W. David Arnold, M.D. presents at the Johns Hopkins Department of PM&R’s Grand Rounds on December 18, 2018.
So it's really a great privilege to be here and to be able to talk about this work. So I was fortunate to kind of grow up in science as this field was developing as most people probably in that room. I have at least heard a little bit about how exciting the progress has been in regards to therapies for patients affected by spinal muscular atrophy. I'm a small part of that. But what I'm going to try to do today is tell you about the whole story kind of from my perspective and the things that I added to the field and kind of where, how that's led me to where I am now and the things that I'm currently interested in. So, uh I do have a few disclosures. So uh very fortunately was involved in the RSTP program and that's really probably the reason I'm able to stand here. I have also gotten funding from NIH uh from the National Institute on Aging. Uh as I'm going to talk about a little bit, I'm interested in aging related loss of fiscal function. I also do some clinical work. Uh So I, if anybody's ever heard of the Tennessee fainting goats. Uh patients also get that disease, my Tony cona and we have been looking at a therapy in patients affected with muscle hyperexcitability disorders, which I submitted grant and was funded for that not really relevant to this talk. And then the Neurological Research Institute at OS U has supported me. And then I have served as a consultant for neuromuscular diseases as well as spinal muscular atrophy for a major pharmaceutical company at Roche. So just about a year ago, we published our work in this area showing the first in human trial of gene therapy and spinal muscular atrophy affected infants. And this is one of the family members that was part of the trial. I have a small little story. I actually have lots of slides. So I I can't tell too many stories but I will. Uh so this child actually was involved in the gene therapy study. I remember seeing this, this little girl and the first time I tested her when she was about six months old thinking, wow, this is maybe one of the first infants I've ever seen that seems somewhat cured from this disease. And I'll tell you about how the measures that I do kind of told me that. But the cool story about this is my niece actually had severe epilepsy. They live in North Carolina but came to Columbus to have evaluation for potential epilepsy surgery. And she was undergoing the brain mapping. The same time when this family was staying in the Ronald mcdonald house and my mom got to hold this little girl when she was there. So it was a really emotional, cool story that it was like full circle. My mom got to meet this family and they were just talking and my mom was like, oh my son's involved in that trial and then they found out I was the common denominator there. So really, really amazing thing to be part of. So, what I'm going to try to do is give you an overview of spinal muscular atrophy. I imagine a lot of people in the room know a good bit about this disease. But I'm going to give you kind of the genetic underpinnings and the mechanisms as far as we know on why low levels of survival, motor neuron protein cause muscle or motor neuron disease. And then I'm gonna tell you about the progress in the field and how it kind of led me to becoming interested in loss of physical function and sarcopenia of aging. So spinal muscular atrophy is generally, it affects infants the most common form. So it results from motor neuron death. And so that causes muscle weakness and atrophy. It affects about one in 10,000 live births. It is an awesome recessive disorder and patients with more mild symptoms can often have this fine tremor which seems very stereotypical between patients. The genetic basis of spinal muscular atrophy is a homozygous loss of the survival motor neuron one gene. So this is usually a large deletion of Exxon seven and eight. But in some cases, you can also have a small mis sense mutation. But generally genetic testing for large deletions will pick up 95% of S MA cases related to this gene. The really interesting thing about this disease is there's actually two genes involved with this disease and patients with SM have homozygous loss of SM one but a backup gene SMN two. And we'll talk about how it does. This modifies the disease. And so the number of copies of SMN two gene modulates how severe the disease is. So there's this inverse correlation, more copies less severe disease phenotype. This is a mapping of how most people characterize disease severity. And I would throw the big caveat here. This was pre therapeutic era. So type zero is relatively or extremely rare, has prenatal onset and infants affected have need for respiratory support at birth. The most common form about 50 to 60% of S is type one that has onset in my experience, generally 6 to 8 to 12 weeks, but it's basically onset before six months. These infants never achieve ability to sit independently and their natural history of death without invasive support is usually less than two years of age. The median survival is closer to about 8 to 10 months without support. And then from there you go to less and less severe phenotypes. The one that I've actually become most interested in in our center is one of the leaders investigating this is type four, type three B or four and that has onset during adulthood. So, a very, very mild phenotype represents maybe 5 to 10% of all S MA s. So it's relatively rare. This is, this is a phenotype that can be relatively challenging to diagnose, it can mimic things like other motor neuron diseases such as A and, and so this is one of the areas that we've really been particularly focused on. So I've told you that loss of SMN is present homozygous in patients with S MA. So when you look at the uh two genes of uh related to S MA, you have SMN one. So that's on both alleles that's lost usually a large deletion, sometimes a small mutation. But then pa all patients, there's never been a patient reported ever with S MA or otherwise, that has loss of both of these genes. So it's incompatible with life seemingly to not have a some version of SMN two. But when you have SMN two, only SMN is lost. The problem is this gene is kind of what I call a wimpy backup version of SMM one. So it's essentially, there's a single nucleotide change in Exxon seven that causes the majority of the transcripts to only produce this short truncated version of the protein about 10 to 15% of SMN two is translated into a full length SMM protein. But the bulk of it 90% or so, is this short truncated version that rapidly degrades? And so basically produces small levels of functional SMM protein where this produces predominantly 100%. But in S MA patients, they don't have this. So when you look across phenotypes of S MA from very severe to more mild, generally, this type zero phenotype is associated with only one copy. So you may ask, how does one copy happen? Well, so some patients have no copies of SMN two, but they have SMN one and that is essentially that's associated with the normal phenotype. So one parent basically had one copy of SMN one and large, probably deletion of SMN one and then no copies of SMN two that's still compatible with normal function at least as far as the superficial level. But then those other patient may have, other parents may have had one copy of SMN two and then the homozygous deletion of SMN one, that's pretty unusual. The most common allele frequency in the population is one allele of S or one copy of SMN two on each allele. But this region of the genome is very unstable. So there's a lot of duplication and so you have some patients that have no SMN one, but they have a very high copy number of SMM two and have extremely mild phenotype, maybe even normal. Uh and So this, this genetically has been shown to basically rescue the phenotype based on this backup wimpy version of this SMN gene. So how did I kind of get interested in spinal muscular atrophy? So I uh as similar to Tay. So Tay and I have worked together for the last several years. He's doing great work in aging and we have that in common, but we both did neuromuscular training after residency. Uh And so after residency in 2008, I did a neuromuscular fellowship at the Ohio State University and then joined the faculty thereafter. So I was really primarily a clinician between 2009 in 2012 or so, I was really in the clinic full time seeing patients. But I was curious about a lot of different things and I had been kind of rubbing shoulders with physician scientists and scientists who were doing work in this field. They had developed an animal model which I'll tell you about. And there was a therapy that was looking really promising in mice. And so what I kind of started wondering is there ways that I could use my clinical skill set to kind of take from the bedside to the bed to help these scientists ask very clinical, clinically translatable questions. Are there are biomarkers. So biomarkers are, it's a buzzword, we hear it all the time, but really it continues to be important even after you have therapies to understand the disease. So they help us with diagnosis. So that's SMN one homozygous deletion, that's S ma prognosis. That could be SMN two copy number. So it gives you a prognosis on how severe the phenotype would be. And then it's possible that you may have markers that can tell you the disease state and how likely is it that you would have a response to a treatment and that may help you stratify a different type of treatment. So, in this case associated with low levels of SMN, but maybe there's other pathways that we could augment at different stages of the disease. And that then these measures can also quantify a response and help us, you know, modulate our treatment accordingly. We use that in autoimmune neuromuscular diseases all the time. And then my, my main goal is to help understand the biology of S MA. So right now, there are several different types of biomarkers. So we have molecular biomarkers, which I mentioned we have measures of muscle mass which are simple. The one caveat there is muscle mass doesn't always equal muscle function and sometimes muscle mass doesn't really correlate with strength that well, because you can have reduced muscle quality. Uh But the main thing that I thought and wondered about is like, can we do some of the same measures that we do in babies with S MA in mice? Because in mice, we have genetic tools that we can manipulate pathways. Uh We could track them basically through the whole phase of the disease, presymptomatic to post symptomatic and, and, and then do interventions to see what kind of effects we could have. So what I thought is maybe I can create a little mini emg mouse lab and test these mice and see what happens. And so that was kind of very simple on the surface. But what was surprising is nobody had done it. So it wasn't anything earth shattering, but it was kind of like my doorway into getting into science. And from there, I tried to push myself farther and farther. But that was kind of what made me curious. Can I do this? So, electro physiologically, we have a lot of measures that we do in the clinic. I do those. I do a lot of emg every week. We take these for granted. There's actually a lot of data that I think in the clinic that we don't even leverage. So we kind of we have simple things. So for instance, CMAP compound muscle, a potential stimulate a nerve, you measure it from a muscle, it's basically a summation of the excitability of muscle fibers. And normally we look at the speed of the response. So how fast the conductivity along the nerve takes from stimulation to recording. And then we simply look at this the size, you know, basically the output and that tells us a lot. But there's a lot of other parameters that are basically untapped that may tell us more things about the disease. But basically what I did is looked at these measures and tried to understand, you know, for instance, C map functional status of the neuromuscular system to cause excitability of the muscle fiber. So connectivity at the NMJ numbers of motor neurons number of axons because this is kind of an indirect measure because of collateral sprouting, the C map can be normal but you can have reduced innervation. So motor unit number estimation is an estimate technique that allows us to get insight into the number of motor neurons that are functionally connected to the muscle that you that you're interested in. Very cool because there's not many things in biology that you can actually track a population of cells in vivo longitudinally and we could apply. And so basically, I was interested in applying this to animal models in the way that we had done in the in the clinic. Now, of course, it's an estimate, but that's very powerful because motor neurons as you know, are non dividing cells. So the motor neurons that we're born with are basically there for our lifespan. So when they degenerate, there's no known pathway for regenerating those. So you can really get a clear understanding of the connectivity of the neuromuscular system longitudinally. The rationale that I had for kind of investigating these measures is we knew from clinical studies that when you looked at compound muscle action potential, simple, just ner stimulation the hand muscle. Here, it correlated very well with motor scales that we use in the clinic that are quite complicated and involved. So this is a simple measure, totally objective that could give us insight into the status of the neuromuscular system. We completed around the time when I was starting to get really interested in mice, we had a multi center study, a collaborator at his lab is just down the hall from mine, Steven Cole. He's MV phd. Kind of my inspiration to try to kind of learn how to be like him. So he, he was not very senior compared to me, but he was kind of like a parallel peer mentored me. And so he designed a multi center study where we did these measures in S MA babe longitudinally. And you can see that when you compare uh healthy infants with uh uh normal SMN gene and then infants that have three or more copies versus two copies, you can see a very striking pattern of motor neuron degeneration there. Simple. The I think the biggest motivation to taking the leap though was the fact that in 1995 we discovered what gene caused this disease in 2010 at OS U and nationwide Children's in Columbus Ohio. The first dramatically successful therapy for S MA in a mouse model was published. So this was an A no associated virus subtype nine to basically deliver cargo of SM and two motor neurons and rescue the phenotype. And this took a mouse that normally lives about on average 13 days to living over 500 days. And so this was published a couple of years before I got interested, but this was really catching, catching momentum. And so I wanted to investigate these therapies using a, you know, in model and measures that kind of paralleled what we do in the clinic. There was this concept at the time that S MA is a very unusual natural history kind of progression disease. So most patients by the time they're diagnosed have striking neuromuscular defects. You can see the previous graphs I showed the longitudinal, those were infants that were diagnosed and then enrolled in a clinical trial. They were already dramatically separated out. But there are unfortunately only a handful, but there are a handful of infants that have been studied pre symptomatically and then tracked over time. I've seen a couple of patients, this three or four patients like this. And this was in the literature at the time where basically these four infants or three infant, four infants had been studied kind of early diagnosis because of a sibling that was also diagnosed with S MA. So they were identified before disease onset and they tracked them over time. So this kind of colored area here was the question is is there this pre symptomatic error time period when possibly we might have more impact of disease modulation. So the mouse study that had come out in 2010, clearly showed that with the later time point, the mice didn't do as well. The big caveat there though is the virus and mice targets more glial cells than motor neurons after day five, after birth. And so there was this question of early versus late therapy, which I was very interested in. So one thing that's always good when you're getting into an area is try to link into a mentor that can help you kind of delve into the field, learn the background. And then quickly you can kind of become an expert by synthesizing the data that's already out there. So my primary mentor, he's a phd, he's a British Arthur Burgess, he got invited to do a review. And so he asked me to be involved with this. And so I basically dug into this area. I had been getting more and more interested in it and tried to synthesize this idea of pre symptomatic versus symptomatic versus late symptomatic treatment. And what we kind of tried to character kind of hypothesized regarding was these different phases of disease. How much effect could we have on the disease phenotype? And so we predicted with early treatment, maybe if there is this presymptomatic window, can we get full effect? And then as you go later and later disease, basically, the horse is out of the barn and you lose more motor neurons, motor neurons are not replaceable and so you may have less and less effective. And so this is kind of what we predicted as I was designing my career development award to try to test these concepts in a mouse using clinically relevant measures. So this is kind of my bedside to bench the bedside approach. So on uh on the uh right here you see a picture of our EMG lab and then on the uh or excuse me left and then right here you see a picture of my lab that I have kind of like a little Minnie mouse EMG lab. We do physiology and various other measures. So over the course of the last three or four years, we've taken this from humans to mice to neonatal mice, we have a pick model which I'll talk about. And then this is one of the clinical trial infants that you can see that he's standing up. So I'll show you a video about that later now. And this is a type one baby that should not live to his second birthday. So our first paper was basically just trying to define the natural history of the disease in mice. And so we showed that actually early on between uh day of birth. So we started measuring at P three. So I had to really miniaturize these measures. So between P three third day of life to P six, the measures corresponded pretty well with wild type. And then you see this divergent and now this was untreated. And then we did early therapy with a no associated virus to deliver the SMM one gene. And then we also use the antisense technology, which I'll talk about to restore SMM protein, which showed that when you treat early, you abolish the disease phenotype in these mice. And this was the first paper that I kind of leverage these techno techniques to understand S MA. So from there, uh we wanted to understand early, delayed and then late treatment to kind of see how that affects between the different phenotypes. So this was using what we call the delta seven mouse model. Uh So this is a mouse that actually was created in 2005 by Arthur Burgess who was my mentor. So he was really familiar with this mouse and the phenotype that it had. And so we had previously described that the electrophysiological phenotype started at P six. And so our goal was to kind of go right after birth, 0 to 2 days P four and then P six kind of see when it's really obviously abnormal. Uh we use anti nucleotide strategy for this approach. So this is basically a nucleotide sequence that binds to an area around the gene that makes the SMN two gene act more like SMM one. So instead of only producing 10 to 15% of the full length protein, it makes it produce more like 100% of the full length protein So increasing SMN levels using this anti strategy. So when we did this uh experiment, we tracked survival. And you can see that the diminishing returns, the later you treat we also we as a good read out for these mice. And you can see that the same kind of strategy there. And then we did electrophysiological connectivity and you have C map and you see that the with C map, it's reduced but it's almost starting to normalize because of collateral re innovation. But with delay, you can see there's loss of motor unit number uh with the later treatment. So this suggested that early treatment is probably really, really critical exactly how this aligns with the natural history in humans may vary between individuals depending on their copy number, depending on other genetic factors. So we know there are modifiers in the region and outside the region of genes. So different patients act different. So just because they have two copies SMN two, it doesn't mean they're always going to act like the classic textbook SM one. But generally that is a pretty good association. So after the mouse model gave us some good data, we actually used a different strategy. So there was a group that was trying to generate a pick model of S MA. So I'm not going to go into this detail too much. But humans are the only vertebrate species with two SMN genes. So mouse model had to have a trans gene from human, put in with SM two pig model, we were hoping to have an SMM two containing pig but because of serious complications that didn't happen. So what the strategy that we took instead is we used a no associated virus. So that's a non integrating viral vector to deliver short hairpin RN A to knock down pig SMN protein. Then we did early and light therapy with anos associated virus to put the human SMN back in. So this short hair pin RN A is specific to pig and then we put the human in to replace it. So we did early and late. So we treat, we injected at P five. So you can see this picture here. So uh I'm not an injection type person. They made me do it in residency. I didn't love it but I learned how to do pig cisterna magnet injections and it was actually really cool to see how it outlines the CS F space. So this was a CS F delivery. Um This virus has specific a relatively good tropism to motor neurons. So it's really good at sending its cargo to motor neurons. And that's the power of this strategy. So we treated early and late without treatment, we had a really nice phenotype. So we did C map and uni so we translated that from human to mouse to a pig. And you can see really striking reduction with knock down and then pre symptomatic treatment, you can see it almost normalizes those measures really good, real simple measure. And then we also did morphology showing ventral root loss and motor neuron loss. Um I'm gonna show you some videos if I can make these work here. Uh So this is basically a ramp. Uh and the pigs uh go for like sweet and condensed milk. It's at the top. So they, they really do like to run to the top here And you see that's a normal looking pig. And you can see early rescue here looks pretty normal. This was now that I've seen these therapies in humans. This used to be super exciting. Now it's like, ok, great. Let's get to the human stuff. Uh uh And, and then here's late rescue and no rescue and you'll see that the late rescue looks pretty good. Uh but definitely has some defects which I'll show you electro physiologically here in a minute. So you can see some some hind limb kind of weakness and slipping there, but overall pretty good. And then no rescue you'll see, unfortunately, cannot get up the ramp really whatsoever. So the weakness started in the hind limb with this phenotype and progresses to the fore limb. When we do electrophysiological measures across control, knock down early treatment, late treatment. You can see there is this kind of diminishing return on the number of axons on ventra root counts and electrophysiological electrophysiological. I thought it was really cool to see how these really matched up with each other. So Munini and anatomical counts, there's really no good way to validate each of those. And so it's more conceptual in a lot of ways, but these really, really gave us a lot of confidence that these might serve as a really good biomarker in clinical trials. So basically, what we showed is there's a clear electrophysiological phenotype and models of S MA pig and, and mouse. When we treat early, we have larger effect. When it's treat late, we have less effect. And if it's very late, we have no effect. And so um the the uncertainty of what would happen when we took it to the clinic was really exciting to me because I was really interested to see how this matched up with the first in human trials. So I wasn't involved in the anti sense oligonucleotide therapies, the Nusa nursing, I'm going to present some of that, but I've been involved primarily in the gene therapy strategies. Gene therapy is earlier days, but it's currently under determination whether it's going to be FDA approved Nusa Nursing. The A O therapy is currently FDA approved as of a couple of years ago. So I'm going to kind of talk about the time dependency of response in the early clinical trials of both of these therapies. So when you look at this, it's kind of a busy slide. But what this is comparing is the different clinical trials of Nusinersen. So Spinraza is the label name. This is FDA approved for Spinal Muscular atrophy all types. But you can see the clear differences. So this is basically a motor scale and hires better. And this is the pre symptomatic therapy versus the symptom symptomatic clinical trials. And so clear difference between those. The question is how close to normal this is getting probably not normal. And so the way they define pre symptomatic, they use basically compound muscle action potential, which I've been doing animal and clinically as a cut off. And they said one millivolt or higher met criteria. In my experience, the one millivolt is not normal. So most of the time when I've studied presymptomatic infants, their C maps have been in the range of five and six. So their definition of presymptomatic may not been fully prey. It may have been early symptomatic. So we really don't know early as possible. So maybe shortly after birth, how sufficient is therapy going to be. But in these cases, you can see dramatic improvement with early and late. I think the coolest thing when this went to the clinic is actually, I think later treatment in humans seems to have more effect than the mice did. So normally we see this loss in translation from mouse to human. But in this case, I actually was really, really surprised that with these late treatments, we still got response, I thought that was really amazing. So one is clinical diagnosed S ma the other is like genetically diagnosed and, and it really depends on like functional status. And the pre and post symptomatic was really only defined in this trial. And they used the compound muscle action potential threshold and they used six weeks of age. So six weeks of age and C map amplitude as kind of their threshold. So less than six weeks and C map greater than one millivolt was their criteria for pre symptomatic. And the other trials were they just couldn't be vent dependent. So basically, it was any stage of the disease. So some of these infants were quite late, six months of age, eight months of age and so on. Does that make sense? Yeah. The same kind of pattern that we saw in Nusa nursing, we also saw in gene therapy where you see this is the low dose cohort. So this was a first in human phase one FDA trial was really cool to be involved with this. I saw these infants at baseline one month post and every three months. So all these visits, uh I got to kind of see this progress in the field and just had to keep it to myself, which was really exciting. But you can see clearly that there's different ages of when these infants were enrolled and it shows a striking association with age and response and there's a paper that's hopefully coming out soon, that's going to detail some of these changes more because just because these were younger didn't always mean that they were more functional. So it didn't even always line up with how functional they were at baseline. It would seem to be really closely closely associated with age. Now, the one thing I would say is these, these two infants were clearly the, uh the ones that were least symptomatic when they were treated and you'll see some videos of uh, one of these infants in a minute. But yes, yes. So chop and 10 is just like a motor function scale and so lower is bad is high is good. Generally, no type one s ma infant ever gets above this 40. So the fact that they're shooting over that and then they're hitting the ceiling here is really just amazing. I mean, it's just, I mean, it's really, really incredible. Yeah. Good questions. No. Yeah. So of the master. So in this, we actually all the data suggests that it's going to persist but we don't know. So, so in this strategy, you're using a associated virus to target motor neurons. We've done a lot of work in mice showing that sin levels really if you increase in motor neurons, that's all you probably need. We've done a series of pre locks p experiments like deleting and increasing in different cell tissues and showing that a few years ago. But motor neurons are non dividing cells, it's a non integrating vector. So in dividing cells, you're going to lose it in non dividing cells, it should persist. Could it become inactivated down the road? It's possible we just simply don't know. But in theory, this could be a one time treatment to restore SMN levels in motor neurons. I follow you once it's more about putting the gene back in. So it's a self complementary, like set up and it's under a continuous promoter. So it's, it's basically the A V delivers the cargo of the gene to the cell and then it's there. So it's that cell is producing its own SMN. Now, in dividing cells, it's a non integrating. So it doesn't go into the genome because there's concerns with that you could hit a oncogen or something like that. But this in motor neurons because they're non dividing. It's just sitting there and it should, unless it becomes inactivated, which we don't know, it should produce smm protein lifespan as long as the cell is alive. Maybe, yeah, maybe you need more requirements or who knows? So this is the girl that my mom got to hold when she was much lower than this. And it's not sorry about the video quality. This was the parents that took it. But she, you know, the it was, well, I thought I was gonna, I thought I clipped that. Sorry about that. Ok. So you can see that she has weakness. I think anybody who deals with pediatrics can see that she's clearly weak, but she was treated uh definitely after symptom onset. And to me, that's amazing that you get this kind of rescue. So I would say she may eventually be pheno typically more similar to like a S ma type three. So partial motor neuron lost. Uh but compared to type one, that's amazing. Now, then this other uh child 14 months here, I can tell you a few months after this he was run away with for me when I was trying to do my measures. So uh this is like right when he first started walking, so now he runs and he can reach up and touch the elevator. I mean, it's just like he looks normal, she looks dramatically improved and now that all patients get this this much improvement. No, and this is because these are the two youngest patients that were treated. So there's a paper as I mentioned coming out that will give more details on these cases. So crazy, exciting progress in the field. I was like really fortunate to be able to grow up in science as this was happening. Uh It taught me how to be critical to work in the lab, to build a team. And then I also got to see miracles like this which were really like really triumphant in clinical trials. The question is how sufficient these therapies are going to be. And I think we still don't know that. So the one thing that I would so people probably in the room, are familiar with polio and post polio. So if you partially lose your motor neuron pool, the other motor neurons can compensate and cloud or reinnervation. So on the surface, you get the tip of the iceberg and you may look kind of normal from a functional standpoint. But as we age, you lose motor neurons. And so it's possible that they could have accelerated aging and they could have like a sarcopenia phenotype. And that's what probably causes patients with S MA to get worse as they get older. At least part part of it, we don't know that for sure. So we still need good measures of motor function to track that because what if their A AV quits working but we don't know it and then their SMN levels are dropping and then all of a sudden, like if anybody's familiar with A LS, you don't really get clinical defects in muscle until you have striking loss of motor neurons. So you could just have this S min levels be dropping and the motor neurons could be dying. But collateral renovation is kind of keeping up with it and all of a sudden devastated. So we still need better markers and better understanding of what happens in these patients. We need to know SMN levels clearly change during age and young development is really high and then it goes down, it seems to be important for neuromuscular junction maintenance and repair. But is there a way to track how much a patient needs. And so that's really challenging. So people are working on that. And then, as I mentioned, could we have an accelerated aging phenotype in the girl that was partially treated but not completely normalized shifting gears here a little bit because S ma kind of spark some interest in the role of motor neuron function as we age. So there's a lot of clinical similarities between spinal muscular atrophy patients and age related muscle weakness and atrophy. And because the motor neuron is a non dividing cell, there's no way to replace it. So that has to do a remarkable task. So if you think about the size of a motor neuron, it's 1 10/1000 of a meter. And then that single motor neuron living in my spinal cord sends an axon all the way to the muscles in my foot. It has to maintain that connection for my whole life. In addition to that motor neurons during contractions, depending on what type of contractions you're doing for a ballistic contraction, it may fire up to 100 times a second to cause the muscle contract. So very energetically active, maintaining the NMJ a meter away and having to do that for my whole life. And so I was really interested what happens in the motor neuron as we get older. And so there are some studies that have looked at this and I'm going to kind of walk through some preliminary data about some of this that we've looked at mainly in mouse models, but a little bit of clinical data that other people have also established. So, sarcopenia, how does it affect individuals and society? So, sarcopenia initially, which is described as loss of muscle mass has become more equivalent to muscle mass and weakness because people have shown clearly that muscle weakness during aging drops faster and that muscle strength is more tightly associated with outcomes. So if you look at mass, the the amount of explanation that mass provides for people related to outcomes and function functionality is not very good compared to muscle strength. So basically, mus muscle uh mass has always equal strength. Sarcopenia has a dramatic impact on individuals large increase in impairment and loss of mobility, but also associated with all cause mortality as well as cancer mortality and the cost of the US is significant. So, prior work, about the time when I was born, I was looking at motor neurons uh and uh how those change across the lifespan. And so as I mentioned, nondividing cell, no way that we know of to replace those. And there is this capacity of motor neurons to re innovate and compensate. So each motor neuron actually has about uh 2 to 3 times the ability to reconnect the muscle fibers. So if a muscle has 200 motor neurons, you can lose up to about two thirds of those and still maintain pretty well your function. So kind of that's related to the post polio polio phenomenon. But during aging, there's known to be loss of motor neurons and motor units and impaired reinnervation through a series of experiments. So uh two groups, as I mentioned uh showed that across the lifespan, there's significant loss of motor neurons. They showed approximately 25% when in including all individuals. But clearly there was variability between individuals. So some individuals had greater than 50% losses is important to note because and actually there's not a lot of clinical data out there, patients that have sarcopenia, not everybody as they get older gets quote like diagnostic meeting criteria of sarcopenia. And most of those criteria are based on cut points of populations, but some do. And in this, in this study, they showed that some individuals seem to have more propensity to motor neuron loss. So is that genetic factors is that environmental factors is that inactivity? What's driving that? And so the big question I have is a sarcoptic patient versus a non sarlo patient? How's their motor neuron count? Motor unit count line up with that subsequent other groups showed that when you look at motor neuron counts and ventral motor axon counts also showed reduction in aging electro physiologically. One of the first studies after motor unit number estimation was uh uh kind of described by Macoma in Canada. They looked at it in across the life span and they showed clear loss in older adults. And so this was kind of the rationale for looking in mice and study single fiber which Tay and I have both been working in great deal with looking at neus junction transmission. So in the clinic, I do a lot of single fiber in our EMG lab, I'm really interested in ne muscular junction disorders and this has been shown to have defects in aging. This hasn't really been looked in really detail. And Tay is actually working on that right now. Uh One study using a different technique uh just recently reported that when you look at sarcoptic individuals versus non sarcoptic individuals, it seems that sarcoptic individuals seem to have more impairment of neuromuscular junction reinnervation. Uh which I found interesting um uh uh because that suggests that maybe you lose motor units and you can compensate in some individuals. But if you can't compensate, then maybe that's why some people have more susceptibility to sarcopenia. So all that as the backdrop, my hypothesis kind of moving into the aging world was whether or not uh loss of connectivity of motor neurons and the motor unit could lead to loss of physical function with aging. Um And really didn't have any hypothesis where that's coming from because the the problem with the motor unit, you have so many different factors that are interacting and a lot of people in these fields are either muscle people or motor neuron people. So what I tried to do is have a systems based approach to look at the whole motor unit and and try to figure out is it coming from the muscle synapse, motor neuron or all of the above? So our approach. So back in about 2013, 2014, when I was working with S MA, the biggest key with a career development award is to always be thinking about the next step because a career development award depending on who you are, that might be your first step towards independence. For me, it was learning how to be a scientist and my independence step was going to be next. So I was trying to think of the next concept that my lab would be investigating primarily. So really investigating motor unit connectivity was that for me. So I back in 2013, 2014, basically got 10 old mice as old as I could get as around 10 months, which is actually not very old and just started tracking them longitudinally to see when their motor unit connectivity degenerated and how that was associated with muscle mass and muscle contractivity across the lifespan. So using these measures, small numbers, but we showed and these are repeat measures in the same mice. So we tracked them about every couple of months and we showed that actually loss of motor un connectivity was a relatively early phenomenon. And so we recently uh uh published, published this, this work and So uh a couple of interesting findings with this. So 20 months of age is relatively young for a mouse, the single motor unit potential, the SM is the size of a single motor unit. And I expected that would go up with the reinnervation during aging, but that wasn't a clear signal. So it made me wonder, is there a loss of reinnervation capacity in aging mice? But clearly the motor unit number reduced, We also looked at uh in vivo imaging with MRI. So I got a small pilot grant to do uh MRI the hein limb and we quantified the muscle, I had an undergraduate student that learned how to do this. Um and in small numbers, but it was very interesting that we saw a very striking association between muscle volume and motor unit number in old mice. And then we next wanted to look at. So sure, it's great mo unit number estimation is a nice electrophysiological measure. But we wanted to kind of branch into actually muscle function. So in the last decade or so, there's a technique that allows in vivo assessment of muscle contractility, you can do this repetitively over time. So kind of like the final common pathway. What we really care about is how is the muscle functioning. So we want to look at muscle function compared to motor unit number. Um And uh we investigated this, this study was for simplicity and and efficiency. This wasn't a longitudinal study So we took cohorts, we had previously shown that motor unit connectivity started at 20 months. So we wanted to take kind of midlife uh early mid and kind of early aging to kind of see how these measures were associated. So this measure is uh in uh it's ridiculously simple. So it's like a little accelerator pedal that the mouse's foot is taped to. And then you stimulate the tibial branch of the cytic nerve to get planar selection. Or you can do the fibular nerve to get Dorsa selection. And you can measure twitch torque or titanic torque. You can see here repetitive twitch responses at five Hertz. As you increase the rate, you get a summation of the twitch responses to get a muscle output. And so this kind of gives us assessment of muscle contractivity, muscle strength as well as forced frequency of relationships and things like that, which would give you insight into muscle fiber type and things like that. And so when we did this measurement, we showed uh at the age where we see the motor unit loss, we do already see motor unit drops. So you can see that that is reduced over time. Motor unit number is reduced. We looked at the association between motor unit number and muscle contractivity seemed to be pretty well correlated. The one big caveat. And I think this is I admit this is a weakness of this is this is all the mice pulled together So there's multiple variables that are changing between the mice. One is age, one is the modi number and the contractivity. So this is a BVA correlation. So you have to consider this could be primarily driven by age rather than the interaction between the two. But still interesting doesn't tell us cause or effect, but still gives us some exciting information to kind of go on. So muscle size and muscle contractivity seem to be connected to motor unit numbers. So the question is, is that maybe a target that we can kind of address? And, and so uh that's basically the approach that we've been taking um in the last few minutes, I'm going to kind of tell you about a very serendipitous situation. Uh uh where kind of both of my worlds kind of came back together. So it was kind of a full circle. So in 2014, there was a group. Uh so I'm around, he's at Columbia. He actually was a post doc fellow with Arthur Burgess. My primary mentor developed these mice that you could turn SMN levels up and down to different time points. And what he showed if SMN levels are low through development, they develop S MA if you allow them to develop for the first three weeks with relatively high levels of SMN, they look normal and then if you turn it down, they develop this kind of early phenotype of aging. They also did experiments where they did purple nerve injury and showed that the mice with low levels of SMN seem to have impaired regeneration. So these studies suggested that there's an early period when SMM levels are really needed, this might be relevant to S MA it also showed that SMN levels seemed to be important for neuromuscular junction, like maintenance as well as repair. And so I was actually, I'm a big runner endurance athlete kind of person. So I was doing a long run out in the woods. It was really bad weather. So I probably shouldn't have been driving. It was super icy and my friends that were going to run with me didn't show up. But I remember right when this thought hit me, I was like, well, we have my stand and the vivarium that have high levels of SMN, how do they age and how do their nerves regenerate? Uh So basically, I went right back, started thinking about the experiments. I uh uh basically transferred them to another protocol that had nerve injury in it and, and did went straight away to do these experiments. Super excited. So let me tell you a little bit about these mice that, that we call a high copy or a high sea mice. Uh So in 2000, so almost 20 years ago, these mice were generated where there's low levels of SMN. And then they have mice that actually have uh 8 to 16 copies of SMN two. So that they have very high levels of SMM protein from this backup gene. And uh basically, these are the mice that were down the vivarium that I was interested in. When I started digging. Actually, this is a scan of uh uh census and lifespan from I think 2003, 2004. They actually, I looked at the average life span of these mice. And when you compare this, it's, you know, it's to take it with a grain of salt. When you compare this to prior lifespan studies, these mice are on the feb background. So those are like white mice. Uh their average life, our median lifespan was between 25 to 50% longer than wild type mice. So, you know, when you over express a protein, you're always worried about toxic effect. I wouldn't say we can argue too strongly that this increases survival, but at least it's not a negative, but maybe it's a positive. They also seem to have generally larger litters and they seem to be bigger and that's the opposite of aging. So aging mice, the meth mouse is tiny, it's a little dwarf mouse. And so this kind of breaks the rules in aging, which is to me, very interesting. Uh So I did a couple of our three different experiments to kind of dig into this and this is all early days. So what I just kind of want to tell you the next step of my journey towards becoming a phys physician scientist. Um So the first experiment, this was an experiment I did in 20 late 2014, early 2015, I believe that's when it was showing basically peripheral nerve injury. So this is a nerve injury of the static nerve and then we track these measures in vivo. So this is c map immunity. At this point, we didn't have muscle contractility. So I couldn't do those measures. So how this aligns to muscle contractility. It's hard to say it was actually really striking how much of effect we have been doing nerve injury in other paradigms. And so we had this kind of optimized. And so I quickly applied this to this question and it was really quite striking. So right now we are generating mice for. So this is transgenic. They always express high levels. What we want to do is have mice that we induce SMM levels after nerve injury. So it's more translational. Next thing small numbers again, I apologize for, I don't know why my graphics have gotten distorted here. But basically, we have four month, 18 month and 26 month mice. And you can see that C map I Muni and aging seem to be improved Muni more than C map. I don't know what to make of that. But that was pretty encouraging for the aging responses. This is just showing you the graph of the survival. Um And then the last experiment which I'm actually most excited about, but it's also the most challenging. So basically, we have generated mice that have about wild type level SMM protein using a combination of knockout and transgenics. So we have done uh ECL Eliza assays for SMM levels and different tissues and shown that these are around 1 to 2 times wild types. So they're a little high at baseline, but we're taking these aging them out for 26 months and then treating them with a small compound R G7 800 which has been developed for inducing SMN levels in S MA patients and treating these mice after they're at a relevant age or 26 months, which is really old for a mouse. This is just a schematic of the of the uh how the drug works and increases full length of expression. This is the drug molecule. We basically, this was generated by Roche but we actually had it generate, we had somebody compound this for us. And so this is parallel to Roche what they're doing. And then we have done these preliminary experiments. So, electrophysiology, small numbers can't get excited about it really no difference there whatsoever, but the red is treated blue is untreated. But then what was really cool is the muscle contractility and these were all done blinded to treatment. We did vehicle and uh RG 7800 using oral gage, but muscle contractility seems to show a pretty striking effect. So it's not surprising that motor unit number wouldn't be changed because those that shouldn't regenerate, but NMJ connectivity might improve. And so when you normalize this, I didn't show this. But when you normalize this body mass and a muscle, it shows the same signal. So what that suggests is maybe improve muscle quality, so improve contractivity because maybe the NMJ connectivity, we also need to look at the muscle effects directly as well. So I'm gonna kind of come to a conclusion here. And um so I think I've being able to show you the excitement in the field of S MA and, and how much of a privilege it is to be able to do this in mice and pigs and in humans and how it's really a great paradigm that I think can be reapplied to other diseases. I would, I would kind of caution that S MA in a lot of ways is primed for success because you have this built in target. But I think it's been amazing to be part of that. And then I think there's still a lot of things we need to do to understand estimate across the lifespan, whether or not it improves connectivity with nerve injury and aging. That's a different story. But I think for SM this is also important. And so what this would tell us, I think is it may be that we don't yet know the optimal level of SMM protein. So for S MA patients, you know, we need to know how high we need to take it before we're going to get the optimal effect and when that needs to be high, does it need to say high through life span? Those are unknown questions. But I think this work will hopefully get us closer to some answers in that regard. And really the main summary of my lab is it seems that motor unit connectivity is important for aging. So that's like a target that we're using to try to develop therapeutics. I also wonder could sarcopenia be a motor unit penia. And so could uni be a good biomarker in aging studies because everybody is focused on muscle mass, but maybe we need better mechanistic related biomarkers. And so we're hoping to move this forward. So right now, we have some experiments where we're trying to look at SMN levels across the life span in motor neurons using laser capture micro dissection. So basically motor neuron slices where you use a laser to get tissue from just motor neurons and kind of enrich sample and see if that changes across the life span because we know that it changes during development. But we want to know when old mice, is it low? And that's really mostly pushed by the reviewers of my grants. Um So lots of people that I've worked with uh over the years and and um very fortunate to have mentors that have let me like this crazy physician that wanted to learn how to pipe pet uh after residency after fellowship, you know, so uh uh I one of my post docs. So um she did a uh uh uh her graduate studies in Arthur Burgess's lab. And so she was a grad student when I made this transition. And now she's a postdoc in my lab. And it's just this really amazing environment to be able to grow in. And so I work with lots of different people that help us develop the virus strategies and the measures and people that have been mentors in neuromuscular field as well as aging. And I'll take any questions. Go ahead. Yeah. Mhm Yeah. It's a great question. So actually in the S MA field, they did an exercise study and they showed that SMN levels go high and also extend survival. We're doing some marathon runner mice studies of motor un connectivity across the life span to look at that question. So I would, I would say the short answer is probably the problem is it's hard to measure. So depending on where it's going. So it's different in different cells. It's different at different life points of the lifespan. So it's really noisy in patients and it's not easy to track. So if you look at their blood, it doesn't always even show a difference between types of. So it's not very reliable. So they're working on better measurements of that. But that's a great question which I think the answer is yes So. Mhm mhm. Are there any ambulance? Yeah. So yeah, so the first question is the function of SMN and, and that would have been probably really good for me to mention sometimes I give talks to people that know this field forward and backwards. So I apologize for that. Second question is whether it's altered in sarcoptic patients or not. So the first question is uh it's really important for SNP assembly as uh assembly. Uh So RN A metabolism uh so in aging, you do have increased species of RNA. So it is possible that that could be a role. They don't know for sure if SMN levels when they're low, if disruption of surp assembly is the mechanism of S MA we don't know why low levels of SMN calls a it's a big huge unknown. It's really amazing. 1000 different mechanisms have been invoked but nobody knows. As far as your second question, I think there's two things I would say to that one. No, people haven't looked at that. The second thing is because we have these differences of genotyping in the population. We could actually look at that genetically in humans to look at dosage of SMNSM one and SM two because SMM one actually is duplicated in some people. So some people have three copies of SMN one and would they be less susceptible to sarcopenia? So those are questions that I hope to be able to address more in the clinic with databases like there's a health ABC database of older adults aging and they have all kinds of phenotypic readouts. But the animal will give the rationale for NIH to let us use that database. So that's, that's down the road. But I hope we can address those questions. Mhm. Has. No, it's actually my backyard. No. So this is OK. So my, my primary goal in the last six years has become, has been to become an independent physician scientist. I got my own independent lab space within the last year. So most of this I was squatting in other people's space, which it's fine. That's actually usually I would say starting out that's the better way to do it. But this year was really an amazing year. But my other big goal and some time on my life span two years ago, I did a relay across the English Channel, but sometime I want to do that solo. So this is Dover England and one of my favorite places. Really cold water. But I hope to go from there to Calais France one of these days. Yeah.
Related Presenters
