3 June 2013. Lifelong schizophrenia-like overexpression of neuregulin 1 (Nrg1) can be reversed in mice, according to a study published May 22 in Neuron. Researchers led by Lin Mei of the Medical College of Georgia in Augusta report that when they restored Nrg1 to normal levels in adult mice, a collection of abnormalities went away, including neurotransmitter imbalances, impairments in working memory, sensory gating, and social interaction. Nrg1's effects on glutamate signaling relied on LIM domain kinase 1 (LIMK1), a molecule implicated in other neurodevelopmental disorders. The results show that the brain, at least in mice, remains pliable enough in adulthood to take advantage of changes in molecular expression.
Finding a neural defect and fixing it might seem like a simple enough strategy, but it bumps against the declining brain plasticity that comes with age. Even if schizophrenia were caused by a single molecular mishap, and researchers could find ways to fix it in adults, the aging brain might not be able to make use of the fix. Moreover, a potentially tangled history of compensatory processes launched during brain development would seem to dim the chances for a simple fix of what is "broken."
For Nrg1 in mice, however, this simple strategy appears to be successful. This could be good news for schizophrenia therapeutics if Nrg1 or other molecules work similarly in humans. Nrg1 is a familiar suspect for schizophrenia and is involved in many aspects of brain development and synaptic signaling (Mei et al., 2008), but its precise role in the disorder remains unclear. While candidate gene studies have associated it with schizophrenia (e.g., Stefansson et al., 2002), Nrg1 has not been highlighted by recent genomewide association studies (see SRF related news story). Postmortem studies have found either too much or too little Nrg1 expression in schizophrenia (e.g., Bertram et al., 2007; Law et al., 2006), and re-creating these abnormalities in mice perturbs neurotransmitter signaling in ways that jibe with the hypotheses about brain pathology in schizophrenia (see SRF related news story).
Too much of a good thing
First authors Dong-Min Yin and Yong-Jun Chen and colleagues follow the too-much Nrg1 lead and generated their own mice that overexpressed Nrg1 in excitatory neurons of the prefrontal cortex and hippocampus. This resulted in Nrg1 levels 50-100 percent higher than those of control mice, apparent at birth and maintained into adulthood. Along with the overexpression came a collection of behavioral anomalies compared to control mice: a decrease in prepulse inhibition, a common measure of how sensory information can shape a later response that is reduced in schizophrenia; less time spent with another mouse, akin to the social withdrawal of schizophrenia; deficits in working memory and spatial memory in maze tests, similar to cognitive difficulties in schizophrenia; and hyperactivity, which may correspond to psychomotor agitation seen in schizophrenia.
Inside the Nrg1-overexpressing brains, the researchers documented synaptic changes as well. In the hippocampus, excitatory synapses released less glutamate while maintaining a similar density of excitatory synapses compared to controls. Inhibitory synapses were also sluggish, though this seemed to stem from a decrease in γ-aminobutyric acid A (GABA-A) receptor density on the postsynaptic side. These findings are consistent with ideas about underactive glutamate and GABAergic signaling in schizophrenia (see SRF Hypothesis).
The fix is in
The researchers had designed their mice with a second transgene that allowed them to turn off the Nrg1 transgene with doxycycline, thereby introducing normal levels of Nrg1 expression. When Yin, Chen, and colleagues turned off the Nrg1 transgene in eight-week-old adults, Nrg1 levels were not different from control mice also treated with doxycycline, and, strikingly, no brain or behavior anomalies were detected. The doxycycline-treated Nrg1 mice matched control levels of prepulse inhibition, social interaction, working and spatial memory, and locomotor activity. Similarly, measures of glutamatergic and GABAergic signaling in the hippocampus of the doxycycline-treated Nrg1 mice were no different from those measured in controls. The findings suggest that even a long-standing (for mice) perturbation in Nrg1 expression can be readily reversed. Conversely, the researchers found that introducing Nrg1 overexpression in adult mice led to similar deficits, which argues that Nrg1 overexpression is sufficient to cause these impairments.
Nrg1 overexpression did not seem to act through Nrg1’s typical receptor partner, ErbB4, to bring about the glutamatergic impairments, however. The glutamate deficits remained in hippocampal slices made from Nrg1-overexpressing mice that limited ErbB4 activity, either by acutely inhibiting it with a drug, or by breeding Nrg1-overexpressing mice without one copy of the ErbB4 gene. Instead, the researchers found a role for LIMK1, which encodes a kinase that interacts with Nrg1’s intracellular side. When the researchers inhibited LIMK1 activity in hippocampal slices made from Nrg1-overexpressing mice, glutamate signaling remained normal. LIMK1 is one of 25 genes within the critical region deleted in Williams syndrome, and duplicated in some cases of autism (Sanders et al., 2011) and schizophrenia (Kirov et al., 2012).
The results bring to mind the rescue of Fragile X-associated phenotypes in adult mouse models, which first showed that long-standing disruptions to the brain were not necessarily permanent (Dölen et al., 2007). The new results point to an adult brain that remains sensitive to changes in Nrg1, and offer up a potential therapeutic strategy for schizophrenia.—Michele Solis.
Yin DM, Chen YJ, Lu YH, Bean JC, Sathyamurthy A, Shen C, Liu X, Lin TW, Smith CA, Xiong WC, Mei L. Reversal of Behavioral Deficits and Synaptic Dysfunction in Mice Overexpressing Neuregulin 1. Neuron. 2013 May 22; 78: 644-657. Abstract