Identification and characterization of neuronal ensembles activated during drug-induced behaviors Bruce T. Hope Repeated drug administration produces behavioral changes in rats as they learn to associate drug effects with stimuli present during drug administration. This form of learning is thought to involve neuroplastic changes in sparsely distributed neurons, called neuronal ensembles, that are activated by drug-associated stimuli. Until now there has been no method to identify these neuronal ensembles for analysis. All methods to date examine neurons in whole brain regions or a class of neurons identified by neurotransmitter or some other chemical characteristic. These methods miss or under-represent alterations in the specific neurons activated during drug-induced behavior. To address this problem, my lab has developed novel methods for identifying and characterizing neuronal ensembles that are activated during locomotor sensitization or self-administration of cocaine and heroin. Identifying these neurons will help us to characterize the molecular and cellular alterations that mediate behavioral responses to cocaine and other drugs of abuse. We bred cfos-lacZ transgenic rats to identify and selectively manipulate these neurons in live tissue. The transgene in these rats contains a c-fos promoter that regulates transcription of the bacterial gene lacZ, which encodes the enzyme beta-galactosidase (Kasof et al. 1995, J. Neurosci. 15:4238-4249). We have observed induction of beta-galactosidase in fixed striatal and prefrontal cortex tissue following acute drug administration and cue exposures. Beta-galactosidase is induced with a similar time course, dose-response relationship, as Fos protein and is co-induced in all Fos-labeled neurons indicating its use as a marker of activated neurons. Beta-galactosidase labeling has enabled us to identify and manipulate the enzyme in live tissue. We used our recently developed Daun02 inactivation procedure with cfos-lacZ rats to selectively inactivate neuronal ensembles in nucleus accumbens and prefrontal cortex that were activated during various conditioned drug behaviors. Daun02 is a suicide substrate for beta-galactosidase that, in culture following bath application of Daun02, kills or inactivates cells that contain the enzyme (Farquhar et al. 2002, Cancer Chemother. Pharmacol.50:65-70). Wehave shown that after cocaine has induced beta-galactosidase, subsequent injection of Daun02 into the nucleus accumbens will inactivate only those neurons activated during cocaine administration and attenuate sensitized cocaine-induced locomotor activity. Daun02 does not inactivate neurons after saline injections to cfos-lacZ rats or in wild-type rats injected with cocaine. Daun02 appears selective for only activated neuronal ensembles because when neurons are activated in one environment and Daun02 is injected into the accumbens, it does not alter locomotor activity in a distinct environment. Selective lesions of activated neuronal ensembles during behavior has never been done before. It allows us to determine causal roles for neuronal ensembles in behavior. We found that context-specific neuronal ensembles: (1) in nucleus accumbens mediate context-specific locomotor sensitization; (2) in ventral medial prefrontal cortex mediate context-specific reinstatement of heroin seeking; (3) in orbitofrontal cortex mediate expression of cue-induced heroin seeking; (4) in nucleus accumbens and ventral medial prefrontal cortex mediate context-induced reinstatement of cocaine seeking. The technique has immense potential for understanding many different forms of learning processes. We have bred cfos-GFP mice that express green fluorescent protein in activated neurons. In collaboration with Dr. Carl Lupica, we have detected the fluorescent signal in sparsely distributed neurons in striatal slices obtained from cfos-GFP mice that had previously received cocaine. Identification of live neurons in striatal preparations that were active during drug-induced behavior has also never been done before. Following sensitization to cocaine, we have found unique electrophysiological alterations that occur only in neurons that contain GFP versus those that do not. We have found altered AMPA/NMDA ratios and spontaneous ESPCs following activation of glutamatergic afferents. We also used our recently developed cfos-GFP rats to examine similar electrophysiological alterations in behaviorally-activated neurons during reinstatement of cocaine and heroin-seeking in rats. We have developed a method to dissociate striatal neurons from adult rat brains following conditioned drug behaviors and sort the activated neurons that contain beta-galactosidase or Fos from the majority of non-activated neurons that do not contain these activation markers. We use fluorescence-activated cell sorting (FACS) to sort these neurons. Following FACS purification, we have found unique alterations in mRNA messages that are different for activated versus non-activated neurons using microarrays and real-time PCR. We have used FACS to identify the unique molecular alterations in activated neurons: (1) in nucleus accumbens during context-specific sensitization to cocaine; (2) in prefrontal cortex during cue-induced reinstatement of heroin seeking; (3) in dorsal striatum following acute methamphetamine; (4) in dorsal striatum during context-induced reinstatement of methamphetamine seeking. To further advance our studies, we developed a new cfos-teto-Cre transgenic rat system along with new viruses to manipulate neuronal ensembles in these self-administration models. Using context-induced reinstatement of cocaine seeking, we demonstrated context-specific selection of neuronal ensembles that were distinct when exposed to different contexts. We also demonstrated context-specific induction of the optogenetic protein halorhodopsin these neuronal ensembles that was used to show selective inhibition of memories associated with drug taking versus other memories not associated with drug taking. We focus our methods on cocaine, heroin, and methamphetamine self-administration models. The methods we have developed in the past several years have enabled us to identify and characterize a unique class of neurons that are selectively activated during drug administration. These neurons appear to be part of the neuronal ensembles that represent stimuli and learned associations between environment, interoceptive cues, and drug effects. Understanding the role of these neuronal ensembles in behavior and the ways that repeated drug administration alters them will help us to understand how drugs of abuse produce the learned behaviors associated with addiction. The long-term goal will be to selectively attenuate or even erase addiction-related memories in human addicts without significant effects on their other memories.
Showing the most recent 10 out of 41 publications
