cellcaldun (lQa3)13, 487-m Q Longman GrcupUK Ltd 1902

Thapsigargin reveals evidence for fMLPinsensitive calcium pools in human leukocytes J.S. RBTNES and J.G. IVERSEN Department of Physiology, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway Abstract - To investiyte the relationship between different intracellular Ca*+ pools, cytosolic free calcium ([Ca *]i) was surveyed by means of a Fura- fluorescence ratio method on single isolated human leukocytes. Both monocytes and neutrophilic granulocytes (PMN) displayed long lasting spontaneous [Ca*‘]i transient changes (l-2 min). In PMN stimulated with the bacterial peptide fMLP we observed transients with shorter duration (IO-30 s) and smaller amplitude often superimposed on the long lasting transients. The time course of changes in [Ca*+Ii was recorded in a large number (149) of single leukocytes prestimulated for 5 min with fMLP and then challenged with thapsigargin (a blocker of Ca*+ uptake in intracellular pools). Statistical analysis of Ca*+& responses revealed that fMLP-sensitive pools contributed to the long lasting [Ca12 transients seen in both leukocyte types. However, the existence of fMLP-insensitive calcium pools may explain the superimposed transients seen in PMN. Thapsipargin was also added together with EGTA (to impede contribution from extracellular Ca +) to 198 fMLP prestimulated and 153 unstimulated PMN. Based on Ca*+ registrations in these ceils and a mathematical model (supposing two separate first order responses) the amount of Ca*’ stored in the various pools and their release kinetics were estimated. The results indicate that fMLP-insensitive calcium pools exist in PMN but not in monocytes. Since the digital imaging technique also depicts cellular motility, an additional finding was that the leukocyte’s ability to sequestrate the Ca*+ from the cytosol seemed important to locomotion.

Abbreviations: FLUX-~:l-[2-(Scarboxyoxazol-2-yl)-6-aminobenzofuran5-oxy]-2-(~-amino-S-methylphenoxy)-ethane-N,N,N’,N’te traacetic acid; FurrA/AM: Fura-2-acctoxymethyl es*, fMLPz N-formyl-methionyl-leucyl-phenylalanine; Inap3: myo-inositol 1.4.5trisphoaphate; PMN: polymorphw nuclear (neutrophilic granulocyte); EGTA: ethylene glycol bk@-aminoethyl ether) NXN’,N’-tetraacctic acid, Tg: thapsigargin, HEPES: 4(2-hydroxyethyl)- 1-piperazine ethanesulfonic acid; [Ca*+]i: cytosolic free calcium; W: endoplasmic reticulum

Changes in cytosolic free Ca*+activity represent an -0-t


of in~&&r

si ds. 2P


changes are brought about both by Ca influx from the outside aud by release from intracellular stores. Functional studies in various cell types have suggested the existence of different types of intracellular Ca*+ stores: (a) InsPwensitive Ca*+ stores, i -e - Ca*+ stores that release Ca*+ in response to InsP3. most likely via the InsP3 receptoc (b) Ca*+/ caffeine-sensitive Ca*+ stores, i.e. Ca*+ stores that




ca2+in response to [Ca2+]ielevations or caffeine, presumably via the ryanodine receptor. and (c) InsPs-insensitive and [Ca2+]i-insensitive Ca2+stores, no2yhysiological release mechanism for this type of Ca store being known. Recent studies have identified InsPs-insensitive pools in various non-muscle cells [l-3], with a release mechanism that is caffeine-sensitive through a ryanodine receptor. In some of these reports the InsP3-insensitive cytoplasmic regions appear to be localised to other regions of the cell than the InsP3- sensitive pools. Furthermore, evidence for other pools that are insensitive to both InsP3 and caffeine, has been reported [4,X. In the present investigation we wanted to clarify the relationship between different Ca2’ pools in human leukocytes. The [Ca2+]imeasurements on PMN were compared with corresponding ones on mononuclears (mainly monocytes), which were monitored simultaneously in the same experiment. The two types of leukocytes are known to be different in several respects. Neutrophils (PMN) am end cells which synthesize only a few proteins and accordingly contain very little ER [3, 61. The remaining non protein synthesizing part of ER may still be a Ca2’ store, though. Another possible Ca2’ store in neutrophils am so called calciosomes, by some authors regarded as the main InsP3-sensitive pool 16, 71. On the other hand, the monocytes have well developed Golgi apparatus and ER suited to their long lasting activity once they have transformed to tissue macrophages. The widely used leukocyte stimulating bacterial peptide fMLP (receptor binding: Kd = 10 nM) induces formation of diacylglycerol and InsPs inside the cell. InsPs functions as second messenger to mobilize Ca2+ from intracellular stores and diacylglycerol activates protein kinase C [8]. We used a supramaximal concentration of fMLP to achieve the highest possible depletion of the Imps-sensitive stores. Thapsigargin (Tg), a sesquiterpene la&one isolated from the root of the plant 7’hapsia arganica L, causes elevation of cytosolic free. Ca1p in several s [4, g-151, including PMN [16]. Tg is not ;i%Y ionophore [16], and it does not affect the generation of InSP3 1111. Tg is a2ytent inhibitor of the endoplasmic reticulum Ca -ATPase in rat may release

hepatocytes [12], and Lytton et al. [15] have recently demonstrated that Tg inhibits specific Ca2+ATPases in intracellular pools. Thus Tg may cause a rise in [Ca2+]l by preventing ATP-dependent uptake of Ca2+ into intracellular stores. The Tg effect should accordingly reflect the turnover of Ca2’ between cytosol and intracellular pools. When EGTA is also added, any influence of Ca2’ inllux across the plasma membrane should be minimized. The present experhnents were designed to investigate the existence of fMLP-insensitive (presumably In&-insensitive) calcium pools in human leukocytes. We did this, by first depleting the fMLPsensitive stores and then exposing the residual stores with Tg and EGTA in combination. Image analysis uncovered two types of leukocytes which responded differently both to fMLP and Tg. By observing [Ca2+]lin a large number of single cells we were able to statistically evaluate the differences and similarities between the two cell types and the effects of the stimulants. The results indicate that fMLP-insensitive calcium pools exist in PMN but not in fMLP sensitive mononuclears (monocytes).

Materials and Methods Materials

The dye Colorrapid used in staining the leukocytes was obtained from Lucema-Chem AG. Standard salt solutions consisted of (mmol/l): NaCl, 136; KCl, 5; MgCla, 1.2; CaClz, 1; Bactodextrose, 11; HEPES 10 or 30; pH 7.35. These salt solutions are in the following referred to as HEPES 10 mM and HEPES 30 rnM, respectively. Hydroxyethyl starch 6% (‘Plasmasteril’) was obtained from Fresenius AG, Bad Homburg, Germany. Thapsigargin from Gibco Ltd. Paisley, Scotland EGTA, poly-L-lysine, fMLP and HEPES f-mmSigma, St Louis, MO, USA; Fura-2/AM from Molecular Probes, Eugene, OR, USA; and Araldite from Ciba-Geigy, Basel, Switzerland. y-Interferon was obtained from Boehringer Mannheim, Germany. Leukocytepreparation and classijkation

Leukocytes were isolated from heparhked blood, in







Photograph of Colonapid

stained leukocytes, showing 2 typical mononuclear leukocytes and 10 typical PMN, represt mting

half the field of view in one experiment.

00 Mean [Ca2’]i behveen 2&30 s after addition of Tg/EGTA displayed as a pseudorolor image of the same fMLP prcstin~~ulatcd leuk :ocyks as shown in (A)


experiments taken from the same healthy volunteer (to minimize interindividual differences). Blood (4 ml) was mixed with 4 ml 30 mM HEPES salt solution and 2 ml ‘Plasmasteril’. After 30 min sedimentation, 3 ml of the supematant was distributed to 10 wells (two small polyethylene cylinders were glued with Araldite to 5 cover glasses, making a total of 10 wells). The cover glasses had in advance been coated with poly-L-lysine (each well was incubated with 0.4 ml 1 mg/ml poly-L-lysinefor 15 min before washing with distilled water). Leukocyte sedimentation was continued in the wells for 15 min while lightly shaken in a 37°C water bath. Leukocytes not attached plus the red blood cells and most of the platelets were removed by gentle washing. The attached leukocytes were kept at 37°C and incubated in 12 @I Fun&!/AM in 10 mM HEPES for 30 min (300 pl in each well). The cells were washed twice and incubated in a medium containing 30% plasma and 70% 30 mM HEPES. All wells were used within 6 h after the vein puncture. At the end of each registration the cells were stained with a routine stain for blood smears (Colorrapid) as displayed in Figure 1A. Two morphologically distinct types of leukocytes are apparent: neutrophilic granulocytes (PMN) and one mononuclear leukocyte type, which mainly consists of monocytes. Lymphocytes were in our experiments not responsive to fMLP (data not shown) and almost compIetely removed by the washing procedure. all


RAM. The imaging and registration software have been developed in our laboratory. The video camera and control signals are concomitant with the real time registration recorded on a video tape (all control signals arc coded on the audio tracks). Sampling of [Ca2+]ifrom all the cells were after the experiment recorded from the video tape (with correction of the slight bias introduced by the video recording) by computing the median of the fluorescence signals from a square covering most of the cell. The median is preferred to the mean because the camera signals might saturate and median is insensitive to saturation as long as at least 50% of the pixel values are below the saturating level. One group of samples was calculated in 100 ms. The position of the square was updated every 6th second to follow the moving leukocyte. Photobleaching of Fura- was reduced by using a photographic shutter interposed between the light source and the microscope 50% of the registration period so that 2.5 ratios were measured in 1 s. The fluorescence signals were smoothed by averaging over a period of 3 s. The microscope table was assembled in an incubator box which kept the cells and solutions at 37°C. The free Ca2’ concentration was calculated using the equation: [Ca’+]= Kd fi (R-Rmin)/(RmarR)

defmed according to [17], where Kd = 224 nM, fi = 6.4, Rmia= 0.2 and Rmax= 5.0, calibrated by using Measurement of [0?‘]1 in moving leukocytes, an isotonic 0.9% NaCl solution with 1 mM [Ca2+] experimentaldesign (Rmax) and nominal zero [Ca2’] + EGTA (Rmin) with the video camera as detector. The exitation The fluorescence measurement instrumentation con- wavelengths were 345 and 385 nrn with wavelength sists of 2 monochromators (PTI A-SCAN 1, Ham- bands of 15 and 20 MI. The autofluorescense was burg, Germany), a Nikon Diaphot-TMD inverted measured from cells containing no Fura- and was microscope with a Fluor 40x objective (N.A. 1.3) subtracted from the fluorescence signal before the where the excitation beams are deflected into the ratio was calculated. objective by a dichroic mirror. Images were collected through a 30 nm HBW filter at 510 nm with an Assay of stimulant application and cell viability intensified CCD video camera (Extended ISIS-M, criteria Photo& Science, Robertsbridge, UK), a Nikon 0.9-2.25 zoom lens placed in front of the video At the beginning of the experiment leukocytes in camera, a Sony u-matic SP video recorder, a frame each well were incubated in 300 pl30 mM HEPES/ grabber (512 x 512 pixel per frame, &bit resolution) plasma solution. fMLP in 200 pl 10 mM HEPES and a 386/33 MHz computer with 13.5 Mbytes was added in less than 2 s to prestimulate the


leukocyte and TgEGTA in 500 pl 10 mM HEPES was added 5 min afterwards in the same way. Control experiments without tMLP prestimulation were achieved by just adding 200 p.l 10 mM HEPES (which did not alter the [Ca2+]ilevel). TgEGTA was added automatically by a computer controlled injection device and the time of the application was recorded on the video tape. Only cells that were clearly diagnosed as mononuclear-s or PMN, were included in the statistical calculations. Furthermore, the average fluorescence signals had to be higher than 22% of the video saturation level, and all samples had values lower than the saturation level. To eliminate artefacts introduced by presumably dying cells (high basal [Ca2+]iand no tMLP response), all leukocytes with a small fMLP response wem excluded. Thus, only cells responding to fMLP by at least 110 nM [Ca2+]lincrease in the first set of experiments (exploring the Tg/EGTA effect on fMLP prestimulated leukocytes) were included. In the second set of experiments (exploring the effect of prestimulation with fMLP) PMN leukocytes that displayed less than 20 nM variation in basal [Ca2+]iwere excluded (to eliminate cells with little spontaneous release). In both sets a total of 3 x 8 wells from 3 different experimental days were observed.


endent, sequestered calcium (see introduction), the increased [Ca2+]ldue to Tg in a Ca2’-free medium includes contribution from agonist-sensitive as well as agonist-insensitive stores. In our calculations the TgiEGTA response was defined as the difference between the minimal and succeeding maximum [Ca2+]l level within 4 min after the Tg/EGTA application. Statisticalevaluationand numerical tools

The interrelation between data was evaluated by computing the linear correlations, and Fisher’s z-transformations were used to estimate significance levels in comparisons between correlations. An approximately binormal distribution was achieved by calculating the logarithm of all data before the correlation and significance analyses were executed (Tables 2 and 3). Statistical evaluation of other data was performed with Ztailed Student‘s t test with unequal variances. The software used in the statistical analysis and the power spectrum calculations was developed by implementing the numerical recipes from [HI. The multiple exponential regression analysis used in Figure 5A was performed on the statistical program BMDP-AR (derivative-free nonlinear regression). All P-values listed arc based on a two-sided alternative. The power spectra were estimated by the Definition of the basal /Ca2’]i, initial jMLP maximum entropy method (16 poles) [18]. We response, relative jMLP response and the thapsi- transformed the [Ca2+]isignals from 3 min before garginlEGTA response the fMLP stimulation until the TgEGTA addition, and to avoid discontinuities the zero line was Basal [Ca2+]iis defined as the mean [Ca2+]llevel defined between the starting value and the value at during a 3 mitt period before the first stimulant the end of the trace. application. As mentioned in the introduction, fMLP is a stimulant that induces formation of diacylglycerol Results and InsPs, thus the fust phase in the fMLP response expresses (at least to some extent) release from the Calcium transientsin human leukmytes InsP3-sensitive pools. The initial fMLP response (used to quantify the contribution from the agonist- Multiple transient elevations of cytosolic free sensitive pools) was taken to be the maximum calcium were observed in single human leukocytes (Ca2+]ilevel during the 160 s period after the fMLP both spontaneously and after stimulation with fMLP. stimulation. A relative fMLP response is defmed as These calcium transients, as observed in monothe initial fMLP response divided by the basal nuclear leukocytes, were often marked and with a duration of several minutes (Fig. 2A), whereas the [Ca2+]ilevel. Since Tg seems to deplete all Ca2’-ATPase dep- calcium transients in human PMN consisted of two



components, long lasting spikes (l-2 min) with transients of shorter duration (10-30 s), often superimposed. The long lasting transient spikes were in general retained also in a Ca2+-freemedium (Fig. 2B). The superimposed transients often looked like the sinusoidal oscillations illustrated in Figure 2C and could also be elicited in ca2’-tiee medium by adding Tg/EGTA (Fig. 3C).


Time (set) Ng. 3 Eflkot of MLP (1 pM) followed by thapsigargin (3 pM) and EGTA (5 ml@ on human leukocytes.

(A) Registration of a

typical response in a mononuclear leukocyte with a large fMLp innease and a small Tfit3TA

effect. (B) Registration of typical

Ph@I response with mom advanced T@GFA imxease.

Time (SW) Ng.


(A) Cytosolic




leukocyte before and ai&

in a single human




and (C) Cytosolic calcium transients in a FMN in spite of the presence

of Ca2+-complexiog

EX3TA Q), and their markedly

changed pattern a&r tMLP stimulation (C)

induced [Ca*+]i

(C) An oscillating Tg/EGTA mspcose in a single PMN

Eflect of thapsigargin in a calciumfree medium on jMLP-prestimulatedsingle leukocytes

Upon stimulation with fMLP at 1000 nM and TgIEGTA at 3000 nM/5 mM, both the magnitude and the shape of the responses we= quite different


in the two types of leukocytes (Pi$; 3). fMLp induced a transient elevated [Ca ]I level in mononuclears (Pig. 3A). The representative trace shows an initial increase to 630 nM; the Tg/EGTA response, however, was just 40 nM. On the other hand, the representative fMLP induced [Ca2+]i increase in PMN was 440 nM (Pig. 3B), whereas the TgEGTA elicited a 270 nM transient increase in the [Ca’+]i. In both cell types the stimulation with Tg/EZGTAinvariably resulted iu a short decrease in the [Ca2+]ilevel (probably due to abolished Ca2’


influx) with a subsequent transient increase. In some of the PMN this subsequent increase showed an oscillatory pattern (Fig. 3C). Replacement of Tg with 20 mM caffeine failed to elicit any [Ca2+]i increase (data not shown). Statisticaldiflerences between the calcium responses of PMN and mononuclears

We compared 71 mononuclears and 78 PMN in three different experiments on separate days. The

Table 1 Effects of fMJ_Pand thapsigargin/E!GTAon PMN and mononuclear leukocytes ICa’* Ii hMJ

Cell type and



number recorded

Net cell











migration (p) 43 (16)











Table 2 Analysis of con-elation between basal, fh4lB and thapsigargiu/EGTA-induced [Ca2’]i responses in human leukocytes Correlation pair: Cell rype

(Number of celle)

Initial fML,P response vs basal [Ca2+]i : Mononuclears PMN, all cells PMN. ‘fast movers’


response vs basal [Cs2+]i

-0.30 a.26 -0.51

P- 0.01 P - 0.02

(71) (78)

0.35 a.04

P - 0.003. PL - 0.02

(71) (78) (153)

a.39 0.01 -0.45

P - 0.001. pb - 0.01

(71) (78) (30)

Relative fMLP response vs thapsigaqin/EGTA Monomdears PMN


Correlation coe~cient





Prestim.mononucletus Prestim. PMN Not prestim. PMN Cellswerez1dysedon3di&rsnte~tal&ys. 8cc foumi in the tsxt. r.Bvel of L+nificance catid

fMLl’wasadded5mktbrfautheT#EGTAadditkf1. by Piir z-ellndmmk

P-diffaent~zaoandPb”‘-differmtfromthefMLepreathnulatsdPMNleukocytes. * Adysi¶ of the 30 fwest moving cells


Thapsigargin reveals evidence for fMLP-insensitive calcium pools in human leukocytes.

To investigate the relationship between different intracellular Ca2+ pools, cytosolic free calcium ([Ca2+]i) was surveyed by means of a Fura-2 fluores...
4MB Sizes 0 Downloads 0 Views