All-trans-retinoic acid upregulates TNF receptors and ... - Nature

ã

Oncogene (2000) 19, 2110 ± 2119 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc

All-trans-retinoic acid upregulates TNF receptors and potentiates TNF-induced activation of nuclear factors-kB, activated protein-1 and apoptosis in human lung cancer cells Sunil K Manna1 and Bharat B Aggarwal*,1 1

Cytokine Research Laboratory, Department of Bioimmunotherapy, Box 143, University of Texas MD Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, Texas, TX 77030, USA

Retinoids modulate the growth and dierentiation eects of TNF but the mechanism is not understood. In this study, we investigated the eect of all-trans-retinoic acid (ATRA) on the cell surface expression of TNF receptors and receptor-mediated signaling in various human lung cancer cell lines. ATRA treatment of cells that express wild-type p53 (A549 and H460), or null p53 (H1299), or mutant p53 (H596) increased the number of TNF receptors, as determined by the speci®c binding of 125Ilabeled TNF to these cells, in a dose- and time-dependent manner. Treatment with 2 mM ATRA for 24 h at 378C produced the maximal increase. Scatchard analysis indicated that the increase induced by ATRA was due to an increase in receptor number and not to an increase in anity. The upmodulation of TNF receptors was also con®rmed by covalent receptor-ligand cross-linking studies. The increase in TNF receptors sensitized H596 cells to TNF-induced activation of NF-kB, AP-1 and apoptosis. A549 cells, however, were completely resistant to TNF-induced activation of NF-kB, AP-1 and apoptosis. Treatment of these cells with as little as 0.5 mM ATRA was eective in converting TNF-resistant cells to TNF-sensitive. Overall our results indicate that ATRA induces the TNF receptors in human lung cancer cells, which sensitizes them to TNF-induced signaling leading to activation of NF-kB, AP-1 and apoptosis. Oncogene (2000) 19, 2110 ± 2119. Keywords: ATRA; TNF; NF-kB; AP-1; apoptosis Introduction Retinoic acid (RA) is an active metabolite of vitamin A and regulates a wide range of biological processes including cell proliferation, dierentiation and morphogenesis (DeLuca et al., 1991). RA is used as a chemopreventative agent in vivo (Benner et al., 1995) and induces apoptosis in a variety of dierent cell types in vitro (Lotan, 1995). The action of both natural and synthetic analogs of retinoids is mediated through speci®c nuclear receptors, called retinoic acid receptors (RAR-a, -b, and -g) and retinoid X receptors (RXR-a, -b, and -g) (Chambon et al., 1996). All-trans-RA (ATRA), for example, activates RAR ± RXR heterodimers to exert its biological action. Retinoids modulate the growth and dierentiation of both

*Correspondence: BB Aggarwal Received 27 December 1999; revised 21 February 2000; accepted 28 February 2000

myeloid and epithelial cells in vitro (Breitmann et al., 1980; Drach et al., 1993; Amos et al., 1989; Abemayor et al., 1989). In combination with various cytokines, ATRA shows a synergistic activity for dierentiation of hematopoietic cells (Hemmi et al., 1987; Trincheiri et al., 1987; Tobler et al., 1987; Peck et al., 1991), but the mechanism by which they synergize is not understood. RA have been shown to modulate the receptors for a number of dierent cytokines (Kilian et al., 1988; Sidell et al., 1988; Falk et al., 1991; Jetten et al., 1980; Scheibe et al., 1992; Winzen et al., 1992; Lotan et al., 1992; Sidell et al., 1991; Zheng et al., 1990). Although retinoids have promising potential for use in the prevention of lung cancer (Benner et al., 1995), a majority of human lung cancer cell lines in vitro are resistant to ATRA (Geradts et al., 1993). In contrast to ATRA, the lung cancer cells respond to human tumor necrosis factor (TNF), an apoptotic cytokine, and suppression of in vivo tumorigenicity of human lung cancer cells by retrovirus-mediated transfer of the TNF cDNA has been demonstrated (Han et al., 1994). TNF also shows synergistic action with interferon (IFN)-g in human lung cancer cells in vitro (Aggarwal et al., 1984; Beaupain et al., 1990). A cytokine that plays an important role in cell proliferation and dierentiation (Aggarwal and Vilcek, 1992), TNF exerts its eects by binding to two dierent TNF receptors with molecular masses of about 60 kDa (p60) and 80 kDa (p80) (Hohmann et al., 1989). Although most cells express both receptors, their relative abundance varies among dierent cell types. The p60 form of the TNF receptor is more prevalent on epithelial cells, whereas the p80 receptor is more abundant on cells of myeloid origin (Hohmann et al., 1989). TNF has been shown to synergize with ATRA for proliferation and dierentiation of dierent hematopoietic cells (Trincheiri et al., 1987; Tobler et al., 1987; Peck et al., 1991). In addition, we have shown that ATRA inhibits TNF production by macrophages stimulated by lipopolysaccharide and interferon (IFN)-g (Mehta et al., 1994). We have also reported that ATRA downregulates both types of TNF receptors in myeloid cells (Totpal et al., 1995). The eect ATRA has on TNF-mediated cellular responses in lung cancer cells has not been described, however. In the present study, we investigated the eect of ATRA on the expression of TNF receptors and on receptormediated cellular responses in human lung cancer epithelial cells. Our results indicate that, in contrast to myeloid cells, ATRA upregulates TNF receptors in all lung cancer cell lines tested and this correlates with

ATRA modulates TNF receptors and cellular responses SK Manna and BB Aggarwal

2111

sensitization of these cells to TNF-induced activation of NF-kB, AP-1, and apoptosis. Results The aim of this study was to examine the eect of ATRA on the TNF receptors and receptor-mediated cellular responses of human lung cancer cells. Treatment of these cells with ATRA under the experimental conditions had no eect on the viability of these cells (data not shown). We used cell lines that diered in p53 status because, the expression of death receptor (DR5), a member of the TNF receptor family, can be mediated through p53 (Wu et al., 1997). The role of p53 is further evident from a recent report that ATRA induces the expression of DR5 in lung cancer cell lines expressing wild-type p53 but not in cell lines with mutant p53 (Sun et al., 1999b). We, used human non-small cell lung cancer cell lines H460 and A549, which possess wild-type p53, cell line H596, which has mutant p53 and A1299, which has a null p53 mutation. ATRA upregulates cell surface expression of TNF receptors The four lung cancer cell lines were treated with dierent concentrations of ATRA at 378C for 24 h, washed, and then analysed for cell surface expression of TNF receptors (Figure 1a). ATRA increased the speci®c binding of 125I-labeled TNF on all lung cancer cell lines in a dose-dependent manner. The maximum increase occurred at 2 mM ATRA given for 24 h. The highest increase in TNF binding was observed with H596 cells (140%), followed by H460 (85%), A549 (70%) and A1299 (40%). Thus TNF receptor expression was induced by ATRA in lung cancer cell lines with wild-type p53, mutant p53, or null p53, suggesting that unlike DR5, p53 does not mediate the induction of TNF receptor expression by ATRA. An increase in TNF receptors was also observed on pretreatment with ATRA (1 mM for 24 h) of lung cancer cell lines, Calu-1 (112%), Calu-6 (98%) and H332 (140%). For most of the subsequent studies, we used lung cancer cell lines A549 and H596. Next we studied the time course of upregulation of TNF receptors by ATRA. The cells were exposed to 2 mM ATRA at 378C for dierent times and then assayed for speci®c binding of 125I-TNF. Binding increased in a time-dependent manner throughout the ATRA treatment (Figure 1b), reaching approximately 140% and 105% for H596 and A549 cells, respectively, after 24 h. To determine the receptor number and anity, the cells were treated with 2 mM ATRA for 24 h at 378C, washed, and then subjected to Scatchard analysis of receptor binding characteristics by incubating cells with variable amounts of 125I-labeled TNF in the absence or presence of a 50-fold excess of unlabeled TNF (Figure 2a). For untreated cells, the receptor numbers/cell and dissociation constant (Kd) were 4095 sites/cell and 0.265 nM, respectively, but for cells treated with 2 mM ATRA, the values were 8332 receptor numbers/cell and 0.275 nM, respectively. These results indicated that ATRA increased the number of TNF receptors,

Figure 1 ATRA induces TNF receptors in various human lung cancer cell lines. (a) Dose-dependent eect of ATRA on TNF receptors: H596, A549, H460, and A1299 cells (0.56106 ml) were cultured overnight in RPMI 1640 medium containing 10% FBS, washed, and suspended in the same medium containing 1% serum for 24 h at 378C with dierent concentrations of ATRA as indicated in the ®gure. The cells were subjected to 125I-TNF binding assay at 48C as described in Materials and methods. Results indicated in the ®gure are the percentage increases of 125ITNF binding above untreated control. (b) Eect of time-course of ATRA on TNF receptors: H596 and A549 cells (0.56106 ml) were cultured overnight in RPMI 1640 medium containing 10% FBS, washed, and suspended in the same medium containing 1% serum with 2 mM ATRA for dierent times as indicated in the ®gure. Then cells were used for 125I-TNF binding assay at 48C as described in Materials and methods. Results indicated in the ®gure are percentage increases of 125I-TNF binding above untreated control

whereas the anity of the receptors did not change signi®cantly (Figure 2b). To further con®rm the ATRA-mediated upmodulation of TNF receptors, we carried out receptor-ligand cross-linking studies on lung cells. 125I-labeled TNF was Oncogene


2112

Oncogene

cross-linked to the cell surface receptors on control and ATRA-treated cells using disuccinimidyl suberate (DSS), and the ligand-receptor complexes were visualized following SDS ± PAGE (Figure 3a). A single major band with an approximate molecular mass of 80 kDa appeared in untreated cells. An identical band was also observed in ATRA-treated cells, but the amount of cross-linked receptor-ligand complex was signi®cantly increased (Figure 3a). This band was

competed o by unlabeled TNF. After subtracting 17 kDa for the molecular mass of TNF, the remaining mass for the receptor is approximately 60 kDa, thus suggesting that the p60 receptor was upregulated by ATRA. An additional minor band observed at around 50 kDa must be a cleaved form of the 60 kDa form of the receptor. We could not, however, visualize the p80 receptor in either control or ATRA-treated cells, most likely because epithelial cells express only p60 receptor

Figure 2 ATRA increases TNF receptor number and not the anity. (a) Saturation kinetics of TNF binding to ATRA-treated and untreated H596 cells. H596 cells (0.16106/ml) were cultured in 12-well plates overnight in RPMI 1640 medium containing 10% FBS, washed, and suspended in the same medium containing 1% serum either in the presence or absence of 2 mM ATRA for 24 h at 378C . Then cells were washed with fresh cold medium containing 0.256106 cells/well; dierent concentrations of 125ITNF were added. For nonspeci®c binding, a 50-fold excess of unlabeled TNF was added 30 min before the addition of labeled TNF at 48C. TNF binding was assayed as described earlier. Speci®c TNF binding was detected after subtraction of nonspeci®c binding from total binding. (b) Scatchard analysis of ATRA-treated and untreated H596 cells. Results in a were calculated and Scatchard analysis done

Figure 3 Receptor-ligand cross-linking shows that ATRA upregulates TNF receptors (a). Induction of TNF receptors by ATRA on H596 and A549 cells: H596 and A549 (26106 in 5 ml) cells were cultured in 60 mm petri dishes overnight in RPMI 1640 medium containing 10% FBS, washed and suspended in the same medium containing 1% serum either in the presence or absence of 2 mM ATRA for 24 h at 378C. Then cells were incubated with 125 I-labeled TNF at 48C in the presence or absence of 50-fold excess cold TNF. Thereafter cells were washed with ice-cold DPBS (465 ml) and ®nally suspended in 1 ml D-PBS along with 1 mg/ml DSS (freshly prepared at 50 mg/ml DMSO). After 1 h at 48C with gentle mixing, cells were washed with ice-cold D-PBS (364 ml) and scrapped and the pellet was lysed with 0.05 ml lysis buer as described in Materials and methods. Cell extract protein (250 mg) was resolved on 8.5% SDS-polyacrylamide gel. The gel was dried and the radioactive band was detected by phosphoImager using Image Quant software. (b) Dose-dependent eect of ATRA on TNF receptors: H596 were exposed to dierent concentrations of ATRA and then processed as indicated in a


(Hohmann et al., 1989). The upregulation of the TNF receptor expression, as revealed by cross-linking, by ATRA was dose-dependent with maximum induction occurring at 1 mM (Figure 3b). ATRA upregulates the p60 form of TNF receptors in lung cells To further con®rm which receptor was in¯uenced by ATRA, we used receptor-speci®c antibodies. As shown in Figure 4a, ATRA induced TNF receptors in H596 cells and anti-p60 antibody abolished TNF binding in both untreated and ATRA-treated cells whereas anti-p80 antibody had no signi®cant aect on the binding. This

Figure 4 ATRA upregulates the p60 form of TNF receptor. H596 and A549 cells were treated with ATRA (2 mM) for 24 h at 378C. Then cells were washed with fresh medium and incubated with anti-TNF receptor antibody (against either p60 or p80) for 1 h at 378C. Thereafter, cells were washed with fresh ice-cold medium, and 125I-TNF binding examined as described earlier on both H596 (a) and A549 (b) cells

suggest that H596 cells express primarily p60 form of the TNF receptors and these are upregulated by ATRA. The type of TNF receptors expressed by A549 cells was also examined (Figure 4b). Like H596, A549 cells also expressed primarily p60 receptors and these were upregulated by ATRA. Thus these results are consistent with a report of Hohmann et al. (1989), that epithelial cells express only p60 form of the TNF receptors.

2113

ATRA upregulates the TNF-mediated NF-kB activation in lung cells The activation of NF-kB is one of the early signals induced by TNF in most cells (Chaturvedi et al., 1994). The activation of NF-kB by TNF in lung cells, however, has not been reported. Therefore, we ®rst examined the ability of TNF to activate NF-kB in H596 and A549 cells. Cells were exposed to various concentrations of TNF for 30 min, the nuclear extracts were then prepared and examined for NF-kB by EMSA (Figure 5a). TNF activated NF-kB in H596 cells in a dose-dependent manner and optimum activation (sixfold) occurring at 1000 pM TNF. Interestingly, however, no activation of NF-kB was observed in A549 cells even with 10 000 pM TNF. As shown above A549 cells do express TNF receptors, but these receptors appeared to be nonfunctional. Various combinations of Rel/NF-kB proteins can constitute an active NF-kB heterodimer that binds to speci®c sequences in DNA. To show that the retarded band visualized by EMSA in TNF-treated cells was indeed NF-kB, we incubated nuclear extracts from TNF-activated cells with antibody (Ab) to either p50 (NF-kBI) or p65 (Rel A) subunits and then conducted EMSA. Antibodies to either subunit of NF-kB shifted the band to a higher mw (Figure 5b), thus suggesting that the TNF-activated complex consisted of p50 and p65 subunits. Neither preimmune serum nor such irrelevant antibodies as anti-cRel or anti-cyclin DI had any eect on the mobility of NF-kB. Excess unlabeled NF-kB (100-fold) caused complete disappearance of the band, indicating the speci®city of NF-kB. The oligonucleotide with mutated NF-kB site failed to compete with the binding of the NF-kB protein to the oligonucleotide with wild-type NF-kB site. To determine the eect of ATRA on TNF-induced NF-kB activation, H596 and A549 cells were treated with 0.5, 1, and 2 mM ATRA for 24 h and then treated with dierent concentrations of TNF for 30 min and examined for NF-kB activation. As shown in Figure 5c, 10 pM TNF had no signi®cant eect on NF-kB activation in H596 cells. However when these cells were treated with even 0.5 mM ATRA, there was signi®cant activation of NF-kB by TNF. ATRA by itself did not activate NF-kB to any signi®cant extent. Thus these results indicate that ATRA enhanced TNF-induced NF-kB activation in H596 cells. As indicated above, TNF was unable to activate NF-kB in A549 cells. What eect ATRA has on the ability of TNF to activate NF-kB in these cells was also examined. As shown in Figure 5d, 100 pM TNF failed to activate NF-kB in these cells. Pretreatment with ATRA, however, sensitized A549 cells to respond to TNFinduced NF-kB activation. Thus induction of TNF receptors by ATRA correlated with sensitization of lung cells to ligand-induced NF-kB activation. Oncogene


2114

Figure 5 ATRA sensitizes lung cells to TNF-induced NF-kB activation. (a) Dose eect of TNF on NF-kB activation in H596 and A549 cells: H596 and A549 cells (26106/ml) were cultured in 6-well plates overnight in RPMI 1640 medium containing 10% FBS, washed, and cultured in the same medium containing 1% serum. After 24 h at 378C, cells were then stimulated with dierent concentrations of TNF as indicated in the ®gure for 30 min; the nuclear extracts were prepared and then assayed for NF-kB as described in Materials and methods. (b) Super-shift and speci®city of the TNF-induced NF-kB. Nuclear extracts were prepared from untreated or TNF (1 nM) treated H596 cells (26106/ml), incubated for 15 min with either dierent antibodies or with unlabeled wild-type or mutant NF-kB oligo and then assayed for NF-kB by EMSA as described in Materials and methods. (c and d) Eect of ATRA on TNF-dependent NF-kB activation. (c) H596 and (d) A549 cells (26106/ml) were cultured in 6-well plates overnight in RPMI 1640 medium containing 10% FBS, washed, and cultured in the same medium containing 1% serum along with 0, 0.5, 1 or 2 mM ATRA. After 24 h at 378C, cells were stimulated with 0, 0.01, 0.1 and 1 nM TNF, as indicated in the ®gure, for 30 min. The prepared nuclear extracts were then assayed for NF-kB as described in Materials and methods

ATRA upregulates TNF-induced AP-1 activation in lung cells TNF is also a potent activator of the transcription factor AP-1. This transcription factor, like NF-kB, is activated within minutes of TNF treatments and probably through a pathway that has overlapping and non-overlapping steps with that of NF-kB (Meyer et al., 1993; Karin et al., 1998). Therefore, we ®rst examined the ability of TNF to activate AP-1 in H596 and A549 cells. Cells were exposed to various concentrations of TNF for 30 min, and the nuclear extracts were prepared and examined for AP1 by EMSA (Figure 6a). TNF activated AP-1 in HP596 cells in a dose-dependent manner, optimum activation (fourfold) occurring at 100 pM TNF. Interestingly, however, no activation of AP-1 was observed in A549 cells even with 10 000 pM TNF. These results are similar to that noted for NF-kB activation. Oncogene

A combination of c-fos and c-jun proteins can constitute an active AP-1 heterodimer that binds to speci®c sequences in DNA. To show that the retarded band visualized by EMSA in TNF-treated cells was indeed AP-1, we incubated nuclear extracts from TNFactivated cells with antibody (Ab) to either c-fos or cjun subunits and then conducted EMSA. Antibodies to either subunit reduced the AP-1 band (Figure 6b), thus suggesting that the TNF-activated complex consisted of c-fos and c-jun subunits. Neither preimmune serum nor such irrelevant antibodies as anti-p50 or anti-cyclin DI had any eect on the mobility of AP-1. Excess unlabeled AP-1 oligo (100-fold) caused complete disappearance of the band, indicating the speci®city of AP-1. To determine the eect of ATRA on TNF-induced AP-1 activation, H596 and A549 cells were treated with 0.5, 1 and 2 mM ATRA for 24 h and then treated with dierent concentrations of TNF for 30 min and examined for AP-1 activation. As shown in Figure 6c,


2115

Figure 6 ATRA sensitizes lung cells to TNF-induced AP-1 activation. (a) Dose eect of TNF on AP-1 activation in H596 and A549 cells: H596 and A549 cells (26106/ml) were cultured in 6-well plates overnight in RPMI 1640 medium containing 10% FBS, washed, and cultured in the same medium containing 1% serum. After 24 h at 378C, cells were stimulated with dierent concentration of TNF as indicated in the ®gure for 30 min, nuclear extracts were prepared and then assayed for AP-1 as described in Materials and methods. (b) Super-shift and speci®city of TNF-induced AP-1. Nuclear extracts were prepared from untreated or TNF (1 nM)-treated H596 cells (26106/ml), incubated for 15 min with either dierent antibodies or with unlabeled AP-1 oligo, and then assayed for AP-1 by EMSA as described in Materials and methods. (c and d) Eect of ATRA on TNF-dependent AP-1 activation. (c) H596 and (d) A549 cells (26106/ml) were cultured in 6-well plates overnight in RPMI 1640 medium containing 10% FBS, washed and cultured in the same medium containing 1% serum along with 0, 0.5, 1 or 2 mM ATRA. After 24 h at 378C, cells were stimulated with 0, 0.01, 0.1 and 1 nM TNF, as indicated in the ®gure, for 30 min. The prepared nuclear extracts were then assayed for AP-1 as described in Materials and methods

10 pM TNF had no signi®cant eect on AP-1 activation in H596 cells. However when these cells were treated with even 0.5 mM ATRA, there was signi®cant activation of AP-1 by TNF. ATRA by itself did not activate AP-1 to any signi®cant extent. Thus these results indicate that ATRA can enhance TNF-induced AP-1 activation in H596 cells. As indicated above TNF was unable to activate AP-1 in A549 cells. What eect ATRA had on the ability of TNF to activate AP-1 in these cells was also investigated. As shown in Figure 6d, 100 pM TNF had minimal eect on AP-1 activation in these cells. Pretreatment with ATRA, however, sensitized A549 cells to respond to TNF-induced AP-1 activation. Thus induction of TNF receptors by ATRA correlated with sensitization of lung cells to ligand-induced AP-1 activation. ATRA upregulates TNF-induced apoptosis in lung cells Among cytokines, TNF is one of the most potent inducers of apoptosis (Rath and Aggarwal, 1999).

Whether ATRA modulates TNF-induced apoptosis in lung cells was also examined. H596 and A549 cells were exposed to various concentrations of TNF for 72 h either in the presence or absence of 2 mM ATRA, and then we examined cell viability by the MTT method. As shown in Figure 7a, TNF induced cytotoxicity in H596 cells, and this eect was potentiated by ATRA. In the absence of ATRA, H596 cells were unaected by 10 pM TNF, but 70% TNF-induced cytotoxicity was observed in the presence of ATRA. As was the case for NF-kB and AP-1 activation in the absence of ATRA, A549 cells were resistant to TNF-induced cytotoxicity even at the highest concentrations tested (Figure 7b). The presence of ATRA sensitized A549 cells to TNFinduced cytotoxicity. TNF-induced cytotoxicity involves the activation of caspases. The activation of caspase 2, -3 and -9 is known to cleave PARP substrate. How ATRA aects the TNFinduced cleavage of this substrate was also examined. Results with H596 cells indicate that ATRA had no signi®cant eect on TNF-induced PARP cleavage Oncogene


2116

Figure 7 ATRA potentiates TNF-induced apoptosis in human lung cancer cells. (a and b) ATRA enhances TNF-mediated cytotoxicity: H596 (a) and A549 (b) cells (56103/0.1 ml) were cultured in 96 well plate overnight in RPMI 1640 medium containing 10% FBS, washed, and cultured in same medium containing 1% serum either in presence or absence of 2 mM ATRA for 24 h at 378C. Dierent concentration of TNF were added and incubated for another 72 h at 378C in a CO2 incubator. MTT dye was then added (100 mg/well) and after 2 h, the cells were lysed with lysis buer containing SDS (20%) and DMF (50%), and absorbance was measured at 590 nm. The result indicated in ®gure was % mean O.D. of triplicate assays. (c and d) ATRA enhances TNF-mediated PARP cleavage: H596 and A549 cells (0.66106/ml) were cultured in 6-well plates overnight in RPMI 1640 medium containing 10% FBS, washed, and cultured in same medium containing 1% serum either in presence or absence of 2 mM ATRA for 24 h at 378C. Thereafter, cells were stimulated with dierent concentrations of TNF for 24 h at 378C, washed, extracted the pellet and 50 mg protein analysed on 7.5% SDS ± PAGE. Western blot was carried out against anti-PARP monoclonal antibody. The bands were located at 116 kDa and 80 kDa for H596 (c) and A549 (d)

(Figure 7c). TNF failed to induce PARP cleavage in A549 cells in the absence of ATRA (Figure 7d), but in presence of ATRA it induced cleavage of PARP. ATRA by itself had no eect on either of the cell lines. Discussion In the present report, we demonstrate that ATRA increases the cell surface expression of TNF receptors on various lung cancer cells lines independently of p53 expression. Scatchard analysis revealed that ATRA upregulated TNF receptor number without signi®cantly changing the receptor anity. This increase correlated with an increase in sensitization of lung cells to ligandinduced activation of NF-kB, AP-1, and apoptosis. This is the ®rst report to show that ATRA can modulate the number of TNF receptors on lung cells. Oncogene

ATRA has been shown to upregulate DR5, another member of the TNF receptor family, on lung cells, but its induction in that case was linked to p53 expression (Wu et al., 1997; Sun et al., 1999b). In our study, induction of TNF receptors by ATRA was p53-independent. Our results are consistent with a previous report, which showed upregulation of TNF receptor number and its mRNA upon treatment of human neuroblastoma SKNBE cells with ATRA (Chambaut-Guerin et al., 1995). Like TNF receptors, ATRA is also known to induce nerve growth factor receptors, another member of the TNF receptor family, on PC12 nerve cells (Scheibe and Wagner, 1992). On leukemic cells, however, we reported that ATRA downregulates TNF receptor number and receptor mRNA (Totpal et al., 1995). Thus the eects of ATRA on TNF receptors appear to be cell type-speci®c. How ATRA does modulate TNF receptors is not clear. It is possible that an increase in TNF


receptor number is due to increase in transcription of the TNF receptor gene mediated through retinoic acid receptor (RAR). Whether the promoter of the p60 TNF receptor gene contains RAR response elements is not known at present. Alternatively, it is also possible that ATRA increases TNF receptor through an indirect mechanism by stabilizing the turnover of the receptors. Upregulation of TNF receptors correlated with modulation of ligand-dependent signaling. In the case of H596 cells, TNF-dependent NF-kB activation was potentiated by ATRA pretreatment. These results are consistent with a report of Harant et al. (1996), who showed that ATRA synergizes with TNF in activating NF-kB-dependent IL-8 gene transcription in the human melanoma A3 cell line. In contrast to H596 cells, A549 cells were completely resistant to TNF-induced NF-kB activation. Only after treatment with ATRA did A549 cells respond to TNF-induced NF-kB activation. The lack of eect of TNF in the absence of ATRA, was not due to a lack of TNF receptors. Why untreated A549 cells were resistant to TNF is not clear. It has been reported that TNF is a potent inducer of manganese superoxide dismutase in A549 cells (Pogrebniak et al., 1991), which can block TNF-induced NF-kB activation (Manna et al., 1998). Whether conversion of A549 cells to a TNF-sensitive phenotype by ATRA is entirely due to induction of TNF receptors is also less likely. The regulation of TNF receptor does not always correlate with the cellular response (Hass et al., 1985; Chan and Aggarwal, 1994; Pandita et al., 1992; Aggarwal and Eessalu, 1987). This relationship varies, depending on the nature of the TNF-dependent response and also on the cell line. For instance antiproliferative eects of TNF require full receptor occupancy (Hass et al., 1985), whereas for TNF-dependent NF-kB activation it was found that only 10 ± 25% of the total receptors on U-937 cells were sucient for full response (Chan and Aggarwal, 1994). It is possible that ATRA modulates TNF signaling also at the post-receptor level independently of receptor induction. Our results on induction of NF-kB by TNF in lung cells may be relevant to previous studies that showed induced expression of various adhesion molecules (ICAM-1, VCAM-1 and ICAM-1) in lung cells by TNF (Sheski et al., 1999), inasmuch as the latter is regulated by NF-kB. Besides increasing NF-kB activation, ATRA also upregulated TNF-induced AP-1 activation. For H596 cells TNF-induced AP-1 activation was upregulated by ATRA; A549 cells were converted from TNF-insensitive to TNF-sensitive phenotype. These results dier from a report of Agadir et al. (1999) who found that ATRA inhibited phorbol ester-induced AP-1 activation in human lung H460 and H292 cells. These results also dier from a report of Moreno-Manzano et al. (1999) who showed abrogation of H2O2-induced AP-1 activation by ATRA in mesangial cells. This suggests a signi®cant heterogeneity in the eects of ATRA even within various lung cancer cell lines. Similar to TNF, ATRA is also a potent inducer of apoptosis in various cell types (Rath and Aggarwal, 1999; Sun et al., 1999a; Lotan et al., 1995; Oridate et al., 1995; Zou et al., 1998). We found that ATRA potentiates TNF-induced apoptosis in H596 cells. In contrast, A549 cells were completely resistant to TNF-induced apoptosis. This diers from a report by Prewitt et al. (1994), who found that TNF caused G0/G1 arrest in these cells,

which led to resistance to doxorubicin-induced apoptosis. Whether sensitization to TNF-induced apoptosis by ATRA is entirely due to induction of TNF receptors is not clear. Because Mn-SOD can also suppress TNFinduced apoptosis (Manna et al., 1998), it is possible that ATRA acts also via downregulation of anti-apoptotic factors including Mn-SOD. The role of NF-kB, both in stimulating and inhibiting apoptosis, has been demonstrated. The mechanisms of activation of NF-kB and apoptosis by TNF have both overlapping and nonoverlapping steps. Expression of antioxidant enzymes blocks TNF-induced activation of NF-kB and apoptosis (Manna et al., 1998, 1999). Thus it is not surprising that ATRA potentiates both NF-kB and apoptosis simultaneously. Overall our results demonstrate that ATRA sensitizes human lung cancer cells to TNF-induced cellular responses, in part through induction of TNF receptors. Although radiation and interferon have been used to potentiate the eects of TNF in lung cancer cells (Aggarwal et al., 1984; Gridley et al., 1996; Suarez Pestana et al., 1996), our results suggest the use of ATRA in combination with TNF for lung cancer. The in vivo suppression of tumorigenicity of human lung cancer cells by retrovirus-mediated transfer of the TNF cDNA (Han et al., 1994) might be enhanced by combination with ATRA.

2117

Materials and methods Materials RPMI-1640 and DMEM were obtained from Whittaker MA Bioproducts (Walkersville, MD, USA). Fetal bovine serum (FBS) and antibiotics-antimycotics (penicillin, streptomycin, and amphotericin B) were obtained from GIBCO (Grand Island, NY, USA). Bacteria-derived recombinant human TNF (speci®c activity 56107 U/mg) was kindly supplied by Genentech, Inc. (South San Francisco, CA, USA). Carrierfree Na 125I was purchased from Amersham (Arlington Heights, IL, USA); PD-10 (prepacked Sephadex G-25 medium) columns were from Pharmacia Fine Chemicals (Piscataway, NJ, USA); iodogen and gelatin were from Sigma Chemical Co. (St. Louis, MO, USA). Double-stranded oligonucleotides having the NF-kB and AP-1 consensus sequence were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Poly (ADP) ribose polymerase (PARP) antibody was purchased from PharMingen (San Diego, CA, USA). Polyclonal antibodies were raised in rabbits against each type of receptor and puri®ed by receptor-anity chromatography as described (Higuchi et al., 1992). All-trans retinoic acid was purchased from Kodak Fine Chemicals (Rochester, NY, USA). MTS (3-(4,5dimethylthiazol -2-yl)-5-(3-carboxymethoxyphenyl) -2-(4-sulfophenyl)-2H-tetrazolium, inner salt) was purchased from Promega (Madison, WI, USA). Cell culture Human non-small cell lung cancer cell lines (NSCLC) H460 and A549 which possess wild type p53, H596, which has mutant p53 (Mitsudomi et al., 1992), and A1299, which has null p53, were kindly provided by Dr Adi Gazdar (University of Texas Southwestern Medical Center, Dallas, TX, USA). Lung cancer cell lines Calu-1 (mutant p53), Calu 6 and H332 were supplied by Dr D Yu (University of Texas MD Anderson Cancer Center, Houston, TX, USA). These cells were grown in RPMI-1640 supplemented with 10% FBS and 50 mg/ml gentamicin. Oncogene


2118

Retinoic acid treatment Solutions of ATRA (1 mM) was prepared in dimethyl sulfoxide (DMSO) and then diluted in the appropriate medium. Cells (0.2 ± 0.56106/ml) were incubated with ATRA in 1% serum containing RPMI-1640 medium supplemented with 50 mg/ml gentamicin at 378C for 0 ± 24 h in 12-well plates or in T25 tissue culture ¯asks. The medium was then removed, and the cells washed, counted for viability, and then examined for TNF receptors. An appropriate DMSO control was run wherever necessary. Radiolabeling of TNF and receptor binding assay Human TNF was iodinated with [125I]-Na by the Iodogen method, puri®ed, and examined for cell surface receptors as described previously (Aggarwal et al., 1985). Brie¯y, cells (0.256106/ml) were cultured for 12 h and then treated with ATRA for 24 h. Thereafter cells were harvested and examined for TNF receptors by incubating with radiolabeled TNF (26105 c.p.m./ml) in the presence or absence of a 50fold excess of unlabeled TNF. After 4 h at 48C, cells were washed four times with phosphate buer saline and then, trypsinised and trypsin-containing cells were counted in a Gamma-counter. To calculate the speci®c binding, labeled TNF binding obtained in the presence of unlabeled TNF (nonspeci®c) was subtracted from labeled TNF binding in the absence of unlabeled TNF (total). Covalent cross-linking of TNF to cell surface receptors The cross-linking procedure was carried out according to the previously described method (Stauber et al., 1989). Either untreated or ATRA-treated human lung cancer cells (26106), were incubated with 125I-TNF (56105 c.p.m.) for 4 h at 48C, washed with ice cold D-PBS (465 ml) and then incubated with 1 mg/ml disuccinimidyl suberate (freshly prepared at 50 mg/ml DMSO) for 1 h at 48C with gentle mixing. Then cells were washed with ice cold D-PBS (365 ml) and scraped from the dish, and the cell pellet lysed with 0.05 ml lysis buer containing 20 mM Tris, pH 8.0, 137 mM NaCl, 10% glycerol, 1% Triton X100, 1 mM sodium orthovanadate, 2 mg/ml leupeptin, 2 mg/ml aprotinin, 1 mM phenylmethylsulfonyl ¯uoride, 0.5 mg/ml benzamidine, and 1 mM dithiothreitol. The lysate was centrifuged, and the supernatant was collected. Cell extract protein (250 mg) was resolved in 8.5% SDS-polyacrylamide gel. The gel was dried and the radioactive bands detected by phosphoImager using Image Quant software. Assay for NF-kB and AP-1 NF-kB was assayed by electrophoretic mobility shift assay (EMSA) following the method described by Chaturvedi et al. (1999). Brie¯y, nuclear extracts were prepared and EMSA were performed by incubating 5 mg of nuclear extract (NE) with 16 fmoles of 32P end-labeled 45-mer double-stranded NF-kB oligonucleotide from the HIV-LTR, 5'-TTGTTACAAG GGACTTTCCGCTG GGGACT TTCCAG GGAGGCGTGG-3' (bold letters represent NF-kB binding site). A double-stranded mutated oligonucleotide, 5'-TTGTTACA AC TCACTTTCCGCTG CTCA CTTTC CAGGG AGGCGTGG-3', was used to examine the speci®city of binding of

NF-kB to the DNA. The speci®city of binding was also examined by competition with the unlabeled oligonucleotide. For the AP-1 assay, the EMSA was similar to that for NFkB but used a 32P end-labeled double-stranded oligonucleotide of AP-1, 5'-CGCTTGATGACTCAGCCGGAA-3', 3'GCGAACTACTGAGTCGGCCTT-5' and the nuclear extract containing 6 mg protein was analysed on a 6.6% native gel. Speci®city of binding was determined routinely by using an excess of unlabeled oligonucleotide. TNF cytotoxicity assay The cytotoxic eect of TNF on cells was determined by the modi®ed tetrazolium salt 3-(4-5-dimethylthiozol-2-yl) 2-5diphenyl-tetrazolium bromide (MTT) assay (Haridas et al., 1998). Brie¯y, 5000 cells/well were pretreated with 2 mM ATRA for 24 h and then treated with indicated concentrations of TNF in a ®nal volume of 0.1 ml for 72 h at 378C. Thereafter, 0.025 ml of MTT solution (5 mg/ml in PBS) was added to each well. After a 2 h incubation at 378C, 0.1 ml of the extraction buer (20% SDS, 50% dimethyl formamide) was added. After an overnight incubation at 378C, the optical densities at 590 nm were measured using a 96-well multiscanner autoreader (Dynatech MR 5000), with the extraction buer as a blank. Immunoblot analysis of PARP degradation TNF-induced apoptosis was examined by proteolytic cleavage of PARP (Haridas et al., 1998). Brie¯y, 26106 cells/ml were pretreated with 2 mM ATRA for 24 h and then stimulated with dierent concentrations of TNF for 24 h at 378C. After treatment, cell extracts were prepared by incubating the cells for 30 min on ice in 0.05 ml of buer containing 20 mM HEPES pH 7.4, 2 mM EDTA, 250 mM NaCl, 0.1% NP-40, 2 mg/ml leupeptin, 2 mg/ml aprotinin, 1 mM PMSF, 0.5 mg/ml benzamidine, and 1 mM DTT. The lysate was centrifuged, and the supernatant was collected. Cell extract protein (50 mg) was resolved in 7.5% SDS ± PAGE, electrotransferred onto a nitrocellulose membrane, blotted with mouse antiPARP antibody, and then detected by chemiluminescence (ECL; Amersham). Apoptosis was represented by the cleavage of 116 kDa PARP into an 85 kDa peptide product.

Abbreviations TNF, tumor necrosis factor; ATRA, all-trans retinoic acid; NSCLC, non-small cell lung cancer cells; FBS, fetal bovine serum; AMF, autocrine motility factor; MTS, (3-(4,5dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt; IL, interleukin; TGF, transforming growth factor; DSS, Disuccinimidyl suberate; BSA, bovine serum albumin; PBS, phosphate-buered saline; EGF, epidermal growth factor; NGF, nerve growth factor; PkC, protein kinase C. Acknowledgments We wish to thank Dr Kapil Mehta for critical reading of the manuscript and Mr Walter Pagel for editorial suggestions. This research was conducted, in part, with a support from The Clayton Foundation of Research.

References Abemayor E and Sidell N. (1989). Environ. Health Perspect., 80, 3 ± 15. Aggarwal BB, Moat B, Lee SH and Harkins RN. (1984). Thymic Hormones and Lymphokine. AL Goldstein (ed.), Plenum Publishing Corp: New York, pp. 235 ± 245. Oncogene

Aggarwal BB and Vilcek J. (eds). (1992). In: Tumor Necrosis Factor: Structure, Function and Mechanism of Action. Marcel Dekker: New York, pp. 624. Aggarwal BB, Eessalu TE and Hass PE. (1985). Nature, 318, 665 ± 667.


Aggarwal BB and Eessalu TE. (1987). J. Biol. Chem., 262, 10000 ± 10007. Amos B, Cliord J and Lotan R. (1989). In: Pharmacology of Retinoids in the Skin. U Reichert and B Shroot (ed.). Karger: Basel, pp. 26 ± 36. Agadir A, Chen G, Bost F, Li Y, Mercola D and Zhang X-K. (1999). J. Biol. Chem., 274, 29779 ± 29785. Benner SE, Lippman SM and Hong WK. (1995). Oncology, 9, 205 ± 210. Beaupain R and Martyre MC. (1990). Anticancer Res., 10, 1061 ± 1066. Breitman TR, Selonick SE and Collins SJ. (1980). Proc. Natl. Acad. Sci. USA, 75, 2936 ± 2940. Chambon P. (1996). FASEB J., 10, 945 ± 954. Chan H and Aggarwal BB. (1994). J. Biol. Chem., 269, 31424 ± 31429. Chambaut-Guerin AM, Martinez MC, Hamimi C, Gauthereau X and Nunez J. (1995). J. Neurochem., 65, 537 ± 544. Chaturvedi M, LaPushin R and Aggarwal BB. (1994). J. Biol. Chem., 269, 14575 ± 14583. Chaturvedi M, Mukhopadhyay A and Aggarwal BB. (1999). Methods Enzymol., 319, 585 ± 601. Drach J, Lopez-Berestein G, McQueen T, Andree M and Mehta K. (1993). Cancer Res., 53, 2100 ± 2104. DeLuca LM. (1991). FASEB J., 5, 2924 ± 2933. Falk LA, Benedetti F, Lohrey N, Brichenall-Roberts MC, Ellingsworth LW, Faltynek CR and Rusetti FW. (1991). Blood, 77, 1248 ± 1255. Geradts J, Chen JY, Russell EK, Yankaskas JR, Nieves L and Minna JD. (1993). Cell Growth Dier., 4, 799 ± 809. Gridley DS, Andres ML, Garner C, Mao XW and Slater JM. (1996). Oncology Res., 8, 485 ± 495. Hemmi H and Breitman TR. (1987). Blood, 69, 501 ± 507. Han SK, Brody SL and Crystal RG. (1994). Am. J. Resp. Cell Mol. Biol., 11, 270 ± 278. Hohmann HP, Remy R, Brockhaus M and van Loon APGM. (1989). J. Biol. Chem., 264, 14927 ± 14934. Higuchi M and Aggarwal BB. (1992). J. Biol. Chem., 267, 20892 ± 20899. Hass PE, Hotchkiss A, Mohler M and Aggarwal BB. (1985). J. Biol. Chem., 260, 12214 ± 12218. Harant H, de Martin R, Andrew PJ, Foglar E, Dittrich C and Lindley JD. (1996). J. Biol. Chem., 271, 26954 ± 26961. Haridas V, Darnay B, Natarajan K, Heller R and Aggarwal BB. (1998). J. Immunol., 160, 3152 ± 3162. Jetten MA. (1980). Nature, 284, 623 ± 629. Kilian PL. (1988). In: Monokines and other Non-Lymphocytic Cytokines. Alan R. Liss, pp. 203 ± 208. Karin M and Delhase M. (1998). Proc. Natl. Acad. Sci. USA, 95, 9067 ± 9069. Lotan R. (1995). J. Natl. Cancer Inst., 87, 1655 ± 1657. Lotan R, Amos B, Watanabe H and Raz A. (1992). Cancer Res., 52, 4878 ± 4884. Mitsudomi T, Steinberg SM, Nau MM, Carbone D, D'Amico D, Bodner S, Oie HK, Linnoila RI, Mulshine JL, Minna JD, et al. (1992). Oncogene, 7, 171 ± 180. Mehta K, McQueen T, Tucker S, Samal B, Pandita R and Aggarwal BB. (1994). J. Leukoc. Biol., 55, 336 ± 342.

Manna S, Zhang HJ, Yan T, Oberley LW and Aggarwal BB. (1998). J. Biol. Chem., 273, 13245 ± 13254. Manna SK, Kuo MT and Aggarwal BB. (1999). Oncogene, 18, 4371 ± 4382. Meyer M, Schreck R and Baeuerle PA. (1993). EMBO J., 12, 2005 ± 2015. Moreno-Manzano V, Ishikawa Y, Lucio-Cazana J and Kitamura M. (1999). J. Biol. Chem., 274, 20251 ± 20258. Oridate N, Lotan D, Mitchell MF, Hong WK and Lotan R. (1995). J. Cell. Biochem., 23, 80 ± 86. Peck R and Bollag W. (1991). Eur. J. Cancer, 27, 53 ± 57. Pogrebniak HW, Prewitt TW, Matthews WA and Pass HI. (1991). J. Thor. Cardiovas. Sur., 102, 904 ± 907. Prewitt TW, Matthews W, Chaudhri G, Pogrebniak HW and Pass HI. (1994). J. Thor Cardiovas. Sur., 107, 43 ± 49. Pandita R, Pocsik E and Aggarwal BB. (1992). FEBS Lett., 312, 87 ± 90. Rath PC and Aggarwal BB. (1999). J. Clin. Immunol., 19, 350 ± 364. Sheski FD, Natarajan V and Pottratz ST. (1999). J. Lab. Clin. Med., 133, 265 ± 273. Sun SY, Yue P, Wu GS, El-Deiry WS, Shroot B, Hong WK and Lotan R. (1999a). Oncogene, 18, 2357 ± 2365. Sun SY, Yue P, Wu GS, El-Deiry WS, Shroot B, Hong WK and Lotan R. (1999b). Cancer Res., 59, 2829 ± 2833. Suarez Pestana E, Bjorklund G, Larsson R, Nygren P, Nilsson K and Bergh J. (1996). Acta Oncologica, 35, 473 ± 478. Scheibe RJ and Wagner JA. (1992). J. Biol. Chem., 267, 17611 ± 17616. Stauber GB and Aggarwal BB. (1989). J. Biol. Chem., 264, 3573 ± 3576. Sidell N, Taga T, Hirano T, Kishimoto T and Saxon A. (1991). J. Immunol., 146, 3809 ± 3814. Sidell N and Ramsdell F. (1988). Cell Immunol., 115, 299 ± 309. Scheibe RJ and Wagner JA. (1992). J. Biol. Chem., 267, 17611 ± 17616. Trinchieri G, Rosen M and Perrusia B. (1987). Blood, 69, 1218 ± 1224. Tobler A, Munker R, Heitjan D and Koeer P. (1987). Blood, 70, 1940 ± 1946. Totpal K, Chaturvedi MM and Aggarwal BB. (1995). Blood, 85, 3547 ± 3555. Wu GS, Burns TF, McDonald ER 3rd, Jiang W, Meng R, Krantz ID, Kao G, Gan DD, Zhou JY, Muschel R, Hamilton SR, Spinner NB, Markowitz S, Wu G and elDeiry WS. (1997). Nature Genet., 17, 141 ± 143. Winzen R, Wallach D, Engelmann H, Nophar Y, Brakebusch C, Kemper O, Resch K and Holtmann H. (1992). J. Immunol., 148, 3454 ± 3460. Zou CP, Kurie JM, Lotan D, Zou CC, Hong WK and Lotgan R. (1998). Clin. Cancer Res., 4, 1345 ± 1355. Zheng ZS and Goldsmith LA. (1990). Cancer Res., 50, 1201 ± 1205.

2119

Oncogene