See Allison Curley's snapshots from the conference.
November 27, 2013. Dopamine neurons command the attention of a large swath of neuroscientists, and many of them have an ongoing interest in the role of dopamine in schizophrenia (see Current Hypothesis by Anissa Abi-Dargham). This interest was on full display at the annual meeting of the Society for Neuroscience in San Diego, California, and particularly at a packed lecture given by Anthony Grace of the University of Pittsburgh, Pennsylvania, early Sunday morning, November 11. Focusing on schizophrenia and depression, Grace discussed his work into how alterations to brain circuitry can rework the ventral tegmental area (VTA), known for its dopamine-containing neurons. For schizophrenia, Grace drew on the MAM model of schizophrenia that he developed with Holly Moore (Moore et al., 2006), which has gained a substantial following among neuroscientists interested in schizophrenia: Prenatal injections of methylazoxymethanol acetate (MAM), a DNA-methylating agent, give rise to schizophrenia-like signs in rodents, including thinned cortex, deficits in sensory-motor gating, learning, social interaction, and hyperactive dopamine signaling.
Grace has found that these MAM-induced changes stem from a hyperactive hippocampus (see SRF related news story). Several synapses away, this produces a slight increase in the number of VTA neurons that are active at rest, which could promote “aberrant salience”—a state hypothesized in schizophrenia in which faulty dopamine signals flag unimportant events as important and provide grist for delusions (Kapur, 2003). Grace argued that fixing the original problem in the circuit—which he blamed specifically on parvalbumin (PV) interneurons—may better calm psychotic symptoms than current antipsychotic drugs, which in his model only block the secondary effect of enhanced dopamine signaling. Grace also discussed newer data suggesting that a different pathway through the amygdala to the VTA may contribute to depression.
But the amygdala may participate in schizophrenia, too, and may even lie upstream from hippocampal pathology, according to Grace’s graduate student Yijuan Du, who gave a talk on Monday afternoon. In MAM-treated mice, Du reported that basolateral amygdala neurons projecting to the infralimbic prefrontal cortex (ilPFC), which also sends input to the hippocampus, had higher firing rates at rest than saline-treated mice. This heightened activation may disrupt hippocampus signals either through direct projections to the amygdala or indirectly through the ilPFC, given previous evidence that amygdala activation disrupts hippocampal interneurons (Berretta et al., 2001).
The enhanced VTA activity in the MAM model can be mimicked by a selective decrease in PV expression in the ventral hippocampus, according to a poster on Monday, November 11. Angela Boley, who works with Grace lab alumnus Daniel Lodge at the University of Texas Health Science Center in San Antonio, presented preliminary results showing that lentiviral knockdown of PV in the ventral hippocampus, without an obvious loss of interneurons (as monitored by expression of the GABA-making enzyme, GAD67), led to a slight elevation in resting activity in VTA neurons, as well as an increase in the number of spontaneously active VTA neurons compared to controls. Animals with this targeted PV knockdown also showed increased locomotion in response to amphetamine compared to controls, which is a measure of enhanced dopamine signaling also seen in MAM-treated animals. This suggests that PV knockdown in the ventral hippocampus is sufficient to recapitulate some of the defining features of the MAM model. Some of these features may also be inherited, apparently. In the same poster session, Stephanie Perez, also of Lodge’s lab, described how MAM-treated mice had offspring that also showed a decrease in PV expression in the ventral hippocampus, as well as an increase in the number of spontaneously active VTA cells. These effects were most pronounced in offspring whose father had been treated with MAM.
But the brain anomalies characterizing the MAM model do not consistently produce behavioral deficits, according to a poster on Sunday afternoon, November 10, from Nadia Malik of Eli Lilly, Windlesham, United Kingdom. She found that different litters of MAM-treated animals vary in their expression of behavioral deficits in reversal learning; for example, in one test, only three out of 19 litters showed an effect. Finding quick ways to prescreen the animals for behavioral deficits would therefore expedite the use of the model in pharma-scale research.
Diverse dopamine cells
The VTA contains a diverse population of neurons, however, which complicates interpretation of changes in its activity. This heterogeneity extends to the DA neurons therein, which may explain how DA neurons can be involved in so many different functions. On Tuesday morning, November 12, Stephan Lammel of Stanford University, Palo Alto, California, delved into this diversity in a symposium devoted to VTA. Lammel’s work finds that DA neurons in the lateral portions of the VTA differ from their medial counterparts in terms of their projection targets, their intrinsic currents, and synaptic properties. Similarly, Lammel has recently found that these regions receive different inputs, which ultimately determine behavioral responses (e.g., Lammel et al., 2012). In the following talk, Elyssa Margolis of the University of California, San Francisco, described differences between DA neurons in VTA in their sensitivity to stimulation by the μ-opioid receptor (MOR). These receptors reside on DA neurons, where their activation excites DA neurons and contributes to reward-related signaling. Margolis said these MOR effects are more pronounced in DA neurons sending projections to the medial PFC and the amygdala, and less so in those projecting to the nucleus accumbens. Jesse Wood of the University of Pittsburgh followed with data from VTA neuron activity in rats learning to associate a cue with a reward. He found that combining information from all VTA neurons was less informative than taking information from smaller groupings of neurons—something that might reflect the region’s diversity—and suggested that information coming out of the VTA stems from changeable groupings of synchronized VTA neurons.
By selectively tinkering with subsets of DA neurons in the VTA, two posters on Wednesday morning, November 13, proposed new models of schizophrenia. Muhammad Chohan of the New York State Psychiatric Institute in New York City presented preliminary efforts to increase burst firing in dopamine-containing VTA neurons that project to the dorsal striatum, an important locus for psychosis. Selectively knocking out a calcium-activated potassium channel called SK3, which contributes to slow and regular firing patterns, in dopamine-containing cells of the midbrain in mice increased bursting and firing rates in the lateral regions of the VTA (including the substantia nigra), which project to the dorsal striatum. When the researchers measured dopamine release in the dorsal striatum with voltammetry, however, they found less dopamine release than in controls. Chohan suggested that this might reflect some kind of compensation, and a more acute takedown of SK3 may result in the expected increase. The converse experiment, in which bursting in medial parts of the VTA was selectively decreased, was presented across the aisle by Abigail Kalmbach at Columbia University in New York City. Selectively knocking out the glutamate receptor subunit NR1 in the VTA decreased bursting by the medial parts of the VTA, and she said experiments are underway to characterize behavior in both animal models.
Other routes to disturbing dopamine
Another interneuron-inspired model of schizophrenia, the ErbB4 knockout mouse, has effects on DA signaling as well, according to a poster in the same session by Miguel Skirzewski of Andres Buonanno’s lab at the National Institutes of Health in Bethesda, Maryland. ErbB4 is the receptor for neuregulin-1, a protein implicated by genetic studies of schizophrenia. ErbB4 resides in PV interneurons, and Buonanno’s group has previously found that neuregulin-1 activation of ErbB4 increases extracellular DA (Shamir et al., 2012). To look into the mechanisms behind this, Skirzewski used microdialysis to sample DA and its metabolites in the dorsal hippocampus in ErbB4 knockout mice. He found that they lacked the usual DA upregulation in response to neuregulin, but when they were stressed with a tail pinch, abnormally high DA levels emerged. The mice also showed enhanced levels of amphetamine-induced locomotion. Together, the results suggest that interactions between neuregulin-1 and ErbB4 shape dopamine signaling.
In a poster on Saturday, November 9, Ryan Ward of Columbia University, New York City, presented recent results from a mouse model carrying a more direct manipulation of the dopamine system. Developed by Eric Kandel and colleagues, these mice overexpress the D2 subtype of DA receptors (D2R) in the dorsal striatum, which simulates the focal increase in dopamine signaling found in people with psychosis (see SRF related news story). Previous work has shown that these mice have impaired motivation (Drew et al., 2007), and according to the data in Ward’s poster, this may influence attentional capabilities. Specifically, in an attention task, normal mice will perform more accurately when they know that they will receive a reward if they get it right; however, the D2R-overexpressing mice were indifferent to this condition, making the same number of errors regardless of reward probability. This suggests that problems with recognizing the reward value of a stimulus may underlie some cognitive impairments in schizophrenia (see SRF related news story).—Michele Solis.