Endometriosis Foundation of America 2014
Tackling complex human diseases: a journey from autoimmunity to endometriosis
- Peter Gregersen, MD
Thank you very much. I am going to be as direct and brief as possible. I knew nothing about endometriosis until Tamer Seckin came and saw me last April. I had not given a thought to it but it immediately struck me as a disease phenotype which was in many ways analogous to what I have been doing for the last two and a half decades in autoimmunity. I just want to give you a flavor for why I think that and what some of the opportunities are for comparing across these diseases and taking similar approaches that are rapidly developing now.
First of all the genetics; the genetics have a lot of similarity in terms of a challenge. There are significant genetics around this disease it is not quite as strong when you just look at familial aggregation as Lupus or type I diabetes or rheumatory arthritis. I caution you not to conclude from that that genetic load is low because this relates to this entire heterogeneous phenotype which likely has many subgroups in the same way that there are subgroups of ovarian carcinoma associated with it. I am sure there are going to be subgroups that have quite strong genetic components. In autoimmunity and in psychiatry and a number of other disorders, type II diabetes there have been massive studies of really now hundreds of thousands of subjects for and case control analysis for autoimmunity there are hundreds of genes. We just published 100 genes just for RA alone last week in Nature. Endometriosis has been led by Stacey and Krina and a number of other people around the world but it is still limited to maybe three, four or five thousand cases and controls and really in the world of high throughput genetics that is not there yet. The more you do, the more genes you find because there are a lot of genes that contribute a small level of risk to these disorders and while the level of risk that they confer may be small the insight into what the pathogenesis is can be enormous.
One of the soft underbellies of modern genetics is that it hits on these big genome wide screens, leads to the publication of “genes that are involved with disease” but frankly many times they miss several genes that could be candidates in the regions and very often the causatives changes are not known. In fact far less than 10 percent are even in coding regions of genes. A lot of these are likely to be regulatory in nature. We definitely do expand this.
I would just like to emphasize to this group particularly there is something very valuable about finding families where there is a lot of endometriosis running in the families and it seems to be more highly penetrant. Those are invaluable families for really identifying highly penetrant disorders, highly penetrant genes for this disorder. I think populations where there is a lot of consanguinity can be extraordinarily valuable for identifying recessive alleles. I think that is one way the geneticist community could branch out further in addition to just expanding the sample sizes.
The other thing of course we have heard today is that the phenotypes are incredibly heterogeneous as are autoimmune diseases. A disease like Lupus has these protean manifestations is highly genetic and in fact may be actually one of the less heterogeneous genetically. Whereas rheumatory arthritis, which people think of as being a thoroughly clean vanilla disease, probably is quite heterogeneous in terms of its underlying genetics. I think one of the things that has driven success to the extent we have had in autoimmunity is to not look at the genetics of the disease but to look at the genetics of what we would call intermediate phenotypes. We are very lucky in autoimmunity to have access to the tissue that is involved, namely it circulates in blood, so we can take it out and ask about intermediate phenotypes that the T-cells and B-cells and monocytes might exhibit that relate to these genes. A corollary of that is that one way I tend to think of autoimmunity is actually a compilation of normal intermediate phenotypes that are coming together to a certain level of threshold that leads to clinical disease.
We have lots of examples of this now where the disease genes that we have identified for autoimmunity actually change rheostat for responsiveness for the immune system. It is not an on/off switch it is just slightly shifting the dose response curve to whatever the biologic stimulus is. As an example of that and how complex it can be I want to use the example of PTPN22, which was really one of the first – it is a phosphate, it is an intracellular phosphatase. It has a relative risk of about 2RA for Lupus, type I diabetes. After HLA it is one of the strongest risk genes that is reasonably common. What it does, what the variant does, it is an amino acid change. It changes the binding of PTPN22 to another kinase and this whole thing acts basically as a rheostat for signaling of T-cell receptor, B-cell receptor and probably a whole bunch of other things.
The interesting thing is that the effect of this mutation is to reduce the strength of this association so it changes the level of the rheostat. In the case of Lupus and RA that confers risk. In the case of Crohn’s disease it actually confers 50 percent protection! And in the case of several infectious diseases it causes relative resistance. That speaks to the fact well why is this a variant in the population? It is probably here because it conferred benefit. It has been selected for under some conditions there is actually an incline of increasing frequency in Europe for this particular allele is quite high in Scandinavian populations. It is not seen at all in Asian populations. It is likely that this allele is in the population because it has done something for us maybe it is resistant to the plague back 1340 or other diseases. One could well imagine it is survival with fetus at some level could be selected for under some conditions. I think we have to think of the alleles that associate with disease as being quite complicated in their action and they are going to tend to act on critical components of the disease, which may be multiple in nature.
Obviously we have heard a lot about Sampson’s Theory today and that immediately appealed to me when I found out about it because the idea – one of the problems we have in autoimmunity yes we have access to all these cells in the peripheral blood but there are 50 different cells that could be interesting there. That you could make up a plausible story about why this gene, PTPN22 for example, which is expressed in almost every hematopoietic cell – you could make up a story that actually the cell of importance is not the T-cell but a monocyte or dendritic cell. We have a problem of an embarrassment of riches for potential cells.
In the case of endometriosis if you buy into this hypothesis of the stem cell it obviously directs you to looking at what is going on in those stem cells. Obviously there is more to it than that, stem cells get into the peritoneum. We just head that it changes the whole biology not only of the peritoneum and the immune response to the peritoneum but maybe of the endometrium itself, which is something that I had not considered and is quite fascinating. Of course there is the biology of pain perception and that can be peripheral, central and so forth.
Really, the focus from my point of view of taking my lessons from autoimmunity is to focus on an intermediate phenotype that is plausible and doable for this disease. And that is an intermediate phenotype that relates to the stem cells, they could be genetic in origin, environmental in origin, there could be epigenetic changes that maybe you are going on because of that peritoneal inflammation. And to look at the phenotypes that we think relate to how the disease acts and it is those phenotypes that then coming together in the presence of other factors will lead to full-blown clinical diseases we recognize. The details of these phenotypes can be quite complicated. This is a signaling pathway that Linda Griffith reported about a year ago looking at sheddases and its control of signaling of membrane tyrosine kinases. You can correlate that with the levels of invasion of these cells by treating the cells with different things, and in fact if you do a principle components analysis, which is something similar to what Linda was talking about, you can actually divide cells into groups according to whether they came from endometriotic or not endometriotic lesions. The ability to have a cell that you can interrogate in this way is incredibly powerful because it is one kind of cell. It is not the peripheral blood which has 50 kinds of cells. I have been extensively involved in gene expression studies over the last decade of peripheral blood in autoimmunity and it is extremely challenging. There is probably only one good example of it being successful and that is identifying the role of interferon in Lupus. Having a single cell to look at is great and in fact you can get the cells. We are now getting those cells growing very nicely out of menstrual effluent. We have developed a few tricks to having that happen well and so then we have documented those cells, they are growing, are consistent with mesenchymal stem cell. We have now started differentiating them to show that in fact they do have stem cell properties.
The idea of ROSE, which was briefly introduced this morning, is to be able to apply this approach to phenotypic characterization on a population level. You cannot do that if you have to biopsy the endometrium. My thought was that we could collect menstrual flow, menstrual effluent from endometriosis patients and controls and obviously we have just learned, or I have just learned more than I was thinking about, of how important those controls are to characterize. We can also take people who do not appear to have disease and look at genetic risk carriers and non-carriers. That is something that we have taken huge advantage of for autoimmunity.
We have a population of two or three thousand people at North Shore that have been fully characterized for 200,000 markers related to autoimmunity and we can call those people in on the basis of those markers and ask does the T-cell behave differently when it has PTPN22 or not? We can do the same thing with endometrial stem cells derived from menstrual effluent. We can take people who have the Wnt4 haplotype or the Wnt4 non-risk haplotype and ask are they behaving differently in whatever assay we are looking at. Whether it is an assay of cellular behaviour or an epigenetic assay or a gene expression assay or what have you, whatever the phenotype is we can relate it back to the known genetics. That is why it is so important to expand our number of genes, expand the populations that we can study. Fortunately I was a bit nervous about his having no experience with this that we would succeed but in fact people are very compliant with using these diva cups, putting it in overnight, sending it in by FedEx. You send it FedEx 24, 36 hours these cells grow just fine. They are happy. It allows you to repeat sample, make sure this phenotype is still there. Obviously if phenotype might change pre and post endometriosis one could imaging doing a cohort study of people who are now not symptomatic and then looking and seeing if there is a change in the function of these cells.
The second part which, you know Tamer was all excited about when he came to see me, was, “Let’s get the lesions. I’m getting you these lesions I can really define them well and I’m overwhelmed with complexity of these lesions”. I sort of backed off a little bit but I think we are going to do that because they are very complex. I think in the end it requires an analysis of single cells, I have had a lot of experience looking at complex cell populations with these methods and it is really hard to disentangle the message. What has happened recently is the development of single cell approaches and we have just got this thing called a fluidigm machine in the lab, which basically is a little microfluidics device, which allows you to feed it a few hundred to a few thousand cells. It will isolate those cells and if you want to click on the video what happens is you deliver the cells – this machine is only about this big. So what happens is this microfluidics device, which is only about this big, you feed it the cells, the cells go in independently. Each capture site is labeled with a number so that after this capture occurs you can take it to a fluorescent microscope. You can see that the cell is actually there, number 79, and in fact has the service label that you want to study so you can select which cells are there or you can correlate it with their phenotypic properties immunologically. It will focus the cell and then will lyse the cell within that chamber and extract the RNA from that single cell. You enter these chambers where first the RNA extraction happens and then it opens up a chamber, you make CDNA from that RNA and then you amplify that CDNA with a set of common primers. What you end up with is at the end the total RNA from that single cell. You can do that for 96 cells at once and then you can screen it by PCR or depot RNAC. There are efforts going on to do this for proteomics. I think this offers tremendous power to look at complex cell populations and I think it will be really fun to apply it to selected lesions starting with relatively simple lesions that we think are going to be able to tell us something about what the nature, for example, of the stem cells are and what kind of monocytes are in there. Obviously it is quite well known that mesenchymal stem cells can influence a type of macrophage that one responds with whether M1 or M2 macrophage that we can look at that within these lesions and whether those predictions are correct.
I just want to thank Christine, actually when I started getting interested last spring I went to Christine and found that she actually thought about this disease and had actually been working on MIF in this disease. That was another benefit aside from that fact that I have worked with Christine for many years. Maya Tevlin is in the back of the room, she is a new scientist who has come from Rockefeller. She had been at Rockefeller for seven years or so. She has just come to start working on this. She has experience with stem cells and has really already developed approaches to doing this. Mary Keogh and Margaret DeFranco are the nurses who were recruiting for this. They are also in the back and I have to thank Tamer for having got me interested in this and introduced me to all the wonderful people at this meeting. So thank you very much.
Dr. William H. Rodgers: I would like the speakers to come up and we have a few minutes for questions. Just to get things started I would like the speakers to comment on – I think many of the speakers have mentioned that they believe that inflammation has a lesion that prevents the elimination of endometriosis. Can any comment on the data supporting that and whether it is possible that inflammation may be permissive for the formation of endometriosis, i.e. that the lesion may be in the inflammatory or immunologic pathways maybe ramped up rather than dysfunctioning, and that is the reason that lesions develop in women? I think that would affect the types of controls and other materials that you might want to use to deal with immunologic studies.
Dr. Akoum: Actually it has been shown that endometrial cells from women with endometriosis are capable of escaping the immune system, so they are resistant to the immune system to…cytosis for instance. This has an explanation why these cells resist and trigger inflammation into the implantation side. This might be an explanation.
Dr. Fazleabas: Part of it also is they take up metastatic properties and as a result they change their ability to, and Peter alluded to this a little bit, is in terms of their immune surveillance. Stuff we published before if you look at the T-reg population for example within the eutopic endometrium versus the ectopic endometrium you see a complete reversal. The eutopic endometrium has a T-reg population that reflects infertility. But the ectopic lesions reflect a T-reg population that is very similar to cancers which are TGF beta driven. I think once that tissue comes up the environment, attaches to the peritoneum even though it is endometrium in origin, it takes on a whole different phenotype.
Dr. Linda Guidice: I agree with that.
Dr. William H. Rodgers: But is it the implantation that is causing the change in phenotype locally or is it that there is a predisposition that the person has to develop that phenotype during implantation? Because all women probably are exposed to retrograde menstruation but only a subset of them develop it. I do not think you can distinguish from the data that is available whether the inflammatory or immunologic changes is a lesion in its degree of response to elimination of lesions or whether it is a response is that it develops and stimulates the development of lesions once they occur. Because it really is a tissue formation process it is not a single cell type. It is forming a lesion, it is stroma and glands and the cell type is differentiating into a tissue. In many animal models embryogenesis macrophages for instance are necessary to be present for things like limb regeneration in amphibians and a number of other animals that regenerate. They are always associated with dramatic inflammatory processes. They are permissive rather than the opposite.
Dr. Christine Metz: One possibility might be the stickiness or the adhesive properties of either the seeding cells or the sites where the cells seed, which may be unrelated to inflammation or related to inflammation.
Dr. Peter Gregersen: I am dying to ask Asgi if there are particular female baboons that are good for this and others that it takes a while for the lesions to develop?
Dr. William H. Rodgers: That was my next question.
Dr. Asgi Fazleabas: This probably goes back to your initial question. No, so far we have never had a problem. Basically, you get different gradations of disease in terms of the number of lesions but I think we just overpower the immune system in terms of being able to clear out the menstrual debris, putting in .8 grams of this tissue, it just overwhelms the immune system – it just attaches. The only difference you see is basically by the number of lesions but they all get lesions.
Dr. Peter Gregersen: It seems to me that that is a very powerful…to ask, how important is the blood in having this? If you just put in the purified stem cells would it work? I realize with a big animal that is a big commitment to do that kind of experiment.
Dr. William H. Rodgers: There certainly are more questions than answers here. I want to get things started. There are questions from the audience.
Audience Member: I have really enjoyed all of the presentations that have been made today and I would like to touch on an issue that has not been raised so far in this meeting but was touched on briefly by our keynote speaker at lunch. I sit on the AMA Commission that looks at issues of race and ethnicity and its relationship to a variety of diseases across the whole spectrum of health care and issues related to health care disparities. I have seen numbers of patients in my 35 years of practice on whom I have made a diagnosis of endometriosis but were followed by other physicians who almost reflexively made a diagnosis of chronic PID and did not even consider endometriosis in their differential, again based on their race and ethnicity. My question is are there any studies, definitive studies that have been done that show any type of relationship, differences in terms of blacks and Hispanics between white population as it relates to endometriosis? Again, I have seen more adenomyosis than actually endometriosis. My second question is in the ongoing research that is going on in this very complex and interesting disease entity are there any blacks and Hispanics included in patients being studied?
Dr. Linda Guidice: That is a great question. Stacey Missmer, I think she is the best person to answer the question.
Dr. Stacey Missmer: In the cohort studies, including the medical professional cohort studies there is a significant, about 40 percent lower rate in women of color and in…than in Caucasian women. However, it still does not address the issue. These are still women who are achieving a surgical evaluation and diagnosis. It does not get at diagnostic bias issues. There…populations there is a threefold increased rate of fibroids – we have done studies asking women for self reports of endometriosis. We find rate geographic regions and racial ethnic differences in knowledge of endometriosis and sort of a community dialogue around all the pain and endometriosis versus endometrial polyps and endometrial biopsies where even fibroids are more uniformly known and discussed and understood. So, there are a lot of factors that I think come into play. But the ideal study, the ideal data collection has not been done yet.
Dr. William H. Rodgers: So, it sounds like very limited data but it is likely that there may be some differences so far. Other questions?
Dr. Liselotte Mettler: I just wonder to look at this wonderful picture that we have there of Hero or Heres covering the eyes. I do not know if it is male or female but presumably a female. So where are we with all the suggestions we have received as I think especially this panel shows things that we have heard 20 years, I mean it is really the genes have been described ten years ago by now they are some more. What really compels you, what is really catching? I did not get this from all the lectures. What is really catching and bringing us ahead?
Dr. William H. Rodgers: Well, I can say what excites me, just very briefly, is the availability of tools that I think can make it possible to answer some of these very difficult questions. We have examples of very powerful genomics that can be brought to bear on endometriosis. We have evidence that there are tissue banks that are available in which you can do those studies. There are some animal models that are well characterized that can be used to test some hypothesis. So I think there are the pieces of the puzzle that are needed to understand what the process is. Endometriosis like arthritis and many other multifactorial diseases are likely very complex diseases. It takes time for science to learn how to deal with that level of complexity.
Dr. Peter Gregersen: Actually I am surprised to hear you say that because I do not think you would have heard the level of sophistication even five years ago using some of the technologies of looking at large data sets, having real mathematicians look at those data sets in various ways. I think Linda just touched on what is going on and there is an enormous amount going on that we can take this multi-level data and analyze. I think this business about being able to isolate and characterize single cells is a year old. I have the sense that there is just an explosion of opportunity that is really quite new applied to this problem.
Dr. William H. Rodgers: It is quite exciting and it is not very much different than the kind of progress that has been made in understanding breast cancer or other cancers. The same quality and type of studies that have been done but they are not popularized as much. You are not reading about them in the New York Times or in popular press. Advocacy is a reason why, it is surprising.
Dr. Stacey Missmer: To follow up what you were just saying…even two years ago, even a year ago, there are many of the tools now being applied to endometriosis because of conversations and increase in the multidisciplinary interest that did not exist. Many of these tools as you just said in breast cancer, in diabetes, immune disease, cardiology has been well utilized and developed into new systems biology techniques that are now being applied to endometriosis. I think we will see an explosion in the next five years of discovery.
The other thing, and this pulls in the theme from the morning that we discussed a lot is that we have to embrace the phenotypic variation in endometriosis. Almost every published study…across endometriosis as one entity and as was discussed in detail this morning we know that ovarian cancer has many subtypes, they have different molecular characteristics, they have different treatment response characteristics. Breast cancer – it is only eight years now that we have understood there was triple negative breast cancer. Individualized medicine is completely dependent on the molecular profile of breast tumor specific markers and looking at breast cancer you would not see differences. We know, even to a lay person who knew nothing before this morning about endometriosis just looking at lesions you know there are differences. Collecting that detail difference is the difference in symptom presentations and what that means is the next important step. The only way to get that is through multicenter large collaboratives.
Dr. William H. Rodgers: I think we have time for one more question.
Audience Member: I think, following what you are saying, that is going to require a lot more funding and I think that has to be the lead focus because they are all out of tools to apply but they all cost a lot of money to….My question to the panel is in thinking about this problem transitioning from a surgical to medical treatment, and this can be both clinical experience and in the lab in terms of what you see in the models, how much reversibility is there? Because it seems like as the disease progresses you get sort of this fibrotic mass that at some point you have to remove and since the incidence really seems to start in the younger population but the samples that we are studying are about the time the patient presents – when in your thought can intervention happen and thinking medical treatments or maybe even in physiological cases where the patient treated themselves or they had symptoms but they do not anymore, what are you seeing and what can we expect should we be thinking about?
Dr. William H. Rodgers: I think that is a very good question. I think restated another way the treatment of advanced endometriosis is maybe different than prevention. Can the panelists comment?
Dr. Linda Guidice: For minimal/mild disease there are some data of sort of serendipitous laparoscopies or second looks without treatment, some older studies that show that there is some reversibility. So just because the disease is there does not necessarily mean it is always going to be there. However, when we start thinking about prevention I think it is really important that we understand what the cause is before we can start even thinking about prevention. You might want to consider just say the administration of contraceptive steroids non-cycling to teenagers but I do not know that that is really the answer to prevent endometriosis on a population level.
It is a really good question but I think we are not quite there yet.
Dr. William H. Rodgers: One more question?
Dr. Liselotte Mettler: …I think we definitely learned that just stage one and two as Linda just mentioned we do not have to be so aggressive with surgery as we have been in the last 20 years. That is my opinion. I think, I do not know...stating that. We should really be keeping back with the surgery not have young patients have five laparoscopies by age of 25.
