Is “PMS-SHANK3 unrelated” truly unrelated to SHANK3? Maybe not.

David relaxing in the car during a drive through the countryside

Originally posted 10 April 2024

In 2022 a group of scientists associated with the Phelan McDermid Syndrome Foundation (PMSF) published a consensus paper that addressed a controversy around what constitutes Phelan McDermid syndrome (PMS). It was an important step toward defining the disorder, providing guidance to geneticists, and working towards an ICD code for diagnosis and medical reimbursement. Although the consensus opinion was not universal, the clarification was welcome. These experts split PMS into two mutually exclusive classes: “PMS-SHANK3 related” and “PMS-SHANK3 unrelated”. It is a simple, clear dichotomy. But genetics are rarely simple, as we shall see.

The thinking at the time was from the perspective of genetic testing. If the test result includes an abnormality of the SHANK3 “coding region”, then it gets the designation PMS-SHANK3 related. All other test results would be PMS-SHANK3 unrelated.

The people who hammered out this definition (distinction) are all experts. They were fully aware of all the possible ways genes can contribute to a disorder. There are dissenters, also experts, who feel that if SHANK3 is not involved, then the name PMS should not be applied. The consensus paper explains that people with interstitial deletions (PMS-SHANK3 unrelated) do not appear to have a syndrome different from PMS. There is also no transition from PMS to some other syndrome as deletion sizes get larger, whether or not SHANK3 is involved. There is no clear evidence that PMS-SHANK3 unrelated is some other type of disorder. The key concern for dissenters is that genes other than SHANK3 do not contribute to the disorder in the same way that SHANK3 causes PMS.

However, what if we can show that other genes on chromosome 22 contribute to PMS through their impact on SHANK3, or their impact on the molecules that interact with SHANK3? If the genes of 22q13.3 (the site of interstitial deletions) have such direct impact on SHANK3, then perhaps the term “PMS-SHANK3 unrelated” is misleading. If the biology of PMS-SHANK3 unrelated is highly related to SHANK3, then a distinction may not be warranted. The latest evidence suggests there are at least six genes that impact SHANK3 (and partner molecules), and these genes contribute to PMS when deleted. I would call these genes SHANK3 related.

The six genes and the details of how they interact with SHANK3 are discussed in a recent paper. The paper is open source (anyone can read it), but very technical. Full disclosure: I helped write the paper, so this blog post is undeniably biased. People who have read my previous blog posts should know that I have long been a proponent of looking closely at the many genes of PMS. The paper is a review of work done by scientists primarily between 2017 and 2024. Science is a continuous process and during this period enough information came to light to explain the tight relationships between six genes (PLXNB2, BRD1, CELSR1, PHF21B, SULT4A1, TCF20) and SHANK3. The first two genes are physically close to SHANK3 on chromosome 22, and thus are deleted in most deletions that include SHANK3. The remaining four genes are fairly evenly spaced across the last 4 Mb (megabases) of the PMS region of chromosome 22.

I have written blog posts about all of these genes at one time or another. I identified PLXNB2 and PHF21B in Why PMS is worse for people with larger deletions, and PMS Gene PHF21B is critical for normal brain development. I wrote about TCF20, a gene that has long been associated with intellectual disability (TCF20 may explain why some big deletions are worse than others). I have flagged the potential importance of BRD1 in several blogs (e.g., Regression and psychiatric dysfunction in PMS). CELSR1 was highlighted in CELSR1: Do some people with PMS have more fragile brains? SULT4A1 has also been on the radar for some time: New science: SULT4A1, oxidative stress and mitochondria disorder. My message has always been that genes important to PMS will emerge once there was sufficient evidence to critically explain why larger deletions have greater impact. SHANK3 has always been the most important gene and it has been the most intensely studied, but from the beginning of PMS research, it was never the only important gene. By “important” I mean important to families.

All seven genes (including SHANK3) impact brain development and all are involved with the process of inflammation. In this case inflammation includes “cellular stress”, “mitochondrial function”, and “recovery from injury”. These are all related processes, and all known to exacerbate PMS. So, impact on early development and response to stress and injury are features common to all of the genes.

Three of the genes (BRD1, PHF21B and TCF20) regulate what other genes do in the brain. For example, all three regulate the activity at synapses, the part of neurons that SHANK3 regulates. In addition, two more genes (PLXNB2 and SULT4A1) are directly involved in synaptic function. In fact, SULT4A1 not only regulates the same glutamate receptor as SHANK3, but it also regulates breakdown of SHANK3 at the synapse.

At this point it should be clear why PMS-SHANK3 unrelated may not really be unrelated to SHANK3. The six genes listed above join SHANK3 in shaping the development of the brain, the response of the brain to insults, and the operation of the synapses – the most important role of SHANK3 in the brain. It should also be clear why people with interstitial deletions typically have symptoms consistent with other cases of PMS. Likewise, it should also be clear why people with larger and larger deletions tend to have more severe PMS (see: Why PMS is worse for people with larger deletions).

The very close association between SHANK3 and at least six other PMS genes has a number of ramifications. First, while the distinction between SHANK3-related and -unrelated is sensible from the point of view of genetic testing, it may not be a good way to think about PMS as a disorder. Second, treatments that target SHANK3 will likely miss other relevant genes that tightly influence SHANK3, and thus may not produce very satisfying results in patients with chromosomal deletions. One way to think of the problem is that SHANK3 treatment in people with deletions essentially expands the number of PMS patients with interstitial deletions. It may be difficult to distinguish between a treatment that is not effective for SHANK3, and a treatment that is not effective because of other genes. Finally, when we are searching for an effective treatment for SHANK3 haploinsufficiency, maybe we should also look at treatments for one or more of the other genes. We may be missing out on developing additional valuable treatments.

arm22q13

Why PMS is worse for people with larger deletions

David uses an adaptive switch to compensate for his poor coordination

Originally posted 27 December 2022

It is colder during wintertime

No one in the continental US would say that the autumn is colder than the winter. There are some days during the autumn that are colder than some days during the winter. But, winter is colder than autumn. Phelan-McDermid syndrome (PMS) is much the same. After many years of study, the scientific consensus is clear. Larger deletions come with greater loss of function, even though there are some cases of small deletions having greater impact than some cases of larger ones. The problems with larger deletions come in many forms. In the most recent study, people with very small deletions or only SHANK3 variants had “fewer delayed developmental milestones and higher cognitive ability”. People with larger deletions were “more likely to have a variety of medical features, including renal abnormalities, spine abnormalities, and ataxic gait” [1]. These same larger deletions greatly disrupt the genetics and profile of immune cells in the blood [2]. The whole genetic profile and metabolism of people with deletions greater than 1 Mb is off kilter: “DNA methylation epi-signature showed significantly different metabolomic profiles indicating evidence of two … distinct clinical subtypes of Phelan-McDermid syndrome” [3]. These new studies have only served to validate what was originally shown nearly 10 years ago [4].

What do scientist look for?

While scientists affiliated with the Phelan-McDermid syndrome community have focused largely on the SHANK3 gene, many scientists outside this community have been quietly advancing our understanding of other critical genes of PMS. I have written blogs on some of these important genes, including CELSR1, SULT4A1, TCF20, PHF21B, and others. When it comes to the core symptoms of PMS, which are all associated with brain dysfunction, three aspects of a gene are important to understand. First, what problems are caused by losing the gene in humans and model animals? Second, what biological impact does losing a gene have during early (pre-natal) development? Third, what role does the gene play in the adult human/animal?

Pleiotropy

When the same gene can impact a person in multiple ways, that is called pleiotropy. For example, losing SHANK3 can cause multiple problems, including low muscle tone, slower or less reaction to pain, GI disturbances, psychiatric illness, and regression. SHANK3 is said to be a pleiotropic gene. Multiple problems can arise from a single gene that gets used at different times in different places. When we think about different times we usually think about how a gene is used during prenatal development, a process that occurs only once in life, and how the gene gets used after birth. Prenatal development lays down the architecture of the brain. In humans, development of the brain continues into the 20s, but the most rapid changes occur before birth. Rapid, but slower changes continue during the first 3 years of life. Things like movement and language rapidly develop in the first few years. But, these changes depend upon the framework laid down in the uterus.

SHANK3 is active during prenatal development. It is involved with axon guidance, the process that neurons use to find each other during development [5] and other processes in early development [6-10]. These processes are distinct from how SHANK3 contributes to synaptic function in the adult brain.

What has come to light in the past few years is the pleiotropy of other PMS genes that contribute both during brain development and in the adult brain.

PLXNB2

One gene that is getting a lot of research attention is PLXNB2. From 2007 through 2014 a series of papers describe how Plexin-B2, the protein produced by the Plxnb2 gene as studied in mice, is critical for normal brain development [11-13].  In 2017 it was shown that the same protein is involved in pain sensation [14]. In the past few years the research has been extended. Loss of Plexin-B2 interferes with normal learning and memory in the adult mouse brain [15]. The role of Plexin-B2 parallels and overlaps the role of Shank3. Both are used at excitatory synapses in the brain. Both are involved in pain pathways in the spinal cord. In humans, these two genes are very near each other on chromosome 22, so almost all people who have PMS from a deletion are missing both genes. We can speculate that these genes exacerbate the damaging influence of each other.

What makes Plexin-B2 especially interesting for PMS is that it affects the brain circuit essential for remembering danger. Specifically, loss of Plexin-B2 in mice interferes with “conditioned fear recall”. That is, the mice will learn to recognize a warning tone when the tone signals a brief foot shock, but 2 days later they have forgotten the meaning of the warning tone. The neurons involved fail to form adequate synapses 2 to 3 days after the training period [15]. These are some of the same synapses where Shank3 operates. When I read this I nearly fell off my chair. (You could say I was shocked.) If there is one thing that my son, David does poorly, it is learning about danger. It took years to explain to David what it means when something is (dangerously) hot. It took equally long to warn him about stepping over an obstacle or curb. We often suppose that our children do not have the same sensitivity to pain as most people. That may be true, but we also might be misled by their inability to incorporate painful or fearful experience into memory.

PHF21B

Larger deletions of chromosome 22 can disrupt the PHF21B gene. This blog is about genes that have a role both in development and adult function. I have already written a blog on how PHF21B is critical for normal brain development. Now there is new research showing that PHF21B regulates synapses during “social” learning [16]. Social learning in mice is when a mouse can distinguish between a new mouse and one that has visited the cage before. Not distinguishing between a familiar mouse and a stranger is a serious inability. The Phf21b protein normally triggers parts of the DNA to build a memory after spending time with a new animal. In animals missing about half of their Phf21b protein, the ability to remember a fellow mouse (“conspecific”) is disrupted. The circuitry that is disrupted involves the Shank proteins. Thus, PHF21B is another gene that is important for both development and adult function, and is intimately associated with SHANK3.

CCDC134

Another gene that has received recent attention is CCDC134. To be honest, I did not pay much attention to this gene until recently. The gene is lost only with the very largest deletions (> 9 Mb), so loss of this gene is rare. CCDC134 had long been suspected to be essential for normal brain development, but mice missing CCDC134 died before they were old enough for behavioral studies. This past year a group in China produced a mouse that was missing Ccdc134 protein only in the cerebellum of the brain [17]. The cerebellum is well known as a critical area for motor coordination. Making a custom knockout mouse did the trick. Mice missing only Ccdc134 and only from the cerebellum develop malformation of the cerebellar Purkinje and granule cells. Mice with these deficits had problems with grip strength, motor coordination and motor learning. If you know a PMS child with a very large deletion who has serious difficulties standing and walking, it could be largely because of this gene. Most people with PMS have problems with the cerebellum. There is evidence that loss of Shank3 impacts the cerebellum in mice [9], but that more prominent malformations occur with deletions in people [18]. Now we know the largest deletions likely add substantially to the problem.

It can get very cold in winter

Larger and larger deletions have greater and greater impact on people with PMS. Even people with interstitial deletions that do not disrupt SHANK3 can have PMS. The combined developmental and adult functions of missing PMS genes, and their overlap with the function of SHANK3 at the synapse, conspire to make larger PMS deletions more detrimental in multiple ways. This blog is about recent finding in 3 PMS genes. Previous blogs have discussed other genes that impact PMS. As science advances we will learn more.

My parents never lived in a warm climate. One December my mother came to visit me at school in Atlanta, GA.  It was an unexpected 80 degrees that day. She asked, “Is it always this warm?” I said, “Yes.” The truth was, a year prior Atlanta had been blanketed with an equally rare snowstorm. Everyone in Atlanta knows it is colder in the winter, even though there can be the rare warm day. At this point, it is quite clear that larger [terminal] deletions of 22q13 are more impactful than smaller deletions or SHANK3 variants. PMS is more impactful than many other neurodevelopmental disorders. There will always be exceptions, but we must fully understand the rule before we can explain the exceptions.

arm22q13

Papers referenced

1.            Levy, T., et al., Strong evidence for genotype-phenotype correlations in Phelan-McDermid syndrome: results from the developmental synaptopathies consortium. Hum Mol Genet, 2022. 31(4): p. 625-637.

2.            Breen, M.S., et al., Large 22q13.3 deletions perturb peripheral transcriptomic and metabolomic profiles in Phelan-McDermid syndrome. HGG Adv, 2023. 4(1): p. 100145.

3.            Schenkel, L.C., et al., DNA methylation epi-signature is associated with two molecularly and phenotypically distinct clinical subtypes of Phelan-McDermid syndrome. Clin Epigenetics, 2021. 13(1): p. 2.

4.            Sarasua, S.M., et al., Clinical and genomic evaluation of 201 patients with Phelan-McDermid syndrome. Hum Genet, 2014. 133(7): p. 847-59.

5.            Halbedl, S., et al., Shank3 is localized in axons and presynaptic specializations of developing hippocampal neurons and involved in the modulation of NMDA receptor levels at axon terminals. J Neurochem, 2016. 137(1): p. 26-32.

6.            Liu, X., et al., SHANK family on stem cell fate and development. Cell Death Dis, 2022. 13(10): p. 880.

7.            Pillerova, M., et al., Neuromotor Development in the Shank3 Mouse Model of Autism Spectrum Disorder. Brain Sci, 2022. 12(7).

8.            Panagiotakos, G. and S.P. Pasca, A matter of space and time: Emerging roles of disease-associated proteins in neural development. Neuron, 2022. 110(2): p. 195-208.

9.            Otazu, G.H., et al., Neurodevelopmental malformations of the cerebellum and neocortex in the Shank3 and Cntnap2 mouse models of autism. Neurosci Lett, 2021. 765: p. 136257.

10.          Mossa, A., et al., Developmental impaired Akt signaling in the Shank1 and Shank3 double knock-out mice. Mol Psychiatry, 2021. 26(6): p. 1928-1944.

11.          Deng, S., et al., Plexin-B2, but not Plexin-B1, critically modulates neuronal migration and patterning of the developing nervous system in vivo. J Neurosci, 2007. 27(23): p. 6333-47.

12.          Vodrazka, P., et al., The semaphorin 4D-plexin-B signalling complex regulates dendritic and axonal complexity in developing neurons via diverse pathways. Eur J Neurosci, 2009. 30(7): p. 1193-208.

13.          Laht, P., et al., B-plexins control microtubule dynamics and dendrite morphology of hippocampal neurons. Exp Cell Res, 2014. 326(1): p. 174-84.

14.          Paldy, E., et al., Semaphorin 4C Plexin-B2 signaling in peripheral sensory neurons is pronociceptive in a model of inflammatory pain. Nat Commun, 2017. 8(1): p. 176.

15.          Simonetti, M., et al., The impact of Semaphorin 4C/Plexin-B2 signaling on fear memory via remodeling of neuronal and synaptic morphology. Mol Psychiatry, 2021. 26(4): p. 1376-1398.

16.          Chin, E.W., et al., The epigenetic reader PHF21B modulates murine social memory and synaptic plasticity-related genes. JCI Insight, 2022. 7(14).

17.          Yin, S., et al., Ccdc134 deficiency impairs cerebellar development and motor coordination. Genes Brain Behav, 2021. 20(7): p. e12763.

18.          Aldinger, K.A., et al., Cerebellar and posterior fossa malformations in patients with autism-associated chromosome 22q13 terminal deletion. Am J Med Genet A, 2013. 161A(1): p. 131-6.

PMS Gene PHF21B is critical for normal brain development

In spite of his poor motor function, David loves walking in the woods.

Originally posted 11 December 2021
Updated 14 June 2022

Multiple studies of Phelan-McDermid syndrome (PMS) have demonstrated worse intellectual outcomes with larger chromosome deletions. Genetic studies on large populations of people with intellectual disabilities show that the measure called “pLI” is useful for predicting IQ as a function of deletion size (see PMS, IQ and why interstitial deletions matter). The high pLI PMS genes are the most import genes of 22q13.3, the chromosome site of PMS problems (see Which PMS genes are most important?). In the past I have written about which PMS genes might cause the structural abnormalities in brain MRIs of PMS patients (see Which genes cause brain abnormalities in Phelan McDermid syndrome?). But, just this past year, scientists have discovered that a high pLI PMS gene is vital to the proper microstructure of the brain. Based on a series of very careful studies in mouse embryos, this group of researchers uncovered the special role of PMS gene, Phf21b. The high pLI of Phf21b (0.987 out of a possible 1.00) already tells us that it is an important PMS gene. The scientific team now tells us why it is so important in their 20+ page scientific paper (Phf21b imprints the spatiotemporal epigenetic switch essential for neural stem cell differentiation.) They solved the mystery of Phf21b and the whole PMS community benefits from their work. The paper is difficult reading, but open to the public.

To understand the new research we need to take a moment to review how our brains develop in the uterus. PMS is a neurodevelopmental disorder. That is, it is a disorder of brain development. Our brains assemble into complex structures with even more complex synaptic connections as they grow. The higher centers for thought, precise movement and proper interpretation of our senses are located in the cortex. The cortex is the part of the brain that we can see in photographs of the human brain. It is the most developed and expanded part of the brain that humans have.

Part of Figure 8
Amitava Basu et al. Genes Dev. 2020;34:1190-1209
© 2020 Basu et al.; Published by Cold Spring Harbor Laboratory Press

I have borrowed part of figure 8C from the paper on Phf21b to show the developing cortex in a mouse. It shows that neurons (the specialized brain cells responsible for thinking and acting) arise as general purpose cells from the ventricular zone (VZ, magenta color) and subventricular zone (SVZ, green color). Together, these are called progenitor cells. They are especially good at reproducing at a high rate to fill the areas above the SVZ. The rapidly growing and reproducing cells actually climb their way into final position. I found this video on the Internet that shows the process (Neurogenesis Animation). The rapid growth is shown 2:39 (min:sec) into the animation. This process creates about 100,000 cells per mm³ in the developing mouse. The process is driven by a large group of specialized genes that fuel the rapid growth and reproduction. These genes are collectively called, cell cycle genes. At a certain very crucial moment in the life of a progenitor cell, the cell stops growing and reproducing. The cell switches from being a progenitor to a neuron, and then spends the rest of its life as a neuron. Phf21b is the gene that tells the progenitor cells when to switch.

How does one gene stop the frenetic activity of progenitor cells? How does it switch off the many cell cycle genes involved in rapid growth and reproduction? The scientists spent some time figuring this out. Phf21b protein is not produced much in progenitor cells until the proper moment. Then, Phf21b protein production suddenly increases. It goes into the nucleus and searches out the precise spot in the DNA that regulates each cell cycle gene of the cortex. Phf21b is an epigenetic switch. It wraps around the regulatory site of each growth gene, and assembles a toolkit of proteins to shut that gene off.

Right side of Figure 8C. See credits (above)

Here is the right half of figure 8C from the paper. The magenta neural progenitor (NPC) cell on the left gets transformed into a neuron (blue cell on the right) by the action of Phf21b (green molecule in the middle). Through several steps involving other molecules (Rcor1, Hdac2, and Lsd1), Phf21b acts as a precision switch. It turns off just the cell cycle genes that promote rapid cortical growth. Phf21b literally hits the brakes on growth to trigger the creation of neurons in the cortex of mammals.

The experiments used by the authors to figure out this exquisite mechanism were complicated and, in some cases, exceedingly delicate. The team of nine researchers represented six institutes in three different countries. Among the many experiments undertaken, they showed the impact of disrupting this mechanism. With insufficient Phf21b, the normal pattern of neurons in the cortex is disrupted.

It has recently been shown that improper regulation of proliferation, both too little or too much, is associated with autism.

In spite of the groundbreaking work, certain questions remain. First, what aspects of the PMS disorder (what phenotypic characteristics) are driven by loss of PHF21B in humans? One hint comes from a paper published in 2014 by Disciglio and colleagues. It describes nine patients with interstitial deletions of 22q13. Although these authors argued that they discovered a new syndrome, the phenotype readily overlaps with PMS, with intellectual disability and speech delay being the two central traits. Many people (including myself) see these simply as cases of PMS (see The four types of Phelan McDermid syndrome). Regardless, of the nine patients in the study, eight are missing PHF21B and the exception (called patient #4) had a deletion that was very near PHF21B (within 84 Kb). That patient was arguably the least affected individual, as well. Thus, PHF21B may have been a major driver of the manifestations (symptoms) seen in these patients. More cases of PMS with interstitial deletions (PMS type 2) are needed to confirm this observation. What would be especially useful would be to find patients with pathological variants of PHF21B. As more genetic testing is done in newborns, especially whole exome or whole genome sequencing, there will be more cases. Why should we expect a growth in cases? We already know the two critical characteristics: 1) loss of one functional copy of PHF21B is pathological (the pLI is greater than 0.9) and 2) loss of one functional copy of PHF21B is compatible with life (there are patients with PMS type 1 that have deletions that include PHF21B because their deletions are greater than 5.81 Mb). About 40% of all identified individuals with PMS are missing this gene (see Understanding deletion size). So, we know these patients are common in PMS.

The second unanswered question is what, exactly, does the loss of PHF21B do to the human cortex once development is complete. The human primate cortex is far more complex than the rodent cortex (see Mashiko et al 2012, Chen et al 2016, Preuss and Wise 2022). Still, the mouse cortex can provide insights into the dysplasia likely to occur in humans. One research direction would be to compare the histology of rodent cortex in a Phf21b knockout mouse with post-mortem or resected brain tissue from a PMS patient with a terminal deletion greater than 5.81 Mb. Perhaps the answer can come from simply comparing cortex samples from patients with terminal deletions smaller than 5.81 Mb and larger. Brain donations to brain banks by patients with rare diseases are a rare and important resource for this type of work.

The third unanswered question is how can loss of a PHF21B gene be rectified? I will not speculate on what approaches might be used to compensate for the loss of a gene so crucial to neurodevelopment. How to fix the loss of neurodevelopment genes is a central problem for all neurodevelopmental disorders. You can try to fix the gene or you can try to compensate for its absence. Perhaps this topic will be the subject of a future blog.

arm22q13