244 THE ANTIOEN-ANTIBODY REACTIONS 



this phenomenon are complicated by the fact that the heat inactivation was carried 

 out in the presence of other serum proteins, and, as the work of Kleczkowski 

 (19416), van der Scheer, Wyckoff and Clarke (1941fl), Bawden and Kleczkowski 

 (19426) Jennings and Smith (1942), and Krejci, Jennings and Smith (1942) shows, 

 heating mixed solutions of antibody globulin and proteins like albumin or casein 

 in the presence of salts (Kleczkowski 1943) produces globulin-albumin and globulin- 

 casein complexes which display no direct antibody activity, but which cannot be 

 considered as denatured in the ordinary sense of the word. 



The study of the antigen-antibody complex offers a more direct approach to 

 the nature of antibody, since protein united with antigen has presumably been 

 selected from the antiserum by reason of its specific combining powers. A certain 

 amount of non-specific material may be adsorbed by the complex — such adsorption 

 occurs in complement fixation — but the amounts are relatively too small to 

 invalidate the assumption that the bulk is in fact antibody. For instance, Heidel- 

 berger and Landsteiner (1923), Marrack and Smith (19316) and Haurowitz and 

 Breinl (1933) have shown that non-specific coloured proteins are not carried down 

 in the precipitate when an antigen reacts with its specific antibody ; and Dean, 

 Taylor and Adair (1935), using the optimal proportions technique and examining 

 a sample of antiserum containing antibodies to purified egg albumin and purified 

 horse-serum albumin, found that either antibody could be specifically precij^itated 

 by the corresponding antigen without affecting the titre of the other. It has 

 been established that antigen-antibody complexes contain considerable quantities 

 of protein from the antiserum, and that this protein has the general characters 

 of serum globulin. Thus the precipitate formed when 2-5 mgm. of Type II pneumo- 

 coccal polysaccharide, which contains no nitrogen and is therefore particularly 

 suitable for work of this kind, reacts with its homologous antiserum, has been 

 found to contain 37 mgm. of serum protein (Felton and Bailey 1926) ; and the 

 antibody titre of a given antiserum may be measured with considerable accuracy 

 by determining the amount of protein that is removed by a soluble or particulate 

 antigen (Heidelberger and Kendall 1929, Heidelberger et al. 1930, Heidelberger 

 and Kabat 193i). 



That the protein so bound has the general properties of globulin is attested by 

 such observations as those of Marrack and Smith (1930, 1931rt), who found that 

 the diphtheria toxin-antitoxin compound showed the same ultra-violet absorption 

 spectrum as serum globulin, or those of Breinl and Haurowitz (1930), who found 

 that a precipitate, containing about 10 per cent, of hsemoglobin and 90 per cent, of 

 material derived from a homologous antiserum, showed the same proportions of 

 tyrosine, cystine, histidine and arginine as did serum globulin when examined by 

 the same technique. 



Antigen-antibody complexes act as antigens, and induce in animals " anti- 

 antibodies " that are specific for antibody in that they combine with the antigen- 

 antibody complex, but not with antigen alone. The antisera will also react with 

 normal serum globulins (Ando 1937 ; see also Marrack and Duff 1938). Another 

 noteworthy feature of such antisera, to which we shall refer later, is their capacity 

 to react with specific precipitates of other antigens, provided that the antibody 

 in these precipitates was produced in the same sj)ecies of animal (Treffers and 

 Heidelberger 1941). However, the properties of antibody protein as revealed in 

 the antigen-antibody complex are open to the objection that the antibody may 

 have been altered by its union with antigen. 



