human t- cell heterogeneity as...

Human T - Cell Heterogeneity as Delineated with a

Specific Human Thymus Lymphocyte Antiserum

IN VITRO EFFECTSONMITOGENRESPONSE,MIXED LEUKOCYTECULTURE, CELL-MEDIATEDLYMPHOCYTOTOXICITY,ANDLYMPHOKINEPRODUCTION

JAMESN. WOODY,AFrAB AimmD, RICaARD C. KNUDSEN,DOUGLASM. STRONG,and KENNETHW. SELL

From the Department of Clinical and Experimental Immunology, NavalMedical Research Institute, Bethesda, Maryland 20014

A B S T R A C T Human peripheral blood lymphocytes(PBL) were evaluated by their responses to phyto-hemagglutinin (PHA-P), concanavallin A (con-A),and pokeweed mitogen (PWM), both before and aftertreatment with an antiserum against human thymiclymphocyte antigens (HTLA) that had been madeT-cell-specific by multiple absorptions with immuno-globulin EAC-positive lymphoblast cell lines (B cells).Cells treated with HTLA were examined for theirability to react in a mixed lymphocyte culture (MLC)and to form killer cells in a cell-mediated lymphocyto-toxicity (CML) system. Sensitized cells were also ex-amined for their ability to respond to purified proteinderivative (PPD) by blastogenesis, migration inhibi-tory factor release (MIF), and lymphotoxin (LT)production, both before and after treatment with HTLAand complement. The HTLA was in itself highly stim-ulatory to PBL. However, with the addition of comple-ment and subsequent cell destruction, a marked decreasein its stimulatory response was noted. PBL treatedwith HTLA and complement exhibited marked inhibi-tion of responsiveness to con-A with little decrease inPHA-P or PWMstimulation except at very high con-centrations of HTLA. MLC reaction was inhibitedonly when responder cells were treated with HTLA+C'. Treatment of stimulator cells with HTLA+ C'did not significantly alter the MLC response. The

Dr. Woody's present address is Department of Zoology,University College, London, U. K

Received for publication 4 October 1974 and in revisedform 14 January 1975.

HTLA+ C'-treated cells failed to form killer cells inthe CML reaction and inhibited PPD-induced blasto-genesis from PPD-sensitized individuals; however,treatment of sensitized cells with HTLA+ C' had littleeffect on the release of MIF and LT. It is suggestedthat subpopulations of T-cells carry surface antigensthat bind with this specific antisera, and that the con-A-responsive cells, the responder cells in the MLC, andkiller T-cells comprise a separate subset from cells re-sponding to PHA-P or PWM, or the MIF- and LT-producing cells.

INTRODUCTION

Murine thymus-derived lymphocytes (T-cells)1 havebeen identified with specific anti-O serum (1). Antiseradirected against other T-cell surface antigens, includingthymic leukemia antigen and lymphocyte-defined antigen,have further identified various subpopulations amongT-cells (2, 3). The use of these antisera has segregated

'Abbreviations used in this paper: B cells, bone marrow-derived lymphocytes; C', complement; CML, cell-mediatedlymphocytotoxicity; con-A, concanavallin-A; EAC, sheeperythrocyte-antibody-complement complex; E rosette, lym-phocyte forming spontaneous rosettes with sheep red bloodcells; FCS, fetal calf serum; HTLA, human thymic lym-phocyte antiserum; LT, lymphotoxin; MIF, migration in-hibitory factor; MLC, mixed leukocyte culture; NIRS, nor-mal rabbit serum; PBL, peripheral blood lymphocyte;PHA-P, phytohemagglutinin-purified; PPD, purified pro-tein derivative; PWM, pokeweed mitogen; SRBC, sheepred blood cell; T-cell, thymus-derived lymphocyte; VBS,veronal-buffered saline with calcium and magnesium.

The Journal of Clinical, Investigation Volume 55 May 1975 956-966956

the responses of lymphocyte populations to mitogens(4)' and to various T-cell-dependent antigens, such assheep red blood cells (SRBC) (5).

Such assays in the human systems have been hin-dered by the lack of specific anti-T-cell antisera, as allpreparations are prepared in heterologous species, withthe majority of antibodies nonspecific anti-species anti-bodies and only a small fraction of anti T-cell-specificantibodies present.

Only recently have various authors described thepreparation of specific anti-human T-cell antisera (6-10) and examined some of the effects of these antiseraon human cells in terms of inhibition of spontaneousSRBC rosette formation, a T-cell marker (11), andeffects on mitogen stimulation and mixed leukocyteculture (MLC) responses (12-14). Our studies con-firm what has been done previously and further outlinefunctional T-cell subclass heterogeneity, attempting togain a clear understanding of these subclasses and theirrole in disease processes.

METHODS

Preparation of human thymic lymphocyte antiserum(HTLA). Rabbit anti-human thymocyte serum was pre-pared by injecting New Zealand white rabbits with 1 X 10'viable fetal thymocytes on day 1, with complete Freund'sadjuvant and footpad injections. The animals were boostedon day 14 with 1 X 10' viable cells given intravenously. Onday 21 the animals were bled by cardiac puncture and theserum was collected and stored at -20'C. Rabbit anti-human thymocyte serum with a cytotoxicity titer of 1: 1,600or greater, prepared in this manner, has been shown to beconsistently immunosuppressive, as tested in the rhesusmonkey skin allograft survival assay (15).

A small pool (400 ml) was selected from three rabbitsproducing highly cytotoxic antisera. The antisera (100 ml)were pooled, heated to 56'C for 30 min and absorbed threetimes with 30 ml of packed human AB red cells carefullywashed free of contaminating leukocytes by repeated re-moval of the buffy coat. The antisera were then dividedinto 100-ml aliquots and absorbed 12 times with 10 mlof packed washed human cultured lymphoblasts from threelines known to have surface immunoglobulin and comple-ment receptors. An attempt was made to use cells con-taining most of the common human lymphocyte antigentypes, since the antisera were prepared from thymocytesfrom more than one donor. The absorptions were done for1 h at 370C on a circular rotator with every third absorp-tion performed overnight at + 4VC. After each set of ab-sorptions, the antisera were ultracentrifuged at 18,000 gfor 1 h and filtered through a 0.22-nm Millipore filter(Millipore Corp., Bedford, Mass.). The antisera were ab-sorbed until no activity was found against the absorbinglymphoblast cell lines either by cytotoxicity or indirect im-munofluorescence. It was found that an additional threeabsorptions with gluteraldehyde-treated human serum wererequired to remove antibodies directed primarily againsthuman albumin and IgG. The antisera were stored at- 70'C for use in the following assays. When other cell

'Kieselow, P. 1974. Personal communication.

types were used for absorptions, the technique was identicalto that outlined above.

Cytotoxicity testing. Peripheral blood lymphocytes wereprepared by the method of B6yum (16) and cytotoxicitytesting was performed by the method of Terasaki and Mc-Celland (17).

Treatment of peripheral blood lymphocytes (PBL) withHTLA. Ficoll-Hypaque-separated PBL were aliquotedinto sterile 10 X 75-mm (Falcon Plastics, Oxnard, Calif.)test tubes and treated with various dilutions of HTLA at37'C for 30 min. Rabbit complement (C'), absorbed withlymphoblasts to remove natural antibodies, was added toeach tube and incubated further for 1 h at 37'C. These cellsuspensions were then relayered onto fresh FicollHypaquegradients and the viable cells recovered and washed threetimes with RPMI-1640 (Grand Island Biological Co., GrandIsland, N. Y.). Cells were counted with an AutocytometerII (Fisher Scientific, Silver Spring, Md.) and the via-bility was determined by trypan blue dye exclusion. Via-bility was always t 90%. For most assays 5 X 10' cellswere pelleted and treated with 0.9 ml of the HTLA dilu-tion and 0.1 ml of rabbit complement (diluted 1:4); thistreatment resulted in 50% cytotoxicity. In experimentswhere HTLA was used as a mitogen, PBL were treatedin vitro with various dilutions of HTLA alone, incubatedat 37'C for 30 min, washed three times with RPMI-1640,recounted, and dispensed into microtiter plates at 2 X 106/culture.

Rosette inhibition studies. The erythrocyte (E) rosettetest for spontaneous SRBC T-cell binding was performedwith fresh (<2 wk) SRBC in Alsever's solution (Micro-biological Associates, Bethesda, Md.). The cells were washedthree times in veronal-buffered saline (Ca++ and Mg++)(VBS). A 0.5% solution of SRBCwas prepared by adding10 ml of buffer consisting of equal portions of VBS andfetal calf serum (FCS) to 50 1l of packed SRBC. TheFCS (10 ml) was absorbed three times with 2 ml packedwashed SRBC to remove any antibodies to SRBC, if pres-ent. 100 Al of 0.5% SRBCsolution was added to 100 Al ofPBL containing 2-4 X 10' cells. The mixture was incubatedat 37'C for 30 min, centrifuged at 200 g for 5 min, andleft at + 4VC for a minimum of 2 h or overnight Thecells were gently resuspended and a small sample placed:on a glass slide with a cover slip. 200 cells were countedunder phase and the percent of rosette-forming cells counted.Cells with three or more adherent SRBCwere counted asrosettes.

The EAC rosette test was carried out by sensitizing 2ml of 2% SRBCwith an equal volume of 1: 64,000 dilutionof human 19S anti-Forssmann (compliments of Dr. R.Wistar, Naval Medical Research Institute), incubating themixture at 37'C for 30 min, and adding fresh mouse serumat a 1: 5 dilution in VBS for 30 min at 37'C as a com-plement source. Reagent made in this manner and storedat + 4C could be used for 7-14 days, after which timethe coated SRBCfailed to. exhibit fluorescence with -a fluo-rescent-conjugated anti-C'3 reagent (Meloy Laboratories,Inc., Springfield, Va.). 100, Al of cells and 100 Adl of EACreagent wereadded and incubated on a rotating platformfor 45 min at 37'C; the cells were resuspended and countedas previously described.

Inhibition studies on both types of rosettes were done byadding 100 1. of the HTLA to the lymphocytes, incubatingthem 1 h at 37'C, and then proceeding with the test asdescribed.

Functional Specificity of a Rabbit Anti-Human T Cell Serum 957

TABLE I

50% Cytotoxicity Titers of HTLA againstVarious Cell Types

Cell type Titer

Thymus cells 1:1,024PBL 1:64MOLT-4* 1:64CLLJ 1:6Lymphoblast§ <1:2

* MOLT-4, a continuous lymphoblast T-cell line.t Cells from a patient with chronic lymphocytic leukemia.§ A continuous lymphocytoblast cell known to be a B-cellline.

Mitogen responses. Responses of lymphocytes to the non-specific mitogens phytohemagglutinin-P (PHA-P) (DifcoLabs,: Detroit, Mich.), concanavallin A (con-A) (Calbio-chem, San Diego, Calif.), and pokeweed mitogen (PWM)(Grand Island Biological Co.) were always performed intriplicate as previously described (18). In brief, 100 ,Al ofPBL containing 2 X 106 cell/ml in RPMI-1640, and 100 Upenicillin/ml, 100 jug streptomycin/ml, and 2 mML-gluta-mine containing 10%o heat-inactivated (560C, 30 min) FCSwere plated into a microtiter plate and 50 A.l of the ap-propriate mitogen was added. The cultures were incubatedfor 48 h at 370C, 5%o CO2 and 95% air. The cultures werethen pulsed with 20 ,.l of media containing 1 ,uCi ['H]-thymidine (sp act 1.9 Ci/mmol, Schwarz-Mann Div., Bec-ton, Dickinson & Co., Orangeberg, N. Y.) per well andharvested 18 h later with a multiple automated sample har-vester. The triplicate samples were counted for 1 min eachin a Packard Tricarb liquid scintillation counter (PackardInstrument Co., Inc., Downer's Grove, Ill.) and the mean,standard error, and percent standard error of triplicatesamples calculated with a Wang 700C advance program-ming calculator (Wang Laboratories, Inc., Tewksbury,Mass.). The percentage of standard error for such experi-ments rarely exceeded 5%o.

MLC. MLC reactivities were measured in triplicate asdescribed in previous publications (19). In brief, stimulatorcells were treated with 25 ,ug/ml mitomycin C (Calbiochem)at 37'C for 30 min to inhibit mitosis. They were washedtwice in RPMI-1640, recounted, and used. 2 X 10' cells/wellwere used as stimulator cells and an equal number as re-sponder cells. Untreated cells and cells treated with eitherHTLA or HTLA+ C' were used as stimulator and re-sponder cells. Cells were incubated for 4 days at 37'C, 5%C02 and 95% humidified air, pulsed with 1 ACi of ['H]-thymidine per well for 18 h and harvested with the multipleautomated sample harvester.

Assay for purified protein derivation (PPD) -induced blas-togenesis. The assay for PPD-induced blastogenesis utilizedPPD (Parke, Davis & Co., Detroit, Mich.) at various con-centrations, usually 5-20 ,ug/ml. The assay technique wasidentical to that of the mitogens, except PPDwas added andthe culture incubated for 5 days rather than 48 h. Whenmigration inhibitory factor (MIF) or lymphotoxin (LT)was assayed, a 48-h incubation was used.

Assay for MIF. Our assay system for MIF is a modi-fication (20) of the David and David technique (21). Inbrief, PBL were collected from patients with known sensi-tivity to tuberculin PPD. 6 X 108 cells were incubated for

48-72 h in a 75 X 100-mm tube with various concentrationsof PPD, before centrifugation and collection of the super-natant fluid. 50 ,ul of peritoneal exudative cells (50 X 106/ml), obtained by oil induction from guinea pigs, were loadedinto plain 75-mm capillary tubes sealed at one end. Thecapillary tubes were centrifuged at 150 g for 10 min andthe capillary cut at the cell-fluid interface. The short pieceof capillary tubing containing the exudative cells was affixedto the inside of a migration inhibition chamber (Mini LabCo., Ville De Laval, P. Q.) with silicone grease, Eachchamber contained two capillary tubes and was sealed witha coverslip.

Fluid (0.5 ml) from the lymphocyte cultures were in-jected into the sealed chambers with a tuberculin syringeand the inj ection ports were sealed with plastic tape toprevent evaporation. The chambers were incubated at 37°Cfor 24-48 h. The image of cover slip, capillary, and cellmigration was proj ected with a Bausch and Lomb pro-jecting prism (Bausch & Lomb Inc., Rochester, N. Y.)onto a sheet of paper and the outline of the migrated cellstraced. The area of migration was determined by planimetryand the percent migration inhibition was determined bythe following formula.

%migration inhibition1 average area of migration with antigen X 100

average area of migration without antigen

LT assay. Sensitized lymphoid cells in the presence ofthe sensitizing antigen, or normal cells when cultured withnonspecific mitogenic agents, cause lymphocyte transforma-tion. The cell-free medium of such activated cultures hasbeen shown to possess a lymphokine termed lymphotoxin(LT) (22). LT activity has been demonstrated by theability of cell-free supernatant fluids to cause the destruc-tion of mouse L-929 fibroblasts, and the secretion of LT hasbeen shown to be a function of T cells (23). The LTassay was performed in a microassay system (24). In thisassay system, the LT activity was measured as a decreasein the uptake of ['H]thymidine by L cells. Briefly, super-natant fluids from sensitized cells stimulated by PPD wereadded to L cells, and the ID5o levels of LT activity werecalculated as the dilution that gave 50% of the counts perminute of the media control values.

Cell-mediated lymphocytotoxicity (CML) assay. Re-sponder cells used in this assay were sensitized in vitro byincubation with mitomycin-C-treated allogeneic lymphocytes,as described in the MLC section. The sensitized cells werethen incubated with the PHA-P-stimulated, 51Cr-labeledallogeneic target cell type originally used as stimulator cells.The cytotoxic activity of the killer cells was calculated bythe percentage of 'Cr release by the following formula:

%0"Cr releasedmean cpm of experimental-mean cpm of spontaneous releasemean cpm of freeze-thaw control- mean cpm of spontaneous release

RESULTS

Specificity of antisera for T-cells. Initial experi-ments were designed to determine the specificity of ourHTLA. We approached the problem by testing theantiserum for cytotoxic activity with complement

958 J. N. Woody, A. Ahmed, R. C. Knudsen, D. M. Strong, and K. W. Sell

against various cell types. The results of this set ofexperiments is shown in Table I. The 1:1,024 titerobserved against human fetal thymocytes was expected,since the sera were selected for their high cytotoxicactivity, and fetal thymocytes were used as the immu-nizing antigen. The titer of activity against PBL,which were 70% T-cells by E-rosette criteria, andMOLT-4 cells, a cultured lymphoblast line with T-cellcharacteristics (25), is somewhat less than that ob-served against thymocytes, but suggests that both cellsare sensitive to the antisera. Cells from a patient withchronic lymphocytic leukemia, by clinical criteria, werealso used in the assay; 85% of these cell exhibited com-plement receptors and were considered to be B-cells.The HTLA exhibited very low levels of activity againstnot only these cells but also against the surface im-munoglobulin-positive B-cell lymphoblast line used toabsorb the antisera.

To identify further the antibodies in HTLA responsi-ble for its cytotoxic activity, a number of absorptionexperiments were performed, and the absorbed seratested for cytotoxic activity against PBL in the pres-ence of complement. The PBL in this case were 70%T-cells by E rosette criteria. The results of these ex-periments are seen in Fig. 1. Absorptions with humanthymocytes removed more of the cytotoxic activitytoward PBL than any other cell type, while absorp-tions with B-cell lymphoblasts removed little or noneof the activity. Absorptions with MOLT-4 cells, thecultured T-cell line, and with cells from a patient withSezary syndrome, a T-cell leukemia, with 90% of thecells forming spontaneous E rosettes, both removedsignificant cytotoxic activity against PBL, but less thanobserved when thymocytes were utilized. The shamabsorption merely diluted the HTLA to the same degreeas the actual absorption, acting as a control to assure

64'

32

o 1:16'!

C) IJQ

1.4

1(2

SHAMLYMPHOBLAST(B)MOLT (T)

StZARY CELL (Tr

THYMUS(Tr

2 No of

FIGURE 1 Cytotoxic properties of HTLA after absorptions.The effect of 20 volumes % absorptions with various celltypes on 50% cytotoxic titer of HTLA against PBL.

lOC

8(7C6C

L' 5C- 4CW3C0 2C

Icle CC

3020100

))

)

))

)

NRS

HTLA

0 2 4 8 16 32 64 128 256

) Z5--- HTLANRS

I

C 0 2 4 8 16 32 64 128 256RECIPRICOL OF DILUTION

FIGuRE 2 HTLA pretreatment and rosette formation. Theeffect of HTLA pretreatment on spontaneous SRBC rosetteformation (upper) and CRL rosette formation (lower)with PBL.

ourselves that the effect observed was not due to dilu-tion alone.

Additional specificity studies utilized the ability ofHTLA to inhibit the formation of either spontaneousE rosettes, a T-cell phenomena, or EAC rosettes, aB-cell phenomena. When PBL were incubated withHTLA, before the addition of SRBCor EAC reagent,it significantly inhibited the formation of spontaneousE rosettes to a dilution of 1: 128-1: 256, whereas littleinhibition of the EAC rosettes was seen except whenthe antisera were used undiluted (Fig. 2). Normal rab-bit serum (NRS) used as a control had little effect onspontaneous rosettes but did inhibit EAC rosettes toa 1: 2 dilution.

Stimulation of PBL by HTLA. From our previousexperiments we concluded that HTLA exhibited T-cellspecificity. To test its functional capacity, we examinedthe ability of this absorbed antisera to act as a specificT-cell mitogen. In these experiments the PBL weretreated with various dilutions of HTLA or HTLAand rabbit complement, as described in the Methodssection. The cells were incubated for 48 h and pulsedwith ['H]thymidine and harvested as described above.The results are seen in Fig. 3. It was found that HTLAalone was a potent mitogenic agent, with as little as 50tzg of HTLA giving 98,000 cpm while PHA-P, a potentT-cell mitogen, gave a peak response of 90,000 cpmwhen similar numbers of PBL's were used in the cul-ture. Cells treated with HTLA and complement did notexhibit the response at any concentration, implyingthat the cells responsible for the activity had beeneliminated with the addition of complement to thesystem.

Functional Specificity of a Rabbit Anti-Human T Cell Serum

II C5

959

180

160

140

120

0 100x

E{, 80

60

40

20

HTLA

C- NRS

---o'll' "on HTLA + C

C 3.1 6.2 12.5 25 50 100 200 mg HTLA

FIGuRE 3 Treatment of PBL with HTLA. Uptake of[8H]thymidine by PBL after treatment with various con-centrations of HTLA alone or with C'.

Alteration of mitogen responses by HTLA and com-plement. After the demonstration that HTLA andcomplement could eliminate the blastogenic response in-duced by HTLA alone, we undertook studies to deter-mine the effect of HTLA and complement on the abilityof cells to respond to various nonspecific mitogens.EBL were treated with HTLA and complement andsubsequently incubated with the nonspecific mitogens.As seen in Fig. 4, the response of the cells to PHA-Pwas not significantly decreased, except at higher con-centrations of antisera, while the response of treatedcells to PWMwas essentially unaltered. The abilityof treated cells to respond to con-A was inhibited even

401

20-

C 1.5

C NRS+C

IiLLLI I IrI3.1 6.2 12.5 25 50 100 200

PHA-P

at very low concentrations of antisera, 6.2 A&g of HTLAcorresponding to a 1: 32 dilution of the antisera.

Inhibition -of MLC response by HTLA and comple-ment. With the previous data suggesting that HTLAand complement would inhibit the PBL response tocon-A at certain dilutions while having little effect onPHA-P responses, we concluded that the PBL couldbe separated into functional subsets using the antiseraat these concentrations. Wewere interested in lookingat other functional parameters thought to be mediatedby T-cells and, therefore, began a series of experimentsto examine these. One such reaction is the one-wayMLC, where cells from one donor (responder cells)are mixed with cells (stimulator cells) from an allo-geneic donor that have been blocked with mitomycin Cso that they are unable to respond by DNA synthesis,but are capable of stimulating the responder cells. Theresults of a dose-response assay treating PBL respondercells with various concentrations of HTLA and com-plement are seen in Fig. 5. Treatment of the respondercells results in a marked decrease in the ability of thesecells to react to the stimulator cells, as measured bythe uptake of [3H]thymidine. This inhibition occurseven at low concentrations of antibody, similar to thelevels noted to inhibit the con-A responses.

The second part of this experiment deals with thetreatment of either stimulator or responder cells withHTLA and complement and the ability of these treatedcells to function in the MLC reaction. The cells weretreated with HTLA at a 1: 10 dilution and complement,which corresponds to a level of HTLA between 12.5and 25 ig; the results are shown in Table II. Controlcells did not show any significant responses as mea-

C 1.5 3.1 6.21252 501002

PWMC 1.5 31 6.2 25 50 100 200g HTLA +C

CON-A

FIGURE 4 HTLA pretreatment and mitogen response. Uptake of ['H]thymidine by PBLafter treatment with various amounts of HTLA+ C' followed by mitogeri stimulation.

960 1. N. Woody, A. Ahmed, R. C. Knudsen, D. M. Strong, and K. W. Sell

130 r

120F

.100

801W)0

'cE0.0 60

sured by the uptake of [5H]thymidine. When untreatedresponder cells are mixed with allogeneic mitomycin-C-treated stimulator cells, such as with Ax Bm or Bx Am,a normal MLC response is observed. When untreatedresponder cells are mixed with allogeneic cells treatedwith HTLA, complement, and mitomycin-C, a moderatedecrease in counts is seen as with Ax Bhm and an in-creased response is noted with Bx Ahm, suggesting thatthe cells inducing stimulation are only slightly affectedby treatment with HTLA and complement. When re-sponder cells are treated with HTLA and complementand mixed with mitomycin-C-treated stimulator cells,such as Ah X Bm and Bh X Am, the MLC response issignificantly decreased, indicating that the respondercells are significantly affected by treatment with HTLAand complement, as was also noted in the experimentsoutlined in Fig. 4.

Inhibition of formation of CMLkiller cells by HTLAand complement. In the previous experiments we wereable to show that treated PBL responder cells wereunable to participate in the MLC reaction, whereastreated stimulator cells worked nearly as well as nor-mal cells. Experiments to evaluate the ability of HTLAand complement treatment to inhibit the generation ofkiller cells and previously activated killer cells in theCML reaction were carried out.

In this series of assays, PBL were used either un-treated or treated with HTLA 1: 10 and complement,as previously described. Both treated and untreated re-sponder cells were sensitized in vitro as in the MLCstudies.

After sensitization, these cells were either directlyused in the CMLassay or treated with HTLA+ C' orNRS+ C' and used in the assay. These effector cellswere then incubated with 'Cr-labeled target cells (stim-ulator cells), stimulated with PHA-P, and the percent

30

°20EC-0 If)

C= NRS+C

_--I

None C 200 100 50 25 12.5 6.25 3.12

,sg HTLA +0 I1.5

FIGuRE 5 HTLA pretreatment of responder cells in MLC.Uptake of ['H] thymidine by responder cells in MLC re-action after pretreatment with various amounts of HTLA+ C'.

of chromium release was calculated as an index oftarget cell lysis by the sensitized cells.

Treatment of responder cells before initiation ofMLC (sensitization phase) with HTLA+ C' led tono significant activation (Table II) and subsequent gen-eration of killer cells (data not shown). As seen inTable III, untreated in vitro-sensitized responder cellsexhibited marked cytotoxic potential, causing 18-49%'ICr release. Treatment of these same sensitized cellswith HTLA and complement nearly eliminated thecytotoxic potential. Treatment of these sensitized cellswith NRS and complement had little effect.

In the set of experiments outlined in Table IV, thesame protocol was followed except mitomycin-C-treatedstimulator cells were incubated with either HTLA andcomplement or NRS and complement before incubationin the MLC reaction. Treatment of the stimulator cellsdid not prevent the formation of cytotoxic lymphocytes.

TABLE IIThe Effect of Pretreatment of PBL with HTLA 1:10 and Complement on the Ability of These

Cells to Act as Stimulator and Responder Cells in the MLCReaction

Untreated cells Treated stimulator cells Treated responder cells

type cPm %SE* type cm %SE* type cOm %SE

A 479 5.4 Ah 257 4.3B 798 1.7 Bh 399 2.0Amt 412 7.7 Ahm§ 185 7.7Bm 93 5.6 Bhm 134 5.8Ax Am 775 3.9 Ax Ahm 639 5.8 Ahx Am 797 6.5Bx Bm 1,295 6.0 Bx Bhm 1,693 4.0 Bhx Bm 1942 11.6Ax Bm 18,587 3.7 Ax Bh 9,076 5.5 Ahx Bm 943 1.8Bx Am 22,436 5.6 Bx Ah, 29,195 3.5 Bhx Am 2219 6.1

* %SE, percent standard error on triplicate samples.t m, mitomycin-C-treated.§ h, HTLA and complement-treated.


40 r

v__

TABLE I IIE8ffect of Treatment of Sensitized Cells with HTLA + C' on

CML-Mediated Chromium Release fromTarget Cells

Responder Stimulator Cr-5 1 No NRS HTLAcell type cell type target cell treatment + C' + C'

A X Bm§ B 49.9* 42.5 -1.8A X Cm C 33.0 21.6 2.3B X Am A 29.6 - 0.7B X Cm C 30.9 25.6 1.2C X Am A 28.5 22.1 -2.1C X Bm B 18.6 11.8 1.9A$ X Am A 2.7 -

BT X Bm B 1.9Ca X Cm C 3.4

* Percent chromium released from target cells.T Untreated controls.§ Mitomycin-C-treated cells.

Effect of HTLA and complement on blastogenesis,MIF, and LT production by sensitized lymphocytes.In the previous experiments we have shown that thecon-A-responsive cells and the MLC responder cellsare sensitive to treatment with HTLA and complement.To ascertain the effect of HTLA and complement treat-ment on the responses of sensitized lymphocytes, a num-ber of experiments were undertaken with cells frompatients sensitive to PPD by skin test.

In the first of these experiments, cells were treatedwith either HTLA alone or with HTLA and comple-ment before exposure to various concentrations of PPD.NRS with or without complement served as a control.The blastogenic response to PPD was measured by theuptake of ['H]thymidine. The results are seen in TableV. The HTLA 1: 5 alone stimulated the cells, as notedamong data for the control samples. PPD induced sig-nificant blastogenesis by itself, which was slightly in-hibited by NRS and complement, but not by HTLAalone. HTLA and complement at both dilutions testedsignificantly inhibited the response of the sensitizedcells to PPD.

To extend the experiments to evaluate MIF and LTproduction by sensitized lymphocytes stimulated withPPD, the following studies were performed. Cells from

TABLE IVEffect of Pretreatment of Stimulator Cells with HTLA + C' on

CML-Mediated Chromium Release fromTarget Cells

Responder Stimulator Cr-5 1 No NRS HTLAcell type cell type target cell treatment + C' + C'

A X Bm* B 49.9* 55.0 62.7A X Cm C 33.7 40.7 45.9B X Am A 29.6 21.5 38.5B X Cm C 30.9 22.7 38.9C X Am A 28.5 32.6 35.9C X Bm B 18.6 21.9 27.6

* Percent chromium released from target cells.t Mitomycin-C-treated cells.

PPD-sensitive donors were treated with various com-binations of sera and complement before exposure toPPD. The supernatant fluid was harvested after 48 hand assayed for MIF and LT production. The resultsare shown in Table VI. None of the sera used as con-trols induced the cells to produce significant amountsof MIF or LT, while cells stimulated with PPD aloneproduced significant amounts of both. Neither NRS,HTLA, or HTLA and complement inhibited the pro-duction of MIF or LT in response to PPD stimulationby sensitized cells.

DISCUSSION

The studies reported here provide an initial attempt toclassify human T-cells into subpopulations by func-tional assays. In the murine system, cells have beenclassified with specific antisera, such as 0, lymphocyte-defined antigens, and thymic leukemia antigen (1-3),both in terms of surface antigens and functional as-says (4).

When animals are immunized with a complex anti-gen, such as a human thymocyte, the heterolgous anti-sera contain not only specific antithymocyte antibodiesbut a great deal of nonspecific antibodies, includingantibodies directed against other lymphocyte antigens,human species antigens, immunoglobulins, and prob-ably other as yet unknown antigens. This is supported

TABLE VThe Effect of Pretreatment of PPD-Sensitive Cells with HTLA + C' on Blastogenic Response to Various

Amounts of PPDas Measured by the Uptake of [H3]Thymidine

HTLA 1:5 HTLA 1:20Control NRS NRS+ C' HTLA 1:5 + C' + C'

cPm

Control 953 878 683 11,596 221 612PPD, 20 sg/ml 38,359 27,404 9,999 34,849 624 2,112PPD, 10 yg/ml 11,797 18,769 12,354 22,709 359 3,889PPD, 5 Ag/ml 7,182 12,560 7,246 19,519 435 1,934


by the fact that many absorptions were required tolend cytotoxic specificity towards T-cells in the HTLAused in this study. The cytotoxic activity seen againstchronic lymphocytic leukemia cells suggests that thesecells carry antigens not removed from the HTLA bylymphoblast absorptions. The antisera thus preparedare only functionally specific.

In some aspects, HTLA can be compared to 0 anti-sera in the murine system, although it probably detectsa group of antigens peculiar to T-cells rather than aspecific antigenic site. It would, however, be expectedto exhibit more specificity toward cells with thymocyte-surface antigens, as it was prepared against thymocytes.It would also be expected to differentiate types ofT-cells expressing more or less thymus-surface antigens.Stobo and Paul (26) and Granger and Kolb (22) havenoted that certain T-cells in the murine system con-tained larger amounts of surface 0 antigens, as detectedwith anti-@ serum. The results obtained by absorbingthe HTLA with various cell types tend to support asimilar conclusion in human cells. Of particular interestare the MOLT-4 and Sdzary cells, considered to beT-cells by their ability to form E rosettes and theirlack of surface immunoglobulin. This is exemplified bytheir inability to remove the cytotoxic activity of theHTLA against PBL as compared to thymocytes. It ispossible, however, that HTLA discriminates betweendifferent functions of T cells rather than physicallybeing cytotoxic for subpopulations of T cells. The ex-pression of thymus surface antigens may be of clinicalsignificance, since it has been shown by Smith (27)that some types of malignant cells fail to form Erosettes; however, they do react with antisera specificfor thymic lymphocyte antigens, suggesting that certainT-cells may modulate or lose their SRBCreceptor sites,making clinical recognition difficult when one uses theE rosette as the sole criterion for T-cell identification.

The blastogenic response of cells treated with HTLAalone, noted in our experiments, indicates that T-cellsare intensely stimulated by this antisera, more thanseen with either PHA-P or con-A. This could prove tobe a functional test of T-cells, somewhat more specificthan mitogen stimulation, as well as an assay forHTLA, since this response is not seen with antiseraraised against B lymphoblasts,' and increases as non-specific antibodies are removed from antithymocyte seraby absorption with B lymphoblasts.

The response of cells treated with HTLA and com-plement to mitogens suggests that the con-A-responsivecells are most sensitive to this antiserum. Only athigher concentrations are the PHA-P-sensitive cellsaffected, while the PWM-responsive cells are not sig-

3Woody, J. N., A. Ahmed, and K. W. Sell. Unpublishedobservations.

TABLE VIThe Response of Sensitized Cells to Pretreatment with HTLA+C',

Followed by Stimulation with PPD, as Measutred by theProduction of MIF and L T

Migration L-cellTreatment of sensitized cells inhibition inhibition

Cell control 0 6.9Cells + NRS 7 10Cells + C' 3 24Cells + HTLA -9 10Cells + HTLA + C' -5 6Cells + NRS+ PPD 68 95Cells + PPD 72 96Cells + HTLA + PPD 65 95Cells + HTLA + C' + PPD 75 95

nificantly decreased. Since the original antisera weremade against thymocytes, one can speculate that per-haps the same relationship between 0 antigen andcon-A-responsive cells in the murine system applies tothe human system. Granger and Kolb (22) noted thatT-cells containing larger amounts of surface 0 antigenwere more sensitive to anti-O serum and more responsiveto con-A. Cells containing less 0 surface antigen werenoted to be more responsive to PHA-P. It is entirelypossible that the reason that the PHA response re-quired higher amounts of anti-HTLA than the con-Aresponse to be inhibited may be a matter of quantityrather than quality. One would only have to postulatethat the PHA conditions were supraoptimal (i.e., re-duction of high proportion of reactive cells is not seenuntil 95% of the cells are eliminated) versus a sub-optimal condition with regard to con-A. However, thiswould seem unlikely since concentration curves of bothmitogen and cell concentrations on HTLA+ comple-ment-treated PBL's failed to reveal any con-A response.

When the responder cells in the MLCreaction weretreated with HTLA and complement, a decrease inactivity was observed, suggesting that the MLC re-sponder cell was sensitive to approximately the sameconcentrations of HTLA as was the con-A-responsivecell. When the stimulator cells were treated with HTLA+ C', only slight differences in MLC responses werenoted, indicating that large numbers of T-cells are notnecessary for stimulation of allogeneic cells. When re-sponder cells were treated with HTLA+ C', a markeddecrease in the MLC response was noted, indicatingthat T-cells are necessary for adequate responses.

A similar phenomenon was noted in the generationof killer cells in the CML reaction, where HTLA+ C'prevented the development of these cells. The sensitivityof the cells to HTLA would suggest that either the


TABLE VI IClassification of Human T-Cells on Basis of Functional Assays

and Sensitivity to HTLA

HTLA-sensitive HTLA-insensitive

Con-A PHA-PPPD, blastogenesis PWMResponder cell, MLC MIFKiller cell, CML LT

same cell or a similar cell participating in the MLCis being affected.

Bach and Bach (28) have suggested that the re-sponder cell in the MLC recognizes different antigensfrom those recognized by the killer cell in the CMLreaction. Our data suggest that either both cells aresensitive to HTLA, by detecting a group of thymicantigens on the T-cell surface that might be expressedby both types of cells, or that the responder cell in theMLC transmits information through mediators neces-sary for the development of the killer cell, hence, with-out the MLCresponder as the afferent limb, the effectorcell cannot be generated.

Of interest were the results shown in Table V andVI, where HTLA and complement inhibited the blasto-genic response of sensitive cells exposed to PPD,whereas the release of MIF and LT from the same cellswas not affected, suggesting that at least two differentcell populations were involved, one undergoing blasto-genesis and the other releasing mediators. It has beensuggested that MIF is preformed and may be releasedfrom cells, as cultured B lymphoblasts are known tomake MIF (29). Whether LT is preformed or is syn-thesized is unknown; however, evidently the cells neednot undergo blastogenesis for this mediator to be re-leased. Whether LT production is a T- or B-cell func-tion in humans is unknown; however, LT productionby PHA or con-A stimulated cells suggests a T-cellfunction.

From our data, summarized in Table VII, we con-clude that the cells responding to con-A, the MLCresponder cells, and the CML killer cells, as well ascells responding to PPD in the sensitized individual,express equivalent amounts of thymic lymphocyte anti-gens as detected by our HTLA. On the other hand,it is unclear at present what types of cells respond toPHP-P or PWMand what types of cells produce MIFand LT. Whether these T-cells exhibit little thymicsurface antigens or are B-cells or some intermediatetype of cell is at present uncertain.

Recently, a goat antihuman thymocyte globulin(GAHTG) has been described by Owen and Fanger(30). There are some similarities in the results ob-tained with this absorbed GAHTGand the HTLA re-

ported in this communication: they both inhibit theresponse of PBL to con-A and allogeneic cells; andthey do not affect the response of PWM. However,there are several differences in the experimental pro-cedures and the results obtained: their experiments areperformed with continuous incubation of cells withGAHTG, whereas in our case the cells were treatedwith HTLA and complement, the dead cells removedby gradient separation, and the cells washed beforeassay; and their antibody is mitogenic to PBL beforeabsorption with chronic lymphocytic leukemia cells, Ibutthis mitogenic property is removed after absorption. Inthe case of HTLA, however, the mitogenic propertywas increased after absorption with lymphoblasts. Theseresults suggest that (a) the mitogenic principal inGAHTGeither resides on B-cells alone or is an anti-body against an antigen common to both T and B cellsand, therefore, it is not really clear whether it stimu-lates T cells or B cells, whereas the mitogenic principalin HTLA could not be absorbed out by absorption withlymphoblasts or with cells from patients with chroniclymphocytic leukemia and, therefore, present on Tcells and also stimulate T cells; (b) the goat mayrecognize certain antigens in human thymus cellsdifferent from those recognized by rabbits; and (c)both goats and rabbits recognize certain similar anti-gens, giving rise to antisera that block the con-Aresponse and the MLC response.

The identification of specific antibodies to subsets ofT-cells may have a great clinical relevance in light ofthe description of a naturally-occurring thymocytotoxicantibody found in the sera of old New Zealand blackmice (31), thought to play a role in the pathogenesisof the autoimmune disease process (32) related to theloss of a thymic suppressor (33). Such antibodies occurin patients with systemic lupus erythematosus and maybe found in other autoimmune diseases. As more islearned about cellular immune dysfunctions in variousdisease states and the pathophysiologic role of variouscell types in immune disorders, the hypothesis givenregarding T-cell subpopulations may tend to groupcertain disorders. One such example is in the case ofsubacute sclerosing panencephalitis, where cells wereshown to liberate LT and MIF on antigenic stimulation,but did not transform (20) to a viral preparation. Thecell types involved and the area in the cellular immunesystem where these factors operate will be most usefulin the further delineation of T-cell subsets.

ACKNOWLEDGMENTS

Wewould like to thank Dr. Stanley Green for the culturedlymphoblasts used in absorbing our antisera, and to AlBlack, Ken Stewart, and Joe Keenan for their technicalassistance. Wealso would like to thank Dr. Richard Wistarfor his technical advice.


This research was supported by the Bureau of Medicineand Surgery Work Unit Nos. MF51.524.013.1002AA2C andM4318.01.0004AGGX. The opinions or assertions containedherein are the private ones of the authors and are not to beconstrued as official. or reflecting the views of the NavyDepartment or the naval service at large. The animals usedin this study were handled in accordance with the pro-visions of Public Law 89-54 as amended by Public Law 91-579, the Animal Welfare Act of 1970 and the principlesoutlined in the "Guide for the Care and Use of LaboratoryAnimals," U. S. Department of Health, Education, andWelfare Publication No. (NIH) 73-23.

REFERENCES

1. Reif, A. E., and J. Allen. 1964. The AKR thymicantigen and its distribution in leukemias and nervoustissues. J. Exp. Med. 120: 413-433.

2. Leckband, E., and E. A. Boyse. 1971. Immunocompetentcells among mouse thymocytes: a minor population.Science (Wash. D. C.). 172: 1258-1260.

3. Boyse, E. A., M. Miyazawa, T. Aoki, and L. J. Old.1968. Ly-A and Ly-B: two systems of lymphocyte iso-antigens in the mouse. Proc. R. Soc. Lond. B. Biol. Sci.170: 175-193.

4. Scher, I., D. M. Strong, A. Ahmed, R C. Knudsen,and K. W. Sell. 1973. Specific murine B-cell activationby synthetic single- and double-stranded polynucleo-tides. J. Exp. Med. 138: 1545-1563.

5. Mosier, D. E., and C. W. Pierce. 1972. Functional ma-turation of thymic lymphocyte population in vitro. J.Exp. Med. 136: 1484-1500.

6. Yata, J., G. Klein, N. Kogayashi, T. Furukawa, andM. Yanagisawa. 1970. Human thymus-lymphoid tissueantigen and its presence in leukemia and lymphoma.Clin. Exp. Immunol. 7: 781-792.

7. Smith, R. W., W. D. Terry, D. N. Buell, and K W.Sell. 1973. An antigenic marker for human thymiclymphocytes. J. Immunol. 110: 8844887.

8. Williams, R. C., Jr., J. R. DeBoard, O. J. Mellbye,R. P. Messner, and F. D. Lindstrom. 1973. Studies ofT- and B-lymphocytes in patients with connective tissuediseases. J. Clin. Invest. 52: 283-295.

9. Bobrove, A. M., S. Strober, L. A. Herzenberg, andJ. D. DePamphilis. 1974. Identification and quantitationof thymus-derived lymphocytes in human peripheralblood. J. Immunol. 112: 520-527.

10. Brown, G., and M. F. Greaves. 1975. J. Immunol. Inpress.

11. Wortis, H. H., A. G. Cooper, and M. C. Brown. 1973.Inhibition of human lymphocyte rosetting by anti-T sera.Nat. NewBiol. 243: 109-1,11.

12. Aiuti, F., and H. Wigzell. 1973. Function and distribu-tion pattern of human T lymphocytes. I. Production ofanti-lymphocyte specific sera as estimated by cyto-toxicity and elimination of function of lymphocytes.Clin. Exp. Immunol. 13: 171-181.

13. Aiuti, F., and H. WigzelL 1973. Function and distribu-tion pattern of human T lymphocytes. II. Presence of Tlymphocytes in normal humans and in humans withvarious immunodeficiency disorders. Clin. Exp. Immunol.13: 183-189.

14. Burger, D. R., B. J. Wilson, A. Malley, and R. M.Vetto. 1974. The effect of antilymphocyte antibody onlymphocyte transformation: Selective suppression of

mitogen, MLC, and antigen stimulation of human lym-phocytes. Transplantation (Baltimore). 17: 541-550.

15. Sell, K. W., J. N. Woody, J. Smith, C. Darrow, andD. Kayhoe. 1973. Evaluation of human cultured lympho-blasts as a source of antigen for production of immuno-suppressive antilymphocyte serum. Transplant. Proc. 5:541-547.

16. B6yum, A. 1968. Isolation of mononuclear cells andgranulocytes from human blood. Isolation of mono-nuclear cells by one centrifugation, and of granulocytesby combining centrifugation and sedimentation at 1.Scand. J. Clin. Lab. Invest. 21 (Suppl. 97): 9-109.

17. Terasaki, P. I., and J. D. McCelland. 1964. Micro-droplet assay of human serum cytotoxin. Nature (Lond.).204: 998-1000.

18. Strong, D. M., A. A. Ahmed, G. B. Thurman, andK W. Sell. 1973. In vitro stimulation of murine spleencells using a microculture system and multiple auto-mated sample harvester. J. Immunol. Methods. 2: 279-291.

19. Thurman, G. B., D. M. Strong, A. Ahmed, S. S. Green,K W. Sell, R. J. Hartzman, and F. H. Bach. 1973.Human mixed lymphocyte cultures. Evaluation of amicroculture technique utilizing the multiple automatedsample harvester (MASH). Clin. Exp. Immunol. 15:289-302.

20. Ahmed, A., D. M. Strong, K W. Sell, G. B. Thurman,R. C. Knudsen, R. Wistar, Jr., and W. Grace. 1974.Demonstration of a blocking factor in the plasma andspinal fluid of patients with subacute sclerosing pan-encephalitis. I. Partial characterization. J. Exp. Med.139: 902-924.

21. David, J. R., and R. David. 1971. Assay for inhibitionof macrophage migration. In vitro methods. In Cell-Mediated Immunity. B. R. Bloom and P. R. Glade,editors. Academic Press, Inc., New York. 249-258.

22. Granger, G. A., and W. P. Kolb. 1968. Lymphocytein vitro cytotoxicity: mechanisms of immune and non-immune small lymphocyte mediated target L cell de-struction. J. Immunol. 101: 111-120.

23. Shacks, S. J., J. Chiller, and G. A. Granger. 1973.Studies on in vitro models of cellular immunity: therole of T and B cells in the secretion of lymphotoxin.Cell. Immunol. 7: 313-321.

24. Knudsen, R. C., A. Ahmed, and K W. Sell. 1974. Anin vitro microassay for lymphotoxin using the mul-tiple automated sample harvester. J. Immunol. Methods.5: 55-63.

25. Minowada, J., T. Ohnuma, and G. E. Moore. 1972.Rosette-forming human lymphoid cell lines. I. Estab-lishment and evidence for origin of thymus-derived lym-phocytes. J. Natl. Cancer Inst. 49: 891-895.

26. Stobo, J. D., and W. E. Paul 1972. Functional hetero-geneity of murine lymphoid cells. II. Acquisition ofmitogen responsivness and of e antigen during the on-togeny of thymocytes and "T" lymphocytes. Cell. Im-munol. 4: 367-380.

27. Smith, R. W. 1975. In Proceedings of the Eighth Leu-kocyte Conference, Academic Press, Inc., New York.In press.

28. Bach, F. H., and M. L. Bach. 1973. Cell-mediated im-munity: separation of cells involved in recognitive anddestructive phases. Science (Wash. D. C.). 180: 403-406.

29. Granger, G. A., G. E. Moore, J. G. White, P. Mat-zinger, J. S. Sundsmo, T. S., S. Shupe, W. P. Kolb, J.


Kramer, and P. R. Glade. 1970. Production of lympho-toxin and migration inhibitory factor by establishedhuman lymphocyte cell lines. J. Immunol. 104: 1476-1485.

30. Owen, F. L., and M. W. Fanger. 1974. Studies on thehuman T-lymphocyte population. I. The developmentand characterization of a specific antihuman T-cell anti-body. J. Immunol. 113: 1128-1137.

31. Shirai, T., and R. C. Mellors. 1971. Natural thymocy-

totoxic autoantibody and reactive antigen in New Zea-land black and other mice. Proc. NatI. Acad. Sci. U. S.A). 68: 1412-1415.

32. Talal, N., and A. D. Steinberg. 1974. The pathogenesisof autoimmunity in New Zealand black mice. Curr. Top.Microbiol. Immunol. 64: 79-108.

33. Barthold, D. R., S. Kysela, and A. D. Steinberg. 1974.Decline in suppressor T cell function with age in femaleNZB mice. J. Immunol. 112: 9-16.


human t- cell heterogeneity as...

Documents