Eur. J. Immunol. 1979. 9: 155-159

Christopher A. Sunderlando, W. Robert McMaster+and Alan F. Williams MRC Immunochemistry Unit, Department of Biochemistry, University of Oxford

Purification of leukocyte-common antigen

Purification with monoclonal antibody of a predominant leukocyte-common antigen and glycoprotein from rat thymocytes A leukocyte-common (L-C) antigen which can be dominant as an immunogen in rabbit anti-rat thoracic duct lymphocyte serum has been purified from rat thymocytes. Initially, an antigenic fragment of 100000 apparent mol. wt. was prepared at 400 to 900-fold purification by lentil lectin affinity chromatography and gel filtration in deoxycholate. Mice were then immunized with this fraction, and a hybrid myeloma cell line secreting antibody to the L-C antigen was prepared by cell fusion. This antibody was used for affinity chromatography and gave pure L-C antigen at 1400-fold puaification compared with thymocytes. The L-C antigen is a major membrane glycoprotein of rat thymocytes and has an apparent mol. wt. of 150000 as determined by electrophoresis on polyacrylamide gels in sodium dodecyl sulfate. The antigen constitutes one of the three thymocyte glycoproteins which stain intensely for carbohydrate with periodic acid Schiff stain. It is present on > 95 % of thymocytes, bone marrow cells and thoracic duct lymphocytes.

1 Introduction When rat thymocyte membrane is electrophoresed on polyacrylamide gels in sodium dodecyl sulfate and stained for carbohydrate, three major glycoprotein bands are seen at apparent mol. wts. of 150000, 84000 and 25000 [l]. The band at an apparent mol. wt. of 25000 is the Thy-1 antigen, now purified and chemically characterized [2]. The other two main glycoproteins are much less well characterized although it is known that the 150000 mol. wt. band binds to lentil lectin, while the 84000 band does not [1]. In the analysis of a number of rabbit anti-rat thoracic duct lymphocyte sera by quantitative radioimmunoassay, it was found that much of the antibody was directed against what appeared to be one antigen. This was called the leukocyte-common (L-C) antigen since it was present on thymocytes, bone marrow cells, peripheral lymphocytes and macrophages, but not in other tissues 131. The L-C antigen could be solubilized from rat thymocyte membrane with sodium deoxycholate and was bound by a lentil lectin affinity column. The mol. wt. of the antigen in deoxycholate was estimated to be 172000 on the basis of its hydrodynamic properties. If this is a major membrane molecule, as is suggested by its dominance in the above xenogeneic immunization, it could be the whole, or a part of, the major thymocyte .glycoprotein band seen on polyacrylamide gels with an apparent mol. wt. of 150000. A similar situation has been found in mouse thymocytes where rabbit anti-mouse thymocyte sera have been shown to immunoprecipitate two major bands from Nonidet-P40 detergent lysates of mouse thymocytes surface-labeled with 12’1 [4]. [I 21781 0 +

155

One of these bands with an apparent mol. wt. of 25 000 is the Thy-1 antigen and the other, termed T200, is a membrane glycoprotein of 200000 apparent mol. wt. In this study, we describe the purification of the L-C antigen from rat thymus, and show that this antigen makes up the 150000 apparent mol. wt. glycoprotein of rat thymocyte membrane.

2 Materials and methods* 2.1 Materials Deoxycholate from Koch-Light Labs., Colnbrook, Bucks., GB, was twice recrystallized from acetone: water (4 :1).Diisopropylfluorophosphate (iPr,P-F) was obtained from Sigma Chemical Co., St. Louis, MO. Pristane (2,6,10,14-tetramethyl pentadecane) was from Aldrich Chemical Co. Inc., Milwaukee, WI.RPMI 1640 tissue culture medium and fetal calf serum were from Flow Labs, Irvine, Scotland. - Wistar rats, BALB/c and D2.C (DBA/2 X BALB/c)F1 mice were purchased from Olac 1976 Ltd., Oxon, GB. Serum 4.2 is a rabbit serum raised against Ig-negative rat lymphocytes from the thoracic duct. Liver-absorbed serum 4.2 was used to follow L-C antigenic activity, as described in [3]. Rabbit antiserum to pure L-C antigen (see Sect. 3.3) was raised by immunizing twice at a one-week interval with 100 pg L-C antigen intramuscularly in complete Freund’s adjuvant. Seven weeks later, the animals were boosted with 100 pg L-C antigen intravenously. The rabbits were then bled weekly and the sera pooled. Purified horse F(ab’), anti-rabbit IgG antibody and rabbit F(ab’)2 anti-mouse IgG antibody were as in refs. [3] and [6].

Fellow of the Sidney Perry Foundation. W. R. M. holds an Exhibition of 1851 Scholarship.

Correspondence: A. F. Williams, MRC Cellular Immunology Unit, Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX 13 RE, GB Abbreviations: FACS: Fluorescence-activated cell sorter iPr2P-F Diisopropyl fluorophosphate L-C antigen: Leukocyte-common antigen PAS: Periodic acid Schiff (reaction) SDS-PAGE Polyacrylamide gel electrophoresis in sodium dodecyl sulfate 0 Verlag Chemie, GmbH, D-6940 Weinheim, 1979

2.2 Preparation and solubilization of membranes

Purified thymocyte membrane was prepared as described [ 11 with modifications. Thymocytes were teased into salt solution

* All procedures were performed at C - 4 “ C unless stated otherwise, and any methods not mentioned were as in ref. [5]. 0014-2980/79/0202-0155$02.50/0

156

Eur. J. Immunol. 1979. 9: 155-159

C. A. Sunderland, W. R. McMaster and A. F. Williams

and filtered, but not washed, prior to addition of Tween 40 to 2.5% with cells at 5 X 108/ml. The mixture was stirred and homogenized [l] and then layered onto 32% sucrose and centrifuged at 24000 rpm for 60 min at 4 " C in a Beckman SW 27 rotor. The membrane was removed from the sucrose interface and diluted with 0.01 M Tris HCl buffer, p H 8 prior to pelleting by centrifugation at 30000 rpm for 60 rnin in a Beckman Type 35 rotor. The pellets were resuspended to 3.3 X lo9 cell equivalents (3.3 mg protein) per ml by homogenization in 0.01 M Tns, pH 8.0, 0.003 M NaN,. Prior to solubilization, iPr,P-F was added to 10 mM, and for solubilization an equal volume of 4% deoxycholate plus 2.5 mM iodoacetamide in 0.01 M Tris, p H 8.0, 0.003 M NaN, was added. The mixture was homogenized, stirred on ice for 60 min and then centnfuged at 25 000 rpm for 90 min in a Beckman SW 27 rotor. The supernatant was removed, and further iPr,P-F was added to 5 mM after each preparative step. Lentil lectin affinity chromatography and gel filtration in deoxycholate was performed as described [5].

2.3 Polyacrylamide gel electrophoresisin sodium dodecyl sulfate (SDS-PAGE) SDS-PAGE was performed in a BioRad (Bromley, Kent, GB) slab gel apparatus using a 10% ( w h ) polyacrylamide separating gel and a 3% ( w h ) stacking gel. Gel thickness was 1.5 mm for protein staining and 3.0 mm for carbohydrate staining [l]. Samples at protein concentrations (given in figure legends) were either taken in an equal volume of 20% glycerol (viv) and 5 % SDS plus 2% dithiothreitol when reduction was required, or were concentrated by precipitation in 15% trichloroacetic acid for 30 min at 0 "C, washed twice with acetone at 0 "C to remove deoxycholate and trichloroacetic acid, and then solubilized in 8 M urea and 10% SDS plus 2% dithiothreitol when required. All samples were boiled for 5 min prior to loading. Gels were stained for protein with Coomassie Blue and for carbohydrate with the periodic acid Schiff (PAS) method as used in ref. [5]. The apparent mol. wts. were determined using the following marker proteins: chymotrypsinogen (26 OOO), ovalbumin (43 OOO), bovine serum albumin (68 000), phosphorylase (94000), @-galactosidase (130000) and myosin (200000).

Sect. 3.1) in complete Freund's adjuvant at an interval of 1 month. Six weeks later, a third 10-pg dose was given intravenously without adjuvant. Spleen cells from the immune mice were fused with the P3NSII1-Ag4-1 nonsecretor myeloma cell line (provided by Dr. C. Milstein) 3 days after the third injection. The methods used in the fusion and selection of hybrids were as described [6,8,9] except that the tissue culture medium was RPM I 1640 plus 10% fetal calf serum. The steps involved are briefly discussed in Sect. 3.4. The resultant cell line is called MRC OX 1 (L-C antigen), and the cells were grown in culture with ['"C] lysine to label the antibody [8]. Analysis of secreted material by SDS-PAGE showed that the heavy chains were of the same mol. wt. as y-chains suggesting the antibody to be of the IgG class. 2.6 Affinity chromatography using monoclonal antibody D2.C (DBA/2 X BALB/c)F1 mice were treated with pristane, and four weeks later given lo7 MRC OX 1 hybrid cells interperitonelly. Three weeks later, ascites fluid was collected and immunoglobulins were precipitated and washed with 16% Na,SO, and coupled to Sepharose 4 B. Coupling and use of the derivatized beads for affinity chromatography were performed as described [5]. The column used contained 10 ml of Sepharose 4 B with 10 mg of protein coupled/ml. 2.7 Labeling with fluorescent antibody and analysis in the

fluorescence-activatedcell sorter (FACS); saturating binding of antibody to thymocytes Thymocytes, bone marrow cells and thoracic duct lymphocytes were incubated with MRC OX 1 (L-C antigen) antibody, washed, incubated with fluorescein-conjugated rabbit F(ab), anti-mouse IgG antibody and washed again. These procedures and analysis of binding with a FACS I1 have been described [61. To measure saturating binding of MRC OX 1 , s X lo6 thymocytes were incubated with saturating amounts of antibody and washed by centrifugation. The amount of mouse Ig bound was then measured by inhibition of a radioimmunoassay for mouse Ig which was calibrated by inhibition with pure mouse IgG [6].

2.4 Indirect binding assays and units of antigenic activity

L-C antigen was assayed by inhibition of indirect radioactive binding assays, as previously described [3], using liver-absorbed serum 4.2 or rabbit anti-L-C antigen antiserum (see Sect. 2. I). To follow antigenic activity during purification, the assays were performed using trace amounts of second antibody to maximize sensitivity, while for analysis of the specificity of the sera, saturating second antibody was used [7]. A unit of antigenic activity is the amount of antigen needed for 50% inhibition of the standard trace binding assay.

3 Results 3.1 Preliminary purification of L-C antigen

Binding assays used in the production of monoclonal antibodies were as in ref. [6] except that glutaraldehyde-fixed thymocytes were used as target cells.

It has been previously shown [3] that the L-C antigen may be solubilized from thymocyte membrane with deoxycholate and that it binds to lentil lectin. In further studies (unpublished), the glycoproteins eluted from lectin were passed through Sephadex G-200 in deoxycholate, and the L-C antigenic activity was found to be associated with glycoproteins of 100000 to 150000 apparent mol. wt. These studies were done in the absence of proteolytic inhibitors, and it became obvious that proteolysis of the major 150000 apparent mol. wt. glyco-

2.5 Production of monoclonal antibody to L-C antigen*

* Cloned hybrid myelomas prepared in the MRC Cellular Im-

BALB/c mice were immunized subcutaneously with two 10-yg doses of the 100000 mol. wt. fraction of glycoproteins (see

munology Unit will be coded MRC OX followed by a number and the name of the antigen involved if appropriate. The anti-(L-C antigen) hybrid is thus celled MRCOX 1 (L-C antigen).

Eur. J. Immunol. 1979. 9: 155-159

Purification of leukocyte-common antigen

protein was occurring. If proteolytic inhibitors were added, the major glycoprotein was not degraded, but in gel filtration it chromatographed with a number of other more minor glycoproteins from which it could not be separated. Thus, the glycoprotein with L-C antigenic activity could not be identified. When autolysis was allowed to occur, o r when the glycoproteins were briefly digested with trypsin (for 10 min at 37°C without deoxycholate), a more homogenous glycoprotein fraction resulted after gel filtration on Sephadex G-200 in deoxycholate. This fraction retained L-C antigenic activity, and purifications of 400 to 900-fold were obtained. The higher purification occurred after the tryptic digest. This glycoprotein fraction was analyzed by SDS-PAGE, and the main band was found at 100000 apparent mol. wt. However, other minor bands were present, and it was obvious that by the above procedures the L-C antigen could not be purified completely either in the intact or partially degraded form.

3.2 Preparation of a monoclonal antibody to L-C antigen To obtain antibody of unequivocal specificity, a monoclonal antibody to L-C antigen was prepared by cell fusion techniques [6, 8, 91. Mice were immunized with the 100000 mol. wt. glycoprotein fraction which was approximately 400-fold enriched for L-C antigenic activity. Spleen cells from one of these mice were then fused with P 3-NSI/1-Ag4-1 nonsecretor myeloma cells using polyethylene glycol. The cells were diluted into 48 wells and hybrids selected by their ability to grow in HAT medium. Eleven wells grew hybrids secreting anti-thymocyte antibody, and, detected by indirect radioactive binding assays, the activity in all these wells could be inhibited by addition of the 100000 mol. wt. glycoprotein fraction. One well of consistently high activity and good inhibition was selected for cloning by serial dilution into soft agar. Single clones were picked from the dilution and grown. Approximately 5 % of these clones were positive in terms of secreting antibody binding to thymocytes. To ensure final purity, selected positive clones were recloned in soft agar, whereupon all the resulting clones secreted anti-thymocyte antibody. Cells from such a reclone, called MRC OX 1 (L-C antigen) were then grown extensively. Inhibition of binding by the 100000 mol. wt. glycoprotein fraction (900-fold purified for L-C antigen) was confirmed, though notably approximately 50 ng of antigen was

157

required for 50% inhibition as compared to 3 ng with rabbit anti L-C antigen serum. Pure L-C antigen (see Sect. 3.3) also showed these different sensitivities of inhibition when binding assays with monoclonal antibody or rabbit antibodies were compared.

To provide large amounts of antibody for affinity chromatography, MRC OX 1was grown as tumors in mice. 3.3 Purification of the L-C antigen using a monoclonal antibody column Purified thymocyte membrane was solubilized in 2 % deoxycholate in the presence of 5 mM iPr,P-F and 1.25 mM iodoacetamide to minimize proteolytic degradation. The extract was then applied to a column containing Sepharose 4 B MRC OX 1 antibody. Less than 5 % of the antigenic activity passed through the column. The column was eluted with 0.05 M diethylaminde-HC1, p H 11.5, 0.5% deoxycholate buffer giving a 50% yield of antigenic activity (Table 1). After addition of further iPr,P-F ( 5 mM), the eluate was concentrated and run on a Sephadex G-200 column in 0.5% deoxycholate buffer. The results of these procedures are shown in Table 1 and Fig. 1. In the antibody column step, a purification of 100-fold was obtained for L-C antigen, and the final purification was 1400-fold relative to lysed cells. After SDS-PAGE, it is evident that the purified preparation contained only one clear protein band of 150000 apparent mol. wt. (Fig. 1A). The PAS-stained gel (Fig. 1B) shows a complete depletion of the 150000 band after passage of the extract through the column, leaving only a faint band of slightly higher apparent mol. wt. The antigenic activities in Table 1 were determined by inhibition of binding assays with rabbit antibody raised against the

Table 1. Purification of L-C antigen with a monoclonal antibody column")

Material

Total protein

(mg) Lysed cells 8880 Purified membrane 305 Deoxycholate extract 200 Antibody column eluate 1.1 After Sephadex G-200 1.2

Total units Yield of antigenic genic activity activity x (%)

of anti-

Relative specific activity

3.70 1.25 1.20

100 33 33

1 10 14

0.62 0.70

17 19

1080 1300

a) Data are from one representative experiment. Units of antigenic activity were assayed using rabbit anti-L-C antigen serum.

Figure I . Affinity chromatography with an MRCOXl (L-C antigen) column. Purified thymocyte membrane extract, prepared in the presence of proteolytic inhibitors, was passed through an MRCOXl antibody-Sepharose 4 B affinity column. The column was washed with deoxycholate buffer and eluted with 0.05 M diethylamine, pH 11.5, in 0.5 % deoxycholate buffer. The eluate was concentrated and chromatographed on Sephadex G-200 in deoxycholate. SDS-PAGE was carried out on the following reduced samples: (I): deoxycholate extract of membrane; (11): extract passed through the column; (111): material eluted from the column; (IV): eluted material after Sephadex G-200 chromatography. Gel A was stained with Coomassie Blue and Gel B with PAS reaction. Amounts loaded were for Gel A: 40 yg for (I) and (11), 4 yg for (111) and (IV); for Gel B: 180 bg for (I) and (11), 7 yg for (111) and (IV). Samples (I) and (11) were precipitated by trichloroacetic acid and acetone-extracted to remove deoxycholate.

Eur. J. Immunol. 1979, 9: 155-159

C. A. Sunderland, W. R. McMaster and A. F. Williams

158

pure 150000 mol. wt. glycoprotein. This was done because a sensitive assay is needed if all the fractions are to be accurately assayed. The assay with monoclonal antibody was much less sensitive (see Sect. 3.2). To check that the different sera were not detecting determinants on different molecules, binding assays using monoclonal antibodies and liver-absorbed serum 4.2 were also inhibited with membrane and the pure glycoprotein. The purification factor obtained was the same as shown in Table 1 using the rabbit anti-L-C antigen serum.

To determine the maximum amount of antibody which could bind to thymocytes, the cells were incubated with saturating amounts of MRC O X 1 antibody, washed and suspended in Triton X 100. The activity of the cell extracts was then measured by inhibition of a radioimmunoassay for mouse Ig. The assay was calibrated with pure mouse IgG. At saturation, 48 000 molecules of MRCOX 1 antibody were bound per cell.

4 Discussion Purification of glycoproteins from membranes as complex as those of thymocytes using the presently established techniques of lectin affinity chromatography and gel filtration is likely to be possible only in a few special cases. Thy-1 antigen was easily purified because of its low mol. wt. [2], but the large thymocyte glycoproteins which bind to lentil lectin could not be resolved by gel filtration. The possibility of using monoclonal antibody columns therefore represents a considerable advance. Here, in a preparation taking in total three days, the L-C antigen was obtained intact and pure with the M R C O X l antibody column giving a 100-fold purification in one step. The yield of antigen from the column was very good, and the eluted fraction had much less contaminating material than had previously been found by purification of Thy-1 antigen using a rabbit anti-Thy-1 antibody column [5]. The favorable elution may be due to the fact that the monoclonal antibody recognizes only one determinant per molecule as distinct from conventional antibodies where more than one determinant could be involved. Also, the monoclonal mouse antibody may have a lower affinity than hyperimmune rabbit antibody since 15fold more antigen was needed for inhibition of binding assays with MRCOXl antibody compared with rabbit anti-L-C antigen antibody.

1 10-6

0

TO-~O

10-9 10-8 1u7 Absorbing protein ( g per assay)

Figure 2. Antisera analyzed with the purified L-C antigen. Serum 4.2 (0-0) and rabbit anti L-C antiserum (A-A) were analyzed for inhibition of their thymocyte binding activity with antibody columnpurified L-C antigen. The antigen at various dilutions (75 $) was pre-incubated with 75 ~1 antiserum, centrifuged, and S O 4 supernatant incubated with 5 X lo6 glutaraldehyde-fixed thymocytes as targets. The cells were washed, and 1251-labeled F(ab’)* horse antirabbit IgG was added in saturating ( S O PI of 54 pgiml and 0.2 LCiipg) amounts in a second incubation. Specific binding is defined as the binding which can be inhibited by pre-incubation of antibody with purified thymocyte membrane. Dilutions of the antisera used were serum 4.2, 1:150 and anti-L-C antigen serum at 1 :200.

3.4 Analysis of antisera To confirm that the purified antigen was the same as that to which a large part of the original rabbit anti-thoracic duct lymphocyte serum (4.2) was directed, this serum was inhibited by the purified L-C antigen in a radioimmunoassay and compared with the inhibition of rabbit anti-L-C antigen serum. Total antibody activity was defined by inhibition with thymocyte membrane. Fig. 2 shows that 50% of serum 4.2 was inhibited by the pure L-C antigen compared with 100% inhibition of the binding of antibody from rabbit anti-L-C antigen serum.

3.5 Labeling of lymphoid cells with MRC OX 1 Originally, absorption studies of liver-absorbed serum 4.2 showed L-C antigen to be present on thymocytes, bone marrow cells, peripheral lymphocytes and macrophages, but not on liver, brain, kidney and red blood cells [3]. Since the specificity of MRC OX 1 antibody was assured, it could be used with confidence to label L-C antigen on the surface of cells. Thymocytes, bone marrow cells and thoracic duct lymphocytes were incubated with MRC O X 1 antibody and then with fluorescein-conjugated anti-mouse IgG antibody. Binding was analyzed using a FACS 11. In each case, > 95 % of cells were labeled; the labeling profiles were quantitatively heterogeneous (results not shown).

The purifications show that the L-C antigen is a major membrane glycoprotein of rat thymocytes. It has an apparent mol. wt. of 150000 and makes up the major PAS staining band of thymocyte membrane at this position. From the purification of 1400-fold it can be calculated that there are 70000 molecules of L-C antigen per thymocyte. This agrees quite well with the value of 48000 molecules of antibody bound per thymocyte in saturating binding studies with MRCOX 1 antibody. The value from antibody binding would be an underestimate if one antibody bound to two molecules of antigen (assuming a single antigenic determinant per molecule). At 70 000 molecules/ cell, there is nummerically much less L-C antigen than Thy-1 antigen (about 600 000 moleculesicell [2]) in rat thymocytes. However, the mol. wt. of the Thy-1 antigen is only 18000 [lo] and, therefore, on the basis of protein weight, these two glycoproteins are likely to constitute roughly the same proportion of the membrane. This is supported by the staining for protein of thymocyte glycoproteins after SDS-PAGE where the L-C antigen band is of similar intensity to that of Thy-1 antigen [l]. Large glycoproteins have also been studied on mouse lymphoid cells, and it seems likely that the rat L-C antigen is equivalent to mouse T 200 glycoprotein studied by Trowbridge and co-workers [4, 111. A monoclonal antibody has been produced to this molecule, and use of this antibody in immunoprecipitation studies clearly showed that T200 is a L-C antigen with an apparent mol. wt. of 190000 on thymocytes [12]. A large thymocyte surface glycoprotein of about 170 000 apparent mol. wt. has also been defined by surface labeling of

Eur. J. Immunol. 1979. 9: 15%165

Mycoplasma in cytotoxicity assays

carbohydrate [13]. Similarly, in other studies, a protein of 200000 apparent mol. wt. has been identified using a monoclonal antibody produced by a hybrid cell line prepared by cell fusion of spleen cells from rats immunized with mouse lymphocytes [14]. Estimates of mol. wt. after SDS-PAGE are inaccurate for large molecules and glycoproteins, and it is possible that the same molecule is being detected in all these cases. However, this may not be so. In the course of purifying rat glycoproteins, a complexity of large glycoproteins was revealed, and in particular the monoclonal antibody column reveals a glycoprotein which is slightly larger than the L-C antigen but not related antigenically (Fig. 1). We thank Dr. Cesar Milstein for the NSI myeloma cell line and for invaluable advice on cell fusion and cloning. Received June 23, 1978; in revised form September 25, 1978.

C. G. Brooks', R. C. Rees+' and R. H. Leachn Cancer Research Campaign Laboratories, University of Nottingham,Nottingham' and Mycoplasma Reference Laboratory, Public Health Laboratory Service, Norwichm

5 References 1 Standring, R. and Williams, A.F., Biochim. Biophys. Acta 1978. 508: 85. 2 Williams, A.F., Letarte-Muirhead, M., Barclay, A.N. and Morris, R.J., ColdSpring HarborSymp. Quant. Biol. 1976. 41:51. 3 Fabre, J. W. and Williams, A. F., Transplantation 1977. 23: 349. 4 Trowbridge, I. S., Weissman, I. L. and Bevan, M. J., Nature 1975. 256: 652. 5 Letarte-Muirhead, M., Barclay, A. N. and Williams, A. F., Biochem. J. 1975.151:685. 6 Williams, A. F., Galfre, G. andMilstein, C., Cell 1977. 12:663. 7 Morris, R. J. and Williams, A. F., Eur. J. Immunol. 1975. 5: 274. 8 Kohler, G. and Milstein, C., Eur. J. Immunol. 1976. 6: 511. 9 Galfre, G., Howe, S. C., Milstein, C., Butcher, G. W. and Howard, J. C., Nature 1977.266: 550. 10 Kuchel, P. W., Campbell, D. G., Barclay, A. N. and Williams, A. F., Biochem. J . 1978.169:411. 11 Trowbridge, I. S., Nilsen-Hamilton, M., Hamilton, R. J. and Bevan, M. J., Biochem. J. 1977.163: 21 1. 12 Trowbridge, I. S., J. Exp. Med. 1978. 148: 313. 13 Gahmberg, C. G., Hayry, P. and Anderson, L. C., J . Cell. Biol. 1976. 68: 642. 14 Springer, T., GalfrC, G., Secher, D. S. and Milstein, C., Eur. J. Immunol. 1978. 8: 539.

High nonspecific reactivity of normal lymphocytes against rnycoplasma-infected target cells in cytotoxicity assays* Several rat tumor cell cultures were deliberately infected with three species of mycoplasma commonly found as contaminants of cell lines grown in vitro, and the effect of mycoplasma infection on the results of cytotoxicity assays was examined. Lymph node cells and spleen cells from normal animals showed an apparently high spontaneous cytotoxic activity against tumor cells infected with either M. arginini or M.hyorhinis, but the reactivity against cells infected with M.orale was not significantly higher than that against uninfected cells. The high reactivity towards tumor cells infected with M. arginini and M . hyorhinis bore a close resemblence to natural cellmediated immunity in that spleen cells were much more reactive than lymph node cells, spleen cells from nude mice were as effective as spleen cells from normal mice, and the reaction crossed both strain and species barriers. However, closer examination revealed that the cytotoxic effects were directly caused by depletion of arginine or other essential nutrients from the medium. These findings imply that a cautious approach should be taken when interpreting certain aspects of spontaneous cell-mediated cytotoxicity, and that the greatest care be taken to ensure that the cells used as targets in any cytotoxicity test are mycoplasma-free.

[I 22011

* This work was generously supported by the Cancer Research Cam0

159

paign of Great Britain. Pre\ent address: Department of Virology, University of Sheffield Medical School, Beech Hill Road, Sheffield 10, GB

Correspondence: C. G. Brooks, Cancer Research Campaign Laboratories, University of Nottingham, University Park, Nottingham, GB Abbreviations: CS: Calf serum FCS: Fetal calf serum ['251]dUrd Labeled iododeoxyuridine LNC: Lymph node cell(s) MEM-H Minimum essential medium supplemented with HEPES buffer NK: Natural killer NS: Not significant PBS: Phosphate-buffered saline SpC: spleen cell(s) 0 Verlag Chemie, GmbH, D-6940 Weinheim, 1979

1 Introduction Following the demonstration by Takasugi et al. [l] that leukocytes from healthy individuals were reactive against cultured tumor cells in a microcytotoxicity assay, there has been a growing interest in natural cell-mediated immunity (recently reviewed by Herberman and Holden [2]). The phenomenon is widespread, being observed in a variety of cytotoxicity tests, with normal lymphocytes from every species of animal examined, and with a multitude of syngeneic, allogeneic, and xenogeneic target cells. Although there is no doubt that many examples of such normal lymphocyte reactivity are of real biological interest, as emphasized by the recent studies 0014-2980/79/0202-0159$02.50/0