D Plenary Lecture II
Ivan Diamond, M.D., Ph.D.
Professor and Vice-Chair, Department of Neurology, University of California at San Francisco and Director, Ernest Gallo Clinic and Research Center
Thank you very much for asking me to address the workshop. First, I will explain how I got started in alcohol research and afterward I will describe some of our more exciting observations.
Fourteen years ago, when the Gallo Center was organized, I had never done an experiment on alcohol in my life. I was a professor at UCSF when Mr. Gallo proposed that we build a center devoted to the neural basis of alcoholism and alcoholic brain disorders. I was asked to create a program that would be unique, that would use new strategies to investigate alcoholism, and that would not merely duplicate what was already going on in alcohol research. How does one start such a program? I surveyed the alcohol research community in the United States and learned that no one was exploiting the power of molecular and cellular neuroscience to investigate alcoholism. Therefore, I decided to develop a program designed to uncover fundamental molecular mechanisms of intoxication, tolerance, and dependence with the expectation that this information would lead to new therapies for alcohol addiction and alcoholic neurologic disorders.
I was not able to identify young investigators in alcohol research who had similar interests. Therefore, with the help of my colleague, Dr. Adrienne Gordon, we recruited faculty who did not have experience in alcohol research, but who recognized an exciting opportunity to use cellular and molecular biology to investigate alcoholism and its medical complications. In addition, because we were based in a clinical department, I wanted to be sure that our work would have relevance to clinical problems. Therefore, we brought together clinical and basic scientists to create a critical mass of investigators, just as Avram Goldstein described in his address this morning.
Our strategy was straightforward. First, we wanted to use simple, homogeneous cellular systems to identify the most important molecular mechanisms that mediate the response of neural cells to alcohol. Whenever we could, we would extend our findings to circulating blood cells taken from alcoholic subjects to confirm the relevance of our observations. Once we were able to measure important cellular and molecular events, I thought we would be in a much better position to apply this information to the brain with all its complexity and heterogeneity.
What about the problem of alcoholism in our society? Two-thirds of the adults over 14 years of age in the United States drink alcohol. If you calculate how much alcohol is produced in America and divide it by the number of people who drink, the quantity of alcohol consumed is roughly equivalent to 10 gallons of whisky per person per year. Now, this is clearly not the case for each American. Instead, about 10 percent of the drinking population, or perhaps 7 percent of the public, consume nearly 50 percent of the alcohol produced in this country. These are the alcoholics we encounter in hospitals and clinics; a conservative estimate is that about 25 percent of hospital beds in the United States are occupied by patients who have alcohol-related problems. In San Francisco, I estimate that as much as 75 percent of patients at San Francisco General Hospital have alcohol-related medical diagnoses. Medical complications involve the liver, heart, and just about every organ system in the body. When you add socioeconomic costs to the medical costs, the burden of alcoholism and alcohol abuse to American society is estimated to be more than $100 billion per year. Parenthetically, this enormous cost to society is much greater than the cost of other major medical problems, such as heart disease or stroke. Yet, the percentage of the National Institutes of Health's budget devoted to alcohol research is trivial when compared to the research budgets for other research areas.
What happens to people who drink excessively? Everyone is aware of the characteristic intoxicated behavior produced by alcohol. The degree of intoxication can be correlated with blood alcohol levels, because there is a very rapid equilibration between alcohol in the blood and in the brain. Therefore, blood alcohol levels accurately reflect brain alcohol concentrations. Alcohol is a sedating agent and, when blood levels reach 500 mg percent, naive individuals can become comatose and even die because of respiratory depression. This happens not because alcohol destroys neurons, but because alcohol depresses neuronal function in the respiratory center of the brain, so that breathing stops.
Yet, not everything is as simple as these correlations suggest. Many years ago, Mirsky found that blood levels can be misleading. Here, I reproduce some of his studies with volunteers given alcohol to drink. After volunteers were given a few drinks in about an hour, intoxication developed at blood levels approximating 170 mg percent. These subjects would be considered "drunk" by most conventional tests. And yet, if the same individuals were given more alcohol and examined 6 hours later, they were considered to be "sober" even though
their blood levels were now much higher, approximating 300 mg percent. Since the blood alcohol level tells us the concentration of alcohol in the brain, this improvement in behavior cannot be due to less alcohol in the nervous system. Instead, something happened in the brain to accommodate and adapt to the presence of ethanol. This is a short-term or acute tolerance. We also encounter more striking long-term tolerance to alcohol in every emergency department across the country. Here I illustrate a study of patients who came to an emergency department for medical care. They were asked the question, "Have you had a drink in the last 6 hours?" If the answer was yes, blood alcohol levels were measured. The results show that the average was 270 mg percent, and a few people had blood ethanol levels greater than 500 mg percent, concentrations that can cause coma or death. What I didn't tell you is that all of these people were considered to be "sober" on crude physical examination. Clearly, these patients were chronic alcoholics with remarkable tolerance to the intoxicating effects of ethanol. If you are interested in the world's record for blood alcohol levels, it was measured at UCLA in a woman who walked into the emergency department after having discontinued drinking 3 days earlier. She had symptoms of withdrawal and her blood level was about 1,500 mg percent. Clearly, some alcoholics exhibit a remarkable ability to tolerate tremendous amounts of alcohol.
Alcoholics begin to experience symptoms and signs of alcohol withdrawal when more than 6 hours elapses after the last drink. That is why alcoholics tend to have a drink first thing in the morning. Alcohol suppresses these symptoms. Perhaps the craving for a drink during alcohol withdrawal is the same craving responsible for alcohol addiction. Alcohol withdrawal is characterized by hyperexcitability, and the most dangerous problem is alcohol withdrawal seizures; these occur in the first 24 to 48 hours. Later, of course, withdrawing alcoholics may develop delirium tremens, a well-known hyperexcitable syndrome with dramatic symptoms and signs.
We begin to wonder how we could approach such clinical phenomena at a cellular and molecular level. We thought it would be possible to identify cellular tolerance as a reduced response to repeated doses of ethanol, or cellular dependence by an abnormal cellular response during ethanol withdrawal that would be corrected by returning alcohol to the cells. We succeeded in both instances.
Robert Messing at the Gallo Center was interested in the molecular basis of hyperexcitability, particularly alcohol withdrawal seizures. He discovered that neural cells in culture adapt to ethanol by increasing the concentration and activity of voltage-dependent calcium channels. Ordinarily, ethanol inhibited calcium flux through these channels, but when ethanol was removed from the cells, the increased concentration of channels mediated a tremendous increase in calcium flux during alcohol withdrawal. This could contribute to neuronal hyperexcitability. The advantage of working with a cellular system is that it is possible to identify the molecular mechanisms that underlie these functional adaptations. For example, Bob Messing discovered that protein kinase C was required for ethanol to induce up-regulation of the voltage-dependent calcium channel.
Indeed, levels of two specific isozymes of protein kinase C, d and e, increased dramatically with chronic exposure to ethanol, and it seems likely that one of these isozymes, probably d, is required to mediate the action of ethanol on calcium channel up-regulation. These results indicate that an ethanol-induced increase in gene expression for a particular protein kinase C isozyme may set the stage for the molecular events that lead to alcohol withdrawal seizures. Ongoing studies in several laboratories now suggest that calcium channel blockers can prevent alcohol withdrawal seizures in experimental animals and alcoholics. The point of this illustration is that it would have been very difficult to identify this specific molecular mechanism if whole brain or heterogeneous brain preparations were used initially. Now, however, it is possible to produce transgenic animals with overexpression or alteration of specific protein kinase C isozymes to confirm their roles in specific adaptive responses to ethanol. This work should suggest that it is not possible to explain neural responses to ethanol merely on the basis of membrane perturbation and fluidity changes. Instead, it seems that regulatory mechanisms affecting specific membrane proteins are very special targets for ethanol in the brain.
Another example of a regulatory system affected by alcohol was developed in collaboration with Adrienne Gordon over the past several years. We have been interested in cyclic AMP signal transduction. When a neurotransmitter reacts with its receptor, as in this cartoon, adenylyl cyclase is activated via a G protein, Gas. The result is an increased production of cyclic AMP which then stimulates protein kinase A activity. With long-term exposure to ethanol, however, neural cells adapt by reducing cyclic AMP signal transduction. This desensitization affects all receptors coupled positively to adenylyl cyclase. The advantage of working with cells in culture was that it allowed investigators to determine the molecular mechanism responsible for these changes. We found that long-term exposure to ethanol caused a selective reduction in gene expression for Gas, and thus decreased production of Gas mRNA and protein. This accounts for heterologous desensitization of signal transduction. Interestingly, these changes at a cellular level mimic physical dependence. Receptor-stimulated cyclic AMP levels are abnormally low during alcohol withdrawal and can be restored to normal by adding ethanol back to the cells.
Identification of these short- and long-term neural responses to ethanol made it possible to discover the molecular mechanisms that regulate these events. This slide provides an overview of the pathway we have identified. Ethanol inhibits a specific adenosine transporter to block re-uptake of adenosine into neural cells. As a result, cells exposed to ethanol accumulate extracellular adenosine, which then reacts with adenosine receptors on the cell surface. In this case, adenosine A2 receptors positively coupled to adenylyl cyclase stimulate the production of cyclic AMP to activate protein kinase A. This results in a heterologous desensitization of cyclic AMP production associated with diminished protein kinase A activity at the cellular membrane. As a consequence of reduced
phosphorylation, the adenosine transporter becomes tolerant to ethanol inhibition. In other words, two molecular mechanisms, one, a model of dependence and another, a model of tolerance, are linked during cellular adaptive responses to ethanol. And yet, each can be studied separately, as I will show later.
I have told you about studies in cells. What does this mean in the real world of alcoholism? We have discovered the same mechanisms at work in alcoholics. The same kinds of changes are demonstrable in lymphocytes from actively drinking alcoholics; there is desensitization of cyclic AMP signal transduction and tolerance of the adenosine transporter to ethanol inhibition. This is most striking in erythrocytes from alcoholics where there is virtually no ethanol inhibition of adenosine uptake. All of these studies illustrate that mechanisms discovered in model cell systems in the laboratory have direct relevance to pathogenetic mechanisms in human beings.
We have pursued cellular tolerance and dependence in studies too extensive to detail here. First, we tried to determine how the transporter, which mediates ethanol tolerance, is regulated during adaptation to ethanol. As this slide illustrates, without going into experimental detail, we have discovered that the sensitivity of the transporter to ethanol inhibition is controlled by the level of protein kinase A-mediated phosphorylation. In turn, this appears to be regulated by protein kinase C (PKC) via a protein phosphatase. A theme emerging from our work and other laboratories doing alcohol research is that the primary targets of ethanol involve regulatory mechanisms, such as protein kinases and protein phosphatases. As a result, we began to think about protein kinase A and how it might be regulated by exposure to ethanol.
Recall that protein kinase A exists as an inactive holoenzyme with regulatory and catalytic subunits. When cyclic AMP binds to the regulatory subunits, the catalytic subunits are dissociated and free to catalyze phosphorylation at different sites in the cell. We were curious how ethanol affected this regulatory mechanism. If we examine color photographs of confocal microscopic images of neural cells showing the catalytic subunit of protein kinase A before and after treatment with ethanol, we see that under normal conditions, most of the enzyme is localized to the Golgi apparatus in the neuron. Cellular localization here is confirmed by concomitant localization with other Golgi markers and there is no evidence of the catalytic subunit in the nucleus of the cell.
In contrast, however, exposure to ethanol produces a dramatic translocation of protein kinase A catalytic subunit to the nucleus. The effect of ethanol is related to concentration and time. High concentrations produce translocation in hours; low concentrations take 2 to 3 days. With this illustration, you can see virtually complete migration of the catalytic subunit of protein kinase A from the Golgi apparatus into the nucleus; the nucleus becomes literally filled with the catalytic subunit and the enzyme remains there as long as alcohol is present. It would be of interest, then, to determine the consequences of protein kinase A translocation. First, a reduction in cytoplasmic protein kinase A explains the reduction in phosphorylation at membrane sites observed during adaptation to
ethanol. Second, nuclear localization of protein kinase A has great implications for changes in the regulation of gene expression. These kinds of changes may well underlie some of the chronic adaptive and sustained responses produced by ethanol in neurons and in the brain. It remains to be determined whether protein kinase translocation occurs in all neurons or is localized to specific neuronal populations in the brain, like the nucleus accumbens.
The long-term consequences of adaptive changes involve changes in gene expression, which probably underlie the development of complex abnormalities such as addiction and alcoholic neurologic disorders. Moreover, changes in gene expression may help to answer a puzzling question in alcohol research: How is it that short-term exposure to alcohol produces functional and metabolic changes whereas long-term exposure causes structural pathology and disease?
Because chronic exposure to ethanol produces changes in cellular and molecular function that require selective changes in gene expression, Michael Miles at the Gallo Center searched for evidence that specific kinds of genes are either ''turned on" or turned off" by ethanol. He has already identified a family of ethanol-responsive genes. In this slide, it is clear that there is a selective increase in gene expression for some genes, including several stress protein genes. In order to study the regulation of ethanol-responsive genes, Dr. Miles coupled the promoter from an ethanol responsive gene to a reporter enzyme, chloramphenicol acetyltransferase (CAT). Now, assays of CAT activity can be used to identify factors that confer ethanol sensitivity. This will take us a long way in determining important regulatory mechanisms that medicate ethanol sensitivity and designing new therapies specifically to prevent or reverse adverse responses.
So, in a few short years, investigators at the Gallo Center have moved from behavioral concepts such as tolerance and dependence to selective effects of ethanol on gene expression and the regulation of signal transduction mechanisms in the cell. Undoubtedly, these changes contribute to altered complex behaviors, such as addiction, and the development of alcoholic brain disorders, such as dementia. This should not be surprising since CNS responses are ultimately regulated by the genes that control neural cell function.
As we accumulate increasing evidence of a role for gene expression in responding to ethanol, and perhaps conferring vulnerability to alcoholism, we can exploit advances in complex organisms in which behaviors can be linked to gene expression more easily than in human beings. Here is a slide highlighting some ethanol-induced behaviors in one such preparation. Acute exposure to ethanol first produces incoordination, then hyperactivity, and finally drowsiness and sleep. After awakening, there is transient incoordination, followed by a complete recovery. Although these behaviors resemble the effects of alcohol in human beings, this preparation is not a mammal, it is a fruitfly. Ulrike Heberlein has developed a Drosophila genetics laboratory at the Gallo Center to identify genes that mediate ethanol responses and perhaps to identify candidate genes
that may be linked to genetic vulnerability for alcoholism in affected patients. We believe this new strategy will be of major importance in alcohol research because we know a great deal about the Drosophila genome, that genes controlling regulatory functions appear to be conserved throughout the animal kingdom, and that regulatory genes appear to be major targets for ethanol. Here, then, is a unique strategy: Exploit the validity of Drosophila as a model for learning, for behavior, and for responses to addictive agents. In the next few slides I want to show you how rapidly Ulrike Heberlein's work has progressed.
First, you need a way to measure ethanol sensitivity in fruitflies. For this, Dr. Heberlein uses something she calls an "inebriometer." The design is based on a model prepared by Howard Nash and Ken Weber to study anesthetic agents. You can see that the apparatus consists of a column about 4 feet, 6 inches long containing a series of incomplete soft platforms placed along the length of the column. The normal fly introduced into this column will remain at the top because of negative geotaxis. By contrast, a dead fly introduced into the top of the column would bounce down the platforms and come out at the bottom. So it is very easy to separate living flies from dead flies.
The column is prepared so that ethanol vapor can be introduced. Now flies become intoxicated, and, as they become increasingly uncoordinated, they fall to different levels in the column, holding on to different platforms as they fall down. When collected over time, the flies elute in a bell-shaped curve. This slide shows how the flies can then be separated by column chromatography according to their ethanol sensitivity. Populations of flies can be generated that are either very sensitive or resistant to ethanol intoxication.
One of the advantages of Drosophila is that it is possible to identify genes that control these kinds of behaviors. It is also possible to introduce random mutations into the genome and then screen for abnormal behaviors. If a behavior is identified, it is then possible to work back to the mutated gene. This approach is theoretically possible in animals, but is much more easily done with thousands of flies. Dr. Heberlein has introduced one mutation per chromosome and screened 50,000 flies. She has already identified more than a dozen mutants with increased or decreased sensitivity to ethanol and she should be identifying the responsible gene(s) in the next few months. Moreover, it is also possible to use available mutants as controls to prove that ethanol sensitivity is mediated by the nervous system and not due to an artifact. Perhaps this is best illustrated by the frequently asked question whether ethanol metabolism contributes to prolonged intoxication and increased sensitivity. Null mutants for alcohol dehydrogenase are available and when studied in the inebriometer, the elution profile is the same as wild-type flies. The only difference is that the eluted flies wake up slowly since they take a long time to clear ethanol because they are deficient in alcohol dehydrogenase. Clearly, sensitivity to ethanol in Drosophila cannot be explained by a failure to metabolize ethanol in the null mutants. In the slides to follow I show you examples of ethanol tolerance, ethanol sensitivity, and the fact that tolerance can be produced and studied for more than 20 hours after
exposure to ethanol. Unusual tolerance to intoxication in flies may be particularly important, since Marc Schuckit has shown that resistance to intoxicating effects of alcohol may be linked to genetic alcoholism and the development of alcoholism 10 years later.
Think of the potential of this strategy: not only is it possible to identify genes that mediate normal and abnormal CNS responses to ethanol, it will be possible to test candidate genes in Drosophila derived from other systems to determine whether they contribute to ethanol-induced changes in neural function. In addition, genes identified in Drosophila can be used to identify genes in human beings. Perhaps those genes play a major role in mediating ethanol responses in alcoholics. Finally, genes that mediate and modify ethanol responses may be candidate genes for genetic alcoholism. The great advantage of Drosophila is that it is possible to move rapidly back and forth from complex behaviors to genes of importance. This is not easily done with mice, rats, or human beings.