hörsal C samhällsvetarhuset, fredagen den 25 februari 1983 k l . 0 9 .00.
ABSTRACT
RECOGNITION REQUIREMENTS AND REGULATORY EVENTS DIRECTING T CELL RESPONSES.
M artin Gull berg, Department of Immunology, University of Umeå, S-901 85 Umeå, Sweden.
The present study has considered c e llu la r and molecular requirements in T c e ll responses. The central ro le of T c e ll growth factors (TCGF) in T c e ll responses prompted us to study the regulatory events d irec tin g TCGF production in le c tin stimulated cu ltures. I t was found th at normal spleen c e lls , activated with Concanavalin A fo r 24 h, develop suppressive c e lls th a t block de novo TCGF production by fresh spleen c e lls . The induction time fo r e ffe c to r suppressor c e lls (nonadherent, L y t-2 -p o s itiv e T c e lls ) was found to be 18 h and to p a ra lle l the termination of TCGF production in s it u . The suppressive mechanism is n either iji s itu absorption of TCGF produced at control rates nor k i l li n g of TCGF producing c e lls . These results suggest th a t suppression of TCGF production is an active process which d ire c tly and reversib ly blocks TCGF-producing c e lls .
This study also indicated th a t ConA induced a very lim ite d p ro life ra tio n of L y t-2 - T helper c e lls (TH) in unselected T c e ll populations. The activatio n and growth requirements of L y t-1 + TH c e lls were d ire c tly investigated and compared with those of L y t-2 + cytotoxic T lymphocytes (CTL), as defined by the selective expression of Lyt d iffe re n tia tio n antigens and functional a c t iv it ie s . This analysis revealed a profound difference in ac tiv atio n and growth requirements between these T c e ll subsets. Thus, while L y t-2 + CTL precursors can be induced to TCGF re a c tiv ity by soluble le c tin s , in the absence of specialized accessory c e lls ,; L yt-2" TH c e ll precursors show a s t r ic t accessory ce ll requirement both fo r a c tiv atio n and p ro lite r a tio n . F in a lly , the low level of TH c e ll e ffe c to r function, detected in a primary responses to allo-MHC-antigens or le c tin s , appears to be due to the develop
ment of suppressive Lyt2+ T c e lls .
The functional relevance of Lyt-2 antigens expressed on CTL membranes was fu rth e r assessed in the la s t part of th is study. Two d is tin c t ac tiv atio n systems were used, namely MHC-antigens, provided as U V -irradiated stim ulator c e lls or polyclonal induction by a 4 h pulse, with le c tin s . Both procedures were shown to s e le c tiv e ly induce L y t-2 + CTL precursors into TCGF re a c tiv ity without leading to m itosis, unless TCGF was added. In both cases i t was found th at monoclonal a n ti-L y t-2 antibodies in h ib ite d the two antigen- dependent phases of CTL responses namely, the i n i t i a l induction step and ta rg e t c y to ly s is . The analogy observed between antigen sp ecific and le c tin mediated indueton and targ e t c y to ly s is , with regard to the s u s c e p tib ility of in h ib itio n by a n ti-L y t-2 antibodies has lead to a general hypothesis on CTL a c tiv a tio n .
Key words : T c e ll growth fac to r production / L y t-2 -p o s itive T suppressor c e l l / L y t-2 -p o s itiv e cytotoxic T lymphocytes / ac tiv atio n requirements of T lymphocytes / blocking of T c e ll functions.
No 96 - ISSN 0346-6612
From the Department of Immunology U niversity of Umeå, Umeå, Sweden
E ditor: the Dean of the Faculty of Medicine
RECOGNITION REQUIREMENTS AND
REGULATORY EVENTS
DIRECTING T CELL RESPONSES
by Martin Gull berg
L is t of abbreviations... 1 Abstract... 2 O riginal Papers... 3
I INTRODUCTION
1. The immune system... 4 2. I n i t i a l trig g e rin g of T lymphocytes... 7 3. Lymphocyte growth fa c to rs ...11
I I OUTLINE OF THE PRESENT STUDY
1. Regulation of T c e ll growth factor production...15 2. Growth and stim ulatory requirements for d iffe re n t T
ce ll populations... 16 3. Recognition requirements involved in induction and
ta rg e t c e ll ly s is of cytotoxic T lymphocytes... 17
I I I GENERAL DISCUSSION
1. Negative regulation in a T c e ll response... 19 2. Possible mechanisms of action of T c e lls capable of
blocking TCGF production...21 3. The ta rg e t c e ll for TCGF...24 4. Growth requirements fo r L y t-2 “ T helper ce ll precursors...26 5. Functional significance of the Lyt series of
d iffe re n tia tio n antigens... 29 6. Blocking of antigen recognition by monoclonal a n ti-L y t-
2.2 an tibo d ies...31 7. Involvment of L yt-2 antigens in le c tin mediated CTL
ac tiv atio n and targ e t c e ll destructio n ... 34 8. P arallelism between antigen and le c tin mediated T c e ll
a c tiv a tio n ...36
IV ACKNOWLEDGEMENTS... 41
REFERENCES 42
L is t of abbreviations:
Class I antigens Class I I antigens ConA
CSF CTL Ig
LAF LPS mAb MHC (H-2) MLC TCGF TRF TH Ts UV
K and D region encoded antigens I region encoded antigens Concanavalin A
Colony stim ulating factors Cytotoxic T lymphocyte Immunoglobulin
Lymphocyte ac tiv atin g fa c to r
t
Lipopolysaccharide Monoclonal antibody
Major h is to c o m p a tib ility complex Mixed lymphocyte culture
T c e ll growth fa c to r T c e ll replacing fa c to r T helper lymphocyte T suppressor lymphocyte U ltra v io le t lig h t
Akademisk avhandling fö r med.sc i.
doktorsexamen vid Umeå U n iv e rs ite t, 1983.
ABSTRACT
The present study has considered c e llu la r and molecular requirements in T c e ll responses. The central role of T c e ll growth factors (TCGF) in T c e ll responses prompted us to study the regulatory events d irectin g TCGF production in le c tin stimulated cu ltures. I t was found th at normal spleen c e lls , activated with Concanavalin A fo r 24 h, develop suppressive c e lls th a t block de novo TCGF production by fresh spleen c e lls . The induction time fo r e ffe c to r suppressor c e lls (nonadherent, L y t-2 -p o s itiv e T c e lls ) was found to be 18 h and to p a ra lle l the termination of TCGF production in s i t u . The suppressive mechanism is neither in s itu absorption of TCfiF produced at control rates nor k illin g of TCGF producing c e lls . These resu lts suggest th at suppression of TCGF production is an active process which d ire c tly and reversib ly blocks TCGF-producing c e lls .
This study also indicated th at ConA induced a very lim ite d p ro life ra tio n o f L y t-2 - T helper c e lls (TH) in unselected T c e ll populations. The a c tiv a tio n and growth requirements of L y t-1 + TH c e lls were d ire c tly in vestigated and compared with those of L y t-2 + cytotoxic T lymphocytes (CTL), as defined by the selective expression of Lyt d iffe re n tia tio n antigens and functional a c t iv it ie s . This analysis revealed a profound d ifferen ce in ac tiv atio n and growth requirements between these T c e ll subsets. Thus, while L yt-2+ CTL precursors can be induced to TCGF r e a c tiv ity by soluble le c tin s , in the absence of specialized accessory c e lls , Lyt-2" TH c e ll precursors show a s t r ic t accessory ce ll requirement both fo r ac tiv atio n and p ro life r a tio n . F in a lly , the low level of TH ceVI e ffe c to r function, detected in a primary responses to a llo-MHC-antigens or le c tin s , appears to be due to the development of suppressive L yt-2+ T c e ll s.
The functional relevance of Lyt-2 antigens expressed on CTL membranes was fu rth e r assessed in the la s t part of th is study. Two d is tin c t a c tiv a tio n systems were used, namely MHC-antigens, provided as UV- irra d ia te d stim ulator c e lls or polyclonal induction by a 4 h pulse, with le c tin s . Both procedures were shown to s e le c tiv e ly induce L yt-2+ CTL pre
cursors into TCGF re a c tiv ity without leading to m itosis, unless TCGF was added. In both cases i t was found th at monoclonal a n ti-L y t-2 antibodies in h ib ite d the two antigen-dependent phases of CTL responses namely, the i n i t i a l induction step and targ e t c y to ly sis. The analogy observed between antigen sp ecific and le c tin mediated inducton and targ e t cy to ly s is , with regard to the s u s c e p tib ility of in h ib itio n by a n ti-L y t-2 antibodies has lead to a general hypothesis on CTL a c tiv a tio n .
Key words: T c e ll growth fac to r production / L y t-2 -p o s itiv e T suppressor cel 1 / Lyt-2-p o si t i ve cytotoxic T lymphocytes / ac tiv atio n requirements of T lymphocytes / blocking of T c e ll functions. .
No 96 - ISSN 0346-6612
This thesis is based on the following publications and manuscripts, which w ill be referred to by th e ir Roman numerals:
I . Gullberg, M., Iva rs, F ., Coutinho, A ., and Larsson, E .- L ., Regulation of T c e ll growth factor production: Arrest of TCGF production a fte r 18 h in normal le c tin stimulated mouse spleen c e ll cu ltures. J . Immunol., 127:407, 1981.
I I . Gullberg, M., and Larsson, E .- L ., Studies on induction and e ffe c to r functions of ConcanavalinA-induced suppressor c e lls abrogating TCGF production. J. Immunol., 128:746, 1982.
I I I . Gullberg, M., and Larsson, E .- L ., ConA-induced TCGF re a c tiv ity is se le c tiv e ly acquired by L yt-2 p o s itive T c e ll precursors.
Submitted for p ub lication .
IV. Gullberg, M., Pobor, G., Bandeira, A ., Larsson, E .-L ., and Coutinho. A., D iffe r e n tia l requirements for activatio n and growth of unprimed cytotoxic and helper T lymphocytes.
Submitted for p ub lication .
V. Gullberg, M., and Larsson, E .- L ., Selective in h ib itio n of antigen induced "step 1" in cytotoxic T lymphocytes by a n ti- Lyt-2 antibodies. Eur. J. Immunol. 1 0 0 6 , 1982.
V I. Gullberg, M. and Larsson, E .- L ., Induction of c y to ly tic T lymphocytes by mitogenic le c tin s is s p e c ific a lly in h ib ited by a n ti-L y t-2 antibodies. Submitted for p ub lication .
I INTRODUCTION
1. The immune system
Ever since the appearance of re p lic a tin g organisms on earth , there has been a continuous competition for " l i f e space" between d iffe re n t forms o f l i f e . Later on, more d iffe re n tia te d creatures appeared which were not able to compete on the level of growth rates, and systems were developed to protect these from invasion of rapid ly re p lic a tin g environmental organisms such as viruses, b a c te ria , fungi and parasites.
The most sophisticated defense system of th is kind is the immune system o f vertebrates.
The immune system consists of a large number of c e lls with d iffe re n t specialized functions. When an infectious agent succeeds in penetrating in to the intern al environment, the f i r s t attack is performed by granulocytes and macrophages, which co n stitute the unspecific part of the defense mechanism. I f the invaders survive the f i r s t unspecific response, macrophages w ill present selected parts of th e ir antigens to T and B lymphocytes, which are the c e lls responsible for the sp ecific immune response. These c e lls resemble true aristocrates in the sense th a t most of them p ers ist throughout th e ir l i f e span doing nothing.
Nevertheless, i f an antigen is present in s u ffic ie n t amounts, clones which recognize i t w ill be activated and s ta rt expanding. Thus, a large pool of lymphocytes are c irc u la tin g in the lymphatic system and each expresses c lo n a lly d is trib u ted receptors. The d iffe re n t clones are, together, sp ec ific for a ll kinds of molecular patterns. I f an antigen enters the system, only a minor fra c tio n of th is pool is able to s p e c ific ia lly recognize it s structure and i n i t i a t e clonal expansion.
This is the most fundamental concept in immunology and i t was f i r s t visualized by Jerne in 1955 (1 ) , and two years la te r modified by Burnet, as the clonal selection theory (2 ).
In mammals the immune system develops in the fe ta l liv e r but a fte r
b ir th , bonemarrow is the major s ite for production of hematopoetic c e lls (3) lymphocytes included. All lymphocytes o rig in ate from a common pi uri potenti al hematopoietic stem c e l l, which divide independently of antigen in the bone marrow and the progeny of which d iffe re n tia te s to specialized c e lls w ithin the mammalian hematolymphoid system ( 4 ,5 ,6 ) . During the d iffe re n tia tio n in the bone marrow, B-lymphocytes precursors acquire immunoglobulins on the c e ll membrane. Immunoglobulins serve both as antigen sp ec ific receptors and e ffe c to r molecules ( 7 ) . As mentioned above, the lymphocyte compartment consists of a large number of c e lls and in the B c e ll system, each clone express immunoglobulin receptors with a given combining s ite or id io typ e. According to the clonal selection theory, one clone is only allowed to express one type of combining s ite . In the lig h t of the recent knowledge concerning the genetic organization of the immunoglobulin genes, i t is possible to state th at the requirement one clone - one sp e c ific y , is f u l f i l l e d ( 8- 10).
In contrast to the B c e lls , maturation of the T c e ll lineage does not occur in the bonemarrow. Precommited T c e ll precursors enter the thymus from the blood stream and the intrathymic environment w ill give ris e to an extensive p ro life r a tio n . During th is p r o life r a tiv e phase w ithin the s u p erfic ia l thymic cortex the b la st c e ll gives ris e to a progeny of small lymphocytes (1 1 ). In the maturation process in the thymus the d iffe r e n tia tin g lymphocytes express sp ecific determinants on th e ir c e ll surface: Thy-1, T la , Gv-1, Lyt and H-2 antigens (1 2 ). The mature T c e lls enter the peripheral lymphoid organs as small resting c e lls , expressing c lo n a lly d is trib u ted receptors.
The antibody d iv e rs ity is probably generated in the bonemarrow while the re p e rto ire of T c e ll s p e c ific itie s is generated during the intrathym ic d iffe re n tia tio n process (13 -18 ).
Mature T c e lls can be divided into three major functional subgroups, namely T helper c e lls (T H -c e lls ), cytotoxic T lymphocytes (CTL) and T suppressor c e lls (T s -c e lls ). Nothing is known with c e rta in ty eith e r about the gene organisation or the biochemical structure of the
clo n a lly d is trib u ted T c e ll receptor. Although some investigators have claimed th at immunoglobulin idiotypes are present on T c e lls , recent molecular genetic studies have p ra c tic a lly excluded p a rtic ip a tio n of the immunoglobulin genes in the T c e ll receptor complex (19-21; L.
Hood, J .F .A .P . M ille r , S. Tonegawa. Personal communications, 1982). I t is c e rta in , however, th at T c e lls recognize antigen in a d iffe re n t way than B c e lls . T c e lls appear to recognize antigens exclusively in association with the products of polymorphic genes at the major h is to c o m p a tib ility complex (MHC). I t is possible, however, th at there exists heterogeneity between d iffe re n t functional T c e ll subsets, also with regard to the antigen sp ecific receptor.
T lymphocytes mediate the c e llu la r immune response and they exert th e ir e ffe c to r functions eith e r by the release of humoral factors or by mechanisms involving c e ll to c e ll contact, as is the case for CTL in the cytolysis of ta rg e t c e lls . Also, TH c e lls trig g e r B c e lls by a d ire c t in te rac tio n (2 2 ,2 3 ). D iffe re n t T c e ll subsets can be distinguished from each other by the expression of certain c e ll surface markers belonging to the Lyt-system (2 4 ). Based on these d if f e r e n tia tio n antigens, T c e lls can be separated into three d iffe re n t sub
sets, namely Lyt-1+, Lyt-23+ and L y t - l +23+ c e lls (2 5 ,2 6 ).
The TH c e lls belong to the L y t-1 + subset, which account for about one- th ir d of a ll peripheric T c e lls , while Lyt-23+ T c e lls constitute approximately 10-15 % o f the T c e ll pool and include both the CTL and the Ts c e lls . The la s t subset, namely L y t - l +23+ T c e lls , which accounts fo r approximately h a lf of a ll T c e lls , seems to include a ll kinds of functional T c e ll a c tiv it ie s depending on the inducing antigen (2 7 ,2 8 ).
Despite the fa c t th at exceptions (27,28) have been reported during .the la s t years, the c la s s ific a tio n outlined above seems to remain useful.
Most immunologists believed for many years th at antigens alone could a c tiv a te lymphocytes and drive them into m itosis. Already 14 years ago, however, c e ll collaboration was demonstrated and, during the la s t 5 years a number of lymphocyte sp ecific growth factors have been described, which p a rtic ip a te in the c e llu la r co llab o rative events and
are required for a ll types of immune responses (2 3 ,2 9 -3 2 ). Lymphocyte growth factors are not sp ec ific for antigen, but apparently sp ec ific fo r each d i f f e r e n t i a t e class of hematopoietic c e lls . Since immune responses are s p e c ific , however, there must e x is t a discrim inatory event before the action of growth fac to rs. This appears to be the case, since only lymphocytes which have recognized antigen s p e c ific a lly w ill be induced to express sp ecific growth receptors (2 3 ,3 3 ). Thus, lymphocyte activatio n comprises two d iffe re n t steps, namely antigen recognition involving the c lo n a lly d is trib u ted receptors, and a c e ll type sp ec ific growth factor which drives the c e ll into m itosis. The i n i t i a l trig g e rin g event w ill be fu rth er discussed in the next part of th is introduction, while the th ird part w ill deal with production and action of lymphocyte growth and maturation fac to rs.
2. I n i t i a l trig g e rin g o f T lymphocytes
Mitchison (1954) and Lawrence (1959) suggested more than 20 years ago, th a t c e ll mediated immunity to bacterial antigens could only be created when these antigens were presented in association with s e lf antigens of the individual (3 4 ,3 5 ).
Rosenthal and Shevach reported in 1973, th at primed T c e lls could be restim ulated to p ro life r a te in^ v itro by adding the antigen together with syngeneic or semiallogeneic accessory c e lls , but not when adding i t together with allogeneic c e lls (3 6 ). Furthermore, blocking of the T c e ll p r o life r a tio n , was achieved by addition of antibodies directed to the MHC antigens present on the antigen presenting c e ll (3 6 ,3 7 ). Taken together, these observations suggested th at T c e lls are activated only i f the recognition of the antigen occurs in association with c e ll surface molecules encoded a t the MHC complex. This phenomenon of obligatory recognition of antigen in association with MHC-encoded products has been termed M H C -restriciton , or in the mouse, H-2 .re s tric tio n .
The MHC complex shows an enormous degree of polymorphism w ithin any given species, and was f i r s t id e n tifie d as being responsible for the phenomenon of g ra ft re je ctio n (3 8 ). The mouse MHC, the H-2 complex, consists of four major regions denominated as K, I , S and D region. H-2 K and D regions encode c e ll surface proteins present on a ll nucleated c e lls in the body, while H-2 I-encoded proteins are p rim arily expressed on B c e lls and antigen presenting accessory c e lls such as macrophages and d en d ritic c e lls . Some of the serum complement components are co n tro lled by the H-2 S region (1 2 ).
Zinkernagel and Doherty found, in 1974, th at virus sp ecific CTL were able to lyse solely syngeneic virus infected targ et c e lls (3 9 ). Further analyses of th is phenomenon have revealed th at th is re s tric tio n mapped to the K and D region of the H-2 complex (4 0 ). M ille r et al^ in 1975 found th at transfer of delayed type h yp ersen sitivity caused by protein antigens, a function mediated by TH c e lls was H-2 I re s tric te d (4 1 ).
Despite some reported exceptions, there is a general agreement th at antigen recognition by CTL is re s tric te d to K and/or D-regions, while TH c e lls recognize antigen in connection with I-A or I-E regions en
coded anti gens (1 5 ,1 7 ,1 8 ,2 7 ,2 8 ,4 2 -4 6 ).
The phenomenon of H-2 re s tr ic tio n of T c e lls is determined, p rio r to the challenge with antigen. Bevan (17) and Zinkernagel et al^ (15) have demonstrated, using FI -► parental chimeric mice, th at the MHC-haplo- type of the parent host governs the MHC re s tr ic tio n of the developing F l T c e lls . The presence of MHC-encoded antigens on the e p ith e lia l c e lls of the thymus ensures th at maturing T c e lls can learn " s e lf preference" during the intrathym ic steps of d iffe re n tia tio n .
An important function, known to be controlled by genes located w ithin MHC-region, is the control of sp ecific immune response to d is tin c t thymus-dependent antigens. These genes show the same degree of poly
morphism as the MHC-region and has been termed "immune response genes"
(Ir-g e n e s ) (4 7 ). I t has been found th at Ir-genes co n tro llin g I-reg io n re s tric te d responses map w ithin the same region (4 8 ), whereas Ir-genes co ntroling CTL responses have been located w ithin the K and D regions
(1 6 ,4 9 ). These findings were the f i r s t indications th a t Ir-gene products and the antigens responsible for H-2 re s tric tio n are the same.
Several mechanisms have been proposed to account for the influence th at MHC antigens have on the response of T c e lls and most of these proposals can be c la s s ifie d into the three following broad categories:
determinant selection th a t imply sp ecific association of antigen and MHC molecules to present unique antigenic determinants (5 0 ,5 1 ), clonal deletions o f s e lf reactive T c e ll clones th at w ill give ris e to a hole in the T c e ll re p e rto ire (52) and f in a lly a MHC sp ec ific T c e ll selection process based on a degree of re a c tiv ity to s e if-MHC-encoded molecules expressed in the thymus (1 4 ,1 6 ,5 3 -5 5 ). The mechanisms proposed above are not mutually exclusive and provide explanations for most experimental observations to date.
The theory of Jerne on the generation of d iv e rs ity (1 4 ), la te r modified to include re s tric te d T c e lls (1 6 ), accounts for these phenomena, namely Ir-gene e ffe c ts , MHC-restriction and MHC-dependent a llo - re a c t iv ity . The basic postulates are a complete germ-line re p e rto ire of T c e lls for MHC determinants of the species and dual s p e c ific ity of T c e lls for MHC-determinants and antigens. The d e ta ils of the "dual s p e c ific ity " shown by re s tric te d T c e lls , however, have been much debated. Von Boehmer e t a l (16) proposed a two receptor model which postulate th a t a T c e ll has two d iffe re n t receptors, one for antigen X, and another for s e lf MHC antigens (F ig . 1 ). One of the problems of a ll dual receptor models, are the resu lts from cold targ e t in h ib itio n experiments. In h ib itio n with unlabeled target c e lls is only obtained when the re s tr ic tin g element and the antigen are presented together on the same c e ll surface (5 6 ). In the competing model, generally refered to as the altered s e lf hypothesis, i t is assumed th at an antigen X associates with the MHC antigen to form a hybrid molecule, or a new- antigenic determinant, and th is complex is then recognized by a single receptor (5 7 ,5 8 ). Since a ll T c e lls are reactive to modified s e lf MHC antigens, i t is lik e ly th a t they crossreact with allo-MHC-encoded antigen, explaining the high frequencies of a llo re a c tiv e T c e lls (5 9 ).
This hypothesis is supported by observations that H-2 re s tric te d k i l l e r c e lls sp ec ific for antigen X, can lyse allogeneic targ e t c e lls
expressing neither the antigen X, nor the re s tr ic tin g s e if-MHC-encoded antigen (1 7 ,6 0 -6 2 ).
Personally, I regard the a lte r e d -s e lf hypothesis as an a ttr a c tiv e model because of it s s im p lic ity and the demonstration th at T c e lls , which are s p e c ific a lly re ac tiv e to allo-MHC-encoded antigens, can indeed recognize foreign antigens in an H-2 re s tric te d manner.
T c e ll
''antigen presenting c e ll dual recognition
T c e ll
‘'antigen presenting c e ll a lte re d - s e lf
Figure 1.
MCH-encoded antigen antigen X
A vertebrate contains between lO ^ -io ^ lymphocytes with various s p é c ific ité s . According to the concept of clonal selection theory, only a few clones w ill be activated and expanded in an immune response, w hile a ll others are kept re stin g . The study of lymphocyte ac tiv atio n co n stitu te s, th e re fo re , a fundamental aspect of immunology. In the study of a sp ecific immune response, defined antigenic structures lik e haptens or synthetical copolymers serve as important to o ls . These antigens however, give rise to clonal responses which are too small to be studied in unprimed lymphocyte populations, from ce ll biological and biochemical points of view.
Polyclonal a c tiv a to rs , also termed mitogens, have been very useful in the study of lymphocyte a c tiv atio n since they overcome th is d if f ic u lt y and stim ulate large numbers of c e lls . Lectins co n stitute a large group among these mitogens and they have been extensively used in d iffe re n t experim ental systems (6 3 - 6 5 ) . These p ro te in s , w ith e x q u is ite
s p e c ific ity for sugar residues on c e ll surface glycoproteins, bind to abundantly represented ta rg e t proteins on a ll lymphocyte c e ll membranes (6 6 ). Mitogenic le c tin s however, s e le c tiv e ly induce T c e lls to p r o life r a te , demonstrating the s p e c ific ity of a c tiv atio n and the d ifferen ce from binding. The increase in immunoglobulin synthesis observed under some conditions is a consequence of co lla b o ra tiv e in t e r actions with le c tin -a c tiv a te d TH c e lls (6 7 -6 9 ).
The demonstration th a t lectin-m ediated T c e ll a c tiv a tio n , as antigen- induced responses, requires the p a rtic ip a tio n of accessory c e lls , has underlined th e ir value as a model system for studying the mechanisms of c e ll cooperation involved in T c e ll trig g e rin g (7 0 -7 3 ). Recent reports have demonstrated th at le c tin induced p ro life ra tio n is a re s u lt of an in tr ic a te team work between d iffe re n t c e llu la r subsets, which allows fo r the conclusion th a t mitogenic le c tin s are in fa c t no true mitogens (3 2 ,3 3 ,7 3 ,7 4 ). Furthermore, only few of a ll the le c tin s , known to s p e c ific a lly recognize sugar determinants on T c e lls , are polyclonal T c e ll activato rs (6 6 ), again showing the difference between binding and trig g e rin g .
I t is s t i l l not established whether le c tin mediated ac tiv atio n of T c e lls is a true immunological reactio n , i . e . , employing c lo n a lly d is trib u te d receptors, or i f le c tin s bypass the requirement of sp ecific recognition by in te ra c tin g with a hypothetical "activatio n s ite " . In view of the amount of studies performed with mitogenic le c tin s , i t is o f importance to find s im ila r itie s or differences between le c tin and antigen induced lymphocyte a c tiv a tio n .
3. Lymphocyte growth factors
As previously o utlined, lymphocyte growth is dependent on the hormone
lik e action of growth fac to rs. Already in 1965, soon a fte r the discovery of polyclonally ac tiv atin g le c tin s , soluble mitogenic factors were found in supernatants of le c tin stimulated cultures (7 5 ,7 6 ). For
many years, many d iffe re n t kinds of biological e ffe cts were a ttrib u te d to these factors and a rather confusing picture of th e ir biological sig nifican ce emerged. A real breakthrough occured in th is f ie ld as Morgan e t a ! (77) observed th a t conditioned media from le c tin stimulated lymphocyte cultures contained growth promoting a c t iv it ie s fo r bonemarrow T c e lls in v i t r o . These findings were soon expanded by G1111s and Smith (7 8 ), who showed th at the same p rincip les could be applied to s p e c ific a lly activated T c e lls , making i t possible to propagate and expand activated T c e lls , and o fferin g the way to the recent development of j j i v itro studies on cloned T c e ll populations (7 8 -8 3 ). The protein responsible for the growth promoting a c tiv ity for T c e lls has been characterized as a glycoprotein with a molecular weight of 15 K 1n human and 30 K in mouse (2 9 ).
Supernatants of le c tin stimulated lymphocyte cultures however, do not only contain T c e ll growth factor (TCGF) but also a m u ltip lic ity of a c t iv it ie s a ffe c tin g , for instance, B c e lls (T c e ll replacing factor (TRF) ) ( 8 4 ) , hematopoietic stem c e lls (colony stim ulating factors ( CSFs) ) , and several other factors a c tiv a tin g macrophages and mast c e lls (8 5 -8 7 ). The l i s t of a c tiv it ie s can be made very long and i t would seem th a t there exists as many factors in such a supernatant as there are assays described for lymphokine a c t iv it ie s .
The nonspecific action of TCGF i t s e l f , indicated th at the process of T c e ll ac tiv atio n could not simply be the generation of growth fa c to rs, as immune responses are s p e c ific . The discovery th at TCGF was non- mi togenic on small unprimed T c e lls made i t possible to introduce an i n i t i a l discrim inatory step, involving c lo n a lly d is trib u ted receptors, which induces TCGF responsiveness (3 3 ,8 8 ,8 9 ). This model of T c e ll a c tiv a tio n was demonstrated by the fin d in g , th at 4-5 h pulse with le c tin or antigen resu lts in induction of TCGF r e a c tiv ity in an unprimed T c e ll population, which is otherwise, re s is ta n t to the mitogenic a c tiv ity of the fa c to r. Induced c e lls are subsequently able to s p e c ific a lly absorb the growth promoting a c tiv ity (3 3 ,9 0 ). Further
more, Robb e t al have recently demonstrated, by using a radiolabeled
homogeneous TCGF preparation, th at induction of TCGF re a c tiv ity gives ris e to more than a 1000-fold increase in TCGF binding sites on T c e lls (9 1 ).
The lectin-dependent induction of resting T c e lls to TCGF r e a c tiv ity does not require accessory c e lls , while production of growth factors is a c e ll cooperative response which involves TH c e lls and I-A + macro
phages in conjunction with the le c tin (3 2 ,3 3 ). The subsequent TCGF dependent p ro life ra tio n takes place independently of the le c tin and of any other accessory c e ll (9 2 ). I-A + macrophages can be substituted for by a d iffe re n tia tio n factor (lymphocyte ac tiv atio n fa c to r, LAF) which enables p u rifie d T c e lls to produce TCGF (3 2 ). LAF is produced by I-A + macrophages, the induction of which is the T c e ll dependent responses of nonprimed populations requires a TH c e l l, d is tin c t from the TCGF producing T c e ll and designated T "inducer" c e ll (74 ).
The model o f T c e ll a c t iv a t io n , which we have proposed, is schematically presented in Fig. 2 (9 3 ).
Tprod.
TCGF ct-la ag
T ind. CTL
CTLp) ►(CTLp M0
CTL Tsuppr]
Several investigators have recently described a B c e ll sp ecific growth fac to r with sim ilar biological proporties as TCGF (2 3 ,3 0 ,3 1 ,9 4 ,9 5 ).
Furthermore, evidence for the existence of d iffe re n t kinds of maturation and d iffe re n tia tio n factors acting on CTL or B c e lls was recently obtained (8 4 ,9 6 ,9 7 ). Even B c e ll p ro life ra tio n induced with
lipopolysaccharide (LPS), which has been believed to be a "true"
mltogenlc process, appears to depend on factors produced by macrophages (9 8 ).
Personally I b elie ve , th at I t Is now the time to leave "the period of crude supernatants" In favour of well defined homogenous protein preparations, or we w ill soon be back to the same type of confusions we thought we l e f t some years ago.
I I OUTLINE OF THE PRESENT STUDY
The broad alms of th is investig ation have been to study the c e llu la r and molecular requirements in T c e ll responses. The regulation of TCGF pro du ctio n appeared to be o f p a r t ic u la r i n t e r e s t , since the a v a ila b ilit y of th is growth factor seems to be the only lim itin g variab le for the clonal expansion of TCGF- reactive c e lls . The i n i t i a l p art of the work presented here is concerned with the analysis of the regulatory events d irec tin g TCGF production by normal spleen c e lls stimulated with Concanavalin A (ConA). These studies focused our in te re s t on the growth and stim ulation requirements of d iffe re n t T c e ll subsets. The c e llu la r subsets were defined both fu n c tio n ally and by the se lective expression of Lyt d iffe re n tia tio n antigens. The functional relevance of L yt-2 antigens, which are present on CTL membranes, was fu rth e r assessed in the la s t part of th is study. This approach also resulted in a comparative study of the mechanisms underlying the a c tiv a tin g properties of allo-MHC-antigens and polyclonally ac tiv atin g le c tin s .
Regulation o f T c e ll growth factor production
The peculiar k in etics of accumulation of TCGF a c tiv ity in ConA stimulated cultures f i r s t attrac te d our a tte n tio n . Thus, a study was performed to evaluate the mechanisms regulating TCGF production in these cu ltures. I t was found th at normal spleen c e lls , activated by ConA for 24 h, develop suppressive c e lls th at were able to block de novo TCGF production by fresh spleen c e lls . The e ffe c to r c e lls mediating suppression of TCGF production have been characterized as nonadherent, L y t-2 -p o s itive, T c e lls . The induction time for these suppressor c e lls , in primary cu ltures, was found to be 18 h and to very c lo s e ly p a r a lle l the te rm in a tio n o f TCGF production in s i t u . Furthermore, incubation of 24 h ConA-activated c e lls , in the absence of
the le c tin during a 24-72 h period, leads to a gradual loss of suppressor a c t iv it y . This a c tiv ity however, could be reinduced by re addition of ConA, with the same kin etics of appearance as in primary conditions. The ConA-activated c e lls fu lly recovered th e ir capacity to produce TCGF a fte r a le c tin free culture period of 72 h, showing th at TCGF responsive c e lls were s t i l l present. Taken together with the fa c t th a t no co rrelatio n was found between suppressive and c y to ly tic a c t iv t i t y , these resu lts show th at in h ib itio n of recovery of TCGF in these supernatants is neither due to in s itu absorption of TCGF produced at control rates nor to k il li n g of TCGF producing T c e lls .
2. Growth and s tim u la to ry requirem ents fo r d if f e r e n t T c e ll populations.
I t is now generally accepted, th at TCGF conditioned medium is s u ffic ie n t to maintain CTL lin e s or clones in^ v itro fo r a prolonged period of time. In contrast, however, the stim ulation requirements for TH c e lls are s t i l l controversial and most of the present studies has been performed on TH c e ll lin es or clones, maintained in v itro fo r a prolonged period of time.
The present study is concerned with the ac tiv atio n and growth requirement of CTL and TH c e lls in an unprimed p u rifie d T c e ll population. As markers for these subpopulations, both the expression of Lyt d iffe re n tia tio n antigens and functional assays for various e ffe c to r c e ll a c t iv itie s have been used. The experimental findings can be summarized as follow s. A profound difference e x is t between CTL and TH c e ll precursors in th e ir ac tiv atio n and growth requirements. Thus, w hile L yt-2+ CTL precursors can be induced to TCGF r e a c tiv ity by soluble le c tin in the absence of adherent accessory c e lls , Lyt-2~ TH c e ll precursors show a s t r ic t accessory c e ll requirement for ac tiv atio n and p ro lite r a tio n . Furthermore, exogeneously added TCGF has no e ffe c t on TH c e ll p ro life ra tio n and effe cto r functions, in contrast to CTL p r o life r a tio n , which is exclusively dependent on th is growth fa c to r.
F in a lly , the low levels of e ffe c to r helper functions, detected in a primary response to allo-MHC-antigens or le c tin s , appears to be due to the development of suppressive L yt-2+ T c e lls . Thus, in order to study TH c e ll growth and d iffe re n tia tio n to effe cto r c e lls , L yt-2+ T c e lls must be removed from an unselected T ce ll population.
3. Recognition requirements involved in induction and ta rg e t c e ll ly s is o f cytotoxic T lymphocytes.
The la s t part of the present investig ation deals with the e ffe c t of monoclonal antibodies (mAb) against c e ll surface molecules at various le v e ls of CTL induction, growth and effe cto r functions. All the mAb used are of the same Ig class and specific for T c e ll d iffe re n tia tio n antigens, expressed in equivalent amounts, such th at sp ecific effe c ts can be evaluated. In th is approch, two d is tin c t systems of a c tiv atio n were used, namely a llo MHC Class I antigens provided as U V -irradiated allogeneic stim ulator c e lls or a 4 h pulse with a p olyclonally a c tiv a tin g le c tin . Both procedures were shown to s e le c tiv e ly induce L y t-2 + CTL precursors to TCGF responsiveness without leading to m itosis, unless exogeneous TCGF is added to the system. Thus, these systems makes i t possible to se le c tiv e ly study i n i t i a l antigen recognition, in te rja c e n t TCGF dependent clonal expansion and targ e t c e ll cytolysis by L y t-2 + CTL. In contrast to a n ti-T h y-1 .2 and a n ti-L y t- 1.2 mAb, a profound in h ib itio n was exerted by a n ti-L y t-2 .2 mAb on both CTL functions which involve antigen recognition, namely the i n i t i a l induction step and ta rg e t c e ll cy to ly s is . The in te rja c e n t TCGF dependent p r o life r a tio n , however, was re sista n t to the same mAb. The p ara lle lism observed in these protocols between antigen-specific and lectin-m ediated a c tiv atio n or targ e t c e ll cy to ly s is , in regard to the s u s c e p tib ility to in h ib itio n by a n ti-L y t-2 .2 mAb, lead to a general hypothesis on CTL a c tiv a tio n . I t is assumed th at Lyt-2 antigens are fu n c tio n a lly associated with sp ecific recognition receptors on CTL and
since lectin-dependent induction is also sensitive to Lyt-2 blocking, i t is postulated th a t le c tin s ac tiv ate CTL upon binding to th is receptor complex as w e ll. Since le c tin s , however, bind to many other surface glycoproteins on T c e lls , but only the Lyt-2 associated sites are trig g e rin g complexes, i t is concluded th at CTL can only be induced by interaction s with sp ecific c lo n a lly d is trib u ted receptors. I t would follow th at "mitogenic" le c tin s bind to c lo n a lly d istribu ted receptors on CTL and th a t such experim ental to o ls may be o f value in investig ation s of T c e ll reperto ires and MHC r e s tr ic i ton.
I I I GENERAL DISCUSSION
In the following sections I have attempted to summarize and discuss my experimental findings in general terms. In ad dition, a b rie f overview is given for each p art of the present work.
Since the publications and manuscripts, on which the thesis is based, already contain the detailed experimental resu lts and separate discussions, there is no reason for me to repeat them here. Rather, I w ill especially concentrate on the more controversial parts of th is work, where c o n flic tin g resu lts have been reported, or d iffe re n t opinions e x is t at present. I think th at i t w ill be quite obvious for the non-immunologists th at immunology can not be noted for it s lack of controversies.
1. Negative regulation in a T c e ll »response
Regulation of immune responses is often thought to be due to the a c tiv ity of suppressor T c e lls . Among T c e lls , however, the Ts subgroup is the most co n tro versial. These c e lls have been a ttrib u te d e ffe c to r functions in both c e ll mediated and humoral immune responses. Various kinds of Ts c e lls , which d iffe r in s p e c ific ity and mechanisms of action have been described in the lit e r a tu r e , for instance; allotype sp ecific Ts c e lls (9 9 ), idiotype sp ecific Ts c e lls (100,101), c a rrie r protein sp ec ific Ts c e lls (102) tumor sp ecific Ts c e lls (10 3,10 4), and Ts c e lls in h ib itin g delayed type h yp ersen sitivity responses (105,106).
Many immunologists are very c r it ic a l in regard to the in te rp re ta tio n of experiments with Ts c e lls and even to accept th e ir existance as a defined member of the immune system. One of the reasons for th is is th a t an immune response th a t has fa ile d to appear, can always be e x p la in e d in many d if f e r e n t ways. Thus, i t is an in h e re n t c h a ra c te ris tic in most systems measuring suppression, th at i t is almost
impossible to include a ll the appropriate controls. Another serious problem in the T suppressor f ie ld is th a t the only c e llu la r marker defining the Ts c e ll population, namely the I-J encoded determinant, has very recently been questioned.
Steinmetz e t al_ (107) have shown, by using overlapping cloned genomic DNA sequences, th at the I-J subregion is confined to a 3,4 kb region of DNA, th at is far too l i t t l e to encode the t r a it s of the I- J region.
From these molecular genetic re s u lts , i t could be concluded th at the I - J subregion may be a genetic a r t if a c t . I t should be noted in th is context, th at ever since the o rig in al description of the I-J region by Murphy e t a \ _ (9 9 ), putative I-J sp ec ific antisera have been extensively used in Ts studies (108). I t is remarkable, however, th a t no gene product has been iso lated and characterized, despite the fa c t th at ce ll lin e s and clones, expressing I-J determinants, have been reported (10 9,11 0), as well as monoclonal antibodies with th is s p e c ific ity ( 111, 112).
Despite the problems outlined above, very few immunologists today dispute the existence of a T c e ll population th at is able to block other T c e ll functions. I t is s t i l l a matter of controversy, however, whether or not these functions are mediated by a specialized T c e ll subpopulation, and the overall importance of th is kind of immune regulation remains an open question. Furthermore, the c e llu la r markers th a t have been reported for Ts c e lls are also present on CTL or TH c e lls , except the elusive I-J determinant.
As in many other cases, also for Ts c e lls mitogenic le c tin s have been explored to ac tiv a te e ffe c to r functions. In several systems, ConA activated T c e lls have been shown to suppress mitogenic and antigen sp e c ific responses (113,114). Furthermore, spleen c e lls from mice in jected with ConA in vivo are markedly suppressed in th e ir capacity to be restimulated in^ v itro with ConA or PHA (115,116).
In the present study, the regulatory mechanisms d irec tin g TCGF production in a ConA driven bulk cu ltu re , were investigated (paper
I , I I ) . We have found th at TCGF production In ConA stimulated cultures is arrested 18-24 h a fte r i n it ia t io n . Cell tra n s fer experiments indicated th a t suppressive c e lls , able to block de novo TCGF production, were induced in such cu ltures. E ffector c e lls mediating suppression were characterized as non adherent, ra d io re s is ta n t, L y t-2 - p o s itiv e T c e lls . The induction phase of these c e lls is radio sensitive and requires 18 h of cu lture with the le c tin .
2. Possible mechanisms o f action o f T c e lls capable o f blocking TCGF production.
Several investigators have proposed th at Ts c e lls exert th e ir e ffe c to r function by secreting soluble fac to rs. Many of these suppressor factors have been characterized as antigen s p e c ific , idiotype and I-J p o s itiv e (103,117,118). As we are very c r it ic a l towards the concept of antigen sp ec ific factors in general, and as the existence of I- J determinants is highly questionable, we do not consider these factors to co n stitute an a ttr a c tiv e general mechanism of suppression.
A soluble in h ib ito r th a t suppresses the p ro life r a tiv e response of T c e lls to alloantigen have been found in MLC supernatants. This fa c to r, with an apparent MW of 10.000, was found to suppress the production but not the function of TCGF (119). Also, a substance has been found in sera from normal mice, th a t in h ib it the function of TCGF (12 0). Our attempts to block TCGF production with culture supernatants from suppressed cu ltures, however, have been without success. Although some in h ib itio n was observed, i t was not at a ll comparable with the potent suppressive e ffe c t exerted by 24 h ConA activated c e lls .
Another plausible mechanism of suppression is le c tin f a c ilit a t e d k i l li n g of TCGF producing TH c e lls in cu ltu re , by activated CTLs.
This, however, is u n likely because 18 h of ConA a c tiv atio n is not s u ffic ie n t to induce e f f ic ie n t cytotoxic a c tiv ity by CTLs and no
c o rrelatio n was found between cytotoxic and suDDressive a c t iv it y . Others have also shown th at suppression in MIC systems by ConA or antigen activated suppressive c e lls , is not due to cytotoxic a c tiv ity ( 121, 122).
Larsson et al_ have shown th at a 4 h pulse with ConA is s u ffic ie n t to render T c e lls capable of absorbing TCGF a c t iv it y (3 3 ). Furthermore, i t has been demonstrated th at a TCGF dependent CTL clone can suppress the generation of c y to ly tic a c tiv ity in MLC cu ltures, presumably by absorbing TCGF (123). I t is c le a r, th ere fo re, th at the suppression of recovered TCGF a c tiv ity in supernatants exerted by ConA activated T c e lls can, at le a s t p a rtly , be a ttrib u te d to absorption. We do not accept, however, th at absorption of TCGF produced at normal rates is the sole mechanism of suppression as proposed by Palacios et al_ (124).
The evidence supporting some other type of suppression mechanism is the f o il owing:
1. The number of activated T c e lls necessary to remove TCGF, in an amount equivalent to th at present in a 24 h ConA supernatant, is hundred fold higher than the number of irra d ia te d Ts c e lls able to abrogate TCGF production (3 8 ,8 9 ).
2. The assumption by Palacios et å ] _ is based on the finding th at the p r o life r a tiv e response, in a mixed culture of fresh and ConA activated c e lls , is restored by addition of preformed TCGF.
However, th is evidence is irre le v a n t as the same re s u lt could be expected regardless of the mechanism(s) behind suppression of TCGF production. Furthermore, i t was reported th at addition of the immuno-suppressive drug cyclosporin A to the primary ConA cu lture abrogates induction of Ts c e lls . From our point of view, th is could also be expected, since i t confirms our observations th at DNA re p lic a tio n is required for induction of effe cto r Ts c e lls ( I I ) .
3. Since ConA activated CTLs absorb TCGF very e f f ic ie n t ly , there should be a clear co rrelatio n between suppressive and cytotoxic
a c t iv it y , and th is is not the case ( I I ) .
4. The maintenance of suppressive a c tiv ity required the continuous presence of ConA, and we have shown th at Ts c e ll a c tiv ity reverts w ith in a period of 48 h to 72 h in the absence of le c tin . These c e lls , however, s t i l l p ro life r a te in response to TCGF, and thus must be competent to absorb TCGF. Even more d ir e c tly , Kumagai e t al have reported th a t spleen c e lls deprived of ConA re ta in functional TCGF receptors for at le a s t 72 h (125).
5. The induction time for Ts c e ll a c tiv ity is between 18-24 h and p a ra lle ls very closely the termination of TCGF production in s i t u , whereas i t has been demonstrated th at a 4 h pulse with ConA is s u ffic ie n t to ac tiv ate in spleen c e lls competence to absorb TCGF (3 3 ).
Taken together, our observations suggest th at suppression of TCGF production is an active process mediated by L yt-2+ T c e lls , which d ir e c tly and reversib ly block TCGF producing c e lls . Such r e v e r s ib ilit y excludes k i l li n g of producer c e lls , and the precise mechanism of in h ib itio n s t i l l remains unknown. There are some indications th a t c e l l - to -c e ll contact is required, since the suppressive a c tiv ity on a per c e ll basis decreases with the c e ll density in te s t c u ltu re . The to ta l TCGF production in control cu ltu res , however, decreses d ra s tic a lly with the c e ll density and these experiments are consequently d i f f i c u l t to perform in a low c e ll density c u ltu re.
Nevertheless, recent resu lts extend these findings and give them larg er b io log ical perspectives. As discussed below, i t is clear th a t ConA induces very lim ite d p ro life ra tio n of L y t-2 - TCGF producing c e lls in an unselected splenic T c e ll population ( IV ) . In te re s tin g ly , however, removal of L y t-2 + c e ll from such a population enables p ro life ra tio n and enrichment of L y t-2 - TCGF producing c e lls . Thus, ConA per se is able to induce L y t-2 - c e lls to p ro life ra tio n and e ffe c to r functions, but i t seems th a t the simultaneous a c tiv atio n of L yt-2+ suppressive c e lls blocks various types of TH c e ll functions, in addition to re s u lt in
p re fe re n tia l growth of L yt-2+ c e lls in cultures of unselected c e lls .
Other phenomena of ConA induced suppression of T c e ll functions th at might be related to th is discussion have also been observed in lim itin g d ilu tio n type of experiments. Thus, an unfractionated spleen ce ll population contains c e lls able to in h ib it growth of a large fra ctio n of re ac tiv e clones stimulated with ConA and TCGF (126). Eichmann e t al^
have observed a sim ilar phenomenon of suppression in frequency determinations of streptococcus A sp ecific TH c e lls , in ConA induced splenic T c e ll populations (127).
3. The ta rg e t c e ll for TCGF
The o rig in al finding by G ill is and Smith (78) th a t tumor sp ecific CTL could be enriched and maintained in the presence of ConA conditioned media, indicated th at CTLs were the targets for the growth promoting a c tiv it y contained in those supernatants, which was then defined as TCGF. I t was la te r reported that TH clones could also be maintained and retained th e ir sp ec ific e ffe c to r function, in the presence of preformed TCGF and in the absence of antigen and accessory c e lls (128,129). In co n trast, however, Larsson et al_ (130,131) found th at the growth of TH c e lls , enriched by an a lte rn a tiv e method, was not supported by TCGF, but th at continuous antigen recognition on competent stim ulator c e lls was required for the maintenance of sp ecific helper function. Most l ik e l y , these c o n flic tin g reports r e fle c t how various iji v itro selection and enrichment techniques can give ris e to T c e lls with d iffe r e n t growth requirements. I t seems th at exogenously supplied TCGF, a t the beginning of enrichment cultures for TH c e lls , resu lts in c e lls th a t are s t r ic t ly dependent for growth on TCGF or some other unknown fac to r present in the TCGF containing preparation used.
At present i t is hard to judge whether i t is the TCGF-dependent or TCGF-independent TH c e ll which is the "physiological" type of helper c e lls . Both types of TH c e lls have been shown to be active in vivo and re co n s titu te T dependent responses in nude mice (132,133). Furthermore, i t has been reported th at some TH clones release TCGF upon antigenic
Stimulation and, therefore may be growth supported by th e ir "own" TC6F (12 9,13 4,13 5). The major problem in th is context is the lack of homogeneous preparations of mouse-derived TCGF. With the TCGF preparations a v ailab le today, i t is impossible to state whether the same, or d is tin c t, molecular e n titie s promote both TH c e ll and CTL p r o lite r a tio n .
In most studies concerning TH c e ll growth, extensively primed TH c e lls or clones enriched and maintained i r ± v itro fo r prolonged periods of time have been used. Obviously, these studies may produce c o n flic tin g re su lts due to the various in_ v itro selection protocols. The discrepencies outlined above make i t important to investig ate a c tiv a tio n and growth requirements for various T c e ll subpopulations in unprimed populations. The experiments described in paper I I I , IV , V and VI demonstrate, th at a fte r i n i t i a l a c tiv a tio n , TCGF dependent growth is almost exclusively acquired by L y t-2 + CTL precursors. The evidence for th is conclusion is the follow ing:
1. A Lyt-2" T c e ll population is not able to i n i t i a t e TCGF dependent exponential growth a fte r a 4 h pulse with ConA ( I I I ) . Furthermore, lim itin g d ilu tio n experiments reveal th at the frequency of ConA inducible TCGF re ac tiv e clones, is 99 percent decreased i f L y t-2 + c e lls are removed from an unselected T c e ll population ( I I I )
2. UV-irradiated allogeneic c e lls , which are nonstimulatory in conventional MLC, s e le c tiv e ly induce L yt-2+ c e lls to in i t i a t e TCGF dependent growth (V ). Under these conditions, solely antigen triggered TCGF dependent c e lls are able to p ro life ra te and mature, since U V -irra d ia tio n , or glutaraldehyde fix a tio n , of stim ulator c e lls does not induce TH c e ll functions in primary cultures
( V ,136,137).
3. T c e ll populations obtained by i n i t i a l activatio n with ConA or UV- irra d ia te d stim ulator c e lls contain, a fte r 5 days of TCGF dependent p r o life r a tio n , a cytotoxic potential comparable to a cloned CTL lin e . Such a TCGF expanded T c e ll population is , on a
per ce ll basis, about 5 times more e f f ic ie n t in target c e ll ly s is , as compared to e ffe c to r c e lls generated in a conventional MLC.
These results strongly ind icate a homogeneous CTL response and th e re fo re , a strong selection for CTL growth by TCGF.
The resu lts outlined above are also supported by the previous demonstration by Larsson et a \ _ (136,139) th a t a ll TCGF reac tiv e clones, derived a fte r polyclonal a c tiv atio n of splenic T c e lls with eith e r ConA or LA, are able to k i l l tumor targets in the presence of PHA. The high frequency of ConA-inducible TCGF-reactive c e lls obtained in these protocols (1 in 5) (111,138,139), indicates th a t both the Lyt-1"23+ and L y t - l +23+ subsets are induced to TCGF r e a c tiv ity , as previously suggested by others (138,140,141). The same is also shown by the decreased TCGF dependent p ro life ra tio n observed in negatively selected Lyt-1"23+ c e ll populations, as compared to the response obtained with unselected T c e ll populations ( I I I ) . The phenotype L y t-1 -2 3 - is used in th is discussion on operational basis only, to indicate T c e ll populations remaining a fte r treatment with monoclonal a n ti-L y t-1 antibodies and complement. I am aware th at the Lyt-1 marker appears to be expressed on a ll T c e lls (14 2), but the c y to to x ic ity data would in d icate large differences in the concentration of th is marker on d iffe r e n t sets of T c e lls .
4. Growth requirements fo r Lyt-2" T helper c e ll precursors
In the previous section, the TCGF reactive c e ll in primary responses was characterized as a L yt-2+ CTL precursor, and the results ind icate th a t CTLs and TH c e lls are not only fu n c tio n ally d is tin c t but also th at they have d iffe r e n tia l requirements for ac tiv atio n and growth. I t is generally accepted th a t la p o s itive macrophages play an essential role in antigen presentation (36 ,1 43 -14 7), but as recently shown and confirmed in th is in v e s tig a tio n , i t appears th at ConA induced acqu isitio n of TCGF r e a c tiv ity is an accessory c e ll independent event (V ,3 2 ,3 3 ). On the other hand, as outlined in the previous section, the
