Sensitivity of Human Renal Cell Carcinoma Lines to TNF, Adriamycin, and Combination: Role of TNF mRNA Induction in Overcoming Resistance

0022-5347/93/1495-1202$03.00/0 THE JOURNAL OF UROLOGY Copyright © 1993 by AMERICAN UROLOGICAL ASSOCIATION, INC.

Vol. 149, 1202-1208, May 1993

Printed in U.S.A.

SENSITIVITY OF HUMAN RENAL CELL CARCINOMA LINES TO TNF, ADRIAMYCIN, AND COMBINATION: ROLE OF TNF mRNA INDUCTION IN OVERCOMING RESISTANCE JEFFREY TAYLOR SAFRIT, ARIE BELLDEGRUN, AND BENJAMIN BONAVIDA* From the Department of Microbiology and Immunology and Division of Urology, UCLA School of Medicine, and Jonnson Comprehensive Cancer Center, University of California at Los Angeles, Los Angeles, California

ABSTRACT

vv e have examined 6 human renal cell carcinorna (Ree) cell lines for their sensitivity and resistance to the cytolytic effect of tumor necrosis factor (TNF) and adriamycin (ADR), alone or in combination. The results of cytotoxicity mediated by TNF and ADR showed no direct correlation as TNF resistant lines were sensitive to ADR while the TNF sensitive lines were resistant. The combination of TNF and ADR resulted in enhanced cytotoxicity against the tumor lines. Induction of TNF mRN A and protein has been suggested as a mechanism of resistance to TNF in certain tumors. Resistant and sensitive lines were capable of upregulating TNF mRNA after treatment with TNF or PMA+ionophore for periods as short as 1 hour, but only the resistant lines were able to secrete detectable levels of TNF protein. Therefore, a positive correlation existed between resistance to TNF and production and secretion of TNF by the cell lines. In the presence of the protein synthesis inhibitor cycloheximide (CHX), the TNF mRNA level in the TNF resistant lines was increased while the sensitive lines required an additional signal, such as exogenous TNF, to upregulate the mRNA. Due to the enhanced cytotoxicity seen with the combination of TNF and ADR, we determined the effect of this combination on the levels of TNF mRNA. As examined in a constitutively TNF expressing line, ADR alone reduced the constitutive mRNA level and, in combination with TNF, reduced the level of induction produced by TNF. This downregulation of TNF mRNA by ADR may playa role in the enhanced cytotoxicity seen with combined TNF and ADR treatment. The present study demonstrates that RCC cell lines differ in their sensitivity and/ or resistance to TNF. Further, ADR and/or TNF resistant RCC lines can be rendered sensitive by combining TNF and ADR. The constitutive and/or inductive secretion of TNF by certain lines and its relationship to tumor pathogenesis as well as overcoming resistance are discussed. KEY WORDS:

carcinoma, renal cell; doxorubicin; tumor necrosis factor; RNA, messenger

Renal cell carcinoma (RCC) has long been recognized as one of the least responsive tumors to traditional therapy. A recent review of cytotoxic chemotherapy in patients with RCC listed numerous trials with 36 different cytotoxic agents that provided an average of less than 5 to 10% antitumor activity. 1 One possible reason for this high degree of resistance to chemotherapy is the overexpression of the multidrug resistance (MDR) phenotype in many renal tumors.2 In fact, normal kidney tissue has a substantial degree of expression of the MDR-1 gene. 3 Although the presence of the MDR phenotype may account for the majority of the resistance mechanisms in RCC, it does not appear to be solely responsible. Mickisch et al. found that 28% of the highly resistant renal tumors did not appear to express the MDR-1 gene. 4 For these reasons, other therapies including biological response modifiers (BRM) have been tested. A number of BRM's, including IFN-a and IL-2 are among the few single agents that have shown any activity against RCC. This includes the use of IL-2 for therapy with lymphokine activated killer cells (LAK), and for the expansion and activation of tumor-infiltrating lymphocytes (TIL).5.6 One such agent now in clinical trials against RCC and other tumors is tumor necrosis factor-alpha (TNF).7-11 Tumor necrosis factor is a cytokine with antitumor activity originally identified in the serum of mice challenged with endotoxin after BCG inoculation. 12 It has been shown to have Accepted for publication November 17, 1992.

* Requests for reprints: Department of Microbiology and Immunology, UCLA School of Medicine, Center for the Health Sciences, 10833 Le Conte Ave., Los Angeles, California 90024. This work was supported in part by a gift from the Borion Research Foundation and in part by a grant from the Concern Foundation.

numerous effects on both normaP3-18 and tumor cells. 12 , 18-20 Tumor necrosis factor's antitumor effects, while not well understood, have been linked to the generation of free radicals 21 and the induction of programmed cell death 22 in the tumor cell. Although TNF would seem to be a promising anticancer agent, many tumor lines are resistant to its cytotoxic effects. This resistance may be due to the overexpression of certain oncogenes,23 the induction of free radical scavengers such as manganese superoxide dismutase (MnSOD),24 or the expression of TNF itself by the tumor cells. 25 -27 Previously we and others have shown that certain MDR-1-expressing drug-resistant tumor lines are more sensitive to TNF than their MDR-1 negative, drug-sensitive parental lines (unpublished data),28 For these and other reasons, it is obvious that a complex relationship exists between the tumor cell and its sensitivity and/or resistance to drugs and TNF. This relationship was recently examined using a large group of histologically diverse tumor lines where a hierarchy of sensitivity and resistance to TNF and other cytotoxic modalities was revealed,29 Among other things, this hierarchy suggests the possibility of common mechanisms of sensitivity and resistance to many cytotoxic modalities, including TNF. The present study was designed to elucidate the cytotoxic effect of TNF and ADR, used alone or in combination, as well as the possible underlying mechanisms of reversing resistance. Specifically, the present study addresses four questions concerning the renal lines' sensitivity or resistance to drugs and TNF: 1) Is there a correlation between the sensitivity and/or resistance of the lines to TNF and adriamycin (ADR)? 2) Is there a correlation between the sensitivity and/or resistance to

1202

OVERCOMING R.ESISTANCE OF RENAL CARCINOlIJiA EY TNF AND ADRIAMYCIN

TNF and the TNF' mRNA and TNF protein levels in these lines? 3) Is there a correlation between the sensitivity and/or resistance to TNF and the expression of the MDR phenotype? 4) Can combination treatment of the renal lines with TNF + ADR alter the TNF or MDR mRNA expression by these lines? These questions were investigated with 6 human RCC lines of different sensitivities to TNF and ADR. MATERIALS AND METHODS

Cell Lines and Media. The renal carcinoma cell lines R4, R4a, R6, and Rll were supplied by Dr. Hans Stotter, Bethesda, Maryland; RCC 30 was supplied by Dr. Kyogo Itoh, Houston, Texas. The above lines were obtained from nephrectomized RCC specimens. The renal adenocarcinoma line 444 was derived from malignant ascites of a patient with metastatic RCC (Dr. Arie Belldegrun). These tumor lines were all confirmed to be RCC by histopathologic examination. 30 The ovarian tumor lines OVC-8 and ADI0 were supplied by Dr. Robert Ozols. These lines, as well as the B-cellline Raji and the pro myeloid U937, were maintained in RPMI 1640 supplemented with nonessential amino acids, glutamine, antibiotics and 10% fetal bovine serum (Gibco/BRL, Gaithersburg, Maryland). All lines were grown in a humidified atmosphere at 37C in 5% CO 2, Adherent cells were harvested by brief treatment with a 0.25% trypsin, 5 X 10- 4 M. EDTA solution and washed in complete medium prior to use. Reagents. The recombinant TNF-a was generously supplied by Smith-Kline-French, Philadelphia, Pennsylvania. Antiserum directed against rTNF was raised in rabbits by intramuscular injection of 50 f.lg. rTNF in complete Freund's adjuvant. The rabbits were boosted 3 weeks later, and after 7 days were bled by venous puncture to test the serum for neutralizing activity. The concentration was then adjusted for the TNF-a Elisa assay. The rabbits' care was in accordance with UCLA guidelines. Alkaline phosphatase linked goat-anti-rabbit IgG was purchased from Caltag (San Francisco, California). Adriamycin was purchased fwm Sigma Chemical Co., St. Louis, Missouri. The TNF cDNA used in making probes for Northern blot analysis was a gift from Smith-Kline-French. The MDR1 cDNA was a gift from Carsten Lincke at The Netherlands Cancer Institute. Northern Blot Analysis. Cytoplasmic RNA from the tumor lines was prepared following the method of Chomczynski and SacchpI: 40 f.lg. per lane of tumor cell RNA was electrophoresed in 1.2% agarose-2.2 M. HCHO gels in IX MOPS buffer. 32 The RNA was transferred to Zeta Probe nylon membranes (Biorad, Richmond, California) in 20X SSC (IX SSC = 150 mM. NaCl, 15 mM. sodium citrate pH 7.0), and 50 to 100 ng. of eDNA probe was labeled with [32 P]dCTP (NEN) by random oligoprimer extension as described by Feinberg and Volgelstein. 33 The nylon filters were UV cross-linked and then prehybridized at 42C from 1 hour to overnight in 50% Formamide (BRL), 5x Denhardts, 0.1% SDS, 100 f.lg./ml. salmon sperm DNA (ssDNA), and 5x SSC. Radiolabeled probe was added at 1 X 106 cpm/m!. of hybridization fluid (6x SSC, 0.5% SDS, 5x Denhardts, 100 f.lg./ml. ssDNA) and the blot was incubated overnight at 65C. Hybridized filters were then washed 2 times for 15 minutes each in 2X SSC/O.I % SDS at room temperature and 2 times for 5 minutes each in O.lX SSC/O.l % SDS at 65C and exposed to KODAK XAR-5 X-ray film. Cytotoxic Assays. The chromium-51-release assay was used to determine the extent of tumor cell lysis perpetrated by rTNF or Adriamycin. Briefly, tumor cells at 1 to 3 X 106 cells per m!. were incubated with 0.1 ml. Na 25I Cr04 (Amersham, Arlington Heights, Illinois 1 mCi./ml.) in a total volume ofl ml. of RPMI with 20% FCS at 37C for 1 to 3 hours. Cells were then washed three times and resuspended at 1 X 105 cells per ml. in culture media. Then 0.1 ml. of 5ICr-labeled tumor cells at 1 X 105 cells per ml. plus 0.1 ml. cytokines, drugs, or toxins was added to 96 well microtiter plates (Costar, #3787, Cambridge, Massachu-

1203

setts) and incubated at 37C for 18 to 22 hours. The assays were harvested by centrifuging the microtiter plates at 200 g. for 5 minutes after which 100 ,ul. of cell-free supernatant was removed and counted for 5ICr content using a Beckman 5500 gamma counter. Triplicate samples were used and mean percent cytotoxicity was estimated using the following formula: %specific release = (cpm experimental release - cpm spontaneous release)/(cpm total - cpm spontaneous release) X 100. Spontaneous release did not exceed 25% of total. Calculations for synergy were done using isobologram analysis as described by Berenbaum. 30 TNF ELISA. Monoclonal antibodies (mAB) for TNF used in this assay were the generous gift of Dr. G. Trinchieri (Wistar Institute, Philadelphia, Pennsylvania).34 The murine IgG 1 mAB, B154.7.1 and B154.9.l were each specific for a different epitope of the TNF molecule. B154.7.1 derived from ascites fluid was purified by affinity chromatography using an antibody directed against B154.7.1. B154.9.1 from ascites was partially purified by 50% ammonium sulfate precipitation. Wells of ELISA plates were coated with 50 f.lL of both antibodies for at least 1 day and were stored as long as 4 weeks at 4C. To set up the assay, coated plates were washed three times and blocked with Elisa PBS containing 1 % BSA for 1 hour. Plates were washed twice and 50 f.ll. of both supernatants and standards were added to the wells. After overnight incubation, plates were washed three times and 50 f.lL of polyclonal anti -TNF antibody was added to each well. After a 2-hour incubation, alkaline phosphatase conjugated goat anti-rabbit (CALTAG LAD) IgG was added to each well and incubation was continued for another 2 hours. Finally, plates were washed and incubated with the substrate (Sigma 104). Plates were read 2 hours at the 405 mm. filter by Titrated Multiskan MCC/240 Elisa reader. TNF binding assay. [125I]-labeled rTNF ([1 25 I]-rTNF) (specific activity approximately 1.5 X 108 cpm/f.lg.) was prepared using Iodobeads (Pierce), following the procedure recommended by the manufacturer. Renal tumor cells (1 X 105/ml.) were incubated in microfuge tubes in 0.5 mL complete media containing 1 nM. I25I_rTNF with or without a 100-fold excess of unlabeled rTNF at 4C for 2 hours. The cells were then washed three times with ice cold medium, and the radioactivity remaining in the pellet was determined. The B-cell line Raji was used as a negative control and the pro myeloid line U937 as a positive control for rTNF binding. Flow cytometry. For flow cytometry 100 f.ll. of tumor cells at 2 X 106 cells per m!. were incubated with ADR for 10 minutes at 37C, diluted with ice cold Ix PBS and immediately analyzed. After analysis the cells were washed twice to remove excess ADR and resuspended in fresh media for 1 hour at 37C before analysis. The above experiments were performed either in the presence or absence of 5 flg./ml. verapamil to block efflux, or 10 ng./ml. TNF to determine the effect of TNF on influx/ efflux. Cells were analyzed on an EPICS C 2981 flow cytometer (Coulter, Hialeah, Florida). RESULTS

Sensitivity of several renal carcinoma cell lines to the cytotoxic effects of rTNF and adriamycin. Previous studies have reported both the existence 35 ,36 and the absence 37 of a correlation between the sensitivity of tumor lines to TNF and drugs. Because of this controversy the present study examined the relationship between the sensitivity of RCC lines to TNF and ADR. The sensitivity of several established human renal carcinoma cell lines to TNF and ADR was examined by the 5ICr-release assay. As can be seen in figure 1, A, of the six lines tested, four were resistant to rTNF «10% lysis) at high doses, while two of the lines were sensitive at moderate doses. For comparison, the TNF sensitive U937 promyeloid cell line is shown. Resistance to TNF may sometimes be attributed to lack of TNF receptor expression. Accordingly, one each of the sensitive and resistant lines, Rll and R4, were tested for the expression of TNF

1204

SAFRIT, BELLDEGRUN AND BONAVIDA

receptors on the cell surface. Both lines expressed the TNF receptors as examined with 125I-Iabeled rTNF in a competitive binding assay and were compared to the extremely low expressing B-cell line Raji and the positive control U937 (fig. 1, B). Calculating the number of receptors per cell using the specific activity of the radiolabeled TNF resulted in approximately 9,700 receptors per R4 cell, 12,000 receptors per Rll cell, and 15,600 receptors per U937 cell (data not shown). Thus the resistance to TNF is not attributed to a lack of TNF receptor expression. We also tested the lines for sensitivity or resistance to the chemotherapeutic drug adriamycin. Interestingly, the two lines sensitive to TNF were very resistant to normally cytotoxic doses of the drug. The TNF -resistant lines, on the other hand, were sensitive to adriamycin at low concentrations (fig. 1, C). Because

T~~F

used in combination \vith P;o..DR has resulted in

enhanced cytotoxicity against various tumor lines,38-41 we examined the effect of this combination on the RCC lines. The combination of TNF and adriamycin resulted in enhanced cytotoxicity in all the lines tested. Figure 2, A-D shows the results of a representative experiment using TNF sensitive (fig. 2, A and B) and resistant (fig. 2, C and D) RCC lines. These results demonstrate that the RCC lines are heterogeneous with regard to their sensitivity to TNF and ADR and that there is no indication of a correlation between TNF and ADR sensitivity and resistance. Expression of TNF mRNA. Because the production of TNF mRNA and/or protein by tumor cells themselves has been 30

b

:g

40 ng/ml 10 ng/ml 1 ng/ml

. . TNF

= =

25

A

B 444

ADR (!-lg./rnll 0.01

0.1

]0

OJ

;-

t

5i t'2

o 80

0,]

0::

0

0,)

<{

70 60 ]0

0.1

50

(1:2

f'raClional inhibi]ory concel1lra)ion ofTNF

C R6

° ]00

00]

0]

]

]0

"~"I(Cg) " 0

D--H~

"f!·E~

.---"==-~=:

90 A.--II

80

001

0 I

t'2 70

60

10

FIG. 2. Cytotoxicity of combination of TNF and ADR vs. RCC line R6 (A) and RCC line 444 (B,C) measured by chromium-51-release as described in Materials and Methods. Cytotoxicity is expressed as percent survival (A,B). Isobologram analysis 33 was then generated from these data (C). Isobologram analysis could not be performed on R6 because of lack of cytotoxicity by ADR.

20

x

0

~

15

*

10

u

R6

R4

b)

R l'

7~,-~,-~,---,-

RC30 444

U937

-~--,-------,

I

o

E Il U

Raji

R11

R4

U937

c)

~

40.-----------------------------, . . ADR 10 ug ml 35 1 ug/ml 30 01 ug/enl

=

=

~ ::!h''. I~

10

~L

_~~I~~-"L----J R4

R4'"'

R6

R 11

RUO 444

U937

FIG. 1. A, cytotoxicity of rTNF - vs. human RCC lines measured by [51 CrJ-release as described in Materials and Methods. U937 cells were used as a TNF sensitive control. B, ['25 IJ-rTNF binding assay to measure the presence of TNF receptor on RCC lines. Cells were incubated with [125IJ-rTNF for 2 hours. Presence of lOO-fold excess unlabeled TNF to determine specific vs. background binding. Results are shown with background cpm subtracted. U937 and Raji cells were used as positive and negative controls respectively. C, cytotoxicity of ADR vs. human RCC lines.

linked to TNF resistance,26 the RCC lines were examined for the presence of TNF mRNA by Northern blot analysis. In a representative experiment, seen in figure 3, A, only TNF resistant R4 cells constitutively expressed detectable levels of TNF mRNA. The remaining lines, however, can be induced to express varying amounts of the mRNA following treatment with 10 ng./ml. rTNF for 1 to 3 hours. Although this induction of TNF mRNA is seen in all the lines, the TNF sensitive R6 and Rll express much lower levels of mRNA than do their TNF resistant counterparts. The kinetics of the induction by TNF treatment is shown in figure 3, B. Using R4 as an example, induction of TNF mRNA occurs within 1 to 3 hours after the beginning of treatment, with most of the increase in expression occurring in the first 2 hours, This same pattern is observed with all RCC lines tested. Although there is an increase in expression after treatment with TNF from 4 to 6 hours, the level is only slightly greater than the constitutive amount. This same pattern of induction of TNF message by treatment of RCC with TNF is seen if the cells are activated with PMA and ionophore. Namely, the resistant lines are stimulated to produce TNF mRNA while the sensitive lines produce little if any (fig. 3, A). These results demonstrate that treatment of TNFresistant RCC cells with TNF results in induction of TNF mRNA. Further, the induction of TNF mRNA is rapid, thus possibly counteracting the early part of TNF-mediated sensitization for cytotoxicity. Effect of cycloheximide on induction of TNF mRNA leuels. Previous studies have shown that TNF biosynthesis is regulated at both the transcriptional and post-transcriptional levels (for a review, see reference 42). This regulation centers around the presence of a (U + A) exclusive element in the 3' -untranslated region in TNF and other cytokines. 43 This element appears capable of directing mRNA degradation as well as suppressing translation. 44 ,45 Thus, the protein synthesis inhibitor cycloheximide has been shown to superinduce the level of TNF mRNA in tumor lines by supposedly inhibiting the production

OVERCOMING RESISTANCE OF RENAL CARCINOMA BY TNF AND ADRIAMYCIN

1205

FIG. 3. A, induction of TNF mRNA by TNF or PMAjionophore analyzed by Northern blot as described in Materials and Methods. RCC lines were incubated with TNF (10 ng.jml.) or PMA (1 mg.jml.) + ionophore (200 nM.) for 2 hours at 37C. Total RNA was then isolated and equal amounts loaded and examined for TNF mRNA as described in Materials and Methods. B, kinetics of TNF mRNA induction by TNF. RCC line R4 was incubated with 10 ng.jml. TNF ± ADR for indicated times. Blot was stripped and reprobed with ~-actin as a control for equal loading in both A and B (data not shown). C, densitometry scan of bands in (B) shows relative levels of mRNA compared to control.

of a labile protein. 27 We therefore tested the renal lines for superinduction of TNF message by cycloheximide. Treatment of the cells with 10 ,ug./ml. cycloheximide for 3 hours induced message in the 4 TNF resistant lines but not in the sensitive lines (fig. 4). These results suggest that a certain amount of TNF message preexists in the resistant lines and is allowed to accumulate perhaps because cycloheximide inhibits the synthesis of a labile protein, as previously noted. 27 If the cells are exposed to the combination of TNF and cycloheximide, all 6 lines superinduce TNF mRNA levels, suggesting that the TNF sensitive lines need to be triggered before any superinduction can occur by cycloheximide. Correlation between the induction of TNF mRNA and secretion of TNF. The above results show that TNF mRNA can be induced in the RCC lines by TNF /PMA + Ionophore. To study the stability and function of the TNF mRNA we examined its translation in the RCC lines. Previous studies have shown that the existence of TNF mRNA does not necessarily predict whether or not TNF protein will be synthesized and secreted. 46 ,47 Supernatants of either control or stimulated cells were therefore examined for the presence of TNF protein. As measured by ELISA and compared with the high secreting ovarian line OVC-B, TNF protein was produced constitutively only by the TNF-resistant R4, which also constitutively expresses the TNF mRNA. Stimulation of the RCC lines with either TNF or PMA and ionophore leads to varying levels of secretion in all the resistant lines (table). Even though the sensitive lines can be induced to upregulate their TNF mRNA, they do not appear to produce any protein whatsoever. These results demonstrate that TNF resistant RCC lines can be induced to translate TNF mRNA and secrete the protein whereas the sensitive lines do not secrete detectable levels of TNF. Expression of MDR-l mRNA. Many renal tumors have been previously shown to overexpress the MDR gene.2, 3, 48 Because the adriamycin resistance pattern of the RCC lines examined is opposite to that seen with TNF, we examined total RNA of each line for the presence ofMDR-1 mRNA. As shown in figure

5, none of the RCC lines appear to express the MDR-1 gene in any detectable amounts, compared with the adriamycin resistant, ovarian line ADIO. Further analysis of drug uptake and efflux of the cells by flow cytometry substantiates the previous results. As shown in figure 6 (column 1, row C), only the MDR mRNA positive ADIO pumps ADR out of the cells. The action of the MDR pump is blocked by the addition of verapamil as shown in row D. The ADR resistant lines, R6 and Rll, exhibited no differences in uptake or efflux of the drug. Since 4 of the 6 lines are adriamycin sensitive, the lack of MDR-l mRNA in these lines was not unexpected. The lack of MDR-1 mRNA in the ADR-resistant lines demonstrates that there is no correlation between the expression of the MDR phenotype and sensitivity or resistance of RCC to TNF. Effect of TNF ± ADR on TNF mRNA levels. The combination of TNF and ADR has previously been shown to cause enhanced cytotoxicity in other tumor lines 38-41 as well as in the RCC lines examined here. Because the induction of TNF may be related to tumor cell TNF resistance (see above results), it was postulated that the effect of the combination of TNF + ADR would downregulate the induction of TNF mRNA. As shown in figure 3, Band C, not only was the expression of constitutive TNF mRNA reduced by ADR alone, but the combination of TNF and ADR reduced the mRNA level induced by TNF itself. These results suggest that the enhancement in cytotoxicity seen with the combination of TNF and ADR may be related to the reduction in protective TNF mRNA levels. DISCUSSION

The present study examines several properties of R' jC sensitivity and resistance to TNF and ADR. Although no direct correlation was found between the sensitivities of the lines to the two cytotoxic agents, enhanced cytotoxicity was seen when TNF and ADR were combined. To elucidate the mechanisms behind this enhanced cytotoxicity, we examined the cell lines for previously suggested mechanisms of resistance to TNF and ADR. Tumor necrosis factor mRNA was induced in each line by treatment with TNF, whereas only the resistant lines were

1206

SAFRIT, BELLDEGRUN AND BONAVIDA

FIG. 5. Analysis of MDR-l mRNA. RCC lines were incubated with or without cycloheximide to stabilize mRNA for 2 hours. Total RNA was then isolated and probed for MDR-l mRNA as described in Materials and Methods. f3-actin was used as a control for equal loading.

ADIO

Cell Line

IIll

I

it

l

I

131~........... .

1

Rll

I

~" ~

121

141

R6

,

~ ~ .

RCC30

~.

i"

~.

_.A"

~

1_ 1.

FIG. 6. Analysis of ADR uptake and efflux by flow cytometry as described in Materials and Methods. Row 1) Control cells background fluorescence. Row 2) Cells + 1 ILg./ml. ADR for 10 minutes at 37C. Row 3) Cells from row 2 washed twice and incubated at 37C for 1 hour. Row 4) Same treatment as in row 3 + 5 ILg./ml. verapamil in all stages. FIG. 4. Effect of cycloheximide on TNF induction. RCC cells were incubated with TNF (10 ng./ml.) ± cycloheximide (10 ILg./ml.) for 2 hours.

TNF secretion by renallines* Cell Line R4 R4a R6

Rll RC30 444 OVC-S***

Medium 30.7 ± o± 0.5 ± 4.5 ± O± o± 57.5 ±

12.lt 5.5 6.2 2.2 7.2 4.5 14.1

PMAjIonophore** S17.2 ± 55.3 ± 1.5 ± 3.2 ± 197.7 ± 65.S ± 1305 ±

50.3 10.1 4.7 2.4 42.3 1S.2 70.5

t pg.jml. ± standard deviation. * Secretion measured by Elisa as described in materials and methods. ** PMA (1 ng.jml.) + A231S7 (200nM.). *** OVC-S included as a positive control. capable of secreting the protein. To probe the regulatory mechanisms involved in the induction of TNF message, the protein synthesis inhibitor cycloheximide was used. Cycloheximide alone induced TNF message in the TNF resistant lines, while the TNF sensitive lines required the addition of exogenous TNF for upregulation of the mRNA. Because ADR + TNF results in enhanced cytotoxicity, we examined the effect of ADR on the induction of TNF message by TNF. Adriamycin downregulated the level of TNF mRNA induced by TNF, thereby suggesting that a reduction in the protective effect attributed to the induction of TNF mRNA in TNF resistant lines may lead to the enhanced cytotoxicity. As a baseline from which to address other questions, the sensitivity of the RCC lines to cytotoxicity caused by TNF in vitro was examined. The RCC cell lines, as a group, proved to have a heterogeneous response to TNF -mediated cytotoxicity, as also reported in previous studies. 49 . 50 Using a small sample

of 6 human renal tumor lines, 2 were sensitive and 4 were resistant to TNF -mediated cytotoxic effects. These findings corroborate those of Heicappell et aI., who found that 5 RCC lines established from a single surgical specimen had differing degrees of sensitivity and resistance to TNF.49 Tumor necrosis factor is emerging as an important cytokine in potential combination therapies directed against human renal cell carcinoma. It has been used with promising results in combination with other BRM's such as IFN-a 50 and with anticancer drugs such as etoposide 51 . 52 against RCC. Because of earlier discrepancies concerning the existence of a correlation between the sensitivity of tumor lines to TNF and certain chemotherapeutic drugs,35-37 we tested the sensitivity of the RCC lines to the chemotherapeutic drug adriamycin. Adriamycin lysed each of the TNF-resistant lines but was unable to kill their TNF -sensitive counterparts, resulting in an inverse relationship of sensitivity and resistance to TNF and ADR. These findings cannot be generalized to the tumor system studied, as various TNF sensitive tumor lines are also drug sensitive (unpublished observation). It was recently reported that some tumor lines that have acquired the multidrug resistance phenotype by coculture in the presence of cytotoxic drugs are more sensitive to TNF than their drug sensitive parental lines. 27 Because of the similar sensitivity and/or resistance pattern seen here with the RCC lines and because renal tumors have previously been noted for high expression of the MDR-l gene, we examined these cells for the presence of MDR-l mRNA. Surprisingly, although the number of lines examined is limited, we were unable to detect any MDR-l transcript in either the TNF sensitive-drug resistant or the TNF resistant-drug sensitive lines. Likewise, the renal cells could not pump out ADR as compared with the MDR mRNA positive ovarian carcinoma line AD10. Thus the

OVERCOMING RESISTANCE OF RENAL CARCINOMA BY TNF AND ADRIAMYCIN

drug resistance of the TNF sensitive renal lines does not appear to be related to the multidrug resistant phenotype. Further experiments are planned to expand on these findings and to elucidate the scope and mechanism of this resistance. The combination of TNF and ADR was tested against the RCC lines to determine whether the cytotoxicity produced by either agent alone could be enhanced. We have obtained enhanced cytotoxicity in a limited number of experiments. In other instances additive cytotoxicity was obtained. In correlation, previous results from our laboratory41 and others38.40-53 have shown synergistic cytotoxicity with this combination of antitumor agents against other human and mouse tumor cell lines. For example, Bonavida et al. demonstrated synergistic cytotoxicity with the combination of TNF and ADR (as well as cisplatin) in vitro against human ovarian tumor celllines. 41 Mutch et al. achieved enhanced cytotoxicity also against human ovarian tumor lines with the combination of TNF and cisplatin. 53 Regenass and coworkers showed similar effects of this combination in an in vivo murine system versus intradermally implanted Meth A sarcoma. 39 Because the mechanisms of the antitumor effects caused by TNF are still debated, we examined the renal lines for the presence of previously suggested attributes of resistance to TNF and other anticancer agents. 23-27 Among these is the production of TNF mRNA and/or TNF protein by the resistant tumor cell. 25-27 Our findings with the renal lines demonstrate that the 4 tumor lines resistant to the cytotoxic effects of TNF can produce and secrete the TNF protein while the 2 sensitive lines do not. These results follow the example noted by Spriggs et al. 26,27 with various tumor lines including breast, ovarian, and lung. Stimulation of the tumor cells' protein kinase mechanisms by PMA/ionophore had a similar effect in that only the resistant lines upregulated their message and secreted detectable protein. To determine whether the regulation of the TNF message in RCC is similar to that seen in other tumor lines, we examined the mRNA for TNF in the presence of the protein synthesis inhibitor cycloheximide. The resistant lines all upregulated their mRNA for TNF after brief exposure to CHX. These results are in keeping with those seen by Spriggs et aI., where cycloheximide probably inhibits the synthesis of an RNAse that would normally degrade the message after a short time, thus suggesting negative regulation by this labile protein. 27 Whether this increase in RCC mRNA for TNF is due to stabilization of the message or enhancement of transcription has yet to be determined in our system. The RCC lines sensitive to TNF were not superinduced by cycloheximide alone; they needed a stimulation signal as well (such as, exogenous TNF) to produce any TNF mRNA. This suggests that TNF message may be differentially regulated in sensitive versus resistant RCC tumor cell lines. Because TNF and other cytokines are regulated at the transcriptional and post-transcriptional levels,42 another possibility is that the induction seen with TNF treatment occurs at the transcriptional level, whereas the cycloheximide induction is post-transcriptional. This would explain in part the inability of CHX alone to up regulate TNF mRNA in TNF-sensitive lines. To further examine possible sites of regulation of the TNF gene in RCC cell lines and to clarify the mechanisms resulting in enhanced cytotoxicity by combination treatment, we determined the effect of TNF + ADR on TNF mRNA levels in these cells. Simultaneous treatment of the RCC line R4 with TNF and ADR reduced both constitutive expression and TNF induced expression of TNF mRN A. This result, while not conclusive' suggests that enhanced cytotoxicity could possibly result from the reduction of the protective effect of TNF mRNA in this line. Further experiments on TNF sensitive and resistant lines, including measurement of the kinetics of this reduction as well as the effect of cycloheximide on the results seen with the combination ofTNF and ADR, are needed to support these

1207

findings. While the role of TNF mRNA induction and its downregulation by ADR is suggestive for enhanced sensitivity, other mechanisms of resistance, such as the induction of MnSOD and the role of topoisomerase, are also operative. In conclusion, we have tested several RCC lines for their in vitro response to TNF and ADR in the hope of elucidating possible resistance mechanisms and the means to overcome them. We have shown that ADR downregulates TNF mRNA induced in these lines, possibly leading to the enhanced cytotoxicity seen with these agents in combination. Future directions include the examination of the specificity of this TNF mRNA downregulation by ADR by determining ADR's effect on other TNF inducible resistance mechanisms such as MnSOD. We are also interested in the correlation between secretion of TNF by RCC tumor lines and the degree of resistance attained in vivo. Because RCC seems to be a good candidate for BRM type therapies,6 more study in this area is essential for the development of an effective therapeutic regimen. Acknowledgments. The authors acknowledge the technical assistance of Jose Trevejo and the secretarial assistance of Cammy Wang. REFERENCES

1. Yagoda, A. and Bander, N. H.: Failure of cytotoxic chemotherapy, 1983-1988, and the emerging role of monoclonal antibodies for renal cancer. Urol. Int., 44: 338, 1989. 2. Fojo, A. T., Shen, D. W., Mickley, L. A., Pastan, I. and Gottesman, M. M.: Intrinsic drug resistance in human kidney cancer is associated with expression of a human multidrug-resistance gene. J. Clin. Oncol., 5: 1922, 1987. 3. Fojo, A. T., Ueda, K, Siamon, D., Poplack, D. G., Gottesman, M. M. and Pastan, 1.: Expression of a multidrug-resistance gene in human tumors and tissues. Proc. Nat. Acad. Sci. USA, 84: 265, 1987. 4. Mickisch, G. H., Kossig, J., Keilhauer, G., Schlick, E., Tschada, R K and Aiken, P. M.: Effects of calcium antagonists in multidrug resistant primary human renal cell carcinomas. Cancer Res., 50: 3670,1990. 5. Quesada, J. R: Biological response modifiers in the therapy of metastatic renal cell carcinoma. Semin. Oncol., 15: 396, 1988. 6. Graham, S. D.: Immunotherapy of renal cell carcinoma. Semin. Urol., 7: 215, 1989. 7. Sherman, M. L., Spriggs, D. R, Arthur, K A., Imamura, K., Frei, E. and Kufe, D. W.: Recombinant human tumor necrosis factor administered as a five days infusion in cancer patients: phase I toxicity and effects on lipid metabolism. J. Clin. Oncol. 6: 344, 1988. 8. Blick, M., Sherwin, S. A., Rosenblum, M. and Gutterman, J.: Phase I study of recombinant tumor necrosis factor alpha in cancer patients. Cancer Res., 47: 2986, 1987. 9. Bartsch, H. H., Nagel, G. A., MuIi, R, Flener, Rand Pfizenmaier, K: Phase I study of recombinant human tumor necrosis factor alpha in patients with advanced malignancies. Mol. Biother. 1: 21,1988. 10. Bartsch, H. H., Pfizenmaier, K., Schroeder, M. and Nagel, G. A.: Intralesional application of recombinant human tumor necrosis factor alpha induces tumor regression in patients with advanced malignancies. Eur. J. Cancer Clin. Oncol. 25: 287, 1989. 11. Aulitzky, W. E., Tilg, H., Gunther, G., Mull, R, Flener, R, Vogel, W., Herold, M., Berger, M., Judmaier, G. and Huber, C.: Recombinant tumor necrosis factor alpha administered subcutaneously or intramuscularly for treatment of advanced malignant disease: a phase I trial. Eur. J. Cancer, 27: 462, 1991. 12. Carswell, E. A., Old, L. J., Kassel, R L., Green, S., Fiore, N. and Williamson, B.: An endotoxin-induced serum factor that causes necrosis in tumors. Proc. Nat. Acad. Sci. USA, 72: 3666, 1975. 13. Old, L. J.: Tumor necrosis factor (TNF). Science, 230: 630, 1985. 14. Scheurich, P., Thoma, B., Ucer, U. and Pfizenmaier, K: Immunoregulatory activity of recombinant human tumor necrosis factor (TNF)-a: induction of TNF receptors on human T cells and TNF -a mediated enhancement of T cell responses. J. Immunol., 140: 531, 1987. 15. Yokota, S., Geppert, T. D. and Lipsky, P. E.: Enhancement of

1208

16.

17.

18.

19. 20.

21.

22. 23.

24. 25.

26.

27. 28.

29.

30. 31. 32. 33. 34.

35.

SAFRIT, BELLDEGRUN AND BONAVIDA

antigen- and mitogen-induced human T lymphocyte proliferation by tumor necrosis factor-a. J. Immunol., 140: 531, 1988. Gamble, J. R., Harlan, J. M., Klebanoff, S. J. and Vadas, M. A.: Stimulation of the adherence of neutrophils to umbilical vein endothelium by human recombinant tumor necrosis factor. Proc. Nat. Acad. Sci. USA, 82: 8667, 1985. Shalaby, M. R., Palladino, M. A., Hirabayashi, S. E., Essalu, T. E., Lewis, G. D., Shepard, H. M. and Aggarwal, B. B.: Receptor binding and activation of polymorphonuclear neutrophils by tumor necrosis factor-alpha. J. Leukocyte BioI., 41: 196, 1987. Sugarman, B. J., Aggarwal, B. B., Hass, P. E., Figari, I. S., Palladino, M. A. and Shepard, H. M.: Recombinant human tumor necrosis factor-a: Effect on proliferation of normal and transformed cells in vitro. Science, 230: 943, 1985. Haranaka, K. and Satomi, N.: Cytotoxic activity of tumor necrosis factor (TNF) on human cells in vitro. Jpn. J. Exp. Med., 51: 191, 1981. Ruggiero, V., Lathaln, K. and Baglioni, C.: Cytostatic and cytotoxic activity of tumor necrosis factor on human cancer cells. J. Immunol., 138: 2711, 1987. Matthews, N., Neale, M. L., Jackson, S. K. and Stark, J. M.: Tumor cell killing by tumor necrosis factor: inhibition by anaerobic conditions, free radical scavengers, and inhibitors of arachitanate metabolism. Immunol., 62: 153, 1987. Schmid, D. S., Honung, R., McGrath, K M., Paul, N. and Ruddle, N. H.: Target cell DNA fragmentation is mediated by lymphotoxin and TNF. Lymphokine Res., 6: 195, 1987. Hudziak, R. M., Lewis, G. D., Refaat Shalaby, M., Essalu, T. E., Aggarwal, B. B., Ullrich, A. and Shepard, H. M.: Amplified expression of the HER2/ERBB2 oncogene induces resistance to tumor necrosis factor a in NIH 3T3 cells. Proc. Nat. Acad. Sci. USA, 85: 5102, 1988. Wong, G. H. W., Elwell, J. H., Oberly, L. W. and Goeddel, D. V.: Manganous superoxide dismutase is essential for cellular resistance to cytotoxicity of tumor necrosis factor. Cell, 58: 923, 1989. Rubin, B. Y., Anderson, S. L., Sullivan, S. A., Williamson, B. D., Carswell, E. A. and Old, L. J.: Nonhematopoietic cells selected for resistance to tumor necrosis factor produce tumor necrosis factor. J. Exp. Med., 164: 1350, 1986. Spriggs, D., Imamura, K, Rodriguez, C., Horiguchi, J. and Kufe, D. W.: Induction of tumor necrosis factor expression and resistance in a human breast tumor cell line. Proc. Nat. Acad. Sci. USA, 84: 6563, 1987. Spriggs, D. R., Imamura, K., Rodriguez, C., Sariban, E. and Kufe, D. W.: Tumor necrosis factor expression in human epithelial tumor cell lines. J. Clin. Invest., 81: 455, 1988. Salmon, S. E., Soehnlen, B., Dalton, W. S., Meltzer, P. and Scuderi, P.: Effects of tumor necrosis factor on sensitive and multidrug resistant human leukemia and myeloma cell lines. Blood, 74: 1723,1989. Safrit, J., Tsuchitani, T., Zighelboim, J. and Bonavida, B.: Hierarchy of sensitivity and resistance of tumor cells to cytotoxic effector cells, cytokines, drugs, and toxins. Cancer Immunol. Immunother., 34: 321, 1992. Berenbaum, M. C.: Criteria for analyzing interactions between biologically active agents. Adv. Cancer Res., 35: 269, 1981. Chomczynski, P. and Sacchi, N.: Single step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem., 162: 156, 1987. Goldberg, D. A.: Isolation and partial characterization of the Drosophila alcohol dehydrogenase gene. Proc. Nat. Acad. Sci. USA, 77: 5794, 1980. Feinberg, A. P. and Vogelstein, B.: A technique for radio labelling DNA restriction endonuclease fragments to high specific activity. Anal. Biochem., 132: 6,1983. Cuturi, M. C., Murphy, M., Costa-Giomi, M. P., Weinman, R., Perussia, B. and Trinchieri, G.: Independent regulation of tumor necrosis factor and lymphotoxin production by human peripheral blood lymphocytes. J. Exp. Med., 165: 1581, 1987. Dollbaum, C., Creasey, A. A., Dairkee, S. H., Hiller, A. J., Rudolph, A. R., Lin, L., Vitt, C. and Smith, H. S.: Specificity of tumor

36. 37. 38.

39. 40.

necrosis factor toxicity for human mammary carcinomas relative to normal mammary epithelium and correlation with response to doxorubicin. Proc. Nat. Acad. Sci. USA, 85: 4740, 1988. Fruehauf, J. P., Mimnaugh, E. G. and Sinha, B. K: Doxorubicininduced cross-resistance to tumor necrosis factor (TNF) related to differential TNF processing. J. Immunother., 10: 165, 1991. Neale, M. L. and Matthews, N: Development of tumor cell resistance to tumor necrosis factor does not confer resistance to cytotoxic drugs. Eur. J. Cancer Clin. Onc., 25: 133, 1988. Alexander, R. B., Nelson, W. G. and Coffey, D. S.: Synergistic enhancement by tumor necrosis factor of in vitro cytotoxicity from chemotherapeutic drugs targeted at DNA Topoisomerase II. Cancer Res., 47: 2403,1987. Regenass, U., Muller, M., Curschellas, E. and Matter, A.: Antitumor effect of tumor necrosis factor in combination with chemotherapeutic agents. Int. J. Cancer, 39: 266, 1987. Watanabe, N., Niitsu, Y., Yamauchi, N., Ohtsuka, Y., Sone, H., Neda, H., !,.,1acda, },1. and Urushizaki, L: Synergistic cytotoxicity

41.

42.

43.

44. 45.

46. 47.

48.

49.

50.

51.

52.

53.

of recombinant human TNF and various anti-cancer drugs. Immunopharm. Immunotox., 10: 117, 1988. Bonavida, B., Tsuchitani, T., Zighelboim, J. and Berek, J.: Synergy is documented in vitro with low-dose recombinant tumor necrosis factor, cisplatin, and doxorubicin in ovarian cancer cells. Gyn. Oncol., 38: 333, 1990. Beutler, B.: Regulation of Cachectin biosynthesis occurs at multiple levels. In: Cytokines and Lymphocortins in Inflammation and Differentiation. Edited by M. Melli and L. Parente. New York: Wiley-Liss, Inc. pp. 229-240, 1990. Caput, D., Beutler, B., Hartog, K, Brown-Shimer, S. and Cerami, A.: Identification of a common nucleotide sequence in the 3'untranslated region of mRNA molecules specifying inflammatory mediators. Proc. Nat. Acad. Sci. USA., 83: 1670, 1986. Shaw, G. and Kamen, R.: A conserved AU sequence from the 3' untranslated region of GM-CSF mRNA mediated selective mRNA degradation. Cell, 46: 659, 1986. Kruys, V., Wathlet, M., Poupart, P., Contreras, R., Fiers, W., Content, J. and Huez, G.: The 3' untranslated region of the human interferon-beta mRNA has an inhibitory effect on translation. Proc. Nat. Acad. Sci. USA, 84: 6030, 1988. Beutler, B., Krochin, R., Milsark, I. W., Leudke, C. and Cerami, A.: Control of cachectin (tumor necrosis factor) synthesis: mechanisms of endotoxin resistance. Science, 232: 977, 1986. Kronke, M., Hensel, G., Schluter, C., Scheurich, P., Schutze, S. and Pfizenmaier, K: Tumor necrosis factor and lymphotoxin gene expression in human tumor cell lines. Cancer Res., 48: 5417,1988. Kakehi, Y., Kanamaru, H., Yoshida, 0., Ohkubo, H., Nakanishi, S., Gottesman, M. M. and Pastan, I.: Measurement of multidrugresistance messenger RNA in urogenital cancers; elevated expression in renal cell carcinoma is associated with intrinsic drug resistance. J. Urol., 139: 862, 1988. Heicappell, R., Naito, S., Ichinose, Y., Creasey, A. A., Lin, L. S. and Fidler, I. J.: Cytostatic and cytolytic effects of human recombinant tumor necrosis factor on renal carcinoma cell lines derived from a single surgical specimen. J. Immunol., 138: 1634, 1987. Kavoussi, L. R., Ruesing, R. A., Hudson, M. A., Catalona, W. J. and Ratliff, T. L.: Effect oftumor necrosis factor and interferon gamma on human renal carcinoma cell line growth. J. Urol., 142: 875, 1989. Baisch, H., Otto, U. and Kloppel, G.: Antiproliferative and cytotoxic effects of single and combined treatment with tumor necrosis factor alpha and/or alpha interferon on a human renal cell carcinoma xenotransplanted into nu/nu mice: cell kinetics studies. Cancer Res., 50: 6389, 1990. Donaldson, J. T., Keane, T. E., Poulton, S. H. and Walther, P. J.: Enhanced in vivo cytotoxicity of recombinant human tumor necrosis factor with etoposide in human renal cell carcinoma. Urol. Res., 18: 245, 1990. Mutch, D. G., Powell, C. B., Kao, M.-S. and Collins, J. L.: In vitro analysis of the anti-cancer potential of tumor necrosis factor in combination with cisplatinum. Gyn. Oncol., 34: 328, 1989.