Immunol. Cell Biol. (1990)68, 277-284
Production of a monoclonal allo-antibody to murine natural cytotoxic cells Y. C Smart, K. L. Stevenson, M. L. Farrelly, J. H. Brien and R. C. Burton Discipline of Surgical Science. Faculty of Medicine, University of Newcastle. NSW 2308. Australia (Submitted
15 May 1990. Accepted for publication 27 July 1990.)
Summary A mouse IgGi monoclonal antibody (1C4), which recognizes a cell surface molecule on murine natural cytotoxic{NC) cells was produced. By flowcytometry. IC4preferentially reacted with less than 5% of fresh CBA spleen cells and 20-50% of CBA-interleukin-3 (lL-3) cells, an in vitro derived NC-like cell line. In vitro treatment of spleen cells from a number of inbred mouse strains either with 1C4 alone or 1C4 coupled to dynabeads markedly decreased or abolished NC activity of the cells against ^'Cr-labelled WEHl-164 targets. Splenic NC activity of these same mouse strains was also reduced or abolished by in vivo administration of IC4. The effect was evident within 2 h of treatment and persisted for at least I week. In contrast IC4 had little or no effect on splenic NK activity against "Cr-labeiled YAC-I targets over the same range of experiments in vitro and in vivo. Results of strain surveys for both in vitro and in vivo reduction of splenic NC activity by 1C4 treatment showed that CBA. C57BL/6, BALB/c and NZB mice were positive and CE and DBA/2 mice were negative, indicating that 1C4 recognizes an allo-antigen on mouse NC cells. This allo-antibody has been designated NC-II, and thus IC4 is an anti-NC-11 monoclonal antibody. INTRODUCTION The first evidence for the existetice of natural cytotoxic (NC) cells, was obtained by Slutrnan e( al. and Paige et al. in 1978 (1,2). They showed thai NC cells were able to kill 3-methylcholanthrene (3-MCA) induced fibrosarcoma target cells in long (24 h) but not short (4 h) in vitro cytotoxicily assays. Subsequent studies with anti-NK-21 (CE anti-CBA) and anti-NK3-1 (C3H anti-St) allo-antisera provided definitive evidence that NC cells vi'ere a separate population of cells to the NK cells which lysed lymphoma (YAC-1) targets in short (4 h) 5iCr release assays (3.4). Further, complement fixing antibodies were detected which identified NC cells in a C3H anti-St serum (4). Correspondence: Professor R. C. Burton. Discipline of Surgical Science, David Maddison Clinical Sciences Building. Cnr. King and Watt Streets, Newcastle, NSW 2300. Australia. Abbreviations u.fed in this paper: NK. natural killer: NC, natural cytotoxic; *'Cr. sodium chromate (Na2^'CrO4); MoAb, monoclonal antibody; TCM. tissue culture medium: FCS, fetal calf serum; i,p,, intraperitoneal; lL-3, interleukin-3: LU, lytic unit; PBS, phosphate buffered saline; FITC. fluorescein isothiocyanale; PEG. polyethylene glycol; HAT, hypoxanthin. aminopterin. thymidine; TNF. tumour necrosis factor; 3-MCA, 3-methytcholanthrene; ELISA, enzyme-linked immunosorbent assay.
Recently. Hackett et al.(5) using the anti-NK1 • 1 monoclonal antibody (MoAb) (6) in cell sorting experiments showed that sorted NK-1-1 ^, NK-1-1~ and unsorted cells all demonstrated similar NC activity, and estimated that 5-10% of total splenic NC activity was mediated by N K - M + NK cells. Bykowsky and Stutman also produced evidence by cell sorting, which indicated that NC activity was mediated by more than one cell type (7). They showed that both T and non-T cells mediated natural cytotoxicity, but that the major NC cell activity was in the Thy-I+, L3T4+ T cell fraction. These observations indicate that natural cytotoxicity is mediated by cells of more than one lymphohaemopoietic lineage. Of interest is the heterogeneity of lymphohaemopoietic cells that can spontaneously lyse tumour targets in v'//ra(1.3), and recent research in our laboratory has concentrated on producing anti-NK and anti-NC cell specific MoAb. This paper reports the production of a MoAb which significantly reduces or abolishes splenic NC activity in vitro znd in v/vo in a number of mouse strains. MATERIALS AND METHODS Mice CE, CBA, BALB/c, C57BL/6. DBA/2, NZB, {BALB/c X CE) Fl and (BALB/c X CBA) Fl mice of
278
Y. C. SMART£T,4Z,.
both sexes were obtained from the Central Animal House, University ofNewcastle and maintained in the David Maddison Clinical Sciences Building Animal Holding under standard conditions. All mice were used at 6-13 weeks of age. For production of MoAb. CE mice were 6 weeks of age at the time of first immunization. Tissue culture medium (TCM)
Dulbecco's modification of Eagle's Medium (Flow Laboratories, Sydney, NSW, Australia) was supplemented with 2 mmol L-glutamine, 20 mmol/L Hepcs, 5X 10~5 mol/L 2-mcrcaptoethanol. 50 jig/mL gentamycin and 10% fetal calf serum (FCS). Immunization and fusion protocol Six week-old CE female mice were given weekly intraperitoneal (i.p.) injections of CBA spleen cells (5X 10'' viable cells in PBS, Dulbecco's phosphate buffered saline) according to our protocol for the production of NK cell specific allo-antisera (3). After 15 or more i.p. immunizations, individual CE mice were given a booster i.p. injectionof 5 X 10^ viable cells and the spleen was excised for fusion 3 days later accord ing to the protocol of Kohler and Milstein (8). Briefly, a single cell suspension of spleen was prepared in TCM without FCS and mixed with 100% viable NS-I myeloma cells in a 10:1 ratio. Fusion was carried out in 50% (w/v) polyethylene glycol 1500 (PEG 1500; Boehringer Mannheim. Germany) with gentle ccntHfugation (300^) for 90s. Following removal of PEG by washing, the cells were resuspended in TCM containing 10% FCS and 2% HAT (hypoxanthine. aminopterin and thymidlne). aliquoted into two 24 well plates containing mouse peritoneal macrophage feeder cells and cultured at 37°C in 5% COj in air. The fused cells were maintained in HAT medium for 14 days. Clones were visible at days 7-10 and initial screening for immunoglobulins (Ig) was commenced on day 12. Hyhridoma selection strategy Hybridomas which secreted significant levels of Ig were detected by the enzyme-linked immunosorbent assay (ELISA) technique using horseradish peroxidase conjugated rabbit anti-mouse Ig (Dakopatts. Denmark) and ABTS (2,2-azino-di-3-ethyl-benzthiazolinc sulfonate) substrate (Amersham, England). The Ig positive supcrnatants were screened against CBA (immunizing strain) spleen cells usingflowcytometry. Supernatants binding less than 10% spleen cells were then tested for reactivity with an NC-like cell line (unpubl. data). This cell line was produced by culturingCBA spleen cells in TCM containing 30% WEHl3B (a promyelocytic leukaemia cell line of BALB/c mice; provided by Professor Don Metcalf, Walter and Eliza Hall Institute) culture supernatant (crude interleukin 3 [lL-3]) using previously described methods (9,10) and has been named CBA-IL-3. This CBA-IL-3 cell line has been maintained in culture for about 2 years and has been shown to have only in vitro NC
activity. Supernatants binding to more than 10% of these cells were tested further for in vitro reduction of CBA splenic NC and NK activity against ^'Cr-labellcd WEH1-164 or YAC-I target cells. A hybridoma which secreted MoAb with significant anti-NC but negligible anti-NK activity was cloned twice by limiting dilution and then injected into pristane-primed and sublethally irradiated (BALB/c X CE) Fl mice to generate MoAb ascites. Two pools of ascitcs were produced which had comparable titres of anti-NC activity, and both were used in the studies described herein. This MoAb was named IC4, and was of IgGi isotype as determined using the mouse Typerkit (BIO-RAD Laboratories. CA. USA). Preparation of spleen effector celts Spleens were harvested from at least two mice for each experiment unless stated otherwise and the cells were kept in ice cold TCM throughout the preparation period. This was found to be crucial for preserving the NK and NC activities of the harvested spleen ceils. Briefly, spleens were aseptically removed and gently teased apart in 5 mL TCM. The cell suspension was passed through a sterile small nylon wool column to remove spleen capsules or connective tissues. The suspension was made up to 10 mL and centrifuged at 250 g for 5 minat4°C. The cell pellet was resuspended in the residual supernatant and then 5 mL bufl'ered ammonium chloride solution was added to lysc the red cells. After 5 min on ice, 4 mL TCM was added, followed by an underlay of 1 mL FCS. The suspension was again centrifuged and the cell pellet was washed with TCM (no FCS) and then resuspended in TCM (no FCS) to 5X lCcell/mL. Two hundred microlitres of cell suspension were aliquoted into each tube for the cytotoxicity assay. release assay Splenic NC activity was assayed by in vitro lysis of 5'Cr-labelledWEHI-164, a 3-mcthy!cholanthrene (3MCA) induced firbrosarcoma of BALB/c mice and an archetype NC cell target, and splenic NK activity was assayed by in vitro lysis of ^'Cr-labelled YAC-I, a Moloney virus induced T lymphoma and an archtype NK cell target as described elsewhere (3,11). Eriefly. three to five two-fold dilutions of effector spleen ceils were assayed on 10" siCr-tabelied WEHI-164 and YAC-I target cells (starting with 100:1 effector:targct ratio) in quadruplicate in 200 fiL of TCM in 96 well microtitre trays. The cells were pelleted at 150^ for 1 min. and incubated at 37°C in 5% CO2 in air and the assays were of 18 h duration. After incubation, the cells were again centrifuged and 100 jiL of supernatant from each well was harvested and the levels of radioactivity measured in a COBRA gamma-counter (PACKARD Instrument Co., Illinois, USA), The standard deviation of the replicates was within 10% of the mean, Lytic units (LU) were calculated from the linear slope of a graph of per cent specific lysis versus effector: target (E:T) ratios. One LU is herein defined as the number of effector cells which mediate 20% specific lysis of the tai^et cells. Results are expressed either as
279
A MONOCLONAL ANTIBODY TO MOUSE NC CELLS
LU per 10^ spleen cells or per cent reduction in LU according to the following formula: % reduction in LU = LU (control) - LU (1C4 treated) ^ ^^^ LU (control)
for 30 min with 10'^ spleen cells prepared from at least two mice. After this incubation, the dynabeads were removed with a magnetic particle concentrator. The negatively selected spleen cells were resuspended to the starting volume and used as effectors against ^'Crlabelled WEHl-164 and YAC-1 target eells.
Flow cytomelric analysis Cells (>90% viability) were harvested from at least two CBA mouse spleens or CBA-lL-3 cell line in culture, and resuspended to 1O*" cells per tube in PBS. Into each tube was aliquoted 100 |jL of hybridoma supernatanl and then the tube was incubated at 4°C for 30 min. After two washes, ihe cell pellets were incubated wilh a secondary fluorescein isothiocyanate (FITC) conjugated goat anti-mouse IgG (heavy and light chain specific, Tago. Burlingame. CA, USA) at 4°C for 30 min. After two more washes the per cent stained cells were measured using a FACSean Automated Flow Cytometer (Becton-Dickinson Immunocytometry System. Mountain View, CA. USA). Negative controls were samples stained only with the secondary fluoreseeinated antibody.
In vivo depletion of splenic NC and NK activity CBA mice were injected i.p. with various doses (5. 20 and 85 ^iL in a total volume of 200 ^L in TCM) of 1C4 ascites. Control mice were injected with 200 (J1 TCM alone. Spleens from two 1C4 treated and two TCM treated mice were harvested at 5 and 24 h after injection, single cell suspensions were prepared from eaeh spleen, resuspended to lO*" eells per tube and assayed for NC and NK activity against *'Cr-labelled WEH1 -164 and YAC-1 targets. From these results the optimal dose (that which elicited maximum reduction ineytotoxieity by 1C4) was established at 20 (JL. which was the amount injected into miee to determine the kinetics of in vivo reduction of NC activity. In this kinetic study mice were sacrificed at 0, 5, 16, 24, 30 and 48 h and 1 and 2 weeks after injection. The spleen cells were tested against WEHl-164 and YAC-1 targets in vitro. Mice injected with TCM alone served as the control at each time point of testing. Data from this study showed that the maximum reduction in splenic NC activity was 24 h after treatment. In a subsequent strain survey, the model used was injeetion of 20 jiL 1C4 followed by In iv7ro assay ofthe cytotoxic aetivity of the spleen 24 h later.
In vitro MoAb depletion of splenic NC and NK activity Spleen cells pre-incubaied wilh !C4 Spleen cells (I X 10') from different inbred mouse strains were incubated with 1C4 ascites. diluted 1:2-1:512. in 10 mL V-bottomed tissue culture tubes, for 30 min at 37°C in 5%CO2 in air. After incubation, the cells were washed twice, adjusted to the starting volume with TCM and the cytotoxicity against ^'Cr-labelled WEHl-164 and YAC-1 target cells was assessed as described above. Cells incubated with TCM alone served as controls. 1C4 added directly lo the effector-target culture Twenty-five microlitres of neat 1C4 ascites was added at 0, 4. 8 and 24 h into replicate cultures (200 (iL) of four two-fold dilutions of CBA spleen effectors and -^'Cr-labelled WEHl-164 targets. The eells were mixed after each 1C4 addition and gently pelleted at 150;? for I min. One hour after the last addition of IC4. the plates were again centrifuged and all the supernatants were harvested for radioactivity measurement as above. The control was untreated spleen cells. In vitro tlyriabead assay Dynabead M-450 magnetic beads eoated with sheep anti-mouse IgG (DYNAL AS. Oslo, Norway) were used at a spleen:bead ratio of 20:1 in attempts to remove splenic NC and NK cells. Briefly, dynabeads were prepared aecording to the manufacturers" instructions, and reacted at room temperature for 2 h on a rotator with 200 |iL of IC4 supernatant. Beads similarly reacted with 200 ^L TCM or 200 \ih OKT3 supernatant (mouse anti-human CD3) served as the control. One percent normal mouse serum was added to the reaction mixture in order to block the unbound sites. The treated dynabeads were washed three times with PBS containing 1 % FCS and then reacted at ^TC
RESULTS MoAb tC4 (IgGi) was initially selected by flow cytometry on two criteria: less than 5% binding to naive spleen cells ofthe immunizing strain, CBA, and 20-50% binding to the CBA-IL3 cell line (Fig. 1). This was followed by a series of
10'
10^
103
10"
FLUORESCENCE INTENSITY
Fig. 1. Superimposed How cytometric histograms of the CBA-IL-3 eell line reacted with 1C4 and the fluorescein isothiocyanate (FITC) conjugated seeondary antibody (a) and the control stained only with FITC conjugated secondary antibody (b).
280
Y, C. SMART
experiments both in vitro and in vivo to determine whether 1C4 was an anti-NC MoAb.
Table 1. Effect of in vitro IC4 treatment of CBA spleen cells on splenic NC activity.
In vitro reduction of splenic NC and NK activity by IC4 Effect of IC4 pretreatment on NC activity of spleen cells Pretreatment of CBA spleen cells with IC4 in vitro abolished NC activity against 5'Cr-labelled WEHl-164 at the 1:2 and 1:4 dilutions of the IC4 ascites and a dose-response was evident (Fig. 2). When tested against YAC-1 targets, these same spleen cells showed only 27% reduction in NK activity at these same 1:2 and 1:4 dilutions of the 1C4 ascites with no further reduction at higher dilutions. Table 1 shows the results of seven experiments in CBA mice with the 1:4 dilution of IC4 ascites. In each experiment, spleen cells were pooled from two or three mice and as can be seen, the reductions in NC activity against WEHl-164 by in vitro pre-
% Reduction in lytic units Lytic units 1C4 treated Experiments Control* 1C4 treated versus control
90
DC o
33 120
161 159
56 55
100 100
100 76 76 65 76
5 110 9 90 J
UJ
70
4 -10,
16
64
256
1024
RECIPROCAL OF 1C4 DILUTION
50
Fig. 3. A typical experiment comparing the reduction in splenic NC activity between the immunizing (CBA o) and recipient (CE«) strains of mice by in vitro treatment with IC4 at 1:2-1:512 dilutions of ascites.
30
o a HI
0 0 0
500 345 140 500
*Spieen cells treated witb TCM.
o
70
z z
357
5 50 O 30 \^ 10
110 r O
1 2 3 4 5 6 7
10 -10
0
4
16
64
256
1024
RECIPROCAL OF 1C4 DILUTION Fig. 2. A typical experiment comparing the effect of in vitro treatment of CBA spleen cells with IC4 (1:21:512) on NC activity against WEHl-164 (o) and on NK activity against YAC-1 (•).
treatment with 1C4 ranged from 65 to 100%. These data also show the wide variations in NC activity per 10** spleen cells observed in repeated experiments, despite the use of age and gender standardized inbred mice of one strain. In this
Table 2. Effect of i>i vitro 1C4 (1:4 dilution) treatment of spleen cells from five inbred mouse strains on splenic NC activity. Mouse strain C57BU6 BALB/c NZB CE
Experiments
Control*
1
200
2
3 1
200 625 370 135 352 142 208 357 217
2
414
1
2
I
2 1 2
*Spleen cells treated with TCM.
Lytic units IC4 treated 0 100 238 185 34 148 100 154 294 161 384
% Reduction in lytic units 1C4 treated versus control
100 50 «2 50 75 58 30 26 18 26 7
A MONOCLONAL ANTIBODY TO MOUSE NC CELLS
regard natural cytotoxicity is similar to natural killing (I I). When the immunizing (CBA) and recipient (CE) strains were compared for differences in NC cell mediated lysis of WEHl-164 following the treatment of spleen cells with 1C4 ascites. NC activity was abolished at the 1:2 and 1:4 dilutions in CBA strain while CE mice demonstrated only 31 and 18% reductions at these same dilutions (Fig. 3). When spleen cells from a number of inbred mouse strains were tested for reduction in NC activity by 1C4 ascites (1:4 dilution), results showed that strains C57BL/6. BALB/c and NZB were characterized by markedly reduced NC activity (50-100%) while the reduction (7-30%) was much less in CE and DBA/2 mice {Table 2).
281
Effect of IC4 coupled dynabead treatment on NC activity of spleen cells When 1C4 was coupled to sheep anti-mouse IgG-coated dynabeads. followed by reaction of the dynabeads with CBA spleen cells and subsequent removal by a magnetic particle concentrator, the remaining negatively selected spleen cell population demonstrated 79 and 100% reduction in splenic NC activity in two experiments as compared with the controls (Table 3). When the experiment was conducted with spleen cells from the strains tested above the results showed more than a 42% reduction in splenic NC activity in C57BL/6. BALB/c and NZB, while CE and DBA/2 mice showed 0 and 11-13% reductions of NC activity respectively {Table 3). When negatively depleted CBA spleen
Table 3. hi vitro depletion of splenic NC cell activity by 1C4 coupled dynabeads. Mouse strain CBA
Experiments
Control*
1
3333 81 370 62 145 106 60 52 588 57 167 133
2 1 2 1 2 I 2 1 2 1 2
C57BU6 BALB/c NZB CE DBA/2
Lytic units 1C4 treated
% Reduction in lytic units IC4 treated versus control
714 0 1S6 31 75 48 33 30 666 63 149 116
79 100 58 50 48 55 45 42 0 0 11 13
•Controls were beads treated with TCM or an irrelevant mouse anti-human MoAb.
cells were tested for NK activity in replicate experiments 0 and 26% reductions of NK activity were observed {data not shown).
120
t-
z
3 O 1-
100
60
>•
z
60
ION
1
40
O
20
1 1 1
0 0
4
8
24
PULSE TIME {h) Fig. 4. Per cent reduetion in splenic NC activity when eultures of effeetor-target cells in a ^'Cr release assay were pulsed with IC4 at times 0, 4, 8 and 24 h. Results represent the mean and s.d. of two experiments.
Effect of adding IC4 to the in vitro assay of natural cytotoxicity When 1C4 was added at various times to assays of CBA splenic effectors and s'Cr WEHl-164 target cells, a time dependent reduction of NC activity was observed {Fig. 4). In two separate experiments, results showed that the effect was maximal (100%) when 1C4 was added at time 0 and 4h and decreased to 48% at 24 h. Effect on NC activity of\n vivo treatment of mice with 1C4 Tbe optimal effective dose of 1C4 and the time of maximum splenic NC activity loss were
Y.C. SMART £r^L.
282
Table 5. In vivo effects of i.p. injections of 20 [tL 1C4 on CBA splenic NK activity.
110
90
t— >_l
70
z
50
REDUCT ION
z o
30
Titne after injection (h)
•O
% Reduction in lytic units Lytic untts IC4 treated Control* 1C4 treated versus control
2 5 24 48
10 -10
20
40
60
80
1 00
233 345 800 182
217 303 833 156
7 12 0 14
*Mice treated with TCM.
DOSAGE OF 1C4 Fig. S. Dose- and titne-response of the effect of in vivo 1C4 administration on in vitro splenic NC cell activity of (CBA X BALB/c)FI tnice. The lytic activity of the spleen cells was measured by the in vitro ^'Cr release assay. Results measured at 5 (o) and 24 (•) hours after IC4 injection are represented as per ceni reduction in lytic unit versus dosage of 1C4. determined by monitoring in vitro the ability of spleen cells from 1C4 treated mice to lyse -'''Crlabelled WEH 1 -164 targets. Results showed that maximum reduction of splenic NC activity was achieved with 20 jiL of neat 1C4, monitored after 24 h of adtninistration (Fig. 5). The kinetics of in vivo reduction of NC activity were investigated by treating mice with a
single injection of 20 fiL 1C4 or TCM and monitoring splenic NC activity at intervals, from 2 h to 2 weeks later. Table 4 summarizes the results of these experiments. The reduction in NC activity occurred rapidly, was evident from within 2 h of injection and the effect was maximal at 24 h, persisted for I week and had almost returned to control levels by 2 weeks after treatment. In contrast, simultaneous monitoring of splenic NK activity over the first 48 h showed that splenic NK activity againt YAC was reduced by no more than 14% throughout this period (Table 5). When tested in a number of mouse strains, a single injection of 20 pL of 1C4 in a total of 46 CBA, C57BL/6. BALB/c and {CBA X
Table 4. In vivo effect of i.p. injections of 20 yiL IC4 on CBA splenic NC cell activity monitored at intervals af\er treatment. Time af^er injection
No. of experiments*
2h 5h 16h 24 h 30 h 48 h 1 week 2 weeks
% Reduction in lytic units 1C4 treated versus conirol Range^ Mean 41-80 54-89 31-98 62-100 50-72 21-94 64-72 0-29
63 68 67 77 59 S8 68
•Pairs of mice were injected with either 1C4 or TCM at the indicated time intervals. ^The range was obtained from the reduction of NC activity when 1C4 treated mice were compared with their TCM treated paired controls. Table 6. Cumulative data showing the effect of in vitro treatment of mice with 20 jiL 1C4 on splenic NC activity. Mouse strain CBA C57BL/6 BALB/c (CBA X BALB/c)FI CE DBA/2
No. of experiments 24 3 4 15 4 4
% Reduction in lytic units IC4 treated versus control Range Mean
50-100 67-70 75-86 52-100 0-23 0
83 69 79 72 t« 0
A MONOCLONAL ANTIBODY TO MOUSE NC CELLS
BALB/c)Fl mice produced a 50-100% reduction in LU. In contrast. CE and DBA/2 mice showed much lower reductions (0-23%) in splenic NC activity (Table 6).
DISCUSSION The development of MoAb to murine T lymphocytes and their subsets has been crucial to the Investigation of T lymphocyte ontogeny and function (12). By analogy, the development of MoAb to murine NK cell subsets should provide a better understanding of the immunobiology of NK and NC cells. This has in fact been the case. Since the production of anti-NK-l-l MoAb in 1984 (6). cell sorting experiments have shown that the NK-11+ subset of NK cells can also mediate natural cytotoxicity (5). Further, an in vivo role for the N K - l l + NK cell subset in controlling the localization and growth of the BI6 melanoma and CT38 colon carcinoma has been established using the chronically anti-NK11 MoAb treated N K - l l " mouse model (13). This same NK-1-1~ mouse model has proven useful in studying the differentiation of murine NK cells (14). Our efforts to produce MoAb to NK and NC cell specific allo-antigens resulted in the generation of an anti-NC MoAh. 1C4, from a CEantiCBA immunization. The MoAb was selected by five criteria: high Ig level by ELISA analysis; low reactivity with naive CBA {the immunizing strain) spleen cells; significant binding to the CBA-IL-3 NC-Iike cell line and IC4 significantly inhibited NC activity of CBA-1L3 cells in a ^'Cr release assay (data not shown); significant in vitro depletion of splenic NC cell activity by IC4 alone and 1C4 coupled to dynabeads; and reduction of splenic NC cell activity by in vivo treattnent with the MoAb. The results of/« vZ/ra treatment of spleen cells with IC4 alone and 1C4 coupled to dynabeads and in vivo treatment of mice with 1C4 showed that reductions in splenic NC activity were strain dependent. We consider that the reductions In NC activity with the use of 1C4 in CE mice of 18-30% with in vitro pretreatment of spleen cells (Table 2), of 0% with dynabead in vitro treatment (Table 3) and 0-23% with in v/vo treatment of mice (Table 6) constitute the "non-specific" backgrounds for these assays. By these criteria. CBA, C57BL/6, BALB/c and NZB mice are positive while CE and DBA/2 are negative for the expression of the cell surface antigen recognized by 1C4.
283
When the 1C4 pretreated spleen cells were tested against YAC-1 targets, there was a modest reduction in NK activity at the 1:2 and 1:4 dilutions of the MoAb (Fig 2). Low levels of reduction in NK activity against YAC were also observed in the dynabead experiments, and after in v/vo treatment of mice (Table 5). We consider that these reductions are unlikely to be biologically significant when compared with the effect on NC activity in the same experiments. We do, however, recognize that the cell surface antigen recognized by 1C4 is probably expressed by a subset of NK cells since some NK cells at least can mediate natural cytotoxicity (5). Therefore, it is possible that a minor component of NK cell mediated lysis of lymphoma tumour targets is mediated by tumour necrosis factor (TNF), which has been shown to be the mechanism by which NC cells lyse target cells (15). The results of the kinetic study of/« vivo IC4 depletion of splenic NC activity taken together with those of the in vitro pretreatment experiments (Tables 1.2, Fig. 2) strongly suggest that 1C4 cannot be simply an anti-TNF MoAb. Preliminary immunoprecipitation experiments indicate that the cell surface structure recognized by 1C4 is not TNF, further supporting this conclusion (unpubl. data). Thus, it is our contention that this cell membrane molecule is likely to be involved in NC-target cell recognition or TNF production. Given that 100% reductions in NC activity were obtained in one or more of each of the different experiments conducted, it is likely that this molecule is present on the cell membrane of all cells which mediate NC activity. However, the possibility that a minor subpopulation of cells with NC activity that does not express the cell surface molecule recognized by IC4 is not excluded. The kinetics of natural cytotoxicity are quite different from natural killing with no measurable lysis in the first 4-6 h of the assay with most target ceil lysis occurring in the final 4 h (3). This suggests that the steps from target cell recognition, through TNF production to TNF release require many hours. Since 1C4 was able to inhibit NC activity by almost half when added 1 h before the assay was terminated, it is likely that IC4 interferes with a late stage of TNF production and/or release. Natural cytotoxicity is mediated by lymphohaemopoietic cells of more than one differentiation lineage (5,7). Thus, although NC cells constitute one subset of the four so far described in fresh murine spleen which spontaneously lyse tumour cells in vitro, it is clear they are quite
284
Y. C. SMART ETAL.
distinct from the NK cells which lyse lymphoma targets (16). We have previously shown that a cytotoxic antibody present in C3H anti-St antiserum recognizes an antigen present on ST and DBA/2 splenic NC cells but not on NZB. C57BL/6, C58, Ma/My, CBA. C3H and BALB/c NC cells (4), It is apparent from results reported here that 1C4 recognizes an as yet undefined antigen which is different from that identified by the C3H anti-St antiserum, although it may be determined by a gene allelic to that which codes for the antigen identified by IC4. This antigen has been designated NC-11. and lC4ananti-NC-ll MoAb.
In summary, we have produced a murine MoAb, anti-NC-11 to mouse NC cells. Just as MoAb anti-NK-ll has contributed to our understanding of the biology of mouse NK cells, we would expect our anti-NC-1 -1 MoAb to contribute new knowledge to the immunobiology of murine NC cells. Acknowledgements We Ihank the National Health and Medical Research Council of Australia, the New South Wales State Cancer Council, the Governtnent Employees Assistance to Medical Research Fund and the Royal Australasian College of Surgeons for financial support.
REFERENCES 1. Stutman, O.. Paige. C. J. and Figarella. E. F. 1978. Natural cytotoxic cells against solid tumours in mice. I. Strain and age distribulion and target cell susceptibility./ Immunoi 121: 1819-1826. 2. Paige, C. J., Figarella, E. F.. Cuttito, M. J.. Cahan, A. and Stutman, O. 1978, Natural cytotoxic cells against solid tumours in mice. II. Some characteristics of the effector cells. 7./mwHno/. 121: 18271835. 3. Burton. R. C, Bartlett, S. P., Kumar, V. and Winn, J. J. 1981. Studies on NK cells H. Serological evidence for heterogeneity of NK cells. ./. Immunol. 127: 1864-1868. 4. Burton. R. C . Bartlett, S. P. and Winn. H. J. 1982. Alloantigens specific for natural killer cells. In NK Cells and other Natural Effector Cells, R. D. Heberman (ed,). Academic Press, New York, pp. 105-112. 5. Hackctt. J. Jr, Bennett, M., Koo, G. C. and Kumar. V. 1986. Origin and differentiation of natural killer cells. III. Relationship between the precursors and effectors of natural killer and natural cytotoxic activity. Immunol. Res. S: 16-24. 6. Koo. G. C. and Peppard. J. R. 1984. Establishment of monoclonal anti-NK-i • 1 antibody. Hyhri£^(>ma 3: 301-303. 7. Bykowsky. M. J. and Stutman, O. 1986. The cells responsible for murine natural cytotoxic (NC) activity: A multi-lineage system, y. Immunoi 137: 1120-1126. 8. Kohler, G.and Mi!stcin,C. 1975. Continuousculture of fused cells secreting antibody of predefined specificity. Nature 256: 495-497. 9. Smart, Y. C, Stevenson, K. L., Thome, R. E., Thomas, W. D., Hsu, L. H. and Burton, R. C.
1989. Expression of natural killer (NK) cell-specific alloantigens on a mouse NK-like cell line. Immunol. Cell. Biol. 67: 239-242. 10. Djeu. J. Y., Lanza, E.. Pastore, S. and Hapcl. A. J. 1983. Selective growth of natural cytotoxic but not natural kilter effector cells in interleukin-3. Nature 306: 788-791. 11. Burton. R. C, Grail. D. and Warner, N. L. 1978. Natural cytotoxicity of haemopoietic cell populations against murine lymphoid tumours, Brit. J. Cancer yi: 806-817. 12. Reinherz. E. L and Schlossman, S. F. 1981. The characterisation and function of human immunoregulatory T lymphocyte subsets. Immunol. Today 2: 69-75. 13. Seaman. W. E.. Sleisenger. M., Eriksson, E. and Koo, G, C. 1987, Depiction of natural killer cells in mice by monoclonal antibody to NK-l-l. Reduction in host defense againsi malignancy without loss of cellular and humoral immunity. / Immunol. 138: 4539-4544. 14. Koo, G. C Dumont, F. J.. Hackett, J. Jr and Kumar. V. 1986, The NK-1 • t ( - ) mouse: a model to study differentiation of murine NK cells, / Immunol. 137: 3742-3747. 15. Ortaldo, J. R.. Mason, L, H., Mathieson, B. J., Liang. S. M., Flick, D. A. and Herbcrman, R. B. 1986. Mediation of mouse natural cytotoxic activity by tumour necrosis factor. Nature, 321: 700702. 16. Burton. R, C . Koo. G, C , Smart. Y. C . Clark, D. A. and Winn, H. J. 1988. Surface antigens of murine natural killer cells. Int. Rev. Cytol. 3: 185210.
Production of a monoclonal allo-antibody to murine natural cytotoxic cells Y. C Smart, K. L. Stevenson, M. L. Farrelly, J. H. Brien and R. C. Burton Discipline of Surgical Science. Faculty of Medicine, University of Newcastle. NSW 2308. Australia (Submitted
15 May 1990. Accepted for publication 27 July 1990.)
Summary A mouse IgGi monoclonal antibody (1C4), which recognizes a cell surface molecule on murine natural cytotoxic{NC) cells was produced. By flowcytometry. IC4preferentially reacted with less than 5% of fresh CBA spleen cells and 20-50% of CBA-interleukin-3 (lL-3) cells, an in vitro derived NC-like cell line. In vitro treatment of spleen cells from a number of inbred mouse strains either with 1C4 alone or 1C4 coupled to dynabeads markedly decreased or abolished NC activity of the cells against ^'Cr-labelled WEHl-164 targets. Splenic NC activity of these same mouse strains was also reduced or abolished by in vivo administration of IC4. The effect was evident within 2 h of treatment and persisted for at least I week. In contrast IC4 had little or no effect on splenic NK activity against "Cr-labeiled YAC-I targets over the same range of experiments in vitro and in vivo. Results of strain surveys for both in vitro and in vivo reduction of splenic NC activity by 1C4 treatment showed that CBA. C57BL/6, BALB/c and NZB mice were positive and CE and DBA/2 mice were negative, indicating that 1C4 recognizes an allo-antigen on mouse NC cells. This allo-antibody has been designated NC-II, and thus IC4 is an anti-NC-11 monoclonal antibody. INTRODUCTION The first evidence for the existetice of natural cytotoxic (NC) cells, was obtained by Slutrnan e( al. and Paige et al. in 1978 (1,2). They showed thai NC cells were able to kill 3-methylcholanthrene (3-MCA) induced fibrosarcoma target cells in long (24 h) but not short (4 h) in vitro cytotoxicily assays. Subsequent studies with anti-NK-21 (CE anti-CBA) and anti-NK3-1 (C3H anti-St) allo-antisera provided definitive evidence that NC cells vi'ere a separate population of cells to the NK cells which lysed lymphoma (YAC-1) targets in short (4 h) 5iCr release assays (3.4). Further, complement fixing antibodies were detected which identified NC cells in a C3H anti-St serum (4). Correspondence: Professor R. C. Burton. Discipline of Surgical Science, David Maddison Clinical Sciences Building. Cnr. King and Watt Streets, Newcastle, NSW 2300. Australia. Abbreviations u.fed in this paper: NK. natural killer: NC, natural cytotoxic; *'Cr. sodium chromate (Na2^'CrO4); MoAb, monoclonal antibody; TCM. tissue culture medium: FCS, fetal calf serum; i,p,, intraperitoneal; lL-3, interleukin-3: LU, lytic unit; PBS, phosphate buffered saline; FITC. fluorescein isothiocyanale; PEG. polyethylene glycol; HAT, hypoxanthin. aminopterin. thymidine; TNF. tumour necrosis factor; 3-MCA, 3-methytcholanthrene; ELISA, enzyme-linked immunosorbent assay.
Recently. Hackett et al.(5) using the anti-NK1 • 1 monoclonal antibody (MoAb) (6) in cell sorting experiments showed that sorted NK-1-1 ^, NK-1-1~ and unsorted cells all demonstrated similar NC activity, and estimated that 5-10% of total splenic NC activity was mediated by N K - M + NK cells. Bykowsky and Stutman also produced evidence by cell sorting, which indicated that NC activity was mediated by more than one cell type (7). They showed that both T and non-T cells mediated natural cytotoxicity, but that the major NC cell activity was in the Thy-I+, L3T4+ T cell fraction. These observations indicate that natural cytotoxicity is mediated by cells of more than one lymphohaemopoietic lineage. Of interest is the heterogeneity of lymphohaemopoietic cells that can spontaneously lyse tumour targets in v'//ra(1.3), and recent research in our laboratory has concentrated on producing anti-NK and anti-NC cell specific MoAb. This paper reports the production of a MoAb which significantly reduces or abolishes splenic NC activity in vitro znd in v/vo in a number of mouse strains. MATERIALS AND METHODS Mice CE, CBA, BALB/c, C57BL/6. DBA/2, NZB, {BALB/c X CE) Fl and (BALB/c X CBA) Fl mice of
278
Y. C. SMART£T,4Z,.
both sexes were obtained from the Central Animal House, University ofNewcastle and maintained in the David Maddison Clinical Sciences Building Animal Holding under standard conditions. All mice were used at 6-13 weeks of age. For production of MoAb. CE mice were 6 weeks of age at the time of first immunization. Tissue culture medium (TCM)
Dulbecco's modification of Eagle's Medium (Flow Laboratories, Sydney, NSW, Australia) was supplemented with 2 mmol L-glutamine, 20 mmol/L Hepcs, 5X 10~5 mol/L 2-mcrcaptoethanol. 50 jig/mL gentamycin and 10% fetal calf serum (FCS). Immunization and fusion protocol Six week-old CE female mice were given weekly intraperitoneal (i.p.) injections of CBA spleen cells (5X 10'' viable cells in PBS, Dulbecco's phosphate buffered saline) according to our protocol for the production of NK cell specific allo-antisera (3). After 15 or more i.p. immunizations, individual CE mice were given a booster i.p. injectionof 5 X 10^ viable cells and the spleen was excised for fusion 3 days later accord ing to the protocol of Kohler and Milstein (8). Briefly, a single cell suspension of spleen was prepared in TCM without FCS and mixed with 100% viable NS-I myeloma cells in a 10:1 ratio. Fusion was carried out in 50% (w/v) polyethylene glycol 1500 (PEG 1500; Boehringer Mannheim. Germany) with gentle ccntHfugation (300^) for 90s. Following removal of PEG by washing, the cells were resuspended in TCM containing 10% FCS and 2% HAT (hypoxanthine. aminopterin and thymidlne). aliquoted into two 24 well plates containing mouse peritoneal macrophage feeder cells and cultured at 37°C in 5% COj in air. The fused cells were maintained in HAT medium for 14 days. Clones were visible at days 7-10 and initial screening for immunoglobulins (Ig) was commenced on day 12. Hyhridoma selection strategy Hybridomas which secreted significant levels of Ig were detected by the enzyme-linked immunosorbent assay (ELISA) technique using horseradish peroxidase conjugated rabbit anti-mouse Ig (Dakopatts. Denmark) and ABTS (2,2-azino-di-3-ethyl-benzthiazolinc sulfonate) substrate (Amersham, England). The Ig positive supcrnatants were screened against CBA (immunizing strain) spleen cells usingflowcytometry. Supernatants binding less than 10% spleen cells were then tested for reactivity with an NC-like cell line (unpubl. data). This cell line was produced by culturingCBA spleen cells in TCM containing 30% WEHl3B (a promyelocytic leukaemia cell line of BALB/c mice; provided by Professor Don Metcalf, Walter and Eliza Hall Institute) culture supernatant (crude interleukin 3 [lL-3]) using previously described methods (9,10) and has been named CBA-IL-3. This CBA-IL-3 cell line has been maintained in culture for about 2 years and has been shown to have only in vitro NC
activity. Supernatants binding to more than 10% of these cells were tested further for in vitro reduction of CBA splenic NC and NK activity against ^'Cr-labellcd WEH1-164 or YAC-I target cells. A hybridoma which secreted MoAb with significant anti-NC but negligible anti-NK activity was cloned twice by limiting dilution and then injected into pristane-primed and sublethally irradiated (BALB/c X CE) Fl mice to generate MoAb ascites. Two pools of ascitcs were produced which had comparable titres of anti-NC activity, and both were used in the studies described herein. This MoAb was named IC4, and was of IgGi isotype as determined using the mouse Typerkit (BIO-RAD Laboratories. CA. USA). Preparation of spleen effector celts Spleens were harvested from at least two mice for each experiment unless stated otherwise and the cells were kept in ice cold TCM throughout the preparation period. This was found to be crucial for preserving the NK and NC activities of the harvested spleen ceils. Briefly, spleens were aseptically removed and gently teased apart in 5 mL TCM. The cell suspension was passed through a sterile small nylon wool column to remove spleen capsules or connective tissues. The suspension was made up to 10 mL and centrifuged at 250 g for 5 minat4°C. The cell pellet was resuspended in the residual supernatant and then 5 mL bufl'ered ammonium chloride solution was added to lysc the red cells. After 5 min on ice, 4 mL TCM was added, followed by an underlay of 1 mL FCS. The suspension was again centrifuged and the cell pellet was washed with TCM (no FCS) and then resuspended in TCM (no FCS) to 5X lCcell/mL. Two hundred microlitres of cell suspension were aliquoted into each tube for the cytotoxicity assay. release assay Splenic NC activity was assayed by in vitro lysis of 5'Cr-labelledWEHI-164, a 3-mcthy!cholanthrene (3MCA) induced firbrosarcoma of BALB/c mice and an archetype NC cell target, and splenic NK activity was assayed by in vitro lysis of ^'Cr-labelled YAC-I, a Moloney virus induced T lymphoma and an archtype NK cell target as described elsewhere (3,11). Eriefly. three to five two-fold dilutions of effector spleen ceils were assayed on 10" siCr-tabelied WEHI-164 and YAC-I target cells (starting with 100:1 effector:targct ratio) in quadruplicate in 200 fiL of TCM in 96 well microtitre trays. The cells were pelleted at 150^ for 1 min. and incubated at 37°C in 5% CO2 in air and the assays were of 18 h duration. After incubation, the cells were again centrifuged and 100 jiL of supernatant from each well was harvested and the levels of radioactivity measured in a COBRA gamma-counter (PACKARD Instrument Co., Illinois, USA), The standard deviation of the replicates was within 10% of the mean, Lytic units (LU) were calculated from the linear slope of a graph of per cent specific lysis versus effector: target (E:T) ratios. One LU is herein defined as the number of effector cells which mediate 20% specific lysis of the tai^et cells. Results are expressed either as
279
A MONOCLONAL ANTIBODY TO MOUSE NC CELLS
LU per 10^ spleen cells or per cent reduction in LU according to the following formula: % reduction in LU = LU (control) - LU (1C4 treated) ^ ^^^ LU (control)
for 30 min with 10'^ spleen cells prepared from at least two mice. After this incubation, the dynabeads were removed with a magnetic particle concentrator. The negatively selected spleen cells were resuspended to the starting volume and used as effectors against ^'Crlabelled WEHl-164 and YAC-1 target eells.
Flow cytomelric analysis Cells (>90% viability) were harvested from at least two CBA mouse spleens or CBA-lL-3 cell line in culture, and resuspended to 1O*" cells per tube in PBS. Into each tube was aliquoted 100 |jL of hybridoma supernatanl and then the tube was incubated at 4°C for 30 min. After two washes, ihe cell pellets were incubated wilh a secondary fluorescein isothiocyanate (FITC) conjugated goat anti-mouse IgG (heavy and light chain specific, Tago. Burlingame. CA, USA) at 4°C for 30 min. After two more washes the per cent stained cells were measured using a FACSean Automated Flow Cytometer (Becton-Dickinson Immunocytometry System. Mountain View, CA. USA). Negative controls were samples stained only with the secondary fluoreseeinated antibody.
In vivo depletion of splenic NC and NK activity CBA mice were injected i.p. with various doses (5. 20 and 85 ^iL in a total volume of 200 ^L in TCM) of 1C4 ascites. Control mice were injected with 200 (J1 TCM alone. Spleens from two 1C4 treated and two TCM treated mice were harvested at 5 and 24 h after injection, single cell suspensions were prepared from eaeh spleen, resuspended to lO*" eells per tube and assayed for NC and NK activity against *'Cr-labelled WEH1 -164 and YAC-1 targets. From these results the optimal dose (that which elicited maximum reduction ineytotoxieity by 1C4) was established at 20 (JL. which was the amount injected into miee to determine the kinetics of in vivo reduction of NC activity. In this kinetic study mice were sacrificed at 0, 5, 16, 24, 30 and 48 h and 1 and 2 weeks after injection. The spleen cells were tested against WEHl-164 and YAC-1 targets in vitro. Mice injected with TCM alone served as the control at each time point of testing. Data from this study showed that the maximum reduction in splenic NC activity was 24 h after treatment. In a subsequent strain survey, the model used was injeetion of 20 jiL 1C4 followed by In iv7ro assay ofthe cytotoxic aetivity of the spleen 24 h later.
In vitro MoAb depletion of splenic NC and NK activity Spleen cells pre-incubaied wilh !C4 Spleen cells (I X 10') from different inbred mouse strains were incubated with 1C4 ascites. diluted 1:2-1:512. in 10 mL V-bottomed tissue culture tubes, for 30 min at 37°C in 5%CO2 in air. After incubation, the cells were washed twice, adjusted to the starting volume with TCM and the cytotoxicity against ^'Cr-labelled WEHl-164 and YAC-1 target cells was assessed as described above. Cells incubated with TCM alone served as controls. 1C4 added directly lo the effector-target culture Twenty-five microlitres of neat 1C4 ascites was added at 0, 4. 8 and 24 h into replicate cultures (200 (iL) of four two-fold dilutions of CBA spleen effectors and -^'Cr-labelled WEHl-164 targets. The eells were mixed after each 1C4 addition and gently pelleted at 150;? for I min. One hour after the last addition of IC4. the plates were again centrifuged and all the supernatants were harvested for radioactivity measurement as above. The control was untreated spleen cells. In vitro tlyriabead assay Dynabead M-450 magnetic beads eoated with sheep anti-mouse IgG (DYNAL AS. Oslo, Norway) were used at a spleen:bead ratio of 20:1 in attempts to remove splenic NC and NK cells. Briefly, dynabeads were prepared aecording to the manufacturers" instructions, and reacted at room temperature for 2 h on a rotator with 200 |iL of IC4 supernatant. Beads similarly reacted with 200 ^L TCM or 200 \ih OKT3 supernatant (mouse anti-human CD3) served as the control. One percent normal mouse serum was added to the reaction mixture in order to block the unbound sites. The treated dynabeads were washed three times with PBS containing 1 % FCS and then reacted at ^TC
RESULTS MoAb tC4 (IgGi) was initially selected by flow cytometry on two criteria: less than 5% binding to naive spleen cells ofthe immunizing strain, CBA, and 20-50% binding to the CBA-IL3 cell line (Fig. 1). This was followed by a series of
10'
10^
103
10"
FLUORESCENCE INTENSITY
Fig. 1. Superimposed How cytometric histograms of the CBA-IL-3 eell line reacted with 1C4 and the fluorescein isothiocyanate (FITC) conjugated seeondary antibody (a) and the control stained only with FITC conjugated secondary antibody (b).
280
Y, C. SMART
experiments both in vitro and in vivo to determine whether 1C4 was an anti-NC MoAb.
Table 1. Effect of in vitro IC4 treatment of CBA spleen cells on splenic NC activity.
In vitro reduction of splenic NC and NK activity by IC4 Effect of IC4 pretreatment on NC activity of spleen cells Pretreatment of CBA spleen cells with IC4 in vitro abolished NC activity against 5'Cr-labelled WEHl-164 at the 1:2 and 1:4 dilutions of the IC4 ascites and a dose-response was evident (Fig. 2). When tested against YAC-1 targets, these same spleen cells showed only 27% reduction in NK activity at these same 1:2 and 1:4 dilutions of the 1C4 ascites with no further reduction at higher dilutions. Table 1 shows the results of seven experiments in CBA mice with the 1:4 dilution of IC4 ascites. In each experiment, spleen cells were pooled from two or three mice and as can be seen, the reductions in NC activity against WEHl-164 by in vitro pre-
% Reduction in lytic units Lytic units 1C4 treated Experiments Control* 1C4 treated versus control
90
DC o
33 120
161 159
56 55
100 100
100 76 76 65 76
5 110 9 90 J
UJ
70
4 -10,
16
64
256
1024
RECIPROCAL OF 1C4 DILUTION
50
Fig. 3. A typical experiment comparing the reduction in splenic NC activity between the immunizing (CBA o) and recipient (CE«) strains of mice by in vitro treatment with IC4 at 1:2-1:512 dilutions of ascites.
30
o a HI
0 0 0
500 345 140 500
*Spieen cells treated witb TCM.
o
70
z z
357
5 50 O 30 \^ 10
110 r O
1 2 3 4 5 6 7
10 -10
0
4
16
64
256
1024
RECIPROCAL OF 1C4 DILUTION Fig. 2. A typical experiment comparing the effect of in vitro treatment of CBA spleen cells with IC4 (1:21:512) on NC activity against WEHl-164 (o) and on NK activity against YAC-1 (•).
treatment with 1C4 ranged from 65 to 100%. These data also show the wide variations in NC activity per 10** spleen cells observed in repeated experiments, despite the use of age and gender standardized inbred mice of one strain. In this
Table 2. Effect of i>i vitro 1C4 (1:4 dilution) treatment of spleen cells from five inbred mouse strains on splenic NC activity. Mouse strain C57BU6 BALB/c NZB CE
Experiments
Control*
1
200
2
3 1
200 625 370 135 352 142 208 357 217
2
414
1
2
I
2 1 2
*Spleen cells treated with TCM.
Lytic units IC4 treated 0 100 238 185 34 148 100 154 294 161 384
% Reduction in lytic units 1C4 treated versus control
100 50 «2 50 75 58 30 26 18 26 7
A MONOCLONAL ANTIBODY TO MOUSE NC CELLS
regard natural cytotoxicity is similar to natural killing (I I). When the immunizing (CBA) and recipient (CE) strains were compared for differences in NC cell mediated lysis of WEHl-164 following the treatment of spleen cells with 1C4 ascites. NC activity was abolished at the 1:2 and 1:4 dilutions in CBA strain while CE mice demonstrated only 31 and 18% reductions at these same dilutions (Fig. 3). When spleen cells from a number of inbred mouse strains were tested for reduction in NC activity by 1C4 ascites (1:4 dilution), results showed that strains C57BL/6. BALB/c and NZB were characterized by markedly reduced NC activity (50-100%) while the reduction (7-30%) was much less in CE and DBA/2 mice {Table 2).
281
Effect of IC4 coupled dynabead treatment on NC activity of spleen cells When 1C4 was coupled to sheep anti-mouse IgG-coated dynabeads. followed by reaction of the dynabeads with CBA spleen cells and subsequent removal by a magnetic particle concentrator, the remaining negatively selected spleen cell population demonstrated 79 and 100% reduction in splenic NC activity in two experiments as compared with the controls (Table 3). When the experiment was conducted with spleen cells from the strains tested above the results showed more than a 42% reduction in splenic NC activity in C57BL/6. BALB/c and NZB, while CE and DBA/2 mice showed 0 and 11-13% reductions of NC activity respectively {Table 3). When negatively depleted CBA spleen
Table 3. hi vitro depletion of splenic NC cell activity by 1C4 coupled dynabeads. Mouse strain CBA
Experiments
Control*
1
3333 81 370 62 145 106 60 52 588 57 167 133
2 1 2 1 2 I 2 1 2 1 2
C57BU6 BALB/c NZB CE DBA/2
Lytic units 1C4 treated
% Reduction in lytic units IC4 treated versus control
714 0 1S6 31 75 48 33 30 666 63 149 116
79 100 58 50 48 55 45 42 0 0 11 13
•Controls were beads treated with TCM or an irrelevant mouse anti-human MoAb.
cells were tested for NK activity in replicate experiments 0 and 26% reductions of NK activity were observed {data not shown).
120
t-
z
3 O 1-
100
60
>•
z
60
ION
1
40
O
20
1 1 1
0 0
4
8
24
PULSE TIME {h) Fig. 4. Per cent reduetion in splenic NC activity when eultures of effeetor-target cells in a ^'Cr release assay were pulsed with IC4 at times 0, 4, 8 and 24 h. Results represent the mean and s.d. of two experiments.
Effect of adding IC4 to the in vitro assay of natural cytotoxicity When 1C4 was added at various times to assays of CBA splenic effectors and s'Cr WEHl-164 target cells, a time dependent reduction of NC activity was observed {Fig. 4). In two separate experiments, results showed that the effect was maximal (100%) when 1C4 was added at time 0 and 4h and decreased to 48% at 24 h. Effect on NC activity of\n vivo treatment of mice with 1C4 Tbe optimal effective dose of 1C4 and the time of maximum splenic NC activity loss were
Y.C. SMART £r^L.
282
Table 5. In vivo effects of i.p. injections of 20 [tL 1C4 on CBA splenic NK activity.
110
90
t— >_l
70
z
50
REDUCT ION
z o
30
Titne after injection (h)
•O
% Reduction in lytic units Lytic untts IC4 treated Control* 1C4 treated versus control
2 5 24 48
10 -10
20
40
60
80
1 00
233 345 800 182
217 303 833 156
7 12 0 14
*Mice treated with TCM.
DOSAGE OF 1C4 Fig. S. Dose- and titne-response of the effect of in vivo 1C4 administration on in vitro splenic NC cell activity of (CBA X BALB/c)FI tnice. The lytic activity of the spleen cells was measured by the in vitro ^'Cr release assay. Results measured at 5 (o) and 24 (•) hours after IC4 injection are represented as per ceni reduction in lytic unit versus dosage of 1C4. determined by monitoring in vitro the ability of spleen cells from 1C4 treated mice to lyse -'''Crlabelled WEH 1 -164 targets. Results showed that maximum reduction of splenic NC activity was achieved with 20 jiL of neat 1C4, monitored after 24 h of adtninistration (Fig. 5). The kinetics of in vivo reduction of NC activity were investigated by treating mice with a
single injection of 20 fiL 1C4 or TCM and monitoring splenic NC activity at intervals, from 2 h to 2 weeks later. Table 4 summarizes the results of these experiments. The reduction in NC activity occurred rapidly, was evident from within 2 h of injection and the effect was maximal at 24 h, persisted for I week and had almost returned to control levels by 2 weeks after treatment. In contrast, simultaneous monitoring of splenic NK activity over the first 48 h showed that splenic NK activity againt YAC was reduced by no more than 14% throughout this period (Table 5). When tested in a number of mouse strains, a single injection of 20 pL of 1C4 in a total of 46 CBA, C57BL/6. BALB/c and {CBA X
Table 4. In vivo effect of i.p. injections of 20 yiL IC4 on CBA splenic NC cell activity monitored at intervals af\er treatment. Time af^er injection
No. of experiments*
2h 5h 16h 24 h 30 h 48 h 1 week 2 weeks
% Reduction in lytic units 1C4 treated versus conirol Range^ Mean 41-80 54-89 31-98 62-100 50-72 21-94 64-72 0-29
63 68 67 77 59 S8 68
•Pairs of mice were injected with either 1C4 or TCM at the indicated time intervals. ^The range was obtained from the reduction of NC activity when 1C4 treated mice were compared with their TCM treated paired controls. Table 6. Cumulative data showing the effect of in vitro treatment of mice with 20 jiL 1C4 on splenic NC activity. Mouse strain CBA C57BL/6 BALB/c (CBA X BALB/c)FI CE DBA/2
No. of experiments 24 3 4 15 4 4
% Reduction in lytic units IC4 treated versus control Range Mean
50-100 67-70 75-86 52-100 0-23 0
83 69 79 72 t« 0
A MONOCLONAL ANTIBODY TO MOUSE NC CELLS
BALB/c)Fl mice produced a 50-100% reduction in LU. In contrast. CE and DBA/2 mice showed much lower reductions (0-23%) in splenic NC activity (Table 6).
DISCUSSION The development of MoAb to murine T lymphocytes and their subsets has been crucial to the Investigation of T lymphocyte ontogeny and function (12). By analogy, the development of MoAb to murine NK cell subsets should provide a better understanding of the immunobiology of NK and NC cells. This has in fact been the case. Since the production of anti-NK-l-l MoAb in 1984 (6). cell sorting experiments have shown that the NK-11+ subset of NK cells can also mediate natural cytotoxicity (5). Further, an in vivo role for the N K - l l + NK cell subset in controlling the localization and growth of the BI6 melanoma and CT38 colon carcinoma has been established using the chronically anti-NK11 MoAb treated N K - l l " mouse model (13). This same NK-1-1~ mouse model has proven useful in studying the differentiation of murine NK cells (14). Our efforts to produce MoAb to NK and NC cell specific allo-antigens resulted in the generation of an anti-NC MoAh. 1C4, from a CEantiCBA immunization. The MoAb was selected by five criteria: high Ig level by ELISA analysis; low reactivity with naive CBA {the immunizing strain) spleen cells; significant binding to the CBA-IL-3 NC-Iike cell line and IC4 significantly inhibited NC activity of CBA-1L3 cells in a ^'Cr release assay (data not shown); significant in vitro depletion of splenic NC cell activity by IC4 alone and 1C4 coupled to dynabeads; and reduction of splenic NC cell activity by in vivo treattnent with the MoAb. The results of/« vZ/ra treatment of spleen cells with IC4 alone and 1C4 coupled to dynabeads and in vivo treatment of mice with 1C4 showed that reductions in splenic NC activity were strain dependent. We consider that the reductions In NC activity with the use of 1C4 in CE mice of 18-30% with in vitro pretreatment of spleen cells (Table 2), of 0% with dynabead in vitro treatment (Table 3) and 0-23% with in v/vo treatment of mice (Table 6) constitute the "non-specific" backgrounds for these assays. By these criteria. CBA, C57BL/6, BALB/c and NZB mice are positive while CE and DBA/2 are negative for the expression of the cell surface antigen recognized by 1C4.
283
When the 1C4 pretreated spleen cells were tested against YAC-1 targets, there was a modest reduction in NK activity at the 1:2 and 1:4 dilutions of the MoAb (Fig 2). Low levels of reduction in NK activity against YAC were also observed in the dynabead experiments, and after in v/vo treatment of mice (Table 5). We consider that these reductions are unlikely to be biologically significant when compared with the effect on NC activity in the same experiments. We do, however, recognize that the cell surface antigen recognized by 1C4 is probably expressed by a subset of NK cells since some NK cells at least can mediate natural cytotoxicity (5). Therefore, it is possible that a minor component of NK cell mediated lysis of lymphoma tumour targets is mediated by tumour necrosis factor (TNF), which has been shown to be the mechanism by which NC cells lyse target cells (15). The results of the kinetic study of/« vivo IC4 depletion of splenic NC activity taken together with those of the in vitro pretreatment experiments (Tables 1.2, Fig. 2) strongly suggest that 1C4 cannot be simply an anti-TNF MoAb. Preliminary immunoprecipitation experiments indicate that the cell surface structure recognized by 1C4 is not TNF, further supporting this conclusion (unpubl. data). Thus, it is our contention that this cell membrane molecule is likely to be involved in NC-target cell recognition or TNF production. Given that 100% reductions in NC activity were obtained in one or more of each of the different experiments conducted, it is likely that this molecule is present on the cell membrane of all cells which mediate NC activity. However, the possibility that a minor subpopulation of cells with NC activity that does not express the cell surface molecule recognized by IC4 is not excluded. The kinetics of natural cytotoxicity are quite different from natural killing with no measurable lysis in the first 4-6 h of the assay with most target ceil lysis occurring in the final 4 h (3). This suggests that the steps from target cell recognition, through TNF production to TNF release require many hours. Since 1C4 was able to inhibit NC activity by almost half when added 1 h before the assay was terminated, it is likely that IC4 interferes with a late stage of TNF production and/or release. Natural cytotoxicity is mediated by lymphohaemopoietic cells of more than one differentiation lineage (5,7). Thus, although NC cells constitute one subset of the four so far described in fresh murine spleen which spontaneously lyse tumour cells in vitro, it is clear they are quite
284
Y. C. SMART ETAL.
distinct from the NK cells which lyse lymphoma targets (16). We have previously shown that a cytotoxic antibody present in C3H anti-St antiserum recognizes an antigen present on ST and DBA/2 splenic NC cells but not on NZB. C57BL/6, C58, Ma/My, CBA. C3H and BALB/c NC cells (4), It is apparent from results reported here that 1C4 recognizes an as yet undefined antigen which is different from that identified by the C3H anti-St antiserum, although it may be determined by a gene allelic to that which codes for the antigen identified by IC4. This antigen has been designated NC-11. and lC4ananti-NC-ll MoAb.
In summary, we have produced a murine MoAb, anti-NC-11 to mouse NC cells. Just as MoAb anti-NK-ll has contributed to our understanding of the biology of mouse NK cells, we would expect our anti-NC-1 -1 MoAb to contribute new knowledge to the immunobiology of murine NC cells. Acknowledgements We Ihank the National Health and Medical Research Council of Australia, the New South Wales State Cancer Council, the Governtnent Employees Assistance to Medical Research Fund and the Royal Australasian College of Surgeons for financial support.
REFERENCES 1. Stutman, O.. Paige. C. J. and Figarella. E. F. 1978. Natural cytotoxic cells against solid tumours in mice. I. Strain and age distribulion and target cell susceptibility./ Immunoi 121: 1819-1826. 2. Paige, C. J., Figarella, E. F.. Cuttito, M. J.. Cahan, A. and Stutman, O. 1978, Natural cytotoxic cells against solid tumours in mice. II. Some characteristics of the effector cells. 7./mwHno/. 121: 18271835. 3. Burton. R. C, Bartlett, S. P., Kumar, V. and Winn, J. J. 1981. Studies on NK cells H. Serological evidence for heterogeneity of NK cells. ./. Immunol. 127: 1864-1868. 4. Burton. R. C . Bartlett, S. P. and Winn. H. J. 1982. Alloantigens specific for natural killer cells. In NK Cells and other Natural Effector Cells, R. D. Heberman (ed,). Academic Press, New York, pp. 105-112. 5. Hackctt. J. Jr, Bennett, M., Koo, G. C. and Kumar. V. 1986. Origin and differentiation of natural killer cells. III. Relationship between the precursors and effectors of natural killer and natural cytotoxic activity. Immunol. Res. S: 16-24. 6. Koo. G. C. and Peppard. J. R. 1984. Establishment of monoclonal anti-NK-i • 1 antibody. Hyhri£^(>ma 3: 301-303. 7. Bykowsky. M. J. and Stutman, O. 1986. The cells responsible for murine natural cytotoxic (NC) activity: A multi-lineage system, y. Immunoi 137: 1120-1126. 8. Kohler, G.and Mi!stcin,C. 1975. Continuousculture of fused cells secreting antibody of predefined specificity. Nature 256: 495-497. 9. Smart, Y. C, Stevenson, K. L., Thome, R. E., Thomas, W. D., Hsu, L. H. and Burton, R. C.
1989. Expression of natural killer (NK) cell-specific alloantigens on a mouse NK-like cell line. Immunol. Cell. Biol. 67: 239-242. 10. Djeu. J. Y., Lanza, E.. Pastore, S. and Hapcl. A. J. 1983. Selective growth of natural cytotoxic but not natural kilter effector cells in interleukin-3. Nature 306: 788-791. 11. Burton. R. C, Grail. D. and Warner, N. L. 1978. Natural cytotoxicity of haemopoietic cell populations against murine lymphoid tumours, Brit. J. Cancer yi: 806-817. 12. Reinherz. E. L and Schlossman, S. F. 1981. The characterisation and function of human immunoregulatory T lymphocyte subsets. Immunol. Today 2: 69-75. 13. Seaman. W. E.. Sleisenger. M., Eriksson, E. and Koo, G, C. 1987, Depiction of natural killer cells in mice by monoclonal antibody to NK-l-l. Reduction in host defense againsi malignancy without loss of cellular and humoral immunity. / Immunol. 138: 4539-4544. 14. Koo, G. C Dumont, F. J.. Hackett, J. Jr and Kumar. V. 1986, The NK-1 • t ( - ) mouse: a model to study differentiation of murine NK cells, / Immunol. 137: 3742-3747. 15. Ortaldo, J. R.. Mason, L, H., Mathieson, B. J., Liang. S. M., Flick, D. A. and Herbcrman, R. B. 1986. Mediation of mouse natural cytotoxic activity by tumour necrosis factor. Nature, 321: 700702. 16. Burton. R, C . Koo. G, C , Smart. Y. C . Clark, D. A. and Winn, H. J. 1988. Surface antigens of murine natural killer cells. Int. Rev. Cytol. 3: 185210.
