J Cancer Res Clin Oncol (1991) 117:239-243 017152169100044Q

C~/icer ~esearch Clinical 9 (~) Springer-Verlag1991

Synergistic cytotoxicity of human recombinant tumour necrosis factor with microtubule effectors

combined

J. Baumgart, B. Schlott, J. Suehnel, W. Vater, W. Schulze, and D. Behnke Institute of Microbiology and Experimental Therapy, Beutenbergstrasse 1I, 0-6900 Jena, Federal Republic of Germany Received 22 October 1990/Accepted 22 January 1991

Summary. Combinations of human recombinant tumour necrosis factor ~ (rhTNF~) with each of four different agents disturbing the microtubule system of the cellular cytoskeleton were tested for synergistic cytotoxic action against murine melanoma B16K and L-M(S) cells. In addition to the known microtubule effectors colchicine, vincristine, and taxol, the influence of the fluorenone-azomethine derivative e-diphenylene-N-{p-[bis-(/~-hydroxyethyl)-amino]-phenyl}-nitrone ( D H P N ) on the rhTNFc~ cytotoxicity was studied. Applying a novel computerbased isobole method [Suehnel J (1990) Antiviral Res 13:23~40] concentration ranges of synergistic, zero, and antagonistic interaction were found after in vitro combination of r h T N F e with each of the drugs tested in a 72-h cytotoxicity assay. In contrast, a 24-h exposure of B16K cells to these combinations still did not inhibit in vitro colony formation to a greater extent than either drug alone. A preliminary in vivo experiment revealed an i n creased antitumour effect after treatment of established subcutaneous melanoma B16 tumours with a combination of r h T N F e and D H P N . Key words: T N F - Microtubule effectors - Synergism

Combination

Introduction T u m o u r necrosis factor e (rhTNFe) is produced by activated macrophages and exhibits cytotoxicity towards various tumour cell lines and transplantation tumours. Therapeutic application of high doses of r h T N F a is, however, limited by adverse side-effects observed in clinical trials (Lenk et al. 1989; Retsas et al. 1989). Recently evidence has been accumulated that the cytotoxic activity

of r h T N F e can be substantially enhanced by its combination with certain anticancer agents. Synergistic effects have been found with combinations of r h T N F e and interferons (Taguchi et al. 1987; Dubois et al. 1989; Balkwill et al. 1987; Broukaert et al. 1986; Fransen et al. 1986), drugs targeted at D N A topoisomerase II (Alexander et al. 1987; Das et al. 1989; Burgers et al. 1989), as well as other antineoplastic agents (Watanabe et al. 1988; Regenass et al. 1987; Fruehauf et al. 1990; Itano et al. 1987; Baumgart et al. 1990). Conflicting results have been reported on the interaction of r h T N F e with microtubuledisrupting agents. Watanabe et al. (1988) and H a and Lau (1988) found a synergistic increase in cytotoxicity when T N F e was combined with vincristine, while Aapro et al. (1986) and Shibayama (1989) reported only a likely additive effect. Alexander et al. (1987) were even unable to detect any interactive effect of these drugs. In the present study, therefore, the cytotoxic activity ofvincristine on r h T N F e cytotoxicity was reinvestigated over a wide concentration range. Moreover, the well-known microtubule-disturbing agents colchicine and taxol as well as a recently described fluorenone-azomethine derivative, D H P N (Oertel et al. 1990) were included in this study. The data obtained from a 72-h cytotoxicity assay with murine melanoma B16K cells and the fibroblast lines L-M(S) and L-M(R) were processed by a novel computer-based isobole method (Suehnel 1990; see Materials and methods). This approach enabled us to arrive at quantitative conclusions on the type and extent of possible interactions of the combined drugs for the whole dose range under study. The cytotoxicity data were further confirmed by a simple colony-forming assay. Initial in vivo experiments were carried out with combinations of r h T N F e and D H P N .

Materials and methods Abbreviations: rhTNFe, human recombinant tumour necrosis fac-

tor e; DHPN, e-diphenylene-N-{p-[bis-(/~-hydroxyethyl)-amino]phenyl}-nitrone Offprint requests to: J. Baumgart

Mice. C57BL/6 males, 6- to 8-week-old,were obtained from the SPF mouse-breeding unit of this institute. Standardized pellet food and acidified tap water were provided ad libitum.

240

O-~ ~--N(CH2 CH2OH)2 Fig. l. Chemical structure of azomethine derivative c~-diphenyleneN- {p-[bis-(/~-hydroxyethyl)-amino]-phenyl}-nitrone (DHPN)

Melanoma B16. The tumour was passaged on syngeneic C57BL/6 mice, 12- to 14-day-old turnouts were mechanically disintegrated and aliquots of 0.2 ml of a 1 : 10 v/v suspension in sterile saline were transplanted subcutaneously on to each experimental animal at day 0. Chemotherapy. Randomized tumour-bearing animals were daily treated with rhTNFe with or without DHPN suspended in sterile 1.5% hydroxyethylcellulose solution from days 7 to 11. While rhTNFc~ was intraveneously injected DHPN was administered intraperitoneally. The doses administered are given in Table 2. Perpendicular axes of tumours were measured twice a week by caliper and volumes were calculated by the formula:

~. Cell lines. The mouse melanoma B16K cell line was cultured in medium 199 (Sifin, Berlin) supplemented with 0.55 mg/ml sodium pyruvate, 0.20 mg/ml L-glutamine, 0.50 gg/ml folic acid, penicillin, streptomycin, and 10% calf serum. The TNF-sensitive and resistant mouse fibroblast lines L-M(S) and L-M(R), respectively, were grown in Eagle's MEM (Sifin, Berlin) supplemented with penicillin, streptomycin, and 10% calf serum. Drugs. Human recombinant tumour necrosis factor c~(rhTNFe) was purified in our laboratory from a recombinant E. eoli strain. The rhTNFe protein lacked the initial three N-terminal amino acids of mature rhTNFe, carried no N-terminal methionine and had a specific activity of 1.1 x 1 0 7 U/mg as determined by the murine L929 cytolytic bioassay (Gase et al. 1990). The fluorenone-azomethinederivative e-diphenylene-N-{p-[bis(/3-hydroxyethyl)-amino]-phenyl}-nitrone (DHPN) (Fig. l) was dissolved in 100% dimethylsulphoxide (DMSO) and subsequently diluted with culture medium to a maximum final concentration of 0.1% DMSO. Taxol was kindly provided by Dr. H. Mueller (Jena) and was isolated from the stem bark of Taxus baccata L. The product contained at least 50% taxol and was dissvoled in DMSO to a maximum final concentration of 0.1% DMSO. Colchicine and vincristine of clinical grade (Gedeon Richter AG, Hungary) were dissolved in isotonic NaC1 solution immediately before use. Mierotubule protein assembly assay. An in vitro assay based on microtubule protein isolated from porcine brain was performed as reported elsewhere (Vater et al. 1983). It was used to determine the influence of 1.4 gM and 2.1 gM DHPN alone and in combination with 20 gg/ml rhTNFe on tubulin assembly. Colony-forming assay. 250 BI6K cells were cultured overnight in plastic Petri dishes. Dilutions of single drugs or combinations with rhTNFc, were then added in triplicate. After an incubation period of 24 h the drug-containing medium was aspirated, the cells were washed gently and cultivation was continued for 7 days in fresh medium. Colonies of more than 30 cells were counted and the results were expressed as the percentage of the mean number of control colonies.

In vitro cytotoxicity assay. As reported previously (Baumgart et al. 1990), cells were treated simultaneouslywith 72 different drug combinations. Single-drug and medium controls (octuplets) were included in each 96-well microtiter plate. Duplicates of I x 1 0 4 freshly trypsinized cells/well were exposed for 72 h to serial twofold dilutions of each of the microtubule effectors and/or rhTNFe. Then the cell layers were stained with 0.016% neutral red, washed gently with saline, and the absorbance was measured at 550 nm after elution of the dye with 50% ethanol containing 0.05 M acetic acid. The mean values were used to calculate the percentage of cytotoxicity for each drug combination according to the formula: cytotoxicity (%)= [A (control)-A (treated)/A (control)] x 100. Data processing. The isobole method was used to evaluate possible interactive effects of drug/rhTNFe combinations on the extent of cytotoxicity. The isobole equation I = xl/X1 +xz/X2 yields the index of interaction L The quantities xl, x2 represent the drug doses used in combination experiments while Xx and )(2 are the doses from experiments with single agents required to produce the same level of cytotoxity. The latter figures can be obtained from the dose/ response curves of the single agents. An interaction index I smaller than 1 indicates synergism, greater than 1 antagonism, and equal to 1 zero interaction. Recently a novel microcomputer-based isobole method was proposed, which fits spline functions to the response data. The program allows one to calculate isoboles for all effect levels from the response surface constructed (Suehnel 1990). Synergistic, zero or antagonistic interaction can then be identified either directly from the shapes of the isoboles or in more complicated cases by invoking the isobole equation. In this paper the combination experiments were analysed by adopting the very same idea of fitting spline functions. However, these functions were not fitted to the response data but instead to the indices of interactions, which were calculated before for all dose combinations. This yields the "surface" of interaction indices. The corresponding contour plot shows lines of equal interaction. Thus, this procedure directly reveals regions of synergistic or antagonistic interaction in the dose range under study. For the calculation of surfaces and contour plots any computer program capable of generating three-dimensional graphics can be used. We applied the programs QGRID, SURF, and TOPO distributed by Golden Software, 1984. The smoothing factor necessary for the spline-fitting procedure was 0.995 in all cases.

Results F i g u r e 2 shows the dose/response curves for all the single drugs tested in a 72-h cytotoxicity assay. It is o b v i o u s t h a t the L-M(S) cell line was a b o u t n i n e t i m e s m o r e sensitive to r h T N F e t h a n the B16K cell line, whereas L - M ( R ) cells were c o n f i r m e d to be resistant to rhTNFc~ a c t i o n (Fig. 2A). The dose/response curves o b t a i n e d for D H P N o n all three cell lines showed a characteristic response plateau, which was reached at a c o n c e n t r a t i o n of a b o u t 0.1 ~g/ml D H P N . A further increase in the drug concent r a t i o n did n o t result in a m a r k e d l y e n h a n c e d i n h i b i t i o n o f cell p r o l i f e r a t i o n (Fig. 2B). Similar characteristics were f o u n d for vincristine, colchicine a n d taxol (Fig. 2C) w h e n they were tested a g a i n s t the B I 6 K cell line. The c o m p u t e r - g e n e r a t e d lines of equal interaction are presented in Fig. 3. C o m p a r i s o n o f the graphs revealed that regions o f synergistic a n d a n t a g o n i s t i c i n t e r a c t i o n were divided by a region o f zero i n t e r a c t i o n ( I = 1 . 0 _ 0.2) c o r r e s p o n d i n g in each case to a " t h r e s h o l d c o n c e n t r a t i o n " o f the m i c r o t u b u l e - d i s t u r b i n g agent (Fig. 3 A - E ) . Higher c o n c e n t r a t i o n s o f the m i c r o t u b u l e i n h i b i t o r comb i n e d with very low doses of r h T N F e certainly resulted

241

@

CYTOTOXICITY 100 (%)

CYTOTOXICITY 100 (%)

/'2

SO

~

~ 0.1

SO

.

CYTOTOXICITY 100 (%)

. 1.0

. 10.0

rhTNF-o( (ng/m[)

o.ol

0.01

0.1

1.0

INHIBITOR (,ug/mr)

0.25

25

0.0(

6.25~

~

1.56 0.19

028

0:78

312 lg5 rhTNF-c~(ng/ml.) (]~)

DHPN (,ug/mr) 1.0

2.0 ~,,\ ,

0.78

'3.12 125 rhTNF-~(ng/ml) (~

0.19

(~

0.'78

OHPN (#g/m[) 1.0

3.12 125 r hTNF-o,:(ng/mt) (~

:!

~0.~. - - ~ - - - ~ ~ 106 __ 0.8- - ~ ~

0.19

COLCHICINE (ng/mt)

|

(ng/rnt) 100

(~)

TAXOL(ng/mt) L,2S

DHPN (,ug/mr)

t'o

Fig. 2. A Cytotoxic activity (%) of rhTNFct on melanoma B16K (o), L-M(S) (A), and L-M(R) cells (V) after an incubation period of 72 h of 1 x 104 cells/well seeded. Results expressed as mean value of at least three independent experiments. B Cytotoxic activity of DHPN on melanoma B16K (o), L-M(S) (zx), and L-M(R) ceils (~7). (Experimental conditions as for A.) C Cytotoxic activity of colchicine (I), vincristine (0) and taxol (D) on melanoma B16K cells. (Experimental conditions as for A.)

VINCRISTINE

0.016 ~ 0.19

oi'~

(~

SO

0HPN (#g/mr) 1.0

|

3:12 12.8 rhTNF-a (ng/m[)

0.01~

0.019

0.078

0.31 1.89 rhTNF-cz(rig/rot)

0.016 1.95

7.80

3.12 125 rhTNF-o~(rig/ml.)

Fig. 3A-F. Computer-assisted graph of lines of equal interaction after combined treatment of different cell lines with microtubule inhibitors and rhTNFct in a 72-h eytotoxicity assay. The graphs are based on interaction indices calculated from the corresponding cytotoxicity data :of combination experiments which can be requested from the authors and the dose/response data of the single drugs (Fig. 2). The numbers indicated are the values of the interaction index. To avoid a distgrtion of the graphs in a few cases the maximum interaction index was defined to be 4.0. The regions of zero interaction (I= 1.0 4-0.2) are hatched. Lines of equal interaction after combined treatment of melanoma B 16K cells with (A) DHPN and rhTNF~, (B) vincristine and rhTNF~, (C) colchicine and rhTNFct, (D) taxol and rhTNF~, and of L-M (S) cells with (E) DHPN and rhTNF~, and (F) L-M(R) cells with DHPN and rhTNF~

242 Table 1. Inhibition of colony formation of melanoma B16K cells

after a 24-h incubation with e-diphenylene -N-{p-[bis-(~-hydroxyethyl)-amino]-phenyl}-nitrone (DHPN) alone and in combination with human recombinant tumour necrosis factor c~(rhTNFz) DHPN (p.g/ml)

rhTNFe (ng/ml)

Colony number a Interaction (% of control) index b I

0.1 0.05 0.01 0.01

10.0 10.0 10.0 100.0

19.7_+4.0 25.9_+3.1 74.9_+ 3.3 65.2 _+7.6

1.28 0.94 0.87 1.68

a Results are expressed as a percentage of the mean colony number of untreated controls (~_+SD, n = 3) b Interaction indices were calculated using the isobole equation and the data of the dose/response curves of the single agents. Colony numbers were reduced to 63.8%; 81.7%; 80.8%; and 85.2% after treatment with 100, 10, 1.0, and 0.1 ng/ml rhTNFc< Colony numbers were reduced to 14.5%; 15.0%; 27.1%; 89.0%; and 100.2% after treatment with 0.2, 0.1, 0.05, 0.02, and 0.01 gg/ml DHPN Table2. Growth inhibition of established subcutaneous B16

melanomas of C57BL/6 mice after treatment with DHPN alone and in combination with rhTNFct DHPN (rag/mouse)

rhTNFc~ (gg/mouse)

0.08 0.08

1.0 1.0

Tumour volume" Percentage (mm3 x 103) of control (%) 2.79 + 0.72 2.96_+0.93 1.85_+0.49 1.08 _+0.21

100 106.1 66.3 38.7 *

9 Results are expressed as mean tumour volumes of treated and control animals on day 17 after tumour transplantation 02-4-SEM; n= 10) 9 Significant relative to control (P