sensors
Article
Broadband Electrical Spectroscopy to Distinguish Single-Cell
Ca2+ Changes Due to Ionomycin Treatment in a Skeletal Muscle
Cell Line
Caroline A. Ferguson 1 , Carmen Santangelo 2 , Lorenzo Marramiero 2 , Marco Farina 3 , Tiziana Pietrangelo 2
and Xuanhong Cheng 1,4, *
Citation: Ferguson, C.A.; Santangelo,
C.; Marramiero, L.; Farina, M.;
Pietrangelo, T.; Cheng, X. Broadband
Electrical Spectroscopy to
Distinguish Single-Cell Ca2+ Changes
Due to Ionomycin Treatment in a
Skeletal Muscle Cell Line. Sensors
2023, 23, 4358. https://doi.org/
10.3390/s23094358
Academic Editors: Cristiano Palego
and Klaus Stefan Drese
Received: 24 February 2023
Revised: 22 April 2023
Accepted: 26 April 2023
Published: 28 April 2023
Copyright: © 2023 by the authors.
Licensee MDPI, Basel, Switzerland.
This article is an open access article
distributed under the terms and
conditions of the Creative Commons
Attribution (CC BY) license (https://
creativecommons.org/licenses/by/
4.0/).
Sensors 2023, 23, 4358. https://doi.org/10.3390/s23094358
1 Department of Bioengineering, P.C. Rossin College of Engineering and Applied Sciences, Lehigh University,
Bethlehem, PA 18015, USA
2 Department of Neuroscience, Imaging and Clinical Sciences, University “G. d’Annunzio” Chieti-Pescara,
66100 Chieti, Italy
3 Department of Engineering of Information, University Politecnica delle Marche, 60131 Ancona, Italy
4 Department of Materials Science and Engineering, P.C. Rossin College of Engineering and Applied Sciences,
Lehigh University, Bethlehem, PA 18015, USA
* Correspondence: xuc207@lehigh.edu
Abstract: Many skeletal muscle diseases such as muscular dystrophy, myalgic encephalomyeli-
tis/chronic fatigue syndrome (ME/CFS), and sarcopenia share the dysregulation of calcium (Ca2+ ) as
a key mechanism of disease at a cellular level. Cytosolic concentrations of Ca2+ can signal dysregula-
tion in organelles including the mitochondria, nucleus, and sarcoplasmic reticulum in skeletal muscle.
In this work, a treatment is applied to mimic the Ca2+ increase associated with these atrophy-related
disease states, and broadband impedance measurements are taken for single cells with and without
this treatment using a microfluidic device. The resulting impedance measurements are fitted using a
single-shell circuit simulation to show calculated electrical dielectric property contributions based
on these Ca2+ changes. From this, similar distributions were seen in the Ca2+ from fluorescence
measurements and the distribution of the S-parameter at a single frequency, identifying Ca2+ as
the main contributor to the electrical differences being identified. Extracted dielectric parameters
also showed different distribution patterns between the untreated and ionomycin-treated groups;
however, the overall electrical parameters suggest the impact of Ca2+ -induced changes at a wider
range of frequencies.
Keywords: broadband impedance spectroscopy; circuit modeling; single-cell characterization;
intracellular calcium
1. Introduction
The concentration of calcium (Ca2+ ) ions in the context of muscle cell operation has
several complex and dynamic functions. The Ca2+ within the cytoplasm skeletal muscle
cells has a variety of deeply important functions, including regulating the mitochondrial
function and controlling signaling pathways between cells. In skeletal muscle systems,
Ca2+ has roles in the excitation–contraction (EC) coupling, mitochondrial function, and as a
signal regulator for protein and degradation [1–3]. These functions are carefully controlled
by the rhythmic movement of Ca2+ from storage in the sarcoplasmic reticulum (SR) to the
cytosol, fostered by a variety of well-characterized proteins in skeletal muscle cells [4,5].
As mentioned, Ca2+ plays a continuous role in muscle contractibility and the development
of mitochondrial abnormalities leading to diseases mostly characterized by symptoms
involving loss of muscle control and fatigue [6]. Natural muscle structure changes with
age, named sarcopenia, or as the result of a genetic disease such as Duchenne Muscular
Dystrophy (DMD), or atrophy resulting from cancer cachexia or acquired immunodeficiency
syndrome (AIDS) [4,5,7,8]. In these diseases, the increased cytoplasmic Ca2+ results in the
https://www.mdpi.com/journal/sensors
 changes with age, named sarcopenia, or as the result of a genetic disease such as Du-
chenne Muscular Dystrophy (DMD), or atrophy resulting from cancer cachexia or ac-
quired immunodeficiency syndrome (AIDS) [4,5,7,8]. In these diseases, the increased
Sensors 2023, 23, 4358 cytoplasmic Ca2+ results in the eventual degradation of muscle filaments and proteolytic 2 of 13

pathways that maintain muscle integrity and mitochondrial polarization that helps with
energy management [9]. Therefore, the ability to distinguish Ca2+ changes intracellularly
eventual
have an degradation
implication for of muscle filaments
potential and proteolytic
diagnostic and therapeuticpathways that maintain muscle
abilities.
integrity
Ions and
suchmitochondrial
as Ca2+ andpolarization
other charged thatparticles
helps with canenergy
impact management
the passage [9].ofTherefore,
intracellular
the ability to distinguish Ca 2+ changes intracellularly have an implication for potential
charges and external electrical stimuli. Electrical diagnosis poses a promising method of
diagnostic and therapeutic
characterizing single-cell abilities.
cytoplasmic changes associated with diseases of muscular in-
Ions such as Ca 2+ and other charged particles can impact the passage of intracellular
tegrity and could develop a unique quantitative diagnostic and evaluative tool. Shifting
charges and external electrical stimuli. Electrical diagnosis poses a promising method
to electrical measurement also offers the opportunity to adopt a rapid and label-free ap-
of characterizing single-cell cytoplasmic changes associated with diseases of muscular
proach to future diagnoses. Developing a modeled understanding of how certain specific
integrity and could develop a unique quantitative diagnostic and evaluative tool. Shifting to
changes in disease progression in skeletal muscle cells impact cell impedance lays the
electrical measurement also offers the opportunity to adopt a rapid and label-free approach
groundwork for generating a disease-specific electrical profile that could be both tunable
to future diagnoses. Developing a modeled understanding of how certain specific changes
and predictive.
in disease Broadband
progression spectroscopy
in skeletal muscle cellsmeasures
impact cellcellimpedance
impedance laysover
the agroundwork
wide range of
frequencies that have been prominent in distinguishing single-cell characteristics,
for generating a disease-specific electrical profile that could be both tunable and predictive. espe-
cially in the
Broadband last decademeasures
spectroscopy [10–13]. cell
By studying
impedancea overbroad bandwidth
a wide range ofoffrequencies
frequenciesthat using
spectroscopy, the impedance measurement can enable the characterization
have been prominent in distinguishing single-cell characteristics, especially in the last of different
cellular[10–13].
decade compartments
By studyingbased on thebandwidth
a broad scattering ofof frequencies
the signal and usingdielectric properties
spectroscopy, the of
the compartment.
impedance measurement Biological tissuethe
can enable and cells have dielectric
characterization properties
of different cellularofcompartments
conductivity (σ)
and permittivity
based on the scattering(ε), whose mainand
of the signal intracellular scatteringof
dielectric properties influences are widely
the compartment. expanded
Biological
tissue
upon and cells have
in review dielectric
articles [14–16].properties
Various of conductivity
studies (σ) and permittivity
have identified frequencies (ε), whose for
of interest
main intracellular
disease scattering
identification [17]. influences are widely
At frequencies in theexpanded
megahertz upon in review
range, articles [14–16].
the dominant scattering
Various studies have identified frequencies of interest for disease
method is β-scattering, making the primary signal changes attributable to cytoplasmic identification [17]. At
frequencies
changes atina the megahertz
single-cell range,
level [16].the
Bydominant
combining scattering method is β-scattering,
this understanding of broadband makingmeas-
the primary with
urements signalelectrical
changes attributable to cytoplasmic
circuit modeling, we studychangestheatphysiological
a single-cell level and[16]. By
dielectric
combining
propertiesthis understanding
of two of broadband measurements with electrical circuit modeling,
cell populations.
we study the physiological and dielectric properties of two cell populations.
2. Materials and Methods
2. Materials and Methods
2.1.Cell
2.1. CellCulture
Culture
L6L6rat
ratmyoblast
myoblastcells
cells(ATCC,
(ATCC, CRL-1458)
CRL-1458) were
were cultured
cultured according
according to American
to American Type Type
CultureCollection
Culture Collection(ATCC)
(ATCC)guidelines
guidelines[18].[18]. The
The cells
cells were
were grown
grown in in Dulbecco’s
Dulbecco’s Modi-
Modified
Eagle
fied Medium
Eagle Medium(DMEM)
(DMEM) base (Corning
base (Corning 10014CV,
10014CV, Corning, NY,
Corning, NY,USA)
USA) supplemented
supplementedwith
with
10%10%FBSFBS (Gibco
(Gibco 10082147,Billings,
10082147, Billings, MT, USA)
USA)and and5%5%Penicillin/Streptomycin
Penicillin/Streptomycin (Life(Life
Technologies 15140122, Carlsbad, CA, USA) and incubated at 37 ◦ C with 5% CO . The
Technologies 15140122, Carlsbad, CA, USA) and incubated at 37 °C with 5% 2CO2. The
media
mediawas wasrefreshed
refreshedregularly
regularlyand andcells were
cells weresplit when
split whenconfluence
confluence on the flask
on the surface
flask surface
was
wasachieved.
achieved.InInFigure
Figure1,1,the
thelayout
layoutofof
how howthese adherent
these cells
adherent areare
cells treated andand
treated used for for
used
biological and electrical characterization is outlined.
biological and electrical characterization is outlined.

Figure1.1. Experimental
Figure Experimentallayout
layoutdescribing
describingthethe biological
biological characterization
characterization andand parallel
parallel electrical
electrical
measurements system.
measurements system.

2.2. Ionomycin Treatment

Cells were detached from the culture flasks and subjected to an Ionomycin treatment
protocol previously established [19,20]. For electrical measurements, spun-down and
Sensors 2023, 23, 4358 3 of 13

detached cells were suspended in 1 µM Ionomycin (MilliporeSigma, Bedford, MA, USA,

56092-81-0) in a solution of phosphate-buffered saline (PBS) with Ca2+ (1.5 mM). The cells
remained in this solution for 10 min at room temperature before rinsing with Ca2+ -free PBS
and preparation for electrical measurement in a solution of 8.5% sucrose and 0.3% dextrose.
Similar steps were conducted on adherent cells after the introduction of the Ca2+ imaging
agents, as shown in Figure 1.

2.3. Ca2+ Imaging

For measurements of the intracellular calcium concentration ([Ca2+ ]i ), the cells were
seeded in a 96-well plate in the culture media described previously at a density of 1000 cells
per well for three days at 37 ◦ C with 5% CO2 . The cells were incubated with physiological
solution containing (mM) NaCl 140, KCl 2.8, CaCl2 2, MgCl2 2, glucose 10, Hepes–NaOH 10,
pH 7.4, named normal external solution (NES) containing 10 mg mL−1 bovine serum albu-
min and 5 µM Fura-2 AM (Molecular Probes, Eugene, OR, USA) for 45 min. Then, the cells
were rinsed by incubation in NES with Ca2+ for 40 min at 37 ◦ C to allow de-esterification
of the dye inside the cells. Ca2+ imaging was performed using epifluorescence microscopy
that assists in the acquisition of the Fura-2 emission fluorescence at both an excitation of
340 nm and then at an excitation of 380 nm. These excitations and measurements were
accomplished using a Xenon lamp and monochromator wavelength switcher Polychrome
II (Till Photonics, Gräfelfing, Germany) to isolate the wavelength of interest. The emission
fluorescent images were collected using a 40× oil objective lens, acquired by an intensified
CCD camera or a C6790 cooled CCD camera (Hamamatsu Photonics, Hamamatsu, Japan),
digitized and stored in an interfaced computer. The software Aquacosmos (version 2.0)
was used to analyze the resulting emission fluorescence ratio (f[340/380]) on selected repre-
sentative cellular areas after the subtraction of background fluorescence. After an initial
baseline was established, 1 µM Ionomycin in 1.5 mM Ca2+ was added for 10 min while
measuring fluorescence. These measurements were accomplished in a temporal sequence
generating spectral measurements showing the cell fluorescence ratio prior to ionomycin
treatment, during treatment of 10 min, following a wash Ca2+ -free PBS, and after incubation
in an 8.5% sucrose and 0.3% dextrose solution. The mean initial values and post-treatment
values were calculated using 25 time points in the initial plateau and the plateau following
the wash, which were used to calculate the ratio. The acquisition time for each fluorescence
emission was 0.5 s and the overall imaging lasted 1.5 h. Additional experiments were
performed using sucrose solution with 10 mM EGTA, a calcium chelator, in order to prevent
any possible external Ca2+ interference post-ionomycin treatment, showing similar results.

2.4. Electrical Measurement

Prior to cell preparation or treatment, a coplanar waveguide (CPW) electrode device
was assembled by attaching a polydimethylsiloxane (PDMS) channel to the CPW surface
aligned with a series gap. Further details on the CPW and channel design can be found
in previous publications [21–23] and a simplified schematic can be seen in Figure 2. The
assembled device was placed on a microscope stage and an isotonic solution of 8.5%
sucrose and 0.3% dextrose in deionized water was passed through to test the assembly
seal prior to the addition of samples. Cells treated with ionomycin and washed with
Ca2+ -free PBS were suspended in the same sucrose and dextrose solution at a concentration
of 500,000 cells/mL and a sample was taken up in a syringe for injection into the channel
assembly. Based on previous work, the cells remain their integrity for up to an hour,
allowing continuous measurement.
Sensors 2023,
Sensors 23, 23,
2023, 4358x FOR PEER REVIEW 4 of 13
4 of 13

Figure 2. Schematics of (a) coplanar waveguide electrodes and PDMS channel, (b) background
Figure 2. Schematics of (a) coplanar waveguide electrodes and PDMS channel, (b) background circuit
circuit model, and (c) single-shell circuit model.
model, and (c) single-shell circuit model.
Once all components were in place, cells were introduced to the channel at a sample
Once all components were in place, cells were introduced to the channel at a sample
flow rate of 0.1 µL/min along with a stream of sheath sucrose at a rate of 0.2 µL/min. An
flow rate of 0.1 µL/min along with a stream of sheath sucrose at a rate of 0.2 µL/min.
electrical dielectrophoretic (DEP) signal of −9 dBm at a frequency of 500 MHz was in-
An electrical
troduced dielectrophoretic
between the series gap (DEP)
in the signal of −9to
electrodes dBm
trapatthe a frequency
cell. Then, of 500 MHz was
a broadband
introduced between the series gap in the electrodes
sweeping signal of −15 dBm from 9 kHz to 9 GHz was applied to measure to trap the cell. Then, a broadband
the
sweeping of −
S-parameters with and without the cell present using a virtual network S-parameters
signal 15 dBm from 9 kHz to 9 GHz was applied to measure the analyzer
with and without
(Keysight E5080A, the Santacell present
Rosa, using The
CA, USA). a virtual network
S-parameters analyzer
include (Keysightcoeffi-
the reflection E5080A,
Santa
cient (S11) and the transmission coefficient (S21). S11 is representative of the powerand
Rosa, CA, USA). The S-parameters include the reflection coefficient (S11) re- the
transmission
flected back coefficient
to the power (S21).
inputS11port
is representative
1 by the system, of the power
while S21 reflected back to the
is representative power
of the
input
power port
that1 passes
by the system,
through while S21 istorepresentative
the system of the
port 2 [24]. These power that passes
measurements through the
were repeated
system to port
after the 2 [24].
release These
of the measurements
previously trappedwere repeated
cell to determine afterthe
thebackground
release of the previously
measure-
trapped
ment of cell to determine
the sucrose solutiontheforbackground measurement
these parameters. of the sucrose
The differential solution
S-parameters withfor
andthese
parameters.
without the cell Thepresent,
differential
ΔS11S-parameters with and
and ΔS21, represent the without the cell to
cell contribution present, ∆S11 and
power reflec-
∆S21,
tion and transmission.
represent the cell The S-parameter
contribution spectrareflection
to power as a function
andof sweeping frequency
transmission. were
The S-parameter
furtheras
spectra fitted to a circuit
a function model to frequency
of sweeping extract impedance associated
were further fittedwith
to a the cell.model to extract
circuit
The raw
impedance signals ofwith
associated ΔS11theandcell.
ΔS21 taken from the Keysight network analyzer for each
cell The
at allraw
201signals of ∆S11 and ∆S21
frequencies appear as spectra.
takenUsing theKeysight
from the impedance measurements
network analyzerfor forthese
each cell
spectra, curve fitting using simulated circuit modeling was used to obtain
at all 201 frequencies appear as spectra. Using the impedance measurements for these spectra, dielectric parame-
ters. fitting using simulated circuit modeling was used to obtain dielectric parameters.
curve

2.5.Simulated
2.5. Simulated Circuit
Circuit Modeling
Modeling
Simulation of
Simulation of aa single-shell
single-shell circuit
circuitmodel
modelwas
wasperformed
performed to to
find thethe
find optimal fit offitthe
optimal of the
cytoplasm resistance and capacitance for each electrically measured cell using
cytoplasm resistance and capacitance for each electrically measured cell using Advanced Advanced
DesignSystem
Design System (ADS,
(ADS, version
version 2023.01)
2023.01)software
softwarebybyKeysight.
Keysight. TheThe
circuit designs
circuit for the
designs for the
simulations can be seen in Figure 2b,c. For each background file, variables
simulations can be seen in Figure 2b,c. For each background file, variables were tunedwere tuned to
fit the properties of the electrode system and solution in a representative circuit
to fit the properties of the electrode system and solution in a representative circuit to to de-
termine the best values to carry on matching the cell itself. Using the optimal background
determine the best values to carry on matching the cell itself. Using the optimal background
characteristics, a circuit modeling the cell layers of membrane and cytoplasm as resis-
characteristics, a circuit modeling the cell layers of membrane and cytoplasm as resistor-
tor-capacitor pairs was used to find the best estimate of the cytoplasm properties. Based
capacitor pairs was used to find the best estimate of the cytoplasm properties. Based on
on previous measurements of L6 cells, the membrane resistance and capacitance were
previous measurements of L6 cells, the membrane resistance and capacitance were kept
kept constant at 235 MOhm and 34.5 pF [25]. The fitting process was automated by
constant at 235 MOhm and 34.5 pF [25]. The fitting process was automated by treating
treating the file number as a variable for each matching pair of background and cell
the file number as a variable for each matching pair of background and cell measurement
files in lists. The goal weights described in Table 1 were assigned by making a shared goal
for each optimization and signal type and minimizing the error to fit all frequencies and a
Sensors 2023, 23, 4358 5 of 13

region of interest. The applied weights were determined by comparing the error generated
by fitting a random smaller subset of spectra to best represent the background and spectra
shape (Figures A1 and A2). The chosen weight values to reduce the overall fitting error for
the differential delta (∆) spectra, which ended up being 5 times the background weights
(Figure A3). The more specific and heavily weighted goals were applied to frequency
ranges where the treated and untreated cells have the largest distinction in the S parameter
spectra. These frequency ranges match theoretical understanding of the frequencies in
which cytosol contribution dominates [26]. In addition, an overall weight increase was
used to value the fit of ∆S11 to compensate for the smaller measured values.

Table 1. Goal and weight parameters (W) assigned to automated simulation circuit models in
Advanced Design System (ADS) software and variables tuned in each optimization.

Circuit Sweep Goal 1 Sweep Goal 2

Optimization Variable
Design Target Freq W Target Freq W
Cg
∆S11 9 kHz–9 GHz 1 ∆S21 9 kHz–9 GHz 1
Cs
Optim 1 Background R0
∆S11 0.9 MHz–9 GHz 2 ∆S21 9 kHz–9 MHz 2 Z0
Zg
∆S11 9 kHz–9 GHz ∆S21 9 kHz–9 GHz Rc
Optim 2 Single-Shell 5 1
∆S11 0.9 MHz–9 GHz ∆S21 9 kHz–9 MHz Cc

For each pair, the background variables were tuned by minimizing the difference of
the ranges described in Table 1 by applying the described goal and weight values. The fit
variables were then applied to fit the cell circuit model based on the remaining goals to find
Rc and Cc . From the extracted variables, the dielectric permittivity (σ) and conductivity (ε)
of the cytoplasm were calculated using Equations (1) and (2), as seen in previous single-shell
fittings [21].
2Cc
εc = (1)
3πr0
2
σc = (2)
πr0 Rc
In this single-shell model, the average cell radius as determined by microscopic image
analysis (r0 ) helps calculate the cytoplasmic dielectric permittivity (εc ) and conductivity (σc ).

2.6. Statistical Analysis

Statistical analysis was performed using the python (3.11.2) package scipy.stats (v1.10.1)
to compare means using an unpaired Students’s t-test. Statistical comparison of distribu-
tions was accomplished using a Kolmogorov–Smirnov test. For both, a p-value of less than
0.05 was considered statistically significant.

3. Results
To determine the Ca2+ change, the fluorescence was monitored over time and the
average f[340/380] in the periods prior to ionomycin addition and after the maximum de-
termined experimental length of 1.5 h, based on viability in sucrose, as shown in Figure 3a.
Although the cell slowly returns to the initial fluorescence level, there was still an elevation
seen in many of the cells. After measuring these average values pre- and post-treatment for
each individual treated cell, the distribution of was analyzed to determine how treatment
changed the f[340/380] or relative Ca2+ concentration, which is shown in Figure 3b. The
distribution of f[340/380] shows a slight peak shift higher after treatment with ionomycin
 Sensors 2023, 23, x FOR PEER REVIEW 6 of 13

Sensors 2023, 23, 4358 6 of 13

post-treatment for each individual treated cell, the distribution of was analyzed to de-
termine how treatment changed the f[340/380] or relative Ca2+ concentration, which is
shown
and in Figure
a notably 3b. The
wider distribution
distribution of f[340/380]
spread. The mean shows a slight peak
and standard shift higher
distribution after
of f[340/380]
signals from pre- and post-ionomycin treatment cells are 0.083 ± 0.008 and 0.084and
treatment with ionomycin and a notably wider distribution spread. The mean ± 0.025,
standard distribution
respectively. While the of f[340/380]
fluorescent signals from
signal pre- is
mean and post-ionomycin
only slightly higher treatment
for thecells
popula-
are 0.083
tion after± the
0.008 and 0.084by
treatment ± 0.025, respectively.
ionomycin, thereWhile the fluorescent
is a significant signal
change in mean is onlyof the
the shape
slightly higher
distribution for the
itself (p =population after the treatment
0.012). However, looking atbythe ionomycin,
distribution thereinisthe
a significant
middle of the
change in the shape of the distribution itself (p = 0.012). However, looking at the distri-
time period during which electrical measurement occurs, there is a significant increase in
bution in the middle of the time period during which electrical measurement occurs,
mean to 0.102 ± 0.041 (p < 0.0009) and a significant shift in the distribution (p < 0.0008).
there is a significant increase in mean to 0.102 ± 0.041 (p < 0.0009) and a significant shift in
This indicates that although the cells return eventually to the original Ca2+ level, there is
the distribution (p < 0.0008). This indicates that although the cells return eventually to the
an increase2+for the period wherein measurement occurs. By treating L6 skeletal muscle
original Ca level, there is an increase for the period wherein measurement occurs. By
cells with ionomycin for a short period of time, Ca2+ mobility led to an extended increase
treating L6 skeletal2+ muscle cells with ionomycin for a short period of time, Ca2+ mobility
in
led to an extended increase in cytoplasmic Ca2+ by shifting the necessary electrical po- [19].
cytoplasmic Ca by shifting the necessary electrical potential for gated channels
Based
tential on
for this
gatedprevious
channelswork, membrane
[19]. Based on thishyperpolarization
previous work, membraneis confined to the first three
hyperpolariza-
minutes of treatment wherein ionomycin can evoke transient
tion is confined to the first three minutes of treatment wherein ionomycin can evoke outward currents such as
Ca 2+ -activated +
K currents
current (IK
transient outward Ca ),as
such asCa
demonstrated
2+-activated Kin our previous
+ current (IKCa),paper [19]. However,
as demonstrated in the
previous
our previous paper [19]. However, the previous work also found that stable membrane min.
work also found that stable membrane potential is re-established within 10
Due to this
potential and the external
is re-established withinaddition
10 min. Dueof calcium,
to this andthethe
slight shiftaddition
external in distribution
of calcium, results
primarily
the slight shift in distribution results primarily from the internalization of externally shift
from the internalization of externally available calcium. Despite the visible
in distribution,
available calcium. there is still
Despite a significant
the visible shiftoverlap, something
in distribution, therealso seen
is still in the distribution
a significant over- of
later electrical measurements.
lap, something also seen in the distribution of later electrical measurements.

Figure 3.
Figure 3. (a)
(a) Example
Example f[340/380]
f[340/380]curve
curveover
overtime
timewith
withhighlighted regions
highlighted of of
regions interest and
interest (b)(b)
and nor-
normal-
malized
ized distribution
distribution of Caof2+Ca
2+-related fluorescence in pre-treatment (Initial) and post-ionomycin
-related fluorescence in pre-treatment (Initial) and post-ionomycin treatment
treatment cells at both the 30 min mark (Post-Treatment Mid) and 1 h mark (Post-Treatment Fin), as
cells at both the 30 min mark (Post-Treatment Mid) and 1 h mark (Post-Treatment Fin), as determined
determined by f[340/380] (n = 89).
by f[340/380] (n = 89).
Following the establishment of the ionomycin treatment as a method of increasing
Following the establishment of the ionomycin treatment as a method of increasing the
the intracellular2+Ca2+ in the L6 muscular cell line, broadband impedance spectroscopy
intracellular
was applied Ca as a in the L6for
method muscular
rapid cellcell capture
line, broadband impedance spectroscopy
and characterization of intracellularwas ap-
plied
properties. The collection of spectra for each cell was examined, then compiled for statis- The
as a method for rapid cell capture and characterization of intracellular properties.
collection of spectra
tical comparison formean
of the each curves
cell was andexamined,
deviations then compiled
in signals for statistical
among treatmentcomparison
types, as of
the mean curves and deviations in signals among treatment types,
shown in Figure 4. In this figure, the mean solid line is accompanied by a lighter shading as shown in Figure 4.
In
region showing the deviation between electrically measured cells at each frequency. the
this figure, the mean solid line is accompanied by a lighter shading region showing
deviation between
From this figure, electrically
while the overlapmeasured
betweencells at each
groups frequency.
is present at mostFrom this figure,
frequency values,while
the overlap between groups is present at most frequency values,
several distinct regions distinguish untreated L6 cells (UNT) and the ionomycin-treated several distinct regions
distinguish untreated
cells (TRT) including L610cells
the MHz (UNT)
to 1 GHz andrange
the ionomycin-treated
for the ΔS11 signal cells
(Figure(TRT) including
4a) and the 1 the
10
MHzMHz to MHz
to 10 1 GHzrange
range forfor
thethe ∆S11
ΔS21 signal
signal (Figure
(Figure 4b).4a) andranges
These the 1 MHz to 10 MHz
are relevant to therange
for the ∆S21
cytoplasm signal (Figure
properties 4b). the
considering These rangesproperties
scattering are relevant to the
of most cytoplasm
biological properties
materials.
At these frequencies,
considering β-scattering
the scattering is thought
properties to dominate
of most biological thematerials.
electrical signal [16].frequencies,
At these While
visually distinct
β-scattering in severaltoareas,
is thought most notably
dominate the MHz
the electrical region,
signal theWhile
[16]. differences are better
visually distinct in
several areas, most notably the MHz region, the differences are better characterized and
explained by the use of circuit modeling to extract the compartment dielectric changes.
 Sensors 2023, 23, x FOR PEER REVIEW 7 of 13

Sensors 2023, 23, x FOR PEER REVIEW 7 of 13

Sensors 2023, 23, 4358 7 of 13

characterized and explained by the use of circuit modeling to extract the compartment
dielectric changes.
characterized and explained by the use of circuit modeling to extract the compartment
dielectric changes.

Figure4.4.Average
Figure Average (a) ∆S11 and
(a) ΔS11 and(b) ∆S21spectra
(b)ΔS21 forfor
spectra untreated
untreatedL6 L6
cells (UNT)
cells andand
(UNT) cellscells
treated with with
treated
1.5
1.5 µM Ionomycin
µM Ionomycin (TRT) to increase cytoplasmic Ca
2+ 2+ for 10 min. Shadows indicate deviation
Figure 4. Average (TRT) to and
(a) ΔS11 increase cytoplasmic
(b) ΔS21 Cauntreated
spectra for for 10 min. Shadows
L6 cells (UNT) indicate
and cells deviation among
treated with
among measured cells while solid lines indicate the mean spectra (UNT, n = 52; TRT, n = 21).
1.5 µM Ionomycin
measured cells while(TRT)
solid to increase
lines cytoplasmic
indicate the mean Ca 2+ for 10 min. Shadows indicate deviation
spectra (UNT, n = 52; TRT, n = 21).
among measured cells while solid lines indicate the mean spectra (UNT, n = 52; TRT, n = 21).
The overlap due to the subtle differences between the UNT and TRT populations can
The overlap due to the subtle differences between the UNT and TRT populations can
also The
be visualized
overlap due from thesubtle
to the impedance
differences measurements
between the similarly
UNT andto TRT
the fluorescent
populations Ca 2+
canCa2+
also be visualized from the impedance measurements similarly to the fluorescent
distribution,
also be visualized shown in Figure 3. Although the duration of electrical measurement is
distribution, shownfrom the impedance
in Figure 3. Although measurements
the durationsimilarly to the
of electrical fluorescent is
measurement Calonger
2+
longer than
distribution, that of
shown calcium
in imaging,
Figure 3. there
Although was theno observed
duration signal
of drift
electrical when comparing
measurement is
than that of calcium imaging, there was no observed signal drift when comparing cells
cells
longermeasured
than atofthe beginning or end of the hour, indicating leakage current, if exists, is
measured at that calcium
the beginning imaging,
or end ofthere was
the hour, no observed
indicating signal
leakage drift whenif comparing
current, exists, is weak
weak after
cells cell
measured cell wash.
atThus, Thus,
the beginning the influence
or end of leakage current on cytoplastic impedance
after
during wash.
the sensing the influence
period is the
neglected. ofofleakage
the hour,
Looking
indicating
current
at one
leakage current,
on cytoplastic
relevant frequency
if exists,during
impedance
point at 190
is
weak
the after
sensing cell
periodwash. Thus,
is neglected. influence
Looking of leakage
at one current
relevant on cytoplastic
frequency point impedance
at 190the MHz,
MHz,
during the
theΔ sensing
S11 signal distribution
period looks Looking
is neglected. similar toattheone spread seenfrequency
relevant in Figure 5, where
point at the
190TRT
the
TRT∆ S11 signal distribution looks similar to the spread seen in Figure 5, where
MHz,population
the Δ S11 signal has a distribution
wider distribution, specifically
looks similar in the increasing
to the spread seen in Figure signal direction.
5, where the
population
The UNT has a wider
population hasdistribution,
a ΔS11 signalspecifically
of 3.5 ± 3.3in× the
10 −4 increasing
compared signal
to the direction.
TRT The
population, UNT
TRT population has a wider distribution, specifically − in the increasing signal direction.
population
which ahas
hadpopulationa ∆S11 signal
significantly of 3.5 ± 3.3 × 10 4
4.5××compared whento the TRTatpopulation, MHz (pwhich
The UNT has ahigher signal
ΔS11 signal of 3.5
of 8.1 ±±3.3 10−4−4compared
410
measured
to the TRT 190population, =
had a significantly
0.021). At this higherthere
frequency, signal is of
a 8.1 ± 4.5 ×
significant 10−in
shift when measured
distribution at
mirroring 190 MHz
the Ca (p
2+ = 0.021).
shift,
which had a significantly higher signal of 8.1 ± 4.5 × 10−4 when measured at 1902+MHz (p =
At this frequency,
which indicates a there is a significant
sensitivity toisionic shift in(pdistribution
changes = 0.017). mirroring
In the MHz the Ca
range, theCa shift, which
electrical
0.021). At this frequency, there a significant shift in distribution mirroring the 2+ shift,
indicates
signal cana sensitivity
penetrate to cell
the ionicmembrane
changes (p = measurements
for 0.017). In the MHz of therange, the electrical
cytoplasm, and in signal
the
which indicates a sensitivity to ionic changes (p = 0.017). In the MHz range, the electrical
can penetrate
lower MHz the
range, cell
there membrane
is minimal for measurements
measurement of
interference
signal can penetrate the cell membrane for measurements of the cytoplasm, and in the
the cytoplasm,
from organelle and in
structurethe lower
or
MHz
lower MHz range, there is minimal measurement interference from organelle structure orlarge
large range,
molecules there is
within minimal
the measurement
cell, making 190 interference
MHz an ideal from organelle
frequency to structure
look at the or
die-
lectric
moleculesproperties
large molecules ofthe
withinwithin thecell,
cytoplasm
the making itself
cell, making 190[15,16].
MHz
190 MHzan ideal
an ideal frequency
frequency to tolook
lookatatthethedielectric
die-
properties of the cytoplasm
lectric properties itself [15,16].
of the cytoplasm itself [15,16].

Figure 5. Normalized Δ S11 signal distribution for both untreated (UNT) and ionomycin-treated
(TRT) cells when measured at 190 MHz.
Figure5.5. Normalized
Figure ∆ S11
Normalized Δ S11 signal
signaldistribution
distributionforfor
both untreated
both (UNT)
untreated andand
(UNT) ionomycin-treated
ionomycin-treated
(TRT) cells when measured at 190 MHz.
(TRT) cells when measured at 190 MHz.
To better understand the implications of the collected ΔS11 and ΔS21 spectra, a sin-
gle-shell circuit model fitsthe theimplications
electrical properties of the cytoplasm resistance (Rc) and
Tobetter
To better understand
understand the implications of
of the
the collected ∆S11and
collectedΔS11 ∆S21spectra,
andΔS21 spectra,a asin-
single-
capacitance
gle-shell (C
circuitc). The circuit model relies on the assumption that the induced increase in
model fits the electrical properties of the cytoplasm resistance (R ) and
shell circuit model fits the electrical properties of the cytoplasm resistance (Rc ) and ca- c
Ca 2+ has no effect on the membrane characteristics or the membrane returns to its typical
capacitance
pacitance
homeostasis
(C(C).c).The
Thecircuit
c quickly.
circuit model
model relies
Themembrane
assumption
relies on
onthe
theassumption
is supported
assumption
by the
that thethe
that induced
induced
gradual returns
return of
increase in in
increase
2+ over
Catypical
Ca
Ca
2+ has no effect
2+ on the characteristics or the membrane to its
timehas
homeostasis
no effect
to the initial on
quickly.
the membrane
concentration
The assumption
characteristics
as theis homeostasis
or the membrane
supported byprocesses
returns
regulate
the gradual return
to its typical
theofindividual
Ca2+2+ over
homeostasis quickly. The assumption is supported by the gradual return of Ca over time
totime
the to the concentration
initial initial concentration
as theas the homeostasis
homeostasis processes
processes regulate
regulate the individual
the individual cells. By
applying the model to each curve, the way properties impact the spectra can be examined.
In the spectra found in Figure 6, representative cells from each treatment group are shown
with simulations closely match the spectral shapes of the experimentally generated curves.
The extracted values of these cells are roughly 0.3 MOhm for Rc for both treatment types
 Sensors 2023, 23, x FOR PEER REVIEW 8 of 13
Sensors 2023, 23, x FOR PEER REVIEW 8 of 13

Sensors 2023, 23, 4358 cells. By applying the model to each curve, the way properties impact the spectra can be8 of 13
cells. By applying
examined. In the the model
spectra to each
found curve, the
in Figure way propertiescells
6, representative impact theeach
from spectra can be
treatment
examined. In the spectra found in Figure 6, representative cells from
group are shown with simulations closely match the spectral shapes of the experimen- each treatment
groupgenerated
tally are shown with simulations
curves. The extracted closely
valuesmatch
of thesethe cells
spectral shapes of
are roughly 0.3the experimen-
MOhm for Rc
and
for both treatment types and 3.89 fF and 3.76 fF for the Cc for the UNT and TRTWhile
3.89
tally fF and
generated 3.76 fF
curves. for
The the C
extractedc for the
values UNT
of theseand TRT
cells arecells, respectively.
roughly 0.3 MOhm for Rc the
cells,
differences
for both were
treatmentminimal
types in
andthe
3.89circuit
fF values
and 3.76 themselves,
fF for the C the
for dielectric
the UNT
respectively. While the differences were minimal in the circuit values themselves, the di-
c parameters
and TRT equate
cells,
toelectric
have significantly
respectively.
parameters different
While equate
the fitting
differences
to have values.
were minimal
significantly in the circuit
different fitting values
values.themselves, the di-
electric parameters equate to have significantly different fitting values.

Figure6.6.Experimentally
Figure Experimentallycollected
collectedspectra
spectra(solid)
(solid)and
andsimulated
simulatedspectra
spectra(dashed)
(dashed)for
forboth
bothionomycin-
iono-
Figure 6. Experimentally
mycin-treated collectedcells
cells and untreated spectra
for (a)(solid)
ΔS11 and (b)
simulated
ΔS21. spectra (dashed) for both iono-
treated cells and untreated cells for (a) ∆S11 and (b) ∆S21.
mycin-treated cells and untreated cells for (a) ΔS11 and (b) ΔS21.
Using the
Using the single-shell
single-shell model
modeland andEquations
Equations (1)(1)
andand (2),(2),
thethefit fit
resistance
resistance (Rc)(R and
c ) and
Using the
capacitance (C c)single-shell
of the modelcan
cytoplasm andalso
Equations
be (1) and
extended to (2), the fit resistance
calculations of the (Rc) and
cytoplasmic
capacitance (Cc ) of the cytoplasm can also be extended to calculations of the cytoplasmic
capacitance (C
permittivity (εcc)) of
andtheconductivity
cytoplasm can also be extended to calculations of thethecytoplasmic
permittivity (ε ) and conductivity(σ(σc).c ).Based
Based on the extracted
on the extracted parameters,
parameters, σc for the
the σ for the
permittivity
UNT and TRT (εcc)populations
and conductivity (σc). Based
were 0.1929 on the
± 0.0002 S/m extracted
and 0.1933 parameters,
± 0.0008 the S/m, forc the
σc respec-
UNT and TRT
UNT and
populations
TRT smallpopulations
were 0.1929 ±±0.0002
were 0.1929
0.0002S/m S/m and ± 0.0008
0.1933 0.0008 S/m, respec-
tively. Though in magnitude, the TRT cells had aand 0.1933 ± higher
significantly S/m, respec-
conductivity
tively.
tively. Though small in magnitude, the TRT cells had a significantly higher conductivity
(p = 8.3Though
× 10−5). small in magnitude,
The extracted the TRT
εc of UNT cellscells
washad11.3a ±significantly
0.34 ε0 andhigher had a conductivity
significantly
(p(p==8.3 × 10−−55).).The Theextracted
extractedεcεof c of UNT cells was 11.3 ± 0.34 ε0 and had a significantly
lower8.3(p×=10 0.0063) average of 10.9 ±UNT
0.13 cells
ε0 forwas
TRT.11.3 ± 0.34
Looking εat
0 and
the had a significantly
distribution of ex-
lower (p =
lower (pσc=and0.0063)
0.0063) average of 10.9 ± 0.13 ε 0 for TRT. Looking at the distribution ofofextracted
tracted εc inaverage of there
Figure 7, 10.9 ±are 0.13
alsoε0 significant
for TRT. Looking
differences at the distribution
in the distribution ex-
of
and
σfitted
c
tracted ε in Figure 7,
cσc and εc inbetween
parameters
there are also
Figure 7,untreatedsignificant
there are also differences
and significant
in the
differences
ionomycin-treated
distribution
in the
cells. distributionparam-
of fitted
Proportionally, ofa
eters
fittedbetween
higher parameters
percentageuntreated
between
of anduntreated
treated ionomycin-treated
cells have and cells.
largerProportionally,
ionomycin-treated
a slightly conductivity cells. anda higher
Proportionally,
a smaller percentage
per-a
ofmittivity
treated
higher cellsthe
percentage
than have a slightly
of treated
untreated cells
cell larger
have conductivity
a slightly larger
distribution. and a smaller and
conductivity permittivity
a smaller thanper- the
untreated
mittivity thancell distribution.
the untreated cell distribution.

Figure 7. Normalized distribution of calculated single-shell model parameters for the (a) cyto-
Figure7.7.
plasmic
Figure Normalized(σ
conductivity
Normalized distribution ofofcalculated
c) and (b) relative
distribution calculated single-shell
cytoplasmic model
permittivity
single-shell model c).parameters
(εparameters forforthethe
(a)(a) cyto-
cytoplasmic
plasmic conductivity (σ c) and (b) relative cytoplasmic permittivity (εc).
conductivity (σc ) and (b) relative cytoplasmic permittivity (εc ).
4. Discussion
4.4.Discussion
Discussion
In Figure 3b, the distribution of the fluorescence related to calcium peaks at a higher
In
ratioInandFigure
Figure 3b, wider
3b,
spreads the distribution
distribution ofofthe
after Ionomycin thefluorescence
fluorescence
treatment, related to
related
consistent calcium
to calcium
with peaks at
of acalcium
peaks
an increase higher
at a higher
ratio
ratio
due andand spreads wider
spreads Basedafter
to treatment. after Ionomycin
Ionomycin
on the treatment,
known actiontreatment, consistent
consistent
of ionomycin with
aswithan increase of calcium
an increase ofiono-
an electroneutral calcium
duetototreatment.
phore,
due treatment.
the cells areBased onthe
expected
Based on the knownaction
toknown
rapidly action ofofionomycin
transport ionomycin asasan
extracellular anelectroneutral
electroneutral
calcium inside, theniono-
re-
ionophore,
phore,
cover
the cells the
are cells
normal are expected
cytosolic
expected totransport
rapidly
toconcentrations
rapidly transport
after the removalextracellular
extracellular of calciumthen
the ionomycin
calcium inside, inside,
[27,28]. then normal
Another
recover re-
cover normal cytosolic
common concentrations
cytosolic concentrations
method of monitoring after
after theCaremoval
2+ the removal
manipulation of the ionomycin
within cells[27,28].
of the ionomycin [27,28].
is nanosecond Another
Anotherpulsed
common
common
method of method
monitoringof monitoring Ca2+ manipulation
Ca2+ manipulation within cellswithin cells is nanosecond
is nanosecond pulsed
pulsed electrical fields
(nsPES), which deliver a small electrical stimulant to generate pores in the cell plasma
membrane, depending on the signal magnitude and polarity [29]. Recent work in this
field to look at Ca2+ increase within cells has shown that the presence of sucrose molecules
can delay the swelling response typically associated with an increase in intracellular Ca2+ ;
however, this can depend on the external concentration introduced and the voltage-gated
Sensors 2023, 23, 4358 9 of 13

channels present on the cell type of study [30,31]. As with our study, these show that
while established in creating a gradient to increase cytosolic Ca2+ , there are less understood
effects on deeper membranes and compartments within the cell after treatment [32]. While
the calcium change is the main factor associated with ionomycin treatment, several long
term changes can be noted, including expression of IL-6 [33,34] or CAIII [35]; however,
these are unlikely to cause an impact in the short term. Due to the rhythmic nature of Ca2+
function in cells, regulation of levels in the cytosol is carefully controlled by several proteins
through storage and release in the sarcoplasmic reticulum. Based on this crucial system, it is
anticipated that while the overall distribution increased, some cells returned to fluorescence
consistent with their initial values. The rapid peak of fluorescence and accompanying
plateau is consistent with previous work observing the rapid flux of cytosolic calcium
before slower recovery after treatment with ionomycin [36].
Similarly, looking at the electrical measurements, the increase is consistent with cy-
tosolic Ca2+ increase and the extracted values are consistent with previous measurements
of similar cellular changes. The values from the UNT cells are comparable in scale to the
previously published fitting value of 0.22 S/m and 9.49 ε0 [21]. The statistical comparison
between these values for the UNT (n = 51) and TRT (n = 20) cells shows that a significant
increase exists for conductivity and a significant decrease exists for permittivity. The small
magnitude of change is consistent with the expected pattern of recovery after the removal
of ionomycin transport complexes. For the maintenance of cell viability, the change in
calcium while established would be minimal in a healthy population. The observed change
in conductivity and permittivity is consistent with an increase in ions, making a more even
distribution of charge throughout the cell cytoplasm. It has been noted earlier that based
on the initial Ca2+ imaging data, not every ionomycin-treated cell sustains the increased
intracellular Ca2+ concentration, leading to the significant overlap between the UNT and
TRT groups. It is also important to note that while this work focuses on the cytosolic
changes, the homeostasis of calcium management also occurs in the sarcoplasmic reticulum.
However, due to the broadband nature of the reported measurements, changes in other
compartments are captured in a wide range of frequencies.
The previous work posits oxidative damage to the mitochondria and flooding of
intracellular reactive oxygen species (ROS) and Ca2+ prior to the start of apoptosis when
exposed to chronic oxidative stress [37]. Physiologically, there is a well-established link
between Ca2+ levels in skeletal muscle and the ability to maintain the balance of ROS and
mitigate the effects of oxidative stress. Not surprisingly, the spectra alteration observed
here from intracellular Ca2+ elevation, an increase in ∆S11 in the MHz range followed by a
dip in the GHz range and a less negative value of ∆S21 in the kHz range, is comparable
to those from L6 cells exposed to long-term oxidative stress [21]. The previous work
identified Ca2+ as a key factor differentiating cells that experienced exposure to oxidative
stress, while in this work, the ability to differentiate calcium through modelling shows
the milder but noticeable contributions Ca2+ makes to the internal dielectric properties.
Alternatively, work using the ratio of impedance at 1 MHz to 300 kHz to characterize
individual cell opacity also showed the ability to measure the changes in neutrophils due
to calcium ionophore exposure [38]. This work has a higher throughput and therefore,
sample size; however, it is limited to blood cells and limited to analysis of size and opacity
to characterize the populations studied. By measuring a full spectrum of frequency values
rather than relying on a smaller number of frequencies, this work shows the capture of
dielectric properties representing complex and multifaceted changes due to elevated Ca2+
levels induced by ionomycin.
The ability to identify cytoplasmic Ca2+ changes has the potential to aid in improving
our understanding of many associated skeletal muscle diseases such as DMD, cachexia,
and the process of sarcopenia development [2]. Considering that earlier published work
also showed that calcium influx is an important factor in the way long-term oxidative
stress changes the electrical properties of muscle cells, the results from this study are not
meant to differentiate between calcium mismanagement and oxidative stress, but rather
Sensors 2023, 23, 4358 10 of 13

explain the contribution calcium mismanagement might have towards the oxidative stress
responses previously observed. This work is also limited by the selectivity of the electrical
spectroscopic method to define specific molecular or ionic contributors with certainty. Due
to the complexity of ion management inside cell models, while the treatment is intended
to alter intracellular Ca2+ concentration, the effects of other ions or molecules may also
contribute to the observed changes in the electrical signal. This limitation of ion sensing
selectivity has been reported in an aqueous solution at high frequencies [39,40]. Because
electrical measurements are widely non-specific and oxidative diseases have complex
and multifaceted effects on muscle cells, the objective is to broaden understanding of
how these effects manifest electrically. As ionomycin treatment is known to alter ion
concentration inside cells without inducing oxidative stress, this study allows us to focus
the manifestation of calcium imbalance in cellular impedance. Additionally, while this
work focuses on skeletal muscle, the potential to monitor neuron resting and altered Ca2+
concentration have implications for many more diseases [41,42]. Typically, measurement of
Ca2+ in vivo relies on the inclusion of fluorescent agents such as the Fura-2 in this work to
look at cytosolic Ca2+ or Mag-Fluo-4 to look at Ca2+ in the endoplasmic reticulum [20,43].
However, these options require cell labeling and complex treatment processes, both of
which are avoided by electrical measurement. The electrical system presented can offer a
wider view of the individual cell dielectric properties at multiple frequencies, more rapid
measurement, and fewer resources required. There is a need to demonstrate a realistic
sensitivity to biological levels of cytoplasm Ca2+ for disease diagnosis or monitoring
applications, which shapes our approach in the future. Going forward, the spectral changes
seen in this work will be used in a further study of ME/CFS clinical samples looking at
how skeletal muscle electrical properties at a variety of frequencies can be correlated with
biological changes to further our understanding of this rare disease.

5. Conclusions
Based on the electrical measurements and corresponding extracted parameters, there
is a subtle change in MHz and GHz spectral patterns that can be correlated with fluorescent-
image-based cytoplasmic Ca2+ levels. The electrically measured differences can be further
described by changes to the dielectric parameters of cytoplasmic permittivity (εc ) and
conductivity (σc ). In this work, it was found that increased cytoplasmic Ca2+ concentration
can be associated with a significant increase in cytoplasmic conductivity and a decrease
in cytoplasmic permittivity. This monitoring system improves the depth of information
available about intracellular conditions and ion study in the cytoplasm. The work presented
here is limited by the lack of comparison with true concentration correlation and selective
sensing of the measurement system to particular ions; therefore, further exploration is
necessary to develop a true system for disease monitoring. That being said, understanding
these Ca2+ levels can help generate understanding and evaluate skeletal muscle disease
progression and treatment effectiveness. In addition, by modeling these changes in the
context of pre-evaluated oxidative changes, important deductions can be made about how
different properties associated with ME/CFS contribute to an overall electrical profile to
move toward a unique and rapid diagnostic tool.

Author Contributions: Conceptualization, C.A.F., T.P. and X.C.; methodology, C.A.F., M.F., T.P. and
X.C.; software, C.A.F.; validation, C.A.F., M.F., T.P. and X.C.; formal analysis, C.A.F.; investigation,
C.A.F., C.S., L.M., T.P. and X.C.; resources, T.P. and X.C.; data curation, C.A.F., C.S., L.M. and
T.P.; writing—original draft preparation, C.A.F.; writing—review and editing, M.F., T.P. and X.C.;
visualization, C.A.F.; supervision, M.F., T.P. and X.C.; project administration, T.P. and X.C.; funding
acquisition, T.P. and X.C. All authors have read and agreed to the published version of the manuscript.
Funding: C.A.F. and X.C. appreciate support through funding from National Science Foundation,
Division of Electrical, Communications & Cyber Systems Grant 1809623. C.S., L.M. and T.P. are
supported through “G. d’Annunzio” University grants.
Institutional Review Board Statement: Not applicable.
 C.A.F., C.S.,
C.A.F., C.S., L.M.,
C.S., L.M., T.P.
L.M., T.P. and
T.P. and X.C.;
and X.C.; resources,
X.C.; resources, T.P.
resources, T.P. and
T.P. and X.C.;
and X.C.; data
X.C.; data curation,
data curation, C.A.F.,
curation, C.A.F., C.S.,
C.A.F., C.S., L.M.
C.S., L.M. and
L.M. and T.P.;
and T.P.;
T.P.;
C.A.F.,
writing—original draft preparation, C.A.F.; writing—review and editing, M.F., T.P. and X.C.; visualiza-
writing—original
writing—original draft preparation,
draft preparation, C.A.F.;
C.A.F.; writing—review
writing—review and editing,
and editing, M.F., T.P.
M.F.,X.C.; and
T.P.fundingX.C.; visualiza-
and X.C.;acquisition,
visualiza-
tion, C.A.F.;
tion, C.A.F.; supervision,
C.A.F.; supervision, M.F.,
supervision, M.F., T.P.
M.F., T.P. and
T.P. and X.C.;
and X.C.; project
X.C.; project administration,
project administration, T.P.
administration, T.P. and
T.P. and
and X.C.;
X.C.; funding
funding acquisition,
acquisition,
tion,
T.P. and X.C. All authors have read and agreed to the published version of the manuscript.
T.P. and
T.P. and X.C.
X.C. All
All authors
authors have
have read
read and
and agreed
agreed to
to the
the published
published version
version ofof the
the manuscript.
manuscript.
Funding: C.A.F.
Funding: C.A.F. and
C.A.F. and X.C.
and X.C. appreciate
X.C. appreciate support
appreciate support through
support through funding
through funding from
funding from National
from National Science
National Science Foundation,
Science Foundation,
Foundation,
Sensors 2023, 23, 4358 Funding:
Division of Electrical, Communications & Cyber Systems Grant 1809623. C.S., L.M. and T.P. 11 of 13
are
Division
Division of
of Electrical,
Electrical, Communications
Communications &
& Cyber
Cyber Systems
Systems Grant
Grant 1809623.
1809623. C.S.,
C.S., L.M.
L.M. and
and T.P. are
T.P. are
supported through
supported through “G.
through “G. d’Annunzio”
“G. d’Annunzio” University
d’Annunzio” University grants.
University grants.
grants.
supported
Institutional Review
Institutional Review Board
Review Board Statement:
Board Statement:
Statement: NotNot applicable.
Not applicable.
applicable.
Institutional
Informed Consent Statement: Not applicable.
Informed Consent
Informed Consent Statement:
Consent Statement: Not
Statement: Not applicable.
Not applicable.
applicable.
Informed
Data
Data Availability
Availability Statement:Data
Statement: Data available
available uponupon request.
request.
Data Availability
Data Availability Statement:
Statement: Data
Data available
available upon
upon request.
request.
Conflicts
Conflicts of Interest:
of Theauthors
Interest:The authorsdeclare
declareno no conflict
conflict of of interest.
interest.
Conflicts of
Conflicts of Interest:
Interest: The
The authors
authors declare
declare no
no conflict
conflict of
of interest.
interest.

Appendix A
Appendix A
Appendix A
Appendix A

Figure
Figure A1.
Figure A1. Background
A1. Background simulation
simulation fitting
Background simulation
simulation fittingto
fitting to experimental
to experimentaldata
experimental datafor
data forrepresentative
for representativecell
representative cellfor
cell for(a)
for (a)S11
(a) S11 and
S11
Figure
and (b) A1.
S21 Background
spectra. fitting to experimental data for representative cell for (a) S11
(b)
andS21
(b) spectra.
S21 spectra.
and (b) S21 spectra.

Figure A2. Delta simulation fitting to experimental data for representative cell for (a) ΔS11 and (b)
Figure A2.
Figure
Figure A2. Delta
Delta simulation
Deltasimulation fitting
simulation to experimental
fitting
fitting to experimental data data
to experimental
data for representative
for representative cell for
for representative
cell for cell
(a) ΔS11
(a) ΔS11 and
for (a)
and∆S11
(b) and
(b)
ΔS21
ΔS21 spectra.
spectra.
(b) ∆S21
ΔS21 spectra.
spectra.

Figure A3. Normalized error for weights applied to Δ curve fitting optimization.
Figure A3.
A3. Normalized
Normalized error for weights applied to Δ
Δ curve fitting optimization.
Figure
Figure A3. Normalizederror
errorfor
forweights
weightsapplied to
applied ∆ curve
tocurve fitting optimization.
fitting optimization.
References
References
References
References
1. Beccafico, S.; Puglielli, C.; Pietrangelo, T.; Bellomo, R.; Fano, G.; Fulle, S. Age-Dependent Effects on Functional Aspects in
1. Beccafico,
Beccafico, S.; Puglielli, C.; Pietrangelo, T.; Bellomo, R.; Fano, G.; Fulle, S. Age-Dependent
Age-Dependent Effects on Functional Aspects in
1. 1. S.;S.;
Beccafico, Puglielli, C.;C.;
Puglielli, Pietrangelo, T.;T.;
Pietrangelo, Bellomo,
Bellomo,R.;R.;
Fano, G.;G.;
Fano, Fulle, S. Age-Dependent
Fulle, S. Effects on on
Effects Functional Aspects
Functional in Human
Aspects in
Human Satellite
HumanCells.
Satellite Cells.
Cells. Ann.
Ann. N. Y.
N. Y. Acad.
Y. Acad.
Acad. Sci.
Sci. 2007,
2007, 1100, 345–352.
1100, 345–352.
345–352. https://doi.org/10.1196/annals.1395.037.
https://doi.org/10.1196/annals.1395.037.
Satellite
Human Ann.
Satellite N. Y.
Cells. Acad.
Ann. N. Sci. 2007, 1100,
Sci. 345–352.
2007, 1100, [CrossRef]
https://doi.org/10.1196/annals.1395.037.
2. Bravo-Sagua, R.; Parra, V.; Muñoz-Cordova, F.; Sanchez-Aguilera, P.; Garrido, V.; Contreras-Ferrat, A.; Chiong, M.; Lavandero, S.
Chapter Five—Sarcoplasmic Reticulum and Calcium Signaling in Muscle Cells: Homeostasis and Disease. In International Review
of Cell and Molecular Biology; Kepp, O., Galluzzi, L., Eds.; Biology of the Endoplasmic Reticulum; Academic Press: Cambridge,
MA, USA, 2020; Volume 350, pp. 197–264.
3. Protasi, F.; Pietrangelo, L.; Boncompagni, S. Calcium Entry Units (CEUs): Perspectives in Skeletal Muscle Function and Disease.
J. Muscle Res. Cell Motil. 2021, 42, 233–249. [CrossRef] [PubMed]
4. Espinosa, A.; Henríquez-Olguín, C.; Jaimovich, E. Reactive Oxygen Species and Calcium Signals in Skeletal Muscle: A Crosstalk
Involved in Both Normal Signaling and Disease. Cell Calcium 2016, 60, 172–179. [CrossRef]
5. Agrawal, A.; Suryakumar, G.; Rathor, R. Role of Defective Ca2+ Signaling in Skeletal Muscle Weakness: Pharmacological
Implications. J. Cell Commun. Signal. 2018, 12, 645–659. [CrossRef] [PubMed]
Sensors 2023, 23, 4358 12 of 13

6. Gerwyn, M.; Maes, M. Mechanisms Explaining Muscle Fatigue and Muscle Pain in Patients with Myalgic Encephalomyeli-
tis/Chronic Fatigue Syndrome (ME/CFS): A Review of Recent Findings. Curr. Rheumatol. Rep. 2017, 19, 1. [CrossRef]
7. Dargelos, E.; Brulé, C.; Combaret, L.; Hadj-Sassi, A.; Dulong, S.; Poussard, S.; Cottin, P. Involvement of the Calcium-Dependent
Proteolytic System in Skeletal Muscle Aging. Exp. Gerontol. 2007, 42, 1088–1098. [CrossRef] [PubMed]
8. Leijendekker, W.J.; Passaquin, A.-C.; Metzinger, L.; Rüegg, U.T. Regulation of Cytosolic Calcium in Skeletal Muscle Cells of the
Mdx Mouse under Conditions of Stress. Br. J. Pharmacol. 1996, 118, 611–616. [CrossRef]
9. Berchtold, M.W.; Brinkmeier, H.; Müntener, M. Calcium Ion in Skeletal Muscle: Its Crucial Role for Muscle Function, Plasticity,
and Disease. Physiol. Rev. 2000, 80, 1215–1265. [CrossRef]
10. Tang, W.; Tang, D.; Ni, Z.; Xiang, N.; Yi, H. A Portable Single-Cell Analysis System Integrating Hydrodynamic Trapping with
Broadband Impedance Spectroscopy. Sci. China Technol. Sci. 2017, 60, 1707–1715. [CrossRef]
11. Bao, X.; Ocket, I.; Bao, J.; Doijen, J.; Zheng, J.; Kil, D.; Liu, Z.; Puers, B.; Schreurs, D.; Nauwelaers, B. Broadband Dielectric
Spectroscopy of Cell Cultures. IEEE Trans. Microw. Theory Tech. 2018, 66, 5750–5759. [CrossRef]
12. Ning, Y.; Multari, C.; Luo, X.; Palego, C.; Cheng, X.; Hwang, J.C.M.; Denzi, A.; Merla, C.; Apollonio, F.; Liberti, M. Broadband
Electrical Detection of Individual Biological Cells. IEEE Trans. Microw. Theory Tech. 2014, 62, 1905–1911. [CrossRef]
13. Grenier, K.; Tamra, A.; Zedek, A.; Poiroux, G.; Artis, F.; Chen, T.; Chen, W.; Poupot, M.; Fournie, J.-J.; Dubuc, D. Low Volume and
Label-Free Molecules Characterization and Cell Monitoring with Microwave Dielectric Spectroscopy. In Proceedings of the 2018
IEEE International Microwave Biomedical Conference (IMBioC), Philadelphia, PA, USA, 14–15 June 2018; IEEE: Piscataway, NJ,
USA, 2018; pp. 82–84.
14. Foster, K.R.; Schwan, H.P. Dielectric Properties of Tissues and Biological Materials: A Critical Review. Crit. Rev. Biomed. Eng.
1989, 17, 25–104.
15. Markx, G.H.; Davey, C.L. The Dielectric Properties of Biological Cells at Radiofrequencies: Applications in Biotechnology. Enzym.
Microb. Technol. 1999, 25, 161–171. [CrossRef]
16. Nasir, N.; Al Ahmad, M. Cells Electrical Characterization: Dielectric Properties, Mixture, and Modeling Theories. J. Eng. 2020,
2020, 9475490. [CrossRef]
17. Grenier, K.; Dubuc, D.; Chen, T.; Artis, F.; Chretiennot, T.; Poupot, M.; Fournie, J.-J. Recent Advances in Microwave-Based
Dielectric Spectroscopy at the Cellular Level for Cancer Investigations. IEEE Trans. Microw. Theory Tech. 2013, 61, 2023–2030.
[CrossRef]
18. ATCC. L6—CRL-1458. Available online: https://www.atcc.org/products/crl-1458 (accessed on 24 October 2022).
19. Fioretti, B.; Pietrangelo, T.; Catacuzzeno, L.; Franciolini, F. Intermediate-Conductance Ca2+ -Activated K+ Channel Is Expressed
in C2C12 Myoblasts and Is Downregulated during Myogenesis. Am. J. Physiol. Cell Physiol. 2005, 289, C89–C96. [CrossRef]
[PubMed]
20. Pietrangelo, T.; Mariggiò, M.A.; Lorenzon, P.; Fulle, S.; Protasi, F.; Rathbone, M.; Werstiuk, E.; Fanò, G. Characterization of Specific
GTP Binding Sites in C2C12 Mouse Skeletal Muscle Cells. J. Muscle Res. Cell Motil. 2002, 23, 107–118. [CrossRef] [PubMed]
21. Ferguson, C.; Pini, N.; Du, X.; Farina, M.; Hwang, J.M.C.; Pietrangelo, T.; Cheng, X. Broadband Electrical Impedance as a Novel
Characterization of Oxidative Stress in Single L6 Skeletal Muscle Cells. Anal. Chim. Acta 2021, 1173, 338678. [CrossRef] [PubMed]
22. Du, X.; Ferguson, C.; Ma, X.; Cheng, X.; Hwang, J.C.M. Ultra-Wideband Impedance Spectroscopy of the Nucleus in a Live Cell.
IEEE J. Electromagn. RF Microw. Med. Biol. 2021, 6, 267–272. [CrossRef]
23. Du, X.; Ladegard, C.; Ma, X.; Cheng, X.; Hwang, J.C.M. Broadband Electrical Sensing of Nucleus Size in a Live Cell from 900 Hz
to 40 GHz. In Proceedings of the 2020 IEEE MTT-S International Microwave Biomedical Conference (IMBioC), Toulouse, France,
14–17 December 2020; pp. 1–4.
24. Caspers, F. RF Engineering Basic Concepts: S-Parameters. arXiv 2012, arXiv:1201.2346v1.
25. Kidokoro, Y. Developmental Changes of Membrane Electrical Properties in a Rat Skeletal Muscle Cell Line. J. Physiol. 1975, 244,
129–143. [CrossRef] [PubMed]
26. Ma, X.; Du, X.; Multari, C.R.; Ning, Y.; Luo, X.; Gholizadeh, V.; Palego, C.; Cheng, X.; Hwang, J.C.M. Reproducible Broadband
Measurement for Cytoplasm Capacitance of a Biological Cell. In Proceedings of the 2016 IEEE MTT-S International Microwave
Symposium (IMS), San Francisco, CA, USA, 22–27 May 2016; IEEE: Piscataway, NJ, USA, 2016; pp. 1–4.
27. Erdahl, W.L.; Chapman, C.J.; Taylor, R.W.; Pfeiffer, D.R. Ca2+ Transport Properties of Ionophores A23187, Ionomycin, and
4-BrA23187 in a Well Defined Model System. Biophys. J. 1994, 66, 1678–1693. [CrossRef]
28. Caridha, D.; Yourick, D.; Cabezas, M.; Wolf, L.; Hudson, T.H.; Dow, G.S. Mefloquine-Induced Disruption of Calcium Homeostasis
in Mammalian Cells Is Similar to That Induced by Ionomycin. Antimicrob. Agents Chemother. 2008, 52, 684–693. [CrossRef]
[PubMed]
29. Sözer, E.B.; Vernier, P.T. Modulation of Biological Responses to 2 ns Electrical Stimuli by Field Reversal. Biochim. Biophys. Acta
(BBA)—Biomembr. 2019, 1861, 1228–1239. [CrossRef]
30. Pakhomova, O.N.; Gregory, B.; Semenov, I.; Pakhomov, A.G. Calcium-Mediated Pore Expansion and Cell Death Following
Nanoelectroporation. Biochim. Biophys. Acta 2014, 1838, 2547–2554. [CrossRef] [PubMed]
31. Burke, R.C.; Bardet, S.M.; Carr, L.; Romanenko, S.; Arnaud-Cormos, D.; Leveque, P.; O’Connor, R.P. Nanosecond Pulsed Electric
Fields Depolarize Transmembrane Potential via Voltage-Gated K+ , Ca2+ and TRPM8 Channels in U87 Glioblastoma Cells. Biochim.
Biophys. Acta Biomembr. 2017, 1859, 2040–2050. [CrossRef]
Sensors 2023, 23, 4358 13 of 13

32. Zhang, J.; Blackmore, P.F.; Hargrave, B.Y.; Xiao, S.; Beebe, S.J.; Schoenbach, K.H. Nanosecond Pulse Electric Field (Nanopulse): A
Novel Non-Ligand Agonist for Platelet Activation. Arch. Biochem. Biophys. 2008, 471, 240–248. [CrossRef]
33. Keller, C.; Hellsten, Y.; Steensberg, A.; Klarlund Pedersen, B. Differential Regulation of IL-6 and TNF-α via Calcineurin in Human
Skeletal Muscle Cells. Cytokine 2006, 36, 141–147. [CrossRef] [PubMed]
34. Holmes, A.G.; Watt, M.J.; Carey, A.L.; Febbraio, M.A. Ionomycin, but Not Physiologic Doses of Epinephrine, Stimulates Skeletal
Muscle Interleukin-6 MRNA Expression and Protein Release. Metabolism 2004, 53, 1492–1495. [CrossRef]
35. Huang, H.; Zhao, Y.; Shang, X.; Ren, H.; Zhao, Y.; Liu, X. CAIII Expression in Skeletal Muscle Is Regulated by Ca2+ –CaMKII–
MEF2C Signaling. Exp. Cell Res. 2019, 385, 111672. [CrossRef]
36. Roufa, D.; Wu, F.S.; Martonosi, A.N. The Effect of Ca2+ Ionophores upon the Synthesis of Proteins in Cultured Skeletal Muscle.
Biochim. Biophys. Acta (BBA)—Gen. Subj. 1981, 674, 225–237. [CrossRef]
37. Pietrangelo, T.; Di Filippo, E.S.; Mancinelli, R.; Doria, C.; Rotini, A.; Fanò-Illic, G.; Fulle, S. Low Intensity Exercise Training
Improves Skeletal Muscle Regeneration Potential. Front. Physiol. 2015, 6, 399. [CrossRef] [PubMed]
38. Petchakup, C.; Tay, H.M.; Li, K.H.H.; Hou, H.W. Integrated Inertial-Impedance Cytometry for Rapid Label-Free Leukocyte
Isolation and Profiling of Neutrophil Extracellular Traps (NETs). Lab Chip 2019, 19, 1736–1746. [CrossRef]
39. Funkner, S.; Niehues, G.; Schmidt, D.A.; Heyden, M.; Schwaab, G.; Callahan, K.M.; Tobias, D.J.; Havenith, M. Watching the
Low-Frequency Motions in Aqueous Salt Solutions: The Terahertz Vibrational Signatures of Hydrated Ions. J. Am. Chem. Soc.
2012, 134, 1030–1035. [CrossRef] [PubMed]
40. Balos, V.; Kaliannan, N.K.; Elgabarty, H.; Wolf, M.; Kühne, T.D.; Sajadi, M. Time-Resolved Terahertz–Raman Spectroscopy Reveals
That Cations and Anions Distinctly Modify Intermolecular Interactions of Water. Nat. Chem. 2022, 14, 1031–1037. [CrossRef]
41. Gleichmann, M.; Mattson, M.P. Neuronal Calcium Homeostasis and Dysregulation. Antioxid. Redox Signal. 2011, 14, 1261–1273.
[CrossRef] [PubMed]
42. Brini, M.; Calì, T.; Ottolini, D.; Carafoli, E. Neuronal Calcium Signaling: Function and Dysfunction. Cell. Mol. Life Sci. 2014, 71,
2787–2814. [CrossRef]
43. Milán, A.F.; Rincón, O.A.; Arango, L.B.; Reutovich, A.A.; Smith, G.L.; Giraldo, M.A.; Bou-Abdallah, F.; Calderón, J.C. Calibration
of Mammalian Skeletal Muscle Ca2+ Transients Recorded with the Fast Ca2+ Dye Mag-Fluo-4. Biochim. Biophys. Acta (BBA)—Gen.
Subj. 2021, 1865, 129939. [CrossRef]

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual
author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to
people or property resulting from any ideas, methods, instructions or products referred to in the content.