el de la mediación celular en la mediación de la respuesta inflamatoria

8/7/2019 el de la mediacin celular en la mediacin de la respuesta inflamatoria

1/28

PARKE-DAVISLECTURETHE ROLE OF CELL-MEDIATEDIMMUNITY IN THE INDUCTIONOF INFLAMMATORYRESPONSES


2/28

The American Journal ofPATHOLOGY

SEPTEMBER 1977 * VOLUME 88 , NUMBER 3

The Role of Cell-Mediated Immunity in the Inductionof Inflammatory ResponsesParke-Davis Award Lecture, 1977Stanley Cohen, MD

Reactions of cell-mediated immunity fall into two broad categories: those that involvedirect participation of intact lymphocytes in the effector mechanism of the reaction andthose that involve mediation by soluble lymphocyte-derived factors known as lympho-kines. The first kind of reaction is essentially limited to lymphocyte-dependent cyto-toxicity, although certain aspects of T cell-B cell cooperation may fall into this categoryas well. The second category appears to comprise the bulk of the so-called cell-mediated immune response and provides a link between this system and the in-flammatory system. Various lymphokines have been shown to exert profound influenceupon inflammatory cell metabolism, cell surface properties, patterns of cell migration,and the activation of cells for various biologic activities involved in host defense.Although substantial information is now available about various physicochemical aswell as biologic properties of lymphokines, purification and characterization data ar e asye t too incomplete to allow us to ascribe al l of these activities to discrete mediatormolecules. Current work involving the development of antibody-based techniques formediator assay may shed light on this issue. Information on the kinds of cells capable oflymphokine production is now available. Contrary to prior expectation, T cells are no tunique in their capacity for lymphokine production. Under appropriate circumstances,B cells and even nonlymphoid cells can do so as well. The unique property of lym-phocytes in this regard appears to relate to their ability to respond to certain specializedsignals such as specific antigen or an appropriate mitogen. Mediator production pe r semay represent a general biologic phenomenon. Although lymphokines have beendefined mainly in terms of in vitro assays, early speculations about their in vivoimportance are proving correct. Evidence for the role of lymphokines comes fromstudies involving detection of lymphokines in tissues, studies involving injection ofexogenous lymphokines, and studies involving suppression of in vivo reactions by

From the Department of Pathology, University of Connecticut Health Center, Farmington,Connecticut.Supported by Grants AI-12477, AI-12225, AI-13258, and CA-19286 from th e National Institutes ofHealth.Delivered at the Sixty-first Annual Meeting of the Federation of American Societies forExperimental Biology, Chicago, Ill., April 4, 1977.Address reprint requests to Dr . Stanley Cohen, Department of Pathology, University of Con-necticut Health Center, Farmington, Ct 06032.

502


3/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 503September 1977various techniques. The use of antilymphokine antibodies has proven useful in th elatter kinds of experiments. Work in many laboratories is beginning to relate thesefindings to clinically relevant situations. A major unsolved problem relates to th eregulation and control of lympholine production and activity. At present only a limitedbody of information is available on this point. This is a potentially fruitful area forfuture investigation since it may provide techniques for manipulating th e immunesystem in ways t ha t a re clinically useful. (Am J Pathol 88:501-528, 1977)

IN SPITE OF AN ENORMOUS BODY of work on th e subject, th e con-cept of cell-mediated immunity remains a puzzling one. Depending uponones' research interests or personal bias, th e phrase conjures a limitedvision of either lymphocyte-dependent cytotoxicity on the one hand or theclassic cutaneous tuberculin-type delayed hypersensitivity reaction on th eother. Even the names represent misnomers. The reaction is no t delayed,at least, no t in comparison with lesions mediated by reaginic antibodywhich have latent periods of 48 to 72 hours. Moreover, the classic timecourse of 24 to 48 hours for delayed hypersensitivity is merely a function ofits traditional site for elicitation, the skin. As we shall see, in other locationssuch reactions may reach maximal intensity in as little as 4 hours. More-over, this class of immunologic reaction is no more or less cell-mediated thanantibody-dependent reactions which are ultimately due to th e participa-tion of a lymphocyte or plasma cell. With the exception of those reactionswhich involve direct cell killing, most reactions of cell-mediated immun-it y are not dependent upon the lymphocyte as effector, but rather onsoluble lymphocyte products known as lymphokines. Finally, "hyper-sensitivity" is obviously no t a good way of describing a normal physiologicprocess.It is even more instructive to look at a typical textbook definition ofdelayed hypersensitivity. It is usually defined as a slowly evolving reactioncomposed predominantly of mononuclear cells accumulating at the testsite. The reaction is indurated, and this is said to be due to the intensecellular infiltration. It is capable of transfer by intact cells but not serum.An up-to-date text will usually mention that the reaction is Tcell-dependent and that migration inhibition factor (MIF) released by theT cells locally plays a critical role. Although there is a sound basis for thisgeneral concept, current knowledge reveals it to be a great over-simplication which is accurate only under very limited circumstances. Infact, the one important principle that we can abstract from the definitionis the relationship between cell-mediated immunity and an inflammatorvresponse. It is this relationship that will be explored in th e presentdiscussion.


4/28

504 COHEN American Journalof PathologyThe Nature of the Cellular Infiltrate

Until little more than a decade ago, it was thought that the bulk of themononuclear cells comprising the infiltrate of delayed reactions repre-sented sensitized cells possessing some sort of immunologic specificitydirected toward the antigen and that there was some mechanism thatserved to attract such cells to the reaction site. However, it is now knownthat only a small number of infiltrating cells are specifically sensitizedcells. Moreover, the bulk of the available evidence suggests that there islittle, if any, preferential accumulation of these sensitized cells at thespecific test site. The evidence for these contentions is based mainly ontransfer experiments; these have been extensively reviewed.' Anotherkind of experiment, which does not depend upon labeling of a pool ofdonor cells of differing specificities, is illustrated in Table 1. In this study,2guinea pigs were immunized with two unrelated antigens at differenttimes. 3H-Thymidine was administered in such a manner that it onlylabeled cells proliferating in response to the first antigenic stimulation.Skin tests were performed with each antigen at separate sites 10 days afterthe administration of the first antigen. As can be seen in Table 1, there wasTable 1-Percentage of Labeled Mononuclear Cells in Skin Reactions of Guinea PigsImmunized With Two Antigens and Given 3H-Thymidine After the First

Guinea pi g No. OCBC TAT1 13.5 5. 02 3. 2 7. 23 12.2 1.14 6. 2 4. 95 8.4 8. 46 5. 3 9. 27 5. 0 4.4Mean 7.7 5.7Standard deviation 1.4 1. 0

TAT OCBC8 7. 5 8. 39 2. 3 2.710 5. 3 7.711 6.3 6.112 4.7 2.113 5.0 5.114 2. 6 1. 5Mean 4.8 4.8Standard deviation 0. 7 0.9

Guinea pigs 1 through 7 were immunized with OCBC on Day 1, given 3H-thymidine on Days1 to 4, an d immunized with TAT on Day 6. In guinea pigs 8 through 14, the reverse sequenceof antigens was used. OCBC = o-chlorobenzoyl chloride, TAT = diphtheria toxoid antitoxin.This table was reproduced wih permission from J Immunol 98:269-273, 1967.


5/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 505September 1977no significant difference in the small number of labeled cells at eachreaction site, regardless of which antigen was given first. Comparison ofthe percentage of labeled cells in each test site in individual animalsshowed that of the 14 animals studied, 7 had higher labeled cell counts tothe first antigen, 6 had higher counts to th e second antigen, and 1 showedno difference. Thus, there is no evidence for preferential accumulation ofspecifically sensitized cells at either site.Another misconception about the n at ur e of th e inflammatorv responseis related to the indurated nature of th e delayed hypersensitivitv reactionsite. This ha s been considered to be a macroscopic reflection of the vastnumbers of infiltrating inflammatory cells. Some time ago, however, it wasshown by ourselves 3 and by Nelson 4that one could suppress cutaneousdelayed hypersensitivity reactions by treating the experimental animalswith anticoagulants. We found that this suppression involved only theeffector arm of th e response and no t the state of immunization. Thesuppression was noted mainly as a decrease in induration; histologicobservation utilizing sections routinely stained with hematoxylin andeosin demonstrated substantial mononuclear cell infiltration in the sup-pressed animals. Moreover, the cellular inflammatorv response to non-specific irritants was no t affected in these animals. These early studieshave been brought into focus by the elegant studies of Dvorak and hisassociates.5 Using modern morphologic investigative techniques, thevhave unequivocally demonstrated deposition of fibrin in delaved hvper-sensitivity sites. In comparative studies involving nonindurated, basophil-rich, cell-mediated lesions, they have shown that the induration of de -layed hypersensitivity reactions is not due to the infiltrating cells bu trather is due to the deposition of fibrin, possibly in association with boundextravascular water.Even the nature of the cellular inflammatory infiltrate itself is no t asclear-cut as is generally believed. It is certainly true that the typicalcutaneous manifestations of cell-mediated immunity in the guinea pigand man is a lesion that is composed predominantly of mononuclear cells,with macrophages representing the great majority of these cells andlymphocytes present in lesser numbers. However, in both the guinea pigand man, it is possible to induce cell-mediated immune lesions in whichthe predominant cell is the basophil. These reactions, known as Jones-Mote hypersensitivitv or cutaneous basophil hypersensitivity (CBH) occurwhen the immunizing stimulus is weak, for example, when antigen isincorporated into incomplete Freund's adjuvant. Also, manv contact al -lergies of man seem to be of the CBH variety.In certain circumstances, the eosinophil may play an important role. If


6/28

506 COHEN American Journalof Pathologyone skin-tests a guinea pig, allows the reaction site to heal, and thenreinjects antigen at that site, then the second reaction is more intense,evolves more rapidly, and is largely composed of eosinophils. This isknown as a "retest" reaction.Even the classic delayed hypersensitivity reaction may have a differentmorphologic appearance in different species. For example, in the mouse,6the lesion contains large numbers of neutrophils. Finally, in certain ex-perimental autoimmune disorders in which cell-mediated immunity isthought to play a pathogenetic role (such as thyroiditis or encephalo-myelitis) the predominant infiltrating cell appears to be the lymphocyte.These examples demonstrate that the typical macrophage-rich reactionrepresents bu t one manifestation of delayed hypersensitivity reactions.The central feature common to al l is that a small number of specificallysensitized lymphocytes interact with antigen locally. As a consequence ofthat interaction, a set of events occurs which leads to the generation of aninflammatory response at the reaction site. Those events are known toinvolve the production and release of soluble mediator substances thatare collectively referred to as lymphokines.The LymphokinesMigration Inhibition Factor

Initially, delayed hypersensitivity reactions such as those described inthe previous section were the focus of most of the studies aiming towardthe elucidation of the mechanisms involved in this class of immunologicphenomena. The subsequent discovery of a variety of assay systems thatrepresented in vitro correlates of delayed hypersensitivity markedly ex-tended the range of experiments that could be performed and thus led tomuch of our current information in this field.In vitro studies on delayed hypersensitivity were initiated by Rich andLewis, who demonstrated that tuberclulin preparations inhibited themigration of cells from spleen fragments or peripheral blood taken fromactively immunized animals. This assay was simplified by George andVaughan, who developed the modern capillary tube method. Bloom andBennett and David independently demonstrated that the inhibition ofmigration of peritoneal exudate magrophages from such tubes was due tothe release of a soluble factor (MIF) from sensitized lymphocytes reactingwith antigen in the exudate suspension. The migration inhibition reactioncould be demonstrated using peritoneal exudate cells from nonimmunizedanimals, provided that a source of MIF was included in the incubatingmediums. The best source proved to be supernatant fluids from cultures of


7/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 507September 1977sensitized lymphocytes incubated with specific antigen. These early stud-ie s are discussed and their references cited in a recent review.7Migration inhibition factor is of great historical importance since it wasthe first of th e nonantibody, lymphocyte-derived soluble factors to bedescribed. It was the first lymphokine. Several important general proper-ties of this class of mediator were first discovered in studies of MIF. Thesefactors are of relatively low molecular weight as compared to conven-tional immunoglobulins, and they lack immunologic specificity. In otherwords, although the lymphokines are generated by specific immunologicreactions between antigen and sensitized lymphocytes, once formed thevthemselves do no t require interaction with antigen for the expression oftheir biologic activity. Thus, they are not antibody-like. The few knownsituations (to be discussed later) in which mediators have apparent speci-ficity are no t exceptions to this role; in those circumstances, the specificityis due to the association of antigenic fragments with the mediators.Three important methodologic principles were explicitly formulated inth e early studies of MIF. First, control supernatants, which were obtainedfrom the sensitized cells incubated in the absence of antigen, were alwaysreconstituted with antigen prior to use. Second, controls for viability werealso included. Third, reversibility of effect was demonstrated. By this ismeant that if the migration-inhibiting preparations are allowed to remainin culture, the cells escape from inhibition and ultimately begin to mi -grate. These last two principles are of great practical importance. As asimple reductio ad absurdum, anvthing that kills a macrophage willprevent it from migrating. Since it is only in the verv recent past thatlvmphokines have begun to be characterized and identified in precisephvsicochemical wavs, the distinction between toxicitv and migrationinhibition remains a very important one to make. A surprisingly largenumber of experiments in the literature are difficult to interpret becausethis distinction was no t alwavs appreciated.Kinds of Lymplxin Actvities

For the most part, the lymphokines have been defined in terms of invitro bioassays. They fall into three main categories: lvmphokines thatdamage their target cells (Iymphotoxins), lymphokines that cause cellproliferation (mitogenic factors), and lImphokines that modulate theinflammatory response. In the first categorv are included not onlv factorsthat kill or injure cells, but also factors that suppress certain of theirbiologic activities. One such example is proliferative inhibitory factor(PIF). Certain "helper" substances involved in T cell-B cell cooperation


8/28

508 COHEN American Journalof Pathologymay fall into the second category; MIF, already described, is a member ofthe third category.Inflammatory LymphokinesThe best studied of the lymphokines involved in inflammatory respon-ses are those that affect macrophages. A number of such lymphokineeffects are listed in Table 2. It can be seen that in general these fall intowell-defined categories. Lymphokines have been described that affect thegeneral metabolic properties of these cells. In addition, there are lympho-kines that affect cell surface properties, lymphokines that affect the migra-tion properties of the cell, and lymphokines that activate th e macrophagesfor phagocytosis and killing functions. These activities have been exhaus-tively reviewed in a number of recent publications 7,8 and will not be

discussed in further detail here. It is obvious, however, that MIF falls intothe category of a lymphokine affecting cell migration properties. Anotherimportant lymphokine in this category is the macrophage chemotacticfactor (MCF) first described by Ward et al.9 By definition, chemotaxis isdirected movement. A chemotactic factor converts random cell movementto enhanced movement in the direction of an increasing concentrationgradient of the factor. The usual assay system involves the Boyden doublechamber technique, in which the test substance is placed in the lowercompartment and the target cell suspension in the upper compartment.The two compartments are separated by a micropore filter with pore sizelarge enough to allow cell migration through the filter. The number ofcells migrating through the filter after a given period of incubation in thetest preparation is compared to th e number migrating in a control prepa-Table 2-Ways in Which Lymphokines Have Been Shown to Affect MacrophagesCell metabolismGlucose oxidationProtein synthesisGlucosamine uptakeCell surfaceLoss of cell coatElectrophoretic mobilityInterfacial tensionCell migrationAmeboid motilityMigration inhibitionChemotaxisCell activation

All of abovePhagocytosisCell-mediated killing


9/28

V ol . 8 8, No. 3 CELL-MEDIATED IMMUNITY 509September 1977ration not containing the test substance. Increased numbers of cells in th epresence of th e gradient are taken as evidence of a positive chemotacticresponse. An important additional control here is to include a chamberwith test substance in both compartments to abolish the gradient and soexclude th e possibility that one is detecting nonspecific enhancement ofrandom motility in the presence of the test substance rather than truechemotaxis.S in ce t he above-mentioned lvmphokine activities ar e defined in termsof bioassay rather than chemical identification, whether or not they areascribable to discrete and therefore theoretically separable molecularspecies remains an open question. Certain lymphokines such as MIF andMCF appear to be different, bu t the case for others such as MIF versusmacrophage activating factor (MAF) remains unclear.The effects of lymphokines on other inflammatory cell types have beenless well studied than their effects on macrophages. There is a migrationinhibitory factor for neutrophils, LIF. This appears to be distinct fromMIF, in part on th e basis of inactivation studies using simple mon-osaccharides.10 Also, there is at least one chemotactic factor for eachgranulocvte type and for lymphocytes. These, mainly on the basis ofindirect evidence, appear to be distinct from one another. One suchlymphokine, the eosinophil chemotactic factor (ECF) ha s unusual proper-ties which are worth describing in detail.As background, it is worth recalling that eosinophils are conspicuouscomponents of inflammatory infiltrates at sites of various kinds ofimmunologic reactions. They are found in th e nasopharynx and bronchi of

patients with allergic conditions such as hay fever or asthma, and in theintestinal tract as a consequence of parasitic infestations. In experimentalsituations, they are present in peritoneal exudates following multipleinjections of foreign protein, in lymph nodes draining sites of antigenadministration, in skin following injection of antigen-antibody complexesinto normal animals, and in skin of delayed hypersensitive animals follow-ing multiple injections of antigen at the same site. In addition, eosinophilsmay be found in certain autoimmune lesions which involve cell-mediatedimmune responses, such as experimental autoimmune thyroiditis. Allthese observations (reviewed in Cohen et al.8) show a relationship be-tween the eosinophil and the immune system, and the last two findingscited suggest that this relationship, in part, involves cell-mediatedimmunologic reactions.The various observations described above led us to the suspicion that alvmphokine with chemotactic activity for eosinophils exists. Preliminarystudies of MIF-rich supernatant fluids from antigen-stimulated lym-


10/28

510 COHEN American Journalof Pathologyphocyte cultures failed to detect such a factor. We soon discovered,17'12however, that these supernatants contained an inactive precursor sub-stance (ECFp) which could be activated by reaction with specific pre-formed immune complexes in vitro to generate a potent eosinophil chemo-tactic factor. Studies us ing antigen-coupled and antibody-coupledimmunoadsorbent columns have demonstrated that ECFp behaves asthough it is associated with a "piece" of antigen and that its activation toECF is associated with a loss of that piece. These findings have led to themodel for activation shown in Text-figure 1. In this model, ECFp com-bines with a free antibody site in the immune complex by coprecipitation.As a consequence of this interaction, a modified, active, antigen-freemediator molecule (ECF) is then released. It should be noted that ECF,on injection into guinea pig skin, produces an inflammatory reaction inwhich the predominant cell type is the eosinophil. We have presented

LYMPHOCYTE

7 ~~Ab+ -Ab-Ag-Ab-+j Ab

ECF PRECURSOR

Cz Abi -Ab-Ag-Ab-

AbD + (A)-Ab-Ag-Ab-ECF

EOSINOPHIL 0

TEXT-FIGURE 1-Diagrammatic representation of th e proposed interaction between eosinophilchemotactic factor precursor (ECFp) and immune complexes to generate ECF activity. (Reprintedfrom J Immunol 111: 1450-1458, 1973, with permission.)


11/28

Vol. 88, No. 3 CELL-MEDIATED IMMUNITY 511September 1977indirect evidence that suggests that ECF may play a role in the cutaneous"retest" reaction 13 and in experimental autoimmune thyroiditis."Kinds of Cells That Make LymphokinesLymOcytsOn the basis of much indirect evidence, mostly based upon the fact thatcell-mediated immune reactions appear to be manifestations of T lym-phocyte function, it was assumed that T cells were th e source of al lIvmphokines. Indeed, this was one of the major unquestioned answers oflvmphokine biology. In 1973, Dr. Takeshi Yoshida, Dr . Hidekichi Sono-zaki, and I explored this problem using surface markers such as th ecomplement receptor on lymphocyte membranes as a basis for cell separa-

tion. Guinea pigs were immunized to a variety of antigens. A tvpicalexperiment is shown in Table 3. In al l cases, T cells could make MIF, asexpected. In the case of antigens such as egg albumin (EA) or bovineserum albumin ( BS A) , either unconjugated or conjugated with dinit-rophenyl (DNP), B cells could not make MIF, again as expected. To oursurprise, when purified protein derivative was used, B cells as well as Tcells made MIF.15 A variety of controls demonstrated that the observedactivitv could not be due to th e contaminant T cells in th e B cell prepara-tions.We next looked at suspensions of T and B cells from nonimmunizedguinea pigs. The results are shown in Table 4. In this situation, T cells donot make MIF, confirming th e nonimmune status of th e experimentalanimals. As before, EA cannot induce B cells to make MIF. However, theB cells once more responded to purified protein derivative of tuberculin(PPD) with MIF production. In all cases, B cell-derived MIF and Tcell-derived MIF were identical by all available physicochemical and

Table 3-Migration Inhibition Factor Production by Sensitized T an d B LymphocytesLymphocytest

Antigen* B cells T cellsEA -2.6 7.2* 30.0 2.44DNP-EA -6.4 4.2 28.1 3.0DNP-BSA 7.6 8.4 42.0 5.6PPD 35.1 5.9 30.9 8.4

EA = eg g albumin, DNP = dinitrophenyl, BSA = bovine serum albumin, PPD = purifiedprotein derivative.*Antigen in culture, 50 ug/mI.t Cell concentration, 6 x 106/ml.t Percent inhibition = (1 - areaexp/areacont) x 100.Adapted from J Exp Med 138:784-797, 1973, with permission.


12/28

512 COHEN American Journalof PathologyTable 4-Migration Inhibition Factor Production by Purified Protein Derivative Stimulationof Lymphocytes Obtained From Normal AnimalsLymphocyte source Antigen* B Cellst T CelIst

Lymph node EA 4.6 8.8$ 1. 0 2.9tPPD 31.25.3 5.8 6.3Spleen EA 9.2 5. 1 9.0 3.8DNPEA 8. 8 4.0 9.3 4.0PPD 31.8 4.0 1. 5 2. 5* Antigen concentration, 50 Ag/ml.t Cell concentration, 10 x 106/ml.t Percent inhibition.Adapted from J Exp Med 138:784-797, 1973, with permission.

biologic characterization procedures. These results were pu t into per-spective for us by the report that commercial preparations of PPD are B-cell mitogens. We repeated these experiments using endotoxin lipo-polysaccharide (LPS), another known B-cell mitogen with identical re-sults. These results have been confirmed in many laboratories. Thus, thecapability of human B cells to make MIF; 16 of human 17 and guinea pig 18B cells to make MCF as well as MIF;1` and of human B cells to make(mitogen-induced) interferon 20 have all been recently described. In mostof these studies, a requirement for B-cell activation by mitogenic factorsha s been demonstrated, although in one report 16 specific antigen wasfound capable of serving as inducing agent as well.In retrospect, studies demonstrating MIF-like activity in long-termlymphocyte cultures may have detected a similar effect since some ofthose lines may have been B cell-derived. We explored this by specificallylooking at a number of B and T cell-derived lines and found that bothwere capable of spontaneous MIF production.21 They also had strongneutrophil chemotactic factor (NCF) activity.The reports by ourselves and others that B cells are capable of produc-ing MIF as well as other lymphokines raise an interesting paradox. Theevidence for the association of cellular immunity and T-cell function isoverwhelming. There is also good evidence that many, if not all, of themanifestations of cell-mediated immunity are due to lymphokine-depen-dent mechanisms.8 Since B cells can make lymphokines in vitro, it is atfirst glance difficult to understand the apparent requirement for T cells incell-mediated immunity. Part of the explanation for this paradox lies inthe difference in the requirements for lymphokine induction by T and Bcells. T cells can be activated by either specific antigen or mitogens. Bcells, with the exception of one report to the contrary,16 appear to requiremitogenic activation. In those instances where specific antigen appears toinduce B cells to produce MIF, this activity requires the simultaneous


13/28

Vol. 88, No. 3 CELL-MEDIATED IMMUNITY 513September 1977presence of T cells,22 and it may well be that under those circumstancesthat the T cells are providing an endogenous mitogenic factor.This difference in th e activation requirements for T and B cells cannotprovide th e whole explanation, since many substances such as PPD canfunction both as antigen and as B cell mitogen. One clue comes from thefact that the conditions under which B-cell lymphokine production occursin vitro involve the use of purified subpopulations of lymphocytes, fromwhich T cells are purposelv removed. In our recent investigations, wetherefore included cell mixing experiments. As before, lymphocytes wereseparated on th e basis of th e presence or th e absence of a complementreceptor. These cells, respectively CRL and NCRL, correspond to B cellsand T cells, subject to th e caveat that each procedure for separating cellsbased upon th e exploitation of surface markers may lead to slightlydifferent subsets of these broad categories. In our hands, these techniqueslead to greater than 90% purity of B cells and greater than 95% purity of Tcells. Since we can use nonimmunized animals as cell donors, pure popu-lations of T cells are incapable of generating MIF on exposure to LPS orPPD, and so contamination of B cells by some T cells is not a problem inthese studies.In the first se t of experiments, we obtained lymphocytes from normalguinea pig spleens and separated them as before. We found that in-cubation of 5 X 10 B cells or 1 X 107 B cells/ml with either PPD or LPSgave no significant difference in MIF production. This allowed a mean-ingful comparison between B cells at 1 X 107/ml and 50:50 mixtures of Tand B cells, each at 5 X 106/ml, and avoided th e necessity for differencesin total cell concentrations in the various cultures.The results of the experiment are seen in Table 5. As before, normal Bcells, incubated with either LPS or PPD, generated MIF. T cells underthese conditions gave activity well within background range. Adding Tcells to the B cells totally removed their ability to make MIF. In theseTable 5-Effect of T Cells on B Cell-Derived Migration Inhibition Factor Activity

Percent migration inhibitionLymphocytes* PPD LPS

T cells 6.2 2.5B cells 30.5 29.7T&Bcells 7.0 3. 8Unseparated suspension 2.5 1. 7Purified protein derivative (PPD) and lipopolysaccharide (LPS) were the agents used toactivate cells.* Cells were obtained from nonimmunized animals and fractionated as described in thetext Total cell concentration 1 x 107/ml.


14/28

514 COHEN American Journalof Pathologyexperiments, it should be noted that the original, unseparated lymphocytesuspensions were also incapable of MIF production.These results suggested that we were dealing with a new class of T-cellsuppressor function. This effect might be exerted either by direct cell-cellinteraction or indirectly by a soluble factor. To explore this question, weincubated normal T cells in the presence or absence of PPD. The controlcultures were reconstructed with PPD prior to us e in the usual manner.Such experimental or control supernatants are devoid of MIF activity. B-cell suspensions were incubated with PPD, with or without either experi-mental or control T-cell supernatant. We found that the experimental,activated T-cell supernatant (but no t the control) destroyed th e ability ofthe B cells to make detectable MIF. Similar results were obtained withLPS-induced activation. This suggested that the T-cell suppressive effectwas mediated by a soluble, T cell-derived factor. By definition, thisrepresents yet another lymphokine which we have defined as MIFIF(MIF inhibition factor). MIFIF appears unique, no t only in terms of it ssuppressive effect on B-cell mediator production, but also because it is theproduct of a T cell that is activated by a putative B-cell mitogen. Themechanism of this effect remains to be elucidated. MIFIF could either actdirectly on the B cells to prevent synthesis or release of MIF or might evenbe cytotoxic for those cells. A alternative possibility is that it could interactwith the MIF as it is formed. However, preliminary evidence obtained byincubating MIFIF-containing supernatants directly with MIF-rich super-natants has shown that MIFIF does no t appear to act directly on MIF orto compete with it for a receptor site on it s target cell. Also, there is noevidence for cytotoxic effect. Thus, the first alternative appears likely.Although a number of prior studies have provided evidence for th eregulation of various manifestations of cell-mediated immunity by sup-pressor cells, MIFIF appears to be the first example of suppressor activitywith respect to the lymphokines themselves. It may well be that mecha-nisms similar to those described here will be found operative with respectto conventional T cell-derived MIF as well and that these mechanismsplay an important role in th e regulation of cell-mediated immune reac-tions in vivo.Nonlymphoid CellsWork in our laboratory ha s demonstrated that the infection of a varietyof cells by viruses such as mumps and Newcastle disease virus in vitro or invivo and simian virus 40 (SV40) in vitro leads to th e production oflymphokine-like substances which we have defined as cytokines (reviewedin Cohen6). Similar data has been obtained by others in th e SV40 system.23


15/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 515September 1977The mediators from these sources have strikingly similar phvsicochemicalproperties to those lymphokines with corresponding biologic properties.As will be discussed later, it is possible to prepare antisera against culturesupematants obtained from antigen-activated cultures of sensitizedlymphocytes.24'25 We found 2S that sepharose bead columns conjugated withan antibody prepared against lvmphocyte-derived MIF were capable ofadsorbing cytokine MIF (derived from SV40-infected African green mon-key kidney cells). Columns conjugated with an antibodv prepared againstcontrol supernatants had no such effect. Moreover, we found that cvtokineMIF could substitute for lymphocvte-derived MIF in certain in vivomodel svstems involving desensitization. These results all suggest theidentity of MIF from virus-infected nonlvmphoid cells to conventionalMIF. Further, they raise the interesting possibility that lvmphokine pro-duction pe r se may be a general biologic phenomenon and that what isunique to the immune system is the way in which lvmphoid cells can beactivated for such production either by mitogen or specific antigen.In Vivo Actfit of Lymphokines

As ha s already been alluded to , verv little is known about the finedetails of lymphokine structure and chemistry. All of th e lvmphokineshave been defined in terms of complex, semiquantitative biologic assavs.Thus, although it is widely held that the lymphokines are th e mediators ofmany of the manifestations of cellular immunity in vivo, it is not surpris-ing that direct evidence for this contention ha s been difficult to obtain.The initial demonstrations that MIF-rich supematants could produceinflammatory infiltrates when injected into guinea pig skin providedevidence on this point, but the significance of those observations wasunclear because of the essentially nonspecific nature of the inflammatorvresponse. Many of the investigations performed bv Dr. Takeski Yoshidaand myself have addressed themselves to this issue. The following dis-cussion will outline some of the approaches we have taken.One can readily induce a nonspecific peritoneal inflammatory exudatein delayed hypersensitive guinea pigs by the intraperitoneal injection ofglycogen. Three or four days after the injection of this irritant, theseexudates consist mainly of mononuclear cells, and of these, macrophagespredominate. The intraperitoneal injection of specific antigen at this timecauses a prompt reduction in the macrophage content of such exudates; 5hours following antigen administration there is a drop of approximately90% in the absolute number of these cells recoverable in the peritoneal


16/28

516 COHEN American Journalof Pathologyfluid. There is a slow recovery of macrophage content over the next 24 to48 hours. This reaction is known as the macrophage disappearance reac-tion (MDR) and has been shown to be a manifestation of cellular immun-ity.The MDR is a consequence of increased macrophage adhesivenessinduced by specific antigen. This leads to clumping and sticking ofmacrophages to peritoneal surfaces, with a subsequent drop in the num-ber recoverable from the peritoneal fluid. The bulk of the experimentalevidence favors the view that the macrophage disappearance reaction isan in vivo analog of the migration inhibition reaction in vitro. It providesan especially suitable model system with which to explore lymphokineactivity in vivo.The results of ou r experiments 27 utilizing this model are summarized inText-figure 2. Group I consists of animals undergoing an active MDR asdescribed above. Group II consists of animals undergoing a passive MDR.These guinea pigs were not immunized. Rather, after the induction of theperitoneal exudate, they were given an intraperitoneal injection of sensi-tized lymphocytes (obtained from an immunized donor) and specificantigen. The control group received only lymphocytes and no antigen.This procedure gives a significant MDR. In Group III, the MDR wassuccessfully transferred by lymphocytes enclosed in micropore chamberswhich remained cell-impermeable during the 5-hour time course of theexperiment. In Group IV, the MDR was passively transferred using onlycell-free MIF-rich supernatants from antigen-stimulated lymphocyte cul-tures. These supernatants were subjected to Sephadex chromatography toexclude the presence of immunoglobulins. Text-figure 3 demonstratesthat none of these procedures had demonstrable effects on the granulocyteor lymphocyte content of the exudates.Groups III and IV provide evidence that the MDR, which (as statedabove) is a manifestation of cellular immunity, is mediated by a soluble,lymphocyte-derived factor.

20 I _ TEXT-FIGURE 2-Macrophage disappearance re-action produced in various ways (control, open col-_imns; antigen, solid columns): in actively immu-_0 nized guinea pigs (2 0 Ag antigen) (I); innonimmuinized recipients by peritoneally-derived,0-_ sensitized lymphocytes (I X 101 cells, 20 Ag antigen)Q. -(II); in nonimmunized recipients by sensitized lym-phocytes in micropore chambers (2 X 106 cells, 40 Agantigen) (III); and in nonimmunized recipients by asoluble, lymphocyte-derived factor (IV). (Reprinted] I IIm I from Cell Immunol 2:341-352, 1971, with per-Group mission.)


17/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 517September 1977

- .go t0l

-JTExT-FICG-RE 3-The effect of various experimen-ta l procedures on the lymphocyte and granulocyte con-tent of peritoneal exudates. The experimental groupsare as defined in Text-figure 2. No significant effect on 20these cell populations is seen. (Reprinted from Cell :Immunol 2:341-352. 1971, with permission.)

10

I II mGroup

Ehtact O Sn Reacn SitesManv investigators have attempted to detect MIF activity in extracts ofdelaved hypersensitivity skin reaction sites. These attempts have beenuniformlv unsuccessful. There are several possible explanations for thisfailure, the most obvious one being that MIF might no t play a role in th eevolution of th e delayed hypersensitivity reaction in th e skin. Another andless heretical possibility is that only small amounts of MIF are producedand released at these sites and that this material is rapidly adsorbed ontocells or other tissue constituents. The adsorbed MIF might no t be releasedinto the medium bv th e relativelv crude extraction procedures currentlv inuse.Our own efforts to find M IF activitv in extracts of skin delaved reactionsites were also unsuccessful. However, armed with the knowledge that theIymphocyte culture supernatants could induce an MDR, we injected someof our skin extract fluids into nonimmunized guinea pigs bearing peri-toneal exudates. To our surprise, we achieved a macrophage appearancereaction with a two- to threefold increase in macrophage content consis-tentlv observed. The MDR is thought to be an in vivo manifestation ofMIF activitv. Thus, it seemed likelv that th e anomalous macrophageaccumulation caused by the skin extracts might be due to the chemotacticactivity of MCF, acting under conditions where no MIF was available toinduce cell stickiness and clumping. This expectation was confirmed 2w-hen the extracts of sites of delayed reactions to various protein antigensor of sites of contact reactions to o-chlorobenzoyl chloride, a potentsensitizing agent, were examined by means of th e Boyden double cham-be r assay svstem for the presence of in vitro chemotactic activitv. Chemo-


18/28

518 COHEN American Journalof Pathologytactic activity toward macrophages and lymphocytes, but not neutro-phils, could be detected in these extract fluids. Extracts prepared fromnormal skin or from skin sites of nonspecific inflammatory reactions di dnot show such activity. Physicochemical properties of these chemotacticfactors were similar to those of respective factors previously obtained fromsupernatant fluids from lymphocytes cultured with specific antigen. Thus,at least two lymphokines, or at least lymphokine-like activities, could berecovered from sites of cell-mediated reactions.Serum Lymphokine Activity in Experimental Animals

Yoshida et al.29 have shown that rats of guinea pigs immunized withvarious antigens in complete Freund's adjuvant showed increases in pe-ripheral blood monocytes for at least 2 weeks following immunization, inagreement with previous studies. However, when specific antigen wasinjected intravenously into such immunized animals, there was a promptreduction in the number of circulating monocytes.The maximal disappearance from the circulation occurred at 6 hoursfollowing antigen administration, and there was a slow return to normalover the next 24 hours. It was shown in these studies that the effect ofantigen on blood monocytes was a function of the state of delayedhypersensitivity of the animals and that this reaction, like the MDR, wastherefore a manifestation of cellular immunity.Because of the analogies between this system and the MDR, Dr. Take-shi Yoshida and I attempted to detect MIF in the serum of such animals.30Guinea pigs were immunized by intramuscular injection of 0.5 mg of theBCG strain of tubercle bacilli and subsequently challenged intravenouslywith 0.5 mg of the BCG. This procedure induces an intense state ofdelayed hypersensitivity, and this state is associated with massive spleno-megaly as well as peripheral lymphadenopathy. Sera were obtained fromthese animals at various times following intravenous challenge. Migrationinhibitory factor activity was detectable in the sera of these animals butnever in unimmunized animals or those controls that were immunized butunchallenged. Activity was maximal at 6 to 12 hours, correspondingapproximately to the times of greatest reduction in the numbers of circu-lating monocytes.Injection of Exogenous LymphokinesThe detection of MIF in guinea pig serum following a systemic anti-genic challenge in delayed hypersensitive animals and the temporal rela-tionship between that serum MIF activity and the drop in circulatingmonocytes in such animals prompted a study of the consequences of


19/28

Vol. 88, No. 3 CELL-MEDIATED IMMUNITY 51 9September 1977administering preformed, exogenous lvmphokines to experimental ani-mals.3W We found that the intravenous injection of MIF-containing, butno t control, supematants into normal animals resulted in a drop in circu-lating monocytes comparable in magnitude to that observed in immu-nized animals challenged with specific antigen. Of even greater interestwas th e observation that lymphokine-treated animals developed anergy; ifimmunized animals were treated with lymphokines systemically, theybecame transiently incapable of mounting a skin reaction to th e testantigen. Appropriate controls excluded participation of antigen or anti-body in the suppressive effect. These results again confirmed that lympho-kines could play in vivo roles.Sem Lymphokiwe Actiin Man

The detection of MIF in th e serum of experimental animals raised th epossibility that MIF could be found in human sera during th e course ofcertain diseases. The abilitv of exogenous MIF-containing supernatants tosuppress cutaneous reactivity suggested that diseases associated withanergy might be good candidates. Accordingly, we initiated studies ofpatients with various lymphoproliferative disorders such as chronic lvm-phatic leukemia, multiple myeloma, Hodgkin's disease, non-Hodgkin'slvmphoma, and the Sezarv svndrome. We found that the large majority ofsuch patients had detectable serum MIF.313' MIF was thus found in B-celldisease as well as T-cell disease, in accordance with the observationsalready discussed on the ability of B cells to make lymphokines. Healthyindividuals and patients with a variety of illnesses such as bronchopneu-monia, adenocarcinoma of the colon, astrocvtoma, diabetes, osteoarthritis,myelofibrosis, etc., have a combined incidence of detectable serum MIFof less than 2%. The other disease process in which we are now findingserum MIF with great frequency is sarcoidosis. It is tempting to speculatethat in these conditions one explanation for the cutaneous anergy fre-quently observed involves the circulating lymphokine itself, in analogywith our studies in experimental animals.One additional situation in which serum MIF is detectable is th eposttransplantation hepatic dysfunction syndrome." These patients havederangements in liver function that are often reflected in episodic varia-tions in chemical parameters such as serum glutamic oxaloacetic trans-aminase (SGOT) levels. Such patients also have transient elevations inserum MIF activity, and the appearance of serum MIF invariably pre-cedes elevations in SGOT. This is illustrated in Text-figure 4. The signifi-cance of this temporal relationship is unclear, bu t the data nonethelessdocument another situation in which a lvmphokine activitv may be de-tected in vivo.


20/28

5u CUOHEN American Journalof Pathology120s~~~~~~~~~~~~~~~~~~~~~~~~~ooogf000-10'4~~~~~~~~~~~~~~~~~0cU-z80 E40 20-60 z-

0~~~~~~~~~~~~~~~~~~~~~~4

10

20 TIME IN DAYSORTHOTOPIC TRANSPLANTATIONTEXT-FIG.URE 4-Relationship between serum MIF and SCOT in a patient with posttransplanthepatic dysfunction. MIF activity is plotted in terms of a migration index; the lower the value, thegreater the activity. (Reprinted from Clin Immunol Immunopathol 3:369-378, 1975, with per-mission.)

The Use of Antilymphokine AntibodiesAs has already been alluded to , there is very little data available on theprecise chemical nature of any of the known lymphokines. As statedabove, they are usually the products of T cells, although in special

circumstances B cells may be also induced to produce lymphokines.'5 Inall these situations, multiple lymphokine activities are generated simulta-neously. Moreover, these substances are present in small amounts andcontaminated by a variety of other products of cellular metabolism. Foral l these reasons, the isolation and characterization of lymphokines hasbeen extremely difficult and indeed, even MIF, the best studied of thesefactors, has been characterized only in terms of general physicochemicalproperties and response to enzyme treatment.The general availability of antibody to any of the known lymphokineswould have obvious importance for their purification and, eventually, forstudies on the mechanisms of cell-mediated immunity. However, suchantibody production has been difficult to accomplish because of theimpurity of the fluids which are the source of such factors. We have foundit possible to circumvent this problem by means of a two-stage immuniza-tion procedure.24 25 Although the details of the procedure are beyond thescope of this discussion, it suffices to state that it is possible to produce anantisera in rabbits which, when conjugated to sepharose bead columns,

P 1 0% -O lIr bl


21/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 521September 1977can selectively remove MIF and MCF activity from activated culturesupernatants. Anticontrol activity does no t have this property. The anti-sera so prepared do no t have lymphocyte or macrophage cytotoxicity.The antilymphokine antiserum ha s proven to be a useful tool in explor-ing the in tnvo activity of lymphokines. As previously indicated, lympho-kine-containing supematants can induce mononuclear inflammatorv infil-trates when injected into guinea pig skin. This activity has been ascribedto a skin reactive factor (SRF) which probably is in reality a mixture of th elymphokines that affect mononuclear cells in vitro. The antilymphokineantibody can remove SRF activity from activated supematants. Of evengreater significance is the observation that injection of antilyrnphokine,but no t anticontrol supernatant antibody can profoundly suppress th eability of a previously immunized animal to respond to local antigenicchallenge with a cutaneous delayed hypersensitivity reaction. Thisobservation strongly implicates lymphokine involvement in the ex-pression of delayed hypersensitivity.The Significance of the LymphoidnesWe have seen that lymphokines and lymphokine-like factors that areidentical or very similar to conventional lymphokines may be produced bya variety of cell types and in a variety of in vitro and in vivo situations. It ispossible to document their participation in th e underlying mechanisms ofa number of experimental models. Their general significance in the over-al l biology of the organism, however, remains to be elucidated. I willtouch briefly on three aspects of cellular immunity which focus on thispoint: the cutaneous delayed hypersensitivity reaction itself, infectiousimmunity, and tumor immunity.D-Hay

The evidence for the participation of lymphokines in cutaneous reac-tions has already been summarized. This includes the detection of chemo-tactic lymphokines in extracts of such reaction sites, th e induction ofinflammatory skin reactions by local injection of lymphokines, th e sup-pression of skin reactions by systemic administration of lymphokines, andthe suppression of skin reactions by local administration of anti-lymphokine antibody. Note, however, that this data provides no informa-tion on which lymphokine or lymphokines are specificallv involved in theskin. Although most textbooks stress th e role of MIF in this regard, this isprobably based on historic grounds, since for many years MIF was theonly lymphokine known. I am aware of no evidence implicating MIF in


22/28

522 COHEN American Journalof Pathologycutaneous reactivity. Indeed, ou r work provides suggestive evidence tothe contrary. As already indicated, extracts of skin reactions have chemo-tactic but no t migration inhibitory activity. The failure to detect MIF isprobably not due to adsorption to cell surfaces, since such adsorption isvery difficult to document convincingly even in vitro. It could be due toselective inactivation or destruction of MIF as compared to MCF. How-ever, other evidence points to a role for MCF rather than MIF. As was alsodiscussed, anticoagulants interfere with the expression of cutaneous reac-tion. In vitro, heparin ha s no effect on the migration inhibition reaction,but we have found that it ca n inhibit MCF activity.34 Finally, by appropri-ate choice of antigen, it is possible to desensitize an animal with respect toeither cutaneous reactivity or the macrophage disappearance reaction(MDR). Similarly, one can prevent the appearance of MCF or MIF insuch animals. Also, one ca n achieve either antigen-specific or antigen-nonspecific desensitization with respect to these parameters, dependingupon experimental conditions. Correlations of these patterns of inhibitionprovide evidence that, whereas the MDR appears to be an in vivo mani-festation of MIF activity, the skin reaction seems to be more dependentupon chemotactic activity.35These results underscore the point that the manifestations of cell-mediated immunity are pleomorphic and the different kinds of reactionsmay result from different patterns of lymphokine activity in each.Infectious ImmunityPerhaps the clearest evidence for the role of lymphokines in infectiousimmunity comes from the work of Mackaness and his co-workers on therole of activated macrophages in resistance to infection by such organismsas Listeria monocytogenes. This response was shown to be dependentupon T lymphocytes which secrete lymphokine mediators capable ofactivating macrophages for this role. This topic, which is beyond the scopeof the present discussion, has been extensively reviewed.36Tumor ImmunityA number of investigations in many laboratories (reviewed in Yoshidaand Cohen 37 ) have provided evidence for several different kinds oflymphokine effects in models involving tumor immunity. Certainly, therole of lymphotoxin in this regard has been exhaustively studied. Likeother lymphokines, lymphotoxins, though induced by specific immuno-logic or mitogenic stimuli, are themselves capable of acting in a non-specific manner. Although lymphotoxins ar e thus usually cytotoxic to avariety of both normal and malignant cultured cells, a few reports show


23/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 523September 1977that tumor cells are more susceptible than normal cells to lymphotoxins.

Lvmphokines may also interact with tumor cells in an indirect mannerby activating inflammatory cells. Just as injection of animals with micro-organisms may lead to increased nonspecific resistance to certain otherorganisms by this mechanism, tumor cells may be killed by activatedmacrophages. For example, Churchill et al. have shown that supernatantsfrom cultures of lymphocytes stimulated by soluble protein antigen acti-vate normal macrophages, either as monolayers or in suspension culture."The responsible lymphokine was called muarophage activating factor(MAF). These "activated" macrophages exhibit enhanced cytotoxic ca-pacity against syngeneic Strain 2 hepatoma and MCA-25 sarcoma cells.Little is known of the mechanism underlying th e interaction of th elymphocyte mediator with the surface of the macrophage which results inenhanced cytotoxicity by the activated cells. However, a number ofmorphologic, biochemical, and functional alterations have been describedwhen macrophages are activated by MAF. These include increased adher-ence to glass plastic; increased ruffled membrane movement; increasedphagocvtotic activity; increased glucose oxidation; decrease in levels ofIysosomal enzyme, acid phosphatase, cathepsin D, and,B-glucuronidase;increase in membrane enzyme adenylate cyclase; increase in in-corporation of glucosamine; and enhanced bacteriostasis to Listeria. It isinteresting that present physicochemical characterization studies cannotdistinguish MAF from MIF, although there still remains the possibilitythat several different factors affecting macrophages exist.As another possible direct effect of lymphokines in addition to cytotoxickilling of tumor cells, one can consider the effects of lymphokines on themobility of tumor cells. It has been thought for many years that themobility of individual tumor cells may be important in local tumor spreadand in the establishment of metastasis. It has, however, proved difficult toexamine the mobility of tumor cells in in vitro systems. Recently, murinelvmphoma cells, mastocytoma cells, and various human tumor cells havebeen shown capable of migration from a capillary tube (reviewed inChurchill et al.-). This is a simple system which allows a quantitativeapproach to the examination of the motility of tumor cells en masse. Thetechnique is similar to that used in studies on macrophage migration invitro as a model for in dvio delayed hypersensitivity.Using this procedure we have shown that lymphokine-containing su-pernatants car. inhibit tumor cell migration in vitro. The responsiblefactor is no t separable from MIF. As is the case for the effect of MIF onmacrophages, the reaction is not associated with cytotoxicity for th e targetcell. The initial studies involved the P815 mastocvtoma line,-" but a


24/28

524 COHEN American Journalof Pathologynumber of other tumor cells obtainable in ascitic form al l behave in asimilar manner. It would be of interest to determine if one could recoverthis factor from tumor-bearing animals. Also, does this factor change themetastatic capacity of tumors when administered exogenously?Concluding RemarksIn this presentation, I have discussed the various manifestations of cell-mediated immunity that are attributable to lymphokine-dependent mech-anisms, with special attention to the lymphokines that act on in-flammatory cells. This represents only a small range of biologic activities.Sensitized lymphocytes ca n also release substances that affect vascularpermeability, and these cells may play a role in collagen synthesis anddegradation as well. Thus, they play a pivotal role in various aspects ofinflammatory and reparative processes. It is important to emphasize thatthis role is entirely independent of their ability to make specific antibody.The surprising discovery that nonlymphoid cells can be induced to makelymphokine-like mediators suggests that the lymphokines may be onemanifestation of a general biologic phenomenon that is important in hostdefense and possibly other aspects of homeostasis. In this view, what isunique to the lymphocyte is that it has acquired some specialized meansfor triggering mediator production not available to other cells. If thisnotion is correct, then in an evolutionary sense, cell-mediated immunitymay be the most primitive expression of immunologic reactivity that hasdeveloped out of even more general and nonspecific basic inflammatoryresponses. The exquisite specificity of the antibody system may representa further evolution of this process. The fact that the cooperation betweenT and B cells for antibody synthesis is in part mediated by lymphokine-like molecules provides some tenuous and indirect evidence for this pointof view.Regardless of the meanRs!bby which this complex system has developed, itis clear that nature ha`s


25/28

Vol. 88 , No. 3 CELL-MEDIATEU IMMUNI I Y 525September 1977

PathogeAnie rritnt

Exogenousfactors\ Endogenousfactors

|IMMUNE RESPONSEI CTIONi; | INFLAMMATORY RESPONSE|I~~~~~~ntibody / ~~~~~~~~~~Cytokines

Cellulor Lymphokines

TEXT-FIGURE 5-Some of t he v ar io us pathsways leading to the induction of inflammatoryresponses.With this remarkable generality of mediator function, it is at first glancedifficult to understand the protective capacity of T lymphocytes in variousclinical and experimental settings. However, I suspect that mediators inthe local environment near a lesion, such as those produced by th einfected cells, may no t be very important in terms of initiating a pro-tective inflammatorv response. After all, one must mobilize cells from th emarrow and arrange for their migration through vessels as well as for theirattraction, immobilization, and activation. The lvmphocyte is a perf ectcell for these activities since it travels with the inflammatory crowd. It is amember of a population capable of rapid expansion, and it is capable ofdelivering a "shot" of mediators when and where it is most needed, atsome proximity to the target cells for those mediators.Another problem that awaits resolution relates to the control mecha-nisms involved in lvmphokine-dependent reactions. As we have seen, oneof the lvmphokines is a mitogenic factor, and it is known that thismediator can act on both T and B cells. Since mitogenic activation is asufficient trigger for lvmphokine production, we should have an endlesscvcle of production, release, and stimulation until our total mass consistsof lvmphokines. While this might be an attractive prospect for immunol-ogists looking for large amounts of mediators for purification and charac-terization studies, it is hardlv a biologically useful end stage. Obviously,regulatory mechanisms must exist. The inhibition of B cell-derived MIF

--. . .9Pr-r% IKAKAB Alkll'V'W


26/28

526 COHEN American Journalof Pathologyby MIFIF may provide a clue to the nature of these regulatory mechan-isms. Much future effort will have to be devoted to this issue.Finally, it should be apparent that most of the discoveries in this fieldhave occurred in spite of the available methodology, rather than becauseof it. It is exceedingly difficult to study phenomena whose detectionrequires complex, crude, time-consuming, and relatively unquantitativebiologic assays. For example, seemingly simple questions relating to theinteraction of mediators with putative surface receptors on target cellshave proven impossible to answer. This has been the impetus for theintroduction of more biochemically and physicochemically based assaysin many laboratories throughout the world and has also led us and othersto devise techniques for raising antisera to various lymphokines. Thesemodern approaches hold promise for the development of radio-immunoassay and single cell assays which will be indispensable for furtherprogress in these areas. Hopefully, this will enable lymphokine biology torest on as firm an immunochemical basis as does the antibody system.References

1. McCluskey RT, Leber PD: Cell-mediated reactions in vivo. Mechanisms of Cell-Mediated Immunity. Edited by RT McCluskey, S Cohen. New York, John Wiley &Sons, Inc., 1974, p 12. Cohen S, McCluskey RT, Benacerraf B: Studies on the specificity of the cellularinfiltrate of delayed hypersensitivity reactions. J Immunol 98:269-273, 19673. Cohen S, Benacerraf B, McCluskey RT, Ovary Z: Effect of anticoagulants ondelayed hypersensitivity reactions. J Immunol 98:351-358, 19674. Nelson DS: The effects of anticoagulants and other drugs on cellular and cutaneousreactions to antigen in guinea-pigs with delayed-type hypersensitivity. Immunology9:219-234, 19655. Dvorak HF, Mihm MC Jr , Dvorak AM, Johnson RA, Manseau EJ, Morgan E, ColvinRB: Morphology of delayed type hypersensitivity reactions in man. I. Quantitativedescription of the inflammatory response. Lab Invest 31:111-130, 19746. Cohen S: Cell-mediated immunity and the inflammatory system. Human Pathol7:249-264, 19767. Remold HG, David JR: Migration inhibition factor and other mediators in cell-mediated immunity.' p 258. Cohen S, Ward PA , Bigazzi PE: Cell cooperation in cell-mediated immunity.' p3319. Ward PA , Remold HG, David JR: The production by antigen-stimulated lym-phocytes of a leukotactic factor distinct from migration inhibitory factor. CellImmunol 1:162-174, 197010. Rocklin RE: Role of monosaccharides in the interaction of two lymphocyte media-tors with their target cells. J Immunol 116:816-820, 197611. Cohen S, Ward PA: In vitro and in vivo activity of a lymphocyte and immunecomplex-dependent chemotactic factor for eosinophils. J Exp Med 133:133-146,197112. Torisu M, Yoshida T, Ward PA , Cohen S: Lymphocyte-derived eosinophil chem-otactic factor. II. Studies on the mechanism of activation of the precursor substanceby immune complexes. J Immunol 111:1450-1458, 1973


27/28

Vol. 88 , No. 3 CELL-MEDIATED IMMUNITY 527September 197713 . Leber PD, Milgrom NI, Cohen S: Eosinophils in delayed hypersensitivity skinreaction sites. Immunol Commun 2:615-620, 197314. Cohen S, Rose NR, Brown RC: The appearance of eosiniphils during th e develop-ment of experimental autoimmune thyroiditis in th e guinea pig. Clin ImmunolImmunopathol 2:256-265, 197415. Yoshida T, Sonozaki H, Cohen S: The production of migration inhibition factor byB and T cells of the guinea pig. J Exp Mted 138:784-797, 197316 . Rocklin RE, MacDermott RP , Chess L, Schlossman SF , David JR: Studies onmediator production by highlv purified human T and B lymphocytes. J Exp Med140:1303-1316, 197417. M1ackler BF, Altman LC, Rosenstreich DL, Oppenheim JJ : Induction of lmpho-kine production by EAC and of blastogenesis bv soluble mitogens during human B-cell activation. Nature 249:834-837, 197418. WTahl SM, Iverson GNM, Oppenheim JJ : Induction of guinea pig B-cell lmphokinesynthesis by mitogenic and nonmitogenic signals to Fc, Ig , and C3 receptors. J ExpM.ed 140:1631-1645, 197419. Bloom BR , Stoner G, Gaffnev J, Shevach E, and Green I: Production of migration

inhibitory factor and lymphotoxin by v non-T cells. Eur J Immunol 5:218-220, 197520. Epstein LB , Kreth HW, Herzenberg LA: Fluorescence activated cell sorting ofhuman T and B lymphocytes. II. Identification of th e cell type responsible forinterferon production and cell proliferation in response to mitogens. Cell Immunol12:407-421, 197421. Yoshida T, Kuratsuji T. Takada A, Takada Y, Minowada J, Cohen S: Lvmphokine-like factors produced by human lymphoid cell lines with B or T cell surface markers.J Immunol 117:548-554, 197622. Bloom BR, Shevach E: Requirement for T cells in the production of migrationinhibitory factor. J Exp Mled 142:1306-1311, 197523. Hammond M1 E, Roblin RO, Dvorak ANM. Selvaggio SS, Black PH, D'vorakHF: NMIF-like activity in simian virus 40-transformed 3T3 fibroblast cultures.Science 185:955-957, 197424. Yoshida T, Bigazzi PE , Cohen S: The production of anti-guinea pig lmphokineantibody. J Immunol 114:688-691, 197525. Kuratsuji T, Yoshida T, Cohen S: Anti-lymphokine antibody. II. Specificity ofbiological activity. J Immunol 117:1985-1991, 197626. Yoshida T, Bigazzi PE , Cohen S: Biologic and antigenic similarity of virus-inducedmigration inhibition factor to conventional, lymphocyte-derived migration inhibitionfactor. Proc Natl Acad Sci USA 72:1641-1644, 197527. Sonozaki H, Cohen S: The macrophage disappearance reaction: Mediation by asoluble lymphocyte-derived factor. Cell Immunol 2:341-352, 197128. Cohen S Ward PA , Yoshida T, Burek CL: Biologic activity of extracts of delayedhypersensitivity skin reaction sites. Cell Immunol 9:363-376, 197329. Yoshida T, Benacerraf B, McCluskey RT, Vassalli P: The effect of intravenousantigen on circulating monocytes in animals with delaved hypersensitihity. JImmunol 102:804-811, 196930. Yoshida T, Cohen S: Lvmphokine activity in vivo in relation to circulating mono-cvte levels and delayed skin reactivity. J Immunol 112:1540-1547, 197431. Cohen S, Fisher B, Yoshida T, Bettigole RE: Serum migration inhibitory activity inpatients with lymphoproliferative diseases. N Engl J Med 290:882-886, 197432. Yoshida T, Edelson R, Cohen S, Green I: Nligration inhibitory activity in serumand cell supernatants in patients with Sezary syndrome. J Immunol 114:915-918.1975.a3. Torisu NI . Yoshida T. Cohen S: Serum migration inhibitory activity in patients withpost-transplantation hepatic dysfunction. Clin Immunol Immunopathol 3:369-376.1975


28/28

528 COHEN American Journalof Pathology34. Cohen S, Yoshida T: Unpublished data35. Sonozaki H, Papermaster V, Yoshida T, Cohen S: Desensitization: Effects oncutaneous and peritoneal manifestations of delayed hypersensitivity in relation tolymphokine production. J Immunol 115:1657-1661, 197536. North RJ: Cell-mediated immunity and the response to infection.' p 18 537. Yoshida T, Cohen S: Lymphokines in tumor immunity. Mechanisms of TumorImmunity. Edited by I Green, S Cohen, RT McCluskey. New York, John Wiley &Sons, Inc., 1977, p 8738. Churchill WH Jr , Piessens WF, Sulis CA, David JR: Macrophages activated assuspension cultures with lymphocyte mediators devoid of antigen become cytotoxicfor tumor cells. J Immunol 115:781-786, 197539. Cohen MC, Zeschke R, Bigazzi PE , Yoshida T, Cohen S: Mastocytoma cell migra-tion in vitro: Inhibition by MIF-containing supernatants. J Immunol 114:1641-1643,1975Acknowledgments

I wish to thank Dr. Baruj Benacerraf for his wise counsel and advice through the years, Dr. TakeshiYoshida for his long-standing collaboration and friendship, and Dr. Marion Cohen, the tumorimmunologist with whom I have the closest and most productive relationship-three children.

el de la mediación celular en la mediación de la respuesta inflamatoria

Documents