Monoclonal Lym-l Antibody-Dependent Lysis of B-Lymphoblastoid Tumor Targets by Human Complement and Cytokine-Exposed Mononuclear and Neutrophilic Polymorphonuclear Leukocytes

Lym-1 is a murine IgG2a monoclonal antibody that recognizes a polymorphic variant of HLA-DR antigens on malignant B cells, with minimal cross-reactivity with normal tissues.  Because it can be safely administered in vivo, a detailed knowledge of its ability to recruit and trigger the antitumor immune effector systems is required to optimize potential serotherapeutic approaches in B-lymphoma patients. By using Raji cells as a model of B-lymphoma targets, we found that Lym-l activates complement-mediated lysis efficiently. Moreover, Lym-l was capable of triggering the antibody-dependent cellular cytolysis (ADCC) by peripheral blood mononuclear cells (MNCs). On the contrary, it failed to trigger neutrophilic polymorphonuclear leukocyte (PMN)mediated ADCC activity. In an attempt to enhance Lym-l ADCC by MNCs and PMNs, nine biologic response modifiers were tested. MNC-mediated Lym-l ADCC was significantly stimulated by interleukin-2 (IL-2) and unaffected by other

Lym-1 is a murine IgG2a monoclonal antibody that recognizes a polymorphic variant of HLA-DR antigens on malignant B cells, with minimal cross-reactivity with normal tissues. Because it can be safely administered in vivo, a detailed knowledge of its ability t o recruit and trigger the antitumor immune effector systems is required t o optimize potential serotherapeutic approaches in B-lymphoma patients. By using Raji cells as a model of B-lymphoma targets, we found that Lym-l activates complement-mediated lysis efficiently. Moreover, Lym-l was capable of triggering the antibody-dependent cellular cytolysis (ADCC) by peripheral blood mononuclear cells (MNCs). On the contrary, it failed t o trigger neutrophilic polymorphonuclear leukocyte (PMN)mediated ADCC activity. In an attempt t o enhance Lym-l ADCC by MNCs and PMNs, nine biologic response modifiers were tested. MNC-mediated Lym-l ADCC was significantly stimulated by interleukin-2 (IL-2) and unaffected by other

YM-l IS
A MURINE IgG2a monoclonal antibody (MoAb) that recognizes a polymorphic variant of HLA-DR antigens present on the surface of B-lymphoma cells and incapable of shedding or undergoing modulation after antibody binding.' It has a relatively low reactivity with normal tissues and cells, including normal B lymphocytes.' Moreover, radioimaging in vivo studies with '231-labeled Lym-l have shown antibody localization to sites of lymphomatous disease.' Finally, by using Raji cells as a model of B-lymphoma target cells, Lym-l was found to trigger the cytolytic activity of complement3 and mononuclear cells (MNCS).~ Owing to these properties, Lym-l is presently considered an attractive candidate for developing MoAb-based therapies of B lymphomas.
A preliminary clinical trial with Lym-l intravenous infusion, performed in 10 patients with refractory lymphoma, showed an evident reduction of lymph node size only in some cases.3 Although a number of factors can contribute to these partial response^,^ the inadequacy of host immune effector systems is likely to play a relevant role. To improve Lym-l antibody-based therapeutic approaches, it is therefore critical to understand whether cell-mediated cytolysis can be enhanced by biologic response modifiers. In this regard, two findings are promising, ie, the capacity of interleukin-2 (IL-2) and y-interferon (7-IFN) to augment Lym-l antibodydependent cytolysis by M N C S~.~ and neutrophilic polymorphonuclear leukocytes (PMNS),~" respectively.
In the present study, we first examined the capacity of Lym-l to activate the lytic potential of the complement system, peripheral blood MNCs and PMNs. Then, in the attempt to increase the cell-mediated cytolytic activity, we tested a panel of nine biologic response modifiers having specific receptors on leukocytes.'"'

MATERIALS AND METHODS
Culture medium and reagents. The following culture medium was used: RPM1 1640 (Irvine Scientific, Santa Ana, CA) supplemented with 10% heat-inactivated (56°C for 45 minutes) fetal calf serum (FCS; HyClone Europe Ltd, Cramlington, NE) and 2 mmol/ Blood, Vol 87, No 12 (June 151, 1996 pp 5171-5178 mediators, including y-interferon (y-IFN), tumor necrosis factor a (TNFa), and granulocyte-macrophage colony-stimulating factor (GM-CSF). On the other hand, PMN-mediated Lym-l ADCC was induced or significantly augmented by various cytokines, such as GM-CSF, TNFru, and yIFN, and chemotaxins, such as formyl peptides (FMLP), complement fragment C5a. and IL-8. Both MNC-and' PMN-mediated ADCC was unaffected by granulocyte colony-stimulating factor (G-CSF) and insulin-like growth factor-I (IGF-l). Finally, only GM-CSF and TNFa augmented the number of PMNs actually engaged in the binding of Raji target cells. The findings presented here, in particular those showing stimulatory activity of biologic response modifiers, may inspire new attempts for developing Lym-l antibody-based approaches t o the therapy of B lymphomas. L glutamine (Irvine Scientific; RPMI-FCS). Hanks' Balanced Salt Solution (HBSS) was from Irvine Scientific. Ficoll-Hypaque was purchased from Seromed (Berlin, Germany). Sodium chromate Cr 51 was from the Radiochemical Center (Amersham, UK). Triton X-100, ethidium bromide, and fluorescein diacetate were purchased from Sigma Chemical CO (St Louis, MO). Heparin was obtained from Roche (Milano, Italy). Giemsa was purchased from Merck (Darmstadt, Germany). Human AB serum, pooled from three normal donors, was stored in aliquots at -80°C immediately after preparation or immediately after heat-inactivation (56°C for 45 minutes). Human CS-deficient serum was purchased from Sigma Chemical CO and stored in aliquots at -80°C. Monoclonal Lym-l antibody was a gift of Prof A.L. Epstein (Los Angeles, CA). A nonsense mouse IgG2a MoAb (Dako S.p.A., Milano, Italy) was used as an isotype control.
Cytokines and chemotaxins. Human recombinant tumor necrosis factor a (TNFa), human recombinant insulin-like growth factor-l (IGF-l), and human recombinant 72 amino-acids IL-8 were purchased from BioSource International (Camarillo, CA). Human recombinant y-IFN was obtained from Genzyme srl (Milano, Italy). Human recombinant C5a and N-formyl-met-leu-phe (FMLP) were purchased from Sigma Chemical Co. Human recombinant IL-2 was a gift of Prof F. Indiveri (University of Genova, Genova, Italy). Human recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF) and human recombinant granulocyte colony-stimulating factor (G-CSF) were kindly donated by Dr G. Guidi (Sandoz Italia, Milano, Italy). Each of these mediators was stored in aliquots 5172 at -80°C and was diluted with culture medium immediately before using.
Heparinized venous blood (10 U/mL heparin) was obtained from healthy volunteers (20 to 37 years old) after informed consent had been obtained. No donor had an infectious disease or was under medication at the time of and for 2 weeks before sampling. MNCs were isolated by centrifugation (400g for 30 minutes) on a Ficoll-Hypaque density gradient, washed with HBSS, and resuspended in RPMI-FCS, as previously described.l The resulting MNC preparations contained 70% to 87% lymphocytes, 13% to 29% monocytes, and less than 1% neutrophils or eosinophils, as determined by morphologic analysis of Giemsastained cytopreps. Cell viability was determined by ethidium bromide-fluorescein diacetate t e d 4 and was greater than 98%. Neutrophilic PMNs were prepared by dextran sedimentation, followed by centrifugation (400g for 30 minutes) on a Ficoll-Hypaque density gradient, as previously described." Contaminating erythrocytes were removed by hypothonic lysis." PMNs resuspended in RPMI-FCS were greater than 97% pure and greater than 98% viable, as determined by the assays described above.
Target cells.
Lymphoblastoid Raji cells'' 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-l antibody was measured by indirect immunofluorescence with flow cytometry using a rabbit antimouse IgG Cytolytic assays. Cytolytic activity of MNCs and PMNs was measured as described elsewhere in detail." Briefly, target cells (2 X IO4) were mixed with MNCs or PMNs at an effector:target ratio of 20: I , with and without Lym-1 MoAb and cytokinedchemotaxins appropriately diluted in RPMI-FCS. The effector:target ratio of 20: 1 was chosen on the basis of preliminary experiments. To test the cytolytic activity of human complement, 2 X IO" cells were mixed with appropriate dilutions of human AB serum or heat-inactivated human AB serum or human CS-deficient serum in the presence or absence of Lym-l or the control IgG2a MoAb. The assays were performed in triplicate and in a final volume of I50 pL, using roundbottom microplates (Falcon, Becton-Dickinson Italia, Milano, Italy). After 14 hours of incubation in a humidified atmosphere of 95% air and 5% COz, the "Cr-release was determined in the cell-free supernatants. A series of experiments was also performed by preincubating (1 hour) either effector cells or target cells with biologic response modifiers. Some experiments were also performed using 20 hours of incubation. The percentage of cytolysis was calculated according to the formula 100 X (E -S)/(T -S), in which E is the cpm released in the presence of effector cells (or complement), T is the cpm released after lysing target cells with 5% Triton X-100, and S is the cpm spontaneously released by target cells incubated with medium alone (< 18%; with 4 exceptions: 20.4%, 2 1 . 4 7~~ 22. I %, and 25.7%).
Target cell binding assay.
The assay was performed under conditions similar to those used for cytolytic assays. Briefly, 10' effector cells were mixed with 2 X lo4 Raji cells in the presence of Lym-l MoAb, with or without cytokines or chemotaxins. This effector:target ratio was chosen on the basis of preliminary experiments (higher effectorxarget ratios lead to cell overcrowding, therefore hampering the reading). Tests were performed in polypropylene 1.5 rnL capped Statistical analysis.
Results were expressed as the mean -+ 1 SD and/or as median with the 95% confidence interval. Statistical differences were analyzed by the Mann-Whitney test. Significance was accepted when P < .05.

RESULTS
When added to "Cr-labeled Raji cells in the presence of 10 pg/mL Lym-l, human serum induced a dose-dependent lysis of Raji target cells as measured by a 14-hour "Cr release assay (Fig 1). The magnitude of human serum-mediated lysis was also dependent on the Lym-l concentration (Fig 2). These data suggest that (1) 10 pg/mL Lym-l and about 20% human serum are sufficient to induce the lysis of the majority of Raji target cells; and ( 2 ) relatively low concentrations of Lym-l (0.1 to 1 pg/mL), combined with 5% to 10% of human serum, are still capable of inducing substantial levels of cytolysis. Time-course experiments showed that the lysis reaches the plateau after 1 hour, ie, the serum-mediated Lym-1 antibody-dependent cytolytic reaction is over at the end of I hour of incubation (data not shown). Heat-inactivated (56°C for 45 minutes) human serum, added to Raji cells in the presence of 10 pg/mL Lym-1, was ineffective (percentage of cytolysis by 5% heat-inactivated serum, 0.7 t 0. MNCs caused negligible lysis (Fig 3). The addition of IO pg/mL Lym-1 resulted in significant cytolysis, ie, MNCs appear to mediate Lym-l ADCC (Fig 3 and its  legend). On the other hand, PMNs were incapable of mediating significant spontaneous and Lym-l antibody-dependent cytolysis (Fig 3 and its legend). As shown in Fig 3, there was intersubject variability in MNC-mediated ADCC, apparently unrelated to the number of monocytes in the MNC populations used (data not shown). Moreover, PMNs from relatively few individuals were found to exert ADCC (Fig 3). Nevertheless, repetitive testing of same donors at various times indicated that the majority had consistent profiles of both MNC-and PMN-mediated ADCC, ie, high responders usually displayed high activity, whereas low responders generally had low activity (data not shown). As summarized in  (Table 1). Although each of these agents was found to enhance ADCC by MNCs from one or more donor, none of them also had significant activity (Table 1) after preincubations with MNCs or target cells (data not shown). Because yIFN and GM-CSF have been previously claimed to enhance Lym-l -dependent MNC cytolysis in a 20-hour assay,6.I6 parallel experiments OTTONELLO ET AL were performed using 14-and 20-hour incubation periods. Both 100 U/mL yIFN and 1 ng/mL GM-CSF also did not significantly stimulate MNC-mediated Lym-l ADCC in the 20-hour assay (percentage of stimulation of MNC ADCC by during 14-and 20-hour incubation periods, respectively; percentage of stimulation of MNC ADCC by GM-CSF, 13.0 ? 6.1 [ P = ,2221 and 9.5 +-8.9 [P = S471 during 14-and 20hour incubation periods, respectively [mean -+ 1 SDI; n =

5).
As shown in Figs 5 and 6, two cytokines (GM-CSF and TNFa) and two chemotaxins (FMLP and C5a) induced or enhanced PMN-mediated Lym-l ADCC significantly (legends to Figs 5 and 6). When added to PMNs plus Raji cells in the absence of Lym-l or to Raji cells incubated with Lym-1 in absence of PMNs, they had no effect on the 5'Cr release from target cells. Stimulation of PMN-mediated Lym-l ADCC by GM-CSF and FMLP was also detected using Lym-1 concentrations less than than 10 pg/mL (Fig 7). The same phenomenon could be observed using TNFw and C5a (data not shown). As for MNC-mediated Lym-l ADCC, other agents were tested for their ability to increase the target lysis by PMNs (Table 2). Among them, yIFN and IL-8 significantly augmented PMN-mediated Lym-1 ADCC (Table 2), but the magnitude of the target lysis was relatively low as compared with that observed using GM-CSF, TNFa, FMLP, and C5a. Other cytokines were completely ineffective, with low intersubject variability (Table 2), and were unable to stimulate PMN-mediated Lym-1 ADCC even after preincubation with PMNs or Raji cells (data not shown). As   Lym-l was also found to trigger MNC-mediated lysis, an event enhanced by IL-2. This finding, coupled with the incapacity of IL-2 to induce LAK activity in the present setting, suggests that the signals delivered by Lym-l and IL-2 converge and synergize to amplify the ADCC activity of MNCs. None of the other cytokines herein tested, including yIFN and GM-CSF, was found to enhance Lym-l ADCC by MNCs, even during a 20-hour assay and after 1 hour of preincubation with either effector or target cells. Therefore, our data do not confirm theleis, two cytokines (GM-CSF and TkFa) and two chemotaxins (FMLP and C5a), per se ineffective in absence of Lym-l, were found to induce Lym-l ADCC by PMNs. Percentage of cytolysis at 20:l PMN:Raji cell ratio. Lym-l concentration was 10 pg/mL. Incubation time was 14 hours.
Moreover, y-IFN and IL-8 displayed significant but very low PMN stimulatory activity. As for the IL-2 enhancement of Lym-l ADCC by MNCs, these results suggest the intervention of a synergistic interaction between Lym-l and each of the aforementioned biologic mediators. However, at present, we have no hint as to the detailed biochemical processing whereby these distinct mediators exert the same stimulatory action. Nevertheless, the enhancement of PMN ADCC by GM-CSF and TNFa is associated with an augmentation of the number of target cell-bound effectors, whereas other mediators (FMLP, C5a, y-IFN, and L-8) stimulate ADCC of bound PMNs without affecting the effector-target conjugate formation. Finally, it is of note that, whereas the results obtained with y-IFN are confirmatory,6 those observed with GM-CSF, TNFq FMLP, and C5a are novel. In fact, to our knowledge, only GM-CSF was previously found to stimulate the MoAb-dependent tumoricidal activity of PMNs, using melanoma, neuroblastoma, or certain colorectal carcinoma target The ability of Lym-l shown here to mediate complementdependent cytolysis efficiently, coupled with its particular reactivity for lymphoma cells,' suggests its possible use for the purging of harvested bone marrow before reinfusion. On the other hand, preliminary clinical studies with Lym-l intravenous infusion in patients with refractory lymphomas have shown low response rates? suggesting that the in vivo activation of complement-and MNC-mediated cytolysis by Lym-l is relatively inefficient as compared with in vitro findings. Nevertheless, the administration of Lym-l might be effective in patients with a low tumor burden. As shown in biopsy specimens, Lym-l can trigger a sort of inflammatory or immune response at tumor sites, at least in a subset of lymphoma patients3 Similarly, certain antitumor and complement-activating MoAbs have been found to induce tumor-destructing inflammatory reactions in melanoma patients." Therefore, although radiolabeled or toxin-conjugated Lym-l may also be effective,23 the possibility of augmenting the activity of immune cells by using biologic response modifiers appears to be a reasonable option to improve Lym-l antitumor effects. Taking into account the intersubject variation of lymphoma cell ability to bind Lym-1,' the combination of the antibody with IL-2 to increase MNC-ADCC might specifically target and potentially eradicate small numbers of residual tumor cells in selected patients. Moreover, the capacity of certain cytokines (GM-CSF and TNFa) and chemotaxins (FLMP and C5a) to trigger PMN-mediated Lyml ADCC raises potentially attractive possibilities to develop new approaches to the serotherapy of lymphomas. Although the infusion of cytokines or other mediators has its own set of problems, the known toxicity of intravenous administration of TNFa,"." the ability of IL-2 to promote the production of TNFcI,~~ and the expected unwanted effects of FMLP and C5a26 might be reduced by conjugation of these molecules with the antibody. As far as GM-CSF is concerned, it has been recently shown that this cytokine potentiates Lym-l ADCC by macrophages." This finding and the present data, coupled with the ability of GM-CSF to increase monocyte-macrophage production28 and PMN production and s u r~i v a 1 , 2~~~~ render the cytokine the best candidate for inducing phagocytes to express their Lym-1 -dependent antitumor potential. Consistent with this possibility, the administration of an antitumor MoAb together with GM-CSF has yielded encouraging clinical responses in patients with colorectal car~inoma.~'