|Home | About | Journals | Submit | Contact Us | Français|
Cosgrove et al. previously reported on the dynamics of GABAA receptor levels during abstinence from alcohol in treatment seeking alcohol dependent individuals (Staley et al. 2005, Cosgrove et al. 2014). GABAA receptor levels increased temporarily within a week of withdrawal and then normalized within about 1 month of abstinence; this transient increase occurred only in alcohol dependent non-smokers, whereas receptor levels remained elevated in smokers at 1 month of abstinence from alcohol. GABAA receptor levels are of relevance to relapse risk, because elevated levels in smokers correlated with greater craving for both alcohol and cigarettes during the first week of withdrawal and with even higher craving for alcohol at 1 month of abstinence in smokers; alcohol dependent non-smokers did not display any such correlations with alcohol craving. These human studies are thought-provoking and raised important questions about the best way of withdrawing from substances during treatment: Will GABAA receptor levels – and potentially craving for substances – change differently if a treatment seeker withdraws from alcohol only (continuing tobacco use) or withdraws from both alcohol and tobacco at the same time?
To answer this question in humans, one would have to study at least 100 individuals during substance use to obtain 4-week follow-up data in 10 abstinent alcoholics who also abstained from smoking for about the same time. This is of course because of the relatively low success rate for smoking cessation that lasts longer than 7 days (a typical clinical trial endpoint for measuring smoking cessation); a few weeks of substance abstinence are needed for effective neuroadaptations. In an effort to address the important question nevertheless, the authors conducted studies instead in a nicotine/alcohol model of non-human primates (NHP). Their results are published in the April issue of Alcoholism Clinical and Experimental Research (Hillmer et al. 2016). The authors describe results of well-thought out and well-conducted GABAA receptor imaging studies in a large population of 30 rhesus macaques that orally and separately self-administered alcohol and/or nicotine for 20 weeks and then were withdrawn from alcohol and/or nicotine to mimic human treatment conditions. The macaques were imaged with [11C]flumazenil PET before any drug administration and then again after 1 day, 8 days and 12 weeks into withdrawal, for a total of 4 serial measures in each animal and an impressive 120 PET studies total. General linear mixed modeling analyses found no effects of withdrawal group on cortical binding potential across time; binding potential is a well-accepted measure of GABAA receptor availability. The finding indicates that it makes no difference to GABAA receptor availability at the different time points if the macaques are withdrawn from alcohol only (but continue nicotine self-administration), are withdrawn from alcohol and nicotine simultaneously, or are withdrawn from alcohol alone. Or in other words, withdrawal condition across the three alcohol-consuming groups does not appear to affect GABAA receptor availability differentially over time. This prompted the investigators to pool the data. The pooled data showed a significant effect of time, with cortical binding potential significantly higher than baseline at 1 and 8 days of withdrawal and then falling back to baseline levels at 12 weeks. In a subsample of macaques, the authors also obtained data at 4 weeks of abstinence, when they detected no difference from baseline levels (the numerically negative values are in the noise level for this type of [11C]flumazenil PET measurements). Thus, the study indicates that GABAA receptor levels normalize after about 1 month of abstinence independent of withdrawal group, which is consistent with the investigators’ previous findings in abstinent alcohol dependent individuals who do not smoke cigarettes (Cosgrove et al. 2014).
In addition, cortical binding potential in the macaques did neither correlate with alcohol nor nicotine quantities consumed during the recent 90 days of the 20-week administration phase. This lack of correlations is important, because the two nicotine-consuming groups self-administered quite different amounts of nicotine during free access (but similar amounts of alcohol). Furthermore, the authors detected no meaningful correlations between binding potential and peripheral inflammatory cytokine levels, and they contributed this lack of correlations to the large variance in the cytokine data and the exploratory nature of the evaluation.
This NHP data addressed a long-standing clinical question: Does it make a difference to substance craving and treatment outcome if individuals withdraw from alcohol and nicotine separately or simultaneously? And the simple answer was no; at least as it concerns GABAA receptor levels and associated craving for these substances, this new study suggests that the withdrawal pattern does not appear to matter. In addition, as nicotine consumption alone and withdrawal from nicotine alone were not associated with significant changes in GABAA receptor levels, the authors concluded that nicotine replacement therapies may not alter GABAA receptor levels inadvertently. These are important findings and conclusions, which may even call for re-interpretation of previously published work as pointed out by the authors. The current findings and their conclusions were obtained through careful, thorough, large-scale NHP laboratory experimentation by experienced investigators. The experiments and analyses appear to have been carried out with proper consideration of perceived limitations by investigators who clearly demonstrate a grasp of the field, as demonstrated by their previous work and by a recent review of the literature on neurotransmitter imaging as an important tool for evaluating the neurochemical basis of alcohol dependence (Hillmer et al. 2015). The statistics in this new report properly considered the potential for non-normally distributed data (both imaging and correlations). The selection of figures beautifully supports the main message(s) of the article, limitations are discussed (e.g., pons as reference region in [11C]flumazenil scans), proper conclusions are drawn, and the studies and analyses have been set well in the context of previous work. All findings are reported in a clear and concise manner and the discussion points follow logically, so that the paper is a joy to read.
The reported results appear to have settled a longstanding clinical question – at least to the degree that the NHP experiments mimic the human condition. However, therein lays the crux: When we compare the authors’ previously reported human PET data (Cosgrove et al. 2014) to their current NHP PET data of similar withdrawal group behavior, differences emerge that should signal caution when interpreting the overall work. The clinical data indicated in abstinent alcohol-dependent non-smokers significant increases in cortical GABAA receptor availability at about 1 week of abstinence (which was associated with higher craving for both alcohol and tobacco) and normalization after 1 month of abstinence; however, no such normalization of elevated cortical GABAA receptor availability was observed in the generally larger group of abstinent alcohol-dependent smokers over the same time, when craving for alcohol was also high. Based on these differences between the nicotine/NHP model and the human condition, the authors conclude that something other than nicotine in cigarette smoke must prevent normalization of cortical GABAA receptor availability during alcohol abstinence in smokers. They briefly mention harmala alkaloids and CO and reference corresponding papers. This part of the discussion amounts to a tacit admission of the inappropriateness of the nicotine model to mimic cigarette smoking in laboratory experiments, leading the authors to recommend confirmation of their results in future human studies. Whether these differences between the human study and the animal model have a bearing on the translation of the animal study results to the human condition is unclear.
Despite these limitations of the nicotine animal model, animal studies clearly do have a place in our research: they allow assessments in an initial drug-naïve state virtually impossible to perform in all but the largest serial human studies, complete control over the experimental environment with isolation of a single pharmacological agent (here nicotine) as it acts on a single biological target (here GABAA receptor), control of drug exposure, and execution of serial studies without drop-out from change in substance use behavior or free will.
Nevertheless, and as much as I admire the execution of a complex study design with 4 different NHP groups and 4 serial assessments, a critical question arises: Are we ignoring smoke at our peril? Although nicotine is an important and critical component of cigarette smoke – one that has abuse potential and keeps smokers hooked on tobacco – it is only one of more than 4000 compounds in cigarette smoke (see e.g., review by Durazzo et al. 2010). Smoking clearly does so much more to the human body than just deliver nicotine. Cigarette smoke is impossible to mimic properly in the laboratory, except, of course, through smoke itself. Cigarette smoke exposure chambers have been used with small animals in previous research (e.g., Burns and Proctor 2013; Colli Neto et al. 2014; Duvareille et al. 2010), and it may be time that they be put to wider use in our substance abuse research communities for a better understanding of the effects and correlates of tobacco smoking, including nicotine dependence.
Seen in the wider scientific context of neurotransmission, the time course of human GABAA receptor levels over 1 month of abstinence from alcohol (and/or nicotine) determined by Cosgrove et al. in their reports (Cosgrove et al. 2014, Hillmer et al. 2016) is consistent with magnetic resonance spectroscopy (MRS) studies of cortical glutamate levels in alcohol dependence. We and others reported normalization of initially elevated cortical glutamate levels (during withdrawal) within alcohol dependent treatment seekers over about 1 month of abstinence (Hermann et al. 2012; Mon et al. DAD 2012). The GABAA receptor data, including the negative binding potential at 4 weeks of abstinence, may also be consistent with a waxing and waning of binding potential during the time course of abstinence, reminiscent of the time course of MRS detectable glutamate levels during alcohol abstinence. Taken together, the glutamate studies paint a picture of possible sinusoidal and damped modulation of glutamate levels until they normalize after about 1 month of abstinence. Corresponding MRS-observable cortical GABA levels have not been investigated as widely or with respect to their time course during withdrawal and extended abstinence, presumably due to greater experimental difficulties in measuring GABA by MRS. The little that is known shows similar cortical GABA levels at 1 and 5 weeks of abstinence in alcohol dependent individuals, independent of smoking history (Mon et al. 2012; Mason et al. 2006).
The new [11C]flumazenil PET report in macaques also speak to an issue that has been discussed anecdotally by substance use researchers involved in treatment. The authors observed dramatic increases in nicotine self-administration by NHP while withdrawn from alcohol after a long period of self-administration of both substances. This is in contrast to findings in human smokers, who generally do not change tobacco consumption during substance abuse treatment (McClure et al. 2015) or extended abstinence from alcohol (Schmidt et al. 2014). The observed increases in nicotine consumption with alcohol abstinence may be unique to the animal model, as the animals had experienced no drug exposure at all before the experiments, whereas their human counterparts over the years have learned to titrate their own nicotine exposure within a relatively narrow range, which they do not even deviate from substantially during substance abstinence.
In summary, this new report by Hillmer et al. describes a nicotine NHP model that addresses neuroadaptations in GABAA receptor availability as a function of nicotine consumption and withdrawal in comorbid alcohol dependence. The NHP report shows no differential effect of withdrawal pattern (alcohol abstinence alone or simultaneous abstinence from nicotine and alcohol) on receptor levels, which have been related to craving for alcohol and cigarettes. Thus, the studies are of strong clinical relevance. They also suggest nicotine-based therapeutics can be used to treat tobacco and alcohol consumption simultaneously in patients with alcohol use disorders without altering GABAA receptor availability. Nevertheless, differences in cigarette/nicotine consumption during abstinence from alcohol and differences in GABAA receptor levels at 1 month of abstinence between smoking alcohol dependent humans and the nicotine/NHP model strongly suggest that nicotine in laboratory experiments is not a good proxy of cigarette smoking. Going forward, we substance use researchers ignore smoking at our peril.
NIH AA10788 (DJM)
“I have no interest or relationship, financial or otherwise, that might be perceived as a potential source of conflict of interest.”
Commentary on “Nicotine and nicotine abstinence do not interfere with GABAA receptor neuroadaptations during alcohol abstinence” by Hillmer at al., this issue of ACER