Dr ChristofLudescher Department of Haematology and Oncology, University of Innsbruck, Anichstrasse 35, A-6020 Innsbruck, Austria.
To better understand the phenomenon of P-glycoprotein (P-170) expression we investigated lymphocyte subpopulations for P-170 function in healthy volunteers. Studies were based on three-colour flow cytometry including the fluorescent probe rhodamine 123 (Rh123), which is transported by P-170. Marked Rh123 efflux was detected in CD8+ T lymphocytes with CD8+/CD45RA+ T cells (naive cells) showing significantly higher P-170 activity as compared with CD8+/CD45RA− cells (P < 0.04). Vice versa, CD8+/CD45RO+ T cells (memory cells) demonstrated less P-170 activity than CD8+/CD45RO− cells (P < 0.04). P-170 function was less prominent in CD4+ T cells, however, Rh123 efflux was higher in the CD4+/CD45RA+ and CD4+/CD45RO− subpopulations (P < 0.025) corresponding to the CD8+ results. Dye efflux differed significantly between activated and non-activated CD8+ and CD4+ as well as CD8+/CD11b+ and CD8+/CD11b− T lymphocytes. Since CD16+ natural killer cells (NK) expressed the highest level of P-170, the NK cytotoxicity against 51Cr-labelled K562 target cells was assayed in the presence or absence of P-170 inhibitors. NK related cytotoxicity was significantly reduced in the presence of R-verapamil and dexnigaldipine-HCP in a dose-dependent manner. The differential expression of P-170 activity in naive and memory T cells together with the reduced NK related cytotoxicity in the presence of MDR-modulators suggest a physiological role of P-170 in immunological functions of these lymphocyte subsets. Consequently, the addition of MDR modulators to conventional chemotherapy as a strategy to overcome drug resistance should consider possible adverse immunosuppressive effects.
Classic multidrug resistance (MDR) is characterized by the expression of the mdrl-gene, which encodes for the transmembrane protein P-glycoprotein (P-170) ( Pastan & Gottesman, 1987). P-170 functions as an efflux pump to decrease the intracellular accumulation of a variety of lipophilic xenobiotics including many anticancer drugs ( Weinstein et al, 1990 ). The clinical importance of P-170 overexpression as a mechanism of drug resistance in malignant cells has been well characterized ( Campos et al, 1992 ; Goasguen et al, 1993 ; Nooter & Sonneveld, 1994; Yuen & Sikic, 1994; Ludescher et al, 1995a ). P-170 has also been identified in several normal tissues such as kidney, liver biliary ducts, adrenal gland, gastrointestinal tract and endothelial cells of the blood–brain barrier ( Gottesman et al, 1991 ). Current data suggest that P-170 may be involved in the process of hormone release with aldosterone and cortisol as physiological substrates ( Ueda et al, 1992 ). There is increasing evidence that normal haemopoietic cells of the blood and bone marrow, particularly natural killer (NK) cells and cytotoxic CD8+ T lymphocytes, show significant levels of P-170 activity ( Chaudhary & Robinson, 1991; Chaudhary et al, 1992 ; Kobayashi et al, 1994 ). The predominant expression of P-170 in blood cells with cytolytic activity suggests a specific immunological function, e.g. concerning CD8+ T-cell or NK-cell mediated cytotoxicity ( Gupta et al, 1992 ). However, few data are available regarding the role of P-170 in normal peripheral blood cells. Since NK cells play an important role in the immune surveillance against malignant cells, ongoing therapeutic strategies which try to reverse clinical drug resistance by adding P-170 inhibitors to conventional chemotherapy might enhance chemotherapy-induced immunosuppression.
In this study we investigated the expression of P-170 by means of a functional three-colour flow cytometric assay. Using the fluorescent dye rhodamine 123 (Rh123) as a measure of P-170 activity ( Ludescher et al, 1992 ; Feller et al, 1995 ) together with monoclonal antibodies (MoAb) we attempted to precisely define the MDR phenotype in different lymphocyte subsets from healthy volunteers. Moreover, NK cell cytotoxicity in the presence or absence of P-170 modulators was investigated to prove a possible inhibitory effect of these agents on NK cell-related cytotoxicity.
MATERIALS AND METHODS
Peripheral blood from healthy volunteers (n = 10) was collected in 10 ml tubes containing heparin as anticoagulant. Mononuclear cells (PBMCs) were isolated by density gradient centrifugation (Lymphoprep, Nycomed AS, Norway) and washed three times in phosphate-buffered saline (PBS). All samples were processed immediately after collection. Viability of cells exceeded 97% as determined by propidium iodide staining (10 μmol/l).
The following MoAbs were used: anti-CD4/Leu3, CD8/Leu2, CD11b/Leu15, CD14/LeuM3, CD15/LeuM1, CD16/Leu11c, CD19/Leu12, CD28/Leu28, CD38/Leu17, CD45RO/Leu45RO, HLA-DR (all from Becton Dickinson, Mountain View, Calif.) and CD45RA/2H4 (Coulter Immunology, Hialeah, Fla.) were conjugated with phycoerythrin (PE). CD4 and CD8 (Sigma Immuno Chemicals, St Louis, Mo.), were conjugated with Quantum RedTM (QR), and CD3/UCHT1 (DAKO, Denmark) was conjugated with RPE-Cy5. Rh123 efflux analysis of total CD4 and total CD8 T cells were always performed together with CD3 staining (RPE-Cy5-conjugated) to exclude non-T cells. All MoAbs were used at saturating concentrations as recommended by the manufacturer and isotype-matched control antibodies were processed in parallel.
Rhodamine 123 efflux assay and flow cytometric analysis
1 × 106 cells/ml were resuspended in Dulbecco's modified Eagle medium buffered with 25 m M 3- N-morpholinopropanesulphonic acid (DMEM-MOPS). For Rh123 efflux studies, aliquots of cell suspension (500 μl) were incubated simultaneously with 150 ng/ml Rh123 and with PE- or QR-labelled monoclonal antibodies (MoAb). Incubation time was 45 min at 4°C. After two washes in ice-cold medium cells were resuspended in Rh123-free medium (DMEM-MOPS) containing either 10 μmol/l R-verapamil (Sigma Chemicals, Munich, Germany) or no MDR inhibitor. The Rh123 efflux/retention in medium was selectively determined for the PE- or QR-labelled cells at several time intervals (up to 2 h). Between measurements the cells were kept at 37°C.
Stained cells were analysed on a FACScan flow cytometer (Becton Dickinson, Mountain View, Calif.) equipped with an air-cooled argon-ion laser emitting at 488 nm. A minimum of 10 000 events (up to 50 000) were acquired and analyses were carried out in duplicate. Data were processed with a CellQuest software (Becton Dickinson, Mountain View, Calif.) Rh123 efflux studies were performed by placing a gate around the PE-or QR-labelled cell subsets and analysing these cells for Rh123 efflux/retention. To determine the fraction of cells showing Rh123 efflux and thus P-170 activity, a marker was set to the left of the fluorescence peak of cells which were co-incubated with verapamil. Results were expressed as ‘percentage positive cells’. Interpretation of results by mean channel fluorescence intensities instead of ‘percentage positive cells’ was comparable.
The cytotoxicity assay was performed using PBMCs as effector cells. The NK-sensitive 51Cr-labelled K562 human erythroleukaemia cell line was used as the target cells. Cells were incubated either without a MDR modulator or with R-verapamil at concentrations from 1 to 20 μmol/l or dexniguldipine-HCl at concentrations from 1 to 10 μmol/l. At these concentrations viability of mononuclear cells exceeded 90% as determined by propidium iodide staining (10 μmol/l). After a 4 h incubation, the release of 51Cr into the supernatants was determined in triplicate at effector/target ratios of 3/1, 6/1, 12/1, 25/1 and 50/1.
The mean percentage lysis was calculated by the following formula: percentage of lysis = [(cpm release in test − cpm spontaneous release)/(cpm maximum release − cpm spontaneous release)] × 100, where spontaneous release represents the count per minute (cpm) of supernatants from the target cells incubated in medium alone and maximum release, the cpm of target cells lysed in 1 N NaOH. Additionally, reactivity against autologous mononuclear blood cells was tested and was negative in all experiments.
Statistical analysis was performed on a PC Apple Macintosh LC II equipped with the software StatView SE + Graphics. Student's t-test and Wilcoxon signed-rank test were used to test for significant differences.
P-170 activity in peripheral blood cells
Function of P-170 was determined by measuring verapamil-sensitive intracellular retention of the fluorescent dye Rh123. Three-colour flow cytometry using PE- and QR-conjugated MoAbs enabled the selective determination of Rh123 efflux or retention in various lymphocyte subsets. Naive and memory T cells revealed different patterns of P-170 function ( Figs 1 and 2). P-170 function in naive CD8+/CD45RA+ and CD4+/CD45RA+ T cells was significantly higher when compared with their CD45RA− counterparts (P < 0.04 and P < 0.05, respectively). Conversely, CD8+ and CD4+ memory cells co-expressing CD45RO showed low Rh123 efflux activity when compared with their CD45RO− counterparts (P < 0.04 and 0.03, respectively).
Analysis of CD3+/CD8+dim and CD3+/CD8+bright T-cell subsets for Rh123 efflux showed a slight tendency towards a lower percentage of Rh123 effluxing CD8+dim T cells compared with the CD8+bright T cells but this difference did not reach statistical significance. Activated CD4+ as well as CD8+ T cells from normal volunteers, characterized by the co-expression of the antigen HLA-DR, showed significantly reduced Rh123 efflux as compared with the non-activated CD4+/HLA-DR− (P < 0.008) and CD8+/HLA-DR− (P < 0.02) counterparts. Expression of the integrin molecule CD11b on CD8+ T cells defined a further subgroup of normal CD8+ lymphocytes with low P-170 function (P < 0.01 when compared with CD8+/CD11b− cells). Finally, the expression of the co-stimulatory receptor CD28 on CD8+ T cells was associated with increased P-170 activity when compared with CD8+/CD28− cells although the difference was not statistically significant (P < 0.06).
Analyses of the major blood cell subsets revealed the highest percentage of Rh123 effluxing cells among CD16+ NK cells (mean 88%, range 78–96), followed by CD8+ suppressor/cytotoxic lymphocytes (mean 83%, range 70–91), CD4+ helper/inducer lymphocytes (mean 29%, range 16–39) and CD19+ B cells (mean 20%, range 17–25). CD14+ monocytes and CD15+ granulocytes lacked P-170 activity. The results of all subset analyses are summarized in I. Rhodamine 1II. Rh12 Tables I and II.
Table I. Rhodamine 1. 23 (Rh123) efflux from peripheral blood cells. The percentages of cells showing dye efflux (mean and range) was determined after a 60 min incubation in Rh123-free medium.
Table II. Rh12. 3 efflux from further CD4+ helper/inducer and CD8+ suppressor/cytotoxic T-cell subsets. Mean and range of the percentages of cells showing dye efflux after 60 min incubation in Rh123-free medium are indicated.
NK-cell activity in the presence of MDR modulators
As the highest percentage of Rh123-effluxing cells was found among lymphocytes with cytolytic activity (NK cells, CD8+ T cells) we tested whether the MDR inhibitors R-verapamil and dexniguldipine-HCl could affect NK cell cytotoxicity. For this purpose we measured the NK cell mediated 51Cr release from target K562 cells in the absence or presence of these two MDR modulators. Since calcium channel blocking activity of drugs might influence cytotoxicity, we selected R-verapamil and dexniguldipine-HCl for our studies. These two agents lack calcium channel activity ( Hollt et al, 1992 ; Hofmann et al, 1991 ). As shown in Fig 3, we found a dose-dependent decrease of lytic activity in the presence of 1, 5 and 10 μmol/l R-verapamil, e.g. 1 μmol/l R-verapamil produced a 24% and 10 μmol/l a 55% reduction of NK-cell-related cytotoxicity at an effector/target ratio of 50:1. The same effect was also observed at lower effector/target ratios (data not shown). Dexniguldipine-HCl was an even more potent modulator of NK cell cytotoxicity: 1 μmol/l reduced the 51Cr release by 36% and 10 μmol/l by 66% (effector/target ratio 50:1, Fig 3B). Lower concentrations of MDR inhibitors were required to totally abrogate Rh123 efflux from NK cells measured by flow cytometry. 1 μmol/l R-verapamil and 0.5 μmol/l dexniguldipine-HCl completely inhibited Rh123 efflux.
Within mononuclear cells of the bone marrow and peripheral blood a quite heterogenous expression of the membrane transporter P-170 has been observed ( Chaudhary et al, 1992 ). Monocytes and granulocytes, for example, lack functional P-170, whereas cells with cytolytic activity such as NK cells or CD8+ T cells show the highest levels of expression ( Klimecki et al, 1994 ). In this study we investigated the activity of P-170 in several lymphocyte subsets by means of a three-colour flow cytometric assay using the fluorescent probe Rh123, emitting at 530 nm. PE-labelled MoAbs, emitting at 585 nm, and QR-labelled MoAbs, emitting at 670 nm. The fluorochrome Rh123 has been proved useful for the determination of the functional status of P-170 ( Ludescher et al, 1993 , 1995b). In a recent study Feller et al (1995 ) demonstrated that Rh123 is a specific and sensitive probe for the detection of low levels of P-170-mediated drug resistance. Our intention was to exactly characterize the differential expression of P-170 in individual blood cell subsets and thus to get additional information about the functional significance of P-170 in the normal lympho-haemopoietic system.
With this flow cytometric assay we detected clear differences concerning the Rh123 efflux pattern between T lymphocytes with the CD45RO and CD45RA phenotype. CD8+ as well as CD4+ T cells co-expressing CD45RA showed significant higher P-170 activity than cells co-expressing CD45RO. This observation is of special interest since CD45RA and CD45RO define naive and memory T cells, respectively ( Mackay, 1991; Trowbridge & Thomas, 1994). Although the profile of cytokine synthesis by human naive and memory T cells does not fall into a strict TH1 (IL-2 and IFN-γ producing) and TH2 (IL-4 and IL-5 producing) pattern, apparent differences of cytokine gene expression have been observed between naive and memory cells. Naive T cells synthesize greater amounts of IL-2 than memory T cells, whereas IL-4 synthesis is virtually restricted to the CD45RO memory T cells ( Akbar et al, 1991 ; Picker et al, 1995 ). The observed differential Rh123 efflux characteristics among CD45+ subsets therefore suggest a possible role of P-170 for the transmembrane transport of cytokines in naive but not memory T cells. Indeed, in a recent study Drach et al (1996 ) were able to show inhibition of IL-2 and IFN-γ release from stimulated normal lymphocytes in the presence of verapamil, tamoxifen and the P-170-specific MoAb UIC2. Our functional Rh123 data are in keeping with these results. Additionally, we detected a reduced Rh123 efflux activity among activated T lymphocytes (characterized by the expression of the HLA-DR antigen) in healthy volunteers. One possible explanation for this observation might be the unidirectional transition from the CD45RA to the CD45RO immunophenotype after activation, which has been considered to parallel T-cell differentiation from naive to memory cells ( Johannisson & Festin, 1995). This phenotype switching, which is accompanied by an unpaired cytokine production, seems to be associated with down-regulation of P-170 activity.
The highest expression of P-170 in circulating peripheral blood cells was found among NK cells. The homology between P-170 and hlyB, which functions as a cytolysin transporter in E. coli ( Gros et al, 1986 ), led to the assumption that P-170 is involved in the secretion or transport of perforin from NK cells. In accordance with a recent study ( Chong et al, 1993 ) we observed a significant inhibitory effect of MDR modulators on NK-cell-mediated cell killing in a 51Cr release assay. Since the process of degranulation is calcium dependent, R-verapamil and dexniguldipine-HCl were chosen for our experiments because these agents show essentially no Ca++ channel blocking activity ( Hollt et al, 1992 ; Hofmann et al, 1991 ). 10 μmol R-verapamil and dexniguldipine-HCl resulted in a > 50% reduction of 51Cr release. Notably, we also found a significantly higher P-170 activity among the CD8+/CD11b− compared with the CD8+/CD11b+ T-cell subset. Since the CD8+/CD11b− T-cell subset is considered to have mainly cytotoxic function ( Fox et al, 1987 ), these data, together with the NK assay results, indicate that an effective target cell killing appears to be P-170 dependent. This corresponds to a previously published study in which an effector function of P-170 in human CD8+ T cells has been suggested ( Gupta et al, 1992 ). Perhaps this could be of clinical importance, since cytotoxic cells, especially NK cells, play an essential role in the immune surveillance against tumour cells ( Robertson & Ritz, 1990). Several data indicate that the prognosis of cancer patients correlates with the status of their immune system ( Tajima et al, 1996 ; Rubbert et al, 1991 ). Inhibition of P-170 by adding MDR-modulating agents to conventional chemotherapy, a strategy that is being increasingly initiated to overcome clinical resistance in tumour patients, might have a negative impact on the immune system and consequently on the patients' outcome. Further studies are necessary to more precisely define the interaction between P-170 and NK or T-cell-related cytotoxicity and to examine the in vivo effect of MDR reversing agents on the immune system in ongoing clinical trials.