The commonly-used epilepsy drug, valproic acid (VPA), can have a highly beneficial effect on some babies born with spinal muscular atrophy (SMA), the number one genetic killer during early infancy. But in about two-thirds of such cases it is either damaging or simply has no effect. Now, for the first time, researchers have found a way to identify which patients are likely to respond well to VPA prior to starting treatment. Their results have major implications, not just for SMA patients, but for other conditions treated with the drug such as migraine and epilepsy, and may even provide the conditions for turning VPA non-responders into responders, the researchers say.

Dr. Lutz Garbes, from the Institute of Human Genetics, University of Cologne, Germany, will tell the annual conference of the European Society of Human Genetics on June 24 that he and his colleagues had analysed blood RNA samples from a small group of SMA patients who had been treated with VPA. They found, as expected, that only about one third of patients responded well. In an attempt to discover whether blood sampling was the most appropriate test method to use, they also looked at VPA response in another tissue -- fibroblasts (a type of skin cell). They found that the response in blood and in skin was the same in 60% of cases.

The researchers then generated pluripotent stem cells from fibroblasts of both a VPA responder and a non-responder, and differentiated them into GABAergic neurons (neurons that produce the amino acid GABA, the chief neurotransmitter in the mammalian nervous system). These neurons, when treated with VPA, exhibited a similar response to that previously found in blood and fibroblasts.

"This indicates for the first time that response to VPA is the same among blood and skin and suggests that monitoring blood for VPA therapy is indeed feasible in central nervous system diseases," says Dr. Garbes. "But, even more importantly, by using the SMA patients' fibroblasts we were able to identify a decisive factor in the suppression of the positive response to VPA treatment. Utilising transcriptome-wide microarray profiling*, we found that high levels of the fatty acid transporter protein CD36 are associated with the lack of positive response to treatment.

"The implications of this discovery are far-reaching. First, we have been able to prove that monitoring blood is a reliable method for doctors to determine response to VPA treatment in many central nervous system diseases, since our findings are not specific to SMA. Second, the identification of CD36 as the crucial factor in suppressing response to treatment provides a simple way of appraising whether a patient will respond to therapy before treatment starts. And third, in the long run we may find a way to target CD36 in order to be able to change a non-VPA responder into a responder."

Knowing that CD36 is a crucial factor here means that the current, potentially dangerous, 'trial and error' approach to VPA treatment is now obsolete, the researchers say. Screening of patients for CD36 prior to treatment would mean that only those who would respond positively to VPA would be given it. This is important because, in some cases, VPA can cause life-threatening side-effects such as impairment of liver, blood cell and pancreatic function, especially in those just starting the treatment. "But we still do not understand how CD36 suppresses response to VPA, only that it does so," says Dr. Garbes. "A greater understanding of its effects could also lead to the detection of even better targets to overcome the problem. "

In the case of SMA, VPA works by inhibiting enzymes called histone deacetylase (HDACs) which are involved in regulating the packaging of DNA. HDACs lead to a denser DNA packaging whereby protein production from genes is reduced. Other enzymes called histone acetyltransferases (HATs) lead to a more relaxed DNA structure, producing more protein. By inhibiting HDACs with VPA, the DNA packaging balance shifts towards the more relaxed structure and thus genes get activated and proteins produced. In SMA, the crucial gene is SMN2, a copy gene of the disease-determining gene SMN1. In healthy individuals, SMN1 is the major source of SMN protein, but SMN2 cannot fully compensate for the loss of SMN1 in SMA patients. By increasing SMN2 activity, it will produce more SMN protein and ameliorate the condition.

"Avoiding needless VPA treatment of non-responders would have a major effect on healthcare costs and improve quality of life for patients," Dr. Garbes will say. "Half of the babies born with SMA will die within two years, but the other half can live to twenty or even longer, so this is an important finding for them. Our findings may also help identify patients who are candidates for VPA treatment in many other diseases of the central nervous system, some of them very common.

"In the EU, approximately 550 SMA babies are born each year, and there are about 311,000 new cases of epilepsy per year. It is estimated that, in Europe, migraine affects up to 28% of people at some time in their lives. We are happy that we may have been able to contribute to the development of personalised medicine for so many people," he will conclude.

*A transcriptome-wide microarray profile provides a way of identifying all the genes that are differentially expressed in distinct cell populations or subtypes, allowing the effects of treatment to be monitored.