a history of very heavy menstrual periods. I'm not sure if that was quantified and proven, but there's a very common anecdote that I hear from the majority of my patients.

I'll talk just a little bit about epigenetics because some of us think that the major defect in endometriosis is primarily epigenetic and epigenetic means how DNA is programmed, the same piece of DNA is programmed to function in different ways. For example, when you look at an individual, when you look at me what you're seeing is mostly protein. And these proteins are made by the same DNA. I carry the same DNA in my skin as in my liver or in my eye, but they all look different, right? The reason is the genes that are in either embryonal stem cells or the basal part of the endometrium, functional part of the endometrium, or the eye, are differentially used, differentially programmed to make different sets of proteins. This has been programmed from birth or even earlier. So, you could imagine then a situation in which if this programming goes wrong, that could have something to do with endometriotic phenotype and I'll explain to you.

The first time we sort of came across this epigenetic problem in endometriosis was we looked at these genes called SF1 and estrogen receptor beta. They are both nuclear receptors. We looked at their promoters and they were appropriately methylated, the DNA was methylated, so they were suppressed. So, they were not supposed to be expressed in normal endometrium and they did not. And then, in endometriosis, we found out that the DNA was pathologically unmethylated and in fact we were really surprised to see that, but it did fit our hypothesis. And these genes were turned on.

Why is it important that SF1 and ERβ are on in endometriotic cells? Because they are the tissues that can make estradiol from cholesterol and then they also mediate the bad actions of estradiol, such that aromatase is expressed in endometriosis. In fact, that was our first finding. In fact, I was a fellow working with one of, I think, friends with both Harry, Alan [inaudible 00:21:52] in [inaudible 00:21:54] and I was learning how to do endometriosis surgery from Alan. And Alan was throwing away all these tissues. He was not even sending them to pathology. I said, "Why don't I take these to my lab and check for aromatase?" Because I was looking at aromatase in the breast. I found more aromatase in these tissues than breast tissue. That's why I turned my attention to endometriosis, in fact.

Anyway, so we found out that prostaglandin E2 stimulates this formation of estradiol from cholesterol. The orchestrator is this nuclear receptor called SF1. To make a long story short, estradiol is not appropriately inactivated in endometriosis because of the deficiency of this other enzyme. And this is because progesterone receptors are low and this enzyme is not activated properly. And this is because ERβ was suppressing progesterone receptor levels. And again, ERβ was working through estradiol to induce COX2 and prostaglandin E2. Again, we made a huge mess of vicious cycle. What comes first? We think that what comes first is the SF1 and ERβ because they are regulated by this really fundamental mechanism of unmethylation. Is this important clinically? Of course. We treat endometriosis pain with COX inhibitors and aromatase inhibitors.

Again, I will not go into detail, but progesterone resistance is also important because progesterone receptor levels in the stroma of endometriosis is very low and the stromal cells are the first line of cells that come in contact with the capillaries, both in the endometrium and in endometriosis. And the stromal cells sends some signals to the epithelium, so that the epithelium can function appropriately for implantation and other reasons. But these mechanisms were along with retinoic acid system disrupted, messed up, in endometriosis.

What we did was we wanted to look at genome-wide abnormalities of epigenetics. The DNA methylation. And we looked at first to see how methylation of DNA could vary with respect to either proximity to CPG island. These DNA methylation sites are called cytosine and guanine, CPG. You could look at whether DNA is methylated close to an island or is methylated close to a transcription start site that starts the expression of a gene. And in a normal situation, the distribution was like that that only 11% of these differentially methylated regions should be close to a transcription start site and their distribution between the island shores and open seas are quite equal.

But when we look at the difference between normal endometrium and endometriosis, this was altered tremendously. The bodies of the genes became more important. The transcription start sites were less important. And then, again, with respect to the CPG islands, open seas became more important than the islands. This was genome-wide abnormality, so we went ahead and looked at some genetic single genes. We found out that the gene named GATA6 was overly expressed in endometriosis because it was not appropriately methylated. Whereas, its counterpart, GATA2 was present in normal endometrium. And again, you're going to ask me what are these GATA2 and GATA6? They are transcription factors, but what we found was when we added GATA6 to a normal endometrium, it recapitulated all the abnormalities, all the steroid abnormalities that we saw over the past 20-25 years. And we were really excited about that.

So, you could probably put the abnormalities in endometriotic cells. One is the steroid-related abnormalities. We thought that GATA6 together with SF1 is very, very important and is one of the master regulators of these abnormalities. The other set of abnormalities, in fact, I would like to ask Rob's opinion on this later on, was about the immune function of these endometriotic cells. When we over-express the GATA6 and compare this with normal endometriosis and endometrium, we came up with 130 differentially expressed genes and 30 of them were part of the immunome.

Again, I mean, when you look at the pathways, the inflammatory response and complement activation pathways, were some of the highest ranked pathways in these differentially expressed gene sets.

We did demonstrate that, for example, these chemokines, CXCL 14 was stimulated by GATA6. Again, it also was important for complement-related pathway genes. It reminded me of one of Rob's early studies. These endometriotic cells made tons of IL-6, for example, by themselves. So, it brings the question that are they transformed in such a way that they act like ovarian granular cells on the one hand. Then also, they act like, for example, macrophages and other immune types of cells. Maybe they don't need to recruit any macrophages from elsewhere, from the blood. They do their own work in terms of making cytokines and also chemokines. I think that's an interesting discussion.

Here are just a few slides about how estrogen and ERβ is involved in inducing inflammation in endometriosis. We found out that estradiol works through ERβ to induce this kinase SGK and this new protein that we found, RERG, which was originally described as a GTPase. But to make a long story short, RERG is phosphorylated by PG2 and it does go back to the nucleus and contributes to ribosome biogenesis. I do not know much about ribosome biogenesis. I think this is one area that should be explored later in endometriosis. Again, SGK1 is known to be a sensor of oxidative stress and is very important for cell survival. Both of them do contribute to inflammation through these pathways. I think this is still an open area. Like, how does estradiol and ERβ increase inflammation?

If you go back to a five week old embryo, and look at the genital ridge, a lot of things are happening at that time. Like, the ovaries are formed. Look, the peritoneal cavity's right here. Coelomic cavity. Mullerian duct is right here. So, these things are almost millimeters away from each other and you could imagine that a mis-programming could go wrong here and it could give rise to the endometriotic phenotype. Again, I mean, this is quite broad and speculative, but I think this is important to note.

If we put everything together in a woman free of endometriosis, retrograde menstruation does happen and these women are ... But these cells are appropriately programmed, so they don't express was they are not supposed to express. And these cells, they apoptosis, they die. They are reabsorbed. On the other hand, in a woman, who is destined to develop endometriosis, she already has some ... A few cells in her probably eutopic endometrium is some epigenetic abnormalities. If one of these cells reach the peritoneum, they could express these survival factors and some of them may, for example, give rise to expression of aromatase. Again, is this important? Sure, because if these three inhibitors were shown to treat endometriotic pain. And this could be one of the mechanisms. And I'll take this one step further. For example, during this process, let's say, there is a corpus luteum and a set of stromal and epithelial cells with some blood was able to populate some portion of this corpus luteum. You could imagine that this could give rise to an endometrioma. Something like this. Or it could be much smaller than this. I mean, it could be more subtle.

When you look at the surface of the endometrioma, there is normal ovary somewhere here. There is extensive fibrosis as you all know. There is endometrial stroma and there is one layer of endometrial epithelium. Then, again, there are the most prominent cell type is these hemosiderin-laden macrophages. Probably these cells, I would argue that, they could probably evolve from some of these stromal cells. It's quite likely. That's speculative though. We know that cancer, like carcinoma by definition, arises from epithelial cells.

So, it is quite possible that when these cells fall into the ovary, this epithelial cell type could be an endometrial cell type or it could carry some of the tubule epithelium with it, like some of the [inaudible 00:35:22] cells. And in the endometrium environment, there are tons of epigenetic abnormalities affecting the stromal cells. I think [inaudible 00:35:36] New England Journal publication and others show that stromal cells are not mutated. The DNA in these cells are not changed in a classical way. So, all the mutations they found were in the epithelial component. Some of these mutations were ARID1A and KRAS. It's quite likely that this constant bathing by these nasty factors like prostaglandin, estradiol, cytokines, MMPs, or exposure of these epithelial cells, can give rise to some accumulation of mutations over the years and eventually just could give rise to various cancers of the ovary. Again, this is total speculation, but the epithelial cells are more wired to accept mutations compared to stromal cells because almost sarcomas are extremely rare either in ovary or elsewhere.

This is my last slide. Again, I would think that evolution of a woman's body, lower brain and reproductive tract, probably did not catch up with her higher functions and recent lifestyle choices. I think at least this should be recognized in any strategy about endometriosis, breast cancer or ovarian cancer. I would encourage broader use of ovulation suppression strategies for the purpose of preventing endometriosis, ovarian cancer and possibly breast cancer, not with birth control pills in this case. Especially in susceptible groups. Birth control pills, they are the most available ones, but they do have significant side effects and I think we should develop new technologies at least to provide an option to suppress repeated episodes of ovulation with less side effects.

This is my lab group who contributed to some of these studies and my collaborators. Thanks so much.