My name is Jillian Wilborn and I've been working for 11 months as a research intern at the University of Wisconsin-Madison in the Zou Lab. Before I begin, I would like to take the time and thank the moderators and the organizing committee of the Endocrine Society Annual Conference. This is such a wonderful opportunity to present my work and get some feedback from all of you guys. Today I'm so excited and honored to be discussing bilateral and unilateral malformations of the male reproductive tract in mice with androgen receptor deletions in mesenchyme. Let's begin. I would like to start this discussion by going over the normal developmental program of the paired bilateral male reproductive tract. In the earliest stages of development, an embryo possesses two sets of gonads. They possess the primitive female tract known as malaria duct, shown in pink, and the primitive male tract known as the wolfine duct, shown in blue. It is important to note here that when I say the words paired and bilateral, it means these structures are developing at the same time on both the left and right sides. In male mice, the testes will develop and begin to secrete androgen. This androgen secretion drives the morphogenesis of the wolfine duct into the epididymis, vas deferens, and seminal vesicles. Knowing this, we are able to establish the idea that androgen action drives the morphogenesis of the wolfine duct. Now you may be curious, how exactly do androgens exert their action? My project is specifically focusing on the epididymis because it is known to play a vital role in human male fertility. Looking at the fetal stage, androgen will bind to its androgen receptors in the mesenchyme, which I have shown in purple. This will then go on to influence the development of the epithelial layers, which I have shown in blue, which leads to the epididymis being coiled during the adult stages. My project is specifically focusing on the action of the androgen receptor in the mesenchyme, and how exactly it drives the morphogenesis of the wolfine duct. I would now like to discuss how we generated our androgen receptor knockout mice, which we needed to generate in order to investigate this question. Our lab identified a mesenchyme specific Cree, OSR2 Cree, and bred an OSR2 Cree male mouse with an XAR flocks, XAR flocks female mouse. It is important to know here that the AR gene is X-linked, so our AR knockout male offsprings only need one copy. In this situation, the OSR2 Cree will go in and locate the androgen receptor gene, which is found in exon 1, and remove exon 1 located between our two inserted LOXP sites. This cross then generates a 25% chance of having an offspring that is positive for OSR2 Cree and an XAR flocks Y male mouse. Now you may be wondering, what is the normal process of wolfine duct coiling? 14.5 days into embryonic development, or E14.5, the wolfine duct is stabilized in male mice. Two days later, at E16.5, coiling is initiated. This coiling continues past E18.5 and into postnatal development. We are now able to address the consequences of ablating the androgen receptors in the knockout mice. As many of you guys know, as we are dissecting, we are looking down at the mouse. So the mouse's left and right sides are opposite of our left and right sides, which you can see indicated in our pictures coming up. We began dissections at PND0, because at this stage, coiling has only already happened for several days, which makes the phenotype easier to observe. Looking at these structures, you can see there's a normal coiling pattern in both the head region and the tail region. This structure is normal, a normal pattern on both the left and right sides, and we observed this pattern in 23 of our control mice. Now I want to look at our androgen receptor knockout mice, of which we observed three overall phenotypes. The first we observed, which is what we had expected, was abnormal coiling occurred on both the left and right sides when you ablated the androgen receptor. In these images, you can see that the cystic structure is really located towards the head region, or the caput region, and the tail region, or the cata region, is characterized by degeneration. We observed this phenotype in seven of our original sample collections. Looking at our next phenotype, we observed that only the right Wolfian duct had this abnormal coiling pattern. Once again, you see the abnormal cystic phenotype located to the head, and you see degeneration of the tail, and this was observed in only one of our mice. The third major phenotype we saw was abnormal coiling only occurred on the left side, but if you compare the right side in this androgen receptor knockout mice to this control, you can see that a normal coiling pattern is maintained throughout the head and tail regions, and this phenotype was observed across nine of our mice. We then wanted to confirm via an immunohistochemistry procedure that our androgen receptors were properly ablated in the androgen receptor knockout mice where the side was developing normally. Looking at the control, you can see that the mesenchyme has the androgen receptors shown by brown, and you can see that the epithelial cells also have the androgen receptors. Looking at the left Wolfian duct, you can see that the androgen receptors have properly been ablated, and you can see that cystic coiling phenotype that I just showed you on the previous slide. Then, looking at the right Wolfian duct, you can see that there's a normal coiling phenotype despite the absence of androgen in the mesenchyme. We also wanted to confirm that the CRE we used, OSR2-CRE, was active on both the left and right sides in these knockout mice, so we generated another test cross where we bred that same OSR2-CRE mice that we used to create the androgen receptor knockout mice with a Rosa TD tomato female. In this situation, the CRE goes in and removes the stop codon that is located before the TD tomato gene, which allows the TD tomato gene to be expressed, and thus we can see expression in our offspring, and that's the cross. So, we then performed another IHC procedure where we observed TD tomato expression, which is indicative of CRE activity. We specifically decided to look at the E14.5 stage because, as I previously mentioned, this is the stage where the Wolfian duct is stabilized before coiling is initiated. In this slide, I have labeled the Wolfian duct, shown in blue, and you can see there's indeed a presence of TD tomato expression, which means that CRE is active at this stage. These results show that our androgen receptor knockout mice who displayed the normal phenotype on that one side, that pattern was not due to the inactivity of CRE or the expression of androgen receptors, but rather some other mechanism is taking place that's allowing for that phenotype to be normal. I would now like to summarize what we have discussed today, in that our experiments have shown when you ablate the androgen receptor in the mesenchyme, three main phenotypes are observed. 39% of our original sample collections had abnormal coiling on both the left and right sides. 50% of the time, abnormal coiling only occurred on the left side, and 11% of the time, abnormal coiling only occurred on the right side. You may not be surprised that these results left a lot of questions in our mind, and the question that I'm most interested in investigating is whether or not the testes are responsible for creating some sort of compensation factor that allows proper development to take place in our knockout mice. In order to investigate this question, I will perform an ex vivo cultural experiment where I will physically separate the testes from the epididymis and grow these structures in two different petri dishes or regular dishes. In this situation, if the phenotype is abnormal on both sides, and that's always the phenotype we observed, we will then move on to performing an RNA sequencing experiment where we will compare the transcriptomes of both the left and right testes. In this situation, if one of the sides develops normally and it has a factor that is significantly upregulated, then we will move on to investigating that factor and seeing if it's indeed the compensation factor that's a lot that's present in the side that develops normally. If this is not the case, I will then compare the transcriptomes of the left and right wolfing ducks for the same reasons. Last but not least, I would like to thank all the members of the Zao lab, Dr. Fei Zhao, McKenna Crossan, Shuai Jia, and Akhil Madanini. Without the help and support of everyone pictured up here, I would not be standing here today and presenting in front of you guys. So thank you, and I will now open up the floor for questions. The presentation is open for discussion. Hello, good morning. Rodolfo Reyes, Buenos Aires Children's Hospital, Argentina. This is a question of ignorance. I would have expected that the epithelial androgen receptor would be involved, given that the expression seems to be pretty higher than mesenchyme. Yes. But you hypothesized that testis factor is compensating. Why not the epithelial androgen receptor? So, as I mentioned, our CRE that we're using is a mesenchyme-specific CRE, so it really is only knocking out that mesenchyme androgen receptor. I'm really curious about the testes, because I think that's the question that we can address first and easiest, and so if we ablate the testes and the androgen receptors, or there's a normal coiling phenotype on only one side, then we can move on to addressing the epithelial cells. But short-term, future direction is really looking at those testes. That's a great point. Okay. Do you think that in cases where this is not a syndromic feature, endocrine disruptors can have any role? That's a great point. Yeah. We haven't specifically looked at endocrine disruptors yet, but I've been talked to or approached by several other researchers who have suggested that, so we're looking at finding a collaborator to potentially address those questions in the future. Great. Any other questions or comments? Okay. Thank you. So, our next speaker will be Dolores Lamb. She'll be presenting on FBFox2 copy number variation disrupts urethral development, causing hypospadias. Hi. Thank you very much. So, these are my disclosures. None of them are relevant to the topic of this talk. So, I'm going to talk to you about genital tract anomalies in humans. So, these are the most common birth defects. On the left side, we're seeing the lower tract anomalies, and there are also, of course, anomalies of the upper urinary tract. And collectively, these anomalies often come in groups in patients, and this is because there's a common embryonic origin, which is the intermediate mesoderm. So, the story that I'm going to tell you about today has to do with copy number variations. These are structural chromosomal anomalies that cause micro deletions or micro duplications, and it's a structural chromosomal anomaly. And if genes are either micro deleted or micro duplicated that are dosage sensitive, it depends on how many copies of the gene are present in the genome, then that impacts the amount of messenger RNA and protein expressed in a cell. So, we had done an analysis of the Decipher database, which is publicly available of patients with a wide variety of different anomalies, and we used a series of very stringent criteria to select candidate genes. And looking at this database, early on, we found 17 hypospadias patients with copy number variations at 22q12, which encompassed the RBFOX2 gene. And this is showing you just a smidgen of the data that we looked at, where we're using locus mapping to try to define the minimal region, because many times, multiple genes are affected in these patients. And so, again, it was RBFOX2. This was the thesis study by Marisol O'Neill. And so, RBFOX2 is a very interesting gene. Quite a bit is known about it. This regulates mesenchymal to epithelial transition, and it does so by messenger RNA splicing. And this information comes to us from studies of both heart and brain development. And important for our study, it's known that this does so regulating mesenchymal to epithelial transition by controlling FGFR2 messenger RNA splicing. And for us, this was very important, because we know that defects in the splicing to the different isoforms of FGFR2 impacts male genital urinary development, resulting in lower tract and upper tract anomalies. So we tested the hypothesis that these RBFOX2 copy number variants, gene dosage changes, alter normal penile development, affecting FGFR2 and other important transcript splicing in the developing penis. So we first looked at the incidence or prevalence of RBFOX2 copy number variations and showed that they were enriched in non-syndromic genitourinary patients. These are pediatric patients that we got from our cohort at Texas Children's. We also looked at patients without gene anomalies. We looked further in the Decipher database of people with various forms of different anomalies. And you can see that in the database of genomic variants, which consists of the general population of individuals with and without gene birth defects, that there's about a 0.1 percent incidence of copy number variations. Very different than what we saw in the pediatric patient with geoanomaly cohorts. So it was tremendously enriched in these patients. This was statistically significant. So as I told you, RBFOX2 regulates alternative splicing, and it does this by being an RNA binding protein. There's a bimodal alternative splicing that occurs that I show you on the right side of the slide. It's important because it's involved in stem cell survival, pluripotency, and differentiation, as well as epithelial versus mesenchymal cell identity. And again, as I told you, RBFOX2 balances the FGF isoforms. And what you're looking at down in the in situ hybridization at the bottom of the screen is work done by my pediatric urology fellow, Jeff White, where he used in situ hybridization to localize, let's see, RBFOX2 in the green and FGFR2 in the developing. And this is a normal mouse penis between embryonic date 14.5 to, I think it's 18.5. And you can see that RBFOX2 is highly expressed up towards the glands approaching where the urethral meatus will be. It's also highly expressed in the developing urethra that is about to be undergoing urethral tubulization. And you can also see, I hope, that some of this is co-localized by the change of the color between the combination of the fluorescence of the red and the green shown together. So again, both of these proteins are co-expressed in the condensing urethral mesenchyme. So then we used confocal microscopy, which was work performed by Victor Ruthig in my lab, who's a postdoc. And here we're looking at RBFOX2 protein, FGFR2, and androgen receptor localization in the developing mouse penis, again, between embryonic date 14.5 and 18.5. So on the glands, RBFOX2 in the green is shown clearly as sort of an epithelial cap. There's also a region of encasement which aggregates at the glandular urethra formation. Now, if we look further at the urethral development itself, you can see that there's a ring formation here, which then migrates approximately during the development of the urethra over between embryonic days 14.5 to 16.5, and eventually ends at the tip of the glands where the urethral meatus will form. In the red is the androgen receptor protein expression. You can see it's widely expressed throughout the whole genital tubercle, but there are areas of very heavy expression. And here on the bottom, you can see the cells where it's colocalizing both in the glands as well as in the shaft of the penis. And this happens to be the ring formation here, which you can see very clearly. And there's also, if we look, there's kind of a net of RBFOX2 expressed over the developing glands of the penis as well as down the shaft. So we created a RBFOX2 haploinsufficient model to mimic the CNVs found in our patients. This was helped very greatly by Tom Cooper at Baylor sharing his RBFOX2 floxed mouse with us. And, again, Marisol O'Neill generated these haploinsufficient mice. Jeff then, together with Marisol, extensively phenotyped these animals using micro-CT. And so here you can see hypospadias present in the developing genital tubercle with the open prepucial foals. And you can see this more clearly over here. I'm sure you're perhaps not used to looking at this type of micro-CT. I think more clearly you can see also that there were other phenotypic anomalies, bilateral hydro-ureteronephrosis. So you can see very clearly the hydronephrosis in the kidneys, right, and, again, present in all of these animals that I'm showing you here. We also found that on the postnatal day one mouse that there was significant bladder distension. So what you're seeing here is bladders filled with urine of these mice because they are unable to urinate. And this, as well as other problems, causes their demise shortly after birth if they survive that long. So then we looked at the human phenotypes to see how the mouse phenotypes could compare. So we had both our copy number variation data, as shown there, right here. We also had whole exome sequence data on a number of these patients. And you can see that as we look across the various types of defects that we had really superb correlation of the phenotypes between the mouse model and the human condition. And so the mouse and the human phenotypes of either the RBFOX2 copy number variations or damaging mutations in RBFOX2 were consistent. So we saw, again, genitourinary anomalies, both upper and lower track. We saw intellectual disability, cardiac defects, which were known, facial and head anomalies, as well as skeletal anomalies and abnormalities of the kidney and urinary tract. So then we looked at the RBFOX2 transcript targets, and they are regionally expressed in the penis. We had RNA-seq data that Marisol and Jeff had worked on. So here we had data from the RBFOX2 null mouse penis compared to wild type. And, you know, they identified 268 significantly altered candidate genes. RNA-seq data, single-cell RNA-seq data was available to us, publicly available to all of us by Dr. Humphrey Yao's lab and Marty Kahn's lab. And so we were able to use that to look more closely at some of the RBFOX2 loss gene candidates based on what was found in the cardiac studies. And we were able to show that they have both normally expressed epithelial versus mesenchymal temporal regionality in terms of the expression based on this single-cell RNA-seq data. This was a special study done by a high school student who actually did programming and everything else to come up with all of this data. And this was done with Victor's mentorship as well. And again, the box here has shifted, it has to go up a bit. But those are many of the candidates that we identified. So again, as I told you, when we did RNA-seq, we saw differential gene expression of a whole series of genes, upregulation here, downregulation here. Importantly for us, some of the top ones that were upregulated were estrogen late response genes, as well as androgen response genes. So these were among the highest upregulated genes. We also saw a differential transcript usage from the mutant versus wild type with changes in the transcript coding potential as well as the transcript domain status. And interestingly to us, 20 years ago, Donald McDonald's lab identified a protein, which they called RTA, today it's known as RBFOX2, which interacts with the estrogen receptor alpha and beta N-terminal activation domain. And it led his lab to suggest that this protein, RBFOX2, is a co-repressor of steroid receptor family. So in conclusion, I've shown you that 22q12 is a hotspot genetic locus for congenital GU anomalies. RBFOX2 is the best candidate in this region. Patients with GU anomalies are more likely to harbor RBFOX2 gene dosage changes relative to the general population. And certainly, RBFOX2 is both regionally and temporally expressed in the developing murine GU tract. And the FGFR2 has overlapping expression profiles in the urethra. Importantly, RBFOX2 haploinsufficiency results in a highly penetrant phenotype, including hydroureteronephrosis, hypospadias, urethral malformations, among other problems. And the RNA-seq data of the penis of the RBFOX2 lost mice, again, showed differentially expressed genes and differentially spliced genes that led to isoform changes in a whole slew of different proteins. And so this is just to acknowledge the support of my laboratory, both by the NIH and by the Dow Foundation, as well as all of the individuals in my lab who contributed to this work. Thank you. Hi. Hey, Dory. Thank you for this beautiful data. I was wondering, and perhaps you said it and I might just have missed it, at the beginning when you were looking at these regions that had been in this Decipher database, were the regions mostly deletions, were they mostly duplications, was it 50-50, and regardless of whether you were deletion or duplication, did you all have the same phenotype? So many times, some of that is hard to distinguish. So remember that you usually have a reciprocal micro-deletion or micro-duplication syndrome. Some of them are impactful, some of them are not. Complicating the story is that sometimes, for example, you could have a micro-duplication in a region that disrupts the gene, right, that is the informative one for us, say, RBFOX2. And so sometimes you have overlap, right, with the phenotypes. Sometimes you have related phenotypes that when you go back and figure out the signaling pathways, for example, if something's impacting Wnt signaling, if you have, you know, a decrease, right, in the one protein competing for this Wnt signaling with Wt1 for Wnt signaling, you may see one phenotype and then you see the opposite of what you would expect, say, with deletion of, say, Wt1 on GU development. So sometimes it's a yin-yang and sometimes it is not. Remember that here, I would say, we had both micro-deletions and micro-duplications, similar body parts affected. Some of the individual phenotypes were slightly different, right, I'm saying in this particular gene, but affecting the same developmental body parts. Great. Thank you. Good. Thank you. Hi, Dory. Nice talk. This is Joni Jorgensen, University of Wisconsin. Thank you. Yeah, I just wondered, looking at your careful consideration of where this RBFOX2 is expressed in the urinary tract development and the kidney and such, do you have any thoughts as to what regulates its expression? So yes, actually, we do. We've actually identified the upstream regulator, and fascinatingly, and we don't have any of this published yet or even totally worked out, but fascinating, when you have deletion of that gene, we see the same phenotypes that we see with our published data on MAZ, on CRKL, BAMP7's a little bit different, that one's androgen action, but again, it's an E2F1, that's the other one. And so it's a, we've identified the one above this that then is then impacting all these other signaling pathways below. Thank you. Yeah, thank you. Good question, Joni. Beautiful talk. I'm Fei Zhao from University of Wisconsin Medicine as well. So I have two questions. One is, did you see sexually dimorphic expression of RBFOX2? And what would happen if you use this methanocytic-specific creatine to knock out this gene? Would you expect the same phenotype? So we did not use, so the second part of the question, we did not do that because we were trying to recapitulate our patient phenotypes, right, and it's affecting every cell in their body. So that's why we did not do that. In terms of the sexual dimorphism, interestingly, we haven't done enough work on RBFOX2, but I can tell you that for almost all the genes we've identified, again, in these signaling pathways, we see female phenotypes as well, sometimes not what you would expect. And I think that the other kind of take-home message for this is, you know, the physicians and urologists all think, oh, this is simple, say, with hypospadias or cryptarch, it isn't easy for me to fix. Nobody pays attention to these birth defects, including geneticists. So that's why it's poorly reported in the literature. In the more recent data for some of this, we have a lot more, say, hypospadias reported for the different genes because people are now becoming more mindful that these kids, a lot of them are actually syndromic. So good take-home message. Great. Thank you. So now we'll move on to Rosella Canamella. So Rosella will be speaking on sperm-carried IGF2 downregulates mitogens released by Sertoli cells, a paracrine mechanism of spermatogenic regulation. So thank you so much. Thank you to the moderators and to the committee of the Endocrine Society for having gave me this wonderful opportunity to discuss the results of this research here. So basically, this is a research that has been conducted by me and a group where I belong. We are basically endocrinologists from the University of Catania, but we have deep expertise in male infertility. And basically, we realize in our clinical practice that the prevalence of male infertility is increasingly worldwide, and that accordingly to the literature, about 8 to 12% of reproductive-aged couples suffer from infertility. And a half of cases can be attributed to the male partner. But worryingly, despite careful diagnostic workup, we are not able to identify the cause of the etiology of the male infertility. And a lot of articles have been published in the past, in the last decades, and they address to the idiopathic male infertility. So they tell that as high as the 10% of male factor infertility is idiopathic. But according to other research, for example, this is a research from a German group that was carried out in as high as 20,000 patients suffering from male infertility, and they were not able to identify the cause as high as the 70%, which is something that is really struggling. And the research has been carried out to try to understand the causes, and therefore, thanks to the progress in molecular biology and in the understanding of sperm transcriptome, we are able to see that the sperm carries thousands of RNAs, and that according to recent data, the sperm RNAs are not silent, but they play a role, and they actually play a role in the spermatogenesis during sperm maturation, during also the early embryo development, and it seems that the sperm is able to carry these transcripts to the embryo, and there can affect the early stage of the embryo development. One target that seems to be very promising is the IGF2, because it is a paternally imprinted gene, and therefore, it is transcribed only by the paternal chromosome. And as we know, the gametes have an haploid genome, and therefore, if we look at the biology of the spermatozoa and the biology of the oocyte, we will see that only the sperm is able to carry the IGF2, while the oocyte cannot carry the IGF2, because it is an haploid cell, and the IGF2, as I said, is a paternally expression. So what usually happens is that the expression of the IGF2 is regulated by the methylation of the H19 de novo methylation region, which is this one. The females usually have an hypomethylation of this region, so that the transcriptional apparatus can access the H19 gene and not the IGF2. So what happens in the female is that the IGF2 is not transcribed, because the H19 de novo methylation region is hypomethylated. But in the paternal gene, what happens is that this region is physiologically hypermethylated, and so the transcriptional apparatus cannot access the H19 gene and has to transcribe the IGF2 gene. The studies have reported that the pattern of methylation of the H19 de novo methylation region is altered in the infertile men, and therefore, it seems that the infertile patients have an abnormal low methylation of the H19 de novo methylation region, which may somehow affect the transcription of the IGF2 gene. So the question is, is the sperm able to transcribe this gene? Because we know that the sperm, the chromatin of the sperm is highly compacted, and we were able to think that the sperm was not able to express, to transcribe its DNA. But it seems that there are some regions of the sperm chromatin which are not compacted, and that can allow the transcriptional apparatus to transcribe the genes. And this is particularly true for the paternally imprinted genes. So we tried, firstly, to see if the IGF2 gene is transcribed in the spermatozoa, and we randomly evaluated four samples that were assessed through SWIMAP. The samples were heterogeneous, so we found that we used both patients with normal sperm concentration and count, also patients with oligozoospermia, and patients with normal sperm motility, but also those with astenozoospermia. And we found that the RNA is transcribed in all of these samples. We then looked at the protein, because you know the IGF2 is a growth factor. So if the sperm carries the IGF2 into the oocytes, it is important to know, because being a growth factor, it can enhance the first mitosis. And therefore, if infertile patients may have low levels of this protein, somehow this may explain at least a part of those cases that are considered idiopathic so far. We looked at the protein, and we found by Western blot that the protein is expressed. We used the sertoli cells as a positive control, and all of the samples actually expressed the IGF2. We also looked at the data of the immunofluorescence, and we found that the IGF2 had cytoplasmic, equatorial, and postacrosomal localization, which is compatible with the possibility of the sperm to carry this protein inside the oocyte. Because through the acrosome reaction, what is behind the acrosome can be released in the cytoplasm of the oocyte. And interestingly, we also found that the population, the sperm population, is quite heterogeneous. So this is something that we also know. We already know that fragmentation of the sperm DNA is different among the sperm that we use for intracytoplasmic sperm injection. And again, for the IGF2 expression, there is some heterogeneity, because there were some sperm that was lacking of the protein. So maybe this may reflect some alteration of the sperm in the imprinting, so in the methylation pattern of that specific spermatozoa. And also, there were differently degrees of expression of the IGF2, because the immunofluorescence was not the same in the sperm population. So there was a heterogeneous expression of the protein. Then, we tried to understand if this protein may have a role in the control of the spermatogenesis. So we are not actually understand the impact on the embryogenesis in this data, but we would like to understand if the amount of IGF-2 produced by the sperm may somehow impact on the spermatogenesis. And to accomplish this, we used a porcine model of surgery cells, which were incubated with the IGF-2. We used three different dosages, and we also used a selective and covalent inhibitor of the IGF-1 receptor, which is the receptor that the IGF-2 used to express to do its function, to exert its function. The outcomes were the mitogens factors, so the GDNF, the FGF-2, and the SGF, that are released by the sertoli cells. Indeed, the FSH stimulate the sertoli cells to release these mitogens, which are involved in the switch of the gonocytes from a quiescent state to a proliferative and differentiating state, and therefore they can start the process of the spermatogenesis. So we also tried to see if the IGF-2 is able to impact on the expression of the FSH receptor, and also we evaluated the sertoli cell proliferation. Interestingly, we found that the IGF-2 is able to modulate the expression of the mitogens in our model. Indeed, incubation with the IGF-2 significantly down-regulated the expression of the GDNF, and also of the FGF-2, and also of the FGF. Interestingly, these factors, these results, were completely reversed when we incubated these cells with the inhibitor, thus showing that the effects were displayed by the IGF-2 and the HF-1 receptor. Regarding the FSH receptor, we found that, similarly, incubation with IGF-2 was able to down-regulate both the expression, but also the expression of the FSH receptor protein, and again, these effects were reversed by the incubation with the inhibitors. Somehow, the HF-2 was able, on one hand, to down-regulate the expression of the mitogens, but on the other hand, also to down-regulate the expression of the receptor of the FSH, which down-sensitized, I mean, the FSH itself, enhanced the production of the mitogens. This is another mechanism that makes the Sertoli cells insensitive or less sensitive to the stimulation of the FSH, and therefore to the production of the mitogens. These results were confirmed also by the immunofluorescence. In fact, as you can see, we found that the expression of the FSH receptor was down-regulated when we incubated the cells with the IGF-2, but this effect was reversed to some extent when we included also in the experimental model the inhibitor. There was no clear evidence on the impact of the IGF-2 on the proliferation of the Sertoli cells. To conclude, we demonstrated that the IGF-2 is expressed both in terms of RNA, but also in terms of protein in human spermatozoa, and we were able to demonstrate that the incubation with the IGF-2 resulted in the down-regulation of mitogens, of Sertoli cells released mitogens, and also in the down-regulation of the FSH receptor. This may represent somehow a negative feedback through which spermatozoa can regulate the process of spermatogenesis, because we can imagine that the spermatozoa, when they are in a high amount, release the IGF-2 into the tubules, and this may be received by the Sertoli cells that down-regulate or reduce the amount of release of the mitogens. Maybe this process, this feedback, could be impaired in patients with oligodospermia, so this is just a mechanism that may somehow be applied in those patients with idiopathic oligodospermia to identify if, at least a part of these patients, may have an impairment of this mechanism. So thank you all for your attention. Thank you, Lucella, for the nice presentation. It's open to questions and answers. Hello, Rodolfo Rey, Buenos Aires, Argentina. So you showed us that the mRNA and protein are present in the spermatozoa. In human spermatozoa, yes. In human, yes. In human spermatozoa. And this is also confirmed in mice models. So do you hypothesise that they are transcribed and translated in human spermatozoa, or could it come from spermatids or spermatocytes? So the hypothesis is that at least a part of the amount of the transcript and the protein can be like a remnant of the spermatogenesis, so they can come also from. But it is not the only hypothesis, because basically these proteins come from an imprinted region, and recent research seems to suggest that imprinted regions may be actively transcribed in also, even in the spermatozoa. Spermatozoa is a really fascinating cell, and we have a lot to discover in this cell. So maybe that they may come from an active process of transcription. Wonderful. Thank you so much. Thank you. Oh, one more question. Go ahead. So I just have a quick question. So I'm wondering if you removed the cytoplasmic droplet when you were isolating the messenger RNA, and how you were able to distinguish that it was actually in the sperm and not the cytoplasmic droplet, because, you know, sperm are not transcriptionally active, right? And you don't have the mechanism, right, to be able to be making proteins. Yeah. No. This was not done in this research. So we just separated the spermatozoa from the sperm fluid by swimming, yeah, and therefore all the population was made of sperm cells. And then we identified the protein and the mRNA, and that's it. This is what we've done. Okay. Because I think that's important to look at. And remember that the round spermatids, as well as the pacotene spermatocytes, have like the highest rate of RNA synthesis in the body, right? So just a little bit of contamination there could totally maybe change the interpretation of the data. Yeah, maybe. Okay. Thank you. Now we're going to welcome Amanda Schwartz from the University of Michigan. Active testosterone treatment impairs in vitro fertilization outcomes in a female mouse model for gender-affirming testosterone with improvement in outcomes following cessation of testosterone treatment. Welcome. This one advances me? That should. Yeah. That's awesome. I'd first off like to thank our moderators and the Endocrine Society for this wonderful opportunity to present our team's research today. On behalf of our team, we'll be presenting our gender-affirming mouse model, looking at the impact of testosterone on IVF outcomes. So today, since we're kind of changing gears a little bit, talking about somatogenesis to transgender care, I'm going to go through a little bit of background, and then I'll chat through study design and results, as well as our conclusions and next steps. So a little bit of background here. In general, the gender-diverse population can be very hard to estimate because many studies unfortunately don't include surveys that ask about gender identity. And when it is included, oftentimes we have a lot of heterogeneity between different gender identity choices. However, when we try to extrapolate from the surveys we do have, and using a conservative estimate about 0.5% of the U.S. adult population, extrapolates to about 1.4 million individuals in the United States and 25 million adults worldwide. Whenever we talk about gender care, I think it's really important to emphasize the idea of a gender transition, which is a very personalized decision and can oftentimes reflect individuals' desires, their dysphoria, but also their access to care. And when an individual chooses to do a gender transition, they may or may not choose to have gender-affirming hormones as part of that transition. And our project focuses on masculinizing hormone therapy, which centers around testosterone. Traditionally, this was an IM injection only, but now it can also be sub-Q or there's transdermal options as well. One place where we are in agreement among major medical societies, including Endocrine as well as ASRM and WPATH, is that before starting any gender-affirming care, we should be giving fertility preservation counseling for all these individuals. The challenge here is that while we know there will be counseling, there's really not a lot of evidence for which to counsel patients in this space. There is a recent large prospective study of trans men in Europe that were initiating testosterone, and they found that at three months, the majority of patients were amenorrheic, and by six months, 91% had ceased cycling. The impact on ovarian reserve, though, is more mixed. There's two large studies in this space. One of them showed no change on testosterone, the other one showed a decrease in AMH. And unfortunately, both of the studies have a more complicated picture, because they're not receiving testosterone alone. In one study, there was also GnRH analog involved, and another study also had progesterone. And similarly, histopathology offers another more cloudy picture. In our mouse model, we saw no change in ovarian function or follicle distribution, which is what most histopathologic studies of trans men on testosterone are showing, but there are some, and particularly a lot of the more historic studies, that actually show a trend towards a PCO morphology, and then a couple that show a trend towards atrophy in the ovary. So it's very unclear. When we look at the evidence as a larger, take a step back and look at a larger picture, there is some concern for a possible detrimental impact of testosterone on fertility, but whether or not, how to quantify that impact and whether or not it's reversible really remains largely unknown. We think about testosterone and IVF, the literature becomes even more minimal, as you can see from the really small size of this table here. There are three case series that are looking at, or that included trans men previously on testosterone that ceased testosterone and then went through IVF, and in two of the studies, we saw no change in the number of eggs retrieved between those trans men on T and those not, but then one of the studies showed fewer oocytes on the previously exposed testosterone group. And interestingly, even in one of the groups that showed no difference in number of oocytes retrieved, there was a need for higher granatotropin, which would indicate a likely higher cost cycle. And then just a couple months ago was the first case series of individuals that underwent IVF without discontinuing their testosterone treatment. And so this was just two people out of UCSF, but it showed that IVF, while continuing, testosterone is feasible, and sure, oocytes were retrieved in both of those individuals. So as a lab, when we looked at the literature, we tried to ask how we can design a study to really give us some more insight into what is the impact of both active testosterone as well as testosterone cessation on IVF outcomes, determine the impact as well as if that impact is reversible. And so in order to design this study, I have our four treatment arms shown here. We used a previously validated model within our lab using 10-week black six female mice, and they were implanted with either a blank celastic tubing implant with ethanol only, or a celastic tubing implant with testosterone dissolved in ethanol. And then groups one and two represent our immediate treatment groups, so they received the implants for 12 weeks and then went through an IVF simulation with the implants in place. And this would be in contrast to groups three and four, which represent our cessation or washout groups. So they also received the 12-week of treatment with implants, and then the implants were removed and there was a 10-day washout period before they underwent IVF. Shown here on the left is our IVF simulation protocol. Because our mice were of more advanced age after being implanted 10 weeks and then having implants in for 12, we used a hypersimulation model. And so they were injected on the evening of day one, then triggered on the evening of day three with HEG, and then we retrieved oocytes 14 hours post-trigger. And then our mice were fertilized with sperm retrieved from a hybrid male strain to optimize fertility. And then on the right here is just depicting our culture and transfer model, in which we did insemination and culture to the two-cell embryo stage. And then two-cell embryos were injected into pseudopregnant mice in a one-to-one ratio, meaning that for each oocytes from a single stimulated mouse were then cultured to two-cell embryos, and those embryos were then transferred to a female pseudopregnant mouse. This first portion of our results just shows the feasibility of our model. So as we anticipated, we saw that all of the mice that had testosterone implants did see cycling within that first week of testosterone treatment. And then on the right here, I'm just showing our hormone levels, and again, we see that as expected, the testosterone treatment groups have testosterone levels that are significantly higher than controls for the duration of the study. The caveat there, of course, is you can see the purple T washout group, the testosterone level does come all the way back down to normal between weeks 12 and 14, which represents that washout period. Next up is our terminal hormones, and so the immediate group, meaning the group with their implants in place at the time of IVF, is represented at the top, and then our cessation, meaning our explanted group, is represented by the bottom. And what we see here is that there was a decrease in the terminal AMH in the testosterone-treated group as compared to control, but this difference was no longer present after the washout period. And then there was no significant difference in estradiol or progesterone levels at time of oocyte retrieval for either group, and then finally, the terminal testosterone levels are as we would anticipate, higher in the immediate group, and then return to normal in the washout group. Next up is our primary outcome, we're looking at our IVF results here, and so we did see a decrease in total number of oocytes as well as mature oocytes in the mice that were receiving active testosterone treatment, and that's correlated to then a decrease in the number of two-cell embryos that were obtained from those stimulated mice. And then the cessation group, we no longer see this difference across the, there's no longer a statistical difference between the control and the testosterone groups, which supports the potential reversibility of the detrimental impact that we found with the active testosterone. We wanted to look further and try and get a little bit of insight into oocyte quality, which of course can be very difficult to measure in the IVF space, but by calculating maturity rate and fertilization rate, we hope to gain a little bit of insight there, and we saw that there was no difference in maturity or fertilization between the treated and the control groups, as well as in the washout period. And then these pictures just represent, in the top rows you can see our mature oocytes on the evening of retrieval, and then below them for each group is the two-cell embryos that are shown. Finally we looked at our embryo transfer outcomes, and so for the testosterone-treated mice, there was a decreased probability of achieving a live birth for a single stimulated mouse as compared to the controls. And then we saw a trend towards a decrease in the total number of pups per litter, as well as a decrease in the live birth rate, but that did not achieve statistical significance. When we look at our results as a whole and the steps up and back to the bigger picture, we did find that there was a negative impact of active testosterone treatment on our IVF outcomes. We saw a decrease in total oocytes, mature oocytes, and two-cell embryos, and then this correlated to a decreased probability of a live birth for a single stimulated female. And then I think what's more promising here, though, is that we saw improvement in our IVF outcomes following the testosterone cessation period. And so these findings would support the potential reversibility of the impact of testosterone on reproductive potential. Many further studies are needed to determine testosterone's impact beyond the two-cell embryo stage, and what our lab is currently working on is embryo development all the way to blastocyst, as well as whether there's any impact on the reproductive potential of F1 offspring. And then ultimately, our goal would be to try to get additional research in order to try to quantify what is the benefit of testosterone cessation before IVF, and so that we can create a patient-centered model where individuals receiving gender-affirming care can help weigh the potential benefit and increase oocyte yield, with the potential negative being that the testosterone discontinuation can be very challenging and dysphoric for those individuals and so trying to weigh whether it's worthwhile at an individual level to consider testosterone discontinuation before undergoing an IVF procedure. And so there's many people I'd like to thank. I have my co-authors here. Of most note, I'd like to thank Dr. Zhu and Dr. Moravec, but Dr. Zhu really, he taught me IVF not only in humans as part of our fellowship, but also has helped me troubleshoot IVF in mice for the last 10 months. So I'm very grateful to him. And then the Center for Reproductive Medicine for my clinical training and the Shikhanov Lab for making this project possible. I'd like to open up to any questions. Yes. Thank you for a fantastic talk. It's really useful for patients. Can I just ask you to speculate whether, as it seems, do you reach a ceiling of effectiveness or do you think if you increased the doses of stimulation, you could overcome the reduced outcomes? Because that would be very important practically. Yeah, so is your question whether we could stem the individuals harder and try to get the same? Yeah, so we, as far as mouse models go, I use the largest like hyper-stimulation option that we have. It's a recombinant option out of Japan, which FSH and PM, and FSH and HGG, sorry. So it's a combination like similar to a menopurine in human, but it's as strong as mouse stimulation protocols go. We were kind of forced into using that. I did a lot of practice, but because our mice were already 22 to 24 weeks, we're also kind of working against age at the same time. So I think your question extrapolated human level will be more challenging because we're probably not using as the same type of, we don't have the same advanced age challenge that we'd be seeing clinically that we saw on our mouse model. So certainly I think that that was that one case series that showed that they got equal results between prior T and no-exponential T, but with higher gonadotropins. So I think at the human level, there is some insight into potentially being able to overcome this lower yield with higher gonadotropin, but I wasn't able to look at that in my mouse model, I think because we were also working against an advanced age. Thank you. And domestic UCLA, this is a wonderful work and keep up the excellent work. I have two questions. One is, do you have any thoughts on why the pregnancy rate was, the live birth rate was lower? Is this, do you have data on implantation per embryo transfer? So we don't. I think that when I look at my results as a whole, it kind of seems like we start with a lower yield, which then correlates to a lower number of two-cell embryos. And then in the BLAST data I'm looking at now, I'm seeing kind of, it's not, we're not finished analyzing yet, but we're trend towards a decreased number of two-cell embryos that persists all the way to the blastocyst phase. So I think that the insight I'm getting is more based on kind of you get less from the start and you end up with less down the line. But I think implantation would be awesome. One thing that we're looking into is MFMs. There's a MFM lab at Michigan that does do ultrasounds. And so we could potentially actually scan the mice in early gestation to determine how many early pregnancy sacks we see. It was just something that I'm thinking about doing with our next arm. Yeah, that's great. The reason I mention that is you're probably aware Janta Sarek in the 90s looked at non-genomic actions of androgen inhibiting the calcium flux induced by estradiol. And that was at a nanomolar concentration. So if that turns out to be an issue, you might want to think about even lower levels of measuring testosterone, because it could be a non-genomic effect. Hopefully it isn't. The second question I have is just more of what your opinion is, because obviously folliculogenesis is much longer in the humans versus the animal model. So what are your thoughts if there is an adverse effect, whether it be on recruitment or selection or whatever that is? What's the length of washout you think is realistic in human physiology? Yeah, so one of the residents at University of Michigan looked into this and presented recently SRI, just a survey study of REIs across the country that do trans care and IVF in trans individuals. And the most common, the majority of providers recommend discontinuation period. And the most common cited in the survey was three months. And so we designed our study. We tried to find a period of time in the mice that would be kind of similar to a three month washout. And that's how we came up with the 10 day. It's 10 day washout, but 14 days once you include the days of STEM. So we kind of used that as a starting point. I think the fact that we see recovery in a short washout begs the next question of, well, could your washout then be shorter? Because for these trans guys and non-binary individuals, the washout is all the qualitative studies that do exist to say that the washout or the discontinuation is the most dysphoric aspect, especially if it's to the point where menses are resumed, which for most people is in the one to three month range. And so when I found these results, I thought to myself, I wonder if a shorter washout potentially could result in the same recovery. Because at the patient level, that would, I think, really impact quality of life. Exactly, congratulations. Thank you. Hi, Joni Jorgensen, Wisconsin. Yes, beautiful study. I have a couple of questions. First, I might've just missed it, but the cycling, how long did it take for the mice to finish or like quit cycling on the treatment? Yeah, so they went into persistent diastereosis by post-op day three in all the treated groups. And there was a PhD in our lab that validated this model. And so they did cytology for like 22 full weeks. And so given that that was already validated, we didn't have any breakthrough. We just followed them for a week post-implant, but that was sort of where that ended up ceasing. Yeah, I'm sorry for that PhD. I know, I was like, that's dedication. That's very important. And then my next question, and an important one, and maybe you don't know all the answers to this, is how are you going to evaluate those blastocysts and early embryos? Yeah, so it's definitely very challenging. We are doing a blinded blastocyst grading system. And then ideally this summer, we're hoping to do a morphokinetic analysis, like by using a embryoscope to try to find timeline, like to try to evaluate the regularity of the divisions and the timelines of the different divisions as means to evaluate. Thank you. Thank you. Thank you so much. Thank you. So our next two talks are on a similar theme. One will discuss risks, and the next we'll discuss benefits. And we have Chana Jayasena here from the Imperial College of London, speaking on adverse cardiovascular events and cause mortality in men during testosterone treatment, individual patient and aggregate data meta-analyses. Welcome. Thank you so much for the kind introduction and the invitation. So I'm here very much as a spokesperson for an international collaborative efforts and testament to all the trialists who took part. So these are disclosures. Okay, so worldwide prescriptions of testosterone are undoubtedly rising, and it's mainly given to men aged 40 to 59, an age group that could plausibly get cardiovascular outcomes. If you look at the effects of testosterone on the cardiovascular system, it has complex effects. It could cause coronary vasodilatation, which has been speculated to be positive, but it also increases muscle mass and insulin sensitivity, which could be positive. But on the other hand, we know it increases hematocrit and could therefore plausibly cause thrombotic risks. And a minority of studies, it's been well documented, have reported increased cardiovascular risk. And then there's been a flurry of other studies failing to or not show risk, and maybe some showing lower risk. So the guidelines reflect the evidence, which is complete uncertainty, opacity over cardiovascular safety of testosterone. And what this does is it impacts on our patients because clinicians are really, really worried about this, and there's enormous variation in prescribing behavior. And so the UK government commissioned this evidence synthesis, which was to try to independently, to get a snapshot of where we are at the moment to try to plan the scope for future studies. And we did an individual patient data meta-analysis to try to actually gain the source data from the trial list of eligible studies. And the advantages of this over previous meta-analyses of aggregate data where you actually have access to unpublished cardiovascular events, you can independently and in a blinded manner adjudicate and decide on outcomes. And also because you will know that a heart attack happened in a 66-year-old man who was a smoker, who had a BMI of 33 and a starting testosterone of X, you can then try to attempt to do some secondary analyses and triangulate whether any of these outcomes have associated factors. In terms of the inclusion criteria of trials, these are randomized trials with a placebo comparator. We were taking men with a broad category of less than 350 nanograms per deciliter, so 12 nanomiles per liter. But we then were able to do post-hoc analysis to see if you dial down that inclusion criterion, do you actually change the outcomes? And we excluded crossover designs, less than three months, combined interventions, and studies restricted to conditions that we thought would affect the outcomes such as HIV, cirrhosis, et cetera. And we did exclude congenital as well, which we felt would be a different condition. Now, I'm gonna be focusing on the primary outcomes here, so mortality and cardiovascular events, and the physiological markers which feed into that, and prostate, et cetera. I'm not gonna be talking about sexual function, which was unsurprisingly positive, and quality of life, which in elements was positive as well. So it's important to say, where serial measurements were made, such as blood pressure, et cetera, we chose the one closest to 12 months. So if it was a three-month study, it was the last one. If it was a three-year study, we took 12 months. And all adverse effects were categorized by two investigators in a blinded manner, and adjudicated where necessary. So we found 17 eligible studies, and we got the source data from 17 out of these 35. And we did a secondary aggregate analysis where we compare the individual patient meta-analysis to see whether there was skew in the studies that weren't being captured, and in none of them was there any discrepancy. Mean treatment duration was 9.5 months, so three to 36 months. And in terms of participant characteristics, these very much kind of mirror the T-trials, so obese, a lot of diabetes, angina, smoking, and prior MIs. And I want to underline that no trials for prior cardiovascular endpoints primary, which is the limitation. So in terms of all-cause mortality, what we can see is that, so for all of these, we're showing an outcome on the left, number of studies, and then the number, so there were six deaths out of 1,621, 0.4% in the testosterone group, versus 12 out of 1,537, 0.8%. So these were small numbers and similar. And then down below shows the categorization of events. So we found very few deaths, and certainly too few to meaningfully analyze. Next, cardiovascular and cerebrovascular outcomes. So what we found, and again, I've highlighted in blue the kind of the headline, which is that we found 120 versus 110 cardiovascular events in each, which was non-statistically significant. As you can see, we found that more than one event happened in people. So there are 120 people affected, and there are 182 events in the testosterone group. And again, I've provided a breakdown of exactly what was observed, and there were no differences. Looking at stroke outcomes, again, very, very similar. So 15 versus seven, no significant difference, unable to comment other than say that they were balanced between them. We couldn't find any difference. I'm not showing you this, but we also looked at the effect of starting testosterone, diabetes, smoking status, and age, and this had no effects on this data. And so this is a secondary analysis, just underlining that if you combine the studies where we had individual patient data with the studies where we didn't, there was no overall effect. Physiological markers. So I've put in green the things that went up. So unsurprisingly, testosterone went up when you gave testosterone, and hematocrit and hemoglobin went up, so no surprises there. If we look at lipids, I've put in red the things that go down, and as has been reported widely, you get these incredibly minor subclinical reductions in HDL, cholesterol, and triglycerides, the clinical significance of which is completely unknown, and probably very minor. There was a slight reduction in glucose, but when we excluded people with a diabetes diagnosis, then this was removed. So we found no differences in glycemic premises or blood pressure. And finally, other outcomes. We found an increased risk of edema and unsurprisingly high hematocrit, but nothing else. So there are clear limitations with this approach, aren't there? So we only had the individual data from 62% of participants, but our aggregate analysis was concordance. None of the trials used any unifying cardiovascular MACE classification because they weren't looking for them, but we were relying on the fact that they would have noticed if one of their patients had died or had a heart attack. And there were a low number of recorded deaths, so we can't really make any meaningful conclusions. And the mean follow-up went down to three months. So again, that's probably too short to be thinking about things like atherosclerosis. But what does this find? Well, there are a few deaths that have happened so far, and there's no current evidence that we could find that testosterone would elevate cardiovascular or cerebrovascular risk, but this is in the short term and using studies that weren't designed to do this. We haven't identified any high-risk groups, and we couldn't find any effects on glycemia or blood pressure, and there are no minor effects on HDL hematocrits and edema are provoked, as we would expect. And the long-term safety is being currently looked by the FDA-traversed trial, and we keenly expect the results shortly. Thank you. So I just wanted to highlight that this was a collaborative effort and to thank everyone who collaborated and came together on this evidence synthesis. Thank you. Okay, thank you for this interesting study, and it's open for discussion now. Hi, Seminario Boston. Thank you for this lovely work and for helping us sort of, when we're sitting in the office clinically with patients, a lot of the male patients that I see have been put on testosterone by another provider. Sometimes their testosterone levels were normal, but yes, testosterone was still initiated for some sexual symptoms. And so I always worry in that group where I, you know, there wasn't clear biochemical evidence for the initiation even of the hormone that, oh my gosh, is the patient at risk or at harm? So I was just wondering if there was any opportunity in your datasets to look back at presenting or initial T levels and parse your results. Thanks, Stephanie, great question. So I didn't explicitly show you the sensitivity analysis. We did look very keenly at starting testosterone, and we couldn't find any difference. And what we don't want is for this to be a kind of a red flag, a green flag to say, right, it's open season for giving testosterone. Because any drug that is not needed is unsafe. But we could not find a trend for increased harm when you became looser in your definition. Okay, great, thank you. Great talk, Stephanie Page from the University of Washington. So, you know, we've seen these meta-analyses before, and it's very progressive of the UK government to actually fund something like this and get the real data. And we all really eagerly await the Traverse trial. So what I'm wondering is, is there a way you can continue this work? So as we in the field continue to do RCTs, rather than retrospectively coming back and trying to dig out the data, is there a mechanism that, you know, once we publish our data, we could add to this repository so we're not constantly with these meta-analyses trying to, because there's more data we can get from these trials. So I'm wondering if there's any impetus to do that with this consortium you've built. I think that's a really great point. And I think what the UK government wants is for this to have the maximum impact for patients. And it clearly needs to be a case. But if that was felt possible, then yes, it would be good to actually see whether we could give this project longevity. That's a great suggestion. Gary Whitted from Adelaide. Thanks, Jen, a really nice presentation. I enjoyed seeing the totality of the data. Insofar as the glucose is concerned, the fasting glucose is a fairly blunt tool. And glycated hemoglobin, I don't think you would expect to see a significant change. That said, I think it's quite interesting that you didn't see anything. In the T4DM study, it really took almost two years of treatment. And it was only in the last sort of few months that you see this effect, which I think is largely driven by the change in body composition. Yeah, no, I think that's a great point. And I think it goes without saying that if you're gonna be increasing muscle mass and give enough time, you would expect these glycemic changes to happen eventually, yeah. Okay, thank you, Cain. Thank you. And we look forward to wrapping our discussion with Fenu Dipika from the New Mexico VA Healthcare System. And she's gonna speak on In Men on Testosterone Therapy. Baseline testosterone levels of less than 264 nanograms per deciliter is associated with greater improvement in body composition, while level of greater than 264 nanograms per deciliter favors improvement in metabolic profile. Thank you everybody, and I would like to thank Endocrine Society for giving me this opportunity. So I'll make it easy and shorten the title a little bit of my presentation. So I'm Dipika from Baylor in Houston, and I'm one of the endocrinology fellows. And I'm gonna be talking about the study analysis, Baseline Testosterone Predicts Body Composition and Metabolic Response to Testosterone Therapy. To start with the background, we all aware male hypogonadism affects body composition, BMD, and cardiometabolic health. There has been emerging role of testosterone therapy for management of obesity in hypogonadal men. As we already know, there are some cross-sectional and a few longitudinal studies out there that have shown the association between low serum testosterone and risk of metabolic syndrome or type two diabetes. An elegant paper by Dinsa, which was a randomized trial for a shorter duration, showed that giving testosterone to hypogonadal men with type two diabetes improved their insulin sensitivity and their body composition. More importantly, I would like to point out that the shift in the diagnosis of hypogonadism from 300 to a threshold of less than 264 per the current end of society guidelines would suggest that the men who were diagnosed with hypogonadism in the past are not considered hypogonadal by the current definition. That could reflect some unmet need. Study done by Snyder a decade ago showed that indeed it is the men who have testosterone of less than 200 that showed the greatest improvement in bone mineral density at the lumbar spine. But this was done using an immunoassay back in 1990. With that, we have no study to date that's comparing or investigating whether baseline testosterone influences body composition and metabolic response to testosterone therapy. And so the objective of my study is to determine if there is difference in changes in body composition, metabolic profile, bone turnover markers, and BMD in men with T less than 264 and compared to those with T greater than equal 264. We hypothesize that testosterone therapy in men with baseline T of less than 264 will result in greater improvements in body composition, metabolic profile, and bone parameters. The methods, this is a secondary analysis of a parent trial that was completed in 2016 on pharmacogenetics of testosterone therapy. Enrolled 105 men who were given testosterone for a total duration of 18 months, and the subjects were divided into those with baseline testosterone of less than 264, 43 men, and greater than equal 264, 57 subjects. The hormonal assessments were done by LCMS, bone turnover markers and adipokines by ELISA, bone mineral density and body composition via DEXA. So here's just a depiction of body composition analysis by Hologic in our research center. All the adipocyte and lean indices are automatically generated by the DEXA as you already would know. We did the BMI and body composition analysis at six, 12, and 18 months duration. Now into the statistical methods, we first determined the testosterone T-cut score that divided our subjects into two groups of body composition at baseline. The figure here shows the regression line, the testosterone cut score of 200 on the immunoassay as shown on the x-axis was equivalent to 262 on the LCMS shown on the y-axis. Since this caused no difference in the separation of our body composition outcomes, we used the cut score of 264, which is defined by the Endocrine Society for diagnosis of hypokonadism, and divided our subjects based on this cutoff into those with T less than 264 and those greater. As I mentioned, the screening of the parent trial, which was done on immunoassay, but for this study, we used LCMS, which was done on the stored samples at the end of the study. Then the analysis of each variable between the two groups was done by analysis of covariance. The results of body composition, bone marrow density were additionally adjusted for age and BMI. Jumping into the results, we had 105 subjects in the parent trial, and 100 of those had testosterone available by a mass spec. 43 in the group with low T, and 57 in the group with higher T. I'll move on to the table, which shows the demographic data. I highlight here the BMI and the weight, which are significantly higher in the group with low baseline T compared to those with higher baseline T. Vitamin D, PTH, and important comorbidities, namely diabetes, hypertension, cardiovascular disease, sleep apnea, were not statistically different between the two groups. The first outcome of our study, body composition, so this table shows changes in total fat-free mass and regional, which includes appendicular and truncal fat-free mass, and the changes are depicted as both mean person change from baseline, as well as absolute change, shown by the symbol delta. What's evident that men with baseline T of less than 264 had a greater increase in total fat-free mass at 18 months. These group of men also had increase in appendicular fat-free mass, both at six and 18 months. The figure here on the right side, on the left side, sorry, the top figure shows the changes in total fat-free mass, and the red line depicts the group with low T and the blue with the higher T. And as we see that the increase in total fat-free mass occurs in both groups with testosterone therapy. But the group with low T shows a significantly higher increase at 18 months. And then the figure in the bottom shows changes in appendicular fat-free mass. Here again, we noticed that the appendicular fat-free mass increases in both groups, but the increase is significantly higher in men with low baseline testosterone. And this is evident at six, as well as 18 months. Moving on to the next parameters of body composition. So fat mass, there were no significant between group differences in changes in fat, total fat mass and regional at appendicular and truncal sites. Similarly, there were no changes between group differences in lean mass at the total, as well as appendicular and regional lean mass. However, within group analysis, not depicted in this table by repeated measures, showed that both groups had reduction of total and regional fat mass and increase in total and regional lean mass. The second outcome is metabolic profile. Strikingly, at 12 months, the group of men with higher baseline T of greater than equal 264, had reduction in their A1c, compared to those with T less than 264. And likewise, fasting blood glucose reduction was evidently higher in the group with higher baseline testosterone compared to those with lower. The significance of this A1c reduction for those with high baseline T becomes borderline at 18 months. Next is the lipid profile. So the total cholesterol seems to show a pattern of non-significant decline. Then we have triglycerides, which seem to have a non-significant increase at 12 and 18 months, probably in response to testosterone treatment and a parallel increase in estradiol. Another striking result is LDL reduction, which is significantly greater in the group with higher baseline testosterone and no change in those with low baseline testosterone. HDL reduction is evident in both groups, but the changes were not significant between the two groups, as well as within the group. Now the table, the continuation of this table is adipokine's data, which again reflects that leptin reduction is noted with the course of testosterone therapy, and the reduction is significantly higher in those who have higher baseline testosterone. For adiponectin, another adipokine of interest, we see there is not a very specific pattern, and there are no significant between-group differences. The figures here, I would say the top right, shows the changes in A1c. So we see the blue line, which is the group with higher testosterone, had a significantly greater reduction in A1c at 12 months. And we see a pattern, a non-significant increase in A1c in the group with low baseline T. The figure on the top, sorry, the top right there, shows the leptin levels. At the levels reduced significantly at 18 months for group with higher baseline testosterone. The figure on the bottom right shows changes in adiponectin, and what's striking is there is a steep decline in those with higher baseline testosterone for adiponectin from 12 to 18 months, but there are no significant between-group differences. And the third outcome of our study was bone. So bone mineral density, there were no significant between-group differences in changes in BMD at the three regions of interest. However, at baseline, men who had low testosterone had higher BMD at lumbar spine and total hip. Likewise, for bone turnover markers, osteocalcin, CTX, and sclerostin, there were no significant between-group differences, although the group of men with low baseline testosterone seemed to have a pattern of increase in osteocalcin and decline in CTX, but this was not significant within group analysis. This is just a supplement table showing changes in hormonal and safety profile. By virtue of our study design, and to keep testosterone at a target range for the guidelines, the group of men who had T less than 264 experienced a greater increase in testosterone in response to T therapy, and this was higher at 18 months compared to those with the high baseline T. And these group also had a higher increase in estradiol at 12 months compared to those with high baseline T. Hematocrit increase was significantly greater for the group of men with low baseline testosterone, and no changes in PSA levels were noted. So our study shows that T therapy results in improvement in body composition parameters in all men, irrespective of their baseline testosterone. However, it is men who have testosterone of less than 264 that were noted to have greater increase in their total and appendicular fat-free mass compared to those with T greater than equal 264. In contradiction to our hypothesis, men with T greater than equal 264 experienced an improvement in several metabolic parameters compared to those with less than 264. And no difference in bone outcomes, BMD, bone turnover markers, were observed between the two groups. Moving on, the significance of our data is that it provides further evidence that T treatment improves body composition to comparable degree in all men, irrespective of their baseline T. And it's important to understand that the current Endocrine Society guidelines recommend testosterone therapy for improvement in body composition only in men with HIV, and against for improvement of glycemic control in hypogonadal men with type 2 diabetes. So although we know that prior studies have suggested some improvement in insulin sensitivity with T therapy, but based on our study findings, we think that giving testosterone remains controversial for just the metabolic benefit. And hence, the strengths of our study is the first one evaluating if the baseline severity of hypogonadism is the one that influences the changes in these outcomes. It's prospective, longitudinal, compared to previously published data. Limitations being it's a secondary analysis, we had limited sample size, and a high dropout that may have contributed to lack of between group differences in some of these outcomes. In conclusion, T therapy results in improvement in body composition, irrespective of baseline T, but T less than 264 is associated with greater improvement in fat-free mass, whereas T greater than equal to 64 favors improvement in metabolic profile. With that, I'd like to acknowledge my mentor, Dr. Raina Villarreal, and Dr. Dennis Villarreal, and the study coordinators and statistician who contributed tremendously for this study. And I'd like to open up for questions. Thank you. Thank you, Dr. DiPicca. And some questions. Thank you for the intriguing data. Gary Whitted from Adelaide, Australia. The T4DM study, which was published in January 2021, looked at over 1,000 men studied over two years correcting for baseline testosterone against similar parameters. And the primary outcome paper focused mainly on the glucose outcome, because that was the primary outcome. There's no relationship between the baseline glucose, however you measure it, or however you divide it, the baseline testosterone, however you measure it, and however you divide it up, and the final outcome of improvements in glycemia, or prevention of diabetes, which was the outcome of the study, suggesting, and consistent with the epidemiology, that higher testosterone levels are protective in other words, it's a pharmacological effect. And I think you see very similar effects with obesity, very similar effects with the lean body mass and the bone and so on. So I wonder, for example, if you look at your hematocrit data, where you got the difference between the two groups, whether this is an effect of differences in baseline parameters that were not adjusted for, and taking a simple dichotomous split with very few subjects where you didn't have the opportunity to look at categories. So comparing the highest and the lowest. So is there a question, I mean, I appreciate the input you gave me about the study and the glycemic, but I quite didn't understand your question on the hematocrit, yeah, the hematocrit. The question is, because of your statement about hematocrit being only applicable in one of the groups, it raises the question about baseline differences between the two study groups. So you have a simple dichotomous split in the testosterone. Did you have any attempt to correct for baseline differences in a variety of parameters? And did you, I know you've only got small numbers, but could you construct categories to look at the highest and the lowest groups of testosterone, even as a sensitivity analysis? No, we did not do the sensitivity analysis, and you're right, we did adjust for the baseline in each, when we were doing the mean person changes and we were calculating absolute changes, and for the metabolic and hormonal profile, we only adjusted for baseline. We did not adjust for any other variables. For BMI and BMD, we adjusted for, additionally for HPMI, yeah. But we did do the logistic regression to see if the magnitude of increase would influence the, you know, the changes in our outcomes, and the p-value was not significant, yeah. Thanks for a really great talk. It was also about the HbA1c. So you, I think you said that the people with the higher testosterone had improvements in HbA1c, but the people with the lower testosterone didn't. I just wondered whether that could be artifactual, because the people with low testosterone were also getting big increments in hematocrit. That's true. And so that the HbA1c may be influenced by an increasing hematocrit in the people who are really hypergonadal. Just wondered, yeah. So if they had increased in hematocrit, so they would potentially have a falsely high hemoglobin A1c. That's an interesting thing to look at, but yeah. Which I think was also seen in the T4DM, that the HbA1c didn't actually improve so much, but it was moving goalposts because the hematocrit was going up as well. That's a good point, thank you for bringing that, yeah. Rodolfo Rey, Buenas Aires Children's Hospital in Argentina. So I have a methodological curiosity. If I understood well, you reclassified the patients according to LC-MS results, is it? Yes, yes. Okay, so did you look at the same, I mean, did you do the analysis with the non- Immunoassays? Yeah, with the immunoassays, because as you know, LC-MS has its, I mean, proven strengths, but you have the disadvantage of aging of samples, which unlike wine, is not always better. So. Yeah, I understand that. And yes, we did both ways. We did with the original samples, which were the immunoassay, and the results were quite different, which I didn't show. Actually, those results showed that the men with low baseline T of less than 200, because that's how we classified with the immunoassay, had improvement. They were all in line with our hypothesis, improvement in body composition, improvement in metabolic profile, and improvement in BMD only at lumbar spine, but no changes in bone turnover markers. When we regrouped our patients using LC-MS, we found controversial results with the metabolic profile specifically. Yeah. So I wonder which is better. Well, for the guidelines now, the standard is LC-MS, but I think my talk is highlighting more about the proportion of individuals that are going underdiagnosed and seeking testosterone from unauthorized providers. Maybe we need to increase our threshold to diagnosis from 264 to 300. That would at least take into account the fifth percentile of the normative distribution curve. And maybe we could address this problem without significant risks. I'm a pediatric endocrinologist. I don't dare discuss the cutoff. It's a methodological question. Exactly. Aged. Aging of the sample, yes. Yeah. I think that is something I can put in my limitations. Thank you. Yeah. Okay, thank you. Thank you. Thank you, everybody. Thank you.

testosterone

IVF outcomes

gender-affirming mouse model

female mice

eggs retrieved

fertilization rates

negative effect

reversible

testosterone treatment

fertility

Amanda Schwartz

University of Michigan

body composition

metabolic profile