The therapeutic potential of antitumor drugs is seriously limited by the manifestation of cellular drug resistance. We used the budding yeast Saccharomyces cerevisiae as a model system to identify novel mechanisms of resistance to one of the most active anticancer agents, cisplatin. We pinpointed NPR2 (nitrogen permease regulator 2) as a gene whose disruption conferred resistance to cisplatin. In addition, we observed a 4-fold cross-resistance of yeast npr2Δ cells (i.e., cells from which the NPR2 gene had been disrupted) to the anticancer drug doxorubicin, in combination with hypersensitivity to cadmium chloride. Furthermore, npr2Δ cells displayed unaltered cellular cisplatin and doxorubicin accumulation and showed an enhanced rate of spontaneous mutation compared with the isogenic parent. These data indicate that the npr2Δ phenotype overlaps that of the sky1Δ cells that we characterized previously (Mol Pharmacol 61:659–666, 2002). Therefore, we generated yeast npr2Δ sky1Δ double-knockout cells and performed clonogenic survival assays for cisplatin and doxorubicin, which revealed that NPR2 and SKY1 (SR-protein-specific kinase from budding yeast) are epistatic. The double-knockout strain was just as resistant to cisplatin and doxorubicin as the single-knockout strain that was most resistant to either drug. In conclusion, we identified NPR2 as a novel component involved in cell kill provoked by cisplatin and doxorubicin, and our data support the hypothesis that NPR2 and SKY1 may use mutual regulatory routes to mediate the cytotoxicity of these anticancer drugs.
The overall results of treatment of metastasized tumors with anticancer drugs are rather disappointing, because of intrinsic and acquired drug resistance. One of the most commonly applied anticancer agents is cisplatin, which forms platinum-DNA adducts that play an important role in the cytotoxicity of the drug. Platinum-based chemotherapy is curative for most patients with advanced testicular cancer (Einhorn, 1997) and active against ovarian, bladder, cervical, head and neck, and lung cancers (Perez, 1998). Regrettably, cellular resistance to cisplatin, either at the onset of treatment or at relapse, is frequently encountered and seriously limits its therapeutic potential (Perez, 1998; Niedner et al., 2001). In vitro studies have revealed numerous resistance mechanisms, including reduced intracellular accumulation, detoxification by glutathione or metallothioneins (resulting in decreased DNA platination), alterations in DNA repair, increased tolerance, and aberrations in pathways modulating programmed cell death (Perez, 1998; Niedner et al., 2001). Unfortunately, the mechanisms identified so far do not satisfactorily explain the unresponsiveness of particular tumors to platinum-based chemotherapy observed in the clinic. Therefore, additional cisplatin resistance pathways (for which the genes involved still need to be pinpointed) may exist.
We are using the budding yeast Saccharomyces cerevisiae as a model system for drug sensitivity and resistance in higher eukaryotes. Over the years, yeast has proven its value as an accessible tool to gain a better understanding of complicated and often integrated mechanisms such as recombination, DNA repair, and checkpoint control in mammalian cells. Human genetic defects can regularly be addressed directly in yeast because of the evolutionary conservation of genes and signal transduction pathways (Resnick and Cox, 2000). With regard to the cellular response to anticancer agents, there are many fine examples that illustrate the power of yeast as a model organism for mammalian cells. To mention a few, the Saccharomyces cerevisiae pleiotropic drug resistance network comprises homologs to the human ATP-binding cassette-type transporters P-glycoprotein and MRP and is an important laboratory model for the clinical problem of multidrug resistance (Kolaczowska and Goffeau, 1999). The yeast nucleotide excision repair (NER) system has been instrumental in the study of NER in mammalian cells (Prakash and Prakash, 2000; Resnick and Cox, 2000). It has long been known that S. cerevisiae NER mutants are hypersensitive to cisplatin (Abe et al., 1994). In line with this, the involvement of NER in cisplatin resistance in patient-derived material has been observed repeatedly (Crul et al., 1997), whereas decreased NER may contribute to the high cisplatin sensitivity of testis tumors (Köberle et al., 1999). The predictive value of yeast in the study of cellular drug resistance in humans is also illustrated by our own research. In previous studies (Schenk et al., 2001; Schenk et al., 2002), we identified SKY1 (SR-protein-specific kinase from budding yeast) as a cisplatin sensitivity gene. Its abrogation conferred cisplatin resistance in yeast, and down-regulation of its human homolog SRPK1 (SR-protein-specific kinase) led to cisplatin resistance in an ovarian carcinoma cell line (Schenk et al., 2001). Importantly, we recently monitored SRPK1 protein expression in a series of testicular germ cell tumors, and found that the expression in tumors from chemorefractory patients was significantly lower than in tumors from patients responding to platinum-based chemotherapy (P. W. Schenk H. Stoop, F. Mayer, C. Bokemeyer, G. Stoter, J. W. Oosterhuis, L. H. J. Looijenga, and K. Nooter, manuscript in preparation1). This indicates that the identification of SKY1 as a cisplatin sensitivity gene in S. cerevisiae may ultimately be of importance to predict the chemoresponsiveness of particular tumors in the clinic.
Our present data indicate that NPR2 (nitrogen permease regulator 2) is a novel yeast drug sensitivity gene whose abrogation leads to resistance not only to cisplatin but also to another anticancer agent, doxorubicin. The overall phenotype that we show for S. cerevisiae npr2Δ cells (i.e., cells from which the NPR2 gene had been disrupted) is highly reminiscent of our previous data for sky1Δ cells (Schenk et al., 2002). Therefore, we generated npr2Δ sky1Δ double-knockout cells and found that NPR2 and SKY1 are epistatic. The cisplatin and doxorubicin resistance of the double-knockout cells was neither additive nor synergistic compared with the single-knockout cells, suggesting that NPR2 may act in mutual regulatory routes with SKY1.
Materials and Methods
Chemicals. Yeast nitrogen base (YNB), Bacto-agar and yeast extract/peptone/dextrose (YPD) broth were obtained from DIFCO Laboratories (Detroit, MI). Cisplatin [cis-diamminedichloroplatinum(II) (Platinol; Platosin)] and doxorubicin hydrochloride (Adriamycin; Doxorubin) were purchased from Pharmachemie (Haarlem, The Netherlands). Oxaliplatin [[trans-(l)-1,2-diaminocyclohexane]oxalatoplatinum(II) (Eloxatin)] was obtained from Sanofi-Synthelabo (Maassluis, The Netherlands). Etoposide (Vepesid) was from Bristol-Myers Squibb (Woerden, The Netherlands), 5-fluorouracil (Fluorouracil-TEVA; Adrucil) from TEVA Pharma (Mijdrecht, The Netherlands), and copper sulfate from Merck (Darmstadt, Germany). Other chemicals including amino acids and additional cytotoxic agents were purchased from Sigma-Aldrich (Zwijndrecht, The Netherlands).
Yeast Strains and Culture Conditions.S. cerevisiae strains used in this study are listed in Table 1. All yeast strains were routinely maintained on selective synthetic YNB medium [0.67% YNB and 2% d-(+)-glucose]. Where appropriate, media were solidified by the addition of 2% Bacto-agar and supplemented with amino acids depending on the auxotrophic requirements. S. cerevisiae was grown at 30°C under vigorous shaking for liquid cultures. Escherichia coli strains MC1061 and DH5α were used as bacterial hosts for plasmids.
Yeast Random Transposon Insertion Library, Screening Strategy, and Characterization of Cisplatin-Resistant Strains. A yeast genomic mini-Tn3::lacZ::LEU2 transposon insertion library (a kind gift from Dr. P. B. Ross-Macdonald) (Burns et al., 1994) was amplified in E. coli strain MC1061 and transformed into leucine-deficient yeast cells to randomly disrupt genes in the S. cerevisiae genome as described previously (Schenk et al., 2001). Homologous recombination between yeast DNA transposon flanking sequences and endogenous genomic sequences is thought to yield transformants in which the original genomic copy has been replaced by the mutagenized version (Rothstein, 1991). NER-deficient rad4Δ yeast strain MGSC131 (Verhage et al., 1996) was used as recipient, because it displays a steep dose-response curve to cisplatin. A total of 3 × 105 leucine-proficient S. cerevisiae MGSC131 transformants were replated at a density of 104 cells per 94-mm dish on selective YNB containing 4 μg/ml cisplatin, and colonies surviving this one-step drug selection were picked and retested in semiquantitative spot assays and quantitative clonogenic survival assays as described previously (Schenk et al., 2001). Upon confirmation of resistance phenotypes, the number of transposons per yeast strain was determined by Southern blotting. For the strains containing single insertions, inverse PCR (Ochman et al., 1988) employing BstYI and outward-directed primers complementary to mini-Tn3::lacZ::LEU2 was directly used to obtain sequences flanking the transposon elements. Suitable PCR products were purified and sequenced using transposon-specific primers, and sequences were finally analyzed employing public databases (Schenk et al., 2001).
Plasmid Construction and Transformation. The low-copy yeast expression vector pYCTEF111 containing the LEU2 auxotrophic marker was obtained by ligating the 5.8-kb PvuII backbone fragment from centromeric plasmid YCplac111 (Gietz and Sugino, 1988) to the 1.0-kb PvuII fragment from pYCTEF (Schenk et al., 2001), containing the constitutive translation elongation factor 1α promoter. A 1.9-kb NPR2 PCR product was obtained employing primers 5′NPR2-S (5′-GAC TAG CCC GGG CTC TAC TAA AGG GAA TGG TCA G-3′) and 3′NPR2-S (5′-GAC TTG AGT CGA CGA ATT TCT CTA ATT TTA ACT CAG C-3′). This fragment was restricted with SmaI and SalI and cloned into SmaI/SalI-digested pYCTEF111, yielding pYCTEF111-NPR2. After propagation in E. coli subcloning efficiency DH5α competent cells (Invitrogen, Breda, The Netherlands), the NPR2 expression plasmid and the empty vector were transformed into the appropriate yeast strains.
Cytotoxicity Assays. Sensitivity to cytotoxic chemicals was determined by a semiquantitative spot assay as described previously (Burger et al., 2000). Briefly, S. cerevisiae cells from freshly streaked plates were inoculated and grown to mid-log phase in selective liquid YNB medium. Different aliquots of cells (as indicated in the figure legends) were then spotted onto selective YNB plates containing various compound concentrations and incubated at 30°C for 3 to 4 days. Sensitivity to ionizing radiation was tested by a similar assay, during which spotting of the yeast cells was immediately followed by irradiation with increasing doses of γ-rays using opposing 137Cs sources (Gamma Cell 40; Atomic Energy of Canada, Ottawa, Canada) at a rate of 1.06 to 1.08 Gy/min. Sensitivity to ultraviolet light was also tested by a semiquantitative spot assay as described previously (Burger et al., 2000), using a germicidal lamp at 254 nm. Sensitivity to cisplatin and doxorubicin was analyzed in detail by a quantitative clonogenic survival assay. Serial dilutions of mid-log phase yeast cells were plated onto selective YNB containing various drug concentrations. Plates were incubated at 30°C for 3 to 4 days, colonies were counted, and percentage survival was calculated based on the number of colonies arising in the absence of cytotoxic agents (Burger et al., 2000; Schenk et al., 2002). Where appropriate, the slopes of the log-linear concentration-survival curves were determined by linear regression analysis using SPSS 10.1 (SPSS Inc., Chicago, IL).
Determination of Cellular Platinum and Doxorubicin Accumulation. Cellular platinum content was determined by atomic absorption spectrometry (AAS) using a flameless spectrometer (type 4110 ZL; PerkinElmer Analytical Instruments, Shelton, CT) as described previously (Burger et al., 2000). A total of 2 × 107 exponentially growing yeast cells were incubated in 5 ml of selective liquid YNB medium containing various cisplatin concentrations for 18 h. After drug exposure, cells were immediately washed three times, and 3 × 107 cells were pelleted and completely lysed in 100 μl of chloroform. After evaporation of residual chloroform, samples were dissolved in 0.2% nitric acid and subjected to AAS to determine total platinum content. Cellular doxorubicin content was measured by high-performance liquid chromatography (HPLC) as described previously (de Bruijn et al., 1999). A total of 8 × 106 exponentially growing yeast cells were incubated in 2 ml of selective liquid YNB medium per Vacutainer tube with Hemogard closure (BD Clinical Laboratory Solutions, Franklin Lakes, NJ) containing various doxorubicin concentrations for 18 h. After drug exposure, cells were immediately washed three times in ice-cold medium, and 1.5 × 107 cells were pelleted and dissolved in drug-free human plasma. Samples were finally subjected to pretreatment and HPLC as described previously (de Bruijn et al., 1999) to determine total doxorubicin content.
Determination of Mutation Rates. The frequency of spontaneous mutations was assessed by a forward mutation rate assay that detects genetic alterations inactivating the arginine permease gene (Tishkoff et al., 1997) [i.e., conversion of the CAN1+ to the can1 (canavanine resistance) mutant phenotype]. A total of 11 parallel cultures per strain were inoculated at low density from freshly streaked plates and grown to stationary phase in liquid YPD broth at 30°C. The viable titers and numbers of canavanine-resistant mutants were then determined by plating different aliquots on YPD broth and synthetic medium containing 40 μg/ml canavanine sulfate, respectively. Mutation rates were finally calculated from the resulting median mutant frequencies by iteration according to Drake (1991).
Generation of npr2Δ sky1Δ Cells. A SKY1 disruption was generated in npr2Δ and BY4742 cells using a sky1Δ::lox-URA3-lox geneblaster (i.e., a 5′ SKY1 flank-lox-URA3-lox-3′ SKY1 flank fragment). First, the flanking sequences of the SKY1 gene were amplified separately from yeast genomic DNA by PCR, employing primers dis1 (5′-CCG AAG CCA TTG TAG GGG AG-3′) and dis2 (5′-CAT GGT GAC CAA TTA TTT CTC AGC GCC AGG TG-3′) for the 5′-flank and dis3 (5′-CAT GGT TAC CTT TAT TTT GCC CTT GCC TTT T-3′) and dis4 (5′-GCC ACA ACG GTC GCA AAG TC-3′) for the 3′-flank. BstEII-specific bases are indicated in bold. The resulting products, containing different BstEII sites, were then digested with BstEII and ligated to a 1.3-kb BstEII fragment containing a URA3 cassette flanked by loxP sequences (Jansen et al., 2002). Finally, the entire geneblaster was amplified by PCR using primers dis1 and dis4 and transformed into S. cerevisiae npr2Δ and BY4742 cells to obtain npr2Δ sky1Δ and BY-sky1Δ cells, respectively. In the resulting yeast cells, SKY1 sequences were deleted from -72 to + 2242 relative to the 2229-bp open reading frame. Disruptions of NPR2 and/or the SKY1 gene in npr2Δ, npr2Δ sky1Δ, and BY-sky1Δ cells were checked by PCR.
Isolation and Characterization of Cisplatin-Resistant Yeast Strains. To identify yeast cisplatin sensitivity genes (whose disruption confers resistance to the drug), we transformed haploid, NER-deficient, cisplatin-sensitive S. cerevisiae MGSC131 cells with a yeast genomic mini-Tn3::lacZ::LEU2 transposon insertion library (Burns et al., 1994) and screened 3 × 105 transformants for cisplatin resistance as described previously (Schenk et al., 2001). In line with the low level of acquired resistance (1.5- to 3-fold) emerging in tumor samples or cell lines obtained before and after cisplatin treatment of patients (Andrews et al., 1990; Niedner et al., 2001), we selected nine library-derived yeast strains that were 2- to 4-fold cisplatin-resistant. These strains were characterized further. The sites flanking the transposon elements (i.e., the loci that had been disrupted in these strains) were identified by means of inverse PCR (Ochman et al., 1988) and sequencing, followed by comparison with public databases. As described previously (Schenk et al., 2001), the YMR216C locus corresponding to the SKY1 gene had been disrupted in five transposon-containing yeast strains, which were all about 4-fold cisplatin-resistant compared with isogenic SKY1+ cells.
In one of the remaining library-derived strains, the single transposon insertion was located between the YEL063C and the YEL062W locus corresponding to the CAN1 and the NPR2 gene, respectively. In this region, CAN1 and NPR2 have a common 5′-untranscribed sequence that may represent a shared functional regulatory segment (Rousselet et al., 1995). Because the original MGSC131 strain carries a can1-100 mutation (Verhage et al., 1996), it seems very unlikely that the cisplatin resistance in the transposon containing strain would have resulted from deregulation of CAN1 instead of the NPR2 gene. The Npr2p gene product is thought to be a regulatory protein for nitrogen permeases, which may act at both the transcriptional and post-transcriptional levels (Rousselet et al., 1995). We tested the transposon-derived mutant in a clonogenic survival assay (Fig. 1A) and found that it was 2-fold cisplatin-resistant compared with untransformed parental MGSC131 cells.
Characterization of Independent Mutant Cells. To exclude that the observed cisplatin-resistant phenotype of the transposon containing yeast strain might have arisen from unrelated mutations induced during the screening procedure (i.e., cisplatin-induced mutations instead of library-derived NPR2 disruption leading to resistance) or that NER deficiency might be involved, we obtained an independent disruption mutant. S. cerevisiae npr2Δ cells were made by disruption of the NPR2 gene in the repair-proficient RAD+ strain BY4742, which also contains an intact CAN1/NPR2 untranscribed regulatory segment and CAN1 coding region (Table 1). Analogous to the original transposon-containing strain, npr2Δ cells were also cisplatin-resistant compared with isogenic BY4742 cells (Fig. 1B). In addition, we tested independent BUH3 npr2 cells that had been generated in a different repair-proficient, CAN1+ background (Table 1). These npr2 mutant cells (Rousselet et al., 1995) were also cisplatin-resistant compared with their isogenic parent CMY375, with a similar level of resistance (Fig. 1C). The data thus clearly indicate that the observed cisplatin resistance was linked to disruption of the S. cerevisiae NPR2 gene, without requirement for NER deficiency or deregulation of the CAN1 gene.
Cross-Resistance of Yeast Cells Containing a Disrupted NPR2 Gene. To learn more about the mechanisms that may underlie the cisplatin resistance phenotype and to determine its specificity, we monitored npr2Δ mutants for cross-resistance to other cytotoxic agents. Semiquantitative spot assays were performed using a range of different classes of chemicals (mostly anticancer drugs), including the cisplatin analog oxaliplatin, heavy metals (cadmium chloride, zinc chloride, and copper sulfate), the methylating agent N-methyl-N′-nitro-N-nitrosoguanidine (MNNG), the antimetabolite 5-fluorouracil, and topoisomerase II inhibitors (doxorubicin and etoposide). In addition, we assessed possible cross-resistance to ionizing radiation and ultraviolet light. Interestingly, disruption of NPR2 conferred resistance not only to cisplatin but also to the anthracycline doxorubicin (Table 2; Fig. 2), with a level of resistance in the 4-fold range. However, sensitivity toward the cisplatin analog oxaliplatin and another topoisomerase II inhibitor, etoposide, was unaltered. Moreover, we did not observe changes in sensitivity to zinc chloride, copper sulfate, MNNG, 5-fluorouracil, ionizing radiation, or ultraviolet light either. Notably, npr2Δ mutants were hypersensitive toward cadmium chloride. In Fig. 2, representative examples of the spot assays performed are shown.
Complementation of Yeast npr2Δ Cells with an Expression Plasmid Harboring NPR2. To further confirm that NPR2 is involved in the cytotoxicity induced by anticancer agents, its coding region was cloned into the low-copy S. cerevisiae expression vector pYCTEF111. The resulting plasmid was transformed into npr2Δ cells, and the cisplatin and doxorubicin sensitivity of the transformed strain was determined versus the appropriate controls. Upon transformation of plasmid pYCTEF111-NPR2 (harboring NPR2 under the constitutive translation elongation factor 1α promoter) into npr2Δ cells, the cisplatin- and doxorubicin-resistant phenotype was fully complemented. The cells became as sensitive to cisplatin (Fig. 3A) and doxorubicin (Fig. 3B) as strain BY4742 transformed with the empty vector, whereas transformation with pYCTEF111 alone left the npr2Δ cells clearly drug-resistant. These data affirm that the observed cisplatin and doxorubicin resistance was linked to disruption of the NPR2 gene, and indicate that the corresponding gene product may be actively involved in the cytotoxicity of these drugs (i.e., that NPR2 is a cisplatin and doxorubicin sensitivity gene).
Cisplatin and Doxorubicin Sensitivity of Yeast Cells Containing a Disrupted DUR3 Gene. The most likely downstream target of NPR2 is the S. cerevisiae DUR3 gene, encoding a putative transmembrane component required for active transport of urea (ElBerry et al., 1993). Rousselet et al. (1995) showed that the level of DUR3 mRNA in BUH3 npr2 cells is strongly increased, indicating that NPR2 might code for a transcriptional regulator of DUR3 (see Discussion). Furthermore, several lines of evidence suggest that the Npr2p protein is also involved in the regulation of the nitrogen permease Dur3p at the post-transcriptional level (Rousselet et al., 1995). Therefore, we initially hypothesized that NPR2 might mediate cisplatin- and doxorubicin-induced cell kill through a DUR3-dependent pathway. To assess the possible role of DUR3 in anticancer drug resistance, we obtained dur3Δ disruption mutant cells derived from strain BY4742 (Table 1) and determined their sensitivity to cisplatin and doxorubicin. In contrast to npr2Δ cells, dur3Δ cells did not show altered drug sensitivities, compared with their isogenic parent (Fig. 4). This finding implies that DUR3 is not involved in the cytotoxic action of cisplatin or doxorubicin.
Platinum and Doxorubicin Accumulation in npr2Δ Mutants. To determine whether the resistant phenotype of npr2Δ cells might be linked to impaired drug accumulation, we monitored cellular platinum and doxorubicin content by AAS (Burger et al., 2000) and HPLC (de Bruijn et al., 1999), respectively. For both agents, we found a clear dose-effect relationship between drug exposure and cellular accumulation. However, S. cerevisiae npr2Δ cells did not display reduced platinum or doxorubicin accumulation compared with isogenic NPR2+ cells (Fig. 5). Therefore, the observed cisplatin and doxorubicin resistance can probably not be explained by decreased drug import or increased drug export.
Yeast npr2Δ Cells Display a Mutator Phenotype. Interestingly, the pattern of drug resistance and hypersensitivity that we found for S. cerevisiae npr2Δ cells is highly reminiscent of our previous data for sky1Δ cells (Schenk et al., 2002). Sky1Δ cells also displayed cross-resistance to cisplatin (but not oxaliplatin) and doxorubicin, in combination with hypersensitivity to cadmium chloride. In addition, the cisplatin- and doxorubicin-resistant phenotype of sky1Δ cells was not linked to impaired drug accumulation either. NPR2 and SKY1 might thus be involved in common cellular pathways. As demonstrated before (Schenk et al., 2002), sky1Δ cells also showed a mutator phenotype (i.e., a 2.4-fold enhanced rate of spontaneous mutation compared with their isogenic parent). To further explore a possible relationship between NPR2-and SKY1-dependent cellular processes, we determined whether npr2Δ cells also display this phenotype. The rate of forward mutation at the CAN1 locus was thus assessed for npr2Δ cells and isogenic BY4742 NPR2+ cells. There was a 2-fold increase of the mutation rate per replication in the npr2Δ strain (rate, 4.8 × 10-8) versus the NPR2+ strain (rate, 2.5 × 10-8), as shown in Fig. 6. These data indicate that, like disruption of the SKY1 gene, loss of NPR2 induces a mutator phenotype.
Cisplatin and Doxorubicin Resistance of npr2Δ sky1Δ Cells. To further explore whether NPR2 and SKY1 might play a role in common pathways, we generated a SKY1 disruption in npr2Δ and BY4742 parental cells, to obtain an npr2Δ sky1Δ double-knockout strain and its isogenic control BY-sky1Δ, respectively. To determine whether NPR2 and SKY1 are epistatic, we tested the npr2Δ sky1Δ cells for their sensitivity toward cisplatin and doxorubicin. Because the double knockout strain showed diminished growth in the absence of cytotoxic agents, we assessed its possible drug resistance in semiquantitative spot assays (Fig. 7, A and B) as well as in quantitative clonogenic assays (Fig. 7, C and D). From our experiments, it is clear that neither the cisplatin nor the doxorubicin resistance displayed by the npr2Δ sky1Δ cells was additive or synergistic, compared with the single-knockout strains. We conclude that the double-knockout cells were as resistant to cisplatin as the single-knockout cell that was most resistant to this drug (i.e., strain BY-sky1Δ) (Fig. 7, A and C) and as resistant to doxorubicin as the npr2Δ strain, which was the most resistant single-knockout strain for the anthracycline (Fig. 7, B and D). In summary, our data indicate that NPR2 and SKY1 are epistatic. In the case of parallel pathways, the drug resistance of the double knockout strain would have been additive or greater than additive. Given the phenotype observed, NPR2 may thus act in mutual regulatory routes with SKY1 to mediate the cytotoxicity of anticancer agents.
In the present study, we aimed to identify and characterize novel cisplatin sensitivity genes whose disruption renders S. cerevisiae cisplatin-resistant. We found that different yeast strains lacking a functional NPR2 gene were all cisplatin-resistant compared with their isogenic parents. In cross-resistance studies, npr2Δ cells were resistant to cisplatin and the anthracycline drug doxorubicin and hypersensitive to cadmium chloride. Whereas Rousselet et al. (1995) previously reported that npr2 mutants are affected in their urea and proline transport capacities, the observed cisplatin and doxorubicin resistance is probably not a direct result of decreased drug import or increased drug export. Npr2Δ cells did not display reduced platinum accumulation compared with isogenic NPR2+ cells. This seems to exclude a role for the high-affinity copper transporter Ctr1p, which was very recently identified as a regulator of cisplatin uptake and sensitivity (Ishida et al., 2002; Lin et al., 2002). In line with this, npr2Δ cells did not show altered sensitivity to copper sulfate. Although the combination of cisplatin resistance and hypersensitivity to cadmium chloride may be poorly understood, it is not unprecedented in either yeast (Perego et al., 1997) or mammalian cells (Farnworth et al., 1990), as discussed previously (Schenk et al., 2002). Our finding that NPR2 disruption led to a 4-fold doxorubicin resistance is an important observation, as unresponsiveness to anthracycline treatment constitutes a tremendous clinical problem (Monneret, 2001). Doxorubicin is thought to inhibit topoisomerase II through DNA intercalation (Pratt et al., 1994). Topoisomerase II is, however, probably not involved in the doxorubicin resistance of npr2Δ cells, because we did not see cross-resistance to etoposide, another well known inhibitor of this enzyme (Pratt et al., 1994). Because the observed resistance cannot immediately be explained by altered doxorubicin accumulation either, a role for the yeast pleiotropic drug resistance network involved in anthracycline efflux (Kolaczowska and Goffeau, 1999) seems unlikely as well. Because DUR3 and its gene product have been regarded as the most likely downstream targets of NPR2 (Rousselet et al., 1995), we assessed a possible role of this putative transmembrane active urea transporter (ElBerry et al., 1993) in anticancer drug resistance. In contrast to yeast npr2Δ cells, dur3Δ cells were neither resistant nor hypersensitive to cisplatin or doxorubicin compared with their isogenic parent. Our data thus indicate that NPR2 exerts its cytotoxic effects in response to either cisplatin or doxorubicin independent of the nitrogen permease DUR3. Although we can not dismiss a tentative role of nitrogen transport in the drug-resistant phenotype of the npr2Δ cells, this suggests that a novel function of NPR2 (unrelated to nitrogen transport) might be involved.
A possible mechanistic lead can be obtained by evaluation of the overall phenotypes induced by disruption of NPR2 and SKY1. The pattern of cross-resistance to cisplatin (but not oxaliplatin) and doxorubicin, in combination with hypersensitivity to cadmium chloride in npr2Δ cells, is highly reminiscent of our previous data for sky1Δ cells (Schenk et al., 2002). In addition, the npr2Δ strain did not show reduced platinum or doxorubicin accumulation and displayed an enhanced rate of spontaneous mutation compared with the isogenic parent, similar to the sky1Δ strain that we characterized earlier (Schenk et al., 2002). We tested npr2Δ sky1Δ double-knockout cells for cisplatin and doxorubicin resistance, and concluded that NPR2 and SKY1 are epistatic. These data suggest that NPR2 may act in mutual regulatory routes with SKY1. The Sky1p protein is believed to phosphorylate several serine-rich proteins, including its well established in vivo S. cerevisiae substrate Npl3p (Yun and Fu, 2000). Although Rousselet et al. (1995) correctly noted that Npr2p (SWISS-PROT accession number P39923) is a serine-rich protein as well, it does not comprise a true consensus site for serine phosphorylation by Sky1p (Yun and Fu, 2000). It is, therefore, not very likely that Sky1p might directly regulate Npr2p through straightforward phosphorylation. According to the Saccharomyces Genome Database (http://www.yeastgenome.org/), Npr2p may be a transcription factor, as inferred from electronic annotation. It is thus conceivable that Npr2p functions as a transcriptional modulator for specific downstream components within a tentative SKY1 network. Although we can not dismiss a possible role of Npr2p in the transcriptional regulation of SKY1, the transcription of SKY1 is probably not regulated by cisplatin. In previous experiments, we did not detect alterations in SKY1 RNA levels upon cisplatin treatment of yeast cells (Schenk et al., 2001). In fact, our data imply that NPR2 and SKY1 are connected in a shared regulatory network instead of just a single hierarchic cascade. Because npr2Δ cells were more resistant to doxorubicin, whereas BY-sky1Δ cells were more resistant to cisplatin, it seems unlikely that one component would simply act directly upstream from the other to mediate the same response (i.e., cell death) to different stimuli (i.e., cisplatin and doxorubicin). Based on the drug sensitivity profile and mutator phenotype of S. cerevisiae sky1Δ cells, we previously proposed that Sky1p might play a significant role in mismatch repair (MMR), base excision repair, and/or Rev3p-dependent pathways (Schenk et al., 2002). By analogy, Npr2p might also be involved in such processes. Because cisplatin and doxorubicin resistance (Drummond et al., 1996; Durant et al., 1999) and an elevated mutation rate (Branch et al., 1995) have been associated with loss of MMR in several cell types, MMR deficiency could, for instance, underlie these phenomena in npr2Δ cells. Because MMR deficiency has also been linked with tolerance to DNA methylation damage in human cancer cells (Branch et al., 1995), one might argue that the absence of cross-resistance to the methylator MNNG contradicts a role for MMR. However, in S. cerevisiae, mutations in MMR genes do not necessarily render the cells more tolerant to MNNG (Bawa and Xiao, 1997), in line with our present data.
Irrespective of the underlying mechanisms, our data identify S. cerevisiae NPR2 as a novel gene involved in cell kill provoked by the anticancer drugs cisplatin and doxorubicin. Based on primary sequence conservation, we found NPRL2/Gene21 as the most credible candidate human NPR2 homolog (using http://www.ncbi.nlm.gov/blast/Blast.cgi). There are three regions of high similarity between the predicted polypeptide sequence encoded by the human NPRL2/Gene21 cDNA (RefSeq NP_006536.1) and Npr2p, with 32% identity (53% similarity) over 152 amino acids, 36% identity (50% similarity) over 119 amino acids, and 36% identity (54% similarity) over 55 amino acids. Strikingly, NPRL2/Gene21 is one of the candidate tumor suppressor genes residing on a 120-kb critical tumor homozygous deletion region of human chromosome 3p21.3 found in lung and breast cancers (Lerman and Minna, 2000). Recently, it was reported that NPRL2/Gene21 is, indeed, one of the candidates whose adenovirus vector-mediated expression results in tumor suppressor activity in vitro and in vivo (Ji et al., 2002). Our next aim will be to see whether this human tumor suppressor gene might also be involved in processes altering sensitivity to anticancer drugs. In addition, it will be highly interesting to assess the possible interplay between NPRL2/Gene21 and other components controlling drug sensitivity, such as the SKY1 homolog SRPK1. The finding that NPRL2/Gene21 resides on human chromosome 3 is of special interest with respect to a tentative role in MMR. Chauhan et al. (2000) recently tested the hypothesis that a single wild-type copy of the key MMR gene hMLH1 on chromosome 3 might be exclusively responsible for the restoration of MMR function in MMR-deficient HCT116 human colon cancer cells, upon transfer of a whole copy of the chromosome. However, profound inhibition of hMLH1 expression did not abrogate DNA MMR activity in the corrected cells, suggesting either that hMLH1 is expressed in large excess compared with that required for functional MMR, or that chromosome 3 contains another uncharacterized MMR gene. Given our data for yeast npr2Δ cells, it is tempting to speculate that the human Npr2p homolog is somehow involved in MMR, and that the introduction of an extra copy of NPRL2/Gene21 on human 3p21.3 thus aids in the restoration of MMR function in HCT116 cells upon whole chromosome transfer.
We thank Dr. Petra Ross-Macdonald (Yale University, New Haven, CT) for providing us with the random mini-Tn3::lacZ::LEU2 insertion library, and Dr. Germain Rousselet (Service de Biologie Cellulaire, CEA, Centre d'Etudes de Saclay, Gif-sur-Yvette Cedex, France) for the CMY375 and BUH3 yeast strains. Furthermore, Robert Oostrum, Peter de Bruijn, and Dr. Walter Loos are acknowledged for expert technical assistance. Finally, we are grateful to Maarten Niemantsverdriet for creating the sky1Δ::lox-URA3-lox geneblaster.
↵1 The data on SRPK1 expression in germ cell tumors were presented in Plenary Session 2 of the 14th EORTC-NCI-AACR Symposium on Molecular Targets and Cancer Therapeutics; 2002 Nov 19–22; Frankfurt, Germany [Schenk PW, Boersma AWM, Brok M, Brandsma JA, Den Dulk H, Burger H, Brouwer J, Stoter G and Nooter K (2002) Mechanisms of cisplatin resistance—role of yeast SKY1 and its human homolog SRPK1. Eur J Cancer 38(Suppl. 7):15.].
This work was supported in part by grant DDHK 97-1397 from the Dutch Cancer Society.
Preliminary parts of this work were presented in abstract form at the 92nd Annual Meeting of the American Association for Cancer Research; 2001 March 24–28; New Orleans, LA [Schenk PW, Boersma AWM, Brok M, Brandsma JA, Den Dulk H, Brouwer J, Burger H, Stoter G, and Nooter K (2001) Cisplatin resistance in vitro: possible sensitivity genes in translational research from yeast to man. Proc Am Assoc Cancer Res 42:abstract 4998].
ABBREVIATIONS: NER, nucleotide excision repair; YNB, yeast nitrogen base; YPD, yeast extract/peptone/dextrose; PCR, polymerase chain reaction; NPR2, nitrogen permease regulator 2; SKY1, SR-protein-specific kinase from budding yeast; SRPK1, SR-protein-specific kinase; kb, kilobase(s); AAS, atomic absorption spectrometry; HPLC, high-performance liquid chromatography; MNNG, N-methyl-N′-nitro-N-nitrosoguanidine; MMR, mismatch repair.
- Received November 20, 2002.
- Accepted May 9, 2003.
- The American Society for Pharmacology and Experimental Therapeutics