The unexpected outcomes of medical research: serendipity and the Australia antigenBlumberg BS, Alter HJ, Visnich S. A new antigen in leukemia sera [J Am Med Assoc 1965;191:541–546]
Article Outline
Abstract: The ‘Australia Antigen’ is found in the sera of some normal individuals from foreign populations. The total absence of the antigen from the sera of normal United States’ subjects and its relatively high frequency in acute leukemia suggests that the presence of the antigen maybe of value in the diagnosis of early acute leukemia. Whether the antigen results from or precedes the leukemia process remains to be seen.
[Abstract reproduced by permission of J Am Med Assoc 1965;191:541–546.]
Many of the major discoveries in science have been the product of chance observations that were pursued by the curious mind to their plausible and provable explanation. Mendel's peas and, at least apocryphally, Newton's apple, come to mind as paradigm shifting observations that respectively opened the fields of genetics and established the laws of gravitation. Fleming's chance observation of the inhibition of bacterial growth in the presence of bread mold ultimately produced the first antibiotic. The unexpected link between the Australia antigen and the hepatitis B virus may be less encompassing than genetics or gravitation, but nonetheless is a classic example of a serendipitous observation pursued to major unexpected payoffs, including the prevention of post-transfusion hepatitis, the development of a vaccine, the prevention of hepatocellular carcinoma and laying the foundation for the discovery of non-A, non-B hepatitis and ultimately the hepatitis C virus.
The story began quite simply in two small laboratories at NIH in the early 1960s. It is important to emphasize at this point that the two scientists on the ‘milestone’ paper being discussed herein had no training in liver disease, no abiding interest in hepatitis and no concept that the collaboration they were embarking on would ultimately result in the discovery of the hepatitis B virus. Baruch Blumberg was a senior investigator and leading geneticist working in the Division of Clinical Research of the National Institute of Arthritis and Metabolic Diseases (NIAMD). His primary interest was genetic differences (polymorphisms) among individuals and populations that might lead to differences in disease susceptibility. I was the second scientist engaged in this study and was just embarking on my academic career serving as clinical associate in the NIH Clinical Center. I was fortunate to have been able to spend my military service in the academically-oriented US Public Health Service. This quasi-military body was sometimes referred to as the ‘yellow berets’ representing the cowardly antithesis to the battle-seeking green berets. Despite this friendly put-down, the yellow berets formed an army of young scientists that become the core as well as the corps of NIH long after the demise of the military draft. I was a hematologist whose research interest at that time was determining the cause of transfusion reactions that were not attributable to antibodies against the cellular elements of blood. The third author on this paper was Sam Visnich, an affable technician, whose scientific career ended during the course of these studies when he drew on his prior military experience to become a commercial airline pilot. I draw these three biographical vignettes to reemphasize that the most important discovery in the field of hepatitis, and the foundation for all other discoveries, derived by chance from individuals who had no special knowledge of virology or hepatology and who were moving in different research and career directions when serendipity changed the course of their scientific lives.
I will relate the remainder of this story from my perspective as I saw it unfold and as I pieced together the final parts of the drama in which I played no direct role. In 1962, in my quest to look for anti-protein antibodies that might result in febrile, urticarial or even anaphylactic reactions, I assumed that persons who were multiply transfused might develop antibodies to plasma proteins that were antigenically distinct from their own. I was not particularly concerned with the genetics of these differences but clearly such differences, if they existed, would be genetically determined and parallel the allelic differences long known to exist for multiple erythrocyte antigens. In these early days, before readily available radioimmunoassays and enzyme immunoassays, my simplistic approach was to look for precipitating antibodies by Ouchterlony agar gel diffusion. Templates were made in the NIH biomedical shop that would cut a circle of six wells surrounding a central well into an agar-filled Petri dish. Serum from a multiply transfused patient was placed in the central well and allowed to diffuse to the periphery where it might encounter and precipitate with antigens diffusing from the outer wells containing serum from normal donors or other defined populations. The gels were allowed to develop for a week and gradually my lab bench was inundated with Ouchterlony plates, a veritable suffusion of diffusion. Early in these studies, Dick Aster, then a USPHS research associate in our blood bank and now one of the world's leading experts on blood platelets, told me that he had just heard an interesting lecture by Baruch Blumberg and that Blumberg was using similar gel techniques to search for protein polymorphisms. Aster suggested that I consult with Blumberg which I promptly did. One of the beauties of NIH is the wealth of expertise that resides on a single campus and the ease with which one can establish intramural collaborations. Within days, Blumberg and I had established an informal collaboration that began the circuitous trail to the Australia antigen.
Blumberg and his British collaborator, Tony Allison, had already discovered a polymorphic system involving human beta lipoproteins and I joined in the effort to better define this system and its clinical relevance. A characteristic of this polymorphism, designated Ag, was that the precipitin lines that developed took up Sudan Black lipid stain and appeared intensely blue on the agar plate. One day in 1964, I detected a precipitin line that barely stained for lipid, but subsequently stained red when a general protein stain, azocarmine, was applied. This finding was reproducible and proved to be the result of an immune interaction between antibody in the serum of a patient with hemophilia and an antigen in the serum of an Australian aborigine. The chance testing of aboriginal sera stemmed form the availability of vast collections of sera that Blumberg had accumulated during various field investigations or had sent to him by collaborators also interested in genetic polymorphisms. Thus, each day we randomly tested different, and often exotic, populations. The line formed against the aboriginal serum was unique and had not been found in years of prior investigations using identical techniques. Because of its staining characteristics we initially called this the ‘red antigen’. Subsequently, we debated whether this should be called the Bethesda antigen for the place where it was discovered or the Australia antigen for the place where the patient resided. Based on then emerging nomenclature for newly discovered hemoglobin variants that used designations based on the origin of the patient sample, it was elected to call this the Australia antigen and the presumed polymorphism was designated Au. Approximately 5 years later, when the association of Au with hepatitis had been established, a contentious nomenclature meeting in New Haven changed the name to Hepatitis Associated Antigen (HAA). In 1972 following EM visualization of the virion (Dane particle) and the realization that the antigen was located on the virion envelope, the name was changed to Hepatitis B Surface Antigen (HBsAg), the name it holds today.
I have often been asked what it was like to have observed the Australia antigen for the first time. I would like to say it was a moment of epiphany, but in truth, it was not. The initial discovery failed to reach the heights of scientific euphoria because the finding was not the culmination of an intensive and directed search for some ‘missing link’ and because the clinical relevance of the finding was then unknown. Nonetheless, it nurtured a subdued level of excitement because it was a novel finding and opened a new channel of investigation. As an aside, there was a laboratory moment shortly before this time that does, and always will, stand out vividly in my mind. When I entered the lab on November 23, 1963, I found Dr. Blumberg, Dr. Thomas London and several others standing around the radio with a look of disbelief and intense sadness on their faces. I listened in horror to find that President Kennedy had been assassinated. It is said that everyone remembers the exact circumstances when they first heard of the death of President Kennedy. That was certainly true for me and it was one of the most profoundly sad moments of my life. For me, the discovery of the Australia antigen will always be tempered by that devastating memory.
Following discovery of the Australia antigen, my first task was to determine its prevalence in normal and disease populations. I tested 700 normal blood donors and found the prevalence to be 0.1%, interestingly the same prevalence as subsequently found with more sensitive RIA and EIA techniques. Indeed, the discovery of Au was predicated on the fact that this antigen was present in the blood in such huge quantities that it could be detected by a technique as insensitive as agar gel diffusion. Had this not been the case, discovery of the hepatitis B virus might have been set back another decade or more. As is now well known, the high concentration of antigen is determined by the excessive production of envelope protein compared to nucleocapsid (core) protein such that most surface antigen fails to be incorporated into the virion and circulates instead as non-replicative particles that outnumber virions by at least a thousand to one. With a background prevalence of only 0.1%, there was reasonable hope that we might find a significantly higher prevalence associated with disease. We thus screened a variety of NIH patient populations and the striking finding was that the prevalence in most patients was similar to that in donors, but that patients with leukemia had a hundred-fold higher prevalence (10%). This was not restricted to a particular type of leukemia, but in those early days of leukemia therapy most patients with leukemia admitted to NIH were equally sick and in retrospect equally likely to require blood transfusion. As evident in the title of the first manuscript on the Australia antigen published in 1965, the association with leukemia was reported and we speculated that the antigen might be related to a leukemia virus, the existence of which had been frequently postulated. Interestingly, we discussed the need to look for viral particles by EM, but did not do so at that time. Given the availability of antisera to Au and the high concentration of particles in Au-positive sera, immune EM performed in 1964 could have shortened the link between Au and its viral origin by about five years.
The Australia antigen paper was published in JAMA without fanfare in 1965 []. The term ‘milestone paper’ certainly did not seem applicable at that time, nor could anyone have prophesized the future implications of this research. By the end of 1964, Blumberg had left NIH to assume a position at the Institute for Cancer Research (ICR) in Philadelphia where the rest of the Australia antigen story unfolded. I spent my remaining early tenure at NIH attempting to biophysically characterize the Australia antigen and showed that it had sedimentation characteristics of a lipoprotein, but was distinct from typical human low- or high-density lipoprotein [2]. This is of historical interest because HBsAg has subsequently been found to complex with beta-lipoprotein and it has been postulated that the virus might enter liver cells via lipoprotein receptors. Clearly, the initial biophysical characterization of the Australia antigen was that of the protein itself and not that of the virion with which it associated.
To complete the story surrounding this discovery, Blumberg and his investigative team at ICR, including Tom London, Al Sutnick and Irving Milman, continued to pursue the meaning of the Australia antigen. Focusing on the association with leukemia and his belief that Au was inherited, Blumberg hypothesized that Au might be found in patients with an inherited predisposition to leukemia, namely patients with Down's syndrome. Remarkably, he found that Down's patients had a very high prevalence of the antigen, exceeding the 10% prevalence found in patients with leukemia. While this finding supported the hypothesis that Au was genetically determined and that it might play a causal or supportive role in the development of leukemia, Blumberg, much to his credit, tested the hypothesis further. He next studied Down's patients in institutional settings compared to those living at home. The findings were conclusive. Down's patients living in large institutions had a prevalence of approximately 30%, compared to 10% for those living in small institutions and less than 1% for those living at home. Further, Down's patients were Au negative at birth. This epidemiologic pattern provided the first clue that Au was an infectious agent whose presence increased in the crowded and unhygienic conditions of large institutions for the mentally retarded. While this provided the first link to an infectious origin of Au, the association with hepatitis wasn't made until a Down's patient who had been Au negative became Au positive and a empiric battery of biochemical tests showed that he had elevated ALT levels. Testing of a large number of Down's patients then showed that the mean ALT level in those who were Au positive was significantly higher than those who were Au negative. The final piece in establishing the link to hepatitis came when Barbara Werner, an investigator in Blumberg's lab, developed clinical hepatitis and tested herself to find that she had become Au positive whereas she had previously served as an Au negative control. Although I was not in Blumberg's lab at that time, the story had come full circle for me because seven years earlier Barbara Werner had come to my lab to learn the Ouchterlony technique and because the high prevalence of Australia antigen in patients with leukemia now made sense based on their high exposure to blood products and their relative immunodeficiency that impaired viral clearance. It was like a Hercule Poirot mystery where all the diverse, seemingly unrelated pieces suddenly fell into place and created a coherent whole. The killer was a virus, the antigen was its cloak, the victims were blood recipients, institutionalized persons and sporadic populations and the weapon was a needle or any conduit that would transmit blood or other bodily fluids from person to person. Case closed.
There are many lessons to be learned from this retrospectively landmark study. The first is the value of funding basic research whose clinical outcomes may not be predictable in advance. Indeed, paradigm shifts by their very essence are not predictable. The second lesson is the value of perseverance. The Australia antigen could easily have been forgotten after the initial publication. It was not the kind of finding that stimulated other investigators to pursue the many loose ends. Observations without proven disease associations or therapeutic implications fail to excite the scientific or clinical community. Blumberg, however, persevered and continued to test hypotheses to elucidate the nature of the Australia antigen. He was helped by serendipitous events, but used these to his full advantage until the nature and meaning of the antigen was at last unraveled and a Nobel Prize adorned his mantle.
I would like to end by placing this finding into perspective. The first description of clinical (epidemic) hepatitis is attributed to Hippocrates, whose oath of medical ethics still has validity 2000 years later. The first description of ‘serum’ hepatitis is attributed to Lurman who in 1885 reported almost 200 cases of jaundice among shipyard workers in Breman Germany who had received smallpox vaccine that had been fortified with a human lymph preparation. Many other descriptions of infectious jaundice have found their way into the medical and historical literature. Adrian Reuben has written a wonderful commentary on the relationship of hepatitis to war [3] and indeed it was World War II that served as a major impetus to enhanced research to identify the presumed viral cause of both infectious and serum hepatitis. These wartime and post-war studies provided brilliant epidemiologic insights and led to improved prevention strategies based on hygienic measures and increased caution in the use of vaccines and blood products. However, for all the effort and directed research on hepatitis between 1940 and 1960, no infectious agent was identified and the field was overtaken with the lethargy of frustration. In essence, 2000 years had passed since the teachings of Hippocrates and hepatitis remained a medical enigma with no defined etiologic agent, untreatable severe outcomes and resistance to preventive vaccine strategies that had eradicated other infectious plagues. Enter into this scene a geneticist, a hematologist and an airline pilot and one does not come up with an equation that portends a major breakthrough. Nonetheless, as described herein, this is the combination that serendipitously found an antigen that proved to be an integral part of the hepatitis B virus, that provided the first hepatitis screening assay for the prevention of type B post-transfusion hepatitis, that became the basis of a highly effective hepatitis B vaccine that not only prevents the disease hepatitis but prevents hepatocellular carcinoma, that allowed for the description of non-A, non-B hepatitis by serologic exclusion and hence, that ultimately set the stage for the discovery of the hepatitis C virus. This is a heady outcome for a single precipitin line that stained the wrong way. Dr. Reuben [3] described this as the ‘Thin Red Line’ in reference to a small defensive line of Scottish Highlanders wearing kilts and red jackets who withstood and overcame the charge of a much larger Russian force during the Crimean war. In this contemporary allegory, the thin red line of antigen discovery overcame centuries of inertia and frustration in the war on hepatitis. That war against the hepatitis B virus has been won in theory, but the war is just starting against the social and economic forces that prohibit the highly effective hepatitis B vaccine from being distributed to the billions of people in developing nations who require it the most. Hepatitis B could be eradicated by a global vaccination program similar to that which exists in the United States and other developed nations. Ironically, we have won the war, but we have not yet won the battle.
References
PII: S0168-8278(03)00157-0
doi:10.1016/S0168-8278(03)00157-0
© 2003 European Association for the Study of the Liver. Published by Elsevier Inc. All rights reserved.
