Monoclonal anti-tumor necrosis factor (TNF) antibodies protect mouse and human cells from TNF cytotoxicity

37

Journal of Immunological Methods, 140 (1991) 37-43 © 1991 Elsevier Science Publishers B.V. 0022-1759/91/$03.50 ADONIS 002217599100191M JIM 05945

Monoclonal anti-tumor necrosis factor (TNF) antibodies protect mouse and human cells from TNF cytotoxicity C y n t h i a J. G a l l o w a y , M e l a n i e S. M a d a n a t a n d G. M i t r a Cutter Biological, Miles, Inc., Fourth and Parker Streets, Berkeley, CA 94701, U.S.A. (Received 12 December 1990, revised received 12 February 1991, accepted 13 February 1991)

Tumor necrosis factor (TNF) is a cytokine produced by macrophages which mediates septic shock. T N F is also associated with anti-tumor activity and is cytotoxic to many cell lines in vitro. While the clinical efficacy of anti-TNF antibodies has yet to be established, such antibodies may prevent T N F action. We have evaluated three murine monoclonal anti-TNF antibodies for their capacity to protect mouse and human cells from the cytotoxic effects of human TNF. Mouse W E H I 164 cells were tested because of their known high sensitivity to TNF. Several human cells lines were also examined: U937, HeLa, MRC5, and ME180. Two antibodies protected mouse and human cells from the cytotoxic action of TNF. The neutralizing activity was also present in the Fab fragment and insensitive to the presence of carbohydrate on the antibody. The third antibody had substantially less TNF-neutralizing activity. The relative ability of these monoclonal antibodies to protect both mouse and human cell lines from TNF-induced cell death may be an effective indicator of their in vivo efficacy. Key words: Tumor necrosis factor; Anti-tumor necrosis factor; Septic shock; Cytotoxic neutralization

Introduction

Tumor necrosis factor (TNF) is a cytokine produced by macrophages which promotes a variety of physiological responses and has a major role as a mediator of inflammation. The role of T N F in septic shock has been established (Beutler and Cerami, 1989). T N F was initially described as a mediator of cachexia (Beutler et al., 1985). T N F has been implicated in the cutaneous and intestinal lesions of graft-versus-host disease (Piguet et al., 1987). Raised serum concentrations are associ-

Correspondence to: C.J. Galloway, Cutter Biological, Miles, Inc., Fourth and Parker Streets, Berkeley, CA 94701, U.S.A.

ated with increased mortality from meningococcal infection (Waage et al, 1987) and gram negative purpura fulminans (Girardin et al., 1988). T N F is associated with tissue remodeling including stimulation of fibroblast growth (Beutler and Cerami, 1987) and cytotoxicity (Matthews and Neale, 1987). While the mechanisms which couple T N F to this variety of functions are mostly unknown, binding to a cell surface receptor seems to initiate T N F action. Antibodies which prevent T N F from binding to its receptor may inhibit T N F action. Antibodies which bind to the T N F receptor have been shown to inhibit cytotoxicity (Shalaby et al., 1990). However, anti-receptor antibodies have also been shown to have a cytotoxic effect similar to

38 TNF, possibly induced by receptor cross-linking (Engelmann et al., 1990). Alternatively, an antiT N F antibody could be used to prevent binding of T N F to its receptor and to promote clearance of T N F from the circulation via antigen-antibody complexes. Anti-TNF antibodies have also been shown to neutralize the direct cell-to-cell killing mediated by the membrane form of T N F (Perez et al., 1990). In fact, antibodies to T N F have been shown to be beneficial in several experimental models of gram negative shock (Tracey et al., 1987; Hinshaw et al., 1990; Silva et al., 1990). While anti-TNF antibodies have proven promising in these experimental systems, their clinical use in human patients remains to be established. The ability of an anti-TNF antibody to be effective in vivo may depend on several diverse factors. These would likely include the affinity of the antibody for TNF, the location of the epitope on TNF, and the ability of the antibody to inhibit TNF-dependent functions. Because different antiT N F monoclonal antibodies would be expected to differ in these and other qualities, they may also differ in their ability to prevent T N F action. Therefore, we have evaluated three murine monoclonal antibodies directed against human recombinant T N F a for their capacity to protect mouse and human cells from the cytotoxic effects of human TNF. The relative ability of these monoclonal antibodies to protect both mouse and human cell lines from TNF-induced cell death may be an effective indicator of their in vivo efficacy.

Materials and methods

Reagents Human recombinant T N F was provided by Dr. P. Olsen, Chiron, U.S.A., and had a specific activity of 2 - 4 x l0 s U / m g in the L929 cytotoxic neutralization assay (3 x 1 0 4 cells/assay). Except where noted, all other reagents were from Sigma, U.S.A. A n tibodies A10G10, A6, and B6 are mouse monoclonal antibodies produced by immunization of B A L B / c mice with human recombinant T N F a (Harlow and Lane, 1988). All are IgG1 antibodies.

Cells and cell culture Mouse W E H I 164 cells (clone 13) were obtained from Dr. T. Espevik, University of Trondheim, Norway. HeLa, U937, and MRC-5 cells were grown in DMEM. W E H I 164 cells were grown in RPMI 1640. ME180 cells were grown in McCoy's 5A medium. All media (Mediatech, U.S.A.) contained 10% fetal bovine serum (Hyclone, U.S.A.), 300 mg/1 glutamine and 1 mM sodium pyruvate. Human cell lines were obtained from American Type Culture Collection (ATCC). Carbohydrate digestions A10G10 was digested with neuraminidase (Boerhinger-Mannheim, F.R.G.) or endoglycosidase F (Endo F, Boerhringer-Mannheim) according to Haselback and Hosel (1990). To determine whether digestion with Endo F or neuraminidase was complete, the remaining glycans were detected with a glycan detection kit (BoehringerMannheim, no. 1210 238). Sambucus nigra agglutinin (SNA) was used to detect terminal sialic acid, and Galanthus nivalis agglutinin (GNA) was used to detect high mannose residues. T N F sensitivity The sensitivity of various cell lines to T N F was determined by finding the T N F concentration which produced the most cell death at various cell numbers. Cells were harvested into RPMI media containing 10% fetal bovine serum, 100 U / m l penicillin, and 0.01 m g / m l streptomycin. A 50 ffl aliquot containing the indicated number of cells was placed into each well of a 96-well microtiter plate which had been previously blocked with RPMI containing 1.0% bovine serum albumin. 50 ffl of T N F in media containing 2 ffg/ml actinomycin D was added and the plate incubated overnight at 37°C. MTI" (3-(4,5-dimethylthiazol-2yl-)2,5 diphenyl tetrazolium bromide) was used to measure cell death (Mosmann, 1983). Each well received 5 /~1 of 5 m g / m l MTT in phosphatebuffered saline. After 3 - 4 h at 37°C, the reaction was stopped using 100 /~1 of 10% SDS in 0.1 N HC1. After allowing at least 4 h to solubilize the M T T precipitate, the plates were read at 570 nm (650 nm reference). Data points were the average of duplicates.

39

Cytotoxic neutralization To measure the ability of anti-TNF antibodies to protect against TNF, anti-TNF antibody was titrated at the concentration of T N F and the number of cells which produced the greatest amount of cell death. Cells were added to 96-well flat-bottomed microtiter plates in 50 /~1 RPMI media containing penicillin-streptomycin. Cells were allowed to recover from harvest during a 4 h incubation at 37°C before addition of T N F and antibody. To prepare antibodies and T N F for addition to cells, the antibodies were serially diluted from 20 /~g/ml into media containing 2 /~g/ml actinomycin. An equal volume of 4 n g / m l T N F in the same media was added to the diluted antibodies. After incubation at room temperature for 2 h, 50 ffl of the a n t i b o d y / T N F mixture was added to the cells. The cells were then transferred to a 37°C incubator for overnight. Cells which survive this treatment by protection with anti-TNF antibody were detected using M T T (see above). Final data is derived from the average of two replicates minus the background value generated by a non-specific antibody of the same subclass.

Results

To determine if A10G10 would neutralize the cytotoxic activity of T N F on a cell line which was

very sensitive to T N F , the ability of human recombinant T N F to kill W E H I 164 cells was examined. The cytotoxic activity of T N F was titrated at various numbers of W E H I 164 cells. After an overnight incubation at 37°C, cell survival was measured with MTT. At 2 x 1 0 4 cells/well, maximal killing is achieved at 4 n g / m l T N F (Fig. la). At extreme concentrations of TNF, the killing activity decreases somewhat. To determine if antiT N F antibody inhibited cell killing, various concentrations of antibody were preincubated with T N F at 4 n g / m l and added to the cells. After an overnight incubation at 37°C, the antibody inhibited TNF-induced cell death with an IDs0 of about 1 ~ g / m l antibody (Fig. lb). The ability of an antibody to prevent T N F cytotoxicity depends on its affinity for T N F and the ability of the antibody to inhibit binding of T N F to its receptor. To determine how distinct anti-TNF antibodies would neutralize the cytotoxic activity of T N F , two other anti-TNF antibodies were tested in the neutralization assay. As shown in Fig. 2a, antibody A6 and A10G10 had comparable activities. Another antibody directed against TNF, B6, was approximately ten times less effective at neutralizing T N F (Fig. la). Most likely the B6 antibody either has a lower affinity for T N F than A10G10, or it is directed to an epitope which less effectively inhibits binding of T N F to its receptor. The Fab of A10G10 was also able to

0.7

0.5 b.

0.6 0.5 c [3

0.4

o AIOGIO/

o 2 x 10~4

0.,3

15:1100~3 -

0.4

1

10E,3

x

0.2

o

3> o Q

JD

0.3

0.1

0.2

0.0

0.1

I

I

I

I

I

I

i

I

I

f

0.00010.01 I 100 10000 ng/mlTNF

0,001

{ 0.01

I 0.1

I 1

I 10

-0.1 100

I..Ig/m I Mob

Fig. 1. a: using M T T to detect cell survival, the sensitivity of W E H I 164 cells to T N F is measured at different n u m b e r s of cells/well. o , 2 x 104 cells per well; O, 1 × 104 cells per well; v, 5 × 103 cells per well; v, 1 × 103 cells per well. b: at the T N F concentration which produced maximal killing of W E H I 164 cells, A10G10 was titrated for neutralization activity.

40 0.6

0.6 0.5

b.

a.

0.5

0.4

0.4

c o

0.5

0.3

o ,m

0.2

0.2

0.1

0.1

0.0

0.0

3> ¢y

8-

<

-0.1 0.001

0.01

0. I

I

10

0.01

0.1

1

10

o

-0.1 O0

~ g / m l Mob

lug/ml Mob

o AIOGIO • A10G10 Fab

o A10G10 • A6 '7 B6

Fig. 2. a: antibodies A10G10 (o), A6 (O), and B6 (v) were c o m p a r e d for neutralization activity using W E H I 164 cells, b: intact A10G10 ( o ) and the Fab of A10G10 (e) were c o m p a r e d for neutralization activity using W E H I 164 cells.

neutralize TNF (Fig. 2b), however it was less active than the intact molecule even when compared on a moles binding site basis. Most likely this indicates an affinity difference between the fragment and the intact molecule. The activity of some antibodies may be effected by carbohydrate modification of the heavy or light chains of the molecule (Walker et al., 1987, 1989; Wallick et al., 1988). To determine whether termi-

nal sialic acids had a role in antibody activity, A10G10 was subjected to digestion with neuraminidase, and tested for TNF-neutralizing activity on WEHI 164 cells. After digestion, removal of sialic acid was verified by lectin blotting with SNA. As shown in Fig. 3a, neuraminidase digestion has no effect on the activity of A10G10. Evidently, terminal sialic acids are not necessary for antibody activity. We next investigated if re-

0.5 b.

O.

0.4

"y

o A10G10 o o

0.5

c

o

0.2 0

0.1

0.4

p

0.3

0.2 o o" ci

0.1 ~

0.0 -0.1

0.5

• I I0

I I00

0.0

Neuramlnidase

I I I00010000

n g / m l Mab

J

I I0

I I00

J I000

I I0000

-0.1

n g / m l Mab

Fig. 3. a: neuraminidase-treated A10G10 (O) w a s c o m p a r e d to intact antibody ( o ) in the neutralization assay on W E H I 164 ceils, b: Endo F treated A10G10 (e) was c o m p a r e d to intact antibody ( o ) in the neutralization assay on W E H I 164 cells.

41

had the smallest proportion of cells killed at all concentrations. In all human cell lines, the killing activity of T N F decreased above the concentration of T N F needed to produce maximal killing, indicating that the affinity of T N F for its receptor may decrease at high ligand concentrations. In the human cell lines, the addition of more cells above 5 X 1 0 4 cells/well did not produce a greater proportion of killed cells. Therefore, it is possible that some proportion of each population is resistant to TNF. To determine if the anti-TNF antibody would also protect human cell lines from the cytotoxic effects of T N F , the ability of A10G10 and B6 to inhibit cytotoxic activity on MRC-5 cells, ME180 cells, and U937 ceils was examined. Because HeLa cells were relatively insensitive to TNF, they were not tested for antibody neutralization. As shown

moval of N-linked glycan chains would effect antibody activity. N-linked carbohydrate was removed by digestion with Endo F, and verified by lectin blotting with GNA. As shown in Fig. 3b, removal of N-linked chains also did not effect antibody activity. To determine if anti-TNF would be effective at protecting human cells from TNF-induced cell death, the ability of T N F to kill human cells was tested on four human cell lines: ME180, HeLa, and MRC5, and U937. As shown in Fig. 4, T N F was cytotoxic to ME180, HeLa, and MRC5 cells at an EDs0 of about 1 n g / m l TNF. U937 cells were more sensitive to the effects of T N F and had an EDs0 of about 0.3 n g / m l TNF. All human cell lines tested were almost 100 times less sensitive to T N F than the mouse W E H I 164 cell line. Hela cells (Fig. 4b) were the least sensitive to T N F and

1.0

1.0

cells] I~

Q.

0.8

o~c

b.

ME180

HeLa

0.8

cells

0.6

0.6

0.4

SQ o.4 ~

0.2

0.2

o .<

I I I I I E I I I I C.

0.8

--

k4RC5 cells

I I I

o '

I I I I d. U937 cells

~

I 0.8

o~ 0.6 c

0.6

g

0.4

So O.4 ~

0.2

0.2

_o .<

0.0

I I I 0 . 0 0 0 1 0.01

I

I 1

I

I 100

I

I I 10000

n g / m l TNF o

I I

0

1 x 10E5

•

5 x 10E4

I I 1 1 1 1 1 t l l 0 . 0 0 0 1 0.01 1 100

10000

o

0.0

n g / m l TNF v 2 x 10E4

.

1 x 10E4

number of cells per well Fig. 4. The sensitivity of four h u m a n cells lines (a, ME180 cells; b, H e L a cells; c, MRC5 cells; and d, U937 cells) to T N F was compared at different n u m b e r s of cells/well ( o , 1 × 105 cells per well; O, 5 x 104 cells per well; v, 2 x 104 cells per well; v, 1 x 104 cells per well).

42

MRC5 cells

ME180 cells

400 n g / m l TNF

400 n g / m l TNF

U937 cells 20 n g / m l TNF 0.4

0.6 0.5

b.

0.3

0.4 3>

g 0

Q

g

co

0.3

0.2

o

S-

0.2

Q

0.1 on

o.1

CD

0.0

0.0

-0.1 -0.2

t

I

I

0.0001 0.01

I

t

I

1

I

I

100

0.001

~ g / r n l Mob

I

I

0.1

I

I

10

I

I

I

1000

I

I

yg/ml

jJg/ml Mob o AIOGIO

I

0.0001 0.01

•

I

1

I

I

E

-0.1

100 Mob

B6

Fig. 5. At the T N F concentration which produced maximal killing of MRC5 cells (a), ME180 cells (b), and U937 cells (c), the neutralization ability of A10G10 ( o ) and B6 (e) was compared.

in Fig. 5a, when A10G10 was added to MRC-5 cells at the m a x i m u m T N F concentration required for killing, A10G10 neutralized T N F action at IDs0 of 1 /~g/ml Mab. A10G10 also neutralized T N F action on ME180 cells at a similar concentration (Fig. 5b). On the monocytic cell line U937, which was more sensitive to TNF, ten times more A10G10 was required to neutralize the T N F (Fig. 5c). As on the mouse cell line, antibody B6 also was neutralizing, but was much less effective than A10G10 (Figs. 5a, 5b and 5c).

Discussion

The ability of three anti-TNF antibodies to neutralize T N F action was investigated in a cytotoxicity assay using the uptake of M T T as an indicator of cell death. Mouse W E H I 164 clone 13 cells were chosen as a sensitive indicator cell because they were almost 100 times more sensitive to TNF-induced cell death than other lines (Espevik and Nissen-Meyer, 1986). The neutralizing activity of two monoclonals, A10G10 and B6, was examined on the mouse cell line and three human cell lines. A10G10 was a more potent neutralizing antibody than B6. The third antibody, A6, was similar in activity to A10G10. Neutralizing activity was also present in the Fab of A10G10 and

insensitive to the presence of carbohydrate on the antibody. A10G10 may protect cells from T N F action by interfering with binding of T N F to the T N F receptor. Antisera raised against amino acids 1-15 and 1-30 of T N F blocked both binding and TNF-induced cytotoxicity, while antisera raised to other portions of the molecule did not inhibit T N F binding (Socher et al., 1987). Therefore it is likely that A10G10 binds T N F near the receptorbinding domain of T N F . Antibodies like B6, which either do not block T N F receptor binding, or do so at lower affinities than A10G10, will also have less effective neutralizing activity. The effectiveness of an anti-TNF antibody in vivo may also depend upon the effector function of the IgG. For example, T N F could be efficiently cleared from the circulation as an antigen-antib o d y complex by Fc receptors on macrophages. A10G10 is a mouse IgG1 antibody and would be expected to bind to the h u m a n F c R I I (Anderson, 1989). The high concentration of antibody needed to protect the monocytic U937 cells in vitro may be a manifestation of Fc binding in the assay system. Because only a few molecules of T N F are needed to cause cell death, perhaps binding of A10G10 to Fc receptors lowers the effective neutralizing activity of the antibody. In an in vivo system with a much larger capacity to bind, inter-

43 nalize, and degrade antigen-antibody complexes, one would expect Fc-related binding to potentiate clearance of TNF. This study indicates that monoclonal TNF antibodies differ in their ability to neutralize TNF cytotoxicity. Such antibodies may also differ in their ability to prevent TNF action in vivo. While t h e a b i l i t y o f a n t i - T N F to n e u t r a l i z e c y t o t o x i c i t y may not be the only quality necessary for a potent therapeutic antibody, in at least two studies, antiTNF antibodies which were effective at cytotoxic n e u t r a l i z a t i o n a l s o p r o v i d e d p r o t e c t i o n f r o m septic s h o c k i n a n i m a l m o d e l s ( L u c a s e t al., 1 9 9 0 ; S i l v a et al., 1990). I n v i t r o s c r e e n i n g o f t h e s e antibodies for their relative ability to prevent TNF-mediated cell d e a t h o f m o u s e a n d h u m a n cells m a y b e a g o o d i n d i c a t o r o f t h e i r e x p e c t e d i n vivo efficacy.

References Anderson, C.L. (1989) Human IgG Fc receptors. Clin. Immunol. Immunopathol. 53, $63-$71. Beutler, B. and Cerami, A. (1987) Cachectin: more than a tumor necrosis factor. New Engl. J. Med. 316, 379-385. Beutler, B. and Cerami, A. (1989) The biology of cachect i n / T N F - a primary mediator of the host response. Ann Rev. Immunol. 7, 625-55. Beutler, B., Mahoney, J., Le Trang, N., Pekala, P. and Cerami, A. (1985) Purification of cachectin, a lipoprotein lipasesuppressing hormone secreted by endotoxin-induced RAW 264.7 cells. J. Exp. Med. 161,984-995. Engelmann, H., Holtmann, H., Brakebusch, C., Avni, Y.S., Sarov, I., Nophar, Y., Hada, E., Leitner, O. and Wallach, D. (1990) Antibodies to a soluble form of a tumor necrosis factor (TNF) receptor have TNF-like activity. J. Biol. Chem. 265, 14497-14504. Espevik, T. and Nissen-Meyer, J. (1986) A highly sensitive cell line, WEHI 164 clone 13, for measuring cytotoxic factor/ tumor necrosis factor from human monocytes. J. Immunol. Methods 95, 99-105. Girardin, E., Grau, G.E., Dayer, J.-M., Roux-Lombard, P., the J5 Study Group and Lambert, P.-H. (1988) Tumor necrosis factor and interleukin-1 in the serum of children with severe infectious purpura. N. Engl. J. Med. 319, 397-400. Harlow, E. and Lane, D. (1988) Antibodies, A Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, pp. 53-244. Haselbeck, A. and Hosel, W. (1990) Description and application of an immunological detection system for glycoproteins on blots. Glycoconjugate J. 7, 63-74. Hinshaw, L.B., TeKamp-Olson, P., Chang A.C., Lee P.A., Taylor, F.B., Murray, C.K., Peer, G.T., Emerson Jr., T.E.,

Passey R.B. and Kuo G.C. (1990) Survival of primates in LD100 septic shock following therapy with antibody to tumor necrosis factor (TNF alpha). Circ. Shock 30, 279-292. Lucas, R., Heirwegh, K., Neirynck, L., Remels, H., Van Heuverswyn, H. and De Baetselier, P. (1990) Generation and characterization of a neutralizing rat anti-rm TNF-a monoclonal antibody. Immunology 71,218-223. Matthews, N. and Neale, M.L. (1987) Studies on the mode of action of tumor necrosis factor on tumor cells in vitro. In: E. Pick (Ed.), Lymphokines, Vol. 14. Academic Press, London, pp. 223-252. Mosmann, T. (1983) Rapid colorimetric assay for cellular growth and survival, application to proliferation and cytotoxicity assays. J. Immunol. Methods 65, 55-63. Perez, C., Albert, I., De Fay, K., Zachariades, N. and Kriegler, M. (1990) A nonsecretable cell surface mutant of tumor necrosis factor (TNF) kills by cell-to-cell contact. Cell 63, 251-258. Piguet, P.-F., Grau, G.E., Allet, B. and Vassalli, P. (1987) Tumor necrosis factor/cachectin is an effector of skin and gut lesions of the acute phase of graft-vs-host disease. J. Exp. Med. 166, 1280-1289. Shalaby, M.R., Sundan, A., Loetscher, H., Brockhaus, M., Lesslauer, W. and Espevik, T. (1990) Binding and regulation of cellular functions by monoclonal antibodies against human tumor necrosis factor receptors. J. Exp. Med. 172, 1517-1520. Silva, A.T., Bayston R.F. and Cohen, J. (1990) Prophylactic and therapeutic effects of a monoclonal antibody to TNFa in experimental gram negative shock. J. Infect. Dis. 162, 421-427. Socher, S.H., Riemen, M.W., Martinez, D., Friedman, A., Tai, J., Quintero, J.C., Garsky, V. and Oliff, A. (1987) Antibodies against amino acids 1-15 of tumor necrosis factor block its binding to cell-surface receptor. Proc. Natl. Acad. Sci. U.S.A. 84, 8829-8833. Tracey, K.J., Fong, Y., Hesse, D.G., Monogue, K.R., Lee, A.T., Kuo, G.C., Lowry, S.F. and Cerami, A. (1987) Anticachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteraemia. Nature 330, 662-64. Tracey, K.J., Vlassara, H. and Cerami, A. (1989) Cachectin/ tumor necrosis factor. Lancet i, 1122-1125. Waage, A., Halstensen, A. and Espevik, T. (1987) Association between tumour necrosis factor in serum and fatal outcome in patients with meningococcal disease. Lancet i, 355-357. Walker, M.R., Lee, J. and Jefferis, R. (1987) Immunogenicity and antigenicity of immunoglobulins: detection of human immunoglobulin light-chain carbohydrate, using concanavalin A. Biochim. Biophys. Acta 915, 314-320. Walker, M.R., Lund, J., Thompson, K.M. and Jefferis, R. (1989) Aglycosylation of human IgG1 and lgG3 monoclonal antibodies can eliminate recognition by human cells expressing FcRI a n d / o r FcRII receptors. Biochem. J. 259, 347-353. Wallick, S.C., Kabat, E.A. and Morrison, S.L. (1988) Glycosylation of a V H residue of a monoclonal antibody against a(1-6) dextran increases its affinity for antigen. J. Exp. Med. 168, 1099-1109.