Spinal Muscular Atrophy, From Bench to Bedside

Darryl De Vivo, MD, Sidney Carter Professor of Neurology, Professor of Pediatrics, and Director Emeritus of the Pediatric Neurology Service at Columbia University Medical Center

Darryl De Vivo, MD

At the 47^th Child Neurology Society Annual Meeting in Chicago, Illinois, Darryl De Vivo, MD, the Sidney Carter Professor of Neurology, Professor Pediatrics and Director Emeritus (1979—2000) of the Pediatric Neurology Service at Columbia University Medical Center, introduced the symposium on spinal muscular atrophy and the transition from reactive care to proactive care in the molecular era. NeurologyLive sat down with De Vivo for an exclusive interview to further delve into his presentation.

It’s a very exciting time in the spinal muscular atrophy space, De Vivo noted—127 years since spinal muscular atrophy was first described and the community finally has nusinersen (Spinraza), an FDA approved drug effective in treatment of this disease. De Vivo added that spinal muscular atrophy was approved this year as the 35^th disease to be added the Recommended Uniform Screening Panel for newborns, opening the door to newborn screening for spinal muscular atrophy at a state level.

While we may be at a point in time where we can treat these genetic diseases, a new set of problems is revealed. Basic science, De Vivo explains, can be used by clinicians to help make informed decisions regarding the examination of infants and young children during development.

NeurologyLive: Can you discuss your presentation "An SMA Primer for Clinicians (1891 to 2018): A Tale of Two Genes?"

Darryl De Vivo, MD: It’s a great pleasure for me to have an opportunity to introduce the symposium on spinal muscular atrophy. I’m going to be talking about the history of spinal muscular atrophy over the last 127 years, and I’m going to divide it into 2 chapters.

The first chapter was the first century from 1891—1991 when physicians, after coming to realize there was such a condition as a motor neuron disease that affected infants and children, wrestled with the idea as to whether the different clinical presentations of the phenotypes were in fact part of the spectrum of one disease or whether they were several different diseases that shared a lot of similarities and therefore they might not necessarily be caused by the same mechanism.

It really wasn’t until 1990 or so when a scientist at Columbia University Conrad Gilliam, PhD, and his colleagues localized the candidate gene for spinal muscular atrophy to a specific region of the human genome, mainly the long arm of chromosome 5. And so, in 1991 there was an important clinical meeting that brought together several physicians and scientists who were interested in the subject of spinal muscular atrophy to try to develop an up to date classification. They based it on the clinical severity and the level of development that the infected infant was able to achieve.

The most severe phenotype involved the infants with the type-1 disease who never were able to sit, the type-2 disease was thought of as kind of the intermediate form of the condition where the infants and children were able to sit and possibly stand but were not able to walk, and the type-3 disease involved the milder phenotype where the children were able to stand and walk. Now, both the type-2 children who were able to sit and the type-3 children who were able to walk might lose their ability at some point later in their disease, but at least in the beginning they were able to either sit or walk, defining the type 2 disease or the type 3 disease, as such.

Those were the accomplishments of the first century since Werdnig first described the entity in 1891. Then the second part of my presentation focuses on the last 27 years, from say 1991 up to the current period of 2018, and most specifically in 1995 when attention focused on the genotype, the molecular basis for the clinical phenotypes. It had become increasingly clear that this phenotypic spectrum in fact was the consequence of the same genetic abnormality, but there was something modifying the expression of the disease, so that some infants had a very severe form and other infants and older children had a much milder form.

In 1995 with the work of Judith Melki and her colleagues in France, drawing on work by other’s leading up to that important observation by Melki and her colleagues, it was recognized that they were in fact 2 genes. One gene, which was originally called SMNT, T standing for telomere later called SMN-1, and a second gene located more proximal to the first gene, called SMNC, C for centromere, and later called SMN2. It turns out that both the SMN1 gene and SMN2 gene were nearly identical; there were only 5 nucleotide differences between SMN1 and SMN2 and only 1 of those nucleotide changes was essential or critical and that was a C to T transition at nucleotide 6 in exon 7.

As a result, that caused alternative splicing of the SMN2 gene, such that only 10% of the transcripts of the SMN2 gene produced full-length SMN protein, 90% of the time it was a truncated unstable gene product that was produced and was rapidly degraded by the cell as such. We suddenly came to realize that there are 2 genes, the second being an inverted duplication of the first gene, a mutation in the first gene SMN1 caused the human disease that we know of as SMA and the SMN2 gene allowed the fetuses to develop prenatally such that they would be born and then depending on the number of copies of the SMN2 gene they would either have the severe form, which unfortunately was the case 60% of the time, or the milder forms, the type-2 or the type-3 forms of the disease. The number of copies of the SMN2 gene determined the phenotypic severity of the disease, as such.

Now, what has happened since 1995? A number of exciting observations have been made, not the least of which has ultimately proved treatment for this disease, an antisense oligonucleotide that corrects the splicing era that exists in the SMN2 gene, and as a result of that increasingly we can convert the SMN2 gene to function like the SMN1 gene, so the therapeutic target is a built-in gene in our genome, an androgynous copy of the SMN1 gene with the splice error built in and the therapy corrects the splicing error and allows us then to have a functioning SMN1 gene with an increased amount of full-length SMN protein.

We’ve come to realize a number of other important observations, for example, this is thought to be a motor neuron disease and the classical neuropathology does show us that the motor neurons wither and die increasingly as a function of the severity of the clinical phenotype. On the other hand, early on, one of the earliest manifestations, particularly of the severest form, type-1 disease, is the loss of tendon reflexes. Normally if you have a pure motor neuropathy you may have mild reduction in the tender reflexes, but you don’t have loss of the tendon reflexes. More commonly if you have a sensory neuropathy, one of the earliest findings is a loss of the tendon reflexes.

Secondly, patients who have this disease tend to have weakness mainly of the girdle muscles and the paraspinal muscles rather than the distal muscles. Again, if this is a neuronopathy, like a neuropathy, you would expect the more distal muscles to be affected more than the proximal muscles, but as it turns out in this disease the proximal muscles are more affected than the distal muscles, so another puzzling issue that we now have an answer to.

Thirdly, why is it that you must really treat as early as possible? Ideally if you treat before the newborn baby develops clinical symptoms the outcome can be extremely gratifying, and the infant can develop normally while remaining on treatment. Alternatively, if they become fully symptomatic it’s very hard to rescue the clinical phenotype, so we give them the same treatment we give to the pre-symptomatic child, but we don’t get the same gratifying response—we may get a little bit of benefit but not nearly the degree of benefit that we get when we treat very early—so that’s a third question.

The questions are: loss of tendon reflexes, why do we see that early; secondly, why do we have what looks like a muscular dystrophy or a myopathy in the setting of a neuropathic process; and thirdly, why can we treat the presymptomatic infants very successfully but have great difficulty rescuing patients who have been chronically affected and symptomatic with the disease, as such.

It turns out that some wonderful work, particularly by our investigators at the Columbia University Motor Neuron Center in New York City, George Mentis, PhD, Livio Pellizzoni, PhD, and Umrao Monani, PhD, all basic scientists, it really shed a lot of light on the questions that I just presented to you a moment ago.

For example, Mentis has shown it is not a cell autonomous disease, it’s not a disease that just affects the motor neurons, in fact it looks like the primary driver for the sensory motor circuit dysfunction starts is a disturbance in the proprioceptive neurons, the dorsal root ganglion cells that converge on the motor neurons, so they are proprioceptive synapses that converge on the motor neuron dendrites and soma that are lost, early in the disease, and as a result, you get dysfunction of the motor neuron, you get loss of the potassium channels that are so important on the surface membrane of the soma of the motor neurons and ultimately you get death of the motor neuron, as such. It looks as though the first event is the DF fomentation of the motor neuro, the loss of the proprioceptive synapses lead to a disturbance in the electrophysiology of the motor neuron, a decreased of a motor firing of the motor neuron and as a result instead of getting a good contraction of the muscle, you can just a little bit of twitching before you get no stimulus at all to the muscle fibers, as such. In my opinion that is the correlate to the observation of the clinician to find a loss of tendon reflexes early in the process.

Secondly, Pellizzoni and Mentis also made another important observation—there are motor neurons in the spinal cord and lower brain stem that appear to be quite vulnerable to SMN insufficiency. There are other motor neurons that seem relatively resistant and when you look at the motor neurons that are vulnerable, they tend to send their connections to the muscles that subserve our girdle muscles in the shoulder girdle and the pelvic girdle, and the paraspinal muscle. The question is why are those particular motor neurons vulnerable to death whereas as the other motor neurons are relatively resistant, even in the face SMN insufficiency. It turns out that with SMN insufficiency, we get upregulation and activation of another gene called P53, but only in those motor neurons where P53 activated is phosphorylated, do we actually get death of the motor neurons. Upregulation and activation of p53 is an important first step, but then the second step needs to be the phosphorylation of the activated P53 in order for the motor neuron to undergo program death.

Now it turns out that these investigators followed it up most recently with a very nice paper, published this year, showing that there are 2 genes that repress the genes P53, there are non-redundant repressors of the P53 gene. When there is SMN insufficiency, that produces a splicing error in those 2 repressive genes, so they lose the capacity to function by repressing P53 and therefore P53 is upregulated and activated. Then in those cells that have the capacity to phosphorylate the activated P53, those cells go onto to die, and those cells are the ones that send their connections to the muscles involving the shoulder girdles, the pelvic girdles, and the paraspinal muscles, and that helps us now understand why a neuropathy looks like a myopathy in the setting of spinal muscular atrophy.

The final point I’d like to make with you is the work of Monani over the last few years where he looked to see the requirements for SMN protein during development. He used the mouse model for spinal muscular atrophy, which he himself developed when he worked with Arthur Burghes, PhD, in Ohio State as a post-doctoral research scientist, and what they had developed was a very effective severe model for spinal muscular atrophy. Monani later showed that during maturation of the neuromuscular junction, where the nerve connects to the muscle, there is a high requirement for SMN, then after the neuromuscular junction is matured there is a significant in the reduction for the requirement of SMN protein to maintain the system. There’s a high requirement during maturation and a decreased low requirement for maintenance of the neuromuscular junctions. That now helps us understand that if you treat very early before the infant is symptomatic, then you can preserve the integrity of the neuromuscular junction and allow effective connections to be made between the nerve and the muscle. If you don’t treat early, you miss that window of opportunity when the neuromuscular junction is maturing, the system is being established, connections are being made, and therefore later on, since you missed that window of opportunity, if you treat then, you can’t recover the loss of function because the system has already been damaged irreversibly with the loss of the motor neurons.

All of this basic science now informs us as to what our findings mean when we examine these infants and young children during development. Early loss of tendon reflexes implies deafferentation of the motor neurons; the myopathic phenotype that we associate with spinal muscular atrophy tells us that there are certain cells that have the capacity to phosphorylate the activated P53 gene and as a result that triggers program cell death and those particular vulnerable motor neurons are the ones that send the impulses to the girdle muscles and the paraspinal muscles; and then finally we know that there’s a higher requirement for SMN during the maturation of the neuromuscular junction and there was a low requirement for SMN to maintain the system after it is full matured. There’s a very nice story from bench to bedside and here we are now in 2018 with an FDA approved drug for this disease, so it’s effective in the treatment of this disease, and this year the federal government has approved spinal muscular atrophy as the 35^th disease that can be added to the newborn panel to screen for a disease for which there is an effective treatment.

It’s a very exciting time and here we are 127 years after the original description by Werdnig in 1891.