WO1994006938A1

WO1994006938A1 - Methods for preventing multidrug resistance in cancer cells

Info

Publication number: WO1994006938A1
Application number: PCT/US1993/008799
Authority: WO
Inventors: Preet M. Chaudhary; Igor Roninson
Original assignee: The Board Of Trustees Of The University Of Illinois
Priority date: 1992-09-17
Filing date: 1993-09-17
Publication date: 1994-03-31
Also published as: AU5162293A; IL107032A0; MX9305737A; CN1092321A

Abstract

This invention is directed to methods for preventing the emergence of multidrug resistance in tumor cells during cancer chemotherapy. In particular, it relates to the use of protein kinase inhibitors to prevent the induction of the multidrug resistance (MDR1) gene by chemotherapeutic drugs. MDR1 gene expression, which results in tumor cell resistance to subsequent treatment with certain chemotherapeutic drugs is shown herein to be induced in response to treatment with various cytotoxic agents. Inhibitors of protein kinases are also shown herein to suppress this cellular response. Therefore, protein kinase inhibitors may be useful in the prevention of MDR1 induction by chemotherapeutic drugs in a variety of tumor cells, when administered prior to and/or simultaneous with cytotoxic drug treatment in cancer patients.

Description

METHODS FOR PREVENTING MULTIDRUG RESISTANCE IN CANCER CELLS

1. INTRODUCTION This invention is directed to methods for preventing the emergence of multidrug resistance in tumor cells during cancer chemotherapy. In particular, it relates to the use of protein kinase inhibitors to prevent the induction of the multidrug resistance (MDRl ) gene by chemotherapeutic drugs. MDRl gene expression, which results in tumor cell resistance to subsequent treatment with certain chemotherapeutic drugs is shown herein to be induced in response to treatment with various cytotoxic agents. Inhibitors of protein kinases are also shown herein to suppress this cellular response. Therefore, protein kinase inhibitors may be useful in the prevention of M-Di?! induction by chemotherapeutic drugs in a variety of tumor cells, when administered prior to and/or simultaneous with cytotoxic drug treatment in cancer patients.

2. BACKGROUND OF THE INVENTION Chemotherapy is a primary form of conventional cancer treatment today. However, a major problem associated with cancer chemotherapy is the ability of tumor cells to develop resistance to the cytotoxic effects of anti-cancer drugs during the course of treatment. It has been observed that tumor cells can even become simultaneously resistant to several chemotherapeutic drugs with unrelated chemical structures and mechanisms of action. This phenomenon is referred to as multidrug resistance. The best documented and clinically relevant mechanism for multidrug resistance in tumor cells is correlated with the expression of P-glycoprotein, the product of the MDRl gene. P-glycoprotein is a broad specificity efflux pump located in the cell membrane, and it functions by decreasing the intracellular accumulation of many lipophilic cytotoxic drugs, including some widely used anticancer agents such as anthracyclines, vinca alkaloids, epipodophyllotoxins, actinomycin D and taxol, thereby rendering cells resistant to these drugs (Pastan and Gottesman, 1991, Annu. Rev. Med. 42:277-286; Roninson (Ed.), 1991, Molecular and Cellular Biology of Multidrug Resistance in Tumor Cells , Plenum Press, New York; Schinkel and Borst, 1991, Seminars in Cancer Biology 2:213-226).

Human P-glycoprotein is expressed in several types of normal epithelial and endothelial tissues (Cordon-Cardo et al., 1990 J. Histochem.

Cytochem. 38:1277-1287; Thiebaut et al., Proc. Natl. Acad. Sci. USA 84:7735-7738), as well as in he atopoietic stem cells (Chaudhary and Roninson, 1991, Cell 66:85-94), and a .subpopulation of mature lymphocytes (Neyfakh et al., 1989, Exp. Cell Res. 185:496-505). More importantly, MDRl mRNA or P-glycoprotein have been detected in most types of human tumors, both before and after chemotherapeutic treatment (Goldstein et al., 198*9, J. Natl. Cancer Inst. 81:116-124; Noonan et al. , 1990, Proc. Natl.

Acad. Sci. USA 87:7160-7164). The highest levels of MDRl expression are usually found in tumors derived from MZ l-expressing normal tissues; e.g., renal, adrenocortical or colorectal carcinomas. In other types of solid tumors and leukemias, MDRl expression prior to treatment is usually relatively low or undetectable, but a substantial fraction of such malignancies express high levels of MDRl after exposure to chemotherapy (Goldstein et al., 1989, ____ Natl. Cancer Inst. 81:116-124). Prior to the present invention, the increase in MDRl expression^' after chemotherapy was usually believed to result from in vivo selection for rare pre-existing tumor cells that were already resistant to chemotherapeutic drugs due to MDRl expression.

Even low levels of MDRl expression have been correlated with the lack of response to chemotherapy in several different types of cancer (Chan et al., 1990, J. Clin. Oncol. 8:689-704; Chan et al., 1991, JL. Engl. J. Med. 325:1608-1614; Musto et al. , 1991, Brit. J. Haematol.77:50-53) . indicating that P- glycoprotein-mediated multidrug resistance represents an important component of clinical drug resistance. Whereas many clinical and pre-clinical studies have addressed pharmaceutical strategies for inhibiting the P-glycoprotein function (Ford and Hait, 1990, Pharmacol. Rev. 42:155-199), prior to the present invention, little was known about the factors that are responsible for the induction or up-regulation of P-glycoprotein expression in tumor cells under conditions relevant to cancer chemotherapy. Understanding such factors could provide insight into the development of methods for preventing the appearance of P-glycoprotein in human tumors, thus reducing the incidence of multidrug resistance in cancer, and leading to more effective chemotherapy of cancer.

Numerous gene transfer studies have demonstrated that elevated expression of the MDRl gene is sufficient to confer the multidrug resistance phenotype (Roninson (Ed.), 1991, Molecular and Cellular Biology of Multidrug Resistance in Tumor Cells, Plenum Press, New York) . For instance, mouse NIH 3T3 cells infected with a recombinant retrovirus carrying human MDRl cDNA became multidrug-resistant in proportion to the density of human P-glycoprotein on their surface; the correlation was not affected by the presence or absence of cytotoxic selection (Choi et al., 1991, Proc. Natl. Acad. Sci. USA 88:7386-7390) . Nevertheless, consistent association of some other biochemical changes with multidrug-resistant cells suggested that such alterations may also play a role in multidrug resistance, possibly by affecting the expression or function of P-glycoprotein. The most prominent of such changes is the increased activity of protein kinase C (PKC) , found in many, albeit not all, multidrug-resistant cell lines obtained after multiple steps of cytotoxic selection (Aquino et al., 1990, Cancer Co mun. 2:243-247; Fine et al., 1988, Proc. Natl. Acad. Sci. USA 85:582-586; O'Brian et al., 1989, FEBS Lett. 246:78-82; Posada et al., 1989, Cancer Commun. 1:285-292). PKC activation has been shown to increase the levels of drug resistance in some drug-sensitive and multidrug-resistant cell lines (Ferguson and Cheng,

1987, Cancer Res. 47:433-441; Fine et al. , 1988, Proc. Natl. Acad. Sci. USA 85:582-586; Yu et al. , 1991, Cancer Commun. 3:181-189) . Although PKC can apparently phosphorylate P-glycoprotein (Chambers et al., 1990, Biochem. Biophys. Res. Commun. 169:253-259; Chambers et al., 1990, J. Biol. Chem. 265:7679-7686; Hamada et al., 1987, Cancer Res. 47:2860-2865), it is unknown whether such phosphorylation is at all responsible for the observed change in drug resistance. While it has been shown that certain PKC inhibitors reversed multidrug resistance in some P-glycoprotein expressing cell lines (O'Brian et al., 1989, FEBS Lett. 246:78-82; Posada et al., 1989, Cancer Commun. 1:285-292; Palayoor et al., 1987, Biochem. Biophys. Res. Commun. 148:718-725) , the available evidence suggests that at least some of the observed effects were due to direct inhibition of P-glycoprotein function by the tested compounds rather than inhibition of PKC-mediated phosporylation (Ford et al. 1990, Cancer Res. 50:1748-1756; Sato et al., 1990, Biochem. Biophys. Res. Commun. 173:1252-1257). The above studies have provided no indication that the PKC-interactive agents could have an effect on the expression rather than phosphorylation or function of P-glycoprotein.

Several laboratories have investigated the factors that regulate MDRl gene expression in normal and malignant cells. An example of apparently normal physiological regulation of an MDRl homolog was found in mouse uterine endometrium, where the expression of a mouse mdr gene was induced by steroid hormones at the onset of pregnancy (Arceci et al., 1990, Mol. Repro. Dev. 25:101-109; Bates et al. , 1989, Mol. Cell. Biol. 9:4337-4344). In rat -liver, the expression of a mdr gene was found to be inducible by several carcinogenic or cytotoxic xenobiotics; similar induction was also observed during liver regeneration (Fairchild et al., 1987, Proc. Natl. Acad. Sci. USA 84:7701-7705; Thorgeirsson et al. , 1987, Science 236:1120-1122). Furthermore, a rodent homolog of MDRl was induced in several cell lines in response to treatment with some cytotoxic drugs (Chin et al., 1990, Cell Growth Diff. 1:361-365). In contrast, no induction of human MDRl gene by cytotoxic drugs was detected in any of the human cell lines tested in the same study. Other investigators have also failed to detect MDRl induction upon treatment with cytotoxic drugs (Schinkel and Borst, 1991, Sem. Cancer Biol. 2: 213-226) . Several studies have indicated, however, that the human MDRl gene may also, under certain conditions, be susceptible to stress induction. Thus, MDRl expression in some human cell lines was increased by treatment with heat shock, arsenite (Chin et al., 1990 J. Biol. Chem. 265:221-226) or certain differentiating agents (Mickley et al., 1989, J. Biol. Chem. 264:18031-18040; Bates et al., 1989, Mol. Cell Biol. 9:4337-4344). Some cytotoxic P-glycoprotein substrates were reported to stimulate transcription of a reporter gene from the human MDRl promoter (Kohno et al., 1989, Biochem. Biophys. Res. Commun. 165:1415-1421; Tani ura et al., 1992, Biochem. Biophys. Res. Commun. 183:917-924) and to increase P-glycoprotein expression in a mesothelioma cell line after prolonged exposure (Licht et al., 1991, Int. J. Cancer 49:630-637). Despite such reports of MDRl induction, however, it was never determined whether short-term exposure to any agents used in cancer chemotherapy could induce expression of the MDRl gene in human cells, and whether such induction could be prevented by any pharmaceutical agents.

Recently, Kioka et al. have reported that the addition of a flavonoid, quercetin, can prevent an increase in MDRl expression in a hepatocarcino a cell line, induced by arsenite, a compound which is not used in cancer treatment, but is known to activate the transcriptional pathway mediated by the heat shock response element of the promoter (Kioka et al. , 1992, FEBS Lett. , 301:307-309). Although not mentioned by the Kioka et al. publication, inhibition of PKC activity is one of the biological effects of quercetin (Gschwendt et al., 1984, Biochem. Biophys. Res. Commun. 124:63), and it is possible therefore that PKC inhibition by quercetin could be responsible, in part, for the observed inhibition of MDRl induction by arsenite. However, it is noteworthy that the ability of quercetin to inhibit transcriptional response mediated by the heat shock response element is believed by those skilled in the art to be unrelated to PKC inhibition (Kantengwa and Polla 1991, Biochem. Biophys. Res. Commun. 180:308-314). Furthermore, the Kioka et al. publication provides no suggestion that non-flavonoid PKC inhibitors would be able to inhibit MDRl induction by arsenite, or that quercetin would be able to inhibit the induction of MDRl expression when used in combination with chemotherapeutic drugs or any other agents that are not known to activate the heat shock response element-mediated pathway.

3. SUMMARY OF THE INVENTION The present invention relates to the use of protein kinase inhibitors for preventing the emergence of multidrug resistance in cancer cells, and an in vitro method for identifying protein kinase inhibitors which would be useful towards the same goal.

This invention is based, in part, on the discovery that anticancer drugs, whether or not transported by P-glycoprotein, can induce the expression of the MDRl gene in human tumor cells of diverse tissue origin. The increase in MDRl gene expression is observed at both the RNA and protein levels. MDRl induction is also observed upon treatment of cells with PKC agonists. Further, this induction by either a cytotoxic drug or a PKC agonist can be prevented by treatment of cells with a protein kinase inhibitor, indicating that a protein kinase- mediated pathway is involved in MDRl gene induction, and that protein kinase inhibitors may be useful in preventing the expression of MDRl gene in cancer cells exposed to chemotherapeutic agents. More specifically, this inhibitory effect is associated with inhibition of PKC as protein kinase inhibitors that are inactive against PKC fail to suppress MDRl induction, while protein kinase inhibitors which have potent effects on PKC efficiently inhibit the response.

The ability of chemotherapeutic drugs to induce MDRl expression in human cells upon short-term exposure in vitro indicates that cancer chemotherapy induces multidrug resistance directly, rather than through selection of pre-existing rare variants. Such direct induction is likely to occur during a patient's course of drug treatment, and it would account, at least in part, for the observed increased incidence of MDRl expression in the treated relative to untreated malignancies. Hence, administration of protein kinase inhibitors prior to and/or simultaneous with the chemotherapy involving cytotoxic drugs may be useful in preventing MDRl induction, and thus, prevent the emergence of multidrug resistant cancer cells, leading to a more favorable therapeutic outcome.

The invention is described by way of examples in which PKC agonists are shown to induce MDRl expression in normal peripheral blood lymphocytes

(PBL) and tumor cells. Additionally, various cytotoxic anticancer drugs are also described to be capable of activating the MDRl gene. Most importantly, protein kinase inhibitors are shown to prevent this MDRl induction mediated by PKC agonists or cytotoxic drugs, especially in tumor cells that have little or no detectable P-glycoprotein prior to treatment. A variety of uses are encompassed by the invention described herein, including but not limited to, the prevention of the appearance of multidrug resistant tumor cells during chemotherapy of cancer.

4. BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 Effects of phorbol ester (TPA) , diacylglycerol (DOG) and staurosporine (Staur) on P-glycoprotein function and expression in H9 cell line. A: Three hr Rhl23 accumulation by untreated and TPA- or DOG-treated cells. B: Three hr Rhl23 accumulation by untreated cells and cells treated with TPA or DOG after pretreatment with staurosporine. C: Staining of untreated and' TPA- or

DOG-treated cells with the IgG2a isotype control. D: Same as in C, stained with UIC2 antibody. E: UIC2 staining of untreated cells and cells treated with staurosporine alone or with TPA or

DOG after pretreatment with staurosporine.

Figure 2 cDNA-PCR analysis of the effects of TPA, DOG and staurosporine on MDRl mRNA expression in different cell lines. In each lane, the upper band (167 bp) corresponds to MDRl , and the lower band (120 bp) to /3₂-microglobulin specific PCR products. A: Effects of TPA or DOG, with or without staurosporine pretreatment, on MDRl mRNA expression in H9 cells. B: Time course of induction of MDRl mRNA in H9 cells by TPA. The two negative control (neg. con.) lanes correspond to PCR carried out with water or reverse transcriptase mixture without RNA in place of cDNA. C: . Induction of MDRl mRNA in K562 cells by TPA or DOG and in MCF-7 cells by TPA.

Figure 3. Flow cytometric analysis of drug-induced MDRl expression.

A. Efflux of P-glycoprotein-transported fluorescent dyes from K562 cells in the absence (left) or in the presence (right) of 30 μM verapamil (VER) . Top panel: Rhl23 efflux from untreated cells (-) and from cells treated with 50 μM

Ara-C (ARA) for 12 hours or with 10 μM Ara-C for 2 or 3 days. Bottom panel: DiOC₂(3) efflux from untreated cells and from cells treated with 1 μg/ l vinblastine

(VBL) for 36 hrs.

B. Increased P-glycoprotein expression in Ara-C treated KG1 leukemia cells. Left: Rhl23 accumulation in untreated cells or cells treated with 10 μM Ara-C for 1.5 days. Right: indirect immunofluorescence labeling of the same cells with anti-P-glycoprotein UIC2 antibody or IgG2a isotype control. C. Contour density maps of K562 cells maintained in drug-free media after exposure to different drugs and analyzed by double labeling using DiOC₂(3) (horizontal axis) and UIC2 antibody (left) or IgG2a isotype control (right) indirectly labeled with phycoerythrine (PE) (vertical axis) . Top to bottom: untreated cells, cells treated with 60 ng/ml Adriamycin for 3 days and grown without drug for 5 weeks, cells treated with 30 μM chlorambucil (CHL) for 5 days and grown without drug for 2 weeks, cells treated with 10 μM Ara-C for 3 days and grown without drug for 5 weeks (this experiment utilized one-half the amount of the secondary antibody used in the other assays) ; Rhl23-dull population of cells treated with Ara-C as above and isolated by fluorescence-activated cell sorting six weeks after removal from the drug.

Figure 4 cDNA-PCR analysis of MDRl mRNA expression in drug-treated cells, In each lane, the upper band (167 bp) corresponds to MDRl , and the lower band (120 bp) to /3₂-microglobulin specific PCR products, amplified in separate 5 tubes.

A. MDRl induction in K562 cells by Ara-C. Cells were exposed to the indicated concentrations of Ara-C for 4.5 days. Cell growth

10 relative to untreated cells was determined by the MTT assay in parallel with RNA extraction.

B. MDRl induction in K562 cells treated with different drugs. The

15 times of drug exposure are indicated. The drugs and their concentrations are as follows: -, untreated cells; DAU, 250 ng/ml daunorubicin; ADR, 500 ng/ml

20 Adriamycin; VBL, 20 ng/ml vinblastine; VP, 1 μg/ml etoposide; MTX, 200 ng/ml methotrexate, CDDP, 3 μg/ml cisplatin; CHL, 50 μM

25 chlorambucil; 5FU, 2 μg/ml

5-fluorouracyl; HU, 30 μM hydroxyurea.

C. MDRl induction in KB-3-1 carcinoma cells, untreated or treated for 2

30 days with 200 ng/ml Adriamycin or

10 μM Ara-C.

D. MDRl induction in EJ carcinoma cells, untreated (-) or treated for 4 days with 10 μM Ara-C.

35 E. Maintenance of drug-induced MDRl expression in K562 cells . Cells were treated for 3 days with 60 ng/ml Adriamycin, 10 μM Ara-C or 200 ng/ml methotrexate and cultured in drug-free medium for the indicated period of time.

Figure 5. Effect of protein kinase inhibitors on MDRl mRNA induction by cytotoxic drugs in H9 cells. In each experiment, the inhibitors staurosporine (ST), H7, Iso-H7 (IH7) or HA1004 (HA) were added Λ twice, the first time immediately prior to the addition of the corresponding drug and the second time after the specified period of time.

A. H9 cells, untreated or treated with 50 μM Ara-C for 22 hr. The inhibitors were added at the indicated concentrations at the beginning of the experiment and 16 hr later.

B. H9 cells, untreated or treated with 200 ng/ml Adriamycin for 22 hrs. Equal amounts of inhibitors (0.03 μM staurosporine, 10 μM H7, HA-1004 and Iso-HV) were added at the beginning of the experiment and 16 hrs later.

C. H9 cells, untreated or treated with 40 ng/ml vinblastine or 200 ng/ml methotrexate for 36 hours. Equal amounts of inhibitors (0.1 μM staurosporine, 50 μM H7) were added at the beginning of the experiment and 24 hrs later.

Figure 6. Vinblastine resistance in Ara-C or Adriamycin-treated K562 cell.

A. Growth inhibition by vinblastine in untreated and Ara-C (ARA)- or Adriamycin (ADR)-treated cells. Cells were treated as in FIG. 3c and grown in the absence of drugs for six weeks. Vinblastine inhibition assay was carried out for 10 days.

B. Growth inhibition by vinblastine in untreated cells and Rhl23-dull and Rhl23-bright populations of Ara-C treated cells. Ara-C treated cells, six weeks after removal from the drug, were stained by the Rhl23 efflux procedure and separated into Rhl23-dull and Rhl23-bright populations by fluorescence- activated cell sorting. The Rhl23-dull population was >60% pure (FIG. 3c) , and the Rhl23-bright population was 90-95% pure. One week after sorting, vinblastine inhibition assay was carried out for 7 days.

5. DETAILED DESCRIPTION OF THE INVENTION

The present invention relates to the use of protein kinase inhibitors to prevent the emergence of the multidrug resistance phenotype ih cancer cells. The discovery of MDRl induction by cytotoxic drugs and the ability of protein kinase inhibitors to prevent such induction are fully described and exemplified in the sections below. For clarity of discussion, the invention is described in terms of two protein kinase inhibitors which have potent effects against PKC activity and a panel of human tumor cell lines. However, the principles may be analogously applied to a wide variety of in vitro cell lines and in vivo tumors treated with any chemotherapeutic drug, using any protein kinase inhibitors.

5.1 INDUCTION OF MDRl GENE EXPRESSION TPA (12-o-tetradecanoylphorbol-13-acetate) , an efficient PKC activator, and diacylglycerol, a physiological stimulant of PKC, are shown in Example 6, infra , to increase MDRl gene expression in normal human PBL, and in cell lines derived from different types of leukemias or solid tumors. The effect of TPA is observed in all the tested cell lines that expressed P-glycoprotein prior to treatment, and in some but not all other cell lines without detectable P-glycoprotein. It is possible, however, that MDRl expression could be induced in the non-responsive cell lines by higher concentrations of TPA than those tested described herein. The observed effects of TPA and diacylglycerol indicate that MDRl expression in human cells may be regulated through a PKC-mediated signal transduction pathway. The increase in MDRl expression in cells treated with the PKC agonists is observed at the level of both P-glycoprotein and MDRl mRNA. The increased steady-state level of MDRl mRNA may reflect either increased transcription or decreased mRNA degradation. It should be noted that the major downstream promoter of the human MDRl gene (Ueda et al., 1987, J. Biol. Chem. 262:505-508) contains the AP-1 site responsible for the stimulation of transcription by TPA (Angel et al., 1987, Cell 49:729-739; Lee et al., 1987, Cell 49:741-752) . The AP-1 site and its surrounding sequences are conserved between the human MDRl gene and its rodent homologs (Hsu et al., 1990, Mol. Cell. Biol. 10:3596-3606; Teeter et al., 1991, Cell Growth Diff. 2:429-437) . The AP-1 sequence of the hamster pgpl gene was shown to be an essential positive regulator of its promoter (Teeter et al., 1991, Cell Growth Diff. 2:429-437) , although the corresponding element of the homologous mouse mdrla (mdr3 ) gene may have a negative regulatory effect (Ikeguchi et al., 1991, DNA Cell Biol. 10:639-649) . Thus, it is possible that the AP-1 element of the human MDRl promoter is directly responsible for the stimulation of MDRl expression by PKC agonists, described herein. The induction of MDRl gene expression by a PKC-mediated pathway may explain the previous finding that multidrug-resistant cell lines selected for increased P-glycoprotein expression frequently contained elevated levels of PKC (Aquino et al., 1990, Cancer Commun. 2:243-247; Fine et al., 1988, Proc. Natl. Acad. Sci. USA 85:582-586; O'Brian et al., 1989, FEBS Lett. 246:78-82; Posada et al., 1989, Cancer Commun. 1:285-292) . Presumably, an increase in PKC activity could represent an early event responsible for increased MDRl gene expression during the selection of such cell lines. This interpretation does not preclude, however, that the phosphorylation of the induced P-glycoprotein by PKC could further increase the P-glycoprotein activity. Evidence for the latter hypothesis comes from the study of Yu et al. (Cancer Commun. 3:181-189), who found that the level of drug resistance in a multidrug-resistant subline of MCF-7 cells, obtained after transfection with MDRl cDNA transcribed from a heterologous promoter, could be increased by the introduction of a vector expressing high levels of PKC . The increased resistance in the PKCα transfectants was accompanied by increased P-glycoprotein phosphorylation, without apparent changes in its expression levels.

PKC plays a central role in various signal transduction pathways, associated with different adaptive, proliferative and differentiative processes. Even though PKC agonists have been found to induce MDRl expression in normal and malignant hematopoietic cells, the same result has not been achieved using hematopoietic growth factors that may also act through PKC-mediated pathways. Furthermore, PKC agonists induce MDRl expression in cell lines of not only hematopoietic but also epithelial origin, indicating that PKC-mediated regulation of MDRl expression may have a general physiological role.

PKC-mediated mechanisms have been implicated in the transcriptional response to DNA damage by UV irradiation or alkylating agents (Kaina et al., 1989, In M. W. Lambert and J. Laval (Ed.), DNA Repair Mechanisms and Their Biological Implications in

Mammalian Cells, Plenum Press, New York; Papathanasiou and Fornace, 1991, pp.13-36 In R. F. Ozols (Ed.), Molecular and Clinical Advances in Anticancer Drug Resistance. Kluwer Academic Publishers, Boston, MA) . PKC activation has also been associated with cellular response to other cytotoxic drugs, such as cytosine arabinoside (Kharbanda et al., 1991, Biochemistry 30:7947-7952) or doxorubicin (Posada et al., 1989, Cancer Res. 49:6634-6639). Thus, PKC-mediated induction of MDRl expression could be a part of a general stress response to different types of cellular damage, including the damage produced by cytotoxic chemotherapeutic drugs.

Indeed, the present invention discloses that MDRl expression in human leukemia and solid-tumor derived cell lines can be induced by short-term exposure to different cytotoxic drugs that are used in cancer chemotherapy (see Example 7, infra) . MDRl induction, both at the RNA and at the protein levels, was observed in a subpopulation of cells treated with either P-glycoprotein-transported agents (Adriamycin, daunorubicin, vinblastine, etoposide) or chemotherapeutic drugs that are not transported by P-glycoprotein (Ara-C, methotrexate, 5-fluorouracil, chlorambucil, cisplatinum, hydroxyurea) . Since MDRl expression does not provide resistance to drugs of the second group, and because MDRl induction could be achieved after short times of drug exposure (less than one cell generation in many cases) , these findings indicate that cytotoxic selection for W-D-l-expressing cells could not be responsible for the observed increase in MDRl expression. MDRl induction became detectable at the same time as visible cell damage, indicating that it was more likely to be an indirect consequence of such damage, rather than a direct response to specific agents.

Most importantly, MDRl expression induced by treatment with cytotoxic drugs, did not disappear after the removal of the drugs, but was maintained for at least several weeks in cells cultured in drug-free media. P-glycoprotein-positive cells, growing in the absence of the drugs, showed no apparent changes in their differentiation. Thus, drug-induced MDRl expression is a stable phenomenon which is not limited to dying or terminally differentiating cells. In addition to increased MDRl expression, drug-treated cells displayed a 2-3 fold increase in resistance to vinblastine, a P-glycoprotein transported drug; such resistance was specifically associated with MDJl-expressing cells. In addition, drug-treated cells showed increased resistance to chlorambucil, a chemotherapeutic drug which is not transported by P-glycoprotein. The latter finding suggests that some other clinically relevant mechanisms of drug resistance may be co-induced with MDRl expression after treatment with cytotoxic drugs.

Taken collectively, these findings suggest that treatment of human tumor cells with various drugs used in cancer chemotherapy can induce MDRl expression directly, rather than by selection of preexisting genetic variants, as previously believed. The resulting increase in multidrug resistance is. stable and may be sufficient to reduce the response to chemotherapeutic drugs both in vitro and in vivo . It seems likely that drug-mediated induction of MDRl expression would occur during cancer chemotherapy, and it may account, at least in part, for the observed increase in the incidence of MDRl expression in drug-treated human tumors. This'invention therefore provides the first documentation of MDRl induction under clinically relevant conditions and suggests that PKC may play a central role in such induction. The latter hypothesis provides a basis for chemotherapeutic protocols that would prevent MDRl induction during cancer chemotherapy through the inhibition of PKC. 5.2 USE OF PROTEIN KINASE INHIBITORS TO PREVENT MDRl INDUCTION

The present invention demonstrates that protein kinase inhibitors, especially those with potent activity against PKC, are capable of preventing the induction of MDRl gene expression in cancer cells. For example, staurosporine, a potent but non-selective inhibitor of PKC (Ruegg and Burgess, 1989, Trends Pharmac. Sci. 10:218-220) , is shown to prevent MDRl induction in P-glycoprotein-negative cells treated with TPA, diacylglycerol and a number of chemotherapeutic cytotoxic drugs, including Ara-C, vinblastine, methotrexate and Adriamycin. Another protein kinase inhibitor, H7, is also shown to prevent MDRl induction by chemotherapeutic drugs. These findings provide the evidence that PKC may be involved in such induction and the possibility of using protein kinase inhibitors to suppress MDRl gene expression. Furthermore, the observation that chemotherapeutic drugs induce resistance in tumor cells to at least one drug not transported by P-glycoprotein implies that other drug-resistance pathways may be co-induced with MDRl expression by chemotherapeutic drug treatment, and raises the possibility that protein kinase inhibitors may be able to suppress the emergence of various mechanisms of drug resistance in tumor cells rather than just MDRl induction.

In some P-glycoprotein-positive cell lines, however, staurosporine, when used alone, significantly increased P-glycoprotein expression. It should be noted that staurosporine is a P-glycoprotein inhibitor and may bind directly to P-glycoprotein (Sato et al., 1990, Biochem. Biophys. Res. Commun. 173:1252-1257) . Additionally, two other P-glycoprotein binding compounds, cyclosporine A and verapamil, that are also known PKC inhibitors, increased P-glycoprotein expression in some of the P-glycoprotein-positive cell lines. It is conceivable therefore that these agents and staurosporine act through a common, presently unknown mechanism. These results suggest that protein kinase inhibitors may be more effectively used to prevent an increase of MDRl in P-glycoprotein-negative or nearly negative tumors than in tumors already expressing P-glycoprotein in a large fraction of tumor cells. It should be noted, however, that the finding that staurosporine increased P-glycoprotein expression in a small number of hematopoietic cell lines does not indicate that augmentation of P-glycoprotein expression by PKC inhibitors is a general property of P-glycoprotein-positive tumor cells, and that patients with P-glycoprotein-positive tumors cannot benefit from the use of protein kinase inhibitors to prevent further drug-induced increase of multidrug resistance in tumor cells.

P-glycoprotein negative solid tumors or leukemias can be identified by the analysis of biopsy material, surgical or hematological specimens of patients' tumors using techniques well known in the art (Roninson (Ed.), 1991, Molecular and Cellular Biology of Multidrug Resistance In Tumor Cells. Plenum Press, New York) . These techniques include but are not limited to immunocytochemical, immunohistochemical or immunofluorescent assays with

P-glycoprotein-specific antibodies; vital staining with P-glycoprotein transported fluorescent dyes; northern dot blot or slot blot hybridization with

MlλRl-specific nucleic acid probes; or cDNA-PCR analysis of MDRl mRNA. Working examples of some of the above assays are described in Examples 6 and 7, infra. It should be noted that some cell lines that appear to be P-glycoprotein negative by protein or

107297.1 function-based assays may still show MDRl mRNA detectable by cDNA-PCR (See Table 1) . This indicates that protein or function-based assays would be preferable as the primary criterion for the identification of tumors that are likely to benefit from the use of protein kinase inhibitors. Alternatively, cDNA-PCR or other methods for MDRl mRNA measurement may be used with the understanding that MDRl mRNA expression at the level of K562 cells or slightly (e.g. 2-fold) higher may still be indicative of P-glycoprotein negative tumors.

Although the protein kinase inhibitors tested herein are known to be non-selective in their inhibitory activities, i.e., their action is not specific for PKC, the studies described herein provide evidence that their ability to inhibit PKC activity may be a critical factor in the prevention of.' MDRl induction. For example, two potent PKC inhibitors, staurosporine and H7, are capable of inhibiting MDRl induction by cytotoxic drugs. In contrast, HA1004, a protein kinase inhibitor that is inactive against PKC, is shown to be totally ineffective in preventing MDRl induction. Hence, it is highly likely that any protein kinase inhibitor that is-capable of inhibiting PKC, irrespective of its specificity for PKC, would be useful in preventing MDRl induction in cancer cells. Accordingly, any protein kinase inhibitor capable of preventing the induction of MDRl by chemotherapeutic drugs as measured by any method described in Example 6, infra , such as fluorescent dye accumulation, cDNA-PCR for MDRl mRNA or staining with P-glycoprotein specific antibody, may be used in the practice of the method of the invention. Such inhibitors may be administered in a cancer patient bearing a solid tumor or leukemia prior to and/or simultaneous with treatment by chemotherapeutic drugs. Any anti-cancer drug commonly used in cancer chemotherapy is encompassed within the scope of this regimen, including, but not limited to, Ara-C, Adriamycin, daunorubicin, vinblastine, etoposide, methotrexate, 5-fluorouracyl, chlorambucil, cisplatin, and hydroxyurea.

A number of compounds capable of inhibiting PKC have been investigated in vitro and in vivo for potential use in cancer chemotherapy. However, it should be noted that while such compounds were found to show selective growth inhibition for tumor relative to normal cells (Powis and Kozikowski, 1991, Clin. Biochem. 24:385-397; Grunicke et al., 1989, Adv. Enzyme Regul. 28:201-216) they have not been shown or suggested to be capable of preventing MDRl expression in cancer cells. In vitro studies have shown that the anti-proliferative effects of PKC inhibitors occurred at approximately the same dose as their PKC inhibitory activity (Grunicke et al., 1989, Adv. Enzyme Regul. 28:201-216). The in vivo tested compounds include staurosporine and its benzoyl derivative CGP 41 251, which were found in nude mice to show anti-tumor effect at 1/10 of their maximum tolerated doses (MTD) (MTD was 1 mg/kg for staurosporine and 250 mg/kg for CGP 41 251) (Meyer et al., 1989, Int. J. Cancer 43:851-856). Other staurosporine analogs shown to have antitumor activity in vivo include UCN-01 (Takahashi et al., 1987, J. Antibiot. 40:1782-1784) and 8-N-(diethylaminoethyl) rebeccamycin (BMY 27557) (Schurig et al., 1990, Proc. Aroer. Assoc. Cancer Res. 31:Abs. 2469) . For the latter compound, the optimal doses for i.p. administration ranged from 12 mg/kg/injection daily x 9 to a single dose of 64 mg/kg. Another group of PKC inhibitors actively

107297.1 investigated as anticancer agents comprises ether lipid analogues, including hexadecylphosphocholine, ET-18-OCH3, ilmofosine, SRI 62-834 and BM 41440 (Powis and Kozikowski, 1991, Clin. Biochem. 24:385-397; Grunicke et al., 1989, Adv. Enzyme Regul. 28:201-216) . Some of these agents have been used in clinical trials. The MTDs established in these trials for oral administration were 200 mg/day for ilmofosine (Berdel et al., 1988, Proc. Amer. Assoc. Cancer Res. 29:Abs. 2050) and 5 mg/kg body weight for BM41440 (Herrmann et al., 1987, LJpids 22:962-966) .

Hexadecylphosphocholine was also used topically for treatment of skin metastases of breast cancer, at a dose range of 0.2 to 38.5 g per patient, administered over 3 to 128 weeks (Unger et al., 1990, Cancer Treat. Rev. 17:243-246). Compounds of this group were also tested as purging agents for autologous bone marrow transplantation (Vogler et al., 1991, Exp. Hematol. 99:557 Abs.) . Another PKC inhibitor, suramin, has been used in the treatment of parasitic diseases, and is being evaluated in clinical trials as an antineoplastic agent. Continuous infusion of suramin at a rate designed to reach a peak of 300 μg/ml at the end of 14 days has shown activity in hormone-refractory prostate cancer (Myers et al.,

1992, J. Clin. Oncol. 10:875-877) . A member of another class of PKC inhibitors, the flavonoid quercetin, was shown to potentiate the anti-tumor effect of cisplatin, a drug which is not transported by P-glycoprotein, in nude mice when administered i.p. at 20 mg/kg (Grunicke et al., 1989, Adv. Enzyme Regul. 28:201-216) .

While none of the above compounds, with the exception of staurosporine described in Examples 6 and 7, infra, have been tested for the ability to prevent MDRl induction by cytotoxic drugs, the results disclosed in the present invention strongly indicate that they are likely to possess such an effect, since all of them are capable of inhibiting PKC. The availability of in vivo animal and clinical trial data for these and other PKC inhibitors enables anyone skilled in the art to use such compounds in combination with conventional anticancer drugs to prevent the emergence of multidrug resistance during chemotherapy. These compounds may be administered into a cancer patient prior to and/or simultaneous with chemotherapeutic drug treatment at a dose range of about 1-250 mg/kg body weight, either by repeated injections, by continuous infusion, or as topical treatment.

In addition, an in vitro assay may be developed for rapid identification of any compound which is capable of preventing the induction of MDRl gene expression by chemotherapeutic drugs. For example, the H-9 or K562 leukemia cell lines may be treated with a test compound for about 30 minutes prior to exposure to 10 μM Ara-C, 200 ng/ml Adriamycin, 200 ng/ml methotrexate or 40 ng/ml vinblastine under regular tissue-' culture conditions for 10-36 hours (though any time of culture over 1 hour may be sufficient, see FIG. 2B) , followed by evaluation of the ability of such compounds to prevent MDRl induction by the drugs, as compared to controls. The compounds, identified by such an assay as being capable of preventing MDRl induction by chemotherapeutic drugs, need not necessarily be protein kinase inhibitors, but may be used for patient treatment in the same manner as protein kinase inhibitors. EXAMPLE: A PROTEIN KINASE INHIBITOR PREVENTS PROTEIN KINASE C AGONIST-MEDIATED MDRl INDUCTION IN NORMAL AND TUMOR CELLS

6.1. MATERIALS AND METHODS

6.1.1. CELL LINES AND DRUG TREATMENT Normal human PBL were obtained from healthy volunteers by venipuncture after informed consent, followed by the isolation of low-density mononuclear cells by density gradient centrifugation in Histopaque-1077 (Sigma, St. Louis, MO) . KG1 cell line was maintained in Iscove's modified Dulbecco medium with 20% fetal calf serum (FCS) and 2 mM L-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin (GIBCO Laboratories, Grand Island, NY). MCF-7, EJ, KB-3-1, HeLa, and HT-1080 cell lines were maintained in DMEM with 10% FCS and 2 mM L-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin. All other cell lines were maintained in RPMI medium with 10% FCS and 2 mM L-glutamine, 100 units/ml penicillin and 100 μg/ml streptomycin.

Stock solutions of 100 μg/ml TPA (Sigma, St. Louis, MO) and 30 mM 1,2-dioctanoylglycerol (DOG or DiC₈) (Molecular Probes, Eugene, OR) were prepared in dimethylsulfoxide (DMSO) and stored at -30°C. Control experiments with DMSO solutions showed that DMSO had no effect on P-glycoprotein function or expression. Different concentrations of TPA were used for the treatment of different cell lines, depending on the observed cytotoxicity. PBL were treated with 1 ng/ml, H9 and K562 cells with 10 ng/ml, KGla and KG1 cells with 100 ng/ml and the other cell lines with 10 ng/ml of TPA. Similarly, different concentrations of DOG were used for different cell lines.. • Thus, PBL were treated with 75 μM DOG, whereas H9 and K562 cells were treated with two 75 μM doses of DOG given 2 hr apart. Cells were exposed to TPA or DOG for 8-12 hr before flow cytometric analysis or RNA extraction. Staurosporine (Sigma, St. Louis, MO) was used at 100 nM concentration for KGla cells and at 30 nM for the other cell lines; it was added to the cells 30 minutes prior to the addition of TPA or DOG.

6.1.2. FLOW CYTOMETRIC ASSAYS

P-glycoprotein activity was assayed by a rhoda ine 123 (Rhl23) accumulation assay. For this assay, drug-treated or untreated cells were washed three times and incubated for 1.5-2 hours at 37°C in media containing 100 ng/ml Rhl23 (Sigma, St. Louis, MO) . Cells were then washed, stained with propidium iodide (PI) and kept on ice until analysis. Cells growing in monolayer were suspended with 20 mM ethylene-diaminetetraacetic acid (Sigma) in phosphate buffered saline (PBS) at pH 7.4 and washed three times prior to Rhl23 staining. In some experiments, the Rhl23 efflux assay (Chaudhary and Roninson, 1991, Cell 66:85-94) was used instead of Rhl23 accumulation. P-glycoprotein expression on the cell surface was analyzed using a P-glycoprotein-specific mouse IgG2a monoclonal antibody (mAB) UIC2 (Mechetner and Roninson, 1992, Proc. Natl. Acad. Sci. USA 89:5824-5828). Mouse IgG2a isotype control antibody was obtained from Sigma. For staining of PBL with UIC2 mAB or isotype control, 10⁶ cells were stained at 4°C with 10 μg of the antibody for 30 minutes and, after two washes, stained for 30 minutes with 10 μg of FITC-conjugated goat anti-mouse IgG2a antibody (Fisher Scientific, Fairlawn, NJ) , diluted 1:2 with PBS plus 2% FCS. Cells were then washed twice with ice-cold PBS plus 2% FCS, stained with PI and kept on ice until analysis. Essentially the same protocol was used for staining of other cell types except that 2 μg of the secondary antibody were used per 10⁶ cells. In some experiments, phycoerythrine (PE) -conjugated goat anti-mouse IgG2a was used as the secondary antibody; no PI was added in such cases. Flow cytometric analysis was conducted on a Coulter Epics 753 Flow Cytometer.

6.1.3. RNA EXTRACTION AND CDNA-PCR ANALYSIS RNA was extracted from approximately 10⁶ cells by a small-scale sodium dodecyl sulfate extraction procedure (Peppel and Baglioni, 1990, BioTechnigues 9:711-713) . cDNA synthesis and polymerase chain reaction (PCR) amplification of MDRl and /3₂-microglobulin cDNA sequences were carried out essentially as described (Noonan et al., 1990, Proc. Natl. Acad. Sci. USA 87:7160-7164; Noonan and Roninson, 1991, pp.319-333 In Roninson (Ed.),

Molecular and Cellular Biology of Multidrug Resistance in Tumor Cells. Plenum Press, New York) , with the following modifications, (i) Taq DNA polymerase was added to the PCR mixtures after initial heating of the samples to 94°C. (ii) The yield of the

/3₂-microglobulin-specific band, obtained after 28 cycles of PCR, was used as the primary criterion for equalizing the starting amounts of the cDNA templates in different preparations, in order to account for differential RNA degradation in cells subjected to different types of treatment. ³²P-labeled PCR products were detected by autoradiography. 6.2. EXAMPLES A functional assay was used to detect changes in P-glycoprotein activity in human PBL treated with different agents that induce lymphoid differentiation or activation, based on flow cytometric analysis of cellular accumulation of Rhl23, which is a P-glycoprotein-transported fluorescent itochondrial dye. In this assay, cells expressing little or no P-glycoprotein stain brightly with Rhl23, whereas cells with higher levels of P-glycoprotein activity appear Rhl23-dull. No apparent effects on P-glycoprotein were observed in cells treated with calcium ionophore A23187, IL-Iα or IL-2. In contrast, treatment of PBL with the phorbol ester TPA caused a significant increase in the number of Rhl23-dull cells. The increase in the Rhl23-dull population could be prevented by the addition of 30 μM verapamil, a P-glycoprotein inhibitor. Since the best-known cellular effect of TPA is the stimulation of PKC, it was also tested whether DOG, a cell-permeable diacylglycerol and a physiological stimulant of PKC, would have an effect on Rhl23 accumulation in PBL. Treatment of PBL with DOG also decreased Rhl23 accumulation by PBL. To determine if the observed effect of PKC stimulants on P-glycoprotein activity could be due to an increase in P-glycoprotein expression, untreated and TPA-treated PBL were stained by indirect immunofluorescence labeling with a monoclonal antibody UIC2 that recognized an extracellular epitope of

P-glycoprotein encoded by the human MDRl gene. TPA treatment markedly increased the levels of P-glycoprotein on the cell surface. The increase in P-glycoprotein was accompanied by a corresponding increase in MDRl mRNA levels in the total population of TPA-treated PBL, as detected by polymerase chain reaction (PCR) amplification of MDRl cDNA sequence. Thus, the TPA-induced increase in P-glycoprotein activity was due at least in part to the activation of MDRl gene expression at the RNA and protein levels. Since PBL comprise a heterogeneous population of many different subtypes, a series of leukemia-derived clonal cell lines were tested for changes in P-glycoprotein expression after treatment with TPA. As summarized in Table 1, all the cell lines that were positive for P-glycoprotein prior to TPA treatment showed a major increase in their P- glycoprotein expression after exposure to TPA. This group included human KG1 and KGla stem-cell like leukemia cell lines, whose relatively high level of

P-glycoprotein was likely to reflect the expression of this protein in normal hematopoietic stem cells (Chaudhary and Roninson, 1991, Cell. 66:85-94) , as well as murine EL4 thymoma and LBRM 33 lymphoma cell lines.

TABLE 1. EFFECT OF TPA ON MDRl EXPRESSION

Cell Line Untreated TPA-treated Assays

Normal cells PBL ++ F,A,R

Human hematopoietic cell lines KG1 (acute myelogenous leukemia) KGla (acute myelogenous leukemia) K562 (chronic myelogenous leukemia) H9 (T-cell leukemia) HL-60 (prorayelocytic leukemia) THP-1 (promyelocytic leukemia) Jurkat, clone E6-1 (T-cell leukemia) Molt-4 (T-cell leukemia) U937 (histiocytic lymphoma)

I

Mouse hematopoietic cell lines

EL4 (thymoma)

LBRM 33, clone 4A2 (lymphoma)

H man solid tumor cell lines EJ (bladder carcinoma) MCF-7 (breast carcinoma) HeLa (cervical carcinoma) KB-3-1 (subline of HeLa) HT 1080 (fibrosarcoma)

MDRl gene expression was evaluated by a functional assay for Rhl23 accumulation (F) , UIC2 antibody staining (A) or cDNA-PCR assay for MDRl mRNA (R) , and expressed as relative values. Cells were considered negative if they expressed no P-glycoprotein detectable by the Rhl23 or UIC2 staining assays and had MD- 1 mRNA level no higher than that of KB-3-1 cells

Among the cell lines that expressed no detectable P-glycoprotein, H9 and K562 leukemia cell lines showed clear-cut induction of MDRl mRNA and P-glycoprotein by either TPA or DOG. Flow cytometric assays showed that the treatment of these cell lines with TPA or DOG resulted in the appearance of a major cell population that expressed P-glycoprotein (FIG. 1) . These changes were paralleled by an increase in steady-state levels of MDRl mRNA in TPA- or DOG-treated cells (FIGS. 2A,B). As shown in FIG. 2B, MDRl mRNA became detectable in H9 cells 2 hrs after the addition of TPA and continued to increase until at least the 5 hr point, indicating a rapid response to TPA, consistent with the possibility of transcriptional activation of MDRl by TPA in these cells.

The increase in MDRl expression after TPA treatment was not limited to hematopoietic cells, but was also observed in some solid tumor-derived cell lines, including EJ bladder carcinoma cells that expressed a low level of P-glycoprotein, and MCF-7 breast carcinoma cells where MDRl expression was undetectable without TPA treatment (FIG. 2C) . As summarized in Table 1, most of the tested P-glycoprotein negative cell lines showed no induction of M£λR2 expression after TPA treatment. It should be noted, however, that these cell lines were only treated with a fixed concentration (20 ng/ml) of TPA, and were not tested for their ability to respond to higher TPA concentrations.

In an attempt to interfere with the induction of MDRl gene expression by PKC agonists, a potent protein kinase inhibitor, staurosporine, was used to treat various cell lines. Unexpectedly, staurosporine alone caused a significant increase in P-glycoprotein expression in the cell lines that were already positive for P-glycoprotein (KG1, KGla, mouse EL4 and LBRM 33 cell lines) . Two additional compounds, cyclosporine A and verapamil, which are known to be P-glycoprotein inhibitors as well as inhibitors of PKC, have also been found to increase P-glycoprotein expression and/or dye efflux in KG1 and EL4 cells. The effect of PKC inhibitors on P-glycoprotein expression in the P-glycoprotein-positive cell lines made it difficult to analyze the interactions between staurosporine and PKC agonists in such cells. However, staurosporine did not induce MDRl expression in the P-glycoprotein-negative H9 cells. The addition of staurosporine to H9 cells 30 minutes prior to TPA or DOG treatment completely abolished MDRl induction by these agents, as evidenced by flow cytometric (FIG. 1) and cDNA-PCR assays (FIG. 2A) . Staurosporine also inhibited the effects of TPA and DOG in normal PBL.

7. EXAMPLE: PROTEIN KINASE INHIBITORS

PREVENT CYTOTOXIC DRUG-MEDIATED MDRl INDUCTION IN TUMOR CELLS

7.1 MATERIALS AND METHODS

7.1.1 FLOW CYTOMETRIC ASSAYS K562 cells were stained with 100 ng/ml Rhl23 or with 10 ng/ml of DiOC^p) in 5 ml of DMEM with 10% fetal calf serum, for 10 minutes at 37°C. After two washes, the cells were allowed to efflux the dye for 3 hrs (for Rhl23) or 2 hrs (for DiOC₂(3)) at 37°C in 5 ml of dye-free media, as previously described (Chaudhary and Roninson, 1991, Cell. 66:85-94) . In double-labeling experiments, 3 ng/ml of DiOC₂(3) in 5 ml of media were used for staining. ^* Each efflux assay was carried out in the presence and in the absence of 30 μM verapamil. KG1 cells were stained with 100 ng/ml Rhl23 in 5 ml of media for 3 hours at 37°C and analyzed without efflux. Indirect immunofluorescence labeling (Chaudhary and Roninson, 1991, Cell.

66:85-94) was carried out using 2 μg of the primary antibody (UIC2 or mouse IgG2a isotype control from Sigma) and 10 μg of the secondary antibody, PE-conjugated F[ab']₂ fragments of sheep anti-mouse IgG, (Sigma) per 2 X 10^s cells. One μg of the secondary antibody per 2 X 10^s cells was used for KG1 cell line. Flow cytometric analysis and flow sorting were carried out as described (Chaudhary and Roninson, 1991, Cell, 66:85-94); non-viable cells were excluded from the analysis on the basis of abnormal size or granularity or, in experiments not utilizing phycoerythrine, by accumulation of propidium iodide.

7.1.2. GROWTH INHIBITION ASSAYS Cells were plated in duplicate in 96-well microtiter plate at 3,000 cells per well, and allowed to grow in increasing concentrations of different drugs. Cell growth after 7-10 days was analyzed by the MTT assay (Pauwels et al. , 1988, J. Virol. Meth. , 20:309-321) .

7.2 EXAMPLES The studies described in Example 6, supra , demonstrate that PKC agonists can induce MDRl expression, suggesting an important role of PKC in the activation of the multidrug resistance response in tumor cells. PKC has been implicated also in cellular responses to different types of cytotoxic stress (Papathanasiou and Fornace, 1991, pp.13-36 In R. F. Ozols (Ed.), Molecular and Clinical ^'Advances in - 35 -

Anticancer Drug Resistance. Kluwer Academic Publishers, Boston, MA) . In particular, PKC is activated by treatment with l-3-D-arabinofuranosylcytosine (Ara-C) , an effective anti-leukemic drug (Kharbanda et al., 1991,

Biochemistry, 30:7947-7952) . Therefore, experiments were performed to test whether Ara-C, which is not transported by P-glycoprotein, would have any effect on P-glycoprotein function in K562 leukemia cells. As illustrated in FIG. 3a, exposure of K562 cells to Ara-C for 12-72 hours led to the emergence of a subpopulation of 3-17% live cells showing efflux of Rhodamine-123 (Rhl23) . Rhl23 efflux was sensitive to the P-glycoprotein inhibitor verapamil. The appearance of Rhl23-dull cells was paralleled by a dose-dependent increase in MDRl mRNA expression relative to /3₂-microglobulin in Ara-C-treated K562 cells, as detected by polymerase chain reaction amplification of cDNA sequences (FIG. 4a) . A number of other chemotherapeutic drugs were also tested for their ability to induce MDRl expression in K562 cells. Adriamycin, daunorubicin, vinblastine, etoposide, methotrexate, 5-fluorouracil, chlorambucil, cisplatin and hydroxyurea were all found to induce MDRl mRNA expression (FIG. 4b) and the efflux of Rhl23 or DiOC₂(3) , another P-glycoprotein-transported dye (Chaudhary and Roninson, 1991, Cell. 66:85-94) , by 3-10% of the treated cells (FIG. 3a) . Only the first four of these drugs are transported by P-glycoprotein (Roninson

(Ed.) , 1991, Molecular and Cellular Biology of Multidrug Resistance in Tumor Cells. Plenum Press, New York) . This, together with the short times of drug exposure required for MDRl induction, indicates that cytotoxic selection for MDfJl-expressing cells could not be responsible for the emergence of the P-glycoprotein-positive subpopulations.

The ability of cytotoxic drugs to induce MDRl expression was not limited to K562 cells. Ara-C increased P-glycoprotein expression in KG1 leukemia cells which contained a significant amount of the protein prior to drug treatment, as seen by Rhl23 accumulation or immunoreactivity with monoclonal antibody UIC2 (FIG. 3a) . Ara-C also activated MDRl mRNA expression in H9 T-cell leukemia (FIG. 5) , KB-3-1 epidermoid carcinoma (FIG. 4c) , and EJ bladder carcinoma cells (FIG. 4d) , though the magnitude of induction was somewhat lower in carcinoma cell lines. In addition, MDRl mRNA expression was induced in H9 cells by treatment with Adriamycin, vinblastine and methotrexate (FIG. 5) , and in KB-3-1 cells with Adriamycin (FIG. 4c) . However, P-glycoprotein induction was not detected in HL60 leukemia cells treated with the same drugs. In all cases, MDRl induction became detectable at the same time as visible cell damage, as evidenced by cell swelling, increased granularity, altered cell shape, and growth inhibition (FIG. 4A) . In addition, continuous passage of some cell lines in the absence of drugs for several months also led to a small increase in MDRl expression, accounting for the variability in the base levels of MDRl mRNA in untreated cells.

Next, it was tested whether drug-induced MDRl expression was maintained after the cytotoxic treatment. For this purpose, K562 cells were treated with cytotoxic concentrations of Ara-C, Adriamycin, chlorambucil or methotrexate for 3-5 days, and then allowed to grow in the absence of the drugs. At different time points, MDRl expression in the surviving cells was analyzed by dye efflux and immunofluorescence labeling with UIC2 (FIG. 3c) or by cDNA-PCR (FIG. 4e) . MDRl expression in a subpopulation of treated cells was maintained for at least several weeks after the removal of the drugs, (up to 11 weeks in the Ara-C treated population) . P-glycoprotein-positive K562 cells showed no significant changes in their size, granularity and expression of di ferentiation-related antigenic markers. The presence of multidrug-resistant cells six weeks after the removal of Ara-C or Adriamycin was also demonstrated by a growth inhibition assay with vinblastine, a P-glycoprotein substrate. Vinblastine resistance, characterized by approximately 2-3 fold increase in the ID₁₀ value, was specifically associated with the Rhl23-dull subpopulation of cells (FIG. 6) . Thus, drug treatment leads to sustained induction of MDRl expression and its associated drug resistance in a subpopulation of treated cells. It was also found that Ara-C and Adriamycin-treated K562 cells were more resistant than the untreated cells to the cytotoxic effect of chlorambucil, a chemotherapeutic alkylating agent which is not transported by P-glycoprotein. This result indicates that other pathways or mechanisms of clinically relevant drug resistance may be co-induced with MDRl expression after treatment with chemotherapeutic drugs.

To determine if PKC was involved in M.D-2 induction by cytotoxic drugs, two PKC inhibitors, staurosporine and H7, were used to block MDRl mRNA induction in H9 cells. The addition of these compounds blocked MDRl induction by Ara-C, Adriamycin, methotrexate and vinblastine, as detected by cDNA-PCR (FIG. 5) and dye efflux assays with Ara-C treated cells. No significant inhibition was observed with Iso-H7, a structural analog of H7 with 10-fold weaker effect on protein kinases (Pelosin et al., 1990, Biochem. Biophys. Res. Commun., 169:1040-1048) . Similar results were observed with Ara-C treated K562 cells. To investigate the specificity of the observed inhibition for PKC, effects of increasing doses of H7 (IC₃₀=6.0 μM for PKC, 3.0 μM for protein kinase A) and HA1004, a non-PKC specific protein kinase inhibitor (IC₅₀=40 μM for PKC, 2.3 μM for protein kinase A) (Hidaka et al., 1984, Biochemistry. 23:5036-5041) were compared. As shown in FIG. 5a, H7 inhibited MDRl induction by Ara-C at 10 μM or higher concentration, but HA1004 showed no significant inhibition even at 60 μM. These results are consistent with a role for PKC in MDRl induction by cytotoxic drugs. The data presented herein demonstrate that different chemotherapeutic drugs, including those that are not transported by P-glycoprotein, can induce MDRl expression directly, rather than by selection of preexisting genetic variants. Drug-induced MDRl expression is limited to a subpopulation of treated cells and is associated with a moderate increase in the resistance to P-glycoprotein-transported drugs (approximately 2-3 fold in the case of K562 cells) . This increase may be sufficient .^'to reduce the response to chemotherapy in vivo and to enhance the selection of genetic mutants with higher levels of drug resistance. Drug-mediated induction of MDRl expression may occur during cancer chemotherapy, and it may largely account for the increased incidence of MDRl expression in treated tumors. Hence, the demonstration that PKC inhibitors can prevent MDRl induction indicates the possibility of using such agents in combination with cytotoxic drugs in cancer chemotherapy in order to achieve a higher degree of eradication of cancer cells. The present invention is not to be limited in scope by the exemplified embodiments, which are intended as illustrations of individual aspects of the invention. Indeed, various modifications of the invention in addition to those shown and described herein will become apparent to those skilled in the art from the foregoing description and accompanying drawings. Such modifications are intended to fall within the scope of the appended claims. All publications cited herein are incorporated by reference in their entirety.

Claims

WHAT IS CLAIMED IS:

1. A method for preventing the induction of MDRl gene expression in cancer cells treated with chemotherapeutic drugs, comprising administering a protein kinase inhibitor to an individual undergoing cancer chemotherapy.

2. The method of Claim 1 in which the protein kinase inhibitor inhibits protein kinase C activity.

3. The method of Claim 1 in which the cancer cells contain little or no detectable MDRl P-glycoprotein, as determined by immunoreactivity with anti-P-glycoprotein antibodies, accumulation or efflux of P-glycoprotein transported dyes, or assays for MDRl mRNA expression.

4. The method of Claim 3 in which the cancer cells are derived from hematopoietic tumors.

5. The method of Claim 3 in which the cancer cells are derived from solid tumors.

6. The method of Claim 1, 2, 3, 4 or 5 in which the inhibitor is administered prior to and simultaneously with chemotherapeutic drugs.

7. The method of Claim 1, 2, 3, 4 or 5 in which the inhibitor is administered simultaneously with chemotherapeutic drugs. 8. A method for identifying inhibitors of MDRl gene induction by chemotherapeutic drugs, comprising:

(a) exposing a tumor cell line to a test inhibitor;

(b) adding a cytotoxic drug to the cells;

(c) culturing the cells for at least one hour in the presence of the test inhibitor and the cytotoxic drug; and

(d) measuring the expression of MDRl gene in the cultured cells by its mRNA level, P-glycoprotein level, or accumulation or efflux of

P-glycoprotein transported dyes.

9. The method of Claim 8 in which the drug is Ara-C.

10. The method of Claim 8 in which the drug is vinblastine.

11. The method of Claim 8 in which the drug is Adriamycin.

12. The method of Claim 8 in which the drug is methotrexate.

13. ._.The method of Claim 8 in which the tumor cell line is H-9 leukemia.

14. The method of Claim 8 in which the tumor cell line is K562 leukemia. 15. A method for preventing the induction of MDRl gene expression in cancer cells treated with chemotherapeutic drugs, comprising administering an inhibitor to an individual undergoing cancer chemotherapy, which inhibitor is identified by the method in Claim 8, 9, 10, 11, 12, 13 or 14.

AMENDED CLAIMS

[received by the International Bureau on 21 February 1994 (21 .02.94) ; original cl aims 1 ,6 ,8 and 15 amended ; new claims 16-29 added ; other claims unchanged (5 pages ) ]

1. A method for preventing the induction of multidrug resistance in cancer cells treated with

5 chemotherapeutic drugs , comprising administering an effective amount of a protein kinase inhibitor to an individual undergoing cancer chemotherapy.

2. The method of Claim 1 in which the 10 protein kinase inhibitor inhibits protein kinase C activity .

3. The method of Claim 1 in which the cancer cells contain little or no detectable MDRl

15 P-glycoprotein, as determined by immunoreactivity with anti-P-glycoprotein antibodies, accumulation or efflux of P-glycoprotein transported dyes, or assays for MDRl mRNA expression.

20 4. The method of Claim 3 in which the cancer cells are derived from hematopoietic tumors.

25

6. The method of Claim 1, 2, 3, 4 or 5 in which the inhibitor is administered prior to chemotherapeutic drugs.

30 7. The method of Claim 1, 2, 3, 4 or 5 in which the inhibitor is administered simultaneously with chemotherapeutic drugs.

35

8. A method for identifying inhibitors of multidrug resistance induction by chemotherapeutic drugs, comprising:

(a) exposing a tumor cell line to a test inhibitor;

(b) adding a cytotoxic drug to the cells;

(c) culturing the cells for at least one hour in the presence of the test inhibitor and the cytotoxic drug; and (d) measuring the expression of MDRl gene in the cultured cells by its mRNA level, P-glycoprotein level, or accumulation or efflux of P-glycoprotein transported dyes.

9. The method of Claim 8 in which the drug is Ara-C.

10. The method of Claim 8 in which the drug is vinblastine.

11. The method of Claim 8 in which the drug is Adriamycin.

12. The method of Claim 8 in which the drug is methotrexate.

13. The method of Claim 8 in which the tumor cell line is H-9 leukemia.

14. The method of Claim 8 in which the tumor cell line is K562 leukemia.

15. A method for preventing the induction of multidrug resistance in cancer cells treated with chemotherapeutic drugs, comprising administering an inhibitor to an individual undergoing cancer chemotherapy, which inhibitor is identified by the method in Claim 8, 9, 10, 11, 12, 13 or 1 .

16. Use of a protein kinase inhibitor for the manufacture of a medicament for preventing the induction of multidrug resistance in cancer cells treated with chemotherapeutic drugs in an individual undergoing cancer chemotherapy.

17. Use of a protein kinase inhibitor as claimed in Claim 16 in which the protein kinase inhibitor inhibits protein kinase C activity.

18. Use of a protein kinase inhibitor as claimed in Claim 16 in which the cancer cells contain little or no detectable MDRl P-glycoprotein, as determined by immunoreactivity with anti-P-glycoprotein antibodies, accumulation or efflux of P-glycoprotein transported dyes, or assays for MDRl mRNA expression.

19. Use of a protein kinase inhibitor as claimed in Claim 18 in which the cancer cells are derived from hematopoietic tumors or from solid tumors.

20. Use of a protein kinase inhibitor as claimed in any one of the Claims 16 to 19 in which the medicament is for administration prior to chemotherapeutic drugs.

21. Use of a protein kinase inhibitor as claimed in any one of Claims 16 to 19 in which the medicament is for administration simultaneously with chemotherapeutic drugs.

22. Use of a protein kinase inhibitor for the manufacture of a medicament for preventing the induction of multidrug resistance in cancer cells treated with chemotherapeutic drugs in an individual undergoing cancer chemotherapy, which inhibitor has been identified by the method of any one of Claims 8 to 14.

23. Protein kinase inhibitors for use for preventing the induction of multidrug resistance in cancer cells treated with chemotherapeutic drugs in an individual undergoing cancer chemotherapy.

24. A protein kinase inhibitor as claimed in Claim 23 which inhibits protein kinase C activity.

25. A protein kinase inhibitor as claimed in Claim 23 wherein the cancer cells contain little or no detectable MDRl P-glycoprotein, as determined by immunoreactivity with anti-P-glycoprotein antibodies, accumulation or efflux of P-glycoprotein transported dyes, or assays for MDRl mRNA expression.

26. A protein kinase inhibitor as claimed in Claim 25 wherein the cancer cells are derived from hematopoietic tumors or from solid tumors.

27. A protein kinase inhibitor as claimed in any one of Claims 23 to 26 which is for administration prior to chemotherapeutic drugs.

28. A protein kinase inhibitor as claimed in any one of Claims 23 to 26 which is for administration simultaneously with chemotherapeutic drugs.

29. Protein kinase inhibitors as claimed in Claim 23 for use in preventing the induction of multidrug resistance in cancer cells treated with chemotherapeutic drugs in an individual undergoing cancer chemotherapy, which inhibitors have been identified by the method of any one of Claim 8 to 14.