Cytotoxicity of the indole alkaloid reserpine from Rauwolfia serpentina against drug-resistant tumor cells.
Background: The antihypertensive reserpine is an indole alkaloid from Rauwolfia serpentina and exerts also profound activity against cancer cells in vitro and in vivo. The present investigation was undertaken to investigate possible modes of action to explain its activity toward drug- resistant tumor cells.
Material and methods: Sensitive and drug-resistant tumor cell lines overexpressing P-giycoprotein (ABCB1/MDR1), breast cancer resistance protein (ABCG2/BCRP), mutation-activated epidermal growth factor receptor (EGFR), wild-type and p53-knockout cells as well as the NCI panel of cell lines from different tumor origin were analyzed. Reserpine's cytotoxicity was investigated by resazurin and sulforhodamine assays, flow cytometry, and COMPARE and hierarchical cluster analyses of transcriptome- wide microarray-based RNA expressions.
Results: P-glycoprotein- or BCRP overexpressing tumor cells did not reveal cross-resistance to reserpine. EGFR-overexpressing cells were collateral sensitive and p53- Knockout cells cross-resistant to this drug compared to their wild-type parental cell lines. Reserpine increased the uptake of doxorubicin in P-glycoproteinoverexpressing cells, indicating that reserpine inhibited the efflux function of P- glycoprotein. Using molecular docking, we found that reserpine bound with even higher binding energy to P- glycoprotein and EGFR than the control drugs verapamil (P-glycoprotein inhibitor) and erlotinib (EGFR inhibitor). COMPARE and cluster analyses of microarray data showed that the mRNA expression of a panel of genes predicted the sensitivity or resistance of the NCI tumor cell line panel with statistical significance. The genes belonged to diverse pathways and biological functions, e.g. cell survival and apoptosis, EGFR activation, regulation of angiogenesis, cell mobility, cell adhesion, immunological functions, mTOR signaling, and Wnt signaling.
Conclusion: The lack of cross-resistance to most resistance mechanisms and the collateral sensitivity in EGFRtransfectants compared to wild-type cells speak for a promising role of reserpine in cancer chemotherapy. Reserpine deserves further consideration for cancer therapy in the clinical setting.
Reserpine is an indole alkaloid derived from the roots of Rauwolfia serpentina and serves as potent antihypertensive drug (Panda et al. 2012) It represents an established second line treatment against hypertension (Milne and Pinkney-Atkinson 2004; Pillay 2009).
Reserpine mediates the depletion of neurotransmitters from postganglionic nerve endings, which consequently lower arterial pressure and total peripheral resistance ultimately leading to decreased heart rates and cardiac output. Its administration together with diuretics effectively lowers arterial pressure and significantly reduces morbidity and mortality related to hypertension (Bakris and Frohlich 1989).
Early studies started at the mid-1950s reported anti-tumor effect of reserpine in vivo independent from its cardiovascular action. Experimental studies in mice bearing advanced leukemia, reserpine increased animals' life span by three-fold (Burton et al. 1956). Other studies reported the anti-tumor activity of reserpine in different mouse sarcomas in vivo (Belkin and Hardy 1957; Nelson et al. 1981).
Since the 1980s, attention was paid toward phenomena related to chemotherapy failure (Efferth 2001; Gillet et al. 2007; Eichhorn and Efferth 2012) and the ATP binding cassette (ABC) transporter, P-glycoprotein (ABCB1), which confers multidrug resistance (MDR) by energy-dependent drug efflux process (Hall et al. 2009). Drugs such as verapamil, quinidine, tamoxifen, progesterone, rapamycin, cyclosporins and others were found to inhibit P-glycoprotein and to overcome MDR (Arced 1993). Interestingly, reserpine was also reported as P-glycoprotein inhibitor. It suppressed photolabeling of P-glycoprotein with a vinblastine analogue in MDR cell lines (Akiyama et al. 1988). Reserpine and other indole alkaloids enhanced the sensitivity of MDR cancer cells toward cytotoxic agents (Beck et al. 1988; Zamora et al. 1988).
In this study, we investigated the role of reserpine in different cancer cell lines in an approach to understand possible molecular modes of action. Since MDR is multifactorial in nature and other mechanisms in addition to P-glycoprotein also contribute to unresponsiveness of tumors, we also investigated several other mechanisms. The ABCtransporter BCRP/ABCG2, the mutated tumor suppressor gene p53 and the activated epidermal growth factor receptor (EGFR) all mediate drug resistance (El-Deiry 1997,2003; Gillet et al. 2007). Therefore, we investigated, whether or not cell lines expressing these genes reveal cross-resistance to reserpine. The intention was to prove, whether reserpine could be used to bypass drug-resistance and to eradicate otherwise unresponsive tumors by reserpine.
Material and methods
Drug-sensitive CCRF-CEM and multidrug-resistant CEM/ADR5000 leukemia cell lines were cultured in RPMI 1640 medium, supplemented with 10% fetal bovine serum (FBS) (Invitrogen) and 1% penicillin (100 U/ml)-streptomycin (100 [micro]g/ml) (PIS) antibiotic (Invitrogen) and incubated in humidified 5% C[O.sub.2] atmosphere at 37[degrees]C.
Breast cancer cells transduced with control vector (MDA-MB231-pcDNA3) or with a cDNA for the breast cancer resistance protein BCRP (MDA-MB-231BCRP clone 23), human wild-type HCT116 [(p53.sup.+/+]) colon cancer cells as well as knockout clones HCT116 ([p53.sup.-/-]) derived by homologous recombination, non-transduced human U87MG glioblastoma multiforme cells and U87MG cells transduced with an expression vector harboring an epidermal growth factor receptor (EGFR) gene with a genomic deletion of exons 2 through 7 (U87MG.AEGFR) were all maintained in DMEM medium, supplemented with 10% FBS and 1% penicillin-streptomycin and incubated under standard conditions as described for leukemia cell lines.
The resistance of the different resistant cell lines has been maintained by using 5000 ng/ml doxorubicin for CEM/ADR5000, 400 [micro]g/ ml geneticin for U87MG.[DELTA]EGFR and HCT116 ([p53.sup.-/-]) and 800 ng/ml of the same compound for MDA-MB-231 BCRP clone 23.
The glioblastoma cell lines were kindly provided by Dr. W. K. Cavenee (Ludwig Institute for Cancer Research, San Diego, CA). Transfected breast cancer cell lines were a generous gift from Dr. B. Vogelstein and H. Hermeking (Howard Hughes Medical Institute, Baltimore, MD). The leukemia cells were kindly provided by Dr. J. Beck (Department of Pediatrics, University of Greifswald, Greifswald, Germany).
The cytotoxicity of reserpine has been investigated using the resazurin reduction assay (Borra et al. 2009). Resazurin is an indicator dye, which is reduced in viable cells to highly fluorescent resorufin, in contrast to non-viable cells, which lost their metabolic capability and are not able reduce resazurin. In total 96-well cell culture plate (Thermo Scientific, Germany) were seeded with 20,000 cells/well in a total volume of 100 [micro]L, and then treated with different concentrations of reserpine diluted in 100 [micro]L medium. Adherently growing cells were allowed to attach overnight and treated after 24 h. The cells were incubated with reserpine for 72 h. Then, 0.01% of resazurin (Sigma-Aldrich, Germany) diluted in double distilled water (dd[H.sub.2]0) was added (20 [micro]L/well) and incubated for another 4 h. Infinite M2000 Pro[TM] plate reader (Tecan, Germany) was used to measure the fluorescence using an excitation wavelength of 544 nm and an emission wavelength of 590 nm. Experiments were performed three times with at least with six replicates per experiment. The 50% inhibition concentrations ([IC.sub.50]) were calculated from dose response curves of each cell using Microsoft Excel 2013 software.
Doxorubicin uptake assay
Flow cytometry has been used to measure the retention of doxorubicin. Doxorubicin is substrate of P-glycoprotein and its inherent fluorescence was used to assess the efflux activity of this drug transporter. CCRF-CEM and CEM/ADR5000 cells were seeded in phenol red-free RPMI 1640 medium in a concentration of 5 x [10.sup.5] cells/well in 12 well plates at a total volume 500 [micro]L. Then, cells were treated with medium containing 10 [micro]M doxorubicin with and without reserpine (15 [micro]M). Cells were incubated at 37[degrees]C in an atmosphere containing 5% C[0.sub.2] for 3 h, which is the time required for maximum doxorubicin uptake (Krishan and Hamelik, 2005). Finally, cells were washed to remove free doxorubicin. Cells were measured on LSR-Fortessa FACS analyzer (Becton-Dickinson, Germany) equipped with argon blue laser, the excitation and emission wavelength of doxorubicin were 488 nm and 610/20 nm, respectively. Only living cells, identified by DAP1 to stain dead cells were considered for the analyses. Data were processed by Flowjo software. Three controls were taken, unstained cells to determine auto-fluorescence, cells treated with doxorubicin alone to measure the efflux efficacy, and in combination with verapamil as positive control for a P-glycoprotein inhibitor.
AutoDock4 (Hetenyi and Van Der Spoel 2002) was used for molecular docking calculations of reserpine. A homology model human ABCB1 based on the crystal structure of murine P-glycoprotein was previously constructed by us (Zeino 2014; Tajima et al. 2014) and used in the present investigation for binding site determination of reserpine. Verapamil was included in our analyses as control drug for a well-known P-glycoprotein inhibitor.
Molecular docking of reserpine to EGFR was done on the ATP-binding site of EGFR kinase domain. The crystal structure of EGFR was retrieved from The Protein Data BANK (Database code PDB1 Ml 7; http://www.rcsb.org/pdb/home/home.do). Erlotinib which is an irreversible tyrosine kinase inhibitor of EGRF was taken as control drug for docking. The 3D structure of reserpine, verapamil and erlotinib were downloaded from PubChem (pubchem.ncbi.nlm.nih.gov).
Drug binding residues of ABCB1 were identified as His61, Gly64, Leu65, Met69, Ser222, Leu304, Ile306, Tyr307, Phe336, Leu339, Ile340, Ala342, Phe343, Gln725, Phe728, Phe732, Leu762, Thr837, Ile868, Gly872, Phe942, Thr945, Tyr953, Leu975, Phe978, Ser979, Val982, Gly984, Ala985, Met986, Gly989, Gln990, and Ser993 (Aller et al. 2009). At the other hand, residues identified as binding sites for EGFR include Leu620, Leu694, Phe699, Val702, Ala719, Lys721, Met742, Leu764, Thr766, Gln767, Met769, Pro770, Cys773, Thr830 and Asp831 (Yadav et al. 2014).
Crude BDP structures of the receptors proteins were refined, energy minimized and polar hydrogens were added and saved as pdbqt files. Then, the grid map parameters were set to cover the defined residues. Numbers of runs and energy evaluations were reset to 250 and 25,000,000, respectively. Docking calculations were performed using Lamarckian Genetic Algorithm. For image visualization of docking results, Visual Molecular Dynamics (VMD) software was used.
COMPARE and hierarchical cluster analyses of microarray data
Messenger RNA expression profiles of 60 human cancer cell lines were deposited at the database of the Developmental Therapeutics Program (DTP) of the National Cancer Institute (NCI) (http://dtp.nci. nih.gov).
Compare analysis (Pauli et al. 1989) was performed to correlate IC50 values for reserpine and microarray-based transcriptomewide mRNA expression levels in the NCI cell line panel. According to gene expression levels, resistance and sensitive candidate genes were determined using standard and reverse COMPARE as described (Villeneuve and Parissenti 2004; Zeeberg et al. 2011; Reinhold et al. 2012).
Pearson's rank correlation test was used (W1NSTAT program, Kalmia) for calculation of significance values (p-values) and ranking correlation coefficients (R-values) as a relative measure of the linear dependency of two variables. The median log10 IC50 value was taken as a cut-off threshold for determination of cell lines being sensitive or resistant to reserpine.
Hierarchical cluster analysis (WARD method) groups objects into clusters according to similarities and closeness between them. For calculation of distances of all variables involved in the analysis, the program automatically standardizes the variables by transforming the data with a mean = 0 and a variance = 1. Then, cluster trees were performed.
Resazurin reduction assay
The cytotoxic effects of reserpine have been investigated in different cancer cell lines (Fig. 1) to gain insight into the mechanisms of action underlying the activity of the compound. Reserpine exerted 50% cell viability inhibition in parental CCRF-CEM cells at a concentration of 14.52 [+ or -] 1.62 [micro]M and of 13.2 [+ or -]1.02 [micro]M in P- glycoprotein-overexpressing, multidrug-resistant CEM/ADR5000 cells. These results clearly show that multidrug-resistant cells did not exhibit cross-resistance to reserpine, although these cells are highly resistant to established anticancer drugs such as anthracyclines, Vinca alkaloids, taxanes, epipodophyllotoxins and others (Efferth et al. 2008).
Treatment of drug-sensitive MDA-MB-231-pcDNA3 and BRCP-transfected, multidrug-resistant MDA-MB-231 -BCRP (clone 23) resulted in [IC.sub.50] values of 34.33 [+ or -] 10.38 [micro]M and 40.6 [+ or -] 6.84 [micro]M, respectively.
U87MG cells transfected with mutation-activated EGFR (U87MG.AEGFR) were much more sensitive toward reserpine ([IC.sub.50]: 9.15 [+ or -] 2.67 [micro]M) than wild-type cells (87.98 [+ or -] 11.84 [micro]M), indicating that EGFR-overexpressing cells display collateral sensitivity (hypersensitivity) to reserpine.
Finally, we compared the responsiveness of p53 wild-type and knockout cells toward reserpine. HCT116 ([p53.sup.+/+]) colon cancer cells were sensitive to reserpine ([IC.sub.50] : 30.07 [+ or -] 7.57 [micro]M), while HCT116 ([p53.sub.-/-]) cells were resistant to reserpine.
Doxorubicin uptake assay
Doxorubicin uptake was measured in terms of fluorescence intensity, which can be taken as a measure for intracellular accumulation of the drug. CCRF-CEM and P-glycoprotein-overexpressing CEM/ADR5000 cells were both incubated with doxorubicin with or without reserpine. CCRF-CEM cells that do not express P-glycoprotein were sensitive to doxorubicin and neither reserpine nor the control drug verapamil shows any effect on doxorubicin accumulation.
By contrast, CEM/ADR5000 cells showed only intracellular low fluorescence intensity of doxorubicin (Fig. 2). The fluorescence intensity of doxorubicin considerably increased after addition of 3.75 [micro]M reserpine (0.25 x [IC.sub.50] value) and even more increased after incubation with 15 [micro]M reserpine ([IC.sub.50] value). The retention effect of reserpine ([IC.sub.50] value) was 2.04-fold higher than of verapamil. It is important to note that there were no auto-fluorescent effects from either tested cancer cell lines.
Molecular docking of reserpine to ABCB1 and ECFR
Reserpine showed a high binding energy for human P-glycoprotein/ABCBl (-9.65 [+ or -] 0.79 kcal/mol), even higher than that of the control drug verapamil (-8.57 [+ or -] 0.13 kcal/mol) (Table 1). Two hydrogen bonds with drug binding residues (Gly226 and Lys234) interacted with reserpine. Interestingly, reserpine bound to the same drug binding site as verapamil (Fig. 3A). The residues involved in this interaction were Ser222, Pro223, Gly226, Ala229, Ala230, Lys234, Phe 303, Ile306, Tyr310, Leu339, Ala342, Phe 343 and Gly 346.
Molecular docking of reserpine to the tyrosine kinase domain of EGFR (PDB 1M17) revealed a low binding energy (-7.32 [+ or -] 1.12 kcal/ mol). Two hydrogen bonds were involved in this interaction (Lys721 and Met769). It is remarkable that reserpine revealed a higher binding affinity than the control EGFR inhibitor, erlotinib (-5.93 [+ or -] 0.3 kcal/mol). Both reserpine and erlotinib shared the same residues involved in hydrophobic interaction (Fig. 3B), i.e. Leu694, Lys721, Glu738, Leu768, Met769, Gly772, Cys773, Asp813, Arg817, Leu820 and Asp831.
Cross-resistance of reserpine to established anticancer drugs
To get a clue on possible modes of actions of reserpine, we correlated the [log.sub.10][IC.sub.50] values of the NCI cell lines to reserpine with those of 87 standard drugs. The cellular responses of 4 out of 7 anti-hormonal drugs significantly correlated with those of reserpine (= 57%). Alkylating drugs were also frequently correlated to reserpine. Seven of 13 alkylating agents (= 54%) revealed significant correlations to reserpine (p < 0.05; R > 0.30). Comparable results were obtained for mTOR inhibitors (2/4 drugs = 50%). Intermediate correlation rates were found for DNA topoisomerase inhibitors (3/8 drugs = 37.5%) antimetabolites (2/15 drugs = 13%), as well as tyrosine kinase inhibitors (1/13 drugs = 8%). No correlations were found for platin compounds, tubulin inhibitors, and epigenetic inhibitors (Fig. 4). These results may indicate that reserpine exerts multiple modes of action, a feature which is frequently observed with phytochemicals (Efferth and Koch, 2011).
3.5. COMPARE and cluster analyses of microarray data
COMPARE and cluster analyses of microarray data have shown that cell lines of different tumor types show different degrees of sensitivity to reserpine (Fig. 5). In an approach to identify molecular pharmacology and mechanism underlying responsiveness to reserpine, candidate genes have been identified using transcriptome-wide COMPARE analysis. Using Pearson test, genes were ranked whose mRNA expression directly or inversely correlate with the log10IC5o values for reserpine of all NCI cell lines (Table 2). Only genes with correlation coefficients of R > 0.5 and R < -0.5 were considered for direct inverse correlations respectively.
These genes belonged to pathways and biological functions that presumably determined responsiveness of tumor cells to reserpine, e.g. cell survival and apoptosis (TRAFh SULFi), EGFR activation (EFEMP1, EPS15LF), regulation of angiogenesis (ANCPTL4), cell mobility (ACTA2, FISPB3, CNNi), cell adhesion (POSTN, MOG), immunological function (CFI, LBP, IU7B), mTOR signaling (PPAPDC3) and Wnt signaling pathway (DIXDC1, SFRP2, CPZ, ZRANB1).
Hierarchical cluster analysis was performed using the mRNA expression values listed in Table 2. As shown in Fig. 6, the resulting dendrogram could be separated into four main branches. The median value of the [log.sub.10][IC.sub.50] values of reserpine was then used as cutoff value to define cell lines as being sensitive or resistant to reserpine. The distribution of sensitive and resistant cell lines is shown in Table 3. Interestingly, a statistically significant distribution pattern across the four dendrogram branches was observed. The majority of cell lines in clusters 1 and 2 were reserpine-resistant, whereas those in clusters 3 and 4 were mainly sensitive (P = 0.00437; [chi square] test). Since only mRNA values but not [log.sub.10][IC.sub.50] values were included into the cluster analysis, the mRNA expression profile alone was sufficient to predict sensitivity or resistance to reserpine. Clusters 1 and 2 contain both sensitive and resistant cell lines to reserpine whereas clusters 3 and 4 contain only resistant cell lines to reserpine.
Cross-resistance and collateral sensitivity of drug-resistant cell lines
Our results showed that reserpine has novel cytotoxic features that may be beneficial for the treatment of drug-resistant tumors. P-glycoprotein-overexpressing cells were not cross-resistant to reserpine, indicating that this compound may be valuable to eradicate MDR tumor cell populations of refractory tumors. The early reports from the 1950s on the anticancer activity of reserpine in vivo (Burton et al. 1956; Belkin and Hardy 1957) did not went into further mode of action analyses, since major molecular mechanisms were still unknown at that time. For instance, apoptosis as important mechanisms of cell death (Kerr et al. 1972) and P-glycoprotein as mediator of multidrug resistance (Juliano and Ling 1976) came only up in the 1970s. Later on, it has been indeed demonstrated that reserpine triggers apoptotic cell death, which explains the anticancer activity of this drug independent of its calcium channel blocking, antihypertensive effects. Reserpine induces cell death by activation of the TRAIL-mediated apoptotic pathway leading to up-regulation of BAX, down-regulation of BCL-2 as well as activation of caspase-3 and caspase-8 (Cantarella et al. 2009). In the present investigation, we found that reserpine reversed doxorubicin resistance of P-glycoprotein overexpressing CEM/ADR5000 cells. Reserpine is an effective modulator for P-glycoprotein, which increases intracellular doxorubicin accumulation by binding with high affinity to this ABC-transporters leading to efficient drug efflux inhibition. These results are in accord to previous data reporting on the photoaffinity labeling of P-glycoprotein and its inhibition by reserpine (Qian and Beck 1990). The inhibition of P-glycoprotein's efflux function by reserpine was subsequently confirmed by other authors (Schlemmer and Sirotnak 1994; Bhat et al. 1995; Jette et al. 1995; Sarver et al. 2002). In addition to these studies, we did not only confirm reserpine's inhibitory activity on P-glycoprotein's function, but we also suggested the amino acids responsible for reserpine's binding in the pharmacophore of P-glycopotein by molecular docking. Taken all these results together, reserpine can be understood as a "two-in-one" drug. It kills tumor cells due to its profound cytotoxicity, and at the same time it inhibits P-glycoprotein's efflux enhancing the activity of standard drugs such as doxorubicin.
It was pleasing to observe BCRP-transfectant tumor cells did also not exert cross-resistance to reserpine. BCRP (ABCG2) represents another ABC transporter that confers resistance to a broad spectrum of anticancer drugs in various tumor types, including human breast carcinoma, colon carcinoma, gastric carcinoma, fibrosarcoma, and myeloma (Doyle and Ross 2003). The fact that many of the established anticancer drugs clinically fail, because of the activity of ABC-transporters such as P-glycoprotein and BCRP illustrates the necessity to identify and develop novel anticancer drugs. Our data give reason to hope that reserpine is a promising candidate drug to improve the effectiveness of cancer chemotherapy.
In this context, it was interesting to observe that reserpine was hypersensitive (collateral sensitive) to tumor cells transfected with a mutation-activated form of the epidermal growth factor receptor (EGFR). This is an important oncogene, which leads not only to carcinogenesis, but also to tumor progression, metastasis and worse prognosis for survival time of cancer patients (Chanprapaph et al. 2014). EGFR modulates cell survival, proliferation, angiogenesis and migration through activation of PI3K, MAPK and STAT3 signaling pathways (Taylor 2012). EGFR overexpression has been found in several tumor types, including glioblastoma, colorectal cancer, head and neck squamous cell carcinoma, non-small cell lung cancer, breast, renal, ovarian, bladder, prostate and pancreatic cancers (Gomez et al. 2013). EGFR overexpression also confers resistance to anticancer drugs (Wykosky et al. 2011). In the present investigation, we were able to identify a potential novel role for reserpine. The results that EGFR-transfected cells were collateral sensitive to this drug compared to non-transfected control cells and that reserpine binds with even higher affinity than erlotinib to the tyrosine kinase domain of EGFR indicates that reserpine may act as EGFR kinase inhibitor. This hypothesis warrants further more detailed investigations in the future.
Another important determinant of anticancer drug resistance is the tumor suppressor gene p53. Mutations in p53 lead to functional inactivation causing deregulation of cell cycle arrest, DNA repair, and apoptosis induction (Liao et al. 2014). Our results with p53-knockout tumor cells showed that a loss of p53 function was associated with decreased cytotoxicity toward reserpine compared to wild-type p53proficient cells.
Cross-resistance pattern of the NCI cell line panel between reserpine and standard drugs
We performed a large correlation analysis between the [log.sub.10][IC.sub.50] values of reserpine and those of 87 standard anticancer drugs with the aim to identify possible modes of action of reserpine. Interestingly, the highest correlation rates were found to anti-hormonal drugs (57%) and DNA alkylating agents (53%) followed by mTOR inhibitors (50%). These results are supported by hints from the literature.
Reserpine has previously been investigated in vivo endocrinic disruptor screening assay and was found to interact with the estrogen (Ohta et al. 2012; Koch et al. 1980) reported chemopreventive effects of reserpine. Prolonged estrogen exposure induce pituitary tumors. The authors found that elevated serum prolactin levels in rats induced by ethinyl estradiol were slightly reduced by reserpine.
Alkylating agents kill cancer cells by alkylation of DNA and subsequent induction of apoptosis. The interaction of reserpine to binds to the DNA repair enzyme MSH2, thereby triggering the MSH2dependent cell-death pathway (Vasilyeva et al. 2009). In this context it is important to mention that despite the inhibition of DNA repair, reserpine did not induce genetic damage and was not genotoxic (Tsutsui et al. 1994; Kevekordes et al. 1999) indicating that reserpine may not be carcinogenic.
The correlation of reserpine to two out of four mTOR inhibitors (sirolimus, temsirolimus) is a novel find and has not been recognized before. The hypothetical inhibitory effects of reserpine on estrogen receptor and related downstream signaling pathways deserve more detailed investigation in the future.
COMPARE and cluster analyses of microarray data
Apart from the low-level cross-resistance of HCT-116 [p53.sup.-/-] cells to reserpine compared to wild-type HCT-116 [p53.sup.+/+] cell, the pother drug-resistant cell lines overexpressing ABCB1, ABCG2, or EGFR did not reveal cross-resistance to reserpine. This indicates that the responsiveness of tumor cells toward reserpine may be determined by other factors. For this reason, we applied COMPARE and hierarchical cluster analyses of microarray-based transcriptome-wide mRNA expressions to find out, whether the expression of other than the classical drug resistance genes mentioned above correlate with reserpine resistance among the NCI cell line panel. COMPARE analysis has been reported as valuable method to identify the mode of action of anticancer drugs using the NCI tumor panel (Wosikowski et al. 2000; Evans et al. 2008; Fagan et al. 2012; Luzina and Popov 2012). Furthermore, microarray hybridization and clustering techniques have been widely applied for mechanistic studies of established and novel drugs in cancer research (for instance see Reinhold et al. 2012; Villeneuve and Parissenti 2004; Zeeberg et al., 2011)
COMPARE analysis revealed genes belonging to different functional classes, whose mRNA expression correlated to [log.sub.10][IC.sub.50] values of reserpine. Remarkably, genes involved in EGFR activation were identified (EFEMP1, EPS15LI). This result and the fact that U87MG EGFR transfected cells were collateral sensitive toward reserpine compared to non-transfectant wild-type U87MG cells indicate that EGFR signaling may indeed influence reserpine's cytotoxicity toward tumor cells. Further experiments shall analyze the signaling processes of EGFR upon reserpine treatment.
Furthermore, mRNA expression of genes involved in the regulation of cell survival and apoptosis (TRAFI, SULF1) and angiogenesis (ANCPTL4) were observed to correlate to cellular responsiveness to reserpine. Both apoptosis induction and the formation of new blood vessels in tumors are important factors determining tumor survival and growth. Therefore, apoptosis-inducing and anti-angiogenic drugs became effective weapons in the fight against cancer (Efferth 2010; Wahl et al. 2011; Cheng et al. 2012; Zihlif et al. 2012; Krusche et al. 2013; Seo et al. 2013). Our data indicate that reserpine kills tumors by induction of apoptosis and inhibition of tumor blood vessel formation.
One gene involved in mTOR signaling appeared in our analysis (PPAPD3). Since reserpine was also correlated with the activity of mTOR inhibitors (sirolimus, temsirolimus), it can be hypothesized that reserpine might disturb mTOR-related signaling routes.
Furthermore, the expression of a number of genes involved in cell mobility (ACTA2, HSP3, CNN1), cell adhesion (POSTN, MOC) and immunological functions (CF/, LBP, IL17B) appeared in the COMPARE analysis. These biological functions are all involved in the regulation of tumor microenvironment and metastasis, indicating that reserpine might inhibit metastatic spread of tumors.
In this context, the WNT signaling pathway is important not only for metastasis, but also for cancer stem-like cells (Efferth 2012). Whether reserpine reveals cytotoxicity toward stem cells warrants further investigations.
The fact that reserpine that revealed less profound cytotoxicity toward P-glycoprotein-, BCRP- or EGFR-expressing but p53-mutated tumor cells raises the question, whether or not these favorable results might be translatable to the clinical situation. Reserpine has been repeatedly shown to inhibit tumor growth in mice (Nelson et al. 1981), indicating that reserpine might not only exert anticancer activity not only in vitro and in animals, but possibly also in human cancer patients. Furthermore, reserpine has been clinically applied for decades to treat cardiovascular diseases. Therefore, using reserpine for cancer therapy as novel indication should not be out of reach.
Conclusion and perspectives
The anticancer activity of reserpine in vivo is known since the 1950s (Burton et al. 1956; Belkin and Hardy 1957). However, reserpine had not been further developed as cancer drug, probably due to high toxicity. The effects of reserpine on the cardiovascular system may appear as severe side effects in normotonic cancer patients, but may even be beneficial in comorbid cancer patients that suffer from hypertonia. Furthermore, reserpine may serve as lead compound to develop novel derivatives with cytotoxicity against drug-resistant tumors, but without activity on the cardiovascular system.
Speculating that reserpine would make its way into clinical oncology, it will be used as part of combination therapy regimens rather than as monotherapy. In this context, the inhibition of P-glycoprotein and increase of intracellular drug accumulation is a pleasing feature leading to improved cancer killing efficiencies. Previously, it has been reported that reserpine does not only improve the activity of anticancer drugs involved in MDR, but also of alkylating drugs probably by interfering with DNA repair mechanisms (Wakusawa et al. 1984).
Although the antihypertensive reserpine has already been described in the 1950s to reveal anticancer activity in vivo, its true potential has not been recognized at that time to our point of view. In the present investigation, we reinvestigated this drug because it reveals surprisingly profound activity toward otherwise drug-resistant tumors.
Cancer patients, preferentially older ones, suffer not only from their tumor but also from high blood pressure. Together with the activity of reserpine against drug-resistant tumor cells, reserpine might be a valuable supplement to combination therapy protocols to treat comorbid patients with cancer and hypertension.
In conclusion, the results of the present investigation speak for a promising role of reserpine in cancer chemotherapy due to its activity toward drug-resistant tumor cells. Reserpine deserves reconsideration and re-evaluation for cancer therapy in the clinical setting.
Conflict of interest
The authors declare they have no conflict of interest.
Received 2 November 2014
Revised 7 December 2014
Accepted 12 January 2015
Abbreviations: ABC, ATP-binding cassette; BCRP, Breast cancer resistance protein; EGFR, Epidermal growth factor receptor; MAPK, Mitogen-activated protein kinase; MDR, Multidrug resistance; MSH2, MutS homologue 2, mismatch repair enzyme; mTOR, Mammalian target of rapamycin; NCI, National Cancer Institute; PI3K, Phosphoinositid-3-kinase; STAT3, Signal transducer and activator of transcription 3.
Akiyama, S., et at, 1988. Most drugs that reverse multidrug resistance also inhibit photoaffinity labeling of P-glycoprotein by a vinblastine analog. Mol. Pharmacol. 33 (2), 144-147.
Aller, S.G., et al., 2009. Structure of P-glycoprotein reveals a molecular basis for polyspecific drug binding. Science 323 (5922), 1718-1722.
Arced, R.J., 1993. Clinical significance of P-glycoprotein in multidrug resistance malignancies. J. Amer. Soc. Hematol. 81,2215-2222.
Bakris, G.L., Frohlich, E.D., 1989. The evolution of antihypertensive therapy: An overview of four decades of experience. J. Amer. College Cardiol. 14 (7), 15951608.
Beck, W.T., et al., 1988. Effect of indole alkaloids on multidrug resistance and labeling of P-glycoprotein by a photoaffinity analog of vinblastine. Biochem. Biophys. Res. Commun. 53 (3), 959-966.
Belkin, M., Hardy, W., 1957. Effect of reserpine. Science 125 (3241), 233-234.
Bhat, U.G., Winter, M.A., Pearce, H.L., Beck, W.T., 1995. A structure-function relationship among reserpine and yohimbine analogues in their ability to increase expression of mdrl and P-glycoprotein in a human colon carcinoma cell line. Mol. Pharmacol. 48 (4), 682-689.
Borra, R.C., et al., 2009. A simple method to measure cell viability in proliferation and cytotoxicity assays. Braz. Oral Res. 23 (3), 255-262.
Burton, R., et al., 1956. Reticular activating system of brain stem and "animal hypnosis" antileukemic action of reserpine. Science 125 (3239), 156-157,
Cantarella, G, Di Benedetto, G, Martinez, G, Loreto, C, dementi, G, Cantarella, A, Prato, A, Bernardini, R, 2009. Amylin prevents TRAIL-mediated apoptotic effects of reserpine in the rat gastric mucosa. Peptides 30 (8), 1466-1472.
Chanprapaph, K., Vachiramon, V., Rattanakaemakorn, P., 2014. Epidermal growth factor receptor inhibitors: A review of cutaneous adverse events and management. Dermatol. Res. Pract. 2014, 734249.
Cheng, J.-J., Tsai, T.-H., Lin, L-C, 2012. New alkaloids and cytotoxic principles from Sinomenium acutum. Planta Med. 78 (17), 1873-1877.
Doyle, L.A., Ross, D.D., 2003. Multidrug resistance mediated by the breast cancer resistance protein BCRP (ABCG2). Oncogene 22 (47), 7340-7358.
Efferth, T., 2010. Personalized cancer medicine: From molecular diagnostics to targeted therapy with natural products. Planta Med. 76(11), 1143-1154.
Efferth, T., et al., 2008. Prediction of broad spectrum resistance of tumors towards anticancer drugs. Clin. Cancer Res. 14 (8), 2405-2412.
Efferth, T., 2012. Stem cells, cancer stem-like cells, and natural products. Planta Med. 78 (10), 935-942.
Efferth, T., 2001. The human ATP-binding cassette transporter genes: From the bench to the bedside. Curr. Mol. Med. 1 (1), 45-65.
Efferth, T., Koch, E., 2011. Complex interactions between phytochemicals. The multitarget therapeutic concept of phytotherapy. Current Drug Targets 12 (1), 122-132. Eichhorn, T., Efferth, T., 2012. P-glycoprotein and its inhibition in tumors by phytochemicals derived from Chinese herbs. J. Ethnopharmacol. 141 (2), 557-570. El-Deiry, W.S., 1997. Role of oncogenes in resistance and killing by cancer therapeutic agents. Curr. Opin. Oncol. 9(1), 79-87.
El-Deiry, W.S., 2003. The role of p53 in chemosensitivity and radiosensitivity. Oncogene 22 (47), 7486-7495.
Evans, A., et al., 2008. Glut-1 as a therapeutic target: Increased chemoresistance and HIF-l-independent link with cell turnover is revealed through COMPARE analysis and metabolomic studies. Cancer Chemother. Pharmacol. 61 (3), 377-393.
Fagan, V., et al., 2012. COMPARE analysis of the toxicity of an iminoquinone derivative of the imidazo[5,4-f]benzlmidazoles with NAD(P)H:quinone oxidoreductase 1 (NQO1) activity and computational docking of quinones as NQO1 substrates. Bioorg. Med. Chem. 20 (10), 3223-3232.
Gillet, J.-P., Efferth, T., Remade, J., 2007. Chemotherapy-induced resistance by ATPbinding cassette transporter genes. Biochim. Biophys. Acta 1775 (2), 237-262.
Gomez, G.G., et al., 2013. Therapeutic resistance in cancer: microRNA regulation of EGFR signaling networks. Cancer Biol. Med. 10 (4), 192-205.
Hall, M.D., Handley, M.D., Gottesman, M.M., 2009. Is resistance useless? Multidrug resistance and collateral sensitivity. Trend. Pharmacol. Sci. 30 (10), 546-556.
Hetenyi, C., Van Der Spoel, D., 2002. Efficient docking of peptides to proteins without prior knowledge of the binding site. Prot. Sci. 11 (7), 1729-1737.
Jette, L, Murphy, G.F., Leclerc, J.M., Beliveau, R., 1995. Interaction of drugs with Pglycoprotein in brain capillaries. Biochem. Pharmacol. 50 (10), 1701-1709.
Juliano, R.L., Ling, V., 1976. A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants. Biochim. Biophys. Acta 455 (1), 152-162.
Kerr, J.F., Wyllie, A.H., Currie, A.R., 1972. Apoptosis: A basic biological phenomenon with wide-ranging implications in tissue kinetics. Brit. J. Cancer 26 (4), 239- 257.
Kevekordes, S., et al., 1999. SOS induction of selected naturally occurring substances in Escherichia coli (SOS chromotest). Mut. Res. 445 (1), 81-91.
Koch, M., et al., 1980. Induction of pituitary tumours and hyperprolactinemia in female rats by estrogens. The effect of apomorphine, reserpine and L-dopa. Arch. Toxicol, 89-91.
Krishan, A., Hamelik, R.M., 2005. Flow cytometric monitoring of fluorescent drug retention and efflux. Methods Mol. Med. Ill, 149-166.
Krusche, B., Arend, J., Efferth, T., 2013. Synergistic inhibition of angiogenesis by artesunate and captopril in vitro; and in vivo. Evid. Based Complement. Alternat. Med. 2013,454783.
Liao, J., et al., 2014. New insights into p 53 functions through its target microRNAs. J. Mol. Cell Biol. 6 (3), 206-213.
Luzina, E.L, Popov, A. V, 2012. Synthesis, evaluation of anticancer activity and COMPARE analysis of N-bis(trifluoromethyl)alkyl-N'-substituted ureas with pharmacophoric moieties. Euro. J. Med. Chem. 53,364-373.
Milne, F.J., Pinkney-Atkinson, V.J., 2004. Hypertension guideline 2003 update. S. Afr. Med. J. 94 (3 Pt 2), 220-225 pp. 209-16, 218.
Nelson, M., Nelson, D.S., Hopper, K.E., 1981.1. Tumor growth in mice with depressed capacity to mount inflammatory responses: Possible role of macrophages. Amer. J. Pathol. 104(2), 114-124.
Ohta, R., et al., 2012. Ovariectomized mouse uterotrophic assay of 36 chemicals. J. Toxicol. Sci. 37 (5), 879-889.
Panda, S.K., et al., 2012. Phyto-pharmacognostical studies and quantitative determination of reserpine in different parts of Rauwolfia (spp.) of Eastern Odisha by UV spectroscopy Method. Asian J. Plant Sci. Res. 2 (2), 151-162.
Pauli, K.D., et al., 1989. Display and analysis of patterns of differential activity of drugs against human tumor cell lines: Development of mean graph and COMPARE algorithm.). Natl. Cancer Inst. 81,1088-1092.
Pillay, T., 2009. A comparison of two methods for measuring anti-hypertensive drug use: Concordance of use with South African standard treatment guidelines. Bullet. World Health Organ. 87 (6), 466-471.
Qian, X., Beck, W.T., 1990. Binding of an optically pure photoaffinity analogue of verapamil, LU-49888, to P-glycoprotein from multidrug-resistant human leukemic cell lines. Amer. Assoc. Cancer Res. 50,1132-1137.
Reinhold, W.C., et al., 2012. CellMiner: A web-based suite of genomic and pharmacologic tools to explore transcript and drug patterns in the NCi-60 cell line seL Cancer Res. 72 (14), 3499-3511.
Sarver, J.G., Klis, W.A., Byers, J.P., Erhardt, P.W., 2002. Microplate screening of the differential effects of test agents on Hoechst 33342, rhodamine 123, and rhodamine 6 G accumulation in breast cancer cells that overexpress P-glycoprotein. J. Biomol. Screen. 7(1), 29-34.
Schlemmer, S.R., Sirotnak, F.M., 1994. Functional studies of P-glycoprotein in insideout plasma membrane vesicles derived from murine erythroleukemia cells overexpressing MDR 3. Properties and kinetics of the interaction of vinblastine with P-glycoprotein and evidence for its active mediated transport. J. Biol. Chem. 269 (49), 31059-31066.
Seo, E., et al., 2013. Antiangiogenic activity and pharmacogenomics of medicinal plants from traditional Korean medicine. Evid. Based Complement. Alternat. Med. 2013, 131306.
Tajima, Y., et al., 2014. Nitensidine A, a guanidine alkaloid from Pterogyne nitens, is a novel substrate for human ABC transporter ABCB1. Phytomedicine 21 (3), 323-332.
Taylor, T.E., Furnari, F.B., Cavenee, W.K., 2012. Targeting EGFR for treatment of gliboblastoma: Molecular basis to overcome resistance. Curr. Cancer Drug Target. 12 (3), 197-209.
Tsutsui, T., et al., 1994. Reserpine-induced cell transformation without detectable genetic effects in Syrian hamster embryo cells in culture. Carcinogenesis 15 (1), 11 - 14.
Vasilyeva, A., et al., 2009. Small molecule induction of MSH2-dependent cell death suggests a vital role of mismatch repair proteins in cell death. DNA Rep. 8 (1), 103-113.
Villeneuve, D.J., Parissenti, A.M., 2004. The use of DNA microarrays to investigate the pharmacogenomics of drug response in living systems. Curr. Topic. Med. Chem. 4 (13), 1329-1345.
Wahl, 0., et al., 2011. Inhibition of tumor angiogenesis by antibodies, synthetic small molecules and natural products. Curr. Med. Chem. 18 (21), 3136-3155.
Wakusawa, S., Miyamoto, IC, Koshiura, R., 1984. Increase of sensitivity and uptake of vinblastine by reserpine in rat ascites hepatoma. Japan. J. Pharmacol. 36 (2), 187195.
Wosikowski, K., et al., 2000. Reduced growth rate accompanied by aberrant epidermal growth factor signaling in drug resistant human breast cancer cells. Biochim. Biophys. Acta 1497 (2), 215-226.
Wykosky, J., et al., 2011. Therapeutic targeting of epidermal growth factor receptor in human cancer: Successes and limitations. Chin. J. Cancer 30 (1), 5-12.
Yadav, I.S., et al., 2014. Ensemble docking and molecular dynamics identify knoevenagel curcumin derivatives with potent anti-EGFR activity. Gene 539 (1), 82-90.
Zamora, J.M., Pearce, H.L., Beck, W.T., 1988. Physical-chemical properties shared by compounds that modulate multidrug resistance in human leukemic cells. Mol. Pharmacol. 33 (4), 454-462.
Zeeberg, B.R., et al., 2011. RedundancyMiner: De-replication of redundant GO categories in microarray and proteomics analysis. BMC Bioinform. 12,52.
Zeino, M., et al., 2014. The ability of molecular docking to unravel the controversy and challenges related to P-glycoprotein--A well-known, yet poorly understood drug transporter. Invest. New Drug 32 (4), 618-625.
Zihlif, M., et al., 2012. Screening the antiangiogenic activity of medicinal plants grown and sold in Jordan. Planta Med. 78 (3), 297-301.
Sara A.A. Abdelfatah, Thomas Efferth *
Department of Pharmaceutical Biology, Institute of Pharmacy and Biochemistry, Johannes Gutenberg University of Mainz, Staudinger Weg 5, 55128 Mainz, Germany
* Corresponding author. Tel.: +49 6131 3925751; fax: +49 6131 3923752.
E-mail address: firstname.lastname@example.org (T. Efferth).
Table 1 Molecular docking of reserpine to P-glycoprotein/ABCBl and tyrosine kinase domain of EFGR. Target Compound Lowest binding Mean binding protein energy (kcal/mol) energy (kcal/mol) ABCB1 Reserpine -9.65 ([+ or -] 0.34) -9.12 ([+ or -] 0.79) Verapamil -8.57 ([+ or -] 0.13) -8.57 ([+ or -] 0.12) EGFR Reserpine -7.86 ([+ or -] 1.57) -7.32 ([+ or -] 1.12) Erlotinib -6.81 ([+ or -] 0.18) -5.93 ([+ or -] 0.3) Target Compound Residues involved Number of residues protein in hydrogen bond involved in hydrophobic interaction ABCB1 Reserpine Gly226 Lys234 14 Verapamil Lys234 15 EGFR Reserpine Lys721 Met769 10 Erlotinib Met769 Cys773 13 Target Compound pKi * (nM) protein ABCB1 Reserpine 92.75 ([+ or -] 42.2) Verapamil 526.79 ([+ or -] 117.04) EGFR Reserpine 531.2 ([+ or -] 361.08) Erlotinib 10230 ([+ or -] 0.34) * pKi: Predicted inhibitory constant. Table 2 COMPARE analysis of genes, whose microarray-based mRNA expression correlated with [log.sub.10] [IC.sub.50] values for reserpine in a panel of 43 cell lines. R-value Gene symbol Genbank Pattern ID accession number Standard compare (resistance genes) 0.666 MYBPH NM_004997 CC183640 0.651 SLC9A9 AI536067 GC66696 0.641 OLFML2B AL050137 GC83013 0.636 BBS9 AF095771 GC152697 0.631 EFEMP1 AI740711 GC159025 0.621 CPI U83411 GC97760 0.618 TRAF1 U19261 GC32739 0.617 FMOl NM_002021 GC181437 0.617 TMEM132B AI435595 GC157072 0.615 CTNNA2 M94151 GC34818 0.614 ANGPTL4 AF169312 GC153504 0.613 SULF1 A1479175 GC157295 0.611 CFI Y00318 GC101200 0.607 CH25H AF059214 GC56092 0.604 HSPB3 NM_006308 GC184572 0.603 ACTA2 X13839 GC35741 0.602 DIXDC1 AF070621 GC26930 0.602 SFRP2 A1246042 GC61488 0.601 POSTN D13666 GC85573 0.599 ZNF385D NM_024697 GC189574 0.596 ATP8B2 AB032963 GC151263 0.595 C10orf72 AL080114 GC83204 0.594 CHI3L1 M80927 GC179212 0.594 CNN1 D17408 GC37204 0.591 COL1A2 AA628535 GC149250 0.591 LBP NM_004139 GC182970 0.591 ODF2 AL138382 GC165390 0.591 CLCA2 NM_006536 GC184752 0.59 PPAPDC3 BC006362 GC172554 0.59 DCTN3 W26651 GC30840 0.589 RBP4 NM_006744 GC184918 0.589 BGN NM_001711 GC181189 0.588 MOG U64565 GC28275 0.586 SMAP2 AI702142 GC71360 0.583 PSMB9 AI758695 GC73064 0.58 LDB2 NM_001290 GC180889 0.578 IL17B NM 014443 GC186079 0.573 EPS15L1 AV710549 GC167431 0.573 LOC284542 BF060736 GC174719 Reverse COMPARE (Sensitive genes) -0.635 C1orf86 H41433 GC13018 -0.632 ZRANB1 AW269836 GC169305 -0.592 MGC70870 AA278816 GC42843 -0.59 CTBP2 AF016507 GC55340 -0.586 CCND1 T89175 GC11864 -0.585 LOCI00287615 AI700233 GC71205 -0.574 CMTM4 W46185 GC15744 -0.552 LARP1B AK027164 GC163652 -0.539 KCNQ4 H26683 GC12683 -0.534 DOCK6 AI198543 GC60349 -0.53 MYL12B U26162 GC191176 -0.526 SMC5 N70436 GC15093 -0.524 LOC285147 H23213 GC87116 -0.524 ZNF205 AF060865 GC37936 -0.523 S1K2 AA142956 GC41241 -0.519 TMED5 N49615 GC14323 -0.519 AGAP1 AB029022 GC37681 -0.518 RBMS2 R99202 GC13107 -0.513 TM9SF3 N93209 GC15318 -0.512 SJ8SIA3 N62107 GC14703 -0.512 HSPC159 AK025603 GC163255 -0.507 SPATA20 H15544 GC11293 -0.506 ZNF652 AA057773 GC17251 0.506 MRPL17 AI141700 GC59526 -0.504 C13orf1 N30354 GC14024 -0.501 TAC1 N47310 GC14485 -0.5 GCOM1 N51713 CC14400 R-value Gene symbol Gene name Standard compare (resistance genes) 0.666 MYBPH Myosin binding protein H 0.651 SLC9A9 Solute carrier family 9 0.641 OLFML2B Olfactomedin-like 2B RNA 0.636 BBS9 Bardet-Biedl syndrome 9 0.631 EFEMP1 EGF-containing fibulin-like extracellular matrix protein 1 0.621 CPI Carboxypeptidase Z 0.618 TRAF1 TNF receptor-associated factor 1 0.617 FMOl Flavin containing monooxygenase 1 0.617 TMEM132B Transmembrane protein 132B 0.615 CTNNA2 Catenin (cadherin-associated protein), [alpha] 2 0.614 ANGPTL4 Angiopoietin-like 4 0.613 SULF1 Sulfatase 1 0.611 CFI Complement factor 1 0.607 CH25H Cholesterol 25-hydroxylase 0.604 HSPB3 Heat shock 27 kDa protein 3 0.603 ACTA2 Actin, [alpha] 2, smooth muscle, aorta 0.602 DIXDC1 DIX domain containing 1 0.602 SFRP2 Secreted frizzled-related protein 2 0.601 POSTN Periostin, osteoblast specific factor 0.599 ZNF385D Zinc finger protein 385D 0.596 ATP8B2 ATPase, class 1, type 8B, member 2 0.595 C10orf72 Chromosome 10 open reading frame 72 0.594 CHI3L1 Chitinase 3-like 1 (cartilage glycoprotein-39) 0.594 CNN1 Calponin 1, basic, smooth muscle 0.591 COL1A2 Collagen, type 1, [alpha] 0.591 LBP Lipopolysaccharide binding protein 0.591 ODF2 Outer dense fiber of sperm tails 2 0.591 CLCA2 Chloride channel accessory 2 0.59 PPAPDC3 Phosphatidic acid phosphatase type 2 domain containing 3 0.59 DCTN3 Dynactin 3 (p22) 0.589 RBP4 Retinol binding protein 4, plasma 0.589 BGN Biglycan 0.588 MOG Myelin oligodendrocyte glycoprotein 0.586 SMAP2 Small ArfGAP2 0.583 PSMB9 Proteasome (prosome, macropain) subunit, beta type, 9 (large multifunctional peptidase 2) 0.58 LDB2 LIM domain binding 2 0.578 IL17B Interleukin 17B 0.573 EPS15L1 Epidermal growth factor receptor pathway substrate 15-like 0.573 LOC284542 Hypothetical protein LOC284542 Reverse COMPARE (Sensitive genes) -0.635 Clorf86 Chromosome 1 open reading frame 86 -0.632 ZRANB1 Zinc finger, RAN-binding domain containing 1 -0.592 MGC70870 C-terminal binding protein 2 pseudogene -0.59 CTBP2 C-terminal binding protein 2 -0.586 CCND1 Cyclin D1 -0.585 LOCI00287615 Hypothetical protein LOCI 00287615 -0.574 CMTM4 CKLF-like MARVEL transmembrane domain containing 4 -0.552 LARP1B La ribonucleoprotein domain family, member 1B -0.539 KCNQ4 Potassium voltage-gated channel, KQT-like subfamily, member 4 -0.534 DOCK6 Dedicator of cytokinesis 6 -0.53 MYL12B Myosin, light chain 12B, regulatory -0.526 SMC5 Structural maintenance of chromosomes 5 -0.524 LOC285147 Hypothetical protein LOC285147 -0.524 ZNF205 Zinc finger protein 205 -0.523 S1K2 Salt-inducible kinase 2 -0.519 TMED5 Transmembrane emp24 protein transport domain containing 5 -0.519 AGAP1 ArfGAP with GTPase domain, ankyrin repeat and PH domain 1 -0.518 RBMS2 RNA binding motif, single stranded interacting protein 2 -0.513 TM9SF3 Transmembrane 9 superfamily member 3 -0.512 SJ8SIA3 ST8 [alpha]-N-acetyl-neuraminide [alpha]-2,8-sialyltransferase 3 -0.512 HSPC159 Galectin-related protein -0.507 SPATA20 Spermatogenesis associated 20 -0.506 ZNF652 Zinc finger protein 652 0.506 MRPL17 Mitochondrial ribosomal protein L17 -0.504 C13orfl Chromosome 13 open reading frame 1 -0.501 TAC1 Tachykinin, precursor 1 -0.5 GCOM1 GRINL1A complex locus RNA R-value Gene symbol Gene function Standard compare (resistance genes) 0.666 MYBPH Structural constituent of muscle 0.651 SLC9A9 Solute hydrogen antiporter activity 0.641 OLFML2B Extracellular matrix binding 0.636 BBS9 Required for ciliogenesis 0.631 EFEMP1 Epidermal growth factor-activated receptor activity 0.621 CPI Metallocarboxypeptidase activity 0.618 TRAF1 Regulation of cell survival and apoptosis 0.617 FMOl N,N-dimethylaniline monooxygenase activity 0.617 TMEM132B Molecular function 0.615 CTNNA2 Structural constituent of cytoskeleton 0.614 ANGPTL4 Protein with hypoxia-induced expression in endothelial cells acts as a regulator of angiogenesis 0.613 SULF1 Diminishes proliferation, and facilitates apoptosis 0.611 CFI Inactivates complement subcomponents 0.607 CH25H Cholesterol 25-hydroxylase activity 0.604 HSPB3 Inhibitor of actin polymerization 0.603 ACTA2 Cell motility 0.602 DIXDC1 Positive effector of the Wnt signaling pathway 0.602 SFRP2 Modulator of Wnt signaling 0.601 POSTN Cell attachment and adhesion 0.599 ZNF385D Nucleic acid binding 0.596 ATP8B2 Nucleotide binding 0.595 C10orf72 Protein binding 0.594 CHI3L1 Important role in the capacity of cells to respond to and cope with environmental changes 0.594 CNN1 Actin binding 0.591 COL1A2 Protein binding 0.591 LBP Affinity enhancer for CD14, facilitating its association with lipopolysaccharides 0.591 ODF2 Structural molecule activity 0.591 CLCA2 May act as a tumor suppressor in breast and colorectal cancer 0.59 PPAPDC3 Negative regulator of myoblast differentiation 0.59 DCTN3 Involved in spindle assembly and cytokinesis 0.589 RBP4 Retinol transporter activity 0.589 BGN May be involved in collagen fiber assembly 0.588 MOG Mediates homophilic cell-cell adhesion 0.586 SMAP2 ARF GTPase activator activity 0.583 PSMB9 Threonine-type endopeptidase activity 0.58 LDB2 Transcription factor binding transcription factor activity 0.578 IL17B Stimulates the release of tumor necrosis factor [alpha] and IL-1-[beta] 0.573 EPS15L1 Role in receptor-mediated endocytosis 0.573 LOC284542 Unknown Reverse COMPARE (Sensitive genes) -0.635 Clorf86 Polyubiquitin binding -0.632 ZRANB1 Ubiquitin thiolesterase activity/positive regulator of Wnt signaling pathway -0.592 MGC70870 Unknown -0.59 CTBP2 Transcription corepressor activity -0.586 CCND1 Cyclin-dependent protein serine/threonine kinase regulator activity -0.585 LOCI00287615 Unknown -0.574 CMTM4 Cytokine activity -0.552 LARP1B Nucleic acid binding -0.539 KCNQ4 Delayed rectifier potassium channel activity -0.534 DOCK6 Guanyl-nucleotide exchange factor activity -0.53 MYL12B Myosin regulatory subunit -0.526 SMC5 DNA double-strand repair -0.524 LOC285147 Unknown -0.524 ZNF205 Transcriptional regulation -0.523 S1K2 Activates insulin signaling pathway -0.519 TMED5 Vesicular protein trafficking, maintenance of Golgi apparatus -0.519 AGAP1 GTPase-activating protein -0.518 RBMS2 Nucleotide binding -0.513 TM9SF3 Unknown -0.512 SJ8SIA3 [alpha]-N-acetylneuraminatea [alpha]-2,8-sialyltransferase activity -0.512 HSPC159 Carbohydrate binding -0.507 SPATA20 Fertility regulation -0.506 ZNF652 Transcriptional repressor 0.506 MRPL17 Structural constituent of ribosome -0.504 C13orfl Protein binding -0.501 TAC1 Active peptides which excite neurons -0.5 GCOM1 Unknown Only correlations with cut-offs of R > 0.5 and R < -0.5 were considered. Gene function information was retrieved from Gene Card database (http://www.genecards.org) Table 3 Separation of clusters of cancer cell lines obtained by hierarchical cluster analysis for reserpine. The [log.sub.10] [IC.sub.50] median value (M) of reserpine was used as cut-off. Partition Cluster 1 Cluster 2 Cluster 3 Cluster 4 Sensitive <-4.80 M 3 5 8 5 Resistant >-4.80 M 8 11 2 0 P = 0.00437 ([chi square]-test)