Content uploaded by Tatyana Pivneva
Author content
All content in this area was uploaded by Tatyana Pivneva on Dec 12, 2017
Content may be subject to copyright.
Cell Calcium 43 (2008) 591–601
Store-operated Ca2+ entry in astrocytes: Different spatial arrangement
of endoplasmic reticulum explains functional
diversity in vitro and in situ
Tatjyana Pivneva a,1, Brigitte Haas b,1, Daniel Reyes-Harob, Gregor Laube c,
Ruediger W. Veh c, Christiane Nolteb, Galina Skiboa, Helmut Kettenmannb,∗
aCytology Department, Bogomoletz Institute of Physiology, Bogomoletz Str. 4, 01024 Kiev, Ukraine
bMax Delbr¨uck Center for Molecular Medicine (MDC) Berlin-Buch, Robert-R¨ossle-Str. 10, 13125 Berlin, Germany
cDepartment of Electronmicroscopy and Molecular Neuroanatomy, Charite University Medicine, Schumannstraße 20/21, 10098 Berlin, Germany
Received 6 July 2007; received in revised form 19 September 2007; accepted 5 October 2007
Available online 3 December 2007
Abstract
Ca2+ signaling is the astrocyte form of excitability and the endoplasmic reticulum (ER) plays an important role as an intracellular Ca2+ store.
Since the subcellular distribution of the ER influences Ca2+ signaling, we compared the arrangement of ER in astrocytes of hippocampus
tissue and astrocytes in cell culture by electron microscopy. While the ER was usually located in close apposition to the plasma membrane in
astrocytes in situ, the ER in cultured astrocytes was close to the nuclear membrane. Activation of metabotropic receptors linked to release of
Ca2+ from ER stores triggered distinct responses in cultured and in situ astrocytes. In culture, Ca2+ signals were commonly first recorded close
to the nucleus and with a delay at peripheral regions of the cells. Store-operated Ca2+ entry (SOC) as a route to refill the Ca2+ stores could be
easily identified in cultured astrocytes as the Zn2+-sensitive component of the Ca2+ signal. In contrast, such a Zn2+-sensitive component was
not recorded in astrocytes from hippocampal slices despite of evidence for SOC. Our data indicate that both, astrocytes in situ and in vitro
express SOC necessary to refill stores, but that a SOC-related signal is not recorded in the cytoplasm of astrocytes in situ since the stores are
close to the plasma membrane and the refill does not affect cytoplasmic Ca2+ levels.
© 2007 Elsevier Ltd. All rights reserved.
Keywords: Astrocytes; Endoplasmic reticulum; Electron microscopy; Ca2+ signaling; Zn2+ ; Capacitative calcium entry
1. Introduction
The astrocyte excitability is based on Ca2+ signaling [1].
Ca2+ signals can propagate as waves over remarkably long
distances through a network of astrocytes both in culture
and in situ [2,3] or can increase locally in response to neu-
ronal activity [4]. Astrocytes express a large repertoire of
metabotropic receptors linked to Ca2+ signaling, e.g. for glu-
tamate or ATP [5]. Their activation is linked to release of
Ca2+ from intracellular stores. The endoplasmic reticulum
∗Corresponding author at: Max Delbr¨
uck Center for Molecular Medicine
(MDC) Berlin-Buch, Robert-R¨
ossle-Str. 10, 13125 Berlin, Germany.
Tel.: +49 30 9406 3325; fax: +49 30 9406 3819.
E-mail address: kettenmann@mdc-berlin.de (H. Kettenmann).
1Equal contribution.
(ER) is an important Ca2+ storing compartment, and crucially
contributes to the Ca2+ signals that influence cell activity.
Metabotropic receptor induced Ca2+ signals usually com-
prise two components, a rapid release of Ca2+ from the ER
and Ca2+influx from the extracellular space through slowly
activating store-operated channels in the plasma membrane,
also referred to as Ca2+-release activated Ca2+ (CRAC) chan-
nels. Depletion of the Ca2+ stores triggers store-operated
Ca2+ entry (SOC). This ‘capacitative’ Ca2+ entry is a ubiqui-
tous phenomenon, found in both excitable and non-excitable
cells (for review see [6]). While the currents related to
SOC have been well characterized, the molecular identity of
the channel itself and the signal that relays the Ca2+ con-
tent of the stores to the activation of SOC in the plasma
membrane, remain elusive. Several members of the tran-
sient receptor potential (TRP) family of ion channels have
0143-4160/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ceca.2007.10.004
592 T. Pivneva et al. / Cell Calcium 43 (2008) 591–601
been proposed as candidates for the ion channels [7]; for
review [6]), and one member, TRPC1, plays a role in astro-
cyte Ca2+ entry [8]. Recently, Orai1/CRACM1 have been
identified as a component central to the store-operated Ca2+
channel activation [9,10].Ca
2+-release activated Ca2+ influx
not only replenishes depleted Ca2+ stores, but also plays
a role in prolonging intracellular Ca2+ responses, thereby
regulating numerous cellular functions, such as gene tran-
scription, proliferation and cytokine release. In astrocytes, for
instance, intracellular Ca2+ levels control glutamate release
[11,12].
The activity of store-operated Ca2+ channels can be mod-
ulated by extracellular zinc [13,14], which is, besides iron,
the most abundant trace element in the body. In mouse astro-
cytes, Zn2+ can inhibit the capacitative Ca2+ influx, thereby
modulating the intracellular Ca2+ response to metabotropic
agonists [15]. In our previous work we demonstrated a
difference between the impact of Zn2+ on ligand-induced
Ca2+ responses in situ and in culture. While in cultured
astrocytes the activity of SOC is linked to a zinc-sensitive
elevated plateau phase, this zinc-sensitive component was
not recorded in astrocytes in hippocampal tissue [15]. One
possible explanation for the different behavior of cultured
and in situ cells could be different structural arrangement
of stores and Ca2+ entry channels in cultured versus in
situ astrocytes. Although there is evidence suggesting that
the ER in astrocytes consists of spatially distinct compart-
ments, at least in terms of release mechanisms [16], little is
known about the structural organization/localization of the
ER. In the present study, we therefore studied the spatial
arrangement of the ER in cultured astrocytes and astro-
cytes in the tissue and addressed the question whether a
distinct spatial organization would affect Ca2+ signaling in
astrocytes.
2. Materials and methods
2.1. Cell culture
Primary cultures of hippocampal astrocytes were prepared
as described previously [15] with some modifications. In
brief, hippocampi were dissected from brains of newborn
NMRI and eGFP/GFAP mice, and carefully freed from blood
vessels and meninges. Tissue was dissociated by trypsiniza-
tion and gentle trituration in the presence of 0.05% DNAase
(Worthington Biochem. Corp., Freehold, USA). After wash-
ing twice, cells were plated in dishes of 35 mm diameter
containing poly-l-lysine (PLL)-coated glass cover slips using
Basal Medium Eagle’s (BME)/10% fetal calf serum (FCS).
One day later, cultures were washed twice with Hank’s bal-
anced salt solution (HBSS) to remove cellular debris and
maintained for 4 days. When reaching subconfluent state,
cellular debris, microglia cells, oligodendrocytes as well as
their early precursor cells were dislodged by manual shak-
ing and removed by washing with HBSS. The purity of the
astrocytes was routinely determined by immunofluorescence
using a polyclonal antibody against glial fibrillary acidic pro-
tein (GFAP, DAKO, Hamburg, Germany), a specific astrocyte
marker. The cultures typically showed more than 90% cells
positive for GFAP.
2.2. Calcium imaging in cultured astrocytes
Calcium imaging was performed as described previously
[17] with slight modifications. Briefly, cultured astrocytes on
coverslips were dye-loaded for 30 min with 5 M Fluo-4-
acetoxymethylester (Fluo-4-AM, Invitrogen, Karlsruhe) in
HEPES buffer containing (in mM) NaCl (150), KCl (5,4),
MgCl2(1), CaCl2(2), HEPES (5) and d-glucose (10); pH
was adjusted to 7.4. Subsequently, coverslips were trans-
ferred to the stage of an upright microscope (Axioskop;
Zeiss, Oberkochen, Germany) and superfused with buffer. In
Ca2+-free solution, CaCl2was omitted, MgCl2was increased
to 2 mM. A fluorescence imaging system (Till Photonics,
M¨
unchen, Germany) was used for excitation and monitor-
ing the emitted fluorescence. Excitation wavelength was set
to 495 nm by means of a monochromator. Intracellular Ca2+
changes were detected by a cooled CCD camera (Sensi-
Cam, PCO, Kelheim, Germany) mounted to the microscope.
Images were sampled at 0.5 Hz using a software developed in
our group combined with the TIDA (HEKA, Lamprecht, Ger-
many) software. F/F0was calculated by averaging the first 15
values of the recording (F0) and dividing all values by this.
Values are given as percentage of increase from baseline. To
display responding cells, the images showing the reaction
were averaged and the background was subtracted.
High resolution calcium imaging in subcellular areas
was performed with a confocal laser scanning microscope
(Noran Odyssey XL confocal microscope Praire Tech-
nologies, WI). The scanner was mounted on the upright
microscope (Axioskop, Zeiss) equipped with a 40×magnifi-
cation, numerical aperture 0.75, water immersion objective.
A wave length of 488 nm was used for optical excitation and
the emission wave length was measured at 530nm. Images
were acquired every 30 ms (33 Hz) for a period of 40 s and
only the raising phase of the response to ATP was considered
for analysis.
To obtain the fluorescence intensity ratio, the resting flu-
orescence value was determined after averaging ten images
of the scan mode. Intensities of all subsequently recorded
images were divided by these resting values. These values
were obtained with the Image Pro Plus 5.0 software and fur-
ther analyzed with the Origin 7.0 software. Astrocyte cultures
that reacted to the application of 10 M ATP were selected
for the experiments.
2.3. Preparation of brain slices, dye-loading and
calcium recordings from astrocytes in situ
All experiments were performed according to the guide-
lines of the German animal protection law. For experiments,
T. Pivneva et al. / Cell Calcium 43 (2008) 591–601 593
10–14 days old NMRI mice were used. Slice preparation,
dye-loading with Fluo-4-AM, and image recording was per-
formed as described in Haas et al. [18]. Briefly, mice were
decapitated and the brains removed. Two hundred and fifty
micrometer thick slices were cut in ice-cold artificial cere-
brospinal fluid (ACSF) using a Vibratome (Microm HM
650 V, Microm International, Germany). ACSF contained (in
mM) 134 NaCl, 2.5 KCl, 2 CaCl2, 1.3 MgCl2, 26 NaHCO3,
1.25 K2HPO4and 10 mM glucose. By continuously gassing
the solution with carbogen (5% CO2, 95% O2), the pH
was adjusted to 7.4. To obtain calcium-free solution, CaCl2
was omitted, MgCl2was increased to 2 mM. The hippocam-
pus was recognized by its specific architecture.Slices were
stored in gassed ACSF for 45 min prior to staining. Dye-
loading with 10 M Fluo-4-AM was performed for 40 min
at room temperature (RT), followed by 10 min incubation
at 37 ◦C.
To selectively label astrocytes, slices were stained with
the red-fluorescent dye Sulforhodamine 101 (SR101; Molec-
ular Probes). We used a staining procedure modified from
Nimmerjahn et al. [23]. Acute slices were immersed in
1 ml standard ACSF containing 30 g (final concentration
50 M) SR101 for 1 min. The solution was quickly removed
and replaced with the conventional staining solution con-
taining Fluo-4-AM as described above. After the staining
period, astrocytes were labeled with both, the red-fluorescent
SR101 and the green fluorescent calcium-indicator dye
Fluo-4.
Hippocampal slices were transferred to the stage of an
upright microscope (Axioskop; Zeiss, Oberkochen, Ger-
many) and intracellular Ca2+ changes were recorded using
a cooled CCD camera as described above. Data analy-
sis and image illustration were performed as described
above.
2.4. Immunocytochemistry
Cultured astrocytes of NMRI or eGFP/GFAP trans-
genic mice were immunostained for calreticulin or Sec61
to show the distribution of endoplasmic reticulum (ER).
Astrocytes on coverslips were rinsed briefly with Dul-
becco’s phosphate buffered saline (DPBS) and then fixed
for 30 min with 2% formaldehyde at RT. Cells were per-
meabilized with 0.1% Triton X-100 in blocking solution
(5% BSA, 5% normal goat serum in 0.1 M PBS) for
30 min at RT. Cells were incubated overnight at 4 ◦C with
anti-calreticulin antibodies (Upstate Biotechnologies, Lake
Placid, USA; 1:50) or anti-Sec61 antibodies (kindly pro-
vided by Thomas Sommer, MDC, Berlin, Germany, 1:1000)
diluted in blocking solution. After extensive rinsing cells
were incubated with secondary antibodies (goat anti-rat
IgG coupled to Cy2 or goat anti-rabbit IgG Alexa 568,
respectively) for 2 h at RT. Cells were rinsed, mounted
with aqua polymount (Polysciences Inc.), and investi-
gated in a fluorescence microscope (Zeiss Axiophot, Zeiss
Germany).
2.5. Electron microscopy
2.5.1. Astrocyte culture
For electron microscopic investigation cultured astrocytes
prepared from newborn NMRI or eGFP/GFAP transgenic
mice were shortly rinsed with DPBS and then fixed for 30 min
with 2% paraformaldehyde, 1.25% glutaraldehyde in 0.1 M
phosphate buffer (PB pH 7.4). After several rinses cells were
postfixed in 1% osmium tetroxide, dehydrated in increasing
series of ethanol, transferred to propylene oxide, infiltrated
with epoxy resin (Plano, Marburg, Germany; Araldite CY
212, DDSA, DMP-30), and flat embedded. Ultrathin sections
were stained with uranyl acetate and lead citrate and studied
with a Ziess 910 electron microscope at 80 kV.
2.5.2. Preparation of tissue
Six weeks old eGFP/GFAP transgenic mice (n= 6) were
anesthetized deeply by pentobarbital and perfused intra-
cardially with saline followed by perfusion with fixative (4%
paraformaldehyde, 0.05% glutaraldehyde and 0.2% of picric
acid in 0.1 M phosphate buffer; pH 7.4) for 20 min. The brains
were removed and 500 m thick, parasagittal slices were
cut with a vibratome (Leica, VT 1000S, Germany). Freeze-
substitution and low temperature embedding in acrylic resins
were carried out as described earlier [19,20]. For cryoprotec-
tion, slices were placed into sucrose solutions (from 0.5 M to
2 M sucrose) in 0.05 M Tris–Maleat buffer. Slices were then
slammed onto copper blocks cooled in liquid N2. This was
followed by freeze-substitution with methanol and embed-
ding in Lowicryl HM 20 resins at −50 ◦C (Chemischer Werke
Lowi GMBH, Germany).
2.5.3. Postembedding immunocytochemistry
Ultrathin sections (75–90 nm thickness) from Lowicryl-
embedded hippocampal area were picked up on formvar-
coated copper grids and were incubated on drops of blocking
solution (50 mM Tris–HCl, pH 7.4, 0.3% NaCl, 10% normal
goat serum (NGS)) for 30 min. Then, sections were floated
over night at 4 ◦C on drops containing the primary antibody,
anti-green fluorescent protein antibodies (GFP; rabbit IgG
fraction, Molecular Probes, Mo Bi Tec.) diluted 1:50 in TBS,
2% NGS. Afterwards, the sections were washed and incu-
bated on drops of goat anti-rabbit IgG coupled to 12 nm
gold particles (Immunotech-Dianova; 1:30) for 2 h at room
temperature.
After extensive rinsing in PB and ultra-pure water, the sec-
tions were contrasted with saturated aqueous uranyl acetate
followed by staining with lead citrate. For negative con-
trol, primary antibodies were either omitted and slices were
incubated in blocking solution or replaced by 5% normal rab-
bit serum. Sections were investigated with a Zeiss electron
microscope (model EM 910; Carl Zeiss MicroImaging, Inc.).
To test for the specificity of immunogold labeling, 20 electron
micrographs (20,000×magnification) from hippocampal
areas were analyzed. Astrocytic versus non-astrocytic areas
were measured using NIH-image (http://rsb.info.nih.gov/nih-
594 T. Pivneva et al. / Cell Calcium 43 (2008) 591–601
Fig. 1. Localization of endoplasmic reticulum (er) in astrocytes in vitro
and astrocytes from hippocampus of eGFP/GFAP transgenic mice in situ.
Astrocytes were prepared for electron microscopic study as described.
(A) Transverse section of cultured astrocytes (in this case prepared from
eGFP/GFAP mice). Both rough and smooth ER (see arrowheads in inset) are
localized in the vicinity of the nucleus (nu). Arrow in inset point to GFAP
fibers. (B) Astrocytes in situ were labeled by eGFP followed by immunogold.
Specificity of the labeling was checked as described in Section 2. For better
image/). The density of labeling was calculated (astrocytes:
6 particles/m2, non-astrocytic areas: 2 particles/m2).
2.5.4. Quantifying the probability of ER localization
ER localization was quantified by putting a rectangular
frame defining the region of interest (ROI, 3mm by 15 mm)
on electron micrographs taken at 12,000×magnification
either from astrocytes in culture or from astrocytes in situ.
At this magnification the ROI equals 0.25 m×1.25 m.
Three reference structures were used: (1) nuclear membrane
with ER structure appearing within the same ROI as a por-
tion of the nuclear membrane; (2) plasma membrane with the
ER structure appearing in the same ROI as a portion of the
plasma membrane and, (3) cytoplasm ER structures being
within the frame with no membrane structures neither close
to nuclear not to the plasma membrane. Care was taken that
the longer dimension of the ROI was oriented in parallel to the
nuclear membrane or plasma membrane. Thus, the distance
between ER and membrane was between 0.05 and 0.125 m.
Micrographs of 41 cultured astrocytes and of 19 astrocytes
in situ were analyzed; the ROI was placed three times on
each reference structure in a given astrocyte by an unbiased
investigator. If ER and the reference structure appeared in the
same ROI, this was counted as positive event. Frequencies of
positive events were calculated for each reference structure,
and mean values determined.
We used Statistica software (v. 5, StatSoft, Tulsa, USA)
and applied the two-tailed Kolmogorov–Smirnov test to
assess the differences between samples [21].P< 0.05 was
considered statistically significant. The mean values were
normalized to the most frequently detected localization and
expressed as % of that.
3. Results
3.1. Differences in ER distribution between astrocytes in
culture and astrocytes in situ
In our previous work we described differences in Ca2+
responses of astrocytes in situ and in culture [15] and hypoth-
esized that one possible explanation for these differences in
shape of the Ca2+ signal after metabotropic receptor stimula-
tion is a different arrangement of the calcium entry and store
sites in cultured astrocytes as compared to astrocytes in situ.
To address this question, we compared the distribution of the
ER between cultured hippocampal astrocytes and astrocytes
visibility gold particles are encircled and astrocytic membrane is marked
by dotted line. (C) Probability of ER localization in different cellular com-
partments in astrocytic cells in situ and in vitro was analyzed as described.
Results were normalized to the most frequent occurrence in a certain com-
partment and are given as %. Two-tailed Kolmogorov–Smirnov test was
applied to assess the differences between samples, and P< 0.05 was con-
sidered as statistically significant. *P< 0.05; **P< 0.01. Bars: A=1m,
B= 0.25 m.
T. Pivneva et al. / Cell Calcium 43 (2008) 591–601 595
Fig. 2. Cytoplasmic expression of eGFP does not affect ER-distribution
in cultured astrocytes. (A) Immunocytochemical labeling of ER in cultured
astrocytes with calreticulin antibodies supports the observation that the ER is
preferentially arranged in the perinuclear area. (B and C) Astrocytic cultures
prepared from two different litters of eGFP/GFAP mice were immunolabeled
with Sec61 antibodies (C) to show the localization of ER. There was com-
parable ER-distribution in eGFP-positive and -negative astrocytes. Arrows
hint at eGFP-positive and -negative cell. Bars: 10 m.
in the intact hippocampus using electron microscopy (EM).
Horizontal sections of astrocyte cultures were analyzed in
the transmission EM and micrographs were taken from a
total of 41 samples showing both, nucleus and cytoplasmic
membrane on the same micrograph (Fig. 1A) at a magnifi-
cation to resolve the ER structures (12,000×). As described
in the method section, we determined the probability of the
ER being either close to the nucleus or to the plasma mem-
brane. The probability ratio between ER localization close
to the plasma versus nuclear membrane is shown in Fig. 1C.
In cultured astrocytes, we most frequently found ER cister-
nae in the areas close to the nuclear membrane, whereas the
probability to detect ER close to the plasma membrane was
only around 30% of that. The difference was found to be sta-
tistically significant (Fig. 1C). Immunocytochemistry with
antibodies to either calreticulin or Sec61 (Fig. 2), proteins
associated with endo(sarco)plasmic reticulum membranes,
confirmed our observation of preferential ER localization in
perinuclear areas. There is strong immunolabeling around the
nucleus whereas in the processes of the cells the labeling is
usually less pronounced (Fig. 2A).
To unambiguously identify the astrocytes’ cytoplasmic
compartments in the hippocampal tissue, we made use of
the eGFP/GFAP transgenic mouse model, in which astro-
cytes express the enhanced green fluorescent protein (eGFP)
under control of the GFAP promoter [22]. Astrocytic com-
partments were identified by postembedding staining with
anti-GFP antibodies. As for cultured astrocytes (see above),
the localization of ER was analyzed in micrographs taken
from 19 different astrocytes. In contrast to cultured astro-
cytes, the ER in astrocytes in situ was preferentially observed
in the vicinity of the plasma membrane while it was found
less frequently close to the nucleus (Fig. 1B). The probability
(after normalization) to detect ER close to the nucleus was
around 20% and we most frequently found ER close to the
plasma membrane (Fig. 1C). Frequencies for nuclear- ver-
sus membrane-associated ER localization were analyzed by
Kolmogorov–Smirnov test as described and turned out to be
significantly different.
To exclude the possibility that the presence of eGFP pro-
tein disturbs the ER arrangement, we analyzed astrocyte
cultures of eGFP/GFAP transgenic animals. Those cultures
normally contain both eGFP-expressing and eGFP-negative
astrocytes (Fig. 2B). Cells were counterstained with anti-
bodies specific for the ER (calreticulin, Ser61). As shown
in Fig. 2B and C, the ER is usually arranged in the perin-
uclear region both in eGFP-negative and eGFP-expressing
astrocytes. These observations also indicate that the cyto-
plasmic abundance of the eGFP protein does not affect the
spatial organization of Ca2+ stores.
3.2. Differences in Ca2+ responses in astrocytes in situ
compared to cultured astrocytes
For comparison of astrocyte Ca2+ responses in culture and
in situ, we used similar experimental paradigms to record
Ca2+ changes. While in our previous study [15] we made use
of the eGFP/GFAP transgenic mouse to identify astrocytes
and the long-wavelength excitation Ca2+ indicator X-rhod-1
(excitation 550 nm), here we used the Ca2+ indicator Fluo-
4 both in hippocampal slices and in hippocampal astrocyte
cultures obtained from NMRI mice. To identify astrocytes,
hippocampal slices were co-loaded with Fluo-4-AM and Sul-
forhodamine 101 (SR101), the latter being a dye which is
specifically taken up by astrocytes [23].Fig. 3A shows that
the cells responding to ATP by an increase in Fluo-4 fluo-
rescence in stratum radiatum of the hippocampus were also
labeled by SR101 (Fig. 2B and C). To confirm the speci-
ficity of SR101 staining in the hippocampus, we also applied
this dye to slices of eGFP/GFAP transgenic mice. Fig. 3D–F
show that virtually all eGFP-positive cells are labeled by the
red SR101 dye. We found, however, that a subpopulation
of SR101 labeled cells were not eGFP-positive. This is in
line with our previous observation that in this eGFP/GFAP
transgenic model not all astrocytes are eGFP-positive [22].
We analyzed Ca2+ responses to ATP and 2-methyl-S-ADP
in astrocytes in cultures. A brief application (30 s) of ATP
(100 M) or 2-methyl-S-ADP (1 M) triggered a biphasic
Ca2+ transient in astrocytes; a rapid increase in intracellular
596 T. Pivneva et al. / Cell Calcium 43 (2008) 591–601
Fig. 3. Fluo-4 loaded cells in slices are identified as astrocytes hippocampal slices obtained from 12 days old NMRI mice were loaded with Fluo-4 and
Sulforhodamine 101 (SR101). (A) Reacting cells at 488 nm illumination in response to the application of ATP. (B) SR101 fluorescence at 578nm (red channel).
(C) The overlay shows that most of the reacting cells are sulforhodamine positive, indicating they are astrocytes. Scale bars in (A–C) correspond to 100 m. (D)
Hippocampal slices from 12 days old eGFP/GFAP mice at 488nm illumination were loaded with SR101, scale bar 50 m. (E) Double-labeled cells (arrows)
are visible in the magnified view in the overlay. (F) Arrowheads indicate SR101 positive cells which do not express eGFP. Scale bar represents 20 m.
Ca2+ followed by a rapid decline to an elevated plateau level.
Ca2+ levels remained elevated for about 2–3 min and then
slowly returned to base level (Fig. 4B and D). The plateau
phase depended on extracellular Ca2+ as it was lacking in the
absence of extracellular Ca2+ (Fig. 4D). We never observed
differences between application of ATP (n= 5, 50–150 cells
each) or 2-methyl-S-ADP (n= 17, 50–100 cells each). Zn2+
reversibly blocked the plateau phase indicative of SOC acti-
vation. Also here, ATP (Fig. 4B, n= 5, 50–100 cells each)
and 2-methyl-S-ADP acted similarly (data not shown, n=3,
50 cells each).
To analyze the involvement of SOC in Ca2+ signals from
astrocytes of hippocampal slices, we used similar paradigms
as in culture. We compared the Ca2+ response to ATP
(100 M, n= 21 slices, 5 animals, Fig. 4F) or 2-methyl-S-
ADP (10 m, n= 3 slices, 2 animals, data not shown) in
the absence or presence of zinc (100 M) in Fluo-4 stained
cells. Fig. 4E displays the reacting cells (background was
subtracted) after ATP application. Fig. 4F gives averaged
responses from 50 cells after control application of ATP and
in the presence of Zn2+. In contrast to responses in cultured
cells, the Ca2+ signal in astrocytes in situ after stimulation
with ATP usually lacked the plateau phase while in cultured
astrocytes it usually lasted for several minutes. Consequently,
Zn2+ had no apparent effect, neither on the amplitude nor
on the time course of the Ca2+ response. We also compared
reactions to 2-methyl-S-ADP in normal versus Ca2+-free
solution similar as we did in cultured hippocampal astro-
cytes (n= 25 slices, 8 animals). Fig. 4H shows average Ca2+
traces obtained from 25 cells in hippocampus responding to
2-methyl-S-ADP in the absence and presence of extracellular
Ca2+.
We used the integral of the Ca2+ transient to further
quantify the contribution of the Zn2+-sensitive component.
Therefore the baseline was subtracted and the response was
normalized to the peak. In vitro we found that the peak
area was decreased in the presence of Zn2+ by approx. 50
%, while in situ there was no difference. This substantiates
that the ATP response lacks the Zn2+-sensitive component
in situ.
Fig. 4. Differences in Ca2+ influx upon purinergic receptor stimulation in vitro vs. in situ. (A) The left image shows the Fluo-4 fluorescence in a hippocampal
astrocyte culture. Due to background subtraction only reacting cells are displayed. Traces in (B) show the averaged response of 100 cells to application of
100 M ATP, either in the presence or in the absence of 100 MZn
2+. The delay in the onset of the ATP response is due to a slow perfusion system, not to a
delayed response after ATP application. Note that the plateau of the purinergic Ca2+ signal disappears in the presence of Zn2+ . The maxima of the peaks are
aligned to better display the differences in the delayed component. (C and D) Similar to (A and B) the averaged reaction of 150 cells to application of 1M
2-methyl-S-ADP (MeSADP) compared to Ca2+-free conditions is shown. (E) Fluo-4 fluorescence of the cells in the slice shown in (F) (after ATP application).
Due to background subtraction only reacting cells are displayed. Cells display the typical shape of astrocytes. (F) Traces from the reaction after 30s application
of 100 M ATP and after 2 min application of 100 MZn
2+ prior to the ATP application. The averaged response of 50 cells in a hippocampal slice is shown.
Note that the difference between the traces is more subtle compared to the responses obtained in culture. (G and H) Similar to (A, E and F) the average
response to 2-methyl-S-ADP (30 s, 10M) from 25 cells in a hippocampal slice is shown, in control and nominally Ca2+ -free solution. (I) The integral of the
Ca2+ transient was calculated after baseline subtraction and alignment of the respective peaks. Note that in the presence of Zn2+ the peak area is significantly
decreased in vitro by approx. 50% (P< 0.05 (paired t-test), while there is no difference in situ.
T. Pivneva et al. / Cell Calcium 43 (2008) 591–601 597
598 T. Pivneva et al. / Cell Calcium 43 (2008) 591–601
3.3. Purinergic receptor activation in cultured
astrocytes elicits a faster [Ca2+]iincrease in perinuclear
areas as compared to cell processes
Our ultrastructural studies indicated a concentration of ER
close to the nucleus in cultured astrocytes. To test whether
the Ca2+ increase due to release from internal stores would
differ within a cell, we studied Ca2+ kinetics with sub-
cellular resolution. Recordings were made in low-density
cultures to exclude that calcium signals originate from closely
apposed cells via gap junction coupling. We sampled images
of Fluo-4-fluorescence (F/F0) at 33 Hz while applying the
P2Y-specific ligand 2-Me-S-ADP (1 M). We then compared
the kinetics of the Ca2+ increase indicative of Ca2+ release
from stores close to the nucleus and in more distal regions
remote from the nucleus. The rising phase of the normal-
ized F/F0was fitted to a sigmoidal function and the time to
obtain 50% of the maximal signal (EC50) was calculated; the
EC50 for regions close to the nucleus was 0.41 ±0.04 s, and
0.60 ±0.07 s for regions in the astrocytic processes (n= 32).
In 75% (24/32) of the recorded cells the Ca2+ response
occurred more rapidly in the perinuclear regions as compared
to more distal regions (Fig. 5). Only in 19% of the cells (6
from 32) the increase in [Ca2+]iwas faster in processes than
close the nucleus (not shown). In 2 out of 32 astrocytes we
observed no significant differences between these two areas.
These results indicate that the Ca2+ release is commonly ini-
tiated close to the nucleus and subsequently spreads within
the cell.
4. Discussion
4.1. Ultrastructural differences between astrocytes in
culture and in situ
In the present study, we found significant differences in
the subcellular distribution of the ER in cultured astrocytes as
compared to astrocytes in tissue. In situ, ER is preferentially
localized close to the plasma membrane while in cultured
cells it is often concentrated around the nucleus. The high
density of ER in the perinuclear area of cultured astrocytes
has been reported by Grimaldi et al. [24]. There is ample
evidence that the geometry of a cell influences the architec-
ture of the ER, and a correlation between cell morphology,
cytoskeletal organization and regulation of Ca2+ homeostasis
has been shown for several cell types [25–27]. Astrocytes in
culture usually appear as flattened polygonal cells without
a complex process pattern and thus display a quite different
morphology as their in vivo counterparts. The morphological
phenotype of cultured astrocytes can be affected by prolonged
exposure to cAMP. This treatment results in differentiation,
namely transformation from the flat polygonal form into a
stellate, process-bearing phenotype. With regard to cell mor-
phology, the cAMP-differentiated astrocytes in vitro are very
much comparable to the astrocytes in the tissue. Interestingly,
as reported by Grimaldi et al. [24] the stellate astrocytes
showed a more uniform distribution of the ER including
within processes and close to the plasma membranes, which
is in agreement with our findings from the EM analysis.
4.2. The eGFP/GFAP-mouse is a valuable tool for
morphological studies at the EM level
We used the eGFP/GAFP transgenic animal model for
identifying astrocyte compartments in the ultrastructural
level [22,28]. In contrast to GFAP, which is present only
in the main processes [29], eGFP is spread throughout the
entire astrocytic cytoplasm including fine cellular structures
surrounding synapses. To visualize eGFP distribution on the
EM level, we used an anti-eGFP antibody. Usually, EM-
immunolabeling procedures are counterproductive for a good
ultrastructure preservation of membranes, and details such as
organelles are difficult to detect. Our postembedding pro-
tocol was sufficient to reveal (rough) ER cisternae in the
tissue slices and at the same time preserve immunogenicity
for specific binding of anti-GFP and immunogold antibodies.
This enabled us to unequivocally identify the astrocytic com-
partment and analyze the morphological arrangement of the
major Ca2+ store, the ER in relation to the cytoplasma mem-
brane. Ultrastructure details, such as cytoskeletal elements
connecting ER cisternae with plasma membrane as shown
by Lencesova et al. [30] for muscle cells or neurons were,
however, not detectable.
4.3. Functional differences between astrocytes in culture
and in situ
These structural differences were paralleled by distinct
Ca2+ responses to metabotropic receptor activation when
comparing culture versus in situ astrocytes. To make the data
between culture and in situ comparable, we used the same
types of animals, same dyes and the same ligands to stimulate
cells. Taking cells into culture, however, may trigger expres-
sion of other proteins than in situ resulting, for instance, in
different patterns of purinergic receptors accounting for the
difference in responses to application of ATP. While there is
no convincing evidence for the presence of functional P2X
receptors in astrocytes from hippocampal brain slices, P2X
receptors are expressed in culture [31]. We therefore also
used a specific agonist to P2Y1, P2Y12 and P2Y13 recep-
tors. We found, however, similar differences between culture
and slices. In cultured astrocytes, Ca2+ responses showed a
biphasic time course, a rapid transient followed by a plateau
phase. The rapid transient signal was due to release from
intrinsic Ca2+ stores while the subsequent plateau phase was
due to capacitative Ca2+, as it could be blocked by zinc [15].
This influx of extracellular Ca2+ is most likely via store-
operated Ca2+ channels, also referred to as CRAC [32].Ca
2+
entry through store-operated Ca2+ channels has been shown
for many cell types including cultured glial cells [32–34].
In contrast, astrocytes in situ lack the second, zinc-sensitive
T. Pivneva et al. / Cell Calcium 43 (2008) 591–601 599
Fig. 5. Analysis of Ca2+ transients in subcellular regions of cultured hippocampal astrocyte Astrocytes were loaded with Fluo-4-AM as described and stimulated
with 1 M 2-Me-S-ADP. The internal Ca2+ level rises faster in perinuclear regions. (A) The upper left image shows the spatial uniformity of resting Ca2+ levels
in Fluo-4 loaded cell, 1 and 2 mark the ROIs. Series of ratio images (F/F0) captured at the times indicated were colour coded indicating the level of Ca2+. (B)
Time course of Fluo-4 fluorescence in the two ROIs depicted in the upper left image of panel (A). Arrows indicate start (t= 0.0 s) and end (2.33s) of recording
interval for the false colour images given in (A). Note that the response is faster in the ROI close to the nucleus (1) than in the ROI close to the cytoplasmic
membrane (2). Right graph shows the fit of the data to a sigmoid function.
response. We have, however, shown previously by applying
different protocols that can activate SOC, that astrocytes in
situ express SOC necessary to refill the stores [15]. This indi-
cates that store-operated Ca2+ entry in astrocytes in situ does
not lead to Ca2+ increases in the cytoplasm. We assume that
this is a general principle since we found this difference also
by activating metabotropic glutamate receptors [15].
Grimaldi et al. [24] also compared Ca2+ responses in the
flat versus the cAMP-differentiated, stellate astrocytes. They
found larger responses in the stellate cells and, in addition,
they reported a more pronounced second plateau indicative
of a SOC component recorded in the soma. Although, as
discussed above, the cultured stellate astrocytes are mor-
phologically more reminiscent of astrocytes in situ they
apparently show functional differences with respect to Ca2+
signaling.
4.4. The location of ER could explain functional
differences
The location of ER in relation to the plasma membrane
could provide an explanation for these functional differences
between astrocytes in culture and in tissue. Both, astrocytes
in situ and in vitro express SOC necessary to refill stores,
but a SOC-related signal is not recorded in the cytoplasm
of astrocytes in situ since the stores are close to the plasma
membrane and the Ca2+ is taken up rapidly without affect-
ing cytoplasmic Ca2+ levels. This indicates that Ca2+ entry
channels and stores are closely associated.
Indeed a direct molecular coupling of ER and store-
operated Ca2+ channels has been suggested [35,16]: in such
“signaling microdomains”, also named PlasmaERsomes
[16], the ER membrane can come in very close contact with
the cytoplasmic membrane. Structurally, these microdomains
resemble very much the triads, a coupling between the
sarcoplasmic reticulum and transverse tubule, in the myofib-
rils of skeletal muscle cells. Microdomains/plasmaERsomes
define a restricted space in the cytosol in which Ca2+ ions can
accumulate without getting off into the bulk cytosol. In astro-
cytes, for instance, these microdomains contain clusters of
isoform specific Na+-pumps and Na+/Ca2+-exchangers [36].
Recently, Golovina [8] demonstrated, by using a membrane
tethered Ca2+ indicator to monitor rapid large changes in
[Ca2+]iclose to the plasma membrane and high resolution
immunocytochemistry, that these microdomains also contain
TRPC1-encoded store-operated calcium entry. This result
600 T. Pivneva et al. / Cell Calcium 43 (2008) 591–601
was obtained in cultured astrocytes and our data would indi-
cate that this cooperation of stores and store-operated Ca2+
entry would even be more efficient in tissue astrocytes.
We cannot exclude that other factors may account for the
differences in SOC-related Ca2+ signal: taking cells into cul-
ture drastically changes their phenotype also with regard to
expression of proteins. It is thus conceivable that astrocytes
in vitro have a higher level of expression of SOCs/CRAC
and therefore have an increased driving force for Ca2+ entry.
In turn, astrocytes in situ may extrude Ca2+ out of the cell
more efficiently, both of which would result in differences
in shape of the Ca2+ signal. In addition, mitochondria are
another huge site for Ca2+ storage [37]. The mechanisms how
mitochondria shape the profile of an intracellular Ca2+ signal
in astrocytes and if there exist differences in mitochondrial
status/physiology as well as their distribution and spatial rela-
tionship with the ER in vitro and in situ, are little understood
and remain to be determined.
4.5. Functional consequences of different ER
localization
The localization of the ER Ca2+ stores will influence all
types of intracellular Ca2+ responses. This will also be rel-
evant for Ca2+ waves, a long-range communication form of
astrocytes. Large Ca2+ waves can be easily elicited in cultured
astrocytes, while they are more difficult to elicit in intact tis-
sue and so far have not been observed in vivo under normal
conditions (for review see [38]). Two mechanisms are dis-
cussed for the propagation of these waves: (Ca2+ dependent)
released ATP activating adjacent purinergic receptors or dif-
fusion of Ca2+/IP3 through gap junctions to coupled cells.
Both mechanisms might be intensified by larger cytosolic
Ca2+ elevations and this might be an explanation for spread
over larger distances in culture.
Longer and higher elevated cytosolic Ca2+ concentrations
could also influence other Ca2+-dependent processes such as
the calcium-dependent release of gliotransmitters, e.g. gluta-
mate. The larger Ca2+ responses could facilitate the release
in cultured cells. Indeed, the extend and impact of gluta-
mate release in situ compared to in vitro is still under active
investigation (e.g. [39]).
Acknowledgements
We thank Irene Haupt for excellent technical assistance;
Dr. Thomas Sommer (MDC) for providing antibodies to
Sec61, and Dr. Anja Hoffmann (Berlin) and Alexander G.
Nikonenko (Kiev) for helpful discussions.
References
[1] A. Verkhratsky, H. Kettenmann, Calcium signalling in glial cells,
Trends Neurosci. 19 (1996) 346–352.
[2] A.H. Cornell-Bell, S.M. Finkbeiner, M.S. Cooper, S.J. Smith, Gluta-
mate induces calcium waves in cultured astrocytes: long-range glial
signaling, Science 247 (1990) 470–473.
[3] C.G. Schipke, C. Boucsein, C. Ohlemeyer, F. Kirchhoff, H. Ketten-
mann, Astrocyte Ca2+ waves trigger responses in microglial cells in
brain slices, FASEB J. 16 (2002) 255–257.
[4] J. Grosche, V. Matyash, T. Moller, A. Verkhratsky, A. Reichenbach,
H. Kettenmann, Microdomains for neuron–glia interaction: parallel
fiber signaling to Bergmann glial cells, Nat. Neurosci. 2 (1999) 139–
143.
[5] A. Verkhratsky, R.K. Orkand, H. Kettenmann, Glial calcium: home-
ostasis and signaling function, Physiol. Rev. 78 (1998) 99–141.
[6] A.B. Parekh, J.W. Putney Jr., Store-operated calcium channels, Physiol.
Rev. 85 (2005) 757–810.
[7] Y. Mori, M. Wakamori, T. Miyakawa, M. Hermosura, Y. Hara, M.
Nishida, K. Hirose, A. Mizushima, M. Kurosaki, E. Mori, K. Gotoh,
T. Okada, A. Fleig, R. Penner, M. Iino, T. Kurosaki, Transient recep-
tor potential 1 regulates capacitative Ca(2+) entry and Ca(2+) release
from endoplasmic reticulum in B lymphocytes, J. Exp. Med. 195 (2002)
673–681.
[8] V.A. Golovina, Visualization of localized store-operated calcium entry
in mouse astrocytes. Close proximity to the endoplasmic reticulum, J.
Physiol. 564 (2005) 737–749.
[9] S. Feske, Y. Gwack, M. Prakriya, S. Srikanth, S.H. Puppel, B. Tanasa,
P.G. Hogan, R.S. Lewis, M. Daly, A. Rao, A mutation in Orai1 causes
immune deficiency by abrogating CRAC channel function, Nature 441
(2006) 179–185.
[10] M. Vig, C. Peinelt, A. Beck, D.L. Koomoa, D. Rabah, M. Koblan-
Huberson, S. Kraft, H. Turner,A. Fleig, R. Penner, J.P.Kinet, CRACM1
is a plasma membrane protein essential for store-operated Ca2+ entry,
Science 312 (2006) 1220–1223.
[11] P. Bezzi, G. Carmignoto, L. Pasti, S. Vesce, D. Rossi, B.L. Rizzini,
T. Pozzan, A. Volterra, Prostaglandins stimulate calcium-dependent
glutamate release in astrocytes, Nature 391 (1998) 281–285.
[12] V. Parpura, T.A. Basarsky, F. Liu, K. Jeftinija, S. Jeftinija, P.G. Haydon,
Glutamate-mediated astrocyte-neuron signalling, Nature 369 (1994)
744–747.
[13] C.R. Jan, S.N. Wu, S.C.J. Tseng, Zn2+ increases resting cytosolic Ca2+
levels and abolishes capacitative Ca2+ entry induced by ATP in MDCK
cells, Naunyn Schmiedebergs Arch. Pharmacol. 360 (1999) 249–
255.
[14] D. Choi, J.Y. Koh, Zinc and brain injury, Annu. Rev. Neurosci. 21
(1998) 347–375.
[15] W. Kresse, I. Sekler, A. Hoffmann, O. Peters, C. Nolte, A. Moran, H.
Kettenmann, Zinc ions are endogenous modulators of neurotransmitter-
stimulated capacitative Ca2+ entry in both cultured and in situ mouse
astrocytes, Eur. J. Neurosci. 21 (2005) 1626–1634.
[16] M.P. Blaustein, V.A. Golovina, Structural complexity and functional
diversity of endoplasmic reticulum Ca(2+) stores, Trends Neurosci. 24
(2001) 602–608.
[17] S. Kirischuk, T. Moller, N. Voitenko, H. Kettenmann, A. Verkhratsky,
ATP-induced cytoplasmic calcium mobilization in Bergmann glial
cells, J. Neurosci. 15 (1995) 7861–7871.
[18] B. Haas, C.G. Schipke, O. Peters, G. Sohl, K. Willecke, H. Kettenmann,
Activity-dependent ATP-waves in the mouse neocortex are indepen-
dent from astrocytic calcium waves, Cereb. Cortex 16 (2006) 237–
246.
[19] A. Baude, Z. Nusser, E. Molnar, R.A. Mcllhinney, P. Somogyi, High
resolution immunogold localization of AMPA type glutamate recep-
tor subunits at synaptic and non-synaptic sites in rat hippocampus,
Neuroscience 69 (1995) 1031–1055.
[20] Z. Nusser, S.G. Cull-Candy, M. Farrant, Differences in synaptic
GABAa receptor number underlie variation in GABA mini amplitude,
Neuron 19 (1997) 697–709.
[21] A.G. Nikonenko, G.G. Skibo, Technique to quantify local clustering
of synaptic vesicles using single section data, Microsc. Res. Tech. 65
(2004) 287–291.
T. Pivneva et al. / Cell Calcium 43 (2008) 591–601 601
[22] C. Nolte, M. Matyash, T. Pivneva, C.G. Schipke, C. Ohlemeyer, U.K.
Hanisch, F. Kirchhoff, H. Kettenmann, G FAP promoter-controlled
eGFP-expressing transgenic mice: a tool to visualize astrocytes and
astrogliosis in living brain tissue, Glia 33 (2001) 72–86.
[23] A. Nimmerjahn, F. Kirchhoff, J.N. Kerr, F. Helmchen, Sulforhodamine
101 as a specific marker of astroglia in the neocortex in vivo, Nat.
Method 1 (2004) 31–37.
[24] M. Grimaldi, A. Favit, D.L. Alkon, cAMP-induced cytoskeleton
rearrangement increases calcium transients through the enhancement
of capacitative calcium entry, J. Biol. Chem. 274 (1999) 33557–
33564.
[25] J.A. Rosado, S.O. Sage, The actin cytoskeleton in store-mediated cal-
cium entry, J. Physiol. 526 (2000) 221–229.
[26] C.M. Ribeiro, J. Reece, J.W. Putney Jr., Role of the cytoskeleton in
calcium signaling in NIH 3T3 cells. An intact cytoskeleton is required
for agonist-induced [Ca2+]isignaling, but not for capacitative calcium
entry, J. Biol. Chem. 272 (1997) 26555–26561.
[27] J.R. Holda, L.A. Blatter, Capacitative calcium entry is inhibited in vas-
cular endothelial cells by disruption of cytoskeletal microfilaments,
FEBS Lett. 403 (1997) 191–196.
[28] R. Jabs, T. Pivneva, K. Huttmann, A. Wyczynski, C. Nolte, H. Ketten-
mann, C. Steinhauser, Synaptic transmission onto hippocampal glial
cells with hGFAP promoter activity, J. Cell. Sci. 118 (2005) 3791–
3803.
[29] E.A. Bushong, M.E. Martone, Y.Z. Jones, M.H. Ellisman, Protoplas-
mic astrocytes in CA1 stratum radiatum occupy separate anatomical
domains, J. Neurosci. 22 (2002) 183–192.
[30] L. Lencesova, A. O’Neill, W.G. Resneck, R.J. Bloch, M.P. Blaustein,
Plasma membrane–cytoskeleton–endoplasmic reticulum complexes in
neurons and astrocytes, J. Biol. Chem. 279 (2004) 2885–2893.
[31] W. Walz, G. Gimpl, C. Ohlemeyer, H. Kettenmann, Extracellular ATP-
induced currents in astrocytes: involvement of a cation channel, J.
Neurosci. Res. 38 (1994) 12–18.
[32] A.B. Parekh, R. Penner, Store depletion and calcium influx, Physiol.
Rev. 77 (1997) 901–930.
[33] J.W. Putney Jr., TRP, inositol 1,4,5-trisphosphate receptors, and
capacitative calcium entry, Proc. Natl. Acad. Sci. U.S.A. 96 (1999)
14669–14671.
[34] E.C. Toescu, T. Moller, H. Kettenmann, A. Verkhratsky, Long-term
activation of capacitative Ca2+ entry in mouse microglial cells, Neuro-
science 86 (1998) 925–935.
[35] M.J. Berridge, P. Lipp, M.D. Bootman, Signal transduction—The cal-
cium entry Pas de Deux, Science 287 (2000) 1604–1605.
[36] M. Juhaszova, M.P. Blaustein, Na+pump low and high ouabain affinity
alpha subunit isoforms are differently distributed in cells, Proc. Natl.
Acad. Sci. U.S.A. 94 (1997) 1800–1805.
[37] V.Y. Ganitkevich, The role of mitochondria in cytoplasmic Ca2+
cycling, Exp. Physiol. 88 (2003) 91–97.
[38] E. Scemes, C. Giaume, Astrocyte calcium waves: what they are and
what they do, Glia 54 (2006) 716–725.
[39] T.A. Fiacco, C. Agulhon, S.R. Taves, J. Petravicz, K.B. Casper, X.
Dong, J. Chen, K.D. McCarthy, Selective stimulation of astrocyte cal-
cium in situ does not affect neuronal excitatory synaptic activity, Neuron
54 (2007) 611–626.
