In a study published this month in the journal, Science, Tyzio et al. (2014) report some exciting findings which help fill in some of the gaps on oxytocin research in autism. Although in my own genomics research it has been clear that mutations in genes for the oxytocin (OXTR) and vassopressin (AVPR1A) receptors are strong candidates for autism risk, this mechanism of action has been vague at best.
You see, most of the core candidate genes for autism seem to code for proteins involved in differentiation, cytoskeletal structure, transcription factors, calcium signaling, cation transport, chromatin remodeling, and various neurotransmitters such as dopamine, GABA, glutamine, and serotonin. But, at least in my mind, while various catecholamine, glutamatergic, and GABAergic pathways have be clearly linked with neuronal growth and development, oxytocin’s and vassopressin’s links have been fuzzier, especially since so much of the research has dealt with development within the postnatal periods. Although this new study doesn’t address the potential role of vassopressin in these autism rodent models, the role which oxytocin plays during labor is wonderfully illustrative as to how this hormone may drive aspects of development.
During early development and up into the neonatal time period, GABAergic interneurons, the neurons most closely associated with “inhibition” of associated glutamatergic neurons, are not actually inhibitory but are instead excitatory drivers of the developing brain. This is due to higher chloride concentrations within GABAergic cells. Up until that point, GABAergic cells produce only low levels of the chloride exporter, KCC2. In other words, they tend to build up a lot of intracellular chloride.
The illustration above shows an excitatory pyramidal neuron receiving input from different kinds of GABAergic interneurons. Image borrowed from here.
One of the great things that oxytocin does is to protect the brain from the acute trauma of labor by helping to instigate that initial shift of GABA interneurons from excitatory to inhibitory by acutely lowering intracellular chloride levels. It’s uncertain how this is done, although there may be elements of calcium-mediated induction of chloride transport which has been reported in other oxytocin research, since the effects seem to be immediate and not solely a result of gene transcription . Interestingly, the diuretic, bumetanide, mimics oxytocin in some ways by blocking the NKCC1 chloride importer, helping to reduce excitatory activity of GABA interneurons and promoting their switch to inhibitors.
Tyzio et al. studied two rodent models of autism, one a valproic acid (VPA) rat model and the other a Fragile X (FXS) mouse model. Reduction of GABA-A receptor (GABA-AR) levels are associated with maturation of the GABAergic system. In both of these animal models, although controls showed adult GABA-AR levels by postnatal days 15-30 (P15-30), GABA-AR levels in VPA and FXS mice were comparatively elevated and KCC2 levels downregulated. In addition, neurons taken from each of these models on the day of birth (P0) and which were exposed to either oxytocin or bumetanide showed significant decreases in intracellular chloride content and also a reduction in GABA-AR levels, promising a possible means for intervention.
Most interestingly, when the researchers applied a GABA-AR agonist (isoguvacine) to VPA and FXS neurons, rather than inhibiting or not affecting the spike frequency, spike frequency increased in both newborn and adolescent rodents. The researchers also found that at P0, VPA and FXS hipocampus exhibited increased frequencies of glutamatergic spontaneous excitatory postsynaptic currents (sEPSCs). What this suggests is that in some human forms of autism, it is possible that maturation of the inhibitory GABAergic system could be delayed or even permanently altered, driving increased neuronal excitation and potentially leading to many of the symptoms associated with the conditions. The researchers also report that bumetanide application helped to control sEPSCs, suggesting that the seat of some of the dysfunction does indeed lie within the GABAergic system in these mouse models. I should however mention in regards to the delay of maturation of the GABAergic system in autism, that successful treatment of autistic symptoms using low-frequency rTMS which specifically targets interneurons suggests, at least in high-functioning individuals, that the GABAergic system does indeed provide inhibition rather than excitation. Were it the latter, low-frequency rTMS would instead promote seizures and probably a worsening of symptoms. So if the the inhibitory nature of the GABAergic system is maturationally delayed in autism, the question is “For how long?”
The researchers went further to test whether bumetanide could help reverse some of the effects seen in VPA and FXS rodents: by giving pregnant mothers bumetanide 1 day before delivery, the researchers were able to restore GABA-AR levels, reverse the effects of isoguvacine, and significantly reduce sEPSCs in their pups. –Although I should point out that levels of sEPSCs were not completely normalized. In contrast, blocking oxytocin receptors the day prior to delivery produced symptoms reminiscent of the VPA and FXS animal models within this study. In addition, these same mice exhibited behaviors similar to the untreated autism animal models.
This has been a fascinating and thorough piece of work to read. It is apparent that the researchers were extremely thorough and clear with their methodological designs and testing and have certainly addressed quite a few very important questions relevant to autism, epilepsy, and neurodevelopment in general.
However, I would also like to take the opportunity to remind the reader that, while it’s tempting to say “This is how oxytocin affects autism risk,” it is necessary to remember that oxytocin likely has multiple functions within the central nervous system. Given its differentiative effects on various stem cell populations, it’s quite possible, perhaps even likely, that it plays roles in neurogenesis, aborization, and synaptogenesis, in addition to its neonatal and postnatal roles. Both the VPA and FXS rodent models have shown extensive effects on neuronal populations including structural changes, not just the dysregulation of the excitatory-inhibitory balance. So while oxytocin or bumetanide application might prove useful treatments for some of the physiological changes in some people with autism, it can’t undo some of the structural characteristics of the condition, especially those which are not subject to plasticity. That doesn’t mean that some problem symptoms can’t be alleviated, but it is best to view these results cautiously, taking into account the breadth of research that has been performed to date. In other words, this is an incredible study but very rarely is any single piece of work a cure-all/explain-all for a condition as complex as heterogeneous autism.