GLIA 16:10&116 (1996)

Pzu Nucleotide Receptor Activation in Rat Glial Cell Line Induces [Ca2+]; Oscillations

Which Depend on Cytosolic pH UWE CZUBAYKO AND GEORG REISER

Instilut fur Neurobiochemie der Otto-van-Guericke- Universittit Magdeburg, Magdeburg, Germany

KEY WORDS Glial cell line, Purinoceptor, Signal transduction, Intracellular pH, Furosemide, Amiloride, Na'-K' -ClF-transporter

ABSTRACT In single rat glioma cells, the signal transduction process activated by the UTP sensitive purinergic nucleotide receptor was studied by determining [Ca2+1, by Fura- 2 fluorescence and measuring pH by BCECF fluorescence to elucidate the control of [Ca2+l, oscillations by intracellular pH. Addition of UTP for long time periods (some min) causes a [Ca2-1, response composed of i) an initial large peak and a following sustained increase (160 s duration), and ii) subsequent regular [Ca2'], oscillations (amplitude 107 nM, fre- quency 1.5 oscillations per min). The maintenance of the [Ca2+], oscillations depends on the continued presence of agonist. The oscillations are abolished by reducing extracellular Ca2+ concentration. The interaction of UTP receptors and bradykinin receptors during the TCa"], oscillations was investigated because previous studies have already shown that the peptide causes comparable [Ca2+], oscillations. During [Ca"], oscillations induced by UTP or bradykinin, long-term admission of both hormones (400-500 s) causes a large initial response superimposed on regular [Ca"], oscillations. Short pulses (12 s) of the second agonist given in any phase of the oscillations induce large [Ca2'], peaks. In both cases, the following oscillations are not disturbed. The influence of cytosolic pH was studied by alkalinizing pH, by application of NH,Cl. [Ca2+], oscillations stop after addition of NH4C1. Recovery of NH4C1-induced alkalinization is reduced by furosemide. To the same degree, the interruption of [CaZ+], oscillations is significantly prolonged in the presence of furose- mide. Thus cytosolic alkalinization suppresses hormone-induced [CaL+], oscillations in rat glioma cells. The understanding of the molecular mechanism of this interference of pH should provide an important contribution for unravelling the function of cytosolic pH in cellular signal transduction. o 1996 Wiley-Liss, Inc.

INTRODUCTION

The existence of receptors for extracellulary occurring ATP, originally proposed on the basis of pharmacological evidence from autonomic neuroeffector cells has now been firmly established (refs. in Dubyak and El-Moatas- sim, 1993). These receptors which specifically recognize purine nucleotides, classified as P2-purinergic recep- tors, are found as a largely heterogeneous species (Zim- mermann, 1994). Subclasses of P2 receptors, Pax, Pay, Pzz, and PZt were defined by pharmacological criteria using sequences of affinity for various nucleotide ago- nists (Fredholm et al., 1994). Recently two types of G- protein coupled purinoceptors were identified by molec- 0 1996 Wiley-Liss, Inc.

ular cloning, the PZp receptor (Webb et al., 1993) and the P21T receptor (Lustig et al., 1993). The latter receptor is characterized by the responsiveness not only to pu- rines (adenine nucleotides) but, with equal potency, also to the pyrimidine nucleotide UTP.

Received June 22, 1995; accepted August 24, 1995

Abbreviations BCECF 2',7'-his(2-carboxyethyl)-5(6)-carboxyfluorcscein HES HEPES buffered saline [Ca"], cytosolic Ca2+ activity

Address reprint requests t o Prof. Dr. G. Iiciser, Otlo-von-~uericke-Universitat Magdehurg, Institut fur Ncurobiochemie, Leipziger Strasse 44, 39120 Msgdc- burg, Germany.

EFFECT OF pH, ON GLIAL [Ca‘+l, OSCILLATIONS 109

Apart from acting via second messenger-inducing re- ceptors, ATP also directly opens ion channels (Edwards et al., 1992; Evans et al., 1992). The ATP-gated ion channel, which has also been molecularly cloned (Brake et al., 1994; Valera et al., 19941, is permeable to Nat, K’, and Ca”. The presumable membrane topology of the receptor i s not characteristic for that of the superfamily of ligand-gated ion channels. This indicates that the function of ATP as a neurotransmitter has evolved during evolution independent from that of clas- sical neurotransmitters like acetylcholine or GABA.

Astrocytes, the main population of glial cells, display receptors for transmitters as varied as those expressed by neurons. However, the function of neurotransmitter receptors on glial cells is still largely unknown (Kimel- berg, 1995). ATP and UTP activate phospholipase C, Ca2+ release, and arachidonic acid formation in astrocytes (Bruner and Murphy, 1993; Pearce and Lang- ley, 1994). Similarly, in Schwann cells (Lyons et al., 1994) and in oligodendrocytes (Kirishuk et al., 1995) PZy receptors trigger CaL- release from internal stores.

The physiological significance of purinoceptors on glial cells still has to be clarified. The nucleotides may have modulatory functions similar to those proposed for the heterogeneous class of glutamate receptors (Hanson, 1995). This involves short-term regulation of metabolism and possibly of cell volume, but on the other hand, also long-term mitogenic effects by enhancing the prolifera- tion rate (Ciccarelli et al., 1994; Neary, 1994). Pathologi- cal processes may be affected since nucleotides are re- leased at large quantities from damaged or stressed cells.

Since continuous addition of a variety of hormones causes oscillations, we tested the effect of the nucleotide UTP in rat glioma cells, which are studied commonly as a model for astrocytes. Using a different clone of this cell line, C6-2B, Munshi et al. 11993) have found that ATP and UTP activate phospholipase C, induce a transient rise in [Ca2-],, and inhibit CAMP accumulation elicited either by stimulation of (3-adrenergic receptors or by di- rect activation of adenylyl cyclase. The similarly high po- tency ofATP and UTP in activating the nucleotide recep- tors in the glioma cells (Lin, 1994; Munshi et al., 1993) indicates the presence of Pzu receptors in these cells.

We report here that continuous exposure of rat glioma cells to UTP causes [Ca2+], oscillations with a frequency of 1.5 per min and an amplitude of 107 nM. The mainte- nance of the oscillation depends on the presence of suffi- ciently high extracellular Ca2+ concentration (above 100 KM). The question of interference of cytosolic pH and [Ca”l, oscillations was investigated by pH shifts in- duced by an NH&l pulse. Alkalinization suppresses the oscillation. This inhibition is prolonged by simultaneous addition of furosemide.

MATERIALS AND METHODS

Polyploid rat glioma cells (clone C6-4-2) of passage numbers between 14 and 32 were cultured as described (Reiser, 1992). [Ca”]: recording from single glioma cells

was carried out according to Reetz and Reiser (1994a) using microspectrofluorometry. In brief, cells grown on circular glass coverslips (12 mm in diameter) with 20 to 90% confluency were removed from the culture dish and placed into 1 ml HBS (145 mM NaC1,5.4 mM KC1, 1.8 mM CaC12, 1.0 mM MgC12, 25 mM glucose, 2 mM Na2HP04, and 20 mM HEPES, pH 7.4, adjusted with nis) supplemented with 2.5 ~J.M FURA-2IAM (for [Ca2+], measurement) or 2 pM BCECF-AM (for pH, measure- ment) for 30 min at 37°C. HBS with different Ca2+ con- centrations was prepared by adding CaCla a t the con- centration indicated to nominally CaLt-free HBS.

Fluorescence was determined from single cells loaded with the dye Fura-2JAM or with BCECF/AM using a photon-counting dual excitation microfluorometric sys- tem (Newcastle Photometrics, Newcastle, UK). Cells were placed in a chamber (0.2 ml) on the stage of an inverted microscope (Nikon Diaphot). The emitted Fura-2 fluorescence light (510 nm) was corrected on- line for background and autofluorescence, and was Sam- pled for time periods of 200 ms at a rate of 1 Hz with alternate excitation at 350 nm and 380 nm. The values were stored on a personal computer. To carry out the Fura-2 measurements, the excitation light passed a neutral-density filter (ND2) and a dichroic mirror with 455 nm cut-off, and emitted light went through a 510 nm barrier filter. [Ca2-] was assessed by converting the ratio F350/F380 using a calibration curve established in vitro. Statistical analysis was carried out by calculating mean values and standard deviations of cytosolic Ca2+ activity.

Fluorescence of BCECF was sampled with excitation alternately at 440 nm and 490 nm. For that purpose, cells were illuminated using a barrier filter (dichroic mirror with 510 nm cut-off) and a neutral density filter (ND 32). Emitted light passed a band pass filter (520 to 560 nm).

In-situ calibration of intracellular pH (pH,) was car- ried out using the H ’ ionophore nigericin to equilibrate pH, and pHuut, according to published procedures (Chail- let and Boron, 1985). Cells were loaded with BCECF/ AM then incubated in K-HBS buffer supplemented with 10 pM nigericin (37°C). K-HBS diffcrcd from regular HBS in so far as KC1 concentration was 145 mM, NaCl concentration 5 mM. The fluorescence signal and the ratio (F,,, nm/F440 ””,) were recorded from several cells, successively superfused with four different K-HBS buff- ers (pH 6.6; 7.0; 7.4; 7.8 adjusted with KOH). The se- quence of buffer changes was repeated in ascending and in descending orders of pH values. The mean values of the ratio obtained from at least five different cells plot- ted versus pH fitted a straight line (r = 0.998). A simple computer program was used to convert the ratio values of BCECF fluorescence to pH. Dye bleaching during experiments was reduced by exposing the BCECF loaded cells intermittently only to UV light by closing the shutter of the illuminating light path.

Fura-2/AM was from Chemie-Technik Munchen (CTM), Munich, Germany; UTP, and furosemide were from Sigma, Deisenhofen, Germany; BCECF/AM was

110 CZUBAYKO AND REISER

500-

400-

300 - 200 -

50 -

A 1.6 -z

t 1.2

3 0.8

; 1.4

3 1.0 0

- .I! 0.6 5

0.4

UTP (1 00 pM) I '

0 200 400 time (s)

B [Ca2'] (nM)

1 UTP(100pM) I

z 0 . z 0 100 200 300 400 500

time (s) n

400

200

100 50

.. .

0.48 8 1 . 8

7 w

0 200 400 600 800 1000 1200 time (s)

Fig. 1. Oscillations of [Ca"], induced by UTP in rat glioma cells. A Induction of [Ca"], oscillations by addition of UTP. Comparable oscillations were seen in a t least 30 cells. B: Interruption ofthe oscilla- tions duringomission of UTP, tracing is representative for three experi- ments. In this experiment the initial response is not shown because UTP has been already applied prior to the recording depicted in graph. C: Suppression of oscillations hy reduct,ion o f extracellular Ca" con- centration to 50 kM. The example shown is typical for nine expcri- ments. Here and in Figs. 2 t o 4: in single cells LCa"1 i s determined by measurement o f the ratio of Fura-2 fluorescence with excitation al 350 nm and 380 nm, as described undcr Methods. During constant superfusion of thc cclls with medium, the time period of the addition of hormoncs or drugs, or of the change of the composition of the medium is indicatcd by a horizontal bar above the traces. Simultaneous addi- tions of two substances are indicated by overlapping bars.

from Gibco, Eggenstein, Germany; and nigericin from Calbiochem, Bad Soden, Germany. EIPA was gener- ously provided by Dr. H.J. Lang, Hochst A.G., Frank- furtiMain, Germany.

RESULTS UTP-Induced [Ca2+li Oscillations

Figure 1A demonstrates that continuous addition of UTP to rat glioma cells induces a fast rise of [Ca2-li

from the resting level of 97 i 15 nM (mean 2 SD) t o a maximal peak value of 430 i 97 nM (n = 25). In most of the cells, after the peak a sustained increase is reached with a duration of 159 f 52 s. After returning to a value of 123 2 24 nM, which is still slightly above resting [Ca2+l, a series of oscillations starts which have an amplitude of 107 ? 49 nM and a frequency of 1.5 2 0.3 oscillations per min (n = 30 cells). The maxi- mal [Ca2'l, reached during the oscillations is 230 nM, which is comparable in size to the plateau value seen during the initial phase. Short pulses (12 s) ofthe nucle- otide UTP or of ATP caused a fast transient rise in [Ca2'1, in ra t glioma cells (data not shown) comparable to the large initial transient caused by continuous addi- tion of the nucleotide. The requirement of the continued presence of the stimulating hormone for maintenance of oscillations is demonstrated in Figure 1B. Addition of UTP-free medium immediately leads to a stop of the oscillations, which restart with readmission of the agonist. In most cases, the oscillations can be observed without apparent damping for a considerable period of time.

The UTP-induced [CaL+], oscillations depend on the concentration of extracellular Ca2 I . During the presence of UTP, a reduction from the usual concentration of 1.8 mM to 50 p,M leads to a ICa"1, close to the resting value and a loss of oscillatory activity (Fig. 1C). These findings are consistent with our previous experiments in which the mechanism of [Ca2' 1, oscillations caused by bradyki- nin in the glioma cells has been investigated (Reetz and Reiser, 1994b).

Since the peptide bradykinin induces [Ca"], oscilla- tions comparable t o those seen here with UTP, we stud- ied the question of whether oscillations caused by one of these hormones are affected by intermittent stimula- tion of the other receptor. In the examples shown in Figure 2A,B the experimental paradigm was long-term (ca. 300 s) addition of one or the other of the two hor- mones and then switching to both hormones. During the [Ca2 1, oscillations, supplementation with a second hormone causes a typical "initial response" consisting of a large spike exceeding in size that of the regular oscillations followed by a plateau. The subsequent oscil- lations seen in the presence of both hormones cannot be distinguished with statistical significance from those elicited by only one hormone. It is noteworthy that switching from exposure to two hormones back to one hormone mostly causes a cessation of oscillations or at least a reduction in the amplitude. This effect is reversible since renewed addition of both hormones causes a restart or a t least an increase in the amplitude of oscillations.

The possible interaction of the receptors was also tested by experiments exemplified in Figure 2C,D. Dur- ing stable oscillations caused either by UTP or by brady- kinin, brief pulses of the second agonist (12 s) are ap- plied. In every case the pulse induces a Ca" spike with an amplitude about twice as large as that of the oscilla- tions. Then the oscillations continue with the same characteristics as before.

[Ca2

+Ii 401 20

0-

100-

1.4

Bk (

10 n

M)

- OD

M

U. \

0.8-

0

v) M b 0

.6.

.C

1.21

1.4

I UT

P (1

00 p

M)

Bk (

10 nM)

Bk

- 0 E

J 1.0

50

-1

i J

i

0 200

400

600

800

time (s)

B

[Ca’

*Ii

(nM

)

500

40

0

300

20

0

1 oa

50

0

00 L? 1.0

lV21

R a

fl

A!

” . -- 0

250

500 750

1000 1250 1500

tim

e (s)

0

ln

Y

0

100

.C

50

[Ca

2+

Ii (n

M)

C

1.6

1.3

1 .o

0.7

0.4

I 1

.4

- 1.2

a

1.0

E 0

M

U. r,

200

4001

E 0.

8 0

Ln M

0

&

Q

k

0.6

cy:

0.4

.r

100

50

UTP

(1

00

pM

) L

d

B

0 300

60

0

900

tim

e (s)

Bk (10

nM

)

0 200

400

600

800

1000

tim

e (s)

Fig.

2.

Inte

ract

ion

of L

JTP

and

brad

ykin

in r

ecep

tors

dur

ing

LCa”

1, os

cilla

tions

in

rat

glio

ma

cclls

. A,

B: S

imul

tane

ous c

onti

niio

us ad

diti

on o

f bot

h ag

onis

ts, f

ollo

win

g a

long

-ter

m a

ppli

cati

on o

f eit

hcr b

rady

ki-

nin

or U

TP,

cau

ses

a si

iper

impo

sed

larg

e tr

ansi

ent r

ise

wit

h u p

late

au t

here

afte

r. R

esul

ts in

A (

B) a

re

typi

cal f

or 1

2 (7

) sim

ilar

exp

erim

ents

. C, D

: At t

he ti

mc

poin

ts m

arke

d by

arr

ows,

pul

ses

of 1

2 s

dura

tion

of

10

nM b

rady

kini

n (m

arke

d “B

”) or

of

100

JLM

CT

P (m

arke

d “U

i w

erc

appl

ied

duri

ng t

he I

Ca’+

l, os

cill

atio

ns. R

ecor

ding

A s

how

s a

segm

ent

with

out

thc

init

ial

resp

onsc

to

brad

ykin

in. R

esul

ts s

imil

ar

to t

hose

in

C (D

) wer

e se

en i

n 7

(15)

expe

rim

ents

.

112 CZUBAYKO AND REISER

Cytosolic pH and [Ca2'Ii Oscillation

The NH4CI pulse method (Roos and Boron, 1981) has been widely used to impose a fast shift in cytosolic pH. The initial diffusion of NH3 through the plasma mem- brane causes an alkalinization of pH,, whereas the entry of NH,' through cation channels has an acidifying ef- fect. Withdrawal of NH4C1 in the extracellular medium induces diffusion of NH3 from the cytosol to the outside compartment and an associated intracellular acidifica- tion. However, it is difficult to rule out the possibility that NH,+ might have effects on proteins involved in the physiological process studied, which are not due to the pH change imposed. In the glioma cells, addition of NH,Cl to the supcrfusion medium causes an immediate alkalinization (Fig. 3A). The mean change at 20 mM NH,C1 is 0.26 -C 0.12 (n = 5) pH units. There is a slow, regulatory recovery process. After withdrawal of NH4C1 an acidification of the cytosol by 0.59 ? 0.15 (n = 5 ) pH units is observed.

The influence of an inhibitor of the Na'-K'-CI- co- transport system, furosemide, was tested. The recovery from the alkaline shift which is 89 f 15% under control conditions is reduced in the presence of furosemide to 32 t- 20%. The acidification of pH, caused by removal of NH,Cl is identical in the absence and in the presence of furosemide, but the recovery from acidification is clearly reduced from 74 i 20% (n = 5 ) under control conditions to 40 t 9% (n = 5) in the presence of furose- mide. UTP itself has no influence on pH, in the rat glioma cells.

The main interest of the present study was to find out whether cytosolic pH affects the UTP-induced [Ca' 'I, oscillations. Addition of NH4C1 suppressed the [Ca2'l, oscillations temporarily for 86 f 37 s (n = 3) (data not shown). In the analogous experiment depicted in Figure 3B, simultaneous addition of furosemide and NH,Cl causes an immediate stop of the oscillations. Furose- mide by itself has no effect on the oscillation. In seven out of ten comparable experiments in the presence of NH4Cl applied in a medium supplemented with furose- mide no [Ca"], oscillations occurred during the time period of observation of 200 to 300 s. In the remaining three experiments, oscillations resumed after 197 ? 40 s, as exemplified in Figure 3C. Withdrawal of NH4C1 causes a sharp, transient rise in [Ca2+], which is not followed by oscillations any more when furosemide is present. In the absence of furosemide after returning from NH4C1 containing medium to NH4C1-free HBS the cells continue to show [Ca2'], oscillations (not shown).

A pH regulatory system present in most cells is the Na+/H- exchanger (Orlowski et al., 1992) which can be blocked specifically by the amiloride analogue EIPA (Kleyman et al., 1988). Figure 4A shows an experiment recording cytosolic pH by BCECF fluorescence. The NH4Cl pulse induces the pH, shift already described in Figure 3A. The second NH&1 pulse, which is given in the presence of 10pM EIPA, demonstrates that the back-regulation of pH, from the alkaline shift is not affected by EIPA. The acidification, however, at the end

of the NH&l pulse is exacerbated by the amiloride ana- logue, from 0.59 2 0.13 pH units (n = 6) in control ex- periments to 0.68 0.07 pH units (n = 3) after the ad- dition of EIPA.

The most pronounced effect of EIPA is seen on pH recovery from an acidification: regulation under control condition is 74 t- 19% (n = 6), yet with EIPA only 9 t 1%. The influence of 1 pM EIPA on [CaZ+], oscilla- tions seems to be moderate (Fig. 4B). The large peak at the end of NH4Cl addition shows some superimposed [Ca2+], oscillations with a frequency lower than those before the NH,C1 pulse which is comparable to control experiments testing NH,C1 pulses. The effect of EIPA on [Ca2+1, oscillations could not be determined at concen- trations of the drug above 10 pM, because the Fura-2 fluorescence signal was significantly influenced. EIPA caused an increase in fluorescence intensity at 380 nm, whereas the signal at 350 nm remained unaffected.

DISCUSSION

[Ca2+], oscillations are generated in various non-excit- able cells upon stimulation with hormones or growth factors which activate the phosphoinositide pathway (Jacob et al., 1988; Berridge, 1993). Models proposed originally for the generation of these oscillations dif- fered in the assumption as to whether or not oscillations in InsPg concentration are necessary for this process (Meyer and Stryer, 1991; Tsien and Tsien, 1990). I t has now been firmly established that the Ins(1,4,5)P1 receptor is intimately involved in the mechanism of agonist-evoked generation of [CaZ' 1, spikes in numerous cell types (Berridge, 1993). The CaLt sensitivity of the InsP, receptor is essential for such oscillations (Peter- sen et al., 1994). InsP3-evoked Ca2- release causes a modest and slow rise in [Ca2'1,, and a t a threshold level Ca2+ markedly enhances the open-state probability of the InsP3-activated Ca2+-release channels, thereby lead- ing to a dramatic rise in [Ca2+],. At a higher level of [CaLtl,, the negative feedback effect of CaL- on the InsP3 receptor becomes important and the channels close.

The physiological function of LCa2+l, oscillations has not yet been explained. CaZt spikes o r oscillations cer- tainly have advantages over regulation of [CaL+], with a graded amplitude: i) the time course of [Ca2-], oscilla- tions could selectively activate Ca2+-binding proteins with particular association and dissociation rate con- stants for CaLt (Meyer and Stryer, 1991), ii) local, short- lasting Ca2' r i se has an advantage over a global [Ca2+], rise, avoiding undesirable CaZi -dependent activation processes elsewhere in the cell, and iii) frequency-modu- lated Ca2+ oscillations are more resistant to noise than graded Ca"signa1s.

Ca2+ oscillations can frequently be sustained only for a limited period of time in the absence of extracellular Ca2'. This effect is generally explained by assuming that the role of Ca" entry is simply to recharge the intracellular stores and to replenish intracellular CaZ ' lost during the [Ca2+], oscillation. Martin and Shut-

A

7 . 5 -

7 . 0 -

6 . 5 -

n

c E

LL zs

E d LL

a c

v

.C

~ m

- furosemide (50pM)

1 . 2

300 - - E 1 .0- C

0

Isr tn

200 - 0.8-

E 0 In M 0.6-

1 0 0 - !5 0 .C

cl

5 0 -. 0 . 4

4 1 1 ' 1 I ' I ' 1 . 1

0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0 1 2 0 0

time (s)

UTP (1 00 pM)

I 1 I I I I

300

2 0 0

1 0 0

5 0

[ C a 2 + I i (nM) B

- 1.0 E

E

0 OD M LL . 0.8

0 In m

0 c m ai

!5 0.6 .C

0 . 4

UTP (100pM)

furosemide (50 pM)

0 200 4 0 0 6 0 0 800

time (s)

Fig. 3. Influence of shifts in intracellular pH and of simultaneous addition of' furosemide on [Ca2+l, oscillations in rat glioma cells. A Change in inlracellular pH by addition of NH&l (20 mM) to the superfusion medium, without or with simultaneous application of the Na'-K -C1- cotransport inhibitor furosemide (50 FM). Cytosolic pH here and in Figures 4 and 5 was determined by the ratio of BCECF fluorescence intensities at 440 nm and 490 nm. For details see text.

B, C: UTP-induced lCaZA1, oscillations are suppressed immediately upon exposure to the alkalinizing agent in the presence of 50 pM furosemide. Result demonstrated in (B) is typical for that obtained in seven out of ten experiments where the oscillations were inhibited. The example shown in C exemplifies the remaining three experiments where oscillation was resumed after a time of interruption (195 2 40 s; n - 3 )

114

7 . 0 -

6.5 -

CZUBAYKO AND REISER

0 d v

E L \

c 0 01 P L L Y

0

m E

.C d

1 0 0 -

5 0 -

7

‘ ? = 6-

5- I

I

0 300 600 9 0 0 1 2 0 0 1 5 0 0

t ime (s)

B [Ca”], (nM)

2 0 0 4 0 . 8 1 UTP (1 00 pM)

i ElPA (1 uM) 0 . 7

0.6

0 . 5

0.4 1 1

0 2 0 0 4 0 0 6 0 0 8 0 0 time (s)

Fig. 4. Influence of shift in pH, and of simultaneous addition of the arniloride analogue EIPA on [Ca’-I, oscillations in ra t glioma cells. A Comparable to Figure 3A, with the exception tha t here the inhibitor of the Na+;HL exchanger, ethylisopropylamilon’de (EIPA, 10 pM) was used. For details see text. B: Addition of NH,Cl causes an interruption of hormone-induced oscillations.

tleworth (1994) have demonstrated that in exocrine cells from the avian nasal gland, Ca2+ entry plays an essential and critical role in the oscillations by promot- ing the repetitive release of Ca2 ’ from agonist-sensi- tive stores.

A comparison was made of the characteristics of [Ca2+1, oscillations in rat glioma cells induced by the nucleotide UTP with those caused by the peptide brady- kinin (Reetz and Reiser, 1994b) or by the protease thrombin (Czubayko and Reiser, 1995). The initial phase of the Ca2+ response induced by 10 nM bradykinin has a 110 nM higher amplitude (460 nM) and a 30 s shorter duration (110 s) than the corresponding values obtained with UTP stimulation. A similar relationship was seen for the oscillations induced by thrombin. Possi- bly, the high value of [Cas+l1 more strongly activates

processes responsible for induction of oscillations to avoid a long duration of toxic [Ca2+], concentrations. The oscillations, despite being elicited by the diverse hormones, have identical frequencies and amplitudes (induction by bradykinin: 115 +- 37 nM, 1.6 -C 0.3 cy- cledmin, n = 26; by thrombin: 108 2 38 nM, 1.3 5 0.2 cycles/min, n = 12; or by UTP: 107 t 48 nM, 1.5 ? 0.3 cyclcslmin, n = 30).

[CaZ-l, oscillations in astrocytes can be induced by the neurotransmitters glutamate (Jensen and Chiu, 1991) or by serotonin (Nilson et al., 1991). We have recently studied ICa”1, oscillations in astrocytes caused by stimulation of purinergic receptors (Reetz and Re- iser, 1995). In astrocytes the time course, kinetics and frequencies of oscillations are different from those seen in the glioma cells. The pattern seems t o be a hallmark

EFFECT OF pH, ON GLIAL [Ca"], OSCILLATIONS 115

of the cell type due to differences in the specific proteins occurring in the respective cells. Glutamate-induced Ca2+-waves in astrocytes have been proposed to serve the purpose of long-range glial signaling (Cornell-Bell e t al., 1990) and also possibly of long-term plasticity (Pasti et al., 1995).

The importance of LCa"]],, in the UTP-induced oscilla- tions in rat glioma cells is shown by the fact that reduc- tion of [Ca2t],x to values of 100 FM or lower immediately suppressed the oscillations. Also the [Ca2+1, oscillations induced by the protease thrombin (Czubayko and Reiser, 1995) o r by the peptide bradykinin (Reetz and Reiser, 1994b) were inhibited by such a reduction of [Ca2+Iex. From our previous experiments we conclude that the initial peak rise of [Ca2'Ii is due to the release of Ca2+ from InsP3-activated Ca2' stores. The sustained rise of [Ca2'li and the following oscillations critically depend on Ca2' influx. However, for the maintenance of oscillations Ca2 influx is necessary to drive Ca2' release from internal stores. The same mechanism seems also to be valid for those oscillations elicited by UTP.

Experiments investigating the interference of UTP and bradykinin (Fig. 3) show that an intermittent rise of InsP3 concentration caused by addition of a second hor- mone can lead to a large transient Ca2' release. This re- sponse resembling the 'initial response' does not affect the following [Ca'' 1, oscillations. The attenuation of 0s- cillations by changing from exposure t o two hormones back to one hormone indicates that a steady elevation of InsP3 could be necessary for maintaining the oscillations.

An important finding of the study presented here is the fact that intracellular alkalinization suppresses the UTP-induced [Ca2'Ii oscillation. Moreover, furosemide delays the regulation of NH4C1-induced alkaline pH and prolongs, to the same degree, the inhibition of [Ca"li oscillations. The experiment using furosemide also gives an indication of the mechanism of pH regeneration after alkaline load; i.e., NH,+ entry. NH,' entry is re- duced by furosemide by blocking the Na'-K+-Cl- co- transport which accepts NH4* instead of K+ (Amlal et al., 1994; Watts and Good, 1994). The lack of an effect of EIPA on [CaZtli oscillations is in agreement with our finding that EIPA does not affect the NH,Cl-induced al- kalinization.

The processes responsible for pH regulation in C6 glioma cells have been studied in detail by Shrode and Putnam (1994). Shrode and Putnam report that 1 mM amiloride causes an acidifying ApH of 0.12 units compa- rable t o the value induced by EIPA in the glioma cells investigated here. The alkalinizing Na- /H ' exchange has also been seen in the present experiments, whereas HC03--dependent pathways do not play a role in our study in which cells were superfused with HC03--free HEPES-buffered medium.

The interaction of pHi and regulation of [Ca2+1, has not yet been explained satisfactorily, since there seem to be subtle differences between different cell types. Tsunoda (1990) has investigated the influence of pHi on cholccystokinin-induced lCa'+li oscillations in ra t pancreatic acinar cells. The inhibition of oscillations by

intracellular alkalinization has been attributed to an interference with Ca2 ' storage. Tsunoda has suggested that the reduced H+ concentration in the cytosol leads to H ' leakage from stores, thus causing Ca" uptake into stores. The reduction in [Ca"Ii interrupts the oscil- lations, On the other hand, high H concentration in the cytosol causes the reverse, namely a Ca2+ efflux from stores. Sauve et al. (1990) explored the influence of pH, on [Ca2'], oscillations in HeLa cells. The observed interruption of [CaZt], oscillations by alkalinization has been attributed to blockage of Ca2' uptake into stores which is necessary for replenishing the stores between the points of' maximal rise. In bovine endothelial cells, Ca2' uptake into stores was blocked by alkaliniaation (Danthuluri et al., 1990).

Also extracellular pH seems to be involved in the regulation of [Ca2+li oscillations. The augmentation of oscillations by extracellular alkalinization and the sup- pression of oscillations by extracellular acidification (not shown here) are most likely due to control of Ca2 ' entry through plasma membrane channels. The effect of pH,, is certainly not related to the control of [Ca2'Ii oscillations by pH, described above. In MDCK cells, spontaneous Ca2' oscillations which depend on Ca2- influx have a sensitivity t o pH,, (Wojnowski et al., 1994) which is opposite to that found in the glioma cells. How- ever, data obtained for bradykinin-stimulated endothe- lial cells (Thuringer et al., 1991) provide a possible ex- planation for the results seen with the rat glioma cells. In the endothelial cells, cytosolic alkalinization by NH4C1 reduced the activity of Ca2' -dependent K' chan- nels. Thus the Ca2' influx was blocked. Thuringer et al. (1991) also described that an alkaline extracellular pH maintained the channel activity, a finding consistent with the stabilization of [Ca2'li oscillations we have seen with a jump of pH,, from 7.4 to 7.8.

Further experiments will have to identify the step which is modulated by pHi in the glial cell line, either the "capacitative Ca2' entry" into the cytosol and the uptake into stores or alternatively, the release through the Imp3-activated CaZ1 channels. "Capacitative Ca2+ entry" activated by emptying Ca" stores (Putney, 1990) has been claimed to be caused by generation of a diffus- ible intracellular messenger molecule, termed Ca2' in- flux factor (Randriamampita and Tsien, 1993). The inhi- bition of [Ca' ' I , oscillations by cytosolic alkalinixation is not agonist-specific, since a comparable effect was seen with [Ca"+Ii oscillations elicited by thrombin or by bradykinin in the glioma cells (Czubayko and Reiser, unpublished observation). Understanding of the molec- ular mechanism responsible for the control of [Ca"], oscillations by cytosolic pH will give important insights into the role of pHi in cellular signal transduction.

ACKNOWLEDGMENTS

This work was supported by a project grant from Deutsche Forschungsgemeinschaft (Re 563/2-4) and by Fonds der Chemischen Industrie.

116 CZUBAYKO AND REISER

REFERENCES

Amlal, H., Paillard, M., and Bichara, M. (19941 NH, transport path- ways in cells of medullary thick ascending limb of rat kidney. </. Biol. Chenz., 269:21962-21971.

Berridge, M.J. (1993) Inositol trisphosphate and calcium signalling. Natirre, 361:315-325.

Brake. A.J., Wagenbach; M.J., and Julius, D. (1994) Sew structural motif for ligand-gated ion channels defined hy an ionotropic ATP receptor. Nature, 371:519-523.

Bruner, G. and Murphy, S. (1993) UTP activates multiple second mes- senger systems in cultured rat astrocytes. Neurosci. Lett., 162:705-108.

Chaillet, R. and Boron. W. (1985) Intracellular calibration of a pH- sensitive dye in isolated, perfused salamander proximal tubules. J . Gen. Physiol., 86:765-794.

Ciccarelli, R., Di Iorio, P., Ballerini, P., Ambrosini, G., Giuliani, P., Tiboni, G.M., and Caciagli, F. (1991) Effects of exogenous ATP and related analogues on the proliferation rate of dissociated primary cultures of rat astrocytes. J. Nei~rosri. R., 39556-566.

Cornell-Bell, A.H., Finkbeiner, S.M., Cooper, M.Y., and Smith, S.J. i 1990) Glutamate induces calcium waves in cultured astrocytes: Long-range glial signaling. Science, 247:470473.

Czubayko, U. and Reiser G. (1995) [Ca2'l, oscillations in single rat glioma cells induced by thrombin through activation of cell surface receptors. NeuroReport, 6:1249-1252.

Danthuluri, N.R., Kim, D., and Brock, T.A. (1990) Intracellular alka- linization leads to Ca2- mobilization from agonist-sensitive pools in bovine aortic endothelial cells. J. B i d . Chem., 265:19071-19076.

Dubyak, G.R. and El-Moatassim, C. (1993) Signal transduction via P,-purinergic receptors for extracellular ATP and other nucleotides. Anz. J . Physiol., 265:C577-C606.

Edwards, F.A., Gibb, A.J., and Colquhoun, D. (1992) ATP receptor- mediated synaptic currents in the central nervous system. Na- ture, 369:144-147.

Evans, R.J., Derkach, V., and Surprenant, A. 11992) ATP mediates fast synaptic transmission in mammalian neurons. Nature, 357:503-505.

Fredholm, B.G., Abbracchio, M.P., Hurnstock, G., Daly, J.W., Harden, T.K., Jacobson, K.A., Leff, P, and Williams, M. (1994)(VI.) Nomen- clature and classificat.ion of purinoceptors. Pharmacol. Rev.,

Hanson E. (1995) Astrocytes in glutamate neurotransmission. FASEB 46:143-153.

J., 350:343-350. Jacob. R., Merritt, J.E.. Hallam. T.J., and Rink. T.J. (1988) Reuetitive

spikes in cytoplasmic calcium' evoked by histamine in hum&n endo- thelial cells. Nature. 335:40-45.

.Jensen, A.M. and Chiu, S.Y. (1991) Differential intracellular calcium responses to glutamate in type 1 and type 2 cultured brain astrocytes. J. Neurosci., 11:1674-1684.

Kimelberg, H.K. (1995) Receptors on astrocytes: What possible func- tions,? Neurochem. Int., 1:2740.

Kirishuk, S., Scherer, J., Kettcnmann, H., and Vcrkhratsky, A. (1995) Activation of P,-purinoreceptors triggered Ca2' release from InsP,- sensi tive internal stores in mammalian oligodendrocytes. J. Phys- id . , 48341-57.

Kleyman, T.K. and Cragoe. E.J. 119881 Amiloride and its analogs as tools in the studv of ion transport. J. Mernbr. Biol., 105:l-21.

Lin. W-W. I 1994) Heterogeneity of nucleotide receptors in NG108-15 neuroblastoma and C6 glioma cclls for mediating phosphoinositide turnover. J. Neurochem., 62536-542.

Lustig, K.D., Shiau, A.K., Brake, A.J., and Julius D. (1993) Expression cloning of an ATP receptor from mouse neuroblastoma cells. Proc. Natl. Acad. Sci. U.S.A., 905113-5117.

Lyons. S.A.; Morell. P.: and McCarthy, K.D. (1994) Schwann cells ex- hibit P21. purinergic receptors that regulate intracellular calcium and are up-regulated by cyclic AMP analogues. J. Neurochrm., 63552-560.

Marlin, S.C. and Shuttleworth, T.J. 11994) Ca2' influx drives agonist- activated [Ca' ' I, oscillations in an exocrine cell. FEBS Lett., 35232-36.

Meyer, TI and Stryer. L. (1991) Calcium spiking. Annu. Reu. Bzophys. Chent.. 20:163-174.

Munshi, R., De Bernardi, M.A., and Brooker, G. (1993) P1,.-purinergic receptors on C6-2B rat glioma cells: Modulation of cytosolic Ca" and CAMP levels by protein kinase C. Mol. Pharmacol., 44:1185-1191.

Neary, J.T., Whittemore, S.R., Zhu, Q., and Norenberg, M.D. (1994) Synergistic activation of DNA synthesis in astrocytes by fibroblast growth factors and extracellular ATP. J. Neurochem., 63:490-494.

Nilsson, M., Hansson, E., and Ronnback, L. (1991) Heterogeneity among astroglial cells with respect to 5HT-evoked cytosolic Ca2- responses. A microspectrofluorimetric study on single cells in pri- mary culture. Life Sci., 49:1339-1350.

Orlowski, J., Kandasamy, R.A., and Shull, G.E. (1992) Molecular clon- ing of putative members ofthe Na/H exchanger gene family. J. B i d . Chem., 256:9331-9339.

Pasti, L., Pozzan, T., and Carmignoto, G. (19951 Long-lasting changes of calcium oscillations in astrocytcs: A new form of glutamate-medi- ated plasticity. J. Bid. Chem., 270:15203-15210.

Petersen, O.P., Pctersen, C.D.H., and Kasai, H. (1994) Calcium and hormone action. Annu. Reo. Physid., 56:297-319.

Pearce, B. and Langley, D. (1994) Purine- and pyrimidine-stimulated phosphoinositide breakdown and intracellular calcium mobilisation in astrocytes. Brain Res., 660:329-332.

Putney, J.W. Jr. (1990) Capacitative calcium entry revisited. Cell Cul- czum, 11:611-624.

Randriamampita, C. and Tsien, R.Y. (1993) Emptying of intracellular Ca" stores releases a novel small messenger that stimulates Cat' influx. Nature, 364:809-814.

Reetz, G. and Reiser, G. (1994a) Cross-talk of the receptors for bradyki- nin, serotonin and ATP shown by single cell Ca2- responses indicat- ing different modes of CaL- activation in a neuroblastoma x glioma hybrid cell line. J. Neurochem., 62390-897.

Reetz, G. and Reiser, G. (199413) Bradykinin-induced oscillations of cytosolic Ca"-activity and of membrane potenLia1 in rat glioma cclls. SOC. Neurosci. Abstr, 20:1350.

Reetx, G. and Reiser, G. (1995) ATP-induced oscillations of cytosolic Ca2+ activity in cultured astrocytes. Proceedings of the 23rd Got- tingen Neurobiology Conference, Vol. I1 Elsner, N. and Menzel, R., eds. Thieme Verlag, Stuttgart, pp. 698.

Reiser, G. (1992) Neuroblastoma cells. ElPctrophysiologicaZ Methods for in Vitro Sludies in Verlebrale Neurobiology. Wiley-Liss, New York, pp. 111-116.

Roos, A. and Boron, W.F. (1981) Intracellular pH. Physiol. Reo., 61996-434.

Sauve, R.; Diarra, A., Chahine, M., Simoneau, C., Morier, N., and Roy, G. (1990) Ca" oscillation induced by histamine H1 receptor stimulation in HeLa cells: Fura-2 and patch clamp. Cell Calcium, 12:165-176.

Shrode, L.1). and Putnam, R.W. (1994) Intracellular pH regulation in primary rat astrocytes and C6 glioma cells. Glia, 12:196-210.

Thuringer, D., Diarra, 4., and Sauve, R. (19911 Modulation by extracel- lular pH of'bradykinin-evoked activation of CaL--activated K' chan- nels in endothelial cells. Am. J . Physiol., 261:H656-H

Tsien, R.W. and Tsien, R.Y. (1990) Calcium channels, stores, and oscil- lations. Annu. Rev. CeZl Biol., 6:715-760.

Tsunoda, Y. (1990) Cytosolic free calcium spiking affected by intracel- M a r pH change. Exp. Cell Kes., 188:294-301.

Valera, S., Hussy, pi., Evans, R.J., Adami, N., Narth, R.A., Surprenant. A., Buell, G. (1994) A new class of ligand-gated ion channel defined by P2x receptor for extracellular ATP. Nature, 371516-519.

Watts, B.A. and Good, D.W. (1994) Effects of ammonium on intracellu- lar pH in rat medullary thick ascending limb: Mechanisms of apical membrane NII,' transport. J. Gen. Physiol.. 103:917-936.

Webb, T.E., Simon, J., Krishek, B.J., Bateson, A.N., Smart, T.G., King, B.F.; Burnstock. G.? and Barnard, E.A. 11993) Cloningandfunctional expression of a brain G-protein-coupled ATP receptor. FEBS Lett., 324:219-225.

Wojnowski, L., Mason, W.T., Schwab, A,, and Oberleithner, €I. (1994) Extracellular pH determines the rate of Ca' entry into madin-darby canine kidney-focus cells. J . Membrane Bid., 138:143-149.

Zimmermann, H. (1994) Signalling via ATP in the nervous system. Trends Neurosci., 17:420-426.