I am Roberto Civitelli from Washington University in St. Louis. I'm interested in the skeletal biology and metabolic bone disorders. And the topic of this session is osteocalcin. The osteocalcin controversy. Osteocalcin, for those who don't know it, is a non-collagenic protein of the bone matrix that was discovered many, many years ago. And its function has escaped knowledge for many years. It is thought to be intimately involved in the organization of the organic matrix. But a few years ago, reports from a few groups pointed out a potential hormonal function of osteocalcin, which of course revolutionized the thinking about the bone as an endocrine organ. Unfortunately, some of the findings originally proposed have not been universally confirmed. And that has generated some diatribes and debates. The scope of this session is to review the data in a transparent way, without controversy initially at least, and present the status of the knowledge for everybody then to discuss and think about it. So without further ado, I am inviting the first of the three speakers. The speakers will talk for about 25 minutes and then we'll have a discussion at the end for about 15 minutes. The first speaker will be Professor Fanxin Long from the University of Pennsylvania. Fanxin is a very good friend. He has been affiliated with Washington University for many, many years, has developed most of his career there. He is an expert in bone biology, bone forming cell biology in particular, bone development, and also bone cell energetics. He will tell us about the basic aspects of osteocalcin and glucose metabolism. Thank you, Roberta. Thanks to everyone again. Thanks for staying for this late session. Nicola, thank you again for inviting me to give me the task of reviewing the basic biology of osteocalcin. I have to say at the outset that osteocalcin per se is not the focus of my lab, but I'm interested in it as a bone biologist because as our chair has already alluded to, it's the most abundant non-collagenous protein in bone. So what I've decided to do then is to spend the first half to review some basic biology of osteocalcin and also the mouse models that have been published in the context of glucose metabolism only. Other speakers will cover other endocrine functions that have been proposed for this protein. In the second half of my talk, I'm going to tell a little story in my own lab that it's not directly related to osteocalcin, but it's somewhat related. As you will see, I hope you'll find it interesting as well. Without further ado, I want to kick off by reviewing some basic biochemistry of osteocalcin. As I already alluded to, it's an old protein that has been fascinating bone biologists for a long time. As you can see that, sorry, I'm trying to learn how to use the pointer here. This is taken from an old, you know, 1989 Peter Hoshka review wrote this about the protein. Obviously, osteocalcin was studied as a protein first, and it is the most abundant non-collagen protein in the bone and it's made mainly by osteoblasts and has another name. It's a BGP or bone gamma carboxyglutamic acid or GLA protein. It's a small protein. Depending on the species, the mature protein is about 45 to 50 amino acids and it binds calcium. The unique thing about this protein is it has three glutamic acid residues. The protein, when it's first made, actually is longer. It's about twice as long before the cleavage of the signal peptide or propeptide. The three glutamic acid residues can be modified by adding a carboxy group in a vitamin K dependent process. That's the carboxylation. That's obviously because the vitamin K involvement is sensitive to warfarin. Warfarin will inhibit that process. Finally, the studies a long time ago have demonstrated that actually the circulating levels of osteocalcin is mainly regulated at the synthesized level, not so much dependent upon bone resorption. That was, from my understanding, basically based on by treating animals with warfarin, preventing the carboxylated osteocalcin from being produced, you can see a precipitous drop in the circulation for this protein, which has nothing to do with the influence on bone resorption. Finally, the circulating osteocalcin is cleared through the kidney. Just to zoom in a little bit on these three carboxy, the glutamic acid residues, as you can see here, typically here are three glutamic acid groups, the side chains, and they can be carboxylated by adding carboxy groups to the three side chains. That's dependent upon the carboxylase inhibited by sensitive to warfarin. The protein was studied for a long time. Then later, obviously, microbiology arrived. Then osteocalcin genes were cloned and discovered. You can see here the different species have a different organization in the genome for encoding this protein. The humans, you and I, have only a single gene. It's on chromosome one. Rats also have a single gene, the BGLAB gene, but it's on a different chromosome, chromosome two. Mice, on the other hand, have three related genes, BGLAB, which is original osteocalcin, and then the BGLAB2 and BGLAB3. They're all located within a very short distance of each other, 16 kilobases on chromosome three, as you can see the map down here. Both BGLAB and BGLAB2 are highly expressed in bone. They have identical amino acid sequence. BGLAB3, even though it's in the same neighborhood, it's actually not expressed much at all in the bone, but it's expressed in kidney and lungs. At the protein level, it differs from BGLAB and BGLAB2 by four amino acids. BGLAB2 and BGLAB3 are somewhat irrelevant for bone biology. I want to now come to cover the mouse studies that have been published by different groups. The first was Dr. Karsandy's group. That was the first genetic study of osteocalcin back, let's see, 26 years ago now. It was the first knocking out this gene and found people were obviously, up to that point, a lot of functions were proposed for this protein, but it turns out there was very minimal functional data gathered from that mouse. That is, there was an increase in bone mass as reported by the group. It was focused on mature mice, six to eight months of age. The methodology used was x-ray radiography and also histology. MicroCT was not quite used at that time. Then at the cellular level, it was discovered that the bone formation was actually caused by an increase of bone formation due to the loss of osteocalcin. The osteoclasts actually are also increased. It was a net gain of bone mass due to increased bone formation, overriding the increased bone resorption. Mineralization seems to be normal based on that study. That was based on dynamic histomorphometry. A couple years later than after that, a few years later than that, that mouse was analyzed for mineralization and it was reported that the hydroxyapatite crystal size was actually increased in those mice. It was based on the Fourier-transformed infrared imaging. Fast forward a number of years, about three years ago, another group used the same mouse strain but bred it to a pure background, C57 Black 6, and it was reported that the crystal size, the mineralization process, actually was altered in this mutant mouse. The mineral to matrix ratio was reduced and the crystallinity was reduced and the carbonate ratio was increased. It's indicating to that the crystals are actually less mature in the absence of osteocalcin. That was based on this mouse generated by Dr. Casanti's group. Initially, the bone mass increased and later it was found that there were some mineralization differences. But most importantly, this is back about 15 years ago. Dr. Casanti's group used this mouse, they generated it. By the way, this mouse was, I forgot to mention that it was generated in the mixed background of 129 C57 Black 6 by homologous recombination, deleting both BG Lab and BG Lab 2. As you can see, the first big discovery from this mouse was that the knockout mice were bigger than the normal lithomates at different ages. As you can see in panel A, the weight with random feeding was bigger and then the serum insulin level was lower. It turns out, based on the extensive studies of GTT and ITT at different ages, and some of these studies did not specify the sex what they were analyzed, but the CND was done with the high-fat feeding for six weeks in three-month-old animals. You can see these mutant animals have a defect in glucose tolerance and also in terms of sensitivity to insulin, ITT shown in panel E. This was really the first study that started a new era of osteocalcin as endocrine signal from bone to, in this case, regulating insulin production in the beta cells. Since that point on, for a long time, that model was actually the only model available in the mouse until a couple of years ago, two groups actually published additional mouse models. And here is shown by the Barton-Williams group. By the way, I'm going to name this knockout mice model by the name of the PI so that it can be a little easier to distinguish. Here, Dr. Williams' group used the CRISPR-Cas9 technology to knockout BGLAP and BGLAP2, again, similar to the previous knockout model, BGLAP3 is intact. And this was in the Black 6 and C3H mixed background, and then back across to Black 6 for only two generations. But the major finding here is that actually, even at the bone level, there's controversy already compared to the previous model. There's no increase in bone mass or strength, but they did find some consistent with the previous study. The carbonate level in the minerals is actually increased, indicating some disruption of the normal mineralization process. And obviously, I'll come to the next slide. The main finding is that these animals do not have a metabolic phenotype. And so I want to point out a couple of caveats for this study already. Again, this is a new technology, CRISPR-Cas9 gene editing technology, and it's known that this technology can generate off-target effects. As a matter of fact, a group sequenced did RNA-seq for the animals of the knockout versus the control. They found 12 genes non-related to the genes the gene intended have changed expression. And also, that change includes the BGLAP3, which, as I mentioned, is not so much related to bone biology, and that is very lowly expressed in bone to begin with, but it was upregulated by eightfold. But still, overall, the RPKM or expression level is very low compared to BGLAP or BGLAP2. So the authors concluded that even though there was this upregulation of BGLAP3, it should not be considered a compensatory mechanism for the lack of bone phenotype or metabolic phenotype, as they reported here. As you can see, they looked at only a couple of time points. At six months of age, in these animals, both male and female, in contrast to what Dr. Cassandre's group reported in their model, there was no size difference, no weight difference, or no fat difference. And because of the lack of phenotype, the group did not do too much of metabolic studies, but you can see these are the studies they've done. They did five, six months of females, and then they did add lip feeding or fasting, but they did not put them on to high-fat challenge, high-fat diet challenge. But the bottom line is that they did not see, they concluded there's no metabolic phenotype in terms of glucose handling. And then, at the same time, published in the same journal by a different group in Japan, Dr. Komori's group, and they found they used the old-fashioned homologous recombination method that was the same as what they used in the original Cassandre study. They see, to delete, similarly, the BGLAP and BGLAP2. BGLAP3, again, is intact. And this time, it was in a pure background, BGLAP6. And they studied these animals extensively. And the phenotype they reported, the only phenotype they reported, is actually in the mineralization, the crystal alignment, as you can see, the final point. There's no change in bone mass, either in bone formation or resorption. And the only thing they saw was the alignment of the crystals, hydroxyapatite crystals, or what disrupted in the mutant. That can affect the mechanical property of the bone. And this was done by Raman maxilloscopy. So it's a little different methodology from what has previously used for looking at the minerals. So again, the bottom line here, the focus here, is that these mice, too, do not have any metabolic phenotype. And this, in my view, is the most extensive study of the potential metabolic phenotype in these mice. They studied four different ages, from young, growing, to adult, as shown here. And also in both sexes, male and female, on the normal diet or high-fat diet. And with GTT, glucose tolerance test, and insulin tolerance test. As you can see, the conclusion from this study is that no glucose metabolic phenotype to speak of in these animals at any age, in either sex, or with or without high-fat diet challenging. So those were the two mice studies that they were contrasting with the previous osteocalcin knockout model. And actually, a few years ago, there was another study done in the rats, also used CRISPR Cas9. This is by Dr. Graham's group. The rationale for using the rats is, as I alluded to, rats, like humans, have only a single BGLAB gene. So they used CRISPR Cas9 in this rat model to delete the BGLAB gene. So what they, by doing this, what they have achieved is multiple lines of rats with different indels and frame shifts. And they all have the same shared outcome, that is, the BGLAB gene is disrupted, no osteocalcin to be measured in the circulation or anywhere in the body. And one caveat about this study is that they, in their phenotypical analysis, they actually mixed all these different animals with different indels of frame shifts because they have the same shared feature of knocking out the osteocalcin protein. And what they have found is that they actually found a little increase in trabecular bone mass, but cortical bone didn't change, and they have found some mechanical strength increased. But in terms of metabolism, they do not see any metabolic phenotype in these rats. No fat change or overall size change. So this is the only slight glucose metabolism phenotype they reported, is that in the males, there was actually a little better glucose handling. As you can see in the last ITT and GTT, the closed circle is the null. As we can see, they do better, actually, than the wild type, which is the open square. So this is the opposite of what you would expect from the mouse study. Knocking out the osteocalcin, it should worsen, it should compromise the glucose handling ability of the animal. But in here, actually, for unknown reasons, these animals showed a little better handling of glucose. So I just want to, at this point, I want to finish this summary of the literature by pointing out some agreements and disagreements from these four different models, three mouse models and one rat model. As you can see, what can be agreed on seems to be that osteocalcin, at the very minimum, is playing some role in the mineral deposition. Although even there, the details can be different. Some of the difference may be contributed to the methodology difference. As you can see, some use a Fourier transform for imaging, and the other uses Raman microscopy. And the disagreements are wider here. We cannot even agree upon whether it's affecting bone mass or strength. But the main disagreement here is that it seems to be from the three groups that study, that knocked out osteocalcin that differ from the original group, that in terms of whether these animals have a defect in glucose metabolism. We don't have all the explanations at this point why these disagreements come about. Some of them could be because clearly these are generated in different labs. I guess probably the number one thing that one could do is to resolve some potential technical differences, maybe swap mice and to change to see if similar results can be achieved. And by swapping mice, you can actually resolve some of the issues of the housing, environmental effects, microbiome, diet in a local environment. That could be some differences. As we know, metabolism is very susceptible to changes like that. And there's some genetic background differences, some are mixed background, some are black six, pure background. And then there's sampling differences. The sample size could have an influence on the outcome. You could have sampling errors, especially if the effect size is small. And if you don't have enough sampling size, you could miss or misinterpret the result. Some look at only one sex, not both. And finally, there are a couple models were generated with CRISPR-Cas9. We cannot forget that there could be some potential complications from that methodology, because they can introduce non-intended changes. And that can only be resolved by sequencing the whole genome. All right, so in the final few minutes, I'm going to transition to a little study we have done in the laboratory. The rationale for that is because besides the knockout studies, the loss of function studies, groups, and particularly Cassandra's group and our collaborators, have demonstrated that by injecting osteocalcin into the animals, you can actually improve the metabolic profile of the animals. As you can see here, this is from Dr. Cassandra's group in the 2008 PNAS paper. The orange line is wild-type mice put on high-fat diet, but injected with osteocalcin, recombinant osteocalcin, through an osmotic pump. You can see the orange line is always doing better between the gray line, which is mice on chow diet, and the blue line is mice on high-fat diet. So on all these parameters, the orange line, which is mice on high-fat diet, but with osteocalcin injection, is doing better in terms of glucose handling and body weight gain in response to high-fat diet. So we were inspired by this study. I want to look at a model we have. We study wind signaling in the laboratory. One of the models we generate is by overexpressing the wind ligand through a genetic approach, we can induce a huge increase in bone mass. So the rationale is that if we induce a large increase in bone mass, we should be expecting to increase the osteocalcin levels in the circulation to a great extent. And we want to see whether that can cause a metabolic benefit in these animals. So that's the rationale. So the genetics are shown here. We generate the ostrich's RTTA. So with doxycycline, you can induce the expression of Wnt7b in a permanent manner, depending upon the activity. So as you can see, they were published in this mouse. And then you can see it's a very robust model. By inducing the Wnt7b expression for one month, you have a huge increase in bone mass previously published. And we use this model. And now to induce Wnt7b phenotypes, the control are missing one of the three genes. Therefore, you cannot induce Wnt7b expression in the presence of dogs. So the method here is that we generate these genotypes and put it on dogs at one month of age for two weeks so that Wnt7b in the animal with the overexpression genotype can overexpress throughout the rest of the time. But the main thing is treating the animals with high-fat diet to induce metabolic problems and control with a regular diet. And we measure different measurements at different time points, at least after three months of the minimum of three months of high-fat. As you can see, there's a huge increase in bone mass. As predicted, these are just some of the microCG data. As you see, if you look at this cortical bone, the panel here, B8-LVTA, is almost 100% bone. There's a huge amount of bone increase. And this was due to bone formation, as you can see, both in male and females. Child diet versus the high-fat diet, there's a lot of increase in bone mass at the end of the experiment. So we measure the osteocalcin. As predicted, the osteocalcin should be much higher than the control. As you can see, both males and females, the uncarboxylated form of osteocalcin, which is hypothesized to be the hormone form, and also the carboxylated osteocalcin are both increased by about three-fourths in the over-expression mass, regardless of the diet. And the main thing is that it did not change the glucose levels in these animals at the end of the experiment. The females are actually sort of not sensitive to the challenge in terms of glucose levels. But the males is diabetic or prediabetic after high-fat diet. But the Wnt7BO expression increased three-fold in osteocalcin level in the circulation, didn't help. And the body weight, again, also didn't help on high-fat diet, regardless of the genotype. And the libid profile also did not help. And these are shown for the high-fat diet mice with or without the over-expression Wnt7B. And GTT, again, did not show, ITT didn't show the insulin sensitivity issues. And glucose handling, almost no change, maybe a slight improvement in the males, but nothing to write home with. So just want to quickly summarize what the little story we had is by over-expressing Wnt7B, we greatly increased bone mass and increased the circulating osteocalcin levels by three-fold. But we did not see a metabolic phenotype. And so we do not know how to explain, reconcile with the previous studies where you can inject osteocalcin into the animals and have a metabolic benefit. But obviously, the technology is different. One is injection, or ours is a chronic over-expression. And in the previous studies, actually, they did not measure in the circulating osteocalcin levels. So with that, I'm going to conclude and thank to you for your attention. And the unpublished work was done by a postdoc, Dr. Song. Thank you. Thank you, Fancine. As I mentioned before, we will take questions at the end of the session after all three speakers have given their talks. So I just need to exit? I think to probably change the slides. So and also, we'll have questions coming from the attendees logged on in the online meeting. So the second speaker is Professor Andrea Isidori from La Sapienza University in Rome, where he is a professor of endocrinology. He's an expert in reproductive endocrinology. And he's the current president of the Italian Society for Andrology and Sexual Medicine. So he's going to take us through another component of the osteocalcin story, and that has to do with male fertility, regulation of male fertility. So. Thank you very much for the kind introduction. So I will start by asking to the audience, what is osteocalcin? Is a hormone, growth cofactor, neurotrophin, or just a bone marker? Now, we endocrinologists tend to be a little bit egocentric, and we see hormones everywhere. We see endocrine glands in every tissue. But I would like you to remind that to be a hormone, must be something measurable, regulated, acting on these same organs and through a specific receptor. So let's see if osteocalcin works like this. We have just heard how osteocalcin is synthesized and released into the circulation. There are two cells contributing to that. Osteoblasts mainly, that is the producer. And there are some controversies whether the osteoblasts can actually affect the circulation of the total pool. So osteocalcin is, yes, indeed, very measurable in the circulation. And it's also regulated throughout life. We can clearly see an increase during adolescence in osteocalcin levels. Then we see some sort of decline or stable level in the middle of life. And then a second growth with aging. And this is true both for the total osteocalcin and the undercarboxylated osteocalcin. So that the ratio between the two remains relatively stable throughout life, except for a slight increase in the percentage of uncarboxylated osteocalcin to the aging. And I would like you to mirror these changes to what occurs in terms of bone metabolism. So probably the increase during puberty likely reflects the achievement of peak bone mass. The relatively stable levels during the middle time is related to probably a lower bone turnover. It can be physiological, but also pathological, like in type 2 diabetes mellitus, as I will show you later on. And the increase with age is likely a reflection of the increased bone turnover. And the increased ratio between uncarboxylated and total osteocalcin likely reflects an increase in osteoclastic activity, although this is controversial. So what about the gonadal phenotype of the osteocalcin depleted mouse? Now, the first representation, the first data on this are for Carcenty Group, published in 2011. And they found that this animal have smaller testicles. They have a lower sperm output. They have low circulating testosterone levels and unexplained high estrogen levels, because you should remember that in males, estrogen are mainly derived from testosterone. So this is a little discrepant finding. They have demonstrated a reduction in the enzymes responsible for steroid hormone synthesis. And these enzymes are controlled by LH. And in fact, they found a compensatory rise in LH. And this mirrors very nicely what happens in the seminal tubules when you see preserved the early stage of spermatogenesis, but reduce the later stage that are more dependent on testosterone, the postmyotic. So this is a clear definition of a lydic cell failure. So a primary hypergonadotropic hypogonadism. And interestingly, if you increase osteocalcin levels, you can get a increased size in the testicles and higher testosterone levels. So exactly the opposite. Similarly, if you treat these animals, the osteocalcin deficient animal with osteocalcin, you can restore the expression of the sterogenic enzyme. Thus, dependently, you can increase testosterone. So it all seems to say that osteocalcin is indeed a lydic cell stimulating hormone. But in order to be so, you need to have a receptor. You remember the key and lock theory for endocrinology. And the first receptor, attention went to this originally orphan receptor, the G-protein couple receptor, C6A, was knockout mice was resembling the same gonadal feature observed in the osteocalcin deficient mice. So a lower endogenital distance, lower testosterone, lower testicular size. And in fact, this receptor is expressed by the lydic cells, both in mouse and humans. It is regulated throughout life and is expressed in the testes, but not in the ovaries. Because one feature of the carcinoid mice is that the gonadal phenotype was only present in the male mouse, but not in the female mice. The problem is that this receptor is expressed in many, many tissue. So it means that osteocalcin must be acting on all these tissue. And even more, the fact that this receptor is very promiscuous. So you can get stimulation or activation by osteocalcin, yes, but also arginine, calcium, even testosterone. This receptor is stored to mediate the non-genomic effect of testosterone. What we know, however, is that in the lydic cells, the main molecular pathway involved is that related to CKMP, PKA, and CREB stimulation. This situation is even farther complicated by the fact that recent data seems to show that there is also a competition, not only from testosterone, but even from sex hormone binding globulin in a competitive binding with osteocalcin in the lydic cells. So it's quite complicated if you think that this is acting on the gonad. And also there is another problem because if osteocalcin stimulates testosterone, then testosterone has an anabolic effect on the osteoblast, then it will lead to a perpetual growth, continuous stimulation. So clearly there must be something else controlling this feedback. Otherwise it's grown up to out of control. So one hypothesis is that sex hormone can affect the osteoblast, and though they release of the undercoagulated osteocalcin from the matrix. And there are some data in this direction because if you deplete your animal of the osteoblast, you lower circulating osteocalcin levels and you can lower testosterone and testicular weight and sperm count. And conversely, if you increase the number of osteoblasts, you increase osteocalcin levels and also you increase testosterone, sperm count and testicular weight. So it seems that osteoblast plays some role in this. But there is also another way to control this because if you increase testosterone with osteocalcin, then this higher level of testosterone will suppress the pituitary gland. So you will lower LH. And this is a very powerful feedback mechanism. So probably the situation will go to another standby unless there is a direct simulation of osteocalcin on the pituitary gland. However, the receptor that we have thought to be the transductor of osteocalcin is not expressed in the pituitary gland, neither in the hypothalamus. And this is very strange because you know, we used to say that the testicle and the brain have a very high homology in the expression of genes. And this is why sometimes we say that men think with their testicles rather than with their brain. But this is not the case for osteocalcin. And indeed, if you give enough LH, so enough testicular stimulation, you can rescue the fertility problems of the osteocalcin deficient mice. If you give enough LH or ACEG, you do not change too much osteocalcin or the osteocalcin processing genes. And most importantly, if you delete LH, you cannot use osteocalcin to whatever level to rescue the phenotype, okay? So apparently in this complex system of feedbacks, we show that there is not a direct control on the pituitary gland, and probably the LH testosterone feedback is dominant in respect to the bone testicular axis. But even worse, there was something about to happen in May, 2020, we were all entering lockdown, but this was not COVID, was something else related to bone as my honorable previous speaker has shown. Two different independent works published from Japanese and American groups, they created different models of knocking down osteocalcin, and both these animal had no gonadal phenotype. So testicides was normal, tendency to a reduction in the model by digital, but not significantly. And the authors were not able to explain how this could be possible, except by invocating different genetic and environmental background. So probably it's time to move from the mouse to the humans then. And even here, the situation is not that very clear. There is no consensus on the relation between osteocalcin, LH, and testosterone in humans, just presenting some study, one in the German, more than 1,000 subjects, a very small positive correlation between osteocalcin and testosterone. In another study from the Amsterdam, a little more older subject, and they found a correlation only in the highest, in the subject with the highest quartile of serum osteocalcin, showing a positive correlation of LH, but not with testosterone. So we looked at what happens in patient with Klinefelter. We just presented yesterday this data. Why? Because Klinefelter syndrome is a model of exhaustion of the LH testicular feedback, because LH is very high, cannot go any higher than that. And in this subject, we found a nice increase in osteocalcin during puberty, as expected, a lowering in adults, a positive correlation between LH and testosterone. But what is most important is that we found that in hypogonadal subject, Klinefelter, osteocalcin tended to be lower, and it was even lower in those who were receiving testosterone replacement therapy. So these open apps, what is the effect of testosterone on replacement on osteocalcin? So we reviewed all the literature just one year ago, just very recently, and we showed that most of the effect of testosterone on bone markers are toward an anti-resorptive effect. However, if you look at the data, there is a very significant heterogeneity. It means that different groups respond differently to testosterone administration in terms of bone markers. And Nicola Napoli very elegantly showed this in a group of hypogonadal subject divided in those who had type 2 diabetes mellitus and those who didn't, because they showed an opposite response in terms of osteocalcin level to testosterone replacement therapy. And they nicely concluded that this is probably because testosterone has both effect at stimulating bone formation but also inhibiting bone resorption. But the former is definitely much stronger in patients with diabetes because they have a particularly state of low bone turnover. And so the complex is now becoming, the scenario is becoming now more complex. So we have a balance, different balance of sex hormone on bone metabolism, and probably bone turnover status is key determinant. And it becomes even more important in those situation in which the dominant way of controlling the axis, so LH testosterone, is for whatever reason impaired, like in diabetes but also in clinical therapy, for example. And also, although it is not expressed in the pituitary gland, maybe osteocalcin can have some brain effect. Why this? Because carcity, from the very beginning, described that their osteocalcin knockout animal had a behavioral phenotype. They exhibit more passivity. They were scared, impaired spatial memory, anxiety, depression. They confirmed that these behavioral phenotypes were related to changes in the neurotransmitters. And this impairment was observed both in male and female mice. And it was not, absolutely not present in the mice knockout for the GPR6 receptor. So there must be something else working here and something that is independent from the sexual and the metabolic feature. Because we have just said that the sexual feature is only exhibited by the male mice. More, what they've done, they injected intrathecally osteocalcin and they were able to correct most of these behavioral features. So they increased the passivity, they reduced anxiety and depression, and they also improved the spatial memory, so cognition. They also went further by doing a postnatal deactivation of osteocalcin, and these animals still had some behavioral problem in terms of memory, although less prominent. But what was very interesting is that some of the structural brain alteration found in the non-invisible knockout of the osteocalcin were no longer present if you do a postnatal deactivation. So this tells us that osteocalcin is important during the embryogenesis of the brain, and in fact by providing osteocalcin to the mother, they were able to rescue the brain impairment in the progeny. And this was clearly also associated with normalization of the neurotransmitters, and also a good response. So maternal pool of osteocalcin seems necessary for the acquisition of memory in the adult offspring by reducing neuronal apoptosis in the hippocampus, and so preventing those structural brain abnormalities. But how if the receptor is not there? Well, there are two receptors that have been associated with this effect. One is another orphan receptor, the GPR158, that's been found in many brain areas that are able to mediate osteocalcin signaling. And in fact, the knockout mice for this receptor more or less overlap with the behavioral phenotype of the osteocalcin knockout mice. And there is also very recently another receptor, but expressed only by oligodendrocytes, the GPR37. So you can see it's the same family of G-protein coupled receptors, in which they control differentiation and melanization. However, it's important to notice that neither in the Mauritian nor in the digital paper, the behavioral phenotype was described. So we do not know if these animals have a phenotype or not. So let's go back to the humans again. I have three studies, and I will just finish. The first study is about 400 subjects, men and women. What they found is that only in women, they were able to find a correlation between osteocalcin and global cognition and executive function that accounts for more or less 5% of the variability of these parameters. But nothing in men, and no correlation with bone density and cognition. But the study was cross-sectional, retrospective. They did measure under carboxylation, and there was no assessment of the gonadal function. Another study, very, very small, I'm just presenting because it's fascinating how we can now study cognition. They do a DTI MRI. So they looked at the signal in MRI intensity, and they found a correlation between altered MRI intensity and parameters of cognition. But of course, the sample size is very, very small, and we cannot yet be so confident in correlating DTI metrics with cognition. And the last study is about 200 patients with type 2 diabetes. This is interesting because they divided the population in those with impaired or normal cognition, and this is a methodological good point. And they found that osteocalcin levels were lower in male with type 2 diabetes and impaired cognition. So once again, there is a gender difference, and it was positively correlated with a variety of index of cognitive function, even corrected after many confounding factors. So I'm back to the first slide. What is osteocalcin? Well, first of all, we should put away our egocentrism, and probably osteocalcin is not a hormone, at least in the traditional way we intended. And yes, it's a very useful bone marker, especially in those situations in which bone turnover is a little bit more complicated by other concomitant factors. What about the health? Well, probably osteocalcin may have a role, especially during skeletal development and or altered bone turnover status, possibly collaborating with testosterone for a coordinated and efficient growth that is more relevant or becomes relevant in which the other dominant way to control androgens, for example, the hypothalamic pituitary gonadal feedback is impaired. And what about the brain? Maybe it might have some effect in neural development, but again, probably as a modulator rather than an effector, possibly supporting survival and myelinization. So I've finished, and thank you for your attention. Thank you, Andrea. Very nice presentation, discussing all the aspects of this osteocalcin function or maybe lack of. So again, hold on to your questions until the end of the next presentation. And then we'll have actually probably a little bit more than 15 minutes at the end to discuss. So the next and last speaker is Professor Anne Schwartz from the University of San Francisco, California, San Francisco, where she is an epidemiologist and has worked for many years on large studies, clinical studies in bone health, primarily focusing on the effect of diabetes and deranged energy metabolism on bone. She has published many seminal studies on these topics and has been an associate editor of JBMR for the last three years, which I appreciate very much. Anne, you have 25 minutes. Great. Thank you, Dr. Civitelli. And thank you to the program organizers, Dr. Napoli in particular, for the chance to be here and talk with you about osteocalcin. I'm going to be taking up the clinical perspective, so looking at human studies. And I'm just going to do a brief little background and then look at cross-sectional studies, spend more time on the longitudinal studies, both observational and secondary analyses of randomized trials. So, of course, now you've had these two presentations looking at the background, but just to remind you, human studies here were following the mouse models, I should say the Carcente models, showing that mice without osteocalcin are obese, insulin resistant, prone to diabetes, and that the active form was under carboxylated osteocalcin. Other piece of background, which you just heard about from the last speaker, is that type 2 diabetes itself has an effect on bone formation. So, there have been small studies with biopsy evidence showing you one here where bone formation has decreased in this, the participant with type 2 diabetes, with reduced mineralizing surface. Also, nice little graph here of these are five type 2 patients looking at their glucose and mineral apposition rate, and they're strongly correlated. And we also, there's meta-analyses looking cross-sectionally. We know that bone turnover markers in general are lower in diabetes osteocalcin, but also P1NP-CTX. And then let me just make one note here, because a lot of the studies I'm going to show you have looked at total osteocalcin kind of as a marker for under carboxylated osteocalcin. It's not completely, there's some legitimacy to that. You can see on the left the relationship between the absolute levels of under carboxylated osteocalcin and total osteocalcin in the serum. This is in about 400 older adults. They're pretty strongly correlated, less so when you look at percent UCOC. So that's on the right, not as strongly correlated with total levels. So let's jump into the cross-sectional studies here. And a lot of evidence here that osteocalcin, total osteocalcin is lower in those with type 2 diabetes. Under carboxylated osteocalcin is also lower, not as many studies, but showing you one here. We move on to fasting glucose. Again, multiple cross-sectional studies showing that there's a negative correlation. So this is consistent with the mouse model. So higher osteocalcin, lower fasting glucose, correlation of negative 0.16. And not as many studies, but plenty of studies here showing the same thing with under carboxylated osteocalcin and fasting glucose. People have also looked cross-sectionally at BMI and at adiposity. I'm showing you here the percent body fat. Again, negatively correlated with total osteocalcin on the top and with under carboxylated osteocalcin, minus 0.18 correlation for both of those. Similar correlation for BMI. I'm not showing you that. But let's stop for a minute here on these cross-sectional results. So they are consistent with the Carcenti model, but we also know that there's this other thing going on, which is that diabetes and higher BMI have an effect on bone formation and bone turnover. And so the cross-sectional studies are basically consistent with both hypotheses and are a very limited usefulness. Nice place to start, but we need more. So let's look at longitudinal studies. So let's look at longitudinal studies. This was a meta-analysis of total osteocalcin and incident diabetes. Three studies included here. And let me just be clear, these studies are measuring osteocalcin at baseline in people without diabetes and then measuring, you know, who develops incident diabetes over the course of follow-up. And here there's actually a suggestion of a protective effect, although it's not statistically significant, that perhaps higher osteocalcin is protective here. Since that meta-analysis I showed you, there have been eight more studies. The largest of those studies actually found no effect. I'm not going to show you all of those because I wanted instead to look at the studies where they've measured under carboxylated osteocalcin in the serum and then looked at incident diabetes. As you've heard, under carboxylated osteocalcin is proposed as the active form. So no published meta-analysis. Somebody here should maybe do one. But we have the largest, I've organized these by size, largest study here on top showing no effect. But you can see if you look at the point estimates, they're tending towards perhaps a protective effect here of higher under carboxylated osteocalcin, although the only significant study is this second one down with a point estimate of .6. And you could see over on the right, we're not exactly comparing apples and oranges with these numbers. So this is just to give you a feel. It's not a meta-analysis, but maybe something here. So let me go on to fasting glucose or change in fasting glucose. So these studies are designed, you know, measure osteocalcin, measure fasting glucose, follow people forward, measure fasting glucose again or multiple times. This meta-analysis basically showing no effect of total osteocalcin on change in fasting glucose. And at least in the study that we did recently in older adults, we found a similar, similar finding for under carboxylated osteocalcin. So not associated with the changes in fasting glucose. Longitudinal results with weight change, a little more interesting. So I'm using here results from a trial and taking advantage of the fact that treatment with PTH increases under carboxylated osteocalcin and lindronate decreases it. So looking at this study, the PTH group, I get this pointer going. Okay. PTH group is A. So figure 1A. And you could see osteocalcin on top under carboxylated osteocalcin on the bottom. There's a big increase at three months. So it's baseline three months, six months. I've got a pointer to the three-month increase over 200%. And the lindronated treated group, it's hard to appreciate, but there's actually a decrease of 29%, which is pretty substantial. So there wasn't any difference between the PTH and lindronate groups in terms of changes in weight or fat mass, but numbers were small. And these authors took advantage of these big dramatic changes in under carboxylated osteocalcin and basically used this as an observational study. They said, let's look at the change over three months in under carboxylated osteocalcin. That's in the quintiles at the bottom. So over to the far right are the people in figure A who had a big increase in under carboxylated osteocalcin. At the left, people had a decrease. And we're looking at the top change in weight. So you can see those people with a big increase in osteocalcin have weight loss. And in B, you're seeing change in fat mass. Again, big increase in the osteocalcin, a drop in the fat mass. So that was interesting and consistent with the Carcente mouse models. But to summarize these observational studies, I would say inconclusive at this point in terms of effect on incident diabetes. Not seeing any evidence that there's an influence on fasting glucose. And I've showed you this one study where there's an association with weight change. But of course, this is just one study. And I also want to point out some difficulties in interpretation. I haven't done this with each of the studies. But we could say in general, of course, big improvement over cross-sectional studies. We know we've measured osteocalcin before we measure our outcomes. But we do have other potential problems. So measurement of under carboxylated osteocalcin is not that easy. And non-differential measurement errors would tend to attenuate any real associations. We also have a problem of confounding. Have we controlled for it? Have we found all the correct confounders? And finally, there's a problem of reverse causality. So hyperglycemia may be having some effect on bone turnover. And we're not getting everybody with exactly the same glycemic levels at the beginning of these studies. So one way to get around this and use a somewhat different approach is using randomized trials. So we can do secondary analyses of randomized trials. Of course, they're not, you know, they're more caveats because they aren't designed to look at outcomes of diabetes or changes in glucose. But still, they have some advantages. They don't require measurement of under carboxylated osteocalcin. And the randomized design does protect us from confounding. Disadvantages, we're testing, remember, we're testing the intervention. We're not really testing what under carboxylated osteocalcin is doing. And just a reminder here, so I'm going to start with the osteoporosis therapy. So anti-resorptive therapies, we're reducing bone turnover. That includes osteocalcin and under carboxylated osteocalcin. From the Carcenti models, we'd predict higher incidence of diabetes, greater increases in fasting glucose and weight gain in the treated group. Anabolic therapies, of course, the opposite expectation. So here's a study that our group did looking at trial of three different anti-resorptives. First here is the FIT trial of alendronate, horizon with zoledronic acid, and freedom looking at denosumab. You can see this is all women, post-menopausal, about 50% overweight or obese. And of course, these women are, to get into these trials, they're osteoporotic either by BMD or by a presence of a vertebral fracture. So what did we find? Well, here's our results for the incident diabetes. So nothing statistically significant for either any of these three trials, but a suggestion of, if anything, a protective effect as opposed to what we predicted, an increased risk of diabetes. No changes consistent with that. We didn't see changes in fasting glucose associated with treatment in any of these. But interestingly, there is maybe something going on with changes in weight. So if you look here at the three different trials, you'll see that there's a positive association here with treatment, statistically significant for FIT and freedom. And I'm going to move on to a later study that was done on zoledronic acid that went out a little further. I wanted to show you the graphs here. So on the top here is the change in weight, and the dashed line is placebo. So this group, this is post-menopausal women, is losing weight on average. And zoledronic acid isn't really causing weight gain, but it's preventing weight loss. And these investigators looked at the change in lean mass and fat mass, and you could see that basically the treatment is preserving fat mass. So this is a side effect that, I don't know, you don't usually see mentioned when people are prescribing osteoporosis therapies. But, you know, you could tell your patient, take this drug and you'll not only preserve bone, you'll preserve your fat mass. Okay, you can see why nobody talks about this, but it is there. Okay, so moving on to the anabolic therapies, a couple of small trials. So this trial, just to show you, there is an increase in osteocalcin and undercarboxylated osteocalcin with PTH treatment. And this trial also found that fasting glucose decreased in the treated group compared with a control, but they didn't find any differences in measures of adiposity. And in contrast, another small trial of PTH, this was in hypoparathyroid patients, they did find a difference in body weight consistent with the Carcenti model, PTH with some weight loss compared to placebo, but they didn't find any difference in glucose during this trial. Okay, so how to summarize the osteoporosis therapies? Well, it seems pretty clear the anti-resorptives are not affecting incident diabetes or fasting glucose. It's pretty strong evidence that the reductions of undercarboxylated osteocalcin are not affecting these outcomes. The anti-resorptive therapies do maintain weight, PTH might cause weight loss. They're both consistent with the hypothesis that higher undercarboxylated osteocalcin levels might prevent weight loss. However, back to, you know, this point that the trials are assessing the intervention, not undercarboxylated osteocalcin. Specifically, the intervention is causing other things to happen as well, and some of those other mechanisms might be what's having the effect on weight. Here's some of the speculations. With anti-resorptive therapy, could it be increasing bone mass? I mean, yes, it does, but that doesn't account for the increase in weight. Certainly wouldn't impact, you know, the maintenance of fat mass. More plausible alternative idea is that it might be affecting sclerostin. Then with PTH therapy and weight loss, there's some concern that there may be side effects that impair appetite in the PTH treated group. So, I want to show you one other set of clinical trial results, and this is looking at vitamin K supplementation. So, vitamin K reduces undercarboxylated osteocalcin, so in some ways similar to the anti-resorptives, but it doesn't really have an impact on, or much of an impact on, total osteocalcin, and it is not suppressing bone turnover in general. You can see on the top here, women and men, this is the change in serum undercarboxylated osteocalcin. The treated group here is the dashed line. These are the people getting vitamin K, and then you can see much less impact on total osteocalcin in the two bottom panels. This trial found no difference in fasting glucose, so these are the results at six months, that increase of three in the men, in the placebo, and, you know, 3.5 in the treated group. Women, 2.3, 2.3, 36 months, same thing. We're not seeing an impact on changes in fasting glucose. There's also no differences in fat mass or lean mass. Appendicular lean mass is on the top in women and men, and you can see the treated group with the dashed line is just parallel to the placebo group with a solid line, and on the bottom is total body fat, and again, no effect of the vitamin K supplementation. So looking at these clinical trials as a group, the anti-resorptive therapies, the anabolic therapies and vitamin K were not seeing an effect on risk of incident diabetes or changes in fasting glucose. Interesting differences here in terms of change in weight, particularly fat mass, which is happening with osteoporosis therapies, at least with the anti-resorptives, but it's not happening with vitamin K. So this suggests that the mechanism for the effect of these osteoporosis therapies on weight is probably not via under carboxylated osteocalcin. So how do we reconcile the mouse and human results? Because we saw these strong results in the Carcenti models, but we're not seeing much solid evidence in the human studies. And this is obviously something we may wanna talk about in the panel. I'll just point out that, of course, mice and humans are different. Mice are useful models, but there's not 100% analogy here. So the other thing here is that there's different degrees of exposure. So much of the mouse models you saw were genetic alterations. Human studies were looking at the effects of physiological levels of osteocalcin, a physiologic range, and the observational studies were changing the levels more dramatically with the interventions, but we're only using those interventions for a few years before we look at our outcomes. So in the future, I definitely think there's a place for additional longitudinal studies and maybe some secondary analyses, maybe another vitamin K trial, but things that would consider these metabolic outcomes. And I certainly think it would be interesting to understand these weight and fat mass changes with osteoporosis therapies. I don't think it has to do with osteocalcin probably, but would be important to understand the interplay of bone and fat. So I wanna acknowledge my collaborators and funding from the NIDDK, and thank you for your attention. Thank you. for finishing a little bit earlier. We can, okay, I'd like to invite the other two speakers to come back to the podium here. We'll have additional chairs for you, don't worry. And we can, yeah, yes, chairs were available, the speaker has to actually take, carry the chair himself. Okay. Okay. Bring your own chair. Are there questions from online audience? Okay. All right. Thank you for your patience. So, we can open the questions from the floor here and then we'll also see if there are questions from online. Please, introduce yourself and go ahead. Thank you. Peter Kamenitsky from Paris. Beautiful talks. I would like to ask you a small question regarding the intriguing data by Frank Uri and Gérald Carcanti on the cognition. So, if we take hyperparathyroidism as a situation that can in some way mimic in human what we observe with increased osteocalcin levels, we clearly have improvement in terms of cognitive functions in these patients treated, replaced by PTH, which obviously is a situation with many confounding factors, especially increased calcemia and maybe PTH action on the brain itself. But what I didn't get from your talk, maybe I missed it, does osteocalcin bypass the brain barrier? And are there any binding studies in mice that would show us where it actually would act in the brain? Andrea, I think this is for you. Yeah, sure. Okay, thanks for the question. Regarding the fact that osteocalcin can pass the barrier, it has been proven for the undercarboxylated form. So, this is the only one that can pass. Regarding the effect of PTH, that's very interesting. There are this data supporting a positive effect on cognition. And this data relates either due to complexity of factors. So, it's difficult to relate this to osteocalcin. Most of the studies suggest it's an issue of vascular inflammation or the effect on precipitation of calcium. So, this is probably more if the data on PTH are related probably to other factors than the effect on osteocalcin. To be honest, most of the study associating cognitive function with osteocalcin were not controlled for PTH. So, maybe it's the time to do that. Thank you. Yes, thank you. Nikki Partridge, New York University. That's a great review from the three of you. Really fabulous. So, actually, my question is really for Fangxing Long. The question, first of all, I mean, and I think it came out, the problems with the background strains of the mice. I just want to continue that because they said C57, Black 6, but were they Jackson or NIH, which is an added problem because it's pretty clear that the NIH sub-strain, I think, are bigger and have more fat. And I don't know that that was ever mentioned in the original papers. That's my first question. The second question is that there's been some speculation that, in fact, when the original deletion was done on the Gersenti mice, was that there's other genes on the other strand that were actually deleted that might explain some of the phenotype that he's seeing in the mouse, particularly the behavioral effects. I just wondered about your response to those two questions. Right. So, the first one has to do with the differences between the N versus J form. So, that is, I think, in the mixed background, I think they're mostly Js, if I remember correctly. I don't think it's J. Right. So, it is a clear difference. The J form, I think the Jackson mice does have a mutation. It's an NAD metabolism, which could potentially play into this. Yeah, that's something to look further in detail to see whether that contributes to the discrepancy. The second question— Was whether there's other genes on the other— Right, right. Other strands. Yeah, right. I know the—I remember that there was that discussion. I just forgot what—yeah, there is another gene. I think it's far prior to the first VGLAB gene. I just forgot what—I think the function of that gene is also not clear, right? Not very well studied. Maybe Andrei. Yeah, if I may add, it's true that the genetic background is probably the key to understanding the discrepancy between the two systems. And there might be other genes involved, and the purity of the carcinogen mice needs to be confirmed. But in support of their funding is the fact that by injecting osteocalcin, you can reverse it. So, even if there are other genes involved, or even if the genetic background is not exactly the one we have, we would like to see the results of injecting osteocalcin in these models are striking. So, you know, from a pharmacological perspective, it still leaves open above what are the discrepancies a lot of interesting things to be done. Thank you. Another comment I'd like to make is that some of the potential reasons for discrepancies, you know, the environment where the mice live, or, you know, differences, small difference in procedures could be overcome by exchanging, you know, the mice and cross check. But unfortunately, that's not been done yet. I hope that whoever is involved is able to come to understanding and, you know, and try to work together because, you know, it's failing to address these potential problems is going to be a burden to, you know, to move forward from here. So, I guess I'm saying that we probably people should collaborate a little bit more openly. Hi, there. Katie Basham at the University of Utah. I work on the adrenal gland, and there was a recent proactive study showing that osteocalcin could potentially promote stereogenesis from the adrenal gland. So, I'm wondering about your WNT7B overexpression studies where you're upregulating osteocalcin. Have you looked at the production of steroids from the adrenals in that? We haven't. We haven't. That's interesting. And my second question, I think, is a little bit more directed to the second speaker. I'm wondering if you can comment on the similarity between some of the receptors. So, like GPR158 versus 37 and how those, you know, there was no expression in the pituitary, for example, of GPR158. Can they compensate for one another? Do you think they have similar functions? So, the question is whether 158 can compensate for 37. Or do they have similar function in binding osteocalcin? Are there any differences between those two receptors? Do they have similar functions? Okay. They do have similar functions. Okay. Robert is supporting on this. Great. Thank you. Any other questions? If not, I have one for Anne. And, you know, you went through several, you know, longitudinal, particularly randomized control trials with different medication, changing bone turnover in different ways. So, that's quite convincing. And, you know, the largest increases in bone turnover, not necessarily osteocalcin, but bone turnover, is in Padgett's disease of bone. And those patients usually respond very well to bisphosphonate. Have you found any data on Padgett's patients? I know those tests are not very frequent. They're actually very far, very few and far between. But I was wondering if you have any knowledge of that. I am not aware of studies. But this would be, it's another example of a place where, you know, it wasn't the intention of the original study, but a place where we could go explore, I think, very usefully. So, and I don't, I have no idea if those trials have published, maybe weight change is looked at. You know, you'd have to go in again to look at fasting glucose or look at instant diabetes. It's probably a need to get some funding to go in. Any other questions? Yes. I am Saul Madozowski from NIH. I wish I could make a scientific comment. I would make a comment that does not pertain to science. I think that this session was very civil and the discussion really was extremely important as well as the presentation. Something that really was needed when the early controversy emerged about this topic that has yet to be resolved. And this is an extremely positive development. The other thing that I think is very important, particularly in the current environment, science is challenged, is not to make comments to the press in which we, very early findings, we put forward these reports claiming incredible outcomes that may never materialize. I think this backfires. And we need to be very careful when we undertake these claims to the press. Thank you. Well, thank you very much for the comment. I appreciate also the value of this type of conversation. I just want to underscore your second comment. Before launching into a full career on a certain finding that's been found, that's been essentially supported by just one mouse model in this case, and a finding that may be exciting but may pose some questions about plausibility also based on what we know from human pathophysiology essentially, I think it would be important to pause and do a second study perhaps to verify the first one. This process is now necessary, for example, clinical trials. All the large-scale studies on systematic review, meta-analysis, studies on Mendelian randomization, now they are very popular. Most journals, I can speak as an editor, now require a discovery cohort and a validation cohort. So this could be transferred to also the clinical world to say that yes, there is one study posing and claiming a new exciting finding, but let's hold our horses until we find a verification study that can be done in different ways or just in different groups of animals. I totally agree. I want to echo Roberto's comment. I think it comes down to the premise, right, the premise for hypothesis. And you could argue that a lot of times we generate the hypothesis based on premise that is in the literature or in your preliminary data. I think a lot of cases we need to be more careful with, especially the premise in the literature. I think NIH is well aware of this premise is an important evaluation criterion for grants. Please. Hi, this is Zelal. I'm a metabolic bone disease fellow from McMaster University. So this is my first time to hear about this topic, and I find it mind-blowing and very interesting. My question is directed to Professor Andrea. You nicely explained the pathways in the male and the association between LH and testosterone to osteocalcin. I would be interested to know how does this work for women or in the female, and whether or not there is a relationship between why women tend to lose bone more quicker than men in their postmenopausal period and whether or not estrogen has any role with osteocalcin or a connection. Thank you for the question. Regarding the fact that there is an issue based on the fact that the knockout model by Carcenty, it was only observed in males and not females. And the reason why is this because the receptor, the GPRC6-alpha is expressed only in lydic cells but not in the ovaries. So probably this is the explanation for the difference between the two. Regarding the other question, whether estrogen could be a modulator, I think so, probably estrogens are a modulator in terms of altering bone turnover. So probably it's clearly an issue. Also, if you look at the physiology of steroid hormone production in male and female is completely different because you have cyclic up and down. So it's also reasonable that osteocalcin doesn't have an influence because you need cyclicity. So if you have a constant stimulation, probably it doesn't help, which is somehow different from what occurs in androgens. They have some circadian rhythm, but more or less it's stable throughout life. So probably this is an evolutionary explanation for the difference between the two. But it's just my opinion. Nicola? I have two comments. One is for Roberto and Fancina's basic scientists. Many of the data that have been shown are basically based on four or five mice per group. Do you think, I don't know, I'm not a basic scientist, but doing clinical research, I know how important is the power. Are the studies powerful enough to see differences and to prove physiology with the four or five mice? Yeah, it is underpowered. Yeah, Nicola, you're putting the finger on a very painful wound, I would say, because many of the, especially the older studies, were not really meeting the criteria that we called rigorous now. In other words, no indication of sample size or power analysis to give us an idea of why seven mice in one study and six in another one, for example. Obviously, when you deal with complicated crosses, transgene mice carrying two or three different transgenes, the numbers cannot be big because it's just impossible to generate hundreds of them. But, you know, the focus on power analysis should be emphasized more and more. And, you know, the NIH now requires all grantees to specify, to be explicit about how many animals are needed for the experiments proposing grant applications and why. So the culture is slowly changing, but it's going to take a while before even preclinical studies will reach, you know, the rigor that we expect from clinical trials. Not just to pay respect to our animals, our friends, mice, but also because we do need rigor. And I share your concern that some of the claims may in fact be based on, let's say, a rigorous interpretation of studies that were underpowered or not designed properly. This may, and now this will open another huge discussion, but Roberto's editor, you know, that right now we are also now basically chased by journals to publish and they want the scoop. So sometimes this is another problem. Sometimes in order to get just the paper published and they can be cited quickly, this increases also the, better lower the quality. I have a question for Andrea, if I can make another comment. The last question. Last question. So in the mouse models, the first model from URI, from the group of Carcenti, surprisingly there was an increase in estrogen, but the expression of aromatase basically was not changed. How would they explain the increase in estrogen? Yeah, this is another discrepancy because they didn't explain that because estrogens are increased, but androgens were not. And so it by definition must be an increase in aromatase activity. Otherwise you cannot explain that. But it was flat. In fact. So it's some sort of strange finding that. Okay. And probably it relates to variability. Once again, as you said, probably, you know, you sampled from two or three mice that had some levels and testosterone probably was sampled in a different course. And then you put all together in one table. It happens, you know, because sometimes. Thank you. All right. Thank you, everybody, for staying until the really very end. And I'm going to close the session and adjourn. Thank you very much. I hope we've accomplished the goal that was to present knowledge as it stands without preconceived dogmas coming in. And that is helpful to move forward from this controversy. Thank you very much. Thank you. Applause.

hormonal function

osteocalcin

glucose metabolism

male fertility regulation

bone metabolism

controversies

uncertainties

neural development

cognition

metabolic health

mouse models

human studies

clinical trials