Eur. J. Immunol. 1989.19: 1153-1155

Short paper

I n vivo target cell killing by cytotoxic T cells 1153

Yang h a , Paul Waring and Arno Miillbacher

Alloreactive cytotoxic T cells induce DNA fragmentation in peritoneal macrophages: evidence for target cell killing by cytotoxic T cells in vivo Division of Cell Biology,

The John Curtin School of Medical Research, The Australian National University, Canberra and Immunobiology Section,

New Haven

This report addresses the question whether cytotoxic T cells can cause target cell death in vivo by examining target cell DNA fragmentation. The results show that alloreactive cytotoxic T cells induced significant DNA fragmentation in peritoneal macrophages in vivo and that the DNA fragment was a multiple of 180 f 30 bp. Furthermore, the effector cells which caused this characteristic DNA fragmentation were CD8’ T cells. These results demonstrate that cytotoxic T cells can cause target cell death in vivo.

Of Medicines,

1 Introduction

Cytotoxic T (TJ cells, defined by their capacity to kill specific target cells in vitro, have been shown to be important effector cells in allograft rejection [ l , 21, viral clearance [3-51, virus- induced immunopathology [6] and tumor rejection [7, 81. However, the mechanisms of T, cell action in vivo still remain controversial [9]. As well as causing target cell lysis in vitro, T, cells have been shown to secrete a variety of lymphokines, such as interleukin 2 [lo], interferon-y [ l l ] and lymphotoxin [12], which may partially or fully account for the above-men- tioned effects. No direct evidence is available as to whether T, cells can kill target cells in vivo.

Recent studies have suggested two possible but not necessarily mutually exclusive mechanism by which T, cells kill target cells. First, T, cells may deliver cytoplasmic granule compo- nents such as perforin to target cells. Perforin has been shown to form channels through the target cell membrane and cause target cell lysis in a way similar to that induced by the activated complement (C) [13]. This interpretation, however, has been challenged by the finding that antiserum specific for the cyto- lytic granule components is unable to block target cell lysis by T, cells [14]. Furthermore, the absence of these granules in sensitized, highly cytotoxic peritoneal T cells [15] and the mor- phological findings [ 161 that target nuclear disintegration pre- cedes target cell lysis are also incompatible with the notion that exocytosis of granule substance are responsible for target cell lysis. Second, T, cells have been shown to induce so-called programmed cell death or apoptosis in target cells [17, 181. One characteristic of this kind of cell death is the degradation of DNA into discrete fragments of 150-180 bp, apparently by internucleosomal cleavage following induction or activation of a specific endonuclease [19]. The distinctive pattern of target cell DNA fragments caused by T, cells allows an analysis of T, cell-induced target cell death in vivo. Such an analysis is not possible using the conventional T, cell assay based on isotope release from target cells.

[I 74451

Correspondence: Yang Liu, Immunobiology Section, Yale University School of Medicine, 310 Cedar Street, New Haven, CT 06510, USA

Abbreviations: C: Complement mAb: Monoclonal antibody(ies) T, cells: Cytotoxic T cells

This report presents evidence that alloreactive T, cells induce a characteristic pattern of DNA fragmentation in peritoneal macrophages in vivo. These results reveal that T, cells kill target cells in vivo.

2 Materials and methods

2.1 Mice

CBA/H or BALB/c mice were bred in the animal breeding establishment in the John Curtin School of Medical Research under specific pathogen-free conditions. Adult female mice were used at the age of 6-8 weeks.

2.2 The generation of alloreactive T, cells

Alloreactive T, cells were generated in vitro, as has been described [20]. Briefly, splenocytes from CBA/H (H-2k) or BALB/c (H-2d) mice were irradiated (2000 rad from a 6oCo source) and were used as stimulators. Naive splenocytes from BALB/c or CBAM mice were used as responders. Stimulator cells and responder cells were mixed at 1 : 5 ratio and cultured (at a density of lo6 cells/ml) in Eagle’s minimum essential medium (EMEM) containing nonessential amino acids, 5% fetal calf seurm and 2-mercaptoethanol M) at 37°C in an humidified atmosphere containing 5% C 0 2 for 4 days. At the end of the culture, dead cells were removed by Ficoll (Phar- macia, Piscataway, NJ) separation. Viable cells were sus- pended in serum-free EMEM prior to the in vitro or in vivo T, cell assays.

2.3 Induction of DNA fragmentation in peritoneal macrophages

Peritoneal macrophages were elicited by i.p. injection of 3% (w/v) thioglycollate medium (Difco, Detroit, MI). Four days after thioglycollate injection, 2.5 x lo7 T, cells were injected i.p. in 1 ml EMEM. Two hours later, peritoneal cells were recovered by peritoneal washing with 10 ml of ice-cold Puck’s solution. Peritoneal cells were washed once with Puck’s solu- tion by centrifugation at 1500 rpm at 4°C for 5 min and 3 x lo6 cells were resuspended in 0.5 ml of ice-cold phosphate-buf- fered saline (PBS).

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1989 OO14-2980/89/0606- 1 153 $02.50/0

1154 Y. Liu, P. Waring and A. Miillbacher Eur. J. Immunol. 1989.19: 1153-1155

2.4 Preparation and analysis of cellular DNA from peritoneal cells

Harvested peritoneal cells (3 x lo6 in 0.5 ml cold PBS) were lysed by addition of 0.5 ml of neutral lysing buffer (0.2 M NaCl, 10 mM EDTA, 20 ml Tris-HC1, 1% sodium dodecyl sulfate, pH 8.0) and 50 p1 pronase solution (10 mg/ml, Cal- biochem-Behring, La Jolla, CA). After a 16-h incubation at 37 "C, the lysate was extracted using phenol and chloroform and the DNA precipitated from the acqueous solution using ethanol (5 vol) at -70°C for 6 h. DNA was centrifuged at 12 000 x g for 30 min, washed 3-4 times with ice-cold ethanol and air-dried. Samples were rehydrated in 10 mM EDTA and treated with 8 units of RNase (Sigma, St. Louis, MO) for 4 h at 37°C prior to applying them to 1.5% agarose gel. Elec- trophoresis using Tris acetate buffer, pH 8, containing 50 pg/ml ethidium bromide was conducted in a submarine apparatus (Pharmacia) at 80 mA for 5-7 h. The gel was photographed under UV light (Transilluminator, Ultra-Violet Products Inc., San Gabriel, CA; Polaroid film # 57). Relative size of the frag- ments was determined using standard containing DNA frag- ments with sizes of 1305, 517, 396, 220, 150 and 75 bp.

2.5 Cytotoxicity assay and antibody treatment of T, cells

Cytotoxicity of either macrophages or alloreactive T, cells was determined using a conventional 'lcr-release assay [21]. T, cells or peritoneal macrophage were used as targets. The dura- tion of the assay was 6 h. Alloreactive T, cells were treated with C ' alone or C ' plus rat anti-mouse Ly-2 monoclonal antibody (mAb; originally provided by F. Fitch, University of Chicago, Chicago, IL) as described [20].

3 Results and discussion

The activity and specificity of the T, cells generated in vitro was tested using a conventional 51Cr-release assay. Fig. 1 shows that T, cells caused specific lysis of macrophage targets. Anti-H-2k (BALB/c anti-CBA/H) T, cells lysed CBAM (H-2k) macrophages but not BALBk (H-zd) macrophages, and vice versa, a11ti-H-2~ (CBAM anti-BALB/c) T, cells lysed BALB/c macrophages but not CBA/H macrophages. These T, cells were then injected i.p. into mice which had received thiogly- collate medium 4 days earlier. The cellular DNA of the peritoneal cells from these mice was analyzed using agarose gel electrophoresis.

Significant DNA fragmentation was observed in peritoneal cells from the mice which had received alloreactive T, cells specific for recipient major histocompatibility complex anti- gens but not from the mice which had received syngeneic T, cells or no effector cells (Fig. 2). Thus, significant DNA frag- mentation was observed in peritoneal cells from two CBA/H (H-2k) mice which had received a11ti-H-2~ T, cells (lanes 11 and 12) but not in peritoneal cells from three individual CBA/ H mice which received a11ti-H-2~ T, cells (lanes 3-5) or two individual CBA/H mice which received no cells (lanes 1 and 2). Similarly, DNA fragmentation was observed in peritoneal cells from three individual BALB/c (H-2d) mice which had received anti- H-2d T, cells (lanes 8-10) but not in the peritoneal cells from two individual BALB/c mice which received syngeneic anti-H- 2k T, cells (lanes 13 and 14) or no effector cells (lanes 6 and 7).

1

-20 4 -0 5 0 5 1 5 2 5 3 5

In E.T

Figure 1. Specific lysis of BALB/c (H-2d; solid circles) or CBA/H (H-2k; open circles) macrophage targets induced by a11ti-H-2~ (unbro- ken lines) and at1ti-H-2~ (broken lines) T, cells. All points shown were means of triplicates with SE <3.2%.

The sizes of those DNA fragments were approximate mul- tiples of 180 k 30 bp.

An experiment was performed to test whether the DNA frag- mentation observed was induced by T, cells during in vitro isolation of the DNA. Anti-H-2k T, cells were mixed with CBAM peritoneal cell at 4°C and these cell mixtures were processed in the same way. No DNA fragmentation was observed under these conditions (Fig. 2, lane 15). Therefore, the DNA fragmentation observed must have been induced intraperitoneally . The possibility that macrophages lysing T, cells by a mecha- nism of backward-killing would be responsible for inducing DNA fragmentation was addressed by labeling either T, cells or macrophages using them as target cells in an in vitro

Figure 2. T,-induced DNA fragmentation in peritoneal cells as deter- mined by agarose gel electrophoresis. Lanes 1 and 2, peritoneal cells from two individual CBA/H mice which received no cells; lanes 3-5, peritoneal cells from three individual CBA/H mice which received 2.5 X lo') a11ti-H-2~ T, cells; lanes 6 and 7, peritoneal cells from two individual BALB/c mice which received no cells; lanes 8-10, peritoneal cells from three individual BALBlc which received 2.5 x lo7 anti-H-2d T, cells; lanes 11 and 12, peritoneal cells from CBA/H mice which received 2.5 X 10' anti-H-2k T, cells; lanes 13 and 14, peritoneal cells from two individual BALB/c mice which received anti-H-2k T, cells; lane 15, peritoneal cells from CBA/H mice mixed with anti- H-2k T, cells (1 : 1) at 4°C. Relative size of the five lowest bands in lanes 8-12 were: 170 bp, 320 bp, 485 bp, 680 bp, 900 bp, approxi- mate multiples of 180 k 30 bp.

Eur. J . Immunol. 1989.19: 1153-1155

48 -

38 - z . (0

v) 28 - z 0 'c

a

._ - ._ : 18-

v ) .

8 -

In vivo target cell killing by cytotoxic T cells 1155

Thus, the data presented above indicate that alloreactive T, cells cause target cell DNA fragmentation in vivo and that the effector cells are conventional CD8+ T, cells. As this DNA fragmentation pattern is a characteristic in vitro manifestation of target cells killed by T, cells, these results demonstrate that T, cells kill target cells in vivo.

The proposition that T, cells can kill target cells in vivo is consistent with several lines of indirect evidence. First, mor- phologically defined apoptotic cells have been observed in vivo during responses which are presumed to be mediated by T, cells, such as rejection of pig liver allograft, graft-vs.-host reactions, fixed drug eruptions and active chronic hepatitis [22]. Second, in a variety of cytotoxic T lymphocyte-mediated viral clearance or virus-induced immunopathology , correla- tions have been repeatedly documented between T, cell- induced target cell lysis in vitro and T, cell function in vivo [3, 4, 231. Two questions should be addressed before the mecha- nism of action of T, cells in vivo can be resolved: (a) can T, cells induce target cell lysis in vivo ? (b) Is target cell killing by T, crucial for T, cell function in vivo? The studies reported here have addressed the first question and provided a basis to address the second question.

-2

In E:T

Figure 3. The cytotoxicity of anti-H-2k T, cells (unbroken lines) and CBA/H (H-2') macrophages (broken lines) as tested on the T, cell (open circles) or the macrophage (solid circles) targets. All points shown were means of triplicates with SE <2.7%.

50

40

'E30 v)

-1 0 --I

1 2 3 4 5 In E:T

Figure 4. Phenotype of the effector cells which caused target cell DNA fragmentation. (a) At1ti-H-2~ T, cells were treated with anti- CD8 mAb plus C ' (open circles) or C ' alone (solid circles) and the T, cell activity tested on CBNH peritoneal macrophage targets. (b) Anti- H-2k T, cells (2.5 X 10') treated with c ' or anti-CD8 mAb plus c' were injected i.p. into CBA/H mice. Peritoneal cells from individual CBA/H mice which received C '-treated (lanes 3-6), antLCD8 mAb plus C '-treated (lanes 7-10) T, cells or no cells (lanes 1 and 2) were harvested and their DNA was extracted and analyzed by agarose gel electrophoresis.

cytotoxicity assay. T, cells were not lysed by macrophages, whereas macrophages were lysed by T, cells (Fig. 3). Thus it is unlikely that the DNA fragments observed were of T, cell rather than of macrophage origin.

To phenotype the effector cells which induced macrophage DNA fragmentation in vivo, the primary anti-H-2k T, cells were treated with rat anti-mouse CD8 mAb plus C ' . This treatment abolished T, cell cytotoxicity in vitro (Fig. 4a). Anti- H-2k T, cells treated with either anti-CD8 mAb plus C ' or C ' alone were injected into CBA/H mice i.p. and the genomic DNA fragmentation in the peritoneal cells from these mice was analyzed using agarose gel electrophoresis. Fig. 4b indi- cates that anti-CD8 mAb plus C ' prevented DNA fragmenta- tion. Hence, the effector cells which induced DNA fragmenta- tion in vivo and caused 51Cr release from target cells in vitro were CD8' T cells.

We thank Allan Sjaarda for technical assistance.

Received February 13, 1989.

4 References

1 Berke, G., Prog. Allergy 1980. 27: 69. 2 Cerottini, J . C. and Brunner, K. T., Adv. lmmunol. 1974.18: 67. 3 Kees, U. and Blanden, R. V., J. Exp. Med. 1976. 143: 450. 4 Yap, K. L., Ada, G. L. and McKenzie, I. F. C., Nature 1978.273:

5 Oldstone, M. B. A., Blount, P., Southern, P. J. and Lampert, P.

6 Zinkemagel, R. M., Haenseler, E., Leist, T., Cerny, A., Hengart-

7 Germain, R. N. , Dorf, M. E., Benacerraf, B., J. Exp. Med. 1975.

8 Schrader, J . W. and Edelman, G. M., J. Exp. Med. 1976.143: 601. 9 Mullbacher, A. and Ada, G. L., Microbial Pathogenesis 1987. 3:

10 Andrus, L., Granelli-Piperno, A. and Reich, E., J. Exp. Med.

11 Lin, Y. L. and Askonas, B. A,, J. Exp. Med. 1981. 154: 225. 12 Paul, N. and Ruddle, N. H., Annu. Rev. Immunol. 1988. 6: 607. 13 Young, J. D.-E. and Cohn, Z. A., Adv. Immunol. 1987. 41: 269. 14 Goldstein, P., Nature 1987. 327: 12. 15 Berke, G. and Rosen, O., J. Immunol. 1988. 141: 1437. 16 Sanderson, C. J . , Biol. Rev. 1981. 56: 153. 17 Russell, J. H., Masakowski, V., Rucinsky, T. and Philips, G.,

18 Duvall, E. and Wyllie, A. H., Immunol. Today 1986. 7: 115. 19 Ucker, D. S., Nature 1987. 327: 62. 20 Miillbacher, A., Ashman, R. B. and Ada, G. L., Scand. J. Immu-

21 Yap, K. L. and Ada, G. L., Immunology 1977.32: 151. 22 Searle, J . , Kerr, J. F. P. and Bishop, C. J., Pathol. Annu. 1982.17:

23 Luckacher, A. E., Braciale, V. L. and Braciale, T. J., J. Exp.

238.

W., Nature 1986. 221: 239.

ner, A. H. and Aithage, A., J. Exp. Med. 1986. 164: 1075.

142: 1023.

315.

1984. 159: 647.

J. Immunol. 1981. 128: 2087.

nol. 1984. 19: 365.

229.

Med. 1984. 160: 814.