0022-3565/01/2991-343–350$3.00
THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS
Copyright © 2001 by The American Society for Pharmacology and Experimental Therapeutics
JPET 299:343–350, 2001
Vol. 299, No. 1
3937/933931
Printed in U.S.A.
Inhibition of Ca2⫹-Induced Relaxation by Oxidized Tungsten
Wires and Paratungstate
RICHARD D. BUKOSKI, SIMONEQUE SHEARIN, WILLIAM F. JACKSON, and MOHAN F. PAMARTHI
Cardiovascular Disease Research Program (R.D.B., S.S., M.F.P.), Julius L. Chambers Biomedical Biotechnology Research Institute, North
Carolina Central University, Durham, North Carolina; and Department of Biological Sciences (W.F.J.), Western Michigan University, Kalamazoo,
Michigan
Received March 8, 2001; accepted June 5, 2001
This paper is available online at http://jpet.aspetjournals.org
ABSTRACT
Recent studies of rat mesenteric arteries using a wire myograph
detected decreased Ca2⫹ and acetylcholine-induced relaxation
responses. Preliminary experiments indicated the reduced responses were associated with the tungsten wire used in the
myograph system. Compared with earlier observations, arteries
mounted on aged 28-␮m tungsten wire showed decreased
maximal Ca2⫹-induced relaxation responses of arteries precontracted with phenylephrine (91.9 ⫾ 1.5 versus 54.8 ⫾ 4.5%,
p ⬍ 0.001) and reduced sensitivity to Ca2⫹ (ED50 ⫽ 1.65 ⫾ 0.07
versus 4.58 ⫾ 0.16 mM, p ⬍ 0.001). Similar shifts were seen for
acetylcholine. When the surface of the wire was cleaned by
abrasion with fine sandpaper, both the ED50 for Ca2⫹ and
maximal relaxation significantly improved. An enhanced sensitivity to Ca2⫹ was also seen when arteries were mounted on
newly purchased 14-␮m tungsten or 14-␮m 24K gold wire with
the rank order: 14-␮m gold ⬎ 14-␮m tungsten ⬎⬎ 28-␮m aged
tungsten wire. Laser Raman spectral analysis of the aged
28-␮m tungsten wire showed that the surface was in an oxidized state that shared spectral characteristics with the paratungstate [W12O42]⫺12 anion. The effect of the paratungstate
anion on arterial relaxation was therefore tested. Paratungstate,
but not the structurally dissimilar tungstate and metatungstate
anions, significantly reduced the sensitivity and magnitude of
relaxation induced by Ca2⫹ and to a lesser extent, relaxation
induced by acetylcholine. To learn whether paratungstate inhibits relaxation through the generation of oxygen radicals, the
effect of the superoxide dismutase mimetic 4-hydroxy-2,2,6,6tetramethylpiperidine 1-oxyl (1 mM) was assessed and found to
have no effect. Since Ca2⫹-induced relaxation is inhibited by
iberiotoxin, the effect of paratungstate on K⫹ channel activity
was assessed. Paratungstate had no effect on currents through
large conductance, Ca2⫹-activated K⫹ channels in whole-cell
recordings from vascular smooth muscle cells, ruling out an
action at the BKCa channel. We conclude that: 1) surface oxidation of tungsten wire commonly used in wire myography
significantly and adversely affects vascular responses to vasodilator compounds, 2) the effect is likely mediated by the paratungstate anion, and 3) the effects of the anion are not associated with free radical generation or K⫹ channel inhibition.
The study of vascular force generation by isolated small
arteries is often performed using a wire myograph, which
enables the measurement of isometric tension and relaxation
responses. Most myographs that are used are variations of
the original system described by Bevan and Osher (1972) and
modified by Mulvany and Halpern (1977). The procedure
consists of inserting thin wires through the lumen of the
vessel and using these wires to attach the segment to “feet”
which are in turn connected to a force transducer and a
translation stage micrometer that permit measurement of
tension and stretching of the segment, respectively.
Over the past several years, our laboratory has used a wire
myograph system to demonstrate that cumulative addition of
extracellular Ca2⫹ to preconstricted mesenteric branch ar-
teries causes dose-dependent, sensory-nerve-dependent relaxation (Bukoski et al., 1997). The sensitivity of this relaxation event for extracellular Ca2⫹, as reflected by the ED50
value for Ca2⫹, lies between 1.5 and 2.0 mM (Mupanomunda
et al., 1999). Over the course of ongoing work, we found that
both the magnitude and the Ca2⫹ sensitivity of the Ca2⫹induced relaxation response became greatly reduced. We
therefore initiated a series of experiments to determine the
underlying cause. The results indicate that a surface oxidation product on the tungsten wires that are used to mount the
vessels on the myograph significantly impairs relaxation of
isolated arterial segments.
This work was supported by Grants HL 54901, HL59868, and HL64761 to
R.D.B. and HL 32469 to W.F.J. from the National Institutes of Health.
Animals. All procedures involving animals were performed in
accordance with approval of the Institutional Animal Care and Use
Experimental Procedures
ABBREVIATIONS: DS, dissociation solution; PSS, physiologic salt solution; PE, phenylephrine; VIS, visible; TEA, tetraethylammonium; BKCa,
large-conductance Ca2⫹-activated K⫹ channel; 4-OH-TEMPO, 4-hydroxy-2,2,6,6-tetramethylpiperidine 1-oxyl; ANOVA, analysis of variance.
343
344
Bukoski et al.
Committee at both universities. Male Wistar rats (8 –10 weeks of
age) were obtained from Harlan Sprague-Dawley (Indianapolis, IN)
and housed in the North Carolina Central University animal care
facility in colony rooms with fixed light/dark cycles and constant
temperature and humidity and provided with Purina rodent chow
and water ad libitum (Purina, St. Louis, MO). Mesenteric tissue was
isolated while the rats were anesthetized with a mixture of ketamine
and xylazine (100:5, mg/kg). Male golden Syrian hamsters (9 –12
weeks of age) were obtained from Charles Rivers Laboratories (Wilmington, MA) and maintained at Western Michigan University under conditions similar to the rats. Hamsters were killed by CO2
asphyxiation followed by cervical dislocation, their cremaster muscles removed and second and third order arterioles isolated by hand
dissection as described previously (Jackson, 2000).
Isolation of Vascular Smooth Muscle Cells. Vascular smooth
muscle cells were isolated enzymatically from hamster cremasteric
arterioles as described previously (Jackson, 2000). Arteriolar segments were placed into 1 ml of dissociation solution (DS, in mmol/l:
140 NaCl, 5 KCl, 1 MgCl2, 0.1 mM CaCl2, 10 HEPES, 10 glucose, 1
mg/ml bovine serum albumin, 10 ␮M sodium nitroprusside, and 10
␮M diltiazem, pH 7.4, with NaOH, 295–300 mOsm) at room temperature. After 10 min of incubation, most of this solution was removed
and replaced with 1 ml of DS containing 1.5 mg/ml papain and 1
mg/ml dithioerythritol and the arteriolar segments incubated at
37°C for 35 min. The papain solution was then removed and replaced
with 1 ml of DS containing 1.5 mg/ml collagenase, 1 mg/ml elastase,
and 1 mg/ml soybean trypsin inhibitor, and the segments were incubated for an additional 16 to 19 min at 37°C. The enzyme-containing solution was then replaced with 4 ml of DS at room temperature
and allowed to settle for approximately 10 min. This solution was
then removed and replaced with 1 ml of fresh DS not containing
sodium nitroprusside or diltiazem. Cells were released from the
segments by gentle trituration (1– 4 strokes) using a 100 to 1000 ␮l
Eppendorf style pipettor and stored in this solution for up to 4 h at
room temperature.
Biophysical Measurements. Isometric force generation by rat
mesenteric branch arteries was measured using previously described
methods (Bukoski et al., 1997). The small intestine and the attending mesenterium was pinned to a dissecting dish filled with ice-cold
physiologic salt solution (PSS) of the following composition (in
mmol/l) 150 NaCl, 5.4 KCl, 1.17 MgSO47H2O, 1.18 NaH2PO4, 6.0
NaHCO3, 1.0 CaCl2, 20 HEPES, and 5.5 glucose, pH 7.4. Branch II
and III segments were isolated by microdissection taking special care
to leave a portion of the omental membrane attached to the adventitial surface of the blood vessel. Following these preparative procedures, vessels were mounted on a wire myograph (Kent Scientific,
Litchfield, CT) using the indicated wires and immersed in PSS
warmed to 37°C and gassed with 95% air/5% CO2. After a 15-min
equilibration period, the segments were stretched to a predetermined length that was equivalent to an internal diameter of 200 to
225 ␮m and allowed to equilibrate for an additional 15 min. After the
equilibration period, the vessels were contracted with 5 ␮M phenylephrine until reproducible contractile responses were obtained
(three to four times).
Relaxation to specific compounds was assessed by cumulatively
adding the agent to vessels that were precontracted to an average of
62% of the maximal response with 5 ␮M phenylephrine (PE). The
percentage of relaxation was calculated taking the magnitude of the
prerelaxation tone as 100% of the amount that the vessel could relax.
When the effect of a specific compound, i.e., [(NH4)10W12O41], on the
relaxation responses was assessed, the vessel was pretreated with
the compound for 10 min, following which contraction was induced
by the addition of PE and the response to the dilator assessed.
Laser Raman Spectral Analysis. Raman spectra of tungsten
wire samples was performed by NAMAR Scientific (Russellville, AR)
and collected using a maximum power of 2 mW of 514.5-nm radiation
from an Omnichrome model 150 argon-ion laser for Raman excitation. Radiation from the sample was collected in a 180° back-scat-
tering geometry by a model BHSM Olympus microscope and a 50⫻
magnification objective (Olympus, Tokyo, Japan). The scattered radiation was directed into a Renishaw System 1000 Raman spectrometer where the laser radiation was filtered using a series of holographic notch filters and the Raman signal dispersed by high
resolution grating (1800 grooves/mm) onto a thermo-electrically
cooled CCD detector (⫺70°C) (Renishaw, Gloucestershire, UK). The
Raman spectra were collected and stored using Renishaw Raman
software operated on a Pentium-based PC. The combined spectral
resolution and reproducibility were experimentally determined to be
better than 3 cm⫺1. Spectral calibration was performed using the
atomic emission lines of a neon arc lamp.
UV/VIS Spectral Analysis. In solution with a neutral pH, the
metatungstate and paratungstate anions are in equilibrium with the
tungstate anion (Pope, 1983). We therefore performed UV/VIS spectral analysis of tungstate compounds to provide information about
the spectral characteristics of dilute solutions of these salts and their
stability over time. Salts were dissolved as described under the
materials section (see below) to a final concentration of 5 ␮M, and
spectral scans from 190 to 320 nm were collected using a Beckman
Coulter, Inc. (Fullerton, CA) DU 640 scanning spectrophotometer.
Electrophysiology. The perforated-patch technique was used to
assess the effects of tungsten compounds on macroscopic K⫹ currents
using voltage clamp protocols as described previously (Jackson et al.,
1997). An aliquot of cell-containing solution was placed on the coverslipped bottom of a 1-ml flow-through chamber, allowed to settle for
5 to 10 min, and then superfused with HEPES-buffered PSS (in
mmol/l): 140 NaCl, 5 KCl, 1 MgCl2, 2 CaCl2, 10 HEPES, 10 Glucose,
pH 7.4, with NaOH at room temp (295–300 mOsm). Patch-pipettes
were constructed from 1-mm i.d. ⫻ 1.5-mm o.d. Corning 7052 glass
tubes (Garner Glass Company, Claremount, CA), fire polished, their
tips filled with pipette solution (see below), and then back-filled with
the same solution containing 240 ␮g/ml amphotericin B as described
previously. Electrical access was gained in 10 to 15 min and was
maximal in 15 to 30 min. Pipette solution contained (in mmol/liter)
100 K-aspartate, 43 KCl, 1 MgCl2, 10 HEPES, 0.5 EGTA, pH 7 to 7.2
adjusted with NaOH, 295 to 300 mOsm adjusted with sucrose. Pipettes had tip resistances of 3 to 4 M⍀ when filled with this solution.
Seals were made on vascular smooth muscle cells by gently touching
the fire-polished tip of the pipette to a cell and then applying gentle
suction by mouth. Seal resistances were all greater than 10 G⍀ for
the studies presented, and no attempt was made to correct for leakage currents.
Currents were measured and membrane potential clamped with a
Warner PC-505A patch-clamp amplifier (Warner Instrument Corp.,
Hamden, CT). The amplifier was controlled by Axograph software
(version 4.5) running on a Power Mac 6500 computer equipped with
an ITC-18 data acquisition interface (Instrutech Corp., Port Washington, NY). Currents were filtered at 1 kHz and sampled at 5 kHz.
All currents reported were normalized to cell capacitance to account
for any differences in cell size. Cell capacitance was estimated by
integration of the capacitative transient elicited by stepping from
⫺60 to ⫺70 mV for 40 ms with the current filtered at 10 kHz and
sampled at 50 kHz. Cell capacitance averaged 22 ⫾ 1.1 pF (n ⫽ 12
cells from seven animals). Access resistance for these same cells
averaged 15 ⫾ 1 M⍀.
Cells were held at ⫺60 mV and then stepped, for 400 ms, to test
potentials from ⫺90 to ⫹60 mV in 10-mV increments. The average
current during the last 200 ms of the test pulse was then measured
at each potential and used to construct current (I)-voltage (V) relationships.
Materials. The tungsten and gold wires were obtained from Wirenetics (Sunnyvale, CA); phenylephrine was obtained from ACROS
Organics (Morris Plains, New Jersey); sodium tungstate (Na2WO4)
and sodium metatungstate [3Na2WO4 䡠 9WO3 (also written as
Na6W12O39)] were obtained from the Aldrich Chemical Co (Milwaukee, WI); ammonium paratungstate [(NH4)10W12O41] was obtained
from CERAC (Milwaukee, WI). All other chemicals were of reagent
Paratungstate and Relaxation
345
grade or better and obtained from the Sigma Chemical Co or Invitrogen (Carlsbad, CA). Phenylephrine was dissolved in PSS containing
100 ␮M ascorbic acid, and the tungstate compounds were dissolved
in acidified water that was then buffered with HEPES (0.1 M final
concentration), and the pH was adjusted to 7.4 with NaOH.
Statistical Analysis. The agonist concentration eliciting 50% of
the maximal relaxation (EC50) was determined from plots of the
percentage of initial tension versus the concentration of agonist. All
data are presented as mean ⫾ S.E.M., and statistical analysis was
performed using the SYSTAT software package (SPSS Science, Chicago, IL). Comparisons among groups were performed using ANOVA
with a repeated measures design when appropriate. A value of p ⬍
0.05 was taken to indicate a statistically significant difference.
Results
In experiments carried out previously, we found that mesenteric branch arteries mounted on 28-␮m tungsten wires
and precontracted with 5-␮M PE, responded to the cumulative addition of extracellular Ca2⫹ with a dose-dependent
relaxation to 8.1 ⫾ 1.46% of the initial tension response to PE
and with an ED50 for Ca2⫹ of 1.65 ⫾ 0.07 mM (n ⫽ 11). These
results were similar to those reported previously for arteries
precontracted with methoxamine (Mupanomunda et al.,
1999). In more recent studies, we found that the response of
isolated mesenteric branch arteries was significantly attenuated, with an obvious reduction in the maximal relaxation
response (54.8 ⫾ 4.5% of initial PE tension, n ⫽ 6, p ⬍ 0.001)
and in the ED50 value for Ca2⫹ (4.58 ⫾ 0.16 mM, n ⫽ 6, p ⬍
0.001). Multiple experimental parameters could have contributed to the reduced relaxation. These include the purity
of the water used to prepare the physiologic salt solution, the
colony or genotype of the rats used as the tissue source, the
purity of the gas mixture used to aerate the buffer, and the
amount of sunlight present in the room. Each of these was
tested as a contributing factor, but none was found to affect
the impaired dilator response.
During this same time period, we initiated experiments to
study relaxation responses of murine arteries and obtained
14-␮m diameter tungsten wire to accommodate the smaller
lumen of these vessels. We subsequently mounted rat arteries on this smaller wire to test the hypothesis that it was the
28-␮m diameter tungsten wire that was attenuating the relaxation response. In these experiments, adjacent vessel segments were mounted on either 14- or 28-␮m tungsten wire
and the relaxation response to either the cumulative addition
of Ca2⫹ or acetylcholine was assessed. Both the magnitude
(83.5 ⫾ 3.9%, n ⫽ 6) and the Ca2⫹ sensitivity (ED50 ⫽ 2.35 ⫾
0.20 mM, n ⫽ 6) of the Ca2⫹-induced relaxation response of
vessels mounted on the 14-␮m tungsten wire were significantly enhanced compared with those mounted on 28-␮m
wire (p ⬍ 0.001) (Fig. 1A).
To learn whether the attenuated response seen when vessels were mounted on 28-␮m wire was limited to the relaxation induced by extracellular Ca2⫹, or whether it extended
to another vasodilator pathway, we assessed the response of
arteries to the cumulative addition of acetylcholine. Both the
maximal relaxation to acetylcholine (66.0 ⫾ 5.3% on 28-␮m
wire versus 97.4 ⫾ 0.8% on 14-␮m wire, n ⫽ 5–7, p ⬍ 0.001)
and the ED50 values for acetylcholine (1.03 ⫾ 0.44 ␮M on
28-␮m wire versus 0.04 ⫾ 0.005 ␮M on 14-␮m wire, p ⫽
0.002) were significantly reduced on the 28-␮m wire (Fig.
1B).
Fig. 1. Response of PE-contracted mesenteric branch arteries to cumulative addition of extracellular Ca2⫹ (A) or acetylcholine (B) when
mounted on aged 28-␮m tungsten wires, 14-␮m tungsten wires, or 28-␮m
tungsten wire after removal of surface oxidation by abrasion with fine
sandpaper. Values are mean ⫾ S.E.M., n ⫽ 6 per group; ⴱ, significant
difference at p ⬍ 0.05.
These findings caused us to consider the possibility that
the 28-␮m wire was contaminated with a vasotoxic substance. We therefore performed several experiments in which
the tungsten wires were soaked in ethanol to remove possible
contaminating organic substances. No improvement in the
responses was observed (data not shown). We next considered the possibility that the surface of the 28-␮m wire had
become oxidized and the oxidation product was inhibiting the
relaxation. As a test, we compared the responses of vessels
mounted on the 28-␮m tungsten wires with those of segments mounted on 28-␮m wire taken from the same stock,
but which had been cleaned by abrading the surface with fine
sandpaper. Sanding the 28-␮m wire restored the relaxation
response of the artery to both the cumulative addition of
extracellular Ca2⫹ and to acetylcholine so that differences in
the maximal response and ED50 values for each dilator were
no longer detected (Fig. 1, A and B).
To learn whether minor surface oxidation might also be
present on the 14-␮m tungsten wire and acting to cause more
subtle inhibition of the responses, we compared relaxation
346
Bukoski et al.
responses of arteries mounted on 14-␮m tungsten wire with
those of adjacent segments mounted on 14-␮m wire made
from gold. Vessels mounted on 14-␮m gold wire had significantly enhanced sensitivity to the vasodilator effect of extracellular Ca2⫹ (ED50 tungsten ⫽ 1.83 ⫾ 0.06 mM versus ED50
gold ⫽ 1.45 ⫾ 0.02 mM; n ⫽ 6, p ⫽ 0.001) while no difference
in the maximal response (92.5 ⫾ 1.1% for tungsten versus
93.7 ⫾ 2.6% for tungsten; n ⫽ 6, p ⫽ 0.686) was detected (Fig.
2A). In contrast, no differences in the vasodilator response to
acetylcholine were detected (ED50 tungsten ⫽ 0.010 ⫾ 0.002
versus ED50 gold ⫽ 0.010 ⫾ 0.002 ␮M; n ⫽ 6, p ⫽ 0.981)
(maximal response for gold 89 ⫾ 2.5% versus 89.3 ⫾ 7.2% for
tungsten; n ⫽ 5–7, p ⫽ 0.890) (Fig. 2B). These results are
consistent with the hypothesis that the more recently purchased 14-␮m tungsten wire was causing a slight but significant inhibition of the response to Ca2⫹, but not the response
to acetylcholine.
Working on the assumption that surface oxidation was
responsible for the reduced response of arteries mounted on
Fig. 2. Effect of mounting vessels on 14-␮m tungsten wires or 14-␮m gold
wires on ligand-induced relaxation. A, response of PE-contracted mesenteric branch arteries to cumulative addition of extracellular Ca2⫹, n ⫽ 6
per group; repeated measures ANOVA gave an overall p value of 0.015; ⴱ,
significant difference at a p value of at least ⬍0.05. B, response to
acetylcholine, n ⫽ 8 for tungsten wire and 5 for gold wire, repeated
measures ANOVA gave an overall p value of 0.177; no significant differences were detected.
the 28-␮m tungsten wire, laser Raman spectral analysis was
performed on a segment of the corroded 28-␮m tungsten wire
and of an adjacent segment that had its surface cleaned with
fine sandpaper. The results of the spectral analysis indicated
that the surface of the corroded wire was in an oxidized state
and that the oxidation product could be removed by sanding
(Fig. 3). The Raman spectra for the tungsten oxide coating
determined at two different places on the wire were compared with reference spectra in the literature (Griffith and
Lesniak, 1969). Analysis of the oxide in one region most
closely matched that of Na12[W12O42]⫺12 although the lack of
lattice vibrations in the low-frequency region indicated a
microcrystalline form. The oxide in the second region most
closely matched that of the paratungstate [W12O42]⫺12 anion,
indicating an amorphous tungstate structure (Fig. 3).
In view of these results, we tested the effect of ammonium
paratungstate [(NH4)10W12O41] and two structurally dissimilar tungsten oxide formulations, sodium tungstate
(Na2WO4) and sodium metatungstate (3Na2WO4 䡠 9WO3) on
the relaxation responses of isolated arteries. These salts were
dissolved in acidified water, then buffered to pH 7.0 to 7.4.
Because the paratungstate and metatungstate anions are in
slow equilibrium with the tungstate anion (Pope, 1983), UV/
VIS spectral scans of dilute (5 ␮M) solutions of ammonium
paratungstate, sodium metatungstate, and sodium tungstate
were collected to determine whether spectral characteristics
of these compounds vary and might provide information
about stability of the salts in solution (Souchay et al., 1972).
Excitation from 190 to 320 nm resulted in the absorption
spectra illustrated in Fig. 4. Sodium metatungstate showed
the most complicated spectrum absorbing from 190 to 290 nm
with a broad peak centered at around 258 nm. Ammonium
paratungstate showed a lower level of absorbance from 190 to
290 nm and lacked the broad peak at 258 nm characteristic of
sodium metatungstate. In contrast, sodium tungstate was
characterized by a much lower level of absorbance at 190 nm,
which was completely absent at 212 nm and higher. These
spectra did not change over a 2-h period, indicating that
appreciable conversion of meta- and paratungstate to tungstate was not occurring during the time interval between
which the solution was prepared and used in the experiment.
Pretreatment of arteries with 1 ␮M ammonium paratungstate significantly inhibited Ca2⫹-induced relaxation (Fig.
5A) such that there was a reduction in both the maximal
response (control ⫽ 84.4 ⫾ 1.8% versus ammonium paratungstate ⫽ 50.7 ⫾ 4.2%; n ⫽ 6, p ⬍ 0.001) and the ED50
value (ED50 control ⫽ 1.91 ⫾ 0.15 versus ammonium paratungstate ⫽ 4.33 ⫾ 0.31 ␮M; n ⫽ 6, p ⬍ 0.001). Similarly,
when the effect of the paratungstate anion on the acetylcholine-induced relaxation response was examined, there was a
significant reduction in both the maximal response (control ⫽
87.3 ⫾ 2.4% versus ammonium paratungstate ⫽ 68.8 ⫾ 8.0%;
n ⫽ 6, p ⬍ 0.05) (Fig. 5B) and sensitivity (ED50 control ⫽
6.9 ⫾ 1.9 versus ammonium paratungstate ⫽ 22.6 ⫾ 0.9.6
nM; n ⫽ 6, p ⬍ 0.037).
In contrast, 100 ␮M sodium tungstate was without effect
on the relaxation response of the arteries to the cumulative
addition of extracellular Ca2⫹ (Fig. 6A), whereas 50 ␮M
sodium metatungstate caused a slight but significant increase in the relaxation response to the lowest concentrations
of extracellular Ca2⫹ (Fig. 6B). To test whether the ammonium cation might be responsible for the attenuated relax-
Paratungstate and Relaxation
347
Fig. 3. Laser Raman spectra of corroded 28-␮m tungsten
wire and a segment of the same wire that was cleaned with
fine sandpaper.
Since it is well accepted that oxygen free radicals can
diminish relaxation responses to certain vasodilators, we
assessed the effect of the superoxide dismutase mimetic 4-hydroxy-2,2,6,6-tetramethylpiperidine 1-oxyl (4-OH-TEMPO)
on Ca2⫹-induced relaxation in the presence and absence of 1
␮M ammonium paratungstate. Pretreatment with this compound did not affect Ca2⫹-induced relaxation or the inhibitory effect of paratungstate on Ca2⫹-induced relaxation (Fig.
8), indicating that neither the superoxide anion nor downstream reactive oxygen species are involved.
Fig. 4. UV/VIS spectra of 5 ␮mol/l solutions of ammonium paratungstate,
sodium metatungstate, and sodium tungstate. Spectra were collected as
described under Experimental Procedures using freshly prepared solutions and after the solutions were incubated at 37°C for 2 h. Abs, absorbance units relative to water.
ation responses, we assessed the effect of 10 ␮M ammonium
chloride (NH4Cl) on Ca2⫹-induced relaxation in two preparations and found that it was without effect (data not shown).
Previously, we showed that Ca2⫹-induced relaxation can
be inhibited by blockers of large-conductance Ca2⫹-activated
K⫹ (BKCa) channels such as tetraethylammonium (TEA) and
iberiotoxin (Bian and Bukoski, 1995; Ishioka and Bukoski,
1999). Therefore, we tested the hypothesis that paratungstate anion might inhibit Ca2⫹-induced relaxation by blocking BKCa channels. We found that 1 ␮M paratungstate anion
had no effect (p ⬎ 0.05) on macroscopic K⫹ currents in vascular smooth muscle cells (Fig. 7A). Likewise, in six cells
from three animals, 1 ␮M sodium metatungstate was without effect (data not shown). As a positive control, we examined the effects of 1 mM TEA, which, at this concentration,
selectively blocks vascular smooth muscle BKCa channels
(Nelson and Quayle, 1995). In contrast to the lack of effect of
the tungsten compounds, this quaternary ammonium BKCa
blocker significantly inhibited currents at positive membrane
potentials (Fig. 7B), consistent with other studies in the
literature (Nelson and Quayle, 1995).
Discussion
We have performed experiments to determine the mechanism(s) underlying the unexpected decrease in the Ca2⫹induced relaxation response of isolated mesenteric branch
arteries associated with aged tungsten wire. The new findings of this study include the demonstration that 1) spontaneous oxidation of tungsten wire significantly inhibits vasorelaxation responses to both Ca2⫹ and acetylcholine; 2) the
major tungsten oxidation product is similar to the paratungstate [W12O42]⫺12 anion; 3) low (1 ␮mol/l) concentrations of
ammonium paratungstate significantly attenuate Ca2⫹ and
acetylcholine-induced relaxation; and 4) substitution of tungsten wire with gold wire significantly enhances the Ca2⫹
sensitivity of the Ca2⫹-induced relaxation event.
The wire myograph technique originally described by
Bevan and Osher (1972) and modified by Mulvany and Halpern (1977) has permitted a large number of studies of vascular physiology and pharmacology on isolated arteries as
small as 100 ␮m in diameter. The basic approach of the
method is to thread thin wires into the lumen of a vessel
segment and to use these wires to affix the vessel by means
of screws to two metal or Plexiglas feet, with one affixed to a
translation micrometer and the other affixed to a force transducer. Although the original method used platinum wires for
mounting vessel segments (Bevan and Osher, 1972), at
present most laboratories use either tungsten (Steeds et al.,
348
Bukoski et al.
Fig. 5. Effect of 1 ␮mol/l ammonium paratungstate on the relaxation
response of PE-contracted arteries mounted on 14-␮m tungsten wires to
the cumulative addition of extracellular Ca2⫹ (A) or acetylcholine (B).
Values are mean ⫾ S.E.M., n ⫽ 6 per group; ⴱ, significant difference at
p ⬍ 0.05.
1997; Liu et al., 1998; Thorin et al., 1998) or stainless steel
wire (Buss et al., 1999; Scotland et al., 1999) presumably
because their tensile strength is greater than that of platinum.
The present report indicates that the tungsten wire that
our laboratory has routinely used for wire myography undergoes spontaneous surface oxidation. Surface oxidation is a
common property of the transition metals (Griffith and
Lesniak, 1969). We used laser Raman spectroscopy to analyze the state of the tungsten wire because, to our knowledge,
it is the only method that is available for detecting oxidation
on the surface of very small metal samples. The results
showed that the surface of the wire was in an oxidized state
and indicated that the oxidation product was similar to the
paratungstate anion. Our findings that the biological effects
of the oxidized wire are mimicked by low concentrations of
ammonium paratungstate, and that two other structurally
unrelated compounds (sodium tungstate and sodium metatungstate; Fig. 9) do not inhibit vascular relaxation, have led
to the conclusion that the surface oxidation product is a
paratungstate-like compound.
Fig. 6. Effect of 100 ␮mol/l sodium tungstate (A) or 50 ␮mol/l sodium
metatungstate (B) on the relaxation response of PE-contracted arteries
mounted on 14-␮m tungsten wires to the cumulative addition of extracellular Ca2⫹. Values are mean ⫾ S.E.M., n ⫽ 5 per group; ⴱ, significant
difference at p ⬍ 0.05.
The paratungstate and metatungstate anions are complex
structures, which in solution are in very slow equilibrium (on
the order of days) with one another. This relationship is
illustrated by Scheme 1 (Pope, 1983).
We therefore thought it important to determine whether
the polytungstates that we tested spontaneously dissociate to
form the tungstate anion at neutral pH. The approach that
Scheme 1
Paratungstate and Relaxation
Fig. 7. A, lack of effect of paratungstate anions on whole-cell K⫹ currents
in vascular smooth muscle cells. Shown are current densities in the
absence (control) and presence of 1 ␮M ammonium paratungstate. Data
are mean ⫾ S.E.M., n ⫽ 6; factorial analysis of variance indicated a
significant effect of voltage (p ⬍ 0.05), but no significant paratungstate or
interaction (voltage ⫻ paratungstate) effects (p ⬎ 0.05). B, inhibition by
TEA of whole-cell K⫹ currents in vascular smooth muscle cells. Shown
are current densities in the absence (control) and presence of 1 mM TEA.
Data are mean ⫾ S.E.M., n ⫽ 6; factorial analysis of variance indicated
significant effects of voltage and TEA, as well as a significant interaction
(voltage ⫻ TEA) between these two treatments (p ⬍ 0.05).
we used was to analyze the UV/VIS spectra of diluted solutions of these compounds (Souchay et al., 1972). The absorption spectra of the meta- and paratungstate anions were
different from the spectrum of the tungstate anion and did
not vary over the 2-h period that was examined, indicating
that they were chemically stable during the course of our
experiments.
Tungsten and sodium tungstate have been studied extensively in the past with special emphasis on the ability of
dietary tungsten and tungstate to inhibit xanthine oxidase
and H2O2 production (Owen and Dundas, 1969; Smith et al.,
1987; Swei et al., 1999) and of tungstate to mimic the serum
glucose clearing action of insulin, presumable by stimulating
glucose 6-phosphatase activity (Barbera et al., 1997; Foster
et al., 1998). To our knowledge, however, there is no published information about the vascular effects of ammonium
349
paratungstate. We therefore believe that our data are the
first to demonstrate that ammonium paratungstate significantly impairs both Ca2⫹ and acetylcholine-induced relaxation.
A question of primary concern is the mechanism by which
the paratungstate anion suppresses the vasodilator effect of
extracellular Ca2⫹. Early studies from Bohr (1962) were interpreted to indicate that relaxation induced by very high
concentrations of extracellular Ca2⫹ was the result of either
a membrane-stabilizing effect of the Ca2⫹ cation or an effect
of the cation to stimulate the smooth muscle cell Na⫹ pump
and induce hyperpolarization (Webb and Bohr, 1978). In
contrast with these earlier reports, collective evidence from
our laboratory supports the following pathway for the high
sensitivity Ca2⫹-induced relaxation of the isolated mesenteric branch artery. Extracellular Ca2⫹ binds to a Ca2⫹sensing receptor located on perivascular sensory nerves; activation of this receptor results in the production and/or
release of a nerve-derived hyperpolarizing vasodilator factor;
this hyperpolarizing factor binds to a receptor on adjacent
smooth muscle cells; and activation of the receptor is coupled
with opening of an iberiotoxin-sensitive KCa channel, which
leads to hyperpolarizing vasodilation (Bukoski, 1998, 2001).
The finding that the paratungstate anion blocks both
Ca2⫹-induced relaxation and attenuates acetylcholine-induced relaxation makes it unlikely that the compound is
acting by inhibiting a sensory nerve Ca2⫹ receptor, since this
receptor is not believed to be active in acetylcholine-induced
relaxation. Moreover, the electrophysiological studies reported here failed to demonstrate inhibition of BKCa channels by the paratungstate anion. Additional evidence obtained using 4-OH-TEMPO rule out the possibility that
paratungstate-generated superoxide anions are chemically
inactivating a nerve-derived vasodilator. These data indicate
that paratungstate must be acting at some other point along
the signaling cascade, either at the receptor for the nervederived hyperpolarizing factor, at a site intermediate to activation of the receptor and opening of the KCa channel, or at
a site distal to KCa channel opening. Further research will be
required to determine the precise mechanism of action.
Our results surrounding the finding that surface oxidation
of tungsten wire impairs arterial relaxation are of interest for
several reasons. One is that it would seem prudent to stop the
use of tungsten wire for mounting arterial segments on wire
myographs. In place of tungsten, we recommend that another
nonreactive wire be used such as gold or tungsten-free stainless steel. A second point of interest is that our analysis has
shown that the paratungstate anion is a potent inhibitor of
Ca2⫹-induced relaxation. This compound may find utility as
a tool in the study of mechanisms of Ca2⫹-induced relaxation.
The third deals with the finding that very low, physiologic
concentrations of extracellular Ca2⫹ are able to elicit nearly
complete relaxation of isolated arteries when they are
mounted on gold wire. The finding that the ED50 for Ca2⫹ is
less than 1.5 mM when studied using gold wire compares
well with our recent demonstration that interstitial Ca2⫹ in
the duodenal submucosa (Mupanomunda et al., 1999) and
the renal cortex (Mupanomunda et al., 2000) ranges from 1 to
slightly more than 2 mM and suggests that a Ca2⫹-activated
dilator system may be functional in select vascular beds
under physiologic conditions.
350
Bukoski et al.
Fig. 8. Effect of 1 ␮mol/l ammonium paratungstate on Ca2⫹induced relaxation in the presence and absence of 1 ␮mol/l of the
superoxide dismutase mimetic 4-OH-TEMPO. Values are
mean ⫾ S.E.M., n ⫽ 4 per group; no effect of 4-OH-TEMPO was
detected.
Fig. 9. Bond and polyhedral representations of tungstate anions. Bond
(A) and polyhedral (B) representations of the oxygen coordination of the
tungstate WO42⫺ anion; C, polyhedral representation of the paratungstate anion; D, polyhedral representation of the metatungstate anion
based on the Keggin structure for tetrahedral heteroatoms. Adapted from
Pope (1983) with permission.
Acknowledgments
We acknowledge the expert administrative assistance of Barbara
Manning and the technical assistance of Monje Baker.
References
Barbera A, Fernandez-Alvarez J, Truc A, Gomis R, and Guinovart JJ (1997) Effect of
tungstate in neonatally streptozotocin-induced diabetic rats: mechanism leading
to normalization of glycaemia. Diabetologia 40:143–149.
Bevan JA and Osher JV (1972) A direct method for recording tension changes in the
wall of small blood vessels in vitro. Agents Actions 2:257–260.
Bian K and Bukoski RD (1995) Modulation of resistance artery force generation by
extracellular Ca2⫹. Am J Physiol 269:H230 –H238.
Bohr DF (1963) Vascular smooth muscle: dual effect of calcium. Science (Wash DC)
19:597–599.
Bukoski RD (1998) The perivascular sensory nerve Ca2⫹ receptor and blood pressure
regulation. An hypothesis. Am J Hypertens 11:1117–1123.
Bukoski RD, Bian K, Wang Y, and Mupanomunda M (1997) The perivascular
sensory nerve Ca2⫹ receptor and Ca2⫹ induced relaxation of isolated arteries.
Hypertension 30:1431–1439.
Bukoski RD (2001) Dietary Ca2⫹ and blood pressure: evidence that Ca2⫹ sensing
receptor activated, sensory nerve dilator activity couples changes in interstitial
Ca2⫹ with vascular tone. Nephrol Dial Transplant 16:218 –221.
Buss NH, Bottcher M, Bottker HE, Sorensen KE, Nielsen TT, and Mulvany MJ
(1999) Reduced vasodilator capacity in syndrome X related to structure and function of resistance arteries. Am J Cardiol 83:149 –154.
Foster JD, Young SE, Brandt TD, and Nordlie RC (1998) Tungstate: a potent
inhibitor of multifunctional glucose-6-phosphatase. Arch Biochem Biophys 354:
125–132.
Griffith WP and Lesniak PJB (1969) Raman studies on species in aqueous solutions.
Part III. Vanadates, molybdates, and tungstates. J Chem Soc A:1066 –1071.
Ishioka N and Bukoski RD (1999) A role for N-arachidonylethanolamine as the
mediator of sensory nerve dependent Ca2⫹-induced relaxation. J Pharmacol Exp
Ther 289:245–250.
Jackson WF, Huebner JM, and Rusch NJ (1997) Enzymatic isolation and characterization of single vascular smooth muscle cells from cremasteric arterioles. Microcirculation 4:35–50.
Jackson WF (2000) Hypoxia does not activate ATP-sensitive K⫹ channels in arteriolar muscle cells. Microcirculation 7:137–145.
Liu R, Lang MG, Luscher TF, and Kauffman M (1998) Propofol-induced relaxation of
rat mesenteric arteries: evidence for a cyclic GMP-mediated mechanism. J Cardiovasc Pharmacol 32:709 –713.
Mulvany MJ and Halpern W (1977) Contractile properties of small arterial resistance vessels in spontaneously hypertensive and normotensive rats. Circ Res
41:19 –26.
Mupanomunda MM, Ishioka N, and Bukoski RD (1999) Interstitial Ca2⫹ undergoes
dynamic changes sufficient to stimulate nerve dependent Ca2⫹-induced relaxation.
Am J Physiol 276:H1035–H1042.
Mupanomunda MM, Tian B, Ishioka N, and Bukoski RD (2000) Renal Interstitial
Ca2⫹. Am J Physiol 278:F644 –F649.
Nelson MT and Quayle JM (1995) Physiological roles and properties of potassium
channels in arterial smooth muscle. Am J Physiol 268:C799 –C822.
Owen EC and Dundas I (1969) Diminution of xanthine oxidase in goat milk after
subcutaneous injection of sodium tungstate (Na2WO4-2H2O) (Abstract). Proc Nutr
Soc 28:59.
Pope MT (1983) Heteropoly and Isopoly Oxometalates. Springer-Verlag, Berlin.
Scotland R, Vallance P, and Ahluwalia A (1999) Endothelin alters the reactivity of
vasa vasorum: mechanisms and implications for conduit vessel physiology and
pathophysiology. Br J Pharmacol 128:1229 –1234.
Smith SM, Grisham MB, Manci EA, Granger DN, and Kvietys PR (1987) Gastric
mucosal injury in the rat. Role of iron and xanthine oxidase. Gastroenterology
92:950 –956.
Souchay P, Boyer M, and Chauveau F (1972) Advances in the solution chemistry of
tungstates, in Contributions to Coordination Chemistry in Solution, p 159, Stockholm.
Steeds RP, Thompson JS, Channer KS, and Morice AH (1997) (1997) Response of
normoxic pulmonary arteries of the rat in the resting and contracted state to NO
blockade. Br J Pharmacol 122:99 –102.
Swei A, Lacy F, Delano FA, Parks DA, and Schmid-Schonbein GW (1999) A mechanism of oxygen free radical production in the Dahl hypertensive rat. Microcirculation 6:179 –187.
Thorin E, Cernacek P, and Dupuis J (1998) Endothelin-1 regulates tone of isolated
small arteries in the rat: effect of hyperendothelinemia. Hypertension 31:1035–
1041.
Webb RD and Bohr DF (1978) Mechanism of membrane stabilization by calcium in
vascular smooth muscle. Am J Physiol 235:C227–C232.
Address correspondence to: Richard Bukoski, Ph.D., Cardiovascular Disease Research Program, Julius L. Chambers Biomedical Biotechnology Research Institute, North Carolina Central University, 700 George St., Durham,
NC 27707. E-mail: [email protected]