"The Discovery of the Anti-Globulin Test" written by A. E. Mourant pages 180 to 183 Vox Sang. 45: 180-83 (1983)

Milestones in Blood Transfusion and Immunohaematology

When the Second World War broke out in 1939, the British Government's medical advisers realized that large numbers of blood transfusions were likely to be needed for military and civilian casualties, and that for this purpose great quantities of blood-group testing serum would be necessary. The only blood group system of clinical importance then known was the ABO system.

A few years before the war a group of workers, supported by the national Medical Research Council, and including Dr. G. L. Taylor as Director and Dr. R. R. Race, had begun work on the genetics of the blood groups. They were housed in Professor R. A. Fisher's Galton Laboratory at University College, London. At the outbreak of war they were instructed to move to the Department of Pathology of Cambridge University, where they were to select human donors whose serum was suitable for use as anti-A and anti-B grouping sera, and to process and distribute such sera. They became known as the Galton Laboratory Serum Unit.

When, in 1940, the rhesus blood groups were discovered by Landsteiner and Wiener, and shown to be of clinical importance, the unit, and Dr. Race in particular, began to carry out fundamental work on the new system. This work [Race and Taylor, 1943; Race et al., 1943, 1944] became the basis of Fisher's CDE hypothesis and notation of the immunology and genetics of the system. The first publication of these was by Race [1944] who, in the same paper, also showed that, in addition to the supposedly normal form of anti-Rh (henceforth to be known as anti-D) antibody, which agglutinated D-positive red cells directly, when they were suspended in a saline medium, there existed a variant, known as 'incomplete' antibody. This could at first be detected only by the 'blocking test' - the blocking of Rh-positive cells, which had been suspended in the incomplete serum, so that they became inagglutinable by the normal antibody.

The existence of this type of antibody, and its blocking properties, were also discovered independently by Diamond [1944] and Wiener [1944]. Soon after this, Wiener [1945] described the 'conglutination' test for incomplete antibodies, the forerunner of the bovine albumin suspension test.

At this time (1944-45) there was, working in the Department of Pathology, a veterinarian immunologist, R. R. A. Coombs, who became interested in the possible nature of incomplete antibody. He tried observing under the microscope the electrophoresis, or migration in an electric field, of red cells suspended in a saline medium, and coated respectively with 'complete' and 'incomplete' antibody, with uncoated cells as a control. It was found that the cells coated with incomplete antibody migrated at a similar speed to those with complete antibody, and very differently from the uncoated cells.

It was at this stage, in July 1945, following the death of Dr. Taylor, and the appointment of Dr. Race as Director, that I joined the Unit in a junior capacity.

Almost simultaneously with my arrival, R. Coombs, by a brilliant feat of intuition, had conceived the principle of the anti-globulin test. He states that he was traveling on an ill-lit wartime train from London to Cambridge, trying to read some papers by Ehrlich on the side-chain theory, and speculating idly (like Kekule on the tetravalent carbon atom) on the behaviour of red cells and antibodies, when he visualized the cells, already coated with molecules of incomplete antibody, which was of course a globulin, but still floating free, becoming linked together by molecules of another antibody, an antiglobulin antibody!

He realized that this idea could be the basis of a practical test, and he formulated in some detail its possible applications -- essentially as the direct and the indirect anti-globulin tests. The general principle had indeed already been proposed and confirmed in practice by Moreschi [1908a, b] but his work was very little known and his method had not got through into the general corpus of immunological techniques. We rediscovered his papers after completing most of our own work, and a note on them was added in proof to the paper by Coombs et al. [1946]. Moreschi had died some years previously. Coombs [1954] subsequently discussed them in some detail, generously acknowledging his full priority.

Coombs theoretical work was followed by many weeks of intensive activity by Coombs, Race, myself, and our technicians. First we had to secure (or prepare) an anti-serum against human globulin. We were fortunate to find that Dr. Muriel Adair, in the neighbouring Physiological Laboratory, did have a stock of sera from rabbits immunized respectively against human globulin, against whole human serum, and against human pseudoglobulin. She gave us a generous supply of these sera, which we proceeded to absorb with human group AB Rh-positive (R1 R2) red cells uncoated with blood group antibodies, until such cells ceased to be agglutinated by the absorbed serum. To the best of my remembrance the original critical experiments, and the majority of the others, were carried out with the anti-human-globulin serum.

Rh-positive red cells (probable genotype cDE/cde) were sensitized with the incomplete form of anti-D, washed three times with physiological saline and then exposed to suitably absorbed and diluted anti-globulin serum.

Sensitized and washed cells suspended in saline gave no agglutination, while the same treated cells, when exposed to the various anti-globulin sera, were agglutinated to a titre of 64 or more. This was confirmed by tests using numerous different incomplete anti-D sera. Homologous cells treated with an incomplete anti-c serum were also shown to be agglutinated by rabbit anti-human globulin serum.

D-positive cells only weakly agglutinated by certain anti-D sera were found to be strongly agglutinated after washing and exposing to a rabbit anti-globulin serum. Similar results were obtained using e-positive cells and the then newly discovered but very weak anti-e serum. The results of this work were published immediately in a short paper [Coombs et al., 1945a] and subsequently in greater detail [Coombs et al., 1945b].

A wide range of positive results, and of negative control tests, had now established the effectiveness and specificity of the anti-globulin test for detecting incomplete and weakly reacting anti-Rh sera, by sensitizing homologous cells artificially with such sera. The next step was to use the test for detecting natural in-vivo sensitization of the cells of infants suffering from haemolytic disease of the newborn. It was already known that though there was marked haemolysis of the cells of such infants, spontaneous agglutination was very rare.

Cord blood specimens were now obtained from numerous infants diagnosed clinically and by routine Rh testing as suffering from haemolytic disease of the newborn, and from normal healthy babies. The red cells from each were washed three times in physiological saline, and then exposed to the anti-human globulin reagent. All the specimens from healthy infants remained unagglutinated while nearly all those from affected infants were agglutinated. Of three babies with somewhat anomalous symptoms, and negative anti-globulin test results, two were finally diagnosed as suffering from jaundice of a non-immunological variety, and one case was unsolved (but the baby survived).

One infant without clinical symptoms but giving a positive result was of probable Rh genotype CDe/cde, as were both its parents. The mother's serum yielded a positive indirect anti-globulin result with the father's red cells. The non-Rh antigen involved in this case was that subsequently known as Kell. Thus at the very outset the test had detected a previously unknown blood group system which has since proved to be of some clinical importance.

The work on the direct anti-globulin test on cord bloods was published by Coombs et al. [1946]. Subsequent work [Coombs and Mourant, 1947], using various human protein fractions to see whether they inhibited the anti-globulin test, showed that the protein fraction responsible for sensitization was gamma-globulin. The 'incomplete' antibody mainly responsible for positive anti-globulin test results is now known as IgG while the 'complete' antibody is IgM.

This was the last paper in which I collaborated directly with Coombs, though I subsequently become both a manufacturer and a user of anti-human-globulin sera. Work on the anti-globulin reaction and its various modifications still goes on, and the history of some of its developments has been described by Coombs [1970, 1981].

Acknowledgement

I am indebted to Prof. R. R. A. Coombs for reading a draft of this paper and suggesting a number of factual corrections. He is, of course, not responsible for the ultimate wording.

References

• Coombs, R. R. A.: Moreschi e alcuni recenti sviluppi nello studio dell' aggluntinazione. Informaz. Med. 9: 126 (1954).
• Coombs, R. R. A.: History and evolution of the anti-globulin reaction and its application in clinical and experimental medicine. Am. J. clin. Path. 53: 131-135 (1970).
• Coombs, R. R. A.: Assays utilizing red cells as markers; in Collor, Immunoassays for the 80's, pp. 17-34 (MTP Press, Lancaster 1981).
• Coombs, R. R. A.: Mourant, A. E.: On certain properties of antiserum prepared against human serum and its various protein fractions: their use in the detection of sensitization of human red cells with 'incomplete' Rh antibodies, and on the nature of this antibody. J. Path. Bact. 59: 105-111 (1947).
• Coombs, R. R. A.: Mourant, A. E.; Race, R. R.: Detection of weak and 'incomplete' Rh agglutinins: a new test. Lancet ii: 15 (1945a).
• Coombs, R. R. A.: Mourant, A. E.; Race, R. R.: A new test for the detection of weak and 'incomplete' Rh agglutinins. Br. J. exp. Path. 26: 255-266 (1945b).
• Coombs, R. R. A.: Mourant, A. E.; Race, R. R.: In vivo isosensitization of red cells in babies with haemolytic disease. Lancet i: 264 - 266 (1946).
• Diamond, L. K.: Progress report to committee on Medical Research of the Office of Scientific Research and Development. OEM. cmr. 384 (1944).
• Landsteiner, K.; Wiener, A. S.: An agglutinable factor in human blood recognised by immune sera for rhesus blood. Proc. Soc. exp. Biol. Med. 43: 223 (1940).
• Moreschi, C.: Neue Tatsachen uber die Blutkorperchen- agglutination. Aentbl. Bakt. 46: 49 - 51 (1908a). Moreschi, C.: Beschleunigung and Verstarkung der Bakterienagglutination durch Eiweisssera. Zentbl. Bakt. 456 (1908b).
• Race, R. R.; Taylor, G. L.: A serum that discloses the genotype of some Rh positive people. Nature, Lond. 152: 300 (1943).
• Race, R. R.; Taylor, G. L.; Boorman, K. E.; Dodd, B.E.: Recognition of Rh genotypes in man. Nature, Lond. 152: 563 (1943).
• Race, R. R.; Taylor, G. L.; Cappell, D. F.; McFarlane, M.: Recognition of a further common Rh genotype in man. Nature, Lond. 153: 52 - 53 (1944).
• Race, R. R.: An 'incomplete' antibody in human serum. Nature, Lond. 153: 771 (1944). Wiener, A. S.: A new test (blocking test) for Rh sensiti- zation. Proc. Soc. exp. Biol. Med. 56: 173-176 (1944).
• Wiener, A. S.: Conglutination test for Rh sensitization. J. Lab. clin. Med. 30: 622 (1945).

Dr. A. E. Mourant, The Dower House, Maison de Haut

return