Eur. J. Immunol. 1979.9: 575-578 Lectin affinity fractionation of cytotoxic T lymphocytes 575

12 Takemori, T. and Tada, T., J. Exp. Med. 1975.142: 1241. 13 Kapp, J. A , , Pierce, C. W., Dela Croix, F. and Benacerraf, B., J.

14 Kontianen, S. and Feldmann, M., Eur. J. Immunol. 1977. 7: 417. 15 Liew, F. Y. and Chan-Liew, W. L., Eur. J . Immunol. 1978.8: 168. 16 Liew, F. Y., Eur. J. Immunof. 1977. 7: 714. 17 Littlefield, J. W., Science 1964. 145: 709.

18 Cotton, R. G. H., Secher, D. S. and Milstein, C., Eur. J.

19 Kohler, G. and Milstein, C., Eur. J. Immunol. 1976. 6: 511. 20 Pearson, T., Galfrt, G., Ziegler, A. and Milstein, C., Eur. J.

21 Luzzati, A. L., Taussig, M. J., Meo, T. and Pernis, B., J. Exp.

Immunol. 1973. 3: 135. Immunol. 1976. 116: 305.

Imnunol. 1977. 7: 684.

Med. 1976. 144: 573.

Arthur Kimura, Anders Om, Goran Holmquist, Hans Wigzell and Bo Ersson+

Department of Immunology, University of Uppsala Biomedical Center and Separation Central, University of Uppsala Biomedical Center’, Uppsala

Unique lectin-binding characteristics of cytotoxic T lymphocytes allowing their distinction from natural killer cells and “K” cells*

Cytotoxic T lymphocytes (CTL) can be selectively depleted from in vitro or in vivo alloactivated populations of T cells on Vicia villosa lectin adsorbents through the lectin-specific interaction with the CTLassociated surface glycoprotein T 145 (Kim- ura, A, Wigzell, H. and Holmquist, G., J . Exp. Med. 1979. 149: 473). Results from these and other experiments have demonstrated the general applicability of this frac- tionation procedure in which no constraints related to antigenic specificitiy of the CTL have been observed. When this fractionation procedure was applied to other compartments of cytolytic cells (natural killer cells and “ K ’ cells), no detectable impact could be seen. This differential lectin binding would appear to offer a means of dissecting the activities of CTL from other compartments of cytolytic lymphoid cells.

1 Introduction

The existence of at least three major tissue-destructive compart- ments of lymphoid cells has been extensively documented for a variety of species (see for review [l]). Cytotoxic T lymphocytes (CTL), natural killer cells (NK) and antibody-dependent “K” cells seemingly all contribute towards the maintenance of a versatile cellular immune effector mechanism. Although a great deal of experimental evidence has been reported to distinguish the cytolytic activities of these three categories of cell types as separate entities, the lineage(s) of both NK and K cells have not been firmly established [l]. Considerable effort has been put forth to identify cell surface antigenic systems which would clearly dissociate these three categories of cell types. Unfortu- nately, approaches in which a particular cell type is selected for by the use of antibody and complement elimination pose

[I 23861

* This work was supported by an NIH contract NOI-CB-64033, NIH grant AI-CA-13485-03 and the Swedish Cancer Society.

Correspondence: Arthur Kimura, Department of Immunology, Uni- versity of Uppsala Biomedical Center, Box 582, S-751 23 Uppsala, Sweden

Abbreviations: Con A: Concanavalin A ADCC: Antibody-dependent cellular cytotoxicity CTL: Cytotoxic T lymphocyte(s) D-G~NAc: N- acetyl-D-galactosamine FCS: Fetal calf serum K cell: Naturally occur- ring Fc receptor-positive cytotoxic cells mediating contactual lysis against IgG-coated target cells NK cell: Natural killer cell NP40 Nonidet-P 40 PBS: Phosphate-buffered saline MLC: Mixed lympho- cyte culture

0 Verlag Chemie, GmbH, D-6940 Weinheim, 1979

further problems in the creation of soluble immune complexes which can be shown to alter the activity of cells participating in antibody-dependent cellular cytotoxicity (ADCC) [2].

We have previously described a CTL-associated cell surface glycoprotein (T 145) whose expression and distribution among activated T lymphocytes is restricted to the Ly-1- 2’ popula- tion of alloactivated T blasts, being undetectable on their pre- cursors and other populations of activated T blasts lacking cytolytic function [3]. During the course of purification proce- dures for T 145, a panel of over 20 lectins were analyzed for their quantitative reactivity against various Nonidet-P 40-sol- ubilized, CTL-derived membrane glycoproteins as a first step towards purification. Unexpectedly, one blood group A-reac- tive lectin derived from the seeds of Vicia villosa was found to display an unusually selective binding to T 145 from among the total pool of surface-labeled glycoproteins [4]. When subse- quently applied to cell fractionation studies, Vicia villosa affin- ity adsorbents could be shown to selectively retain those T cells capable of expressing cytolytic activity [4]. In this study, we present further evidence for the discriminatory power of Vicia villosa by the comparison of its reactivity for other cellu- lar compartments of cytolytic activity.

2 Materials and methods

2.1 Animals

Male and female C B N H mice (6-8 weeks of age) used in these studies were raised and maintained in our breeding facilities, Department of Immunology, Uppsala.

0014-2980/79/0707-0575$02.50/0

Eur. J. Immunol. 1979.9: 575-578 576 A. Kimura, A. om, G. Holmquist, H. Wigzell and B. Ersson

2.2 Conditions for the generation of CTL

In vitro conditions for the allogeneic sensitization of “purified” T cells have been described in detail elsewhere [5]. In brief, spleen T cells, purified by passage through Ig-anti-Ig- coated glass bead columns [6], were incubated in culture with 2000 R irradiated allogeneic or trinitrophenylated (TNP) [7] syngeneic cells in HEPES-buffered Eagle’s Hanks’ amino acid-supplemented media (EHAA) with 0.5% normal mouse serum from the responding mouse strain [8 ] . After 5 4 days in culture, the cells were harvested and the activated T blasts purified by Ficoll density centrifugation [9]. In vivo generated alloimmune T cells were obtained from spleens of CBA/H mice 17-21 days after a single intraperitoneal (i.p.) injection of lo7 EL4 (H-2b) tumor cells. Spleen cells from these mice were then further purified by passage through Ig-anti-Ig affinity col- umns [6].

2.3 Cell-mediated cytotoxicity testing

T cell-mediated cytolysis against 51Cr-labeled target cells was performed in v-bottom Cooke microtiter plates as described earlier [5]. An in vitro maintained line of the C 57 BL/6 lym- phoma E L 4 (H-zb) and CBA/H concanavalin A (Con A) blasts served as target cells. Percent specific lysis was deter- mined after 4 h, as described by Cerottini and Brunner [lo].

Spontaneous cytotoxicity mediated by natural killer (NK) cells was measured against an in vitro maintained line of the H-2” lymphoma YAC-1. The cellular characteristics of this cell type have been described in detail [ l l] . Data are presented as per- cent specific lysis according to the formula and methods described by Kiessling et al. [12].

ADCC was measured against IgG antibody-coated chicken erythrocytes as described [ 131. Control release values included medium plus target cells, antibodies plus target cells and effec- tor cells plus target cells. 5 1 ~ r release from controls never significantly differed between groups and ranged from 3-5 % in our 4-h assays.

2.4 Preparation and use of lectin affinity adsorbents for cell fractionation studies

The A, specific lectin from Vicia villosa seeds was extracted and purified on human blood group A affinity adsorbents as described previously [4]. Vicia villosa lectin and twice recrys- tallized ovalbumin (Worthington Biochemical Corp., Freehold, NJ) were directly polymerized onto glutaraldehyde- activated Biogel P 2 or P 6 beads in 0.1 M phosphate buffer, pH 7.6, according to the method of Weston and Avrameas [14] with an extended coupling period of 72 h at 4°C. The beads were extensively washed with phosphate-buffered saline (PBS, pH 7.4 without Ca++ or Mgff), and remaining active groups were blocked by a 2-h room temperature incubation with 0.1 M &-aminocaproic acid/PBS, pH 7.0. The functional activity and degree of Vicia villosa coupling was determined by the binding of NaB3H4-labeled human blood group A substance [4].

Cell preparations to be fractionated on the affinity columns were washed twice in PBS and resuspended to a concentration of lo7 cells/ml in PBS plus 0.1% gelatin.Ovalbumin and Vicia villosa affinity adsorbents were packed in Bio-rad (Bio-Rad

Labs., Richmond, CA) Econo-columns (1 x 10 cm) and equilibrated in PBS plus 0.1% gelatin at least 1 h before use to minimize nonspecific adherence of cells. The cells were passed at room temperature with a flow rate of 0.5-1.0 ml/min, and unbound cells were washed through the column with excess PBS plus 0.1% gelatin. Bound cells were eluted by allowing the eluting buffer (0.1 M N-acetyl-D-galactosamine (D-Gal- NAc) in PBS plus 0.1% gelatin) to enter the gels and incubate for 5 min at room temperature, followed by gentle mechanical stirring of the beads with a pasteur pipette. The bound eluted cells were collected from the column at a flow rate of approxi- mately 3 ml/min, washed twice in PBS plus 5% fetal calf serum (FCS) and counted for use in the cytotoxicity assays.

3 Results

3.1 General applicability of Vicia villosa affinity adsorbents for the fractionation of CTL

We have previously shown that CTL, generated in a number of different mouse strain combinations, are selectively retained on Vicia villosa affinity columns [4]. The results shown in Table 1 further demonstrate that successful fractio- nation of the CTL compartment can occur without noticeable restraints posed by the antigenic specificity of the CTL. Here, CTL, generated against either allogeneic or modified self determinants, are selectively retained on Vicia villosa affinity adsorbents and can be recovered in a highly enriched func- tional form by the addition of the competitive sugar, D-Gal- NAc. The majority of the cells ( 5 7 4 8 % ) present after either type of activation pass through the column and display only marginal cytolytic activity.

3.2 Differential behavior of CTL vs. NK and K cells on Vicia villosa affinity adsorbents

Cell surface counterparts of the CTL-associated T 145, which allow the selective determination of cytolytic effector cells from their precursors, have not yet been identified for NK and K cells. It was thus of considerable interest to examine the behavior of NK and K cells on Vicia villosa affinity adsorbents, to determine whether the lectin-specified reactivity for CTL would be extended to other cellular compartments of cytolytic activity or constitute a selective reagent for CTL. The results of these experiments are shown in Fig. 1. As seen in Fig. 1 A, the cytolytic activity of in vivo alloactivated T cells can be fractionated in an identical fashion as the in vitro generated CTL shown in Table 1, yielding a noncytotoxic column-passed cell fraction and a highly active sugar-eluted population.

Repeated attempts to fractionate NK cell activity from either the naturally occurring levels found in normal spleen cell sus- pensions or the enhanced NK activity obtained shortly after interferon induction [15] have consistently shown the cytolytic activity of the Vicia villosa-passed cell fraction to be of either the same or slightly higher activity than the unpassed or con- trol column-passed cells. The results shown in Fig. 1 B with interferon-“activated” NK cells provides a more direct com- parison with the alloactivated CTL shown in Fig. 1A and offers greater sensitivity for the detection of significant differ- ences between the various cell fractions, and yet no fractiona- tion is seen.

Eur. J. Immunol. 1979. 9; 575-578

Table 1. Affinity fractionation of CTL induced by allogeneic or modified self activation

Lectin affinity fractionation of cytotoxic T lymphocytes 577

Group Effector T cell population")

I CBARI antic57 BU6

I1 CBA/H anti-CBARI-TNP

Affinity columnb)

None Ovalbumin

V. viUosa

None Ovalbumin

V. vilbsa

CeU fraction').

Unfractionated Unbound

Bound, eluted Unbound

Bound, eluted Unfractionated

Unbound Bound, eluted

Unbound Bound, eluted

Input cells recovered

(%I -

83.7 4.3

57.1 28.3 -

82.7 6.3

68.3 18.0

Percent specific Iysis at:d)

20: l 1 O : l 5 : l 2 .5: l

83.6 69.3 40.1 18.8 85.7 65.5 38.4 15.7 ND ND 37.2 16.5 4.3 1.2 0.5 0.7

87.3 85.4 73.1 46.5

18.4 14.7 10.3 7.1 20.2 16.4 12.0 8.6 ND 18.2 14.4 7.9 0.8 1.2 0.7 0.1

36.7 24.9 18.9 14.6

a) CBA/H T cells, used as the responding cell population, were purified by passage of spleen cells through Ig-anti-Ig columns and reacted in culture against 2000R-irradiated spleen cells of the appropriate type in EHAA tissue culture media plus 0.5% normal mouse serum.

b) Ovalbumin and Vicia villosa were coupled to glutaraldehyde-activated Biogel P 6 beads. Affinity adsorbents were packed into columns with an approximate bed volume of 1 x 10 cm, equilibrated in PBS plus 0.1% gelatin and used for the fractionation of 2 x lo7 MLC-derived lymphocytes.

c) MLC-derived lymphocytes were washed twice in PBS and resuspended to a cell concentration of 107/ml in PBS plus 0.1% gelatin. Cells were passed through the column at a flow rate of 0.5-1.0 mYmin in PBS plus 0.1% gelatin, and the unbound cells were collected. Cells bound to the column were eluted with 0.1 M D-GalNAcPBS plus gelatin and mechanical stirring of the beads. Cells were washed twice in PBS and adjusted to the appropriate concentration in medium plus 5% FCS for titration of cytolytic activity.

d) Cell-mediated cytotoxicity was determined at the various lymphocyte: target cell ratios in a 4-h assay against "Cr-labeled EL4 tumor cells (group I) or TNP-conjugated CBNH Con A blasts (group 11). Percent specific lysis was calculated according to Cerottini and Brunner [lo] with spontaneous release values of 6.3% (group I) and 18.4% (group 11). In group 11, percent specific lysis against unmodified CBNH ConA blasts was less than 3% at all ratios. ND = not determined.

6 1'2 25 5 0 : l 12 25 50 1OO:l 3 6 12 25:l Lymphocyte to target ratio

Figure 1 . Differential fractionation of T vs. NK and K cell cytolytic activity upon filtration through Vicia villosa affinity adsorbents. (A) T cell cytotoxic activity was assayed from CBA/H spleen cells obtained 17 days after i.p. injection of lo7 allogeneic EL4 tumor cells and further purified to isolate T cells by filtration through Ig-anti-Ig col- umns. (B) NK cell activity of CBNH spleen cells from mice given an oral dose of Tilorone (Lilly Research Laboratories) 16 h previously [15]. (C) K cell activity against IgG-coated chicken erythrocytes (CRBC) from spleens of normal CBAiH mice. Details for the condi- tions of cell fractionation through the affinity columns are as given in Table 1. (0) Cells passing through ovalbumin column, ( 0 ) cells bound and eluted from the ovalbumin column, (A) cells passing through the Vicia villosa column, (A) cells bound and eluted from the Vicia villosa column. Cytotoxicity was measured in a 4-h assay against "Cr-labeled (A) EL4, (B) YAC and (C) CRBC.

Finally, experiments of the type shown in Fig. 1 C have clearly shown that the dominant K cell activity found in the spleens of normal mice, as measured against IgG-coated erythrocytes, cannot be retained on Vicia villosa affinity adsorbents. The cellular complexity of the K cell system, as recently reviewed [l], has prevented an extensive testing of all other experimen- tal systems where K cell activity is expressed, and has limited our studies to the classical systems where K cells of mononuc- lear nature have been implied.

4 Discussion

Our results permit us to draw the conclusion that the T 145 glycoprotein, which allows the selective retention of CTL from heterogeneous populations of activated T cells through adher- ence to Vicia villosa affinity columns ([4] and this study), is unavailable for interaction on the surface of NK and K cells. Thus, filtration through such lectin-coated columns would seemingly provide a means of discriminating between cytolytic activities expressed by CTL vs. other cytotoxic compartments in situations where mixed effector cell populations may be implied. Although the actual participation (if any) of the T 145 surface molecules in the cytolytic behavior of the CTL has not been determined, the data presented here indicate that it, in itself, is not a necessary requirement for all mononuclear cells to be able to mediate contactual lysis.

The exellent technical assistance of Ms. Berit Olsson is gratefully ac- knowledged.

Received January 12, 1979.

578 R. F. Richter and H. Ambrosius Eur. J. Immunol. 1979. 9: 578-580

5 References 8 Peck, A. B. and Bach, F. H., J. Immunol. Methods 1973.3: 147. 9 Nabholz, M., Vives, J., Young, H. M., Meo, T., Miggiano, V.,

Rijnbeek, A. and Shreffler, D. C., Eur. J. Immunol. 1974.4: 378. 10 Cerottini, J.-C. and Brunner, K. T., Adv. Immunol. 1974. 18: 67. 11 Kiessling, R. and Wigzell, H,, Immunol, Rev, 1979. 44,.164, 12 Kiessling, R., Klein, E. and Wigzell, H., Eur. J. Immunol. 1975.5:

1 Perlmann, P. and Cerottini, J.-C., in Sela, M. (Ed.), The Antigens, Academic Press, New York 1979, vol. 6, in press.

2 fiessling, R., PetrBnyi, G., Karre, K., Jondal, M., Tracey, D. and Wigzell, H., J. Exp. Med. 1976. 143: 772. Kimura, A. K. and Wigzell, H., J, Exp. Med. 1978. 147: 1418. 3,-

1 IL . 4 Kimura, A., Wigzell, H., Holmquist, G., Ersson, B. and Carlsson,

5 Kimura, A. and Wigzell, H., Contemp. Top. Mol. Immunol. 1977.

6 Wigzell, H., Scand. J. Immunol. 1976.5: (Suppl. 5) 23. 7 Shearer, G. M., Eur. J. Immunol. 1974. 4: 527.

P., J. Exp. Med. 1979. 149: 473.

6: 209.

13 Perlmann, P. and Perlmann, H., Cell. Immunol. 1970. 1: 30. 14 Weston, p. D. and Avrameas, s.9 BiOChem. BiOPhYs. Res. Corn-

15 Gidlund, M., Om, A., Wigzell, H., Senik, A. and Gresser, I., mun. 1975. 45: 1574.

Nature 1978. 273: 759.

Roland F. Richter and Herwart Ambrosius

Section of Biosciences, Laboratory of Immunobiology , Leipzig

Anti-idiotypic antibodies of IgM-type produced in carp (Cyprinus carpi0 L.)

Carp synthesize highly specific anti-idiotypic antibodies of the IgM class. Anti-idioty- pic antibodies could be elicited in these animals by a human IgM myeloma protein and were detected in a passive hemagglutination assay. The agglutination was com- pletely inhibited by approximately 0.15 pg homologous antigen, whereas a 100000- fold excess of a heterologous IgM myeloma protein of the same L chain type did not produce any inhibition. A possible subgroup specificity of carp anti-idiotypic antisera can be excluded.

1 Introduction

Immunoglobulins (Ig) have structures which can be serologi- cally subdivided into iso-, allo- and idiotypes. While iso- and allotypic determinants are the structures of the constant part of Ig, the idiotypic determinants (idiotopes) are located in the variable portion of the antibody molecule and characterize individual antibody specificities [l].

Antibodies directed against idiotypic determinants are called anti-idiotypic antibodies (aId). It is possible to produce these aId antibodies both in xenogeneic and in allogeneic systems [2, 31. aId antibodies have also been elicited in the autologous system [4, 51. According to the network theory by Jerne [6], autologous aId antibodies regulate immunological reactions. It has indeed been possible to detect idiotype-specific, suppres- sive and stimulating effects of aId [7-91. However, such exper- iments have been camed out almost exclusively by using xeno- aId. The experiments conducted by Eichmann and Rajewsky have shown that aId IgGl obtained from guinea pigs had a stimulating effect in mice, whereas aId IgG2 caused idiotype- specific suppression [lo].

It is an open question whether this phenomenon reflects a physiological regulatory principle in the sense that different

[I 23751

Correspondence: Roland F. Richter, Laboratory of Imrnunobiology, Section of Biosciences, 701 Leipzig, TalstraBe 33, G.D.R.

Abbreviations: aid Anti-idiotypic (antibody) PBS: Phosphate-buf- fered saline H W S : Human whole serum HGG Human garnma-globu- lin GAD: Glutaraldehyde

0014-2980/79/0707-0578$02.50/0

classes of autologous aId perform different functions in the regulation of the immune response. Animal species having only one isotype, such as bony fish 1111, represent an interest- ing model for studying the problem of immunoregulation by aId. The antibodies of bony fish have many structural and functional features in common with the IgM of mammals. This offers a convenient opportunity to study the regulatory prop- erties of aId IgM. In order to deal with this problem, it has first to be established that it is generally feasible to produce aId with a high degree of specificity in a fish species [12].

2 Materials and methods

2.1 Myeloma proteins

One protein was isolated from the serum of Patient Ey with Waldenstrom’s macroglobulinemia. We obtained immuno- electrophoretically pure IgM by means of multiple euglobulin precipitation followed by gel chromatography on BioGel A- 1.5 m, 1W200 mesh (BioRad Laboratories, Richmond, CA). After reduction of the IgM paraprotein with 1% 2-mercap- toethanol and subsequent alkylation with iodoacetamide, only p and L chains were detectable in sodium dodecyl sulfate poly- acrylamide gel electrophoresis. From the serum of Patient St. with a double paraproteinemia, we obtained IgM by way of euglobulin precipitation, and, in addition, IgG by DEAE-cel- lulose (Whatman DE 52) column chromatography from the supernatant after concentration by ammonium sulfate precipi- tation. Both paraproteins were filtered through a Sephadex G- 200 column. The L chain type of the various myeloma proteins was determined with monospecific antisera (Behringwerke, Marburg, FRG) .

0 Verlag Chemie, GmbH, D-6940 Weinheim, 1979