Home > News & Opinion > News > 2013 > Drug fixes cellular defects in autism-related disorder

Drug fixes cellular defects in autism-related disorder

Popularity tracker
Virginia Hughes
2 December 2013

Culture shock: Neurons derived from people with Phelan-McDermid syndrome (red) show significantly reduced excitatory signals compared with controls (green).

A new stem-cell model of Phelan-McDermid syndrome points to a possible treatment for the rare autism-related disorder, according to a study published 16 October in Nature1.

Phelan-McDermid syndrome is characterized by intellectual disability, autism, seizures, sleep trouble and motor problems. It is caused by glitches in the 22q13 chromosomal region, which includes SHANK3, a well-known autism candidate gene.

Several mouse models lacking SHANK3 have shown relatively subtle abnormalities, however. Each of these models has a different set of brain and behavioral defects, which researchers are only beginning to understand.

"We like mice, and we think they provide interesting insights into disease, but at the same time, we know that they are not always really great predictors of the human disease," says Ricardo Dolmetsch, global head of neuroscience at the Novartis Institutes for Biomedical Research in Cambridge, Massachusetts. "In the case of Phelan-McDermid, that's definitely true."

Dolmetsch and his colleagues used chemical soups to turn skin cells from individuals with Phelan-McDermid syndrome into neurons. Compared with controls, these neurons have fewer synapses, or junctions between neurons, and dampened transmission of electrical messages between them, the study found.

When the abnormal cells are exposed to a hormone called insulin-like growth factor 1 (IGF1), however, their synaptic signaling is restored to normal levels.

The results agree with those from a previous study reporting that IGF1 improves symptoms in mice lacking SHANK3. They also lend support to a small clinical trial testing IGF1 on children with Phelan-McDermid syndrome.

"I'm optimistic, but I don't know if IGF1 is going to be the answer," says Catalina Betancur, director of research at the Pathophysiology of Central Nervous System Disorders unit at INSERM in Paris, who was not involved in the study.

What she's more optimistic about, she adds, is that the new cellular model of Phelan-McDermid syndrome will allow researchers to screen many more drugs.

"The promise of being able to screen a lot of pharmacological compounds to identify a potential treatment is a huge thing," Betancur says. "So there is hope, where a few years ago there was absolutely none."

SHANK3 surprises:

Dolmetsch's team took skin samples from two people with Phelan-McDermid syndrome and used chemicals to reprogram them into stem cells. Using a different chemical broth, the researchers then turned these so-called induced pluripotent stem (iPS) cells into neurons.

The researchers grew these abnormal neurons with control cells in the same culture dish, and used fluorescent labels to distinguish one set from the other.

Phelan-McDermid neurons show significantly reduced signaling at excitatory synapses, which boost the electrical signals between cells, but no differences in inhibitory synapses, which dampen those signals, the study found. Many other studies have also reported an imbalance of excitatory and inhibitory signaling in autism.

The researchers tried two ways to rescue these defects. First, they used a virus to deliver SHANK3 to the cells, which effectively reversed the synaptic deficits. This is an important finding because SHANK3 is just one of about 25 genes in the Phelan-McDermid region, and researchers have debated whether other genes are also key players in the syndrome.

"It’s great to be able to demonstrate that all the synaptic deficits that we’re seeing in those cells are in fact due to SHANK3," Betancur says.

The second rescue method used IGF1. Before this study, researchers had assumed that IGF1 works by boosting SHANK3 expression at synapses. But Dolmetsch's team found exactly the opposite: IGF1 dramatically suppresses the expression of SHANK3 in both the Phelan-McDermid neurons and controls.

"We were very puzzled by this," Dolmetsch says. "It eliminates expression of SHANK3, and yet at the same time it's rescuing synaptic transmission. How can that be?"

In mice, most synapses express SHANK3 and its sister proteins, SHANK1 and SHANK2. But Dolmetsch's team discovered that's not the case in people.

Looking in samples of human fetal brain tissue, the researchers found that some synapses have SHANK3 protein and others have SHANK1 and SHANK2. "It turns out that these SHANK3-containing synapses seem to be less mature," Dolmetsch says.

He hypothesizes that SHANK3 is involved in the early formation of synapses, and that IGF1 is somehow involved in the maturation of synapses that don't contain SHANK3. That could explain how the drug boosts synaptic signaling in people with Phelan-McDermid.

"It's striking, and quite paradoxical" that IGF1 lowers the expression of SHANK3, says Thomas Bourgeron, director of the Human Genetics and Cognitive Functions Unit at the Institut Pasteur in Paris, who was not involved in the study. "It leaves more questions than answers, but I think this paper is a very good starting point."

Dolmetsch began his iPS studies about six years ago, when the technology was in its infancy. Since then he has collected skin samples from about 60 individuals with Phelan-McDermid syndrome, he says, and will use them to try to replicate the findings from this study.

His team has reprogrammed cells from individuals with other rare types of autism, including Timothy syndrome, DiGeorge syndrome, Dravet syndrome and deletions of chromosomal region 16p11.2. "We're very interested in doing drug screening in a very serious way," he says.

Bourgeron and his colleagues have also made iPS neurons from four people who carry SHANK3 mutations.

Bourgeron says he is optimistic about finding treatments with this approach because these individuals still have one working copy of the gene. "If we could just find a molecule that boosts a little bit this good copy, maybe we could restore some function," he says.

News and Opinion articles on SFARI.org are editorially independent of the Simons Foundation.


1. Shcheglovitov A. et al. Nature Epub ahead of print (2013) PubMed


Name: Sam
2 December 2013 - 3:47PM

How many of autism cases have Phelan-McDermid condition. What is the ratio here: 2 between every 100 autism cases.

They target what they have or what they know...which looks like not much

Name: richard at nasa
2 December 2013 - 11:01PM

Well Sam here is a news flash- a study released a few weeks ago by researchers at Duke found 15% of post mortem brain tissue from people with Autism had abnormalities in shank-3 proteins. Is that big enough for you? Not much? You are correct that they target what they know-and that leads to more and more. Just as the other shanks have now been implicated in autism-the time is now to find how the other ASD suspect genes play a role, answer what genetic or epigentic factors cause the 2% to become 15% and use the PMS model to investigate the other ASD suspect genes. If they keep hitting it in 15% increments....... The alternative is to keep searching the environment for causes- hm.... Vaccines are out- so we better call the lawyers to see who else we can sue. Science or Law-which one do you think will do more for our kids?

Name: Andrew R. Mitz, PhD
3 December 2013 - 4:04PM

"It leaves more questions than answers, but I think this paper is a very good starting point."

As usual, Thomas Bourgeron has it right. Alex Shcheglovitov and Ricardo Dolmetsch have brought preclinical research with iPS to a new and exciting level. This Nature paper is very dense and not easy to grasp without having lots of background in the field and spending lots of time mulling over the results. I respectfully disagree with two statements made in Virginia’s article, and wish to explain why based on the data in the paper. Here goes.

The first statement I disagree with is, “When the abnormal cells are exposed to a hormone called insulin-like growth factor 1 (IGF1), however, their synaptic signaling is restored to normal levels.” iPS neurons cannot be used to test for “normal synaptic signaling”. They can test for spontaneous events and membrane properties (Figures 1 and 2). This is valuable information and predicted a paucity of receptors (Figure 3). However, the most important functions of a synapse are long-term potentiation and long-term depression. Neuroscientists believe that these two mechanisms are essential for learning. The hallmark traits of Phelan-McDermid syndrome (PMS) are poor motor learning (motor problems, as Virginia puts it) and poor book learning (intellectual disability). This Nature paper showed that abnormally low synaptic currents could be boosted by IGF1. At the same time, IGF1 depressed SHANK3. SHANK3 is very involved in long term potentiation and long term depression. In fact, SHANK3 is crucial to the mTOR pathway that is the smoking gun in two other autism-related syndromes, tuberous sclerosis and fragile X. (See Virginia Hughes’s article of December 1, 2011 “Tuberous sclerosis, fragile X may be molecular opposites”). Thus, IGF1 could make patients more autistic, not less, by interfering with normal SHANK3 function. One of the more interesting results in this paper is that, although the patients are missing 50% of their SHANK3 genes (one of two is lost in PMS), there is only a 25% drop in SHANK3 (Figure 4C). Adding IGF1 makes things worse, not better. Dr. Dolmetsch rightly calls it a puzzle. His hypothesis about synaptic maturity is very speculative, at best. IGF1 reduces the amount of SHANK3 in his normal (“control”) iPS neurons, which is clearly pathological.

The second statement I disagree with is, "It’s great to be able to demonstrate that all the synaptic deficits that we’re seeing in those cells are in fact due to SHANK3" I’m not sure where this idea comes from, but the paper shows the exact opposite. Figure 4J shows that more than half (57%) of the neurons could not be rescued by adding SHANK3 back into the neurons. In those neurons, AMPA receptors increased, but the NMDA receptors current remain at zero. Thus, adding SHANK3 to PMS neurons made them worse by creating a AMPA/NMDA imbalance. The authors of the paper speculate that other SHANK3 isoforms might be to blame, but their own data (Figure 4b) does not support that suggestion. The obvious alternative is that another gene is preventing the neurons from functioning normally. What gene? The obvious candidate is MAPK8IP2, which is deleted in the iPS neurons (Supplementary Figure 1). This gene has been shown to regulate the AMPA/NMDA balance. Thus, the paper clearly demonstrates that SHANK3 is not the main cause of synaptic deficits (43% versus 57%) and that another gene, perhaps MAPK8IP2 may be equally or even more important. This is, perhaps, the most important take home message of the paper.

Thank you Thomas Bourgeron. This paper is a very good start, but it raises more questions than it answers. It certainly spotlights what is wrong with the “SHANK3 theory of PMS” and why IGF1 might make individuals with PMS worse off, rather than better.

Add a Comment

You can add a comment by filling out the form below. Plain text formatting.

Question: What is 10 + 4 ?
Your answer: