by Lewis Thomas
Although the city’s officers never insisted that the Health Research Council restrict its sponsorship to research projects directly relevant to the city’s health problems, and a large amount of valuable basic research was accomplished, unrelated to any special disease, everyone involved was aware of the general intention of the fellowship fund. Deliberate efforts were made, therefore, to launch work on several practical problems. One of these was heroin addiction.
By the early 1960s, heroin was generally agreed to be the public health problem of greatest concern to the city, and a formidable economic problem as well. More than five hundred deaths occurred each year among adolescents and young adults alone, and a considerably larger number from homicides and accidents related to heroin. Acute hepatitis spread through the addict population and then beyond, due to contaminated needles. Bacterial endocarditis, pulmonary disease, malaria, and chronic liver disease were attributed to heroin use. Babies were born addicted, at high risk of early death. As to the cost, it was figured that well over $1 billion was lost to city residents each year from thefts to raise funds for heroin. Each addict required at least $50 a day to sustain an average sort of habit, at the prices of that time.
I had become interested in the problem and written some worried essays about it, because of seeing so much of it at first hand among adolescents in the Bellevue wards. I was appointed chairman of an ad hoc committee of the Health Research Council charged with studying the matter and returning several months later with recommendations. But this was in the spring of 1962, at a time when I had just accepted a sabbatical appointment as visiting professor of bacteriology at the University of Edinburgh, starting in July. The committee met several times, during the spring, finding the heroin problem more complicated at each meeting and finding, as well, very few openings for real research work. We were not even sure it was a medical problem as much as a social dilemma. I persuaded Eric Simon, a young Ph.D. biochemist in the Department of Medicine at NYU, to take an interest in opiate addiction as a biochemical problem, and placed a laboratory and fellowship at his disposal. It was a considerable gamble for him; to leave his present line of work and take on a brand-new, obviously intractable problem was a high risk. He took it on with some zest, however, and kept at it for the next twenty years. Simon spotted, early on, the best of openings: was there a special receptor for morphine and its derivatives somewhere in the brain? Using compounds labeled with radioactive tracers injected into rats, he discovered that specific receptor cells do indeed exist, in particular cells of the midbrain, and that the attachment of the drug to the surfaces of these cells could be prevented by morphine antagonists. The work of Simon and a few others led, eventually, to the discovery that the brain manufactures its own morphine-like compound, endorphin, to fit with those same receptors.
My other, still more indirect contribution to the heroin problem was to go off to Edinburgh for my sabbatical on schedule. This meant that someone else would have to serve as chairman of the ad hoc committee in the year of my absence. I telephoned Vincent Dole, a friend of mine and an elegant investigator at the Rockefeller Institute with broad interests and experience in metabolic disease, and asked him please to become chairman. He demurred at first, pointing out sensibly that his work was nowhere near the heroin problem and he knew nothing about the matter. But then he agreed to chair the meetings, and off I went to Scotland. During that year, Dole served first as a conscientious chairman, then as a fascinated student of the field, and, before the year was out, as an irresistible, unstoppable scientist at work. He dropped his other research and plunged into laboratory investigations of all aspects of heroin addiction, emerging before long with the idea—and then a clinic for testing the idea at Rockefeller—that methadone might be the ideal drug to block the episodic highs of heroin addicts without other side reactions, thus enabling the addicts to get off heroin and on to other ways of living. Dole’s work still stands as one of the stunning successes of clinical research, and remains the most practical and effective way of dealing with heroin addiction.
As a sort of spin-off, Dole and his wife, Marie Nyswander, became interested in the city’s prisons. I visited the House of Detention in downtown Manhattan with Vincent one afternoon several years later, where he was running a methadone clinic, and discovered that he had good friends on every cell block. Prison life had turned into a separate, half-intellectual, half-emotional concern for Dole. I suspect that sooner or later he will turn his attention more closely, and emerge again with a good idea for changing the jails. It would fit nicely into the original objective of the Health Research Council—to enlist good scientific minds for worrying about matters crucial for New York’s future—but in the meantime the council languished and died for want of money at the time of the great New York budget crisis in the mid-1970s. A ghostly form of the institution floated off to Albany and still exists, on paper, as a state council, but the city has lost it, perhaps for good. I hope not, though. When the money returns, if it does, and the city once again becomes affluent enough to be ambitious, and if anyone asks me, I would vote to restore the Health Research Council—still at a dollar per citizen—before doing anything else.
14
ENDOTOXIN
Dr. Oliver Wendell Holmes once laid out the dictum that the key to longevity was to have a chronic incurable disease and take good care of it. Even now, 150 years later, this works. If you have chronic arthritis you are likely to take a certain amount of aspirin most days of your life, and this may reduce your chances of dropping dead from coronary thrombosis. When you are chronically ill, you are also, I suppose, less likely to drive an automobile, or climb ladders, or fall down the cellar stairs carrying books needing storage, or smoke too much, or drink a lot.
In research, something like the Holmes rule holds if you readjust the words. The key to a long, contented life in the laboratory is to have a chronic insoluble problem and keep working at it. But this does not mean staying out of trouble. On the contrary, it means endless, chancy experiments, one after another, done in puzzlement. It is worth it, for this is the way new things are uncovered, whether or not—and usually it turns out not—they illuminate parts of your problem. But nothing ever gets settled once and for all when you work this way.
The best example I know of is the near-century-long, still uncompleted study of endotoxin, in laboratories all around the world. I know about it at first hand, for I have profited from it in my own laboratory because of having trapped myself into starting work on endotoxin thirty-five years ago and having never succeeded in stopping, never even settling conclusively any major part of the matter at hand.
Endotoxin had its beginning as a biomedical problem with the first attempts, early in this century, to make a vaccine against typhoid fever. Typhoid vaccine quickly became famous, but not for any remarkable capacity to prevent typhoid fever. Indeed, although it remains in more or less routine use to this day, in much the original form of a crude suspension of heat-killed typhoid bacilli whose walls are rich in endotoxin, its effectiveness has always been marginal at best. The most spectacular and deeply interesting property of typhoid vaccine has, from the beginning, been its capacity to cause fever. The term “pyrogen” was instantly coined, and the massive literature on what started out as an inconvenient side issue began to accumulate. Today, the bibliography of scientific papers dealing with endotoxin and its biological and chemical properties numbers in the tens of thousands. From time to time long reviews of the field are written, usually from one or another specialized point of view—for example, the chemical structure of the lipopolysaccharide molecule in the bacterial cell wall, which is, in fact, the endotoxin—and the accumulated reviews themselves comprise a formidable body of reading, far beyond any single reader’s endurance.
For all the intensity and scale of the work, it is still not known how endotoxin causes fever. Even more puzzling is the array of other intoxicating effects, ranging from scattered hemorrhages throughout many o
rgans of the body to overwhelming shock (similar in its manifestations to traumatic shock) and death. Many of the studies have been done with rabbits, which are the most susceptible of conventional laboratory animals, but it is known from observations in human beings injected with tiny quantities of endotoxin, and also from the reactions of patients suffering from typhoid or related infections, that man is even more vulnerable, perhaps the most sensitive to endotoxin of all animals. Less than one millionth of a milligram will produce shaking chills and high fever in a normal adult.
The center of the puzzle is that endotoxin is really not much of a toxin, at least in the ordinary sense of being a direct poisoner of living cells. Instead, it seems to be a sort of signal, a piece of misleading news. When injected into the bloodstream, it conveys propaganda, announcing that typhoid bacilli in great numbers (or other related bacteria) are on the scene, and a number of defense mechanisms are automatically switched on, all at once. When the dose of endotoxin is sufficiently high, these defense mechanisms, acting in concert or in sequence, launch a stereotyped set of physiological responses, including fever, malaise, hemorrhage, collapse, shock, coma, and death. It is something like an explosion in a munitions factory.
This is the reason, or one reason anyway, why the problem of endotoxin is so engrossing. It provides a working model for one of the great subversive ideas in medicine: that disease can result from the normal functioning of the body’s own mechanisms for protecting itself, when these are turned on simultaneously and too exuberantly, with tissue suicide at the end.
You can observe this sort of thing in one of the tricks that experimental pathologists have played with endotoxin for many years, known as the Shwartzman phenomenon since its original discovery by Gregory Shwartzman in 1928. A small quantity of endotoxin is injected into the abdominal skin of a rabbit, not enough to make the animal sick, but just enough to cause mild, localized inflammation at the infected site, a pink area the size of a quarter. If nothing else is done the inflammation subsides and vanishes after a day. But if you wait about eighteen hours after the skin injection, and then inject a small non-toxic dose of endotoxin into one of the rabbit’s ear veins, something fantastic happens: within the next two hours, small, pinpoint areas of bleeding appear in the prepared skin, and these enlarge and coalesce until the whole area, the size of a silver dollar, is converted into a solid mass of deep-blue hemorrhage and necrosis.
The Shwartzman reaction is a kind of pathologic ritual, with tight rules. The skin must be injected first, then the intravenous injection, and the time interval between the injections is crucially fixed at between eighteen and twenty-four hours. When the intravenous injection is made earlier or later, nothing happens. If the injections are made in reverse, first by vein and then in the skin, nothing happens. If both injections are made into the skin, nothing happens.
I heard about this phenomenon when I was a third-year medical student at Harvard in 1936, and began reading everything I could find about it and about endotoxin. I began to think up things I’d like to do with it, and became obsessed with possibilities. Here was a way that a particular area of tissue could undergo its own devastating disease at its own hands, so to speak, by its own devices. It required only a special combination of events, each of which, by itself, was harmless. It seemed a model for all sorts of events that might be occurring in the infectious diseases then commonplace and insoluble problems for medicine. I wondered, among other things, what would happen to a rabbit if both injections were given by vein, spaced eighteen to twenty-four hours apart. Could the internal organs, any of them, be “prepared” for the reaction? And then, of course, I made the usual illuminating discovery, the commonest epiphany in research: someone else had done my experiment. Two German pathologists, several years earlier, had given both injections by vein, and the outcome was an extraordinary disease called bilateral cortical necrosis of the kidneys. The photographs were impressive; I can still remember them on the right-hand page of the pathology journal: the outer portion of both kidneys entirely destroyed by dense black zones of hemorrhage and necrosis. This phenomenon is now known as the generalized Shwartzman reaction.
Several days after seeing that paper, I had the experience which turned my mind toward research in experimental pathology. It was a Thursday afternoon, and I was sitting at a conference table in Professor Tracy Mallory’s office at the Massachusetts General Hospital, at the weekly seminar in Mallory’s elective course in advanced pathology. I forget what was being talked about, but I remember leaning back in my chair, bumping my head against a heavy glass jar on the shelf of tissue specimens behind me, and knocking it over. I picked it up to replace it, and saw that it contained a pair of human kidneys with precisely the same lesion as the one in the photograph. The label said that the organs were from a woman who had died in eclampsia, with severe bacterial infection as well.
The M.D. program was not then, and still is not, very satisfactory training for research in biomedical science. Then, as now, the Ph.D. program provided a much more rigorous and profound experience in science, with a better grounding in the basic fields of biology needed for medical research. Earning an M.D. has, however, one enormous advantage which makes up in part for its deficiencies. After four years of medical school it is impossible to think about a problem in biology, or to read a paper, without having part of one’s mind trying at the same time to make connections with human disease. The intrinsically amazing aspects of the generalized Shwartzman phenomenon were enough by themselves to catch my interest and make me hanker to work on it, but the human tissue in that glass jar sent me over the line, and from that day on I was resolved to turn myself, one way or another, into an investigator of that queer reaction.
Death is the most familiar of events in living organisms, long before the creature as a whole dies. If you were a conscious member of the replicating mass of developing cells in an embryo, you would probably be horrified by the wholesale slaughter of cells on every side. The piecing together of a fetus involves a great deal of obsessive editing. Structures are meticulously put together, like the forerunner of the kidney, and then as though in an afterthought these cells are destroyed and a new, more advanced kind of kidney is installed. In the process of building the modular columns of nerve cells for the cortex of the brain, many more neurones are produced than can be fitted into the needed circuits, and these redundant cells must be killed off before the final, perfect electronic device can be put in operation. During all of adult life, dying and replacing goes on at a great pace in many vital organs; the lining cells of the intestinal tract, the blood cells, and the skin are the busiest, but there is even one class of proper brain cells, the olfactory neurones in the lining of the nose, where death and replacement by new cells takes place at about three-week intervals.
With such mechanisms at work everywhere, it would be no surprise to learn of occasions when the self-destructive devices are fumbled with, turned loose at the wrong times, with disease as the outcome.
In the Shwartzman phenomenon, cell death is caused by a shutting off of the blood supply to the target tissue. After the second injection, the small veins and capillaries in the prepared skin area become plugged by dense masses of blood platelets and white cells, all stuck to each other and to the lining of the vessels; behind these clumped cells the blood clots, and the tissue dies of a sort of strangulation. Then the blood vessels suddenly dilate, the plugs move away into the larger veins just ahead, the walls of the necrotic capillaries burst, and the tissue is filled up, engorged by the hemorrhage.
The same sort of sequence takes place in the kidney in the generalized Shwartzman phenomenon. The smallest vessels, those in the glomerular capillary tufts, are plugged by tiny thrombi, the surrounding cells die of oxygen deprivation, the vessels themselves burst, and bilateral cortical necrosis is the end result.
My colleagues and I tried our hand at sorting out the participants in these catastrophic events, over the next ten years, a
t Johns Hopkins, Tulane, and the University of Minnesota. The first nice observation was that when the circulating white blood cells are lifted out of the picture, just before the expected occlusion of vessels, the phenomenon is completely prevented; this is accomplished by treating the rabbits with precisely the dose of nitrogen mustard needed to eliminate the leukocytes for a period of twenty-four hours. Next, we found that the temporary inhibition of blood clotting, by a properly timed injection of heparin, prevented the phenomenon. Cortisone didn’t work, which was an odd thing in view of the fact that cortisone will totally protect animals against the lethal, shock-producing effects of endotoxin.
We never did succeed in learning how the white blood cells contribute to the phenomenon, beyond their obvious role in blocking the flow of blood, nor did we find out how the blood vessels were induced to burst their walls.
But in the course of trying, we ran into a number of other things, unrelated, as it turned out, to the Shwartzman phenomenon but interesting enough in themselves, worth, as the Michelin travel guide says, a detour. One of these was the surprising action of papain. It occurred to us that the release of a proteolytic enzyme by damaged tissue cells might be one way of rupturing small vessels, and we guessed that such an enzyme might be the sort most active in the reduced acid environment which we knew existed in the rabbit’s prepared skin. So, without much hard thinking, we injected small amounts of a plant enzyme of this type, papain (from papaya latex), into rabbit skin, and within an hour had a fair copy of the hemorrhagic necrosis of the local Shwartzman phenomenon.