SYNGAP1 Prevalence: Why We Are Sure That SYNGAP1-Related Intellectual Disability is Under-Diagnosed?

Posted on:

Mike is the Managing Director of SRF; he founded the organization with his wife Ashley to help their son Tony and others like him.

Executive Summary

Driving towards therapeutics requires collaboration among many parties; at SRF, we are trying to identify who these therapies will help (identify patients) and who will help us get us to that point (families and scientists).  In that spirit, this article is to help those trying to understand the science around incidence of mutations in SYNGAP1.  If you do not read beyond this paragraph, the takeaway is that we are likely under-diagnosing SynGAP1-related Intellectual Disability by orders of magnitude.  This means that there are far too many people who are not participating in the current collaboration, both to understand the disease better and to benefit from eventual treatment. More work needs to be done to understand how many orders of magnitude.

This article has three parts. The first walks through the previous literature about Syngap prevalence in certain studies.  It was these studies that got us to the standard sound bite in the SynGAP community: “SynGAPians comprise ~1% of all cases of ID.”  

The second part introduces a recent paper by Dr. Lal’s group  and shares a remarkably high predicted incidence of Syngap mutations.  Ranking Syngap against other well known NDD (neuro developmental disorder) & DEE (developmental epileptic encephalopathies) diseases, suggests Syngap is comparable in size to Dravet and larger than CDKL5, Rett or Angelmans Syndrome.   These are important comparisons since all of these diseases have therapies in development. To our knowledge, SynGAP does not, yet.

Finally, we peel the onion on that high SYNGAP1 incidence and look at how it breaks down between missense and PTV (protein-truncating variant) mutations.  Here we realize that a higher number of predicted missense mutations than we were expecting actually begs more questions.

Introduction

SRF Patient Advocates will often state that while the #SynGAPCensus indicates 535 identified patients worldwide at the end of 1Q20, the real number is much higher.  What scientific evidence do we have for this claim?

Until a recent paper from a genetics group led by Dr. Dennis Lal, two recent scientific papers validated the view that the Syngap population is much larger than currently diagnosed.

First, the standard talking point in our community was most recently repeated in Gamache’s 20 year review of SynGAP: “SYNGAP1 mutations may account for up to 1% of all cases of nonsyndromic ID.”  As non-syndromic ID accounts for ~2.3% of the global population (Lemke 2020), then 1% of 2.3% is 0.023%, of humanity.  With a global population of 7.7 billion then there are ~1.8 million SynGAPians & we’ve counted just over 535.  In the US with a population of 330 million, we should have ~76,000, and we’ve only counted 162.

Second, in a recent meeting to review an ICD-10 code for SynGAP, the CDC brief estimated that  “>1 million individuals predicted to be affected world-wide, making pathogenic SYNGAP1 variants more prevalent than Fragile X syndrome.”

A brief review of the data that supported these assertions is useful:

EARLY STUDIES LOOKED AT COHORTS OF PEOPLE WITH INTELLECTUAL DISABILITIES AND FOUND SIMILAR PERCENTAGES WITH SYNGAP MUTATIONS.

  • Hamdan in 2009 noted that 3% of patients in his study with “non-syndromic mental retardation” had SYNGAP mutations.
  • Berryer in 2013 found that 9 of 186 [~5%] NSID  (non-syndromic intellectual disability) patients had SYNGAP mutations.
  • Samocha in 2014 found that 3 of 151 [~2%] patients had SYNGAP mutations.

THE DDD STUDY IN THE UK REINFORCED BOTH THE IMPORTANCE & RELATIVELY HIGH INCIDENCE OF SYNGAP.

  • 2015 study of 1,133 patients found 7 with SynGAP mutations [0.6%]; this was the fifth most identified gene in the study.
  • 2017 study called SynGAP one of the “six most significantly associated genes”
  • Noting here that this was a much larger cohort and the incidence came down from 2-5% to 0.6%.  Read on.

MORE RECENT STUDIES VALIDATE SYNGAP1 AS A TOP DIAGNOSTIC GENE.

  • Wright in 2018 also found SYNGAP1 to be the 6th most diagnostic gene after RID1B, SATB2, SCN2A, ANKRD11, MED13L.
  • Truty, reviewing 9,413 patients tested with the Invitae panel found that SYNGAP1 was the 10th highest incidence gene, accounting for 2.5% of positive diagnoses; notably, however in addition to the 39 hits, there were another 79 VUSs (“variants of unknown significance”).
  • Brimble, in a case report, pointed to an example of a VUS that was found to be pathogenic with RNA testing.  How many other VUSs were actually pathogenic but not identified as such?

A RECENT STUDY SUPPORTS THE 0.5% – 1% RANGE

  • Johannesen et al in 2020, sequenced 200 patients with Epilepsy and ID in Denmark.  46 patients [23%] had a genetic cause discovered; one 26 year-old had a SYNGAP1 mutation.  1 in 200 is 0.5%.

Are all our missing patients just mild phenotypes?

As Vlaskamp et. al. noted when commenting on their own excellent study, a detailed review of 57 SynGAP-affected patients, they “might have underestimated the prevalence of milder phenotypes due to ascertainment bias, given our recruitment via Facebook.”  As a population, our experience has been that those patients who have milder phenotypes (relatively speaking) are less likely to be directed towards genetic testing.  But how many are we missing?  Are they all just mild phenotypes?  55 / 57 of the patients in Vlaskamp’s study have seizures and 56/57 have intellectual disability– are there many others out there without these afflictions, but with a mutated SYNGAP1 gene that is causing disease?  It’s not clear.

Someone boiled a data ocean

Until April 2020, what we’ve written so far is about as much as we knew.  But then a group of geneticists led by Dr. Dennis Lal published a paper called A catalogue of new incidence estimates of monogenic neurodevelopmental disorders caused by de novo variants that also had a commentary from Lemke in the same publication.  It was only seven pages long, one table, one chart.  But the magic was in a Supplementary Table, a link can be found at the end of the first paragraph in the results section.  After you unzip the attachment and open the Excel file you find detailed estimates for incidence for 101 Established NDD (Neuro Developmental Disease) genes on the second tab.  To save you the trouble we’ve put the data in a Google sheet.

The last author on this paper is Dr. Dennis Lal.  He actively engages with rare disease groups and is someone SRF has followed since we met him last year.  The  affiliations on this paper is a list of leading medical institutions: the Cleveland Clinic, Cologne, Boston Children’s and the Broad.  We asked Dr. Lal to explain in simple terms how the estimates were derived, and he kindly obliged. Essentially he took the average number of de novo mutations per person and looked at each gene’s length adjusted by the characteristics of each potential mutation.  Based on that, he was able to generate incidence estimates of mutations on each gene.  These incidences are represented as a number per 100,000 conceptions. This is our simplification of his simplification — his work is  much more complex.

Only a few genes have these estimates already in the literature, but there were some.  The paper includes a table that compares existing estimates to those from their analysis.  For all genes but one, there was no significant difference.  This suggests that the approach is valid.

Screen Shot 2020-05-04 at 11.26.56 AM.png

So what was the number for SYNGAP1?  6.107 per 100,000 (the range was 5.7 to 6.5).  That means that in the USA, which had a birth cohort of 3.8M in 2018, there would be over 230 SynGAPians born every year.  It also means that the previously mentioned incidence of 0.023% might be ~4x too large — but even if you cut the projections to a quarter  (6.107/100,000 =  0.006%, of humanity), globally we’d expect  ~0.46 million SynGAPians vs. a count of 535;  In the US we’d expect  ~20,000, and we’ve only counted 162) the comparison between the existing diagnoses and the expected numbers suggests that we are radically under-diagnosed.  

But wait… the kinds of mutations matter

When you double click on the incidence predicted by Dr. Lal’s group, that incidence of 6.107 for a SYNGAP1 mutation is broken down into two kinds of mutations: PTV and missense. This is where it becomes interesting.

Let’s take a step back. Patients with SynGAP have one functioning gene and one that is not functioning properly (this is why it is called a haploinsufficiency). The one that is not functioning properly could be a PTV (“protein truncating variant”), which means a mutation that results in the protein being shorter than it should be.   As a result, the body looks at the protein and can’t make sense of it, and therefore disposes of it (this process is sometimes called NMD or “nonsense mediated decay”).  

Alternatively, the gene that is not functioning properly could be a “missense mutation,” which happens when the typo in the gene still causes a protein to be produced, but produced in such a way that it causes a change in the function of the protein.  The impact of these mutations isn’t uniform, and there are many open questions as to what damage this mis-functioning protein can cause.

The predictions from Dr. Lal’s group, however, estimated the PTV incidence is 0.787 & the missense incidence is 5.320.  They predict almost seven missense mutants for each nonsense mutant — to SynGAP parents, who typically hear more about nonsense patients, this is surprising.

Where are all the missense patients?

When a genetic diagnosis is made, the mutation that it causes is entered in a database called ClinVar.  If you have enough bioinformatics training you can work with that, but for rare parents, a site called Simple ClinVar does the trick.  It was built by Dr. Lal and his apparently tireless team.  

If you go to Simple ClinVar, type in SYNGAP1 and sort for pathogenic or likely pathogenic mutations you find 132 distinct mutations, 21 of which were missense.  So in the literature today, for mutations that cause or likely cause disease, we have 4 mutations that weren’t missesene for every mutation that was — this feels approximately right based on the numbers in our community.

Screen Shot 2020-05-13 at 3.26.22 PM.png

However, Remember, Lal’s model predicted that there are 7 missense mutations for every PTV, so this clinical data is at odds with the genetic model.  But maybe the missense mutations aren’t getting picked up as being pathogenic? However, even if you don’t filter for pathogenicity and look at the big picture, you see 370 known variants, 129 of which are missense.  That one missense mutation for every ~2 PTV, instead of  7:1.  Clearly patients with missense mutations are not getting diagnosed — why could this be?

Screen Shot 2020-05-13 at 3.41.56 PM.png

Noting that in Clinvar there are 444 SYNGAP1 variants noted, not 370.  When Simple Clinvar is updated, we will write an update for this article.

Where are the Syngapians?

Dr. Johannes R. Lemke wrote a commentary on the predictions from López-Rivera, et al. (Lal’s group) in which he put forward an answer.  One paragraph sums it up, so I quote it here in its entirety:

“However, despite the thorough statistical calculations, the list of incidence estimates is still vulnerable to a few factors that could potentially cause the expected/calculated incidences to deviate from clinically observed frequencies. A gene may be ranked with a somewhat higher incidence if it has high genetic constraint scores and is extraordinarily large (e.g. DYNC1H1 with 4646 amino acids, which is first on the list, or KMT2D with 5537 amino acids, which is third). The same overestimation may occur if the associated phenotypic spectrum comprises embryonically lethal phenotypes that will escape detection by routine screening of affected living individuals. By contrast, the predicted incidence may be underestimated if a gene is rather small or its associated phenotype lies towards the mild-to-normal end of the phenotypic spectrum, with the result that mildly or subclinically affected individuals may be recruited among healthy control populations.

Said differently, larger genes will be higher on the list just because the size of the gene increases that chance that one of few de novo mutations everyone has lands on that gene.  Syngap is 1345 amino acids long so it is not that big.  But the two points that follow are what matter.

Recall the comment we made earlier: that missense mutations, by nature of the way they operate, each cause the affected protein to be expressed in a different way. As such, both elements of Lemke’s argument could be true for SynGAP patients affected by missense mutations.

On one hand, we may have lots of people out there with Syngap missense mutations that aren’t causing pronounced disease; the “associated phenotype lies towards the mild-to-normal end of the phenotypic spectrum.  And as a result we aren’t getting them diagnosed.  This brings up questions of whether the phenotype is mild, or if it just takes longer to manifest–median seizure onset is at age 2 per Vlaskamp–and then we have people who are presenting with symptoms later in life and as a result are not receiving genetic testing.

On the other hand, there may be missense mutations that are “embryonically lethal” and these Syngapians don’t even make it to birth.  This hypothesis would be theoretically supported by the fact that, although all PTV’s behave the same way (conceptually), different missense mutations may have different impacts. It could be that some of the missense mutations are embryonically lethal, whereas others (those of the patients we know and love) cause the disease known as SYNGAP1-Related Intellectual Disability or SYNGAP1-Related Developmental and Epileptic Encephalopathy (Holder 2019).

Conclusion – We need more research

When I first saw Dr. Lal’s number I was glad to see that we had some data to validate the anecdotes from the cohort studies that there are many more SynGAPians.  Finally, there was a major scientist publishing what we believe.  But at the end of all this thinking, we have new and more nuanced questions, and the same next step: Do more research.  There are cohorts in the UK, Denmark and others where large populations have been sequenced and we could imagine a study where we seek out undiagnosed Syngap mutations and work to better understand their effects.  This feels important since we still keep finding SynGAPians that were missed.  A quick review of the weekly profile by SRF will highlight a few people diagnosed in their 20s or 30s.  So we keep working for our patients and all the ones not yet diagnosed who we are every more certain are out there.

See slides associated with this article here.  Please send comments to the author at Mike@SyngapResearchFund.org.

BIBLIOGRAPHY

Berryer MH, Hamdan FF, Klitten LL, et al. Mutations in SYNGAP1 cause intellectual disability, autism, and a specific form of epilepsy by inducing haploinsufficiency. Hum Mutat. 2013;34(2):385‐394. doi:10.1002/humu.22248

Brimble E, Lee-Messer C, Nagy PL, Propst J, Ruzhnikov MRZ. Clinical Transcriptome Sequencing Confirms Activation of a Cryptic Splice Site in Suspected SYNGAP1-Related Disorder. Mol Syndromol. 2019;9(6):295‐299. doi:10.1159/000492706

Deciphering Developmental Disorders Study. Large-scale discovery of novel genetic causes of developmental disorders. Nature. 2015;519(7542):223‐228. doi:10.1038/nature14135

Deciphering Developmental Disorders Study. Prevalence and architecture of de novo mutations in developmental disorders. Nature. 2017;542(7642):433–438. doi:10.1038/nature21062

Gamache TR, Araki Y, Huganir RL Twenty Years of SynGAP Research: From Synapses to Cognition. J Neurosci. 2020 Feb 19;40(8):1596-1605. doi: 10.1523/JNEUROSCI.0420-19.2020.

Hamdan FF, Gauthier J, Spiegelman D, et al. Mutations in SYNGAP1 in autosomal nonsyndromic mental retardation [published correction appears in N Engl J Med. 2009 Oct 29;361(18):1814. Perreault-Linck, Elizabeth [corrected to Perreau-Linck, Elizabeth]]. N Engl J Med. 2009;360(6):599–605. doi:10.1056/NEJMoa0805392

Holder JL Jr, Hamdan FF, Michaud JL. SYNGAP1-Related Intellectual Disability. 2019 Feb 21. In: Adam MP, Ardinger HH, Pagon RA, et al., editors. GeneReviews® [Internet]. Seattle (WA): University of Washington, Seattle; 1993-2020. www.ncbi.nlm.nih.gov/books/NBK537721/

Johannesen, K. M., Nikanorova, N., Marjanovic, D., Pavbro, A., Larsen, L. H., Rubboli, G., & Møller, R. S. (2020). Utility of genetic testing for therapeutic decision‐making in adults with epilepsy. Epilepsia. doi:10.1111/epi.16533

Kim, Jee Hae, et al. “SynGAP: a Synaptic RasGAP That Associates with the PSD-95/SAP90 Protein Family.” Neuron, vol. 20, no. 4, 1998, pp. 683–691., doi:10.1016/s0896-6273(00)81008-9.

Lemke, Johannes R. “Predicting Incidences of Neurodevelopmental Disorders.” Brain, vol. 143, no. 4, 2020, pp. 1046–1048., doi:10.1093/brain/awaa079

López-Rivera JA, Pérez-Palma E, Symonds J, et al. A catalogue of new incidence estimates of monogenic neurodevelopmental disorders caused by de novo variants. Brain. 2020;143(4):1099‐1105. doi:10.1093/brain/awaa051

Pérez-Palma, Eduardo, et al. “Simple ClinVar: an Interactive Web Server to Explore and Retrieve Gene and Disease Variants Aggregated in ClinVar Database.” Nucleic Acids Research, vol. 47, no. W1, 2019, doi:10.1093/nar/gkz411.

Samocha KE, Robinson EB, Sanders SJ, et al. A framework for the interpretation of de novo mutation in human disease. Nat Genet. 2014;46(9):944‐950. doi:10.1038/ng.3050

Truty, R, Patil, N, Sankar, R, et al. Possible precision medicine implications from genetic testing using combined detection of sequence and intragenic copy number variants in a large cohort with childhood epilepsy. Epilepsia Open. 2019; 4: 397– 408.

Vlaskamp DRM, Shaw BJ, Burgess R, et al. SYNGAP1 encephalopathy: A distinctive generalized developmental and epileptic encephalopathy [published correction appears in Neurology. 2019 Nov 12;93(20):908]. Neurology. 2019;92(2):e96–e107. doi:10.1212/WNL.0000000000006729

Wright CF, McRae JF, Clayton S, et al. Making new genetic diagnoses with old data: iterative reanalysis and reporting from genome-wide data in 1,133 families with developmental disorders. Genet Med. 2018;20(10):1216–1223. doi:10.1038/gim.2017.246