Bioelectromagnetics 24:517^523 (2003)

Effect of a 125 mT Static Magnetic Field on the

Kinetics of Voltage Activated Naþ Channels
in GH3 Cells
Arthur D. Rosen*
Department of Biological Sciences, Purdue University,West Lafayette, Indiana
Voltage activated Naþ channels were examined in GH3 cells, using the whole cell patch clamp method.
Channel currents were recorded before, during, and after a 150 s exposure to a 125 mT static magnetic
field. There was a slight shift in the current–voltage relationship and a less than 5% reduction in peak
current during magnetic field exposure. More pronounced, however, was an increase in the activation
time constant, tm, during and for at least 100 s following exposure to the field. This change in tm was
seen primarily at lower activation voltages. No change was noted in the inactivation time constant, th.
Changes were clearly temperature dependent, being evident only at and above 35 8C. These findings
are consistent with the hypothesis that reorientation of diamagnetic anisotropic molecules in the cell
membrane are capable of distorting imbedded ion channels sufficiently to alter their function. The
temperature dependence of this phenomenon is probably due to the greater ease with which a liquid
crystal membrane can be deformed. Bioelectromagnetics 24:517–523, 2003.
2003 Wiley-Liss, Inc.

Key words: sodium channels; DC magnetic fields; diamagnetism; magnetic anisotropy;

thermotropic phase transition; cell membrane

INTRODUCTION These experiments employed the whole cell patch

Static magnetic fields (SMF), with intensities of
10 mT–1.4 T, have been shown to influence a variety of
biological systems. Changes have been described in
spontaneous neuronal activity in the central nervous
system [Rosen and Lubowsky, 1990], in the function of
membrane-bound receptors in hemopoietic stem cells
[Peterson et al., 1992], in neurotransmitter release at the
neuromuscular junction [Rosen, 1992], and in neuronal
action potentials in dissociated cultures of dorsal root
ganglia neurons [McLean et al., 1995]. Many of the
biological phenomena attributed to SMFs have been
explained by changes in membrane ion transport, with
calcium channels being most often implicated.
It has been suggested [Rosen and Lubowsky,
1990] that SMFs induce a partial realignment of
diamagnetically anisotropic molecules within the cell
membrane, thereby distorting imbedded ion channels
sufficiently to alter their function. Support for this
hypothesis has come from experimental studies that
examined calcium channel function both indirectly
[Rosen, 1994] and directly [Rosen, 1996] with fields of
up to 125 mT. If other ion channels were found to be
similarly influenced by SMFs, this would lend addi-
tional support to the functional significance of realign-
ment of diamagnetic molecules within the cell
membrane. The present study was designed to examine
the kinetics of voltage activated sodium channels
during and following exposure to a 125 mT SMF.
clamp method, using cultured GH3 cells, to examine
the response of voltage activated sodium channels. The
GH3 cell line is of rat pituitary origin and exhibits
large tetrodotoxin (TTX) sensitive sodium currents,
characteristic of a neuronal phenotype [Matteson and
Armstrong, 1984], but does not elaborate neuritis in
culture. This obviates any potentialcomplication involv-
ing spatial clamp of rapidly activated inward sodium or
calcium current. Since GH3 cells are maintained in
culture and are anchoring in nature, they are especially
amenable to whole cell patch clamp methodology.
MATERIALS AND METHODS
Cell Cultures
Proliferating GH3 cells were grown in 100 mm
polystyrene culture dishes at 37 8C in a 5% CO2
environment. The growth medium consisted of DMEM
supplemented with 10% fetal calf serum and penicillin/
————— —
*Correspondence to: Dr. Arthur Rosen, Department of Biological
Sciences, Purdue University, West Lafayette, IN 47907.
E-mail: arosen@bilbo.bio.purdue.edu
Received for review 2 August 2002; Final revision received 18
December 2002
DOI 10.1002/bem.10124
Published online in Wiley InterScience (www.interscience.wiley.com).

2003 Wiley-Liss, Inc.

518 Rosen
streptomycin. As the cells approached confluence, they
were split and replated at 10% of their initial density. At
the same time a number of 35 mm polystyrene culture
dishes were established to provide sufficient dishes of
cells for recording on any given day. All recordings
were carried out between two and five days of plating to
minimize any variability due to cell health.
Temperature Control
Bath temperature was controlled with a 34.2 mm
diameter circular thermofoil heater (Minco Products,
Minneapolis, MN) with an 11.4 mm diameter central
opening. This device was cemented to a larger epoxy
glass composite base and fit neatly under a 35 mm
culture dish, while allowing visual access to those cells
situated over the central aperture. Since the heater itself
was only 0.05 mm thick, it did not make physical
contact with the floor of the dish, thus eliminating the
possibility of movement associated with thermal
expansion of the heater’s resistive elements. DC current
for the heater was regulated by a proportional temp-
erature controller (Omron Electronics, Schaumburg,
IL) with an analog output. This arrangement prevented
the introduction of switching transients into the
recording system. Temperature sensing was by means
of a small insulated thermocouple, situated 1–2 mm
from that cell from which recordings were being carried
out. The overall stability of this system was
0.1 8C
from the set point.
Exposure System
The electromagnet used consisted of a 2700 turn
coil, wound on a 2.4 cm2 soft iron C-core with an air gap
of 44 mm. The coil was powered by a computer
controlled current source and the system was capable of
generating static fields with flux densities up to 125 mT.
The magnet was mounted on a specially constructed
aluminum stage designed to fit the Zeiss Axiovert
100 microscope. All ferromagnetic materials were
removed or replaced on this microscope. Tissue culture
dishes, placed on the thermoregulated portion of
the microscope stage, were oriented horizontally
between the poles of the magnet. Recordings were
carried out on those cells within the 650 mm diameter
microscope field, which was fixed in position and
centered midway between the magnet poles. Precise
measurements of the magnetic field generated by this
device revealed nonhomogeneity of less than 0.1%
within the area from which recordings were made
(Fig. 1). The microscope illuminator was operated from
a well filtered DC source and no stray fields could be
detected from it or from the thermoregulation device.
Heat generated by the magnet coils was dissipated well
away from the recording area.
Fig. 1. Spatialdistribution of staticmagnetic fieldusedinthis study.
Measurements made normal to the inter-pole axis (Yplane) at the
level of the floor of the tissue culture dish. Mid-pole field of125 mT.
Recordings made at X ¼ 0
325 mm,Y ¼ 0
325 mm.
Recording Conditions
Immediately prior to recording, the growth medi-
um in a 35 mm culture dish was slowly exchanged with
a 37 8C solution of 140 mM NaCl, 5 mM MgCl2, and
10 mM HEPES, titrated with NaOH to a pH of 7.2. The
bath temperature was allowed to stabilize for 15 min
at the selected temperature prior to recording. There
was no exchange or circulation of the medium during
recording.
Patch pipettes were pulled from borosilicate glass
to an outer diameter of approximately 2 mm and lightly
fire polished. The electrode was filled with a solution
consisting of 120 mM CsCl, 10 mM K-EGTA, and
10 mM HEPES, titrated with KOH to a pH of 7.2. These
electrodes had resistances of 2–4 MO and, combined
with the use of series resistance compensation, voltage
clamp errors due to high access resistance were
minimized. Membrane currents were recorded with an
Axopatch 1-D patch clamp amplifier. Command volt-
ages for this amplifier were generated with HEKA Pulse
software, running on a Macintosh 840AV computer.
Current responses were digitized at 40 kHz with an
Instrutech ITC-16 A/D converter and processed for
display by the Pulse software.
Under slight positive pressure, the electrode tip
was lowered to the cell surface while its resistance was
continuously monitored by measuring the current in
response to application of a 10 mV, 10 ms pulse. Once
the electrode contacted the cell, a small negative
pressure was applied to form a seal of at least 1 GO.
Following seal formation, the tissue within the pipette
was broken with additional negative pressure. The
membrane was initially clamped to a holding potential
of 80 mV, and inward sodium current was assessed by
the application of a series of 100 ms pulses, in 5 mV

increments, from the holding potential to þ40 mV.
Correction for any linear leakage currents was accom-
plished with a P/4 subtraction protocol [Bezanilla and
Armstrong, 1977]. The leak-subtracted data were
displayed, along with on line analysis of the current–
voltage relationships for that cell.
Data acquisition was delayed until the current–
voltage relationship stabilized. This typically occurred
5 min after the cell was clamped, the time needed for the
electrode solution to dialyze the intracellular fluid.
Once stable, currents obtained during the application of
the test pulse sequence were recorded. This was
immediately followed by exposure to a 125 mT SMF
for 150 s. The test pulse sequence was repeated during
exposure, beginning 100 s after the field was turned on,
100 s after the field was turned off, and 200 s after the
field was turned off. Data were stored for off line
analysis. At the conclusion of recordings from a given
cell, 1 mM of TTX was added to the preparation and test
pulses again applied.
Data Analysis
Current–voltage relationships were computed
on line. Activation and inactivation time constants were
computed on the stored data by fitting the sodium
current at each voltage step with a general Hodgkin–
Huxley formalism with four gate particles, three
controlling activation (m) and one controlling inactiva-
tion (h). Each gate makes an independent first order
transition between the open and closed state. The
probability of all particles being in the open state is m3h,
and this predicts sodium conductance. This model
provided an excellent fit to the data. The time constants
for gate activation (tm) and inactivation (th), which
define channel kinetics, are derived from the Hodgkin–
Huxley equations [Hille, 1992].
RESULTS
Sodium current recordings were obtained from
18 cells using the whole cell configuration of the patch
clamp technique. Studies were carried out at seven
temperatures, from 25 to 37 8C in 2 8C increments, with
a minimum of two cells examined at each temperature.
Cells were of fairly uniform size with an average
diameter of 15 mm, and stable recordings were usually
possible for periods of 10–15 min. In every cell studied,
there was a complete block of inward current following
exposure to TTX.
A typical response to a 100 ms voltage step
sequence at 35 8C is shown in Figure 2. The current–
voltage plot revealed the sodium current threshold to be
35 mV. Peak current was initially observed at 5 mV,
but during the first 5 min after clamping, a 10 mV shift
SMFs and Naþ Kinetics 519
toward a more negative value was seen, which then
remained constant. This shift in peak is expected
because of dialysis of the intracellular fluid. Since the
threshold voltage remained at 35 mV, the activation
slope increased in the current–voltage plots. In addit-
ion, a steep activation limb in the current–voltage plot
was found to be characteristic of recordings made in
GH3 cells at temperatures at and above 35 8C [Rosen,
2001]. For the recording conditions of the present study
(no internal Naþ and replacement of internal Kþ with
Csþ), outward current flow is not seen because the
reversal potential is considerably greater than the
þ40 mV maximum depolarizing pulse employed
[Matteson and Armstrong, 1984]. Current recordings
described in this report are consistent with those
described in the literature for GH3 cells [Fernandez
et al., 1984; Matteson and Armstrong, 1984].
In those cells (N ¼ 6) where recordings were
carried out at 35 and 37 8C, exposure to a 125 mT SMF
was associated with a slight shift in the current–voltage
relationship and a less than 5% decrease in the peak
current. Both of these alterations were transient, return-
ing to preexposure values within 100 s after the field was
turned off. More impressive, however, was a decrease
in the activation rate for lower activation voltages,
indicated by an increase in the value of the activation
time constant, tm. Increased tm values were evident
during magnetic field exposure and persisted for at least
100 s after the field was turned off. This is shown in
Figure 3 for the same cell shown in Figure 2. These
changes were evident in all cells where recordings
were carried out at both 35 and 37 8C and was slightly
more prominent at the higher temperature. The magni-
tude of the effect of SMF exposure was evaluated by
calculating the mean percent change from control
values of tm, at each activation voltage, for all cells in
the 35/37 8C group. These calculations are presented in
Figure 4. Statistical analysis (Wilcoxon matched pair
signed rank test) revealed a significant (P < 0.05)
difference in tm during exposure and for 200 s after
the field was turned off. In those cells (N ¼ 12) where
recordings were carried out between 25 and 33 8C, no
changes in the current–voltage relationship, peak
current amplitude, or tm were seen with SMF exposure
(data not shown).
There was no change in the inactivation time
constant th during SMF exposure, regardless of the
temperature. This is illustrated in Figure 5, where values
were computed from the same cell shown in Figure 2.
DISCUSSION
In the presence of an adequate static magnetic
field gradient, diamagnetic anisotropic molecules will
520 Rosen

Fig. 2. Inward voltage activated sodium current in GH3 cell at 35 8C with corresponding current^
voltage plots. Holding potential was 80 mV. Left column: Superimposed current traces for activa-
tionvoltages of 80 to þ40 mV, in 5 mV steps. Right column:Peakinward current foreach activation
voltage. The figure shows records before exposure to a 125 mT static magnetic field (CONTROL),
duringexposure (SMF), and100 sfollowingexposure (POST).Thereisa 5 mV shift in current^voltage
relationship and a 0.28 nA (4.5%) reduction in peak current during exposure. These changes
returned to controlvalues during postexposure trial.

experience translational movements. In a homogene- field may be expressed by the ratio, b, of magnetic
ous field, however, orientation to the minimum free energy to thermal energy,
energy state will occur. For individual molecules,
the effect is small and easily offset by the randomiz- NDx H 2
ing effect of thermal energy. It has been shown b¼ ; ð1Þ
[Maret and Dransfield, 1977] that for domains of kB T
interacting molecules aligned along a common axis, the
aggregate molecular anisotropy can be sufficient to where N is the number of interacting molecules, each
overcome the effects of thermal energy. The probability with a diamagnetic anisotropy of Dx, H is flux density,
of such a molecular domain rotating in a magnetic kB is Boltzmann’s constant, and T is absolute

Fig. 3. Channel activation time constants (A) before 125 mT SMF
exposure, (B) during exposure, (C) 100 s after exposure, and (D)
200 s after exposure. There are increased values of tm, at lower
activation voltages, during and following exposure. tm computed
from cell shown in Figure 2.Recordings at 35 8C.
temperature. Molecular orientation will occur when
magnetic energy exceeds thermal energy, i.e., when
b > 1. Within the physiologically permissible range,
the effect of temperature on thermal energy is small; a
SMFs and Naþ Kinetics 521
Fig. 4. Percent change in channel activation constants (tm) when
compared to controlvalues (A) during exposure to125 mT SMF, (B)
100 s following exposure, and (C) 200 s following exposure. Mean
valuesforallrecordingsmadeat 35 and 37 8C (N ¼ 6); verticalbars
are standard errorofthe mean.Nonparametric analysis (Wilcoxon
matched pair signed rank test) was used to evaluatelevelof statis-
tical significance at each activation voltage. *, Marked points:
P < 0.05.
10 8C difference in temperature will result in only a
3.3% change in b.
The enhanced diamagnetic property of biological
membranes, with their highly ordered phospholipid
bilayer structure, is well established [Neugenbauer
et al., 1977; Boroske and Helfrich, 1978; Braganza
et al., 1984; Speyer et al., 1987]. Although the value of
b, calculated from the above equation, is for freely
mobile diamagnetic structures, molecular reorientation
within a phospholipid bilayer represents the balance
between the theoretical value of b and those inter-
molecular forces that limit movement. At low tempera-
tures, the membrane exists in a gel (Lb) phase with tight
522 Rosen
Fig. 5. Channelinactivationtime constants, (A) before125 mTSMF
exposure, (B) during exposure, (C) 100 s after exposure, and (D)
200 s after exposure. No change is seen in values of th. th is com-
puted from cell shown in Figure 2.Recordings at 35 8C.
alignment of individual molecules, maintained by both
intermolecular forces, and the excluded volume effect
of packed lipid molecules. With increasing tempera-
ture, an abrupt rotameric disordering of the lipid acyl
chains occurs, marking the transition to the liquid
crystal (La) phase. The effect of this nonlinear transition
on the membranes physical properties has recently been
described [Rosen, 2001]. At any temperature at or
above the phase transition temperature, the more fluid
liquid crystal membrane will be more readily deformed
in a magnetic field and such deformation could influ-
ence the intramembranous portion of embedded ion
channels by introducing conformational changes to that
portion of the protein involved in one or more of the
coupled steps involving activation.
Sodium channel activation is due to a sequence
of voltage induced changes in the S4 segment of the
a-subunit [Caterall, 1988]. This subunit is a 260 kDa
protein located entirely within the membrane where it
would be vulnerable to the effects of diamagnetic
reorientation of phospholipid molecules. Maximum
SMF effect was evident with membrane depolarization
to 5 mV or less. Higher depolarizing voltages are
associated with more robust channel activation that
would be expected to be less sensitive to membrane
deformation. Fast inactivation of the sodium channel
has been shown to involve an intracellular linkage
between domains III and IV. This 53 amino acid linkage
[Patton et al., 1992] operates by occluding the channel
pore in the membrane’s cytoplasmic domain. Such an
intracellular mechanism would be unaffected by
physical changes within the membrane.
The time course for the effect of SMFs on the cell
membrane is noteworthy. In previous studies [Rosen,
1994, 1996], maximum effects during exposure to a
125 mT field occurred in 100 s. The present study
demonstrated a significant effect persisting for at least
100 s after the field was turned off. Both of these
observations are consistent with SMF induced mem-
brane deformation being a relatively slow process. The
orientation of diamagnetic structures in a magnetic field
is a function of their anisotropy and geometry, as well as
field strength and the nature of the suspension medium.
Rotation of ensembles of oriented diamagnetic mole-
cules in a homogeneous magnetic field may be describ-
ed by a nonlinear differential equation [Hong et al.,
1971], and from that equation, the time course for
rotation becomes
NDx H 2
ln y1  ln y0 ¼ t ð2Þ
z
where y is the angle between the molecular axial
diamagnetic vector and the field direction and t is the
time required for rotation from y0 to y1. z is the
rotatory frictional coefficient, whose value is estab-
lished by the size and shape of the ensemble as well as
the viscosity of the suspension medium.

The changes observed in sodium channel kinetics,
during exposure to a 125 mT SMF, are similar to those
noted in calcium channels. This, therefore, is consistent
with the hypothesis that slow realignment of a mem-
brane’s diamagnetically anisotropic molecules can
influence the activation of imbedded ion channels while
leaving unaffected those portions of the channel that
do not lie within the membrane per se. The obvious
temperature dependence of this phenomenon appears
to be related to the membrane thermotropic phase
transition, with the less viscous liquid crystal mem-
brane being more easily deformed in the presence of
these fields.
It is worth emphasizing that many inorganic and
virtually all organic molecules have some degree of
diamagnetic anisotropy. These molecules in isolation
will exhibit preferred orientation in the presence of
fields with magnitudes of at least several Tesla [Maret
and Dransfield, 1985]. They do not, however, exhibit
significant orientation in less intense fields. In many
biological systems, diamagnetic molecules are arrang-
ed in a highly ordered fashion and functionally linked
with parallel diamagnetic vectors. It is this structural
property that allows for the summation of individual
molecular anisotropies and, therefore, preferred orien-
tation of the ensemble at lower field strengths.
REFERENCES
Bezanilla F, Armstrong CM. 1977. Inactivation of the sodium
channel. I. Sodium current experiments. J Gen Physiol 70:
549–566.
Boroske E, Helfrich W. 1978. Magnetic anisotropy of egg lecithin
membranes. Biophys J 24:863–868.
Braganza LF, Blott BH, Coe TJ, Melville D. 1984. The super-
diamagnetic effect of magnetic fields on one and two
component multilamellar liposomes. Biochim Biophys Acta
801:66–75.
Catterall WA. 1988. Structure and function of voltage-sensitive ion
channels. Science 242:50–61.
Fernandez JM, Fox AP, Krasne S. 1984. Membrane patches and
whole-cell membranes: A comparison of electrical properties
in rat clonal pituitary (GH3) cells. J Physiol 356:565–585.
SMFs and Naþ Kinetics 523
Hille B. 1992. The Hodgkin–Huxley model describes permeability
changes. In: Ionic channels of excitable membranes. 2nd
edition. Massachusetts: Sinauer. pp 44–58.
Hong FT, Mauzerall D, Mauro A. 1971. Magnetic anisotropy and
the orientation of retinal rods in a homogeneous magnetic
field. Proc Nat Acad Sci USA 68:1283–1285.
Maret G, Dransfield K. 1977. Macromolecules and membranes in
high magnetic fields. Physica 86–88B:1077–1083.
Maret G, Dransfield K. 1985. Biomolecules and polymers in high
steady magnetic fields. In: Strong and ultrastrong magnetic
fields and their applications. New York: Springer-Verlag.
pp 143–204.
Matteson DR, Armstrong CM. 1984. Na and Ca channels in a
transformed line of anterior pituitary cells. J Gen Physiol 83:
371–394.
McLean MJ, Holcomb RR, Wamil AW, Picket JD, Cavopol AV.
1995. Blockade of sensory neuron action potentials by a static
magnetic field in the 10 mT range. Bioelectromagnetics
16:20–32.
Neugenbauer DC, Blaurock AE, Worcester DL. 1977. Magnetic
orientation of purple membranes demonstrated by optical
measurements and neutron scattering. FEBS Lett 78:31–35.
Patton DE, West JR, Catterall WA, Goldin AL. 1992. Amino acid
residues required for fast Naþ-channel inactivation: Charge
neutralizations in the III–IV linker. Proc Natl Acad Sci USA
89:10905–10909.
Peterson HP, vonWangenheim KH, Feinendegen LE. 1992.
Magnetic field exposure of marrow donor mice can increase
the number of spleen colonies (CFU-S 7d) in marrow
recipient mice. Radiat Environ Biophys 31:31–38.
Rosen AD. 1992. Magnetic field influence on acetylcholine release
at the neuromuscular junction. Am J Physiol 262:C1418–
C1422 (Cell Physiol 31).
Rosen AD. 1994. Threshold and limits of magnetic field action at
the presynaptic membrane. Biochim Biophysica Acta 1193:
62–66.
Rosen AD. 1996. Inhibition of calcium channel activation in GH3
cells by static magnetic fields. Biochim Biophysica Acta
1282:149–155.
Rosen AD. 2001. Nonlinear temperature modulation of sodium
channel kinetics in GH3 cells. Biochim Biophysica Acta 151:
391–396.
Rosen AD, Lubowsky J. 1990. Modification of spontaneous unit
discharge in the lateral geniculate body by a magnetic field.
Exp Neurol 108:261–265.
Speyer JB, Sripada PK, Das Gupta SK, Shipley GG, Griffin RG.
1987. Magnetic orientation of sphingomyelin–lecithin
bilayers. Biophys J 51:687–691.