Home > News & Opinion > In Brief > 2011 > Clinical research: Blood test could recommend drug treatment

Clinical research: Blood test could recommend drug treatment

Popularity tracker
Jessica Wright
13 July 2011

Crystal ball: A blood test could help predict whether a child who has autism will respond to risperidone, a drug approved to treat the disorder.

Children who express a subset of genes — including one located in the autism-associated chromosomal region 16p11.2 — at higher or lower levels than normal benefit from treatment with the antipsychotic drug risperidone, according to a study published 7 June in the Pharmacogenomics Journal1.

Testing gene expression profiles using a simple blood test could help determine whether risperidone treatment will be beneficial for a child with autism, the researchers say.

A large meta-analysis published in May reports that risperidone is one of only two effective treatments for autism, alleviating aggression and repetitive behaviors. But it also has troubling side effects that limit its usefulness, including weight gain and repetitive involuntary movements.

Another study, published last October, suggests that certain individuals may be better able than others to tolerate risperidone and benefit from its effects. Researchers identified a genetic variant in P-glycoprotein — a protein that controls the passage of molecules through the cell membrane — that predicts a positive response to risperidone treatment. Another variant in cytochrome P450, which regulates how quickly drugs are metabolized, is associated with less-than-usual weight gain.

In the new study, researchers broadly analyzed gene expression in 42 children with autism before the children began taking risperidone. Of these children, nine showed the most improvement in behavior after eight weeks of risperidone treatment and eight showed the least, based on the irritability measure of the Aberrant Behavior Checklist.

The researchers looked for genes with opposite expression patterns in children with the highest and lowest levels of response. They found 89 genes that fit this pattern.

For example, GBP6, RABL5 and NFKBID are expressed at low levels in the high responders and at high levels in the low responders. RNF213 and RNF40 show the opposite pattern: They are expressed at high levels in children who respond well to risperidone treatment.

The level of expression of these five genes also correlates with the degree of improvement observed overall.

RNF40 codes for a protein that degrades SNAP25, also called GRIP1, an autism-associated protein that functions at the junctions between neurons. It is located in the chromosomal region 16p11.2. Deletions and duplications of another section within this region are associated with autism.

Correction: This article has been modified from the original. It has been changed to clarify the chromosomal location of the RNF40 gene.

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


  1. Lit L. et al. Pharmacogenomics J. Epub ahead of print (2011) PubMed


Name: Jon Brock
13 July 2011 - 6:02PM

In principle I think this is a great idea. If you're going to give drugs to kids with autism, at least make sure you're giving them to the subset of kids who would actually benefit.

But the worry is that, with so many genes to choose from, you'd probably expect 90-odd genes to differentiate between any two groups of 8 or 9 people - even if they were selected at random rather than according to drug response.

Unless I'm missing something, these results are pretty meaningless until someone does a replication study and shows that a large number of the same genes are identified again.

Name: Jessica Wright
13 July 2011 - 6:10PM

Jon, that is a very good point, which is why I think not much is made of the 89 genes as a whole in the paper.

Instead, the authors emphasize the five genes I mentioned because not only is their expression opposite between the two groups, but their expression levels correlate with severity for each of the 42 individuals studied.

RNF40 is also more likely to be relevant because it may have some functional relevance to autism.

Replication is of course important for any study like this.

Thanks so much for your continued careful reading of SFARI!

Add a Comment

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

Question: What is 10 + 4 ?
Your answer: