Chronic alcohol alters rewarded behaviors and striatal plasticity

Abstract

Chronic intermittent ethanol (CIE) alters neural functions and behaviors mediated by the dorsolateral striatum (DLS) and prefrontal cortex. Here, we examined the effects of prolonged (16-bout) CIE on DLS plasticity and DLS-mediated behaviors. Ex vivo electrophysiological recordings revealed loss in efficacy of DLS synaptically induced activation and absent long-term depression after CIE. CIE increased two-bottle choice drinking and impaired Pavlovian-to-instrumental transfer but not discriminated approach. These data suggest prolonged CIE impaired DLS plasticity, to produce associated changes in drinking and cue-controlled reward-seeking. Given recent evidence that less-prolonged CIE can promote certain dorsal striatal-mediated behaviors, CIE may drive chronicity-dependent adaptations in corticostriatal systems regulating behavior.

There is mounting evidence that chronic alcohol causes profound alterations in neural mechanisms mediating plasticity, which may underlie associated changes in behavior (Everitt & Robbins 2005; McCool 2011; Holmes, Spanagel & Krystal 2013). Recent work has shown, e.g. that mice subjected to relatively limited chronic intermittent ethanol (CIE) exposure (eight daily bouts) is sufficient to disrupt synaptic plasticity in the dorsolateral striatum (DLS) and facilitate various forms of operant learning (DePoy et al. 2013). Interestingly, more prolonged CIE exposure (16 daily bouts or more) compromises plasticity-mediating mechanisms in the medial prefrontal cortex and impairs regulation of fear and ethanol-seeking (Holmes et al. 2012; Meinhardt et al. 2013). These previous findings suggest that the impact of CIE on corticostriatal functions may vary depending on the chronicity of exposure. However, the effects of prolonged (16-bout) CIE on DLS plasticity and DLS-mediated behaviors are currently unclear.

To address this gap in the literature, we began with ex vivo electrophysiological recordings made from cortical glutamatergic inputs to synapses in the DLS (Brigman et al. 2013) after mice were exposed to CIE (16 daily bouts) (Fig. 1a). Each daily bout of exposure comprised 16 hours of continuous exposure to vaporized ethanol (achieving 175 mg/dL blood ethanol concentrations) followed by 8 hours of withdrawal, and was designed to mimic drinking patterns characteristic of alcoholism (Koob 2003; Heilig & Egli 2006; Griffin, Lopez & Becker 2009) (for methods in full, see Supporting Information). All recordings (as well as the behavioral tests described below)were performed 3 days after the final bout—a time point at which acute withdrawal has abated and mice do not exhibit signs of heightened anxiety-like behavior, motor impairment or ethanol tolerance (Holmes et al. 2012).

Figure 1.

Prolonged CIE impairs dorsal striatal synaptic plasticity. (a) Location of recordings made from synapses in DLS. (b) Population spike magnitude increased with stimulation amplitude, but CIE mice showed significantly lower values than air mice. (c) High-frequency stimulation (HFS) induced LTD in air, but not CIE, mice after each of three sets of trains compared with baseline. (d) Example traces. n = 6–8/group. Data are means ± SEM. *P < 0.05 versus air, #P < 0.05 versus baseline

As expected, population spike magnitude increased in response to the application of higher stimulation amplitudes (F14,182 = 205.30, P < 0.01). However, CIE mice showed significantly lower values than controls exposed to air vapors (F1,13 = 5.44, P < 0.05) (Fig. 1b), consistent with a CIE-induced loss in efficacy of synaptically induced activation in the DLS. Next, we examined long-term depression (LTD) induced by local afferent high-frequency stimulation (HFS) of glutamatergic inputs to DLS. Indicating LTD in air controls, population spike amplitude was significantly decreased after the first [t(13) = 2.49, P < 0.05], second [t(13) = 2.34, P < 0.05] and third [t(13) = 2.29 P < 0.05] set of three HFS trains, as compared with pre-HFS baseline. By contrast, HFS completely failed to induce LTD in CIE mice (Fig. 1c,d), similar to the effects of more limited (eight-bout) CIE found previously (DePoy et al. 2013). Loss of LTD could reflect constrained rescaling of DLS synapses in response to cortical input and, together with adaptations in other regions, notably enhanced N-methyl-D-aspartic acid receptor-dependent long-term potentiation in dorsal medial striatum (DMS; Wang et al. 2010), may aggregate to cause changes in dorsal striatum (DS)-mediated behaviors.

To test whether changes in DS plasticity were associated with alterations in behaviors known to be mediated by the DS (Wang et al. 2010), we first assessed ethanol drinking using a 24-hour access two-bottle choice ethanol drinking procedure (Holmes et al. 2012). We found that CIE mice showed significantly higher ethanol consumption (CIE × time interaction: F6,78 = 3.21, P < 0.01) and preference (CIE × time interaction: F6,72 = 2.50, P < 0.05), as compared with air controls, specifically on the second, third and fourth 2-day blocks following vapor exposure (Fig. 2a,b). These data replicate prior studies showing that CIE, including more limited (eight-bout) exposure, increases ethanol drinking and seeking (Griffin et al. 2009; DePoy et al. 2013; Meinhardt et al. 2013) and establish a behavioral correlate of the altered striatal plasticity produced by the chronic ethanol exposure.

Figure 2.

Prolonged CIE impairs PIT and increases ethanol drinking. (a,b) CIE mice had significantly higher ethanol consumption and preference than air mice after the vapor chambers. (c) Air, but not CIE, mice pressed the active lever more in the presence of the CS+ than the CS− or during the intertrial interval (ITI), and both groups pressed the active lever more than the inactive lever regardless of CS. (d) Air and CIE mice made more head entries into the magazine in the presence of the CS+ than the CS− or during the ITI, although air mice made more head entries than CIE mice. n = 6–12/group. Data are means ± SEM. *P < 0.05 versus air, #P < 0.05 versus CS+

Based on prior evidence that eight-bout CIE produces alterations in DLS-mediated operant learning paradigms (DePoy et al. 2013), we next examined the effects of the 16-bout CIE procedure on the DS-dependent, Pavlovian-to-instrumental transfer (PIT) task (Corbit & Janak 2007). The ‘general PIT’ paradigm we used assays the ability of an auditory stimulus (CS+) previously associated with availability of a food reward to increase instrumental responding for the reward (Lederle et al. 2011). We found that on a non-rewarded PIT probe test conducted 3 days after vapor exposure, air controls pressed the active lever significantly more in the presence of the CS+ than the CS− or the no-CS intertrial interval (ITI) (F2,22 = 9.60, P < 0.01) and generally favored the active lever over the inactive lever (Fig. 2c). This discriminated PIT effect was absent in the CIE group, which generally showed more active than inactive lever responding, but did not respond more vigorously during the CS+ than the CS− or ITI (Fig. 2c). However, analysis of the number of head entries into the food magazine, a measure related to discriminated Pavlovian approach behavior, revealed that both the air controls (F2,22 = 8.45, P < 0.01) and the CIE mice (F2,22 = 8.20, P < 0.01) made more head entries during the CS+ than the CS− or ITI (Fig. 2d).

The current PIT data indicate deficient PIT, but intact discriminated approach, following CIE, a finding which is in close agreement with that previously reported in rats repeatedly exposed and withdrawn from a liquid ethanol diet (Ripley et al. 2004). A similar loss of PIT is also produced by inactivation of the DLS (Corbit & Janak 2007), raising the possibility that CIE-induced abnormalities in DLS plasticity might contribute to the deficits in this behavior. It should be noted, however, that while drug addiction is typically associated with increased drug-seeking in the presence of reward-predictive cues (Everitt & Robbins 2005), the effects of exposure to drugs of abuse on PIT in rodents have been found to be complex. For example, PIT can be either enhanced or impaired by a psychostimulant such as amphetamine, depending on when the drug is given relative to conditioning and testing (e.g. Wyvell & Berridge 2001; Hall & Gulley 2011).

The chronicity of drug exposure appears to be another critical factor. We previously found that an eight-bout CIE regimen facilitated, rather than impaired, certain behaviors that are dependent upon the dorsal striatum, including a form of visual choice learning (DePoy et al. 2013). Taken together with the current findings, these data suggest that whereas relatively limited CIE may drive adaptive alterations in the dorsal striatum that favor certain behaviors, including choice learning, more prolonged CIE results in a degradation of some dorsal striatum functions and associated behaviors, such as PIT. This model is likely to be an oversimplification, however. First, both the eight- and 16-bout CIE regimens produced loss of DLS LTD, despite the differing behavioral outcomes. Second, the complex behavioral processes affected by CIE are known to recruit multi-region brain networks that encompass not just the dorsal striatum, but also the prefrontal cortex, nucleus accumbens and amygdala (Holmes, Marchand & Coutureau 2010; Graybeal et al. 2011). Indeed, preliminary work has already shown that 12- or 16-bout CIE regimens disrupt orbitofrontal and medial prefrontal cortex-mediated behaviors (Badanich, Becker & Woodward 2011; Holmes et al. 2012; Kroener et al. 2012). Thus, the behavioral effects of CIE probably reflect the emergent consequences of functional alterations across multiple brain regions. The challenge for future work will be to further parse these effects to determine how chronic ethanol dynamically alters how the brain regulates rewarded behavior.