CELLULARIMMUNOIBGY 115,420-428 (1988)

Cytolytic Activity of Murine Anti-dog Lymphoma Monoclonal Antibodies with Canine Effector Ceils and Complement

CARLOS ROSALES, K. ANN JEGLUM,~ MARIA OBRGCKA,$ AND ZENON STEPLEWSKI*,’

*The Wistar Institute, 34th at Spruce Streets, Philadelphia, Pennsylvania 19104; tschool of Veterinary Medicine, University of Pennsylvania, 38th at Spruce Streets, Philadelphia, Pennsylvania 19104;

and SNYGENE Corporation, Malvern, Pennsylvania 19087

Received February 24,1988; accepted March 28, 1988

Peripheral blood leukocytes (PBL), nonadherent lymphocytes, and adherent monocytes separated from freshly isolated blood of 15 dogs were analyzed for their ability to mediate anti- body-dependent cell-mediated cytotoxicity (ADCC) in combination with murine anti-tumor monoclonal antibodies (MAbs). Canine monocytes isolated from most donors by adherence to gelatin-fibronectin-coated plastic surface presented high ADCC activity against the canine lymphoma 17-71 tumor cell line in combination with antilymphoma MAbs 23 1 (IgG2.) and 234-2a (h&). Canine lymphocytes generally showed lower ADCC activity than total PBL or monocytes. Canine PBL effector cells showed high ADCC activity against the human colorectal carcinoma SW948 cell line using the Y-6-specific MAb isotype switch variants 55-2 IgGs, 55-2 IgGi ,5 5-2 IgGzi,, and 55-2 IgGzD . Analysis of the role of murine MAb isotypes on ADCC activity against tumors by canine cells using anti-human tumor class-switch variant MAbs and a panel of anti-canine lymphoma MAbs of different IgG subclass revealed the highest ADCC activity with MAbs of the IgG2, and IgGs subclasses. IgG2. antilymphoma MAbs were also able to lyse tumor cells in complement-dependent cytotoxicity (CDC) assay. These results suggest the poten- tial value of MAbs of IgGs and IgG2, subclasses in immunotherapy against canine lymphoma. 0 1988 Academic Press, Inc.

INTRODUCTION

Canine lymphoma is a spontaneous tumor in an outbred animal and is a valuable model for the investigation of anticancer therapeutics ( 1,2). It is the most common hemopoietic tumor in the dog. Multicentric lymphadenopathy and hepatospleno- megaly are the most common clinical presentations. Extranodal lymphomas occur in less than 15% of the cases (3). The National Cancer Institute’s Working Formula- tion of histologic classification for non-Hodgkin’s lymphoma correlates with progno- sis in dogs as in man (4-6). Both the canine and non-Hodgkin’s lymphoma in hu- mans respond to the same cytotoxic drugs (7-9). In addition, immunotherapy as an

’ To whom correspondence should be addressed.

420

0008-8749/88$3.00 Copyright 0 1988 by Academic FTa, Inc. AU rights of reproduction in any form reserved.

CYTOTOXICITY OF CANINE EFFECTOR CELL 421

adjunct to chemotherapy has been shown to prolong survival and remission duration in dogs with lymphoma ( 10).

We recently reported the production and characterization of murine-derived monoclonal antibodies (MAbs)* against canine lymphoma (11). Murine MAbs have been successfully applied in inhibiting the growth of human tumor cells in nude mice (12, 13). In addition, positive results of immunotherapeutic trials with MAbs in hu- man cancer have been reported ( 14- 16). The mechanisms involved in antibody-de- pendent tumor cell destruction are not well characterized. Lysis of tumor cells may result from the interaction of MAbs with one or more effector cells including lympho- cytes, monocytes, macrophages, polymorphonuclear cells, natural killer, or killer cells. In order to apply murine MAbs for immunotherapy of canine tumors, it is necessary to define which effector cells will be able to utilize which MAb isotype for efficient tumoricidal activity. In the present study, we compared the ability of antilymphoma MAbs to mediate tumor cell lysis by canine effector cells. We also examined canine antibody-dependent cellular cytotoxicity (ADCC) activities in con- junction with different murine MAb isotypes.

MATERIALS AND METHODS

Monoclonal antibodies. Isolation and characterization of antitumor MAbs BR55- 2 (IgG3), 231 (IgGZp), 234 (IgGi), 254 (IgGj), 212 (IgG1), 215 (IgM), and 216-1 (IgM) have been described ( 11, 17). IgGi , IgGza, and IgGzb class-switch variants of MAb 55-2, originally IgG3, and 234-2a (IgG&, a switch variant of IgG+ecreting hybrid- oma 234 were isolated by sib selection (18, 19). Other MAbs used are 37-7, IgG*, (anti-human DR), H24B5, IgGx, (anti-influenza virus), and PJ, IgG, (from mouse myeloma P3X63/Ag8) (20). MAbs used in ADCC assays were purified by affinity chromatography columns of protein A-Sepharose (Pharmacia) (2 1) from ascites flu- ids produced in pristane-primed mice. Undiluted tissue culture supernatants were also used as a source of MAb for some assays.

Cell lines. Canine lymphoma cell line 17-7 1 established in our laboratory ( 11) was maintained in minimum essential medium (MEM)/lO% fetal bovine serum (FBS), 20 mM glutamine. Human leukemia cell lines Raji, Daudi, and K562 were main- tained in RPM1 1640/10% FBS medium. SW948 and SW1 116 (human colon carci- nomas) were maintained in L-15/10% FBS medium. All cell lines are continually maintained at the Wistar Institute.

Canine efictor cells. Blood was obtained from 15 random breeds of dogs of differ- ent ages and sexes. Peripheral blood leukocytes (PBL) were separated from normal canine heparinized blood by density gradient centrifugation on Ficoll-Hypaque (22, 23). Cells collected from the interface were washed three times with phosphate- buffered saline (PBS) without Ca*+ and Mg*+, and resuspended in RPM1 1640/ 10% FBS medium containing 20 mM glutamine, 100 U/ml penicillin, and 100 pg/ml streptomycin. Enriched monocyte fractions were obtained by adherence selection of

’ Abbreviations used: ADCC, antibody-dependent cell-mediated cytotoxicity; CDC, complement-de- pendent cytotoxicity; CRC, colorectal carcinoma; FcR, receptor for the Fc fragment of Ig; FBS, fetal bovine serum; MAb, monoclonal antibody; MEM, minimum essential medium; NK/K cell, natural killer/killer cell; PBL, peripheral blood leukocyte: PBS, phosphate-buffered saline; PMN, polymotphonuclear leuko- cyte; T/E ratio, target/effecter ratio.

422 ROSALES ET AL.

PBL on gelatin-fibronectin-coated flasks (24). The population consisted of 90 to 95% of monocytes, as defined by morphological criteria. Nonadherent leukocytes (lym- phocytes) were obtained following a 1 -hr adsorption on gelatin-fibronectin-coated plastic flasks; this population was >95% depleted of adherent cells.

Antibody-dependent cell-mediated cytotoxicity. Tumor cell suspensions were pel- leted in a 15-ml tube and labeled with [“‘In]indium oxine (25) (indium oxyquino- line, 1 mCi/ml, Amersham) for 10 min at 10 &i/ lo6 cells. Labeled cells were placed in U-bottom 96-well Costar microtiter plates in triplicate at 1 X lo4 cells/well contain- ing RPM1 1640/ 10% FBS medium, followed by addition of 100 ~1 purified MAb (final concentration 100 pg/ml) or 100 ~1 MAb-containing tissue culture supernatant. Anti-influenza MAb of the IgGza subclass was used as negative control. A target/ effector (T/E) ratio of l/20 was used. Plates were covered and incubated at 37°C in 5% CO2 and 95% air for 18 hr.

Complement-dependent cytotoxicity (CDC). Complement lysis assays were per- formed as described by Herlyn and Koprowski (26) with slight modification. Briefly, “‘In-labeled target cells were plated in U-bottom 96-well plates in 50 ~1 of medium at 2 X lo4 cells/well, followed by addition of 50 ~1 MAb solution with shaking. Plates were incubated 1.5 hr at room temperature, and 50 ~1 of cold rabbit serum (Cappel Laboratories) diluted l/5 in RPM1 1640 medium was added as a source of comple- ment. After 45 min of incubation at 37°C in 5% CO*, 50 ~1 of cold medium was added with shaking and plates were centrifuged at 1200 t-pm for 6 min. One hundred microliters of supernatant was collected and radioactivity counted.

For all assays, percentage specific cytotoxicity was calculated using the formula

% specific release = S-SR x 100 If2 - SR ’

where S is the cpm in supematants from target cells incubated with effector cells and MAb or complement, SR is the spontaneously released cpm in supematants from target cells incubated with medium only, and Z is the total cpm (input) added to each well.

RESULTS

Antibody-Dependent Cell-Mediated Cytolysis

Figure 1 shows the dual dependence of ADCC on the presence of MAb and effector cell density. Human CRC SW948 cells were lysed by canine PBL at different T/E ratios in the presence of MAb BR55-2 of IgG*, isotype. In vitro, canine PBL kill tumor cells in ADCC most efficiently at a T/E ratio of l/20. Higher ADCC values were obtained when the cells were incubated for 18 hr. These conditions were used in all subsequent experiments.

PBL, lymphocytes, and monocytes obtained from 15 dogs were screened for ADCC activity using two target cell-MAb systems: CRC SW948 cells and anti-CRC MAbs 55-2 of all IgG subclasses which bind to this cell line, and dog lymphoma 17- 71 cells in combination with antilymphoma MAbs 212 (IgGi), 231 (IgGz,), 254 (IgG,), 234 (IgG1), and 234-2a (IgG2, switch-variant of hybridoma 234). Most canine effector cell preparations showed high ADCC activity on both tumor cell lines with the corresponding MAb (Fig. 2). However, considerable individual variation in

CYTOTOXICITY OF CANINE EFFECTOR CELL 423

A 60 -

50 .'

40 .'

x 30..

Lysis

20 " / -b 10 '. 6 OU --

0 1110 1120 1130 1150

% Lysis

0 1110 1120 l/30 1150

Ratio

FIG. 1. Dependence of ADCC on MAb 55-2 (IgGza) and the number of canine effector cells. Killing of SW948 target cells by effector cells at various T/E ratios in absence @) or presence (A) of MAb 55-2 at 100 &ml in a 6-hr (A) and 18-hr (B) ADCC assay. Effect of antilymphoma MAb 23 1 on CRC SW948 is also shown (0).

ADCC capacity was observed, with values ranging from very low levels to extensive lysis (50-60s) of target cells.

All IgG anti-dog lymphoma MAbs showed some ADCC activity with the different canine blood cell populations. For the same donor, the ADCC values by PBL were always equal to or higher than those found with the nonadherent lymphocytes. Ad- herent, monocyte-enriched fractions gave consistently higher values than those asso- ciated with PBL or lymphocytes (see Fig. 3 for representative results). In every case, highest ADCC values were obtained using MAb 234-2a (IgG2, variant). Strong ADCC activity was observed using MAbs 23 1 (IgG& and 254 (IgG& whereas MAbs 212 (IgGr) and 234 (IgGJ were less effective in killing target cells in combination with monocytes. ADCC activity with canine PBL and MAb 23 1 was higher than with

424 ROSALES ET AL.

‘OT A B .

60 .. .

% . 50 -.

Ly.sk . . . I

40 -. . . 9.1 - .

. . 30 -. l :

. . . l

. .

I .

20.. . : , *a *’ .h .

. .: .

. . : .

:.

10.. ;; .:* ‘.: . ‘ii : - 1 : *. ..A

L ai. . .*. 9 .

0, it. .,. i.. . , . . .* :i: .:. :.. . *. . .: . , . . .

212 231 254 234 234-za IgGl IgGPa lgG3 lgG1 IgGPa

MAb

55-2 55-2 55-2 55-2 IgG1 IgGPa IgGPb lgG3

FIG. 2. ADCC activity of canine PBL (m), lymphocytes (a), and monocytes (A) obtained from normal dogs. ADCC of canine 17-7 1 lymphoma cells (A) and human CRC SW948 cells (B) in the presence of antilymphoma MAbs 2 12,23 1,234,254, and 234-2a and different subclasses of MAbs 55-2.

any other of the MAbs except 234-2a. ADCC values obtained using MAbs 234 or 234-2a with canine lymphocyte effector cells were similar and slightly higher than those of the other antilymphoma MAbs (2 12, 23 1, and 254). ADCC activity on SW948 cells with anti-human CRC MAbs 55-2 and canine PBL effector cells was usually higher (Fig. 2) than that obtained with anti-canine lymphoma MAbs. Lym- phocytes and monocytes presented similar ADCC values for the IgG3 and IgG2, MAb subclasses (Fig. 4).

Dependence of Canine Cell-Mediated ADCC on Antibody Isotype

Anti-human colorectal carcinoma MAbs 55-2 representing the four murine IgG subclasses were evaluated for their capacity to mediate ADCC with canine effector

212 IgGl

231 IgGPa

254 234 234-2a lgG3 IgGl IgGPa

MAb

FIG. 3. ADCC on 17-7 1 lymphoma cell line with canine PBL (FBI), lymphocytes @I). and monocytes (Sa). ‘L’In-labeled target cells were mixed with 100 ~1 of MAb-containing tissue culture supematant and effector cells at T/E ratio of l/20 for 18 hr.

CYTOTOXICITY OF CANINE EFFECTOR CELL 425

60 T 50

Lysis 40

30

20

10

0 55-2 lgG2a

MAb

55-2 lgG3

FIG. 4. ADCC on “‘In-labeled CRC SW 1116 target cells and canine lymphocytes (O), and canine mono- ) as effector cells. Cells were mixed at a T/E ratio of I/ 20 and incubated for 18 hr in the presence

of MAbs 55-2.

cells against human CRC target cell lines (Fig. 5). MAbs of the IgG2,, IgGZb , and IgG3 isotypes showed significant ADCC activity and there is no apparent difference in their ability to mediate ADCC with all effector cell subpopulations tested. Much lower ADCC activity was observed with MAb 55-2 IgGl . Anti-canine lymphoma MAbs 23 1 and 234-2a, both IgGza and MAb 254 (IgG3) gave the highest ADCC values on canine lymphoma 17-7 1 target cells, whereas MAbs 2 12 and 234, both IgG, , were less effective in mediating ADCC (Fig. 6).

Complement-Dependent Cytolysis

Antilymphoma MAbs were tested for their potential to activate complement and lyse dog lymphoma 17-7 1 cells. MAbs 23 1 and 234-2a (both IgG2,) were able to medi-

60

55-2 IgGI 55-2 lgG2a 55-2 IgGPb 55-2 lgG3 MAb

FIG. 5. ADCC on CRC SW948 cells with canine PBL as effector cells. “‘In-labeled target cells were mixed with effector cells at a T/E ratio of l/20 for 18 hr in the presence of MAbs 55-2.

426 ROSALES ET AL.

40

Lysls

30

20

10

0 212

IgGl 231 254 234 234.2a

IgGPa lgG3 IgGl IgGPa MAb

FIG. 6. ADCC on 17-71 lymphoma cell line. Canine monocytes were mixed with “‘In-labeled target cells at a T/E ratio of l/20 and antilymphoma MAbs 212 (IgGJ, 231 (IgG*,), 234 (IgG,), 254 (IgG& and 234-2a (IgG&

ate tumor cell lysis with complement (Fig. 7). MAb 2 12 (IgG1) did not cause activa- tion ofcomplement at all, and MAbs 234 (IgGI) and 254 (IgG,) showed only minimal activity. MAb 2 15 (IgM) activated complement to lyse the tumor cells, but MAb 2 16- 1 (also IgM) did not. The anti-DR MAb 37-7 (IgG& included as a positive control showed high CDC activity.

DISCUSSION

In the present study, we demonstrate that freshly isolated canine blood cells medi- ate significant ADCC with murine antitumor MAbs. Canine killer cells (nonadherent

60

Lysis 50

40

30

20

10

0 0 212 215 231 254 216-1 234 234-2a 37-7

IgGl IgM IgGPa lgG3 IgM IgGl lgG2a IgGPa

MAb

FIG. 7. CDC assay of canine lymphoma 17-7 1 cell line ( ), human lymphoma Raji cell line (9, and human lymphoma Daudi cell line (0) with murine anti-canine lymphoma MAbs 2 12,2 15,23 1,234,254, and 234-2a and anti-human DR MAb 37-7.

CYTOTOXICITY OF CANINE EFFECTOR CELL 427

lymphocytes) were able to lyse tumor cell lines, although not always very efficiently, and canine monocytes were always better in mediating tumor lysis in combination with MAbs. This suggests that different antibody-dependent effecters may be in- volved in the lysis of different tumor cell types. We have also observed efficient killing of canine tumor cell line by murine thioglycolate-activated macrophages in the pres- ence of MAbs (unpublished). Macrophages have been implicated as the main effector cell type mediating antibody-dependent tumor inhibition in nude mice xenografted with human tumor cells ( 12, 13). However, the role of macrophages in tumor regres- sion of cancer patients who received MAbs remains unclear.

In some human monocyte-murine MAbs ADCC systems, it has been shown that certain target tumor cell lines give high ADCC values when target cells are preincu- bated with MAbs for 2 to 3 hr prior to the addition of effector cells, while others require simultaneous incubation of MAb with target and effector cells to obtain opti- mal ADCC activity (27). We did not evaluate the influence of pretreatment with MAb of effector and target cells in ADCC; however, in the systems used in this study, high ADCC values were obtained when cells were incubated simultaneously with MAbs for 18 hr.

Conclusions about the generality of canine cell ADCC activity await further testing of additional tumor cell lines. However, in the case of the anti-canine lymphoma MAbs, no cell line other than the canine tumor cell line 17-7 1 (11) is currently avail- able that expresses the antigens recognized by these antibodies. There are, however, several different human tumor cell lines that express the antigen defined by anti- human carcinoma MAb BR55-2. We found significant ADCC activity by canine effector cells and MAb 55-2 on both human CRC cell lines tested (SW948 and SW1116).

The basis of the significant variations observed among canine donors in ability to mediate ADCC is unclear but might reflect differential FcR expression, different ra- tios of competent cell subpopulations able to mediate ADCC in circulation, and/or differences in ability to elaborate various cytolytic mediators.

In this study, antitumor MAbs of the IgG za, IgGZb, and IgG3 isotypes presented highest ADCC activity in vitro with canine effector cells. In the case of canine mono- cytes, it appears that IgG2, and IgG3 MAbs are better activators of ADCC and that these cells account for most of the ADCC activity observed in vitro (Fig. 6). It is not known which of these IgG subclasses might cooperate optimally with canine mono- cytes or macrophages in vivo. In the murine systems, MAbs of the IgG2, and IgG3 subclasses reportedly have better tumor growth-inhibiting activity in vivo ( 12, 13,28- 32). In dogs, antilymphoma MAbs 23 1 (IgG2,), 234-2a (IgG2,), and 254 (IgG3) gave the highest levels of ADCC. MAbs 23 1 and 234-2a were also the only ones active in complement-mediated killing of 17-7 1 cells. These results, together with previous findings that murine IgG2, and IgG3 can bind the high-affinity monocyte FcR (33) and that IgGZa MAbs activate murine and human macrophages (34) and inhibit tu- mor growth in nude mice ( 12) suggest that MAbs of these murine IgG subclasses also have therapeutic potential in dogs with lymphoma.

Identification of parameters associated with high ADCC activity will provide infor- mation needed in understanding and developing future immunotherapeutic strategies.

428 ROSALES ET AL.

ACKNOWLEDGMENTS

This work was supported by Grant CA 108 15 from NIH. The technical and editorial assistance of Joan McDonald-Smith, Jeanne Liptock, Cynthia Biever, Marina Hoffman, and Maria Marinelli is greatly appre- ciated.

REFERENCES

1. Gillette, E. L., In “Design of Models for Testing Cancer Therapeutic Agents” (I. J. Fidler and R. J. White, Eds.). Van Nostrand Reinhold, New York, 1982.

2. Johnson, R. E., Cameron, T. P., and Kinard, R., Cancer Res. 28,2562, 1968. 3. Cuoto, C. G., In “Current Veterinary Therapy.” Saunders, Philadelphia, 1986. 4. Non-Hodgkin’s Lymphoma Pathologic Classification Project, Cancer49,2 112, 1982. 5. Carter, R. F., Valli, V. E. O., and Lumsden, J. H., Canad. J Vet. Res. 50, 154, 1986. 6. Carter, R. F., Harris, C. K., and JegIum, K. A., In “Proceedings, American College of Veterinary

Pathologists,” 1986. 7. Madewell, B. R., Amer. J. Vet. Res. 36,1525, 1975. 8. MacEwen, E. G., Brown, N. O., and Patnik, A. K., JAVMA 178, 1178, 198 1. 9. Cotter, S. M., J. Amer. Anim. Hosp. Assoc. 19, 159, 1983.

10. Jeglum, K. A., Young, K. M., Barnsley, K., Whereat, A., McGrath, D., and Huston, C., J. Rio/. Re- sponse Mod. 5,168, 1986.

11. Steplewski, Z., Jeglum, K. A., Rosales, C., and Weintraub, N., Cancer Zmmunol. Immunother. 24, 197,1987.

12. Herlyn, D., and Koprowski, H., Proc. Natl. Acad. Sci. USA 79,4761, 1982. 13. Herlyn, D., Steplewski, Z., Herlyn, M., and Koprowski, H., Cancer Res. 40,7 17, 1980. 14. Sears, H. F., Herlyn, D., Steplewski, Z., and Koprowski, H., J. Biol. Response Mod. 3, 138, 1984. 15. Sears, H. F., Mattis, J., Herlyn, D., Hayry, P., Atkinson, B., Ernst, C., Steplewski, Z., and Koprowski,

H., Lancet 1,762, 1982. 16. Miller, R. A., Maloney, D. G., Wamke, R., and Levy, R., N. Engl. J. Med. 306,5 17, 1982. 17. Steplewski, Z., Blaszczyk, M., Herlyn, D., Herlyn, M., and Koprowski, H., In “Monoclonal Antibodies

and Breast Cancer” (R. L. Ceriani, Ed.), pp. 134- 149. M. Nijhoff, Amsterdam, 1985. 18. Spira, G., Bargellesi, A., Teillaud, J., and Scharff, M., J. Zmmunol. Methods 74,307, 1984. 19. Steplewski, Z., Spira, G., Blaszczyk, M., Lubeck, M. D., Radbruch, A., Illges, H., Herlyn, D., Rajew-

sky, K., ScharlT, M., and Koprowski, H., Proc. Natl. Acad. Sci. USA 82,8653, 1985. 20. Kohler, G., and Milstein, C., Nature (London) 256,495, 1975. 2 1. Ey, P. L., Prowse, S. J., and Jet&in, C. R., Immunochemistry 15,429, 1978. 22. Boyum, A., Nature (London) 204,793, 1964. 23. Atkinson, K., Deeg, H. J., Storb, R., Weiden, P. L., Gerhard-Miller, L., Torok-Storb, B. J., Seigneuret,

M., and Thomas, E. D., Exp. Hematol. 8,821, 1980. 24. Freundlich, B., and Avdalovic, N., J. Immunoi. Methods 62,3 1, 1983. 25. Wiltrout, R. H., Taramelli, D., Holden, H. T., In “Manual of Macrophage Methodology” (H. B.

Herscowitz, H. T. Holden, J. A. Bellanti, and A. Graffard, Eds.), pp. 337-344. Dekker, New York/ Basel, 1981.

26. Herlyn, D., and Koprowski, H., Znt. J. Cancer 27,769, 198 1. 27. Lubeck, M. D., Kimoto, Y., Steplewski, Z., and Koprowski, H., Cell. Zmmunol. 111, 107,1988. 28. Bernstein, I. D., Tam, M. R., and Nowinski, R. C., Science 207,68, 1980. 29. Kirch, M. E., and Hammerling, U., J. Immunol. Methods 127,805, 198 1. 30. Young, W. W., and Hakomori, S. I., Science211,487,1981. 31. Badger, C. C., and Bernstein, I. D., J. Exp. Med. 157,828, 1983. 32. Key, M. E., and Haskill, J. S., In “Macrophage-Mediated Antibody Dependent Cytotoxicity” (H. S.

Koren, Ed.). Dekker, New York, 1984. 33. Lubeck, M. D., Steplewski, Z., Baglia, F., Klein, M. H., Dorrington, K. J., and Koprowski, H., J.

Immunol. 135,1299,1985. 34. Steplewski, Z., Lubeck, M. D., and Koprowski, H., Science 221,865, 1983.