2427 A Unique Human Ovarian Carcinoma Cell Line Expressing CD34 in Association with Selection for Multidrug Resistance Hans Minderman, Ph.D.' Udo Vanhoefer, M.D., Ph.D.* Karoly Toth, M.D., Ph.D.' M. Denice Minderman, B.SC.' YOUCef M. Rustum, Ph.D.' ' Department of Experimental Therapeutics, Roswell Park Cancer Institute, Buffalo, New York. * Departmenit of Internal Medicine, West German Cancer Cent'er, University of Essen, Essen, Germany. Presented in part at the 85th Annual Meeting of the American Association for Cancer Research, San Francisco, California, April 10-13, 1994. Supported in part by Grant DHP-115 from the American Cancer Society, Cancer Center Core Grant CA16056, and NIH program project grant CAI 3038. Dr. R. J. Scheper of the Department of Pathology, Free University Hospital, Amsterdam, The Netherlands, provided the LRP (LRP-56) and MRP (MRPm6) monoclonal antibodies. Dr. Udo Vanhoefer was sponsored by a special research fund of the West German Cancer Foundation. Dr. M. L. Slovak of the Department of Cytogenetics, City of Hope National Medical Center, Duarte, California, provided the HT1080 and the HT1080Dr4 Ice11 lines. Dr. W. R. Greco and Mr. J. C. Parsons of the Biomathematics Department provided statistical assistance with the curve-fitting of the doseresponse curves. Dr. M. R. Bat?rof the Division of Medicine, Ros- C? 1996 American Cancer Society BACKGROUND. The role of P-glycoprotein (Pgp) in multidrug resjstance (MDR) is uncontested. Expression of Pgp on hematopoietic cells has been correlated with CD34 expression. For acute myeloid leukemia, the prognostic value of Pgp for clinical response is at best equivalent to that of CD34. The current study investigated whether expression of CD34 can be associated with selection for drug resistance. METHODS. Several established MDR cell lines were screened by flow cytometry for expression of CD34. Human ovarian carcinoma cells (A27801, which simultaneously expressed CD34 and Pgp, were identified. Subsequent cloning resulted in a new cell line (A2780-Dx5c)that expressed CD34 in the absence of Pgp. Involvement of non-Pgp-mediated MDR mechanisms was assessed by immunohistochemistry (MRP and LRP), enzyme activity studies (glutathione pathway), cross-resistance patterns, and Northern blot (type IIcv topoisomerase). RESULTS. A2780-Dx5c was cross-resistant to doxorubicin, daunorubicin, idarubicin, and W-16. However, unlike the Pgp-expressing cells, it was not cross-resistant to vincristine or amsacrine. The drug resistance was correlated with a decreased level of type IIa topoisomerase in the A2780-Dx5c cell line compared with the parental cell line. No evidence was found of involvement of MRP, LRP, or the glutathione pathway with drug resistance in this cell line. CONCLUSIONS. A new cell line of nonhematopoietic and nonvascular endothelial origin that expresses CD34 in association with selection for MDR was cloned. A study of MDR mechanisms in this cell line revealed that reduced type IIcv topoisomerase levels were likely responsible for the MDR observed. A study of the causal relation between the selection of drug resistance and the expression of CD34 may provide insight into why CD34 correlates with poor clinical response in patients with acute myeloid leukemia. Cancer 1996; 78:2427-36. 0 1996 American Cancer Society. KEYWORDS: CD34, multidrug resistance, P-glycoprotein, MRP, LRP, topoisomeraseIl-a. T he CD34 antigen is commonly known for its association with normal huinan hematopoietic progenitor cells' and acute leukeCD34 is expressed in vascular endothelial cells5 m i a ~ . In ~ addition, -~ and vascular neoplasms.6 Increasing evidence supports that CD34 well Park Cancer Institute, provided the bone marrow samples. cer Drug Center, Roswell ParkCancer Institute, Elm and Carlton Streets, Buffalo, NY 14263. Address for reprints: Hans Minderman, Ph.D., Department of Experimental Therapeutics, Grace Can- Received June 18, 1996; revision received August 8, 1996; accepted August 8, 1996. 2428 CANCER December 1, 1996 I Volume 78 / Number 11 is expressed in a limited number of other cell types. Screening of 357 neoplasms of nonvascular derivation6 showed CD34 expression in epithelioid sarcoma (8 of 16 cases) and rare cases of leiomyosarcoma (3 of 221, whereas carcinomas were found to be negative with the exception of 1 case of squamous cell carcinoma. Notably, CD34 was found not to be expressed in ovarian carcinoma. Recently, CD34 expression was also described on a fibrohistiocytic tumor (dermatofibrosarcoma p r o t ~ b e r a n s )The . ~ human CD34 gene has been isolated and characterized at a molecular l e ~ e l . Two ~ . ~ alternative forms of cDNA encoding human CD34 were characterized" and expression of a full length and a truncated form of CD34 on human KMT2 and KGla cells was recently described," but their functional significance has yet to be determined. In the context of leukemia and predictors for clinical response, expression of CD34 on hematopoietic progenitors correlates closely with expression of P-glycoprotein ( P ~ P ) , ' ~a- 'marker ~ associated with multidrug resistance (MDR).l6,I7 A multicenter trial conducted by the Southwestern Oncology Group involving 155 patients with de novo acute myeloid leukemia (AML) correlated clinical response with expression of Pgp and CD34.l' From this study, CD34 appeared to be a better single prognosticator for decreased complete remission rate than MDR-1 gene expression (encoding for Pgp). To further evaluate the association of CD34 with drug resistance, several MDR cell lines and their parental lines were screened for expression of Pgp and CD34. CD34 was found to be expressed on a MDR cell line of nonhematopoietic and nonendothelial origin. This article describes cloning and characterization of this human ovarian carcinoma cell line with regard to drug sensitivity, Pgp, multidrug resistance protein (MRP), lung resistance protein (LRP), and CD34 expression, activities of chemical detoxification enzymes of the glutathione pathway and levels of the key target enzyme of topoisomerase-I1 inhibitors: topoisomerase-11-a. MATERIALS AND METHODS Cell Lines The cell lines used were A2780 (human ovarian carcinoma), MCF7 (human breast carcinoma), KB 3-1 (human cervical carcinoma), HL60 (human promyelocytic leukemia), HT1080 (human fibrosarcoma), and MDR sublines derived from these cell lines (A2780-Dxl, Dx3, -Dx5, MCF7-Adr, KB 8-5, HLGO/Adr, and HT1080/ Dr4) and KG1 (human acute myelogenous leukemia). Detailed procedures for culturing and selection of the MDR strains have been published p r e v i o ~ s l y . ' Cell ~-~~ cloning was performed applying standard limiting dilution techniques in 96-well microculture plates. Bone Marrow Samples After informed consent, bone marrow aspirates were obtained from patients presenting with AML as part of an Institutional Review Board-approved protocol. Nucleated cells obtained after lysis with ammonium chloride buffer (8.29 g NH,Cl, 1 g KHC03, 37 mg K,EDTA per liter, pH 7.3) were cryopreserved and thawed using standard cryopreservation procedures. Total Growth Assay Total growth was assessed by measuring total protein (sulforhodamine-B [SRBI assay). Three hundred exponentially growing cells were seeded per well into 96well tissue culture plates (Falcon #3072, Becton Dickinson, Lincoln Park, NJ). The cells were allowed to attach for 24 hours at 37 "C, 5% CO, in air, in a fully humidified atmosphere. Doxorubicin, daunorubicin, and idarubicin (Adria Laboratories, Columbus, OH), and vincristine (Sigma, St. Louis, MO) were added to the desired final concentrations from stocks made in phosphate-buffered saline (PBS). Stock solutions of amsacrine (m-AMSA) (Lot no. MK5-27-1, gift from the Drug Synthesis and Chemistry Branch, National Cancer Institute, Bethesda MD) and W-16 (Bristol-Myers, Syracuse, NY;freshly prepared) were prepared in 100% dimethylsulfoxide (DMSO) so that the final DMSO concentration in the cultures was < 0.25%, which was determined to have no effect on cell growth. After 2hour drug exposure, cells were washed twice with PBS, resuspended in RPMI1640 + 10% fetal bovine serum (Gibco, Grand Island, NY), and cultured for an additional 4 cell doubling times, except for cells exposed to vincristine, in which case the drug was present continuously. Cell monolayers were fixed in trichloroacetic acid for at least 1 hour, and washed with distilled water using an automated microplate washer (model EL403; Bio-Tek Instruments, Winooski, VT), after which SRB (0.4%weightlvolume [wlv] in 1%volume/ volume acetic acid) was added for 10 minutes at room temperature. Plates were subsequently washed with 1% acetic acid and air dried. Protein-bound dye was solubilized in 10 mM Tris base. Absorbance at 570 nm was measured by an automated Bio Kinetics reader [model EL340; Bio-Tek Instruments). All cytotoxicity assays were performed at least in triplicate. Curve fitting to assess ID50 concentrations was performed as described p r e v i o ~ s l y . ~ ~ lmmunohistochemical Staining for von Willebrand Factor (Factor Vlll) Paraffin sections of cell pellets were incubated with a primary antibody against factor VIII (Lipshaw Immu- CD34 Expression on Drug Resistant Cells/Minderrnan et al. non, Pittsburgh, PA) and antibody binding was visualized with a universal streptavidin/biotin staining kit (Lipshaw Omnitags, Pittsburgh, PA). The staining procedure was performed according to the manufacturer's instructions. Endothelial cells in paraffin sections of normal human brain served as a positive control. lmmunoperoxidase Staining for MDR Markers Paraffin sections of cell pellets were stained for Pgp expression using the JSB-1 monoclonal antibody (MoAb) (Accurate Chemical and Scientific Corp, Westbury, NYI applying an immunoperoxidase sandwich Staining of cytospin method i1S described previo~sly.~' preparations with LRP-563Land MRPm632monoclonal antibodies (MoAbs) (gift from Dr. Scheper) was performed after air drying overnight and 10-minute fixation in acetone. Isotype controls (mouse immunoglobulin IIgIG2b for LRP-56 and mouse IgGl for MRPm6) were used in place of and at the same concentration as the primary antibodies. A polyvalent biotinylated secondary antibody and streptavidin-peroxidase reagent (Shandon-Lipshaw, Pittsburgh, PA) were used to visualize the binding of the primary antibodies. lmmunofluorescent Labeling The MRKl6 MoAb (1 pg/106 cells) (gift from Dr. Tsuruo) was used to detect Pgp expression. An indirect labeling method using goat antimouse F(ab'), conjugated with phycoerythrin (PE) or fluorescein (Caltag, San Francisco, CA) was used to visualize MRK16 binding. Expression of CD34 was visualized with the PEconjugated HPCA-2 MoAb (Becton Dickinson, San Jose, CA), the fluorescein isothiocyanate-conjugated QBEND-10 MoAb (AMAC, Inc), or the biotinylated 128 MoAb33(CellPro, Inc., Bothell, WA). All MoAbs were used according to the manufacturer's instructions. The biotiriylated MoAb was visualized using streptavidin-RED613 (Gibco, Gaitersburg, MD). Flow Cytometry Flow cytometry was performed with a FACSCAN (Becton Dickinson, San Jose, CA) using the 488-nm line of a 50 mW argon laser running at 15 mW output for excitation. Fluorescein emission was collected through a 530130 band-pass filter whereas phycoerythrin emission was collected through a 585/42 band-pass filter. RED613 emission was collected through a 650 longpass filter. Analysis was performed using the Win List software (Verity Software House Inc., Topsham, MN). Flow cytometric cell sorting was performed with a FACSTAR (Becton Dickinson). 2429 Enzyme Assays Cells in exponential growth were harvested with 1 mM EDTA in PBS, washed twice with ice-cold PBS, and lysed by sonication in 5 mM dipotassium hydrogen phosphate buffer. The enzyme activities of the glutathione-S-transferases (substrate: l-chloro-2,4-dinitrobenzene), glutathione peroxidase (substrate: butyl hydroperoxide), glutathione reductase, and glucose 6phosphate-dehydrogenase were determined at 25 "C in the 120,000 g supernatant fluid as described previou~ly."~~' Results are expressed as nmol min- ' per mg protein extract. Reverse Transcriptase Polymerase Chain Reaction Amplification and Northern Blot Hybridization for Topoisomerase-11-a Levels Total RNA was extracted by a single-step guanidine isothiocyanate-phenol-chloroform extraction.38 One yg total RNA of HL60 cells was used for topoisomerase-11-a cDNA probe synthesis with an RNA polymerase chain reaction (PCR) kit (Perkin-Elmer, Branchburg, NJ). Reverse transcriptase was performed using random hexamers as primer according to the instructions of the manufacturer. The PCR reaction was performed in a GeneAmp 2400 PCR system (Perkin-Elmer, Oak Brook, IL) for 30 cycles of denaturation (95 "C, 60 seconds), annealing (52 "C, 60 seconds), and extension (72 "C, 90 seconds). The topoisomerase-11-a specific PCR primers (position 12-59939were 5'-ACCATTGCAGCCTGTA-3' (forward) and 5'-GCTCTTCCCATATTATCC-3' (reverse), respectively.4o The PCR products were gel purified and the band of the expected size of 587 base pairs was isolated. Northern blots containing 15 pg total RNA of untreated A2780 and A2780/DX5c cells were prehybridized at 65 "C (1 hour) in 0.5 M dipotassium hydrogen phosphate buffer containing 7% (w/v) sodium dodecyl sulfate, 1% (w/v) bovine serum albumin, and 1 mM EDTA. Hybridization was performed overnight at 65 "C with a [a-"PIdATP random prime labeled topoisomerase-11-a cDNA fragment corresponding to nucleotides 12 to 599. After autoradiography, blots were stripped and reprobed with a ["PI-labeled cDNA for G3PDH (Clontech Laboratories Inc., Palo Alto, CA) to determine variation in RNA loading. RESULTS Screening of MDR Cell Lines for Pgp And 0 3 4 Expression Several MDR cell lines and their parental cell lines were analyzed for Pgp expression and CD34 expression by flow cytometry. Positive cells were determined by comparing the signal intensity with that of an IgG isotype control. The KG-1 cell line served as a positive CANCER December 1,1996 I Volume 78 I Number 11 2430 110 , I A2780 wt 100 90 -0 8 80 70 9 60 s -%me!! 50 ,S 40 5 30 O 20 +A2780wt 10 -0-A- A2780-Dx1 0 0 -10 A2780-DXl A2780-DX3 U A2780-DX5 A2780-Dx3 0.001 L 0.01 0.1 A2780-Dx5 L 1 10 h h, L!%%l!L I Doxorubicin pglml FIGURE 1, Total growth of A2780 cells after 2-hour exposure to doxorubicin assessed by the sulphorhodamine-B assay (mean 2 95% confidence interval). control for CD34 expression. Pgp expression could be demonstrated in the MDR substrains of the A2780, MCF-7, and KB cells. As expected," no Pgp expression was detectable in the HL60 cell line. In addition to the KG-1 cell line, CD34 expression was only detected on the A2780 cell lines. Because CD34 and Pgp expression were only detected simultaneously on the A2780 cell lines, subsequent studies were focused on these cells. Characterization of MDR A2780 Cell Lines Dmg resistance Sensitivity of the A2780 cell lines to doxorubicin at 2hour exposure was evaluated by the SRB assay. Compared with the A2780-wt, the A2780-Dxl, A2780-Dx3, and A2780-Dx5 are 4-fold, 8-fold, and 30-fold resistant respectively, (Fig. 1) (Table 1). Pgp expression Figure 2 and Table 1 show the flow cytometric characterization of the A2780 cell lines in more detail. A heterogenous expression of Pgp was observed for all three resistant substrains, represented by the existence of two major subpopulations. The average level of expression (intensity of the fluorescence signal) for each individual subpopulation was similar for A2780-Dxl, A2780-Dx3, and A2780-Dx5, but their relative contribution to the total cell population was different for A2780-Dx1 compared with A2780-Dx3 and A2780-Dx5 (Table 1). MRKl6 + FIGURE 2. Flow cytometric evaluation of P-glycoprotein expression (left column) and CD34 expression (right column) on A2780-wt (first row), A2780-Dx1 (second row), A2780-Dx3 (third row), and A2780-Dx5 (fourth row). Dotted lines: isotype control (immunoglobulin [Ig]G2a for MRKl6 and IgG1 for HPCA-2); solid lines: specific monoclonal antibodies. GO34 expression The A2780-wt cell line showed a very low expression of CD34 with HPCA-2 (Fig. 2, right column), evident from the tailing to the right of the histogram. Similar low expression compared with the A2780-wt was observed for A2780-Dx1 and A2780-Dx3 but significant increase of expression was detectable in the A2780Dx5 cells (Fig. 2, right column, fourth row (Table 1). Comparing reactivity of the A2780-Dx5 cell line with the HPCA-2, QBEND-10, and 12-8 MoAbs revealed high reactivity with HPCA-2, low reactivity with QBEND-10, and no reactivity with 12-8 (Fig. 3). To confirm reactivity of all three MoAbs with CD34 expressed on hematopoietic cells, the human myelogenous leukemia cell line KG1 and leukemic blasts from a patient with AML were used as positive controls. Although the pattern of reactivity and resolution was different, possibly related to the use of different fluorochromes and recognition of different epitopes, all three MoAbs showed high reactivity with these controls (Fig. 3). Flow Cytometric Screening of AML Bone Marrow Samples for CD34 Expression To determine the occurrence of the HPCA-2+, 12.8CD34 phenotype in clinical material, five bone marrow CD34 Expression on Drug Resistant CellslMinderman et al. 2431 TABLE 1 Relation between Sensitivity of A2780 Cells to 2-Hour Exposure to Doxorubicin and Immunophenotype with Regard to CD34 and P-GlycoproteinExpression A2780-wt A2780-Dxl A2780-Dx3 A2780-Dx5 %Pat WPgjtt R CD34 % I& IC50 (PglmL) 0 4 0 29 65 80 8 9 10 42 1 0.05 0.2 0.4 1.5 12 9 Pm:p-dycaprotein; I&: immunodobulin 1 0 0 Resistance factor 1 4 8 30 G; t:weak; tt:strone; KG-1 A2780-Dx5 F-QBEND10 3 PE-HPCA2 3 810-12.8/ STREPTlAVlDlN .) FIGURE 3. Flow cytometric evaluation of three different anti-CD34 monoclonal antibodies (MoAbs) on acute myeloid leukemia blasts (top row), KG1 cells (middle row), and A2780-Dx5 (bottom row). The anti-CD34 MoAbs used were QBEND-10 (left column), HPCA-2 (middle column), and SP12.8 (right column). Dotted line: auto fluorescence. samples fiom AML patients were compared for their reactivity with the HPCA-2 and 12.8 MoAbs (Table 2). The samples were selected for high expression of CD34 (assessed by HPCA-2). Both concordance and discordance were observed, but when discordance was observed (Patients 1 and 2) the percentage of CD34 positive cells was higher when HPCA-2 was used. These findings suggest the existence of two phenotypes of CD34, one being the HPCA-2+,12.8' phenotype expressed on hematopoietic cells, and the other the HPCA-2+,12.8- phenotype expressed on the A2780 cell lines and some hematopoietic cells. lmmunohistochemical Staining for von Willebrand Factor (Factor VIM) To exclude the possibility of admixture of the A2780 cell lines with cells of endothelial origin, the cells were screened for the presence of factor VIII. All cell lines proved to be negative (data not shown). Cloning and Sorting of A2780-Dx5 Cells One hundred ninety-two wells were seeded at a dilution of 0.1 cell per well. Microscopic evaluation for cell number per well limited the number of potential clones to 32. From these 32 wells, 17 potential clones 2432 CANCER December 1, 1996 I Volume 78 / Number 11 MRKl6 + HPCA-2 .) FIGURE 4. Flow cytometric irnrnunophenotyping of A2780Dx5 clones with regard to P-glycoprotein (left column) and CD34 (right column) expression. Clone A2780-Dx5a (top row), A2780-Dx5b (second row), and A2780-Dx5c (bottom row). The third row represents the cell line from which A2780 Dx5c was derived through flow cytometric cell sorting. Dotted line: auto fluorescence TABLE 2 Concordance and Discordance between Assessment of the Fraction of CD34-ExpressingCells in Acute Myeloid Leukemia Bone Marrow Samples by HPCA-2 and 12.8 Staining ~~~~~ Parient 1 Patient 2 Patient 3 Patient 4 Patient 5 % CD34 (HPCA-2) % CD34 (12.8) 85 81 65 50 87 64 33 ti0 50 91 Discordant Discordant Concordant Concordant Condordant could be expanded to cell numbers large enough to allow flow cytometric screening for Pgp and CD34 expression. Three clones of interest were obtained, of which the flow cytometric profiles are displayed in Figure 4. Unlike the parent A2780-Dx5 cell line, clone A2780-Dx5a (Fig. 4 top row) displayed a single population of Pgp-expressing cells that coexpressed CD34, whereas clone A2780-Dx5b (Fig. 4, second row) displayed a single population of Pgp-expressing cells in the absence of CD34. No clones were obtained displaying CD34 expression in the absence of Pgp. Therefore, the cell population displayed in Figure 4, third row, was chosen for flow cytometric cell sorting because this population showed an almost equal distribution of Pgp-expressing and nodlow Pgpexpressing cells in the presence of CD34. Sorting was performed based on Pgp expression and the cells were subsequently analyzed again for CD34 and Pgp expression (Fig. 4, bottom row). Thus, a cell line was obtained displaying CD34 expression in the absence of Pgp, which was designated A278O-Dx5c. The absence of Pgp expression was confirmed by immunohistochemical staining with the JSB-1 MoAb applying the immunoperoxidase sandwich method (data not shown).3"Nomenclature and phenotype of the newly derived clones are summarized in Table 3. CD34 Expression on Drug Resistant Cells/Minderman et al. TABLE 3 Nomenclature and Phenotype of A2780 Cell Lines P-glycoprotein A2780-Dx5 (parental cell line) A2780-Dx5a A2780-Dx5b A2780-Dx5~ TABLE 4 Sensitivity of A2780 Cell Lines to Doxorubicin, Daunorubicin, Idarubicin, Vincristine, VP-16,and Amsacrine CD34 (HPCA-2) t, Heterogenous t , Homogenous t t t, Homogenous - - t PgP CD34 Dox, 2h t positive, ~ 2433 negdtive Cloning and Sorting of A2780-wt Cells Because the initial flow cytometric profile of the A2780-wt cells revealed a low expression of CD34 (Fig. 2, second column) an attempt was made to subclone an A2780-Wt cell line that would not show the low CD34 expression. Both limiting dilution and flow cytometric cell sorting failed to produce a cell line with a CD34 staining pattern distinct from the original A2780wt cell line. Therefore, the A2780-wt cells should be considered a clonal cell population with an intrinsic low reactivity with the HPCA-2 MoAb. Dnr, 2h Ida, 2h Vcr, cont. VP-16, 2h M-AMSA, 2h A2780 wt ~ 2 7 8 0D - -It -1tt tt tt Fold resistance Fold resistance Fold resistance 1 1 1 1 1 1 47.1= 2.V 2.P Activity of Glutathione Pathway Enzymes No significant differences were found in the activities of glutathione peroxidase, glutathione reductase, or glutathione-S-transferases between A2780-wt and A2780-DX!ic cells. In addition, the activity of glucose6-phosphate dehydrogenase, which is related to the - ND 1.4 0.8 2.Y 1.1 6a 209.6 6.5a 4.V TABLE 5 Antioxidant Enzyme Activities in Cell Lines A2780-wt and A2780Dx5c nmol m i d per mg proteina Enzymes lmmunohistochemical Screening for Pgp, LRP, and MRP Expression Using the A2780-Dx5 as a positive control for Pgp expression, the HLGO/Adr cell line as a positive control for MRP expression, and the HT1080Dr4 cell line as a positive control for LRP and MRP expression, no evidence was found for the expression of these MDR markers on the A2780Dx5c cell line. A2780 Dx5c Pgp: p-glycoprotein; dox: doxorubicin; Dnr: daunoruhicin: Ida: idarubicin, Vcr: vincristine, 2 h 2-hour; cont: continuous; M-AMSA: amsacrine. a Significanily different from A2780 wt in 2 3, P < 0.05). ~ Cross-Resistance of A2780-Dx5c Cell Line The sensitivity of the newly obtained cloned cell lines to doxorubicin, daunorubicin, idarubicin, vincristine, m-AMSA, and VP-16, was assessed by a total growth assay (SRB). As expected, because of the presence of Pgp, the A2780-Dx5a and A2780-Dx5b cell lines showed cross-resistance to all drugs tested (data not shown). The A2780-DX5c cell line displayed a low (2.6fold) but significant ( P < 0.05) resistance to doxorubicin compared with the A2780-wt cell line and only displayed cross-resistance to daunorubicin, idarubicin, and VP-16. The results are summarized in Table 4. X ~ ~- ~27no-w A278ODx5c 44.8 (212.5) 50.4 (k9.2) 13.5 (24.8) 72.5 (24.2) 43.9 (i1.4) 48.5 (1-11.2) 10.8 (23.6) 54.6 (214.3) -~ Glutathione-S-transferasesb Glutathione reductase Glutathione peroxidase' Glucose 6-phosphate dehydrogenase a The results are presented as the mean values (standard deviation) from two independent experiments in duplicate. l-chloro-2,4-dinitrobenzene as substrate. ' Butvl hvdrooeroxide as substrate. reduced nicotinamide adenine dinucleotide phosphate generation by the pentose phosphate pathway, was not altered. The activities of the glutathione-dependent enzymes are summarized in Table 5. Northern Blot for Topoisomerase-ll-cu Levels The A2780-DX5c and A2780-wt cell lines were compared for differences in RNA levels of topoisomerase11-a. Blots of total cellular RNA extracts were hybridized with a topoisomerase-11-a specific cDNA probe using a G3PDH probe for RNA loading control. The radiolabeled topoisomerase-11-a probe hybridized with a -6.3-kilobase mRNA in both cell lines, a transcript size comparable to that previously described for topoisomerase-11-a mRNA.40However, a lower level of the topoisomerase-11-a transcripts was detected in the A2780-DX5c cell line compared with the A2780-wt cell line. (Fig 5). 2434 CANCER December 1,1996 I Volume 78 / Number 11 FIGURE 5. Northern blot analysis of cellular RNA extracts from A2780wt and A2780-DX5c cell lines. The blots were hybridized with a topoisomerase-ll-cl specific cDNA probe while a G3PDH probe served as a loading control. DISCUSSION Studies designed for determining which (immuno)phenotypes predict the clinical response of AML demonstrated that both Pgp e x p r e ~ s i o n ' ~and ~ ' ~CD34 expressi0n14, 18,41-43 are negatively associated with clinical response. The combination of both has shown to be the best prognostic parameter.14 However, as single prognosticators, CD34 expression is at least of equalI4 or better value than Pgp.'* Although the CD34 protein has been characterized at the molecular level, its function has not been fully revealed. Most data reported so far indicate a role in cell adhesion with a possibility that different glycosylated forms of the protein might serve different adhesive functions.44It is not clear how this function correlates unfavorably with (clinical) drug response. The current study is the first to describe overexpression of CD34 associated with selection of MDR cells of nonhematopoietic and nonvascular origin. Using a panel of three anti-CD34MoAbs different reactivity of the A2780 cell lines was observed compared with established CD34-expressing cells (KG-1). The most notable difference was observed with the 12.8 MoAb, which showed no reactivity with A2780 cells, whereas KG-1 cells were clearly positive. It is verified that the existence of the HPCA-2+, 12.8- CD34 phenotype is relevant in bone marrow samples from AML patients by determining discordance between staining with HPCA-2 and 12.8, which is in agreement with reports by others.45 It has further been demonstrated that the A2780Dx5c clone obtained from the A2780-Dx5 cell line expresses CD34 in the absence of Pgp and displays a low but significant level of MDH compared with the sensitive wild type cell line. The drug sensitivity studies showed that clone A2780-Dx5c was cross-resistant to the anthracyclines and VP-16. This cross-resistance pattern, together with the demonstration of a decreased level of the primary target enzyme of topoisomerase-I1 inhibitors, strongly suggests that the MDR observed in this cell line is related to the decreased topoisomerase-11-a levels. Involvement of other established non-Pgp-mediated resistance mechanisms (MRP, LRP, and glutathione pathway) was ruled out. The current data confirmed that in vitro selection of drug resistance by continuous drug exposure can result in a collateral expression of multiple resistance mechanisms in the resultant resistant cell population. Evidence was presented that selection of drug resistance can lead to a stable induction of CD34 expression. However, the results did not rule out that for the induction to occur, a low level of unstable CD34 expression (like that observed in the A2780-wt cell line) is a prerequisite. Because decreased topoisomerase11-(Ylevels were detected in the resistant cell line, an active role of CD34 in drug resistance in this cell line model is not likely. However, it will be useful to determine the causal relationship between selection of drug resistance and induction of CD34 overexpression because expression of CD34 may have been collaterally induced with expression of a marker that is actively involved in drug resistance. In this regard, the c-myc protooncogene is of particular interest because it is known that c-myc can play a role in development of drug r e ~ i s t a n c ethat , ~ ~ mRNA turnover is reduced in AML,47and that c-myc expression is highest in the CD34-expressing cells.48The presence of c-myc-like motifs upstream of the CD34 encoding regions may suggest a possible regulatory role of c-myc in the expression of CD34.8 REFERENCES 1. 2. 3. Civin CI, Strauss LC, Brovall C, Fackler MJ, Schwartz JF, Shaper JH. Antigenic analysis of hematopoiesis. 111. A hematopoietic progenitor cell surface antigen defined by a monoclonal antibody raised against KG-la cells. / Immunol 1984;133(1):157-65. Civin C. Human mono-myeloid cell membrane antigens. Exp Hematol 1990; 18:461-76. Goldstein LJ, Galski H, Fojo A, Willingham M, Lai S-L, Gazdar A, et al. Expression of a multidrug resistance gene in human cancers. / Nut1 Cancer Inst 1989;81:116-24. 0 3 4 Expression on Drug Resistant Cells/Minderman et al. 4. 5. 6. 7. 8. Carlo-Stella C, Mangoni L, Almici C, Frassoni F, Fiers W, Rizoli V. Growth of CD34+ acute myeloblastic leukemia colony-forming cells in response to recombinant growth factors. Leukemia 1990;4:561-6. Fina L, Molgaard HV, Robertson D, Bradley NJ, Monaghan P, Delia D, et al. Expression of the CD34 gene in vascular endothelial cells. Blood 1990;752417-26. Traweek ST, Kandalaft PL, Mehta P, Battifora H. The human hematopoietic progenitor cell antigen (CD34) in vascular neoplasia. A m J Clin Patkol 1991;96:25-31. Aiba S, Tabata N, Ishii H, Ootani H, Tagami H. Dermatofibrosarcoma protuberans is a unique fibrohistiocytic tumour expressing CD34. Br J Dermatol 1992;127:79-84. He X-Y,Antao VP, Basila D, Marx JC, Davis BR. Isolation and molecular characterization of the human CD34 gene. Blood 1992:79f9):2296-302. ,~~ _._ shmonsDL. sattehWite m. T~~~~ DG. Seed B. ~ ~ cloning of a cDNA encoding CD34, a sialomucin of human hematopoietic stem cells. J Immunol 1992;148(1):267-71. Nakaniura Y, Komano H, Nakauchi H. Two alternative forms of cDNA encoding CD34. Exp Hematoll993; 2~236-42. Krause DS, Fackler MJ, Smith OM, May WS. Two forms of CD34 protein are expressed in human KMT2 and KGla cells. Proceedings 35th Annual Meeting of the American Society of Hematology 1993:llOa. Chaudhary PM, Roninson IB. Expression and activity of Pglycoprotein, a multidrug efflux pump, in human hematopoietic stem cells. Cell 1991;66:85-94. Mason J, Bessent E, Mullen M, Shea T. P-glycoprotein expression on mobilized peripheral blood progenitor cells correlates with CD34. Proc A m Assoc Cancer Res 1992;33:199. Campos L, Guyotat D, Archimbaud E, Calmard-Oriol P, Tsuruo T, Troncy J, et al. Clinical significance of multidrug resistance P-glycoprotein expression on acute nonlyrnphoblastic leukemia cells at diagnosis. Blood 1992;79(2):473-6. Baer MR, Minderman H, Frankel SR, Herzig GP, Rustum Y, Stewart CC, et al. High frequency of CD34 and MDRl expression in acute myeloid leukemia evolved from antecedent bone marrow disorders. Proc A m Soc Clin Oncol 1993; (12 ):315. Juliano RL, Ling V. A surface glycoprotein modulating drug permeability in Chinese hamster ovary cell mutants. Biockim Biophys Acta 1976;455:152. Endiccltt JA, Ling V. The biochemistry of P-glycoprotein mediated multidrug resistance. Annu Rev Biochem 1989;58: 137. Willman CL, Kopecky KJ, Appelbaum F, Grever MR, Head DR, Elias L, et al. Biologic parameters that predict treatment response in de novo acute myeloid leukemia (AML):CD34, but not multidrug resistance (MDR) gene expression is associated with a decreased complete remission (CR) rate and CD34-positive patients more frequently achieve CR with high-dose cytosine arabinoside. Proc A m SOC Clin Oncol 1992;11262. Jamali MAA, Yin M-B, Mazzoni A, Bankusli I, Rustum YM. Relationship between cytotoxicity, drug accumulation, DNA damage and repair of human ovarian cancer cells treated with doxorubicin: modulation by the tiapamil analog Roll2933. Cancer Chemotker Pkarmacol 1989;25:77-83. Soule HD, Vasquez J, Long A, Brennan M. A human cell line from a pleural effusion derived from a breast carcinoma. J Natl Cancer Inst 1973;51:1409-16. Fairchild CR, Ivy SP, Kao-shan C-S, Whang-Peng J, Rosen N, Israel MA, et aI. Isolation of amplified and overexpression ~ 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. ~ 2435 DNA sequences from adriamycin-resistant human breast cancer cells. Cancer Res 1987;47:5141-8. 22. Akiyama S, Fojo A, Hanover JA, Pastan I, Gottesman MM. Isolation and characterization of human KB cell lines resistant to multiple dmgs. Somat Cell A401 Genet 1985;11:117-26. 23. Bhalla K, Hindenburg A, Taub RN, Grant S. Isolation and characterization of an anthracycline-resistant human leukemic cell line. Cancer Res 1985;45:3657-62. 24. Marsh W, Sicheri R, Center MS. Isolation and characterization of adriamycin-resistant HL60 cells which are not defective in the initial intracellular accumulation of drug. Cancer Res 1986;46:4053-7. 25. McGrath T, Latoud C, Arnold ST, Safa AR, Felsted RL, Center MS. Mechanisms of multidrug resistance in HL60 cells: analysis of resistance associated membrane proteins and levels of MDR gene expression. Biochem Pkarmacol 1989;38:3611-9. 26. l McGrath ~ T, Center ~ MS. Adriamycin d ~ resistance In HL60 cells in the abscence of detectable P-glycoprotein, Biochem Biopkys Res Commun 1987;145:1171-6. 27. Marsh W, Center MS. Adriamycin resistance in HL60 cells and accompanying modification of a surface membrane protein contained in drug-sensitive cells. Cancer Res 1987;475080-6. 28. Slovak ML, Hoeltge GA, Dalton WS, Trent JM. Pharmacological and biological evidence for differing mechanisms of doxorubicin resistance in two tumor cell lines. Cancer Res 1988;48:27937. 29. Greco WR, Park HS, Rustum YM. An application of a new approach for the quantitation of drug synergism to the combination of cis-diamminedichloroplatnum and 1-beta-D-arabinofuranosylcytosine. Cancer Res 1990;50:5318-27. 30. Toth K, Vaughan MM, Slocum HK, Arredondo MA, Takita H, Baker RM, et al. New immunohistochemical “sandwich” staining method for mdrl P-Glycoprotein detection with JSB-1 monoclonal antibody in formalin-fixed, paraffin-embedded human tissues. A m J Patkol 1994;144(2):227-36. 31. Scheper RJ, Broxterman HJ, Scheffer GL, Kaaijk P, Dalton WS, van Heijningen HM, et al. Overexpression of a Mr 110,000 vesicular protein in non-P-glycoprotein-mediated multidrug resistance. Cancer Res 1993;53:1475-80. 32. Flens MJ, Izquierdo MA, Scheffer GL, Fritz JM, Meijer CILM, Scheper RJ, et al. Immunochemical detection of the multidrug resistance-associated protein MRP in human multidrug-resistant tumor cells by monoclonal antibodies. Cancer Res 1994;54:4557-64. 33. Andrews RG, Singer JW, Bernstein ID. Monoclonal antibody 12-8 recognizes a 115-kd molecule present on both unipotent and multipotent hematopoietic colony-forming cells and their precursors. Blood 1986;67:842-5. 34. Habig WH, Pabst MJ, Jakoby WB. Glutathione s-transferase. The first enzymatic step in mercapturic acid formation. J Biol Chem 1974;249:7 130-9. 35. Giinzler WA, Kremers H, Flohe L. An improved coupled test procedure for glutathione peroxidase (EC 184.108.40.206) in blood. Z Klin Ckem Klin Biochem 1974;12444-8. 36. Smith IK, Vierheller TL, Thorne CA. Assays of glutathione reductase in crude tissue homogenates using 5,5’-dithiobis (2-nitrobenzoic acid). Anal Biochem 1988;175:408-13. 37. Deutsch J. Glucose-6-phosphate dehydrogenase. In: Bergmeyer UB, Bergmeyer J, Grass1 M, editors. Methods of enzymatic analysis. 3rd edition. Deerfield Beach, FL.: Verlag Chemie, 1983:190-7. 38. Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 1987;162156-9. 2436 CANCER December 1, 1996 / Volume 78 / Number 11 39. Tsai-Pflugfelder M, Liu LF, Liu AA, Tewey KM, Whang-Peng I, Knutsen T, et al. Cloning and sequencing of cDNA encoding human DNA topoisomerase I1 and localization of the gene to chromosome region 7q21-22. ProcNatZAcad Sci USA 1988;85:7177-81. 40. Harker WG, Slade DL, Parr RL, Feldhoff PW, Sullivan DM, Holguin MH. Alterations in the topoisomerase-alpha gene, messenger RNA, and subcellular protein distribution as well as reduced expression of the DNA topoisomerase-beta enzyme in a mitoxantrone-resistant HL60 human leukemia cell line. Cancer Res 1995;55:1707-16. 41. Campos L, Guyotat D, Archimbaud E, Devaux Y, Treille D, Larese A, et al. Surface marker expression in adult acute myeloid 1eukemia:correlations with initial characteristics, morphology and response to therapy. Br J Haematol 1989; 72:161-6. 42. Borowitz MJ, Gockerman JP, Moore JO, Civin CI, Page SA, Robertson J, et al. Clinicopathologic and cytogeneticfeatures of CD34 [MYlO)-positive acute nonlymphocytic leukemia. A m J Clin PathoZ 1989;91:265-70. 43. Geller RB, Zahurak M, Hurwitz CA,Burcke PJ, Karp JE, Pian- 44. 45. 46. 47. 48. tadosi S, et al. Prognostic importance of immunophenotyping in adults with acute myelocytic leukemia: the significance of the stem-cell glycoprotein CD34 (MY-10). Br 1 Haematol 1990;76:340-7. Baumhueter S, Singer MS, Henzel W, Hemmerich S, Renz M, Rosed SD, et al. Binding of L-selectin to the vascular sialomucin CD34. Science 1993;262:436-8. O’Kane Murphy B, Jackson ID, Bociek G, Costas P, Korth D, Kuszynski C, et al. CD34 analysis of GM-CSF mobilized peripheral blood stem cells. Blood 1993;82(10):488a. Niimi S, Nakagawa K, Yokota J, Tsunokawa Y, Nishio K, Terashima Y, et al. Resistance to anticancer drugs in NIH3T3 cells transfected with c-myc andlor c-H-ras genes. BrJ Cuncer 1991;63:237-41. Baer MR, Augustinos P, Kinniburgh AJ. Defective c-myc and c-myb RNA turnover in acute myeloid leukemia cells. Blood 1992;79(5):1319-26. Kastan MB, Stone KD, Civin C. Nuclear oncoprotein expression as a function of lineage, differentiation stage, and proliferative status of normal human hematopoietic cells. Blood 1989;74: 1517-24.