Flow Cytometric Analysis of the Multiple Drug Resistance Phenotype a Krishan, a Sauerteig, I Andritsch and L Wellham

Flow cytometric analysis of the multiple drug resistance phenotype A Krishan, A Sauerteig, I Andritsch and L Wellham

Department of Radiation Oncology, University of Miami Medical School, Sylvester Comprehensive Cancer Center, Miami, FL, USA

Laser flow cytometry is increasingly used for quantitation of use of flow cytometry for functional analysis of drug retention cellular fluorescent drug retention, effect of efflux blockers and and the effect of efflux blockers can be fraught with the danger for expression of drug resistance related cellular surface markers. Several intrinsic and extrinsic factors can affect the results that artifacts may lead to erroneous conclusions. In the present obtained from drug retention functional assays and lead to arti- study, we have a panel of well-characterized cell lines, facts. In the present study, we have used a panel of well-charac- fluorochromes and efflux blockers to identify some of the terized parental and drug resistant cell lines, fluorochromes problems in the use of flow cytometry for drug resistance and efflux blockers to identify the possible sources of artefacts related marker expression and functional assays. in flow cytometric analysis of the multiple drug resistance Besides monitoring of cellular drug retention and the effect phenotype. Keywords: efflux; anthracyclines; P-glycoprotein; MDR; flow cyto- of efflux blockers, laser flow cytometry is a useful tool for metry; fluorescence detection of the marker proteins responsible for drug efflux. In a previous publication we reported on the flow cytometric and immunocytochemical comparison of three anti-P-glyco- Introduction protein antibodies (C219, JSB1 and MDR-Ab1) which recognize the cytoplasmic component of the P-glycoprotein.10 Cellular resistance may be a major factor for failure of cancer However, a major part of the P-glycoprotein is believed to chemotherapy in refractory patients. Studies over the last dec- reside in the cell membrane with only a small portion on the ade have identified multiple drug resistance (MDR) as a cellu- outer cell surface. Recently, three new monoclonals which recognize the outer cell membrane resident portion of the P- lar protective mechanism which confers resistance to a variety 11–13 of chemically unrelated natural products used in cancer glycoprotein have been described. In the present study, chemotherapy. It is evident from the numerous studies pub- we have compared expression of these three monoclonals lished during the last decade that MDR is multifactorial and (MRK16, UIC2 and 4E3) in a panel of human tumor cell lines involves not only drug efflux, but also intracellular drug distri- with MDR1 related drug efflux. bution, detoxification, DNA synthesis and repair.1,2 Rapid drug efflux mediated by a cell membrane resident P- glycoprotein (P-gp) has been shown to reduce cellular drug Materials and methods retention and chemosensitivity. Recent studies have identified MRP and LRP as two other drug resistance related proteins Human lymphocytes from the venous blood of a donor were involved in cellular drug transport, subcellular distribution retrieved by separation on a Ficoll–Hypaque gradient and resistance.3,4 (Histopaque 1077; Sigma Chemical, St Louis, MO, USA) The Besides immunocytochemistry which depends on the use of mononuclear cells were washed and resuspended in mini- monoclonal antibodies for identification of cellular resistance mum essential medium and 4% heat inactivated fetal bovine markers, fluorescent microscopy and confocal microscopy serum. The lymphocytes, at a starting concentration of 1 × 106/ml, were cultured in flasks containing 18 ml of cell and laser flow cytometry have been used to study transport ␮ and subcellular distribution of fluorescent drugs. As cellular suspension and 2 ml of 200 g/ml phytohemagglutinin-P retention and distribution is related to cytotoxicity, these (Difco Laboratories, Detroit, MI, USA). After 72 h incubation ‘functional studies’ have the potential to supplement the with mitogenic stimulation, cultures were analyzed for immunocytochemical data and report on the functional cell cycle distribution and incubated with the various drug activity of the marker proteins.5–8 The ‘bulk’ methods such as combinations. Northern, Western and Southern blot analysis which use tissue extracts or homogenates, although highly refined and sensitive, suffer from the disadvantage that individual cells Cell lines with positive marker expression cannot be identified in a heterogenous cell population.9 Murine leukemia cell lines, P388 and P388/R84: These To quantitate markers and mechanisms involved in cellular cell lines were cultured in RPMI-1640 medium (Gibco/BRL resistance requires knowledge about a variety of sophisticated Life-Technologies, Grand Island, NY, USA) supplemented with 10% fetal bovine serum, 100 units/ml of penicillin, laboratory techniques involving immunocytochemistry, flow ␮ cytometry and molecular biology. Unlike immunocytochemi- 100 g/ml of streptomycin, and 2-mercaptoethanol × −5 stry where one can identify cells with positive or negative (5.5 10 M). The doxorubicin-resistant cell line, P388/84, is reactivity, laser flow cytometry can determine cellular drug 84-fold more resistant than the parental line. retention but does not allow visual examination of the cells. Thus in spite of its sensitivity, convenience and rapidity, the CEM and CEM/VLB100: Human acute lymphoblastic leukemia cell line and its vinblastine resistant subline were kindly provided by Dr William Beck.14 The resistant subline was Correspondence: A Krishan, Experimental Therapeutics Division (R-71), University of Miami Medical School, PO Box 016960, Miami, maintained in the RPMI-1640 medium supplemented with FL 33101, USA 100 ng/ml of vinblastine. Prior to testing, cells were grown in Received 4 October 1996; accepted 3 February 1997 vinblastine-free medium for 7 to 10 days. MDR phenotype quantitation A Krishan et al 1139 SW620, SW620/Ad300, and HCT-15: These human colon son17 was used for staining of cells with UIC2 monoclonal adenocarcinoma cell lines were kindly provided by Drs Susan antibody. E Bates and Antonio T Fojo of the Medical Branch, National Cancer Institute.15,16 The SW620/Ad300 cell line is 76-fold more resistant to doxorubicin than the parental line. Prior to Results and discussion experimentation, the SW620/Ad300 cell line was grown in doxorubicin-free medium for 7 to 10 days. Heterogeneity and retention of different fluorochromes

In most of the earlier drug efflux functional assays doxorubicin Malignant melanoma (FCCM-2) and renal cell carcinoma or daumomycin was used for determination of cellular drug (FCCR-2): These cell lines were established in our labora- retention. Rhodamine 123 (Rh-123) was introduced as a fluo- tory from patient biopsies obtained for diagnostic purposes. rochrome for functional assays due to its excellent fluorescent 18 19 Cells were maintained in RPMI-1640 medium supplemented qualities. Frey et al described the use of fluorochromes, with growth factors and 15% fetal bovine serum. SY-38 and SY-3150 for monitoring of drug retention and efflux. However, it is clear that fluorochromes differ in their binding characteristics as well as in their transport and cellular Reagents retention properties. Based on data published in one of our earlier papers,6 one can argue that cellular retention data can not be extrapolated between different fluorochromes and one Monoclonal antibodies, C219 and 4E3, were purchased from should preferably test cells against the drug which is being Signet Lab (Dedham, MA, USA). MRK16 monoclonal antibody used for treatment of the tumor. was a gift from Dr Takashi Tsuruo (Tokyo University, Tokyo, Data in Figure 1 is from an experiment where normal Japan) or purchased from Kamiya Biomedical (Thousand human peripheral blood lymphocytes (panels a–d) or their Oaks, CA, USA). UIC2 monoclonal antibody was kindly pro- cultures (after 72 h growth in the presence of mitogen, phy- vided by Dr Eugene Mechetner of Ingenex, Menlo Park, CA, tohemagglutinin (PHA), panels e–h) were incubated with dif- USA. ferent fluorochromes to monitor drug retention, efflux and het- Propidium iodide (PI), Rhodamine 123 (Rh-123), verapamil erogeneity. In cells incubated with DNR (Figure 1a), a single (Vpl), dipyridamole (persantin), bovine albumin, vinblastine predominant population can be recognized. In cells incubated sulfate, and anti-mouse IgG2a were purchased from Sigma (St with Rh-123, a second distinct population (arrow) can be seen Louis, MO, USA). Goat anti-mouse FITC conjugated IgG2A (Figure 1b). In cells incubated with SY-38, a pronounced het- was obtained from Boeringer Mannheim Biochemicals erogeneity in dye retention was evident (Figure 1c). In con- (Indianapolis, IN, USA). Doxorubicin hydrochloride (Dox, trast, cells incubated with SY-3150 (Figure 1d) did not show Adriamycin, NSC-123127) was purchased from Adria Labora- the second set of cells similar to those identified by arrows in tories (Columbus, OH, USA). Daunorubicin hydrochloride Figure 1b and c. (DNR, Cerubidine, NSC-821151) was obtained from Wyeth In 72 h cultures of mitogen-stimulated lymphocytes Laboratories (Philadelphia, PA, USA). Prochlorperazine edys- (Figure 1e–h) incubated with the different fluorochromes, no ilate (Compazine, Cmz) and trifluoperazine (Stelazine, Tfp) major subpopulations were seen and the fluorescence profile were purchased from Smith Kline Beecham (Philadelphia, PA, was more of a continuum, especially in cells incubated with USA). SY-38 and SY-3150 were kindly provided by Dr Tom SY-38 and SY-3150 (Figure 1g and h). Observations from this Frey of Becton Dickinson Immunocytometry System, San Jose, experiment indicate the need for choosing the proper fluor- CA, USA. ochrome and identifying the subpopulations which differ in dye retention either due to efflux or other characteristics. Drug retention experiments Drug retention and efflux in normal human peripheral Cells were harvested and resuspended at an approximate con- blood lymphocytes centration of 106 per ml of medium. Fluorochromes and efflux blockers from stock solutions were added and the incubation Drug efflux and sensitivity to efflux blockers (which inhibit continued for 30 to 60 min in a humidified atmosphere of efflux and enhance cellular drug retention and 95% air and 5% CO2. chemosensitivity) is predominantly seen in drug resistant tumor cells. P-glycoprotein expression has been reported in a variety of normal cells such as capillary endothelial cells of P-gp expression brain and testes, adrenals, liver and CD34-positive bone mar- row stem cells.20,21 Although normal human peripheral blood For staining of cells with the C219 antibody, the methods earl- lymphocytes were earlier reported to have no P-gp or mdr1 ier described were used.10 The method described by Arceci et mRNA expression, subsequent studies have shown that a sig- al12 was used for determination of cell surface P-glycoprotein nificant percentage of peripheral blood lymphocytes have P- expression with 4E3 monoclonal antibody. For staining with gp expression and efflux which can be blocked by P-gp anti- MRK16 antibody, cells were washed with Ca++- and Mg++-free bodies. Dot plots in Figure 2 are of human peripheral blood phosphate-buffered saline (PBS), resuspended in 10% goat lymphocytes (unstimulated and PHA-stimulated) incubated serum and incubated on ice for 15 min. After incubation cells with the fluorochrome SY-38 and eflux blockers verapamil were pelleted and resuspended in 50 ␮l of FITC labeled goat (Vpl) or dipyridamole (Dpd) to show heterogeneity of drug anti-mouse IgG for 30 min on ice in the dark and 10% goat retention and sensitivity to efflux blockers. serum in PBS was added to a total volume of 0.5 ml. A modi- Figure 2a and f show forward angle scatter (FSC) vs 90° side fication of the method published by Chaudhary and Ronin- scatter (SSC) dot plots of mononuclear cells isolated on the MDR phenotype quantitation A Krishan et al 1140

Figure 1 Dot plots of human peripheral blood lymphocytes (a–d) and their cultures after 72 h of incubation with mitogen, phytohemagglutinin (e–h). Dot plots a–h are of forward side scatter (FSC) vs drug ﬂuorescence.

gradient. The panels on the left side (a–e) are of cells after in panel 2b (arrows) increased their drug retention in the isolation on a Ficoll–Hypaque gradient while those on the presence of Vpl (panel 2c). right (f–j) are of cultures incubated with PHA for 72 h to stimu- Figure 2d shows SY-38 retention of cells identiﬁed in gated late DNA synthesis and proliferation. DNA distribution histo- R1 region of dot plot in Figure 2a. In 2e, we have the same

grams of unstimulated lymphoblasts had a predominant popu- population (R1 gated) from cultures incubated with SY-38 and lation of cells with G0/G1 DNA content while the stimulated Vpl. A comparison of these two panels shows that most of the cultures had approximately 39% of the population with S- cells in the R1 gate which had low SY-38 retention were sensi- phase DNA content (data not shown). Gate R1 in panel 2a tive to the efflux blocking action of Vpl or other efflux blockers identifies small lymphocytes and R2 gate in panel 2f has a (eg Dpd, data not shown). In contrast to unstimulated lympho- large number of PHA stimulated lymphoblasts with increased cytes, PHA-stimulated lymphoblasts shown in Figure 2 g–h forward angle light scatter. did not significantly increase SY-38 retention after incubation In cells incubated with fluorochromes, daunomycin or Rh- with Vpl or Dpd. A comparison of dot plots in Figure 1 i and 123, minimal heterogeneity in drug retention was seen and j, with and without the efflux blocker, dipyridamole, shows efflux blockers did not increase cellular fluorescence (data not that the mean channel fluorescence value of cells identified

shown). In contrast, cells incubated with the fluorochromes as large blasts in region R2 (Figure 2f) marginally increased on SY-38 or SY-3150 had pronounced heterogeneity in fluoro- incubation with the efflux blocker, Vpl (data not shown) or chrome retention and response to efflux blockers. Dpd (Figure 2j). Figure 2b shows FSC vs fluorescence dot plots of cells (total One can draw the following conclusions from this experi- non-gated population) incubated with SY-38 alone or as in ment: (1) The degree of heterogeneity and response to efflux panel 2c in the presence of efflux blocker verapamil (Vpl). A blockers depends on the fluorochrome used for monitoring comparison of these two dot plots suggests that cells recorded drug retention and efflux. Daunomycin and Rh-123 are not as MDR phenotype quantitation A Krishan et al 1141

Figure 2 Dot plots are of unstimulated (a–e) and mitogen stimulated (f–j) peripheral blood lymphocytes incubated with fluorochrome SY-38. good as SY-38 for identifying the drug effluxing peripheral than that of the subpopulation in unstimulated lymphocytes. blood lymphocytes. (2) In normal human peripheral blood These observations are important in the study of leukemic lymphocytes, a significant (50%) population with reduced SY- populations where one may have subpopulations of normal 38 or SY-3150 retention can increase dye retention by incu- lymphocytes with pronounced drug efflux intermixed with bation with efflux blockers. (3) In stimulated lymphoblasts, leukemic cells. It is evident that the multiparametric flow cyto- drug efflux and sensitivity to efflux blockers is less pronounced metry utilizing other phenotypic markers may be essential to MDR phenotype quantitation A Krishan et al 1142 distinguish between the tumor and the normal cells. In Figure 3a and e dot plots of murine leukemic P388 cells The reader is referred to earlier publications identifying the and the drug-resistant P388/R-84 incubated with daunomycin phenotypic markers of the drug effluxing lymphocytes in the show pronounced heterogeneity in drug retention. The peripheral blood.22,23 P388/R84 population had at least five distinct subpopulations based on the FSC vs SSC dot plots of Figure 3e. Figure 3b shows that single parameter histograms which are commonly Heterogeneity of drug retention and contribution of used to show drug retention and modulation, do not accu- dead or membrane permeable cells rately reflect the presence of this heterogeneity. To discrimi- nate between the cells with low retention (presumed to be Earlier studies have described heterogeneity in retention of resistant), cells with high retention and cells with damaged fluorochromes and the differential response of subpopulations membranes (which can not efflux the drug), we added isotonic to the efflux blocking activity.6 Experimental data in Figure 3 propidium iodide (25 ␮g/ml in Hanks’ balanced salt solution) shows that contamination of the samples with membrane to the cultures to mark the membrane permeable cells. Dot permeable cells can be a major source of artifacts in drug plots of Figure 3c and the histogram in Figure 3d, show that retention functional assays. this procedure can enhance distinction between the mem-

Figure 3 Dot plots and histograms are of P388 and P388/R84 cells showing FSC vs daunomycin retention in P388 (a,b) and P388/R84 (e,f) cells. In cells incubated with isotonic propidium iodide (c,d and g,h), one can clearly identify dye permeable cells. MDR phenotype quantitation A Krishan et al 1143 brane permeable and normal cells as the former increase their SY-38 Ͼ SY-3150 Ͼ Rh-123 2.7 ␮m Ͼ DNR. Thus, not only drug (DNR and PI) retention and are recorded with increased was the choice of the indicator dye important but also the dye cellular fluorescence. concentration was critical for obtaining maximum discrimi- In P388/R84 cells (Figure 3e–h) heterogeneity is more pro- nation between retention in the sensitive and the resistant nounced as most of the drug-resistant live cells efflux DNR cell lines. while the membrane permeable cells (in Figure 3g–h) stain In the CEM: CEM/VLB100 paired cell line, the order was SY- with propidium iodide and stand out as distinct populations 3150 Ͼ SY-38 Ͼ Rh-123 Ͼ DNR. In SW620:Ad300 paired with high cellular fluorescence. This experiment shows that cell line, SY-3150 was the best indicator dye followed by SY- one can easily distinguish between the real heterogeneity of 38. Rh-123 and DNR showed less than . log difference in drug retention in a tumor population and the artifacts created the mean fluorescence channel values of the sensitive and the by the presence of cells with dye permeable (leaky) mem- resistant cells. branes by the procedures described.

Do fluorochromes differ in their potential for Need for titration of different efflux blockers differentiating between the sensitive and the resistant sub-populations? In general drug resistant cell lines are resistant to the cytotoxic activity of higher concentrations of efflux blockers. However, We often encounter tumor cells which differ in their capacity in several tumor specimens, we have noted a significant to transport and retain different fluorochromes. These differ- increase in the number of dead cells upon incubation with an ences may be determined by differences in influx, modes of efflux blocker and a fluorochrome. To avoid this artifact, one transport (passive vs carrier mediated), energy requirements must titrate concentration of the efflux blocker against a parti- for influx, efflux, cell cycle stage, proliferation24 and the selec- cular cell line or tumor specimen to identify concentrations tivity of the efflux mechanism. which will block efflux and enhance retention without damag- Data in Figure 4 was generated from three different pairs of ing the cell membrane (which may result in enhanced drug parental and drug-resistant cell lines incubated with DNR, Rh- retention and artifacts). 123 (0.27 and 2.7 ␮m), SY-38 and SY-3150. Cultures were incubated under identical conditions and analyzed to gener- ate list mode data on FSC vs cellular drug fluorescence. Mean channel fluorescence values were determined for each combi- Best combination of a fluorochrome and efflux nation and plotted as a ratio of parental vs the resistant cell blockers line values. Data in this chart shows that in the P388: P388/R-84 paired cell line, the lower Rh-123 concentration (0.27 ␮m) gave the We have earlier reported that different subpopulations in a maximum discrimination between drug retention (determined tumor may not have similar responses to an efflux blocker. on the basis of mean channel fluorescence) of the two cell Thus there can be subpopulations which will respond to one lines. The next best indicator dyes were in the following order: particular efflux blocker while other subpopulations may not be affected (for example see Figure 5 in Ref 6). Similarly, we have often noted that a heterogenous tumor cell population may reveal different subpopulations based on the fluorochrome and the efflux blocker used to detect the fluorescent cells (eg Figure 6 in Ref. 6). This phenomenon was further illustrated in Figure 1 of human peripheral blood lymphocytes where depending on the particular fluorochrome used, different subpopulations can be recognized. Data obtained from a combination of a fluorochrome and efflux blocker can become further complex as one deals with heterogenous cell populations often seen in human tumors.25 Figure 5 which shows single parameter histograms of three

different drug resistant cell lines (P388/R84, CEM/VLB100 and SW620/Ad300) with different fluorochromes (DNR, Rh-123, SY-38, SY-3150) and efflux blockers (Vpl, Cmz, Dpd) illus- trates this quandary. In P388/R84 cells, incubated with DNR or Rh-123, both Cmz and Dpd blocked efflux and enhanced retention (Figure 5a,b). In cells incubated with SY-38 (Figure 5c), or SY- 3150 (Figure 6d) Dpd was not as good as Cmz in blocking efflux.

In CEM/VLB100 cells, a lot more variability in the effect of different efflux blockers on the retention of different fluorochromes was seen (Figure 5e–h). In this cell line, Dpd was uni- formly better in blocking efflux and increasing retention of the Figure 4 Chart showing ratio of mean channel fluorescence values four different fluorochromes. In SW620/Ad300 cells, a similar of parental and drug resistant cells incubated with the different fluor- heterogeneity in response to efflux blockers was seen ochromes. (Figure 5i–l). MDR phenotype quantitation A Krishan et al 1144

Figure 5 Drug retention profiles of the three drug resistant cell lines P388/R84, CEM/VLB100 and SW620/Ad300 incubated with different fluorochromes and efflux blockers.

monoclonals MRK16, UIC2 and 4E3 have been reported in the literature. The following experiments were undertaken to determine reactivity of these antibodies against a panel of drug resistant cell lines and to illustrate some of the problems in data quantitation.

In Figure 6 histograms of CEM/VLB100 (a), SW620/Ad300 (b), FCCR-2 (c) and HCT-15 (d) cells reacted with the monoclonal antibody 4E3 (solid color) and isotype controls show that this antibody identiﬁed 44 to 97% of the population as P-gp positive in the three cell lines (a,b,d). The FCCR-2 cell line (Figure 6c) had a low percentage (1.7%) of positive cells. The UIC2 monoclonal (Figure 7) gave excellent discrimi-

nation of the P-gp positive cells in CEM/VLB100 (Figure 7b) with minimal reactivity to the parental cell line (Figure 7a). In FCCM-2 melanoma cell line (Figure 7c), UIC2 gave better results (9% positive) than the 4E3 or MRK-16 antibodies which identified 1.3 and 4.2% of the cells respectively as positive (data not shown). In FCCR-2 cells (Figure 7d), UIC2 identified 38% of the population as positive compared to the low reactivity of 4E3 and MRK-16 which identified 21% of the cells Figure 6 Drug-resistant cell lines reacted with the different anti-P- as positive. gp antibodies. CEM/VLB100 (a), SW620/Ad300 (b), FCCR-2 (c) and MRK16 had very good reactivity with all the four indicator HCT-15 (d) cell lines after incubation with an isotype control (open) or the 4E3 monoclonal antibody (solid) are shown. cell lines. Histograms in Figures 6 to 8 and data in Table 1 should serve to illustrate some of the problems in reporting and comparison of statistical data from isotype and antibody

Flow cytometric quantitation of P-glycoprotein treated samples. In histograms such as those of CEM/VLB100 expression cells (Figure 6a, 7b and 8a) which have excellent separation of the isotype and positive histograms, it is relatively easy to Several earlier studies have used the C219 antibody for evalu- determine percentage of positive cells (after electronic gating ation of P-gp expression by laser flow cytometry with variable of 95% cells from the isotype controls). In contrast, histograms results. Three new cell surface P-gp epitope-recognizing such as those of HCT-15 cells in Figure 8c and of MDR phenotype quantitation A Krishan et al 1145 SW620/Ad300 in Figure 8b pose a problem due to the presence of ‘long trailing ends’ of the histograms. One of the major concerns in quantitating P-gp expression in tumor cells analyzed by laser flow cytometry is lack of a common agreed upon method which could allow for data comparison. We have used and advocated the following anal- tyical procedure for reporting of data for comparative purposes. After data in list mode is collected from both the isotype controls and the antibody reacted samples, electronic gates are set on the isotype control to exclude 95% of cells recorded in the FITC or fluorescence channels. These gates are then copied on the antibody-stained samples and the number of cells failling outside the gate are recorded as being positive. Simultaneously, the mean fluorescence channel value of the isotype and the antibody reacted specimen is determined to arrive at a ratio of the two. This procedure has the advan- tage of not only determining as to what percentage of cells have positive expression, but also indicating the intensity of antigen reactivity. As most of the data analysis systems on the currently available flow cytometers have the technical capa-

Figure 7 CEM (parental, a), CEM/VLB100 (b), FCCM-2 (c) and bilities of setting electronic gates and rapidly determining the FCCR-2 (d) cell lines incubated with isotype control or UIC2 anti-P- statistics, these procedures should provide analytical data of gp antibody. value for comparative studies. This may be especially important when one wants to compare antigen expression in sequential clinical samples of a patient over a period of time.

Acknowledgements

These studies were supported by NIH grant CA 57488.

References

1 Gottesman MM. How cancer cells evade chemotherapy: sixteenth Richard and Hinda Rosenthal Foundation Award Lecture (review). Cancer Res 1993; 53: 747–754. 2 Ling V. P-glycoprotein and resistance to anticancer drugs. Cancer 1992; 69: 2603–2609. 3 Cole SPC, Sparks KE, Fraser K, Loe DW, Grant CE, Wilson GM, Deeley RG. Pharmacological characterization of multidrug resistant MRP-transfected human tumor cells. Cancer Res 1994; 54: 5902–5910. 4 Scheffer GL, Wijngaard PLJ, Flens MJ, Izquierdo MA, Slovak ML, Pinedo HM, Meijer CJLM, Clevers HC, Scheper RJ. The drug resistance-related protein LRP is the human major vault protein. Nature Med 1995; 1: 578–582. 5 Krishan A, Ganapathi R. Laser flow cytometric studies on intracellular fluorescence of anthracyclines. Cancer Res 1980; 40: Figure 8 Reactivity of cell lines CEM/VLB100 (a), SW620/Ad300 (b), HT-15 (c) and FCCR-2 (d) with MRK16 anti-P-gp antibody. 3895–3900. 6 Krishan A, Sridhar KS, Davila E, Vogel C, Sternheim W. Patterns of anthracycline retention modulation in human tumor cells. Cytometry 1987; 8: 306–314. 7 Durand RE. Flow cytometry studies of intracellular adriamycin in Table 1 Comparison of the three anti-P-glycoprotein antibodies multicell spheroids in vitro. Cancer Res 1981; 41. 3495–3498. 8 Herweijer H, Van Den Engh G, Nooter K. A rapid and sensitive CEM CEM SW620/ HCT-15 FCCM-2 FCCR-2 flow cytometric method for the detection of multidrug-resistant VLB-100 Ad300 cells. Cytometry 1989; 10: 463–468. 9 Noonan KE, Beck C, Holzmayer TA, Chin JE, Wunder JS, Andrulis

a b IL, Gazdar AF, Willman CL, Griffith B, Von Hoff DD. Quantitative 4E3 97 /17.2 54/37.1 44/2.25 1.3/0.9 1.7/1.4 analysis of MDR1 (multidrug resistance) gene expression in human UIC2 2.1/1.1 98/313 — — 9/1.2 38/2.6 tumors by polymerase chain reaction. Proc Natl Acad Sci USA MRK-16 — 98/403 54/25 20/60 4.2/0.9 21/4.7 1990; 87: 7160–7164. 10 Krishan A, Sauerteig A, Stein JH. Comparison of three commer- aIndicates percent of population with positive reactivity. This data cially available antibodies for flow cytometric monitoring of P- was generated by electronic exclusion (gating) of 95% of positive glycoprotein expression in tumor cells. Cytometry 1991; 12: cells from the isotype control. 731–742. bRatio of mean fluorescence channel values of the isotype and the 11 Hamada H, Tsuruo T. Purification of the 170- to 180-kilodalton antibody stained sample were derived from dot plots of FSC vs membrane glycoprotein associated with multidrug resistance. J fluorescence shown in Figures 6, 7 and 8. Biol Chem 1988; 263: 1454–1458. MDR phenotype quantitation A Krishan et al 1146 12 Arceci RJ, Stieglitz K, Bras J, Schinkel A, Baas F, Croop J. Mono- in flow-cytometric dye-efflux assays for multidrug resistance. Cyto- clonal antibody to an external epitope of the human Mdr1 P-gly- metry 1995; 20: 218–227. coprotein. Cancer Res 1993; 53: 310–317. 20 Cordon-Cardo C, O’Brien JP, Casals D, Rittman-Grauer L, Biedler 13 Mechetner EB, Roninson IB. Efficient inhibition of P-glycoprotein- JLM, Bertino JR. Multidrug-resistance gene (P-glycoprotein) is mediated multidrug resistance with a monoclonal antibody. Proc expressed by endothelial cells at blood–brain barrier sites. Proc Natl Acad Sci USA 1992; 89: 5824–5828. Natl Acad Sci USA 1988; 86: 695–698. 14 Beck WT, Mueller TJ, Tanzer LR. Altered membrane glycoproteins 21 Schlaifer D, Laurent G, Chittal S, Tsuruo T, Soues S, Muller C, in vinca-alkaloid resistant human leukemic lymphoblast Charcosset JY, Alard C, Brousset P, Mazerrolles C, Delsol G. (CCRFCEM). Cancer Res 1979; 39: 2070–2076. Immunohistochemical detection of multidrug resistance associa- 15 Leibovitz A, Stinson JC, McCombs WB III, McCoy CE, Mazur KC, ted P-glycoprotein in tumour and stromal cells of human cancers. Mabry ND. Classification of human colorectal adenocarcinoma Br J Cancer 1990; 62: 177–182. cell lines. Cancer Res 1976, 36: 4562–4569. 22 Klimecki WT, Futscher BW, Grogan TM, Dalton WS. P-glyco- 16 Tibbets LM, Chu MY, Hager JC, Dexter DL, Calabresi P. Chemo- protein expression and function in circulating blood cells from therapy of cell-line derived human colon carcinomas in mice normal volunteers. Blood 1994; 83: 2451–2458. immunosuppressed with antithymocyte serum. Cancer 1977; 40: 23 Chaudhary PM, Mechetner EB, Roninson IB. Expression and 2651–2659. activity of the multidrug resistance P-glycoprotein in human per- 17 Chaudhary PM, Roninson IB. Expression and activity of P-glyco- ipheral blood lymphocytes. Blood 1992; 80: 2735–2739. protein, a multidrug efflux pump, in human hematopoietic stem 24 Ramachandran C, Mead D, Wellham L, Sauerteig A, Krishan A. cells. Cell 1991; 66: 85–94. Expression of drug resistance associated mdr-1, GST␲, and topo- 18 Lampidis T, Munck J, Krishan A, Tapiero H. Reversal of resistance isomerase II genes during cell cycle. Biochem Pharm 1995; 49: to rhodamine 123 in adriamycin-resistant Friend leukemia cells. 545–552. Cancer Res 1985; 45: 2626–2631. 25 Krishan A. Heterogeneity of anthracycline retention and response 19 Frey T, Yue S, Haugland RP. Dyes providing increased sensitivity to efflux blockers in human tumors. Cytometry 1995; 21: 72–75.