Inhibition of BRCA1 Leads to Increased Chemoresistance to Microtubule-Interfering Agents, an E€Ect That Involves the JNK Pathway

Inhibition of BRCA1 leads to increased chemoresistance to microtubule-interfering agents, an eect that involves the JNK pathway

SteÂ phane Lafarge1, ValeÂ rie Sylvain1, Marc Ferrara2 and Yves-Jean Bignon*,1

1Laboratoire d'Oncologie MoleÂculaire, Centre Jean Perrin, BP 392, 63011 Clermont-Ferrand cedex 1-France; 2UNCM, INRA-Theix, 63122 Saint Genes-Champanelle-France

We have developed ribozymes (Rz) that inhibit BRCA1 line mutations have been described, and the majority of expression in order to study the role of this gene in these result in truncated protein. However, decreased chemosensitivity. Two Rz, targeting positions 358 or BRCA1 mRNA levels have been observed in sporadic 5282 of the BRCA1 mRNA, were cloned into the breast cancer cases, showing the possible involvement retroviral vector LXSN and lipofected into the breast of BRCA1 in both sporadic and hereditary breast cancer cell-line HBL100. We obtained 79 ± 99% inhibi- cancers. BRCA1 is located at 17q21 and encodes a tion of BRCA1 expression, as determined by real-time 220 kDa nuclear phosphoprotein (Chen et al., 1996; quantitative PCR and by Western blotting. Decreased Runer and Verma, 1997) with several motifs: an N- expression of BRCA1 led to sensitivity to the DNA terminal RING ®nger domain, two BRCT domains damaging agents cisplatin and etoposide, resistance to and a transactivating domain in the C-terminus the microtubule-interfering agents (MIA) taxol and (Chapman and Verma, 1996), suggesting that BRCA1 vincristine. The molecular mechanism of resistance to can act as a transcription factor (reviewed in Welcsh et MIA was investigated further by determining the status al., 2000). Moreover, BRCA1 is a component of the of the JNK pathway. We found that JNK1 expression RNA polymerase holoenzyme II (Schlegel et al., 2000; was elevated, while JNK2 expression was decreased in Scully et al., 1997a) and can interact with several Rz-expressing clones compared to controls. We have proteins involved in the regulation of transcription, quanti®ed the mRNA levels of BRCA1, JNK1, 2, MEK- including CREB binding protein/p300 (Pao et al., 4, -7 and c-jun after treatment with MIA. Vincristine 2000), RNA helicase A (Anderson et al., 1998) and treatment of control cells resulted in transcriptional CtIP/CtBP (Li et al., 1999; Wong et al., 1998; Yu et repression of BRCA1, while the JNK1, 2, MEK-4, -7 al., 1998). It has been shown that BRCA1 can associate and c-jun genes were induced. In Rz-treated cells, only with p53 both in vitro and in vivo to stimulate its JNK1 and MEK-4 were expressed and none was induced transcriptional activation of the p21waf1/cip1 (Somasun- after MIA treatment. We then studied the phosphoryla- daram et al., 1997), mdm2 (Ouchi et al., 1998) and bax tion of c-jun, a downstream eector of the JNK promoters. BRCA1 can also transactivate the expres- pathway. We observed a strong increase in phosphory- sion of p21waf1/cip1 in a p53 independent manner lated c-jun after MIA treatment of the control cells but (Somasundaram et al., 1997) and modulate the not in BRCA1-Rz treated cells, suggesting inhibition of transcriptional activity of c-myc (Wang et al., 1998a). the JNK pathway. These results show that the BRCA1- All these elements make BRCA1 a candidate for a JNK pathway is involved in the cytotoxic response to transcriptional coactivator. MIA treatment, and inhibition of BRCA1 leads to BRCA1 is also involved in maintaining genome transcriptional modi®cations of the JNK pathway. integrity: it becomes activated by phosphorylation and Oncogene (2001) 20, 6597 ± 6606. relocalizes after exposure to DNA double strand break damage (Scully et al., 1997b). It interacts with the Rec Keywords: BRCA1; microtubule-interfering agents; A homologue hRad51 (Scully et al., 1997c) and JNK; ribozyme BRCA2 (Chen et al., 1998) in subnuclear foci, and activates homologous recombination repair of double- strand breaks (Chen et al., 1999). It can also interact Introduction with the hRad50-hMre11-p95/nibrin complex after irradiation by forming nuclear foci (Zhong et al., BRCA1 (Miki et al., 1994) is involved in 20 ± 45% of 1999). Recently, it has been found that BRCA1 can be inherited breast cancer cases and 80% of families phosphorylated by ATM (Cortez et al., 1999), hCds1/ predisposed to breast and ovarian cancer. Only germ- Chk2 (Lee et al., 2000) and ATR (Chen, 2000) in response to gamma-irradiation. Li et al. (1999) have also shown that CtIP interacts with BRCA1 and is involved in mediating transcription-coupled DNA *Correspondence: Y-J Bignon; E-mail: [email protected] repair (TCR) of p21waf1/cip1. So BRCA1 seems to be Received 22 March 2001; revised 4 June 2001; accepted 5 July 2001 implicated in the TCR machinery and modulation of Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6598 its expression leads to modi®cations of TCR and radio- gene expression. Targeting a speci®c mRNA is possible and chemo-resistance. It was demonstrated that the by modifying the Rz sequence, which is composed of a upregulation of BRCA1 expression leads to increased catalytic core and two ¯anking sequences (Figure 1). resistance to cisplatin in the SK-OV-3 human ovarian Rz can cleave any sequence containing a NUX triplet cancer cell line (Husain et al., 1998), while Abbott et al. (N: A, U, C, G and X: A, C, U), although GUC seems (1999) showed that restoration of BRCA1 in the to be the most ecient (Runer et al., 1990). Other BRCA1 negative HCC1937 human breast cancer cell factors are important in the choice of the triplet, line restores radioresistance (Tomlinson et al., 1998). including the accessibility of the triplet site according BRCA1 is also involved in controlling centrosome to the secondary and tertiary structure of the target duplication through a physical interaction with the mRNA and the G/C composition of the ¯anking centrosome (Xu et al., 1999). Moreover, Hsu and sequences which in¯uence the dissociation of the White (1998) have shown that BRCA1 interacts during duplex Rz/mRNA after cleavage. Several genes have mitosis with gamma-tubulin, a component essential for been inhibited by antisense Rz leading to modi®ed nucleation of microtubules. This suggests that BRCA1 phenotypes and resulting in therapeutic applications may be a regulator of mitotic-spindle assembly. (reviewed in Rossi, 1999; Vaish et al., 1998). Recently, BRCA1 has been linked to the c-Jun N- To study the function of BRCA1, we developed terminal kinase (JNK) pathway. Harkin et al. (1999) ribozymes to inhibit its expression and determined showed that enhanced BRCA1 expression led to the phenotypic changes created by the absence of the modulation of the expression of several genes, BRCA1 protein. Our ®rst goal was to check the including GADD45, and also to the activation of sensitivity of ribozyme-treated cells to chemotherapeu- JNK by phosphorylation, which correlated to an tic drugs to determine the role of BRCA1 in the increase in JNK-dependent apoptosis. Others showed response to drug treatments including genotoxics and that BRCA1 enhanced apoptosis through signalling mitotic spindle poisons. pathways involving H-Ras, MEKK-4, JNK, Fas/Fas ligand interaction and the activation of caspase-8 and 9 (Thangaraju et al., 2000). Results The hammerhead ribozyme (Rz) was initially described in satellite RNA of tobacco ringspot virus Design and efficiency of ribozymes in a cell-free assay and cleaves in cis in vivo (Forster and Symons, 1987). This capacity to cleave RNAs in a sequence-speci®c Before in vivo utilization, the activity of each Rz was manner is now used to inhibit gene expression by using checked in vitro against a synthetic RNA containing trans-acting Rz (Uhlenbeck, 1987). The capability of a the target site to con®rm the catalytic activity of the Rz to form a duplex with a mRNA, cleave it, and Rz. The eciency of cleavage is the quantity of target dissociate from the duplex in order to cleave an cleaved by the Rz. The results show that the Rz another molecule is an attractive strategy to inhibit BRCA1 ± 358 and 5282 were more ecient than

Figure 1 Synthesis of the target by the hammerhead ribozyme targeting the BRCA1 mRNA at the position 358 and its predicted secondary structure in association with the target mRNA. The T7 RNA polymerase promoter is in bold, restriction site in lower case, catalytic core in italic and the ¯anking sequence underlined

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6599 BRCA1 ± 4666 (Table 1). The low activity of BRCA1 ± tative real-time RT ± PCR showed that H100LX3.1, 4666 may be explained by the higher G/C composition H100LX3.2 and H100LX5.3 expressed less than 1% of of the ¯anking sequence. Expecting a good correlation the parental BRCA1 mRNA level, while H100LX3.3 between Rz activity in vitro and in cultured cells, we expressed 20% of the parental level (Figure 2). In a decided to keep the Rz BRCA1 ± 358 and 5282 and to second RNA extraction of H100LX3.3 after several clone these two Rz into the retroviral vector LXSN. weeks of culture, no BRCA1 transcript could be The sequences of our ribozymes were checked against detected; therefore we designated the two extractions the genome data base for possible cross-reaction to H100LX3.3.1 and H100LX3.3.2. other genes and no signi®cant homology to the Western-blotting of BRCA1 showed that no protein cleavage sites was detected. We also tested these could be detected for the four sub-clones (Figure 2). ribozymes on Brca1 mRNA, the murine homologue Even H100LX3.3.1, with a decrease of only 79% in the of BRCA1, and found that our ribozymes were unable amount of mRNA, had no detectable BRCA1 protein. to cleave the murine sequence, demonstrating their speci®city. Sensitivity to DNA damaging agents: etoposide and CDDP Down regulation of BRCA1 expression Untreated BRCA1-Rz cells showed enhanced prolifera- We screened six stable cell clones containing either tion compared to the negative controls HBL100 and LXSN BRCA1 ± 358 or LXSN BRCA1 ± 5282. We H100LXSN (Figure 3a), which is consistent with data studied further the four of them which showed the published by Rao et al. (1996). most signi®cant decrease in BRCA1 mRNA: HBL100 We tested the eect of two DNA damaging LXSN 358.1 ± 358.3 (hereafter referred as H100LX3.1 ± agents: CDDP (an intercalating agent) and etoposide 3.3) and HBL100 LXSN 5282.3 (H100LX5.3). Quanti- (an inhibitor of topoisomerase II) on cellular

Table 1 Summary of the composition of the Rz and their cleavage eciencies Position in the Cleavage efficiency Cleavage efficiency BRCA1 mRNA Exon Target sequence % G/C Tg/Rz* 1/5 (%) Tg/Rz 1/200 (%) 358 6 AUUUA GUC AACUUGU 21 46 97 4666 15 UGGGA GUC UUCAGAA 43 0 50 5282 19 ACCCA GUC UAUUAAA 29 53 96

The eciency is the percentage of target cleaved by the Rz in 90 min (* ratio of target to Rz of 1/5 or 1/200)

Figure 2 BRCA1 expression in controls and BRCA1Rz clones determined by real-time quantitative PCR, using 18S an internal control. All results are normalized to the HBL100 value. BRCA1 expression was also determined by Western blotting using the C- 20 antibody from Santa-Cruz Biotechnology and anti-actin A4700 from Sigma

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6600

Figure 3 Determination of sensitivity to drug treatment by the sulphorhodamine proliferation test. Cells were untreated (a)or treated for 24 h with 15 mM CDDP (b), 0.25 mM etoposide (c), 2.5 mM vincristine (d)or15mM taxol (e), then cultivated in fresh medium. The value at each time point was normalized to the T24 value (OD nh/OD 24 h)

proliferation. The proliferation of the BRCA1-Rz of BRCA1 leads to the activation of JNK by clones was inhibited by CDDP and etoposide phosphorylation. showing the sensitivity of the BRCA1 de®cient cells We performed real-time quantitative RT ± PCR on (Figure 3b,c). JNK1 and JNK2 to determine if reduction of the level of BRCA1 modi®ed the expression of JNKs (Figure 4). The mRNA expression of JNK1 was elevated in a Resistance to mitotic-spindle poison: taxol and vincristine range of 154 ± 266% in BRCA1-Rz clones. JNK2 Taxol and vincristine are microtubule-interfering expression was decreased following the same expression agents (MIA) which act respectively by inhibiting the pro®le as the BRCA1 mRNA. Notably, in extraction depolymerization of the mitotic spindle and by H100LX3.3.1 where the inhibition of BRCA1 was the inhibiting the polymerization of tubulin. For both least, there was little decrease in JNK2 mRNA, drugs, we observed chemoresistance in the BRCA1- whereas in the extraction H100LX3.3.2, BRCA1 ribozyme clones compared to controls (Figure 3d,e): mRNA was barely detectable and JNK2 fully re- the control cells were unable to grow, while the pressed. ribozyme-treated cells continued to proliferate. So inhibition of the expression of BRCA1 led to resistance Modulation of BRCA1, JNK1, 2, MEK-4, -7 and c-jun to mitotic-spindle poisons independently of their level expression during MIA treatment of action. To further investigate the MIA resistance of BRCA1- ribozyme treated cells, we treated H100LXSN and Status of JNK: quantification of JNK1, 2 mRNA H100LX3.1 cells with vincristine and quanti®ed the MIA cause cell death through the induction of G2/M mRNA levels of BRCA1, JNK1, 2, MEK-4, -7, which arrest and initiation of apoptosis via the activation of are kinases of JNK1, 2 and c-jun which is a down- the c-Jun N-terminal kinase/Stress-activated protein stream eector of JNK (Figure 5). kinase by several MAPkinasekinases (Lee et al., 1998; H100LXSN control cells exhibited decreased BRCA1 Wang et al., 1998b, 1999, 2000). As described expression after 24 h of vincristine treatment, which previously by Harkin et al. (1999) increased expression returned to basal level within 24 h after the end of

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6601

Figure 4 Status of JNK1 and JNK2 expression in controls and in BRCA1Rz clones determined by real-time quantitative PCR, using 18S an internal control. All results are normalized to the HBL100 value

Figure 5 Determination of mRNA level of BRCA1, JNK1, 2, MEK-4, -7 and c-jun genes by Real-Time quantitative PCR using of a reference gene (18S ribosomal RNA) as an internal control for PCR. Quanti®cations were done by the comparative Ct method and reported to the value of H100LXSN at T0. Treatment by 2.5 mM vincristine were done for 24 h between T0 and T24 and then cultivated in fresh medium

treatment. In the BRCA1-ribozyme treated cells, slightly in H100LX3.1. The expression pattern of JNK2 BRCA1 mRNA was undetectable throughout the was similar to JNK1 in H100LXSN control cells, but experiment. After treatment with vincristine, JNK1 there was no detectable expression of JNK2 in BRCA1- mRNA levels increased in H100LXSN and decreased ribozyme treated cells.

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6602 We obtained parallel results for MEK-4, -7, the the decreased expression of these genes re¯ects a kinases of JNK1, 2. For MEK-4, the basal level biological response to BRCA1 inhibition due to its expressed by the H100LX3.1 was increased twofold transcriptional functions. compared to H100LXSN. After vincristine treatment, The ®rst phenotype evaluated was cellular sensitivity MEK-4 was induced in control cells and suppressed in to dierent chemotherapeutic drugs. Two kinds of ribozyme-treated cells. Expression of MEK-7 was also drugs were used: DNA damaging agents and micro- induced in control cells, but like JNK2, was undetect- tubule-interfering agents. For the ®rst group, our able in H100LX3.1. results support a role for BRCA1 in the response to c-jun exhibited pro®les similar to that of JNK2: DNA damage and are consistent with the results of induction in control cells and repression in H100LX3.1. Husain et al. who showed that BRCA1 was over- expressed in CDDP-resistant MCF-7 cells and that inhibition of BRCA1 in this cell line restored sensitivity Induction by phosphorylation of c-jun by MIA treatment to CDDP (11): (Husain et al., 1998). The Rz-dependent We used H100LXSN cells and H100LX3.1 cells to sensitivity of HBL100 cells to etoposide and CDDP study the phosphorylation of c-jun after MIA treat- suggests that ribozymes targeting BRCA1 could be ment. c-jun is a downstream eector of the JNK used to sensitize tumour cells to certain DNA pathway, and becomes phosphorylated by JNK in damaging agents. This result must be investigated response to a variety of treatments. H100LXSN further in an animal model to evaluate the usefulness control cells exhibited a strong increase of p-c-jun after of BRCA1 ribozymes as a therapeutic alternative in 24 h of MIA treatment, which was maintained 24 h tumours resistant to some classic chemotherapeutic after treatment (Figure 6). In BRCA1-Rz cells, the treatments or to correlate BRCA1 expression level to basal level of p-c-jun was close to the control level, but chemosensitivity/resistance. induction by MIA was strongly attenuated and We also tested the microtubule-interfering agents delayed. This result showed the reduced capacity of taxol and vincristine. These are poisons of the mitotic BRCA1-de®cient cells to induce downstream factors of spindle and act by inhibiting either the depolymerization the JNK pathway. or the polymerization of tubulin. For both drugs, the ribozyme-treated cells were more resistant than the parental cells. Understanding the mechanism of this Discussion resistance may help us to understand and also predict cancer cases resistant to MIA treatment. JNK is involved Our two Rz targeted dierent regions of the BRCA1 in a signal transduction cascade inducing cell death in mRNA, and both had strong activities in vitro and response to defects in the mitotic spindle. Several studies after transfection into cells. The ®nding that two have shown that MIA activate this pathway, and lead to dierent Rz targeting respectively the 5' and the 3' the phosphorylation of proteins such as p53, c-jun or ends of the BRCA1 mRNA, and neither of which was Bcl2, and to the initiation of apoptosis. homologous to anything aside from BRCA1 in the How might BRCA1 be involved in MIA resistance? GCG database, resulted in identical chemosensitivity Previous studies indicate that BRCA1 is an upstream phenotypes demonstrates the speci®city of these Rz for regulator of JNK and is involved in mitotic-spindle the BRCA1 mRNA. The dierent levels of BRCA1 and assembly. Our results are consistent with all these JNK expression found for the two extractions of elements. The strikingly dierent eect on the tran- H100LX3.3 may be due to modi®cation of transgene scription of JNK1 and JNK2 suggest that these two expression. Several genes were strongly underexpressed homologous genes have dierent functions at the basal in Rz clones, and as the quantitative PCR was level and in response to MIA, although very little data performed using an internal 18S control in a multiplex is available to suggest what this dierence may be. The PCR, this cannot be explained by a defect in RNA JNK kinases, MEK-4 and MEK-7 also exhibited isolation, reverse transcription or in PCR. The dierent expression patterns. Here again, little is sequences of our ribozymes have been checked for known about the dierence between these two genes possible cross-reaction to other genes, no signi®cant and their involvement in a speci®c phosphorylation of homology to the cleavage site was detected. Therefore JNK. During MIA treatment, the control cells exhibited an increase in the mRNA levels of all four genes, whereas in H100LX3.1 a decrease of JNK1 and MEK-4 mRNA levels was observed, while JNK2 and MEK-7 mRNA remained undetectable. The transcriptional level may help us to distinguish a possible association between MEK-4-JNK1 and MEK-7-JNK2. Given the expression patterns in H100LX3.1, we may begin to develop an hypothesis on the resistance to MIA. It seems that BRCA1 induces elements of the Figure 6 Western blot of phosphorylated form c-jun after 24 h of vincristine treatment with anti-p-c-jun (KM-1) primary anti- JNK pathway, which initiate apoptosis after MIA body (Santa Cruz). Treatment by 2.5 mM vincristine were done for treatment. Thus, the BRCA1 de®cient cells failed to 24 h between T0 and T24 and then cultivated in fresh medium induce this pathway resulting in a resistance to MIA.

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6603 This hypothesis was con®rmed by analysis of induction the failure to induce this pathway resulting in the of c-jun phosphorylation by MIA in H100LX3.1 which uninitiation of apoptosis and to cell-cycle progression. was strongly attenuated and delayed compared to Resistance to MIA may also be explained by their control cells. decreased action on microtubules, for example by Control cells exhibited decreased BRCA1 expression modi®cations of drug withdrawal or bioavailability of after 24 h of vincristine treatment, which returned to MIA. Some of the proteins which may be involved in basal level within 24 h after. Similar results have been this pathway are MRP (multidrug resistance protein) found in studies of BRCA1-expressing breast, ovarian or P-glycoprotein, which act by reducing drug and prostate cancer cell lines with various cytotoxic accumulation. Microtubule-associated proteins (MAP) drugs (Andres et al., 1998; Fan et al., 1998a,b). In the are also linked with sensitivity to MIA. Expression of MCF-7 breast cancer cell line, down-regulation of MAP4 is known to be repressed by p53 but also BRCA1 and BRCA2 mRNA levels was detected after correlated with eciency of MIA. Overexpression of treatment with adriamycin, campothecin and the MIAs MAP4 increased microtubule polymerization and taxol and vincristine. In a prostate cancer cell line, sensitivity to Taxol, and also decreased microtubule treatment with adriamicin and UV decreased the binding, which increased sensitivity to vinca alkaloids mRNA levels of BRCA1 and BRCA2, as well as (Zhang et al., 1999). MAP2 may also potentiate the p300 and hRad51, although no eect was observed with eect of MIA; it has been shown that the level of MIAs in this study. BRCA1, which is involved in proteolysis of MAP2 modi®ed the eect of taxanes sensing DNA damages, transcription coupled DNA (Fromes et al., 1996). To date, no link between BRCA1 repair and also in promoting apoptosis via the JNK and microtubule-associated proteins has been estab- pathway, has the same expression pro®le after lished, but, like p53, BRCA1 may modulate MAP treatment by DNA damaging agents or by MIA. expression leading to a modi®cation of MIA eciency. Two hypotheses can explain how BRCA1 eects Two main conclusions can be drawn from our JNK pathway after MIA treatment. The ®rst involves results. First, the inhibition of BRCA1 expression leads the interaction of BRCA1 with the centrosome and to sensitivity to some cytotoxic drugs. Ribozyme speci®cally with gamma-tubulin, a component of the targeting BRCA1 may be useful to sensitize tumour mitotic spindle. The function of this interaction is cells to genotoxic drugs. These results have to be unclear, although Hsu and White (1998) suggest that con®rmed in vivo in an animal model. Secondly, a new BRCA1 may be a regulator of mitotic spindle assembly function may be ascribed to BRCA1: a role in sensing and of the G2/M checkpoint. Moreover, Xu et al. disarray in the mitotic spindle with transduction of this (1999) have shown that BRCA1 is involved in the signal through the JNK pathway. An element that control of centrosome duplication and chromosome merits further investigation is the dierent expression segregation. MIA act by inhibiting either the depoly- patterns of JNK1 and JNK2, MEK-4 and MEK-7,at merization or the polymerization of tubulin and then the basal level and after MIA treatment. The relation- the dynamic nature of the mitotic spindle is blocked. ship and the real functions of these four genes may The sensor which signals dysfunction of the mitotic help us to explain the resistance to MIA caused by spindle is unknown. Wang et al. (1999) suggest the BRCA1 inhibition. existence of a surveillance mechanism signalling the All these results may be useful in explaining functional integrity of microtubules to nuclear tran- resistance to MIA treatment in some breast cancer scription factors. The accumulated phenotypic and cases. Clinical studies may help us to determine the localization data suggest that BRCA1 is a good relationship between BRCA1 expression in breast candidate for this sensor position on the mitotic cancer tumour cells and eciency of MIA treatment. spindle to detect microtubular disarray and activate BRCA1 expression may be a predictive marker for the JNK cascade, leading to initiation of apoptosis. In chemosensitivity/resistance to treatment by MIA such this case BRCA1 may induce signal-transduction as Taxol. pathways indicating the disarray of the spindle leading to the arrest of the cell-cycle and the initiation of apoptosis. In the absence of BRCA1, such spindle disarray goes unsignalled and the cell cycle proceeds Materials and methods unchecked. This hypothesis is supported by the results Cell culture of Xu et al. (1999) who found that 30% of brca1D11/ D11 cells (lacking exon 11) contained multiple centro- The human breast cell line HBL100 was obtained from somes, leading to unequal chromosomal segregation ATCC and was maintained in McCoy's Medium containing and aneuploidy suggesting an important role for 10% fetal bovine serum, 1% glutamine and 1% penicillin- BRCA1 in centrosome duplication. streptomycin, under an atmosphere of 95% air, 5% CO2 at 378C. The second hypothesis involves BRCA1's function as a transcriptional co-activator implicated in transcription coupled DNA repair. Thangaraju et al. (2000) Design and construction of ribozymes and in vitro RNA targets have shown that BRCA1-dependent apoptosis used the To design the anti-BRCA1 ribozyme, we ®rst mapped the 56 JNK pathway and that taxol treatment induced this GUC triplets of the BRCA1 mRNA. We retained all the pathway. Here again, the lack of BRCA1 may result in sites not located in exon 11 (because of a splice variant

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6604 lacking this exon) and with a G/C composition of less than Cloning and transfection 45% in the two ¯anking sequences of seven nucleotides (for better dissociation of the Rz-mRNA duplex). Nine sites The Rz BRCA1 ± 358 and 5282 were cloned into the corresponded to these criteria. Finally, we looked at the retroviral vector LXSN using two unique restriction sites secondary structure of the target site using MFold (Jaeger et EcoRI and XhoI. The cloning of both Rz was veri®ed by al., 1989; Zuker, 1989) to determine the accessibility of the sequencing using two primers ¯anking the cloning site target for the Rz. This tool was helpful in choosing between (LXSNSeqF: 5'-CTCCGCCTCCTCTTCCTC-3' and LXSN- the dierent candidates. With all these data, we selected SeqR: 5'-CTAATTGAGATGCATGCTTTGC-3') done on an three Rz, one in the 5' region of the gene (position 358 in ABI prism 310 sequencer. Transfection of LXSN as a the mRNA) and two in the 3' region (positions 4666 and negative control, LXSNBRCA1 ± 358 and LXSNBRCA1 ± 5282). 5282 into cultured cells was performed using lipofectin1 (Life For each ribozyme, a pair of oligonucleotides were Technologies, Gibco ± BRL) according to the protocol synthesized, one containing the T7 RNA polymerase recommended: transfection of 26105 cells with 2 mgof promoter and a part of the sequence of the Rz (Figure 1). vector and 10 ml of lipofectin1 in medium without serum The two primers contained a complementary sequence of for 24 h, cultivation in complete medium for 72 h, followed 12 nts and also two restriction sites (EcoRI and XhoI) for by selection of clones with 0.8 mg/ml G418. After 3 weeks of cloning into the retroviral vector LXSN. PCR was performed selection, individual colonies were expanded and RNA and by mixing the primer pairs forming a hemiduplex (948C± protein were extracted for the analysis of BRCA1 expression. 30 s, 408C ± 30 s, 728C ± 1 min for 15 cycles). The sequences Total RNA extraction was done using Trizol1 (Life of the primers for the Rz are: BRCA1-358: 5'-GCTAAT- Technologies, Gibco ± BRL), then 1 mg of total RNA was ACGACTCACTATAGGGCgaattcACAAGTTCTGATGA-3'; reverse transcribed using First-Strand cDNA Synthesis kit 5'-ATctcgagATTTAGTTTCGTCCTCACGGACTCATCAG- (Pharmacia Biotech) with random hexamers. Protein extrac- AACTT-3'. BRCA1-4666: 5'-GCTAATACGACTCACTA- tion was performed using lysis buer (150 mM NaCl, 20 mM TAGGGCgaattcTTCTGAACTGATGA-3';5'-ATctcgag- Tris HCl (pH 8.0), 0.2 mM EDTA, 1 mM PMSF, 10% TGGGAGTTTCGTCCTCACGGACTCATCAGTTCAG-3'. glycerol, 1% Triton X-100) with protease inhibitors (com- BRCA1-5282: 5'-GCTAATACGACTCACTATAGGGCgaa- pleteTM, mini, EDTA-free, Roche). RNA and protein were ttcTTTAATACTGATGA-3';5'-ATctcgagACCCAGTTTC- quanti®ed by spectrophotometry. GTCCTCACGGACTCATCAGTATTA-3'.(T7RNApoly- merase promoter in bold, restriction sites in lower case, Quantitative RT ± PCR using real-time PCR TaqMan1 ¯anking sequences underlined and catalytic core in italics). technology The synthesis of DNA template for the RNA targets was performed with two primers ¯anking the GUC target site, The quanti®cation of transcription by Real-Time PCR takes with the forward primer containing the T7 RNA polymerase advantage of the 5' nuclease activity of Taq DNA promoter for in vitro transcription. The primers are (T7 RNA polymerase. The design of two primers ¯anking a labelled polymerase promoter in bold): for RNA target for ribozyme probe containing both a ¯uorescent reporter dye and a BRCA1-358: 5'-GCTAATACGACTCACTATAGGGCCAA- quencher dye was done with the Primer Express software (PE GGAACCTGTCTCCACAAAGTG-3';5'-CGTCTTTTGA- Biosystems). The Tm of the probe was 108C higher than that GGTTGTATCCGCTG-3' for RNA target for ribozymes of the primers for complete probe annealing. During the BRCA1-4666 and BRCA1-5282: 5'-GCTAATACGACTCAC- extension phase, the Taq DNA polymerase degrades the TATAGGGCGCGAAATCCAGAACAAAGCACATC-3'; probe, leading to the emission of the reporter dye once 5'-TTCTCTTGCTCGCTTTGGACC-3'. The PCR was per- separated from the eect of the quencher dye. The formed at 948C ± 30 s, 558C ± 30 s, 728 ± 1 min for 35 cycles. ¯uorescence emitted is directly proportional to the number of amplicons produced during each cycle. The cycle threshold (Ct) is the number of the cycles needed to reach the In vitro transcription ¯uorescence detection threshold and depends on the number The three Rz and the two RNA targets were synthesized by of DNA templates at the start of the PCR. The use of 18S T7 polymerase transcription using 1 mg of DNA template. ribosomal RNA as a reference gene provides an internal The reaction was carried out using the RiboMax kit control for PCR quanti®cation by the comparative Ct (Promega) (7.5 mM NTP, 80 mM HEPES/KOH pH 7.5, method. The PE Applied Biosystems 7700 Sequence Detector 20 mM MgCl2,2mM spermidine, 40 mM DTT and 600 units can discriminate dierent reporter dyes used for the test gene of T7 RNA polymerase) for 3 h at 378C. The DNA template and the reference gene, making possible the quanti®cation of was eliminated by RQ DNase (two units at 378C for 15 min). two genes simultaneously during a multiplex PCR. The The RNA transcripts were puri®ed by phenol/chloroform comparative Ct method subtracts the average 18S Ct value extraction, precipitated with isopropanol and quanti®ed by (each multiplex PCR is done in triplicate) from the average spectrophotometry. test gene Ct value, this result represents the DCt. This DCt is speci®c and can be compared to the DCt of a calibration sample (for example the control cell line HBL100). The In vitro cleavage reaction amount of target, determined by normalization to the The RNA target and the Rz were mixed at dierent molar endogenous reference (18S) and relative to the calibrator, is ratios (1/4, 1/10, 1/20, 1/40, 1/120, 1/200) in a reaction 27DDCt with DDCt the subtraction of the HBL100 DCt from mixture containing 50 mM Tris-HCl pH 8.0 and 30 mM the DCt of the sample, assuming that the eciency of both MgCl2 for 90 min at 378C. The reaction was stopped by PCRs is close to one. For all probes the quencher dye was 6 addition of 95% formamide, 0.05% bromophenol blue and carboxy-tetramethylrhodamine (TAMRA), reporter dye was 0.05% xylene cyanol, and heated at 658C for 5 min. RNA 6 carboxy ¯uorescein (FAM) for BRCA1 and VIC for 18S fragments were separated on 6% polyacrylamide gels for use in multiplex PCR. containing 8 M urea. The gel was stained with ethidium The expression of BRCA1 and 18S was quanti®ed by bromide and analysed using NIH Image 1.6 software. TaqMan technology (two primers and a ¯uorescent probe),

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6605 Table 2 Primers and probes for the real-time quanti®cation Genes Forward primers Reverse primers TaqMan probes

BRCA1 CAGAGGACAATGGCTTCCATG CTACACTGTCCAACACCCACTCTC AATTGGGCAGATGTGTGAGGCACCTG 18S CGGCTACCACATCCAAGGAA GCTGGAATTACCGCGGCT TGCTGGCACCAGACTTGCCCTC JNK1 CCAGGACTGCAGGAACGAGT CCACGTTTTCCTTGTAGCCC JNK2 ATGACCCCTTACGTGGTGACA CATGATGCAACCCACTGACC MEK-4 TGTGACTTCGGCATCAGTGG GCGCTTGGGTCTATTCTTTCA MEK-7 AGGTGTGGAAGATGCGCTTC ACCGGCCACGTCATTGCCG c-jun TCGACATGGAGTCCCAGGA GGCGATTCTCTCCAGCTTCC

For BRCA1 and 18S the TaqMan technology was used (two primers and one probe) whereas the other genes were quanti®ed by the use of SYBR green labelling (two primers). For the probes, the quencher dye was 6 carboxy-tetramethylrhodamine (TAMRA), reporter dye was 6 carboxy ¯uorescein (FAM) for BRCA1 and VIC for the 18S for use in multiplex PCR

whereas JNK1, 2, MEK-4, -7 and c-jun were quanti®ed by sensitive cells at T168). Cells (HBL100, HBL100 LXSN and using SYBRgreen. The measure of the ¯uorescence emitted HBL100 Rz clones) were plated at 5000 cells/well in 96-well by the association of SYBRgreen with the PCR products plates. Twenty-four hours after plating (T24), cells were quanti®ed the test gene ampli®cation. 18S reference ampli®- treated with drug for 24 h (T48) and after changing to fresh cation was performed in a parallel reaction. For sequence of medium cells were cultivated for 7 days. Quanti®cation of primers and probes see Table 2. proliferation was done at T24, T48, T96, T168 and T216 Ampli®cation was performed in 25 ml containing 12.5 mlof using the sulphorhodamine B proliferation test: cells were 26 MasterMix, 200 nM primers (and probe where appro- ®xed with 50% trichloroacetic acid for 1 h then washed with priate) for the target gene (50 nM for 18S) and 5 ng of cDNA water and stained with sulphorhodamine B (SRB) (Sigma), a (508C, 2 min for AmpErase uracil-N-glycosylase activation bright pink aminoxanthene dye which binds to basic amino (for TaqMan technology); 958C, 10 min for Amplitaq Gold acid residues. Then cells were washed with 1% acetic acid activation and 40 cycles 958C, 15 s; 608C, 1 min). and the SRB solubilized with 10 mM Tris HCl (Sigma). SRB was quanti®ed using a spectrophotometer at 530 nm. The background was measured at 640 n and subtracted from Western-blot analysis M the 530 nM value. All the determinations were normalized to Protein were separated on 4 ± 12% SDS ± PAGE and the T24 samples, which is the control of plating before transferred to ImmobilonP membrane (Millipore). Mem- treatment (OD nh/OD 24 h). For quantitative RT ± PCR and branes were blocked for 1 h in PBS-0.01% Tween with 5% determination of phosphorylated c-jun after vincristine nonfat dry milk and immunoblotted with 1/200 diluted treatment, 26106 cells were treated with 2.5 mM vincristine primary antibodies for 1 h: anti-BRCA1 (C-20), anti-p-c-jun for 24 h, washed in PBS and then cultivated in fresh medium (KM1) from Santa-Cruz Biotechnology or anti-actin A4700 for 24 h. RNA and protein extractions were done at T0 (Sigma). The secondary antibodies were anti-mouse or anti- which is the basal level (start of the treatment), T24 (end of rabbit IgG conjugated to horseradish peroxydase (Santa-Cruz the treatment), T48 (and T72 for RNA extraction). Biotechnology) used at 1/2000 dilution for 45 min and detected with the ECL chemiluminescent detection system (Amersham Pharmacia Biotech).

Acknowledgements Drug treatments and quantification of proliferation We thank N Uhrhammer for critically reading the manu- The four drugs tested were: 0.25 mM etoposide (VP16), 15 mM script. S Lafarge and V Sylvain were supported by doctoral taxol, 2.5 mM vincristine, all from Sigma, and 15 mM cisplatin fellowships from Association pour la Recherche contre le (CDDP) from Rhone Poulenc Roher. Taxol, vincristine, Cancer, Villejuif, Paris. This work was partly supported by etoposide were prepared in DMSO, CDDP was prepared by the Ligue Nationale Contre le Cancer and its departmental the supplier (doses corresponding to IC50 of the most committees of Puy de DoÃmeandCantal.

References

Abbott DW, Thompson ME, Robinson-Benion C, Tomlin- Chen JJ, Silver D, Cantor S, Livingston DM and Scully R. son G, Jensen RA and Holt JT. (1999). J. Biol. Chem., 274, (1999). Cancer Res., 59, 1752s ± 1756s. 18808 ± 18812. Chen Y, Farmer AA, Chen CF, Jones DC, Chen PL and Lee Anderson SF, Schlegel BP, Nakajima T, Wolpin ES and WH. (1996). Cancer Res., 56, 3168 ± 3172. Parvin JD. (1998). Nat. Genet., 19, 254 ± 256. Cortez D, Wang Y, Qin J and Elledge SJ. (1999). Science, Andres JL, Fan S, Turkel GJ, Wang JA, Twu NF, Yuan RQ, 286, 1162 ± 1166. Lamszus K, Goldberg ID and Rosen EM. (1998). Fan S, Twu NF, Wang JA, Yuan RQ, Andres J, Goldberg ID Oncogene, 16, 2229 ± 2241. and Rosen EM. (1998a). Int. J. Cancer, 77, 600 ± 609. Chapman MS and Verma IM. (1996). Nature, 382, 678 ± 679. Fan S, Wang JA, Yuan RQ, Ma YX, Meng Q, Erdos MR, Chen J. (2000). Cancer Res., 60, 5037 ± 5039. Brody LC, Goldberg ID and Rosen EM. (1998b). Chen J, Silver DP, Walpita D, Cantor SB, Gazdar AF, Oncogene, 16, 3069 ± 3082. Tomlinson G, Couch FJ, Weber BL, Ashley T, Livingston Forster AC and Symons RH. (1987). Cell, 49, 211 ± 220. DM and Scully R. (1998). Mol. Cell., 2, 317 ± 328.

Oncogene Inhibition of BRCA1 leads to resistance to microtubule-interfering agents S Lafarge et al 6606 Fromes Y, Gounon P, Veitia R, Bissery MC and Fellous A. Scully R, Chen J, Ochs RL, Keegan K, Hoekstra M, (1996). J. Protein Chem., 15, 377 ± 388. Feunteun J and Livingston DM. (1997b). Cell, 90, 425 ± Harkin DP, Bean JM, Miklos D, Song YH, Truong VB, 435. Englert C, Christians FC, Ellisen LW, Maheswaran S, Scully R, Chen J, Plug A, Xiao Y, Weaver D, Feunteun J, Oliner JD and Haber DA. (1999). Cell, 97, 575 ± 586. Ashley T and Livingston DM. (1997c). Cell, 88, 265 ± 275. Hsu LC and White RL. (1998). Proc. Natl. Acad. Sci. USA, Somasundaram K, Zhang H, Zeng YX, Houvras Y, Peng Y, 95, 12983 ± 12988. Wu GS, Licht JD, Weber BL and El-Deiry WS. (1997). Husain A, He G, Venkatraman ES and Spriggs DR. (1998). Nature, 389, 187 ± 190. Cancer Res., 58, 1120 ± 1123. Thangaraju M, Kaufmann SH and Couch FJ. (2000). J. Biol. Jaeger JA, Turner DH and Zuker M. (1989). Proc. Natl. Chem., 275, 33487 ± 33496. Acad. Sci. USA, 86, 7706 ± 7710. Tomlinson GE, Chen TT, Stastny VA, Virmani AK, Lee JS, Collins KM, Brown AL, Lee CH and Chung JH. Spillman MA, Tonk V, Blum JL, Schneider NR, Wistuba (2000). Nature, 404, 201 ± 204. II, Shay JW, Minna JD and Gazdar AF. (1998). Cancer Lee LF, Li G, Templeton DJ and Ting JP. (1998). J. Biol. Res., 58, 3237 ± 3242. Chem., 273, 28253 ± 28260. Uhlenbeck OC. (1987). Nature, 328, 596 ± 600. Li S, Chen PL, Subramanian T, Chinnadurai G, Tomlinson Vaish NK, Kore AR and Eckstein F. (1998). Nucleic Acids G, Osborne CK, Sharp ZD and Lee WH. (1999). J. Biol. Res., 26, 5237 ± 5242. Chem., 274, 11334 ± 11338. Wang Q, Zhang H, Kajino K and Greene MI. (1998a). Miki Y, Swensen J, Shattuck-Eidens D, Futreal PA, Harsh- Oncogene, 17, 1939 ± 1948. man K, Tavtigian S, Liu Q, Cochran C, Bennett LM, Ding Wang TH, Popp DM, Wang HS, Saitoh M, Mural JG, W, Bell R, Rosenthal J, Hussey C, Tran T, McClure M, Henley DC, Ichijo H and Wimalasena J. (1999). J. Biol. Frye C, Hattier T, Phelps R, Haugen-Strano A, Katcher Chem., 274, 8208 ± 8216. H, Yakumo K, Gholami Z, Shaer D, Stone S, Bayer S, Wang TH, Wang HS, Ichijo H, Giannakakou P, Foster JS, Wray C, Bogden R, Dayananth P, Ward J, Tonin P, Fojo T and Wimalasena J. (1998b). J. Biol. Chem., 273, Narod S, Bristow PK, Norris FH, Helvering L, Morrison 4928 ± 4936. P, Rosteck P, Lai M, Barrett JC, Lewis C, Neuhausen S, Wang TH, Wang HS and Soong YK. (2000). Cancer, 88, Cannon-Albright L, Goldgar D, Wiseman R, Kamb A and 2619 ± 2628. Skolnick MH. (1994). Science, 266, 66 ± 71. Welcsh PL, Owens KN and King MC. (2000). Trends Genet., Ouchi T, Monteiro AN, August A, Aaronson SA and 16, 69 ± 74. Hanafusa H. (1998). Proc. Natl. Acad. Sci. USA, 95, WongAK,OrmondePA,PeroR,ChenY,LianL,SaladaG, 2302 ± 2306. Berry S, Lawrence Q, Dayananth P, Ha P, Tavtigian SV, Pao GM, Janknecht R, Runer H, Hunter T and Verma IM. Teng DH and Bartel PL. (1998). Oncogene, 17, 2279 ± (2000). Proc. Natl. Acad. Sci. USA, 97, 1020 ± 1025. 2285. Rao VN, Shao N, Ahmad M and Reddy ES. (1996). XuX,WeaverZ,LinkeSP,LiC,GotayJ,WangXW,Harris Oncogene, 12, 523 ± 528. CC, Ried T and Deng CX. (1999). Mol. Cell., 3, 389 ± 395. Rossi JJ. (1999). Chem. Biol., 6, R33 ± R37. Yu X, Wu LC, Bowcock AM, Aronheim A and Baer R. Runer DE, Stomo GD and Uhlenbeck OC. (1990). (1998). J. Biol. Chem., 273, 25388 ± 25392. Biochemistry, 29, 10695 ± 10702. Zhang CC, Yang JM, Bash-Babula J, White E, Murphy M, Runer H and Verma IM. (1997). Prod. Natl. Acad. Sci. Levine AJ and Hait WN. (1999). Cancer Res., 59, 3663 ± USA, 94, 7138 ± 7143. 3670. Schlegel BP, Green VJ, Ladias JA and Parvin JD. (2000). Zhong Q, Chen CF, Li S, Chen Y, Wang CC, Xiao J, Chen Proc. Natl. Acad. Sci. USA, 97, 3148 ± 3153. PL, Sharp ZD and Lee WH. (1999). Science, 285, 747 ± Scully R, Anderson SF, Chao DM, Wei W, Ye L, Young RA, 750. Livingston DM and Parvin JD. (1997a). Proc. Natl. Acad. Zuker M. (1989). Science, 244, 48 ± 52. Sci. USA, 94, 5605 ± 5610.

Oncogene