11 – SYNGAP1 in the Developing Human Cortex

Event Time

November 6, 2020

Description

Here are our introductory comments:

Dr. Stephan Sanders Webinar 11/6/2020 at 12pm EST

We are very excited to continue the SRF webinar series. The goals of the series are:

  • getting you closer to the science 
  • making you aware of the research that is been done and the opportunities to participate
  • and empowering your communications with clinicians 

Our next webinar that will be taking place is Thursday December 10th at 12CST. This will be with Dr. James Holder of Baylor College of Medicine.

Our talk today with Dr. Sanders is titled “SynGAP1 in the Developing Human Cortex”. Dr. Sanders speaks to us today with experience in a number of different roles related to patients with developmental disabilities and autism spectrum disorder. He was first a pediatrician in his home country, the United Kingdom. In 2007, he and his wife moved to the United States where he signed on for a postdoctoral research position at Yale in Dr. Matthew State’s lab and completed his PhD work. He is currently an Assistant Professor at UCSF in the Department of Psychiatry.

Dr. Sanders’ work in genetics has been instrumental in developing our understanding of how de novo mutations are associated with autism spectrum disorder. Working with the Simons Simplex Collection, exome sequencing was used to develop early methods for identifying genes linked to autism. Further down the line, Sanders lab has prioritized characterizing and studying in depth genes which are strongly associated with ASD. Notably, his lab’s work on SCNA2 has combined genetic data with knowledge about the protein’s structure to understand both the gain and loss of function mutations and the differing phenotypes that they produce.

Sanders has continued to contribute to elucidating genetic links to autism by combining multiple cohorts including copy number variant data and exome data. In 2015 he published a paper identifying 71 risk loci for autism and SynGAP1 was included as a risk gene. His lab is broadening our understanding further by developing and applying new statistical methods to analyze non-coding regions of the genome to illuminate the pathways in gene regulation which may contribute to autism.

His pioneering work in genetics has benefitted many patient communities and we are appreciative for him sharing with us today about that as well as some of his specific knowledge on SynGAP1.

At the end of this presentation, you will have an opportunity to get your questions answered. We’d love to hear from you – please write your question in the chat.

For those of you just joining us, welcome to our talk today by Dr. Stephan Sanders entitled “SynGAP1 in the Developing Human Cortex”. You will be able to find a recorded version of this talk on the SRF website as well as on SRF’s YouTube Channel.

THIS IS FOR TRANSCRIPT ONLY:

:03Awesome glad to have you with us Stephan it’s  good to have you here I know it took us a while  

0:09to get you on the webinar series  but now that you are we’re very excited.   So the goals of our series are to get our families  closer to the science, make them aware of research  

0:19that’s being done and opportunities to participate,  and to empower communications with clinicians. The next  

0:25event that we’re going to be having is December  4th and that is our second annual Syngap Round   table we would love to have you all there  and you can of course sign up for it  

0:34on the web link on the SRF webpage so our talk  today with Dr. Sanders and it’s titled “SYNGAP1 in  

0:42the Developing Human Cortex”. Dr Sanders speaks to  us today with experience in a number of different  

0:48roles related to patients with developmental  disabilities and autism spectrum disorder   he was first a pediatrician in his home country  the United Kingdom in 2007 he and his wife moved  

0:57to the United States where he signed on for a  postdoctoral research position at Yale and Dr.  

1:02Matthew State’s lab and completed his PhD work  he is currently an assistant professor at UCSF  

1:08in the department of psychiatry Dr Sanders work  in genetics has been instrumental in developing  

1:14our understanding of how de novo mutations  are associated with autism spectrum disorder  

1:19working with the simon simplex collection exome  sequencing was used to develop early mass for  

1:25identifying linked to autism further down the  line Sanders lab has prioritized characterizing  

1:30and studying in-depth genes which are strongly  associated with ASD notably his lab’s work on  

1:36scna2 has combined genetic data with  knowledge about the protein structure  

1:41to understand both the gain and loss of function  mutations and the differing phenotypes that they   produce saunders has continued to contribute  to elucidating genetic links to autism  

1:51by combining multiple cohorts including copy  number variant data and exome data in 2015 he  

1:57published a paper identifying 71 risk low cipher  autism and SYNGAP1 was included as a risk gene.  

2:04His lab is broadening our understanding further  by developing and applying new statistical methods   to analyze non-coding regions of the genome  to illuminate the pathways in gene regulation  

2:13which may contribute to autism his pioneering work  in genetics has benefited many patient communities  

2:18and we are appreciative for him sharing with  us today about that as well as some of his   specific knowledge on SYNGAP1. At the end of the  presentation you’ll have an opportunity to get  

2:27your questions answered and we’d love to hear from  you. You can write your questions in the chat for   those of you joining us just now welcome to our  talk today by Dr Stephan Sanders entitled “SYNGAP1  

2:38in the developing human cortex”. You’ll be  able to find a recorded version of this talk   on the SRF website as well as on SRF youtube  channel and with that Dr Sanders i’d like to  

2:48pass things over to you wonderful thank you very  much for such a lovely and kind introduction  Overview of talk

2:53it’s a real honor to to have a chance to  talk with you and thank you to carol for matthews  

2:59for linking us up over the last few years i’ve  found working with family groups some of the most   rewarding and interesting work we’ve done and so  SYNGAP1 is a gene which we’re sort  

3:12of newly focusing on and so very very excited  to have a chance to talk to you so let’s start  

3:19off there we go i’m starting with the important  thing i have no conflicts of interests i  

3:24receive funding from from multiple private and  public entities all of which is focused towards  

3:30something to do with neurodevelopmental  delay and advancing health initiatives.  Exome sequencing identifies de novo mutations in protein coding genes

3:35so a quick overview of the of sort of how how  I got to be interested in SYNGAP1 and other  

3:43disorders this was really the killer experiment  exome sequencing which 10 years ago was very  

3:49new and very exciting but now amazingly is  quite commodified but the really important  

3:54experiment was the families so on the far left the  simon simplex collection collected both parents  

4:01a child affected with autism and critically  an unaffected sibling control and from these  

4:07individuals along with lots and lots of data  um we extracted DNA from that we take out  

4:14the coding regions which make proteins that  gets sequenced and from this we can identify  

4:20all the variants in the genes of the children but  also critically we can identify the variants which  

4:26are in the children but not in the parents and  these are called de novo mutations they are rare  

4:32each of us has fewer than or about maybe about one  of these in the coding region of our genome many  

4:38of which do nothing but a few of them have fairly  devastating effects on the genes they affect  

4:44and these demand mutations because they’re so  rare and because they’ve not been they’ve not gone   through natural selection they can have very very  dramatic effects on symptoms and what we found was  

4:54that there are more of these mutations in children  with autism than in children without autism  

5:00we now know that’s not because there’s an increase  in mutation rate it’s simply that because these  

5:05cause disorders if you look at individuals with  disorders you find more of these mutations it’s  

5:10an ascertainment issue and so then the question is  well which mutations are the ones causing autism  

5:16and which the ones which are just the sort of part  of being a human just the normal variation we have  

5:23and to answer that the what we did was to look  at the where these mutations fall within genes  In cases, these mutations accumulate in genes that contribute to risk of a disorder

5:30so here’s the region of chromosome 2. Each of  these dark blue lines represents the gene and  

5:36here are the mutations found in those genes in  the controls and you can see they’re fairly evenly  

5:42split there’s lots of controls there’s quite  a lot of dots but we see them just scattered   across the genes but when we look in the cases we  see that same scattering there’s some random ones  

5:52but we also see these mutations stacking up  and these stacking up is because these are  

5:57causing risk and therefore we see them in the  same gene again and again and again and then  

6:02a lot of the work over the last few years has  been asking statistically what counts as enough  

6:08of these mutations in the same gene to answer  the question that this gene causes autism itself  Analysis of 35,584 samples identifies 102 genes associated with ASD

6:16and the latest iteration about this was  published earlier this year is looking   at 35,000 exome samples and from this  we find 102 genes which cross the line.  

6:27Now normally when we talk about this we sort  of talk about how important the different   lines are and you know all of these genes are  important some have better evidence and others  

6:35for SYNGAP1 forget all that i mean it’s just one of the top three genes.  

6:40This is an extremely robust finding, it’s  been replicated, we’ve seen in other cohorts  

6:46we really really can be absolutely sure  that SYNGAP1 is associated with autism   and developmental delay and seizures. We’ve seen  it in all three of those and of course you as a  

6:56family group probably convincing of that but when  we’re writing grants being able to write the line  

7:02SYNGAP1 is one of the top x genes in autism  is really really important because we need  

7:07to think about which of these genes is going  to get the most funding and actually starting   off the ones at the top is a good idea because  you’re sure you’re looking at the right thing  

7:17there are more individuals affected and  they tend to have bigger effect sizes   so we’re continuing to do this gene discovery  efforts we’re going to double the sample size  

7:27probably hopefully we publish next year we’ll  see um and we and far more genes coming along  

7:33but SYNGAP1 remains one of the top genes  there and my other the other gene i’ve focused  

7:38a lot on has been SCN2A for the same  reason it’s top and recently SLC6A1 too.  

7:44All three of these genes interestingly share  the features of being autism and seizures   and I think that’s important from a therapeutics  point of view because seizures represent a very  

7:54strong end point which makes it easier  to understand if therapy is working.  

8:00So i want to talk today about some insights we’ve  gleaned firstly across all of these genes together  

8:07and then secondly looking in the human cortex  and seeing what we can learn about SYNGAP1 in  

8:12particular. So we’ll start off with the the  questions here so this is really guided   by a conversation we had 12 years ago so we so 12  years ago we ran the first version of this and we  

8:23and SCN2A was the gene which came out sorry  eight years ago and we ended up talking to an  

8:29expert in therapeutics and that person said well  that’s great it’s great you’ve got a gene target   all I need to know is what does it do when does  it do it and how much do you need to rescue  

8:40for us to treat it and we looked at this and  thought well you know don’t know don’t know and  

8:46don’t know so we’ve then spent a few years trying  to answer these questions about where when what  

8:51and how by looking at all of these genes  together on the logic that they are all are  

8:56causing similar symptoms therefore that the point  where they converge is likely to be important.  

9:03So let’s start off with functions. So if I take those  102 genes I can ask maybe the simplest question  

9:08what do they do? What’s their role in the human  genome? Lots and lots of these genes are involved  ASD genes fall into two major functional groups: gene expression regulation and neuronal communication

9:13in the regulation of other genes and so that’s  all the ones in purple here. 58 genes. Basically  

9:20their job is to regulate the expression of  other genes. Then the other big group is neuronal  

9:27communications. These are genes which are important  in the synapse and neurons talking to each  

9:34other and firing and this of course is where  SYNGAP1 falls. What we do not know is what  

9:42the interaction between these two groups is and if  I could go 50 years into the future and ask a   single question or maybe two questions the first  one would be what is the interaction of these two  

9:52groups? Because I believe that would give us the  answer at a very fundamental level of what autism  

9:57and developmental delay are in the brain and the  second question I’d ask is: when can you reverse it?Many neuronal communication genes associated with ASD are specific to the brain

10:06And so we can then go on and start asking  other questions about where these genes   fall. So we can look in adult data from from lots  and lots and lots of tissues across the body (this  

10:17is a resource called GTex) and simply ask where  are these genes expressed? What we find is that  

10:23overall, autism genes are found in the brain.  Now that might not sound that revolutionary  

10:29but important to know that that’s  the case (we can always think of metabolic causes).  

10:35Secondly though, we find that if you look at the  genes which are involved in neural communication  

10:41we find that they are very, not only expressed in  the gene but specific to the brain. So we don’t   find- we generally don’t find them in other  tissues and secondly it’s the cortex rather than  

10:51the cerebellum which seems to have the strongest  expression. So that suggests, and there’s other data I’m  

10:57not showing here, but that suggests the cortex  is a very very important part in pathology.

11:06If we look at SYNGAP1, which is a neuronal  communication gene, its expression pattern  SYNGAP1 is expressed throughout the body, including. brain and the pituitary gland

11:12is a little bit surprising. On the good side, as we  would expect, it is expressed in the brain. What is  

11:18interesting to me is it’s also expressed in many,  many, many other tissues. I don’t particularly  

11:26have a sort of inference from what that  means. You could maybe think that it means that  

11:34non-neuronal or non-brain phenotypes are  maybe a bit more common in SYNGAP1 than others  

11:40but i i don’t know that’s the case and i’d  be very interested to hear from this group   if that is something which you think might  be the case the other big surprise to me is  

11:49how highly it is expressed in the pituitary  gland and so the pituitary is involved in the  

11:56excretion of many many other  hormones and many regulatory factors   interesting that it’s so highly  expressed there that is an unusual pattern  

12:05I don’t particularly again know what  that means in terms of human biology

12:12so we can then go for the next step forward  so we’re fairly sure it’s the brain that’s not a   surprise but it is elsewhere. It seems likely  the cortex is involved but then we need to get  

12:22down to the question of cell type. So within the  cortex there are many many many different types   of cells and this is a new type of  data which is coming out where we can look at  

12:32the gene expression in every single cell rather  than sort of soup made from the from the tissue.  

12:39And when we look at the cells in the cortex we can  divide them into sort of two main groups firstly   there’s the neurons which fire and send signals  to each other and then the supporting cells called  

12:50the glia which are also very important. Within  the neurons we can divide them in this data  

12:56set which is from the human developing cortex  into the developing neurons in the (top left),  

13:02mature excitatory neurons (bottom), mature inhibitory  neurons (top right) and then the glia which are in  

13:08these different circles and we can then take our  autism data, our autism gene lists, and ask well,  

13:15which of these cell types are enriched  for the genes which we’re seeing expressed and  

13:21we see a very clear result there that it’s the  both the inhibitory and the excitatory neurons  

13:28can seem to be enriched for these signals so  we believe from this we can say not only is  

13:33it probably cortex but maybe other regions but  it’s probably excitatory and inhibitory neurons.  

13:40And some of the genes are very specific  to one of these. So for example SLC6A1 is  

13:46only in inhibitor neurons where SATB2 and TBR1 are very much excitatory  

13:52So what about SYNGAP1? Well so here on the  left is that same map. On the right is the is  

13:59what we see when we ask the expression of SYNGAP1.  So blue represents a cell where we’re not seeing  

14:04it expressed and brown represents a cell where we  are seeing syndicate when expressed and the answer  

14:10is that it’s not particularly clear it looks like  there’s not very many cells expressing SYNGAP1  

14:16in the fetal brain maybe there’s some signal  up in radial glia up here and then maybe some  

14:22in the excitatory neurons here but it’s fairly  scattered and it’s really not a clear picture  

14:29we can do the same experiment in um the brain in  childhood and this is data set for a monoligstine  

14:35at ucsf and here there’s far more cells it’s a the  technology is advancing rapidly so there’s many  

14:42many more cells on this plot and now we’re looking  at cells the neurons and glia later in development  

14:49and what we can see is there’s a very strong  enrichment in the brown for SYNGAP1 now  

14:56all of these have sort of not particularly helpful  labels but what this means is that it’s mostly  

15:02excitatory neurons that’s the l23 l56 l4 and l56  here as well but also some in inhibitory neurons  

15:11there’s maybe a little bit of signal in nasal  stripes as well but it seems that here SYNGAP1  

15:17fits that neuronal signature mostly excitatory  but some inhibitory as well so this is going to  

15:23be important if if we get to a stage of having  something to use as a therapy asking a question  

15:30where do you give it what are you trying to  target what cell type you’re trying to get it into

15:36so we’ve been working on some on a data set which  is trying to extend these resources and so this is  

15:44as gene expression data from the developing human  cortex and while the samples we’ve seen before  

15:51have involved maybe 10 maybe up to 40 individuals  this is a data set with 176 brain samples in it  

15:58giving us a much much clearer view  of where and when genes are expressed  

16:03the samples are arrayed across a wide  wide range of developmental time points  ASD genes are highly enriched in the brain, especially the developing human cortex

16:10they tend to peak in regions  partly due to tissue availability

16:16when we take the 102 autism genes what we  find is there are basically two patterns  

16:23the first pattern is that of the gene  expression regulation genes these are the genes   which regulate other genes and what we find  with those is that are expressed very very  

16:32highly in the developing human brain so that’s  the purple line here and then they go down  

16:39um after birth and into childhood in  contrast the neuronal communication genes  

16:46go rise during fetal development reaching a peak  around birth and then basically plateau after that  

16:56and so one of the questions we keep asking  ourselves when we stare at this is well which of   these is when autism and developmental delay take  place is it early in fetal development is it where  

17:08they cross over is it later when the neuronal  communication comes up because we infer that that  

17:14question of causation will have some impact on the  answer about how how much you can reverse symptoms  

17:22later in life and so this is why that’s the second  question i would want to ask in the future when  

17:27can we reverse symptoms now i must say in this  last week i think we’ve had some encouraging news  

17:35on this front in Angelman syndrome where in five  individuals treated between the ages of two and  

17:41fif five and fifteen we’ve seen some improvement  in their symptoms with a new therapy it caused  

17:50some side effects it caused some muscle weakness  but the actual neural symptoms seem to improve   and to me the only experiment which is ever  going to really answer this question of when  

18:01can we reverse symptoms is by doing this in humans  maybe in non-human primates but i think the human  

18:07experiment is is probably the one which is going  to answer this question for us and we simply don’t  

18:12know the answer so here’s the expression  of SYNGAP1 so we’re on the same scale here  SYNGAP1 has an unusual expression trajectory across development; it falls during childhood

18:20so each dot represents a brain sample the blue  line is sort of the best fit line across them the  

18:28black line in the middle the vertical one that’s  birth and on the left we have fetal development   on the right we have childhood and early adulthood  the higher the dot is the more it is expressed in  

18:40the human cortex at that time and so SYNGAP1 also  has a slightly unusual pattern here so it goes  

18:48like most of the neuronal communication genes it  goes up during fetal development reaching a peak  

18:54somewhere around here but it has much much more  of a fall later in childhood now again i can’t  

19:02draw that the inference i want to what does  this mean about when therapy can be given   how much is reversed or what impact that has in  symptoms um i would be interested to hear any um  

19:12experiences of what the pattern is with seizures  with this fool does SYNGAP1 become less important  

19:18that in later childhood that that’s i think gonna  be an important question to answer i’m very very  

19:25keen on seeing multiple modalities of data telling  me the same thing i generally feel if you see it  

19:30once it’s interesting if you see it twice it  might be true um over here is some data from  

19:37done in the mouse brain and so p4 is day four  after birth p14 is day 14 after birth and ad is  

19:46adult and this is looking at SYNGAP1 expression  over time and what we find is that there’s a peak  

19:52around p14 and this pattern here really links up  quite nicely what we’re seeing in the human cortex  

19:59and this is good because it suggests  that the mouse gene expression profile   matches that of the human and it also looks more  likely that this pattern we’re seeing is real

20:10so let me move on to the next question what is  the what are you trying to do what would fixing  

20:17SYNGAP1 look like now this i expect is a topic  you’re very familiar with but let me give my  

20:25my two cents on this i’d be interested to hear  how that tallies with with your own experience  

20:31so the first question to ask is does it  increase does the mutation increase or decrease  

20:37the function of the gene so a simple way of asking  this is to plot on the x-axis that they are going  

20:44along the bottom along the bottom mutations which  are protein truncating variant which mean that  

20:50you’re missing one copy of the gene versus the  number of mutations which are missense which  

20:56means you’re altering the function of the gene now  those mis-sense function mutations can increase or  

21:03decrease function whereas the protein truncating  variance can pretty much only decrease function  

21:11and so when we plot it like this we can in a  very simplistic way say that those which go  

21:16up here are gainer function and those  which go to the right are loss of function  

21:22and that’s where Syngap falls with a fairly clear  result maybe not quite as clear as arabic 1b but  

21:28still it looks very much like loss of function is  the predominant effect and that’s a really really  

21:35important thing to know in scn2a we found a mixed  result the gain of function causes early seizures  

21:41the loss of function causes autism developmental  delay with later onset seizures whereas in gap one  

21:47i think it’s a it’s a clearer picture which  from a therapeutic point of view is again a   really really important point forward and so that  means that a successful therapy would presumably  Increasing expression or function should improve symptoms in SYNGAP1 loss-of-function mutations

21:59increase the expression of SYNGAP1 and broadly  i see three approaches towards doing this  

22:09the simplest one is by taking a medication  just like taking ibuprofen or acetaminophen but  

22:15something which targets SYNGAP1 with the view to  potentiating its function now this is not my area  

22:22of expertise i’m i’m this is the sort of realm  of um biochemists chemists and drug companies  

22:29performing screens to try and identify these small  molecules it looks encouraging to me that SYNGAP1  

22:36is maybe more unique than some of the other  genes which we’re looking at which might make   that more of a successful strategy but really not  my area of strength in the middle we have a class  

22:46of compounds called antisense oligonucleotides  now these have been around for about 30 years  

22:51but they’ve been developing and improving over  that time and in the last two three years we  

22:58have seen some very very very dramatic results  from these as a class of drugs so particularly  

23:04in spinal muscular atrophy in baton syndrome and  i think now we can also add angelmans to that list  

23:11we’ve seen the ability of these to get into  bits of the brain or bits the central nervous   system to do what they’re meant to do and lead to  symptom improvement and in spinal muscular atrophy  

23:22these have had a life-saving effect  really really dramatic results

23:28they are they are quite easy to design to decrease  the expression of a gene those are called the  

23:34gapmas but of course that’s the opposite of what  we’re trying to do um it is much much harder to  

23:41make them increase expression it is possible in  some genes so there’s been some very interesting  

23:47express um data from from scn-1a dravet syndrome  where with clever targeting of asos they’ve been  

23:54able to increase the expression of SYNGAP1 in  the brain this really relies on having a gene  

24:01which has um versions of it which are expressed  which don’t make protein and i must say it’s not  

24:08clear to me that syndicate one is one of those  but that that might be an area for um exploration  

24:15the other thing you can do is potentially skip  exons which contain the protein truncating variant  

24:21now this relies on on two things firstly the axon  needs to be non-essential so it needs to be a bit  

24:28of the gene which isn’t for example the ras domain  and secondly it needs to be in frame so if you if  

24:36you skip it it needs to make sure that it doesn’t  change all of the exons which follow it otherwise  

24:41it will still be back to the protein truncating  variant so that was so as a therapeutic strategy  

24:48this would only ever treat a tiny minority  of cases but it might be a comparatively  

24:55easy therapy to develop because there’s  precedent because it’s reversible and  

25:01because some of the dynamics of giving these  in the brain have already been worked out   and on the right we have um sort of proper  gene therapy so this this really involves  

25:12some form of delivery agent for example a virus  going into the brain delivering a package which  

25:18aims to increase SYNGAP1 it’s ration i mean in  sin gap one i see three approaches towards this  

25:26the first one would be simply putting the  SYNGAP1. mrna inside the package so there  

25:32is more of it expressed the trouble there is the  g needs to be less than 4700 base pairs because  

25:38of the physical size of the aav and the long iso  form of SYNGAP1 is about six thousand so so this  

25:46so for the long ice form of SYNGAP1 this might  not be possible for the short it probably is  

25:53even it’s a very simplistic notion it’s it’s  like sort of just increasing it in a scatter shot  

26:01manner hoping that that replaces a sophisticated  mechanism and it’s even if you could just put  

26:06it in there it’s unclear if that would be the  right thing to do but that would be an avenue   forward there’s a new concept called crispr a um  which a colleague of mine at ucsf has been at the  

26:19forefront of developing in um in vivo i’m going  to give some more details on that next few slides  

26:24and then i’m sure you’ve heard of crispr  which of course recently won a nobel prize   which would involve editing the defect the trouble  here is that that editing would be different for  

26:34almost every single patient and so from  an fda point of view that involves you   know at the moment trying to create 150  drugs rather than just one and that that  

26:43seems while it might be possible it seems very  very difficult to imagine that um taking place  CRISPRa can activate the remaining allele to restore function in haploinsufficient disorders

26:51so crispr a the nice the nice thing about  crispr a is it is possible to imagine  

26:57a single therapy which would work for most  individuals with with a syngap1 mutation  

27:04in particular individuals with a splice site  or a protein truncating variant mutation  

27:10missense is less clear but even there it may have  some benefit and so while crispr cuts crispr a  

27:18activates and the way it does that let’s take  here our um hypothetical haploid insufficient  

27:24gene so that’s a SYNGAP1 would be an example of  this a gene where missing one copy is a problem

27:32we put in an sgrna this is a little um  small amount of of rna very very small  

27:38which targets a region upstream of the  transcription start site where expression  

27:43begins so what we where this gets designed  is an enhancer region or a promoter region  

27:50you then put in the cas9 so this is the the  protein which does the work in in crispr  

27:57but it’s a dead cas9 that’s  what the d is for it doesn’t cut   all it does is find this this rna and then go and  bind to it and links to it is this vp64 or there  

28:10are some other proteins which can be do and what  this does is it sends a message to the gene to   start expressing and so the way this works is that  you go to the normal the unmutated copy of SYNGAP1  

28:23crispr a goes in there increases the expression  of it and then that hopefully rescues the deficit  

28:32so here’s an example of this working in mice  and so here the gene being targeted is sim one  

28:39sim one causes developmental delay mild  development delay but it also causes really quite  

28:45dramatic obesity so these two mice here the one on  the left it has is missing one copy of SYNGAP1 and  

28:52has eaten itself into this very very plump state  a single injection was given of this um crispr a  

29:01into the brain um after birth and then seven  weeks later the mice were photographed and you can  

29:09see here that this um this single injection has  resulted in the obesity phenotype going away and  

29:15this mouse is a normal size and so this in vivo in  a mouse is a rescue of a haploid condition with a  

29:24single injection and what i think is particularly  exciting about crispr a is that it seems to  

29:32certainly in this in this gene it seems to inherit  some of the important regulatory features so for  

29:39example sim1 is particularly expressed in the  hypothalamus and that was still true even in the  

29:45crispr a treated one it didn’t for example start  to get expressed in other tissues and so while  

29:51giving while simply replacing the gene with aav is  very simplistic and you know you might have it in   the wrong place at the wrong time in the wrong  circumstances crispr a may capture some of that  

30:02nuance however there’s no precedent yet of giving  crispr a into a human let alone into a human brain  

30:10and so the path there towards regulatory approval  is a longer one and a more complicated one  

30:15that said this looks like it has the potential  to be a a successful therapy in these conditions  

30:23and we’ve been working at ucsf on this in scn2a  and have seen similarly promising results in mice

30:32so i’d like then to move on to the last part which  is talking about isoforms and so this is an area  

30:38where we think we’ve been able to leverage some  of the um the data we’ve developed to try and get  

30:44some insight into the isoforms in instant gap one  and this has been a new area of research for me  

30:50and i think this is an important topic in terms  of therapeutics for this group and it may be  

30:56that you’ve already had talks on this topic  already i’d just like to know that alicia and   lindsay have been two trainees in my lab who’ve  really really helped with the analysis of this  SYNGAP1 has long and short isoforms; few patient mutations are seen in the axons at the start or end

31:09so this is a paper i’m sure which is familiar to  this group um of looking at the known mutations  

31:16in SYNGAP1 this is from 2016 and lining them  up with the with the SYNGAP1 gene and asking  

31:22where do they fall and so this paper i  think makes two important observations  

31:29the first one is that there are multiple isoforms  of SYNGAP1 now that that was known already  

31:35but the important thing is here is  displaying mutations alongside those   let’s we’ll come back to this a little bit more  detail in a couple of slides but here you can  

31:44easily appreciate there are long isoforms which  start here and then there’s a short isoform called  

31:51b which starts here and a very short isoform  called c which starts there now all of these  

31:58are the same gene all of these are SYNGAP1 but  SYNGAP1 the gene has many many functional units  

32:05which can be put in different places and that’s  what we call an isoform so when you look at the   protein that gets made it looks like it’s optional  whether these three exons here are included or not  

32:17similarly at the other end there are exons over  here which may or may not be included now when i  

32:23look at the mutations which are known so far there  are a few things which um which i noticed straight  

32:30away the first one is as we’ve seen already most  of them are protein truncating variants there are  

32:36surprisingly few missense variants but the second  thing which caught my eye was that if you look at  

32:44the ends of it there are very very few mutations  they all seem to cluster within the middle  

32:50now for a missense mutation that wouldn’t surprise  me at all it looks like all the really important  

32:55functional domains tend to be in the middle and  the missense variant needs to disrupt a functional  

33:00domain to have an impact but a protein truncating  variant which is the majority of these doesn’t  

33:07it doesn’t really matter where they fall they’re  going to disrupt the gene so the fact that we’re  

33:13not seeing protein truncation variants are not  many many protein truncating variants at the start  

33:18and the end is interesting and suggests that there  might be some insights to be gleaned from this

33:27i said it’s always nice to see replication and  so in 2019 there was this updated paper so the  

33:33the dots at the bottom represent the mutations  have been previously published this paper adds  

33:39a stack of new mutations which i’m sure represent  many of the individuals in the family group above  

33:45the gene are the protein truncating variants below  the gene are the missense and again this pattern  

33:51that most of them are protein truncating variants  and also that there are few in the first three  

33:57maybe first four and the last four exons even  for those which are there this paper is trying to  

34:04distinguish them by phenotype which is what the  color is showing and green is being the milder   phenotypes even those individuals who do have  mutations in these first three or last four they  

34:14seem to have milder phenotypes on average than the  other individuals and so it looks on the 20 000  

34:22foot view to me it looks like there’s something  funny going on at the start and end of this gene

34:29so digging back into literature um back to 2001 we  can see some work done in rats which starts to ask  

34:36the question or start to describe these different  patterns by which the syngap1 gene is made  

34:43to then make a protein and what this paper does  is to identify and then name these differences  

34:51at the start and the end now this was done in 2001  graphics have progressed a lot since then and so  

34:59this paper is is pretty hard to uh to interpret  but there’s a very very nice figure made in 2018  

35:06which really captures the um the main points of  this and that is that you’ve got this bit in the   middle which you need like this is present in  all the different isoforms if you’ve not got  

35:16this you’re not really um looking at a SYNGAP1 proper protein but at the start you’ve  

35:21got three options you’ve got a long a short  and a very short which are called a b and c  

35:28and at the end you’ve got a bit more variety  but which they’ve sort of condensed these four  

35:34colors here and there’s a very short one labeled  beta a slightly longer one gamma slightly longer  

35:40than that alpha one and the longest is alpha  two and so you can have combinations of these  

35:46three at the start and four at the end and  that leads you to make a functional protein

35:53so this obviously adds a layer of  complication to an otherwise simple story but  

35:59i think it’s the level of complication that’s  important because there’s this other paper   in 2012 which asks the question well what what do  the different isoforms do to neurons in the brain  SYNGAP1 isoforms have opposing effects on synaptic strength; which effect needs to be increased?

36:12and this is a summary which they make of this  but the the quick summary of it is is that the  

36:20long isoform and the alpha one tend to decrease  synapse strength whereas the short isoform and  

36:28the alpha 2 increase synapse strength so  while all of these are loss of functions  

36:35we have these two different effects at the level  of isoforms which are opposing and so while the  

36:43question to what do i need to do is the gene  i need to increase it that’s simple but the   question is to which isoform or which function  do i need to increase that seems more complicated  

36:55does a therapy increase synaptic strength or  does it decrease it or does it need to capture  

37:01the complexity of doing both at the right time  and again i’m afraid i don’t have a simple answer   to this but I can show you some data which maybe  illuminates what happens in the normal development

37:14when we look in humans we find that the story is  

37:19similar but maybe a little bit more complicated.  Each one of these represents an isoform so if I  

37:27just skip back we go to this paper here here we’ve  got five different isoforms but in humans it looks  

37:34like there are 11 different isoforms which make  a functional protein now i think we can probably  

37:39discount this top one it doesn’t seem to be  expressed very much doesn’t seem to do very much  

37:45but these remaining 10 are important and the  question is when and where do we see these being  

37:52expressed so let’s just go through and familiarize  familiarize ourselves with how this maps to the  

37:58old papers and this is often a challenge in  genetics is trying to use the same terminology  

38:05as was used 20 years ago to actually be able  to understand the insights which have followed   but i’m fairly confident these are mapped  the right way so if you remember we’ve got  

38:15the short and the long isoform so that’s  at the start here we’ve got the long the “A”  

38:20the short the “B” and then we’ve got this very  short one called “C” now in the human genome  

38:26we don’t see any evidence for that “C” isoform in  the list of transcripts i think we need to go  

38:33back to the raw data and just check that that  is true but it looks like we only have an a   and b so a long and a short and i’ve colored here  the long ones in red and the short ones in blue

38:48let’s start with this top one here this  is an unusual isoform in that it start  

38:55it stops early so here instead of  going all the way on there and making   some of these important parts like the RAS it  stops early now i would be very interested to know  

39:07what happens to this transcript i can tell you  in the human brain it’s not expressed very much   but i would also if i had to guess this  isoform gets destroyed and doesn’t make protein  

39:19i would be very interested to know if that’s true  if you could if that is true it might be possible  

39:24to design an aso here which which forces it to  then keep expressing the rest of the exons that  

39:32might increase expression with an ASO this this  might represent what’s called the poisoned exon  

39:37but i think we need to do an experiment to see  if that’s the case so this might be an important  

39:42isoform for therapeutics we then have these two  short isoforms and then these multiple long ones  

39:50and they differ by the end where we’ve got  these different ending bits here and remember  

39:56the blue ones the short ones increase synapse  strength the red ones decrease synapse strength  

40:03so the first thing we can do is we can look at  the exons which are in common across the long  

40:09and the short and ask what does their what’s their  expression pattern look like in the human cortex  

40:16we go back to our brain of our collection 176  samples and this time instead of showing the   gene in its totality we can show each  exon in those bits which are in common  

40:28and what we see here encouragingly is  they all have basically the same pattern   they start it’s expressed  highly throughout development  

40:36starts here goes up during mid-fetal development  peaks at about six months and then goes down  

40:41and this as i’ve pointed out before fits  with what we’ve seen in the mouse.

40:48We can then ask the same question for the  exons which are unique to the long isoform  

40:55in red or unique to the short isoform in blue

41:01the red ones interestingly we see the same pattern  but it’s muted it’s fairly flat across development  

41:08there’s maybe a sort of 50% increase over  development but it’s fairly flat and here the that  

41:14same mcmahon paper seems a slight increase but  it’s not dramatic in contrast the short ice form  

41:23has a much much more dramatic effect here there’s  like an eightfold difference between early fetal  

41:29childhood and late and early adulthood and this  again fits with what’s seen in the mouse there’s  

41:34a more dramatic difference seen in the magnus  cortex in a time period which seems to be the same  

41:42so we can plot these two together here’s in red  this is this the long isoform and here in blue  

41:48is the short one and so and then on the right we  can look at the ratio of the two and we can see  

41:56the ratio differs over development and so this  then raises the question to me well which one of  

42:03these is important now of course the thing we know  is when diagnosis are made and that tends to be  

42:09fairly early on due to  seizures or developmental delay   and so it’s tempting to look at this short  isoform and ask the question is this the one  

42:19which is not present enough  to go and lead to symptoms because if so  

42:29this could potentially be good news because that  would suggest that what needs to be done   is strengthening the synapse and secondly the  short isoform is smaller and could potentially  

42:39be delivered by AAV as gene replacement. Now  both those statements make a lot of assumptions  

42:45which I have no data to back them up but I  think it raises important hypotheses to be tested

42:53and so and this is just putting that back  to this pattern here the a isoform decreases  

42:59synaptic strength but remains flat the b  isoform increases synaptic strength and  

43:05has this dramatic surge during fetal development  the other important observation here of course is  

43:11that the longer isoform is is throughout  developments expressed at a higher levelSummary

43:18and so in summary the key things we’ve talked  about here SYNGAP1 is one of the leading causes  

43:24of autism, developmental delay, and epileptic  encephalopathy. These three conditions share  

43:30a great deal of overlap I think the strength lies  in thinking together but when writing a grant to  

43:35try and secure funding you can make a very very  bold statement about the importance of SYNGAP1  

43:41in terms of the number of people affected the  effect sizes and how critical it is as being a  

43:47linchpin of these disorders and that really I  think puts its researchers in this field in a very  

43:52strong position when we look at the human cortical  data we see evidence that cortex is important and  

43:59we see that it particularly is excitatory  and inhibitory neurons which are involved   with a slightly higher expression than excitatory  neurons I suspect it is something to do with the  

44:08communication between these two neuron types is  one of the underlying important components of this  

44:15mutations are predominantly loss of  function this is really important in other   in other disorders there’s a lot of work to be  done to try and puzzle this out in sin gap one  

44:25i think you’ve got a fairly strong evidence that  loss of function is the predominant effect you  

44:30might need to be a little bit cautious of that  assumption with some of the mismutations but   by and large most individuals loss  of function seems to be the effect  

44:38which means that a successful therapy  would increase SYNGAP1 action or levels

44:45but then there’s this added complication of these  long and short isoforms which different function  

44:51the long isoform is more abundant but the shorter  isoform undergoes this surge in mid-fetal gesta  

44:57mid-fetal development to six months after  birth which maybe coincides with the symptoms  

45:04and disorder onset encouragingly we see similar  trajectories in the human cortex and the mouse  

45:12there’s always this question about what  model should be used should we be using   cell lines organoids non-human primates um or mice  that differs for what you’re looking at but for  

45:22the for the point of view of gene expression  the mouse seems like a pretty good model  

45:28and i think this raises a really important  question about future therapeutics   if crispr a was to be developed for SYNGAP1 where  would you target it it needs to hit to go to a  

45:39promoter and enhancer and those promoted enhancers  differ between the long and the short isoform  

45:45which one needs to be targeted and i think this  is a research question which needs to be answered   if i had to guess based on this data  i would probably go with the short  

45:54but that might not be the right answer i think  that needs to be answered and then secondly i’ve  

46:00shown you data for the long and short iso form  i’ve not shown you answers for these um different  

46:06end parts of the gene that’s a little bit more  complicated it’s not easy to treat the data i’ve  

46:12got in the same way to answer that question and  so me and alicia and lindsay need to go and sit  

46:18down and get a bit more granular with this data  and see if we can get similar answers for these  

46:24end ones which also have functional impacts it  might be that through exxon skipping you could  

46:30nudge it towards one of these rather than the  others which again might be a therapeutic strategy

46:37and i’d like to end by just thanking all  the wonderful people i’m involved in this   who i work with of course my lab particularly uh  lindsay and alicia who’ve been working on this  

46:47um kevin bender has been integral to scn2a work  and electrophysiologists at ucsf and nadav has  

46:54been um doing amazing work on chris bray and with  that i would like to turn it over to questions

47:05that was quite a talk i’m gonna make that  required reading when people call me up   to ask about Syngap i’ll be like go watch  this and then if you still have questions  

47:14i probably can’t answer them but um thank  you so much dr sanders that was lovely and  Questions & Answers

47:23obviously somebody um lindsay alicia and  yourself put in a good bit of work so beautiful  

47:29slides thank you um i’ve invited dr schleck to  talk he’s a he’s an md and he’s on our board  

47:36and um he’s he’s remarkably well read on the Syngap science so i’m gonna pass it to him because  

47:43uh i don’t even want to try to manage the q a on  this one so Hans if you if you’re not muted do you  

47:48want to throw in some questions and i’m probably  gonna invite J.R. who’s also a parent with some   strong science background to ask questions but  go ahead yeah can you guys hear me okay yeah oh  

48:00super so let me start by just saying thank you dr  Sanders um it’s wonderful to have a brilliant mind  

48:07interested in our kids and this protein because it  is quite fascinating as a protein so just a basic  

48:14question i’m just struck by how multi-functional  the protein is in terms of the various domains  

48:21just how much it’s responsible for sort of encoded  within the peptide are there is there an analogy  

48:29to other brain proteins that we can make where  we say that they have such differential effects  

48:36so many isoforms something a parallel that that we  can look at for comparison yeah so having multiple  

48:45isoforms 10 is the rule rather than the exception  for these genes um there’s only a few genes  

48:53which are so simple that this is not an issue so  so generally transporters and channels tend to be  

48:59a little bit simpler simply because if you don’t  if you don’t wrap it around and make a hole in the  

49:05membrane it’s not a channel so the channels and  transporters are generally simpler but even there  

49:10we’ve been doing some similar work on scn2a  and there’s some fairly dramatic findings um  

49:16most of the genes involved have multiple isoforms  which take place i think syngap is maybe a little  

49:24bit unusual already advanced in that someone has  actually gone and done this work and answered the  

49:31question of what or at least the first step  in the question of what do these isoforms do   in most genes in many of the genes we’re starting  off with what does the gene do um let alone  

49:43what do the different isoforms do for that role  so sitting at one i’d say you’re you’re ahead   of the curve but that’s just that adds complexity  to the problem but you’re far from alone in this  

49:53this this is this is how you build something  as complex as a human brain out of 20 000 units  

50:00um it’s because those 20 000 different genes  actually encode 200 000 different um functional  

50:08bits it’s a bit like lego you know you you build  the first bit of the kit and that’s great you’ve   built your little spaceship but then you look  on the back and oh i can make that spaceship  

50:17into a dinosaur um or at least or maybe modify  my spaceship to have two boosters on the back  

50:23that’s where the functionality comes from in  SYNGAP1 there’s one other little bit to this  

50:28and that seems to be that neuronal activity also  seems to regulate some of those expression changes  

50:34and so it may be that part of the way that it  it sort of acts as a controlling mechanism for  

50:41learning and as the neuron fires that maybe  damps down or increases synaptic strength  

50:48across these genes i i like to look at patterns  across them i think all of these genes need  

50:54to be seen as as a as a group together you know  basically even though the symptoms differ slightly  

51:00there’s far more in common between them than there  is different and it’s that point where they all do   the same thing which is going to be the point  where we can develop a therapy for everyone  

51:10and to me learning seems to be that place the  the ability to make a synapse stronger or weaker  

51:15and SYNGAP1 is absolutely  at the heart of that process   and and in writing proposals and grants for  this gene trying to emphasize that that you are  

51:25absolutely the center of biology i think is an  important point to be made to funding agencies

51:32you’ve been willing to sort of express insight or  you know sort of stick your neck out with a guess  

51:38and so forth have you looked at this anti-sense  this naturally occurring as1 of SynGAP and are  

51:47there insights to be gained there it targets  the region around exxon 15 16 17. absolutely so  

51:55that that gene is in our data i i have i have  seen it and been interested but i’ve not yet  

52:00dug into it and so that’s that’s on my my list of  things to pull out the other one which caught my  

52:06eye and looking through this there’s a micro rna  called mir-5004 it’s in the introns um i think  

52:15it’s like seven between 17 and 18 something like  that and its expression pattern mirrors SYNGAP1  

52:21so micrornas are regulatory let me give i’ve  missed a big picture point here every single  

52:28gene therapy which has made it into the  clinic so far has targeted targeted the   regulatory machinery of a gene in the brain so  that’s true for Spinal Muscular Atrophy for Angelman  

52:39i think it’s true also for some of the work  and to share muscular dystrophy so i think   understanding how the brain and  the cells control isoforms and expression levels  

52:50i suspect that is the route forward for therapy  and so micro RNAs and antisenses i think those are  

52:57really really important and so you’re absolutely  right to highlight that i’m going to i will work   out the expression profile of the antisense it  wasn’t in the first cut of the data but it will  

53:06make sure it goes in the second i also want to  figure out the c isoform and see if that exists in   humans and but that micro RNA, working at what that  targets i think would be an important priority.  

53:17And it seems like micro RNAs are getting much more  attention and potentially also used as biomarkers  

53:22because they can be detected in plasma. Absolutely  absolutely um that it would be very it’d be very  

53:29exciting if that panned out i i suspect seeing as  the seeing as the loss of SYNGAP1 occurs at the  

53:38RNA level it’d be a really interesting question if  that micro RNA tracks the the nonsense mediated decay.  

53:46In a similar vein the double knockout mice  

53:55they fall out late in gestation they universally  expire within five six seven days of being born  

54:03any insights to be gained there i mean it to  me at least it seems almost like gestation  

54:10maternal circulation is providing potentially  enzyme replacement or or SynGAP replacement  

54:18so this that is a pattern we’ve seen across most  autism genes so the vast majority you lose one  

54:24copy and you develop autism on developmental delay  you lose two copies and it’s fatal um to the best  

54:31of my knowledge we have never seen a double mutant  in a human in any of these genes suggesting that  

54:38and humans are generally a bit more sensitive to  these things than mice um so that’s but in terms  

54:44of um the question as to why they die that’s  a really interesting one i must say i’d not um  

54:55i’d not particularly i’d not particularly  questioned it given that they do such um  

55:01such challenging such integral things to neurons  the fact that it’s not compatible with life i  

55:07find i find not that surprising you’re right  if there was something being passed over from   maternal circulation that could be interesting um  but we see the same thing with scn one earth 2a  

55:19where it’s very very hard to imagine the channel  crossing the placenta crossing the brain barrier  

55:24getting into the cells so i i think it’s unlikely  but it’s a really interesting thing to check  

55:30yeah i mean at least in the mecp2 literature in  the past year and a half i think hollis klein  

55:38out in california showed you could provide  mecp enriched exosomes and correct circuit  

55:48sort of deficiencies and and pathology um so it’s  just tantalizing to think that there could be  

55:55a modality there to help kids i think  that’s an underlying question of how  

56:00simple can the therapy be is it enough to  just scatter shot increase expression in a   few cells or do you need to increase expression  by 50 by 100 in the right in every single cell  

56:13and this is one of those big unknowns you know  what how sophisticated does a therapy need to be  

56:20one of the things we might tackle in in the  near future is a humanized syn gap mouse  

56:27do you think that’s worthwhile i  mean just off the top of your head

56:33it does the as1 the naturally occurring  antisense does not occur in mice so there’s  

56:40potential insights to be gleaned there serves  as a a little bit closer to the human as a model  

56:47yeah that is a really interesting question um  i’m encouraged that the expression profile is  

56:54similar in humans than mice i i think the honest  answer to this is i’ve not looked at the homology  

57:00enough to know how different it is um there’s a  dozen amino acid differences at least from my wifi  

57:09yeah and are they in functional domains scattered  it seems from beginning to end yeah does the  

57:20the anti-sense i think that is that is  really an interesting avenue forward  

57:26so um the absence of that in the mouse is a shame  i think does the rat have the antisense in it

57:34i’m not certain the other model which a  colleague ucsf has been very interested in  

57:39is the vole so the um to mice and rats are quite  close evolutionary voles are much further away  

57:44but are very very similar in terms of the  husbandry and and processing um so we’re  

57:50going to see if there’s any evidence of the  antisense in that in that model the other nice   thing of folds is they’re inherently more sociable  they form they’re very key they’re full little  

57:58repair bonds um and so that that we’ve found  that there’s generally a more sensitive model   of behavior but then maybe with the seizures  that’s not such a problem in some gap one um  

58:09yeah i i i that’s a really interesting question i  don’t i don’t think i have a strong instinct for  

58:16that i mean the the closer you get to human is  better whether it’s worth the expense of doing   that i i think i’m gonna have to pass on that one  i’m afraid okay well thank you again pleasure i’m  

58:29i’m thrilled to see two brilliant minds chatting  about Syngap makes me happy um jr do you have a  

58:35question and then i see a question in the audience  i’m going to promote Dr Kadam to speak she’s a   researcher at hopkins who has um done a lot with  our asian community and i’m absolutely thrilled  

58:44to see her here but j.r do you want to go yes  thank you can you hear me yeah go yeah um so  

58:50thank you so much for your talk i wanted to ask  you some more about the isoform ratios and um so  

58:56i’m assuming those are in uh neural tissue that  you were looking right yes this is this is human  

59:02cortex right so i’m wondering based on that um  expression slide you showed at the beginning which  

59:08i think was from the broad gtex site there’s  obviously a bunch more tissues are there any  

59:15more amenable to doing kind of human biopsies on  patients so my point being can you can you analyze  

59:24the isoform ratio in a different tissue that  then would be amenable to taking a biopsy from  

59:31a patient to see how the syngap uh  you know when you have the syngap  

59:39ver you know uh pathogenic variant how does  that change your isoform ratios yeah that’s  

59:44a really good thought so i’m just i’m staring  at my slide trying to find blood um because you   know if you want you want it wasn’t very blood’s  pretty low that’s pretty low so but there were um

59:59yeah well let’s let’s just let’s imagine that  we found a tissue with a biopsy um so it’s the  

1:00:05the question then becomes whether the splicing  patterns are brain specific or neuron specific  

1:00:12so splicing can be very very um can differ wildly  between tissues it can also be quite similar  

1:00:19i think there’s a very um there is an easy  experiment for us to do to answer that question  

1:00:25that’s simply to go and take the gtex data  and do the same analysis there so one of the   um one of the things we’ve we’ve realized it is  um it’s difficult to look at isoforms and short  

1:00:36read sequencing data and we don’t have long read  sequencing data because you’re trying to sort of  

1:00:42guess how they all join together it’s like trying  to build a lego model looking at five bricks   together and working out how the rest of it looks  um but what it’s very good at doing is looking at  

1:00:50the ratio between two exons so i think we could  apply that to gtex and we could probably make  

1:00:56a list of tissues which seem to show similar  dynamics that’s an interesting question um let  

1:01:02me let me put that on the on the list and see what  we come up with okay great yeah because ultimately  

1:01:08the question is what’s going wrong right and so  yeah you obviously need a base of what’s right but  

1:01:14what’s going wrong in um our patients is really  important um and i think linked to that it does  

1:01:20the ratio change in the face of synthetic  one mutation because because these changes  

1:01:26are activity dependent and also there’s maybe  there might be some regular auto regulation going  

1:01:32on it’d be very interesting to know i mean even  in the mouse does the ratio change dramatically

1:01:40right so um i had one other question with  this with the uh crispr a with the sgrna um  

1:01:47does that so i’m assuming that sgrna must  distinguish between the two alleles it’ll just  

1:01:52yeah and so is is that because i’m sorry sorry  it doesn’t distinguish between two alleles sorry   it can distinguish the two isoforms or the iso the  long and short ice form but it doesn’t distinguish  

1:02:01the two alleles what you do is you up regulate  both um one of them still decay the other one okay  

1:02:08increases the the amount and that’s why it’s maybe  not such a good therapy for mis-sense mutations  

1:02:14because you also increase now it depends if  the it depends if it’s the absolute amount or  

1:02:22the relative amount of functional non-functional  protein that matters so in scn1 and 2a it looks  

1:02:28like there’s a ceiling to how much you can make  so increasing a missense doesn’t really help  

1:02:34in singing one it might do because it seems  to have a slightly more maybe permissive use

1:02:41okay and then my last question is about the  developing brain and where you see syngap   expression do you have any ideas as to the  cellular location is it is it doing sort of  

1:02:52signpost pathfinding signal transduction stuff or  i mean i assume it’s not doing um is it is there a  

1:03:01post-synaptic density at that time that it’s in or  is it doing some other function in the cell yeah  

1:03:07it’s a good question so the the the pattern it  it um shows mirrors that are synapse development  

1:03:14and so i think that that increase we see from mid  feet longwoods is probably something to do with  

1:03:21with developing and forming synapses before  that there’s not much in the way of neuronal  

1:03:27communication going on and so maybe there’s  a different role there early in development  

1:03:32and i think that’s that’s um speaks towards  it being in many other tissues which don’t   particularly have neurons and then in terms of  its actual function within the synapse i mean that  

1:03:42that enters into sort of proper neuroscience and  i’m you know I’m but a humble geneticist  

1:03:48so but i the one thing which also caught my eye  is the lack of mis-sense mutations in a single  

1:03:55functional domain it’s not like they’re all in  wrath so they’re all in the the sh um they’re  

1:04:01scattered across the whole thing suggesting that  you need the whole thing working not just say the  

1:04:06ras domain um but i think there’s clearly some  important neuroscience experiments there to do  

1:04:14and in the same way they have genes of multiple  isoforms they also have multiple functions   and this is why we need to look at all of  these genes together because you need to  

1:04:24find the point where the genes are all doing  something similar functionally because that’s   likely to be the true causation rather than just  something which it does which is not not sensitive

1:04:35okay great thank you thank you jr and then with  the last question um dr cut and i just want to  

1:04:42point out don’t make an ad for SRF we have this  team of parent doctors and scientists with hans  

1:04:48and j.r and dr dahia and marty dealt with before  and six others so as you and your team want to  

1:04:54talk more about Syngap you don’t have to talk to  normal people like me our our parents physicians  

1:05:00and scientists as you can see here are formidable  um dr kadam do you have a question yes hi hi dr  

1:05:07sanders uh that was fantastic talking my brain is  buzzing because i’m not a genesis and but i’m a  

1:05:14systems neuroscientist and uh my lab is epilepsy  research lab and that is how Syngap comes into  

1:05:20our domain for trying to figure out mechanisms  that result in maybe the early life seizures   and uh one of my i mean my first question that i  i put in is tied to the last thing that you kind  

1:05:31of addressed that the short form isoforms that  increase in the profile would indicate that they  

1:05:37may be tied to the uh you know the development  of synapses during maturation and the pruning  

1:05:44and and you may be familiar with the red  thing Rett syndrome and macp to data about   what that effect it might have on pruning  of synapses and how that might affect  

1:05:53uh evolution of autism or intellectual disability  so since you you kind of seem that you acknowledge  

1:05:59that that might be happening my other question  then based on what we have seen and published  

1:06:06related to what you showed a  little bit about the expression in   both in excitatory neurons and inhibitory  neurons uh including pervaluan uh interneurons  

1:06:15and we record cortical oscillations uh in in these  gap mice and we find that it’s very it’s very  

1:06:21disrupted and power of albumin uh pv internal  dysfunction is heavily tied to uh corticolor  

1:06:28cortical gamma oscillation my question was then  how would we or how do you see us going forward  

1:06:33once we get more information on these short form  isoforms in the synapses uh that when this hypo  

1:06:39insufficiency would of be further affected  by the firing rate property of that neuron  

1:06:45and as you know interneurons have our brain are  called fastbiking so given the same pathology how  

1:06:50would the dynamics of the firing pattern of that  neuron in a immature brain then just foresee uh  

1:06:57affecting um the synaptic homeostasis so i i know  i’ll give you my biased opinion i would think  

1:07:04that it would affect interneurons much more than  the excitatory neurons given what we understand  

1:07:11so far what do you think about that so i think  i’m a bit out of my my domain of comfort um  

1:07:18so the it is very it is hard to extrapolate from  rna patterns to protein levels let alone the  

1:07:29impact of proteins on firing cells oh we know  because we have these eegs and we’re trying to  

1:07:34look at all the data that you guys are generating  and we cannot make any connections exactly and  

1:07:40it really and this this is what so in scn2a the  thing which has made it sort of so much fun and  

1:07:45so i i think we’ll be able to progress is having  like kevin bender who’s an electrophysiologist  

1:07:50or you know the dav is a genomic engineer and  needs to bring the genetics it really it really  

1:07:56understanding these disorders needs  cross-disciplinary interaction um   there’s no way i’m ever really gonna know enough  neuroscience to be able to sort of interpret  

1:08:05the wiggly lines that get made um but the  insights of that is just trying to tie  

1:08:10all those bits together um and i think the  the thing i would say comes from this is that  

1:08:18it is easy to imagine increasing the short or  long isoform to be able to create a model system  

1:08:24where you could go and answer that question  um you know i’m sure your intuition is right  

1:08:30but of course like with all of this and even  with the expression profiles it needs to be   um tested to see how that impacts things it’d  be really interesting a simple experiment  

1:08:39of trying to increase the long or the short  isoform and seeing what that does to seizures   what it does to behavior i think that would  be quite interesting and maybe that’s why  

1:08:48crispr a maybe has a role here as both an  experimental system along with you know  

1:08:54potential therapy could you also do it with an  aav system just simply replacing the short isoform  

1:09:01yes uh yeah in the translation to therapeutics and  since you said about you know the packaging and   the size limit maybe the short form would be more  aminable to kind of that kind of delivery yeah yes  

1:09:13if this synapse strong’s enough association of the  short form and uh you know what we i’m trying to  

1:09:20tie for which you know now that’s why my brain is  buzzing what experiments could we do to prove that  

1:09:25that that that the same pathology would be much  worse in a fast packing uh interneuron than a  

1:09:32neuron and when we talk about seizures coming  early in life versus the same child because   we know that you know they will go through some  stages and the epilepsy will evolve parents know  

1:09:40this different types of seizures come on as the  circuit is enjoying that maybe the therapeutics  

1:09:47at least not related to uh replacing the protein  would be different at the early time point versus  

1:09:52the later time yeah in the sense of because all  of these children are on some type of medication  

1:09:58for stopping their seizures and if we are  understanding from these mechanisms ends up being   true as we get more fine-tuned in our experiments  then the same drug there would be a different  

1:10:07class of drugs you would target at a younger  child having seizures yeah then older child  

1:10:13so that’s why you know just you what your talk was  really um i mean i just have 100 more questions  

1:10:18no answers but uh yes exactly hypothesis  generating indeed um but i think it might be  

1:10:28i always think like the rain is hideously complex  but it’s built on lots and lots of simple bits   it might be there’s a really simple answer here  that the long isoform does sort of things which  

1:10:37they it does in all the cells the short isoform  does stuff in the synapse does the stuff which   leads to thera leads to problems and actually the  short term synapse it’s short ice form is is what  

1:10:47is needed to be replaced i mean maybe the simple  answer there but i think it’s it’s tractable it’s  

1:10:52not an easy experiment but it is attractive  experiment yeah yeah yes well thank you thank  

1:10:57you for your doc i enjoyed it thank you thank  you it’s fun to talk to the group i want to um  

1:11:03i have a lot of questions too but none of them are  as smart as what’s been said hopefully we’ll have   a chance to speak again i want to call attention  to J.R.’s question in the maybe you can articulate  

1:11:12it but i want to also reference i’m sure you’ve  seen the the paper uh in brain earlier this year  

1:11:19by dennis law’s group it was a little yeah so  when that came out estimating frequencies yeah  

1:11:26yep i mean i i jumped all over him because living  at the heart of this organization i keep seeing  

1:11:33the parents show up and i keep saying this is a  radically under diagnose that disease and as we   had a deeper appreciation for the breadth of the  phenotype a lot of children i think are just not  

1:11:43getting caught and not getting sequence because  the phenotype’s relatively mild so they’re getting   an autism diagnosis maybe they’re getting doped  up with these metabolisms being put in a corner  

1:11:53all due respect to those wonderful drugs but i  think there’s a lot of our kids being missed and   so when that paper came out i literally tracked  dennis lal down and i was like explain this to me  

1:12:02like kind of um and the thing that struck me in  that paper and he said to me very kindly although  

1:12:08i’d obviously asked a stupid question he said  well why is there so many more missense mutants   than what we see in clinic because because in the  population we have a lot of ptds and a handful of  

1:12:17missense but his predictions is the opposite  if you double click on his supplementary table  

1:12:226.1 something for a hundred thousand but the  ratio between misheads and ptvs is like five   to one whereas if you look at clinvar it’s it’s  more like two-thirds ptv i said why is that  

1:12:34and then the commentary on that was like well  they’re either so mild we’re not catching them   or they’re so bad they’re pathogenic they’re not  pathogenic the kids aren’t the they’re not even  

1:12:44being born yeah and we don’t know and then and  then J.R. is asking a similar question in the chat  

1:12:50like how do we think about all these missense  mutations and then the assumption that i’ve heard   and again i’m generally the dumbest person in the  room but what i’ve heard from my researchers is  

1:12:59we’re assuming these mis-ends are just becoming or  just behaving like ptvs and everyone’s haploid is   efficient it feels like a bit of a simplifying  assumption maybe it’s the truth who knows  

1:13:08um so i i would point you to jr’s  question but welcome any commentary on on  

1:13:14the discrepancy between what lala at alpert you  know are predicting with majority missense and  

1:13:20what we’re seeing which is kind of the opposite  yeah absolutely i’m just looking up in schema to  

1:13:26see um which is the um the schizophrenia resource  just see how many missings have come up there  

1:13:32uh i’m going to dig into that a bit more now  so not very many it this this is a mystery to  

1:13:38me where are the misinterpretations i it is the  right question um i think the two possibilities  

1:13:44you’ve articulated there are right it’s either  ascertainment through diagnosis or fatal um  

1:13:50most it is very easy to break a gene it’s hard to  make it do something extra so there’s there may  

1:13:56be a small number of gain of function mutations i  find it very easy to imagine that they are fatal  

1:14:02um but there’s not going to be many of those  there’s like to be a large number of loss of   function mutations quite why you um why those  don’t make it to the clinic is a surprise  

1:14:14that this the temptation is to think that it’s a  milder phenotype and that seems the most obvious  

1:14:20or the sort of the hypothesis to beat that  would have the implication that it’s not   ras or the other domains it’s all of them you you  need all four to be missing or five to be missing  

1:14:32missing just one of them is bad but it’s not  terrible and so that it’s actually the way   those coordinate together that would be a simple  ex simplistic explanation to it in terms of where  

1:14:43they are um in in nomad which is looking at  the population there are far fewer mis-sense  

1:14:50variants you’d expect by chance suggesting  that mis-sense variants do do something bad  

1:14:56um i i don’t think the answer is autism because  we take all comers to autism and we’re seeing  

1:15:02the same issue of being being more ptvs i  i think it must be milder phenotypes but  

1:15:08quite how you find those i don’t know the thing  that what would be a lovely experiment would be to  

1:15:14sort of functionally analyze every possible missed  sense very within sin gap one but you know it’s  

1:15:19difficult because what’s your assay what’s the  what’s the end point that you believe matters  

1:15:25and i i’m not sure we’re entirely there yet it’s  easier than something like a transporter yeah  

1:15:31because this is this is the various biotechs  and drug companies we’re talking to are all like   what would we measure as an endpoint that is  yeah this is why hans is so focused on biomarkers  

1:15:42go ahead please and seizures i mean in terms  of the endpoint seizures is is hard to beat  

1:15:48as an endpoint it’s quantifiable it’s it’s you  can you can verify that it is a true seizure  

1:15:54um and so that the fact that there’s a high  rate of seizures should put it in the top tier  

1:16:00of ones for for drug companies yeah i’ll let dr  kadam talk about that if she wants but we have a  

1:16:07radical breadth of seizures as well we have  everything from the either mycolonia to  

1:16:12really um i mean i i would like to add uh so  we have uh just new like heart of the press  

1:16:18unpublished data that we’re going to talk about  at the nih-sponsored meeting in two weeks so  

1:16:24our lab i think is the first lab now to record  from uh new needle pops from the sungap model  

1:16:29that rick jupine’s lab has done and uh we are  recording a bunch of electrographic only seizures  

1:16:37and not to sound an alarm or anything but  and it’s not really surprising is it because  

1:16:42we know that many development disorders have  very early life seizures but this was not we  

1:16:49know there’s a subset of Syngap patients who have  epilepsy diagnosis within the first year of life  

1:16:57so maybe four months or seven months when it’s a  clinical diagnosis because it will be a clinical  

1:17:03observation but a concept in epilepsy research  is these um subclinical seizures are what we call  

1:17:09electrographic only seizures and there is no way  to diagnose a kid with those unless you do an eeg  

1:17:15and what we are recording from the pubs for  the very first time is a a pretty high burden  

1:17:22of electrographic only seizures and uh we won’t go  to the details because we’re doing some additional  

1:17:28controls the response to the traditional drugs  is uh very surprising they’re not responsive yeah  

1:17:34there’s already clinical reports of fifty  percent of syngap patient uh spin gap   patients with seizures being refractory  to first-line standard treatments and  

1:17:44and that’s why i was that’s why my brain is on  fire on understanding these iso forms specific  

1:17:49uh stuff about what would and that’s what we  are bread and butter’s understanding mechanisms   underlying refractoriness right because there are  20 plus drugs on the market most of them have come  

1:18:00come up in the last couple of decades uh but  they all focus on one mechanism it’s ion channels  

1:18:06and not on chloride co-transporters which is  what we are looking at but also so the idea  

1:18:12of gaba being depolarizing and that’s why we  are interested in interneurons and how SynGAP   may be affecting interneuron function early in  life and as you know may know that uh that is  

1:18:22one system that is maturing right up to teenagers  in human and also in the mice that’s the longest  

1:18:27maturing system and uh you know it’s early  it’s already time in the Syngap community and  

1:18:33you know the family foundations are doing we  need to start staging uh just like they did in   red the stages in this syndrome in the sense of  how many fall into this early life kind of uh  

1:18:46high seizure burden uh phenotype versus late  onset seizures and what does that mean to the  

1:18:51because it’s reflecting what the pathology is  related to every other portion of sengab may be   the intellectual disability autism and all the  others uh like you said centigap is expressed  

1:19:01all in all the cell types so that you know  what are the systemic uh symptoms associated   with some gap apple insufficiency so as we  start getting ready to put publish this data  

1:19:12uh you know one thing about as being a  translational research is not to sound alarmist  

1:19:17because you know as scientists we see the result  we report it as it is but but i am surprised by  

1:19:24the load of the subclinical seizures we are seeing  in this very young population of sengab mice uh  

1:19:29that no one has looked at frankly before which  is kind of a paradox because it’s a developmental   disorder we better be looking at the ages that  are translational and even we when we publish  

1:19:39we looked at the our pilot in adult because that’s  where we were starting in that sense we have quite   a new group to send gap specific research and  there is a singaporean clinic at kennedy krieger  

1:19:49from our early paper now you know we’re  trying to move to this clinical trial   but then now with this whole uh any success of  any clinical trial will be as good as you know  

1:20:00uh what stage the incoming recruited patients are  at yeah and that it’s not even what we’re learning  

1:20:07from the aso and sma studies uh it has has to  even even without with the very good hypothesis  

1:20:14and sound scientific data of how an asu might  work uh the new reports coming from that patient  

1:20:21group of how variable the success rate is  is already telling us that it’s not going  

1:20:26to be it’s going to be about fine-tuning it  for every child every patient yeah and and  

1:20:32how severe they are what kind of staging and side  comorbidities they have is going to determine  

1:20:37what your best science is going to be translated  into and to start thinking of parallel thinking   all those things right now let me know like when  we know better we should do better that that  

1:20:47to start staging this as the science move forward  that when we start clinical trials we don’t just  

1:20:53recruit the most number of patients but we know  who’s coming in at what stage yeah and when did  

1:20:58they start seizing because you know repeated  seizures themselves have their own pathology   absolutely you can kindle animals and a normal  animal can become epileptic so now let’s say you  

1:21:09have a child that’s having electrographic seizures  only 400 a day that that would have its own  

1:21:15pathogenesis and then when you do a clinical trial  that excludes that and then your clinical trial   fails you cannot really blame the sciences the  poor planning of understanding the staging and and  

1:21:25i’m glad that it’s getting there now with you know  with this huge i mean we cannot even discount how   much these family foundations help in the sense of  collecting this data the natural history uh which  

1:21:35used to be ignored or not well um classified in  previous clinical trials so so i’m you know i’m  

1:21:42kind of excited that these two things will happen  at the same time and as as new therapies come in  

1:21:48we do it a little bit more smarter because  science is smart the way it gets applied sometimes  

1:21:54it’s not as smart yeah i think um in scn2a we  found that it was amazing how much progress we  

1:22:03made by essentially staring at what parents  said and staring at the gene um i think yes  

1:22:11in retrospect it was you know it was really easy  um sitting at one less easy there’s quite a lot  

1:22:18of variation and we sort of we know where  a lot of that variation comes from it comes   from the environment and common variation but then  there’s other variation which comes from syngap1  

1:22:26and that is tractable if we can just work out what  are the bits which which matter and but um i think  

1:22:33in terms of clinical design definitely detailed  recording of phenotypes really really important  

1:22:39it’s hard to guess beforehand which ones matter  yes and which ones are going to guide success but  

1:22:46i would also i wouldn’t underestimate just having  a big study you know putting my statistical hat on  

1:22:52all problems like this with a big enough sample  size come out in the wash a hundred percent but  

1:22:58when we speak of rare dog rare disorders mostly  profiles all fall into that ditch because you know  

1:23:05the international cohort of of collecting every  possible diagnosed patient into the study uh yes  

1:23:11and we will get there hopefully but that has been  that has been the ditch that most previous trials   have been falling into yeah um thank you so much  dr cotton i’m really thrilled you joined this  

1:23:24chat i want to just make two points and maybe  a question but i realize we’re already well  

1:23:30over time and i’m very appreciative of everyone’s  time first point is i think you know you started   talking about autism and i’ll observe as a as  a someone who’s engaged with a lot of parents  

1:23:41our kids really struggle to get an autism  diagnosis once the seizures are under control   and i’m and as dr kadam alluded to that involves  a lot of aeds which involves a lot of side effects  

1:23:51which i think we tend to gloss over but  the parents struggle mightily with behavior  

1:23:56and then they’re told to get an aba and then they  go to do an [ __ ] and our wonderful children walk  

1:24:01into the room look at the clinician and they’re  very used to going to hospitals by then and   they look and they make they lock eyes with the  clinician and they smile beautifully and their  

1:24:10odds of them getting an autism diagnosis just only  when you have a parent rock up with scars on their  

1:24:16face and bruises on their arms talking about how  the last behave tantrum is a train wreck does  

1:24:21somebody who’s in the [ __ ] autism diagnosis  mode sort of think well they’re awfully social  

1:24:30but this parent needs aba and this kid has enough  other phenotypes that we should probably give them   the diagnosis so a long way of saying our kids  are kind of our kids are not classically autistic  

1:24:40the way people describe an autistic child and  i think that’s an important observation about   the phenotype um and the other thing i and i  welcome a comment on that i’m just sharing it  

1:24:50yeah especially after your lovely comment on scn2a  and how you looked at the kids and looked at the   genes two other points one we are um doing a what  we’re calling a digital natural history study  

1:25:00which deeply offends some of our sap like ingrid  shepherd’s like it’s not a natural issue study   unless you fly everyone to me and i test them we  love ingrid but it’s it’s three digits cheaper  

1:25:12to just take everyone’s medical records and  normalize them so we’re doing that right now   and we’ve got 50 in the u.s and we’re going to  enroll another 50 in december so that’s a decent  

1:25:21cohort and we’re going to keep going um so we’re  going to have hundreds but sometime next year  

1:25:26yeah the third point i just want to ask you  how did you i mean you didn’t put together  

1:25:32this lovely presentation on singet because we  asked you to right so you’ve got this depth of   knowledge on scn2a what was it about singap that  caught your attention was it just that it was on  

1:25:40that the top of that chart and you said we got to  double click here that’s my last question then i   promise you’re sure the last one’s easy say it’s  the top of the chart yeah i’m strong statistics  

1:25:50background there’s a reason it’s at the top so you  know follow the data um that that’s interesting  

1:25:55i think i think the fact that seizures are an  important point i i see the root towards therapy   in these is easiest in the disorders where sieges  are prominent because you have a built-in endpoint  

1:26:05um i think the insights will spread to that all  of the genes but we’ve got to pick a few which  

1:26:10we’re going to drive forward and sim gap one to me  looks like one of those um in terms of autism you  

1:26:16know speaking to the parents in the room like  you know you are trying to get an order from   diagnosis because you get more resources you need  to find the friendly clinician local to you who’s  

1:26:26going to listen to you and give you the diagnosis  you need to get the resources you want you know  

1:26:31i must say from a science point of view the  distinction between autism developmental delay to   me it looks like almost a distraction i mean it’s  that they are so so close to each other yes there  

1:26:42are distinctions there are people who are very  autistic without any of the other developmental   delay but they’ve not got the rare variants um  so i you know just do what you need to do to get  

1:26:53the autism diagnosis it’s not going to it’s not  going to damage the research or anything like that  

1:26:59good good find that find the friendly clinicians  the ones who listen and make sense and that’s  

1:27:05another area of collaborating with the other  resources because this is the same thing happens   in developmental other developmental delays um so  yes and there was a there was a second question  

1:27:16i think i was just plugging the natural general  natural history study in natural history yeah   it’s essential if you’re going to ever get to  the clinic you need an out of history absolutely  

1:27:23important and um and jenny thank you for those  comments very interesting similar things these  

1:27:28are not unique to us in gap one that what i would  love to see is someone quantifying the sleep gut  

1:27:36behavior high pain tolerance across genes because  they all have them but do some have them more  

1:27:42than others and that’s what i need to to know  how to interpret the data i i would volunteer  

1:27:48myself meaning srf and say stxbb1 and a number of  other rare groups because i’ve been fantasizing  

1:27:54about like i don’t know if it’s called a spider  graph or a radar graph but you take those those  

1:27:59symptoms and you relatively because for us i  mean some i think as fox g1 mom was like oh  

1:28:05yeah behaviors aren’t a big deal for us and i was  like really yeah and i just understanding that at  

1:28:10a relative level would be absolutely fascinating  absolutely it that is an absolutely critical   experiment the family groups need to come together  and collect the same data collect it in the same  

1:28:20way i i think there will be really important  insights from that but it’s you will need big   big numbers and really good data collection like  citizen well thank you so much i’ve been laughing  

1:28:33you might have seen me looking off screen and  chuckling i’ve been getting all kinds of excited   text messages from parents and people watching  about the quality of this so really compliments  

1:28:41you and your team and thank you again thank you  great pleasure to talk to you and uh contact

1:28:47cheers you