methods to detect p-glycoprotein-associated multidrug ...immunostaining, flow cytometry, northern...

[CANCER RESEARCH 56, 3010-3020, July I. 1996]

Methods to Detect P-Glycoprotein-associated Multidrug Resistance in Patients'Tumors: Consensus Recommendations1

William T. Beck,2 Thomas M. Grogan,2 Cheryl L. Willman, Carlos Cordon-Cardo, David M. Parham,3John F. Kuttesch,4 Michael Andreeff, Susan E. Bates, Costan W. Berard, James M. Boyett, Nathalie A. Brophy,

Henk J. Broxterman, Helen S. L. Chan, William S. Dalton, Manfred Dietel, Antonio T. Fojo, Randy D. Gascoyne,David Head, Peter J. Houghton, Deo Kumar Srivastava, Manfred Lehnert, Catherine P. Leith, Elisabeth Paietta,Zlatko P. Pavelic, Lisa Rimsza, Igor B. Roninson, Branimir I. Sikic, Peter R. Twentyman, Roger Warnke, andRonald WeinsteinSi. Jude Children's Research Hospital, Memphis. Tennessee 38105 Â¡W.T. B.. D. H.. P. J. H.. D. M. P.. J. F. K.. J. M. B.. C. W. B.. D. K. S.I: University of Arizona College of

Medicine and Ancona Cancer Center. Tucson. Arizona 85274 Â¡T.M. G.. W. S. D., L R.. R. We.]: University of New Mexico. Albuquerque. New Mexico 87131 [C. L W.. C. P. LI;Memorial Sloan-Ketlering Cancer Center, New York, New York 10021 1C. C-CJ; M. D. Anderson Cancer Center, Houston, Texas 77030 IM. A.]: Medicine Branch, National

Cancer Institute. Bethesda. Maryland 20892 IS. E. B., A. T. F.J: Stanford University Medical Center. Stanford, California 94305 Â¡N.A. B., B. I. S.. R. Wa.J; Free UniversityHospital. Amsterdam 1081. the Netherlands IH. J. BJ: The Hospital for Sick Children, Toronto, Ontario M5C 1X8, Canada /H. S. L C.Â¡:CharitÃ©of Berlin. Humboldt UniversityD-10117. Berlin, German\ [M. D.]; Montefiore Medical Center, Albert Einstein Cancer Center. Bronx, New York 10467 Â¡E.P.I: University of Cincinnati College of Medicine.

Cincinnati. Ohio 45267 Â¡Z.P. P.J: University of Illinois College of Medicine, Chicago, Illinois 60612 [1. B. R.]; Medical Research Council Centre, Cambridge University,Cambridge CB2 2QH. England Â¡P.R. T.]: University of British Columbia, Vancouver V5Z 4E6. British Columbia. Canada Â¡R.D. G.]; and Department C of Internal Medicine,Kantonsspital. CH-9007 St. Gallen. Switzerland Â¡M.LÂ¡

ABSTRACT

Multidrug resistance iMDKi. especially that associated with overex-pression otMDKi and its product, P-glycoprotein (Pgp), is thought to play

a role in the outcome of therapy for some human tumors; however, aconsensus conclusion has been difficult to reach, owing to the variableresults published by different laboratories. Many factors appear to influence the detection of Pgp in clinical specimens, including its low andheterogeneous expression; conflicting definitions of detection end points;differences in methods of sample preparation, fixation, and analysis: useof immunological reagents with variable Pgp specificity and avidity andwith different recognition epitopes; use of secondary reagents and chromogens; and differences in clinical end points. Also, mechanisms otherthan Pgp overexpression may contribute to clinical MDR. The combinedeffect of these factors is clearly important, especially among tumors withlow expression of Pgp. Thus, a workshop was organized in Memphis,Tennessee, to promote the standardization of approaches to MDRl andPgp detection in clinical specimens. The 15 North American and Europeaninstitutions that agreed to participate conducted three preworkshop trialswith well-characterized MDR myeloma and carcinoma cell lines that

expressed increasing amounts of Pgp. The intent was to establish standardmaterials and methods for a fourth trial, assays of Pgp and MDRl inclinical specimens. The general conclusions emerging from these effortsled to a number of recommendations for future studies: (a) althoughdetection of Pgp and MDRl is at present likely to be more reliable inleukemias and lymphomas than in solid tumors, accurate measurement oflow levels of Pgp expression under most conditions remains an elusivegoal; (/>i tissue-specific controls, antibody controls, and standardized

MDR cell lines are essential for calibrating any detection method and forsubsequent analyses of clinical samples; (c) use of two or more vendor-standardized anti-Pgp antibody reagents that recognize different epitopes

improves the reliability of immunological detection of Pgp; (</) samplefixation and antigen preservation must be carefully controlled; in mul-

tiparameter analysis is useful in clinical assays of MDRl/Pgp expression;

Received 1/12/96; accepted 5/13/96.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

1This study represents a series of multi-institutional trials to assess sources of varia

bility in assay to detect P-glycoprotein in tumors from patients. These trials culminated ina final trial and workshop held at St. Jude Children's Research Hospital (Memphis, TN)in May 1994 to forge consensus recommendations. See "Appendix" for grani support

information.- To whom requests for reprints should be addressed. Dr. Beck is now at the Cancer

Center (M/C 569), University of Illinois at Chicago. 900 South Ashland Avenue. Room3352, Chicago. IL 60607-7173. Phone: (312) 355-0827; Fax: (312) 355-0194. T. M. G.,Department of Pathology. University of Arizona College of Medicine. 1501 NorthCampbell Avenue. Tucson. AZ 85274. Phone: (520) 626-2212; Fax: (520) 626-6081.

' Dr. Parham is now at the University of Arkansas. Little Rock. AR.4 Dr. Kuttesch is now at the M. D. Anderson Cancer Center, Houston, TX.

i/1 immunostaining data are best reported as staining intensity and thepercentage of positive cells; and (g) arbitrary minimal cutoff points foranalysis compromise the reliability of conclusions. The recommendationsmade by workshop participants should enhance the quality of research onthe role of Pgp in clinical MDR development and provide a paradigm forinvestigations of other drug resistance-associated proteins.

INTRODUCTION

Considerable effort has been expended to design and execute clinical trials testing methods for the detection and reversal of MDR5-thatform of resistance to "natural product" anticancer agents-in cancer

patients (1-9). Because these studies have not yielded consistentresults in terms of Pgp expression and reversal of MDR (8-14), and

given the continued interest in such research, it seems appropriate toask whether the problems encountered might reflect inadequate experimental support. In general, the MDRl gene and its product, Pgp,are overexpressed de novo in certain malignancies (1, 2, 15) and afterchemotherapy with Pgp substrates in others (16-19). However, aber

rant expression of the protein does not always correlate with theclinical status of the patient or the outcome of therapy (e.g., Refs. 15and 20), nor have attempts to circumvent clinical MDR with Pgp"modulators" produced uniform results (8-14). Because of these

inconsistencies, which no doubt stem from difficulties in detectingPgp, as well as other biochemical and pharmacological factors contributing to resistance, the role of this protein in clinical MDR remainsunclear.

A major problem in assessing the significance of MDRl and Pgpoverexpression in clinical MDR has been the variability in measurements of these factors (10). Although most laboratories can reliablydetect high levels of expression of both MDRl and Pgp. regardless ofthe method or reagent(s) used, lower levels of expression have provedmore difficult to quantify. Variable quantitative results have beenreported for similar clinical specimens by different laboratories usingimmunostaining, flow cytometry, Northern blotting, RNase protection, RNA in situ hybridization, Western blotting, and rt-PCR (1-6),

and few studies have attempted to compare multiple methods (e.g.,Refs. 3 and 4). Even with the same assay (e.g., immunostaining) anddisease, findings have varied according to the anti-Pgp antibodies andsecondary detection systems selected (16, 20). Moreover, the cross-

' The abbreviations used are: MDR. multidrug resistance; AML. acute myeloid leu

kemia; ICC/IHC, immunocytochemistry/immunohistochemistry; Pgp, P-glycoprotein; PE,phycoerythrin; S:N. signal:noise: rt-PCR. reverse transcription-PCR: NCI. National Cancer Institute: ALSAC. American Lebanese Syrian Associated Charities; SWOG. Southwest Oncology Group.

3010

on March 21, 2020. © 1996 American Association for Cancer Research.cancerres.aacrjournals.org Downloaded from

http://cancerres.aacrjournals.org/

MDR METHODS DETECTION WORKSHOP

reactivity of some anti-Pgp antibodies has proven to be an especiallyconfounding variable (21, 22). Other sources of variability includeinadequate controls and the failure to calibrate assays with standardized cell lines.

Because of the urgent need to assess the role of MDRl/Pgp in thedevelopment of clinical MDR, we organized an international workshop on MDR detection methods. As plans were being formulated, itbecame clear that mere measurement of MDR! or Pgp in identicalsamples by many laboratories (the original goal) would not providethe guidance needed for resolution of critical issues. Thus, the goals ofthe workshop were expanded to incorporate the generation of standards against which analyses of clinical specimens could be judged, aswell as the consideration of problems related to sample handling,specimen fixation, and data analysis. Ultimately, the decision wasmade to establish a large, interactive cooperative group that wouldmeet exclusively to forge recommendations on MDR detection andrelated matters. To the best of our knowledge, an equivalent "hands-on" workshop to detect markers of drug resistance has not been

attempted, although we have learned that a French group, headed byProf. J. Robert (Bordeaux, France) is addressing similar issues (notedin Ref. 23).

The detection of MDR] and Pgp in clinical specimens has reliedheavily on antibody and cDNA probes (Refs. 1-6 and referencestherein; Ref. 24). Although immunostaining is routinely applied inpathology laboratories, its use to detect Pgp in tumor specimensposes special problems. Rather than asking whether a specimencontains or lacks Pgp, as is common when attempting to find adiagnostic marker, one must quantify the extent of Pgp expression,which in experimental systems is directly correlated with thedegree of tumor cell MDR (25-27), although quantifying thenumbers of Pgp molecules required to confer drug resistance to atumor cell remains elusive. Thus, the theme of the workshopbecame "improved detection of MDRl/Pgp in tumor specimensthrough comparison of standard techniques." Specific objectives

were to (a) test the logistics of tumor and sample distribution, Pgpdetection methods, and data interpretation; (b) identify pitfalls inthe system; (c) determine if a resource laboratory can distributesamples with adequate preservation of Pgp; and (d) determine ifimmunostaining and other methods and the interpretation of resultsshow substantial variability among and within laboratories. Theoverall goal was to lay the groundwork for moving MDRl/Pgpfrom the laboratory to the clinic, ultimately as a target for specifictherapies. The results and recommendations presented heresummarize the entire agenda of the St. Jude MDR Workshopon Methods to Detect P-Glycoprotein-associated MultidrugResistance.

MATERIALS AND METHODS

Cell Lines. Well-characterized cell lines served as calibration kits in each

of the comparative trials of the workshop. As models for hematologicalcancers, we used the 8226 myeloma cell line and its well-characterizedMDR-expressing derivatives 8226/Dox 1, 6. and 40 (27), which were selected

for increasing resistance to doxorubicin (Dox 1 and 6 express low levels ofresistance). The solid tumor model was the KB cell line and sublines selectedfor increasing levels of MDR expression [KB3-1, KB8-5-11 (26), and KBVI

(28)]. Each cell line was provided to participating investigators in all trials.Ultimately, the lines served as calibration standards for the trials involvingclinical samples. Depending on the objective of the trial, cells were providedeither as fixed material (cytospin preparations, 8226 cells: paraffin sections,KB cells) or as live cells in culture (8226 cells).

Antibodies. The seven anti-Pgp antibodies used in comparative analyses(Table 1) were generous gifts of commercial vendors: C219, C494, JSB-1, and

4E3 through Dr. Ronald J. Casciato (Signet Laboratories. Inc.. Dedham. MA);UIC-2 through Drs. Eugene Mechetner and Gregory Reyes (Ingenix, SouthSan Francisco, CA); MRK-16 through Colin Getty (Kamiya Biochemicals,Thousand Oaks, CA); and HYB-241 through Dr. Lana Grauer (Hybritech. LaJolla, CA). Fluorescein- or biotin-labeled second antibodies and enzyme-

chromogen detection systems were obtained from standard commercialsources, and each participating laboratory used its own "second-step" reagents

and detection methods. The characteristics of these antibodies, their originalcitations, the concentrations provided by the vendors, and the concentrationsused in the trials are given in Table 1.

Antibodies were titrated by three of the workshop participants; they weretitrated for cytospin and histological preparations by Dr. T. M. Grogan (Tucson, AZ), using 8226 cells; for histological sections by Dr. C. Cordon-Cardo(New York, NY), using normal kidney and renal tumors; and for flow cytom-

etry by Dr. C. L. Willman (Albuquerque, NM). using transformed myeloidblast cells. Once approximate titrations were known, the commercial vendorsprovided the same lot of each antibody for all investigators, with recommendations for dilutions. Antibody optima for immunohistochemical assays weredetermined by standard microscopic morphometric grading, and optimal S:Nratios were assessed by image analysis (35).

Solid Tumor "Buttons." Vinblastine-resistant and vinblastine-sensitive

KB cells were prepared by Dr. T. M. Grogan as formalin-fixed, paraffin-

embedded solid tumor specimens to test histological methods in Trial 2.Further, in Trial 4, "sausages" of KB3-1, KB-8-5-11, and KBV1 cells were

placed on the same slides as the clinical specimens to view positive andnegative controls simultaneously.

Immunocytochemical and Immunohistochemical Staining. Workshopparticipants were free to use their standard methods for these procedures, all ofwhich have been published elsewhere (summarized in Table 2).

Tissue and Tumor Selection. Samples were selected and coded in onereference laboratory (C. L. Willman) from three patients with primary de novoAML: they were chosen for study based on variable expression of MRK16reactivity. All samples contained 78% blasts (determined morphologically) andhad been stored for 1-3 years at â€”¿�135Â°C(15) as sterile, viable cell suspen

sions. Coded samples were distributed frozen on dry ice to the workshop

Table 1 Characteristics of workshop antibodies

Antibodies"

CharacteristicReferenceCloneIsotypePurificationVendorLot

no.Concentration(/ig/ml)CytospinsTiterConcentration

(/ig/ml)TissueIHCConcentration

(^ig/ml)4E3294E3.16IgG2aAP"SignetFC3308251/5510C49430C494IgG2aAPSignetGK3359501/5-1/105-1010C21930C219IgG2aAPSignetBC3107401/5810JSB-131JSB-1IgGlAPSignetAR4017501/10-1/20510UIC232UIC2IgG2aAPIngenex1.91/5004MRK1633MRK16-MIgG2aAPKamiya793500HYB24134HYB241IgGHybridtech2.52.5" Spaces indicate missing information or no further manipulations.* AP, affinity purified.

3011




Table 2 Summary of methods used by workshop participants

MethodICCmeFrozen

sectionsParaffinblocksAntigen

retrievalFlow cylometry

PgpdetectionPgp

function (dye or drugefflux)rt-PCRMultiple

methodsInvestigatorsGrogan

ChanTwentyman

Chan.DaltonPavelicRimsza.

GroganParham.KutteschTwentymanPaielta

Leith.WillmanBroxterman

Bates. FojoLeith.WillmanBales.

Dalton. FojoDaltonWillmanWamke.

SikicDietelBates,

FojoReferences36

3738

37.394035.

3641.4238431544,45

46152,

3, 4fr4849504

51-543

participants by overnight courier (North America) or 2-day delivery (Europe).

The expression of Pgp was characterized by the reference laboratory as high,meaning positive staining for MRK16 (MRK16+) and the ability to effluxeither DÃ•OC2or rhodamine 123 (efflux+ ). as detailed in Ref. 16. at levelssimilar to these seen with 8226/Dox6 (Dox6) cells [i.e.. MRK16+/efflux + ,>95% Ml blasts (~Dox6)]; intermediate, meaning MRK16+/efflux + . >90%blasts (<Dox6); or low, meaning gated blasts MRK16-/efflux-, <90%blasts (â€”80%) with contaminating T cells (samples collected from peripheral

blood).Samples of renal cell carcinoma and normal kidney tissue from the same

case were either embedded in OCT compound (Miles. Inc., Elkhart, IN) andfrozen in liquid nitrogen or fixed in buffered formalin and embedded inparaffin. Tissue sections for immunohistochemical studies were distributed toparticipating institutions by overnight (North America) or 2-day (Europe)

delivery.For immunohistochemical testing of rhabdomyosarcoma, a single case that

had shown only focal Pgp staining was selected. Paraffin sections of formalin-

fixed tissue and cryostat sections of fresh tissue preserved at â€”¿�70Â°Cwere

prepared, with care taken to avoid thawing of the frozen sections. The specimens were then shipped to the participants on dry ice by overnight delivery.

Data Collection and Analysis. The Cellular Analysis System 200 and theCell Measurement 2.0 program (Becton Dickinson, Mountain View, CA) wereused in quantitative imaging (35). The 2.0 program measures several cellularproperties, including nuclear features (nuclear size and shape, and summed andaverage absorbance) and cytoplasmic features (size, shape, and summed ab-

sorbance). Because of the nonnuclear location of Pgp. we used the summedcytoplasmic density measurement, centering the cells of interest in the microscopic field and analyzing 100 consecutive cells. The S:N ratio was calculatedas the mean of the summed cytoplasmic densities of 100 KBV-1 cells dividedby the mean of the summed cytoplasmic densities of 100 KB3-1 cells.

References for flow cytometric methods for measuring Pgp levels andfunctional activity are cited in Table 2, as are references for rt-PCR, primers

for MDR1, and control probes.

RESULTS AND DISCUSSION

Trial 1. Logistics of Sample Distribution and Pgp Detection byImmunocytochemistry. For this trial, we used 8226 cell lines asmodels of hematological tumors for standardization and comparisonof Pgp detection methods. Coded cytospin preparations on glassslides, fixed with acetone (4Â°C),were sent from Tucson, AZ, to

participating laboratories by overnight express, to be stained andanalyzed within 24 h, with repeat analyses performed in each of thenext 3 days. Stained slides were initially analyzed microscopically byindividual laboratories, and then sent back to Tucson for quantitationby image analysis. A representative slide is shown in Fig. 1. Theapparent cytoplasmic staining of Pgp in these cells is an artifact of thecytochemistry method: total cells, not sections, are examined from thetop, so it appears to be cytoplasmic staining, when in fact it is plasmamembrane staining. Virtually all of the 11 institutions participating inthis trial were able to demonstrate appropriate Pgp staining density inthe cell lines, although there was considerable variation among institutions, especially when the parental drug-sensitive cell line and the

Fig. 1. Staining of 8226 cell lines withJSB-I anti-Pgp antibody. Myeloma variant

cell lines expressing Pgp (27) are shown.Pgp expression was graded as follows: A, 0(8226/S): ÃŸ. + (8226/DoxI); C, + +(8226/Dox6); and D. + + + (82267Dox40).

3012



1Â«755025"t?

nUoS

501QÃ“

25VM*Ã®C0)â€¢o

â€”¿�0is.on

10Â°^

75502500.6S0si:3.7

^i,1.7Doxl

DoxÃ³Dox452.60.1:nDoxl

DoxÃ³DoxÂ¿982.659.4417.1!â€¢u

S Doxl DoxÃ³Dox40

100

75

50

25

100

75

50

25

100

75

50

25


96.4

2.0 Â«â€¢¿�" 1 1 1

16.6

68.5

Doxl DoxÃ³Dox40

6 ... 98-6

78.2

90.5

S DoxlDoxÃ³Dox40

1073.153.C528.9n4.6

S DoxlDoxÃ³Dox40

1007$502503 96.8100]63.2A6.71

â€”¿�13.675502504

9S.O80.554.014.5S

Doxl DoxÃ³Dox40 S Doxl Dox6Dox4C1Â«755025

07

85,7

62.238.13.610C755025

0852.0SUn68.385.1

100)

Ãœ, 75

50

25

S Doxl DoxÃ³Dox40 S Doxl DoxÃ³Dox40

1164.8

0.07 2.25 5.15

S Doxl DoxÃ³Dox40

Fig. 2. Trial 1: multi-institutional comparison of Pgp densities by immunocytochemistry. Eleven institutions (1-11) produced these results, expressed as the mean Pgp densityexpressed as mean absorbance values per 200 cells for each of the 8226 variant cell lines. See "Materials and Methods" for details of quanlitation of Pgp density. Both the selection

of antibodies and methods of detection were left to the discretion of the individual laboratories.

two cell lines with low levels of resistance, Doxl and Dox6, wereused (Fig. 2). Thus, the data underscore the importance of truenegative controls. Indeed, when results were quantified as S:N ratio,the value for Dox40 was consistently higher than that for Doxl (Fig.3). Staining of consecutive slides over 3 days yielded highly reproducible staining densities (data not shown), indicating that under theseconditions of fixation and transportation, Pgp "travels well."

In this trial, 8226 cells in acetone-dipped cytospins provided a

reliable transferable standard (reproducible results obtained for atleast 3 days) and therefore should be useful in multi-institutional

studies of cellular Pgp. These fixed cells can be sent thousands ofmiles to many different laboratories without appreciable effects on theresults of qualitative and quantitative measurements of Pgp.

At the outset of the workshop, we questioned whether relativedensity determinations based on S:N ratios would provide reliablemeasurements for assay optimization among large numbers of institutions. Our findings suggest that by setting the S:N ratio as shown inFig. 3, A and B, one can establish a quantitative benchmark that willfacilitate comparisons among institutions. This is a novel observationin Pgp research and will be the subject of a separate publication.6

Trial 2. Evaluation of KB Cells as Models of Solid Tumors inStandard Immunohistochemical Assays of Pgp Expression. Wedeemed it important to have a solid tumor cell line resource that wouldallow testing of the effects of fixation, antigen retrieval, and, ultimately, optimization of histolÃ³gica! Pgp assays. To adapt Pgp+and Pgpâ€” KB cells for this purpose, we subjected them to the

entire fixation procedure. To address the issue of poor tissuefixation, a common problem in clinical evaluations of MDR, we

s L. Rimsza and T. M. Grogan, manuscript in preparation.

prepared a tissue-fixation sausage consisting of pellets of variablyfixed KB3-1 (Pgp-) and KBV] (Pgp + ) cells that were individu

ally embedded in master paraffin blocks and sectioned as shown inFig. 4. The purpose was to develop internal Pgp+ and Pgp-

validation controls for clinical specimens with unknown Pgp expression. Thus, KB3-1 (Pgp-) and KBV1 (Pgp+) cells were

grown in cell culture, centrifuged, fixed in colloidon bags andembedded with paraffin for histolÃ³gica! sectioning,6 after which

samples were coded and provided to all participating laboratoriesfor Pgp staining and analysis. Investigators were asked to use anantibody (e.g., JSBI or C494) that is known to stain Pgp informalin-fixed, paraffin-embedded tissues (24, 31, 37),

although the recognition of other epitopes by C494 is acknowledged (22).

From the immunohistochemical staining results in Fig. 5A, it isclear that Pgp can be detected by this method in KB cells. Imageanalysis of cells stained with C494 antibody in two different laboratories (laboratories 1 and 10) yielded sizable S:N ratios for KBVI(Pgp+) and KB3-1 (Pgpâ€”) cells. In contrast, a third laboratory

(laboratory 5) used a second antibody, HYB241, not known to reactwith Pgp under these conditions of fixation, that produced a weaksignal (Fig. 5B).

Thus, although we did not solve all of the problems inherent inantibody staining for Pgp in solid tumors, trial 2 accomplishedseveral important goals. First, it validated KB cells as a legitimatesolid tumor model that can be used productively without theexpense and technical difficulties of xenograft models. Second,and perhaps most important, we were able to show that a PgpÂ±control sausage placed on the same slide with a clinical specimenof unknown Pgp content will permit direct validation of Pgp

3013




A. Dox40 / 8226S B. Dox 1 / 8226 S

Fig. 3. Trial I: comparative S:N ratios by imageanalysis. The ratios shown are the S:N ratios ofeither Dox40 versus S or Dox I versus S cells. S:Nratios were quantitatcd as described in "Materialsand Methods."

500

400-

200-

20

20

10-

Institution: 4 7 6 1110 2 1 5 Institution: 11 10 7 6 24 1 5

staining. The success of this method in trial 2 led to its adoption intrial 4, described below.

Trial 3. Comparative Test of Different Methods for Measurement of Pgp and MDR1. A major aim of the workshop was tocompare results of immunostaining and flow cytometric identificationof Pgp, flow cytometric assessment of Pgp function by dye efflux, andPCR measurement of MDRI mRNA expression. This goal was pursued in trial 3 with use of suspension cultures of 8226/S and 8226/Doxcell lines individually and as admixtures. Briefly, coded flasks of8226/S and 8226/Dox lines were provided to investigators for tests ofspecificity. Sensitivity testing of assays was conducted with admixtures of 90% S and 10% Dox40 cells and 99% S and 1% Dox40 cells.Investigators were encouraged to apply whatever methodology theypreferred to determine the amount or level of Pgp or MDR1 expression, as well as Pgp function. The value of standardized antibodyusage was assessed by testing both a single antibody across multiplemethods and multiple antibodies within a single method (data notshown).

Results of these multiple comparisons are summarized in Fig. 6.The larger number of tests performed with immunostaining compared to flow cytometry, rt-PCR, or dye efflux can be attributed tothe use of as many as seven different anti-Pgp antibodies bydifferent laboratories. There was no evidence in any of the subtrialsthat a particular antibody provided a detection advantage in immunostaining procedures. It is clear that each assay was capable ofcorrectly identifying drug-sensitive cells, as well as those withstrong expression of MDR. Perhaps of greater importance, theassays were sufficiently sensitive to distinguish marginal situationsin which drug-resistant cells made up only 1 or 10% of the totalcell population (Fig. 6, E and F)', however, the variable detection

rates with immunostaining, flow cytometry, and dye efflux suggest

that these methods may yield a significant percentage of falsenegative results. The same caveat applies to immunostaining andflow cytometry used with a cell line expressing low levels of Pgp,and to all assays used with a cell line characterized by faintexpression of Pgp and the MDR1 gene. Some laboratories relied onPCR to perform a quantitative analysis. Indeed, results fromone participating laboratory demonstrated that PCR could quanti-tate a > 160,000-fold range of expression of MDRÃŒmRNA, ranging from 0.01 in 8226/S to 1692 in 8226/Dox40. However, thesensitivity of the PCR assay in most situations is offset by theinability of this method to distinguish between Pgp in normalversus neoplastic cells, demonstrating again the need for multipa-rameter analysis.

Trial 4: Testing of Clinical Samples. In this comparison, investigators used previously validated reagents and cell lines to analyzePgp and MDRI expression in specimens of AML, renal tumors,normal kidney and brain, and alveolar rhabdomyosarcomas.

AML. Table 3 comparesthe detection of Pgp expression in AMLby flow cytometry and immunocytochemistry, using a panel of fiveanti-Pgp antibodies. The results obtained with the MRK-16 antibody,in contrast to the others, correlated best with dye-efflux data in thesesamples and thus represent the actual expression of Pgp. Flow cytometry was clearly superior in distinguishing specimens with low orintermediate levels of Pgp expression. In contrast to the spectrum ofstaining seen with MRK-16, identical results were produced by theC494, C219, and JSB-1 antibodies. The occasional false positiveresult yielded by immunochemical analysis with C219, C494, JSB-1,or 4E3 antibodies probably relates to the admixture of normal cells inAML specimens.

The correspondence of MDRI mRNA detection in AML specimensby rt-PCR with the functional studies (Table 4) was not always clear,

KB 3-1(PGP -)

Fixed

KB CellCulture

CollodionBag "Sausage"

CompositeFixation Slide

KBV-1(PGP+)

12 hr

24 hr

48 hr

Fig. 4. Schema of solid tumor test system for variable fixation of Pgp.The sausage technique shown here is detailed in the text and was usedin the immunohistochemical staining procedures for solid tumors intrial 3.

3014




A. IHC Staining B. Image Analysis

positive negative

W " Â¿ *"'

Ratio of CytoSDIOO KBVI toCytoSD 100KB3-1

20000

100

Â£_Â«SxR-Ã KBV-1 KB 3-1 5 1 10

InstitutionFig. 5. Trial 2: Immunohistochemical detection of Pgp in a solid tumor KB cell model. Cell buttons were prepared for embedding in paraffin by the sausage method illustrated in

Fig. 4. Pgp density was quantitated by image analysis as detailed in "Materials and Methods."

because in some instances PCR identified negative samples as positive, perhaps because of the normal blood elements shown to expressPgp (55-57) in leukemic cell samples.

Solid Tumors. To assess the interlaboratory variability of Pgpdetermination in solid tumor specimens, the participating laboratoriesexchanged both snap-frozen and paraffin-embedded samples of nor

mal kidney, renal cell carcinoma, and alveolar rhabdomyosarcoma,which then were then fixed, air dried, and stained with anti-Pgpantibodies at the concentrations listed in Table 1. Staining of 80-90%

of tumor cells in both frozen and paraffin-embedded sections of

rhabdomyosarcoma was achieved with the C494 antibody (Fig. 7A);comparable results were obtained with C219. Variable staining wasseen with 4E3, whereas HYB241 failed to identify specimens containing Pgp (Table 5).

Although the HYB241, C219, 4E3, UIC2, and MRK-16 antibod

ies uniformly stained normal kidney, mainly in the proximal tubules (Fig.IB), they produced variable staining when applied to tissue partiallyeffaced by renal carcinoma cells. For example, UIC2,4E3, MRK-16, and

Fig. 6. Trial 3: histograms comparing differentassay results for Pgp presence or function in 8226cell lines. For each test set (A-F), there is anexpected result (ER), with the histogram showingthe fraction by each method obtaining this result.A. 8826/S cells: ER = 0. ÃŸ.Dox I cells; ER = +.C, Dox 6 cells; ER = + +. D. Dox40 cells;ER = + + + . E, 90% 8226 S cells and 10% Dox40 cells; ER = +/0. F. 99% 8226 S cells and 1%Dox 40 cells; ER = +10. Eleven institutions con

tributed to the immunocytochemical results,whereas from 3 to 5 separate institutions contributed to the PCR. flow cytometric. and efflux re-

Achbvtig

A (8226 S) ERÂ«=

100

ICC PCR Flow Efflux

â€”¿�IIâ€”IIâ€”I 100D (Dox 40) ER =+++1001mgi0â€”28

294 4445 8

ICC PCR Flow Effluxâ€¢¿�ERs expected results

3015

B (Dox 1) ER

100

1531m

-2--2-T4 5l 4| iâ€”¿�i iâ€”¿�i

C (Dox 6) ER

100

ICC PCR Flow Efflux ICC PCR Flow Efflux

/822690%S \\8226 10% Dox 40/ =WO F /822699%S

\ 8226 1%Dox40 ER = WO

100

ICC PCR Flow Efflux ICC PCR Flow Efflux




Table 3 Comparison of flow cytometric and immunocytochemical detection of Pgp in AML samples using a five-antibody panel

MethodFlow

cytometry

ICC (cytospins)AML

patient1

231

23MRK-16*+

(4/5)+ (4/5)- (1/4)- (0/5)- (0/5)- (0/4)4E3Â±

(1/2)- (0/2)

Â±(1/2)- (1/5)-(2/5)- (0/5)Antibody"C494+

(2/2)+ (2/2)+ (2/2)Â±(3/6)Â±(3/6)Â±(4/6)C219+

(2/3)+ (2/3)+ (2/3)-(2/6)-(2/5)-(2/5)JSB-1+

(1/1)+ (1/1)+ (1/1)-(1/5)-(1/4)-(2/5)

"The data are summarized as the overall outcome of staining ( + , positive results by >50% of the participating laboratories; -, <50% positive; Â±,50%). Numbers in parentheses

denote the actual fraction of laboratories with a positive finding.Results obtained with the MRK-16 antibody agree with dye-efflux functional studies and therefore represent the true variability of the different methods.

Table 4 Variability of MDRI/Pgp detection by rt-PCR and functional assays inidentical samples of leukemia cells

AMLpatient1

23Functional

study"rt-PCR"+

(4/6)+ (4/7)-(3/7)RHO''+

(2/2)+ (2/2)Â±(1/2)DNR+

(3/3)-(0/2)- (0/3)DIOC2+

(1/1)+ (1/1)-(0/1)Hoechst+

(1/1)+ (1/1)-(0/1)

" The data are presented as the overall study outcome for a particular assay ( + , positiveresults by >50% of the participating laboratories; -. <50%; Â±,50%). Numbers in

parentheses denote the actual fraction of laboratories achieving a certain result.* RHO. rhodamine; DNR. daunorubicin.

HYB241 failed to stain renal tumors altogether (Fig. IB), whereas JSB-1

identified a few scattered cells that were regarded as background lymphocytes, macrophages, or both. By contrast, the C219 and C494 antibodies stained 80-90% of the tumor cells. Thus, recognition of Pgpexpression in renal cell carcinoma may be difficult with immunohisto-

chemical techniques. The variable results in this trial likely reflect thedependence of antibody sensitivity on fixation conditions and otherfactors, as shown in trial 2.

CONSENSUS RECOMMENDATIONS

The major goals of this workshop were to determine the inter-and intrainstitutional variability of measurements of Pgp andMDR1 levels in clinical specimens, to form a working group ofinvestigators committed to improving the quality of research in

MDR, and to establish guidelines for the detection of Pgp andMDR1 in clinical specimens. We believe that all of these objectiveswere met.

The following are consensus recommendations for the detection ofMDRlfPgp expression and Pgp function in clinical specimens, arrivedat by analysis of results in the workshop trials. It will be seen thatmany of the statements on immunohistochemical analysis, below, arethe same as those that follow for flow cytometry. This duplication wasconsidered necessary because of readers who might be searching forinformation on specific methods. An outline form of presentation wasselected to enhance information retrieval.

I. Detection of Pgp by ICC/IHCA. Overall Assessment. ICC/IHC has intrinsic advantages for

the detection of Pgp expression in leukemias and solid tumors, including: (a) delineation of microanatomic details; (b) recognition oftumor cell heterogeneity; (c) delineation of admixed normal cellsbearing Pgp; and (d) opportunity to correlate histology and phenotype.Subjectivity is the primary disadvantage.

B. Antibody Reagents. Because some commercially availableanti-Pgp antibodies recognize epitopes on other molecules (21, 22), atleast two antibodies recognizing separate epitopes (one surface, onecytoplasmic) should be used; only limited confidence can be placed instudies using single antibodies (6).

C. Quality Control. Antibody manufacturersare strongly encouraged to control lot-to-lot variability in their analytic reagents, to

A. Rhabdomyosarcoma

C494

B. Renal Carcinoma

UIC2

Paraffin sectionwith C494

Negative Controlwithout C494

Normal KidneyControl

Renal Carcinoma

Fig. 7. Trial 4: Detection of Pgp by immunohistochemical staining in solid tumors and normal kidney. Staining of Pgp with antibodies C494 and UIC2 are shown in representativeparaffin-embedded sections of rhabdomyosarcoma. normal kidney, and renal carcinoma. See "Materials and Methods" and text for further details.

3016




Table 5 Assessment of Pgp in paraffin blocks and frit-en specimens

Immunostaining with antibody"

SampleRhabdomyosarcoma

FrozenParaffinRenal

tumorsFrozen

ParaffinNormalkidneyC494+

(5/5)+(5/5)Â±(2/4)

+(4/5)+(5/5)C219+

(3/3)+(3/3)+

(2/5)+(4/5)+

(5/5)JSB-IND*NDÂ±(2/4)

-(0/3)''+

(5/5)UIC2NDND-(0/2)

-(1/3)-d/3)4E3+

0/1)-(0/1)-(0/1)

-(0/1)+

(1/1)HYB

241-(0/1)-(0/1)ND

NDNDrt-PCRNDND4.0eND14.6r

"The data are presented as the overall study oulcome for a particular detection method ( + . positive result by >507r of the participating laboratories; -. <50% positive; Â±,50%).

Numbers in parentheses denote actual fractions of laboratories achieving a particular result.* ND, not determined.' Arbitrary units.'' A few, scattered Pgp+ cells.

provide information on antibody purity and specificity, and to standardize their preparations of anti-Pgp antibodies with well-characterized A/D/?//Pgp-expressing cell lines, as suggested by the workshop.

(Also see sections II.B.l and 1I.B.2, below).D. Controls. Well-characterized MD/?//Pgp-expressing cell

lines, antibody specificities, and fixation-control tissues should be

matched to the specific tumors and organ sites under investigation. Itis critical that cell lines reflect the physiological or pathophysiologicallevel and pattern of site-specific Pgp expression in vivo. For example,

we favor the 8226/Dox6 cell line over the 8226/Dox40 line for studiesof myeloma and other hematological malignancies, in part because itslevel of drug resistance and Pgp expression is lower and more clinically relevant. Finally, a "gold standard" control cell line should be

used in multiparameter MDRl/Pgp analyses. Although the 8226/Doxseries is well suited for this purpose (27), other well-characterizedPgp-expressing cell lines, such as the CEM/VLB series of increasing

but low levels of resistance, would also be appropriate (e.g., Refs. 25,26, and 45).

E. Standardization of Tissue Handling and Preparation1. Cytospin Preparations. We recommend cytospin prepara

tions of cells freshly dipped in ice-cold (4Â°C)acetone for 10 min after

air drying. The slides should then be stored in a cool, dry environmentuntil the time of assay.

2. Preparations of Snap-Frozen Sections. We recommend

that freshly acquired tissue samples be embedded in OCT compoundand snap-frozen in liquid nitrogen alone or in isopentane (at â€”¿�150Â°C)

quenched in liquid nitrogen.3. Paraffin-embedded Tissues. Standardized fixation should

be used with paraffin-embedded tissues in prospective studies. We

recognize the limitations of data interpretation in retrospective analyses based on paraffin-embedded tissues sections; hence, the condi

tions for optimization of Pgp detection are important and will bereported in detail elsewhere.6 In general, we recommend a brief (4 h)

tissue fixation in 10% neutral buffered formalin for Pgp assays notrequiring antigen retrieval. Antigen retrieval methods are requiredwhen longer fixation periods and other fixatives are used.

F. Data Analysis. Distinguishing between Pgp expression intumor versus normal cells (e.g., Refs. 55-57) is a major challenge one

faces when analyzing the results of immunohistochemical assays.Moreover, the relative importance of the level of Pgp expression (i.e.,intensity of antibody staining) and the percentage of Pgp+ cells iscurrently unknown. We therefore recommend that both types of databe collected in any study of Pgp expression.

1. The intensity of antibody staining should be reported asnegative ( â€”¿�), low (+ ), intermediate ( + + ), high (+ + + ), or ultrahigh

2. We strongly recommend that arbitrary minimal cutoffpoints not be used in data analysis.

G. Multiparameter Assays. All workshop participants encouraged the use of multiparameter assays in studies of tumor cell Pgpexpression. Ideally, the methods should include assessments of protein level (immunochemistry or immunoblot analysis), mRNA (rt-

PCR or RNase protection assay), and functional activity (drug or dyeefflux). Pgp+ cell lines, such as 8226/Dox or CEM/VLB, should beused as standards to validate methods and control for variability.

II. Flow Cytometric Detection of Pgp Expression. Most of thesecomments refer to the analysis of Pgp and MDR I in hematologicalspecimens (bone marrow, blood, and disaggregated lymph nodes).Evaluation of solid tumors, although addressed, was not a major topicin this workshop because of the special problems posed by analysis ofPgp expression in such tissue. Nonetheless, some of the followingrecommendations also apply to specimens of disaggregated solidtumors, as noted.

A. Sample Handling. Flow cytometric assessment of Pgpshould be performed on fresh samples; on transported samples thathave been held for no longer than 24 h, either at room temperature oron wet ice; or on samples that have been cryopreserved as cellsuspensions frozen in 20-90% fetal bovine serum and 10% DMSO at- 135Â°C(15). These methods of sample handling also allow mRNA to

be isolated for analysis of MDRI by rt-PCR.

B. Antibody Reagents1. Epitopes. For flow cytometric studies, anti-Pgp antibodies

that have been shown to be specific for Pgp and that recognizeexternal epitopes are recommended. These presently include MRK16(33), UIC2 (32), 4E3 (29), and HYB241 (34). Whereas MRK16 andUIC2 can detect Pgp in the majority of leukemia cases that arepositive for MDRI by rt-PCR, HYB241 may only recognize this

protein in a subset of MRK+ cases of myeloid leukemia (Ref. 15 andreferences therein). Anti-Pgp antibodies that recognize external

epitopes are preferred because they allow correlation of Pgp expression with other cell surface antigens by multicolor flow cytometry andwith functional measurements of dye efflux/accumulation. They alsoallow analysis of Pgp expression on leukemic cells [see section II.D(Gating), below] with exclusion of residual normal Pgp+ cells (e.g.,T cells, CD56+ cells) and dead cells that have taken up antibodynonspecifically.

2. Quality Control. Antibodies used for Pgp detection mustbe rigorously controlled for quality. We recommend that commercialvendors affinity-purify Pgp-specific antibodies; test antibody purityby SDS-PAGE or other suitable methods; and test antibody specificityin ELISA assays, as well as on a series of standardized drug-sensitiveand increasingly drug-resistant Pgp-expressing cell lines, such as the8226/Dox series (27), the CEM/VLB series (25), or the KB-series (26,

45). We strongly recommend that manufacturers provide completestandardization information with each new lot of antibodies, and weencourage the development of new high-affinity anti-Pgp antibodies.

3017




3. Antibody Staining and Fluorochrome Detection. An-ti-Pgp antibodies should be directly conjugated to either a fluoro-

chrome that emits in the red (e.g., PE; Refs. 15 and 58) or theduochrome reagent streptavidin-Texas Red-PE (15, 58) rather than to

FITC. Detection with a secondary antibody conjugated to a redfluorochrome instead of to FITC is preferred. PE and the duochromereagent are regarded as superior for at least two reasons (58): (a) theformer reagent has a much higher quantum efficiency than FITC andpermits more reliable detection of low but significant levels of Pgpexpression in hematological samples (58). Indeed, some participatinginvestigators reported that a fraction of leukemic samples were Pgp+with use of a PE conjugate but negative with FITC conjugates,particularly cells with low Pgp antigen density; and (b) becausenormal cells autofluoresce in the blue-green (FITC) range (58), use of

dyes that emit in the red allows detection of a wider range of channelshifts above the background control (58), thereby permitting greaterS:N ratios and allowing for detection of cells with very low Pgpexpression.

C. Controls1. All antibody reactions should have an isotypically matched

control, used at the same concentration as the primary anti-Pgpantibody. A "no-primary-antibody" control is also required.

2. As recommended for immunocytochemical staining, well-characterized Pgp-expressing cell lines are required to validate and

standardize flow cytometric assays and to calibrate instruments.Workshop participants recommended the use of the low-level Pgp-expressing doxorubicin-resistant series of 8226 myeloma cell lines(27), although, as discussed above, any well-characterized cell line

series of low Pgp expression (e.g., some CEM/VLB sublines; Ref. 25)can be used. Because of the very low Pgp levels in most leukemicsamples, compared to those in Dox6 cells, participants recommendedthe development of more Pgp-expressing sublines with resistance

levels between those of 8226/S and Dox6. Participants suggested,moreover, that standardized Pgp-MDR cell lines be maintained and

distributed from a single reference laboratory or company.D. Gating. Electronic windows (gating) should be used to ana

lyze Pgp expression by cells with the size and granularity of blastcells. This method allows selective study of Pgp expression on leu-

kemia/lymphoma cells and eliminates residual normal cells that express Pgp+, a potentially significant problem in quantifying Pgpexpression (15).

E. Data Analysis. There are many ways to evaluate and reportflow cytometric data, including the percentage of positive cells, assessment of the mean channel shift (difference in mean channelbetween the control and the experimental sample), and the D value(Kolmogorov-Smirnov statistic; Ref. 15). Each of these measurements

has its strengths and weaknesses (15), precluding recommendation ofa preferred method; however, the participants agreed to report data ascontinuous variables rather than establish a cutoff point for assessingpositivity.

III. Measurement of Pgp FunctionA. Dye or Drug Efflux.

1. Measuring dye or drug efflux in the presence and absenceof a Pgp modulator (e.g., cyclosporin, PSC388, or verapamil) ispreferable to measuring dye or drug accumulation only (15, 44-46).

2. Several dyes (DIOC2, Rhol23, or Hoechst) or drugs(daunorubicin or doxorubicin) may be used to assess Pgp function (15,45, 46); however, some dyes appear to yield better functional estimates than do the actual drug substrates (15), probably because ofmore favorable uptake/efflux kinetics. Because the basis for thisdiscrepancy remains unclear, we recommend that both methods beapplied to clinical specimens.

3. Because most functional measurements are performed withflow cytometry, it is desirable to correlate the functional parameter(dye efflux) with Pgp protein expression as detected with anti-Pgp

antibodies. With this approach, it is possible to detect drug efflux inthe absence of Pgp expression, an indication of other (non-MDR)

mechanisms of drug resistance (45).B. Data Analysis. See Section U.E.

IV. rt-PCRA. Assessment of Tumor Cell Percentage. Analysis by rt-PCR

may be the most specific and most sensitive assay for MDR1 mRNA,but it does not allow identification of the actual cells (tumor ornormal) that express the gene. Thus, we recommend that all rt-PCR

assays be accompanied by a determination of the tumor cell content ofthe sample undergoing analysis.

B. Amplification Reaction. All rt-PCR assays should be per

formed in the exponential range and not the plateau range of theamplification reaction. This allows detection of MDRl expression inthe dominant cell population of a sample, rather than in rare subpopulations. Because competition can occur between different primer setsin a multiplex reaction, it is essential to show no competition betweenprimer pairs.

C. Internal Standards. Both ÃŸ-2microglobulin (50) and his-

tone 3.3 (49) afford reliable internal standards to ensure that samplescontain high-quality RNA suitable for PCR amplification. We do notrecommend the use of ÃŸ-actin,as its expression is variable and candecrease in drug-resistant cells (59).

D. Controls. As with flow cytometry, it is essential to establishPCR assays with negative controls and with a standard series ofdrug-sensitive and increasingly resistant cell lines (see section II.C).

FUTURE CONSIDERATIONS

One of the most important achievements of the St. Jude workshopwas to reach consensus conclusions on means to standardize methodsfor the study of Pgp expression in clinical material. It was also clearthat hematological malignancies offer the best current opportunity forreliable detection of this protein, and that much work remains to bedone before solid tumors can be profitably examined for Pgp over-

expression. Given the complexity of solid tumors, in which clinicaldrug resistance can have pharmacodynamic and pharmacokinetic elements, as well as biochemical elements, any future workshop onMDR must focus on the detection of Pgp in these types of neoplasms.It will also be important to expand the workshop agenda to includeother markers of MDR, such as MRP (45; 60-63), LRP (64, 65),

topoisomerase II (66), and others. Although multiparameter analysesare clearly superior in studies of Pgp levels and function, there is stilla need to develop novel strategies that would integrate the specificityof molecular methods with methods of cellular delineation (e.g.,immuno-PCR or immuno-/'n situ, or in situ PCR) to improve the

identification of resistance-associated proteins.

Many unresolved questions cannot be addressed with confidenceuntil gene/protein detection methods and the understanding of resistance mechanisms have improved. For example, although we believethat we can now reliably measure Pgp, at least in hematologicalmalignancies, accurate measurement of Pgp expression under mostconditions remains an elusive goal. Indeed, it is still not clear whatmethods are most appropriate for determining the MDR status ofpatients' tumors. Clarification of this issue should allow determination

of the numbers of Pgp molecules (or other markers of drug resistance)required to confer resistance to the tumor or tumor cell. This isdirectly related to our present ignorance about the level of, e.g., Pgpdetection that correlates with the acquisition of drug resistance by apatient's tumor cells. Once a consensus regarding the "truth" of Pgp

3018




expression in a tumor has been reached, it will be possible to improvethe design of retrospective and prospective studies of how MDR]expression is related to clinical outcome. Another critical question iswhether or not discrepancies among different Pgp/MDRl assays reflect intrinsic differences in basic tumor cell biology. Our hope is thatthese and other issues will be topics at future MDR workshopsattended by laboratories worldwide.

ACKNOWLEDGMENTS

The authors express their gratitude to St. Jude Children's Research Hospital,

its director, Dr. Arthur Nienhuis. and its former director. Dr. Joseph Simone,for their unwavering support of this project. We are also indebted to membersof the Pathology Department of the University of Arizona Cancer Center fortheir many contributions and to Ingenex, Kimiya Biochemicals, and Signet forgenerous gifts of antibodies. Special thanks go to Dr. Ron Casciato of Signetfor preparing uniform lots of antibodies.

We owe an enormous debt to John Gilbert, Consulting BiomÃ©dicalEditor,for his assistance in condensing and clarifying the manuscript. Many otherindividuals contributed mightily to the organization and conduct of the workshop, and to the preparation of its proceedings. The following people from St.Jude Children's Research Hospital and Memphis deserve special mention and

gratitude: Dolores Anderson, Fabrienne Holloway, and Sandy Lewis for theirtireless efforts throughout; Dr. Jesse Jenkins and Hallie Holt (Department ofPathology); John Zacker, Linda Rawlinson, and Jerry Luther (BiomÃ©dicalCommunications Department); Marianna Quinn and her cafeteria staff; OlaElion and her housekeeping staff; and Margaret Pratt (Regency Travel) forarranging transportation for the participants. We are also grateful to JeanYarab, Mary Guzman, and Liz Vela (University of Arizona) and to Dr. I-MingChen (University of New Mexico) for preparation and distribution of allleukemic specimens. For other technical support, we thank Karen Redman(Medical Research Council, Cambridge, United Kingdom), Nicole Feller (FreeUniversity Hospital, Amsterdam, the Netherlands), and Michelle Anderson andRichard Wee (British Columbia Cancer Agency, Vancouver, British Columbia,Canada).

APPENDIX

This study was supported by: NCI Grants CA-57639 and CA-55164(M. A.); NCI Grants CA-30103 and CA-23099, NCI Cancer Center Support(CORE) Grant CA-21765, and ALSAC (W. T. B.); NCI Grant CA-21765and ALSAC (C. W. B.); NCI Grant CA-21765 and ALSAC (J. M. B.); NCIGrant CA-52168 and American Cancer Society Grant DHP-76 (N. A. B.);Dutch Cancer Society Grant NKB-VU 93-626 (H. J. B.); grants from theNational Cancer Institute of Canada (H. S. L. C.); H. S. L. C. is a ResearchScientist of the National Cancer Institute of Canada; NCI Grants CA-47538and CA-47179 (C. C-C); NCI Grants CA-43043 and CA-17094 (W. S. D.);British Columbia Health Research Foundation Grant 59 (94-1) (R. D. G.);NCI Grant CA-32102, supporting the SWOG Lymphoma and MyelomaBiology Programs (T. M. G.); Technology Transfer Grant from the University of Arizona, Department of Pathology (T. M. G.); Cancer CenterSupport (CORE) Grant CA-21765, NCI Grant 5U10 CA-24507, and ALSAC (D. H.); Cancer Center Support (CORE) Grant CA-21765, NCI GrantCA-23099. and ALSAC (P. J. H.); NCI Grants CA-23099 and CA-21765and ALSAC (J. F. K.); Helmut Horten Foundation Grant SNF 31-37521.93(M. L.); NCI Grants CA-32102 (supporting the SWOG Leukemia BiologyProgram) and CA-60433 (C. P. L.); NCI Grants CA-14958, CA-21115, andP30 CA-1330 and Phi Beta Psi sorority (E. P.); NCI Grants CA-21765 andCA-24507 and ALSAC (D. M. P.); URC-1990-91 (Z. P. P.); NCI GrantCA-32102 (L. R.); NCI Grant CA-40333 (I. B. R.); NCI Grant CA-52168and American Cancer Society Grant DHP-76 (B. I. S.); NCI Grants CA-21765 CA-23099 and ALSAC (D. K. S.); United Kingdom LeukaemiaResearch Fund (P. R. T.); 2 HMA595 (R. Wa.); NCI Grant CA-32102(R. We.); NCI Grants CA-32102 (supporting the SWOG Leukemia BiologyProgram) and CA-60433 (C. L. W.); and by Eli Lilly, Glaxo, Ingenex,Oncotech, Pfizer, Sandoz, Signet, and Warner-Lambert/Parke-Davis.

REFERENCES

1.Goldstein, L. J.. Galski, H., Fojo. A., Willingham. M., Lai, S-L.. Gazdar. A.. Pirker,R.. Green, A.. Crist. W., Brodeur, G. M.. Lieber. M., Cossman, J.. Gottesman. M. M.,and Pastan, I. Expression of a multidrug resistance gene in human cancers. J. Nati.Cancer Inst.. SI: 116-124. 1989.

Fojo, A. T., Ueda, K., Slamon. D. J.. Poplack. D. G., Gottesman, M. M., and Pastan,I. Expression of multidrug-resistance gene in human tumors and tissues. Proc. Nati.Acad. Sei. USA, 84: 265-269. 1987.

Herzog, C. E., Trepel, J. B., Mickley, L. A., Bates, S. E., and Fojo. A. T. Variousmethods of analysis of mdr-1/P-glycoprotein in human colon cancer cell lines. J. Nati.Cancer Inst., 84: 711-716, 1992.

Brophy. N. A.. Marie, J. P., Rojas, V. A., Wamke. R. A., McFall. P. J.. Smith, S. D.,and Sikic, B. I. Marl gene expression in childhood acute lymphoblastic leukemiasand lymphomas: a critical evaluation by four techniques. Leukemia (Baltimore). 8:327-335, 1994.

Chan. H. S. L., DeBoer. G., Haddad, G.. Gallic, B. L., and Ling, V. Multidrug resistancein pediatrie malignancies. Hematol. Oncol. Clin. N. Am.. 9: 275-318. 1995.van der Heyden, S., Gheuens. E.. De Bruijn, E., Van Oosterom. A., and Maes. R.P-glycoprotein: clinical significance and methods of analysis. Crit. Rev. Clin. Lab.Sci.. 32: 221-264, 1995.

Lum, B. L.. Fisher, G. A., Brophy, N. A., Yahanda, A. M., Adler, K. M., Kaubisch,S.. Halsey, J., and Sikic, B. I. Clinical trials of modulation of multidrug resistance.Pharmacokinetic and pharmacodynamic considerations. Cancer (Phila.). 72: 3502-3514. 1993.Raderer. M.. and Scheithauer. W. Clinical trials of agents thai reverse multidrugresistance. A literature review. Cancer (Phila.), 72: 3553-3563. 1993.

Fisher. G. A., and Sikic, B.I. Clincial studies with modulators of multidrug resistance.Hematol. Oncol. Clin. N. Am.. 9: 363-382, 1995.

Sikic, B. I. Modulation of multidrug resistance: at the threshold. J. Clin. Oncol., II:1629-1635, 1993.Arceci, R. J. Clinical significance of P-glycoprotein in multidrug resistance malignancies. Blood, 81: 2215-2222, 1993.Kaye, S. B. P-glycoprotein (P-gp) and drug resistance: time for reappraisal? Br. J.Cancer, 67: 641-643, 1993.Murren. J. R.. and Hait. W. N. Why haven't we cured multidrug resistant tumors?

Oncol. Res.. 4: 1-6, 1992.

Bates. S. E., Meadows, B.. Goldspiel, B. R.. Denicoff, A.. Le. T. B.. Tucker, E.,Steinberg, S.. and Elwood, L. J. A pilot study of amiodarone with infusionaldoxorubicin or vinblastine in refractory breast cancer. Cancer Chemother. Pharma-col.. 35: 457-463, 1995.

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. Correlation of multidrug resistance (MDR1)protein expression with functional dye/drug efflux in acute myeloid leukemia bymultiparameter flow cytometry: identification of discordant MDR-/efflux+ andMDR+/efflux- cases. Blood. 86: 2329-2342, 1995.Chan. H. S. L., Thomer. P.. Haddad. G.. and Ling, V. Immunohistochemical detectionof P-glycoprotein: prognostic correlation in soft tissue sarcoma of childhood. J. Clin.Oncol.. 8: 689-704, 1990.

Chan. H. S. L.. Haddad, G.. Thorner. P. S., DeBoer. G., Lin, Y. P., Ondrusek, N.,Yeger, H., and Ling, V. P-glycoprotein expression as a predictor of the outcome oftherapy for neuroblastoma. N. Engl. J. Med.. 325: 1608-1614. 1991.

Pirker. R.. Wallner, J., Geissler, K., Linkesch, W.. Haas, O. A.. Bettelheim. P.,Hopfner, M., Scherrer, R., Valent, P., Havelec, L., Ludwig. H., and Lechner, K.MDRI gene expression and treatment outcome in acute myeloid leukemia. J. Nati.Cancer Inst.. 83: 708-712, 1991.

Marie. J. P., Zittoun. R.. and Sikic. B. I. Multidrug resistance (MDRI) gene expression in adult acute leukemias: correlations with treatment outcome and in vitro drugsensitivity. Blood, 78: 586-592, 1991.Kuttesch, J. F.. Parham, D. M., Luo. X.. Meyer. W. H., Bowman, L.. Shapiro, D. N.,Pappo. A. S.. Crist, W. M., Beck, W. T., and Houghton, P. J. P-glycoprotein

expression at diagnosis does not correlate with response or outcome in childhoodrhabdomyosarcoma. J. Clin. Oncol.. 14: 886-900, 1996.

Thiebaut, F., Tsuruo, T.. Hamada. H . Gottesman. M. M., Pastan. I., and Willingham.M. C. Immunohistochemical localization in normal tissues of different epitopes in themultidrug transport protein P170: evidence for localization in brain capillaries andcrossreactivity of one antibody with a muscle protein. J. Histochem. Cytochem., 37:159-164, 1989.Rao. V. V.. Anthony. D. C.. and Piwnica-Worms, D. MDR! gene-specific monoclonalantibody C494 cross-reacts with pyruvate carboxylase. Cancer Res.. 54: 1536-1541.1994.Hamblin. T. J. Meeting report. 1st international conference on reversal of multidrugresistance in cancer. Leuk. Res., 19: 427-428. 1995.

Toth, K.. Vaughan, M. M.. Slocum. H. K.. Arredondo, M. A.. Takita. H.. Baker.R. M.. and Rustum, Y. M. New immunohistochemical "sandwich" staining method

for mdrl P-glycoprotein detection with JSB-I monoclonal antibody in formalin-fixed,paraffin-embedded human tissues. Am. J. Pathol., 144: 227-236. 1994.Beck. W. T., Mueller, T. J., and Tanzer. L. R. Altered surface membrane glycopro-teins in Vinca alkaloid-resistant human leukemic lymphoblasts. Cancer Res., 39:2070-2076. 1979.Akiyama, S-L, Fojo, A.. Hanover, J. A.. Pastan. L, and Gottesman. M. M. Isolationand genetic characterization of human KB cell lines resistant to multiple drugs.Somatic Cell Mol. Genet., //: 117-126, 1985.Dalton, W. S.. Dune, B. G. M., Aliens. D. S.. Gerlach. J. M.. and Cress, A. E.Characterization of a new drug resistant human myeloma cell line that expressesP-glycoprotein. Cancer Res., 46: 5125-5130. 1986.

3019




28. Shen, D-W., Cardarelli, C.. Hwang, J.. Cornwell, M., Riehen, N.. Ishii. S., Pastan, I.,and Gottesman, M. M. Multiple drug-resistant human KB carcinoma cells independently selected for high-level resistance to colchicine, adriamycin, or vinblastine showchanges in expression of specific proteins. J. Biol. Chem., 261: 7726-7770, 1986.

29. Arceci, R. J., Stieglitz, K., Bras. J.. Schinkel, A.. Baas, F., and Croop. J. Monoclonalantibody to an external epitope of the human mdrl P-glycoprotein. Cancer Res., 53:310-317, 1993.

30. Kartner, N., Evemden-Porelle. D., Bradley. G., and Ling. V. Detection of P-glyco

protein in multidrug resistant cell lines by monoclonal antibodies. Nature (Lond.),316: 820-823, 1985.

31. Scheper, R. J., Bulle. J. W. M., Brakke, J. G. P.. Quak, J. J., van der School. E.. Balm,A. J. M., Meijer, C. J. L. M., Broxterman, H. J.. Kuiper. C. M.. Lankelma, J.. andPinedo, H. M. Monoclonal antibody JSB1 detects a highly conserved epitope on theP-glycoprotein associated with multidrug resistance. Int. J. Cancer. 42: 389-394,

1988.32. Mechetner, E. B., and Roninson. I. B. Efficient inhibition of P-glycoprotein-mediated

resistance with a monoclonal antibody. Proc. Nati. Acad. Sci. USA, 89: 5824-5828,

1992.33. Mamada, H., and Tsuruo, T. Functional role for the 170- to 180-kDa glycoprotein

specific to drug-resistant tumor cells as revealed by monoclonal antibodies. Proc.Nati. Acad. Sci. USA, 83: 7785-7789, 1986.

34. Meyers, M. B.. Rittmann-Grauer, L., O'Brien, J. P., and Safa, A. R. Characterization

of monoclonal antibodies recognizing a Mr 180,000 P-glycoprotein: differentialexpression of the M, 180,000 and M, 170,000 P-glycoproteins in multidrug resistanthuman tumor cells. Cancer Res., 49: 3209-3214, 1989.

35. Rimsza, L., Weinstein, R.. Rangel, C., and Grogan, T. Optimization of P-glycoprotein

detection in formalin fixed paraffin embedded tissues utilizing antigen retrievalmethods. Anticancer Drugs, J: 11-12, 1994.

36. Grogan, T., Dalton, W., Rybski. J., Spier, C., Meltzer. P., Richter, L., Gleason, M..Pindur, J., Cline, A., Scheper. R.. Tsuruo, T., and Salmon S. Optimization ofimmunocytochemical P-glycoprotein assessment in multidrug-resistant plasma cellmyeloma using three antibodies. Lab. Invest., 63: 815-824, 1990.

37. Chan, H. S. L., Bradley, G., Thorner, P., Haddad. G., Gallic, B. L., and Ling, V. Asensitive method for immunocytochemical detection of P-glycoprotein in multidrug-resistant human ovarian carcinoma cell line. Lab. Invest., 59: 870-875, 1988.

38. Cordeil, J. L., Falini, B.. Erber, W. N., Ghosh, A. K., Abdulaziz, Z., MacDonald. S..Pulford, K. A., Stein, H., and Mason. D. Y. Immunoenzymatic labeling of monoclonalantibodies using immune complexes of alkaline phosphatase and monoclonal andanti-alkaline phosphatase (APAAP complexes). J. Histochem. Cytochem.. 32: 219-229, 1984.

39. Haddad, G., Thomer, P. S., Bradley, G., Dalton, W. S., Ling. V.. Chan, H. S. L. Asensitive multilayer immunoalkaline phosphatase method for detection of P-glycoprotein in leukemic and tumor cells in the bone marrow. Lab. Invest., 71: 595-603,

1994.40. Pavelic. Z. P., Sever, Z., Fontaine, R., Baker, V. V., Reising, J., DentÃ³n, D. M.,

Pavelic. L. J., and Khalily. M. Detection of P-glycoprolein with JSB-1 monoclonalantibody in B-5 fixed and paraffin-embedded cell lines and tissues. Sel. Cancer Ther..7: 49-58, 1991.

41. Giorno, R. A comparison of two immunoperoxidase staining methods based on theavidin-biotin interaction. Diagn. Immunol.. 2: 161-166. 1984.

42. Gown, A. M., de Wever, N.. and Baltifora. H. Microwave-based antigenic unmasking: a revolutionary new technique for routine immunohistochemistry. Appi. Immu-nohistochem.. /: 256-266. 1993.

43. Paietta, E.. Andersen, J., Racevskis, J., Gallagher, R., Bennett, J., Yunis. J.. Cassileth,P., and Wiemik, P. H. Significantly lower P-glycoprotein expression in acute pro-myelocytic leukemia than in other types of acute myeloid leukemia: immunological,molecular and functional analyses. Leukemia (Baltimore). 8: 968-973, 1994.

44. Broxterman. H. J.. Feller. N., Kuiper. C. M., Boven, E.. Versantvoort, C. H. M.,Teerlink. T., Pinedo. H. M., and Lankelma, J. Correlation between functional andmolecular analysis of mdrl P-glycoprotein in human solid-tumor xenografts. Int. J.Cancer, 61: 880-886, 1995.

45. Feller. N., Juiper, C. M., Lankelma, !.. Ruhdal, J. K., Scheper, R. J., Pinedo. H. M.,and Broxterman, H. J. Functional detection of MDRl/PWO and Af/?/VP190-mediatedmultidrug resistance in tumor cells by flow cytometry. Br. J. Cancer, 72: 543-549,1995.

46. Lee. J. S., Paull. K., Alvarez, M., Hose, C., Monks, A., Grever, M., Fojo. A. T.. andBates, S. E. Rhodamine efflux patterns predict P-glycoprotein substrates in theNational Cancer Institute Drug Screen. Mol. Pharmacol., 46: 627-638. 1994.

47. Chomczynski, P., and Sacchi, N. Single-step method of RNA isolation by acidguanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem.. 762: 156-

159, 1987.48. Murphy, L. D., Herzog, C. E.. Rudick, J. R., Fojo, A. T., and Bates, S. E. Use of the

polymerase chain reaction in the quantitation of mdr-1 gene expression. Biochemistry, 29: 10351-10356, 1990.

49. Futscher, B. W., Blake, L. L., Gerlach, J. M., Grogan, T. M., and Dalton, W. S.Quantitative polymerase chain reaction analysis of mdrl mRNA in multiple myelomacell lines and clinical specimens. Anal. Biochem., 213: 414-421. 1993.

50. Noonan. K. E., Beck, C.. Holzmayer. T. A.. Chin, J. E.. Wunder. J. S.. Andrulis, I. L.,Gazdar, A. F.. Willman, C. L.. Griffith, B., von Hoff, D. D., and Roninson, I. B.Quantitative analysis of MDRI (multidrug resistance) gene expression in humantumors by polymerase chain reaction. Proc. Nati. Acad. Sci. USA. 87: 7160-7164,

1990.51. Dietel. M., Arps, H.. Lage, H., and Niendorf, A. Membrane vesicle formation due to

acquired mitoxantrone resistance in human gastric carcinoma cell line EPG85-257.Cancer Res., 50: 6100-6106, 1990.

52. Dietel, M. What's new in cytostatic drug resistance and pathology. Pathol. Res. Pract.,

187: 892-905, 1991.

53. Dietel, M., Herzig. I., Reymann, A.. Brandt, I., Schafer. B., Bunge. A., Heidebrecht.H. J., and Seidel, A. Secondary combined resistance to the mullidrug resistancereversing activity of cyclosporin A in the cell line F4-6RADR-CsA. Cancer Res.Clin. Oncol., 720. 263-271, 1994.

54. Holm, P. S.. Scanlon. K. J.. and Dietel. M. Reversion of multidrug resistance in thep-glycoprotein-positive human pancreatic cell line (EPP85-181RDB) by introductionof hammerhead ribozyme. Br. J. Cancer, 70: 239-243, 1994.

55. Chaudhary, P. M., Mechetner, E. B., and Roninson, I. B. Expression and activity ofthe multidrug resistance P-glycoprotein in human peripheral blood lymphocytes.Blood. //: 2735-2739, 1992.

56. Drach. D., Shourong. Z., Drach, J., Mahadevia, R., Gattringer, C.. Huber, H., andAndreeff, M. Subpopulations of normal peripheral blood and bone marrow cellsexpress a functional multidrug resistant phenotype. Blood, 80: 2729-2734, 1992.

57. Klimecki, W. T., Futscher, B. W., Grogan, T. M., and Dalton, W. S. P-glycoproteinexpression and function in circulating blood cells from normal volunteers. Blood. 83:2451-2458, 1994.

58. Willman, C. L. Flow cytometric analysis of hÃ©matologie specimens. In. D. M.Knowles (ed.). Neoplastic Hematopathology, p. 169. Baltimore, MD: Williams &Wilkins, 1992.

59. Taylor, J. J., and Heasman. P. A. Control genes for reverse transcriptase/polymerasechain reaction (RT-PCR). Br. J. Haematol., 86: 444-445, 1994.

60. Cole, S. P. C., Bhardwaj, G., Gerlach, J. H., Mackie, J. E., Grant, C. E., Almquist,K. C., Stewart, A. J., Kurz, E. U., Duncan, A. M. V., and Deeley, R. G. Overexpres-sion of a transporter gene in a multidrug-resistant human lung cancer cell line.Science (Washington DC), 258: 1650-1654. 1992.

61. Grant, C. R., Valdimarsson, G., Hipfner, D. R., Almquist, K. C., Cole, S. P. C..Deeley, R. G. Overexpression of multidrug resistance-associated protein (MRP)increases resistance to natural product drugs. Cancer Res., 54: 357-361, 1994.

62. Kruh, G. D., Chan, A., Myers, K., Gaughan, K., Miki. T.. and Aaronson, S. A.Expression complementary DNA library transfer establishes mrp as a multidrugresistance gene. Cancer Res., 54: 1649-1652, 1994.

63. Schuurhuis, G. J.. Broxterman, H. J.. Ossenkoppele. G. J., Baak. J. P. A., Eekman,C. A., Kuiper. C. M., Feller, N., van Heijningen, T. H. M., Klumper, E., Pieters, R.,Lankelma. J.. and Pinedo, H. M. Functional multidrug resistance phenotype associated with combined Overexpression of Pgp/MDRI and MRP together with 1-ÃŸ-D-arabinofuranosylcytosine sensitivity may predict clinical response in acute myeloidleukemia. Clin. Cancer Res., /: 81-93, 1995.

64. Scheffer, G. L., Wijngaard, P. L. J., Flens, M. J., Izquierdo, M. A.. Slovak, M. L.,Pinedo. H. M., Meijer, C. J. L. M., Clevers, H. C., and Scheper. R. J. The drugresistance-related protein LRP is the human major vault protein. Nat. Med.. /:578-582, 1995.

65. Izquierdo, M. A., van der Zee, A. G. J., Vermorken, J. B., van der Valk, P., Belien.J. A. M., Giaccone, G., Scheffer, G. L., Rens, M. J., Pinedo, H. M., Kenemans, P.,Meijer. C. J. L. M., de Vries, E. G. E., and Scheper, R. J. Drug resistance-associatedmarker LRP for prediction of response to chemotherapy and prognoses in advancedovarian carcinoma. J. Nati. Cancer Inst., 87: 1230-1237, 1995.

66. Beck, W. T., Danks, M. K., Wolverton, J. S., Chen, M.. Grenzen, B., Kim, R., andSuttle. D. P. Resistance of mammalian tumor cells to inhibitors of DNA topoisomer-ase II. Adv. Pharmacol.. 29B: 145-169, 1994.

3020



1996;56:3010-3020. Cancer Res William T. Beck, Thomas M. Grogan, Cheryl L. Willman, et al. Resistance in Patients' Tumors: Consensus RecommendationsMethods to Detect P-Glycoprotein-associated Multidrug

Updated version

http://cancerres.aacrjournals.org/content/56/13/3010

Access the most recent version of this article at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/56/13/3010To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]



methods to detect p-glycoprotein-associated multidrug ...immunostaining, flow cytometry, northern...

Documents