FMLP- and TNF-stimulated monoclonal Lym-1 antibody-dependent lysis of B lymphoblastoid tumour targets by neutrophils

Human neutrophils, incubated with Cr51-labelled B lymphoblastoid Raji cells in the presence of the anti-target monoclonal antibody (mAb) Lym-1 plus formyl-methionyl-leucyl-phenylalanine (FMLP) or tumour necrosis factor alpha (TNF-α), were found to induce significant Cr51 release, i.e. significant cytolysis. The lytic process was inhibited by mAb IV.3, specific for the Fcγ receptor (FcγR) type II. The mAb 3G8, which reacts with FcγR type III, was ineffective. Moreover, the lysis was inhibited by the anti-CD18 mAb MEM-48. These data suggest that FMLP/Lym-1 as well as TNF-α/Lym-1 cytolytic systems strictly require FcγRII and CD18 integrins. As the lysis induced by TNF-α/Lym-1 was prevented by pertussis toxin (PT), PT-sensitive G-proteins are likely to intervene in post-FcγRII signal transduction. Both the FMLP- and the TNF-α-dependent systems were also found to be equally susceptible to inhibition by various inhibitors of kinases (genistein, staurosporin, 1-(5-isoquinolinnylsulphonyl)-2-methylpiperazine and wortmannin). On the contrary, an inhibitor of protein kinase C (bis-indolyl-maleimide, BIM) was effective only in the FMLP/Lym-1 cytolytic system. Therefore, it appears that signals delivered by FMLP or TNF-α, BIM-sensitive and insensitive respectively, converge and synergize with those from G-protein-coupled FcγRII and, probably, CD18-integrins to promote the expression of the neutrophil cytolytic potential. © 1999 Cancer Research Campaign


Neutrophil preparation
Heparinized venous blood (heparin 10 U ml -1 ) was obtained from healthy volunteers (20-45 years old) after informed consent. No donor had an infectious disease or was under medication either at the time of sampling or for 2 weeks before sampling. Neutrophils were prepared by dextran sedimentation, followed by centrifugation (400 g, 30 min) on Ficoll-Hypaque density gradient, as previously described (Ottonello et al, 1996). Contaminating erythrocytes were removed by hypotonic lysis (Ottonello et al, 1996). PMNs resuspended in RPMI-FCS were > 97% pure viable, as determined by the assays described above.

Target cells
Lymphoblastoid Raji cells (Ottonello et al, 1996) were used as targets in the cytolytic assays. The Raji cell line was grown in RPMI-FCS and subcultured every 3 days. The capacity of these cells to bind Lym-1 antibody was measured by indirect immunofluorescence with flow cytometry using a rabbit anti-mouse IgG F(ab´) 2 polyclonal antibody conjugated with FITC (Dako) (Ottonello et al, 1996). For cytolytic assays, 4 × 10 6 Raji cells were labelled with 100-200 µCi sodium chromate Cr 51 by incubating for 1 h at 37°C (final volume 0.5 ml, medium: RPMI-1640 plus 5% FCS). After washing, labelled cells were resuspended in RPMI-FCS.

Cytolytic assays
Cytolytic activity of neutrophils was measured as described elsewhere in detail (Dallegri et al, 1984;Ottonello et al, 1996). Briefly, target cells (2 × 10 4 ) were mixed with neutrophils at an effector:target ratio of 20:1, with and without 10 µg ml -1 Lym-1 mAb (Epstein et al, 1987) and/or 1 µM FMLP or 1 ng ml -1 TNF-α appropriately diluted in RPMI-FCS. The effector:target ratio of 20:1 was chosen on the basis of preliminary experiments, also taking into account previous observations (Ottonello et al, 1996). In fact, the per cent cytolysis at effector:target ratios of 20:1 and 40:1 was 28.33 ± 6.43 and 32.5 ± 3.08 (mean ± 1 standard deviation (s.d.), n = 3, P > 0.05) in the TNF-α system, whereas it was 23.97 ± 10.82 and 26.87 ± 7.34 (mean ± 1 s.d. n = 3, P > 0.05) in the FMLP system. Experiments were carried out in the absence or presence of the various mAbs and reagents used to probe the cytolytic process. The assays were carried out in triplicate and in a final volume of 150 µl, using round-bottom microplates (Falcon, Becton-Dickinson Italia, Milano, Italy). After 14-h incubation in humidified atmosphere of 95% air and carbon dioxide, the Cr 51release was determined in the formula 100 × (E-S)/(T-S), where E is the cpm released in the presence of effector cells, T is the cpm released after target cells with 5% Triton X-100, and S is the cpm Figure 1 Neutrophil-mediated cytolysis in the absence or presence of 10 µg ml -1 Lym-1 and/or 1 µM FMLP or 1 ng ml -1 TNF-α. Cr 51 -labelled Raji target cells were at 2 × 10 4 . The neutrophil:Raji cell ratio was 20:1. The incubation time was 14 h. (A) The lysis in the presence of both Lym-1 and FMLP was significantly higher than that in the presence of FMLP or that in the presence of Lym-1 alone, P < 0.001. (B) The lysis in the presence of both Lym-1 and TNF-α was significantly higher than that in the presence of TNF-α or that in the presence of Lym-1 alone, P < 0.001 spontaneously released by target cells incubated with medium alone (< 18%).

Immunofluorescence analysis
Neutrophils (10 6 cells) were incubated for 30 min at 4°C in the presence of FITC-labelled mAbs towards FcγRI, FcγRII, FcγRIII or control mAbs. Polyclonal human IgG (4 mg ml -1 ) was added during incubation to inhibit possible non-specific binding of mAbs to high affinity FcγR for IgG. After incubation, the cells were washed in phosphate-buffered saline (PBS) plus 1% BSA and resuspended in PBS for analysis on a Coulter flow cytometer. To compare results, relative fluorescence intensities (RFI) were calculated as the ratios between the linear fluorescence intensity (FI) obtained with the relevant mAb and the FI obtained with the control mAb.

Statistical analysis
Results were expressed as mean ± 1 s.d. and/or a median with the 95% confidence interval (CI). Statistical differences were analysed by the Mann-Whitney test. Significance was accepted when P < 0.05.

Effect of inhibitors of distinct signalling pathways on neutrophil ADCC activity stimulated by FMLP and TNF-α
In order to understand if different post-receptor signal transduction pathways underlie neutrophil cytolytic activity in the two model systems, i.e. the Lym-1/TNF-α vs the Lym-1/FMLP system, the effect of various inhibitors was studied. GST, an inhibitor of tyrosine kinase (Rollet et al, 1994), staurosporin (STP) and H-7, which have been shown to affect the activity of various protein kinases (Ginis and Tauber, 1990) and WMN, which inhibits both phosphatidylinositol-3-kinase and phospholipase D (Vlahos et al, 1995), were found to equally suppress neutrophil activity in Lym-1/TNF-α and Lym-1/FMLP system (Figure 4). Moreover, pertussis toxin (PT, 2 µg ml -1 ) was found to equally inhibit TNF-α-and FMLP-exposed neutrophils (% Cr 51 release, TNF-α system: 16.00 ± 9.47 and 2.78 ± 4.54 in the absence and presence of PT respectively, mean ± 1 s.d., n = 6, P = 0.0152; FMLP system: 10.73 ± 7.1 and 0.01 ± 0.04 in the absence and presence of PT respectively, mean ± 1 s.d., n = 6, P = 0.0022). On the contrary, BIM, an inhibitor of protein kinase C (Toullec et al, 1991), suppressed the lysis induced by Lym-1/FMLP-activated neutrophils without affecting the lysis induced in parallel experiments by Lym-1/TNF-α ( Figure 5). The cytolytic activity of Lym-1/TNF-α-stimulated neutrophils was also unaffected by a dose of BIM ten times higher than that used in experiments reported in Figure 5 (data not shown).

DISCUSSION
A preliminary clinical trial with Lym-1 intravenous infusion, carried out in some patients with refractory lymphoma, showed an evident reduction of lymph node size only in some cases (Hu et al, 1989). Although a number of factors can contribute to these partial responses, the inadequacy of host immune effector systems is likely to play a relevant role. In order to improve Lym-1 antibodybased therapeutic approaches, it is therefore critical to understand whether cell-mediated cytolysis can be enhanced by biological response modifiers. The present study shows that mAb Lym-1, per se ineffective or endowed with a very low activity, interacts synergistically with FMLP or TNF to trigger neutrophil ADCC towards B lymphoblastoid tumour targets. These findings confirm and extend our initial observations (Ottonello et al, 1996). Moreover, the present data suggest that the two systems, i.e. FMLP-and TNF-α-dependent Lym-1 ADCC, share a variety of characteristics including certain receptor and post-receptor requirements. These findings may be a starting point to develop anti-tumour immune reactants, such as anti-tumour mAbs conjugated with TNF-α or FMLP to be considered for in vivo administration (Obrist et al, 1991).
The two cytolytic systems appear to be strictly dependent on the intervention of FcγRII without the involvement of FcγRIII. This directly proves the actual role of FcγRII as cytolytic trigger in neutrophil mAb-dependent ADCC. In fact, this type of receptor was identified as major trigger molecule for neutrophil ADCC by using hybridoma target cells expressing antibodies to various neutrophil surface antigens (Graziano and Fanger, 1987;Elsässer et al, 1996). In this system, neutrophils were indeed able to lyse hybridoma cells expressing antibodies to FcγRII but not those bearing mAb specific for FcγRIII (Graziano and Fanger, 1987;Elsässer et al, 1996). Consistent with our findings, FcγRII but not FcγRIII were found to cluster at the effector-target interface during neutrophil ADCC towards sheep erythrocytes (Petty et al, 1989). Moreover, Valerius and co-workers provided evidence for a crucial role for FcγRII in the lysis of mAb sensitized glioblastoma cells by normal neutrophils (Valerius et al, 1993), whereas Repp and coworkers described the intervention of this type of receptor in neutrophil-mediated lysis of Daudi lymphoma cells opsonized with specific rabbit anti-serum (Repp et al, 1991). On the other hand, the participation of additional cell surface molecules is required for optimal activity, as suggested by the ability shown herein of an anti-CD18 mAb to suppress Lym-1 ADCC. This indicates the intervention of β 2 integrins, generally thought to strengthen adhesion between effector and target cells. The finding is consistent with previous evidences for β 2 integrins intervention in neutrophil ADCC systems carried out with polyclonal antitarget antibodies Kohl et al, 1984). Finally, our data are in agreement with those of other authors (Kushner and Cheung, 1992) showing the requirement for CD18 integrins in neutrophil-mediated mAb-dependent lysis of tumour cells. Using melanoma and neuroblastoma cell lines as targets, Kushner and Cheung have shown that mAb-ADCC by neutrophils requires both FcγRII and FcγRIII Cheung, 1989, 1992). Also, other authors have proposed a role for FcγRIII in neutrophil mAb-mediated tumour lysis (Gavioli et al, 1991). No final explanation for the discrepancies about the role of FcγRIII in ADCC between these findings and our present conclusions is available. Nevertheless, it is of note that, in front of the incapacity of anti-FcγRIII F(ab´) 2 fragments to inhibit neutrophil Lym-1 ADCC, the same but entire anti-FcγRIII mAb (3G8) inhibited the lysis efficiently. This is consistent with the recently shown ability of 3G8 mAb to block the ligand-binding site of FcγRII with its Fc portion (Flesch et al, 1997). Therefore, the inhibition of ADCC by 3G8 mAb observed by Cheung (1989, 1992) might reflect the blockade of FcγRII. Similarly, neutrophil-mediated FcγRII-dependent phagocytosis was found to be susceptible to inhibition by native mAb 3G8 (Flesch et al, 1997). On the other hand, it is known that chemoattractant-stimulated neutrophils undergo shedding of FcγRIII (Huizinga et al, 1990;Tosi and Zakem, 1992), a phenomenon balanced at least in part by a concomitant translocation of receptors from intracellular storage compartments (Tosi and Zachem, 1992). In substantial agreement with these studies, the prolonged exposure to FMLP or TNF-α resulted in a partial down-regulation of the FcγRIII expression. Nevertheless, owing to the considerable levels of FcγRIII expression even after 14 h incubation with FMLP and TNF-α, i.e. ~ 40% of the values detectable on cells incubated in medium alone, it seems unlikely that a stimulus-induced loss of FcγRIII can account for the herein found inability of this receptor to play a role in Lym-1 ADCC.
It has been shown that FcγRI, inducible in neutrophils by interferon gamma (IFN-γ) and granulocyte colony-stimulating factor (G-CSF) (Buckle and Hogg, 1989;Repp et al, 1991;Elsässer et al, 1996;Valerius et al, 1997) is effective in activating neutrophil lytic potential, as demonstrated in reverse cytotoxic assays against anti-FcγRI mAb-producing hybridoma targets (Elsässer et al, 1996) and in tumour cell lysis mediated by bi-specific mAbs with one specificity for FcγRI (Elsässer et al, 1996;Valerius et al, 1997;Würflein et al, 1998). In agreement with other authors (Buckle and Hogg, 1989), neutrophils from healthy donors did not express FcγRI. This receptor was, however, detected after 14-h incubation, although the amount was very low, confirming previous findings (Buckle and Hogg, 1989). Nevertheless, neither FMLP nor TNF-α were found to affect neutrophil FcγRI expression and, on the other hand, neither FMLP-nor TNF-α-stimulated Lym-1 ADCC were inhibited by the anti-FcγRI mAb 197. Therefore, this Fc receptor has no role under our conditions. This consistent with the known inability of normal neutrophils to lyse anti-FcγRI mAb-expressing hybridoma cells (Elsässer et al, 1996).
Since FMLP and TNF-α were herein found to be devoid of effects on FcγRII expression but capable of triggering FcγRIImediated ADCC, the present data are consistent with the possibility that both FMLP and TNF-α act on post-Fc receptor signal transduction systems. It is well-known that post-receptor intracellular signalling pathways leading to specific neutrophil functional responses involve various kinases, phospholipases, calcium and certain signal-transducing proteins such as, for instance, Gproteins (Nishizuka, 1995). The inhibitory activity of pertussis toxin in the FMLP-ADCC system can conceivably involve inhibition of toxin-sensitive and FMLP receptor-coupled G-proteins (Snyderman and Uhing, 1992). As neither TNF-α-dependent cell stimulation nor β 2 -integrin signalling are known to involve pertussis toxin-sensitive pathways (Dinarello, 1992;Hynes, 1992), the inhibitory activity of the toxin in the TNF-α-dependent ADCC system might be attributed to the blockade of the pertussin toxinsensitive src-like tyrosine kinase fgr pathway which has been previously shown to be associated with the FcγRII signal transduction (Hamada et al, 1993;Zhou et al, 1995). This FcγR has been indeed shown to initiate transmembrane signals that can involve pertussis toxin-sensitive pathways (Gresham et al, 1987;Feister et al, 1988). On the other hand, the protein kinase C inhibitor BIM was found to suppress FMLP-but not TNF-α-dependent ADCC, suggesting that it selectively interferes with FMLP signal transduction. On the contrary, other chemicals, including inhibitors of tyrosine kinase and phosphatidylinositol-3-kinase were equally effective in the FMLP-and TNF-α-system, suggesting that the two ADCC conditions share common activating circuits.
In conclusion, taking into account the present observations, Lym-1 ADCC can be envisaged as a process involving the coordinated intervention of various neutrophil receptors. In other terms, signals delivered by FMLP or TNF, BIM-sensitive and insensitive respectively, converge and synergize with those from G-protein coupled FcγRII and presumably β 2 -integrins to induce the activation and expression of the neutrophil cytolytic potential.