Aims: This is a review of preclinical studies covering alcohol-induced brain neuronal death and loss of neurogenesis as well as abstinence-induced brain cell genesis, e.g. brain regeneration. Efforts are made to relate preclinical studies to human studies. Methods: The studies described are preclinical rat experiments using a 4-day binge ethanol treatment known to induce physical dependence to ethanol. Neurodegeneration and cognitive deficits following binge treatment mimic the mild degeneration and cognitive deficits found in humans. Various histological methods are used to follow brain regional degeneration and regeneration. Results: Alcohol-induced degeneration occurs due to neuronal death during alcohol intoxication. Neuronal death is related to increases in oxidative stress in brain that coincide with the induction of proinflammatory cytokines and oxidative enzymes that insult brain. Degeneration is associated with increased NF-κB proinflammatory transcription and decreased CREB transcription. Corticolimbic brain regions are most sensitive to binge-induced degeneration and induce relearning deficits. Drugs that block oxidative stress and NF-κB transcription or increase CREB transcription block binge-induced neurodegeneration, inhibition of neurogenesis and proinflammatory enzyme induction. Regeneration of brain occurs during abstinence following binge ethanol treatment. Bursts of proliferating cells occur across multiple brain regions, with many new microglia across brain after months of abstinence and many new neurons in neurogenic hippocampal dentate gyrus. Brain regeneration may be important to sustain abstinence in humans. Conclusions: Alcohol-induced neurodegeneration occurs primarily during intoxication and is related to increased oxidative stress and proinflammatory proteins that are neurotoxic. Abstinence after binge ethanol intoxication results in brain cell genesis that could contribute to the return of brain function and structure found in abstinent humans.
Introduction: Alcohol-Induced Brain Damage
The characteristic feature of alcohol use disorders is the consumption of dangerous amounts of alcohol despite the knowledge that problems occur during drinking. Excessive levels of intake, the characteristic of binge drinking, leads to the development of tolerance and dependence that then further promotes excessive consumption but is also associated with changes in brain physiology, structure and function. In general, human alcoholics, both men and women, show significant volume loss in cortical and subcortical brain structures that includes both gray and white matter shrinkage. These widespread deficits occur in the absence of major nutritional deficiencies, although nutritional deficiencies can cause neurodegeneration and could contribute to alcoholic degeneration. Both postmortem and in vivo imaging studies of brain morphology find abnormal reduced brain volumes of gray and white matter across multiple regions. The frontal lobes are the most insulted region in the alcoholic brain with the superior frontal cortex showing significant neuronal loss (Kubota et al., 2001; Sullivan and Pfefferbaum, 2005). The frontal lobes regulate complex cognitive skills such as working memory, temporal ordering, discrimination and reversal learning that underlie judgment, attention, risk taking, motivation, mood and wanting. Disorders in these behaviors are central if not causal to the consumption of dangerous amounts of alcohol despite the knowledge of negative consequences. Accordingly, chronic alcoholics demonstrate impaired judgment, blunted affect, poor insight, social withdrawal, reduced motivation, distractibility, attention and impulse-control deficits (Parsons, 1987; Oscar-Berman and Hutner, 1993; Sullivan et al., 2000b; Sullivan and Pfefferbaum, 2005). It has been proposed that progressive increases in ethanol consumption lead to alterations in brain structures that reduce behavioral control promoting further alcohol abuse and neurodegeneration (Crews et al., 2004).
Interestingly, abstinence from ethanol is associated with a recovery in brain volume loss. Pfefferbaum et al.(1995) have shown using longitudinal MRI that in recovery or relapse, short-term abstinence (1 month) can increase cortical gray matter volume and long-term (1 year) abstinence shrinks the enlarged third ventricle. Similarly, neuropsychological deficits improve with abstinence particularly working memory, visual-spatial abilities and certain motor abilities. MR spectroscopy reveals improvement in neuronal integrity during abstinence (Sullivan et al., 2000a,b; Parks et al., 2002; Meyerhoff et al., 2004). Regeneration of the brain during abstinence has been proposed to underlie some of these improvements in executive function and impulse control that occur with sustained abstinence (Crews et al., 2004).
The mechanisms of alcohol-induced brain damage and abstinence-induced regeneration are complex. The extent of neurodegeneration and potential for recovery and regeneration varies by brain region and is dependent on many factors, including pattern of intake, genetics and age. This review will focus particularly on a binge model of alcohol dependence first developed to study physical dependence (Majchrowicz, 1975). Later studies indicated that the binge model induced neurodegeneration (Crews et al., 2004), reduced neurogenesis (Nixon and Crews, 2004) and caused cognitive deficits (Obernier et al., 2002b). More recently, studies showed evidence of regeneration during abstinence following binge treatment: striking increases in neurogenesis (Nixon and Crews, 2004) and cell genesis (Nixon et al., 2008) mimic the regeneration of human brain in abstinence. This review will cover these studies and make an effort to relate them to findings in humans.
Binge Model of Alcohol Dependence
The binge ethanol model of alcohol dependence has been studied widely as a reliable model for ethanol-induced brain damage (Crews et al., 2004). Rats receive ethanol intragastrically three times a day for 4 days. The daily dose of ethanol is ∼9 g/kg/day, which in rats results in blood levels between 250 and 400 mg%. The first dose of 5 g/kg causes significant intoxication, and future doses are adjusted depending on the level of intoxication, i.e. the subsequent dose is lowered for those showing intoxication from the previous dose. Tolerance rapidly occurs with most rats showing high blood levels, but little intoxication on Days 3–4. This model mimics the high blood ethanol levels and heavy alcohol consumption common among alcoholics. For example, 4–10% of emergency room patients have blood alcohols over 250 mg% (Urso et al., 1981; Teplin et al., 1989; Cartlidge and Redmond, 1990; see also Crews et al., 2000, introduction). Further, binge drinking is one of the most common patterns of intake with 40–60% of alcohol abusers drinking in this manner (e.g. Robin et al., 1998). Thus, the high blood alcohol levels and pattern of intake in the rat binge model are consistent with that commonly found in alcohol abusers.
Many investigators have shown that binge ethanol administration induces brain damage (Crews et al., 2004). Brain damage increases progressively after 2 days beginning in the frontal olfactory bulb followed by additional brain regions showing increasing damage with time. Most of the studies investigating binge ethanol-induced brain damage found peak damage just after the last of the 4-day doses (Fig. 1). Significant damage in adults is found in limbic association regions, namely agranular insular cortex, anterior piriform cortex, perirhinal cortex, entorhinal cortex and hippocampal dentate gyrus (Fig. 2), particularly ventral dentate gyrus. In these regions, dark cell degeneration, a necrotic form of cell death with shrunken soma, is the predominant form of neuronal death (Obernier et al., 2002a). In addition, binge ethanol treatment reduces hippocampal neurogenesis as discussed later (Nixon and Crews, 2002). In general, brain degeneration in the binge ethanol treatment model is diffused and widespread, similar to that found in human alcoholics. Interestingly, studies examining adolescent rats versus adults (Crews et al., 2004) and a genetic model of alcoholism, the P rat (Crews et al., 2004), have found increased binge-induced brain damage, particularly frontal and limbic cortical damage in both genetic and adolescent models of alcoholic risk. These studies suggest that the binge model is a reasonable model of human alcoholic neurodegeneration.
Cognitive impairments in alcoholics are linked to alcoholic neurodegeneration and alcohol dependence. Following binge alcohol treatment in adult rats, cognitive function has been examined using the Morris water maze (Fig. 3). Although initial learning of this task was similar between controls and binge-treated rats, as was activity, relearning tasks showed deficits in trials to criterion, greater time in the original (or wrong) quadrant and greater entries into the wrong quadrant. These deficits are consistent with disrupted frontal lobe function and perseveration (Obernier et al., 2002b). In addition, analysis of microglial density suggested that damage had occurred according to persistent elevations of PK11195 binding in frontal cortical regions such as the anterior piriform cortex (Obernier et al., 2002b). Studies in alcoholic human brain have also found increased microglial density in cortical regions (He and Crews, 2008). Thus, the rodent binge model shares both neurodegenerative and cognitive deficits found in human alcoholic brain.
Mechanisms of Alcohol-Induced Brain Damage: Degeneration Occurs During Intoxication
A number of studies have focused on the mechanisms of alcohol-induced brain damage. Early studies, mostly in vitro culture models, suggested that chronic ethanol inhibited glutamatergic N-methyl-d-aspartate (NMDA) receptors that in time resulted in NMDA supersensitivity, an effect only revealed upon the removal of alcohol (Chandler et al., 1993a, 1993b). These in vitro studies and others suggested that during withdrawal, neurotoxicity occurs through the NMDA receptor (Prendergast et al., 2000, 2004). Multiple groups studying the in vivo binge ethanol model therefore investigated the effectiveness of NMDA receptor antagonists at blocking or reducing binge-induced brain damage. Various NMDA receptor antagonists, including MK801 (dizocilpine), memantine and DNQX failed to reduce binge ethanol neurotoxicity and surprisingly, some doses even increased degeneration (Collins et al., 1998; Corso et al., 1998; Crews et al., 2004; Hamelink et al., 2005). Other hypotheses suggested that ethanol intoxication increased the brain density of voltage-gated calcium channels, which then contributed to excessive excitability during withdrawal; however, calcium channel blockers did not protect against binge-induced brain damage [see Crews et al. (2004) for review]. The in vivo binge studies continued to suggest that ethanol-induced brain damage occurred during intoxication and required relatively high blood levels. Blood levels clearly cycle during treatment so the impact of withdrawal was hard to assess. Time course studies of binge-induced brain damage indicated that damage was increasing during ethanol treatment and began to subside with withdrawal (Fig. 1). Most studies used agyrophilic silver stain, which identifies cells irreversibly committed to cell death ∼1–3 days after the insult (Switzer III, 2000). Thus silver staining may increase over a several-day course after cell death. In the binge model, silver stain appears after 2 days and peaks just after the last dose of ethanol (Day 4) but then subsides progressively over the first week of withdrawal as silver-stained material is cleared, likely by phagocytes (Fig. 1; Collins et al., 1996, 1998; Crews, 2000; Obernier et al., 2002a; Crews et al., 2004). The progressive loss of cell death markers during ethanol withdrawal over a week suggests that little to no degeneration occurs during withdrawal in this model. Withdrawal hyperexcitability and seizures, if present, occur during the first 24 h of withdrawal. Cell death markers would be expected to appear within days during the following week if damage occurred during withdrawal in this model. Further, the findings that damage occurs during intoxication when ethanol should block NMDA receptors and the lack of effect of NMDA antagonists on damage suggest that the mechanism of ethanol-induced brain damage is not glutamate excitotoxicity. Thus, ethanol-induced brain damage in the binge model occurs during intoxication.
Human studies also support that neurodegeneration occurs during intoxication with most studies supporting a recency and frequency of heavy drinking as the best indictor of alcoholic brain damage (Parsons, 1987; Sullivan and Pfefferbaum, 2005). Over a 5-year interval the degree of excessive drinking in alcoholics corresponded with gray matter loss, particularly in the frontal lobes (Pfefferbaum et al., 1995). Frontal cortical choline-containing compounds measured by MRI were decreased in alcoholics with significant correlations between alcohol consumption in the last 90 days and decreases in anterior cingulated cortex and frontal white matter (Ende et al., 2006). Other studies have suggested that the white matter may recover rapidly in abstinence (Shear et al., 1995; O’Neill et al., 2001) or be more vulnerable to damage during relapse (Pfefferbaum et al., 1995). Further, efforts to relate neurodegeneration to lifetime alcohol consumption have not found a relationship. Thus, recency and frequency of intoxication best predict alcoholic neurodegeneration consistent with damage during intoxication.
Mechanisms of Alcohol-Induced Brain Damage: Oxidative Stress and Pro-Inflammatory Cytokines Alter Transcription Factors in Brain
Many findings suggest that ethanol-induced brain damage is related to oxidative stress from proinflammatory enzymes activated during ethanol intoxication. Studies in organotypic hippocampal-entorhinal cortex slice (HEC) cultures have found that ethanol sensitizes neurons to insults. HEC slice cultures maintain all the cellular elements of brain as well as the intact structures and interconnections of brain. Transcription factors, such as cAMP responsive element-binding protein (CREB) and nuclear factor κB (NF-κB), regulate the gene expression. During the presence of ethanol, there are changes in protein transcription with increased DNA binding of NF-κB and reduced DNA binding of CREB (Figs 4 and 5). CREB family transcription factors are activated by phosphorylation and promote neuronal survival, protecting neurons from excitotoxicity and apoptosis through regulating the transcription of pro-survival factors (Lonze and Ginty, 2002; Mantamadiotis et al., 2002). Conversely, NF-κB is a transcriptional factor that is known widely for its ubiquitous roles in inflammatory and immune responses (O’Neill and Kaltschmidt, 1997). The balance in expression and activation of these transcription factors—and thus the balance of prosurvival versus proinflammatory states—suggests a mechanism by which alcohol induces brain damage in alcoholic neuropathology.
Due to its pivotal roles in neuronal plasticity and long-term memory, CREB family transcription factors have been implicated in addiction and drug abuse (Lonze and Ginty, 2002; Nestler, 2002). Ethanol activation of oxidative stress coincides with the loss of CREB prosurvival transcription (Figs 4 and Fig. 5). In vivo, binge ethanol treatment results in a decrease in pCREB immunoreactivity in brain during ethanol intoxication, an effect that is most pronounced in brain regions showing degeneration (Bison and Crews, 2003). The reduction in pCREB coincides with the peak of neurodegeneration in the in vivo binge. In HEC cultures, ethanol reduces levels of the CREB-regulated gene, brain-derived neurotrophic factor (BDNF), with reductions in BDNF coinciding with sensitivity to neurodegeneration (Zou and Crews, 2006). Thus, reduced CREB transcription contributes to ethanol neurotoxicity (Fig. 4). Thus, ethanol loss of trophic signals combined with the induction of oxidative stress proinflammatory signals (detailed below) appears to underlie the degeneration and atrophy that occur in alcoholism. The NF-κB transcriptional factor is well known for its role in inflammatory and immune response signaling, as well as in control of cell division and apoptosis (O’Neill and Kaltschmidt, 1997). The molecular composition of the NF-κB family is composed of several members, including p50, p52, p65/RelA, RelB and c-Rel, that form homo- and heterodimers capable of transmitting receptor signals to the nucleus (O’Neill and Kaltschmidt, 1997; Ghosh et al., 1998). NF-κB is activated by stimuli of oxidative stress, cytokines and glutamate (O’Neill and Kaltschmidt, 1997; Mattson and Camandola, 2001; Hinoi et al., 2002), and is the key in the induction of proinflammatory cytokines and enzymes that promote a proinflammatory cascade by further activating NF-κB transcription. In the central nervous system, NF-κB proteins are ubiquitously expressed in neurons and glia (Kaltschmidt et al., 1993; Mattson and Camandola, 2001) where, in addition to regulating physiological processes, they participate in pathological events (O’Neill and Kaltschmidt, 1997; Grilli and Memo, 1999; Pizzi et al., 2002). Increased NF-κB levels have been reported in dying neurons following trauma or ischemia (Terai et al., 1996a; Bethea et al., 1998; Schneider et al., 1999) as well as in brains of patients with Alzheimer's disease and Parkinson's disease (Terai et al., 1996b; Kaltschmidt et al., 1997). Activation of NF-κB transcription is associated with increases in pro-inflammatory cytokines with tumor necrosis factor-α (TNFα) being the prototype. Cytokines are large families of secreted proteins that interact in complex cascades or patterns regulating immune, hormonal and nervous systems. Systemic cytokines are transported from serum into brain, particularly for TNFα, IL1®, NGF, IL6 and IFN© (Banks, 2005). A role for cytokines in alcoholic neuropathology is suggested by several studies [see Crews et al. (2006b) for review]. Szabo's studies of human monocytes from alcoholics found greater amounts of TNFα spontaneously and in response to challenge (Crews et al., 2006b), and acute ethanol increases cytokine induction by TLR2 and TLR4 ligands (Oak et al., 2006). Thus, ethanol through both systemic and central nervous system mechanisms induces a proinflammatory cascade in the brain that contributes to neurotoxicity.
Both in vivo and in vitro evidence support the involvement of a proinflammatory cascade including increased NF-κB-driven induction of oxidative stress enzymes as a key factor in alcohol-induced brain damage. TNFα can directly potentiate glutamate neurotoxicity by inhibiting glutamate uptake through NF-κB mechanisms (Zou and Crews, 2005). In human astroglial cells, which normally regulate extracellular glutamate concentrations, ethanol enhances NFκB-DNA binding and the induction of iNOS (Davis and Syapin, 2004; Davis et al., 2005). Similarly, we found that ethanol induces COX2 (Knapp and Crews, 1999), iNOS and NADPH oxidase gp91, and increases reactive oxygen species producing enzymes that are downstream of NFκB. NADPH oxidase is a multimeric enzyme composed of multiple subunits that in the active enzyme catalyze the transfer of one electron from NADPH to oxygen, giving rise to superoxide. Ethanol significantly increases the brain expression of NADPH oxidase subunits, gp91phox and p67phox, that persists for at least 8 days of abstinence (Qin et al., 2008). Thus, ethanol promotes a proinflammatory and anti-survival environment through the activation of proinflammatory transcription factors and the inhibition of prosurvival transcription factors.
Considerable support for the proinflammatory mechanism of alcohol brain damage also derives from studies showing protection against alcohol-induced brain damage with antioxidants. Collins’ work first suggested inflammation as a mechanism of damage (Collins et al., 1998); however, Eskay clearly established that anti-oxidants protect against binge-induced brain damage (Hamelink et al., 2005). We have also found that administration of the anti-oxidant butylated hydroxytoluene (BHT), during binge ethanol treatment, blocked COX2 induction and neuronal cell death as well as reversing ethanol inhibition of neurogenesis (see Crews et al., 2006a). Furosemide, a diuretic, has been found in some studies to protect against ethanol-induced brain damage and was suggested to act by reducing inflammatory brain swelling (Collins et al., 1998). However, other studies did not find protection by furosemide or the anti-CSF swelling treatment, mannitol (Crews et al., 2004). Eskay's group, in a more thorough study, explained these discrepancies by relating free radical scavenging activity to multiple protective agents including BHT, furosemide and cannabidiol, an anti-oxidant cannabinoid (Hamelink et al., 2005). In vitro, we have shown that BHT reduced NF-κB activation (DNA binding) in HEC slices and reduced ethanol-TNFα, glutamate and/or H2O2-stimulated neuronal cell death (Zou and Crews, 2004). We have also tested the hypothesis that ethanol sensitized neurons to insults by decreasing pCREB by experimentally increasing pCREB activation (DNA binding) using phosphodiesterase IV inhibitor rolipram and examining toxicity. Rolipram treatment increased CREB-DNA binding and markedly reduced ethanol-TNFα, glutamate and H2O2-stimulated neuronal cell death in HEC slices (Zou and Crews, 2004). Further, rolipram administered with binge ethanol treatment in vivo increased pCREB immunohistochemical expression and protected against binge-induced neurodegeneration (Pluzarev et al., 2008). Thus, anti-oxidants that block NF-κB induction of proinflammatory transcription or block the ethanol reduction in pCREB transcription can inhibit ethanol neurodegeneration. Interestingly, there is crosstalk between these transcription factors since blocking ethanol NF-κB induction with BHT also blocks the reduction in pCREB by ethanol (Zou and Crews, 2004). Further, adult neurogenesis is inhibited by ethanol and several studies have found that anti-oxidants also blunt ethanol inhibition of neurogenesis (Herrera et al., 2003; Crews et al., 2006a). Thus, inhibitor studies strongly support the ethanol-induced oxidative stress; proin- flammatory environment causes alcoholic neurodegeneration.
Ethanol Inhibition of Neurogenesis: A Mechanism of Neurodegeneration
The discovery that alcohol also inhibits the ongoing genesis of neurons and glia has highlighted a new possible mechanism by which alcohol results in brain volume/tissue loss or neurodegeneration. Lack of cell generation may be a key mechanism of neurodegeneration (Nixon, 2006). Indeed, in many neurodegenerative diseases, the lack of ongoing cell generation by stem cells has been hypothesized to contribute to tissue loss (see Armstrong and Barker, 2001). New neurons from neural stem cells are constitutively produced in at least two regions of the normal, adult brain. Neural stem cells in the subventricular zone of the anterior lateral ventricles produce olfactory bulb neurons (Alvarez-Buylla and Garcia-Verdugo, 2002; Curtis et al., 2007) and neural stem cells (NSCs) or progenitor cells in the subgranular zone of the hippocampal dentate gyrus produce dentate gyrus granule cells (Altman and Das, 1965; Eriksson et al., 1998). Evidence continues to emerge for the role of neural stem cells/adult neurogenesis in neurodegenerative disorders. Early studies on stress showed that increased adrenal steroids both inhibited neurogenesis and resulted in dentate gyrus degeneration (Cameron and Gould, 1994). Studies in a variety of neurodegenerative disorders have shown a direct correlation between inhibition of neurogenesis and dentate gyrus cell or volume loss and most importantly this degeneration often occurs without overt increases in cell death (for review, see Eisch, 2002). Further, by mere inhibition of NSC proliferation/adult neurogenesis, hippocampal-dependent behaviors and functions are impaired (Gould et al., 1999; Nilsson et al., 1999; Shors et al., 2001; Snyder et al., 2001; van Praag et al., 1999a, 1999b), whereas increased neurogenesis improves hippocampus-related behaviors (Santarelli et al., 2003).
Four primary processes comprise the formation of new neurons, or neurogenesis: proliferation, differentiation, migration and survival (Kempermann et al., 2004). Elegant work in vitro has shown repeatedly that alcohol alters NSCs, including their ability to form colonies, proliferate, differentiate and survive (Crews et al., 2003; Hao et al., 2003a, 2003b; Tateno et al., 2005), which has been extended to in vivo models of alcohol exposure. Most in vivo models report that alcohol intoxication inhibits NSC proliferation (e.g. Nixon and Crews, 2002; Jang et al., 2002a, 2002b). The percent inhibition is strikingly similar among studies and exposure patterns with most reporting a 50 ± 10% decrease in the number of proliferating cells in adults (for review, see Table 1 in Nixon, 2006). Acute alcohol exposure dose dependently reduces NSC proliferation in adolescent rats with high doses completely inhibiting progenitors [see Fig. 1 and Crews et al. (2006c)]. Multiple markers covering all stages of neurogenesis have been examined in the 4-day binge model and support the loss of progenitor cell proliferation immediately after the last dose (Nixon and Crews, 2002, 2004; He et al., 2005; Crews et al., 2006a). Several groups have consistently shown that progenitor cell survival is also reduced, which suggests another mechanism by which alcohol reduces neurogenesis in rats (Nixon and Crews, 2002; Herrera et al., 2003; He et al., 2005). Further, ethanol treatment during adult neurogenesis blunts the growth of the progenitor's dendritic arbor (He et al., 2005). Taken together, these studies indicate that ethanol reduces neurogenesis during intoxication contributing to neurodegeneration through loss of cell generation. Intriguingly, inflammatory processes may inhibit neurogenesis (Monje et al., 2003). Thus, ethanol activation of proinflammatory cytokine–oxidative stress cascades likely inhibits neurogenesis as well as mediating the other necrotic degenerative processes (Figs 1 and 4).
Glia and Alcoholic Neurodegeneration
Neuronal loss may not account for all volume loss across the alcohol-damaged brain (Harper et al., 1987; Harper and Kril, 1989; Kril et al., 1997). Glia may also contribute to alcoholic neurodegeneration. Alcohol causes astroglia to degenerate, leaving a void in trophic and metabolic support, and then neurons degenerate (Kimelberg and Aschner, 1994). With a loss of astroglia comes a reduced ability to take up excess glutamate, buffer K+ (ion homeostasis) and eliminate free radicals (e.g. Dringen, 2000). Glia may be more sensitive than neurons to the effects of alcohol (Miguel-Hidalgo et al., 2002). Careful studies in postmortem human hippocampus found a statistically significant loss of 37% of the glial cells in alcoholic hippocampus that included a reduction of astrocytes and oligodendrocytes but no loss of neurons (Korbo, 1999). Studies of astrocyte cultures have found acetaldehyde to be toxic to astrocytes (Holownia et al., 1997, 1999). Long-term chronic alcohol exposure in vivo decreases an intermediate neurofilament that is a characteristic of astrocytes, glial fibrillary acidic protein (GFAP), in the cerebellum of male and females rats (Rintala et al., 2001). The loss of GFAP expression suggests a loss of astrocytes (Rintala et al., 2001) consistent with the finding that the number of astrocytes identified by giemsa staining in human hippocampus is reduced in alcoholics (Korbo, 1999; Miguel-Hidalgo and Rajkowska, 2003; Miguel-Hidalgo, 2005; Matsuda-Matsumoto et al., 2007). Alternatively, other studies have reported no effect on glia including stereological studies performed on rat brain tissue immediately after 40 weeks of ethanol-containing liquid diet (Dlugos and Pentney, 2001) and on human alcoholic brains (Fabricius et al., 2007). Findings are complicated by the fact that glia proliferate and/or increase in response to brain damage. The term reactive gliosis refers to damage-induced increases in the proliferation of glia and observable increases in glia markers, coupled with altered morphology and function. Several studies have found that ethanol treatment produced reactive astrocytosis across multiple brain regions as evidenced by increased GFAP immunoreactivity, increased S100ß expression and/or increased glial DNA, all markers of gliosis (Rosengren et al., 1985; Franke et al., 1997; Satriotomo et al., 1999; Baydas and Tuzcu, 2005; Gonca et al., 2005). Thus, the role of glia in alcoholic degeneration and regeneration remains uncertain due to rapidly changing roles and responses.
Abstinence and Brain Regeneration
Human studies provide substantial evidence that sobriety results in improvement during abstinence (Carlen et al., 1978; Sullivan et al., 2000b). Studies have found improvement in component processes of nonverbal memory, visuospatial abilities, attention, as well as gait and balance within months of sobriety (Sullivan et al., 2000a, 2000b). Longitudinal neuroimaging studies of brain structure find increases in cortical gray matter volume within 1 month of sobriety followed by increases in white matter and reversal of ventricular enlargement with longer abstinence (Pfefferbaum et al., 1995; O’Neill et al., 2001). Very long abstinence has been reported to resolve most neurocognitive deficits associated with alcoholism (Fein et al., 2006). Recent studies have found associations between the recovery of memory and lateral ventricle shrinkage and the recovery of ataxia and fourth ventricle shrinkage (Rosenbloom et al., 2007). Thus in general, studies suggest that recovery of function during abstinence is associated with ultrastructural changes in brain that include growth of both white and gray matter as well as improved overall function.
In our studies of the binge model of alcohol dependence we found that abstinence results in increased cell proliferation in multiple brain regions (Nixon and Crews, 2004; Nixon et al., 2008; Figs 6 and 7). Cell proliferation increases across multiple brain regions after just 1 day of abstinence (Nixon et al., 2008). After 2 days, cell genesis is increased in cortical regions, including both gray and white matter areas and across many non-cortical brain regions. We found that dentate gyrus proliferation increases from binge ethanol-inhibited levels to control levels after ∼1 day of abstinence that is followed by further increases in proliferation several fold more than in controls that persists for more than 1 week. Two bursts of proliferation occur in the first week of abstinence. The first happens around 2 days after the last dose of ethanol and is the most widespread. We have shown that the majority of these cells are proliferating microglia (Nixon et al., 2008), which is a hallmark of microgliosis (Raivich et al., 1999). Interestingly, these new microglia persist for months and migrate to many brain regions. At 7 days after the last dose, a second but regionally specific burst occurs in the dentate gyrus subgranular zone. The burst of proliferation at 7 days in dentate gyrus, as indicated by increased BrdU+ cells, results in increased doublecortin expression, a marker of immature neurons, peaking around 14 days of abstinence (Fig. 6). At 5 weeks of abstinence when progenitors have had time to mature into neurons, abstinent animals have more BrdU+ cells that have become new dentate gyrus granule cell neurons (Nixon and Crews, 2004). The burst of cell proliferation occurs as the degeneration and fragments of dying neurons clear (Fig. 7). Many brain regions other that the dentate gyrus show increased formation of new cells. Following moderate self-administration, alcohol-preferring rats (P rats) show new cell genesis across the brain including abstinence-induced increases in cell genesis of progenitors in the hippocampus and substantia nigra (He and Crews, 2008). Additional studies are needed to fully understand the mechanisms of increased proliferation and subsequent increase in new progenitors, microglia and neurons during abstinence.
Interestingly, the finding of new progenitors and microglia during abstinence is consistent with recent findings in postmortem alcoholic brain. Brains of human alcoholics show increased levels of the cytokine MCP1 (CCL2) and increases in brain microglial markers (He and Crews, 2008). Further, studies in animals have reported that chronic ethanol can cause a long-term increase in the density of [3H]PK-11195, a ligand for the peripheral benzodiazepine receptor, that is enriched in microglia (Syapin and Alkana, 1988, Obernier et al., 2002b). Microglia exist in multiple states that can range from phagocytic inflammatory to trophic secretion of growth factors. The ‘activated’ microglia marker, ED-1, is not expressed in human or rat brain during abstinence. This is supported by the lack of amoeboid-shaped microglia in any of the new abstinence induced microglia (Riikonen et al., 2002; Crews et al., 2006a; Nixon et al., 2008). Although microglial proliferation is a sign of microglial activation, our studies suggest that the abstinence-induced microglia are not activated in a proinflammatory state. Similarly, human alcoholic brain shows increased microglia markers but not signs of increased proinflammatory microglia (He and Crews, 2008). It is possible that the new microglia are trophic microglia that secrete factors that stimulate the brain regeneration during abstinence.
In summary, during the first week of abstinence following a 4-day binge ethanol treatment, there is a burst of brain cell proliferation in many regions of brain, particularly hippocampus and cortex. Proliferating cells labeled with BrdU largely differentiate into microglia in non-neurogenic brain regions. In the dentate gyrus, proliferating cells differentiate over several weeks mostly into neurons. The cell-, microglia- and neuronogenesis during abstinence represent a unique and long-term change in brain cell structure that persists for long periods and perhaps permanently.
Regeneration: Mechanisms of Abstinence-Increased Cell Genesis and Brain Growth
Regrowth of the brain during abstinence from alcohol likely mimics general mechanisms involved in brain growth and plasticity. Exercise and environmental enrichment are known to protect the brain from neurodegenerative insults, age-related dysfunction and to improve learning ability. Exercise can also block ethanol inhibition of cell proliferation (Crews et al., 2004). One possible mechanism that contributes to increased cell growth during abstinence is the response to cell death, e.g. degeneration stimulated regeneration. Binge-induced degeneration is widespread in cortico-limbic areas; however, postbinge cell genesis occurs in many more brain regions not previously thought to be damaged by binge ethanol treatment. Studies of cell genesis after brain insults have found cell genesis in regions not directly damaged. For example, entorhinal cortex lesions cause increased cell proliferation 3 days later in the dentate gyrus on both the lesion side and the contralateral hippocampus. This suggests that proliferation occurs in projection areas of the dying neurons as well as in the cell body regions (Hailer et al., 1999). Similarly, focal electrolytic lesions of the dentate gyrus result in increased proliferation on postlesion days 1 and 5 in both the ipsilateral and contralateral dentate gyrus (Ernst and Christie, 2006). The lesion-induced increase in proliferation in the dentate gyrus resulted in more neurons, e.g. increased neurogenesis, at 6 weeks. However, a lesion in the frontal cortex stimulated proliferation, but the cells did not differentiate into neurons (Ernst and Christie, 2006). Some studies have suggested that quiescent progenitors exist across the brain that, when stimulated, differentiate depending upon their surrounding environment, with a special ‘neurogenic niche’ regulating neurogenesis in the dentate gyrus and subventricular zone where progenitors constantly become neurons (Wurmser et al., 2004). Simple differentiation can stimulate progenitors, as well as NMDA antagonists (Cameron and Gould, 1994). And though differentiation and neurodegeneration occur during alcohol intoxication, cell- and neurogenesis are inhibited during intoxication. Thus, one mechanism of abstinence regeneration may be a response to intoxication-induced cell death that becomes apparent only during abstinence after the alcohol has cleared (Fig. 7).
Another mechanism of abstinence-induced regeneration may be an increase in pCREB transcription that rebounds from suppressed levels during abstinence. A variety of studies have suggested that pCREB transcription increases plasticity and survival of neurons that have been insulted (Walton and Dragunow, 2000; Mabuchi et al., 2001; Hara et al., 2003). Studies have linked increased pCREB formation and transcription to synaptic prosurvival NMDA receptors (Hardingham and Bading, 2002). Stimulation of these synaptic NMDA receptors may occur during the high extracellular glutamate levels characteristic of the alcohol withdrawal hyperexcitable syndrome. This hyperexcitability is associated with increased glutamate release during withdrawal (De Witte et al., 2003) and synaptic glutamate release is associated with increased activation of kinases that increase pCREB formation and drive BDNF synthesis and secretion (Papadia and Hardingham, 2007). This suggests that alcohol withdrawal-induced synaptic glutamate release might activate synaptic NMDA receptors that are trophic. Interestingly, chronic ethanol has been found to increase the sensitivity of NMDA receptors (Chandler et al., 1994), and the dendritic spines and postsynaptic density proteins within excitatory synapses (Carpenter-Hyland et al., 2004; Mulholland et al., 2008). This could increase excitatory synaptic responses stimulating growth. Increased NMDA stimulation will increase neurogenesis (Joo et al., 2007) similar to that found during ethanol abstinence. The binge model of alcohol dependence shows decreased pCREB histochemistry during intoxication, when NMDA receptors are blunted by ethanol (Fig. 8; Bison and Crews, 2003). Thus, as mentioned earlier, pCREB is reduced during intoxication when degeneration occurs and neurogenesis is inhibited (Fig. 4). Abstinence from ethanol following the binge results in a marked increase in pCREB histochemistry at 3 days of abstinence that is consistent with the cell genesis beginning across brain a few days later (Fig. 8). Thus, it is possible that during abstinence, withdrawal-induced glutamate release stimulates supersensitive synaptic NMDA trophic receptors that increase pCREB transcription resulting in brain growth. The increase in pCREB could increase plasticity, cell growth, cell proliferation and neurogenesis. Thus, another possible mechanism of abstinence-induced increases in neurogenesis, cell genesis, cell growth and plasticity is the increase in synaptic glutamate release and activation of synaptic glutamate receptors during abstinence.
The binge model of alcohol dependence effectively models the degeneration and regeneration associated with alcohol dependence and recovery as well as alcohol-induced relearning deficits due to perseverative-like behavior, consistent with the loss of executive functions in alcoholism. The binge model demonstrates that alcohol-induced degeneration involves brain shrinkage that likely occurs due to multiple factors inhibition ongoing cell generation, cellular atrophy and cell death. The combination of our in vitro models followed by in vivo experiments have allowed us to show that alcohol shifts the balance of proinflammatory versus pro-survival signaling toward inflammation during ethanol intoxication. Ethanol activation of NF-κB transcription increases proinflammatory cytokines and oxidative stress inducing enzymes that promote degeneration. Ethanol reductions in pCREB transcription sensitize neurons to insults, including oxidative stress, that likely contribute to the degeneration and loss of neurogenesis. Inhibition of NF-κB transcription or increased pCREB transcription blocks binge-induced degeneration. Human alcoholics show increased brain cytokines and microglial markers consistent with these processes contributing to the loss of gray and white matter as well as the cognitive dysfunction. Human studies find brain damage associated with the most recent and frequent drinking consistent with the binge studies indicating that ethanol-induced brain damage occurs during intoxication.
Human abstinent alcoholics show a partial reversal of brain cognitive deficits and a regrowth of brain. Similarly, the binge-induced damage clears during abstinence and a burst of new cells signals a regrowth of brain. Cell genesis is found across the brain and results in the formation of new neurons, microglia and progenitors after months of abstinence. Our binge models also demonstrate that in abstinence, glutamate action on synaptic NMDA receptors shifts the balance of signaling toward survival and growth. Thus, an increase in pCREB transcription was observed and it represents one mechanism of the return of plasticity and increased cell genesis associated with abstinence. Additional studies are needed to fully elucidate the processes that contribute to the degeneration and regeneration associated with dependence and abstinence. It is hoped that understanding these mechanisms will lead to new approaches to block the progression to dependence and improve the success of abstinence.