Antibody-Dependent, Cell-Mediated Cytotoxicity by an Anti-Class II

Home , Cytotoxicity

[CANCER RESEARCH 50. 2991-2996. May 15. 1990| Antibody-dependent, Cell-mediated Cytotoxicity by an Anti-Class II Murine Monoclonal Antibody: Effects of Recombinant Interleukin 2 on Human Effector Cell Lysis of Human B-Cell Tumors William C. Biddle,1 James Pancook, Martin Goldrosen, Tin Han, Kenneth A. Foon, and Louis Vaickus

Divisions of Clinical Immunology [W. C. B., J. P., M. G., K. A. F., L. V.Â¡andHematological Oncology Â¡T.H.], Roswell Park Cancer Institute, Buffalo, New York 14263

ABSTRACT exists as to which effector cell type is most effective at ADCC (4, 5). This is partly attributable to the wide spectrum of Lym-1 is an IgG2a murine monoclonal antibody that reacts with variant antibodies (and species of origin) and target cells used in in Class II molecules expressed on B-cell malignancies. Lym-1 was shown vitro assays. A considerable amount of our present knowledge to mediate antibody-dependent cellular cytotoxicity (ADCC) of human concerning ADCC is based on previous studies using heterolo- effector cells against a variety of malignant B-cell lines. Tumor cell lysis was Lym-1 specific because (a) the reaction was dose dependent with gous antisera, irrelevant targets (e.g., chicken red blood cells), significant ADCC detectable at Lym-1 concentrations as low as 1 MH/nil: or nonhuman effector cells, and direct extrapolation to the (b) tumor targets not expressing the Lym-1 antigen were unaffected; (c) human system may not be valid. Murine MoAbs of the IgG2a an isotype-matched irrelevant monoclonal antibody and an IgGl anti- and IgG3 subclass are felt to be the best mediators of ADCC Class II monoclonal antibody failed to mediate ADCC; and (d) addition (6-8). There is no a priori method to determine MoAb activity of Protein A (which binds avidly to Lym-1) blocked ADCC by 90 to in ADCC based on isotype alone, however. In fact, not all 100%. Peripheral blood mononuclear cells obtained from normal donors IgG2a MoAbs mediate ADCC, and their functionality may be as well as from cancer patients were able to interact with Lym-1 to elicit based on antigen density, kinetics of antigen modulation, and ADCC. Recombinant interleukin 2 (rIL-2) enhanced non-antibody-me interaction with Fc receptors, among others. Indeed, it has been diated tumor lysis and Lym-1 ADCC with an optimal concentration of recently reported that IgG2a and IgG3 antiidiotype MoAbs 100 units/ml. Pulse treatment of normal peripheral blood mononuclear cells with rIL-2 was able to augment Lym-1 ADCC but was less effective were unable to mediate ADCC against lymphoma targets with than having the rIL-2 present through the assay. Peripheral blood human effector cells (9). mononuclear cells obtained from patients being treated with high doses IL-2 has been reported to enhance effector cell function in of rIL-2 administered by continuous i.v. infusion demonstrated Lym-1 ADCC following in vivo or in vitro administration in the murine ADCC levels which were higher than normal individuals and which were system (10, 11). Similarly, in vitro exposure to IL-2 has been further augmented by in vitro incubation with rIL-2. shown to result in increased ADCC in human model systems (12). This is, however, not a universal finding. Treatment with IL-2 was shown to increase non-antibody-mediated cytotoxicity INTRODUCTION by human effector cells, but ADCC against bladder carcinoma Lym-1 is an IgG2a Mo Ab2 that recognizes variant DR mol cells was not affected (13). Thus, it is critical to evaluate each ecules present on a variety of B-cell lymphomas and leukemias. new therapeutic MoAb individually to determine whether it can Expression of the Lym-l-Ag appears to be highly restricted in mediate ADCC and to learn whether this reaction can be that Lym-1 reacts with only a limited number of normal cells augmented by recombinant cytokines. This knowledge may help and with lower avidity as compared with tumor cells. The Lym- to optimize the clinical utility of MoAbs used in their uncon l-Ag does not appear to modulate nor be shed into the circu jugated or conjugated forms. lation (1). These attributes make Lym-1 an attractive candidate MoAb for immunotherapy. Previous trials using unaltered MoAbs against B-cell malig MATERIALS AND METHODS nancies have met with limited success (reviewed in Ref. 2). In Human Cell Lines and Culture Conditions. The human Burkitt's contrast, successful treatment of xenografted human tumors in lymphoma cell lines Raji, Jijoye, and EB3 and the K562 erythroleuke- murine models has been demonstrated using unconjugated mia cell line were obtained from the American Type Culture Collection MoAbs and appeared to be largely dependent upon the anti (Rockville, MD). The cell lines designated MO 1011, MO 1018, MO body's ability to mediate ADCC (3). Improvements in the 1046, and MO 2007 were established from tumor cells obtained from clinical utility of MoAbs may be attained by further understand patients with B-cell CLL and PLL, respectively. These cell lines were ing this antitumor mechanism. Although much is known about generated by Epstein-Barr virus transformation and retain many of the ADCC interactions, many areas are still unresolved. phenotypic, molecular, and cytogenetic characteristics of the original tumor cell (14). They were kindly provided by Dr. Tin Han, Roswell In general, the ability to lyse tumor cells depends on the Park Cancer Institute, Buffalo, NY. Cell lines were cultured in CM MoAb subclass, type of effector cell, degree of antigen expres consisting of RPMI 1640 medium, 10% heat-inactivated fetal bovine sion, and susceptibility of the target cell to lysis. Controversy serum, and standard supplements. Cell lines were maintained at 37Â°C

Received 9/20/89; revised 1/10/90. in a humidified 95% air, 5% CO2 environment. All tissue culture The costs of publication of this article were defrayed in part by the payment reagents were obtained from GIBCO, Grand Island, NY. of page charges. This article must therefore be hereby marked advertisement in Monoclonal Antibodies. The Lym-1 monoclonal antibody was sup accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1To whom requests for reprints should be addressed, at Clinical Immunology, plied by Dr. F. DÃ¼rr,Lederle Laboratory, Division of American Cy- Roswell Park Cancer Institute. Elm and Carlton Streets, Buffalo, NY 14263. anamid Corporation, Pearl River, NY. MoAb L227 (IgGl) is reactive 1The abbreviations used are: MoAb. monoclonal antibody; ADCC, antibody- with nonpolymorphic DR, DP, and DQ molecules, whereas MoAb dependent cellular cytotoxicity; CM, complete medium; FcR, Fc receptor; FITC, L243 (IgG2a) recognizes nonpolymorphic DR (15). fluorescein Â¡sothiocyanate:LU. lytic unit; LAK. lymphokine-activated killer cell; An isotype matched (IgG2a) control MoAb (UPC-10) was obtained Lym-l-Ag. Lym-1 antigen; PBMC. peripheral blood mononuclear cells; rIL-2. recombinant interleukin 2; SpA, Staphylococcus aureus Protein A; IL-2, interleu from Sigma Chemical Corp., St. Louis, MO. All antibodies were used kin 2; PLL, prolymphocytic leukemia; CLL. chronic lymphocytic leukemia; E:T, as purified immunoglobulin in phosphate-buffered saline. MoAb purity effectortarget. was determined to be greater than 90% by sodium dodecyl sulfate- 2991

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1990 American Association for Cancer Research. EFFECTS OF rlI.-2 ON HUMAN EFFECTOR CELL LYSIS OF B-CELL TUMOR polyacrylamide gel electrophoresis. Antibodies were sterilized by filtra effectors required to cause 33% target lysis was extrapolated (1 lytic tion through 0.2-/JIT1filters prior to use. unit). Results are expressed as lytic units (LU33) per IO7effector cells Biological Response Modifier Treatment Protocols. The 1L-2/LAK (17). protocols consisted of continuous i.v. infusion of rIL-2, RO 23-6019 Protein A Inhibition of Lym-1 ADCC. SpA and FITC-SpA were (Hoffmann-LaRoche, Nutley, NJ), 6 x IO6 units/m2 daily for 5 days purchased as lyophilized powders (Sigma Corp.). Lym-1 (5 Mg) was (Days 1 to 5), then daily leukapheresis for 3 days (Days 7 to 9), and used to saturate the Lym-1-Ag binding sites on 2.5 x IO5 Raji cells by infusion of IL-2 and LAK cells for 3 days followed by IL-2 alone for 2 incubation on ice for 20 min. After washing, varying amounts of FITC- days (Days 10 to 14). Cells from Patient 1 were obtained on Day 13 of SpA were added for an additional 20 min. Binding of FITC-SpA to Cycle 1; Patient 4 on Day 7 of Cycle 1; Patient 5 on Day I of Cycle 3; Lym-1 was determined by flow cytometry based on mean log fluores and Patient 7 on Day 14 of Cycle 1. The IL-2/a-interferon protocol cence and the percentage of positive cells as compared with control consisted of concomitantly administered rIL-2 at 3 x IO6units/m2/day MoAbs. This method was used to determine a saturating concentration continuous i.v. infusion with a-interferon (Roferon-A, RO 22-8I8I/ of SpA. A 4-fold higher concentration of unlabeled SpA (20 ^g) was 002; Hoffman-LaRoche) at 5 x IO6units/m2/day by i.m. or s.c. injec added in the ADCC assays to determine whether binding of SpA to the tion, both for 4 days. Cells from Patient 2 were obtained after 3 Fc portion of Lym-1 could inhibit Lym-1 ADCC. complete cycles; Patient 3 after 2 cycles. The IL-2/cyclophosphamide Statistical Analysis. Healthy controls and cancer patients were com protocol consisted of cyclophosphamide (400 mg/m2) administered by pared using a Mann-Whitney test. Levels for various combinations of bolus i.v. infusion (prior to the first and third infusion of IL-2) followed effector cells, rIL-2, and Lym-1 were compared using a Wilcoxon signed by rIL-2 (3 x IO6 units/m2/day continuous i.v. infusion for 4 days. rank test. In all cases, a P value of 0.05 or less was considered Cells from Patient 6 were obtained on the third day of rIL-2 therapy statistically significant. during the second cycle. Effector Cell Preparation. After informed consent was obtained, peripheral blood was collected by venipuncture from normal volunteers RESULTS and from cancer patients enrolled in clinical protocol studies. Heparin- Determination of Optimal Lym-1 and IL-2 Concentrations. ized blood samples were subjected to density gradient separation on PBMC were initially tested in the presence of varying concen Ficoll-Hypaque (specific gravity, 1.007; Pharmacia, Piscataway, NJ) trations of Lym-1 for lytic activity against Raji tumor cells in according to the method of Boyum (16). Briefly, blood was diluted 1:1 with RPMI 1640 and layered on Ficoll-Hypaque at a 2:1 ratio and 16-h ADCC assays (Fig. iA). Compared with non-antibody- centrifuged at 800 x g for 20 min at 4Â°C.Themononuclear cell interface mediated cytotoxicity, Lym-1 was effective (from 0.1 to 30 fig/ was removed and washed twice in RPMI 1640. Cells were resuspended ml) in increasing both the percentage of cytotoxicity and the in CM and counted, and viability was determined by trypan blue LU33 of these effector cells. Lym-1 treatment alone in the exclusion after lysis of RBC by treatment with 0.4% acetic acid. absence of effector cells showed negligible cytotoxicity (not Viability was always greater than 95%. shown). In the absence of Lym-1, effector cells at the highest Target Cell Labeling and Preparation. Target cells were kept in log E:T ratios tested (40:1) routinely lysed less than 15% of target growth phase before use and used only if viability was greater than cells. Based on these findings a Lym-1 concentration of 10 fig/ 95%. Chromium labeling was done by adding 100 Â¿iCiof"Cr in saline ml was chosen for all subsequent studies. (specific activity, 250 to 500 mCi/mg of chromium; Amersham Corp., Arlington Heights, IL) to 5 x IO6 target cells suspended in an equal rIL-2 was tested for its ability to enhance Lym-1 ADCC. rlL- 2 was shown to be effective in augmenting Lym-1 ADCC in the volume of 50 HIMTris buffer containing 1 itiM Na2PO4, 300 mM NaCl, and 10 mM KC1 (pH 7.4). After l h at 37Â°C,target cells were washed range of 1 to 1000 units/ml (not shown). rIL-2 treatment at all twice with RPMI 1640 and resuspended in CM. dosages tested (1 to 1000 units/ml) in the absence of effector rIL-2 Treatment. Effector cells were treated for 3 h prior to ADCC cells failed to exhibit any cytotoxicity against any target cells assay with rIL-2 at concentrations ranging from 1 to 1000 units/ml. tested. rIL-2 (100 units/ml) was determined to be optimal and From these experiments it was determined that the optimal dose of shown to enhance Lym-1 ADCC at all antibody concentrations rIL-2 was 100 units/ml. This concentration was used in all subsequent tested (Fig. \B). This effect was particularly evident at Lym-1 experiments. In some experiments, rIL-2 was removed after the 3-h concentrations between 1.9 and 30 Mg/m' w'th approximately pretreatment by washing the effector cells twice with CM prior to assay. a 4 to 5-fold enhancement of lytic activity compared with these This was done to determine if a relatively brief exposure to rIL-2 prior same Lym-1 concentrations without IL-2 (Fig. \A). to the ADCC reaction was adequate for optimal lysis. Both of these rIL-2 treatments were compared in 4- and 16-h ADCC assays. Effector Effect of rIL-2 Exposure Time on Cytotoxicity. Lytic activities cells receiving no IL-2 were treated in a similar fashion in all experi of normal donor PBMC in 4- and 16-h cytotoxicity assays were ments. compared to determine the optimal conditions for cytotoxicity. ADCC Assay. 5lCr-labeled target cells (1 x 10") were added to round- Prolonged exposure (16 h) between PBMC and Raji target cells bottomed microtiter trays at E:T ratios ranging from 40 to 2.5:1. resulted in significantly higher cytotoxicity (Table 1). Both MoAbs diluted in CM were added where appropriate at the desired Lym-1 ADCC and non-antibody-mediated cytotoxicity were concentrations. Tumor cell lysis was measured in both 4- and 16-h increased in all groups 2 to 4 times by prolonged incubation of assays. Release of 5'Cr was determined by absorption of culture medium PBMC effector cells with Raji cell targets. Statistically signifi onto harvesting frame filters (Skatron Corp., Sterling, VA) and count cant differences were observed between effector cells treated ing in a gamma counter (Beckman Model 5500b). All samples were with Lym-1 and/or rIL-2 and incubated with targets for either done in triplicate, and the mean values were used to calculate results. 4 or 16 h. Lym-1 was capable of mediating ADCC in 4-h assays, The net difference between targets receiving 1% Triton X-100 (maximal and its effects were enhanced by rIL-2 treatment. However, release) and spontaneous release (no treatment) was used as the denom inator for assessing lytic activity. The percentage of specific lysis was optimal activity was shown by 16-h incubation. Although rlL- determined by the following equation 2 could upregulate non-antibody-mediated cell lysis in 4 h and by 3-h pretreatments of effector cells, it was most effective with E â€”S % of specific lysis = â€” x 100 16 h of continuous treatment. Similarly rIL-2 effects on Lym- M â€”Ã> 1 ADCC were most marked with the longer exposure time, and where M is maximal release in cpm, S is spontaneous release, and E is significant (P < 0.05) increases in lysis were observed. A similar experimental release. For the calculation of LU, specific lysis percent increased susceptibility to PBMC lysis following 16-h assays age values from at least 5 E:T ratios were used. A linear regression line was observed with other malignant B-cell target lines (not of best fit for the killing curve was calculated, and the number of shown). Thus, in subsequent experiments a 16-h ADCC assay 2992

â€¢(60 Fig. 1. Lym-1 ADCC is dose dependent and enhanced by the addition of rIL-2. Lym-1 -m (0.03 to 30 Mg/ml) Â¡n'he absence (A) or in the Tg presence (B) of 100 units/ml of rIL-2 was -,O ,40 . tested for ADCC activity in 16-h assays against Raji tumor cells. D, percentage of cytotoxicity; 150 â€¢LU33/107 cells. Lym-1 alone in the absence ,20-R Â»?50 of effector cells showed negligible activity. Ef a40 fector cells alone (no Lym-1) routinely resulted .00--" g in less than 15'i lysis of target cells (A). All IO Ã–.â€¢ IO 40â€¢<"302010 results shown are from experiments using ef- |302010 80(-6040 fectortarget ratios of 40:1 and are the mean values of triplicate samples representative of several experiments (n = 3).

-20

-1â€¢â€¢,OO9O8060

30 75 1905 0201 O03 O 75 19 05 0.1 003 O Lym-1 (/Â¿g/mL) Lym-1 (/Â¿g/mL)

Table I Comparison of 4- and Â¡6-hADCC assays Results shown are the mean lytic unit (LU 33/107 cell) values (n = 8) comparing the same normal individual PBMC activity in 4- and 16-h ADCC assays against Raji target cells. Effector cells with or without Lym-1 (10 ^g/ml) received rIL-2 (100 units/ml) for only 3 h prior to assay or were pretreated for 3 h. and rIL-2 remained continuously during the assay. Treatment with Lym-1 or rIL-2 alone or in combination in the absence of effector cells resulted in no lytic activity. Statistical analysis of various groups is shown in parentheses with a P value of 0.05 or less considered statistically significant. LU 33/107 cells

Effector cells 4-h assay 16-h assay (b)45.1 < 0.05 for a and b) No treatment (a) c)(P<0.01< 0.05 for a and d)(P<0.01 for band Lym-13-h 19.7(c)3.3 (d)13.5(f) (P<0.05d)(P fore and rIL-2 < 0.05 for e and f ) (e) g)(/><0.01for e and h)(P<0.0\ for fand 3-hrlL-2 +Lym-1Continuous 43.3(g)6.6 92.6(h)28.8 (f<0.05h)(P for g and rIL-2 < 0.05 for i and j) (i) for Â¡andk)2.0 (j) forj and 1)(P Continuous rIL-2 + Lym-10.1 58.9 (k)(P 145.6(1)(P<0.01 (P < 0.05 for k and 1) was used. A direct individual comparison of normal donors populations were effective against Lym-1-Ag-positive Raji, receiving only brief 3-h pretreatment or continuous rIL-2 ex PLL, and CLL targets. Both non-ADCC lysis and Lym-1 posure in 16-h assays is shown in Fig. 2. The duration of rIL-2 ADCC were increased by the addition of rIL-2. In some in treatment was demonstrated to be an important parameter for stances the total lytic activity in the presence of Lym-1 and rlL- optimal cytotoxicity. Even a brief pulse treatment (3 h) of 2 can be ascribed to rIL-2 increasing the capacity of effector effector cells with rIL-2 resulted in increased lytic activity cells to lyse tumor targets independently of MoAb. However, against Raji cell targets, although pretreatment followed by in other cases greater than additive effects between rIL-2 and continuous exposure was optimal. rIL-2 treatment not only Lym-1 treatment were noted. increased Lym-1 ADCC but also increased non-antibody-me Lym-1, an isotype-matched control antibody (UPC-10), and diated cytotoxicity. This enhanced lytic activity attributed to other anti-DR MoAbs were compared to determine the role of rIL-2 appeared to be an additive phenomenon in some samples; isotype and antigen reactivity in ADCC (Table 3). Lym-1 me however, greater than additive effects between Lym-1 and rlL- diated ADCC in all donors tested. UPC-10, however, failed to 2 were noted in most donors. There was variability between mediate ADCC in any of the donors tested. MoAb isotype was donors both in the magnitude of Lym-1 ADCC and in their important since L243 (IgG2a) was able to mediate ADCC response to rIL-2 treatment (not shown). However, repetitive similar to Lym-1, whereas L227 (IgXÃ¼l) showed negligible target testing of these same donors at later times indicated that the cell lysis despite binding to strongly expressed antigens on Raji majority had consistent cytotoxicity profiles; i.e., a high re cells (not shown). Taken together these results demonstrate the sponder was usually high, and a low responder generally had specificity of Lym-1 ADCC and the importance of Lym-1-Ag low activity. Regardless of this variability, continuous exposure expression, level of expression, and MoAb isotype. to rIL-2 never failed to enhance total lytic activity. To determine whether Lym-1 ADCC required an interaction Specificity of Lym-1 ADCC. Lym-1 ADCC was a specific between Fc receptor-bearing effector cells and the Fc portion reaction between Fc receptor-positive effector cells and Lym-1 - of Lym-1, Protein A was added to the assay (Fig. 3). Protein A Ag-bearing tumor cells. Table 2 shows representative results inhibited Lym-1-mediated target cell lysis in the presence or obtained against a panel of target cell lines which differentially absence of IL-2 by 90 to 100%. Protein A inhibited cytotoxicity express Lym-1-Ag. Lym-1 failed to elicit any additional cyto by specifically blocking Fc-Fc receptor interactions, since it had toxicity other than that generated by native untreated effector no effect on non-antibody-mediated tumor cell lysis. Protein A cells against the EB3, Jijoye, and K562 cell lines which express had no effect on "Cr release of target cells (Fig. 3, Column 3). little or no Lym-1-Ag, respectively. The same donor PBMC Comparison of Normal and Cancer Patient ADCC. The PBMC 2993

260 Table 3 Comparison of different MoAbs in ADCC Results expressed are representative of several normal donors (n = 5) tested for ADCC Â¡n16-hassays using Raji cells as targets. The percentage of cytotoxicity 240 - was obtained at 40:1 effectortarget ratios. All MoAbs were used at a final concentration of 10 Mg/ml. Effector cells were pretreated for 3 h with rIL-2, and ADCC was performed with continuous rIL-2 exposure throughout the assay. The 220 addition of any MoAb without effector cells did not cause lysis of target cells above spontaneous release levels. 200 33/1 07 MoAbLym-lL243L227UPC-10None*IsotypeIgG2aIgG2aIgGlIgG2aÂ»ofDRcytotoxicityVariantspecificity cells1071320NDÂ°0 ISO DRNonpolvmorphic DRNonpolvmorphic 160 DR.DP, " DQNone494714124LU o 140 J ND, not done. * Effector cells alone without MoAb addition. IO '20 CO 3 -> 100 60h

No Continuous 3 Hour Lym-t Lym-l Lym-t IL-2 IL-2 IL-2 +Continuous + 3Hr IL-2 IL-2 Fig. 2. Effect of IL-2 exposure on PBMC lytic activity against Raji target cells in a 16-h ADCC assay. PBMC isolated from normal donors were tested for lytic activity in the absence or presence of Lym-l (10 Mg/ml) Â±IL-2 (100 units/ml). PBMC were treated with IL-2 for 3 h or continuously throughout the assay. Target:effector ratios ranged from 40 to 2.5:1. Results are the mean valuation of triplicate samples and are expressed as LU33/10' cells.

Table 2 Lym-l ADCC against various target celt lines ADCC was performed using 16-h assays with effector cells pretreated with IL- Fig. 3. Protein A inhibition of Lym-l ADCC. The effects of Protein A (SpA) 2 for 3 h and with continuous rIL-2 exposure throughout the assay. Results are treatment on Lym-l ADCC were determined in 16-h assays against Raji tumor expressed as LU 33/107 cells. cells. Lym-l (10 Mg/ml) was added to wells containing target cells in the presence cells'Target LU 33/107 or absence of Protein A (20 fig). The mean percentage of lysis of triplicate samples from untreated or rIL-2 (100 units/ml)-treated PBMC is shown. The tar- -Ag + + + Lym-l get:effector ratio was 40:1. Results are representative of 3 normal donors tested. cellRaji originB/BL expression"2-3+ IL-226 Lym-l43 + IL-284 donors (mean LU33, 118 versus 33, respectively). The mean MO 2007 B/PLL 3+ 10 30 46 76 ADCC activity observed with the patient group was also ap MO 1046 B/CLL 3+ 0 11 35 63 MO 1001 B/CLL 3+ 6 198 35 64 proximately 2 times higher than normal donors. The total lytic MO 1018 B/CLL 2-3+ 1 19 40 activity of cancer patients, although twice as high as normals, 0-1 + EB3 B/BL 14 435 14 42 appears to be an additive phenomenon resulting from the sum JiJOYE B/BL 0-1 + 9 8 8 K562Lineage/My/CMLLym-l 0E"1 68E NDE 50E ND of non-ADCC and ADCC lytic activity (P = 0.0630). In con "Lym-l-Ag expression is an estimate based on mean log fluorescence and trast, IL-2 treatment of normal PBMC appears to be greater number of positive cells as determined by flow cytometry of cells (1 x 10Â°)stained than additive in that total mean lytic activity (127 LU) of non- with Lym-l (5 fig) and a fluoresceinated goat F(ab')Â¡anti-mouse antiserum. * E, effector cells alone: E + IL-2. effector cells plus 100 units/ml of IL-2; E ADCC (33 LU) and Lym-l ADCC (45 LU) is greater than the + Lym-l, effector cells plus 10 fig/ml of Lym-l; E + Lym-l + IL-2, effector cells sum totals (78 LU) (P = 0.0007). Thirteen of 15 normal plus Lym-l and IL-2; B, B-cell; BL, Burkitt's lymphoma; MY, myeloid; CML, individuals tested were considered to have greater than additive chronic myelogenous leukemia; ND, not done. ' All values rounded to the nearest whole number. response to Lym-l and IL-2. Eight had percentages greater than 100% (ranging from 110 to 328%), five were between 40 of cancer patients undergoing rIL-2 treatment in clinical pro and 86%, and only two individuals were less than 20% above tocols were examined for their ability to mediate Lym-l ADCC what would be considered directly additive responses. (Table 4). No significant differences between normals and can cer patients were observed in native effector cell lytic activity DISCUSSION (P = 0.2666) or for Lym-l ADCC (P = 0.0659). Additional testing will be necessary to determine if indeed a statistically Antibody-dependent cellular cytotoxicity represents a poten significant difference can be obtained by increasing sample size. tial efferent arm of cell-mediated immunity which can be ex However, healthy donors and cancer patients differed signifi ploited to develop novel biological regimens for cancer therapy. cantly for rIL-2-treated PBMC (P = 0.0008) and for Lym-l Although much is already known about ADCC from the murine plus rIL-2 treatment (P = 0.0085). Patient effector cells were system, only recently has the feasibility of using human effector more responsive to in vitro stimulation with rIL-2 than normal cells with murine MoAbs against human malignant targets been 2994

Table 4 Cancer patient Lym-I ADCC ADCC assays (16 h) were performed with patient or normal PBMC in the presence or absence of Lym-1 (10 Mg/ml) or rIL-2 (100 units/ml). Effector cells receiving rIL-2 were pretreated for 3 h followed by continuous exposure throughout the assay. Lym-1, rIL-2. or combination treatment in the absence of effector cells resulted in no significant lysis of Raji target cells. Target:effector cell ratios ranged from 40 to 2.5:1 for calculation of lytic units. cellsPatient1234567Normals LU 33/1 07

+Lym-1+ Lym-10.310116274493effectorE + IL-2 E + IL-2338317245238177173117229127+

15)" (n = E, effector cells alone;Age/sex50/M54/M59/M45/M47/M67/M58/MDiagnosisRenalColonColonMelanomaLungRenalRenalAv.Av.TreatmentIL-2/LAKIL-2/IFNIL-2/IFNIL-2/LAKIL-2/LAKIL-2/CTXIL-2/LAKE+ IL-2, effector cells plus 100 units/ml of 1L-2; E + Lym-1,E" cells plus2341307419012240391183310ÃÃg/mlofLym-722051104053793385451;E + Lym-1E IL-2, effector cells plus Lym-1 and IL-2; IFN, interferon; CTX, cyclophosphamide. addressed. To define which MoAbs may be most useful in non-ADCC lysis of human colon carcinoma (12). In addition, ADCC, each should be screened in vitro using as clinically a distinct subpopulation of murine LAK cells that are FcR relevant a situation as possible. In this study, we demonstrate positive has been shown to be expanded (not induced) in re that human PBMC can effectively mediate ADCC with Lym-1 sponse to IL-2 (20, 21). Although both FcR-positive and FcR- against human B-lymphoma/leukemia targets. ADCC was negative populations demonstrated LAK activity, only FcR- Lym-1 dose dependent, and significant in vitro lysis was ob positive cells were able to mediate ADCC. Murine LAK cells served with Lym-1 levels that are readily obtained in vivo (18). which participate in ADCC have, in fact, been shown to effec Specificity of Lym-1 ADCC was shown because tumor cell lysis tively lyse LAK-resistant murine hepatoma (22). Results from was restricted only to malignant human B-cell lines expressing all these studies suggest that LAK cells in combination with Lym-l-Ag. Lym-1 binding to both target cells and Fc receptor- IL-2 and Lym-1 may be beneficial in that both MoAb-depend bearing effector cells was a requirement for tumor cell lysis. ent ADCC and LAK cell activity may be synergistic in mediat Although both normal donors and cancer patient PBMC ing antitumor responses. showed varied levels of Lym-1 ADCC, the activity of all donors The mechanism by which IL-2 can amplify effector cell tested was amplified with clinically obtainable levels of rIL-2.3 function is not understood. Upregulation of effector cell Fc Maximal in vitro enhancement of ADCC was observed with receptor number and/or expansion of FcR-positive cells may rIL-2 concentrations (100 units/ml) and exposure times (3 h) be one mechanism for increased ADCC. Expansion of FcR- less than that required for optimal in vitro LAK cell generation. positive cells appears to be more likely in that it has been It is intriguing that enhancement of both antibody-mediated reported that IL-2 can increase the number of FcR-positive and non-antibody-mediated lysis by rIL-2 was observed with cells in the murine system, whereas FcR density is not affected, short exposure times, and this suggests that the regulatory and these FcR-positive cells are capable of mediating ADCC signals involved have rapid kinetics. Our results are consistent (20, 21). Our studies have similarly failed to show any ampli with a previous report which demonstrated that PBMC pre fication of human PBMC Fc receptor expression as gauged by treated with rIL-2 (1 to 10 units/ml) for 3 h had enhanced flow cytometric analysis with MoAbs reactive with CD64, CDw32, and CD 16 (anti-Fc -y-RI, II, and III, respectively; data ADCC (12). Prolonged exposure (1 to 6 days) of effector cells to rIL-2 (100 to 1000 units/ml) in that study resulted in not shown). We have not yet addressed whether or not expan sion of human PBMC Fc 7-R-positive cells is a consequence of increased LAK activity with concomitant impairment of murine MoAb (NR-CO-04)-mediated ADCC against human colon can IL-2 treatment. Another purported mechanism for the in creased reactivity with IL-2 is the induction of cytolysin in the cer (12). Of interest is the finding that PBMC from cancer patients cytoplasmic granules of murine LAK cells which may be related receiving continuous rIL-2 therapy were shown to further re to their effectiveness in target cell lysis (23). spond to in vitro exposure of low-dose rIL-2 and effectively The limited therapeutic efficacy observed in prior passive lysed tumor targets by both ADCC and non-ADCC mecha serotherapy studies using unconjugated MoAbs may be partially nisms. The combination of these distinct MoAb-dependent and attributed to inadequate involvement of host effector cells (24- -independent pathways for lysis was superior to those observed 28). Similar findings were reported in a recent clinical trial using Lym-1 monoclonal antibody serotherapy in patients with in normal donor PBMC. Although normal donors exhibited markedly less total lytic activity following rIL-2 treatment, the refractory B-cell malignancy (18). Although treatment was well tolerated, high serum levels of Lym-1 were obtained, and no ADCC lytic compartment was enhanced proportionally more patients developed anti-mouse antibody; clinical responses were than in cancer patients. This suggests that (a) the level of activation of the rIL-2 responding effector cell and/or (b) ex minimal. Clinical responses appeared to directly correlate with pansion of effector cell subsets is an important determinant in the number of infiltrating lymphocytes supporting the idea that Lym-1 mediates tumor regression by ADCC. Other antitumor the ultimate lytic pathway (i.e., antibody or non-antibody me diated) that is generated. In support of this, the duration and mechanisms possibly involved include nonspecific phagocytosis dose of rIL-2 have been shown to be important in determining of Lym-1-coated tumor cells by macrophage and complement- whether large granular lymphocytes participate in ADCC or mediated cytotoxicity. Our results suggest that combination therapy of Lym-1 and 3 Data on file. Hoffmann-LaRoche. Inc.. Nutley, NJ. rIL-2 may potentiate effector cell populations. The in vivo 2995

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1990 American Association for Cancer Research. EFFECTS OF rIL-2 ON HUMAN EFFECTOR CELL LYSIS OF B-CELL TUMOR relevance of combination therapy using rIL-2 and MoAb and lymphoma subjects are cytotoxic for cells coated with antibody deriva tives displaying human Fc gamma. Blood, 72: 1985-1991, 1988. against human B-lymphoma was recently demonstrated (19). 10. Berinstein, N., and Levy. R. Treatment of a murine B cell lymphoma with The antitumor effects of an anti-CD 19 (CLB-CD19), IgG2a monoclonal antibodies and 1L-2. J. Immunol., 139: 971-976, 1987. MoAb were potentiated by rIL-2, and the increased therapeutic 11. Eisenthal, A., Lafreniere, R., Lcfor, A. T.. and Rosenberg, S. A. Effect of anti-BI6 melanoma monoclonal antibody on established murine B16 mela effect of this regimen in a xenotransplantation model of human noma liver mÃ©tastases.Cancer Res.. 47: 2771-2776, 1987. B-lymphoma was attributed to increased ADCC. Antitumor 12. Morgan. A. C., Sullivan. W., Graves. S.. and Woodhouse, C. S. Murine monoclonal IgG3 to human colorectal tumor-associated antigens: enhance activity was observed only with the CLB-CD19, IgG2a class ment of antibody-dependent cell-mediated cytotoxicity by interleukin 2. switch variants, suggesting that MoAbs of this isotype, such as Cancer Res.. 49: 2773-2776. 1989. Lym-1, may be most useful for inducing ADCC against B- 13. Bubenik. J., Kieler. J.. Tranholt, V., Steven, K., Indrova. M., and Toulcova, A. Activation of killer cells from blood of urinary bladder carcinoma patients lymphoma. by short-term treatment with recombinant IL-2. Folia-Biol., 32: 183-194. 1986. Recruitment of additional effector cell types and/or the use 14. Han, T.. Dadey, B., Sheedy, D., Xiao, H., Block, A. W., O'Donnell, A., of other cytokines may also result in better treatment modali Bhargava. A.. Fitzpatrick. J., Roth. M. S.. Foon. K.. and Henderson, E. S. ties. Granulocytes, the most abundant leukocyte and monocyte/ Establishment and characterization of leukemic cell lines from patients with macrophage, also participate in ADCC (4, 5, 29, 30). Granu- adult B-cell leukemias. Blood. 72 (Suppl. 1): 721, 1988. 15. Vaickus, L., Jones, V. E., Morton, C. C, Whitford, K., and Bacon. R. N. locyte-macrophage colony-stimulating factor has been shown Antiproliferative mechanism of anti-Class II monoclonal antibodies. Cell. to enhance human granulocyte ADCC against human mela Immunol.. 119: 445-458, 1989. 16. Boyum. A. Separation of blood leukocytes, granulocytes. and lymphocytes. noma and neuroblastoma targets (31). We have demonstrated Tissue Antigens, 4: 269-274, 1974. that recombinant 7-interferon potentiates not only granulocyte- 17. Pross, H. F., Baines, M. G., Rubin, P.. Shragge, P., and Patterson, M. S. mediated Lym-1 ADCC, but ADCC and non-ADCC tumor Spontaneous human lymphocyte-mediated cytotoxicity against tumor target cells. IX. The quantitation of natural killer cell activity. J. Clin. Immunol.. lysis by PBMC is increased as well (32). Human monocytes/ /.-51-63, 1981. macrophages have also been shown to have increased effector 18. Hu, E.. Epstein, A. L., Naeve, G. S., Gill. I., Martin, S., Sherrod, A., Nichols. function and potentiation of ADCC by either y-interferon or P.. Chen. D.. Mazumder. A., and Levine, A. M. A Phase la clinical trial of Lym-1 monoclonal antibody serotherapy in patients with refractory B cell macrophage colony-stimulating factor (33, 34). The role of malignancies. Hematol. Oncol.. 7: 155-166. 1989. Lym-1 in monocyte/macrophage ADCC will be addressed in 19. Vuist. W. M. J.. Buitenen. F. V., de Rie, A.. Hekman, P.. Rumke. P., and Melief, C. J. M. Potentiation by interleukin 2 of Burkitt's lymphoma therapy future studies. Taken together, it appears that recombinant with anti-pan B (anti-CD19) monoclonal antibodies in a mouse xenotrans cytokines have the potential to increase host effector cell func plantation model. Cancer Res., 49: 3783-3788. 1989. tions when used in the correct combination with monoclonal 20. Eisenthal, A., and Rosenberg. S. A. The effect of various cytokines on the in antibodies. This study provides a rationale for using Lym-1 in vitro induction of antibody-dependent cellular cytotoxicity in murine cells. J. Immunol.. 142: 2307-2313. 1989. combination with rIL-2 for future clinical trials. 21. Eisenthal. A., Shiloni. E.. and Rosenberg. S. A. Characterization of IL-2 induced murine cells which exhibit ADCC activity. Cell. Immunol., 115: 257-272. 1988. ACKNOWLEDGMENTS 22. Kawase, I.. Komuta. K.. HarÃ¡.H.. Inoue. T.. Hosoe, S., Ikeda. T., Shirasaka. T., Yokota, S.. Tanio, Y., Masuno. T.. and Kishimoto. S. Combined therapy The authors wish to thank Louise Zgoda for expert technical assist of mice bearing lymphokine-activated killer-resistant tumor with recombi nant interleukin 2 and an antitumor monoclonal antibody capable of inducing ance in this study. We also thank Gail E. Walsh for her excellent antibody-dependent cellular cytotoxicity. Cancer Res.. 48: 1173-1179, 1988. secretarial skills and typing of this manuscript. 23. Henkart, P., Cheung, C. Y., Yang, J.. and Rosenberg. S. A. Cytolytic and biochemical properties of cytoplasmic granules of murine lymphokine-acti vated killer cells. J. Immunol.. 137: 2611-2617. 1986. REFERENCES 24. Foon. F. A.. Schroff, R. W., Bunn, P. A., et al. Effects of monoclonal antibody therapy in patients with chronic lymphocytic leukemia. Blood, 64: 1. Epstein, A. L.. Marder. R. J.. Winter. J. N., Stathopoulis. E., Chen, F.-N., 1085-1093. 1984. Parker, J. W.. and Taylor, C. Two new monoclonal antibodies. Lym-1 and 25. Dillman, R. O., Shawler, D. L., Dillman, J. B.. and Royston. I. Therapy of Lym-2. reactive with human B-lymphocytes and derived tumors, with im- chronic lymphocytic leukemia and cutaneous T cell lymphoma with T101 munodiagnostic and immunotherapeutic potential. Cancer Res.. 47: 830- monoclonal antibody. J. Clin. Oncol., 2: 881-891, 1984. 840. 1987. 26. Oldham. R. K., Foon, K. A., Morgan, A. C., et al. Monoclonal antibody 2. Foon. K. A. Biological response modifiers: the new immunotherapy. Cancer therapy of malignant melanoma: in vivo localization in cutaneous metastasis Res..Â«: 1621-1639, 1989. after intravenous administration. J. Clin. Oncol.. 2: 1235-1244, 1984. 3. Herlyn, D., Powe, J., Ross. A. N., Herlyn. M., and Koprowski. H. Inhibition 27. Ritz, J., Pesando, J. M.. Sallan, S. E., el al. Serotherapy of acute lympho- of human tumor growth by lgG2a monoclonal antibodies correlates with blastic leukemia with monoclonal antibody. Blood. 58: 141-152, 1981. antibody density on tumor cells. J. Immunol.. 134: 1300-1304. 1985. 28. Ball. E. D.. Bernier. G. M.. Cornwell, G. G.. Mclntyre, O. R., O'Donnell, J. 4. Steplewski, Z., Herlyn, D.. Manel. G.. and Koprowski. H. Hypothesis: F.Jr., and Fanger, M. W. Monoclonal antibodies to myeloid differentiation macrophages as effector cells for human tumor destruction mediated by antigens: in vivo studies of three patients with acute myclogenous leukemia. monoclonal antibody. Hybridoma. 2: 1-5. 1983. Blood, 62: 1203-1210, 1983. 5. Ortaldo, J. R., Woodhouse, C., Morgan, A. C., Herberman. R. B., Cheresh, 29. Levy, P. C.. Shaw, G. M., and LoBuglio. A. F. Human monocyte, lymphocyte, D. A., and Reisfeld. R. Analysis of effector cells in human antibody-dependent and granulocyte antibody-dependent cell-mediated cytotoxicity toward tumor cellular cytotoxicity with human murine monoclonal antibodies. J. Immunol.. cells. J. Immunol., 123: 594-599. 1979. Â¡38:3566-3572, 1987. 30. Gale. R. P.. and Zighelboim, J. Polymorphonuclear leukocytes in antibody- 6. Kipps. T. J., Parham, R., Punt, J.. and Herzenberg. L. A. Importance of dependent cellular cytotoxicity. J. Immunol.. 114: 1047-1051, 1975. immunoglobulin isotype in human antibody-dependent cell mediated cyto 31. Kushner. B. H., and Cheung, N.-K. V. GM-CSF enhances 3F8 monoclonal toxicity directed by murine monoclonal antibodies. J. Exp. Med.. 161: 1-17, antibody-dependent cellular cytotoxicity against human melanoma and 1985. neuroblastoma. Blood, 73: 1936-1941, 1989. 7. Gore. M. E., Skilton, R. A., and Coombes, R. C. Antibody-dependent-cellular 32. Vaickus, L.. Biddle. W.. Cemerlic. D.. and Foon. K. Gamma-interferon cytotoxicity against cultured human breast cancer cells mediated by human augments Lym-1-dependent granulocyte-mediated tumor cell lysis. Blood, in effectors using monoclonal and polyclonal antibodies. Br. J. Cancer, 48:877- press. 879, 1983. 33. Mufson. R. A., Aghajanian, J., Wong. G., Woodhouse. C., and Morgan, A. 8. Woodhouse, C. S., and Morgan. A. C. Murine monoclonal IgG3 antibodies C. Recombinant human macrophage stimulating factor enhances the activity to human colorectal tumor-associated antigens: production and characteriza of human macrophages in antibody dependent cell mediated cytotoxicity. tion of antibodies active in both antibody-dependent cellular cytotoxicity and Cell. Immunol.. 119: 182-192. 1989. complement-mediated cytolysis. Cancer Res.. 49: 2766-2772. 1989. 34. Shen. L., Guyre, P. M., Ball. E. D.. and Fanser, M. W. Glucocorticoid 9. Dearman. R. J., Stevenson. F. K.. Wrightham. M., Hamblin. T. J.. Glennie. enhances gamma interferon effects on human monocyte antigen expression M. J., and Stevenson, G. T. Lymphokine-activated killer cells from normal and ADCC. Clin. Exp. Immunol.. 65: 387-395. 1986.

2996

Downloaded from cancerres.aacrjournals.org on September 26, 2021. © 1990 American Association for Cancer Research. Antibody-dependent, Cell-mediated Cytotoxicity by an Anti-Class II Murine Monoclonal Antibody: Effects of Recombinant Interleukin 2 on Human Effector Cell Lysis of Human B-Cell Tumors

William C. Biddle, James Pancook, Martin Goldrosen, et al.

Cancer Res 1990;50:2991-2996.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/50/10/2991

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/50/10/2991. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.