Surface Globulins of T and B Lymphocytes

The concept that lymphocytes recognize antigens by means of a surface receptor molecule of immunoglobulin (Ig) character has come to be widely accepted (1-3). Direct study of such receptors in the mouse by means of antiglobulin reagents labeled with either fuorescent (4, 5) or radioisotope (3, 6-8) markers has revealed a dense coating of surface immunoglobulins on nonthymus-derived or B lymphocytes 1 (5, 8), but the question of receptors on the surface of thymus-derived or T lymphocytes remains more controversial. On the one hand, immunofluorescent (5) and some radioimmunolabeling (6) techniques have failed to show immunoglobulin on the surface of T lymphocytes, On the other hand, treatment of T lymphocytes with specific rabbit anti-mouse globulin reagents has inhibited lymphocyte-antigen interactions such as rosette formation (9, 10) and radioactive antigen suicide (11). Moreover, in some (1214) but not all (15) reports, in vitro binding of antiglobulin onto T lymphocytes has inhibited their capacity to mediate graft-vs.-host activity (12, 13), delayed hypersensitivity responses (12), or T cell dependent helper functions (14). Recently, Bankhurst et al. (15) were successful in labeling up to 37% of some T lymphocytes with l~SI-labeled rabbit anti-mouse d-chain IgG using exceptionally long radioautographic exposure periods. This prompted us to devise a "sandwich" radioimmunolabeling method, in which mouse lymphocytes were reacted first with a rabbit anti-mouse Ig reagent and then with a 12SI-labeled sheep anti-rabbit globulin (SARG). We hoped that extra sensitivity would be obtained which might allow labeling of an even higher proportion of T cells. Two further considerations favored a sandwich approach. First, it was a convenient way to compare directly the lymphocyte-binding power of a wide variety of different rabbit antisera without necessitating either fractionation or labeling of the rabbit sera. Secondly, we wished to study the metabolic