@

Cytometry 13:693-702 (1992)

1992 Wiley-I,iss, Inc.

Signal Transduction in Monocytes and Granulocytes Measured by Multiparameter Flow Cytometry Fridtjof Lund-Johansen and Johanna Olweus Department of Pathology, The Gade Institute, University of Bergen, Haukeland Hospital, N-5021 Bergen, Norway Received for publication January 8, 1992; accepted March 15, 1992

The novel calcium indicator fura red and the oxidative burst indicator dihydrorhodamine (both excited at 488 nm) were used in combination with multiparameter flow cytometry to allow simultaneous kinetic measurements of calcium fluxes and oxidative bursts in monocytes and granulocytes. Using this method it was possible to obtain direct evidence for the following cell type- and stimulus-specific differences in signal transduction pathways: 1) n-formyl-methionyl-leucylphenylalanine (FMLP)/cytochalasinB-induced oxidative burst is several-fold higher in granulocytes than in monocytes although the calcium fluxes have similar amplitudes in the two cell types; 2) stimulus-induced calcium fluxes in granulocytes are mainly due to release from intracellular stores, whereas monocytes mobilize calcium mainly by influx from the medium; 3)the FMLPkytochalasin Binduced calcium flux in monocytes is less sensitive to the G-protein inhibitor pertussis toxin than the flux in granulocytes; 4) in contrast to FMLPkytochalasin B,

Following stimulation, human monocytes and granulocytes produce reactive oxygen intermediates by a process termed the oxidative burst (1). This cell function is important for the killing of ingested microorganisms and for cell-mediated cytotoxicity, and defective oxidative burst is associated with a n increased rate of infections and early death (1).Although several intracellular reactions that are associated with receptor-mediated activation of phagocytes have been described, the post-receptor mechanisms leading to activation of the oxidative burst are yet poorly understood (8).

A commonly used approach for investigating phagocyte signal transduction is to examine the effects of various stimuli and inhibitors on oxidative burst or calcium fluxes in purified cell populations (4,lO16,19,22,23). Few inhibitors are, however, specific for

the protein kinase C activator phorbol myristate acetate (PMA) induces an oxidative burst that is not preceded by a cytoplasmic calcium flux; 5) the PMAinduced oxidative burst can be triggered in monocytes and granulocytes that are depleted of intracellular calcium ions, whereas that induced by FMLPkytochalasin B can not; 6) the G-protein inhibitor pertussis toxin blocks an early event in the signal transduction pathway of FMLP/cytochalasin B, as shown by inhibition of both calcium fluxes and oxidative burst; and 7) 100 nM of the protein kinase inhibitor staurosporine blocks the FMLPkytochalasin B-induced respiratory burst by interfering with a step downstream to cytoplasmic calcium fluxes, whereas only 10-20 nM is necessary to block PMA-induced oxidative burst. o 1992 Wiley-Liss, Inc. Key terms: Fura red, dihydrorhodamine, calcium, oxidative burst, pertussis toxin, staurosporine

one process only (18), and when only downstream responses such a s the oxidative burst are measured, this complicates the interpretation whether a potential inhibition is specific, or due to a general cytotoxic effect.

Address reprint requests to Fridtjof Lund-Johansen, Department of Pathology, The Gade Institute, University of Bergen, Haukeland Hospital, N-5021 Bergen, Norway. Abbreviations used in this paper: PBS-Ca. Dulbeccos phosphate buffered saline with calcium and magnesium; PBS, Dulbeccos phosphate buffered saline without calcium and magnesium; PCS, Fetal calf serum; FMLP, n-formyl-methionyl-leucyl-phenylalanine; EGTA, acid; Ethylene glycol-O,O'-bis~2-aminoethyl)-N,N,N',N-tetraacetic AM-ester, acetoxymethylester; DHR 123, dihydrorhodamine 123; PMA, phorbol myristate acetate; CytB, cytochalasin B; PBS-FCS, PBS with 2 8 FCS and 5 mM glucose; PBS-FCS-Ca, PBS with calcium, magnesium, 2% FCS and 5 mM glucose; R 123, rhodamine 123.

694

LUNDJOHANSEN AND OLWEUS

Furthermore, when monocytes and granulocytes are studied separately, possible celltype-specific differences in signal transduction may be difficult to recognize and interpret. Flow cytometry may be a valuable tool for the investigation of signal transduction, allowing studies of subsets of cells within the same sample (12,15) and simultaneous measurement of two different cellular responses (3,9,20). The present paper reports a multiparameter assay for the study of phagocyte signal transduction. Two new indicators that are compatible with a 488 nm excitation source were applied, the red fluorescent calcium indicator fura red (7) and the green fluorescent oxidative burst indicator dihydrorhodamine 123 (17). By combined measurement of light scatter, two-color fluorescence and time, kinetics of cytoplasmic calcium fluxes and oxidative burst activity were monitored in monocytes and granulocytes simultaneously. Inhibitors of phagocyte activation were evaluated for effects on early (calcium fluxes) and late (oxidative burst) steps of phagocyte signal transduction pathways.

FALS

90 LS

FIG.1. Recognition of granulocytes and monocytes by measurements of cellular light scatter. Right histogram: Discrimination of leukocyte subsets by 90" light scatter (90 LS) (MON, monocytes; GRAN, granulocytes). Left histogram: forward angle light scatter (FALS) gated from the monocyte population.

trifuged again and finally washed once in PBS containing 2% FCS and 5 mM glucose (PBS-FCS).

Measurement of Fluctuations in Leukocyte Cytoplasmic Calcium Concentration Leukocyte cytoplasmic calcium changes were moniMATERIALS AND METHODS tored by using the calcium indicator fura red (7). This Reagents indicator is efficiently excited at 488 nm, emits fluoDulbecco's phosphate-buffered saline (PBS) with or rescence with a maximum at 660 nm, and undergoes a without calcium and magnesium (PBS-Ca or PBS, redecrease in fluorescence intensity upon binding of calspectively) and fetal calf serum (FCS) were from Gibco cium ions (7). Fura red was loaded into the cells as its (Paisley, Scotland, UK). Pertussis toxin, and n-formylacetoxymethyl ester and trapped in the cytoplasm folmethionyl-leucyl-phenylalanine(FMLP) were from lowing deesterification by intracellular esterases. LeuSigma Chemicals (St. Louis, MO, USA). Ethylene glykocytes (107/ml) were incubated in PBS-FCS containcol-O,O'-bis(2-aminoethyl)-N,N,N',N-tetraacetic acid ing 1-10 p M fura red-AM under gentle agitation. (EGTA)-acetoxymethyl ester (AM), A-23187, dihyFollowing 25 min of incubation at 3TC, leukocytes drorhodamine 123 (DHR 123), and fura red-AM were were centrifuged and resuspended in PBS-Ca containfrom Molecular Probes (Eugene, OR, USA). EGTA, ing 5 mM glucose and 2% FCS (PBS-FCS-Ca) to a conphorbol myristate acetate (PMA), and cytochalasin B centration of 1-5 x lo6 celldml. The cells were then (CytB) were from Fluka (Buchs, Switzerland). Staurokept at 20°C under constant gentle agitation until flow sporine was from Boehringer Mannheim (Mannheim, cytometric measurement. This temperature was choGermany). Staurosporine (100 pM), CytB (10 mM), sen to minimize leakage of indicator from the cytoFMLP (20 mM), DHR 123 (30 mM), and fura red-AM (2 plasm and to reduce time-dependent priming of cells mM) were dissolved in dry DMSO (Merck, Darmstadt, observed following a long incubation time a t 37°C (not Germany) and stored in aliquots at -20°C (or -70°C shown), Immediately before flow cytometric analysis, for DHR 123). The solvent concentrations were less 200 pl of the leukocyte suspension was transferred to a than 0.5% in final dilutions. All salts used in labora- 0.7 ml polystyrene microcytometry tube (Robbins Scitory-made solutions were analytical grade from Merck. entific, CA, USA), and centrifuged for 3 s in a n Eppendorf centrifuge 5415C (Hamburg, Germany). The suLeukocytes pernatant was removed and the cells resuspended in Venous blood from normal volunteers was drawn 400 pl of PBS-FCS-Ca. The tube was then placed in a into heparin vacutainers (Becton Dickinson, San Jose, laboratory-made flow cytometry sample chamber (see CA, USA), and the red cells were immediately lysed by below) connected to a waterbath holding 37°C. Monocytes and granulocytes were discriminated by diluting the blood 1 : l O with a solution containing 0.8% NH,Cl, 0.08% NaHC03, and 0.08% EDTA, pH 6.8, flow cytometric measurements of cellular forward an15°C. This solution was prepared new for each experi- gle and 90" light scatter (Fig. 1). Basal cytoplasmic ment by diluting stock solutions (10 x concentration) calcium level was determined for each cell type by with distilled water, and a new stock was made every gated analysis of fura red fluorescence versus time week. Following lysis of red cells, leukocytes were cen- from the populations displayed on the light scatter histrifuged at l60g for 5 min, the pellet was resuspended tograms (Fig 1). After 40 s the analysis was interin 10 ml of the lysing solution, and the cells were cen- rupted, the stimuli added, and the measurement con-

FCM MEASUREMENT OF C E L L SIGNAL TRANSDUCTION

tinued. Fluctuations in cytoplasmic free calcium concentration were recognized as alterations in fura red fluorescence intensity over time. Before the next sample was analyzed, the sample insertion needle of the flow cytometer was thoroughly wiped off, and PBSFCS was flushed through the system for 3 s to remove remaining stimulants. As FMLPiCytB caused a decrease in granulocyte 90” light scatter, the monocyte region was set narrow and close to the lymphocyte population to avoid “contamination” of granulocytes in the monocyte region during activation (Fig. 1). In order to determine the source of the mobilized calcium ions, extracellular calcium ions were chelated by adding 200 p1 PBS-FCS with 4 mM EGTA to 200 p1 of the leukocyte suspension less than 1min before running the sample.

Measurement of Oxidative Burst Cellular oxidative burst was quantified by using the fluorescent indicator DHR 123 (17). DHR 123 is a nonfluorescent, cell membrane-permeable compound. During the oxidative burst, the intracellular DHR 123 is irreversibly converted to the fluorescent compound, rhodamine 123 (R 123) (6,17). R 123 is membrane impermeable, and so accumulates in the cells. The cellular oxidative burst is measured as a function of cellular R 123 fluorescence intensity (17). The specificity of this indicator for oxidative burst has been demonstrated in experiments with cells from patients with chronic granulomatous disease (6). To minimize spontaneous oxidation of DHR 123 to R 123 over time, the indicator solution was kept refrigerated during the experiment, and prewarmed to 37°C less than 60 min before use. Leukocyte suspensions (1-5 x lo6 cellsiml PBS-FCSCa) were kept at 20°C as described above. The cells were transferred to a microcytometry tube, centrifuged for 3 s with an Eppendorf 5415C centrifuge, resuspended in PBS-FCS-Ca containing 30 pM DHR-123, and placed in the flow cytometry sample chamber (37°C). Monocytes and granulocytes were discriminated by measurement of cellular light scatter as described above. Basal levels of cellular R 123 fluorescence were recorded for 40 s before stimuli were added and the measurement continued. Increases of twofold or more in cellular R 123 fluorescence relative to control were considered as significant stimulus-induced oxidative bursts. After running each sample, remaining stimulants were removed as described above. Combined Measurement of Calcium Fluxes and Oxidative Burst Oxidative burst and calcium fluxes were measured as described above except that fura red-loaded leukocytes were pelleted and resuspended in prewarmed PBS-FCS-Ca (or PBS-FCS with EGTA) containing 30 pM DHR 123 less than 1 min before running the sample. Monocyte and granulocyte fura red and R 123 fluorescence was gated versus time to four separate cytograms. Before two-color analysis was performed, a

695

sample was run with DHR 123 only, in order to allow electronic correction of fluorescence overlap. Overlap of fura red fluorescence into the R 123 channel was negligible.

Calcium Depletion of Leukocytes For depletion of intracellular calcium ions, fura red loading was performed in PBS-FCS with 2 mM EGTA and 10 pM EGTA-AM. Following loading, cells were resuspended in PBS-FCS with 2 mM EGTA. Inhibition of Protein Kinase C Activity With Staurosporine Fura red-loaded leukocytes were transferred to a microcytometry tube and centrifuged as described above. The cells were then resuspended in 400 pl of prewarmed PBS-FCS-Ca containing 25 nM staurosporine and 30 pM DHR 123 and kept at 37°C for 2 rnin prior to addition of the stimuli (5). Treatment of Leukocytes With Pertussis Toxin Leukocytes were suspended in PBS-FCS only, or PBS-FCS containing 2 pg/ml pertussis toxin, and incubated a t 37°C with constant gentle agitation. After 35 rnin or 2 h and 5 min, PBS-FCS containing 50 pM fura red-AM was added (20 pl/ml cell suspension), and the incubation continued for 25 min. The cells were then washed once, resuspended in PBS-FCS-Ca, and kept at 20°C until flow cytometric measurement. Flow Cytometry A Coulter Epics V flow cytometer (Coulter Electronics, Luton, UK) interfaced t o a CICERO IBM-PC-based computer system (CYTOMATIONInc., Colorado, USA) was used for the measurements. The excitation source was the 488 nm line of an argon laser (power set to 125 mW). The filter setup was a 590 nm dichroic short pass beamsplitter (Coulter Electronics), 530 ? 15 nm band pass filter for R 123 fluorescence (Omega Optics, Brattelboro, USA) and 630 nm long pass filter (Coulter Electronics) for fura red fluorescence. Deionized water was used as sheath fluid. To avoid inaccuracy in the signals commonly associated with logarithmic amplifiers, R 123 fluorescence was recorded in linear mode, and digitally converted into logarithmic units (3.5 decades) by the CICERO software.

Flow Cytometry Sample Chamber A flow cytometry sample chamber was made to enable measurement of cell activation at physiological temperature (i.e., 37°C).A polycarbonate rod (General Electric) was cut to fit within the sample vial holder of the Coulter Epics V and hollowed out like a flat-bottomed reagent tube (Fig. 2). Plastic tube connectors (Portex, Hithe, England) were fitted into holes that were made in the lower part of the chamber (Fig. 2), and these were connected to tubing carrying water

696

LUNDJOHANSEN AND OLWEUS

Medium with calcium

Sample inserlion needle

MON

GRAN

Ruhber cap

.*.

Ruhhei- ring ( l m m thick)

Tube connectors

2 4

IWI

6

8

2

4

6

8

Calcium-free medium MON

FIG.2. Schematic illustration of flow cytometry sample chamber with (right) and without (left) sample and Coulter Epics V flow cytometry insertion needle and rubber cap.

GRAN

back and forth to a water bath holding 37°C (not shown).

Data Analysis and Statistical Methods Data analysis was performed using CYCLOPS software from CYTOMATION Inc. Significance of differences was determined by paired Student's t-tests. RESULTS Measurement of Changes in Cytoplasmic Calcium Levels of Monocytes and Granulocytes With Fura Red Following the addition of the calcium ionophore A23187 to fura red-loaded leukocytes, there was a rapid and permanent decrease of monocyte and granulocyte red fluorescence (Fig. 3).When extracellular calcium ions were chelated by diluting the sample 1:l with PBS-FCS containing 4 pM EGTA, there was an increase in fura red fluorescence in both cell types (Fig. 3). This increase, indicating efflux of calcium ions from the cytoplasm, had different kinetics in the two cell populations. In monocytes, a rapid and subsequently a slow phase of calcium efflux was observed, whereas in granulocytes, there was a single slow phase (Fig. 3). Further differences in the calcium responses were seen when extracellular calcium ions were chelated before adding the ionophore. In calcium-free medium, the ionophore-induced increase in monocyte calcium concentration was immediately followed by a decrease t o near prestimulated levels (Fig. 3). In granulocytes, the decrease was slower and did not reach prestimulated levels (Fig. 3). The amplitudes of the signals were unchanged when fura red loading concentrations from 110 pM were used (not shown). When FMLP (4 FM) and CytB (1 pM) was added to fura red-loaded leukocytes, there was a rapid decrease in the fluorescence intensity in both cell types (Fig. 4A). The mean ratio of fura red intensity of stimulated to prestimulated cells was 0.42 f 0.02 and 0.48 2 0.02

I

2

4

6

8

'

I

2

4

I

6

I

I

8

TIME min FIG. 3. Flow cytometric determination of fluctuations of leukocyte cytoplasmic calcium concentrationwith fura red. Left:Monocyte fura red fluorescence displayed versus time. Right: Granulocyte fura red fluorescence displayed versus time. The ionophore A23187 (short arrows) and PBS with 4 pM EGTA were added at indicated time points (long arrow). The figure shows data from a single experiment and is representative of four performed.

for monocytes and granulocytes, respectively (mean ? SEM; n = 6). Monocytes and granulocytes responded simultaneously, but with different kinetics as the calcium concentration stayed a t high levels for a longer period of time in monocytes than in granulocytes (Fig. 4A). Addition of 100 ng/ml (160 nM) PMA did not induce sudden alterations in Fura red fluorescence of any cell type (Fig. 5).

Kinetic Measurement of Oxidative Bursts in Monocytes and Granulocytes With DHR 123 Stimulation of leukocytes with FMLPiCytB led to a rapid increase in the R 123 fluorescence of monocytes and granulocytes (Fig. 4A). After 8 min the mean relative R 123 fluorescence of FMLP/CytB-stimulated to nonstimulated cells was 12.6 +- 2.3 and 104.4 15.3 for monocytes and granulocytes, respectively (mean t SEM, n = 5). The response in granulocytes was more rapid than in monocytes (Fig. 4A). In contrast, monocytes responded more rapidly when PMA was used as a stimulus (Fig. 5), although the amplitude of the response was higher in granulocytes (Fig. 5).

Indicators used separately

Indicators used together

(two different samples)

A

(one sample)

B MOIV

GRAN

I I I 1 I I I I

I

-.

- 1 ._ I I

I I I

MON

GRAN 1

I

2

4

6

8

2

4

6

8

2

4

6

8

2

4

6

8

2

4

6

8

2 4

6

8

TIME min

TIME min Calcium-free medium

C

MON

No stimulus

GRAN

MON

GRAN

n

w c4 c p: 3

F4

k 2

2

4

6

8

2

4

6

8

4

6

8

2

4

6

8

2

4

6

8

r, t-4

3

%r

1 I M E niin FIG.4. Flow cytametric measurement of FMLP/CytB-induced cytoplasmic calcium fluxes and oxidative burst with fura red and DHR 123. Monocytes and granulocytes were recognized by light scatter measurements (see Fig. 1).FMLP (4p M ) and CytB (1 p M ) was added at indicated time points (arrows). A: DHR 123 (30 pM) and fura red used separately. Fura red-loaded (top) or unloaded cells (bottom) were suspended in PBSFCS-Ca without (top) or with (bottom) DHR

TIME min 123. ED: Fura red and DHR 123 (30 pM) used in combination. B: Cells were suspended in PBS-FCS-Ca. C: Cells were suspended in PBS-FCS-Ca and diluted 12 with PBS-FCS + 4 mM EGTA immediately before measurement. D Cells were suspended in PBS-FCS-Ca. WS was added at indicated time points (arrow). Fura red was measured in linear mode and R 123 in log mode. The figure shows data from a single experiment and is representative of four performed.

LUNDJOHANSEN AND OLWEUS

698

MON

25nM staurosporine

GRAN ,

t

I

GRAN

MON

p!

SL

L

I

A

2 4

6

8

r

Llmml4 2

4

6 8

I I I I I

2

4

6

8

TIME min FIG. 5. Flow cytometric measurement of PMA-induced changes in cytoplasmic calcium concentration and oxidative burst with fura red and DHR 123. Monocytes and granulocytes were recognized by light scatter measurements (see Fig. 1). Fura red-loaded cells were suspended in PBS-FCS-Ca containing 30 p M DHR 123. PMA (100 ngiml) was added at indicated time point (arrow). Fura red was measured in linear mode and R 123 in log mode. The figure shows data from a single experiment and is representative of four performed.

Combined Measurement of Cytoplasmic Calcium Fluxes and Oxidative Burst The kinetics and amplitudes of the FMLPiCytB-induced decrease in fura red fluorescence were not significantly altered in the presence of 30 p,M DHR 123 (Fig. 4A,B). Furthermore, the oxidative burst was not changed by loading cells with fura red (Fig. 4A,B). When the kinetics of the two responses were compared, it was evident that the calcium flux reached its peak before initiation of the oxidative burst (Fig. 4B). The results obtained when PMA was added to cells in the presence of both indicators (Fig. 5) were also indistinguishable from those obtained when the two indicators were used separately (not shown).

TIME rnin FIG. 6. Flow cytometric measurement of FMLPiCytB-induced calcium fluxes and oxidative burst in the presence of 25 nM staurosporine. Fura red-loaded leukocytes were incubated at 37°C with PBSFCS-Ca containing 30 pM DHR 123 and 25 nM staurosporine for 2 min prior to addition of the stimulus (arrow). Fura red was measured in linear mode and R 123 in log mode. The figure shows data from the same experiment as the one in Figure 4 and is representative of four performed.

5.8%(range 2.1-12.2) for monocytes and granulocytes, respectively (n = 10). When the first six samples were compared with the last six, an increase in the amplitude of the granulocyte oxidative burst of 9.4 k 2.0%ih was observed during a 5 h time period. No significant change in the background levels of R 123 fluorescence were observed during the 5 h period (not shown).

Effect of Calcium Chelators on Calcium Fluxes and Oxidative Burst Induced by FMLPICytB and PMA When extracellular calcium ions were chelated by adding EGTA less than 1min before running the samReproducibility of Measured Values for Calcium ple, the amplitude of the FMLP/CytB-induced calcium Fluxes and Oxidative Burst response in monocytes was inhibited by 38.8 -+ 4.0% The reproducibility of the measurements of cytoplas- (mean i SEM, n = 5) (Fig. 4A,C). Kinetic measuremic calcium fluxes and oxidative burst were exam- ments revealed that the second slow phase of elevated ined. The mean standard deviation of the FMLP/CytB- calcium concentration was abolished in monocytes induced calcium fluxes measured in a 5 h period, were (Fig. 4A,C). In granulocytes, the amplitude of the cal6.7% (range 3.1-12.2) and 5.5% (range 2.5-8.7) for cium flux was inhibited by 15 2 3%, and removal of monocytes and granulocytes, respectively (n = 10). For extracellular calcium ions had less dramatic effects on six samples successively analyzed at any time point, the kinetics of the calcium flux (Fig. 4A,C). Removal of the mean standard deviation of the FMLRCytB- extracellular calcium ions led to a 54.7 2 5.6% and a induced oxidative burst was 8.7% (range 2.5-14.2) and 47.1 t 10.2% inhibition of the oxidative burst in mono-

699

FCM MEASUREMENT OF CELL SIGNAL TRANSDUCTION

25nM s t a u r o s p o r i n e

MON

GRAN

I.

MON

GRAN

- II rr)

c-3

,"

fzl

I

2 I

2

I

I

4

I

I

6

I

I

I

8

2

4

6

8

TIME min FIG.7. Flow cytometric measurement of phorbol myristate acetate (PMA)-induced oxidative burst in the presence of 25 nM staurosporine. Fura red-loaded leukocytes were incubated at 37°C with PBSFCS-Ca containing 30 pM DHR 123 and 25 nM staurosporine for 2 min prior to addition of the stimulus. R123 was measured in log mode. The figure shows data from the same experiment as the one in Figure 5 and is representative of four performed.

4

6

8

2

4

6

8

I -l TIME min FIG. 9. Inhibition of the FMLPiCytB-induced calcium fluxes and oxidative burst by pertussis toxin. Cells were treated with pertussis toxin (2 Fgiml) for 2.5 h and resuspended in PBS-FCS-Ca. FMLP/ CytB was added at indicated time point (grey arrow).

tors did not significantly inhibit the response induced by PMA (not shown). 100

0 12.5

25

1

50

Staurosporine concentration nM

FIG.8. Dose-response curve for the inhibitory effect of staurosporine on granulocyte and monocyte oxidative burst. Cells were preincubated at 37°C with PBS-FCS-Ca containing 30 pM DHR 123 and indicated concentrations of staurosporine for 2 min prior to the addition of phorbol myristate acetate (PMA) or n-formyl-methionylleucyl-phenylalaninekytochalasinB (FMLPICytB). Monocytes and granulocytes were discriminated by light scatter measurements and the oxidative burst measured as a function of R123 fluorescence intensity. The degree of inhibition was calculated by the following formula: 100 x [l - (F staurosporine - F background/F control - F background)] where F staurosporine is the R 123 fluorescence of the cells stimulated in the presence of indicated concentrations of staurosporine, F background is the R 123 fluorescence at the same time point for unstimulated cells, and F control is the R 123 fluorescence a t the same time point for stimulated cells that were not incubated with staurosporine. The bars indicate SEM, n = 4.

cytes and granulocytes, respectively. In several experiments, granulocytes could be divided into distinct slowly and fast responding populations in the presence of EGTA (Fig. 4C). If intracellular calcium ions were depleted by loading the cells in the presence of 2 mM EGTA and 10 pM EGTA-AM, FMLPiCytB-induced calcium responses and oxidative bursts were abolished in both cell types (not shown). In contrast, calcium chela-

The Protein Kinase C Inhibitor Staurosporine Inhibits the Oxidative Burst but Has no Effect on Cytoplasmic Calcium Fluxes Staurosporine inhibited the oxidative burst of FMLPiCytB in a dose-dependent manner (Figs. 6, 8). The degree of inhibition was similar in monocytes and granulocytes (Figs. 6,8).In contrast, staurosporine had little or no effects on calcium fluxes induced by FMLPi CytB (Fig. 6).Staurosporine was approximately tenfold more potent as a n inhibitor of the oxidative burst when cells were stimulated with PMA instead of FMLPlCytB (Figs. 6-8). Pertussis Toxin Inhibits FMLP/CytB-Induced Calcium Fluxes and Oxidative Burst but not PMA-Induced Oxidative Burst Following treatment with 2 pgiml pertussis toxin for 2.5 h, granulocytes failed to respond with increased calcium concentration and oxidative burst following addition of FMLPiCytB (Fig. 9). In contrast, a n increase in monocyte calcium concentration was invariably observed, even in the presence of EGTA (Fig. 9 and data not shown). No oxidative burst was, however, seen in pertussis toxin-treated monocytes. Pertussis toxin did not lead to significant inhibition of the PMAinduced oxidative burst in any cell type (not shown).

700

LUNDJOHANSEN AND OLWEUS

of cytoplasmic calcium fluxes and oxidative burst activity could be monitored independently and simultaneously in the two cell types. As cytoplasmic calcium fluxes and activation of the oxidative burst represent early and late steps of phagocyte receptor-mediated activation pathways, respectively, this assay is well suited for studies of cellular signal transduction. The efficiency of fura red as an indicator of leukocyte cytoplasmic calcium levels was examined by monitoring effects of the calcium ionophore A23187 on monocyte and granulocyte fluorescence intensity. In cal+rrnll+ 2 4 6 8 2 4 6 8 cium-containing medium the fura red fluorescence of both monocytes and granulocytes was permanently reduced after addition of the ionophore. This decrease in . ...'.1.. . fluorescence was rapidly reversed by specifically chem lating calcium ions with EGTA, thereby demonstrating $1 that fura red fluorescence efficiently reflects changes p! in cytoplasmic calcium levels. An increase in cytoplasmic calcium levels was also seen when EGTA was added prior to stimulation with ionophore. This increase was smaller than in the presence of calcium 2 4 6 8 ions, and is likely to reflect release of calcium from TIME rnin intracellular stores (14). Thus, flow cytometric monitoring of leukocyte fura red fluorescence allows meaFIG. 10. Inhibition of granulocyte calcium fluxes and oxidative surement of rapid changes in cytoplasmic calcium levburst following suboptimal treatment with pertussis toxin (2 kg/ml for 1 h at 37°C).Leukocytes were treated with pertussis toxin (right) els as well as determination of the source of the or phosphate-bufferedsaline (PBS) only (left) and loaded with fura mobilized calcium ions. red. During flow cytometric measurement, the cells were suspended Due t o its long emission wavelength, fura red could in PBS-FCS-Ca containing 30 pM DHR 123. Fura red and R 123 be measured in combination with an indicator of oxifluorescence was gated from granulocytes (see Fig. 1) and displayed dative burst, DHR 123. With the instrument and filter versus time on two separate cytograms. FMLPiCytB was added at indicated time points (arrows). The figure shows data from a single settings used in this study, there was no overlap of fura experiment. red fluorescence into the detector used for R 123. In addition, the calcium fluxes and the oxidative bursts observed when using the two indicators in combination, did not differ significantly from those seen when Pertussis Toxin Inhibits Calcium Fluxes and they were used separately. This finding suggests that Oxidative Burst in Granulocytes by an the indicators do not interfere with cell function, and On-Off Mechanism that fluorescence energy transfer is negligible. ComIf the cells were treated with pertussis toxin for 1 h bined measurement of fura red and R 123 fluorescence instead of 2.5 h, distinct populations of strongly and made it possible to obtain direct evidence that FMLP/ weakly responding granulocytes were recognized after CytB-mediated oxidative burst is preceded by an instimulation with FMLP/CytB (Fig. 10). The strongly crease in cytoplasmic calcium concentration, whereas responding cells had calcium fluxes and oxidative that induced by PMA is not. bursts comparable to untreated cells, whereas the The results showed that the calcium fluxes were reweakly responding cells were close to the unstimulated producible within a time period of 5 h, demonstrating cells (Fig. 10). The amount of responding cells varied in good cell viability and little redistribution of fura red to different experiments, but a distinct population of cells calcium-rich cell compartments over time. The backthat responded similar to untreated cells was always ground levels of cellular R 123 fluorescence were also seen in samples treated with suboptimal doses of per- stable within the 5 h period. A stable, low-background tussis toxin (Fig. 10). Monocytes, on the other hand, fluorescence is largely due to the fact that DHR 123 responded as one single population following treatment may be added immediately before stimulation, thereby with any dose of pertussis toxin (data not shown). eliminating the spontaneous intracellular oxidation of dye that would occur over time with the more comDISCUSSION monly used indicator dichlorofluorescein (2,171. The R This paper reports an assay for the measurement of 123 fluorescence of stimulated cells was in some expercytoplasmic calcium fluxes and oxidative burst in iments found to increase during the 5 h period, but a t monocytes and granulocytes. Using the two novel in- any time point, a series of six samples could be run dicators fura red and dihydrorhodamine 123 in combi- successively with less than 10%SD of the amplitudes of nation with multiparameter flow cytometry, kinetics the oxidative bursts. It is therefore suggested that

PBS 60 min

Pertussis toxin 2pg/ml 60min

FCM MEASUREMENT OF CELL SIGNAL TRANSDUCTION

when dose-response relationships of compounds that interact with phagocyte activation are investigated, parallel series should be run, rather than parallel samples. When samples are run in this serial mode, the method should allow sensitive detection of inhibition or stimulation of the oxidative burst within a 5 h period. To determine the role of calcium ions for activation of monocytes and granulocytes by various stimuli, calcium fluxes and oxidative burst were measured in the presence of both extracellular and intracellular calcium chelators. When extracellular calcium ions were removed by adding EGTA immediately before running the sample, there was moderate inhibition of the calcium fluxes and the oxidative burst induced by FMLP/ CytB. When both extracellular and intracellular calcium was removed with EGTA and EGTA-AM, FMLPI CytB-mediated calcium responses and oxidative burst were abolished, whereas the PMA-induced oxidative burst was unchanged in both cell types. These results harmonize well with those reported by others (14,16) and demonstrate that high levels of cytoplasmic calcium ions are not obligatory for activation of the NADPH oxidase enzyme complex. Removal of calcium ions therefore most likely interferes with the signal that couples the ligand-occupied FMLP receptor to enzyme activation. Further, the calcium that is mobilized by release from intracellular stores seems to be sufficient for activation of the NADPH oxidase by FMLPI CytB, but for maximal oxidative burst, influx of calcium via plasma membrane channels is necessary. The usefulness of the assay in detecting how modulators influence signal transduction was evaluated by use of staurosporine and pertussis toxin. The protein kinase C inhibitor staurosporine had no effects on FMLP/CytB-mediated calcium fluxes, but inhibited partially the FMLPICytB-mediated oxidative burst, and abolished the oxidative burst induced by PMA. Staurosporine therefore most likely inhibits the oxidative burst by interfering with a step in the signal transduction pathway that is downstream to the calcium responses, as suggested by others (5,19,21). The large difference in the inhibitory activity of staurosporine on the oxidative burst induced by FMLPICytB and PMA suggests that the FMLP/CytB-induced oxidative burst is largely independent of protein kinase C. The inhibitory effect of higher concentrations of staurosporine observed in this and earlier studies (5,231 could be due to interference with other kinases, such as protein tyrosine kinase (18,24). When used at concentrations below 15 nM, it seems, however, that staurosporine has good specificity for protein kinase C in human monocytes and granulocytes. In contrast to staurosporine, pertussis toxin inhibited both calcium fluxes and oxidative bursts induced by FMLPICytB in granulocytes, but had no effect on the oxidative burst induced by PMA. These results are compatible with the view that pertussis toxin, by inducing ADP-ribosylation of the alfa-i subunit of GTPbinding proteins, inhibits an early step in the stimu-

701

lus-response coupling of the granulocyte FMLP receptor (11). Interestingly, pertussis toxin inhibited the granulocyte oxidative burst and calcium fluxes in an on-off fashion. Omann and Lowes reported a similar inhibitory activity of pertussis toxin for FMLPICytBmediated calcium responses in granulocytes. As oxidative burst was measured on whole cell suspensions in their study, a corresponding reactivity pattern was not demonstrated for this response (13).The distinct inhibitory patterns of calcium chelators, staurosporine, and pertussis toxin demonstrated in the present paper show the applicability of multiparameter flow cytometric measurements in studies of cell signal transduction. Simultaneous measurement of monocytes and granulocytes revealed that the two cell types differed in early as well as late responses to activation. An EGTAinsensitive calcium flux was observed in monocytes but not in granulocytes following treatment with pertussis toxin. This may suggest a cell-type-specific difference in the G-proteins that couple the FMLP receptor to phospholipase C. A cell-type-specific heterogeneity in cytoplasmic calcium regulation is also likely from the results showing higher permeability to calcium ions of monocyte plasma membranes. Monocyte membranes are probably also more permeable to phorbol esters, since they responded more quickly than granulocytes following stimulation with PMA. In addition to the apparent differences in early signal transduction, there was also a large difference in the amplitude and kinetics of the oxidative burst in monocytes and granulocytes following stimulation with FMLP/ CytB or PMA. In conclusion, this study describes an assay for simultaneous measurement of cytoplasmic calcium fluxes and oxidative burst in leukocyte subsets. The assay allows evaluation of how modulating substances affect early and late steps in signal transduction and detection of cell-type-specific differences in response patterns.

LITERATURE CITED 1. Babior BM: Oxygen-dependentmicrobial killing by phagocytes. N Engl J Med 298:721-725,1978. 2. Bass DA, Parce WJ, DeChatelet LR, Szejda P, Seeds MC, Thomas M: Flow cytometric studies of oxidative product formationby neutrophils: A graded response to membrane stimulation. J Immunol 130:1910-1917, 1983. 3. Davies TA, Weil GJ, Simmons ER Simultaneous flow cytometric measurementsof thrombin-inducedcytosolic pH and Ca2+ fluxes in human platelets. J Biol Chem 26511522-11526, 1990. 4. Della Bianca V, Grzeskowiak M, Rossi F: Studies on molecular regulation of phagocytosis and activation ot the NADPH oxidase in neutrophils. J Immunol 144:1411-1417, 1990. 5. Dewald B, Thelen M, Wymann P, Baggiolini M: Staurosporine inhibits the respiratory burst and induces exocytosis in human neutrophils. Biochem J 264:879-884, 1989. 6. Emmendtirfer A, Hecht M, Lohmann-MatthesM, Roesler J: A fast and easy method to determine the production of reactive oxygen intermediates by human and murine phagocytes using dihydrorhodamine 123. J Immunol Methods 131:269-275, 1990. 7. Haugland H Long wavelength Ca2' indicators. Bioprobes 13: 4-6, 1991.

702

LUNDJOHANSEN AND OLWEUS

8. Lambeth JD: Activation of the respiratory burst oxidase in neutrophils: On the role of membrane-derived second messengers, Ca + and protein kinase C. J Bioenerg Biomembr 20:709-733, 1988. 9. Lazzari KG, Proto PJ, Simons ER: Simultaneous measurement of' stimulus-induced changes in cytoplasmic CaZ+and in membrane potential of human neutrophils. J Biol Chem 261:9710-9713, 1986. 10. Naccache PH, Gilbert C, Caon AC, Gaudry M, Huang CK, Bonak VA, Umezawa K, McColl SR Selective inhibition of human neutrophil functional responsiveness by erbstatin, an inhibitor of tyrosine protein kinase. Blood 76:2098-2104, 1990. 11. Ohta H, Okajima F, Ui M Inhibition by islet activating protein of a chemotactic peptide -induced early breakdown of inositol phospholipids and Ca2 ' mobilization in guinea pig neutrophils. J Biol Chem 260:15771-15780, 1985. 12. Omann G, Harter JM: Pertussis toxin effects on chemoattractantinduced response heterogeneity in human PMNs utilizing Fluo-3 and flow cytometry. Cytometry 12:252-259, 1991. 13. Omann GM, Lowes MM: Graded G-protein uncoupling by pertussis toxin treatment of human polymorphonuclear leukocytes. J Immunol 146:1303-1308, 1991. 14. Pozzan T, Lew DP, Wollheim CB, Tsien RY: Is cytosolic ionized calcium regulating neutrophil activation? Science 221:14131415, 1983. 15. Kabinovitch PS, June CH, Grossmann A, Ledbetter JA: Heterogeneity among T cells in intracellular free calcium responses after mitogen stimulation with PHA or anti-CD3. Simultaneous use of indo-1 and immunofluorescence with flow cytometry. J Immunol 137:952-961,1986.

+,

16. Rossi F, Della Bianca V, Grzeskowiak M, Bazzoni F: Studies on molecular regulation of phagocytosis in neutrophils. J Immunol 142:1652-1660, 1989. 17. Rothe G, Oser A, Valet G: Dihydrorhodamine 123: A new flow cytometric indicator for respiratory burst activity in neutrophil granulocytes. Natunvissenschaften 75:354-355, 1988. 18. Ruegg UT, Burgess G M Staurosporine, K-252 and UCN-01:Potent but nonspecific inhibitors of protein kinases. TIPS. 10:218220, 1989. 19. Sako T, Tauber AI, Jeng AY, Yuspa SH, Blumberg PM: Contrasting actions of staurosporine, a protein kinase C inhibitor on human neutrophils and primary mouse epidermal cells. Cancer Res 48:4646-4650, 1988. 20. Szejda P, Parce JW, Seeds MS, Bass DA: Flow cytometric quantitation of oxidative product formation by polymorphonuclear leukocytes during phagocytosis. J Immunol 133:3303-3306, 1984. 21. Tamaoki T, Nomoto H, Takahashi I, Kato Y, Morimoto M, Tomita F: Staurosporine, a potent inhibitor of phospholipid/Ca + + dependent protein kinase. Biochem Biophys Res Commun 135:397402, 1986. 22. Thelen M, Peveri P, Kernen P, van Tscharner V, Walz A, Baggiolini M: Mechanism of neutrophil activation by NAF, a novel monocyte-derived peptide agonist. FASEB J 2:2702-2706, 1989. 23. Watson F, Robinson J , Edwards SW: Protein kinase C-dependent and independent activation of the NADPH oxidase of human neutrophils. J Biol Chem 266:7432-7439, 1991. 24. Yamashita Y, Hasegawa-Sasaki H, Sasaki T: Suppression by staurosporine of Ca2+ mobilization triggered by ligation of antigen-specific receptors on T and B lymphocytes. FEBS Lett 288: 46-50, 1991.