AGGLUTINATION PHENOMENA IN CANCER N. WATERMAN AND L. DB KROMME (Laboratory of the Antoni van Leeuwenhoekhuie, Amsterdam) In the course of our investigations into the cytolysis of cancer cells by different sera and organ extracts and incidently by a new technical procedure, a series of agglutination phenomena were observed. Our cytolytic experiments confirmed to a great extent the facts demonstrated by Freund and Kaminer (3). As a rule normal serum of an adult has a lytic action upon washed tumorcells which is lacking in carcinomatous serum, or organ extract (e.g. skin after tarring), whether human or animal. It is not within the scope of our present article to discuss in detail these highly interesting phenomena and we refer to our more exhaustive papers (1927). We would however again lay stress upon our technic, which is derived from Freund and Kaminer s, and which must be closely followed, if the same results are to be obtained. This technic consists in mincing living tumor with scissors, and mixing it with a solution of the following composition: 1 NaCl 0.6y0 NaH2P04 1.0% ph = 5.4. The emulsion is then filtered through eightfold sterile gauze and washed at least 5 times in the centrifuge with the same highly acid fluid. The result is a suspension of nearly completely isolated cancer cells, which remain in a perfect condition in this solution for many days and may even show mitotic figures. When testing the lytic properties of a body fluid, a drop of this cell suspension is mixed with 1520 drops of the fluid; the contents of the tubes are mixed by shaking with a glass 60
AGGLUTINATION PHENOMENA IN CANCER 61 bead and the concentration of tumor cells is determined by counting in the Burker chamber before and after remaining 24 hours in the thermostate. We have tested the accuracy of this technic by calculation of probabilities and found it fairly satisfactory for highly colloidal fluids. For more watery fluids as, for instance, salt solution dilutions of the same ph as serum (7.7), the counting may depend more or less on secondary circumstances, e.g. the diameter of the capillary pipettes used. For this reason the final tumor cell suspensions were made up in the Freund solution to which human serum in a concentration up to 1/50 was added and then heated to 58" C. In this more colloidal fluid, the distribution of the cells is more equal and they keep still better. The addition of the serum, however, gave rise to the agglutination phenomena, which are the subject of our communication. In undiluted sera agglutination phenomena never occur; nor in dilute sera, when the reaction is too acid. But when sera are diluted, and the resulting ph is less than 6.4, then cancer cells are powerfully agglutinated. The grade of agglutination is so high, that the cells form an absolutely coherent clot, with sharp outlines when examined by the microscope. In the mass of cells spots of disintegration (cytolysis) are visible. We will now show the details of these phenomena. A. THE INFLUENCE OF DILUTING Dilutions of serum 1 to 1 up to 1 to 500 were made with a phosphate buffer solution of the following composition : 1 part primary sodium phosphate 1 S8rensen. 9 parts secondary sodium phosphate 50 cc. physiological sodium chloride solution. One drop of suspension of cancer cells in Freund's acid solution was mixed with 15 drops of the serum dilutions, resulting in a ph 7.4. The result after incubation for 18 hours at 37" C. is shown in Table I. Of course a control test without serum but of the same ph shows no agglutination.
62 N. WATERMAN AND L. DE KROMME The limits of the dilutions vary with the species of tumor as well as with the kind of serum. There may even occur differences in sensitivity in the same tumor as well as in the serum of the same individual. Consequently the optimum dilution at which the reaction is at its maximum, is not always the same. So, there exist for instance marked differences in activity of the sera of the same individuals in relation with the seasons. It is an interesting fact, that the agglutinative power as a rule is diminished in winter time, so that the titer may be lowered from l/loo in summer (August) to 1/10, even 1/5 in November. In autoexperiments these changes were followed exactly. It is probable that some kind of activation (light?) plays a remarkable r61e here. This question deserves an extensive study. Once however the agglutination is positive, the phenomenon is so striking and complete, that it can be judged by the naked eye only. TABLE I Tumor used: Twort carcinoma. Serum: human Serum Dilution with Phosphate Buffer of PH 7.7 1/1.... 1 /5... 1/10... 1/25... 1/50... 1/100... 1/200... 1/500... I Agglutination (Normal Serum) August + I November f Control with no serum but of the same ph show no agglutination. B. THE INFLUENCE OF PH The agglutination depends not on the chemical composition of the buffer (e.g. the phosphate ion), but only on the resulting ph. A borate buffer of the same ph 7.4 and even physiological salt solution wherein the serum ph prevails, show the same
AGGLUTINATION PHENOMENA IN CANCER 63 phenomena. The critical point at which agglutination becomes manifest is ph 6.3. The phenomenon is favored by more alkaline reaction. Below ph 6.3 agglutination is not observed. C. RACE SPECIFICITY Not all animal species show agglutinaiiive action upon cancer cells (Twort carcinoma). Only in the serum of man, mouse, rat, rabbit and possibly cow have agglutinins been demonstrated. Guinea pig and horse serum have not shown the phenomenon. The serum agglutinins are not racespecific; thus human sera will agglutinate cancer cells of mice and vice versa. So far no qualitative difference has been found between normal and carcinomatous human serum. It is of interest to note that, as in the case of cytolysis, blood from the human umbilical cord has no agglutinative property. Like other immunobiological principles the cancer cell agglutination substance is acquired in the course of life. The difference in agglutinative power between the different species, might prove of phylogenetic interest. D. THE CELL SPECIFICITY One of the first questions which arise, is concerning the specificity of the phenomenon, that is whether all tumors.are agglutinated and if this is the case also with normal cells. Table I1 shows that all human carcinomata examined so far TABLE I1 Serum Examined Serum of man 1/50 (fresh) 1 Under investigation. Cell Species 7.4 Various mamma carcinomata Larynx carcinoma Rectum carcinoma Cells of leukaemia of mice Twort carcinoma Schmidt carcinoma Red blood cells Epithelial cells of vagina and bladder Lymphocytes Agglutination f +
64 N. WATERMAN AND L. DE KROMME are agglutinable, and that most but not all transplantable animal tumors are strongly agglutinated. Thus the Twort carcinoma is easily agglutinable, the Schmidt carcinoma, a tumor of nearly the same histological structure, is hardly agglutinable. This different behavior is connected with the different rate of autolysis of the tumors. When a suspension of Twort cells remains a long time in the salt solution (for instance, more than 24 hours at room temperature) the agglutinability even of the Twort carcinoma may be lost. Of normal cells there have been tested, red blood celis, lymphocytes, vaginal and bladder epithelium, the latter being easily obtained in sterile fresh urine. It seemed of no use to test the epithelia of such highly organized parenchymata as the liver or kidney. Red blood cells are not agglutinated under the conditions given. Ordinary epithelia are not agglutinated. The lymphocytes show a different and not easily understood behavior, which needs further investigation. First, the lymphocyte suspension is very sensitive to alkaline reaction by itself; the suspensions are flocked easily already at a ph of 7.3. Further even sera, which show no agglutinative action on cancer cells, as those of the guinea pig, have a cytolytic action on lymphocytes. Therefore the behavior of lymphocytes deserves further investigation. It is not certain that one ia dealing here with the same phenomenon as in cancer cells. It seems very important, that ordinary epithelia are not agglutinated, a type of cell which may give rise to carcinoma. DISCUSSION In the preceding pages we have demonstrated that in some species, namely man, mouse, rat, rabbit, (cow?), there normally exists agglutinins for tumor cells. It was known, that human serum possessed cytolytic properties and the fact was studied long ago by Freund and Kaminer (3) and more recently by ourselves (1) and Lumsden (2). The demonstration however
AGGLUTINATION PHENOMENA IN CANCER 65 of the agglutinative property which, judged by our present knowledge, goes hand in hand with the cytolysis, seems to be a further step toward8 the elucidation of the immunology of the tumorprocess, and not least, because the simple technic, by which the phenomenon is to be observed, allows of research on a large scale. As umbilical serum lacks cytolytic as well as agglutinative properties, we must conclude, that these immunological factors are normally acquired during life. It is not certain at the present time whether this chemical substance may be compared with the ordinary hemolysins and other immune substances. The possibility of extracting cytolytic substances from skin, spleen and lymphatic glands (l), justifies the expectation, that it might also be possible to prepare these agglutinins from the organs mentioned and possibly to identify their chemical constitution. REFERENCES 1. WATERMAN, N. AND DE KROMME, L.: Biochem. Ztschr., 1927, clxxxii, 377; ibid., 1927, clxxxviii, 65. 2. LUMSDEN, TH.: The Lancet, 1926, ii, 112; ibid., 1927, i, 116. 3. FREUND AND KAMINER: Biochem. Grundlagen der Disposition fur Karzinom, Vienne, 1925.