GYNECOLOGIC

ONCOLOGY

38, 333-339

(1990)

Synergy Is Documented in Vitro with Low-Dose Recombinant Tumor Necrosis Factor, Cisplatin, and Doxorubicin in Ovarian Cancer Cells’ BENJAMIN BONAVIDA, PH.D., TOSHIHARU TSUCHITANI, M.D., JACOB ZIGHELBOIM, M.D., JONATHAN S. BEREK, Department

of Microbiology and Immunology, School qf Medicine, Jonsson

and Division Comprehensive

of Gynecologic Cancer

Center.

AND

M.D.

Oncology, Uni\fersity

Department of Obstetrics and Gynecology, of Cal$wniu, Los Angeles 90024

UCLA

Received January 22. 1990

INTRODUCTION

Ovarian carcinomas have been shown to be sensitive or insensitive to the in vitro exposure of several cytotoxic drugs and cytokines. Because of the potential for cytokines to enhance the efficacy of chemotherapeutic agents and to improve their therapeutic index, the optimal dose and schedule of the combination of these agents have been studied. We examined the cytotoxic effect of a combined modality using a variety of concentrations of recombinant tumor necrosis factor (rTNF) (a cytotoxic cytokine) with Adriamycin (ADR) and cisplatin (CDDP) on human ovarian carcinoma cell lines. Cytotoxicity was determined in a 24-hr “Cr-release assay and confirmed in a S-day viability culture assay. Five cell lines were used: PA-l, 222, OVCAR-3, SKOV3, and OVCAR-8. Doses of rTNF that were minimally cytotoxic resulted in significant cytotoxicity and synergy when used with optimal or suboptimal concentrations of ADR or CDDP. This synergy was observed in four cell lines. Interestingly, the rTNFand drug-resistant SKOV-3 cell line was sensitive to the synergistic effect of Adriamycin and rTNF. The synergistic effect that was obtained was specific to rTNF, while the combined use of ADR and CDDP or recombinant interleukin-2 and cytotoxic drugs had no synergistic effect on tumor cell lysis. Further, the addition of anti-TNF antibody abrogated the synergistic effect seen with rTNF and the cytotoxic drugs. These studies demonstrate clearly that significant synergistic antitumor cytotoxic activity against human ovarian carcinoma cell lines can be achieved with combinations of low doses of rTNF and ADR or CDDP, suggesting their possible adaptation in viva for cancer therapy. Further, the studies suggest that rTNF and the cytotoxic drugs tested may share a common lytic pathway. Since rTNF used alone has been relatively inactive in clinical trials, its potential activity may be apparent only when combined with conventional cytotoxic chemotherapeutic agents and when administered in relatively low concentration. 0 1990 Academic Press, Inc.

Recombinant

tumor necrosis factor y (rTNF)

has at-

tracted much attention because of its strong antitumor activity and high target cell specificity [I]. rTNF shows various degrees of cytotoxic and cytostatic activity against some. tumor cells and no detectable activity in others [2-51. No explanation exists for this differential susceptibility to rTNF. Although little is known about underlying molecular mechanisms responsible for rTNFmediated activity, it has been reported that some biological response modifiers or cytotoxic agents are able to modulate rTNF-mediated activity. For example, yinterferon (rIFN) has the ability to enhance the sensitivity of some tumor cells to rTNF [6]. This mechanism for rIFN augmentation of rTNF-mediated activity currently is considered to result from an increase in the number of rTNF receptors on target cells [7-91. Also, it has been reported that some inhibitors of protein or RNA synthesis greatly augment rTNF-mediated cytotoxicity by inhibiting the synthesis of some proteins that protect cells from the cytotoxicity [IO-131. Recent studies in our laboratory have demonstrated that combination treatment of human ovarian cell lines with rTNF and cytotoxic drugs resulted in a highly significant cytotoxic effect. In the present study we have concentrated on the combination of rTNF, Adriamycin, and cisplatin because the latter two have been among the most widely used anticancer agents for ovarian cancer. We have designed the present study to test the hypothesis that rTNF, Adriamycin, and cisplatin share a common pathway in their lytic mechanism and a complementary or additive cytotoxic effect could be achieved when they are used in combination. Furthermore, be-

’ Supported in part by a grant from the Concern Foundation of Los Angeles and the Ramona Moskovitz Cancer Research Fund. 333

0090-8258/90 $ I.50 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

Copyright 0

334

BONAVIDA

TABLE 1 Sensitivity of Human Ovarian Carcinoma Cell Lines to rTNF and Adriamycin rTNF @g/ml)

PA- I 222 OVCAR-3 SKOV-3 OVCAR-8

ET AL.

cell lines were all obtained from the American Type Culture Collection (ATCC). Culture Medium and Reagents

ADR (a/ml)

1

IO

100

1

10

100

* 2 -

+ + -

N.D. N.D. -

k -c -

2+ 2+ ? -

3+ 3+ + -

Note. -, I0 but 20 but 30 but ~40; 3 +, >40; N.D., not determined.

cause rTNF in very low doses has shown augmentation of in vitro cytotoxicity against ovarian cancer cells in preliminary studies in our laboratory, we sought to determine if low-dose rTNF could do this consistently against several ovarian cancer cell lines. When combined, synergistic effects of rTNF, Adriamycin, and cisplatin were found. MATERIALS AND METHODS Tumor Cells

Five human ovarian carcinoma cell lines, PA-l, 222, SKOV-3, OVCAR-3, and OVCAR-8, were used. All are plastic-adherent cells. Trypsinization is required to detach the cells from plastic tissue culture flasks and to make single-cell suspensions. The time required for trypsinization depends on the cell line used. Ten minutes of trypsinization is needed for PA-l, 222 and OVCAR-3; more than 30 min is required for SKOV-3 and OVCAR8. The cell line 222 is a relatively fresh ovarian carcinoma that was established in our laboratory and has been maintained in culture for over 2 years. The remaining four

The culture medium used consisted of RPMI-1640 (M. A. Bioproducts) supplemented with 10% heat-inactivated fetal bovine serum (GIBCO), 1% L-glutamine (GIBCO), 1% nonessential amino acids (GIBCO), 1% penicillin-streptomycin (GIBCO), and 1% fungizone (GIBCO). Purified human recombinant tumor necrosis factor (rTNF, Lot 3056-55) was a gift obtained from Genentech, South San Francisco, California. Adriamycin was purchased from Sigma Chemical Company, St. Louis, Missouri, and cisplatin from Bristol Laboratories, Syracuse, New York. “Cr-Release

Assay

Tumor cells were radiolabeled with Na, “CrO, (Amersham), by suspending 5 x 10’ cells in 1 ml RPMI-1640 containing 10% FBS to which 100 &i of “Cr was added. The cells were incubated for 2 hr at 37°C in 5% COZair, washed three times in HBSS, and resuspended in tissue culture medium at the concentration of 2 x lo5 cells/ml. “Cr-labeled target cells were placed in individual wells of 96-well round-bottomed microtiter plates (Corning Glass Works, Corning, NY) and tests were done in triplicate. A target cell suspension of 0.1 ml and 0.1 ml of reagent solutions (when mixed, 0.05 ml each of TNF, Adriamycin, and cisplatin were used) were added to each well, and incubated at 37°C in 5% C02air for 24 hr. After incubation, plates were centrifuged at 200g for 10 min, and 100 ~1 of cell-free supernatant was collected from each well. The 5’Cr content (exp) of each supernatant was determined using a Beckman 5500 gamma counter. Maximum “Cr release (max) was determined from the supernatant of cells that had been

6o a 50

TNF and Acwa

5

60

*

x c + 0 s 5 x

45

30 15

*

10' TNF(ng/ml)

p( 0.001

! p(

p( 0.001

0.01

0 10-2

10-l

100

10'

TNF(ng/ml)

FIG. 1. Cytotoxicity of a constant dose of ADR with varying concentrations of TNF (toe2 to IO’ rig/ml) against PA-1 ovarian carcinoma cell line. Both agents were added simultaneously to the radiolabeled tumor cells and cytotoxicity was measured as described under Materials and Methods. The doses of ADR added were (a) IO-’ pg/ml and (b) IO pg/ml.

rTNF AND CHEMOTHERAPY

SYNERGY IN OVARIAN

335

CANCER

cc,g 30 0 z +$ 20

+ 0 2 5

x

x

10

! p( 0.01 # p( 0.05

0

45

30 15

*

p( 0.001

! p( 0.01 0

10-3

10-2 TNF ("g/z)

-1

100

10-3

10-2 TNF(ng,:)

-1

100

FIG. 2. Cytotoxicity of rTNF and ADR used in combination was measured against the 222 ovarian carcinoma cell line. The concentrations of rTNF varied in the range lo-’ to 10” rig/ml. The doses of ADR added were (a) lo-’ pg/ml and (b) I pg/ml.

TNF(ng/ml)

TNF (rig/ml)

FIG. 3. Cytotoxic activity of rTNF in combination with a constant dose of ADR was measured against the OVCAR-3 ovarian carcinoma cell line. The doses of ADR added were (a) 10 Fg/ml and (b) 100 pg/ml.

0

10 -1

::F 0 (ng,:;

1

IO2

0

10 -1

IO0

10'

10 2

TNF (rig/ml)

FIG. 4. Cytotoxicity of the same modality as in Fig. 3 was measured against the SKOV-3 ovarian carcinoma cell line. The doses of ADR were (a) 10 pg/ml and (b) 100 fig/ml.

336

BONAVIDA

ET AL.

paper using a multiple cell harvester (MASH 11, M.A. Bioproducts). After being dried at room temperature, the glass-fiber filter paper was placed in the scintillation vial with 2 ml of liquid scintillator (Ecoscint, National Diagnostic), and [3H]leucine incorporation was determined using a liquid scintillation counter (Beckman LS 7500). All experiments were set in triplicate, and the mean value was used for calculation of the percentage uptake of control. The results were obtained with the following formula:

TNF (rig/ml)

FIG. 5. Cytotoxicity of a constant concentration of ADR (10 pg/ml) with various concentrations of rTNF (10-l to 10’ rig/ml) against the OVCAR-8 ovarian carcinoma cell line.

lysed by the addition of 2.0% Triton X-114, and spontaneous release (spont) was determined from target cells incubated in medium alone. The mean value in triplicate was used for calculation of percentage specific lysis. The percentage specific release was calculated by the following formula: Percentage specific release = (w-b, - wmspont)/(wmmax - wmspont) x 100. Protein Synthesis Protein synthesis was determined by incorporation of [3H]leucine (New England Nuclear, Boston, MA) into the tumor cells. Tumor target cells were incubated for 24 hr at 37°C in 5% CO,-air, with 1 &i [3H]leucine present during the last 18 hr. Cell suspension (0.1 ml) and reagent solutions (0.1 ml) comprising 0.05 ml of rTNF and 0.05 ml of a single anticancer agent or cycloheximide were added to each well of 96-well flatbottomed microtiter plates (Nunclon) at a final concentration of 2 x lo4 cells/O.2 ml per well. When rTNF or a single chemotherapeutic agent was used alone, 0.05 ml of culture medium was added. After incubation, the supernatant was discarded and trypsin was added to each well. The cells were harvested onto the glass-fiber filter

O-O

Adriomycin

n

Cycloheximide

-m

Percentage uptake of control = CPM,,,/CPM,,,,,,

The synergistic effect of TNF, Adriamycin, and cisplatin was statistically evaluated with the Student t test. The cytotoxicity of combination treatment was compared to the sum of cytotoxicity of each agent used alone. RESULTS Sensitivity and Resistance of Human Ovarian Carcinoma Cell Lines to rTNF and Adriamycin Five human ovarian cell lines were tested by the “Crrelease assay to determine their sensitivity to various concentrations of rTNF and Adriamycin. As shown in Table 1, PA-l and 222 are both vulnerable to rTNF and Adriamycin, and their vulnerability behaves in a dosedependent fashion. OVCAR-3 is completely insensitive, even to the highest concentration of rTNF (100 rig/ml), but relatively sensitive to Adriamycin, although less sensitive than PA-l or 222. SKOV-3 and OVCAR-8 exhibit a high degree of resistance to rTNF and Adriamycin. Synergistic Cytotoxicity against PA-l and 222 Ovarian Carcinoma Cell Lines of ADR in Combination with rTNF The cytotoxic effect of various concentrations of rTNF used in combination with a constant dose of Adriamycin

60 :; :: % 5

\ \

100.

Statistics

.-.

CHX + TNF(lOng/ml)

TNF( 1 Ong/ml):

gz s+ .g-Lo\o \.

x

15% cytotoxicity

45

30

R 15

0 01

‘\ , 10-4

10-J

10-2

10-l

100

10’

h/ml)

Cycloheximide(pg/ml)

FIG. 6. (a) Protein synthesis inhibition of PA-l by various concentrations of ADR and CHX (10m4to lOI pg/ml). (b) Augmentation of rTNF-mediated cytotoxicity against PA-l by various concentrations of CHX.

rTNF AND CHEMOTHERAPY

Q

150

a e 7 8 120 ‘;; !

SYNERGY IN OVARIAN

O-O

Adriomycin

n

Cycloheximide

-W

TNF(l Ong/ml): $

337

CANCER

4% cytotoxicity

30T

B 0 xg

20.1 -0

/ ‘I!

07

10°

lo2

10’

n-w 10 -1

I 10-l

/

.--y/o

bg/m~)

I loo

10’

102

Cycloheximide(pg/ml)

FIG. 7. (a) Protein synthesis inhibition of OVCAR-3 by various concentrations of ADR and CHX (IO-’ to IO2 pg/ml). (b) Augmentation of rTNF-mediated cytotoxicity against OVCAR-3 by various concentrations of CHX.

against PA-I and 222 is presented in Figs. I and 2. As Figs. 3a and b. All concentrations of rTNF used (IO-’ to 10’ rig/ml) are noncytotoxic for OVCAR-3 targets. shown in Fig. lb, rTNF concentrations of lo-* and 10-l rig/ml are by themselves noncytotoxic to PA-l targets; However, when used in combination with 10 or 100 however, when they are used in combination with 10 pug/ml Adriamycin, significant synergy was observed. We also investigated whether lower doses of Adriapg/ml Adriamycin, significant synergy is produced. For 222, synergy was seen at all rTNF concentrations used mycin produced synergistic cytotoxicity when used in the same concentrations of rTNF. No significant synin combination with 1 pg/ml Adriamycin (Fig. 2b). We investigated whether the combination of various ergistic cytotoxicity was obtained when 10-l or 10’ pg/ml of Adriamycin and any dose of rTNF were comconcentrations of rTNF with a low dose of Adriamycin (lo-* pg/ml) that is not cytotoxic by itself would lead bined (data not shown). to a synergistic effect. Significant synergy was seen at all concentration of rTNF used for both PA-l and 222 Effect of rTNF and Adriamycin on SKOV-3 and OVCAR-8 (Resistant to Both Agents) targets (Figs. la and 2a). These results demonstrate that noncytotoxic high doses of Adriamycin synergize the All concentrations of rTNF and Adriamycin used alone cytotoxic activity of rTNF (10” and 10’ rig/ml) and that were not cytotoxic for SKOV-3 and OVCAR-8 targets. significant cytotoxicity was obtained when noncytotoxic, We then investigated whether the combination of various low doses of rTNF (~10~’ rig/ml) and Adriamycin concentrations of rTNF with a constant dose of Adria(