Miracle Cure

by William Rosen

  One of them was a nineteen-year-old onetime medical student from the University of Kiel named Gerhard Domagk. At the beginning of hostilities, he had enlisted in a Leibgrenadier Regiment from Frankfurt-von-Oder that was, literally, decimated at Ypres: more than one soldier in ten killed or wounded. With nothing salvageable of his unit, he was transferred to the eastern front and the German army’s medical department, named, with apparently no irony, the “Sanitary Service.” As might be expected, the field hospital to which Domagk was assigned in the Ukraine lacked something on the sanitary front. It also lacked any real weapon against the infectious diseases that killed as many soldiers during the First World War as gunfire: cholera, typhus, gangrene, dysentery, and a hundred more. Domagk would spend the rest of the war seeing medical impotence close up, treating patients and assisting in surgeries. By the end of 1918, and the Armistice, he had had more than enough of war, but not of medicine. Three years later, he graduated from Kiel as a fully accredited physician, and went to work in the Baltic port city’s main hospital.

  He recognized quickly that he was far better suited to a career as a researcher than as a clinician, and joined the Pathological Institute at the University of Greifswald, as a lecturer, a privatdozent, in 1924. There, under the institute’s director, the pathologist Walter Gross, whom he would follow to the University of Münster, he began his studies of the most powerful weapon (really, the only weapon) against infectious disease: the vertebrate immune system. In a series of experiments, he injected hundreds of mice with a known pathogen, the bacterium known as Staphylococcus aureus, which causes everything from skin infections to pneumonia, and then extracted cells from the lining of the animals’ livers known to be part of the innate immune system* to see how many of the staph bacteria they had gobbled up.

  Credit: Wellcome Library, London

  Gerhard Domagk, 1895–1964

  Domagk’s experiments resulted in two significant findings. First: The performance of the Kupffer cells improved with exposure to the staph pathogen (though he couldn’t know the reason, which wasn’t discovered until the 1980s: pattern recognition proteins known as Toll-like receptors that identify different sorts of toxins produced by bacteria and tailor the response by white blood cells sent to destroy them). The second revelation was the big one: A staph cell that had been weakened beforehand by exposure to antiseptics was easier for the Kupffer cells to destroy. The insight would determine the next thirty years of Domagk’s life. In 1927, he followed his boss Walter Gross once again to a job in the same industry that had served as patron to Ehrlich and Behring: the dye business.*

  Two years before, the German chemical and dye business had made a giant leap in the consolidation it had been pursuing before and especially after World War I. As far back as 1904, Carl Duisberg, the managing director of Bayer, had been lobbying his competitors on the value of combination, in the form of a fifty-eight-page memo laying out the advantages: lower costs, shared patents, lower risk, and much higher profits. He persisted through the First World War (during which Bayer’s U.S. subsidiary, known primarily for its Bayer Aspirin brand, was confiscated as enemy property) and the hyperinflation of the 1920s, in 1925 finally calling a conference—never one to shy away from grandiosity, he named it the “Council of the Gods”—at which the eight largest chemical firms in Germany, including Agfa, BASF, and Hoechst/Cassella, merged into one, to be known as the “Community of Interest of Dye Businesses”—in German Interessengemeinschaft Farbenindustrie, or I. G. Farben. Carl Bosch of BASF was its chief executive, Duisberg its chairman of the board.

  I. G. Farben was the largest chemical company in the world, and one of the largest of any sort, comparable in size to American companies like General Motors or U.S. Steel, with interests not only in dyes, but in photographic film, industrial solvents, and, with the acquisition of BASF, fertilizers. Bosch had developed a method for synthesizing ammonia—the Haber-Bosch process, the discovery of which would produce two separate Nobel Prizes in Chemistry, one for Fritz Haber in 1918, the other for Carl Bosch in 1931—that today produces 150 million metric tons of fertilizer annually and is responsible for feeding nearly half of the world’s population. Most relevantly for Domagk, I. G. Farben was also Germany’s largest manufacturer of pharmaceuticals, such as they were.

  They weren’t much. Though Behring’s antiserums and Ehrlich’s arsenicals were widely used (and highly profitable), and Bayer’s antiprotozoal drugs Plasmoquine and mepacrine (also known as Atabrine) were starting to be used as malaria treatments, the pharmacopoeia available to fight infectious disease wasn’t significantly greater than it had been when George Washington awakened on the last morning of his life. Nonetheless, the possibility of antibacterial drugs was still promising enough that Heinrich Hörlein, head of pharmaceutical research at Bayer, set Domagk up in a newly refurbished research lab that Carl Duisberg had built in Elberfeld, a Westphalian town just east of Düsseldorf.

  Duisberg and Hörlein believed they knew why no successful antibacterial drugs had appeared since Ehrlich’s Neosalvarsan a decade previously. It was the same argument behind the formation of I. G. Farben itself: scale. If Ehrlich had tested dozens of different recipes in order to find the antisyphilis treatment, Bayer would try hundreds. Or thousands. The way to make trial and error work was simply to increase the number of trials to the point that an effective compound would be guaranteed to emerge: to do for chemical innovation what Henry Ford had done for automobile manufacture.*

  And so they did. Beginning with Domagk’s arrival in 1927, two chemists from Bayer’s tropical medicine group, Josef Klarer and Fritz Mietzsch, started producing coal tar–based compounds with at least some antibacterial properties on the test-tube-and-ring stand equivalent of the assembly line at Ford’s River Rouge factory. By 1931, they had delivered more than three thousand such compounds to Domagk’s lab. There, the bacteriologist exposed them, one by one, to the family of Streptococci bacteria, pathogens responsible for diseases ranging from the familiar and moderate—strep throat, impetigo—to deadly diseases like toxic shock, streptococcal pneumonia, meningitis, and such exotica as the flesh-eating disease known as necrotizing fasciitis. Most of the compounds supplied for testing against strep were created by modifying chemicals that, as with Ehrlich’s early researches, were originally dyes that showed some affinity for a particular bacterium by staining it.

  Meanwhile, Domagk was working at an equally feverish clip to isolate a strain of Streptococci that was a consistent and reliable killer of laboratory mice. His purpose was to create a kind of ideal experimental bacterium, which could rapidly show the effectiveness—or ineffectiveness—of each of the compounds supplied by Klarer and Mietzsch. Over the first few years, his supercharged strep produced thousands of rodent corpses, each one accompanied by autopsy notes documenting the particular compound used to treat it, the progress and symptoms of the disease(s) at time of death, and the method by which the bacterium and the compound had been exposed to the unlucky rodent. The notes also, of course, had a place for showing which compound had a significant effect against the strep bacteria.

  For years, the last space remained stubbornly blank. A decade later, Iago Galdston, a physician and the secretary of the Medical Information Bureau of the New York Academy of Medicine, wrote, “By 1930 it was the universal opinion of physicians that nothing could be discovered which would be effective against the ordinary diseases produced by bacteria.”

  Less than a year later, in 1931, Mietzsch and Klarer started modifying azo dyes, believing them to be less toxic than aniline dyes, and therefore more promising. By summer, the promise appeared to be borne out. One of the azo-derived compounds—KL-487, for “Klarer, #487”—killed at least some of Domagk’s strep strain, without overwhelmingly toxic side effects. Others followed: KL-517, KL-529. The protocol employed looks, in retrospect, a lot like fairly arbitrary tinkering. The first attempts involved successively adding side chains composed of chlorine ato
ms, one at a time, to the azo base. When the possibilities of chlorine were exhausted, they moved on to arsenic, then to iodine.

  Eventually, on the advice of Hörlein himself, Klarer tried sulfur as the missing ingredient in an azo compound. The first one he tried, KL-695, used another important chemical included in the dye business known as sulfanilamide—formally para-amino-benzene-sulfonamide—that had been around since 1909. In late 1932, Domagk treated some more of his unfortunate mice with a compound that integrated sulfanilamide with an azo dye.

  After four years of failure, it’s not hard to imagine the exultation when the results of the new compound were revealed. Domagk had administered the sulfanilamide-plus-azo compound to twelve mice infected with his superstrep strain, with fourteen untreated mice as a control. Within a week, all fourteen untreated mice were dead, most within two days. All twelve that received the compound survived. Whether administered intravenously or orally, it completely cured strep infections in mice. By the time Klarer and Mietzsch had delivered KL-730 to Domagk, the conclusion was inescapable. The lab had found the world’s first successful antibacterial drug.

  Or, actually, drugs. Virtually all azo dyes with a sulfanilamide side chain worked on strep infections, which meant that patenting all conceivable variations would be a legal nightmare for Bayer. Unlike tangible goods, which can be sold only once, intellectual property can be sold to multiple consumers without diminishing inventory, a horn of plenty that never empties.* This means that patents are valuable to inventors because they alone gain the legal right to bar anyone else from profiting off the invention. In the early days of patents, legal protection stopped at national borders, making patents far more valuable to inventors working in a large country like France or Britain than in a relatively small one like the Netherlands or Switzerland. Patents also behave very differently on mechanical inventions than on chemicals. No one can sell a new engine while keeping its components secret from a competitor, but it is considerably harder to reverse engineer a dye or a drug. One consequence is that, although patent offices first started granting licenses in the seventeenth century, the Swiss, the Dutch, and, before the 1870s, the Germans, who specialized in chemical innovation, tended to avoid even seeking them. So long as secrecy could be preserved while dyes and other chemicals were being sold, patents offered far more risk than reward.

  As chemistry matured into a science that could decipher the secrets of drugs and dyes, techniques for analyzing the components and structures of new chemical compounds made it relatively easy to reproduce a competitor’s proprietary molecule. And once secrecy offered no protection to novel chemicals, patents became as necessary for them as they had long been for machinery. Because of their different histories with patent protection, the Swiss, Dutch, Germans, and even the French still guarded chemical innovation with systems very different from the British and American model. German chemical patents didn’t protect a novel product like a new chemical compound, but rather the process by which it was synthesized. Bayer didn’t need to patent every effective variant of the sulfanilamide-plus-azo compounds, but they did have to publish a detailed description of the method by which all of them were created, which was itself a risky proposition. There are multiple methods of chemical synthesis, and a competitor could, if clever enough, create a sufficiently different one, and so benefit from Bayer’s discovery while investing only a fraction of Bayer’s time and money. So when Bayer applied for a new patent on Domagk’s discovery in 1932, the application was quite deliberately as obscure as possible about the creation of KL-730 in order to protect its exclusivity.

  For the next two years, the new drug, which had been named Streptozon, was tested on both animal and human subjects. By the time the patent was granted, in 1934, Domagk had demonstrated that the drug worked against a number of nonstrep infections as well, including spinal meningitis, some strains of pneumococci, and gonorrhea. As a result, Bayer changed the name of the new drug. They called it Prontosil.*

  In 1935, Domagk finally announced his discovery in an article for the journal Deutsche Medizinische Wochenschrift, and the number of human trials expanded to include subjects in Great Britain. There, the first results were considerably less encouraging than the ones reported from the German tests, probably because in the United Kingdom it was tested on strains of streptococci different from the supercharged version developed by Domagk. Notably, however, one of them was S. pyogenes, the primary cause of what was then known as puerperal, or childbed, fever.

  The oldest known medical texts record the dangers, and frequency, of fevers suffered by women within hours of giving birth. The risks of contracting such a fever, however, skyrocketed with the growth of hospital births during the nineteenth century. By the time the Hungarian physician Ignaz Semmelweis, then working at Vienna General Hospital, published a study of childbed fever in 1847, it was attacking as many as four new mothers in ten. Tellingly, Semmelweis discovered that the risks of childbed fever were significantly lower in home births than obstetrics wards. The cause, in the days before Lister, was the way physicians practiced their craft: never washing their hands.* Improved hygiene and antisepsis reduced the numbers of victims substantially, but did not eliminate the risk of disease. Though Semmelweis and others had made puerperal fever less common, their technique was prevention, not cure. Throughout the 1920s, physicians tried Salvarsan and other arsenicals, failing miserably. As late as 1936, it was still attacking up to three hundred out of every ten thousand new mothers in the United States . . . and killing forty-nine of them. Until, that is, the discovery that Prontosil stopped S. pyogenes even more effectively than Ehrlich’s magic bullets targeted T. pallidum. Virtually overnight, mortality from childbed fever fell from 20 to 30 percent to 4.7 percent.

  By then, the power of Prontosil had already been demonstrated to the Bayer scientists in the most personal way possible. In December 1935, a needle was driven into the hand of Gerhard Domagk’s six-year-old daughter, Hildegarde. Lasting injury from the trauma was unlikely, but her risk of infection considerable, as the following days proved. An abscess formed, and Hildegarde’s temperature skyrocketed, peaking at more than 104°. Streptococci had entered the girl’s bloodstream. A year earlier, she would very probably have lost her arm, and possibly her life. But this was the year of Prontosil. A week after starting the therapy, her infection was beaten.

  The effectiveness of Prontosil was no longer in doubt. The mechanism by which it performed its magic, however, remained a mystery. Did the sulfa chain activate the azo dye, or the other way around? Why did Prontosil cure streptococcal infections in infected animals, but fail to kill strep bacteria in a test tube? Throughout Europe and the United States, biochemists tried to solve the puzzle, including Leonard Colebrook at St. Mary’s Hospital in London, and Ernest Fourneau, head of pharmaceutical research at the Institut Pasteur in Paris, both of whom immediately requested samples of Prontosil for testing. Colebrook’s request was granted. Fourneau’s was not.

  Fourneau’s history with Bayer generally, and Hörlein particularly, was less than collegial. Despite Pasteur’s early experience conducting research on behalf of the French wine and cheese industries, France had scrupulously avoided strong links between commerce and chemistry. The premier research facility in the country, the Institut Pasteur, had originally been funded by donations from French families with no expectation of a return on their investment. As a result, the institute, employing a few dozen researchers at most, was unable to produce anything like the number of new chemical compounds that were created by hundreds of Bayer and I. G. Farben scientists and technicians. Nonetheless, Bayer regarded Fourneau as a formidable competitor. For one thing, he was a superb chemist; the author of more than two hundred scientific papers, Fourneau had already developed a synthetic alternative to cocaine for his then-employer, Camille Poulenc of Établissement Poulenc Frères,* before joining the Institut Pasteur as its director of research in 1911. For another, he had the French patent system
on his side. Since Bayer drugs and dyes had no legal protection outside the borders of Germany, they were fair game for anyone with the energy to decode even the most opaque patent. Fourneau was nothing if not energetic; he devoured German academic journals and patent filings, attended German scientific meetings and trade fairs, and worked long hours building a file showing every niche where a homegrown drug might replace an import. In 1921, he had permanently alienated Duisberg and Hörlein when he reverse engineered Bayer’s proprietary treatment for sleeping sickness, originally introduced under the brand name Germanin, but sold in France by Poulenc as Stovarsol.

  Scarcely surprising, therefore, that when Hörlein received Fourneau’s request for samples of Prontosil, he stalled. Fourneau did not. He assembled a team of French chemists to copy the existing drug using the obscure language of Bayer’s patent, and by the middle of 1935, Rubiazol, the French version of Prontosil, was on the market.

  To Bayer, this was an expected, though exasperating, cost of doing business. What followed, however, was startling—and disastrous. On November 6, 1935, Daniel Bovet, a member of Fourneau’s team engaged in testing Rubiazol, infected forty mice with streptococci. The population was then broken into ten groups of four mice each: one an untreated control group, another with the original version of Prontosil/Rubiazol, and seven more with new products just synthesized at the Institut Pasteur. Bovet would later write, “I only had seven new products and we had an extra group of four mice. Why, I asked, not just try the product common to all these products?” The common product was the side chain that Klarer and Mietzsch had added to their azo dye: sulfanilamide.