Department of Physiology & Biophysics, University of Illinois at
Chicago College of Medicine, Chicago, Illinois
 |
Introduction |
Gabapentin (GBP) is
currently used with high efficacy and a favorable side effect profile
in the treatment of refractory epilepsy (Ramsay, 1994
) and as an
analgesic agent in the treatment of diabetic neuropathy (Backonja et
al., 1998), direct nerve injury (Rosenberg et al., 1997), and
postherpetic neuralgia (Rowbotham et al., 1998). Although the analgesic
and anticonvulsant properties of GBP may be related, no consensus
exists to explain the diverse laboratory and clinical findings.
Discerning the underlying mechanism behind GBP's actions has proven
difficult. Interestingly, GBP was found to bind to the 
2/
subunit
of brain voltage-gated Ca2+ channels (VGCC) with
high affinity (Kd 38 nM)
(Suman-Chauhan et al., 1993; Gee et al., 1996).
VGCC are composed of the pore-forming, voltage-sensing 1 subunit and
the auxiliary subunits 2/ and 
. Additionally, 
subunits are
found in channels isolated from skeletal muscle (Eberst et al., 1997)
and brain (Letts et al., 1998). For more than a decade, it was believed
that the 2/ subunit was encoded by one gene (gene 1) (Ellis et
al., 1988), but recently two other genes have been identified (genes 2 and 3) (Klugbauer et al., 1999). GBP is a compound that binds to the
2/ subunit of Ca2+ channels. However, it is
not known whether GBP can attenuate Ca2+ currents
by binding to the 2/ subunit. Thus, in this study we examined the
effects of GBP on Ca2+ currents recorded from
dorsal root ganglion (DRG) cells, the mediators of pain perception. In
addition, we compared possible effects of GBP on
Ca2+ currents from skeletal and cardiac cells
since they possess fewer types of Ca2+ currents
(and thus represent a simpler system) and because no side effects in
muscle have been reported.
DRG, skeletal, and cardiac cells were isolated from newborn mice and
studied with the whole cell configuration of the patch clamp technique.
Our results demonstrate that GBP, at a clinically relevant
concentration of 10 µM (Taylor et al., 1998), significantly decreases
Ca2+ currents in DRG cells but only in the
presence of a depolarizing prepulse or with repetitive stimulation.
Additionally, we found that 50 µM GBP significantly reduces
Ca2+ currents in skeletal myotubes but does not
affect Ca2+ currents in cardiac myocytes.
Furthermore, we examined 2/ gene expression via RT-PCR and cDNA
sequencing and found that DRG, skeletal, and cardiac muscle cells
possess multiple 2/ subunit isoforms, suggesting that VGCC from
these cells may vary with respect to the effects of GBP. Thus, this
report demonstrates that, as with the initial studies, GBP does not
decrease Ca2+ currents in neuronal cells
stimulated with simple voltage steps. However, by using a prepulse or
with repetitive stimulation, we were able to record changes in
Ca2+ currents. Reduced Ca2+
currents in DRGs and skeletal muscle in the presence of a prepulse provides insight into function of the 2/ subunit in VGCC, which previously has been difficult to elucidate. Furthermore, our results suggest that the action of GBP on the 2/ subunit could contribute to its analgesic action.
| |
Materials and Methods |
The experiments conducted were approved by the Animal Care and
Use Committee of the University of Illinois at Chicago and followed the
principles set forth in the Guide for the Care and Use of
Laboratory Animals [NIH Publication 85-23 (revised 1985), National Institutes of Health, Bethesda, MD].
DRG Cultures.
DRG neurons were cultured using methods
previously outlined (García et al., 1996). In brief, DRGs were
removed from newborn mouse pups (postnatal day 0), which had been
anesthetized, decapitated, and placed in cold rodent physiological
saline: (in mM) 146 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 HEPES, 11 d-glucose, pH 7.4. Isolated ganglia were incubated for 15 min at 37°C in 1 ml of
PIPES-buffered saline: (in mM) 120 NaCl, 5 KCl, 1 CaCl2, 1 MgCl2, 25 d-glucose, 20 PIPES, pH 7.0, containing 0.1% type XI
trypsin (Sigma, St. Louis, MO), and 0.01% (w/v) DNase I (Sigma). After
incubation, tissue was rinsed three times with 1 to 2 ml of Neurobasal
media (Life Technologies, Grand Island, NY) containing B27 protein
supplement, 100 µg/ml streptomycin, and 60 µg/ml penicillin. Once
washed, ganglia were triturated with a Pasteur pipette, followed by
more vigorous trituration with a fire polished pipette. Dissociated cells were plated onto 35-mm dishes coated with
poly(L-lysine) (1 mg/ml in 0.15 M boric acid, pH
8.4) and maintained at 37°C in humidified air (95% air and 5%
CO2). Dissociated cells from a single animal were
cultured onto several dishes and allowed to incubate overnight. All DRG
cells were used in patch clamp experiments within 24 h of plating
to reduce variability between cultures.
Skeletal and Cardiac Muscle Cultures.
Primary cultures of
skeletal and cardiac muscle tissues were prepared using protocols
previously described by Beam and Knudson (1988) and Bean and Rios
(1989), respectively. Skeletal and cardiac muscle of newborn mice (at
postnatal day 0) were removed and finely minced. The small pieces of
muscle were incubated at 37°C for 30 to 45 min (skeletal muscle) or
30 min (cardiac muscle) in Ca2+-,
Mg2+-free Rodent Ringer (in mM): 155 NaCl, 5 KCl,
11 glucose, 10 HEPES, pH 7.4, containing collagenase type IA (1 mg/ml)
(Sigma). Dissociated muscle was triturated with a Pasteur pipette in
plating medium (v/v, 80% Dulbecco's modified Eagle's medium with 4.5 g/l glucose, 10% horse serum, and 10% calf serum). Large debris were
removed from the solution by filtration and centrifugation, and a
suspension of single myocytes was obtained. Cultures were maintained in
a 37°C incubator with a gas mixture of 95% air and 5%
CO2. Skeletal myotubes and cardiac myocytes were
studied at 7 to 10 days or 1 to 2 days, respectively, after initial plating.
Calcium Current Measurements.
Ca2+
currents were measured from DRG, skeletal, and cardiac muscle cells in
culture using the whole cell configuration of the patch clamp technique
(Hamill et al., 1981). Data acquisition was synchronized with pulse
generation by a PC-controlled 12-bit AD/DA Digidata 1200A converter
(Axon Instruments, Foster City, CA). Linear components of the membrane
were subtracted digitally by appropriate scaling and subtracting
control currents that do not activate ionic conductances (P/8
protocol). Membrane capacitance was measured by integrating the area
under the capacity transient before series resistance compensation. To
normalize for differences in total membrane area, current densities
were calculated by dividing total current by the membrane capacitance
of the cell and are expressed as pA/pF. Data acquisition and processing
were performed using pClamp 7.0 software (Axon Instruments). Recording
electrodes were pulled from borosilicate glass and had resistances
between 4 and 6 M
(DRGs and cardiac myocytes) or 1.6 and 2 M
(skeletal myotubes) when filled with a solution containing (in mM): 140 Cs-aspartate, 5 MgCl2, 10 Cs-EGTA, and 10 HEPES,
pH 7.4 adjusted with CsOH. To avoid channel rundown, 5 mM Mg-ATP was
added to the intracellular solution in DRG and cardiac myocyte
preparations. The extracellular solution contained (in mM): 145 tetraethylammonium chloride, 10 CaCl2, 10 HEPES, and 0.001 tetrodotoxin, pH 7.4 adjusted with CsOH. GBP was
kindly provided by Parke-Davis Research Laboratories (Warner-Lambert
Co., Ann Arbor, MI). Currents in DRGs and cardiac myocytes were
elicited with 250-ms test pulses delivered from a holding potential of
80 mV. Test pulses were applied in 10-mV increments from 60 to +60
mV. To inactivate low-voltage activated (LVA) currents, test pulses
were preceded by a 1-s prepulse to 50 mV. Following the application
of the prepulse, cells were briefly repolarized to 80 mV (cardiac and
DRG cells) or 50 mV (skeletal myotubes) for 25 ms before a series of
test depolarizations were run. The protocol for skeletal muscle cells
was similar except for the following variations: the pulse duration was
100 ms, test pulses were run from 40 to +60 mV, and a 1-s prepulse to
30 mV was applied. To study the inactivation rate of the macroscopic Ca2+ current in DRG cells, individual traces were
fitted to an exponential function. The fitting procedure started at the
peak current amplitude and finished just before the end of the test pulse.
RT-PCR and cDNA Sequencing.
Total RNA was obtained from DRG,
skeletal, and cardiac muscle cells in culture after 1, 7, and 2 days,
respectively, using a commercial kit (RNeasy Mini Kit, Qiagen,
Chatsworth, CA). The RNA eluate was reverse-transcribed using 50 U of
MuLV reverse transcriptase and reverse primers specific for 2/
genes 1, 2, and 3 (for sequences see below). PCR was performed with
AmpliTaq DNA Polymerase (PerkinElmer, Norwalk, CT) and gene specific
primers 26 to 28 nucleotides in length. The total PCR volume was 100 µl, including 20 µl of RT reaction and 2.5 units of AmpliTaq DNA
polymerase. PCR products were size-fractionated by 2% agarose gel
electrophoresis. DNA fragments of the expected lengths were
subsequently isolated after gel electrophoresis and re-amplified using
standard PCR techniques. To verify the identity of these PCR products,
individual cDNAs were sequenced using an ABI Prism big dye terminator
cycle sequencing reaction kit (PE Applied Biosystems, Norwalk, CT), AmpliTaq DNA polymerase, and the reverse or forward primer. The primers
for the 2/-1 gene were the following: (forward) GGC CGG ATC CGC
AAT TGA TCC TAA TGG and (reverse) GAA GGC TGC AGA TCA TTG CAG TAT TC.
Primers for the 2/-2 gene were the following: (forward) AAC TTC
TTC TAC ACC CGA AAG and (reverse) TTA TAG GAT GCG TTC ACC GAG. The
primer sequence for the 2/-3 gene was taken from Klugbauer et al.
(1999): (forward) GGC ACA GAT GTC CCA GTT AAA GA and (reverse) TGT ATA
GTA GTA GTC ATT GGT CAT. These primers amplified DNA products between
300 and 480 bp. Positive controls for each of the three 2/ genes
under study included a plasmid vector containing 2/-1a cDNA
(kindly provided by Drs. Klugbauer and Hofmann, Universität
München, Munich, Germany), a plasmid vector containing rat
2/-2 cDNA (kindly provided by Drs. Chu and Best, University of
Illinois at Urbana-Champaign), and brain tissue for 2/-2 and
2/-3 genes.
Statistical Analysis.
Data are expressed as mean ± standard error (S.E.) and were analyzed using analysis of variance with
repeated measures (ANOVA) or analysis of covariance (ANCOVA) where
appropriate. Significance for all data was set at p 
0.05. Statistical analysis was performed using Statistica 5.1 (StatSoft
Inc., Tulsa, OK).
| |
Results |
Calcium Currents Recorded from DRG Cells and Effect of
Gabapentin.
The mechanism of action of GBP as an analgesic is
currently unknown. GBP is unique because it specifically binds to the
2/ subunit of Ca2+ channels. Thus, our
overall aim was to explore whether GBP acting on the 2/ subunit
can modulate Ca2+ currents.
Typical Ca2+ currents recorded with 10 mM
Ca2+ in the extracellular solution from DRG cells
are shown in Fig. 1A. LVA currents (Fig.
1A, top) produced by T-type Ca2+ channels have
smaller amplitudes than do high-voltage activated (HVA) currents and
decay rapidly during the 250-ms test pulses. Traces with larger
amplitude and slower decay correspond to HVA currents (Fig. 1A,
bottom), which in DRG cells are mediated by N-, L-, P/Q-, and R-type
Ca2+ channels (Scroggs and Fox, 1992). When test
depolarizations are preceded by a prepulse to 50 mV, LVA currents are
largely inactivated, whereas HVA currents remain practically unaffected
(Fig. 1A, bottom). Figure 1B shows the current-voltage (I-V)
relationships obtained by measuring the maximum amplitude of the
currents normalized to cell capacitance for each test depolarization.
I-V curves for control DRG cells shown in Fig. 1B were obtained in the
absence and presence of a prepulse as indicated. Current density in
cells stimulated with a prepulse are significantly smaller only at
potentials 10 mV due to inactivation of the LVA current by the
prepulse.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 1.
Typical calcium currents recorded from control DRG
cells. A, LVA and HVA currents are shown in top and bottom,
respectively, in the absence (left) and presence (right) of a 1-s
prepulse to 50 mV. LVA currents, elicited with test pulses from 60
to 10 mV, were decreased, while HVA currents, elicited with test
pulses from 0 to 60 mV, remained unaffected. B, I-V relationships of
control DRG cells are demonstrated as a function of membrane potential.
Currents obtained from control cells stimulated with a prepulse ( )
(n = 37) were significantly smaller
(*p < 0.05) compared with those evoked in the
absence of a prepulse ( ) (n = 46) only at
negative membrane potentials (40 to 10 mV) due to inactivation of
the LVA current.
|
|
To test the effect of GBP on Ca2+ currents, DRG
cells were exposed to 10 µM GBP. This concentration is within the
range found in serum of human patients and is therefore clinically
relevant (Vollmer et al., 1986; McLean, 1994; Taylor et al., 1998;
Jiang and Li, 1999). Because acute perfusion of GBP did not produce significant modifications, we incubated the cells with 10 µM GBP to
allow for a better interaction of the drug with the 2/ subunit. Typical currents obtained from DRG cells treated with GBP for 60 min
are shown in Fig. 2A. GBP was also
present in the extracellular recording solution during these
experiments. Using the same protocol outlined above, we observed that,
similar to control cells, the LVA component of the macroscopic
Ca2+ current is significantly reduced in the
presence of a prepulse. However, unlike control cells, the HVA current
density is also reduced (Fig. 2A, bottom). Figure 2B shows
Ca2+ currents recorded at +20 mV from a
GBP-treated cell in the absence and in the presence of a prepulse. From
these traces, it is clearly seen that the current is smaller in the
presence of a prepulse after the cell has been exposed to GBP. The
average I-V relationship of Ca2+ currents
obtained from all cells treated with GBP is shown in Fig. 2C in the
presence and absence of a prepulse. We observed that in the presence of
a prepulse, GBP significantly reduced the HVA component of
Ca2+ currents by 25 ± 3.2%
(n = 40) (range 22-27%) in DRG cells at membrane
potentials between +10 and +50 mV. Interestingly, GBP does not cause
any significant modification of Ca2+ currents in
DRG cells in the absence of a prepulse.
View larger version (22K):
[in this window]
[in a new window]
|
Fig. 2.
Typical calcium currents recorded from 10 µM
GBP-treated DRG cells. A, LVA and HVA components are shown in top and
bottom, respectively, in the absence (left) and presence (right) of a
1-s prepulse to 50 mV. LVA currents are reduced in the presence of a
prepulse, yet unlike control cells, HVA currents are also reduced. B,
single calcium current elicited at +20 mV from GBP-treated cells,
demonstrating differences in current amplitude in the presence and
absence of a prepulse. C, I-V relationships of GBP-treated DRG cells
are demonstrated as a function of membrane potential. LVA and HVA
currents obtained from GPB-treated cells stimulated with a prepulse
() (n = 41) are significantly smaller
(*p < 0.05) compared with absence of prepulse
() (n = 44).
|
|
To compare Ca2+ currents from control and
GBP-treated DRG cells, we plotted the current-voltage relationships
obtained in the absence of a prepulse in Fig.
3A. No significant differences were detected between the groups under these conditions. However, as demonstrated in Fig. 3B, GBP causes a significant reduction of the HVA
Ca2+ current at membrane potentials between 0 and
+60 mV with the application of a prepulse. These results demonstrate
three findings. First, that the reduction in the LVA region was due
only to the presence of the prepulse because current densities were
identical in GBP and control cells at negative membrane potentials.
Second, GBP affects HVA Ca2+ channels after
incubation with the drug, a phenomenon reminiscent of the action of
ryanodine on the ryanodine receptor/Ca2+ release
channel (Rousseau et al., 1987; García et al., 1991). Third,
the action of GBP was observed only after the membrane has been
depolarized with the prepulse, a mechanism of action resembling the
effect of dihydropyridines on L-type Ca2+
channels (Bean, 1984).
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 3.
I-V relationships of control and GBP-treated DRG
cells. A, no prepulse: 10 µM GBP () (n = 43)
did not affect current densities recorded with a regular test protocol
in the absence of a depolarizing prepulse in comparison with control
cells ( ) (n = 46). B, prepulse: current
densities recorded in the presence of a 1-s prepulse to 50 mV
demonstrates the reduction of current (*p < 0.05)
with GBP treatment () (n = 41) in comparison
with control DRG cells () (n = 37). Note that
only the HVA currents were reduced by the concomitant application of
GBP and a prepulse and that no reduction of LVA currents due to GBP was
observed.
|
|
Repetitive Stimulation Potentiates Reduction of Calcium Current in
Gabapentin-Treated Cells.
GBP has achieved success in the
treatment of both pain and refractory epilepsy. In these conditions,
neurons are subjected to repetitive depolarizations and
Ca2+ channel activation. Thus, we tried to mimic
the recurrent electrical activity by stimulating DRG cells with a train
of 10 depolarizing pulses to +10 mV for 125 ms from a holding potential
of 80 mV. The interval between these test depolarizations was set at
5, 10, 15, or 30 s. Figure 4A shows
Ca2+ currents elicited with a train of stimuli at
10-s intervals in control (top) and GBP-treated (bottom) DRG cells.
Ca2+ currents from control cells were unaffected
by the train and the 1st and 10th traces essentially overlapped. In
contrast, the amplitude of Ca2+ currents recorded
from GBP-treated cells was progressively reduced by each subsequent
pulse in a train, as indicated by the arrows.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 4.
Repetitive stimulation of DRG cells. A, DRG cells
were stimulated 10 times to +10 mV for 125 ms from a holding potential
of 80 mV. Sample DRG traces using a 10-s time interval are shown in
control (top) and GBP-treated (bottom) cells. B, pulse intervals varied
between 5 and 30 s and sustained current ratio
(I10/I1) is shown as a function of interval
duration. We observed a significant decrease (*p < 0.05) in current ratio at each interval examined in GBP-treated ()
(n = 12-23) compared with control cells ( )
(n = 8-20).
|
|
To provide a more quantitative analysis with this protocol, we compared
Ca2+ current amplitude elicited with the first
(I1) and the last (I10) pulses and expressed this value as the ratio
I10/I1. Hence, a ratio of 1 indicates that the current during the first pulse and the last pulse
has the same amplitude, whereas a ratio <1 indicates that the current
during the last pulse is smaller. The
I10/I1 ratio for the
maximum Ca2+ current was plotted as a function of
interval duration in Fig. 4B. We observed that the current ratios
obtained from GBP-treated DRG cells are significantly different from
control cells at all interval durations examined (*p < 0.05 with ANCOVA). In addition, the ratios obtained with control cells
at interval durations of 5 and 10 s are smaller than the current
ratio at 30 s and reflect voltage-dependent inactivation of
Ca2+ channels. Nevertheless, GBP is still able to
further reduce current amplitude at these same intervals. This finding
supports our results obtained with a single depolarizing prepulse and
indicates that the action of GBP requires previous activation of
Ca2+ channels to reduce
Ca2+ influx.
Mechanism of Calcium Current Block by Gabapentin in DRG Cells.
Block by Ca2+ channel ligands is a well described
phenomenon in which drug binding and subsequent blockade is enhanced as
channels are shifted to the inactivated state (Bean, 1984; Sanguinetti and Kass, 1984). We hypothesized that such a mechanism may be at play
with GBP treatment since we observed a decrease of
Ca2+ currents in DRG cells only in the presence
of a depolarizing prepulse or with repetitive stimulation. Thus, to
understand further how GBP may be affecting HVA
Ca2+ channels, we performed a detailed analysis
of Ca2+ current properties under several conditions.
We first analyzed the inactivation rate
(
inactivation) of Ca2+
currents as a function of membrane potential in the absence or presence
of GBP. HVA Ca2+ currents were fit to an
exponential function. In all cases, the best fit was obtained with a
single exponential. Figure 5 shows the
values of inactivation obtained in the absence
of a prepulse (Fig. 5A, ) or when the test pulses were preceded by a
prepulse (Fig. 5B, ). The time constant of HVA
Ca2+ current inactivation was not significantly
altered with GBP treatment either in the absence (Fig. 5A, ) or
presence (Fig. 5B, ) of a prepulse. We then measured the
inactivation rate of Ca2+ currents as a function
of interval duration between pulses in repetitively stimulated cells.
Similarly to the quantification of current reduction by GBP using this
protocol (see Fig. 4B), we calculated the ratio of
inactivation for the 10th and 1st pulse values
(10/1). The ratio of inactivation is
shown in Fig. 5C. In control cells, the 10/1 ratio did not vary
significantly from a value of 1 during any time interval examined,
indicating that current inactivation was independent of interval
duration. In contrast, in GBP-treated cells, the 10/1 ratio was
slightly decreased with the 10-s time interval and the decrease became significant with the 15-s interval. A decrease in the ratio indicates that there is an enhanced rate of inactivation during the
course of successive depolarization with GBP treatment.
View larger version (41K):
[in this window]
[in a new window]
|
Fig. 5.
Time constant of inactivation () and fractional
current in DRG cells. inactivation plotted as a function
of membrane potential in DRG cells in the absence (A) and presence (B)
of a prepulse. No differences were detected between control ()
(n = 46) and GBP-treated cells ()
(n = 44) (no prepulse) and control ()
(n = 37) and GBP-treated cells ()
(n = 41) (with prepulse). C, ratio of
inactivation10/inactivation1 in control
() (n = 8-20) and GBP-treated cells ()
(n = 12-23) as a function of interval duration
during repetitive stimulation. The only observed reduction in the ratio occurred at the 15-s interval duration in GBP-treated cells. D,
fractional current demonstrated a significant decrease in current in
GBP () (n = 8) with the second pulse (in the
presence of a prepulse) and recovery of current during the third pulse
(without a prepulse). No significant differences were detected in
control cells () (n = 7) during the course of
the three pulses. *p < 0.05 denotes significant
difference compared with control cells. Fractional current plotted
relative to pulse number in a representative cell [control, open
symbols (E) and GBP-treated, closed symbols (F)]. Symbol shapes
correspond to the same interval durations for both groups: 30-s (),
15-s (), 10-s (), and 5-s () interval durations. GBP decreased
current amplitude both with successive depolarizations and as the
interval duration decreased.
|
|
An increased effectiveness of Ca2+ current block
with depolarization has been well documented for dihydropyridines
(Bean, 1984). Nitrendipine decreases cardiac L-type
Ca2+ currents more effectively when the membrane
potential is less negative, indicating that this drug preferentially
binds to the inactivated state of the channel. Accordingly, when the
membrane potential is hyperpolarized, dihydropyridine binding is less
and the block is smaller. The effect of nitrendipine on cardiac
Ca2+ channels is reminiscent of the effect of GBP
on DRG Ca2+ currents. We also observed that for
GBP to produce a reduction of Ca2+ current, we
had to depolarize the membrane potential with a prepulse. To study
further this effect, we examined Ca2+ currents
with a three-pulse protocol. The first and third pulses were single
depolarizations to +20 mV; the second pulse, also to +20 mV, was
preceded by a 1-s prepulse to 50 mV. Currents elicited during the
second and third pulses were measured and compared with the current
elicited during the first pulse. The fractional current was plotted as
a function of test pulse (i.e., first, second, and third) in Fig. 5D.
The graph demonstrates that no significant alterations in fractional
current occurred with the series of test pulses in control cells.
However, two important features were observed with GBP treatment: as
expected, relative current was decreased during the application of the
second test pulse, which included the depolarizing prepulse. In
addition, the current returned toward basal levels with the application of the third pulse, which did not include the prepulse. Such a mode of
action observed with these series of test protocols suggests that GBP
binding may be affected by the state of the HVA
Ca2+ channels. Furthermore, the recovery of
Ca2+ current during the third test pulse
indicates that a previous depolarization is required to produce the
attenuation in Ca2+ current and suggests that the
fraction of channels opened during the second pulse is smaller because
the inactivated Ca2+ channels are more sensitive
to GBP.
GBP also blocked Ca2+ currents by a use-dependent
mechanism, as demonstrated in the last two panels of Fig. 5. The
relative amplitude of Ca2+ currents recorded
during successive test pulses in a train of 10 depolarizations to +10
mV is shown in Fig. 5, E and F, for representative control and
GBP-treated cells, respectively. Currents from pulses 2 to 10 were
normalized to the amplitude of the current during the first pulse. The
different symbol shapes represent the time interval between the test
pulses for both groups. The relative amplitude of the
Ca2+ current in control conditions remained
unchanged, while it slightly decreased for time intervals 15 s (Fig.
5E). In contrast, GBP enhanced the reduction of the fractional current
both with successive depolarizations and as the interval duration
decreased (Fig. 5F). Moreover, the effect of GBP was readily noticeable
as early as the second or third pulses and it was clearly evident by
the fifth pulse. From this pulse on, the percentage reduction of the
current was smaller. This effect indicates that the build up of
Ca2+ current block by GBP from pulse to pulse is
frequency-dependent. Thus, our results indicate that GBP reduces
Ca2+ currents by binding to the inactive state of
the channel and by promoting accumulation of block during repetitive stimulation.
Calcium Currents from Skeletal and Cardiac Muscle Cells.
Skeletal and cardiac muscle cells possess only a single HVA
Ca2+ current, mediated by L-type channels.
Because 2/ subunits also form part of muscle L-type channels, we
wanted to investigate whether GBP has a similar effect in muscle
Ca2+ currents as that found in DRG cells.
Presently, the effects of GBP on Ca2+ channels in
muscle cells have not been examined. Figure
6A shows typical
Ca2+ currents recorded from a skeletal myotube
after 7 days in culture. Currents were elicited with 100-ms test pulses
from 40 to 60 mV in the absence (left) and presence (right) of a 1-s
prepulse to 30 mV. Figure 7A shows
similar traces elicited with test depolarizations from 60 to 60 mV in
a cardiac myocyte after 2 days in culture. The left panel in Fig. 7A
shows currents in the absence of a prepulse, while the right panel
illustrates currents in the presence of a prepulse to 50 mV.
Comparison of all the current traces elicited from skeletal and cardiac
muscle cells revealed that the presence of a prepulse did not cause
significant alteration of the HVA Ca2+ current.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 6.
Ca2+ currents recorded from skeletal
myotubes (SM). A, typical Ca2+ currents recorded
from a control SM after 7 days in culture. Currents were elicited with
100-ms test pulses from 40 to 60 mV in the absence (left) and
presence (right) of a 1-s prepulse to 30 mV. The presence of a
prepulse did not significantly alter HVA current in these cells. B, I-V
relationships from SM in the presence of a prepulse in 50 µM
GBP-treated () (n = 24) and control cells ()
(n = 41). We observed a significant reduction in
Ca2+ currents (*p < 0.05) when
comparing GBP and control cells in the presence of a 1-s prepulse.
|
|
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 7.
Ca2+ currents recorded from cardiac
myocytes (CM). A, typical Ca2+ currents recorded
from a control CM after 2 days in culture. Currents were elicited with
250-ms test pulses from 60 to 60 mV in the absence (left) and
presence (right) of a 1-s prepulse to 50 mV. The presence of a
prepulse did not significantly alter HVA current in these cells. B, I-V
relationships from CM in the presence of a prepulse in 50 µM
GBP-treated () (n = 21) and control cells ()
(n = 23). We did not observe significant
alterations in Ca2+ current densities when comparing GBP
and control cells in the presence of a 1-s prepulse.
|
|
The concentration of GBP used in DRG cells (10 µM) did not
significantly affect currents in muscle cells, and for this reason, we
increased the concentration to 50 µM. Muscle cells were incubated with GBP for 1 h prior to Ca2+ current
measurements. GBP was also maintained in the external recording
solution. Figures 6B (skeletal myotubes) and 7B (cardiac myocytes) show
the I-V relationships obtained in the presence of a prepulse. We
observed that GBP significantly reduces Ca2+
currents in skeletal muscle in test pulses from 10 to 30 mV in comparison with control myotubes. In contrast,
Ca2+ current amplitude in cardiac myocytes was
not significantly altered by GBP, although there was a tendency to a
decrease. Higher concentrations of GBP (up to 100 µM) did not produce
additional reductions in Ca2+ currents in
skeletal myotubes and failed to significantly alter cardiac myocytes
currents (data not shown). Our data indicate that GBP differentially
affects L-type Ca2+ channels in skeletal and
cardiac muscle cells.
Identification of 2/ Subunit Isoforms in DRG and Muscle
Cells.
HVA channels in DRG, skeletal, and cardiac muscle cells all
contain 2/ subunits. Using homogenates from adult mouse tissues, Angelotti and Hoffmann (1996) found that the 2/-1 subunit mRNA is
alternatively spliced in three regions. Splicing of the 2/ subunit gives rise to five isoforms, 2/-1a to 2/-1e. Brain expresses the "b" isoform, skeletal muscle expresses the "a"
isoform, and heart expresses isoforms "b" to "e". Recently, two
further genes encoding novel 2/ proteins (2/-2 and
2/-3) have been identified (Klugbauer et al., 1999), but their
presence in specific tissues in the mouse is not fully described.
Therefore, we examined the relationship between 2/ gene
expression and the reduction in current by GBP in DRG, skeletal, and
cardiac muscle cells using RT-PCR and cDNA sequencing.
Total RNA was isolated from cells maintained in culture for the same
length of time as cells used in Ca2+ current
measurements. The different isoforms of the 2/-1 gene could be
detected by using a single pair of amplification primers that anneals
outside the alternatively spliced regions. Specific primers for genes 2 and 3 were also used. A plasmid vector containing the 2/-1a cDNA
as well as cDNAs from genes 2 and 3 from adult mouse brain
homogenates were used as positive controls for the RT-PCR experiments.
Adult mouse brain was used as an additional control for genes 2 and 3 because it is known to express high levels of these proteins.
Furthermore, in most cases we simultaneously searched for the three
genes in cells from the same culture dish. We then isolated the PCR
products and re-amplified individual cDNAs. After re-amplification, we
sequenced each cDNA to confirm their identity. Sequences of each cDNA
were then compared with the published sequences for each gene of
interest [GenBank accession numbers U73483-U73487 (gene 1a-e),
AF169633 (gene 2), and AJ010949 (gene 3)]. Figure
8A shows the PCR products obtained for
gene 1 from skeletal myotubes (lane 2), DRG cells (lane 3), cardiac
myocytes (lane 4), and control 2/-1a cDNA (lane 5). RT-PCR
results for gene 2 are demonstrated in Fig. 8B: skeletal myotubes (lane
2), DRG cells (lane 3), cardiac myocytes (lane 4), and adult brain
homogenate (lane 5). Figure 8C shows the PCR products obtained for gene
3 from skeletal myotubes (lane 2), DRG cells (lane 3), and adult brain
homogenate (lane 4). We found that each of the cell types examined
expresses multiple isoforms of 2/ mRNA. However, the identity of
the isoforms was different in the three cell types. Additionally, the
isoforms found in neonatal cells differed from those found in adult
tissues. DRG cells express only the b isoform of 2/-1, as well as
2/-2 and 2/-3. In contrast, skeletal myotubes express
multiple 2/-1 isoforms (2/-1a, 2/-1d, and
2/-1e), 2/-2, and 2/-3. The d isoform of 2/-1 always appeared as a faint band in the gels and was found in all skeletal myotube cultures, while the a and the e isoforms were found
less frequently. Cardiac myocytes also expressed three isoforms arising
from gene 1 (2/-1c, 2/-1d, and 2/-1e) and 2/-2. However, 2/-3 was not detected in cardiac myocytes, as opposed to
DRG and skeletal myotubes. As seen from Fig. 8A, there were always more
than one 2/-1 isoform in skeletal myotubes and cardiac myocytes.
View larger version (50K):
[in this window]
[in a new window]
|
Fig. 8.
RT-PCR of skeletal, DRG, and cardiac cells. Expected
base pair sizes for PCR amplification are listed for each of the gene 1 isoforms. A, 2/-1: skeletal muscle (S) possessed isoforms
2/-1a (488 bp), 2/-1d (416 bp, appearing always as a faint
band), 2/-1e (431 bp) (lane 2); DRG cells (D) expressed
isoform 2/-1b (452 bp) (lane 3); cardiac muscle (C) expressed
three isoforms: 2/-1c (437 bp), 2/-1d (416 bp), and
2/-1e (431 bp) (lane 4). Control cDNA (2/-1a, 488 bp) is
shown in lane 5. B, 2/-2: skeletal myotube (SM), DRG, and cardiac
myocyte (CM) cells each possessed gene 2, as shown in lanes 2, 3, and
4, respectively. Control RT-PCR experiment using adult mouse brain
homogenates (B) is shown in lane 5. C, 2/-3: SM and DRG cells
possessed gene 3, as shown in lanes 2 and 3, respectively. CM cells did
not possess gene 3. Control RT-PCR experiment using adult mouse brain
homogenates (B) is shown in lane 4.
|
|
| |
Discussion |
To explain the clinical efficacy of GBP in vivo, it has been
proposed that GBP may affect ion channels and therefore modulate neuronal excitability. Among ion channels, the most likely candidates affected by GBP are Ca2+ channels since isolation
of the GBP-binding protein from brain microsomes identified the
2/ subunit as a target for this drug (Gee et al., 1996). The
partial amino-terminal region of the GBP-binding protein was identical
to the L-type Ca2+ channel 2/-1 subunit of
adult rabbit skeletal muscle (Hamilton et al., 1989). However, the
effect of GBP on neuronal Ca2+ currents is
controversial. To date, there is only one report showing a
Ca2+ current reduction with GBP of up to 34% in
rat pyramidal neocortical cells, 12% in striatum medium spiny cells,
and 10% in large globus pallidus cells (Stefani et al., 1998). In
contrast, GBP did not modify Ca2+ currents
recorded from human hippocampal granulose cells (Schumacher et al.,
1998). This apparent controversy may be explained by the fact that
tritium-labeled GBP binding is higher in the cortex and hippocampal CA1
pyramidal cell layer (Taylor et al., 1998) and because 2/-1 is
predominant in cortex and very little is present in striatum and globus
pallidus, as determined by Klugbauer et al. (1999). This indicates the
presence of an exquisite sensitivity to GBP that is dependent on
neuronal type and/or localization. Because GBP has been extensively
used as an analgesic agent, we selected DRG neurons, the mediators of
pain perception, as a model cell.
In this article, we demonstrated that 10 µM GBP reduced HVA
Ca2+ current amplitude in DRG cells. This
concentration is within the clinically effective range (10-100 µM)
found in the serum of human patients (Vollmer et al., 1986; McLean,
1994; Taylor et al., 1998; Jiang and Li, 1999). Interestingly, the
effect of GBP was noticeable only after the cells were incubated with
the drug and only after the membrane was depolarized with either a prepulse or a train of stimuli. Our data show that the effect of GBP
depends on the state of the channel and suggest that it preferentially
binds to inactivated Ca2+ channels. In addition,
our data show that the block of HVA Ca2+ currents
by GBP depends on the frequency of stimulation and that it accumulates
from pulse to pulse during repetitive stimulation. Inactivation of
Ca2+ currents during test depolarizations
(inactivation) was affected to a lesser extent
since the only significant change was observed during the 15-s interval
duration in the repetitive stimulation protocol. Taken together, we can
hypothesize that GBP reduces the macroscopic Ca2+
current by binding more effectively to the inactive state of the
channels and by a use-dependent mechanism. The action of GBP is
remarkable and novel since it exerts its effects on
Ca2+ channels through the 2/ subunit, while
all the other Ca2+ channel blockers bind to the
1 subunit.
The decrease of Ca2+ currents caused by GBP (an
average of ~25%) is comparable to the effect of other anticonvulsant
agents that modulate Ca2+ influx such as
phenytoin and carbamazepine (Schumacher et al., 1998). These findings
may explain the efficacy of GBP as an analgesic, since DRG neurons are
depolarized during pain perception, and implicates the ability of the
2/ subunit to modulate Ca2+ channels with a
subsequent reduction in neurotransmitter release. Our results also
agree with a GBP-induced reduction of high-frequency action potentials
from mouse spinal cord and neocortical neurons reported by Wamil and
McLean (1994). Trains of pulses potentiated the effect of GBP, while
hyperpolarization of the cells reversed the effect. Thus, our study and
the work of Wamil and McLean (1994) reveal a use-dependent mechanism
for the attenuation of Ca2+ currents by GBP.
These investigators additionally observed that the effect of GBP was
not immediate and the IC50 for a 10- to 60-min
incubation was 19 µM, similar to the concentration and time used in
the present study. The need for incubation in the present study and
Wamil and McLean's (1994) work demonstrate that the action of GBP
resembles the interaction of ryanodine with the ryanodine
receptor/Ca2+ release channel of the sarcoplasmic
reticulum. For example, when ryanodine is applied to cells or single
channels incorporated in lipid bilayers, the slow association rate
delays drug binding. Therefore, the preparations must be incubated in
ryanodine for several minutes to observe an effect. In fact, it has
recently been reported that maximum GBP binding to solubilized cerebral cortex membranes takes about 60 min at 30°C and more than 2 h at
4°C due to slow association and dissociation rate constants (Taylor
and Bonhaus, 2000).
We explored the effect of GBP on muscle cells because they possess a
less diverse population of Ca2+ channels than do
neuronal cells and because the GBP has no side effects on muscle. In
skeletal and cardiac muscle cells, the L-type Ca2+ channel is the only component of HVA
Ca2+ currents. Skeletal myotubes inconsistently
express an LVA Ca2+ current mediated by T-type
channels. Experiments with muscle cells demonstrated that GBP caused a
reduction only in skeletal myotubes and not in cardiac myocytes.
Similar to the effects in DRG cells, Ca2+
currents in skeletal myotubes were reduced only when the test pulse was
preceded by a depolarizing prepulse and after the cells were incubated
for 1 h in the presence of the drug. However,
Ca2+ currents from skeletal myotubes were less
sensitive to GBP than were DRGs, as a concentration 5 times higher was
required for current reduction. A possible explanation for the
differential effect of GBP is the fact that DRG, skeletal, and cardiac
muscle cells express different isoforms of the 2/ subunit, as
evidenced by our RT-PCR experiments. Comparing the presence of 2/
isoforms between neonatal and adult tissues in each of the three cell
types reveals that muscle cells are more variable at an early stage of
development than are DRG cells. The presence of 2/-3 as well as
multiple isoforms of 2/-1 derived from skeletal myotubes differs
significantly from adult skeletal muscle, which does not express
2/-3 and contains only isoform a of 2/-1. The difference in
expression of 2/ subunit gene isoforms between neonatal and adult
mouse tissues may represent developmental regulation of 2/ genes.
Furthermore, the observed differences in 2/ subunit isoform
expression between DRG, skeletal, and cardiac muscle cells may account
for the lack of effect of GBP on Ca2+ currents
recorded from cardiac myocytes. GBP may preferentially bind to and
affect different isoforms of the 2/ subunit, which then alters
Ca2+ currents. In addition to the variability in
2/ isoform expression, the three cell types used in our study
possess different 1 isoforms, the pore-forming and voltage-sensing
subunit. Neuronal cell HVA currents are comprised of multiple
Ca2+ channels mediated by N, L, P/Q, and R
channels (Scroggs and Fox, 1992), each with different pore-forming
subunits and therefore varying kinetic properties. The pore-forming
subunits in HVA channels from skeletal muscle and cardiac muscle are
each composed of a single isoform, 1S and 1C, respectively. Thus,
another possibility to explain the differential effect of GBP on
Ca2+ currents is that the pharmacological
activity of GBP in reducing Ca2+ influx may not
be strictly dependent upon the highly variable 2/ subunit
isoforms but may occur through the specific interactions of the
various 1 and 2/ subunits, which together act to modulate the
biophysical properties of these channels.
In our experiments, the predominant effect in neurons was mediated by a
~25% reduction in Ca2+ currents. Thus, our
experiments suggest that pain sensation may be modulated by
Ca2+ currents and that the analgesic effect of
GBP is possibly mediated by its action on Ca2+
channels through the 2/-1 subunit. In addition, we also
determined that depolarizing prepulses enhance GBP's actions, which is
a likely explanation for its clinical efficacy.
We thank Tord D. Alden, M.D. for help with the statistical
analysis and Kelly D. García, D.V.M., Ph.D. for comments on the manuscript.
This work was supported by a grant from the National Science
Foundation (J.G.). K.A. was partially supported by a training grant
from the National Institutes of Health (T32 DK07739).
GBP, gabapentin;
DRG, dorsal root
ganglion;
VGCC, voltage-gated calcium channels;
LVA, low-voltage
activated;
HVA, high-voltage activated;
I-V, current-voltage;
RT-PCR, reverse transcription-polymerase chain reaction;
PIPES, 1,4-piperazinediethanesulfonic acid;
ANCOVA, analysis of covariance;
inactivation, inactivation rate;
bp, base pairs.
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
