Autocytotoxic and autosuppressor T-cell lines generated from autologous lymphocyte cultures

Autocytotoxic and Autosuppressor T-Cell Lines Generated from Autologous Lymphocyte Cultures Karen Rosenkrantz, Bo Dupont, Donna Williams, and Neal Flomenberg

ABSTRACT Limiting dilution analyses have demonstrated both the generation and suppression of autocytotoxic cells following in vitro stimulation with autologous peripheral blood mononuclear leukocytes (PBL). Therefore, in order to isolate and characterize the autocytotoxic lymphocytes, interleukin 2-dependent cell lines were derived from autologous mixed lymphocyte microcultures. The cell lines were screenedfor cytolytic activity against autologous phytohemagglutinin-activated lymphoblasts, autologous and allogeneic B-lymphoblastoid cell lines (B-LCL), and the natural killer target K562. Of 189 cell lines analyzed, 26 demonstrated cytotoxicity against autologous target cells. Cell surface phenotyping of all cell lines indicated that they were of T lymphocyte lineage. Two autocytotoxic T-cell clones were subsequently derived in similar fashion. Cell lines were also screenedfor autoregulatory activiiy. Two cells lines were identified that inhibited the generation of autocytotoxicity. Neither of the autoregulatory lines was capable of directly lysing an autocytotoxic line, suggesting that these autosuppressor cells exert their inhibitory effect by a mechanism other than direct lysis of the autocytotoxic effictor cell. These findings indicate that through the application of limiting dilution analysis and in vitro cell culture techniques, au~t11,1w~oJtww i..ylW~rdIIg, yl,= ~gl~14gi4~ilUflJ ~ r l

6~ IIUI~U

~¥1~ a&Igl2~(~ t0 a~al~ze

tjrje

cellular interactions involved in maintaining self-tolerance.

ABBREVIATIONS B-LCL, B-lymphoblastoid cell line EBV Epstein-Barr virus HS human serum IL-2 Interleukin-2 LAK lymphokine activated killer cell

MoAb NK PBL PHA

monoclonal antibody natural killer cell peripheral blood mononuclear leukocytes phytohemagglutinin

INTRODUCTION Immunologic self-tolerance is a fundamental characteristic of healthy organisms [ 1]. The mechanisms whereby immunologic responses against self are prevented or eliminated are largely unknown. Many self-reactive cells appear to be eliminated or aborted in the thymus [2,3]. However, there is evidence that selftolerance is in part maintained by peripheral mechanisms as well. Deletion of

From the Human Immunogenetics Laboratory and Effector Lymphocyte Biology Laboratory, Memorial Sloan-Kettering Cancer Center, New York, New York. Address reprint requeststo Karen Rosenkrantz, M.D., Box 41, MemorialSloan-Kettering CancerCenter, 1275 York Avenue, New York, 10021. ReceivedNovember24, 1986; AcceptedMarch 31, 1987. Human Immunology 19, 189-203 (1987) © Elsevier Science Publishing Co., Inc., 1987 52 Vanderbilt Ave., New York, NY 10017

189 o198-8859/87/$3.5o

190

K. Rosenkrantz specific T-cell subsets may lead to autoimmune ~iseases in experimental animals [4]. In addition, cell lines and clones have been identified that down-regulate autoaggressive responses both in vitro [5,6] and in vivo in animal models of autoimmune disease [7]. Previous work has indicated that both autoaggressive cells and cells capable of down-regulating autoaggressive cells are generated in mixed lymphocyte culture [8]. Cells capable of lysing autologous target cells are demonstrable in limiting dilution analyses at low responder cell doses, while at higher responder cell doses autocytotoxicity disappears. This biphasic response suggests that a low frequency regulatory cell is capable of preventing autoaggressive responses. In this report, we describe the isolation and preliminary identification of both the autocytotoxic and autosuppressor cell populations. Our results indicate that T cells capable of autocytotoxicity and autosuppression do indeed exist in the peripheral blood of healthy human donors.

MATERIALS A N D METHODS Limiting dilution analysis assay LDA assays were performed in V-bottom Linbro microtiter plates (Flow Laboratories, Walkersville, MD) in 0.2 ml supplemented RPMI-1640 (Gibco, Grand Island, NY)with 10% autologous (responder) serum and an IL-2 source. Wells contained graded concentrations of responder PBL and 5 × 104 irradiated stimulator PBL (4000 rads) or BLCL (8000 rads). On the seventh day, 100 Izl of media were removed and replaced with 100 Izl of supplemented RPMI with 10% pooled human serum or autologous serum and IL2. On the tenth day, individual wells were assayed for cytotoxicity in a chromium release assay by adding 3 x 103 5aCr (SmCi/ml, New England Nuclear, Boston, MA) labeled target cells to each well. The supernatants were harvested after 4 hr. PHA-activated lymphoblast target cells were prepared by culturing PBL in 1% PHA-M (Difco Laboratories, Detroit, MI) for 4 days [8]. The cytolytic activity of an individual well was scored as positive when the release of ~:Cr exceeded the mean of the control plate (stimulator cells alone) by 3 standard deviations. The Rankit test was used to demonstrate that the counts obtained from a set of eight control plates (stimulators alone) were distributed normally [8]. Peripheral blood mononuclear leukocytes (PBL) were isolated by Ficoll-Hypaque density gradient centrifugation (Lymphoprep, Hicksville, NY) [9]. Interleukin-2 (IL-2)-containing human-conditioned media was produced from PBL that had been pulsed for 2 hr with purified phytohemagglutinin (PHA-HA-16, Burroughs Wellcome, Research Triangle Park, NC) and phorbol myristic acetate (Sigma, St. Louis, MD) as described [10].

Isolation of autocytotoxic and autosuppressor cell lines and clones. Heparinized blood was obtained from a healthy Epstein-Barr Virus (EBV)-seronegative individual (HLA-A26,28,BI4,w41,w6,Cw8,DR1,3). Cell lines were established in 96 well microtiter plates by incubating 5000 PBL per well with 50,000 autologous irradiated (4000 rads) stimulator PBL per well in RPMI-1640 + 10% autologous serum and 15% human-conditioned media. After 10 days, the contents of individual wells were expanded into cell lines through cultured in RPMI-1640 + 10% pooled human serum (HS) and ILo2 (LYMPHOCULT, Biotest, East Fairfield, NJ). Autologous irradiated (4000 rads) feeder PBL ( 1.0 x 106 PBL/ml) were added to the cultures at weekly intervals. To generate autocytotoxic clones, PBL were obtained from another individual (HLA-AI,B8,w6,Cw7,DR3,4). Cultures were prepared in 96-well plates, as de-

Autocytotoxic and Autosuppressor T-Cell Lines

191

scribed above. On day 10, the contents of the wells were pooled. The cells were counted and cloned via limiting dilution at a density of 0.3 cells per well. Clones were expanded as described above, utilizing human-conditioned media as the IL2 source. The probability that a growing culture was truly derived from a single precursor cell was greater than 98% [ 11 ].

B-lymphoblastoid lines. B-LCL were derived from normal donor lymphocytes by transformation with EBV as previously described [ 12]. B-LCL were maintained in culture in RPMI 1640 + 15% HS using standard tissue cu!ture techniques.

Monoclonal antibodies. For the cytotoxicity blocking studies, the following monoclonal antibodies (MoAbs) were utilized in the form of ascites fluids: the MoAb W6/32 recognizes a monomorphic determinant on HLA class I antigen [13] (American Tissue Type Culture, Rockville, MD); the MoAb SY2 recognizes a common determinant on HLA-DR/DQ molecules (S.Y. Yang, unpublished observations); the MoAb 8B reacts with a non-HLA antigen present on all peripheral blood B lymphocytes and EBV-transformed B-LCL (Y. Morishima, unpublished observations). MoAbs used in phenotypic analysis were: anti-Leu 4 recognizing the pan-T-cell antigen CD3 [T, p19-29], anti-Leu 3 reacting with the CD4 [T, p55] antigen, anti-Leu 2 recognizing the CD8 IT, p32-33] antigen, anti-Leu 11 reacting with an antigen (CD16) associated with the Fc receptor on natural killer cells, neutrophilic granulocytes, and anti-Leu 7 reacting with a subset of natural killer cells. The above antibodies were purchased as Fluorescein isothiocyanate conjugates (Becton Dickinson Monoclonal Center, Inc., Mountain View, CA.). The MoAb OKT11 reacts with the CD2 [T, p50] antigen and was obtained from Ortho Pharmaceuticals (Raritan, NJ).

Cytotoxicity assays. Effector cell lines or clones were washed with RPMI 1640 prior to their use. 5 x 106 target cells in 0.2 ml of medium were labeled with 250 ~ c i N~ . w 2 Lrs]c,-ln. (5 ,,~F;/,,~l, 5~Cr; ~ o w ,_.,~ . . . . . ~,u,.,,~, ,.,v~cu,, ~,,~/,u, 1.5 hr at 37°C, followed by two washes in RPMI 1640 + 5% HS. Target cells and effector cells were resuspended in RPMI 1640 + 10% HS at a concentration of 3 x 104 cells per ml and 7.5 x 105 cells per ml, respectively. Triplicate 0.1 ml aliquots of target cells and effector cells were dispensed into round bottom microtiter wells (Linbro, Hamden, CT). After incubation for 4 hr at 37°C, the plates were centrifuged (200g for 5 rain). The supernatants were harvested (Skatron Harvesting System; Flow Laboratories, Walkersville, MD) and counted in a gamma counter. Spontaneous release was determined by incubating target cells in medium alone. Maximum release was determined by adding 5% Triton-X100 (New England Nuclear) to 0.1 ml of the target cell suspension. Percent cytotoxicity or percent specific lysis was calculated as: .

.

.

.

.

.

~..-~

j

v q

aa*

~..~u

az*a

~.Xlq,

experimental 5~Cr release (cpm) - spontaneous 51Cr release (cpm) ...... maximum 51Cr release (cpm) - spontaneous 51Cr release (cpm)

x

100,

where cpm is counts per minute. Spontaneous release did not exceed 25% of the ma::imum releasable counts for any of the target cells used [14].

Inhibition of cell mediated cytotoxicity by MoAbs. Target cells were resuspended at 6 x 104/ml and antibodies were appropriately diluted in RPMI 1640 + 10% HS. Fifty microliter aliquots of target cells and antibodies (or media controls) were plated in triplicate. Plates were shaken momentarily to ensure adequate mixing of cells and antibody and incubated at room temperature for 45 rain. Thereafter,

192

K. Rosenkrantz the remainder of the assay was performed as described above. In all cases, the spontaneous S~Cr release of the target cells in the presence of antibody was determined to ensure that the observed changes in target cell lysis were not due to direct effects of antibody on the target cells.

lmmunofluorescent characterization of T-cell surfacephenotype. After thorough washing in phosphate-buffered saline with 1% bovine serum albumin and 0.02% sodium azide, 5 x l0 s cells were incubated with Fluorescein isothiocyanate conjugated MoAb for 30 rain at 4°(2. After extensive final washing, cells were examined using an Epics-C cell sorter (Coulter Electronics, Hialeah, FL).

Microwell cytotoxicity assay for autosuppression. Cell lines were screened for autosuppressor activity using an adaptation of the limiting dilution analysis assay [8]. Five-thousand responder PBL per well and 50,000 irradiated stimulator autologous PBL per well were cocultured in 96-well microtiter plates in RPMI 1640 containing 10% autologous serum and 15% human-conditioned media. Cell lines (10,000 cells/well) were added to these microwells. The plates were incubated for 7 days at 37°(2. Individual wells were assayed for cytotoxicity by adding 3000 S~Cr-labeled target cells (5 mCi/ml; New England Nuclear) to each well as previously described [8]. The cytolytic activity of an individual well was scored as positive when the release of S~Cr exceeded the mean cpm of the control plate (stimulator cells alone) by three standard deviations. RESULTS

Identification of Autocytotoxic Cell Lines Figure 1 illustrates a typical limiting dilution analysis assay of the cytotoxic re~ sponse generated toward autologous PHA-activated lymphoblasts following in vitro sensitization with autologous PBL stimulator cells. Raw CPM from each microtiter well are shnwn in rh~ *lprr~r... po,,i,-,, ,~¢...,ho..,.figure ...hao...,,.,I.,.,. . . ..~o,,,,--..,* ... .... dose response curve is shown in the lower portion. As previously described [8], limiting dilution analyses of autocytotoxic responses exhibit typical biphasic curves. Strong autocytolytic responses are generated at low responder ceil doses, with S~Cr release in some of the microtiter wells reaching levels that are threefold above background. At high responder cell doses, autocytotoxicity is inhibited and S~Cr release returns to baseline. This contrasts with allocytotoxic responses in which all wells exhibit positive responses at identical high responder cell doses. This biphasic response is consistent with the actions of two counterpoised cell populations, one capable of autocytotoxicity and the other capable of inhibiting autocytotoxic responses [8,11]. In order to isolate and characterize cells with autocytotoxic and autoregulatory activity, microcultures were established with 5000 responder PBL per tissue culture well. This responder cell dose was selected since multiple limiting dilution analysis assays had indicated that the point of maximum autocytotoxicity occurred at responder cell doses of 5-10,000 cells per well [8]. After 10 days, individual microcultures were expanded into cell lines through culture in lymphocyte conditioned medium. One hundred and eighty-nine cell lines were screened for cytolytic activity against autologous PHA lymphoblasts and autologous B-LCL. Twenty-six of these cell lines exhibited strong autocytolytic activity (--20% lysis at an effector/target ratio of 25/1) against an autologous target cell: either autologous PHA lymph° oblasts (range 25-31% cytotoxicity) and/or autologous B-LCL (20-64% cyto-

Autocytotoxic and Autosuppressor T-Cell Lines

193

RESPONDERCELL NUMBERPERWELL (x 10-5) 0 1200 |

625 •

1.25

2.5

5 .

I0

30 ~

ggg'61

50 _ ~ -_

I

I D

"

G.

"

,~ f

3 ~

o ~3

•

t.

~o

. ib

1

r

:

"

"

v

|

m ,-,'1

•

>. I-.- 4000 x O IO I>O

|

m

6000 CellNumber(x 10-3) i

IO 2 ~ 30 40 5O

FIGURE 1 Raw data from a typical limiting dilution analysis experiment are shown in the top half of the figure. Each dot represents 5=Cr reiease observed in an individual microwell. The top of the boxed area represents the mean 5:Cr release observed in the cont:ol plate (responder cell dose = 0). The bottom of the boxed area represents three standard deviations above this value. Points within the box were thus scored as negative and points below the box as positive. The resuitant dos~ response cut ve is illustrated in the bottom of the figure.

toxicity). There were 16 additional lines that exhibited between 10 and 20% autocytotoxicity against an autologous target cell. Each of the cell lines described hereafter has a unique identification number with a prefix, indicating its cytolytic pattern. In the prefix, P = lysis of autologous PHA lymphoblasts, B - lysis of autologous B-lymphoblastoid cell lines, N K = lysis of the N K sensitive target K562, and N C - no cytolytic activity toward any target. Eleven cell lines exhibiting strong ( > 2 0 % ) autocytotoxicity were then studied in more detail (Table 1). Nineteen noncytotoxic cell lines (hereafter N__CCcell lines) and cell lines that lysed only K562 (NK cell lines) were also studied to provide a basis for comparison. Four types of autocytolytic patterns were observed. Three cell lines lysed autologous PHA-activated lymphoblasts (PBNK, P N K lines). Two of these cell lines also demonstrated cyto!ytic activity against K562 and autologous B-LCL. However, the autocytotoxicity directed against PHA-lymphoblasts was not absolutely associated with the autocytotoxicity against

194

K. Rosenkrantz

TABLE 1 Cytotoxic profile of cell lines" Targets

Cell line t y p e

B-LCL

PHA lymphoblast

K562

PBNK

PNK BNK

+ +

+ + -

+ + +

B

+

-

-

4

NK

-

-

+

II

NC

-

-

-

8

m

Number observed 2 1 4

Cell lines were screened for cytotoxiciw against an autologous B-LCL autologous PHA-lymphoblasts, and K562. +: greater than 20% lysis of the target cell at an effector/target ratio of 25/1.

B-LCL. One cell line, PNK 32, demonstrated 31% lysis of PHA-activated lymphoblasts at an effector/target ratio of 25:1, while lysis of the autologous B-LCL was only 4% under the same experimental conditions (Figure 2). This cell line also demonstrated 17% cytotoxicity against natural killer target K562. Eight autocytotoxic cell lines were identified that lysed autologous B-LCL but not autologous PHA !ymphoblasts. Four cell lines (BNK cell lines) which demonstrated between 26 and 41% cytotoxicity against the autologous B-LCL also demonstrated between 10 and 75% cytotoxicity against K562, the natural killer cell target. However, four other cell lines (B lines) which exhibited between 23 and 50% cytotoxicity against the autologous B-LCL had no cytolytic activity

FIGURE 2 Cytolytic activity of cell lines that lyse autologous PHA-activated lymphoblasts. Cell lines were screened for cytolytic activiw against autologous PHA-activated lymphoblasts, B-LCL, and T-cell lines as described in Materials and Methods. They were also tested against two allogeneic B-LCL: PLH (HLA-A3,Bw47,DR7) and DBB (HLAA2,Bw57,DR7) as well as K562. The effector/target ratio in each case was 25/1. TARGET CELL

PNK 52

PBNK 55

PBNK 96

PHA ACTIVATED LYMPHOBLAST BLCL PNK32

il

BNKI32~

B43

NK 88LLI NC170LJ

BI6

BNK66

°I

~,

NKI 2 NK5-i NC45 BLCL PLH BLCL DBB P ~ s K562 I , I, 0 2O

NT

,II O

20

I 0

2'0

|

!

40

PERCENTAGE CYTOTOXICITY (SICr RELEASE )

|

Autocytotoxic and Autosuppressor T-Cell Lines

193

against K562. Autocytotoxicity activity, therefore, appeared to be distinct from the natural kiUer-like activity. This is in agreement with previous experiments utilizing limiting dilution analysis in which the kinetics of the cytotoxicity against autologous targets are clearly distinguishable from those against the natural killer cell target K562 [8]. As K562 is also readily lysed by lymphokine activated killer (LAK) cells, these findings also indicate that :hese autocytotoxic populations do not represent classical LAK cells. Phenotypic analysis of the autocytotoxic cell lines demonstrated that these cell lines were composed of T cells: virtually 100% of the cells in each culture were positive for the two pan-T lymphocyte antigens, CD3 (p 19-29) Leu 4 and CD2 (p50) T11. Phenotyping with antibodies detecting the T-cell subset antigens CD4 (p55) Leu 3 and CD8 (p 32-33) Leu 2 as well as the natural killer cell markers Leu 7 and CD16 revealed heterogeneous patterns consistent with the polyclonal origin of the cell lines. In addition, the phenotypic data indicated that, with respect to the lymphocyte antigens studied, there were no differences between cells capable of lysing autologous targets and cells that were not (data not shown). Target Specificity of the Autocytotoxic Cell Lines We screened the various autocytotoxic cell lines for lyric activity against other autologous T-cell lines and against themselves. None of the cell lines, including those which lysed PHA-activated lymphoblasts, were able to lyse autologous ILexpanded T-cell lines. In this panel study, cell line PNK 32 (Figure 2), for example, exhibited 31% cytotoxicity against autologous PHA-lymphoblasts at an effector/target ratio of 25/1, but was unable to ~yse any of eight autologous T-cell lines against which it was screened. Similarly, cell line PBNK 35 demonstrated 25% cytotoxicity against autologous PHA-lymphoblasts, but was also unable to lyse any of the eight autologous T-cell lines against which it was screened. PBNK 96 exhibited similar lytic activity toward a smaller target cell panel. PNK 32 (Figure 2), B 43 (Figure 3A), and BNK 66 (data not shown) were studied to ascertain whether they were capable of lysing themselves. None of the three demonstrated self-lysis, including cell line PNK 32, which demonstrated 31% cytotoxicity against autologous PHA-lymphoblasts. To assess whether the autoaggressive cell lines were directed at an HLAencoded gene product, cell line B 43 was studied in detail. As illustrated in Figure 3A, this cell line was capable of strongly lysing an autologous B-LCL but not autologous PHA-acrivated lymphoblasts or K562. In repeated testing, the percent lysis of autologous B-LCL at a 25/i effector/target ratio varied only between 28 and 37%. This cell line was then tested in MoAb-blocking studies (Figure 3B). While a preincubation of target cells with w6/32 MoAb against HLA class I molecules) or with 8B (a MoAb detecting a pan-B cell antigen) had no effect on autocytotoxicity, SY2 (a MoAb directed against a monomorphic epitope present on HLA-DR and DQ molecules) inhibited the autocytotoxic response. The blocking data indicates that this cell line is directed against or restricted by HLA class II molecules. We also asked whether the autoaggressive cell lines exclusively recognized self determinants or whether these cell lines were also capable of recognizing determinants present on allogeneic cells. To address this issue, T-cell clones were developed. Only autologous PBL were used in the initial priming and subsequent feedings. The T-cell clones were tested for autocytotoxic activity against autologous B-LCL. Two autocytotoxic clones were identified. These clones exhibited patterns of cytolytic activity similar to those exhibited by the autocytotoxic Tcell lines and expressed the phenotype of a mature T cell (CD2 +, CD3 +, CD4 +,

196 A A U

PERCENTAGE

TARGETCELLS 0

T

CYTOTOXlCITY

I0

20

I

I

•

I

O L O

PHA~-ACTIVATED LYMPHOBLASTS

G

B 45

U

BNK 6 6

S

40

I

BLCL

0

(51Cr RELEASE)

30

A

L L 0 G E

BLCL PLH I BLCL DBB K562

N

E

I I

I

C

B

A rid If) ti.I /

,,* 5 0

~

40

IU x 0 tO F>. u

o

ANTI HLA CLASS I

5O

A

ANTI B-CELL

W IZ w u w

IO

X

m NO

ANTIBODY

X

ANTI HLA CLASS I! i

I

1/1000 I/IOO

I

I/IO

DILUTION OF ANTIBODY F I G U R E 3 Cytotoxic activity of B lines. Cell line B 43: anti-autologous HLA class II. (A) Analysis of cytotoxic pattern. Cytolytic profile of cel! !ine B 43 at an effector/target ratio of 25/I. (B) Blocking with MoAbs. Fine specificity of cell line B 43. Cell line B 43 was tested for cytotoxicity against the autologous B-LCL in the absence (shaded bar) or presence of w632, anti-HLA class I (0); 8B, anti-B cell (A); and SY2, anti-HLA class II (X) MoAbs at an effector/target ratio of 25/1. The MoAbs were utilized in the form of ascites fluids at the dilutions shown.

Autocytotoxic and Autosuppressor T-Cell Lines CLONE NC 2

CLONE BNK 3

!97 CLONE B 7

60

TARGET CELLS L\\\~ KRZ MR3 ¢'727] K562

~40 t~r

¢3 2O

,v

x\

,,\,

0

FIGURE 4 Cytolytic activity of clones. T-cell clones were screened against: KRZ, an autologous B-LCL (HLA-A1, B8, DR3, 4); MR3, an allogeneic B-LCL (HLA-A2,BT, DR2); and K562 at an effector/target ratio of 25/1.

C D 8 - ) (Figure 4). Clone B 7 capable of lysing an autologous B-LCL (20% lysis at an effector/target ratio of 25/1), but not an allogeneic B-LCL and not K562. In contrast, clone B N K 3 was capable of lysing an autologous B-LCL (57% lysis), an allogeneic B-LCL (55% lysis), and K562. Clone NC 2 represents a noncytolytic clone isolated from the same priming. These studies with cloned T cells suggest that a single cytotoxic cell is capable of recognizing determinant(s) on both autologous and allogeneic target cells. Identification of Cell Lines Capable of Autosuppression It seemed likely that cell lines without direct autocytotoxicity might contain cells capable of suppressing autocytotoxic responses. Of the various types of autocytotoxic responses described above, we chose to study suppression of the autocytotoxic response against autologous PHA-lymphoblasts. Cell lines were tested for their capacity to interfere with the development of autocytotoxicity in a limiting dilution assay. It is demonstrated in Figure 5 that, without adding cell lines, the fraction of wells displaying autocytotoxicity ranged from 0.3 to 0.6 (gray shaded area, Figure 5). As a control, two cell lines (PNK 32 and PBNK 35) which had previously demonstrated direct cytotoxicity against autologous PHA-activated lymphoblasts were added to the minicultures. When those two c. .o. i. l .l i.n.o. ~. . .w . .o.r.o

...-..~...~aclA°A,~ll. tk...~. . . . ..,.**olle i n - -

9 A ~! uA~.x. x. .I.u. .u.a. . I---x

tx ~ t ~ uJ

. .U. t. .u.l. .. y t u iiy. . .U. L: ~

i'~5pofises

on day 7. In contrast, two of 14 additional cell lines tested (NK 18 and BNK 2 l) reduced the number of cells, demonstrating autocytotoxicity to almost zero (p <0.01). To begin to study the mechanism underlying the suppression, we tested the suppressor cell lines for their ability to lyse the autocytoxic cell line PNK 32 (Figure 6), which itself directly lysed autologous PHA-lymphoblasts and was capable of enhancing the autocytotoxic response measured in the limiting dilution assay. Neither of the two suppressor cell lines was capable of lysing PNK 32. Moreover, these autosuppressor cell lines were also incapable of lysing other autologous T-cell lines. This suggested that autosuppressive cell fines exert their function by a mechanism other than direct lysis of the autocytotoxic effector cell. Both suppressor cell lines were able to lyse K652. In addition, cell line BNK 21 demonstrated cytotoxicity against the autologous B-LCL (Figure 6). Phenotypic studies on the cell lines were performed to detect whether there were any differences between the cell lines that were capable of suppressing cytotoxic responses and those that were not. The autosuppressor cell lines expressed the T lymphocyte antigens CD3 (p19-29) Leu 4 and CD2 (p50) T11. In addition, both regulatory cell lines contained subpopulations expressing the CD4 and CD8 surface molecules. The fluorescence pattern for these cell surface antigens was biphasic, suggesting that each cell line was composed of at least two distinct subpopulations. The autosuppressive cell lines NK 18 also contained

198 ADDED CELL LINE" NONE

AUTOCYTOTOXIC (PHA BLAST) NONCYTOTOXlC m Z

"10

Z

Z

Z

Z

CYTOTOXlC (BLCL OR K562)

Z

Z

Z

Z

Z

Z

Z

Z

Z

Z

Z

II iiiiiiiiii;!i:iiiiii;ii!ii;i iiii i !iiiiiiiii

FIGURE 5 Addition of effector cell lines to autocytotoxicity assay. Cell lines were screened for autoregulatory activity. In the control group, 5000 responder PBL were added to each well with 50,000 irradiated autologous PBL. To screen for autosuppression, cell lines (10,000 cells/well) were added with the 5000 responder PBL to individual wells with the autologous stimulator cells. After 7 days, individual wells were assayed for cytotoxicity against 5'Cr-labeled autologous PHA-activated lymphoblasts. In the control groups (without added cell lines) the fraction of wells displaying autocytotoxicity ranged from 0.3 to 0.6 (gray shaded area). Each bar represents a group of 24 test wells. The asterisks denote cell lines identified that inhibit the autocytotoxicity (~', p < 0.01).

FIGURE 6 The cell lines with autosuppressor activity were screened for cytolytic activity at an effector/target ratio of 25/1. NK 18

TARGET CELL

BNK 21

PNK 321-] PHA ACTIVATEDFI LYMPHOBLAST ir J BLCLIL

i[

BI6 / B43 BNK66 NKI2 NK53 NKSI

3

i A¢ NC-,~

BLCL PLH BLCL DBB K562 <~1

I

I ,

O

'

20'

0

20

PERCENTAGE CYTOTOXICITY

,

i

40

,

,

,

I

60

(5'Cr RELEASE)

Autocytotoxic and Autosuppressor T-Cell Lines

199

subpopulations expressing the Leu 7 and CD 16 surface molecules. On the other hand, cell lines N K 12, N K 33, and BNK 16 which did not suppress cytotoxicity in the limiting dilution assay, also contained Leu7- and CD 16-positive populations (data not shown). Therefore, with respect to lymphocyte antigens studied, there was no distinct phenotypic pattern in the autosuppressor cell populations. DISCUSSION Limiting dilution analysis experiments have indicated that cells capable of autoaggression and cells capable of controlling autoaggression are present in the peripheral blood of healthy individuals [8]. The present studies indicate that under appropriate experimental conditions, these populations can be separated and propagated. These findings thus demonstrate the existence of the autocytotoxic and regulatory populations initially suggested from limiting dilution experiments. These populations appear to be related to classical T lymphocytes in that all of the cell lines and clones isolated uniformly expressed the CD3 antigen, which is noncovalently associated with the T-cell receptor. Distinct types of auttocytotoxic cell lines were identified based upon their capacity to lyse different combinations of target cells. In previous limiting dilution analysis experiments utilizing autologous PBL as stimulator cells, a biphasic response can be demonstrated toward either autologous PHA-lymphoblasts or autologous B-LCL. In this study, it is suggested that at least some of the cytotoxic lymphocytes capable of lysing autologous PHA-lymphoblasts may be distinct from those capable of lysing autologous B-LCL. Three cell lines were identified that lysed autologous PHA-lymphoblasts. One of these was unable to lyse autologous B-LCL. On the other hand, eight cell lines were identified which lysed autologous B-lymphoblastoid cells, but not autologous PHA-activated lymphoblasts. The target antigens recognized by the autocytotoxic cells have not been established. The !ytic activity of one cell line, B 43, appeared to be directed at or restricted by HLA class II molecules. Seven of the cell lines as well as one Tr. P. I. I. . .f i. n. n. . e. . •l y ~~.e d k,-~.-k au,~,,vsuu~ o,,t . . . . . . . . ~,,,d allogeneic n-L.~.t.. ~ " " " Previously described autocytotoxic cells have demonstrated a similar range of target specificity: cytotoxic T cells directed against autologous lymphob!asts have been identified [15-22], as have cytotoxic cells with broad reactivity against EBV-transformed cells of autologous and/or allogeneic origin [16,23,24]. It is possible, but unlikely, that the cell lines capable of lysing only B-LCL and not PHA-lymphoblasts were directed at EBV-associated antigens. Cytotoxic cells with specificity for autologous EBV-transformed cells have been reported following in vitro stimulation with autologous EBV-transformed cells or utilizing PBL from EBV-seropositive donors [25-27]. Only autologous PBL were used as stimulator cells and feeder cells in our present studies, and the donor we selected has repeatedly been and continues to be seronegative for anti-EBV antibodies over a 5-year period. This suggests that these cell lines were not directed at a viral product on the B-LCL Similarly, it is unlikely that the lysis of autologous PHA-lymphoblasts in this study was simply the result of lectin-induced nonspecific cytotoxicity [28-32], as not all cell lines capable of cytolytic activity lysed the autologous PHA-activated lymphoblasts. The ~nteracdon of the PBNK/PNK effector cells with these target cells thus appears to be independent of the PHA used for target cell activation. Autocytolytic activity may be seen in bulk cultures of LAK cells [33]. While some LAK cells may potentially lyse autologous PHA lymphoblasts or B-LCL, several of the lines and clones described in this report failed to lyse to NK/LAK

200

K. Rosenkrantz sensitive target K562. These cells are thus distinct from classical NK or LAK populations. The biological significance of the autocytotoxic cell populations is not presently known. These autoaggressive cells may perform specific functions such as intercellular communication and regulation. Alternatively, some of the autocytotoxic cells may represent true aberrant "forbidden clones" with antigenic receptors directed at serf-determinants and without any specific physiological function. Finally, these cytotoxic cell lines may be directed at alloantigens or at foreign antigens in association with self, but their recognition structures may bind to selfantigens in the absence of foreign antigens with sufficient affinity to allow the lysis of autologous target cells. Autoreactive antibodies occur in the initial response against a foreign antigen [15]. Under the pressure of stimulation by antigen, antibodies with higher affinity for the antigen appear and the relative number of self-reactive antibodies diminishes. Some of the autocytotoxic cells we have described may represent cellular counterparts of such autoantibodies. Two regulatory cell lines were identified in these studies that completely suppressed the development of autocytotoxicity. It is unlikely that suppression of autocytotoxicity was caused by overgrowth in the microcuhure wells or by absorption of IL-2 [35,36]. Other IL-2 depenent cell lines tested in parallel, such as PNK 32 and PNK 35, enhanced autocytotoxicity such that all microtiter wells exhibited positive responses. In addition, exogenous IL-2 was added in excess to the limiting dilution assay. It is presently unknown whether these cells directly exert the suppressor effect or induce the development of other suppressor effector populations [37-40]. The regulatory cells appear to act by a mechanism other than direct lysis of autoaggressive cells. Specifically, neither of the suppressor cell lines tested were able to lyse PNK 32, a cell line that markedly increased autocytotoxicity in this assay and that directly lysed autologous PHA-lymphoblasts. In summary, autoaggressive cell lines can reproducibly be generated as predicted from the limiting dilution analysis assay--under conditions in which autoregulatory populations are diluted out. These autocytotoxic lymphocytes are a . . ~ c ~ , ~ - , ~ . c u u s popmauon of cells capable of iysing a variety of selt-target cells. The lytic activity of these cells may reach levels comparable to that of allocytotoxic lymphocytes. In addition, autoregulatory cells can be propagated and characterized independently. In this case, autosuppressive cell lines were identified that down-regulate cytotoxic cells that lyse autologous PHA-activated lymphoblasts. The elimination of an anti-self-response may be in part Eased on the clonal abortion or deletion of anti-self reactive cytotoxic lymphocytes [2,3], but it appears that a peripheral autoregulatory mechanism capable of controlling autocytotoxic cells exists as well. The demonstration of both autocytotoxic cells as well as a mechanism for their regulation strengthens the concept that autocytotoxic cells are not an in vitro culture artifact, but rather that cells capable of autocytolysis represent a normal constituent of peripheral blood in healthy individuals. Disturbances in the interactions between autoaggressive and autoregulatory cell populations might be expected to play a role in the pathogenesis of diseases, including autoimmune disorders, neoplastic transformations, and graft versus host disease. ACKNO~(q.EDGMENTS

This work was supported by grants from the U.S. Public Health Services, National Institute of Health, NCI-CA-22507, CA-08748, CA-33050, CA-23766, CA-44322, and a g'ant from the XOMA corporation. Dr. Karen Rosenkrantz was supported by a Clinical Scholar's

Autocytotoxic and Autosuppressor T-Cell Lines

201

Award supported by NIH grant CA-09512. Dr. Neal Flomenberg was supported by a Clinical Scholar's Award from the Norman and Rosita Winston Foundation. We wcmld like to express our appreciation to Carol Bodenheimer for technical assistance. We thank Richard O'Reilly and Keith Reemtsma for their continued help and support, and Louis Rozos for help in preparing the manuscript.

REFERENCES 1. Jerne NK: The natural-selection theory of antibody formation. Proc Natl Acad Sci USA 41:849, 1955. 2. Burner FM: The clonal selection theory of acquired immunity. London, England, Cambridge University Press, 1959. 3. Nossal GJV, Pike BL: Evidence for the clonal abortion theory of B lymphocyte tolerance. J Exp Med 141:904, 1975. 4. Sakaguchi $, Fukuma K, Kuribayashi K, Masuda T: Organ-specific autoimmune diseases induced in mice by elimination of a T cell subset. I. Evidence for the active participation of T cells in natural self-tolerance; defect of a T cell subset as a possible cause of autoimmune disease. J Exp Med 161:72, 1985. 5. Fink PJ, ,Rammensee HG, Bevan MJ: Cloned cytolytic T cells can suppress primary cytotoxic responses directed against them. J Immunol 133:1779, 1984. 6. Claesson MH, Miller RG: Functional heterogeneity in allospecific cytotoxic T cell clones. I. CTL clones express strong and-self suppressive activity. J Exp Med 160: 1702, 1984. 7. Cohen IR, Holoshitz J, van Eden W, Frenkel A: T lymphocyte clones illuminate pathogenesis and effect therapy of experimental arthritis. Arthritis Rhuem 28:841, 1985. 8. Rosenkrantz K, Dupont B, Flomenberg N: Generation and regulation of autocytotoxicity in mixed lymphocyte cultures: evidence for active suppression of autocytotoxic cells. Proc Natl Acad Sci USA 82:4508, 1985. 9. Boyum A: Isolation of mononuclear cells and granulocytes from human blood. Scand J Clin Lab Invest 21 (Suppl 97):77, 1968. 10. Metier SC, Schlossman SF, Reinerz EL: Clonal analysis of human cytotoxic T lymphocytes: T4 + and T8 + effector T cells recognize products of different major histocompatibility complex regions. Proc Nad Acad Sci USA 79:4395, 1982. 11. Lefkovitz I, Waldmann H: Limiting dilution analysis of cells of the immune system. Cambridge, England, Cambridge University Press, 1979. 12. Hansen JA, Fu SM, Antonelli P, Kamoun M, Hurley JN, Winchester RJ, Dupont B, Kunkel HG: B-lymphoid cell lines derived from HLA-D homozygous donors. Immunogenedcs 8:51, 1979. 13. Barnstable CJ, Bodmer WF, Brown G, Galfre G, Milstein C, Williams AF, Zicder A: Production of monoclonal antibodies to Group A erythrocytes, HLA, and other cell surface antigens. New tools for genedc analysis. Cell 14:9, 1978. 14. Schendel DJ, Wank R, Dupont B: Standardization of the human in vitro cell-mediated iympholysis technique. Tissue Antige~s 13:112, 1979. i 5. Miller RA, Kaplan HS: Generation ofcytotoxic lymphocytes in the autologous mixed lymphocyte culture. J Immunol 121:2165, 1978. 16. Tomonari K: Cytotoxic T cells generated in the autologous mixed lymphocyte reaction. I. Primary auto!ogous mLxed lymphocyte reaction. J Immunol 124: 1111, 1980.

202

K. Rosenkrantz 17. Ahman A, Katz DH: Existence of T cells manifesting self reactivity indistinguishable from alloreactivity. J Immunol 125:1536, 1980. 18. Heuer J, Bruner K, Opalka B, Kolsch E: A cloned T-cell line from a tolerant mouse represents a novel antigen-specific suppressor cell type. Nature 296:456, 1982. 19. Bach FH, Leshem B, Kuperman OJ, Bunzendahl H, Alter BJ, Zarling JM: Allosensitization to obtain anti-tumor immunity. In: The potential role of T cells in cancer therapy. A Fefer, AL Goldstein, Eds. New York, Raven, 1982, pp. 7-19. 20. Reimann J, Miller RG: Generation of autoreactive cytotoxic T lymphocs,tes under limiting dilution conditions. J Immunol 131:2128, 1983. 21. Clayberger C, Dekruyff RH, Cantor H: Immunoregulatory activities of autoreactive T cells: an I-A specific T cell clone mediates both help and suppression of antibody responses. J Immunol 132:2237, 1984. 22. Hess AD, Horwitz L, Beschorner WE, Santos GW: Development of graft-vs.-host disease-like syndrome in cyclosporine-treated rats after syngenic bone marrow transplantation. J Exp Med 161:718, 1985. 23. Seely JK, Golub SH: Studies on cytotoxicity generated in human mixed lymphocyte cultures. I. Time course and target spectrum of several distinct concomitant cytotoxic activities, j lmmunol 120:1415, 1978. 24. Zarling JM, Bach FH: Sensitization of lymphocytes against pooled allogeneic cells I. Generation of cytotoxicity against autologous human lymphoblastoid cell lins. J Exp Med 147:1334, 1978. 25. Sugamura K, Hinuma Y: In vitro induction of cytotoxic T lymphocytes specific for EFstein-Barr virus transformed cells: kinetics of autologous restimulation. J Immunol 124:1045, 1980. 26. Fukukawa T, Hirano T, Sakaguchi N, Teranishi T, Tsuyuguchi I, Nagao N, Naito N, Yoshimura K, Okubo Y, Tohda H, Oikawa A: In vitro induction of HI.A-restricted cytotoxic T lymphocytes against autologous Epstein-Barr virus transferred B lymphoblastoid cell line. J Imm,mol 126:1697, 1981. ")"7 'iVT--II----~ ~ 1Di L A 11~ n . . . . tit" it2 . . . . -_ tit A. Epstein-Barr vicas-~ec ific ,..... /-I. WV ~ltligtlktK" , ~ L .:I & . I l:l.~ t.U.I I. .f]kO, J~UW~" lq, I r ' ~ p ~ l L ~ l l l /]rI, A ,ytutoxic T-cell clones restricted through a single HLA antigen. Nature 297:413, 1982. 28. Forman J, Moiler G: Generation of cytotoxic lymphocytes in mixed lymphocyte reactions. I. Specificity of the effector cell. J Exp Med 138:672, 1973. 29. Bevan M, Cohn M: Cytotoxic effects of antigen- and mitogen induced T cells on various targets. J Immunol 114:559, 1973. 30. Bonavida B, Bradley TP: Studies of the induction and expression of T cell-mediated immunity. V. Lectin-induced nonspecific cell-mediated cytotoxicity by alloimmune lymphoblasts. Transplantation 21:94, 1976. 31. Green Wit, Ballas ZK, Henney CS: Studies on the mechanism of lymphocyte-mediated cytolysis. XI. The role of lectin in lectin-dependent cell-mediated cytotoxicity. J Immunol 121"1566, 1978. 32. Bonavida B, Katz J: Studies of the induction and expression of T cell-mediated immunity. XV. Role of non-MHC papain-sensitive target st~ctures and Lyt 2 antigens in allogeneic and xenogeneic lectin dependent cellular cytotoxicity (LDCC). J Immunol 135:1616, 1985. 33. Sondel, PM, Hank J, Kohler PC, Chen BP, Minkoff DZ, MolendaJA: Destruction of autologous human lymphocytes by lnterleukin 2-activated cytotoxic cells. J Immunol 137:502, 1986. 34. Schwartz R: Autoantibodies and normal antibodies: Two sides of the same coin. Harvey Society Lecture 3, November 1985.

Autocytotoxic and Autosuppressor T-Cell Lines

203

35. Palacios R, Moiler G: T cell growth factor abrogates Concanavalin-A induced suppressor cell function. J Exp Med 153:1360, 1981. 36. Gunther J, Haas W, yon Boehmer H: Suppression ofT cell res~onses through competition for T cell growth factor (Interleukin 2). Eur J lmmunol 12:247, 1982. 37. Gershon RK: Immunoregulation circa 1980: some comments on the state of the art. J Allergy Clin Immunol 66:18, 1980. 38. Germain RN, Bennacerraf B: A single major pathway of T-lymphocyte interactions in antigen-specific immune suppression. Scand J Immunol 13:1, 1981. 39. Morimoto C, Reinerz EL, Borel Y, Schlossman S: Direct demonstration of the human suppressor inducer subset by anti-T cell antibodies. J Immunol 130:157, 1983. 40. Damle NK, Engleman EG: Immunoregulatory T cell circuits in man. Alloandgenprimed inducer T cells activate alloantigen-specific suppressor T cells in the absence of the initial antigenic stimulus. J Exp Med 158:159, 1983.