Antiviral agent Cidofovir restores p53 function and enhances ... - Nature

ã

Oncogene (2002) 21, 2334 ± 2346 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Antiviral agent Cidofovir restores p53 function and enhances the radiosensitivity in HPV-associated cancers Bassam Abdulkarim1,4, Siham Sabri2,4, Eric Deutsch1, Heddia Chagraoui2, Laurence Maggiorella1, Jerome Thierry3, FrancËois Eschwege1, William Vainchenker2, Salem ChouaõÈ b3 and Jean Bourhis*,1 1 Laboratoire UPRES EA N827-10 `RadiosensibiliteÂ-RadiocarcinogeÂneÁse humaine' and UniteÂ METSI, Institut Gustave-Roussy, 94805 Villejuif, France; 2INSERM U362, Institut Gustave-Roussy, 94805 Villejuif, France; 3INSERM U487 Cytokines et immunologie des tumeurs humaines, Institut Gustave-Roussy, 94805 Villejuif, France

High-risk human papillomaviruses (HPVs) have been associated to the development of cervical and some other human cancers. Most of them express E6 and E7 oncoproteins, able to bind to p53 and retinoblastoma (pRb) tumor suppressor proteins respectively and neutralize their function. Restoration of these pathways by blocking E6 and E7 expression would provide a selective therapeutic eect. Here, we show that a clinically approved antiviral agent Cidofovir reduced E6 and E7 expression in cervical carcinoma Me180 and head and neck squamous cell carcinoma HEP2 cells at the transcriptional level. Cidofovir induced the accumulation of active p53 and pRb associated to induction of cyclin dependent kinase inhibitor p21WAF1/CIP1 in Me180 and HEP2 cells. p53 induction was also shown in Hela HPVpositive cervical carcinoma cell line. In addition, S phase cell cycle accumulation with concomitant decrease of cyclin A expression were associated to the antiproliferative activity of Cidofovir in HPV-treated cells. Combining Cidofovir to irradiation both in vivo and in nude mice xenografts resulted in a marked radiosensitization in HPV-positive cells, which was not observed in virus negative cells. This study provides the basis for a new anticancer strategy to enhance the antitumor eect of ionizing radiation in HPV-related cancers, without increase deleterious eects. Oncogene (2002) 21, 2334 ± 2346. DOI: 10.1038/sj/ onc/1205006 Keywords: HPV E6/E7; p53; human carcinoma; Cidofovir; ionizing radiation Introduction Over 90% of human cervical carcinoma are associated with high-risk HPV (human papillomavirus), mainly *Correspondence: J Bourhis, Laboratoire UPRES RadiosensibiliteÂRadiocarcinogeÂneÁse humaine, Institut Gustave Roussy, 39 rue Camille Desmoulins, 94805 Villejuif ceÂdex France; E-mail: [email protected] 4 These authors contributed equally to the work Received 1 June 2001; revised 24 September 2001; accepted 1 October 2001

the serotypes 16 and 18 (Lazo, 1999; zur Hausen et al., 2000) which are also involved in 15 ± 20% of head and neck squamous cell carcinoma (HNSCC) (Gillison et al., 1999). The high-risk HPVs encode two transforming proteins E6 and E7 able to immortalize primary human keratinocytes and to extend the life span of infected squamous epithelial cells (DiPaolo et al., 1993; Schiman, 1995). The continued expression of these viral oncoproteins is directly involved in growth regulation, since expression of the antisense E6/E7 mRNA is sucient to decrease cell growth in cervical carcinoma cells (von Knebel-Doeberitz et al., 1988). E6 and E7 oncoproteins exert profound eects on tumor suppressor proteins p53 and pRb (Villa, 1997). P53 and pRb normally control pathways that regulate the cell cycle and protect the integrity of the genome. Unlike most of human cancers in which p53 and pRb are mutated, in cervical carcinoma and derived cell lines, E6 and E7 oncoproteins bind to wild-type p53 and pRb respectively and neutralize their function. Indeed, E6 protein binds to p53 and promotes its degradation via the ubiquitin proteolytic pathway, inhibiting its growth function (Schener et al., 1991). Similarly, the E7 protein targets hypophosphorylated members of pRb family for ubiquitinmediated proteolysis, resulting in decreased pRb levels and an inappropriate release of active E2F transcription factor in cells expressing the viral oncoproteins (Smith-McCune et al., 1999). As a consequence of these interactions, cells expressing E6 and E7 proteins display aberrant cell cycle checkpoint control and exhibit high rate of mutagenesis (Demers et al., 1996; Hickman et al., 1994). Although p53 and pRb protein levels are low in cervical carcinoma, their signaling pathways are still functional (Butz et al., 1995; Wu et al., 2000). Repression of viral oncoproteins expression would hence restore growth control in these cells by targeting E6 and E7. For this purpose, several approaches have been used such as antisense strategies (Steele et al., 1993; He et al., 1997), delivering HPV E2 gene to directly repress E6 and E7 transcription (Hwang et al., 1996; Goodwin, 2000;

Combined effects of Cidofovir and IR B Abdulkarim et al

Francis et al., 2000), targeting E6 oncoprotein by peptide aptamers (Butz et al., 2000) and inhibiting E6/E7 transcription by nuclear export inhibitors (Hietanen et al., 2000). Although these strategies demonstrated that down-regulating E6/E7 could restore growth control, their ecacy was limited due to insucient gene delivery into tumors or to cytotoxic eects induced in normal cells. Given the critical role of viral oncoproteins in HPV-related cancers, antiviral strategies provide an attractive approach to target speci®cally HPV-expressing cells. Recently, experimental studies reported that antiviral agent Cidofovir [(S)-1-[3-hydroxy-2-phosphonyl methoxy propyl] cytosine `HPMPC'], and acyclic nucleoside phosphonate (ANP), showed antiproliferative activity against HPV-infected cells (Andrei et al., 1998; Johnson and Gangemi, 2000) and emerged as a selective drug for HPV-infected cells. However, the mechanisms by which Cidofovir exerts this inhibitory eect remain unknown. Also, the use of this drug in combination with conventional anticancer therapies has not been reported. To gain new insights into the molecular pathways underlying Cidofovir activity, we investigated in HPVpositive cervical carcinoma and HNSCC cell lines, the eect of this antiviral agent on E6/E7 expression and its consequence on p53 and pRb tumor suppressor pathways. Furthermore, we investigate whether Cidofovir could modulate both in vitro and in nude mice xenografts the tumor sensitivity to ionizing radiation (IR), the most commonly used cytotoxic agent in HPVs-related cancers. In this study, we show that Cidofovir downregulates E6 and E7 oncoproteins at the transcriptional level, with subsequent reactivation of p53 and pRb expression. For the ®rst time, we provide evidence that an antiviral agent can eciently restore tumor suppressor pathways and enhance the therapeutic eect of IR both in vitro and in vivo. Taken together, these ®ndings represent a novel therapeutic approach to target speci®cally HPV-positive cells and enhance the therapeutic ratio in high-risk HPV-related cancers.

Results Cidofovir down-regulates the constitutive and the radiation-induced E6/E7 expression We used immunoblot analysis to assess the expression of E6 and E7 oncoproteins after Cidofovir exposure (1 to 10 mg/ml) in Me180 and HEP2 cell lines. Densitometric analysis revealed that the amount of E6 protein in Me180 (Figure 1a) and HEP2 (Figure 1b) cells decreased by 30 and 25% respectively at 3 days. This eect was more pronounced at 6 days (60 and 80% respectively). In addition, similar reduction in HPV E7 expression was observed in Me180 and HEP2 cells at 3 days (40, 60% respectively) and 6 days (65, 85% respectively) (Figure 1a,b).

2335

Figure 1 Down-regulation of E6 and E7 oncoproteins by Cidofovir exposure in Me180 and HEP2 cell lines. Western blot analysis of E6 and E7 oncoproteins in Me180 (a) and Hep2 (b) cell lines before and after Cidofovir treatment (10 mg/ml) for 3 and 6 days. Blots were incubated with monoclonal antibodies to E6 (C1P5, abcam) or E7 (Ab-1, Oncogene research) and detected with ECL reagent. To assess protein loading, blots were probed with an antibody speci®c for b-actin (AC-40, Sigma)

We further performed semi-quantitative RT ± PCR to analyse E6/E7 mRNA levels in HEP2 cell line after Cidofovir (1 to 10 mg/ml), ionizing radiation (3, 6 and 9 Gy) and combined treatment. Using speci®c primers, we also analysed the eect of these treatment on HPV transcripts in the same cell line. The quanti®cation of RT ± PCR was determined in the linear range of ampli®cation de®ned by preliminary experiments to be between 20 ± 30 cycles. The relative levels of E6/E7 mRNA and HPV transcripts were normalized against b2 microglobulin mRNA. In HEP2 cells, HPV transcripts were reduced by 44 and 79% at 3 and 6 days respectively. Similar reduction was observed in E6/E7 mRNA levels after Cidofovir exposure (Figure 2a). The mRNA species (4,6 kb) corresponding to E6/ E7 mRNA were reduced by 50 and 70% at 3 and 6 days respectively, compared to untreated cells. Conversely, HPV transcripts levels measured 24 h after exposure to IR (9 Gy) were enhanced by threefold and the level of E6/E7 mRNA by up twofold. However, the radiation-induced increase of E6/E7 mRNA was abolished when cells were pretreated with Cidofovir (Figure 2b). Oncogene


2336

Figure 2 Cidofovir reduced the constitutive and the radiationinduced E6/E7 mRNA expression. Semi-quantitative RT ± PCR of HPV transcripts and E6/E7 mRNA from HEP2 cells in the absence and presence of Cidofovir as described in Materials and methods. (a) Quanti®cation of the mRNA levels after normalization for b2-microglobulin mRNA in Cidofovir treated cells (10 mg/ml) for 3 and 6 days. (b) The relative levels of HPV transcripts and E6/E7 mRNA after IR (3, 6 or 9 Gy for 24 h) and the combined treatment (Cidofovir 10 mg/ml for 6 days and IR 9 Gy for the last 24 h). The values are reported as the relative level over untreated cells. C33A HPV-negative cell line was used as a negative control. HPV transcripts and E6/E7 mRNA levels in HEP2 cells was performed as described in Materials and methods

Induction of p53 and pRb proteins following Cidofovir exposure of HPV-expressing cells E6 and E7 oncoproteins are known to target ubiquitinmediated proteolysis of p53 and pRb respectively (Schener et al., 1993; Boyer et al., 1996). Using p53 functional assay in yeast (Flaman et al., 1995), we Oncogene

con®rmed that p53 gene was wild-type in Me180 and HEP2 cells and mutated in C33A (data not shown). Consistent with these data and with the reduction of E6 and E7 expression after Cidofovir exposure, we used immunoblotting to assess p53 and pRb expression. The level of p53 expression was almost undetectable in untreated cells and was slightly increased by IR in Me180 (Figure 3a) and HEP2 cells (Figure 3b). The amount of p53 protein was markedly increased following Cidofovir exposure (3 days) in Me180 (10-fold) and HEP2 (sevenfold). Combining Cidofovir to IR showed similar induction in Me180 (Figure 3a) and HEP2 cells (Figure 3b). We further investigated the eect of Cidofovir on p53 level in Hela HPV-positive cervical carcinoma cell line expressing p53 wild-type and E6 and E7 oncoproteins. We performed a dose response (1, 10, 100 mg/ml) in Hela cells treated with Cidofovir for 3 days. We found that p53 was at the threshold level of detection and this induction was observed even at 1 mg/ ml (Figure 3c). In HPV-negative cervical cancer cells C33A, expression of p53 was not altered by Cidofovir (Figure 3a). As previously shown, E7 oncoprotein binds to hypophosphorylated p105 form resulting in its degradation and inappropriate release of E2F transcription factor (Berezutskaya et al., 1997). We therefore examined the eect of Cidofovir on pRb protein expression. Cidofovir exposure clearly increased pRb protein level in Me180 cells compared to untreated cells (4 ± 5-fold). Interestingly, the ratio of hypophosphorylated active form related to hyperphosphorylated form was higher in Cidofovir than in irradiated cells (1.7 versus 0.8) (Figure 3a). This result suggests that Cidofovir induced a preferential accumulation of hypophosphorylated form as a consequence of E7 decrease, while the low level of active form after IR could be due to the increase of E7 oncoprotein. In HPV-negative cervical cancer cells C33A, expression of pRb was not aected by Cidofovir (Figure 3a). Nuclear accumulation of p53 induced by Cidofovir in HPV-positive cervical carcinoma cells Previous studies have shown that p53-mediated eects on p53 responsive promoters involved its localization in the nucleus (El-Deiry et al., 1993; Crook et al., 1998). To address this question, we performed immuno¯uorescence assays on Me180 cells containing wild-type p53 and harboring E6, and C33A cells which contain mutant p53 protein and do not express E6 protein. Figure 3b illustrates a slight induction of p53 by IR, and its localization within the cytoplasm and the nucleus of Me180 treated cells. Consistent with our ®ndings (Figure 3a), we observed a marked increase of p53 staining in Cidofovir-treated cells (Figure 3d). Interestingly, p53 was mainly located within the nucleus and displayed punctate nuclear staining suggesting a functional activation following Cidofovir treatment. On the other hand, in C33A cell line p53 exhibited cytoplasmic and nuclear localization which


2337

Figure 3 Cidofovir induced reactivation of p53 and pRb in HPV-positive cells. (a) Western blot analysis of cell extracts from cultures treated cells for 3 days with Cidofovir and/or IR Equal amounts of whole cell lysates from Me180 and C33A cells were subjected to SDS ± PAGE and immunoblotted by using antibodies against p53 (DO7, Pharmingen) or pRb (G3-245, Pharmingen). Lane 1: untreated cells; lane 2: IR 3Gy for 24 h; lane 3: Cidofovir (10 mg/ml) for 3 days; and lane 4: (Cidofovir 10 mg/ml for 3 days and IR 3 Gy for the last 24 h). (b) Immunoblotting of p53 in HEP2 cells treated with Cidofovir and/or IR (c) Dose response of Cidofovir (untreated cells, 1, 10 and 100 mg/ml) following drug exposure during 3 days in Hela cell line. (d) Immuno¯uorescence microscopy of p53 in Me180 cells treated by IR (3 Gy for 24 h) or Cidofovir (10 mg/ml for 3 days). (e) Immuno¯uorescence microscopy of p53 in C33A cells treated by IR (3 Gy for 24 h) or Cidofovir (10 mg/ml for 3 days). Nuclei are showed in the Dapistained cells. In the right panel, arrowheads show p53 concentrated in the nuclei of Cidofovir treated Me180 cells or in the cytoplasm of C33A cells following ionizing radiation. Data are representative of three independent experiments Oncogene


2338

was not modi®ed by Cidofovir exposure (Figure 3e). In contrast with the eect found in Me180 cells, IR induced a perinuclear accumulation of p53 with prominent cytoplasmic aggregates in C33A cells (Figure 3e). Cidofovir induced expression of p53-responsive gene p21WAF1/CIP1 Further search for evidence of p53-mediated transactivation, we performed Western blot analysis of p53 target gene, p21WAF1/CIP1. We investigated whether the level of p21WAF1/CIP1 protein was altered by Cidofovir and/or IR treatment in ME180, HEP2 and C33A. p21WAF1/CIP1 expression was very low in Me180 (Figure 4a) and HEP2 (Figure 4b) cell lines. The protein level was markedly increased by Cidofovir alone in Me180 and HEP2 cells (seven and 2.5-fold respectively) and combined treatment (nine and threefold respectively) (Figure 4a,b). In contrast, p21WAF1/CIP1 expression was not altered by Cidofovir and/or IR in C33A cells (Figure 4c). We next studied the localization of the CDK inhibitor p21WAF1/CIP1 after Cidofovir exposure. Immuno¯uorescence microscopy of untreated Me180 cells showed weak staining, with restricted cytoplasmic localization in cells expressing p21WAF1/CIP1 protein (Figure 4d). The same pattern of staining was found in irradiated cells (data not shown). In agreement with p21WAF1/CIP1 induction observed in Western blot analysis, Cidofovir induced a striking increase of p21WAF1/CIP1 expression with persistence cytoplasmic localization, but also appearance of strong nuclear staining (Figure 4d). Cidofovir induced S phase arrest with inhibition of cyclin A expression in HPV-positive cells Since p53 constitutes an important cell cycle checkpoint regulator, it was essential to examine whether the dierence in p53 status could alter the ability of cells to progress through the cell cycle in Cidofovir treated cells. Using ¯ow cytometry, G2/M block was found in Me180 irradiated cells (3 Gy/24H). In contrast, after 3 days with Cidofovir alone (10 mg/ml), Me180 cells accumulated in S phase (50%) with corresponding decrease in G1 phase. This eect was maintained after 4 days of Cidofovir exposure (data not shown). The combined treatment induced an increase in G1 with concomitant decrease in S phase indicating that cells were blocked in G1 phase (Figure 5a). In C33A cells, G2/M block was found in response to IR, but no signi®cant block was found in S phase (30 versus 25%) after Cidofovir alone or combined treatment (Figure 5b). In these cell lines, there was no signi®cant increase (55%) in apoptotic cell death (subG1 peak) in response to irradiation (3, 6, and 9 Gy) and/or Cidofovir at a concentration between 1 and 10 mg/ml. This result was further con®rmed using TUNEL assay (data not shown). In addition, we sought to determine the eect of Cidofovir on cyclin expression involved in S phase.

Oncogene

Using immunoblot analysis (Figure 5c), expression of cyclin A in Me180 cells decreased after Cidofovir exposure (®vefold) with less pronounced eects with combined treatment. In contrast, expression of cyclin E was increased after Cidofovir or combined treatment (threefold). In C33A cells, cyclin A and E protein levels were unaected by Cidofovir exposure (Figure 5d). Cidofovir inhibits cell proliferation and markedly enhances radiosensitivity in HPV-positive cancer cells in vitro To analyse the consequences of E6/E7 down-regulation and cell cycle alterations, we studied cell proliferation and clonogenic survival assays in Me180, HEP2 and C33A cell lines (Figure 6). Cidofovir caused inhibition of cell proliferation in a dose-dependant manner (1, 10, 50 and 100 mg/ml) in Me180 cells. As shown in Figure 6a, Cidofovir (10 mg/ml) eectively inhibited the growth in Me180 cells by 40 and 70% at 3 and 5 days. Similar results were obtained in HEP2 cells (data not shown). Of note, cell cycle arrest and the biochemical changes observed in this study were induced by Cidofovir at 10 mg/ml. In contrast, no signi®cant growth inhibition occurred in C33A cells, and was not increased even with high concentration of Cidofovir (Figure 6b). Clonogenic survival assay (Figure 6c) showed that Me180 cells exhibit higher inhibition of colony formation than C33A cells at dierent concentrations of Cidofovir. At high concentrations (100 mg/ml), no viable colonies were detected after exposure to the compound in Me180 cells. We also used clonogenic survival assay to determine whether Cidofovir would aect sensitivity to irradiation (Figure 7). A strong eect of Cidofovir on radiation-induced clonogenic cell death was observed in HPV-positive cells. Exposure of the two HPVpositive cell lines (Me180 and HEP2) to Cidofovir prior irradiation was more eective than irradiation alone. In terms of cell killing, this eect was seen at doses of irradiation as low as 0.5 Gy or 1 Gy (Figure 7a,b). Interestingly, the surviving fraction at 2 Gy (SF2) which constitutes a clinically relevant dose was markedly reduced with combined treatment than with irradiation alone (2.3 vs 29.6% for Me180 and 2.5 vs 28% for HEP2 cells). In contrast, in C33A cells the combined treatment provided a low improvement compared to irradiation alone (35 vs 27%) and this eect was not enhanced at higher dose of irradiation (4 Gy) (Figure 7c). Synergistic effect of Cidofovir and IR on HPV+ human cancer xenografts in vivo Having established the radiosensitizing potential of Cidofovir in vivo, we next investigated its antitumor activity in combination with IR in xenografts nude mice. Treatments were initiated when the mean tumor volume was 144 mm3+14 in HEP2 and 124 mm3+24 in Me180 cells. In saline treated group, the mean tumor


2339

Figure 4 Cidofovir induced expression and nuclear localization of p21Cip1/Waf1. Equal amounts of whole cell lysates from Me180 (a) and HEP2 (b) cells were subjected to SDS ± PAGE and immunoblotted by using antibodies against or p21 (Ab-1, Oncogene research). Lane 1: untreated cells; lane 2: IR 3Gy for 24 h; lane 3: Cidofovir (10 mg/ml) for 3 days; and lane 4: (Cidofovir 10 mg/ml for 3 days and IR 3 Gy for the last 24 h). (c) C33A cells were similarly treated and immunoblotted for p21 proteins as described in Materials and methods. (d) Immuno¯uorescence analysis of p21Cip1/Waf1 illustrates cytoplasmic localization in untreated Me180 cells and induction of p21Cip1/Waf1 associated to its nuclear localization following Cidofovir exposure (10 mg/ml for 3 days) (as pointed out with arrowheads in the right panel)

volume increased by 465% within 26 days (Figure 8a). In mice treated with Cidofovir or irradiation alone, mean tumor volume increased by 185 and 209% respectively in Me180 xenografts. By contrast, when Cidofovir and irradiation were combined, the tumors dramatically regressed by 70% in the same period (Figure 8a). In nude mice bearing HEP2 xenografts (Figure 8b), the mean tumor volume increased by 332% within 28 days in saline treated group. A moderate inhibition of tumor growth was found with Cidofovir or radiotherapy alone. By contrast, when Cidofovir and radiotherapy were combined, the tumors regressed by 47%. Combining Cidofovir and IR produced a

statistically signi®cant tumor growth delay both in Me180 and HEP2 xenografts (P50.001). In all the combined groups, tumors were found to be in complete remission, more than 40 days after the treatment. Interestingly, no deleterious eect (skin) was observed in these combined groups. Discussion Carcinogenesis of HPV-infected cells, is a process generally associated with the integration of the viral genome in cancer cells, resulting in the loss of expression of the viral E2 gene and the persistence of E6 and E7 Oncogene


2340

a

b

Figure 5 Induction of S phase cell cycle arrest and reduction of cyclin A expression. Flow cytometry analysis of Me180 (a) and C33A (b) cells exposed to Cidofovir drug (10 mg/ml, 3 days) and/or ionizing radiation (3 Gy). The percentage of G1, S and G2/M phases are indicated in each column. Expression levels of cyclin A (BF683, Santa Cruz) and cyclin E (HE12, Santa Cruz) were determined by immunoblotting analysis of lysates from Me180 (c) and C33A (d) cells treated as indicated for FACS pro®les. The blots were stripped and re-probed with a monoclonal b-actin antibody. Data are representative of three independent experiments

oncoproteins expression (Jeon et al., 1995). In cervical carcinoma cells, p53 and pRb tumor suppressor genes are frequently wild-type, but their function is disrupted by the expression of E6 and E7 oncoproteins. In the present study, we show for the ®rst time that a clinically approved antiviral agent Cidofovir (Safrin et Oncogene

al., 1997) was able to inhibit E6 and E7 expression (Figure 1) and to restore p53 and pRb pathways (Figures 3 and 4). In HPV-treated cells, this antiviral agent accumulates in its biologically active form (HPMPCdiphosphate), which inhibits cellular DNA polymerase required in viral gene transcription (Pisarev et al., 1997,


2341

Figure 6 Antiproliferative eect of Cidofovir in Me180 and C33A cell lines. (a) and (b) time-course of cell proliferation using dierent concentrations of Cidofovir in Me180 and C33A cells respectively. (c) Me180 and C33A cells were incubated with increasing doses of Cidofovir (1, 10, 50 and 100 mg/ml) for 24 h and assayed for clonogenic survival assay as described in Materials and methods. Values shown represent the mean of triplicate determinations calculated from a single experiment. Experiments were repeated at least three times

reviewed in Abdulkarim et al., 2001). The ®nding that inhibition of both HPV transcripts and E6/E7 mRNA occurred upon Cidofovir exposure (Figure 2) are agreement with the inhibition of host DNA polymerase.

Figure 7 Eect of Cidofovir on the sensitivity of human cervical carcinoma cells to IR. (a) Me180, (b) HEP2 and C33A (c) cell lines were treated with either no drug or Cidofovir (5 and 10 mg/ ml) for 24 h, followed by the indicated single doses of radiation. Cell survival was determined by colony formation assay. Colonies were stained with methylene blue and the surviving fraction was corrected for cell death attributable to drug alone. Data shown are from a representative experiment. Each experiment was repeated at least three times

IR is the most commonly used cytotoxic agent for cervical carcinoma and HNSCC cells. Despite this, few studies have investigated the eects of IR on the expression of E6/E7 oncoproteins. Santin et al. (1998) Oncogene


2342

Figure 8 The combination of Cidofovir and ionizing radiation induced regression of HPV+ tumors in vivo. Nude mice harboring Hep2 (a) or Me180 (c) xenografts tumors were treated with PBS (open circles), IR alone (7 Gy on day 3 and 5) (®lled circles), or Cidofovir alone (100 mg/kg, day 1 to 5, intra-peritoneally) (open lozenges), or combined treatment (®lled lozenges). Data show the mean tumor volume+s.e.m. expressed in relation to the starting tumor volume. Eight xenografts were used in each group, in two independent experiments. (b) and (d) illustrate the change in mean tumor volumes at 26 and 28 days in ME180 and HEP2 xenografts respectively and expressed as a per cent change in tumor volume

have reported that high doses of IR (12.5 ± 100 Gy) could increase E6/E7 expression in cervical carcinoma cell lines. In accordance with this study, we showed that clinically relevant doses were sucient to increase E6/E7 mRNA expression (Figure 2). Interestingly, combining Cidofovir to IR can eectively inhibit the radiation-induced up-regulation of E6/E7 expression (Figure 2). Using antisense E6/E7 oligonucleotides, the level of E6/E7 has been related to the proliferative capacity of cervical carcinoma cells (Steele et al., 1993; Hu et al., 1995). The radiationinduced increase expression of E6/E7 would confer a growth advantage to HPV-tumor cells, which could Oncogene

escape to local IR. The combination of this antiviral agent and IR may hence be useful to prevent E6/E7 oncoproteins radiation-induced up-regulation and tumor failure. Furthermore, we show that E6 and E7 decrease upon Cidofovir exposure is associated with p53 accumulation and its concomitant nuclear translocation in HPV-positive cells (Figure 3). As previously described, p53 acts as a transcriptional activator of a variety of target genes following its stabilization and re-localization into the nucleus (El-Deiry et al., 1993; Crook et al., 1998). In accordance with these studies, we also demonstrate the increase of p21 expression in


both Me180 and HEP2 cell lines suggesting the restoration of p53 function (Figure 4). Previous studies reported the induction of p53 upon proteasome inhibition in ME180 and Siha cervical carcinoma cells (Freedman and Levine, 1998; Mantovani and Banks, 1999). However, p53 protein failed to localize into the nucleus suggesting that physical interaction between E6 and p53 is able to sequestrate p53 in the cytoplasmic compartment. Our data provide evidence that inhibition of E6 expression is sucient to p53 stabilization and nuclear translocation in Me-180 cells (Figure 3). The E7 oncoprotein has been shown to overcome a DNA-damage-induced G1 cell cycle arrest (Slebos et al., 1994). It has been reported that E7 binds to p21WAF1/CIP1 and blocks its inhibitory function, resulting in abrogation of DNA damage-induced cell cycle arrest (Funk et al., 1997; Jones et al., 1997a). In our study, restoration of cell cycle arrest in HVP-positive cells is probably due to Cidofovir-induced p21WAF1/CIP1 expression. Indeed, E7 decrease would restore p21WAF1/ CIP1 function. The nuclear localization of p21WAF1/CIP1 was correlated to its cell growth inhibiting activity (Goubin and Ducommun, 1995) while the cytoplasmic form was involved in protection against apoptosis (Asada et al., 1999). Nuclear localization of p21WAF1/ CIP1 induced by Cidofovir supports this potential role in cell cycle inhibition in HPV-treated cells (Figure 4). The importance of pRb function is emphasized by the fact that E7 protein can bind and inactivate pRb, allowing progression into S phase without a mitogenic stimuli (reviewed in Moran, 1993). Besides the well documented importance of pRb in G1/S transition, recent reports based on over-expression of non-pophosphorylated pRb mutants suggested a novel function of active form of pRb in the inhibition of S phase progression (Chew et al., 1998; Knudsen et al., 2000). Cytotoxic agents, including cisplatin and etoposide can inhibit S phase progression and this arrest was observed with pRb-positive but not with pRb-negative cells (Knudsen et al., 2000). The antiproliferative activity of pRb is mediated by its ability to inhibit transcription of genes required for cell progression (reviewed in Kaelin, 1999). It has been shown that active pRb inhibits the G1/S transition by repressing cyclin A promoter activity and result in reduced cyclin A protein levels (Knudsen et al., 1999). This cell cycle inhibition can be abrogated directly through ectopic expression of cyclin A or indirectly through ectopic expression of cyclin E, which acts upstream to abrogate pRb function. In agreement with these results, we demonstrated in Me180 cells that Cidofovir induced active hypophosphorylated form of pRb (Figure 3) with concomitant cyclin A reduction likely through reduction of E7 expression. Our results are also in accordance with a previous report showing that the reduction of cyclin A is correlated to E2F repression in HPV-positive cells infected by E2 gene (Goodwin et al., 2000). The increase of cyclin E protein levels with the reduction of cyclin A suggests that cell cycle arrest after Cidofovir exposure correlates with the accumulation in early S phase.

The response to irradiation in the two cell lines (Me180 and C33A) could be explained by their respective p53 status. p53-induced genes (p21, Bax, GADD45 and Fas) can determine whether a cell exposed to DNA damaging agents will progress through the cell cycle, arrest in G1 or G2, repair DNA damage or die via cell apoptosis (Levine, 1997). Me180 cells which harbor a wild-type p53 exhibited a G1/S arrest in response to Cidofovir combined with IR. In contrast, C33A cells expressing a mutated p53 gene, continued to progress into S phase and G2/M after the combined treatment. Therefore, it was important to analyse the eect of Cidofovir and/or IR on the ability of these cells to undergo apoptosis. Previous studies have shown that Cidofovir can induce apoptosis in Epstein barr virus-positive cells and polyomavirus-induced hemangiosarcomas (Neyts et al., 1998; Liekens et al., 2001). In our experiments, we did not detect increased of apoptosis after Cidofovir treatment and/or irradiation. Cidofovir was proven to be `cytostatic' at doses used in this study and the reduction in cell growth in HPV-positive cell lines was most probably due to cell cycle arrest without increase of apoptosis. It is noteworthy that in previous attempts to inhibit E6/E7 oncogene expression by antisense or ribozymes, growth inhibition of HPV-positive cells was reported rather than apoptosis (von Knebel-Doeberitz et al., 1988; Steele et al., 1992; He and Huang, 1997; Alvarez-Salas et al., 1998). This could be due partly to the concomitant reduction of proapoptotic E7 levels which promotes apoptosis in cells expressing p53 wildtype (Tan and Ting, 1995; Jones et al., 1997b; Stoppler et al., 1998; Venturini et al., 1999). Several reports have shown that restoration of wild type p53 expression increases the radiosensitivity in dierent human carcinoma models (Gallardo et al., 1996; Spitz et al., 1996). The synergistic eect observed when combining Cidofovir and IR observed in vitro and in mice xenografts is likely due to the restoration of p53 pathway and reestablishment of G1/S checkpoint by Cidofovir in HPV-positive cells. We are currently investigating whether E6 transfected cell lines harboring wild-type or mutated p53 could suppress p53-radiosensitization in Cidofovir treatedcells (Cai et al., 1997). With respect to cancer therapy, the viral etiology of cervical carcinoma allows the development of therapeutic approaches that could eliminate selectively HPV-positive cells by speci®cally targeting viral oncoproteins without aecting normal cells. The ®ndings that in HPV-positive cancer cell lines degradation of p53 depends entirely on the action of E6 (Hengstermann et al., 2001) constitutes an attractive target for the development of novel strategies for the treatment of cervical cancers. Moreover, combining Cidofovir and IR is of particular interest since radiotherapy remains the main cytotoxic agent used in cervical and head and neck carcinoma. This study provides the basis for a novel anticancer strategy in HPV-related cancers which needs to be further investigated for clinical applications.

2343

Oncogene


2344

Materials and methods Cell lines and cell culture Four cervical carcinoma cell lines Me180, Hela, C33A and head and neck squamous cell carcinoma (HNSCC) cell line hEP2 used in this study were purchased from ATCC. Me180 cells were grown in Mc'Coy medium, and Hela, HEP2 and C33A were maintained in MEM medium, supplemented with 10% FCS. Me180, Hela and HEP2 cell lines contain integrated HPV-18 sequences. The status of p53 was studied by functional test in yeast in these cell lines (Flaman et al., 1995).

Antiviral compound Cidofovir (HPMPC) was obtained from Pharmacia Upjohn (les Yvelines, France). The half life of this compound is 48 h in vivo. To keep stable concentration of Cidofovir in tissue culture medium, we added this drug three times per week by removing all medium and replacing it with fresh medium containing new Cidofovir. The cell lines were treated with various doses of Cidofovir (1 ± 10 mg/ml) for 3 or 6 days. Cells were harvested, washed in PBS and lysed for RNA or protein extraction.

Immunoblot analysis Cells were washed with PBS, and lysed in lysis buer containing Nonidet P-40 for 30 min on ice. Cell debris was removed by centrifugation and protein concentration was determined by Bio-Rad protein assay (Biorad). Equal amounts of proteins were boiled for 5 min in SDS sample buer (40 mM Tris-HCl, pH 7.4, 5% glycerol, 5% mercaptoethanol, 2% SDS, 0.05% bromophenol blue) and subjected to SDS-polyacrylamide gel electrophoresis. The proteins were electrotransfered to nitrocellulose membranes (Hybond, Biorad Laboratories, Hercules, CA, USA) and blocked overnight at 48C in PBS, 0.1% Tween, containing 5% nonfat dried milk. Blots were incubated with monoclonal antibodies: anti-E6 (clone C1P5, abcam), anti-E7 (clone Ab-1, Oncogene Research), anti-p53 (clone DO7, Pharmingen), anti-pRB (clone G3-245, Pharmingen), anti-p21 (clone Ab-1, Oncogene Research), anti-cyclin A (clone BF683, Santa Cruz) or anticyclin E (clone HE12, Santa Cruz) and detected with appropriate secondary antibodies conjugated to horseradish peroxydase. Filters were developed using enhanced chemiluminescence ECL system (BioRad). Anti-b-actin (clone AC-40, Sigma) was used to control protein loading. The densitometric analysis of the bands was performed using MAC BAS software.

Immunofluorescence microscopy Semi-quantitative reverse transcriptase-polymerase chain reaction (RT ± PCR) Total RNA was prepared using a commercial guanidine thiocyanate/phenol method (Trizol, GIBCO BRL) as described by the manufacturer. Total RNA (2 mg) was reverse transcribed at 428C for 45 min in 20 ml of 50 mmol/l Tris-HCL, pH 8.3 containing 75 mmol/l KCl, 3 mmol/l MgCl2, 5 mmol/l dithiothreitol (DTT), 1U Rnase inhibitor, xU Super script II reverse transcriptase, 500 nmol/l dNTPs, and 300 ng of random hexamers. The mixture was then denaturated at 708C for 15 min. 2.5 ml of these complementary DNA (cDNA) products were ampli®ed with 1.5 U Taq (ATGC) in the presence of speci®c primers for the mRNA of interest. Reactions were performed in 50 ml of 10 mmol/l Tris-HCl, 50 mmol/l KCl, 0.01% gelatin, 1.5 mmol/l MgCl2, 200 mmol/l of each dNTP and 15 picomoles of sense and antisense primers. In order to amplify the HPV-18 E6/E7 open-reading frame, the composition of PCR primers were 5' - ATGGCGCCCTTTGAGGATCC - 3', nt 105 ± 134; and 5' - TTACTGCTGGGATGCACACC - 3', nt 907 ± 888 (Kamradt, 2000). The RNA isolated was also ampli®ed for HPV RNA with PCR, and the primers used were: 5' - TGTCAAAAACCGTTGTGTTCC - 3' and 5' GAGCTGTCGCTTAA TTG CTC - 3' (Chiba et al., 1996). b2 microglobulin was used as an internal control. PCR conditions were as follows: for HPV, 60 s at 948C, 2 min at 608C, 30 s at 728C; for E6/E7 60 s at 948C, 1 min 30 s at 648C, 3 min at 728C using a Gene Amp PCR system 2400 (Perkin Elmer). All the RT ± PCR were done in the linear range of ampli®cation de®ned by preliminary experiments to be between 20 ± 30 cycles. Samples containing distilled water and samples without RT enzyme were included as negative controls. Aliquots (10 ml) were removed from the mixture PCR and the ampli®ed products were electrophoresed on a 2% agarose gel and stained with ethidium bromide. The intensity of ampli®ed bands was analysed by the software MAC BAS. Oncogene

Me180 and C33A cell lines were grown on glass coverslips, ®xed for 10 min with 2% paraformaldehyde in PBS and permeabilized with Triton (0.2%) for 4 min. Coverslips were blocked for 1 h with 20% heat-inactivated normal goat serum in PBS. Fixed cells were then incubated with monoclonal anti-p53 (clone DO7, Pharmingen) or antip21WAF1/CIP1 (clone Ab-1, Oncogene Research) diluted with 3% bovine serum albumine (BSA) or IgG1 as a control for 1 h. The slides were washed and incubated for 1 h with TRITC anti-mouse antibody (Jackson Immunoresearch). Coverslips were mounted with antifading Vectashield (Vector) containing DAPI to visualize nuclei and samples were examined by ¯uorescence microscopy (Nikon).

DNA content analysis by propidium iodide (PI) staining Flow cytometry analysis of PI-stained cells was performed to demonstrate the eect of Cidofovir and/or IR on cell cycle progression. Brie¯y, cells were harvested at dierent times, washed, and ®xed in 70% ethanol overnight at 48C. Before ¯ow cytometry, cells were washed and stained with 1 ml of PI (15 mg/ml) containing 2.5 mg/ml Rnase. DNA content was determined by a FACScan ¯ow cytometer (Becton Dickinson) and the proportion of cells in a particular phase of cell cycle was determined by CellQuest software. Clonogenic survival and cellular proliferation assays To study the inhibitory eect of Cidofovir on growth, cells (105/ml) were incubated with various concentrations of Cidofovir (1, 10, 50 and 100 mg/ml) for 3, 5 and 7 days in 96 well-plates. After culture, cell number and viability were evaluated by staining with trypan blue and counting using light microscopy. To evaluate cell killing for HEP2, Me180 and C33A, colony forming assay was performed by plating the appropriate number of cells into culture dishes in triplicate.


Exponentially growing cells were incubated with cytostatic concentration of Cidofovir (1, 5 and 10 mg/ml) for 24 h. The cells were trypsinized and resuspended in fresh medium before plating them into culture dishes. After 4 h to allow cell attachment, the dishes were irradiated. The cells were irradiated (0 ± 4 Gy) at room temperature, with a 137Cs grays source at a dose rate of 1.45 Gy min71. After 14 days, the number of colonies (450 cells) was scored. Colonies were stained with cristal violet and the surviving fraction was normalized to the surviving fraction of untreated cells. Growth of human tumor xenografts in vivo and experimental design Female Swiss nu/nu mice were housed throughout experiments in sterile isolators and fed ad libitum with irradiated food (UAR, Villemoisson, France) and ®ltered water (n886/609/ CEE of the European Community). 36106 cells were injected subcutaneously (s.c.) into the ¯ank of animals aged between 6 and 8 weeks. At day 0, tumor volume was determined by direct measurement, calculated by the formula (D.d2). Mice were randomly assigned to treatment groups (6 ± 8 in each group) and tumor volume was measured twice a week

until death or for at least 40 days. To study the potential interaction between Cidofovir and irradiation, animals were assigned in four groups in Me180 and HEP2 xenografts: (1) Control group received PBS, (2) Cidofovir (100 mg/kg, day 1 to 5), (3) irradiation (7 Gy on day 3 and 5) and (4) Combined group receiving Cidofovir (100 mg/kg, day 1 to 5) and irradiation (7 Gy on day 3 and 5). Cidofovir was given intraperitoneally (i.p.) and the nephrotoxicity was prevented by probenecid i.p. administration before Cidofovir treatment.

2345

Statistical analysis in vivo data are reported as the percentage of original (day 0) tumor volume and plotted as fractional tumor volume (s.e. statistical signi®cance was determined by Kruskal/Wallis and Mann-Witney U-tests.

Acknowledgements We are especially grateful to Dr S ChouaõÈ b for his constructive discussion about the experiments and for the critical reading of the manuscript.

References Abdulkarim B and Bourhis J. (2001). Lancet Oncology, 2, 622 ± 630. Andrei G, Snoeck R, Piette J, Delvenne P and De Clercq E. (1998). Oncol. Res.,10, 523 ± 531. Alvaraz-Salas LM, Cullinan AE, Siwkowski A, Hampel A and DiaPaolo A. (1998). Proc. Natl. Acad. Sci. USA, 95, 1189 ± 1194. Asada M, Yamada T, Ichijo H, Delia D, Miyazono K, Fukumuro K and Mizutani S. (1999). EMBO. J., 18, 1223 ± 1234. Bartek J, Bartkova J and Lukas J. (1997). Exp. Cell. Res., 237, 1 ± 6. Berezutskaya E, Yu B, Morozov A, Raychaudhuri P and Bagchi S. (1997). Cell Growth Dier., 8, 1277 ± 1286. Boyer SN, Wazer DE and Band V. (1996). Cancer Res., 56, 4620 ± 4624. Butz K, Denk C, Ullmann A, Schener M and Hoppe-Seyler F. (2000). Proc. Natl. Acad. Sci. USA, 6, 6693 ± 6697. Butz K, Shahabeddin L, Geisen C, Spitkovsky D, Ullmann A and Hoppe-Seyler F. (1995). Oncogene, 10, 927 ± 936. Cai Z, Capoulade C, Moyret-Lalle C, Amor-Gueret M, Feunteun J, Larsen AK, Paillerets BB and Chouaib S. (1997). Oncogene, 15, 2817 ± 2826. Chew YP, Ellis M, Wilkie S and Mittnacht S. (1998). Oncogene, 29, 2177 ± 2186. Chiba I, Shindoh M, Yasuda M, Yamazaki Y, Amemiya A, Sato Y, Fujinaga K, Notani K and Fukuda H. (1996). Oncogene, 12, 1663 ± 1668. Crook T, Parker GA, Rozycka M, Crossland S and Allday MJ. (1998). Oncogene, 16, 1429 ± 1441. Demers GW, Espling E, Harry JB, Etscheid BG and Galloway A. (1996). J. Virol., 70, 6862 ± 6869. DiPaolo JA, Popescu NC, Alvarez L and Woodworth CD. (1993). Crit. Rev. Oncog., 4, 337 ± 3360. El-Deiry WS, Tokino T, Valculescu VE, Levy DB, Parsons R, Trent JM, Lind D, Mercer WE, Kinzler KW and Vogelstein B. (1993). Cell, 75, 817 ± 825.

Flaman JM, Frebourg T, Moreau V, Charbonnier F, Martin C, Chappuis P, Sappino AP, Limacher IM, Bron L and Benhattar J. (1995). Proc. Natl. Acad. Sci. USA, 92, 3963 ± 3967. Francis DA, Schmid SI and Howley PM. (2000). J. Virol., 74, 2679 ± 2686. Freedman DA and Levine AJ. (1998). Mol. Cell. Biol., 18, 7288 ± 7293. Funk JO, Waga S, Harry JB, Espling E, Stillman B and Galloway DA. (1997). Genes Dev., 11, 2090 ± 2100. Gallardo D, Drazan K and McBride W. (1996). Cancer Res., 56, 4891 ± 4903. Gillison ML, Koch WM and Shah KV. (1999). Curr. Opin. Oncol., 11, 1 ± 9. Goodwin EC and DiMaio D. (2000). Proc. Natl. Acad. Sci. USA, 97, 12513 ± 12518. Goodwin EC, Yang E, Lee CJ, Lee HW, DiMaio D and Hwang ES. (2000). Proc. Natl. Acad. Sci. USA, 97, 10978 ± 10983. Goubin F and Ducommun B. (1995). Oncogene, 10, 2281 ± 2287. He Y and Huang L. (1997). Cancer Res., 57, 3993 ± 3993. Hengstermann A, Linares LK, Ciechanover A, Whitaker NJ and Schener M. (2001). Proc. Natl. Acad. Sci. USA, 98, 1218 ± 1223. Hickman ES, Picksley SM and Vousden KH. (1994). Oncogene,9, 2177 ± 2181. Hietanen S, Lain S, Krausz E, Blattner C and Lane DP. (2000). Proc. Natl. Acad. Sci. USA, 97, 8501 ± 8506. Hu G, Liu W, Hanania EG, Fu S, Wang T and Deisseroth AB. (1995). Cancer. Gene. Ther., 2, 19 ± 32. Hwang ES, Naeger LK, DiMaio D. (1996). Oncogene, 15, 795 ± 803. Jeon S, Allen-Homann BI and Lambert PF. (1995). J. Virol., 69, 2989 ± 2997. Johnson JA and Gangemi JD. (2000). Antimicrob. Agents. Chemother., 43, 198 ± 205.

Oncogene


2346

Oncogene

Jones DL, Alani RM and Munger K. (1997a). Genes Dev., 11, 2101 ± 2111. Jones DL, Thompson DA and Munger K. (1997b). Virology, 239, 97 ± 107. Kaelin Jr WG. (1999). Bioessays, 21, 950 ± 958. Kamradt MC, Mohideen N, Krueger E, Walter S and Vaughan AT. (2000). Br. J. Cancer, 82, 1709 ± 1716. Knudsen ES, Buckmaster C, Chen TT, Feramisco JR and Wang JY. (1998). Genes. Dev., 12, 2278 ± 2292. Knudsen KE,, Fribourg AF, Strobeck MW, Blanchard JM and Knudsen ES. (1999). J. Biol. Chem., 274, 27632 ± 27641. Knudsen KE, Booth D, Naderi S, Sever-Chroneos Z, Fribourg AF, Hunton IC, Feramisco JR, Wang JY and Knudsen ES. (2000). Mol. Cell. Biol., 20, 7751 ± 7763. Lazo PA. (1999). Br. J. Cancer, 80, 2008 ± 2018. Levine AJ. (1997). Cell, 88, 323 ± 331. Liekens S, Verbeken E, De Clercq E and Neyts N. (2001). Int. J. Cancer, 92, 161 ± 167. Mantovani F and Banks L. (1999). Oncogene, 18, 3309 ± 3315. Moran E. (1993). Curr. Opin. Genet. Dev., 13, 261 ± 291. Neyts J, Sadler R, De Clercq E, Raab-Traub N and Pagano JS. (1998). Cancer Res., 58, 384 ± 388. Pisarev VM, Lee SH, Connelly MC and Fridland A. (1997). Mol. Pharmacol., 52, 63 ± 68. Safrin S, Cherrigton J and Jae HS. (1997). Rev. Med. Virol., 7, 145 ± 156. Santin AD, Hermonat PL, Ravaggi A, Chiriva-Internati M, Pecorelli S and Parham GP. (1998). Cancer, 83, 2346 ± 2352.

Schener M, Munger K, Byrne JC and Howley PM. (2000). Proc. Natl. Acad. Sci. USA, 88, 5523 ± 5527. Schener M, Huibregtse JM, Vierstra RD and Howley PM. (1993). Cell, 75, 495 ± 505. Schiman MH. (1995). J. Nat. Can. Inst., 87, 1345 ± 1347. Slebos RJ, Lee MH, Plunkett BS, Kessis TD, Williams BO, Jacks T, Hedrick L, Kastan MB and Cho KR. (1994). Proc. Natl. Acad. Sci. USA, 91, 5320 ± 5324. Smith-McCune K, Kalman D, Robbins C, Shivakumar S, Yuschenko L and Bishop JM. (1999). Proc. Natl. Acad. Sci. USA, 96, 6999 ± 7004. Spitz FR, Nguyen D, Skibbe JM, Meyn RE, Cristiano RJ and Roth JA. (1996). Clin. Cancer Res., 2, 1665 ± 1671. Steele C, Cowsert LM and Shillitoe EJ. (1993). Cancer Res., 15, 2330 ± 2337. Steele C, Sacks PG, Adler-Storthz K and Shillitoe EJ. (1992). Cancer Res., 52, 4706 ± 4711. Stoppler H, Stoppler MC, Johnson E, et al. (1998). Oncogene, 17, 1207 ± 1214. Tan TM and Ting RC. (1995). Cancer Res., 55, 4599 ± 4605. Venturini F, Braspenning J, Homann M, Gissmann L and Sczakiel G. (1999). Nucleic Acids Res., 27, 1585 ± 1592. Villa LL. (1997). Adv. Cancer Res., 71, 321 ± 341. von Knebel-Doeberitz M, Oltersdorf T, Schwarz E and Gissmann L. (1988). Cancer Res., 48, 3780 ± 3786. Wu L, Goodwin EC, Naeger LK, Vigo E, Galaktionov K, Helin K and DiMaio D. (2000). Mol. Cell, 20, 7059 ± 7067. zur Hausen H. (2000). J. Natl. Cancer Inst., 92, 690 ± 698.