Endometriosis Foundation of America
Endometriosis 2013 / From Sampson to the 21st Century
Serdar Bulun, MD
Thank you for providing me with this opportunity. Last night I told Dr. Seckin that this is one of the best endometriosis meetings I have ever attended. It is a wonderful blend of clinicians, patients, MD and PhD researchers. It really is an honor to be here.
Endometriosis, as we all know, is an enigmatic disease and we are still nowhere near a place where we have satisfying outcomes; opportunities and options for patients. In contrast to previous speakers most of the data that I will present relates to pain. Most of the time when I interview my patients, say teenagers, they start with severe dysmenorrhea often starting at menarche. By her 20s she has chronic pelvic pain now prevalent between periods. Late 20s and 30s a first laparoscopy is performed and by the 30s and 40s multiple, and usually ineffective, follow up laparoscopies and experimental medication are done. Most of the time the patient goes through some of these medications, like oral contraceptives in the beginning, then other medications are added. She goes through several courses of GnRH agonist, surgery and finally a hysterectomy and bilateral salpingo oophorectomy is performed in most of these patients.
I do not have to explain to this crowd about what endometriosis is but it is for a pathologist the presence of stroma, glands and bleeding. Most of the data and conclusions that I will present to you sort of stem from some work that we performed over the past 20 years from surgeries performed by top endometriosis surgeons. The stromal component, in our mind, is the major focus of the disease and what I am going to talk about.
For example when you look at an endometrioma in the ovary this is the fibrous tissue associated with it and this tissue imitates, mimics endometrium, eutopically located endometrium. What we have done for the past 20 years is take these cells, the stromal cells, and looked at their behaviour in culture dishes and sometimes in in vivo models.
This is very misleading that the endometriosis is hidden in the cul-de-sac. If you make an incision right at this junction and go further you are going to find that there is deep, infiltrating endometriosis; sometimes involving the bowel that is there. Most of the time what we have done is take these primary cells, brought them to the lab, looked at their molecular behaviours and went back and in vivo confirmed what we found, again, in the disease situations.
One interesting observation is that for pelvic pain purposes ovulation is the enemy. Once you suppress ovulation for endometriosis a lot of patients find there is partial relief. I am not going to go over the details of ovulation but ovulation as a side effect, as you know, involves menses and by evolution humans, like other primates, developed menses. Each time a pregnancy does not occur this unbelievably inflammatory event repeats itself. On the endometriosis side the hormones that are secreted by the ovary also contributes to persistence, growth and inflammation. This pertains because once we stop ovulation through either GnRH agonists or oral contraceptives or high dose progestins the symptoms and the chances that endometriosis symptoms or appearance will recur is lessened. Suppressing ovulation has been the target of a lot of existing treatments.
This could be controversial but I thought that 95 percent of all the endometriosis cases that we see in the pelvis could be explained by the hypothesis which was popularized by Sampson. In eutopic endometrium a number of cells reach the pelvic cavity through retrograde menstruation and it may populate some of these regions that we observe endometriosis. For example, peritoneal endometriosis could happen through this mechanism and by the same token a hemorrhagic cyst or corpus luteum may be populated by these cells and turn into an endometrioma, which was very elegantly demonstrated actually by Vercellini and colleagues. Also, the rectovaginal nodules, the majority of them that we see probably occur through recurrent deposition of these cells in this pouch over the years.
I will talk a little bit about ovarian steroids. They are relevant because, before I get to this - they are almost - they have holistic effects on the endometrium, if you will. Some of us in this room perform IVF. If you take a post-menopausal woman who has no ovarian activity, if you treat this woman with estrogen for 10 to 14 days, whatever your protocol is, and add progesterone on top of this, then put an embryo into the endometrium, it takes. It implants. In this case none of us added cytokines, growth factors or any of those other molecules that are involved in implantation. If you think that implantation is the ultimate function of an endometrium, providing exogenous estrogen and progesterone is enough. That is sufficient to demonstrate to you that it is a very complex mechanism and it is probably very short sighted to pass beyond estrogen and progesterone and say, "Yes, I can just pinpoint one downstream molecule and explain how these molecules work". I am still struggling to explain how it works.
Most of us in the room are familiar with progestins such as a Depo-Provera like MPA and Norethindrone acetate, which is Aygestin. On the other hand anti-progestins became available to treat some of these diseases, such as fibroids and endometriosis. However, they never made it to the clinic. However, some who mixed agonist, antagonist compounds such as RU486 or ulipristal acetate have recently become available and in a paradoxical manner, both these progestins and anti-progestins are somewhat beneficial for endometriosis.
One interesting thing to emphasize is that when a steroid arrives at the eutopic endometrium the capillary first makes contact with the stroma. For example, in endometrial cancer you do need estrogen receptors in the stromal cells to cause epithelial proliferation, which is very interesting to note. The estrogen receptors in epithelium do not play a major role in this process.
By the same token, the protector effects of progesterone for endometrial cancer again occurs through stromal progesterone receptors and again when you look at the histology and anatomy of the endometrium these molecules first make contact with these two of their receptors in stroma. This is important to note. This emphasizes the role of paracrine interactions in endometrium and this also applies to endometriosis.
I mentioned that some of these effects of progesterone and estrogen are sort of holistic. We took this on and we wondered how on earth progesterone shows these effects. We looked at the entire human genome. When you introduce RU486 or progesterone we saw that the progesterone receptor binds to some 7,000 to 32,000 sites across the genome. We are not talking about a single promoter of a gene. This really makes it extremely complex. The same concept also applies to estrogen and its receptor.
We attempted to understand at the gene level in individual genes how progesterone and antiprogestins act and it is very interesting that when you put progesterone under in vitro circumstances it is a very lousy hormone and probably you do need an...situation to study it better. It does not cause binding of this receptor to a promoter for example of a tumor suppressor gene. We think that this is part of the reason why symptoms persist. On the other hand when you introduce a compound like RU486 which has a much higher affinity to its receptor we demonstrated that it does cause a reconfiguration of the promoter of this gene including DNA bending for example. This is the reason, possibly, why these beneficial genes for the patient's health, if you will, are activated and symptoms may improve.
To make a long story short we demonstrated that in the case of endometrium progesterone receptors in the stroma act as a mediator so that it would produce, cause production of, some paracrine compounds including retinoic acid, which would act on the epithelium and turn on some key genes like this enzyme. This enzyme would inactivate estradiol to... Whereas in endometriosis because progesterone receptors are reduced in this tissue this does not occur. Can this be applied to some of the conditions? I accept the data that Dr. Herschlag just showed and I am still a little bit skeptical that the endometria of these patients with endometriosis are defective and maybe there might be a ten to 15 percent chance to improve IVF rates. Some of these clinical studies may not demonstrate this. We think that this mechanism does not happen in endometriosis and to an extent it is also compromised in the eutopic endometria of patients with endometriosis giving rise to progesterone resistance. It remains to be seen. Some of the best experiments are sometimes done by just inhibiting estrogen production and repeating IVF and see what happens. I think I would still leave that door open.
I would like to say a few words about stem cells and epigenetics. This is the endometrium and most of us are familiar with the anatomy of the endometrium. What really makes a difference in the primates are these spirals arteries. What happens is for each menstruation these spiral arteries constrict and they go through coagulation. That is why this functionalis portion of the endometrium sort of disappears. This is some hypothesis which was developed in a collective manner, if you will, including Rob Taylor and Len Dekker in Germany and many other investigators whose names I cannot remember. The idea is that there are a lot of stem cells in the basalis level of the endometrium. It is very simple. Like in some women if this separation occurs fairly high then this portion of endometrium would not have a lot of stem cells in it. On the other hand, if the menstruation occurs at a deeper level for example, then some of these stem cells may be carried through the tubes into the peritoneal cavity, and that may play a role. We do not have any direct evidence for this but it is an attractive theory.
Just a few words about epigenetics and stem cells; this is really interesting because we have the same DNA, for example in our retinal cells in the eye, and also in our endometrium. The DNA is exactly the same, the code is the same. I put it down as like genes this line. However, what really makes the difference is that promoters of these genes contain these CG repeats, such as in the FMR genes. Some of you are familiar with this. If these are open, unmethylated, it produces a protein. If they are methylated they are closed and they are not functional at that time. Stem cells use this mechanism greatly. These are like, if you will, the embryonal stem cells in an inner cell mass, let's say, and they make very few proteins in the beginning compared with some of the other cells and they are select proteins. A stem cell is sort of like a loosely used term and we also use some of the progenitor cells, like in endometrium, we call them tissue stem cells, like the basalis cells. You can envision that some of the other genes are now open and they make some of these proteins. You can also think that in mature cells the functionalis level because this is differentiated, new genes have opened up, keep in mind that proteins are the ones that you phenotype to cells. I look the way I look because of the proteins that I make, not necessarily because of my genes. The retina of the eye would be totally different. This is a very interesting concept and it really explains how stem cells function and differentiate and why we look the way we look.
The endometrium of a disease-free woman has normal molecular and phenotypic features, such as they do not express aromatase and they express low levels of COX2 to make some prostaglandins. The endometrium of women with endometriosis on the other hand starts to act up, they express some aromatase and increased levels of COX2 whereas endometriotic lesions, ectopic lesions express large amounts of aromatase and COX2. This was our original observation some 20 years ago. Epigenetic and stem cell theory provided a wonderful basis and explanation for this as we move forward.
We find this interesting vicious cycle in which prostaglandin E2 turns on the entire steroidogenic cascade as well as aromatase. This feeds back on estradiol production and COX2 formation. What we found was this transcription factor SF1 was the orchestrator in this whole deal. If you take SF1 out this cycle disrupts. We know that the cycle is important because in your practices if you give a patient a COX inhibitor or aromatase inhibitor they provide partial pelvic pain relief.
We tried to get to the bottom of this and what we found was in normal endometrium the DNA methylation sites that regulate these key two transcription factors SF1 and ER Beta were methylated and shut down. Whereas in endometriosis they were inappropriately open and they gave rise to production of these sorts of bad, if you will, proteins for the disease that kept them on.
We put these all together and again show that cholesterol is converted to estradiol in the disease, prostaglandin E2 feeds on this cycle through SF1 which is the orchestrator of these gene activations. As I mentioned before the enzyme that inactivates estradiol is deficient. This is because of the deficiency of progesterone receptor and retinoic acid and now we found that this ER Beta, the other estrogen receptor, is the key. Its levels are increased by more than 100 fold in endometriosis. The level of SF1 is increased by 12,000 fold in endometriosis. These are like big... ER Beta causes COX2 and prostaglandin E2 formation and they also feed up on each other. The confusing picture is what comes first and we think that SF1 and ER Beta are the initiators of this interesting molecular event because their genes are methylated under normal conditions and not in endometriosis.
If we revisit Sampson's hypothesis from a molecular perspective you could imagine this in normal, like more than 90 percent, women, because it is almost like we know all women menstruate in a retrograde fashion. During menses a few cells may go into the peritoneum but these cells are appropriately programmed epigenetically. For example, the SF1 and the ER Beta promoters would be shut down and these cells would not survive and go away. If you look at a patient destined to develop endometriosis, like a young girl, we know that these women have severe pain even with their first menses, right? I mean when I interview my patients with severe endometriosis almost all of them have this history.
There might be a few sleeper cells embedded in the endometrium. These bad cells may go to the peritoneum. They might have the survival epigenetic factors to stay there and hang and maybe proliferate just a little bit and differentiate sometimes into the epithelial cells because they have this stemness embedded in them. It is an interesting hypothesis and again, it remains to be seen whether it is true or not.
Finally, again, if you really think of a five week old embryo, 46XX embryo, a lot of things are going on in this embryo. If you take a cut right through here a bunch of cells are leaving the yolk sack area, these are really epiblastic cells in fact and they are going to the genital ridges. The distances between the mullerian duct at this time from the genital ridges is very, very tiny, teeny, we are talking about microns or millimeters. You can imagine that and look at the peritoneal cavity. You can imagine that a lot things, epigenetically, can go wrong. These could be environmental effects, these could be genetic predispositions and if these individuals had, if you will, the endometriotic phenotype epigenetically they would be destined to develop endometriosis later in life through different things.
This is the last slide. I just wanted to put this up as there is still hope to use these mechanisms. As you know very recently they induced pluripotent stem cells - they have been used extensively to treat some of these diseases. Basically, what happens is, it is a wonderful technology. You can take a patient's somatic cell and you can add these genes to it to reprogram them to act as stem cells or they get really close. They are not exactly inner cell mass type of stem cells but they get close to it.
What you may want to do for example you can differentiate them in vitro and screen them for therapeutics. You can screen these cells for either individual therapeutic options or general therapeutic options for endometriosis. More interestingly, you do not have to take a skin biopsy. For example, we are gynecologists and we can take an endometrial biopsy and we can isolate the stromal cells. We can reprogram these cells and we can put them back in the endometrium if our hypothesis is correct.
It is interesting, currently most of these are used to correct the base...changes the mutations in DNA, put them back and put it back into a patient's either diseased tissue or overall germ cell transmitted disease condition.
I told you there is a lot of hope so this maybe one of the ways that we may move forward among others. I review speculations that I have and that stem from these molecular observations I put them up yesterday. I will go over them again. I think the symptomatology of endometriosis represents a spectrum of a broad chronic disease involving a basic pathology in pelvic tissues including the endometrium. The disease becomes initially manifest with the so-called primary dysmenorrhea in teenage years and advances to laparoscopically visible endometriosis. Sampson's hypothesis, in my opinion, explains the majority of the cases of peritoneal endometriosis, rectovaginal nodule and ovarian endometriomas.
In epigenetically susceptible women, the risk of development of endometriosis and its symptoms increases in direct proportion with the number of ovulatory menses. Long term suppression of menses with oral contraceptives in young women with primary dysmenorrhea should decrease the risk of symptomatic endometriosis. I think it would be clinically beneficial to view pelvic symptoms associated with endometriosis as a spectrum and broaden its definition to a "systemic disease characterized with pelvic pain responsive to hormonal suppression or surgical resection of endometriotic tissue".
Again, I would like to acknowledge all the people who have done this work in my lab but, in particular, my collaborators both scientifically and endometriosis surgeons that I closely collaborated with over the years. Without their help and insight for this disease I do not think these studies could have been performed. For example, the first time I really learned endometriosis was when I was scrubbed with Alan Johns in Fort Worth. You know he was resecting all these tissues and throwing them away and I said, "Why don't you give them to me and we'll take them to the lab and look at them". That type of thing. This is how it all started.
Thank you so much!
Victor Gomel, MD: Serdar, can we ask you a few questions? We have time; there were speakers that were not here, so we have a few minutes. That was a very interesting presentation. I admire your research in the subject, really. Maybe we can ask you a few questions and one that I want to ask you since you brought up the ovulation issue. We know that when ovulation occurs in the...in the peritoneal cavity you have a tremendously high level of estrogen present. I mean, that must do something!
Serdar Bulun, MD: Right, right. Yes, this is very interesting I think when we look at our bodies as humans, women I don't think from an evolutionary perspective their tissues were designed to handle incessant ovulation. Maybe 20 or 30 in a woman's lifetime but not, I do not know, thousands of them. I think ovulation itself is a traumatic event that predisposes towards the endometriotic phenotype, I would argue, in a sense.
Victor Gomel, MD: Yes, I agree because in nature women would be pregnant probably once a year, or a year and a few months if we did not have all the prevention of pregnancies that we use.
Serdar Bulun, MD: Right, I mean from that perspective I think continuous oral contraceptives...I argue with some of my patients, we have a lot of these conversations, may be a more of a physiologic state than recurrent menstruation. It is an interesting discussion point.
Avner Hershlag, MD: I have a question, first of all, thank you for this amazing presentation and for embarking on such a voyage to really understand endometriosis with your very sophisticated research. I feel that all too often we clinicians kind of bypass the path of physiologic questions because we jump right to therapy. And, you know, look what happened for example to the physiology of the fallopian tube. It kind of pretty much, all the studies, pretty much disappeared because we created a bypass through IVF.
The evidence that you have shown on the molecular basis for the alterations and changes in receptor activity is very impressive. Connecting that to the clinical findings that in most endometriosis patients it seems like it is the egg or embryo site rather than the endometrium; this kind of, I think, is asking for studies that would try to do both in the endometriosis patients. That is that you know. I am wondering if there are such studies, or if you are planning, almost suggesting, to collaborate with you on that. And that is to, for example, do endometrial biopsies prior to IVF cycles and actually look at the endometrium, see what it is lacking and then see how that endometrium performs in vitro fertilization.
Serdar Bulun, MD: Right, I mean, I only have anecdotal data. We have not...I could not convince even my clinical group in Northwestern to do some of these studies. For example, one idea might be to - it almost looks like estradiol production in endometrium might be toxic for ovulation right at that implantation window because there is a very active enzyme named 17 Beta HSE type 2. All this enzyme does is get rid of any estradiol molecule in the epithelial cell, right at that window. It would be interesting to target that through either aromatase inhibitors or other means to see what happens and see if we can improve pregnancy rates, for example. That would be an interesting study.
Avner Hershlag, MD: I would be really curious. You know there are studies that show that an endometrial biopsy on the cycle prior to IVF actually increased pregnancy rates. So, if we did a study, a prospective study, where we did biopsies on endometriosis patients and with controls, you know women who do not have endometriosis and studied all the parameters that I assume you have the tools to study, and compare it to what happens with the implantation rate, and perhaps even look at some molecular parameters in IVF such as the follicular fluid, etc., I think that could shed new light on this. Because the question is how do you connect the molecules to the actual - I mean - how important are they in the general scheme of things? It might be that treatment alters that as well.
Serdar Bulun, MD: Let's talk more on that.
Avner Hershlag, MD: Yes.
Audience Member: Hi, I think that research into the molecular basis of endometriosis is really interesting and a really important part of things. I am just wondering if you believe that it is a combination of molecular, the molecular defects that you described in combination with Sampson's hypothesis that causes endometriosis? Then, why would it be that when a patient has a good quality excision surgery they do not usually develop new endometriosis? Presumably they would still have all of those molecular defects.
Serdar Bulun, MD: That is a good question. Again, I think the data is very variable. The outcomes of these studies are directly related to the expertise of the surgeon. But if we look at the United States in general, the recurrence of pain after, even in a treatment naïve patient is two to five years. If it is a patient who has already been treated, either through medical treatment or laparoscopic surgeries, it is six months to two years. It is a complex question. But I think when you look at the entire population the picture is not that rosy.
Victor Gomel, MD: Any other questions?
Audience Member: Thank you, it is very interesting and thought provoking. I was just wondering if a drug could be developed that actually causes methylation. You could talk better about demethlyation after endometrial cells on a molecular level because that would basically be the answer to the treatment of endometriosis medically, right?
Serdar Bulun, MD: Right. Absolutely, I mean, current demethylating agents are very broad effecting agents and I was just talking to another speaker last night about this. You almost need to develop drugs that affect methylation of a particular gene but not the others. I am sure it is doable. It just will require time.