Locomotion of human lymphoid cells

Locomotion of human lymphoid cells

CF.I.T.UI.ARI hrivTTh-or.OGY33, 257-267(1977) Locomotion I. Effect of Culture of Human Lymphoid and Con A on T and Non-T Cells Lymphocytes Iluma...

749KB Sizes 0 Downloads 18 Views

CF.I.T.UI.ARI hrivTTh-or.OGY33, 257-267(1977)

Locomotion I. Effect of Culture

of Human


and Con A on T and Non-T

Cells Lymphocytes

Iluman peripheral lymphocytes \vere separatrd from ~vholc Moot1 on a I;ic-~111 Hypaque gradient. They were then depleted of monocytes, separated into T a11t1 WII-‘I‘ fractions, and assayed for locomotor responses toward casein and entlotoxili-activatetl serum in Boyden chambers. Non-T cells sho~ved higher random motility than did T cells. Culture prior to assay was necessary in order to demonstrate locomotor activity of T cells, but this requirement, although desirable, was not essential for non-T lymphocytes. It was not necessary for Con A to be present in the culture medium or for either T or non-T lymphocytes to be in blast form to show locomotion.

INTRODUCTIOS Until recently, most experiments on chetnotaxis have concentrated on the bvelldelineated and predictable behavior of neutrophils and monocytes, although ‘l\‘artl and his colleagues (1) reported brieflv that rat lymphocytes did show some locon~otioii toward a factor released by antigen-stimulated guinea pig l~nipli n0dc.s. These cells, however, did not respond to factors which were chemotactic for neut rophils, such as coli7l’lement-derived factors and bacterial culture filtrates. Schreiner and Unanue (2) noted spontaneous motility with T cells and locomotor activity of B lyinphocytes toward anti-imiiiunoglobulin, but did not demonstrate chemotactic behavior when the cells were incubated at 37°C for 4 hr. In 19i.5, Russell and co-workers (3) emphasized the necessity for using lymphocytes in a Mastlike or activated form. These workers reported that blast cells from the lymph nodes of mice the skin of which had been painted with the contact sensitizer, oxazolone. showed a high degree of random motility. Furthermore, they showed that cultured cell lines of B blasts had clearly defined chemotactic activity toward those salne chemoattractants which attract neutrophils and monocytes. hlore recently, LI:ilkinson et al. (4) showed that separated normal human lymphocytes would, after culture for 72 hr with PHA, move in a clearly defined chernotactic way as assessed by varying the concentration of chenioattractant above and below the filter, lvith the result that a gradient effect on cell locomotion was apparent. Here we pm-sue this approach to lymphocyte locomotion : LVe assessthe necessity for culturing lymphocytes, with and without mitogen stimulation, and compare the migration of 1 To whom all correspondence 2 Present address : Bacteriology (;I 1 6KT. Scotland.

Copyright AJJ righti

should and

@ 1977 by Academic Press, Inc. of reproduction in anv form reserved.

be addressed. Immunology












separated populations of lymphocytes : T- and non-T-lymphocyte from monocytes, in the presence of casein and endotoxin-activated MATERIALS


populations freed human serum.


Cell Sources Human peripheral blood lymphocytes were separated from heparinized venous blood taken from normal volunteer donors, using Ficoll-Hypaque gradients. The plasma-Ficoll interface layer containing monocytes and lymphocytes was washed twice with RPMI-1640 medium (Gibco, Grand Island, N.Y.) before further separation and culture. Separation

of Monocytes

and Lymphocytes

from Ficoll-Hypaque


Removal of adherent cells by Sephadex G-10 filtration. Mononuclear cells from Ficoll-Hypaque gradients were filtered through columns of Sephadex G-10 beads (Pharmacia, Uppsala, Sweden) according to the method of Ly and Mishell (5) and were eluted with RPM1 containing 20% heat-inactivated fetal calf serum. Removal of monocytes was checked morphologically using tetrachrome staining. This procedure selectively depletes lymphocytes of adherent latex-ingesting cells, leaving the relative proportions of T and B cells unchanged (6). Removal of adherent cells by nylon wool filtration. T cells were prepared from Ficoll mononuclear cells by filtering them through nylon wool columns according to a modification of the method of Julius et al. (7) for separating murine T lymphocytes. Cells which did not adhere to the column were, in this case, eluted with RPMI-1640 medium containing 20% heat-inactivated FCS, and more than 90% of the nonadherent cells were considered to be T cells as judged by their ability to form rosettes with neuraminadase-treated sheep red blood cells. E-Rosette (sheep red blood cell rosettes) forwzation for separation of T and non-T cells. Lymphocyte-monocyte mixtures from Ficoll-Hypaque separations were washed twice with Hanks’ solution (Gibco) and were adjusted to a concentration of 5 x 106/ml in Hanks. E rosettes were formed by incubating the mononuclear cells at 25°C for 60 min in a 1% suspension of sheep erythrocytes (E), pretreated with neuraminidase, in 20% FCS (y-globulin free, heat inactivated). Neuraminidase-treated sheep E rosettes were prepared by incubating a 1% suspension of washed cells in Hanks’ solution containing 10 ~1 of neuraminidase (Behring Diagnostics, Sommerville, N.J.) per ml for 30 min at 37°C. E-rosetting lymphocytes (T cells) were separated from non rosetting lymphocytes and monocytes by Ficoll-Hypaque sedimentation. The pellet contained E-rosetted T lymphocytes and the interfacial layer contained non-T lymphocytes and monocytes. T cells were recovered from the pellet by lysing the sheep erythrocytes with lo/O NHkCl solution. In some experiments adherent cells (mainly monocytes) were removed from the non-T population by their ability to stick to plastic culture trays after incubation for 1 hr at 37°C in RPM1 containing 20% heat-inactivated FCS. The nonadherent cells (mainly B cells) were removed carefully after gentle rinsing of the adherent layers with RPMI. Non-T cells, after removal of monocytes, were considered in our experiments to be B-cell enriched, containing between 40 and 60% immunoglobulinbearing cells.









Separated and unseparated lymphocytes prepared as previously descril)etl were washed twice in RPMI-1640 medium, resuspended in RPM1 containing 1.i$C pooled normal human serum (heat inactivated), penicillin (100 pg/ml) , streptomycin (100 pg/ml), and fresh glutamine (2 rnJd>, and then cultured with and without mitogen (concanavalin A) for various periods of time (up to 72 hr) at 37°C in an atmosphere of 5% COZ. Con A (Pharmacia Uppsala, Sweden), when added to the cultures, was used at a concentration of 10 ~g/ml. All cells were washed t\\-ice in Gey’s solution and were resuspended at a concentration of 10” cells/ml before assay. Locomotor* Assay Casein (Merck, Darmstadt, West Germay) or endotoxin-activated human serum (EAS), prepared by incubating fresh normal human serum with Escherichia coli endotoxin (Difco, Detroit, Michigan) at a concentration of 50 pg/ml for 30 min at 37”C, was used to promote locomotion. Casein and EAS were used at the respective concentrations of 1 mg/ml and 10% in Gey’s solution unless otherxvise stated. Gey’s solution was used as the negative control in all experiments. The tests were carried out in modified Boyden chambers as described by F1’ilkinson (8). Cell locomotion was assayed by the leading-front method (9) which measures the distance, in micrometers, that cells migrate from the upper compartment, through micropore filters, toward a chemoattractant placed in the lower compartment. In all experiments, filters of S-pm pore size (Millipore Corporation, Bedford, Mass.) were used and incubated for 3 hr, except when otherwise stated, at 37°C in 576 COZ. Distinction between directional migration (chemotaxis) and enhanced nondirectional locomotion (chemokinesis) was made by varying the concentration of chemoattractant above and below the filter using a “checkerboard” type of experiment. Details of these experiments are given in the Results section. In some later experiments the percentage of cells moving in the filter was estimated as well as the distance moved. The number of cells was counted in a defined area of the field under a 40x flat-field objective, both on top of the filter and at lo-pm intervals into the filter. Percentage of cells moving was: 1 (No. of cells in filter)/( No. of cells within and on top of filter) ] x 100. RESULTS The Inflztcnce of Culture Period and Presence of COH.4 on Lymphoc~ltc Miyation Lymphocytes from the plasma-Ficoll interface layer were cultured for 24, 48, and 72 hr in the presence or absence of Con A prior to the chemotaxis test. After washing they were allowed to migrate into Millipore filters for varying times from 30 to 180 min (Fig. 1) in the presence of a gradient of 10% endotoxin-activated serum. Lymphocytes from all three culture periods migrated toward activated serum, although the distance travelled into the filter was greater when the lymphocytes had been cultured for 48 or 72 hr than for 24 hr. There was little or no movement of cells before 30 min of incubation. The cells which had been cultured for 48 or 72 hr reached their maximum distance by 135 min, while those cultured for 24 hr did not reach their maximum until 180 min. To our surprse, however, the presence of












FIG. 1. The influence of culture period and presence of Con A on lymphocyte migration. Lymphocytes were cultured for 24, 48, and 72 hr in the presence (+) or absence (-) of Con A. ( O---O ) Lymphocyte migration toward 10% endotoxin-activated serum ; (O---O) lymphocyte migration towards Gey’s solution (negative control).

Con A in the culture medium did not appear to influence the subsequent migration of lymphocytes, since previous publications have emphasized the necessity for lymphocytes to be in blast form (3) or to be activated by culture with a mitogen (4), in order that locomotor responsescan be demonstrated. Nevertheless, it was possible that the presence of Con A in the culture fluids would alter the nature of the locomotor responsefor it is known (3, 10) that lymphocytes, like other leukocytes, demonstrate both chemotactic (directional) locomotion and chemokinetic (enhanced random) locomotion toward chemo.attractants. To test this, experiments were set up in which the concentration of EAS was varied above and below the filter whereby, according to Zigmond and Hirsch (9), the influence of chemoattractants on either the rate of cell locomotion or the orientation of their locomotion within the filter could be distinguished. Lymphocytes cultured for 24 and 72 hr (with and without Con A) were tested, therefore, in various absolute concentrations and various concentration gradients of EAS (Tables 1 and 2). In Tables 1 and 2, the values in boldface from upper left to lower right represent the migration of the cells in the absenceof a gradient but in increasing absolute concentrations of EAS. If the locomotion of lymphocyte populations increased as the absolute concentration of chemoattractant increased, then the cells showed chemokinesis. Values in parentheses are estimates of what the migration would be from one concentration to another if the cells were moving chemokinetically without any chemotaxis [the calculations necessary to do this are given in an appendix of the paper of Zigmond and Hirsch (9)]. H ow much the observed migration differs from the calculated value is a measure of the effect of the gradient on cell locomotion, and, therefore, is a measure of chemotaxis. In all four checkerboard assays (Tables 1 and 2) using endotoxin-activated serum to stimulate locomotion and in four further checkerboard assays using casein (details not included), there was increased locomotion as the absolute con-

C‘hu~kcrlm~rd Assay Concentration

of the Effect of Varying of EXS on Locomotion

the Conccntratioll Gradient of Cultured I.ymphocytes

and .\bsc)li~lc (2-l hr 1

E:ndotoxin-;lctivnted scr~In1 above filter (%)


“11 :\ 0 0.5 .z.s 0..5 O..i


37 74 80 (87) 83 (89) 98 (99)

8.1 (70) 86 97 (94, 106 (100)

x.5 181 ) 79 (85) 96 102 ( 100)

x0 (82) 00 (92) 10.3 iwj 108

76 92 (94) 10-i (9.5) 9.5 (96)

96 (80) 100 101 (92) 102 196)

102 (90) 92 (97) 89 91 (98)

96 (81) 90 (93 J 89 (92 I 102

:\ 0 0.5 3 ..5 6.5 9.5


centration of chemottractant increased, i.e., the cells showed chemokinesis. There was also some deviation of the observed from the calculated values, i.e., there was some evidence of chemotaxis although the deviations occurred more frequently ill the positive than in the negative gradient. The two checkerhoards in which there was the most consistent evidence of chemotaxis rather than chemokinesis uwv i11 TABLE Checkerboard Assay Concentration Endotoxinactivated serum above

of the Effect of Varying of EM on Locomotion Lymphocyte

filter ~-~~~..----


0 -con

2 the Concentration Gradient of Cultured Lymphocytes migr;ltiou

Endotoxin-activated ..-..-~ ~~~--.

serum ~~..~

and Absoltlte (72 hr)

(wm in 3 hr) below






68 99 (91) 99 (99) 108 (91)

79 (73) 98 113 (10-t) 1 17 (103 )

97 (86) 101 (101) 107 113 (100)

98 (731 101 (98) 109 (10-l I 98

83 99 (102) 100 (105) 102 (99)

100 (89) 110 100 (10.5) 108 (104)

105 (89) 106 (106) 104 102 (101)

103 187) 103 (106) 112 (1011 104


L\ 0 0.5 3.5 6.5 9.5



:I 0 0.5 .z .5 6.5 9.5






the populations of lymphocytes cultured with Con A for 72 hr and it occurred whether endotoxin-activated serum or casein was used to stimulate locomotion. Locomotion


Separated [email protected]


The population of mononuclear cells in the plasma-Ficoll interface layer contains not only both T and B lymphocytes and their various subsets, but also lymphoid cells such as K cells and monocytes. Thus, it is a very heterogeneous population of cells all possibly having different locomotor properties. T cells, however, comprise the majority of cells in the Ficoll layer, and cultivation for 72 hr with Con A will presumably have selectively activated T cells and increased their proportion in the final cell population. It seemed reasonable to suppose that we were measuring the migration of activated T cells, but, nevertheless, it was necessary to exclude the effect of minority populations. As a preliminary step monocytes were removed from the interface cell layer obtained from Ficoll-Hypaque gradients by filtering over Sephadex G-10 beads before culture, and this effectively reduced the number of contaminating monocytes to less than 1% as assessed by tetrachrome staining. The filtrate from the Sephadex column was then passed through nylon wool to remove other adherent cells (B and K cells) to give almost pure T cells. The filtrates from both columns together with a sample of unseparated cells were then cultured for 72 hr with Con A before assaying their capacity to migrate toward casein and activated serum (Table 3). All three samples moved toward casein and EAS, demonstrating first that our previous findings were not due solely to monocytes in the plasma-Ficoll interface and, second, that the T cells show locomotor activity. In the next experiment (Table 4) different methods of cell separation were used. The lymphoid cells were separated into T and non-T populations by rosetting techniques (see Materials and Methods), and the non-T cells were further freed from monocytes by utilizing the ability of monocytes to adhere to plastic at 37°C. The various cell populations were then cultured with Con A as before and were assayed for their migration capacity. Here we calculated the percentages of cells migrating through the filter and measured the distance migrated (Table 4). In both this and TABLE


Locomotion of Cultured Lymphocytes before and after Removal of Adherent Cell Populations” Migration (pm in 3 hr)

Cell type Casein (0.8 mg/ml)

Unseparated lymphocytes (control) Lymphocytes (T + B + K) (monocyte depleted by passing over Sephadex G-10 column) T lymphocytes (all adherent cells removed by passing over Sephadex G-10 and nylon wool column) e All cells cultured for 72 hr with Con A.

Endotoxinactivated serum (10%)


90 f

4 (30)

118 f

5 (32)

35 f

3 (13)

83 f

8 (13)

101 f

7 (24)

17 i

4 (7)

97 f

4 (45)

26 f

2 (21)

105 f 5 (65)





TABLE Locomotion


of Cultured

T and Non-T Kosetting


Lymphocyte Techniquesa







in 3 hr)

Endotoxinactivated serum (10%)

Casein (0.8 mg/ml)

----Unseparated control Non-T lymphocytes* (monocyte depleted) T-lymphocytes” -__--





114 zk 2 (41)

109 zk 3 (48)

46 f

10.5 f 97 f

110 It 4 (35) 91 * 4 (31)

52 f ‘4 (14) 18 f 2 (7) --~.. .~ -.

4 (35) 9 (57)

u All cells cultured for 72 hr with Con A (10 rg/ml). Percentage given in parentheses. * Separated from T cells by E rosettes; separated from monocytes c Separated by formation of E rosettes.

of cells migrating by plastic

3 (22)

in each



subsequent experiments (Table S), separated T cells showed less random movement in Gey’s fluid than both non-T and unseparated control cells, as assessed either by the distance travelled or by the percentage of cells moving. There was clear evidence of migration of separated T cells and the non-T population which contained neither monocytes nor more than a small percentage of contaminating T cells which could have responded to stimulation with C’on A. The Efect of Culturing Cd Poplilatiom


Carl A on the Locolnotion

of Separated


In the next two experiments (Table 5) we compared and contrasted the migt-ation of cell populations separated by the different techniques used previously and cultured for 72 hr with or without Con A. In neither experiment did the presence TABLE The

Effect of Culturing with Non-T) Lymphocyte Cell




Con A on the Locomotion Populations toward Casein A __-~ Casein (0.8 mg/ml)

Unseparated tInseparated ,. I* ?‘t, Non-T Non-T* T" 'I-0

control control


‘I Separated * Senarated


+ +

by passage bv resetting

85 85 102 106

through nylon techniaues.

f f f f

(21) (14) (42) (48) wool.

Migration _~ ~.-

Casein (1 my/ml)

123 118 89 107

+ -

of Unseparated and and Endotoxin-Activated

f zt f f

2 3 5 4


(pm in 3 hr) .~-..--. --

Separated (T and Serum


Endotoxinactivated serum (lOr,‘,l


(30) (29) (31) (28)

32 32 21 22 92 f 98 f 96 f 100


(16) (19) (44) (24)

zt zk f f

1 1 2 2

(13) (21) (17) 119)

39 zk




14 f 19 *

(16) (1.1)





of Con A significantly alter the distance travelled or the percentage of cells moving, and this applied not only to control unseparated and non-T cells but also to T cells separated by either method. The influence of the presence of Con A during culture was also equivocal in two checkerboard experiments with separated T cells. Cells cultured with Con A showed very high chemokinesis, but little chemotaxis toward endotoxin-activated serum when incubated for 180 min (Table 6). Those cultured without Con A showed some chemotaxis in the positive but not in the negative gradient (Table 6). When the incubation period was reduced to 135 min the result was reversed (details not included). Cells cultured with Con A showed more evidence of chemotaxis than those cultured without Con A. Until now our experiments have focused on the locomotor properties of unseparated and separated populations after culture, usually in the presence of mitogen, and we assumed that any cells in the Ficoll-plasma interface population demonstrating locomotor behavior without prior culture were monocytes, but we now tested that assumption on separated populations. T- and non-T-lymphocyte populations were allowed to migrate for various periods of time, either immediately after separation or after 72 hr of culture without Con A (Table 7). For purposes of comparison a control unseparated sample containing 25% monocytes was also tested without prior culture. The results with the purified T-cell population were clear cut. There was virtually no migration at 30 min; thereafter, the cultured cells moved in increasing proportions and increasing distances up to the maximum at 180 min. The noncultured T cells showed virtually no movement at all. The control uncultured cells did, however, manifest some movement which could have been due to monocytes, but which persisted up to 3 hr when one might have expected all monocytes to have



Checkerboard Assay of the Effect of Varying the Concentration Gradient Absolute Concentration of EAS on Locomotion of Cultured T Lymphocytes (72 hr)a Lymphocyte

Endotoxinactivated serum above (%I


filter 0






(pm in 3 hr) below






73 83 (85) 101 (94) 102 (96)

90 (75) 89 107 (100) 106 (103)

103 (82) 98 (93) 105 93 (103)

104 (78) 118 (99) 100 (104) 103


102 (109) 110 109 (108) 106 (106)

105 (109) 107 (109) 108 112 (104)

108 (107) 98 (107) 84 (106) 103

A 0 0.5 3.5 6.5 9.5



A 0 0.5 3.5 6.5 9.5

a T lymphocytes

28 108 (110) 109 (117) 99 (104) prepared

by filtration





Influence Cell

of Culture on the Migration ----__-~-I_ type

of T and



in the Presence

of k.\S

Time of incubation


10 f I)

2 Il.41

SD 78 zt 6 (2.3) 90 f 6 124) 89 f 5 (16)

s r)

114 zt 6 134)

1~3 f 9 jz 0 88 zt 50 f 0 96 f i7 f 28 f

2 (1.2) 1 10.8) 3 18) 3 (12, CO) 5 (17, 3 (28) 6 17)

10 I1 9 i 6 f .13 k 3Of2 0 1.5 & 17 f 7 f

(0.7, 2 (0.4) 2 cO.,?l 2 18) (9) IO! 2 16~ 2 (0) 1 IO.81

migrated through the filter. But there was also evidence for movement in the UIIculturetl non-T cells which had less than 1 s contaminating monocytes. The tlistatices covered nere not as great as in previous experiments lvith culturetl tiott-‘1 cells (Table 7), hut the percentage of cells moving was much the sanie. Thus, it must be concluded that either 13 lymphocytes or K cells shmv locotmtion without l)rior culture.

All saiiiples of control unseparated cells that had lwm cultutwl with Coti .\ uuit;titirtl large nutiil~ers of Mast cells some of \vhich were still clearly e\~itlent on the top of the filter after assay. There kvere far fe\vet- I)last-like cells in the tmt-T ant1 vveti in the T population culturetl with Con .A, and in all satiilks cttlttire~l \vitliotit (.oti .A there \vere virtually no Mast cells. Severtheless, there n2s Very little tlifieretice in the tmtiil~er of cells entering the filters many of \\,hicli were sn~all Iyiiiphorytes. \2’e nttist conclude that assutiiptioti of 1ocotiiotor tiiorpholog~ is not restrictecl to blast cells. There were, however, vet-y clear differences in appearance of cells 0ti the filter hetween non-T ant1 ‘II‘ fractions. The non-T Id typical locomotor mcq~ltolt ~gy. \v;ts hlreatl (nit. ant1 had ;I large cytoplastii to tttirlear mtio, after ttiigr:ttic,ii to1\2rtl caseiti or eti(lotoxin lmtli 011 top aid \vitliiti the filter, \\~liercas ‘I‘ 1ytq~l1~ )cytes \\wc utiifortiily t-ottntl small Iytiipliocytcs on t(q), ~vh~cli ~tretclwtl out u,itltitt





DISCUSSION It is now clear that lymphocytes as well as phagocytic cells will show locomotor responses (24, 11) toward chemical substances. We have confirmed the observations of Wilkinson et al. (4) that human lymphocytes will move in the presence of two chemoattractants, casein and endotoxin-activated serum, that are known to attract both monocytes and polymorphs. Previous publications have stressed the necessity for lymphocytes to be in blast-like form (3) or to be activated by culture with a mitogen such as PHA or Con A (4). It would appear that prior culture does affect the subsequent locomotor response of normal human lymphocytes, since unseparated lymphocytes moved further after 48-72 hr than after 24 hr in culture. Subsequent experiments revealed, however, that the dependence on prior culture was different for T and non-T lymphocytes. It was essential that T cells be cultured in order to show locomotor responses toward chemoattractants, but separated non-T cells (a B-cell-enriched population) would move without prior culture, although they migrated faster and further after culture. Furthermore, it was not necessary for the mitogen, Con A, to be present in the culture fluid. Although results with unseparated lymphocytes indicated that prior culture with Con A had some influence on the type of locomotion in that it encouraged chemotactic movement, in the checkerboard experiments, this conclusion could not be extended to separated T cells. There was no consistent evidence that Con A in the culture fluid predetermined that the cells would behave chemotactically rather than chemokinetically. The results presented here emphasize the importance of using appropriate assay methods for distinguishing chemokinesis from chemotaxis in studying cell locomotion and for using those terms precisely as defined by Keller et al. (12). The fact that Con A was not essential during culture throws into doubt previous conclusions (3, 4) that lymphocytes have to be in blast form in order to migrate. Only a very modest proportion of the cells cultured without Con A were blasts, yet 30-40% of the cells moved when placed on filters. Morphological examination of the cells in and on the filters revealed that many small lymphocytes had moved into the filters with typical locomotor morphology. Recent studies on the locomotor responses of mouse lymphocytes toward antigen (10) have also demonstrated that lymphocytes showing locomotor activity were not restricted to large or blast-like forms. It is obvious that much more needs to be done in order to define the conditions necessary for demonstrating, in a reproducible way, locomotion in human lymphocytes. In particular it is necessary to investigate whether serum albumin stimulates chemokinesis in human lymphocytes, as has been shown with mouse lymphocytes (lo), since it is possible that the presence of serum albumin in both the EAS and casein used here might have masked any chemotactic responses of lymphocytes. But there is no doubt that experiments of this type could be very useful for investigating the factors which affect extravasation of lymphocytes in &JO. For example, the fact that both T and non-T lymphocytes respond to such nonspecific stimuli as casein and endotoxin-activated serum is relevant to the study of how lymphocytes are attracted into sites of inflammation and the gut mucosa, where the absence of any evidence of antigenic attraction in cell traffic studies remains a major problem (13, 14).







The authors would like to thank Dr. Robert ‘4. Good for stimulating discussions while this work was in progress, Dr. Peter Wilkinson for help in preparing the manuscript, Mr. I.cc J. Nelson and Ms. Elisa B. Valera for excellent technical help, Dr. Dolores J. Schendel for advice in separating T cells, and Ms. Nancy Salmon for secretarial assistance. This work was supported in part by an Immunobiology Training Fellowship from the J. M. Foundation (G.J.O.) and NIH Biomedical General Research Grant Nos. 5507RR0534 (D.M.\‘.P.) , AI-1 1843 ( N 1 H ), C.4-08748. and C.4-17404 and the Zelda R. Weintrauh Foundation.

IIEFEIIENCES 1. Ward, P. A., Offen, C. D., and Montgomery, J. R., Fed. E’roc.. 30, 1721, 1971. 2. Schreiner, G. F., and Unanue, E. R., J. Zlrr~rlztwl. 114, 809, 1975. 3. Russell, R. J., Wilkinsoq P. C., Sless, I;., and Parrott, D. M. \‘., .Valfrrc (Lorrdolr) 256. 646, 1975. 4. Wilkinson, P. C., Roberts, J. C., Russell, R. J., and McI,oughlin, M, Clilr. E.rfi. Zlrrwltrfwl. 25, 280, 1976. 5. Ly, I. A., and M&hell, R. I., J. Zwrrrzr?zol. Mcflzods 5, 239, 1974. 6. Berlinger, N. T., Lopez, C., and Good, R. A., Nnturc (Lo~do+a) 260, 145, 1976. 7. Julius, M., Simpson, E., and Herzenherg, L., Ezw. J. Z1nwwzo1. 3, 645, 1973. 8. Wilkinson, P. C., 111. “Chemotaxis and Inflammation” (P. C. Wilkinson, Ed.), p. 33. Churchill J,ivingstone, Edinburgh, 1974. 9. Zigmond, S. H., and Hirsch, J. G., J. Exp. Med. 137, 387, 1973. 10. Wilkinson, P. C., Parrott, D. %I. V., Russell, R. J., and Sless. I;.. J. Esp. M‘,d. 145. 1158. 1977. 11. Wilkinson, P. C., Russell, R. J., Pumphrey, R. S. H., Sless, F., and Parrott, D. 1,I. V., III “Future Trends in Inflammation II” (J. P. Giroud. D. 11. Willoughby, and D. F’. L’elo, Eds.), p. 243. Birkhauser, Verlag, Basel, 1975. 12. Keller, H. U., Wilkinson, P. C., Abercromhie. M., Becker, E. L., Hirsch, J. G., Xliller, hf. E. Ramsey, W., Scott, and Zigmond, S. H., Cli~a. Exp. 1~1zww101. 27, 377, 1977. 13. Parrott, D. M. V., Rose, M. L., Sless, F., deFreitas, A., and Bruce, R. G., III “Future Trends in Inflammation II” (J. P. Giroud, D. A. Willoughby, and D. P. Velo, Etls.). p. 32. Birkhauser, Verlag, Base], 1975. 14. Rose, M. I.., Parrott, D. M. V., and Bruce, R. G.. L‘rall. Z~r~r/w+zijl. 27. 36, 1976.