Cell-Specific Modulation of Drug Resistance in Acute Myeloid

Cell-Specific Modulation of Drug Resistance in Acute Myeloid. Leukemic Blasts by Diphtheria Fusion Toxin, DT388-GMCSF. Arthur E. Frankel,*,† Philip ...
0 downloads 0 Views 117KB Size
Download PDF
490

Bioconjugate Chem. 1998, 9, 490−496

Cell-Specific Modulation of Drug Resistance in Acute Myeloid Leukemic Blasts by Diphtheria Fusion Toxin, DT388-GMCSF Arthur E. Frankel,*,† Philip D. Hall,‡ Chris McLain,† Ahmad R. Safa,§ Edward P. Tagge,| and Robert J. Kreitman⊥ Departments of Medicine, Surgery, Experimental Oncology, and Pharmaceutical Sciences, Medical University of South Carolina, Charleston, South Carolina 29425, and Laboratory of Molecular Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20892. Received February 4, 1998; Revised Manuscript Received April 10, 1998

Radiochemotherapy-resistant blasts commonly cause treatment failure in acute myeloid leukemia (AML), and their resistance is due, in part, to overexpression of multidrug resistance (mdr) proteins. We reasoned that targeted delivery of protein synthesis inactivating toxins to leukemic blasts would reduce the cellular concentrations of relatively short half-life resistance proteins and sensitize the cells to cytotoxic drugs. To test this hypothesis, we employed human granulocyte-macrophage colonystimulating factor fused to truncated diphtheria toxin (DT388-GMCSF). The human AML cell line HL60 and its vincristine-resistant sublines, HL60Vinc and HL60VCR, were incubated in vitro for 24 h with varying concentrations of toxin. Doxorubicin was added for an additional 24 h, and cell cytotoxicity was assayed by thymidine incorporation and colony formation in semisolid medium. DT388-GMCSF sensitized HL60Vinc and HL60VCR but not HL60 to doxorubicin. Combination indices for three log cell kill varied from 0.2 to 0.3. In contrast, pretreatment with doxorubicin followed by toxins failed to show synergy. At least in the case of the vincristine-resistant cell lines, modulation of drug resistance correlated with reduction in membrane P-glycoprotein concentrations based on immunoblots with C219 antibody, flow cytometry with MRK16 antibody, and cell uptake of doxorubicin. These observations suggest clinical trials of combination therapy may be warranted in patients with refractory AML. Further, targeted toxins may represent a novel class of cell-specific modulators of drug resistance for a number of malignancies.

INTRODUCTION

There were 9200 new cases of acute myeloid leukemia (AML) in the U.S. in 1996 (1), and, even with complete remission rates of 50-70% and effective consolidation and intensification regimens including allogeneic bone marrow transplantation, over 80% of patients will die from complications of the disease and its treatment (2). Radiochemotherapy resistant blasts are a frequent cause of treatment failure in AML patients (3). In many cases, multidrug resistance (mdr) is due in part to overexpression of resistance proteins which alter drug accumulation, metabolism, and apoptotic thresholds. Modification of resistance protein function has been hampered by the finding that many resistance proteins are expressed by normal human tissues, and noncell specific modulation leads to significant toxicities. For example, the substrate analogues (quinine, cyclosporine, and PSC833) of the P-glycoprotein drug efflux transporter resistance protein have been tested in AML therapy trials and have produced toxicities to marrow and other organs, marked alterations in cytotoxic drug pharmacodynamics and minimal effects on disease-free survival or overall sur* Address correspondence to this author at the following address: Hollings Cancer Center Room 311, 86 Jonathan Lucas St., Charleston, SC 29401. Tel: 803-792-1450. Fax 803-7923200. † Department of Medicine. ‡ Department of Pharmaceutical Sciences. § Department of Experimental Oncology. | Department of Surgery. ⊥ Laboratory of Molecular Biology.

vival (4). Hence, there is a need for new reagents with a unique mechanism of action to modulate drug resistance specifically in AML blasts. Three resistance proteins (P-glycoprotein, glutathione S-transferase π, and Bcl2) have been shown to have rapid turnover (24 h or less) (5-7), and the blast concentrations of these proteins correlate with clinical response rates and remission durations (8-10). These proteins could serve as targets for cell-specific inhibitors of protein synthesis such as targeted toxins. Targeted toxins are polypeptide drugs consisting of tumor cell-directed ligands covalently linked to protein synthesis inactivating peptide toxins. A number of studies both in vitro and in vivo have shown supraadditive cell killing with combinations of immunotoxins and cytotoxic drugs (11-20). Anti-CD19ricin A chain and anti-CD22-ricin A chain conjugates with either cyclophosphamide, doxorubicin, or camptothecin incubated with Burkitt’s lymphoma cells show synergistic cell kill (11). Similarly, anti-CD19-blocked ricin conjugate combined with etoposide or doxorubicin yielded synergistic cell killing of a B-lymphoma cell line and a retrovirally transfected B-lymphoma cell line expressing mdr-1 (12). Finally, anti-p43 epithelial antigenrecombinant ricin A chain conjugate combined with thiotepa or cisplatin showed synergistic cell killing of ovarian carcinoma cell lines (13). When anti-CD19-ricin A chain and anti-CD22-ricin A chain were administered with doxorubicin, cyclophosphamide, or camptothecin to scid mice bearing Burkitt’s lymphoma cells, schedule dependence for supraadditive cell killing was observed

S1043-1802(98)00015-9 CCC: $15.00 © 1998 American Chemical Society Published on Web 06/09/1998

Cell-Specific Modulation in Leukemic Blasts

(14). Only when immunotoxins were given before or at the same time as cytotoxic drugs was synergy seen. With modulating doses of anti-p43-ricin A chain conjugate on ovarian carcinoma cells and anti-CD19-blocked ricin on mdr-1 expressing B-lymphoma cells, concentrations of resistance proteins decreased (12, 13). The ovarian carcinoma cells showed decreases in glutathione Stransferase activity, and the B-lymphoma had decreases in membrane P-glycoprotein. While additional studies report coalitive interactions of immunotoxins with cytotoxic drugs in T-acute lymphoblastic leukemia (15), B-lymphoma (16), colon carcinoma (17), and B-acute lymphoblastic leukemia (18), no similar efforts to modulate drug resistance in AML with targeted toxins have been reported. We chose to use human granulocyte-macrophage colonystimulating factor fused to truncated diphtheria toxin (DT388-GMCSF) to inhibit protein synthesis in myeloid blasts. We and others had previously reported potent protein synthesis inhibition by this fusion toxin on leukemic blasts (19-21). We reasoned that short halflife resistance protein concentrations should be reduced, sensitizing the cells to cytotoxic drugs. Further, GMCSF receptors are expressed on the majority of AML patients’ blasts (24). We chose doxorubicin as the cytotoxic drug for study since anthracyclines are a critical component of AML induction regimens. The cell lines selected were HL60 and its vincristine-resistant sublines HL60Vinc and HL60VCR (25, 26). HL60 is a well-characterized human AML cell line displaying GMCSF receptors. The latter two sublines show different degrees of multidrug resistance, and each overexpresses P-glycoprotein. The goals of the present study were to examine the ability of the fusion toxin to modulate doxorubicin sensitivity in the three cell lines, determine the schedule dependency of any observed modulation, and correlate modulation with changes in P-glycoprotein concentration. We believed such a study was a requisite step prior to examination of resistance modulation with fresh leukemic blasts. MATERIALS AND METHODS

Materials. DT388-GMCSF was prepared as previously described (20) and stored at -20 °C until used. DT390mGM-CSF was a kind gift of Dr. Daniel Vallera and was stored at -20 °C until used (38). Cell Culture. HL60 human leukemia cells were a gift from M. Center, Kansas State University. The cells were grown in RPMI1640 plus 15% fetal calf serum. All media and components were purchased from Irvine Scientific (Santa Ana, CA). HL60Vinc, a vincristineresistant subline of HL60 (26), was a gift from M. Center and maintained in RPMI1640 plus 15% fetal calf serum. The HL60VCR vincristine-resistant subline was developed by stepwise exposure of HL60Vinc to increasing concentrations of vincristine (A. R. Safa, unpublished data) and maintained in RPMI1640 plus 15% fetal calf serum plus 1 µg/mL vincristine (Eli Lilly, Indianapolis, IN). WT19 mouse leukemia cells were a gift of Dr. Andrew Kraft, University of Colorado Medical Center (41). The cells were grown in RPMI1640 plus 15% fetal calf serum with 10% WEHI cell conditioned medium. GMCSF Receptor Density. Ten aliquots of two million cells in RPMI1640 plus 2.5% bovine serum albumin (BSA) plus 0.2% sodium azide plus 20 mM Hepes, pH 7.2, in a total volume of 150 µL were incubated with different concentrations of [125I]GMCSF (Amersham, Arlington Heights, IL) in the presence or absence of 1500

Bioconjugate Chem., Vol. 9, No. 4, 1998 491

ng of unlabeled human GMCSF (Immunex, Seattle, WA) for 30 min at 37 °C with shaking. The cell suspensions were then layered onto 200 µL of phthalate oil mixture (1.5 parts dibutylphthalate and 1 part dioctylphthalate) in 1.5 mL Eppendorf tubes, centrifuged at 12 000 rpm in a microfuge at room temperature for 1 min and both cell pellets and cell supernatants counted in an LKBWallac 1260 multi-gamma counter gated for 125I with 50% counting efficiency. Scatchard plots of specific bound/ free versus specific bound counts per minute were made. Receptor density/cell was calculated as the value for the x-intercept divided by (specific activity in microcuries per microgram times the cell number times 4.2 × 10-8). Kd was calculated as the x-intercept times 2.7 × 10-13 divided by the product of the y-intercept and the specific activity. Thymidine Incorporation Cytotoxicity Assays. A total of 5 × 104 cells (HL60, HL60Vinc, HL60VCR, or WT19) were incubated in 100 µL RPMI1640 plus 15% fetal calf serum in Costar 96 well flat-bottomed plates. Twelve different concentrations of DT388-GMCSF or DT390mGM-CSF were added to each column in 50 µL of media and the cells maintained at 37 °C/5% CO2 for 24 h. Then, eight different concentrations of doxorubicin (Pharmacia, Kalamazoo, MI) were added to each row in a total volume of 25 µL, and incubation continued for another 24 h at 37 °C/5% CO2. Then 1 µCi [3H]thymidine (NEN DuPont, Boston, MA) in 25 µL of medium was added to each well and incubated at 37 °C/5% CO2 for an additional 18 h. Cells were then harvested by a Skatron Cell Harvestor (Skatron Instruments, Lier, Norway) onto glass fiber mats and 3H counts per minute was counted in an LKB liquid scintillation counter gated for 3H. The calculated IC90, IC99, and IC99.9 were the concentrations of toxin, doxorubicin, or combination which inhibited thymidine incorporation by 90, 99, or 99.9% compared to control wells. To determine whether the cytotoxic effect of combination treatment was additive, antagonistic, or synergistic, an analysis using the combination index (CI) for each level of cytotoxicity was calculated using the equation: CI ) A/Ae + B/Be, where A and B are the concentrations of toxin and drug that in combination kill 90, 99, or 99.9% of cells and Ae and Be are the concentrations of the toxin and drug, respectively, extrapolated from their individual killing curves capable of killing 90, 99, or 99.9% of cells alone (27). For CI > 1, antagonism is present; for CI ) 1, additivity is present; for CI < 1, synergy is present. Clonogenic Cytotoxicity Assays. To evaluate inhibition of colony formation as another measure of cell kill, 70 µL aliquots containing 2 × 104 cells were collected from the wells with toxin and doxorubicin described above after 48 h incubation. The cells were dispersed in 3 mL of RPMI1640 plus 15% fetal calf serum with 50 ng/mL GMCSF (Immunex), GCSF (Amgen, Thousand Oaks, CA) and 0.2% agarose (SeaPlaque, FMC Bioproducts, Rockland, ME) in 35 mm gridded polystyrene Petri dishes. After 20 min of solidification at 4 °C, dishes were incubated 7 days at 37 °C/5% CO2 in a humidified incubator. Colonies containing >20 cells were counted on an inverted microscope. IC90, IC99, and IC99.9 were the concentrations of toxin, doxorubicin, or combination, which inhibited colony formation by 90, 99, or 99.9% compared to control wells. CIs were calculated as for thymidine incorporation inhibition. Reversed Schedule Thymidine Assay. Aliquots of 100 µL of medium with 5 × 104 cells were incubated first with 25 µL of medium containing eight different concentrations of doxorubicin. After 24 h, 50 µL of medium

492 Bioconjugate Chem., Vol. 9, No. 4, 1998

containing 12 different concentrations of DT388-GMCSF was added, and incubation continued another 24 h at 37 °C/5% CO2. [3H]Thymidine was then added (1 µCi in 25 µL/well) and incubation and harvesting done as described above. CIs were again calculated as above. Cell Uptake of Doxorubicin. Aliquots of one million cells in 100 µL RPMI1640 plus 15% fetal calf serum with or without 50 uM verapamil (Sigma) were mixed with or without 2.8 ug/mL doxorubicin for 60 min at 37 °C/5% CO2 and then washed with PBS/BSA/azide and then resuspended in 500 µL of PBS/BSA/azide and run on the EPICS-XL with scatter gating and rhodamine fluorescence detection. P-Glycoprotein Content of Cells. A total of 5 × 106 HL60, HL60Vinc or HL60VCR cells in 10 mL of RPMI1640 plus 15% fetal calf serum with or without modulating concentrations of toxin (4 × 10-9 M DT388GMCSF) were incubated 48 h at 37 °C/5% CO2. Cells were then harvested by centrifugation at 1000 rpm for 7 min, washed twice with 10 mL of PBS, and resuspended in 10 mM NaCl, 10 mM Tris- HCl, pH 8.0, 1 mM PMSF, and 50 µg/mL leupeptin, and left on ice for 10 min. Cells were then sonicated, and nuclei and debri removed by centrifugation at 2500 rpm for 15 min. Membranes were isolated by centrifugation at 25 000 rpm for 60 min and resuspended in 1 mM Tris, pH 8.0, 0.1 mM EDTA, 1 mM PMSF, and 50 ng/mL leupeptin. Protein concentrations were determined by Bradford assay using the Bio-Rad protein assay kit and BSA standards. Aliquots equivalent to 1.3 × 106 cells (50-100 µg of protein) were diluted in reducing sample buffer and subjected to 8% reducing SDS-PAGE along with Bio-Rad kaleidoscope high molecular weight protein standards. Proteins were transferred to nitrocellulose sheets using a Bio-Rad Semi-dry Transblot apparatus in Towbin buffer following the recommendations of the manufacturer. The blots were blocked with 10% nonfat dry milk solution/PBS/0.1% Tween-20/0.1% BSA/0.1% sodium azide for 3 h at room temperature with gentle shaking, washed three times with PBS plus 0.05% Tween-20 and once with PBS, incubated with 2 ug/mL C219 anti-P-gp monoclonal antibody (Signet Labs, Dedham, MA) or 1:1000 rabbit anti-GLUT1 polyclonal antibody (a kind gift of Dr. Ian Simpson, NIDDK) in PBS plus 0.05% BSA overnight, rewashed, and incubated with 1:2000 goat anti-mouse Ig conujugated to horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA) or 1:3000 donkey anti-rabbit Ig conjugated to horseradish peroxidase (Amersham). The blots were again washed, and immune complexes detected by chemiluminescence (Amersham) following the manufacturer’s recommendations. Lumograms were scanned on an IBAS automatic image analysis system (Kontron, Germany) and relevant band intensities compared. KB-V1 vincristine-resistant epidermoid carcinoma cells were used as a control for P-gp and were a gift of Dr. Mark Willingham, Wake Forest University, Winston-Salem, NC. A separate assay for the concentration of P-glycoprotein on the surface of leukemic blasts was performed by flow cytometry as previously described (39). Briefly, one million HL60 and HL60VCR cells in 5 mL of RPMI1640 plus 15% fetal calf serum with or without 4 × 10-8 M DT388-GMCSF were incubated 60 h at 37 °C/5% CO2. Cells were harvested by centrifugation at 1000 rpm for 7 min, washed once with 0.5 mL of Hank’s balanced salt solution (Gibco) with 1% fetal calf serum and 0.05% sodium azide (HFN), reacted in 100 µL of HFN with or without 4 µL MRK16 antibody (Kamiya, Seattle, WA) for 20 min on ice, rewashed, resuspended in 100 µL HPN

Frankel et al.

with 2 µL biotinylated goat anti-mouse Ig (Southern Biotech, Birmingham, AL), and incubated for 20 min on ice. The cells were washed again, resuspended in 1 mL of goat serum for 10 min on ice, rewashed in HFN, resuspended in 100 µL of HFN plus 1.5 µL of Texas Redstrepavidin (Molecular Probes, Eugene, OR), and incubated 20 min on ice. Finally, the cells were washed and suspended in 1 mL of PBS plus 0.2% bovine serum albumin and 0.01% sodium azide and run on an EPICS XL flow cytometer (Coulter, Hialeah, FL) with gating for live cells and detection of the FL2 fluorescence. Background fluorescence with isotype matched control antibody was subtracted. RESULTS

To evaluate the ability of targeted toxins to modulate doxorubicin resistance in AML cell lines, we first had to define the variables that affect sensitivity of AML blasts to toxins and drug. In this study, parameters that we evaluated included the cell surface GMCSF receptor density and the amounts and function of P-glycoprotein on the three cell lines. GMCSF Receptor Density and Ligand Affinity. Scatchard plots of [125I]GMCSF binding to AML cell lines revealed high-affinity receptor densities of 320/cell, 123/ cell, and 172/cell for HL60, HL60Vinc, and HL60VCR based on linear regression (Figure 1). The Kds were calculated as 2 × 10-11, 1.4 × 10-11, and 8 × 10-12 M for HL60, HL60Vinc, and HL60VCR, respectively. P-Glycoprotein Concentrations. Immunoblots with C219 included KB-V1 control and showed relative densities for HL60, HL60Vinc, and HL60VCR of 0.07, 0.28, and 2.75 based on equal cell loading on SDS-PAGE (Figure 2 and Table 5). Immunoblots of control protein, GLUT1, revealed bands at 45 and 55 kDa for HL60, HL60Vinc, and HL60VCR and bands at 43 and 53 kDa for KB-V1. Flow cytometry of HL60 and HL60VCR cells for surface P-gp showed minimal P-gp on HL60 cells (20% above background) and marked P-gp on HL60VCR cells (360% above background). Doxorubicin Uptake by AML Cell Lines. The mean channel rhodamine fluorescence of AML cell lines increased from 3 without doxorubicin incubation (autofluorescence) to 15 for HL60 cells, 10 for HL60Vinc, and 7 for HL60VCR cells with 1 h incubation with 5 × 10-6 M doxorubicin. Cell Cytotoxicity Measured by Thymidine Incorporation of DT388-GMCSF, DT390mGM-CSF, and Doxorubicin Alone. Table 1 shows the IC90, IC99, and IC99.9 for inhibition of thymidine incorporation after one or two day incubations with DT388-GMCSF and doxorubicin alone for each of the three AML cell lines. Doxorubicin sensitivity is most for HL60 cells, intermediate for HL60Vinc cells, and least for HL60VCR cells. Forty-eight hour incubation increased toxicity of DT388GMCSF but not doxorubicin. DT390mGM-CSF was nontoxic to HL60 and HL60VCR cells (IC50 > 4 × 10-8 M), but showed potent cytotoxicity to WT19 mouse leukemia cells (IC50 ) 5 × 10-12 M). Cell Cytotoxicity Measured by Clonogenic Assay with DT388-GMCSF and Doxorobucin Alone. Table 2 displays the IC90, IC99, and IC99.9 for inhibition of colony formation in semisolid medium for DT388-GMCSF and doxorubicin on each of the AML cell lines. There was an excellent correlation with the 48 h thymidine incorporation inhibition assay results. Modulation of Doxorubicin Resistance with DT388-GMCSF. The effects of preincubation with

Cell-Specific Modulation in Leukemic Blasts

Bioconjugate Chem., Vol. 9, No. 4, 1998 493

Figure 1. Scatchard plots of [125I]GMCSF binding to AML cells. The experiment was performed as described in the text. The specific activity of the [125I]GMCSF used was 97 µCi/µg at the test date. A, HL60; B, HL60Vinc; C, HL60VCR. Kds were 2 × 10-11 M, 1.4 × 10-11 M, and 8 × 10-12 M for HL60, HL60Vinc, and HL60VCR, respectively. Table 1. Inhibition of Thymidine Incorporation of AML Cell Lines by DT388-GMCSF and Doxorubicina 24 h incubation (M) toxin/drug DT388-GMCSF IC90 IC99 IC99.9 doxorubicin IC90 IC99 IC99.9

48 h incubation (M)

HL60

HL60Vinc

HL60VCR

HL60

HL60Vinc

HL60VCR

4 × 10-11 6 × 10-8 4 × 10-7

1 × 10-8 8 × 10-8 4 × 10-7

4 × 10-7 1 × 10-6 3 × 10-6

6 × 10-11 6 × 10-9 4 × 10-8

2 × 10-9 4 × 10-8 4 × 10-7

1 × 10-10 5 × 10-9 4 × 10-8

7 × 10-8 2 × 10-7 5 × 10-7

1 × 10-6 1 × 10-5 5 × 10-5

5 × 10-6 2 × 10-4 3 × 10-4

2 × 10-8 2 × 10-7 7 × 10-7

7 × 10-7 3 × 10-6 5 × 10-5

5 × 10-6 9 × 10-5 3 × 10-4

a Each assay performed at least twice. IC values above 4 × 10-8 M for DT388-GMCSF and 2 × 10-4 M for doxorubicin projected from curves. Methods as in text.

Table 2. Inhibition of Colony Formation by AML Cell Lines by DT388-GMCSF and Doxorubicina (M) toxin/drug DT388-GMCSF IC90 IC99 IC99.9 doxorubicin IC90 IC99 IC99.9

Figure 2. Immunoblot of AML cell membrane extracts reacted with C219 monoclonal anti-P-glycoprotein antibody. The experimental method is described in the text. Kladeiscope high molecular mass marker migration is depicted by arrows (203, 126, and 71 kDa). Lane 1, KB-V1 cells; lane 2, HL60VCR cells; lane 3, HL60Vinc cells; lane 4, HL60 cells.

DT388-GMCSF on doxorubicin sensitivity of thymidine incorporation and colony formation for the AML cell lines was analyzed by isobolograms (Figure 3) and calculation of the combination index (CI). CIs for colony inhibition and thymidine incorporation inhibition were similar. Modulation was greater for three log cell kill than one or two log cell kill. Concave isoboles were observed for three log cell kill of HL60Vinc and HL60VCR. CIs for

HL60

HL60Vinc

HL60VCR

8 × 10-12 4 × 10-10 4 × 10-8

2 × 10-9 4 × 10-9 2 × 10-8

5 × 10-10 6 × 10-9 6 × 10-8

5 × 10-8 9 × 10-8 1 × 10-7

5 × 10-7 1 × 10-6 5 × 10-6

9 × 10-6 2 × 10-5 4 × 10-5

a Assays performed as described in text. IC values above 4 × 10-8 M for DT388-GMCSF, and 2 × 10-4 M for doxorubicin extrapolated from curves.

the different cell lines are shown in Table 3. In each case, the CI was less for three log cell kill than one log cell kill. The scheduling of toxin and drug were critical. Higher CIs at three log cell kill were observed when drug treatment preceded toxin exposure except in the case of HL60 cells (Table 4). Effects of Modulating Concentrations of DT388GMCSF on P-gp Concentrations. At modulating concentrations of toxins (4 × 10-9 M DT388-GMCSF), the concentrations of P-gp in HL60Vinc and HL60VCR were reduced 6-27-fold relative to untreated cells based on

494 Bioconjugate Chem., Vol. 9, No. 4, 1998

Frankel et al.

Table 3. Combination Indices for AML Cell Cytotoxicity of DT388-GMCSF and Doxorubicina cell line HL60 IC90 IC99 IC99.9 HL60Vinc IC90 IC99 IC99.9 HL60VCR IC90 IC99 IC99.9 a

thymidine incorporation

colony formation

0.7 0.8 1

1 0.7 0.5

0.4 0.2 0.2

1 0.2 0.2

0.6 0.2 0.2

0.5 0.4 0.3

Assays and calculations of CIs as described in text.

Table 4. Schedule Dependency of Modulationa cell line HL60 IC90 IC99 IC99.9 HL60Vinc IC90 IC99 IC99.9 HL60VCR IC90 IC99 IC99.9

ratio CI drug first/ CI toxin first 1.4 0.5 0.4 1.0 0.7 5.0

Figure 3. Isobologram for concentrations of DT388-GMCSF and doxorubicin yielding a 99% reduction in thymidine incorporation of HL60VCR cells. Cells were incubated for 24 h with DT388-GMCSF and then doxorubicin added for an additional 24 h, followed by an 18 h labeling with [3H]thymidine, harvesting and counting. Thymidine incorporation was compared to untreated cells. The concentration of DT388-GMCSF alone producing 99% reduction in thymidine incorporation, Ae, was 4 × 10-8 M. The concentration of doxorubicin yielding a 99% reduction in thymidine incorporation, Be, was projected at 130 µg/mL. The dotted line shows the additivity isobole. Concave lines beneath the additivity isobole show synergy, and convex lines above the additivity isobole show antagonism. The CI was 0.2.

1.3 5.0 5.0

a Values for CI with toxin first or drug first obtained as described in text. IC99.9 ratios of CI for HL60Vinc and HL60VCR were reproducibly lower than IC99.9 ratios of CI for HL60.

Table 5. P-Glycoprotein Concentrations at Modulating Doses of DT388-GMCSFa cell line

ratio density with and without toxin

HL60Vinc HL60VCR

0.17 0.06

a Immunoblots developed with C219 anti-P-gp monoclonal antibody and calibrated with 1.3 × 106 cell membrane extract of untreated cells. Modulating concentrations were 4 × 10-9 M for DT388-GMCSF for 48 h incubation. HL60 cells had no detectable P-gp under all incubation conditions.

densitometry of immunoblots (Figure 4 and Table 5). Control protein GLUT1 concentrations were altered by toxins under the conditions producing maximal modulation of drug resistance but to a much lesser degree (Figure 5). Flow cytometry showed a 2-fold reduction in P-gp surface membrane concentration of HL60VCR cells after fusion toxin exposure. Examination of toxin- and doxorubicin-treated apoptotic cell populations revealed that toxin pretreatment was associated with increased intracellular concentrations of doxorubicin (Table 6). DISCUSSION

Novel treatment strategies for patients with relapsed or refractory AML have focused on circumventing multidrug resistance. Allogeneic bone marrow transplant has produced a significant fraction of durable remission in patients less than 55 years old, in part due to a graftversus-leukemia effect (28). However, most AML patients are older than 55 years old. Chemotherapeutic agents which are not substrates for the common resis-

Figure 4. Immunoblot of AML cell membrane extracts reacted with C219 antibody. Standard protein markers were the same as Figure 4. Lane 1, HL60VCR control; lane 2, HL60VCR treated with 4 × 10-9 M DT388-GMCSF for 48 h; lane 3, HL60Vinc control; lane 4, same as lane 2 with HL60Vinc.

tance phenotypes have been tested. The nucleoside analogue fludarabine produced responses in relapsed patients in combination with cytosine arabinoside (29). The topoisomerase I inhibitor topotecan showed activity as a salvage agent alone and with cytarabine, but at the cost of moderate to severe oropharyngeal mucositis (30). The hypomethylating agent decitabine yielded remissions in drug-resistant AML patients, but the patients had prolonged myelosuppression (31). Finally, nonspecific modulators of P-gp such as cyclosporine have been combined with anthracyclines as salvage therapy (32). Responses were seen, but the modulation of P-gp in liver and kidney led to altered drug pharmacokinetics and enhanced toxicities to marrow, kidney, and central nervous system. The novel antibody-drug conjugate, anti-CD33-calicheamicin, showed minimal toxicities with

Cell-Specific Modulation in Leukemic Blasts

Bioconjugate Chem., Vol. 9, No. 4, 1998 495

Figure 5. Immunoblot of AML membrane extracts detected with anti-GLUT antibody as described in the text. Low molecular mass prestained protein standards (105, 71, and 41.8 kDa) are shown by arrows. Membrane extracts in lanes 1-4 are the same as in Figure 4. Relative densities of the 45 kDa GLUT1 bands were 1, 1.5, 1, and 0.8, respectively, for lanes 1-4. Table 6. Doxorubicin Intracellular Concentrations in AML Cell Lines mean rhodamine fluorescence channela cell line HL60

condition

control DT388-GMCSF HL60Vinc control DT388-GMCSF HL60VCR control DT388-GMCSF

media doxorubicin + alone doxorubicin verapamil 3 4 4 3 3 4

15 34 10 17 7 61

15 32 13 25 19 68

a Cells were incubated for 60 min with 5 µM doxorubicin with or without 50 µM verapamil, washed, and run on an EPICS-XL gated for apoptotic cells by forward and 90° light scatter. Modulating concentrations of DT388-GMCSF as described in Table 5.

complete remissions in 4/35 patients, but most patients failed to respond due to active drug efflux (35). Thus, new reagents with unique mechanisms of action which can directly modify the apoptotic threshold of resistant malignant blasts are needed. Targeted toxins are a novel class of therapeutics which may be useful for relapsed/refractory AML. Previous clinical trials of targeted toxins in T-cell and B-cell lymphomas have shown a number of partial and complete responses; however, clinical benefit has been limited by toxicity to vascular endothelium (33). We had previously engineered a targeted toxin for AML, DT388-GMCSF (19, 20), but we anticipated vascular leak syndrome as the dose-limiting toxicity in man (34). We sought to overcome both the problem of drugresistant leukemic blasts and endothelial injury of targeted toxins by administering combinations of targeted toxins and cytotoxic drugs. We hypothesized that targeted toxins would be synergistic with cytotoxic drugs based on previous studies in vitro with lymphoma and ovarian carcinoma cells (11-14). Thus, with combination therapy, lower doses of both targeted toxin and cytotoxic drug would be needed. Both vascular leak and some of the toxicities of cytotoxic drugs may be ameliorated. To test our hypothesis, we performed a systematic study of the optimal delivery of DT388-GMCSF with doxorubicin on three AML cell lines.

The first interesting finding was the variations in sensitivity to fusion toxin independent of receptor status. The resistance of the HL60Vinc cells to DT388-GMCSF may be due to antiapoptotic protein expression as we have previously reported (19). The results of the clonogenic assays in our study paralleled the results of the thymidine assays. Specificity was confirmed by the lack of toxicity of the diphtheria fusion toxin, DT390mGMCSF, for HL60 and HL60VCR cells. Synergy was observed primarily at higher log cell kill (IC99.9) reflecting a mechanism requiring significant inhibition of protein synthesis. This observation is consistent with the requirement to reduce the intracellular concentrations of some short half-life proteins to modulate resistance. The less pronounced synergy on HL60 cells versus HL60Vinc and HL60VCR suggests a significant impact of mdr-1 on this process. This finding is also compatible with our model of modulation by reducing short half-life resistance proteins. Further support for the model was observed with the immunoblots, flow cytometry, and doxorubicin uptake studies showing reduced antigen and functional activity for mdr-1 in the toxin modulated cells. The differences in the concentrations of P-gp after fusion toxin exposure assayed by immunoblots or flow cytometry may reflect different rates of metabolism of intracellular and cell surface protein. The dependence of synergy on toxin followed by cytotoxic drug, rather than the reverse, is also evidence for the proposed molecular mechanism. Our findings confirm the schedule dependency in vivo reported by Ghetie et al. (14). Our experiments suggest a potential for targeted toxins such as DT388-GMCSF as cell-specific modulators of drug resistance in relapsed/refractory AML patients. However, further studies will be needed to confirm these observations on fresh resistant AML blasts. Since drug resistance contributes to treatment failure for a number of human malignancies (36) and targeted toxins have been designed for many different cell types (37), this approach, both as detailed in this study and in previous reports (11-20), may yield additional cancer therapeutic strategies. ACKNOWLEDGMENT

We wish to thank James Nicholson for imaging analysis and Dr. Ian Simpson, NIDDK, for antibody to GLUT1. LITERATURE CITED (1) Parker, S., Tong, T., Bolden, S., and Wingo, P. (1997) Cancer statistics, 1997. CasCancer J. Clin. 47, 5-12. (2) Schiffer, C. (1997) Acute myeloid leukemia in adults. Cancer Medicine (J. Holland, E. Frei, R. Bast, D. Kufe, D. Marton, and R. Weichselbaum, Eds.) pp 2617-2633, Lea & Febiger, Philadelphia. (3) List, A. (1996) Role of multidrug resistance and its pharmacological modulation in acute myeloid leukemia. Leukemia 10, 937-944. (4) Solary, E., Witz, B., Caillot, D., Moreau, P., Desablens, B., Cahn, J. Y., Sadoun, A., Pignon, B., Berthou, C., Maloisel, F., Guyotat, D., Casassus, P., Ifrah, N., Lamy, Y., Audhuy, B., Colombat, P., and Harousseau, J. L. (1996) Combination of quinine as a potential reversing agent with mitoxantrone and cytarabine for the treatment of acute leukemias: a randomized multicenter study. Blood 88, 1198-1204. (5) Muller, C., Laurent, G., and Ling, V. (1995) P-glycoprotein stability is affected by serum deprivation and high cell density in multidrug-resistant cells. J. Cell Physiol. 163, 538-544. (6) Shen, H., Ranganathan, S., Kuzmich, S., and Tew, K. D. (1995) Influence of ethacrynic acid on glutathione S-transferase pi transcript and protein half-lives in human colon cancer cells. Biochem. Pharm. 50, 1233-1245.

496 Bioconjugate Chem., Vol. 9, No. 4, 1998 (7) Blagosklonny, M., Alvarez, M., Fojo, A., and Neckers, L. (1996) Bcl-2 protein downregulation is not required for differentiation of multidrug resistant HL60 leukemia cells. Leuk. Res. 20, 101-112. (8) Pirker, R., Wallner, J., Geissler, K., Kinkesch, W., Haas, O, Bettelheim, P., Hopfner, M., Scherrer, R., Valent, P., Havelec, L., Ludwig, H., and Lechner, K. (1991) MDR1 gene expression and treatment outcome in acute myeloid leukemia. J. Natl. Cancer Inst. 83, 708-712. (9) Russo, D., Marie, J. P., Zhou, D. C., Faussat, A. M., Melli, C., Damiani, D., Michelutti, A., Michieli, M., Fanin, R., and Baccarani, M. (1994) Evaluation of the clinical relevance of the anionic glutathione S-transferase (GST pi) and multidrug resistance (mdr-1) gene coexpression in leukemias and lymphomas. Leuk. Lymph. 15, 453-461. (10) Maung, Z., MacLean, F., Reid, M., Pearson, A., Proctor, S., Hamilton, P., and Hall, A. (1994) The relationship between bcl-2 expression and response to chemotherapy in acute leukaemia. Br. J. Haematol. 88, 105-115. (11) Ghetie, M. A., Tucker, K., Richardson, J., Uhr, J. W., and Vitetta, E. S. (1994) Eradication of minimal disease in severe combined immunodeficient mice with disseminated Daudi lymphoma using chemotherapy and an immunotoxin cocktail. Blood 84, 702-718. (12) O’Connor, R., Liu, C., Ferris, C. A., Guild, B. C., Teicher, B. A., Corvi, C., Liu, Y., Arceci, R. J., Goldmacher, V. S., Lambert, J. M., and Blattler, W. A. (1995) Anti-B4-blocked ricin synergizes with doxorubicin and etoposide on multidrugresistant and drug-sensitive tumors. Blood 86, 4286-4294. (13) Lidor, Y. J., O’Briant, K. C., Xu, F. J., Hamilton, T. C., Ozols, R. F., and Bast, R. C. Alkylating agents and immunotoxins exert synergistic cytotoxic activity against ovarian cancer cells. (1993). J. Clin. Invest. 92, 2440-2453. (14) Ghetie, M. A., Podar, E. M., Gordon, B. E., Pantazis, P., Uhr, J. W., and Vitetta, E. (1996) Combination immunotoxin treatment and chemotherapy in scid mice with advanced, disseminated Daudi lymphoma. Int. J. Cancer 68, 93-105. (15) Uckun, F., Gajl-Peczalska, K., Myers, D., Ramsay, N., Kersey, J., Colvin, M., and Vallera, D. (1987) Marrow purging in autologous bone marrow transplantation for T-lineage acute lymphoblastic leukemia: efficacy of ex vivo treatment with immunotoxins and 4-hydroperoxy-cyclophosphamide against fresh leukemic marrow progenitor cells. Blood 69, 361-377. (16) Weil-Hillman, G., Uckun, F., Manske, J., and Vallera, D. (1987) Combined immunochemotherapy of human solid tumors in nude mice. Cancer Res. 47, 579-588. (17) Pearson, J., FitzGerald, D., Willingham, M., Wiltrout, R., Pastan, I., and Longo, D. (1989) Chemoimmunotoxin therapy against a human colon tumor (HT-29) xenograft in nude mice. Cancer Res. 49, 3562-3571. (18) Jansen, B., Kersey, J., Jaszcz, W., Gunther, R., Nguyen, D., Chelstrom, L., Tuelahlgren, L., and Uckun, F. (1993) Effective immunochemotherapy of human t(4;11) leukemia in mice with severe combined immunodeficiency (SCID) using B43 (anti-CD19)-pokeweed antiviral protein immunotoxin plus cyclophosphamide. Leukemia 7, 290-301. (19) Frankel, A. E., Hall, P. D., Burbage, C., Vesely, J., Willingham, M., Bhalla, K., and Kreitman, R. J. (1997) Modulation of the apoptotic response of human myeloid leukemia cells to a diphtheria toxin-granulocyte-macrophage colony stimulating factor (GMCSF) fusion protein. Blood 90, 3654-3661. (20) Kreitman, R. J., and Pastan, I. (1997) Recombinant toxins containing human GM-CSF and either Pseudomonas exotoxin or diphtheria toxin kill gastrointestinal cancer and leukemia cells. Blood 90, 242-250. (21) Perentesis, J., Waddick, K., Bendel, A., Shao, Y., Warman, B., Chandan-Langlie, M., and Uckun, F. M. (1997) Induction of apoptosis in multidrug-resistant and radiation-resistant acute myeloid leukemia cells by a recombinant fusion toxin directed against the human granulocyte-macrophage colonystimulating factor receptor. Clin. Cancer Res. 3, 347-360. (22) Roy, D., Griffin, J., Belvin, M., Blattler, W. Lambert, J., and Ritz, J. Anti-MY9-blocked ricin: an immunotoxin for

Frankel et al. selective targeting of acute myeloid leukemia cells. (1991) Blood 77, 2404-2410. (23) Kabral, A., Bradstock, K. F., Grimsley, P., Favaloro, E. J. (1988) Immunophenotype of clonogenic cells in myeloid leukemia. Leuk. Res. 12, 51-58. (24) Kelleher, C., Wong, G., Clark, S., Schendel, P., Minden, M., and McCulloch, E. (1988) Binding of iodinated recombinant human GM-CSF to the blast cells of acute myeloblastic leukemia. Leukemia 2, 211-231. (25) Dalton, W. T., Ahearn, M., McCredie, K. B., Freireich, E. J., Stass, S. A., and Trujillo, J. M. (1988) HL-60 cell line was derived from a patient with FAB-M2 and not FAB-M3. Blood 71, 242-245. (26) McGrath, T., Latoud, C., Arnold, S. T., Safa, A. R., Felsted, R. L., and Center, M. S.(1989) Mechanism of multidrug resistance of HL60 cells. Analysis of resistance associated membrane proteins and levels of mdr gene expression. Biochem. Pharm. 38, 3611-3618. (27) Berenbaum, M. J. (1989) What is synergy? Pharm. Rev. 41, 93-123. (28) Forman, S. J., Schmidt, G. M., and Nademanee, A. P. (1991) Allogeneic marrow transplantation as therapy for primary induction failure of patients with acute leukemia. J. Clin. Oncol. 9, 1570-1585. (29) Gandhi, V., Estey, E., Keating, M. J., and Plunkett, W. (1993) Fludarabine potentiates metabolism of cytarabine in patients with acute myelogenous leukemia during chemotherapy. J. Clin. Oncol. 11, 116-129. (30) Seiter, K., Feldman, E. J., Halicka, H. D., Traganos, F., Darzynkiewicz, Z., Lake, D., and Ahmed, T. (1997) Phase I clinical and laboratory evaluation of topotecan and cytarabine in patients with acute leukemia.J Clin Oncol 15, 44-59. (31) Kantarjian, H. M., Estey, E. H., and Keating, M. A. (1996) New chemotherapeutic agents in acute myeloid leukemia. Leukemia 10 (Suppl. 1), S4-S8. (32) List, A. F., Spier, C., Greer, J., Wolff, S., Hutter, J., Dorr, R., Salmon, S., Futscher, B., Baier, M., and Dalton, W. (1993) Phase I/II trial of cyclosporine as a chemotherapy-resistance modifier in acute leukemia. J. Clin. Oncol. 11, 1652-1667. (33) Frankel, A. E., FitzGerald, D., Siegall, C., and Press, O. W. (1996) Advances in immunotoxin biology and therapy: a summary of the fourth international symposium on immunotoxins. Cancer Res. 56, 926-936. (34) Hall, P. D., Kreitman, R. J., Willingham, M. C., and Frankel, A. E. (1998) Pharmacology and pharmacokinetics of DT388-GM-CSF, a fusion toxin consisting of a truncated diphtheria toxin (DT388) linked to human granulocytemacrophage colony stimulating factor (GM-CSF), in C57BL/6 mice. Toxicol. Appl. Pharm. (In press). (35) Sievers, E., Appelbaum, F., Spielberger, R., Forman, S., Flowers, D., Berger, M., Shannon-Doras, K., and Bernstein, I. (1997) Dose escalation phase I study of recombinant engineered human anti-CD33 antibody-calicheamicin drug conjugate (CMA-676) in patients with relapsed or refractory acute myeloid leukemia (AML). Blood 90, 504a. (36) O’Brien, M. L., and Tew, K. D. (1996) Glutathione and related enzymes in multidrug resistance. Eur J. Cancer 32A, 967-980. (37) Pastan, I. H. (1997) Tumor immunotoxins: technology closes in on potential. Adv. .Oncol. 13, 3-12. (38) Chan, C. H., Blazar, B. R., Eider, C. R., Kreitman, R. J., Vallera, D. A. (1995) A murine cytokine fusion toxin specifically targeting the murine granulocyte-macrophage colonystimulating factor (GM-CSF) receptor on normal committed bone marrow progenitor cells and GM-CSF-dependent tumor cells. Blood 86, 2732-2740. (39) Leith, C. P., Chen, I. M., Kopecky, K. J., Appelbaum, F. R., Head, D. R., Godwin, J. E., Weick, J. K., and Willman, C. L. (1995) Correlation of multidrug resistance (MDR1) protein expression with functional dye/drug efflux in acute myeloid leukemia by multiparameter flow cytometry: identification of discordant MDR-/Efflux+ and MDR+/Efflux- cases. Blood 86, 2329-2342.

BC980015A