ARTICLE
Auteur(s) : Yuriko Furukawa, Nahoko Kasai,
Keiichi Torimitsu
NTT Basic Research Laboratories, NTT Corporation, Kanagawa,
Japan
It is well known that Mg2+ plays an important role in
the inhibitory function to glutamate release [1], the
Ca2+ channel [2], and the NMDA receptor [3, 4]. However,
few reports have investigated the effect of Mg2+ on
neural activity. Recent studies indicate that an Mg2+
deficient diet has significant health effects since Mg2+
is an important ion in relation to biological function.
[Ca2+]i increases when extracellular
Mg2+ is removed from the experimental solution
([Mg2+]o removal) [5]. The amount of
[Ca2+]i is 10 000 times less than that
of [Ca2+]o. However, when cells are
stimulated, [Ca2+]i exhibits a transient
increase as a result of the Ca2+ influx from outside the
cell and the Ca2+ released from inside organella, such
as the endoplasmic reticulum (ER). The change in
[Ca2+]i caused by stimulation is mainly
divided into two types. One is the Ca2+ spike that
indicates a rapid change in [Ca2+]i, and the
other is Ca2+ oscillation that indicates slow periodic
changes. Ca2+ oscillation is known as a signal induced
by fertilization that initiates the differentiation and development
of an egg, namely the beginning of life. It is also involved in
various physiological phenomena, such as hormone and digestive
enzyme release, immunity and gene expression [6-8], and neural
circuits.
Here we investigated the effect of [Mg2+]o
on neural activity and its mechanism using a 64-channel
microelectrode array (MEA) [9] and glutamate measurements performed
with an enzyme-modified-MEA multi-array sensor [10-12]. We also
investigated the difference in the effects of
[Mg2+]o on cortical and hippocampal neurons
based on [Ca2+]i measurements with a
fluorescence probe.
Materials and methods
Neuron culture
This study used primary dissociated culture cells and slice
cultures grown in a serum-free medium [13, 14]. An 18-day Wistar
rat embryo (E18) was used for the culture. Cortex or hippocampus
was removed from the rat brain, and then treated with 2.5 mg/mL
trypsin (Invitrogen Corporation, Carisbad, CA, USA) for 10 min
at 37˚C. Then cells were centrifuged at 1,000 rpm for 5 min
and triturated with a pipette. The culture was carried with a
neurobasal medium (Invitrogen) that consisted of 0.074 mg/mL
L-glutamine, 25 μM glutamate, 50 μg/mL gentamycin (Invitrogen) and
2% B27 supplement (Invitrogen). The cells were filtered with a lens
paper, and then the cell density was adjusted to 1-5 x
106 cells/mL. The cells were plated on glass bottom
culture dishes (35 mmø, uncoated, No. 0, (Matsunami Glass. Ind.,
Ltd, Osaka, Japan) pre-coated with 20 μg/mL laminin (Sigma-aldrich
Corporation, St. Louis, MO, USA) and 100 μg/mL poly-D-lysine (M.W.
70 000-150 000, Sigma), and cultured for 1-4 weeks at 37˚C under
conditions of 5% CO2 and saturated humidity.
For a slice culture, we used postnatal 8-11th (P8-11)
Wistar rat hippocampus. The hippocampus was sliced to a thickness
of 300 μm using a Vibratom (Leica, Heidelberg, Germany). The
slices were cultured on a 64ch microelectrode array (MEA) for 5-7
days at 34˚C, in 5% CO2 and with saturated humidity
after a one-hour recovery period. A medium (DMEM:HEPES = 2:1)
including penicillin and streptomycin, 20 ng/mL BDNF and NGF was
used for the culture.
Electrical activity measurement
A 64-channel MEA was used for the neural electric activity
measurement. The array was fabricated by conventional
photolithography, using quartz substrates with a sputtered layer of
150 nm indium tin oxide (ITO). After the ITO layer had been wet
etched to form the electrode patterns, the surface was passivated
with a silicon-based positive photoresist, except for the
electrodes. The array was composed of 64 electrodes arranged in an
8 x 8 square grid. The recording electrodes were 30 x 30 μm in
size and were 100-150 μm apart. The electrode impedance was
around 100 kΩ at 1 kHz. The extracellular signals detected by the
64 electrodes were amplified using a specially designed 64-channel
amplifier (bandwidth 0.1-10 kHz, NF Corp.) and stored at a rate of
3 kHz or more.
Flow cytometric membrane potential measurement
The membrane potential change was measured with flow cytometry
(cell analyzer, EPICS XL ADC, Beckman Coulter Inc. Fullerton, CA,
USA). Cells were stained with 2.5 μM DiBAC4(3) (Dojindo,
λex = 495 nm, λem= 517 nm) for 10 minutes after being stained for
one hour with 5 μM propidium iodide (PI, Sigma) for dead cell
screening [15, 16]. Suspended cells (1-5 × 106 cells/mL)
were then measured at 10 000 cells/time. Two major
characteristics can be measured with this method; one is the
forward angle light scatter (FS) and the other is fluorescence
(FL). Since the FS value changes with cell size, we could use the
former to estimate the latter. On the other hand, the dye we used –
DiBAC4(3) – can accumulate inside the cell depending on
the membrane potential. Since depolarization of the membrane
potential causes an increase in fluorescence, we could estimate the
membrane potential by measuring the fluorescence intensity change.
Spatial and temporal distribution measurement
of glutamate release
The glutamate concentration induced by
[Mg2+]o removal was measured in real time for
a rat hippocampal slice by using the enzyme-modified-MEA
multi-array sensor. Glutamate oxidase was used to detect glutamate
released from the cells. Consequently, the hydrogen peroxide
generated was finally detected as an electrical current through
horseradish peroxidase and osmium polymer (Os) mediator. This
enzyme reaction results in highly selective and sensitive glutamate
measurement. We measured the spatial distribution and the release
of glutamate at various regions of the hippocampus in real time
using this multi-array sensor.
Effect of low [Mg2+]o
on [Ca2+]i
The primary dissociated neurons from the rat cortex were grown on
glass dishes with a serum-free culture medium. The
[Ca2+]i responses were measured using 10
μg/mL fluo4-AM (Invitrogen, λex = 494 nm, λem = 516 nm) after one
hour of staining, and the [Ca2+]i change was
measured as a fluorescence intensity change by using a confocal
laser microscope (MRC1024MP, [Bio-Rad laboratories, Inc., Tokyo,
Japan] equipped with a Zeiss Axiovert 135 [Carl Zeiss Inc., Jena,
Germany], Kr/Ar laser). The stained cells were washed with HEPES
buffer (Mg 2.0 mM) before use. Low Mg2+ HEPES buffer (Mg
1.9-0 mM) was used to measure the effect of a low Mg2+
concentration.
In this study, we investigated the [Mg2+]o
response based on a [Ca2+]i oscillation
measurement. The measurement was carried out on different culture
days in vitro (DIV) to provide an understanding of the
developmental change. The oscillation frequency is defined as the
number of oscillations in 100 seconds, and a large numerical value
indicates that the oscillation number is large.
Results and discussions
Change in membrane potential caused by low
[Mg2+]o
The change in the electrical activity of rat cortical neurons was
measured by removing Mg2+ from an external solution
using the MEA (figure
1A). After the Mg2+ had been removed, the
spontaneous electrical activity increased (figure 1B) and the evoked
electrical activity indicated highly condensed activity (figure 1C).
Figure 2
shows the membrane potential change caused by a low Mg2+
concentration obtained with flow cytometry. Figure 2A shows the
fluorescent distribution of FL1 (at 525 nm) and FL3 (at
620 nm) for rat cortical neurons. Region [a] (figure 2A) shows cells
stained with 2.5 μM DiBAC4(3), and region [b] (figure 2A) shows
cells stained with 5 μM PI for dead cell screening. Figure 2B shows the change
in the fluorescence distribution induced by different
Mg2+ concentrations. [a] (figure 2B) shows density
plots at 2.0 mM [Mg2+]o and [b] (figure 2B) shows density
plots at 0.5 mM [Mg2+]o. Figure 2C shows a histogram
of total cell number vs fluorescence intensity for [a] (figure 2A). A typical
fluorescence change for the membrane potential in cortical neurons
is shown. The red line is the fluorescence intensity for 2.0 mM
[Mg2+]o and the blue line is the fluorescence
intensity for 0.5 mM [Mg2+]o. The peak
shifted to the right indicates an increase in fluorescence
intensity caused by membrane depolarization. Figure 2D indicates the
correlation between [Mg2+]o and membrane
potential. [a] and [b] (figure 2D) show the
changes in cortical and hippocampal neurons, respectively. The
membrane potential depolarized when [Mg2+]o
was decreased in both the cortical and hippocampal neurons. But the
tendency of the membrane potential increase in the cortical neurons
was different from that in the hippocampal neurons. The tendency
for the hippocampal neurons was slightly more gradual than that for
the cortical neurons.
Spatial and temporal distribution measurement
of glutamate release by [Mg2+]o
removal
Figure 3 shows
real-time glutamate release measurements at various regions of the
hippocampus obtained using the multi-array sensor. As the
Mg2+ blockade of the NMDA receptor is removed by
[Mg2+]o removal, the membrane potential is
depolarized. This induces the glutamate release. The release was
observed on dentate gyrus (DG), CA1 and CA3, but the response was
different in each region. The effects of 500 μM MK801 (TOCRIS,
Bristol UK), NMDA receptor antagonist, and 500 μM CNQX (TOCRIS),
non-NMDA receptor antagonist on glutamate release were measured.
The results indicate that MK801 inhibited the response at CA1 but
did not affect DG or CA3. However, CNQX inhibited the response at
DG and CA3 but did not affect CA1. This suggested that the receptor
distribution was different in each region of the hippocampus.
Neural activity in various regions of the hippocampus might be
controlled/modulated differently by Mg2+.
Effect of low [Mg2+]o
on [Ca2+]i
Although hardly any neural activities are detected immediately
after the culture has started, the neurons start to generate
spontaneous activity after 3-4 days of culture in vitro (DIV)
according to the MEA measurement [9, 17]. Then, the neural
activities of the individual channels of the MEA begin to
synchronize, depending on the increment of the synapse numbers of
the input and output of the cell. As we reported previously,
periodic [Ca2+]i changes were observed at
this culture stage [5]. With this as a basis, we measured the
[Ca2+]i from 5 DIV.
Figure 4
shows [Mg2+]o vs
[Ca2+]i responses depending on the culture
day. The [Ca2+]i responses were measured
using a confocal laser microscope with a fluorescent probe. We
investigated the sequential change in the
[Ca2+]i responses induced by low
[Mg2+]o (Mg 1.9-0 mM). A contour line
graph showing culture period vs [Mg2+]o and
the ratio of the number of responsive cells to the total number of
cells is shown in figure
4A. The [Ca2+]i responses, which
depended on [Mg2+]o, were detected from 5 DIV
in both cortical and hippocampal neurons, but the responses for the
two regions were different. The [Ca2+]i
oscillations exhibit a peak at 12 DIV in higher
[Mg2+]o of 1.2 mM in cortical neurons. But no
response was detected when the culture period was long. The
oscillations were observed only in low
[Mg2+]o of almost 0 mM. On the other hand,
the [Ca2+]i oscillations in the hippocampus
were the most intense for the 14 DIV when
[Mg2+]o was 1.8 mM. The oscillations were
observed in comparatively high [Mg2+]o even
if the culture period was long. Figure 4B shows the
oscillation frequency (times/100 seconds) with a bar graph. The
difference between cortical neurons and hippocampal neurons was
also noticeable in terms of oscillation frequency. The frequency
was high when [Mg2+]o was low (0 mM), and it
did not depend on the culture period in the cortical neurons. The
frequency was high even if the culture period was long and
[Mg2+]o was high, and the frequency was high
for 12-16 DIV particularly in the hippocampal neurons. Overall, the
oscillation frequency of the cortical neurons was low compared with
that of the hippocampal neurons. As the oscillation frequency
indicates the response intensity, it is suggested that hippocampal
neurons are much more sensitive to [Mg2+]o
than cortical neurons.
The [Ca2+]i responses induced by
[Mg2+]o removal were measured when the
glutamate receptor was inhibited by the addition of 500 μM MK801
(TOCRIS) and 500 μM CNQX (TOCRIS) for cortical and hippocampal
neurons (figure
5). The [Ca2+]i oscillation was
induced because the Mg2+ blockade of the NMDA receptor
is removed. The glutamate release was inhibited by MK801 or CNQX
(figure 3). The
[Ca2+]i responses induced by
[Mg2+]o removal were inhibited with the
addition of MK801, but no noticeable inhibition was observed in
CNQX in cortical neurons. On the other hand, the
[Ca2+]i responses induced by
[Mg2+]o removal were inhibited by MK801, and
greatly inhibited by CNQX in hippocampal neurons. These different
responses to [Mg2+]o in cortical and
hippocampal neurons indicate a good correlation with the receptor
distribution. This suggests that the [Mg2+]o
response might be caused by the glutamate receptor response.
However, Ca2+ channels, such as a voltage gated
Ca2+ channel (VGCC), might be involved in these
different responses. Further investigation of the difference in
[Mg2+]o response between cortical and
hippocampal neurons is required.
Conclusion
We investigated the effect of low [Mg2+]o on
neural activity using MEA-based neural activity measurement,
real-time glutamate measurement using a multi-array sensor and
optical measurement with a fluorescence probe, using a confocal
laser microscope and a flow cytometer. The spontaneous and evoked
electrical activity was increased and the glutamate release was
also increased by [Mg2+]o removal. The
glutamate release was different in different regions of the
hippocampus. The membrane potential was depolarized by
[Mg2+]o removal, and the change in the
[Mg2+]o dependent membrane potential differed
for cortical and hippocampal neurons. The effect of low
[Mg2+]o on [Ca2+]i also
differed for cortical and hippocampal neurons. The
[Ca2+]i responses were noticeably different
with respect to culture period and Mg2+ concentration.
In cortical neurons, the [Ca2+]i oscillations
were the most noticeable at 12 DIV. The responses were measured
throughout the entire culture period at low
[Mg2+]o, namely 0 mM, in cortical neurons. On
the other hand, in hippocampal neurons, the
[Ca2+]i oscillations were the most intense at
14 DIV. The responses were observed at high
[Mg2+]o, namely 1.8 mM, and for a long
culture period in hippocampal neurons. The results suggested that
hippocampal neurons are more sensitive to
[Mg2+]o than cortical neurons. The glutamate
receptor distributions in the cortex and hippocampus may be
different, because the [Ca2+]i responses
caused by removing [Mg2+]o with the addition
of MK801 or CNQX were different in cortical and hippocampal
neurons. Further investigation is required for an understanding of
the mechanism of the effect of [Mg2+]o on
neural function.
Acknowledgments
This research was supported in part by a Grant-in-Aid for
Scientific Research B20360014 from JSPS.
References
1 Smith DA, Connick JH, Stone TW. Effect of changing
extracellular levels of magnesium on spontaneous activity and
glutamate release in the mouse neocortical slice. Br J Pharmacol
1989; 97: 475-82.
2 Zhang A, Fan SH, Cheng TP, Altura BT,
Wong RK, Altura BM. Extracellular Mg2+
modulates intracellular Ca2+ in acutely isolated
hippocampal CA1 pyramidal cells of the guinea-pig. Brain Res 1996;
728: 204-8.
3 Mayer ML, Westbrook GL, Guthrie PB.
Voltage-dependent block by Mg2+ of NMDA responses in
spinal cord neurons. Nature 1984; 309: 261-3.
4 Nowak L, Bregestovski P, Ascher P,
Herbert A, Prochaintz A. Magnesium gates
glutamate-activated channels in mouse central neurons. Nature 1984;
307: 462-5.
5 Robinson HP, Torimitsu K, Jimbo Y,
Kuroda Y, Kawana A. Periodic synchronized bursting and
intracellular calcium transients elicited by low magnesium in
cultured cortical neurons. J Neurophysiol 1993; 70: 1606-16.
6 Iino M, Kasai H, Yamazawa T. Visualization of
neural control of intracellular Ca2+ concentration in
single vascular smooth muscle cells in situ. EMBO J 1994; 13:
5026-31.
7 Keizer J, Li YX, Stojilkovic S, Rinzel J.
InsP3-induced Ca2+ excitability of the
endoplasmic reticulum. Mol Biol Cell 1995; 6: 945-51.
8 Miyawaki T, Mizushima A, Hirose K,
Yamazawa T, Bezprozvanny I, Kurosaki T, et al.
Ca2+-sensor region of IP3 receptor controls
intracellular Ca2+ signaling. EMBO J 2001; 20:
1674-80.
9 Jimbo Y, Kasai N, Torimitsu K, Tateno T,
Robinson HP. A system for MEA-based multisite stimulation.
IEEE Trans Biomed Eng 2003; 50: 241-8.
10 Kasai N, Jimbo Y, Niwa O, Matsue T,
Torimitsu K. Real-time multisite observation of glutamate
release in rat hippocampal slices. Neurosci Lett 2001; 304:
112-6.
11 Kasai N, Jimbo Y, Torimitsu K. Electrochemical
monitoring of glutamate release at multiple positions in a rat
hippocampal slice. Anal Sci 2002; 18: 1325-7.
12 Torimitsu K, Furukawa Y, Kasai N. Real-time
detection of neurotransmitter release and its spatial distribution.
Nippon Yakurigaku Zasshi 2003; 121: 349-56.
13 Fukuda J, Yamaguchi K, Akimoto S, Tada Y.
NGF-dependent and -independent growth of neurites from sympathetic
ganglion cells of the aged human in a serum-free culture. Neurosci
Res 1985; 2: 460-71.
14 Bottenstein JE, Skaper SD, Varon SS,
Sato GH. Selective survival of neurons from chick embryo
sensory ganglionic dissociates utilizing serum-free supplemented
medium. Exp Cell Res 1980; 125: 183-90.
15 Lopez-Amoros R, Castel S, Comas-Riu J,
Vives-Rego J. Assessment of e. coli and salmonella viability
and starvation by confocal laser microscopy and flow cytometry
using rhodamine 123, DiBAC4(3), propidium iodide, and
CTC. Cytometry 1997; 29: 298-305.
16 Wickens HJ, Pinney RJ, Mason DJ, Gant VA.
Flow cytometric investigation of filamentation, membrane patency,
and membrane potential in escherichia coli following ciprofloxacin
exposure. Antimicrob Agents Chemother 2000; 44: 682-7.
17 Torimitsu K, Kasai N, Furukawa Y. Effect of
magnesium on neural activities in vitro. Clin Calcium 2004; 14:
26-33.
|
Version imprimable
Version PDF
