in vitro analysis of drug resistance in tumor cells from patients with acute myelocytic leukemia

Med Oncol. & Tumor Pharacother Vol 9 No. 2 pp. 65-74, 1992 0736-0118/90 $3.00 + .00

I N V I T R O A N A L Y S I S O F D R U G R E S I S T A N C E I N T U M O R C E L L S F R O M P A T I E N T S

W I T H A C U T E M Y E L O C Y T I C L E U K E M I A

JORGEN KRISTENSEN, 1'2 BERTIL JONSSON, 2 CHRISTER SUNDSTROM, 3 PETER N Y G R E N 4 and

ROLF LARSSON 1.

lDivision of Clinical Pharmacology, 2Departments of Medicine, 3pathology and 4Oncology, University Hospital, Uppsala University, S-75I 85 Uppsala, Sweden

(Accepted 21 May 1992)

A 72 hours fluorometric microculture cytotoxicity assay (FMCA) was used for the study of chemotherapeutic drug resistance in tumor cell suspensions from patients with acute myelocytic leukemia (AML). A marked heterogeneity with respect to sensitivity was observed for a panel of cytotoxic drugs tested in 76 samples from 60 patients with treated or untreated AML. Primary resistance to vincristine (Vcr) and prednisolone in untreated AML was observed as well as 'acquired' resistance to several other antileukemie drugs. Cross resistance patterns for AML active drugs revealed significant positive relationships between anthracyclines, VP 16 and amsacrine ( A ~ ) , whereas mitoxantrone (Mitox) was more weakly correlated. Sensitivity to cytosine arabinoside was unrelated to the anthracyclines, VP16, Amsa and Mitox but showed a significant relationship to 6-thioguanine. Several resistance modifying agent% including the novel non-immunosuppressive cyclosporin A analogue PSC 833, were able to potentiate the effects of doxorubicin and Vcr at concentrations achievable in the clinic. However, the pattern of activity was heterogenous and the frequency of responsive samples was higher in relapse compared to de novo cases. Individual in vitro~in vivo correlations based on quartile distributions of all accumulated drug sensitivity data from AML patients indicated a high specifity with respect to the identification of drug resistance. The results suggest that the FMCA may provide clinically valuable information on chemotherapeutic drug resistance in AML~

Key Words: Drug resistance, Acute non-lymphocytic leukemia, Acute myelocytic leukemia, Drug sensitivity assay

INTRODUCTION

Resistance to chemotherapy in AML: may result from the inability of a particular drug or combination of drugs to kill a required number of leukemic cells, from the ability of leukemic cells to regrow after chemotherapy or from a combination of these factors. ~ Most interest has been focused on the first of these factors also termed 'classical drug resistance'. For AraC, perhaps the clinically most active drug in AML, mechanisms for resistance at the cellular level have been reported to encompass both trans- port defects and altered conversion of AraC to the activated triphosphate, z One mechanism for resistance to other important antileukemic drugs, including the anthracyclines, vinca alkaloids and etoposides, has been termed MDR. 2'3 MDR defines a cellular phenotype where the development of in vitro resistance to one of these 'natural product' derived cytotoxic drugs, confers cross resistance to the other members of the group. This type of

"To whom correspondance should be addressed

65

resistance is believed to be mediated by the membrane bound pgpl70, transcribed from the MDR1 gene and which actively extrude the drugs from the cell. 2'3 Indeed, many fresh AML samples have shown an increased MDRI expression 4-7 but contradictory results have also been reported) Atypical forms of MDR lacking pgpl70 and with retained sensitivity to vinca alkaloids have been described 9 and suggested to be mediated by alterations of a nuclear enzyme, topoisomerase 2, which is a potential target for clinically active AML drugs such as the an- thracyctines, Mitox, etoposides and Amsa. ~~ Although important and suggestive, most of these observations on specific patterns of cytotoxic drug resistance have origi- nated from studies of laboratory derived resistant cell lines. Considerably less is known regarding drug resistance profiles in clinical AML specimens.

We have previously described a sensitive and reproduc- ible semiautomated fluorometric assay, FMCA, for studies of sensitive and drug resistant cell lines. ~2 The method shows good correlation to the MTI" assay ~3 and was found

66 J. Kristensen et al.

Table 1: List of drugs used in the FMCA. Thc origin, concentration ranges tested and solvents used are listed for each drug.

Drug Origin EDCC* Stock solvents

Ara-C Sigma 0.5 lag ml -t PBS

Doxorubicin Adriamycin| Farmitalia 0.5 lag ml -I SW

Daunorubicin Cerubidin| Rohnc-P 0.1 gg ml -t SW/PBS

VP-16 Vepeside| Bristol-Myers 5 lag m1-1 PBS

Vincristine Oncovin| Lilly 0.1 P-g ml-~ PBS

Melphalan Alkeran| Wellcome 2.5 lag m1-1 acid etOH

Prednisolon Precortalon| Organon 10 lag m1-1 PBS

Mitoxantrone Novan~onc| kederle 0.5 lag m1-1 PBS

6-TG Sigma 10 ~tg ml -t NaOH/SW

m-Amsa Amsakrin| Bristol-Myers 0.5 lag mt -1 SW

Verapamil Sigma 1.25 lag ml -l DMSO/PBS

Quinine Sigma 10 p.g ml -I PBS

Cyclosporin A Sandoz, Basel 1 lag m 1 - 1 etOH/PBS

SDZ-PSC-833 Sandoz, Basel 1 lag mt -1 etOH/PBS

*EDCC; Empirically derived cut-off concentrations for in vitro-in vivo comparison were established from dose response curves for each chemotherapeutic drug as previously described. |4 SW = sterile water, etOH = absolute ethanol. CsA, PSC 833, Ver and Quin were tested at concentrations achievable in the clinic and/or which showed drug resistance reversing effect in MDR cell lines.

to be applicable for chemosensitivity testing of tumor cell specimens from patients with leukemia. ~3-~5 In the present study we have pursued this approach and employed the FMCA for the study of chemotherapeutic drug resistance and its reversal in tumor cell samples from patients with AML.

M A T E R I A L S AND METHODS

Leukemic cell samples

A total of 76 leukemic cell samples were obtained from peripheral blood or bone marrow from 60 patients with newly diagnosed or relapsed AML diagnosed according to French-American-British (FAB) criteria. Samples were from de novo AML (39 initial, 22 at relapse), AML following myelodysplastic syndrome (7), AML following treatment for other malignant disease (3), AML following myelofibrosis (2), or from patients with myeloid blast crisis following chronic myeloid leukemia (4). Mononu- clear cells were obtained by 1.077 g ml -~ Ficoll-lsopaque (Pharmacia, Uppsala, Sweden) density gradient configu- ration. Viability was determined by trypan blue exclusion test and the density gradient centrifugation generally yielded cell suspensions of > 85 % leukemic cells as judged by May-Grtinwald-Giemsa stained cytocentri- fugate preparations. Culture medium RPMI 1640 medium

(Flow, Herts, England) supplemented with 10 % heat-in- activated FCS, 2 mM glutamine, 50 g ml -~ streptomycin and 60 gg ml -~ penicillin was used throughout. Cells were cryopreserved in culture medium containing 10 % DMSO

and 50 % FCS by initial freezing for 24 hours at -70 ~ followed by storage in liquid nitrogen. Both fresh and cryopreserved samples were used in this study. Cryopre- served samples generally show dose-response curves similar to those of corresponding fresh samples as dis- cussed previousty.~4

Reagents and drugs

FDA (Sigma Chemical Co, St. Louis, MO) was dis- solved in DMSO (Sigma) and kept frozen (-20 ~ as a stock solution (10 mg ml -~) protected from light. Chemotherapeutic drugs were obtained from various sources and were diluted and tested at the concentrations indicated in Table 1. Experimental plates were prepared with 20 gl/well of drug solution at 10 • the desired final concentration with the aid of a programmable pipetting robot (PROPET'I~; Perkin Elmer, Norwalk, CT). The plates were stored frozen a t -70 ~ until further use. Under these conditions no apparent change in drug activity was observed for 2 months as judged by repeated testing of cryopreserved samples. Final maximal solvent concentra-

tions of 0.1% had no effect on cell survival. The experiments were performed with continuous drug exposure.

Equipment

The 96 well scanning fluorometer (Fluoroscan 2; Flow) is equipped with a xenon lamp and broadband interference filters exciting fluorescence at 485 nm for FDA. The emitted light from a vertical light path of each well is sequentially read at 538 nm. One plate is read in approximately 1 minute. In most experiments, cells, media and drugs were added to the wells by the pipetting robot, PROPETIE. Addition of buffer and fluorochrome was performed with the aid of an automated 96 well dispenser, Microdrop (Flow).

FMCA procedure

The principal steps of the assay procedure have been described previously, p-'14 Day ! 180 gl of the leukemia preparation (5 x 105 cells/ml culture medium) were seeded into the wells of V-shaped 96 well experimental microtiter plates (Nunc, Roskilde, Denmark) prepared as described above. Six blank wells received only culture medium and six wells with cells but without drugs served as control

(Fig. 2). The culture plates were then incubated at 37 ~ in humidified atmosphere containing 95 % air and 5 % CO> At the end of the incubation period (72 hours), the plates were centrifuged (200 g, 7 minutes) and the medium removed by flicking the plate. After one wash with PBS,

200 gl of PBS containing FDA (10 btg ml -~) was added columnwise to control, experimental and blmtk wells. Subsequently the plates were incubated for 1 hour before reading the fluorescence in the Fluoroscan 2. The fluorometer was blanked against wells containing PBS including the fluorescent dye but without cells. Quality criteria for a successful assay included FDA signal in control cultures of > 5 x mean blank values, mean CV in control cultures of < 30 % and > 80 % of leukemic cell prior to incubation. The success-rate was > 85 % and the most common cause of failure was a low FDA signal in controls and a too low proportion of leukemic cells. In the latter half of the study, immediately following analysis of fluorescence, control wells were exposed to a mixture of Fast green (1%, Sigma), Nigrosin (0.5 %, Sigma) a.nd 25,000 formaldehyde fixed chicken erythrocytes/well for 10 minutes at room temperature. The cellular content of the wells was subsequently cytocentrifuged onto slides, counterstmned with May-Grtinwald Giemsa and the sur- viving cells were evaluated morphologically and counted by standard light microscopy. No apparent selection of normal cells was observed during culture.

Drug resistance in acute leukemia 67

Quantification of FMCA results

The results obtained by the indicator FDA are presented as survival index (SI) defined as fluorescence in per cent of control cultures (Indicator test/Indicator control, with blank values subtracted) at a EDCC defined as described previously ~4 and listed in Table 1. These values were subsequently positioned in plots in relation to the sensitivity for all AML:s tested under the szune condition.

Clinical evaluation

Patients were treated according to local protocols of the Leukemia Group of Middle Sweden and without knowl- edge of assay results. In vivo sensitivity for AML was

defined as CR (i. e. < 5 % leukemic blasts in bone marrow

aspirate for M1 and M5a and < 10 % leukemic cells for other groups of the FAB classification and with a > 50 % marrow cellularity) after maximally three courses of an induction or reinduction chemotherapeutic regimen. For a few patients not evaluated by bone marrow examinations a > 50 % decrease of peripheral leukemic cells after a single course of chemotherapy was required in order to be judged sensitive in vivo. Resistance was defined ,as all cases not meeting the above criteria and where discontinu- ation or change of treatment after 1 or 2 courses of induction therapy was due to non-responsive disease as judged by the treating physician.

In vitro - in vivo comparison

Patients were generally treated with drug combinations whereas, in most cases, only single agents were tested in vitro. The in vitro - in vivo comparisons were conse- quently based on the sensitivity of the most active single agent actually given in vivo. For each of these drugs the SI values were positioned in relationship to the sensitivity of all other AML samples in terms of 1 of 4 quartiles, each encompassing 25 % of the observations. A patient with a sample in the first quartile thus had at least one drug given in vivo corresponding to the 25 % in vitro most sensitive AML samples. The probability of clinical response was then estimated for each quartile. Cyclophosfamide, which requires in vivo activation, could not be tested and Mel was used as the alkylator indexJ 6 Sensitivity was calculated: true positives / true positives + false negatives. Specifity was calculated: true negatives / true negatives + false positives.

Statistical calculations

The relationship between SI values for various drags was estimated by Pearson's coefficient of correlation. The level of significance for these correlations was set to 99 %.


48, and 78 and 47 for Vcr and Pred, respectively. For AML only a few samples have SI values below 50 % whereas this was the case for half of the ALL samples. The re- sponses of normal mononuclear cells were similar to those of the AML samples (open symbols).

Previously treated versus untreated AML samples

Samples from previously untreated patients (n = 34) showed a lower median and a lower proportion of samples in the higher range of SI values (> 75 %) for Dox, 6TG and VP16 as compared to samples from previously treated patients (n = 22; Fig. 3). However, it should be noted that there was a significm~t overlap of SI values which is compatible with the relatively high CR rate observed also in AML at relapse. For AraC no apparent difference was noted.

Fig. I. Effect of antileukemic drugs at EDCC on SI in 62 AML samples. Each bar encompasses 90 % of the observations. The solid line within the bar indicate the median value. The two broken lines delimits 25 % of the observed SI values. The symbols indicate the position of two patients with de novo AML.

RESULTS

Selective drug sensitivity in AML

There was a marked heterogeneity between different samples with respect to in vitro sensitivity to the various drugs at EDCC as illustrated in Fig. 1. The sample from AML 1 showed a high in vitro sensitivity for anthracyclines, VP16 and Amsa corresponding to the 25 % most sensitive samples whereas the sensitivity to AraC and 6TG was lower. The sample from the second patient (AML 2) revealed another pattern with high sensitivity to the latter drugs corresponding to the 25 % most sensitive samples whereas the SI values for anthracyclines, VP 16 and Amsa were located above the median. Note that the sensitivity to anthracyclines, VP16 and Amsa in AML 1 did not confer parallel sensitivity to Mitox.

Primary resistance to Vcr and Pred

In Fig. 2 the results from 32 samples from previously untreated AML are compared with those from 25 samples from untreated ALL in response to Vcr and Pred. A marked difference in the distributions for SI values is evident with median values for AML and ALL of 80 and

Cross resistance patterns of AML active drugs

The relationships between SI values at EDCC for individual drugs are shown in Fig. 4 and Table 2 (points ranging from 55-60). As expected, SI values for Dnr and Dox were significantly correlated (R = 0.71, P < 0.001), whereas no relationship was observed between Dnr and AraC (R = 0.14, n.s.). The anthracyclines were signifi- candy related to VP16 and Amsa ( r> 0.5, P < 0.001) but more weakly to Mitox (r = 0.33, P < 0.01) and not to Vcr or AraC (P > 0.01). Mitox and Amsa were positively correlated and this was also the case for AraC and 6TG. Vcr was only weakly associated with 6TG whereas Pred lacked correlation to any other drug except for a weak relationship to AraC (Table 2).

Effect of RMA:s

Several RMA:s were tested for their ability to potentiate the cytotoxic effect of Dox and Vcr at EDCC in 38 samples (21 initial, 16 relapse). A large variability was observed as exemplified in Fig. 5. Whereas some preparations were insensitive to the different RMA:s (right) others could be significantly sensitized (left). 9 out of 21 (43 %) and 8/18 (44 %) of initial AML samples showed a more than 50 % decrease in SI to at least one RMA in response to Dox and Vcr, respectively (Table 3). The corresponding result for the relapse patients were 12 out of 17 (71%) and 11/16 (69 %). Heterogeneity also seemed related to both type of RMA and the chemotherapeutic drug employed (Table 3). The frequency of responders to PSC 833 was clearly increased in the relapse patients compared to initial AML which was not the case for CsA, Vet and Quin. In combination with Dox, PSC 833 was the most active single RMA in relapsed patients. However, PSC 833 appeared less active in combination with Vcr (Table 3).


Figure 2a AML Patient [A]

Figure 2b AML Patient [B]

F igure 2c

Figure 2d Patient [C]

Fig. 2. Effect of Vcr and Pred at EDCC on SI in 32 previously untreated AML samples (right) and 25 ALL samples (left). Each bar encompasses 90 % of the observations. The solid line within the bar indicate the median value. The two broken lines delimits 25 % of the observed SI values. The symbols indicate the position of normal mononuctear cells.


Table 2.Relationship between SI values of individual drugs at EDCC expressed as correlation coefficients for 53-60 observations on 62 samples from patients with AML

Dnr Amsa Dox Mitox 6TG VPl6 Vcr Pred AraC

Dnr X 0.52"* 0,71"* 0.33" 0,30 n.s. 0.69"* 0.04n,s. 0.21 n.s. 0.13 n.s

Amsa X X 0.52,* 0.51"* 0.19 n.s. 0.41" 0.03 n,s. 0, i1 n.s. 0.06 n.s.

Dox X X X 0.26 n.s. 0.39" 0.65** 0.29 n.s 0.22 n.s, 0.22 n,s.

Mitox X X X X 0.12 n,s. 0,14 n.s. 0.18 n.s. 0, t l n.s. 0,11 n.s.

6TG X X X X X 0.20 n.s. 0.38* 0.31 n,s. 0.47**

VP16 X X X X X X 0.18 n.s. 0.26 n.s. 0.19 n.s.

Vcr X X X X X X X 0.26 n.s. 0,33*

Pred X X X X X X X X 0.19 n.s,

AraC X X X X X X X X X

n.s. = non-significant, = P < 0.01, = P < 0.001

Table 3, Frequency of initial and relapsing AML samples responding with a more than 50 % decrease in SI to individual RMA:s in combination with Dox and Vcr. Total number of

restxmders refer to the number of samples responding to at least one RMA*

Dox

+V +CsA +PSC +Q +VQ

Total number of responders

Initial 2 5 3 5 5 9/21 (43%)

Relapse 3 4 6 3 6 12/17 (71%)

Vcr

+V +CsA +PSC +Q +VQ

Total number of responders

Initial 1 6 1 6 7 8/18 (44%)

Relapse 0 4 3 5 6 11/16 (69%)

*Concentrations of chemotherapeutics and RMA:s as in Table 1

In vitro - in vivo comparison

In 46 cases the in vitro result, expressed as the position of the most active drug in terms of the 1--4 quartiles, were related to clinical outcome as defined in material and methods (Fig. 6). Whereas 89 % of the cases with at least one drug in the first quartile responded clinically, the corresponding response rate for patients with the most active drug above the median was 17 %. Using the median as cut-off, the sensitivity and specifity of the test was 0.94 and 0.63, respectively. The predictive accuracy for both a positive and a negative test was 0.83.

DISCUSSION

A general problem with in vitro drug sensitivity assays is related to the selection of cut-off levels and drug concentrations for in vitro-in vivo comparison. In previous studies w e t4 a n d o t h e r s 17-I9 have employed an arbitrarily

defined cut-off (often 30 %) for distinguishing sensitive from resistant cell populations. The underlying assumption is that all drugs with similar clinical activity will have a similar distribution with respect to the cell survival index measured in vitro. The validity and generalibility of this assumption is not well established and may be seriously undermined depending on factors related to the selection


Fig. 3. Effect of antileukemic drugs at EDCC on SI in 34 previously untreated (A) and 22 previously treated (B) AML samples. Each bar encompasses 90 %. The solid line within the bar indicate the median value. The two broken lines delimits 25 % of the observed SI values.

A w

03 v

x o

1 2 0

1 0 0 -

8 0 -

6 0 -

4 0 -

2 0 -

0 -

0

- - y = 11.868 + 0.67491x R= 0.71137 y = 37.875 +

B A

O O OO O O o

0 001<~ o OJo

0 0 ~/~00 0 0 0

~ 3 _ ~ 0 0 0 0 0 0 ,-~ 0

0 0

I I t I

2 0 4 0 6 0 8 0 Dnr (Sl)

O

1 2 0

1 0 0 -

8 0 -

s

O

~ 4 0 -

20-

0-

0

0.15523x R= 0.14475

O

o O

co

0 i

1 0 0 8 0 1 0 0

0 O0 O o 0

0 0 0 0 0 0

O 0 C]) 0

o oO o cg o ~ o 0

0 0 0 0 0 0 0

~ oo @ 0

2 0 4 0 6 0 Dnr (Sl)

Fig. 4. Relationship between SI values for Dnr and Dox (A) and Dnr and AraC (B) at EDCC for 60 samples from patients with AML (r = 0.71, P < 0.001 in A, r = 0.14, n.s. in B).

of cut-off concentrations, i.e. achievable plasma concentrations taken from the literatureJ 6 the procedure of em- pirical derivation o f concentrations shown to provide maximal scatter of survival indices; ~4'~7 or differences in

drug specific mechanisms for in vitro versus in vivo cell kill. 2~ The new principal for calibration of assay results employed in the present paper may have some advantages in this respect since the results are normalized for differ-


A M L 1

100 4 .......................... ~ .............................. } i

~ q

8 0 - - x .- - ~

~ ! 60-- +

4 0 ti ..... : . . . . .

2 0 - _ ,

O - Dox V c r

1 2 0 -

1 0 0 -

8 0 - e~

t--

E 6o-

4 0 -

2 0 -

O_

A M L 2 /

! i , r ~

; I

] . i . . . . . . . . .

. . . . . . . . . 1

j i i

i

D o x V c r

0 con t ro l

[ ] Cs&

c> PSC 833

x Ver

+ Quin

z~ Ver+Quin

Fig. 5. Effect of different RMA:s on the sensitivity to Dox and Vcr in a RMA sensitive (left) and a RMA resistant (right) patient with AML. V = Ver, Q =Quin.

ences in SI distributions between drags at the chosen cut-off concentrations and the St for each drug will be related to other samples of the same category, q~ais approach may allow specific cut-off limits to be designed for each drug depending on the known clinical response rate.

Using the present approach we were able to detect primary resistance to Vcr and Pred by comparing SI distributions of initial de novo AML patients not previously treated with chemotherapy with those of ALL which is compatible with clinical experience. Acquired resistance to Dox, VP16 and 6TG was also noted after comparison of SI distributions from previously untreated AML and those having received previous chemotherapy. However, there was a significant overlap and only a marginal difference was noted for AraC. These observations are in accordance with the relatively high CR rates obtained also in relapsing patients. 2~

Much interest has been focused on the role of pgp 170 in mediating clinical resistance to anthracyclines in AML. A high expression in a variable proportion of both de novo and relapsed AML has been observed in several studies > 7.22.23 and has been related to poor prognosis and short remission durations, s'6'22 However, the frequency of expression has been highly variable between studies and some investigators have failed to detect any significantly increased MDR 1 expression.* 'Cross resistance profdes' as defined by the relationships between SI values obtained for different drugs at EDCC in vitro may provide important additional information on drug resistance in AML. As would be expected there was a significant relationship between Dnr and Dox but not between anthracyclines and AraC. The lack of correlation between anthracyclines and

Vcr does not support the role of the classical MDR phenotype as the dominant mechanism of resistance in AML. However, such a correlation may be masked by preexist- ing Vcr resistance. In fact, isolated Vcr resistance has been described in leukemic cell systems. > Atypical MDR with altered topoisomerase activity is another alternative but the weak relationship between the topoisomerase 2 tar- geted anthracyclines + VP16 on one hand and Mitox on the other is not entirely in accordance with this hypothesis. Low level of anthracycline resistance selected for in established cell lines have been reported to confer incom- plete cross resistance to Mitox in vitro.25 Furthermore, the lack of high grade cross resistance between an~racyclines and Mitox is also compatible with clinical experience. 26'27 Multifactorial or entirely novel mechanism(s) of resistance are potential alternatives for the observed cross resistance patters in AML.

In the present study RMA:s showed a variable effect on AML samples. In approximately 45 % of the initial AML samples addition of RMA:s apparently increased the effect of Dox or Vcr, implying a potential role of RMA:s also in the treatment of initial AML. The frequency of responsive samples was even higher in tlae relapsed patients (around 70 %), most apparent in the case of PSC 833 in response to Dox. The effect of PSC 833 in combination with Vcr was less active compared to the effect of Csa and Quin, also in the relapsed group. PSC-833 is a novel non-immunosuppressive CsA derivative reported to display a 10- fold increase in efficacy compared to CsA for the specific reversal ofpgp 170 mediated MDR in selected cell lines. 28 These observations support the presence of increased frequency of pgp 170-expression in relapsed patients but may


parameters undoubtly exits for the antileukemic drugs 33 but is seldom accounted for in clinical studies of drug resistance. Further research into the determinants of leukemic cell regrowth potential after chemotherapy and tile parallel monitoring of leukemic drugs in plasma may prove rewarding in future improvements of the predictive accuracy of short-term in vitro drug resistance assays.

Fig. 6. Minimum survival of leukemic ceils from responders and non-responders to chemotherapy according to the in vivo response criteria described in Materials and Methods. The value in terms of quartile position for the most active drug(s) actually given to the patient is presented for47 cases with AML. A patient with a symbol in the 1 quartile had at least one drug corresponding to the 25 % most sensitive samples.

also indicate that RMA (CsA) sensitive mechanisms other than pgp 170 may be operative in clinical AML samples. Since also pgp 170 negative AML samples have been reported to be sensitized by RMA:s, 23 the direct in vitro

testing of their reversing potential may be an advantageous way of selecting optimal RMA:s for clinical application.

Accumulating evidence from studies of AML samples using short-term drug sensitivity assays have shown a high p r e d i c t i v e a c c u r a c y with respec t to drug resistance 14'16'19'29'30 which also was the case in the present

study. However, the overall predictive accuracy for a positive test seems less impressive. It should be noted that all false positives in the present study were from relapsing patients. The poorer prognosis of this group of patients may thus result from alternative or additional factors other than 'classical' drug resistance. 3~ Indeed, based on quan- titative marrow examinations at different time points before and after chemotherapy, Preisler et al. have suggested that an increased ability of leukemic cells to regrow after termination of therapy is an alternative cause of clinical drug resistance to be taken into consideration. ~'3z Another reason for a high false-positive rate may be that inadequate drug concentrations are achieved in plasma in some patients with truly drug sensitive tumor cells. A large inter- individual variability with respect to pharmacokinetic

A B B R E V I A T I O N S

AML acute myelocytic leukemia. ALL acute lymphocytic leukemia. FDA fluorescein diacetate. FMCA fluorometric microculture cytotoxicity assay. MTI" 3-(4,5 dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium

bromide. Disc Differential staining cytotoxicity. EDCC empirically derived cut-off concentration. SI survival index. DMSO dimethyl sulfoxide. AraC cytosin arabinoside. Vcr vincristine. Dnr daunombicin. Dox doxorubicin. Pred prednisolone. Amsa amsacrine. Mitox mitoxantrone. Mel melphalan. VPI6 etoposide. Ver verapamil. Quin quinine. CsA cyclosporin A. PBS phosphate buffered saline. CV coefficient of variation. FCS fetal calf serum. CR complete remission. MDR multidrug resistance. pgp 170 170 kDa P-glycoprotein. RMA resistance modifying agents.

Acknowledgement - - This study was supported by grants from the Swedish Cancer Society (2695-B89--OIX). We also express our sincere gratitude to Lotta Sandberg and Anneli Kraft for expert technical assistance.

R E F E R E N C E S

1. Preisler H: Determinants of response and prediction of response to chemotherapy in acute myelogenous leukemia. BaiUiere's Clinical Hematology 4, 68 ( 1991).

2. Ross D: Cellular and pharmacologic aspects of drug resistance in acute myeloid leukemia. Current Opinion in On- cology 3, 21 (1991).

3. Beck W T: ]'he cell biology of multiple drug resistance. Biochem Pharmacol 36, 2879 (1987).

4. Sato H, Gottesman M, Goldstein L, Pastan I, Block M, Sandberg A, Preisler H: Expression of the multidrug resistance gene in myeloid leukemias. Leukemia Res 14, 11 (1990).

5. Pirker R, Wallner J, Geisster K, Linkesh W, Haas O, Bet- telheim P, Hopfner M, Scherrer R, Valent P, Havelec L,

74 J. Krislensen et al.

Ludwig H, Lechner K: MDR1 gene expression and treatment outcome in acute myeloid leukemia. J Natl Cartcer lnst 83, 708 ( 1991 ).

6. Marie J, Zittoun R, Sikic B: Multidrug resistance gene expression in adult acute leukemias: Correlation with treatment outcome and in vitro drug sensitivity. Blood 78, 586 (1991).

7. Herweijer H, Sonnevald P, Baas F, Nooter K: Expression of MDRI and MDR3 multidrug resistance genes in human acute and chronic leukemias and association with stimula- tion of drug accumulation by cyclosporine. J Natl Cancer Inst 82, 1133 (1990).

8. Ito Y, Tanimoto M, Kumazawa Y, Okumura M~ Morishima Y, Ohno R, Saito H: Increased P-glycoprotein expression and multidrug resistance gene amplification are infre- quently found in fresh acute leukemia cells. Cancer 63, 1534-1538 (1989).

9. Beck W T, Cirtain M C, Danks M K, Felsted R L, Safa A R, Wolverton J S, Suttle D P, Trent J M: Pharmacological, molecular, and cytogenetic analysis of "atypical" multidrug -resistant human leukemic cells. CancerRes 47, 5455 (1987).

10. Defile A, BatraJ, Goldenberg G: Direct correlation between DNA topoisomerase 2 activity and cytotoxicity in adriamycin- sensitive and - resistant P388 leukemia cell lines. Cancer Res 49, 58 (1989).

11. Hall A, Cattan A: Drug resistance mechanisms in leukemia, Bailliere's Clinical Hematology 4,655 ( 1991 ).

12. Larsson R, Nygren P: Pharmacological modification of multi-drug resistance (MDR) in vitro detected by a novel fluorometric microculture cytotoxicity assay. Reversal of resistance and selective cytotoxic actions of cyclosporin A and verapamil on MDR leukemia T-cells. lnt J Cancer 46, 67 (1990).

13. Larsson R, Nygren P, E].cberg M, Slater L: Chemotherapeu- tic drug sensitivity testing of human leukemia cells in vitro using a semiautomated fluorometric assay. Leukemia 4,567 (1990).

14. Larsson R, Kristensen J, Sandberg C, Nygren P: Laboratory determination of chemotherapeutic drug resistance in tumor cells from patients with leukemia using a fluorometric microculture cytotoxicity assay (FMCA). lnt J Cancer 50, 177 (1992).

15. Nygren P, Larrson R: Differential in vitro sensitivity of human tumor and normal cells to chemotherapeutic agents and resistance modulators. Int J Cancer 48, 598 (t991).

16. Kern D, Weisenthal L: Highly specific prediction of an- tineoplastic drag resistance with an in vitro assay using suprapharmacologic drug exposures. J Natl Cancer Inst 82, 582 (1990).

17. Weisenthal L, Dill P, Finkelstein J, Duarte T, Baker J, Moran E: Laboratory detection of primary and acquired drug resistance in human lymphatic neoplasms. Cancer Treat Rep 70, 1283 (1986).

18. Bird M, Bosanquet A, Forskitt S, Gilby E: Semi-micro adaptation of a 4-day differential staining cytotoxicity (Disc) assay for determining the in-vitro chemosensitivity of haematological malignancies. Leukemia Res 10, 445 (1986).

19. Sargent J, Taylor C: Appraisal of the MTF assay as a rapid test of chemosensitivity in acute myeloid leukemia. Br J Cancer 60, 206 (1989).

20. Weisenthal L M, Su Y Z, Duarte T E, Nagomey R A: Non-clonogenic in vitro assays for predicting sensitivity to cancer chemotherapy, in Liss A R: Prediction of response to cancer therapy, pp.75-92. (t 988).

21. Heinemann V, Jehn U: Rationales for a pharmacologically optimized treatment of acute nonlymphocytic leukemia with cytosine arabinoside. Leukemia 4~ 790 (1990).

22. Sato H, Preisler H, Day R, Raza A, Larson R, Brownman G, Goldberg J, Volger R, Grunwald H, Gottlieb A, Bennett J, Gottesman M, Pastan 1:MDR1 transcript levels as an indication of resistant disease in acute myelogenous leukemia. Br J Haematology 75, 340 (1990).

23. Solary E. Bidan J, Cairo F, Chauffert B, Caillot D, Mugneret F, Gauville C, Tsumo T, Carli P, Guy H: P-glycoprotein expression and in vitro reversion of doxorubicin resistance by veraparnil in clinical specimens from acute leukemia and myeloma. Leukemia 5, 592 (1991 ).

24. Haber M, Norris M D, Kavallaris M, Bell D R, Davey R A, White L, Stewart B W: Atypical multidrug resistance in a therapyqnduced drug-resistant human leukemia cell line (LALW-2): resistance to vinca alkaloids independent of P-glycoprotein. Cancer Res 49, 5281 (1989).

25. Hill B, Hosking L, Shellard S. Whelan R: Comparative effectiveness of mitoxantrone and doxorubicin in overcom- ing experimentally induced drug resistance in murine and human turnout cell lines. Cancer Chemother Pharmaco123, 140 (1989).

26. Hiddcman W, Kreutzmann H, Straif D, Ludwig W, Mertelsmann R, Donhuijsen-Ant R, Lengfelder E, AJ'lin Z, Buchner T: High-dose cytosine arabinoside and mitoxantrone: a highly effective regimen in refractory adult acute leukemia. Blood 69, 744 (1987).

27. Ho A, Lipp T, Ehninger G, Illiger H, Meyer P, Freund M, Hunstein W: Combination of mitoxantrone and etoposide in refractory acute myelogenous leukemia: an effective and well-tolerated regimen. J Clin Oncol 6, 213 (1988).

28. Gaveriaux C, Boesch D, Jachez B, Pollinger P, Payne T, Loor F: SDZ PSC 833, a non*immunosuppressive cyclosporin analog, is a very potent multidrug-resistance modifier, d Cell Pharmacot 2, 225 (1991 ).

29. Bosanquet A: Correlations between therapeutic response of leukemias and in vitro drug sensitivity assay. Lancet 1, 711 (1991).

30. Kirkpatrick D, Duke M, Goh T: Chemosensitivity testing of fresh human leukemia cells using both a dye exclusion assay and a tetrazolium dye (MTr) assay. Leukemia Res 14, 459 (1990).

31. Lepri E, Liberati M, Portuese M, Menconi E, Santucci A, Barzi A: In vitro chemosensitivity of newly diagnosed and relapsing acute non-lymphoid leukemia patients. Tumori 77, 25 (1990).

32. Preisler H, Raza A, Larson, R, LeBeau M, Brownman G, Goldberg J, Grunwald H, Volger R, Verkh L, Singh P, Block A, Sandberg A: Protooncogene expression and the clinical characteristics of acute nonlymphocytic leukemia: A leukemia intergroup pilot study. Blood 73, 255 (1989).

33. Evans R, Relling M: Clinical pharmacokinetics-pharma- codynamics of anticancer drugs. Clinical Pharmacokinetics 16, 327 (1989).

in vitro analysis of drug resistance in tumor cells from patients with acute myelocytic leukemia

Documents