Extracted Text
s41598-017-09892-w.pdf
1ScieNtific REPOrTs | 7: 9421 | DOI:10.1038/s41598-017-09892-w
www.nature.com/scientificreports
Enhancement of mesenchymal
stem cell chondrogenesis with
short-term low intensity pulsed
electromagnetic �elds
Dinesh Parate
1
, Alfredo Franco-Obregón
2,3
, Jürg Fröhlich
2,4
, Christian Beyer
4
,
Azlina A. Abbas
5
, Tunku Kamarul
5
, James H. P. Hui
1,6
& Zheng Yang
1,6
Pulse electromagnetic �elds (PEMFs) have been shown to recruit calcium-signaling cascades common
to chondrogenesis. Here we document the e�ects of speci�ed PEMF parameters over mesenchymal
stem cells (MSC) chondrogenic di�erentiation. MSCs undergoing chondrogenesis are preferentially
responsive to an electromagnetic e�cacy window de�ned by �eld amplitude, duration and frequency
of exposure. Contrary to conventional practice of administering prolonged and repetitive exposures
to PEMFs, optimal chondrogenic outcome is achieved in response to brief (10 minutes), low intensity
(2 mT) exposure to 6 ms bursts of magnetic pulses, at 15 Hz, administered only once at the onset of
chondrogenic induction. By contrast, repeated exposures diminished chondrogenic outcome and could
be attributed to calcium entry after the initial induction. Transient receptor potential (TRP) channels
appear to mediate these aspects of PEMF stimulation, serving as a conduit for extracellular calcium.
Preventing calcium entry during the repeated PEMF exposure with the co-administration of EGTA or
TRP channel antagonists precluded the inhibition of di�erentiation. This study highlights the intricacies
of calcium homeostasis during early chondrogenesis and the constraints that are placed on PEMF-based
therapeutic strategies aimed at promoting MSC chondrogenesis. The demonstrated e�cacy of our
optimized PEMF regimens has clear clinical implications for future regenerative strategies for cartilage.
Articular cartilage is an avascular tissue with low potential for self-repair. When le untreated, lesions of the
articular cartilage can lead to osteoarthritis
1–3
. e success of any technology aimed at repairing chondral defects
will thus be based on its ability to produce tissues that most closely recapitulate the mechanical and biochemical
properties of native cartilage. To this end many technologies have been advanced yet, none are without draw-
backs. e �microfracture� technique is commonly plagued by the formation of bro-cartilaginous tissue of low
dexterity
4
. Autologous chondrocytes implantation and osteochondral autogra transplantation are limited by
scarce cartilage production, low proliferative capacity of chondrocytes, chondrocyte de-di erentiation and com-
plications due to donor site morbidity
5
. Stem cell-based approaches are also being actively pursued in hopes of
improved outcome. Mesenchymal stem cells (MSCs) support chondrogenic di erentiation and are an attractive
cell source for cartilage tissue engineering. However, the neocartilage formed by conventional MSC-based repair
1
Department of Orthopaedic Surgery, Yong Loo Lin School of Medicine, National University of Singapore, NUHS
Tower Block, Level 11, 1E Kent Ridge Road, Singapore, 119288, Singapore.
2
Department of Surgery, Yong Loo Lin
School of Medicine, National University of Singapore, NUHS Tower Block, Level 8, IE Kent Ridge Road, Singapore,
119228, Singapore.
3
BioIonic Currents Electromagnetic Pulsing Systems Laboratory, BICEPS, National University
of Singapore, MD6, 14 medical Drive, #14-01, Singapore, 117599, Singapore.
4
Institute for Electromagnetic Fields,
Swiss Federal Institute of Technology (ETH), Rämistrasse 101, 8092, Zurich, Switzerland.
5
Tissue Engineering
Group (TEG), National Orthopaedic Centre of Excellence for Research and Learning (NOCERAL), Department of
Orthopaedic Surgery, Faculty of Medicine, University of Malaya, Pantai Valley, Kuala Lumpur, 50603, Malaysia.
6
Tissue Engineering Program, Life Sciences Institute, National University of Singapore, DSO (Kent Ridge) Building,
#04-01, 27 Medical Drive, Singapore, 117510, Singapore. Correspondence and requests for materials should be
addressed to A.F.-O. (email: suraf@nus.edu.sg) or J.H.P.H. (email: james_hui@nuhs.edu.sg) or Z.Y. (email: lsiyz@
nus.edu.sg)
Received: 13 April 2017
Accepted: 28 July 2017
Published: xx xx xxxx
OPEN
www.nature.com/scientificreports/ 2ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
methodologies commonly contain a mixture of =bro- and hyaline cartilage
6–8
that do not achieve the biochemical,
mechanical or functional properties of native cartilage.
MSCs can be di+erentiated along di+erent cell lineages of mesodermal origin including osteoblasts, chondro-
cytes, skeletal myocytes or visceral stromal cells
9
. Chondrogenic induction of MSCs entails proliferation, conden-
sation, di+erentiation and maturation
10
, necessitating endogenous transcriptional and developmental regulators,
cell-cell and cell-matrix interactions that, in turn, are modulated by environmental stimuli including mechanical
forces, temperature and oxygen levels
10–13
. A common objective is to recreate as closely as possible the in vivo
environmental conditions in vitro so that the rate and quality of chondrogenic development is enhanced and
the functionality of the repaired tissue improved. To this end, various environmental stimuli such as hypoxia,
mechanical, electric and electromagnetic stimulation are currently being explored
14–16
.
Mechanical stimulation can be applied in a semi-controlled manner with the use of bioreactors designed to
impart shear, compression, tension, or pressure on developing tissues. Appropriately applied mechanical stim-
ulation positively in×uences MSC-induced chondrogenic di+erentiation, ECM deposition and the mechanical
properties of the generated cartilage
17–20
. At the cellular level the transduction of mechanical signals (mech-
anotransduction) involves their conversion into biochemical responses, often with the assistance of mech-
anosensitive calcium channels
21–24
. Electromagnetic =eld (EMF)-stimulation has been shown to promote cell
di+erentiation via the modulation of extracellular calcium entry via plasma membrane-embedded cation chan-
nels
25–27
, raising the intriguing possibility that EMFs may be recruiting related pathways.
Studies examining time-variant or pulsing electromagnetic =elds (PEMFs) have alluded to a bene=t over
articular chondrocytes or cartilaginous tissue in vitro , particularly with reference to chondrocyte proliferation,
extracellular matrix (ECM) deposition, secretory activity and in×ammatory status
28–35
. Studies have also exam-
ined the e+ects of PEMF-treatment over the chondrogenic di+erentiation of stem cells derived from bone mar-
row
36–38
, adipose
35
, umbilical cord Wharton jelly
39
, synovial ×uid
40
or peripheral blood sources
40
. Re reported
consequences of PEMF-stimulation over chondrogenesis, however, are largely inconsistent. Some studies report
modest enhancements in the gene expression of Sox9, aggrecan, type II collagen (Col 2) as well as deposition
of sulfated glycosaminoglycan (sGAG), typically on the order of 2-folds
35, 36
, whereas other studies show little
to no e+ect
38–41
. On the extreme end of the spectrum, Wang et al .
37
reported inhibition of both Sox9 and Col 2
expression concomitant with induction of hypertrophy and mineralization in response to exposures of 3 h per
day at an amplitude of 1 mT. Obvious di+erences in stimulation protocols likely underlie reported discrepancies.
Existing EMF studies have typically employed exposure durations between 30 minutes to 8 h per day and more
consistently in the low milli Tesla amplitude range (3–5 mT). Empirical determination of the appropriate expo-
sure and signal parameters for a speci=c biological response and given tissue are essential as there are indications
that cell responses to magnetic =elds obey an electromagnetic eβcacy window de=ned by a speci=c combination
of frequency, amplitude and time of exposure that gives rise to optimum cell response
42, 43
. Here, we systemat-
ically characterized the e+ects of PEMF exposure over MSC chondrogenic di+erentiation by varying the =eld
amplitude, exposure duration and dosage with an emphasis on determining the briefest and lowest amplitude
electromagnetic exposure to render a developmental outcome. Given that both mechanical stimuli and calcium
entry
21, 22
in×uences chondrogenic di+erentiation, we investigated the ability of PEMF exposure to in×uence cal-
cium homeostasis during early induction of MSCs into the chondrogenic lineage, in particular that attributed to
the Transient Receptor Potential (TRP) family of cation-permeable channels, which has been broadly implicated
in cellular mechanotransduction
23, 44
. We show that brief and single exposures to low amplitude PEMFs were
most e+ective at stimulating MSC chondrogenesis. Our results also implicate the involvement of calcium in×ux
and the mechanosensitive TRP channels, TRPC1 and TRPV4, in the chondrogenic development stimulated by
targeted PEMF exposure.
Results
E�ect of PEMF intensities and exposure durations on MSC chondrogenesis. We =rst sought to
determine the magnetic =eld amplitude and duration of exposure at which MSCs undergoing chondrogenic
induction are most responsive using starting conditions preliminarily tested in MSCs for chondrogenic regen-
eration
39
. MSC pellets in chondrogenic di+erentiation medium were subjected to single exposures to PEMFs
of 10 min duration at intensities ranging in amplitudes between 0–4 mT (Fig. 1A), then subjected to exposure
durations between 5 and 60 min at 2 mT intensity (Fig. 1B), applied on the =rst day of chondrogenic induction.
RNA analysis monitoring MSC chondrogenic progression at 7 days post-induction showed greatest increases in
response to 10 min exposures applied at an amplitude of 2 mT as evidenced by enhancements in Sox9, aggrecan
and Col 2 mRNA expression. By contrast, lower (1 mT) or higher (>3 mT) amplitude of PEMFs (Fig. 1A), or
briefer (5 min) or longer (>20 min) durations of exposure (Fig. 1B), resulted in overall smaller e+ect sizes. ±e
same EMF e cacy window translated to the expression of cartilaginous ECM macromolecular proteins (Fig. 1C).
In response to 2 mT amplitude pulsing, a 3-fold increase in Col 2 protein was detected 21 days a er chondrogenic
induction, whereas no increase was detected with exposure to 3 mT. Moreover, a 2-fold increase in sGAG was
detected in response to exposure to 2 mT PEMFs, whereas 3 mT PEMFs produced a signi=cantly smaller increase.
±e relative ine+ectiveness of prolonged exposure to PEMFs was also corroborated at the protein level. Sixty min
exposures to 2 mT PEMFs did not elicit signi=cant increases in Col 2 formation than 10 min exposures (Fig. 1C).
With reference to sGAG production, PEMF amplitudes greater than 2 mT, or exposure durations of one hour,
produced inferior results to 10 min exposures at 2 mT.
Dosage e�ects of PEMFs over MSC chondrogenesis. We next investigated the e+ect of repetitive
exposures to PEMFs. MSCs were exposed to PEMFs at 2 mT for 10 min/day once, twice or thrice on days 1, 2
and 4 following chondrogenic induction (Fig. 2A). RNA analysis a>er 7 days of di+erentiation showed that a
single exposure produced the greatest and most consistent increase in the expression of chondrogenic markers
www.nature.com/scientificreports/ 3ScieNtific REPOrTs }� 9421 D �10.103~ s415~-01}-0~2-w
(Fig. 2A). In another series of experiments, MSCs pellets were exposed once on the =rst day of chondrogenic
induction, or weekly for 3 consecutive weeks (Fig. 2B). RNA analysis a>er 21 days of chondrogenic di+erentia-
tion showed that a single exposure to 2 mT PEMFs for 10 min given on day 1 of induction gave the greatest and
most consistent increase in expression of chondrogenic markers relative to no exposure (0 mT) (Fig. 2B). ±ree
weekly exposures either rendered no additional bene=t (Col 2) or gave similar results to control (0 mT) (Sox9 and
aggrecan). Moreover, the amount of ECM produced was inversely related to the total number of exposures. Single
exposures produced > 2-folds and ~1-fold increases in Col 2 and sGAG, respectively (Fig. 2C), whereas triple
weekly exposures for three weeks (9 total exposures) completely precluded an increase Col 2 and sGAG forma-
tion. ±e change in the amount of DNA across samples varied less than 0.2-fold, although reaching signi=cances
at 2 mT, indicating that cell proliferation was only modestly a+ected within our pellet culture system (Fig.−2C).
E�ect of PEMF treatment to deposition of ECM. ECM deposition in response to PEMF-exposure was
also analyzed using Safranin O staining for proteoglycan and immunohistochemical staining for type II collagen
(Fig. 3). Stained images of day 21 samples showed an enhanced deposition of proteoglycan and type II collagen in
samples exposed only once to 2 mT for 10 min as compared to control (0 mT). By contrast, MSC samples exposed
for longer (60 min), to greater amplitude (3 mT) or repeatedly (3x, 9x) yielded comparable, or inferior, ECM
deposition to control.
Ca
2+
entry pathways implicated in transducing the e�ects of PEMFs over MSC di�erentia-
tion.
To investigate whether PEMF-stimulated MSC chondrogenesis was depended on calcium in×ux, EGTA
(2 mM) was co-administered to the culture medium during PEMF−exposure and summarily replaced a>erwards
with age-matched chondrogenic control media. RNA analysis at day 7 showed that the inclusion of EGTA sig-
ni=cantly decreased the mRNA expression of Sox9, Col 2 and aggrecan in PEMF-treated samples (Fig.−4A),
Figure 1. E+ects of PEMF amplitude (A) and exposure duration (B) on MSC chondrogenesis. Real-time
PCR analysis of cartilaginous markers expression a>er 7 days of di+erentiation was normalized to GAPDH and presented as fold-changes relative to levels in undi+erentiated MSC. (C) Quanti=cation of cartilaginous extracellular matrix macromolecules (Col−2 and sGAG) generated a>er 21 days of chondrogenic di+erentiation of MSC subjected to distinct PEMF parameters. All data shown are mean ± SD, n = 6 from 2 independent
experiments. *Denotes signi=cant increase, or decrease, compared to non-PEMF control.
#
Denotes signi=cant
decrease compared to 2 mT (A), or 10 min (B) PEMF exposure.
www.nature.com/scientificreports/ 4ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
indicating that PEMF-exposure stimulates calcium in×ux. Conversely, transiently supplementing the di+eren-
tiation medium with elevated extracellular Ca
2+
(5 mM CaCl
2) enhanced the mRNA expression of Sox9, Col 2
and aggrecan in otherwise non-exposed samples, and moreover, accentuated chondrogenic gene expression in
PEMF-treated samples. ±ese results corroborate that calcium in×ux is part of the upstream signalling cascade
recruited by PEMFs contributing to chondrogenic induction.
To reveal the Ca
2+
in×ux pathway recruited by PEMFs, we pharmacologically dissected the contribution of
candidate channels utilizing 2-APB (100 µM) or Ruthenium Red (RR, 10 µM) as TRPC or TRPV cation chan-
nel antagonists, respectively, or Nifedipine (1 µM), as a dihydropyridine-sensitive, L-type voltage-gated calcium
channel (VGCC) antagonist. Calcium channel antagonists were included into the di+erentiation medium 10 min
Figure 2. Dosage e+ects of PEMFs over MSC chondrogenesis. (A) MSCs were exposed once (1x), twice (2x) or
thrice (3x) per week. (B) MSCs were subjected to either a single exposure on day 1 of chondrogenic induction (1x) or once per week for 3 weeks (3x). Real-time PCR analysis of cartilaginous markers expression at 7 (A) or 21 days (B) a>er the induction of di+erentiation was normalized to GAPDH and presented as fold-changes relative to level in undi+erentiated MSCs. (C) Quanti=cation of cartilaginous ECM macromolecules generated
during chondrogenic di+erentiation of MSCs in response to distinct PEMF dosing as indicated. MSC pellets
were subjected to either a single PEMF exposure given on day 1 of chondrogenic induction (1x) once per week
for 3 weeks (3x), or thrice weekly for 3 weeks (9x). Data represents the mean ± SD, n = 6 from 2 independent
experiments. *Denotes signi=cant increase compare to non-PEMF (0 mT) control.
#
Denotes signi=cant decrease
compared to single PEMF (1x) exposure.
www.nature.com/scientificreports/ 5ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
before exposure to PEMFs and removed immediately a>erwards with age-matched control chondrogenic media.
Both 2-APB and Ruthenium Red completely inhibited the PEMF-triggered up-regulation of chondrogenic
genes, whereas Nifedipine had no signi=cant inhibitory e+ect (Fig.−4B). Chondrogenic inhibition by 2-APB and
Ruthenium Red was also observed in non-exposed samples, indicating that TRPC- and TRPV-mediated calcium
entry are similarly involved in constitutive chondrogenesis upon induction. By contrast, VGCC-mediated Ca
2+
entry does not appear to play a predominant role in the early induction of chondrogenesis.
We next investigated the expression pro=les of TRP channels (TRPC1, TRPC6, TRPV1, TRPV4, TRPV6)
previously implicated in chondrogenesis and correlated these to our PEMF-induced chondrogenic responses.
Amongst the panel of candidate TRP channels, the expression of TRPC1 and TRPV4 most closely correlated with
our delineated magnetic e cacy window governing chondrogenesis with reference to PEMF amplitude, duration
Figure 3. Histological analysis of pellets exposed to PEMFs of distinct amplitude, duration and dosage. Pellets
were harvested at day 21, sectioned and subjected to Safranin O or type II collagen immunohistochemistry
staining. Images presented were represenation of n = 3, taken at 100× magni=cation.
Figure 4. Investigation of calcium entry pathways implicated in the PEMF-e+ect. (A) Involvement of Ca
2+
in×ux in mediating the e+ects of PEMF-induced MSC chondrogenic di+erentiation. MSCs were exposed for 10 min at 2 mT alone (control, white bars), or in the presence of 2 mM EGTA (dark grey bars) or 5 mM
CaCl
2 (hatched bars) transiently added to the culture media. EGTA and CaCl
2 were included to the bathing
media 10 min before exposure and replaced with age-matched media control cultures 10 min a>er exposure.
(B) Involvement of candidate calcium channels in mediating the e+ect of PEMs over MSC chondrogenic di+erentiation. Control MSC chondrogenic di+erentiation medium (white bars) was supplemented with Nifedipine (1 µM, light grey bars), Ruthenium Red (RR, 10 µM, black bars), or 2-APB (100 µM, dark grey bars)
10 min before exposure and replaced with age-matched media control cultures 10 min a>er exposure. Real-
time PCR analysis was performed on day 7 of di+erentiation. Data represent the means ± SD, n = 6 from 2
independent experiments. *Denotes signi=cant increase, or decrease, compared to non-PEMF (0 mT) control.
#
Denotes signi=cant decrease relativeto 2 mT PEMF treatment.
www.nature.com/scientificreports/ 6ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
and dosage (Fig. 5 and Suppl. Figures 1 and 2). ±ese results corroborate an involvement of TRPC1 and TRPV4
in the PEMF-induced enhancement of chondrogenic di+erentiation of MSC we observed.
E�ect of recurring calcium in�ux on MSCchondrogenesis. We next investigated whether calcium
entry, particularly that via TRPchannels underlies the inhibitory e+ect observed with repeated PEMF exposures.
MSCs were exposed once, twice or thrice to 2 mT PEMFs for 10 min or, alternatively, exposed for 10 min to
aged-matched control di+erentiation media containing elevated extracellular calcium (5 mM CaCl
2) in lieu of
PEMF exposure. RNA analysis at day 7 showed that MSCs treated once with PEMFs, or transiently administered
elevated calcium, on day 1 exhibited enhanced chondrogenesis to comparable levels. By contrast, subsequent
exposures to elevated calcium, on days 2 and 4 suppressed chondrogenesis mirrored the e+ect of multiple expo-
sures to PEMFs (Fig. 6A and Suppl. Fig. 3A).
Analogously, precluding calcium entry (with EGTA) also exhibited dichotomous e+ects if applied during
the =rst versus the second or third exposition to PEMFs, although in opposite direction to that observed with
calcium administration or PEMFs. Whereas EGTA added during the initial exposure to PEMFs (1x) prevented
PEMF-induced chondrogenesis, EGTA applied during the second or third exposure partially counteracted the
Figure 5. Expression pro=les of TRPC1 and TRPV4 in response to determined PEMF eβcacy window
regulating MSC chondrogenesis. Real-time PCR analysis of TRPC1 and V4 exposed to di+erent (A) intensities, (B) durations, and (C) dosages of PEMFs. Data represent the means ± SD, n = 6 from 2 independent
experiments. *Denotes signi=cant increase, or decrease, compared to non-PEMF (0 mT) control.
#
Denotes
signi=cant decrease relative to 2−mT (A), 10 min (B), or single (1x, C) PEMF treatment.
www.nature.com/scientificreports/ 7ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
inhibition of di+erentiation exerted by serial PEMF exposure (Fig.−6B and Suppl. Fig. 3B). Notably, impeding cal-
cium entry with transient application of EGTA during both the second and third PEMF exposure was capable of
almost completely reversing the inhibition of chondrogenesis observed with repeated PEMF exposures, suggest-
ing that PEMFs are activating disparately functioning calcium mechanisms at early (day 1) and later stages (> day
2) of chondrogenic-induction that confer opposite e+ects over chondrogenesis. In contrast to the bene=cial e+ect
of calcium in×ux induced by PEMF at the initial stage of chondrogenesis, subsequent induction of calcium in×ux
by repeated pulsing at later stages of chondrogenesis was suppressive of MSC chondrogenesis. ±e contribution of
TRPC- and TRPV-mediated calcium entry to the chondrogenic-inhibition observed with repeated PEMF expo-
sures was investigated by co-administering 2-APB (100 µM) or Ruthenium Red (RR, 10 µM), respectively, during
PEMF exposure. As observed with transient EGTA application, antagonism of TRPC1/V4-mediated calcium
entry during the =rst exposition to PEMFs was strongly inhibitory of PEMF-induced chondrogenesis, where-
asTRPC1/V4 antagonism during subsequent PEMF expositions was somewhat less protective than EGTA over
di+erentiation (Fig.−6B), implicating other yet to be determined calcium pathways in the later calcium-dependent
inhibitory phase of chondrogenic progression. Notably, Ruthenium Red (TRPV4 antagonist) was capable of
reverting the inhibition of di+erentiation and expression of TRPC1 expression in response to repeated PEMFing,
whereas 2-APB (TRPC antagonist) was unable to revert the inhibition of di+erentiation and TRPV4 expression
in response to repeated PEMFing, suggesting that TRPV4-mediated calcium entry antagonizes TRPC1 expres-
sion leading up to di+erentiation suppression. ±e dichotomous e+ects of precluding calcium entry by EGTA,
Ruthenium Red or 2-APB when applied during the =rst, or the second and third, exposition to PEMFs was
corroborated at the protein level. Preventing calcium entry during the initial exposure to PEMFs (1x) prevented
Figure 6. (A) MSC chondrogenic di+erentiation in response to multiple exposures to PEMFs or exogenously elevated calcium. MSCs were subjected to either PEMF stimulation alone (white bars) or with transient supplementation of CaCl
2 alone (5 mM; hatched bars), once (1x), twice (2x) or thrice (3x) in a week. Dotted
lines refers to expression level of non-treated controls. Real-time PCR analysis was performed on day 7 of chondrogenic di+erentiation. *Denotes signi=cant increase relative to non-PEMF (0 mT) control.
#
and
+
denote signi=cant di+erences relative to respective single (1x) exposure (white and hatched bars, respectively). P = PEMF treatment, Ca = CaCl
2 supplementation. (B) MSC chondrogenic di+erentiation in response to
multiple exposures to PEMFs alone (white bars) or in combination with calcium chelator (EGTA) or TRP channel antagonists. EGTA (2 mM; dark grey bars; “E”), Ruthenium Red (10 µM; RR, black bars; “R”) or 2-APB
(100 µM; light grey bars; �C�) was added to the MSC di erentiation medium during PEMF expoure applied
once (1x), twice (2) or thrice (3x) per week. EGTA, RR and 2-APB were included 10 min before exposure
and replaced with media harvested from age-matched chondrogenic control cultures 10 min a>er exposure.
*Denotes signi=cant increase compare to non-PEMF (0 mT) control.
#
Denotes signi=cant decrease compared to
single PEMF exposure (1x).
+
Denotes signi=cant di+erence compared to respective PEMF control (white bar).
P = PEMF treatment, E = EGTA, R = Ruthenium Red (RR), C = 2-APB. Data shown are means ± SD, n = 6 from
2 independent experiments.
www.nature.com/scientificreports/ 8ScieNtific REPOrTs }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
PEMF-induced cartilaginous Col 2 and sGAG formation, while blocking calcium entry at later exposures coun-
teracted the inhibition of di+erentiation exerted by serial PEMF exposure (Fig.−7). Voltage-gated L-type calcium
channels, on the other hand, do not appear to be strongly implicated in the response as the expression of its subu-
nits (CACNA1C and CACNA2D1) was not perturbed by PEMF or calcium treatment (Suppl. Fig. 3).
Discussion
Pulsed electromagnetic =elds (PEMFs) have been demonstrated to be in×uential in numerous biological functions
including progenitor cell fate determination and di+erentiation. PEMF-based therapies have been previously
shown to enhance chondrocyte and cartilage explant anabolism while also limiting the catabolic consequences of
in×ammatory cytokines
29–31, 33–35, 40
. PEMF exposure has been also reported to enhance the chondrogenic induc-
tion of stem cells
34–37, 45
. Nevertheless, inconsistent and con×icting results plague the scienti=c literature in this
area of study, with PEMF exposures typically being applied on the order of hours per day for several days or
weeks at a time. Here we report a high-e cacy of unprecedentedly brief (10 min applied once) PEMF expo-
sure at inducing MSC chondrogenesis. We consistently detected increases in Sox9, Col 2 and aggrecan mRNA
(>2-folds) in response to lone exposure to 2 mT PEMFs applied at the commencement of induction for only
10 min (Fig. 1). ±ese increments in mRNA later translated into increased chondrogenic ECM protein forma-
tion (> 2-fold) a>er 21 days of di+erentiation. By contrast, stimulation with greater amplitudes (> 3 mT), longer
exposures (> 20 min) or more frequently (> 2x/week) rendered no additional bene=t, or was even less e+ective
at promoting chondrogenesis at both the gene and protein levels. Although higher PEMF amplitudes and longer
duration exposures were capable of augmenting aggrecan mRNA expression and macromolecular sGAG forma-
tion, the levels achieved were no better than those from samples treated only once with 2 mT PEMFs for 10 min.
Col 2 expression was especially susceptible to overstimulation, being negatively impacted by exposures > 2 mT or
longer than 10 minutes. To the best of our knowledge, all published EMF studies examining chondrogenesis have
employed exposure durations between 30 min to 8 h
28, 35, 37, 38
. For instance, Mayer-Wagner et al.
38
using PEMF of
15 Hz, 5 mT, exposed MSCs undergoing chondrogenesis for 45 min every 8 h for a total of 21 days and observed
less than a 2-fold increase in type II collagen expression, with no detected e+ect on Sox9 or aggrecan expression.
Wang et al.
37
using 1, 2, and 5 mT PEMFs at the frequency of 75 Hz exposed MSCs for 3 h per day for 4 weeks
and instead observed a loss of cartilaginous phenotype associated with increased cartilage-speci=c extracellular
matrix degradation in the later stage of chondrogenic di+erentiation.
Given that most conventional PEMF exposure paradigms employ a multiple exposure strategy and have
reported positive chondrogenic outcome
35
, we sought to determine the minimal number of exposures necessary
to promote chondrogenesis (Fig. 2). We found that exposing MSCs once at the commencement of chondrogenic
di+erentiation (1x) was necessary and suβcient to induce chondrogenic gene expression (Fig.−2A), which was
sustainable for up to 21 days post chondrogenic-induction (Fig. 2B). ±e superior e+ect of a single pulse was
also con=rmed at the level of sGAG and Col 2 protein deposition (Figs−2C and 3 ). Indeed, in response to 3
Figure 7. Quanti=cation of cartilaginous ECM macromolecules generated by chondrogenically di+erentiated
MSC in response to single or three weekly PEMF exposures alone (white bars) or in combination with calcium chelator (EGTA) or TRP channel antagonists as indicated. EGTA (2 mM; dark grey bars; “E”), Ruthenium Red
(10 µM; RR, black bars; “R”) and 2-APB (100 µM; light grey bars; “C”) were included once during single PEMF
exposures, or twice during the second and third PEMF exposure. EGTA, RR and 2-APB were added 10 min
before exposure and replaced with media harvested from age-matched chondrogenic control cultures 10 min
a>er exposure. *Denotes signi=cant increase compare to non-PEMF (0 mT) control.
#
Denotes signi=cant
decrease compared to single PEMF exposure (1x).
+
Denotes signi=cant di+erence compared to respective
PEMF control (white bar). P = PEMF treatment, E = EGTA, R = Ruthenium Red (RR), C = 2-APB. Data
represent means ± SD, n = 3.
www.nature.com/scientificreports/ 9ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
exposures per week (10 min pulsing) for 3 consecutive weeks (9x treatments), ECM deposition was unchanged,
or even inhibited, relative to unexposed samples. Ours is likely the =rst report to demonstrate an e+ectiveness
of lone, 10 min, low amplitude PEMF exposures over MSC chondrogenesis, while concomitantly demonstrating
the counter productivity of prolonged or repeated exposures. ±e possibility that prolonged or repeated PEMF
exposures were merely cytotoxic, rather than truly inhibitory to chondrogenesis, was ruled out by our =nding
that total DNA content across all treatments was largely unchanged, despite lower Col 2 yield. In addition, the
amount of sGAG was either unchanged or higher than that in control non-pulsed samples, further indicating
that prolonged/repeated PEMF exposure did not adversely in×uence cell viability. Finally, the PEMF paradigm
demonstrated here to best promote chondrogenesis (2 mT applied once for 10 min) did not alter the expression of
osteogenic genes, Runx2 and ALP (Suppl. Fig. 4). Provocatively, osteogenic markers did increase following 20 min
exposure to 2 mT PEMFs, thereby substantiating our assertion that reduced chondrogenic expression is not a
re×ection of cell death, but likely deferred chondrogenesis towards osteogenesis. Our demonstration of the high
e cacy of brief and early PEMF exposure might thus help explain the existing inconsistencies and the relatively
weaker responses previously reported
28, 30, 32, 34–38
.
Chondrogenesis is known to be modulated by calcium signaling cascades of speci=c temporal sensitivity
24, 46
.
±e dependence of chondrogenesis on extracellular Ca
2+
was =rst alluded to with the demonstration that elevated
extracellular Ca
2+
promoted chondrogenic di+erentiation in chick limb bud-derived cultures
47, 48
. Moreover,
Sox9, the master transcription factor of chondrogenesis, is subject to Ca
2+
-calmodulin regulation
49
. Elevation
in cytoplasmic calcium downstream of calcium in×ux has been demonstrated in response to electric =eld (EF)
or EMF stimulation during MSC-derived osteogenesis or chondrogenesis
27, 50, 51
. We show that MSC chondro-
genesis depends on the presence of extracellular Ca
2+
, whereby a transient (10 min) elevation of extracellular
Ca
2+
or brief (10 min) exposure to PEMFs (Figs 4A and 6A) enhanced MSC chondrogenic di+erentiation in an
additive manner. Previous studies have also revealed that chondrogenesis is positively responsive to intracellular
Ca
2+
within a tightly controlled concentration window
24, 46
. A 1.25-fold increase in cytosolic Ca
2+
concentration
was shown to promote di+erentiation, whereas a moderately greater increase (1.5-fold) negatively in×uenced
in vitro chondrogenesis
48
. It is thus feasible that high amplitude or prolonged PEMF exposures elevate cyto-
plasmic calcium levels beyond the bene=cial threshold for MSC chondrogenic di+erentiation
46
. It is also well
documented that the spatial and temporal patterns of intracellular free Ca
2+
concentration play important roles
in the regulation of various cellular processes, governed not only by absolute Ca
2+
level, but also by periodic
oscillatory changes of cytosolic Ca
2+
concentration
48, 52
. MSCs undergoing chondrogenesis increase their fre-
quency of Ca
2+
oscillations (waves) in the early stages of di+erentiation
48
, coinciding with the initial period of
cellular condensation during the =rst 2�4 days
10
. Conversely, sustained elevations of extracellular calcium inhibit
chondrogenesis
48
, demonstrating a temporal requirement for calcium. Here we show that transient pulsing with
elevated calcium recapitulates the temporal characteristic of the inhibitory actions of repeated PEMF exposures
(Fig. 6A). Moreover, preventing calcium entry (with EGTA) during repeated PEMF exposure precludes the inhi-
bition (Figs 6B and 7), de=ning a developmental change in calcium-sensitivity following calcium-dependent ini-
tiation of chondrogenesis. In this respect, single brief exposition to PEMFs de=ned by a speci=ed electromagnetic
window applied during the early stages of MSC chondrogenesis may be su cient to provide the correct catalytic
rise in intracellular Ca
2+
to optimally promote the initiation of chondrogenesis. Conversely, higher exposure
intensities or multiple exposures could result in excessive or sustained calcium in×ux that may instead disrupts or
interrupts MSC-induced chondrogenesis, respectively.
Ca
2+
in×ux via membrane-associated cation channels is a key event in initiating chondrogenesis, that can be
potentially mediated by either TRP channels and/or voltage-gated calcium channels (VGCC)
24, 44, 46
. ±e transient
receptor potential (TRP) channels are a diverse and widely distributed family of cation channel broadly impli-
cated in cellular mechanotransduction
23, 53–57
. ±e TRPC and TRPV subfamilies have been broadly implicated in
calcium homeostasis, ascribed mechanically-mediated gating
44, 57–61
, as well as implicated in the developmental
programs of diverse mechanosensitive tissues
62, 63
. Previous studies have shown that blocking TRPV4 during
the initial stages of induction inhibited chondrogenesis
60, 64
. TRPV4-mediated Ca
2+
signaling is also a positive
regulator of Sox9 and as such, has been shown to promote chondrogenesis
65
and in transducing the mechanical
signals that support cartilage extracellular matrix maintenance and joint health
44, 59
. TRPC1 is expressed during
early chondrocyte expansion
66
, as well as being involved in the proliferation of mesenchymal stem cells
67
. We
detected time, intensity, and PEMF dosage-dependent up-regulations of both TRPV4 and TRPC1 that closely
correlated with the PEMF-induced expression pattern of chondrogenic markers (Fig. 5). Blocking TRPC1 and
TRPV4 channels with 2-APB
68
and Ruthenium Red
64
, respectively, in the early stage of di+erentiation e+ectively
inhibited chondrogenesis, implicating these TRP channels in the initiation of chondrogenesis, and indicating that
PEMFs recruit the activity of these channels to enhance chondrogenesis. Notably, blocking calcium-permeation
through TRPV4 channels reverses the inhibition on chondrogenic di+erentiation and TRPC1 expression during
repeated PEMF exposure, whereas blocking TRPC1 channels was unable to revert the inhibition on di+eren-
tiation and expression of TRPV4 in response to repeated PEMF exposure, suggesting that TRPV4-mediated
calcium entry antagonizes TRPC1 expression and is an essential step in initiating di+erentiation (Figs−6B and 7).
TRPV4-mediated calcium entry may thus increase a>er the induction of di+erentiation (> 2 days) serving to cur-
tail TRPC1 expression and thereby promote di+erentiation by inhibiting TRPC1-medited proliferation (Suppl.
Fig. 2).
An involvement of voltage-gated calcium channels was more difficult to establish. A predominant role
for L-type VGCCs (CACNA1, CACNA2D1) in transducing PEMF’s effects was not supported given that a
chondrogenically-e+ective dose of, Nifedipine, a L-type VGCC antagonist
53–55
, had no signi=cant e+ect on the
PEMF-induced upregulation of MSC chondrogenesis (Fig. 4B). Moreover, the expression level of the L-type chan-
nel was not correlated with changes in calcium (Suppl. Fig. 3). Ca
2+
in×ux via the low-threshold T-type VGCC
had been previously implicated in tracheal chondrogenesis
56
. ±e expression of T-type VGCC (CACNA1H) was
www.nature.com/scientificreports/ 10ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
induced by lone early exposure to PEMFs or transient calcium administration, and was suppressed by repeated
exposures to PEMF or extracellular calcium, mirroring the expression pattern of chondrogenic markers under
identical conditions (Suppl. Fig. 3A). ±e induction of the T-type calcium channel in response to PEMF/calcium
exposure more likely re×ects chondrogenic di+erentiation, rather than a fully determinant role in PEMF-induced
chondrogenesis, as its expression was in the majority of conditions unchanged (relative to control) by removal of
extracellular calcium during PEMF exposure (Suppl. Fig. 3B). Our strongest data support the interpretation that
TRPC1 and TRPV4 play a more predominant role, although not necessarily exclusive, in transducing the chon-
drogenic e+ects of PEMFs. Further work will require to fully disentangle the intricasy of calcium homeostasis
during the chondrogenic developmental process.
In summary, we have provided comprehensive characterization of the e+ects of PEMFs over MSC chondro-
genic di+erentiation. MSCs undergoing chondrogenic induction are preferentially responsive to a well-de=ned
window of PEMF stimulation of particular amplitude (2 mT), duration (10 min) and dosage (once on day 1 induc-
tion). By contrast, treatment with higher amplitude PEMFs, longer exposure durations or repeated expositions,
as are more common in the =eld, are generally counterproductive, helping explain the lack of resolution in the
=eld. Our results indicate that PEMFs mediate their e+ect by activating calcium in×ux through mechanosensitive
calcium TRP channels. ±e unprecedented eβcacy of our low amplitude, exceptionally brief and non-invasive
PEMF-exposure protocol over MSC chondrogenesis has broad clinical and practical implications for the ultimate
translation of related PEMF-based therapeutic strategies for stem cell-based cartilage regeneration.
Methods
Human bone marrow MSCs culture and chondrogenic di�erentiation. Primary human mesen-
chymal stem cells (MSCs) were purchased from RoosterBio Inc. (Frederick, MD), supplied at passage 3. ±e
MSCs was further expanded in MSC High Performance Media (RoosterBio Inc.) at 37 °C in 5% CO
2 atmosphere.
±e expanded MSCs were used at passage 5�6. Chondrogenic di+erentiation of MSCs was induced through 3D
pellet culture as previously described
16, 69
. Brie×y, 2.5 × 10
5
cells were centrifuged to form pellets and cultured in a
chondrogenic di+erentiation medium containing high glucose DMEM supplemented with 4 mM proline, 50 µg/
mL ascorbic acid, 1% ITS-Premix (Becton-Dickinson, San Jose, CA), 1 mM sodium pyruvate, and 10
−7
M dexa-
methasone (Sigma-Aldrich, St Louis, MO), in the absence of antibiotics, for up to 7 or 21 days in the presence of
10 ng/mL of transforming growth factor-β 3 (TGFβ 3; R&D Systems, Minneapolis, MN). To investigate an involve-
ment of calcium in×ux or of calcium channels in transmitting the e+ects of PEMFs, cells were pre-incubated in
chondrogenic media supplemented with elevated calcium, EGTA or particular calcium channel antagonist for
10 minutes prior to pulsing. Ten minutes a>er exposure to PEMFs the supplemented chondrogenic media was
replaced with age-matched chondrogenic media (0 mT) cultures. To attenuate extracellular calcium in×ux, 2 mM
ethylene-bis(oxyethylenenitrilo) tetraacetic acid (EGTA; Sigma) was added to the bathing media as noted. To pro-
mote extracellular Ca
2+
in×ux the bathing media was supplemented with 5 mM CaCl
2 (Sigma). To block calcium
permeation through dihydropyridine-sensitive, L-type voltage-gated calcium channel (VGCC), Nifedipine (1 µM,
Sigma) was added to the bathing media. 2-aminoethoxydiphenyl borate (2-APB, 100 µM, Sigma) and Ruthenium
Red (10 µM, Merck Millipore) were administered as indicated to block calcium entry via TRPC and TRPV chan-
nels, respectively. Aminoglycoside antibiotics such as streptomycin were excluded in all MSC expansion and
chondrogenic di+erentiation media to avoid interference with mechanosensitive ion channels
70
.
PEMF Exposure system. ±e ELF-PEMF (extremely low frequency � pulsed magnetic =eld) delivery sys-
tem has been described previously
43
. For the purposes of this study a barrage of magnetic pulses of 6 ms duration
was applied at a repetition rate of 15 Hz and at ×ux densities between 1�4 mT. Each 6 ms burst consisted of a
series of 20 consecutive asymmetric pulses of 150 µs on and o+ duration with an approximate rise time of 17 T/s.
±e background magnetic ×ux density measured in the chamber was below 1 µT between 0 Hz to 5 kHz. ±e coil
size, position and individual number of windings were numerically optimized by a CST low frequency solver
for low =eld non-uniformity over a wide frequency range taking into consideration the shielding capacity of the
µ-metallic chassis. ±e measured =eld non-uniformity did not exceed 4% within the uniform exposure region of
the coils.
PEMF treatment. To investigate the optimum dosage of PEMF, MSCs in a 3D pellet culture were exposed to
PEMFs of di+erent exposure durations, dosage and the magnetic ×ux amplitude. MSCs were subjected to PEMFs
of 1–4 mT amplitude with exposure times ranging between 5 to 60 min on the day of chondrogenic induction,
applied once or multiple times as indicated in the respective =gure legend. Cell pellets to be treated once with
PEMFs (1x) were exposed on =rst day of chondrogenic induction. Two scenarios of multiple exposures were
administrated (Fig. 2). Firstly, multiple exposures were administrated during the course of a week; double expo-
sures (2x) were applied on days 1 and 2; triple exposures (3x) on days 1, 2 and 4. Alternatively, multiple exposures
were applied on a once a week basis, for up to three week. Non-exposed (control) cells were placed within the
PEMF device without current ×ux to produce a magnetic =eld to ensure that all cells were subject to the same
climatic and mechanical conditions.
Real time PCR analysis. Chondrogenic cell pellets were digested in 0.25% Type II collagenase (Gibco,
Life Technologies) followed by centrifugation. Total RNA was extracted using the RNeasy
®
Mini Kit (Qiagen,
Germany). Reverse transcription was performed with 100 ng total RNA using iScript
™
cDNA synthesis kit
(Bio-Rad, USA). Real-time PCR was conducted using the SYBR
®
green assay on ABI 7500 Real-Time PCR System
(Applied Biosystems, Life Technologies, USA). Real-time PCR program was set at 95 °C for 10 min, followed by
40 cycles of ampli=cations, consisting of a 15 s denaturation at 95 °C and a 1 min extension step at 60 °C. Primer
sequences used in this study were according to previous publication
16
and presented as Supplementary Table 1.
www.nature.com/scientificreports/ 11ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
Re level of expression of the target gene, normalized to GAPDH, was then calculated using the 2
−ΔΔCt
formula
with reference to the undi+erentiated MSC. Results were averaged from triplicate samples of two independent
experiments.
ECM and DNA quanti�cation. Samples harvested were digested with 10 mg/mL of pepsin in 0.05 M acetic
acid at 4 °C, followed by digestion with elastase (1 mg/mL). A Blyscan sulfated glycosaminoglycan (sGAG) assay
kit (Biocolor Ltd., Newtownabbey, Ireland) was used to quantify sGAG deposition according to manufactur-
er’s protocol. Absorbance was measured at 656 nm and sGAG concentration was extrapolated from a stand-
ard curve generated using a sGAG standard. Type II Collagen (Col 2) content was measured using a captured
enzyme-linked immunosorbent assay (Chondrex, Redmond, WA). Absorbance at 490 nm was measured and
the concentration of Col 2 was extrapolated from a standard curve generated using a Col 2 standard. Values for
sGAG and Col 2 content obtained were normalized to the total DNA content of respective samples, measured
using Picogreen dsDNA assay (Molecular Probes, OR, USA). Quadruplicates of each group were analyzed from
two independent experiments.
Histological and immunohistochemical evaluation. Samples were =xed in formalin, dehydrated, par-
a n embedded, and cut into sections of 5 µm. For Safranin-O staining, the sections were incubated in hema-
toxylin (Sigma-Aldrich), washed and stained with fast green (Sigma-Aldrich), before staining with Safranin-O
solution (AcrosOrganics). For immunohistochemistry, ultra-vision detection kit (±ermo scienti=c) was used.
Endogenous peroxidase in the sections was =rst blocked with hydrogen peroxide before pepsin treatment for
20 min. Samples were treated with monoclonal antibodies of collagen type II (Clone 6B3; Chemicon Inc.) fol-
lowed by incubation with biotinylated goat anti-mouse (Lab Vision Corporation). A mouse IgG isotype (Zymed
Laboratories Inc.) was used as control for immunohistochemistry studies.
Statistical analysis. All experiments were performed in biological replicates (n = 3 or 4) and results
reported as mean ± standard deviation (SD). Statistical analysis was carried out by Students t-test for comparison
between two groups using the Microso> Excel so>ware. ±e level of signi=cance was set at p < 0.05. All quantita-
tive data reported here were averaged from at least two independent experiments.
References
1. Poole, A. R. et al . Composition and structure of articular cartilage: a template for tissue repair. Clin Orthop elat es S26–33 (2001).
2. Goldring, M. B. & Goldring, S. R. Osteoarthritis. J Cell Physiol 213, 626–34 (2007).
3. Ge, Z. et al. Osteoarthritis and therapy. Arthritis heum 55, 493–500 (2006).
4. Buckwalter, J. A. & Mankin, H. J. Articular cartilage: degeneration and osteoarthritis, repair, regeneration, and transplantation. Instr
Course Lect 47, 487–504 (1998).
5. Vasara, A.I. et al . Indentation sti+ness of repair tissue a>er autologous chondrocyte transplantation. Clin Orthop elat es , 233–42
(2005).
6. Wakitani, S. et al . Mesenchymal cell-based repair of large, full-thickness defects of articular cartilage. J Bone Joint Surg Am 76,
579–92 (1994).
7. Wakitani, S. et al . Human autologous culture expanded bone marrow mesenchymal cell transplantation for repair of cartilage defects
in osteoarthritic knees. Osteoarthritis Cartilage 10, 199–206 (2002).
8. Kuroda, R. et al . Treatment of a full-thickness articular cartilage defect in the femoral condyle of an athlete with autologous bone-
marrow stromal cells. Osteoarthritis Cartilage 15, 226–31 (2007).
9. Pittenger, M. F. et al . Multilineage potential of adult human mesenchymal stem cells. Science 284, 143–7 (1999).
10. Goldring, M. B., Tsuchimochi, K. & Ijiri, K. e control of chondrogenesis. J Cell Biochem 97, 33–44 (2006).
11. Mahmoudifar, N. & Doran, P. M. Chondrogenesis and cartilage tissue engineering: the longer road to technology development.
Trends Biotechnol 30, 166–76 (2012).
12. Woods, A., Wang, G. & Beier, F. Δegulation of chondrocyte di+erentiation by the actin cytos<eleton and adhesive interactions. J Cell
Physiol 213, 1–8 (2007).
13. DeLise, A. M., Fischer, L. & Tuan, R. S. Cellular interactions and signaling in cartilage development. Osteoarthritis Cartilage 8,
309–34 (2000).
14. Toh, W. S., Spector, M., Lee, E. H. & Cao, T. Biomaterial-mediated delivery of microenvironmental cues for repair and regeneration
of articular cartilage. Mol Pharm 8, 994–1001 (2011).
15. Klein, T. J., Malda, J., Sah, R. L. & Hutmacher, D. W. Tissue engineering of articular cartilage with biomimetic zones. Tissue Eng Part
B ev 15, 143–57 (2009).
16. Raghothaman, D. et al. Engineering cell matrix interactions in assembled polyelectrolyte =ber hydrogels for mesenchymal stem cell
chondrogenesis. Biomaterials 35, 2607–16 (2014).
17. Responte, D. J., Lee, J. K., Hu, J. C. & Athanasiou, K. A. Biomechanics-driven chondrogenesis: from embryo to adult. FASEB J 26,
3614–24 (2012).
18. Mouw, J. K., Connelly, J. T., Wilson, C. G., Michael, K. E. & Levenston, M. E. Dynamic compression regulates the expression and
synthesis of chondrocyte-speci=c matrix molecules in bone marrow stromal cells. Stem Cells 25, 655–63 (2007).
19. Grad, S., Eglin, D., Alini, M. & Stoddart, M. J. Physical stimulation of chondrogenic cells in vitro : a review. Clin Orthop elat es 469,
2764–72 (2011).
20. Cheing, G. L., Li, X., Huang, L., Kwan, R. L. & Cheung, K. K. Pulsed electromagnetic elds (PEMF) promote early wound healing
and myo=broblast proliferation in diabetic rats. Bioelectromagnetics 35, 161–9 (2014).
21. Steward, A. J., Kelly, D. J. & Wagner, D. R. e role of calcium signalling in the chondrogenic response of mesenchymal stem cells to
hydrostatic pressure. Eur Cell Mater 28, 358–71 (2014).
22. Pingguan-Murphy, B., El-Azzeh, M., Bader, D. L. & Knight, M. M. Cyclic compression of chondrocytes modulates a purinergic
calcium signalling pathway in a strain rate- and frequency-dependent manner. J Cell Physiol 209, 389–97 (2006).
23. Eijkelkamp, N., Quick, K. & Wood, J. N. Transient receptor potential channels and mechanosensation. Annu ev Neurosci 36,
519–46 (2013).
24. Matta, C. & Zakany, R. Calcium signalling in chondrogenesis: implications for cartilage repair. Front Biosci (Schol Ed) 5, 305–24
(2013).
25. Pall, M. L. Electromagnetic =elds act via activation of voltage-gated calcium channels to produce bene=cial or adverse e+ects. J Cell
Mol Med 17, 958–65 (2013).
www.nature.com/scientificreports/ 12ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
26. Ross, C. L. et al . ±e e+ect of low-frequency electromagnetic =eld on human bone marrow stem/progenitor cell di+erentiation. Stem
Cell es 15, 96–108 (2015).
27. Petecchia, L. et al. Electro-magnetic =eld promotes osteogenic di+erentiation of BM-hMSCs through a selective action on Ca(2+)-
related mechanisms. Sci ep 5, 13856 (2015).
28. Schmidt-Rohlfing, B., Silny, J., Woodruff, S. & Gavenis, K. Effects of pulsed and sinusoid electromagnetic fields on human
chondrocytes cultivated in a collagen matrix. heumatol Int 28, 971–7 (2008).
29. Vincenzi, F. et al. Pulsed electromagnetic =elds increased the anti-in×ammatory e+ect of A(2)A and A(3) adenosine receptors in
human T/C-28a2 chondrocytes and hFOB 1.19 osteoblasts. PLoS One 8, e65561 (2013).
30. De Mattei, M. et al . E+ects of pulsed electromagnetic =elds on human articular chondrocyte proliferation. Connect Tissue es 42,
269–79 (2001).
31. Sakai, A., Suzuki, K., Nakamura, T., Norimura, T. & Tsuchiya, T. E ects of pulsing electromagnetic elds on cultured cartilage cells.
Int Orthop 15, 341–6 (1991).
32. Hilz, F. M. et al . In×uence of extremely low frequency, low energy electromagnetic =elds and combined mechanical stimulation on
chondrocytes in 3-D constructs for cartilage tissue engineering. Bioelectromagnetics 35, 116–28 (2014).
33. Boopalan, P. Δ., Arumugam, S., Livingston, A., Mohanty, M. & Chittaranjan, S. Pulsed electromagnetic =eld therapy results in
healing of full thickness articular cartilage defect. Int Orthop 35, 143–8 (2011).
34. Mueller, M. B. & Tuan, R. S. Functional characterization of hypertrophy in chondrogenesis of human mesenchymal stem cells.
Arthritis heum 58, 1377–88 (2008).
35. Chen, C. H. et al . Electromagnetic fields enhance chondrogenesis of human adipose-derived stem cells in a chondrogenic
microenvironment in vitro. J Appl Physiol (1985) 114, 647–55 (2013).
36. Lim, H. L. et al . Dynamic Electromechanical Hydrogel Matrices for Stem Cell Culture. Adv Funct Mater 21 (2011).
37. Wang, J. et al. Pulsed electromagnetic =eld may accelerate in vitro endochondral ossi=cation. Bioelectromagnetics 36, 35–44 (2015).
38. Mayer-Wagner, S. et al. E+ects of low frequency electromagnetic =elds on the chondrogenic di+erentiation of human mesenchymal
stem cells. Bioelectromagnetics 32, 283–90 (2011).
39. Esposito, M. et al. Di+erentiation of human umbilical cord-derived mesenchymal stem cells, WJ-MSCs, into chondrogenic cells in
the presence of pulsed electromagnetic =elds. In Vivo 27, 495–500 (2013).
40. Ongaro, A. et al. Electromagnetic =elds counteract IL-1beta activity during chondrogenesis of bovine mesenchymal stem cells. J
Tissue Eng egen Med 9, E229–38 (2015).
41. Spaas, J. H. et al . Chondrogenic Priming at Reduced Cell Density Enhances Cartilage Adhesion of Equine Allogeneic MSCs - a
Loading Sensitive Phenomenon in an Organ Culture Study with 180 Explants. Cell Physiol Biochem 37, 651–65 (2015).
42. Berg, H. Problems of wea< electromagnetic =eld e+ects in cell biology. Bioelectrochem Bioenerg 48, 355–60 (1999).
43. Crocetti, S. et al. Low intensity and frequency pulsed electromagnetic =elds selectively impair breast cancer cell viability. PLoS One
8, e72944 (2013).
44. O’Conor, C. J., Case, N. & Guilak, F. Mechanical regulation of chondrogenesis. Stem Cell es Rer 4, 61 (2013).
45. Sun, L. Y. et al . E+ect of pulsed electromagnetic =eld on the proliferation and di+erentiation potential of human bone marrow
mesenchymal stem cells. Bioelectromagnetics 30, 251–60 (2009).
46. Fodor, J. et al. Store-operated calcium entry and calcium in×ux via voltage-operated calcium channels regulate intracellular calcium
oscillations in chondrogenic cells. Cell Calcium 54, 1–16 (2013).
47. San Antonio, J. D. & Tuan, R. S. Chondrogenesis of limb bud mesenchyme in vitro : stimulation by cations. Dev Biol 115, 313–24
(1986).
48. Matta, C. et al. Cytosolic free Ca2+ concentration exhibits a characteristic temporal pattern during in vitro cartilage di+erentiation:
a possible regulatory role of calcineurin in Ca-signalling of chondrogenic cells. Cell Calcium 44, 310–23 (2008).
49. Argentaro, A. et al. A SOX9 defect of calmodulin-dependent nuclear import in campomelic dysplasia/autosomal sex reversal. J Biol
Chem 278, 33839–47 (2003).
50. McCullen, S. D. et al . Application of low-frequency alternating current electric =elds via interdigitated electrodes: e+ects on cellular
viability, cytoplasmic calcium, and osteogenic di+erentiation of human adipose-derived stem cells. Tissue Eng Part C Methods 16,
1377–86 (2010).
51. Xu, J., Wang, W., Clark, C. C. & Brighton, C. T. Signal transduction in electrically stimulated articular chondrocytes involves
translocation of extracellular calcium through voltage-gated channels. Osteoarthritis Cartilage 17, 397–405 (2009).
52. Kawano, S. et al. Characterization of Ca(2+ ) signaling pathways in human mesenchymal stem cells. Cell Calcium 32, 165–74 (2002).
53. Curtis, T. M. & Schol=eld, C. N. Nifedipine bloc<s Ca2+ store re=lling through a pathway not involving L-type Ca2+ channels in
rabbit arteriolar smooth muscle. J Physiol 532, 609–23 (2001).
54. Wen, L. et al. L-type calcium channels play a crucial role in the proliferation and osteogenic differentiation of bone marrow
mesenchymal stem cells. Biochem Biophys es Commun 424, 439–45 (2012).
55. Zhang, J., Li, M., Kang, E. T. & Neoh, K. G. Electrical stimulation of adipose-derived mesenchymal stem cells in conductive sca olds
and the roles of voltage-gated ion channels. Acta Biomater 32, 46–56 (2016).
56. Lin, S. S. et al . Cav3.2 T-type calcium channel is required for the NFAT-dependent Sox9 expression in tracheal cartilage. Proc Natl
Acad Sci USA 111, E1990–8 (2014).
57. Guila<, F., Leddy, H. A. & Liedt<e, W. Transient receptor potential vanilloid 4: ±e sixth sense of the musculos<eletal system? Ann N
Y Acad Sci 1192, 404–9 (2010).
58. Maroto, R. et al. TRPC1 forms the stretch-activated cation channel in vertebrate cells. Nat Cell Biol 7, 179–85 (2005).
59. Eleswarapu, S. V. & Athanasiou, K. A. TRPV4 channel activation improves the tensile properties of self-assembled articular cartilage
constructs. Acta Biomater 9, 5554–61 (2013).
60. Phan, M. N. et al . Functional characterization of TRPV4 as an osmotically sensitive ion channel in porcine articular chondrocytes.
Arthritis heum 60, 3028–37 (2009).
61. Hung, C. T. Transient receptor potential vanilloid 4 channel as an important modulator of chondrocyte mechanotransduction of
osmotic loading. Arthritis heum 62, 2850–1 (2010).
62. Kurth, F. et al. Transient receptor potential vanilloid 2-mediated shear-stress responses in C2C12 myoblasts are regulated by serum
and extracellular matrix. FASEB J 29, 4726–37 (2015).
63. Clapham, D. E., Δunnels, L. W. & Strubing, C. ±e TΔP ion channel family. Nat ev Neurosci 2, 387–96 (2001).
64. Somogyi, C. S. et al . Polymodal Transient Receptor Potential Vanilloid (TRPV) Ion Channels in Chondrogenic Cells. Int J Mol Sci
16, 18412–38 (2015).
65. Muramatsu, S. et al. Functional gene screening system identi=ed TΔPV4 as a regulator of chondrogenic di+erentiation. J Biol Chem
282, 32158–67 (2007).
66. Gavenis, K. et al. Expression of ion channels of the TRP family in articular chondrocytes from osteoarthritic patients: changes
between native and in vitro propagated chondrocytes. Mol Cell Biochem 321, 135–43 (2009).
67. Torossian, F., Bisson, A., Vannier, J. P., Boyer, O. & Lamacz, M. TRPC expression in mesenchymal stem cells. Cell Mol Biol Lett 15,
600–10 (2010).
68. Lievremont, J. P., Bird, G. S. & Putney, J. W. Jr. Mechanism of inhibition of TRPC cation channels by 2-aminoethoxydiphenylborane.
Mol Pharmacol 68, 758–62 (2005).
www.nature.com/scientificreports/ 13ScieNtific REPOrTs }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
69. Yang, Z. et al. Improved mesenchymal stem cells attachment and in vitro cartilage tissue formation on chitosan-modi ed poly(L-
lactide-co-epsilon-caprolactone) sca old. Tissue Eng Part A 18, 242–51 (2012).
70. Winegar, B. D., Haws, C. M. & Lansman, J. B. Subconductance block of single mechanosensitive ion channels in skeletal muscle
bers by aminoglycoside antibiotics. J Gen Physiol 107, 433–43 (1996).
Acknowledgements
Re study was supported by University of Malaya HIR-MoE Grant (Reference number � UM.C/625/1/HIR/
MOHE/MED/32 account number – H20001-E000071) and Singapore-MIT Alliance for Research and Technology
(SMART) Foundation (ING14085-BIO). Dinesh Parate was supported by NUS Research scholarship.
Author Contributions
D.P. performed experiments, analyzed data and drafted the manuscript. A.F.O., J.F. and C.B. provided
technological expertise and contributed to the fabrication the PEMF facility. A.A.A., T.K., J.H.P.H. and A.F.O.
provided funding and critical reading of the manuscript. A.F.O. and Z.Y. designed the study, analyzed data and
provided critical revision of the manuscript. All authors reviewed the manuscript.
Additional Information
Supplementary information accompanies this paper at doi:10.1038/s41598-017-09892-w
Competing Interests: Re authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional a liations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. Re images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
� Re Author(s) 2017
www.nature.com/scientificreports
Enhancement of mesenchymal
stem cell chondrogenesis with
short-term low intensity pulsed
electromagnetic �elds
Dinesh Parate
1
, Alfredo Franco-Obregón
2,3
, Jürg Fröhlich
2,4
, Christian Beyer
4
,
Azlina A. Abbas
5
, Tunku Kamarul
5
, James H. P. Hui
1,6
& Zheng Yang
1,6
Pulse electromagnetic �elds (PEMFs) have been shown to recruit calcium-signaling cascades common
to chondrogenesis. Here we document the e�ects of speci�ed PEMF parameters over mesenchymal
stem cells (MSC) chondrogenic di�erentiation. MSCs undergoing chondrogenesis are preferentially
responsive to an electromagnetic e�cacy window de�ned by �eld amplitude, duration and frequency
of exposure. Contrary to conventional practice of administering prolonged and repetitive exposures
to PEMFs, optimal chondrogenic outcome is achieved in response to brief (10 minutes), low intensity
(2 mT) exposure to 6 ms bursts of magnetic pulses, at 15 Hz, administered only once at the onset of
chondrogenic induction. By contrast, repeated exposures diminished chondrogenic outcome and could
be attributed to calcium entry after the initial induction. Transient receptor potential (TRP) channels
appear to mediate these aspects of PEMF stimulation, serving as a conduit for extracellular calcium.
Preventing calcium entry during the repeated PEMF exposure with the co-administration of EGTA or
TRP channel antagonists precluded the inhibition of di�erentiation. This study highlights the intricacies
of calcium homeostasis during early chondrogenesis and the constraints that are placed on PEMF-based
therapeutic strategies aimed at promoting MSC chondrogenesis. The demonstrated e�cacy of our
optimized PEMF regimens has clear clinical implications for future regenerative strategies for cartilage.
Articular cartilage is an avascular tissue with low potential for self-repair. When le untreated, lesions of the
articular cartilage can lead to osteoarthritis
1–3
. e success of any technology aimed at repairing chondral defects
will thus be based on its ability to produce tissues that most closely recapitulate the mechanical and biochemical
properties of native cartilage. To this end many technologies have been advanced yet, none are without draw-
backs. e �microfracture� technique is commonly plagued by the formation of bro-cartilaginous tissue of low
dexterity
4
. Autologous chondrocytes implantation and osteochondral autogra transplantation are limited by
scarce cartilage production, low proliferative capacity of chondrocytes, chondrocyte de-di erentiation and com-
plications due to donor site morbidity
5
. Stem cell-based approaches are also being actively pursued in hopes of
improved outcome. Mesenchymal stem cells (MSCs) support chondrogenic di erentiation and are an attractive
cell source for cartilage tissue engineering. However, the neocartilage formed by conventional MSC-based repair
1
Department of Orthopaedic Surgery, Yong Loo Lin School of Medicine, National University of Singapore, NUHS
Tower Block, Level 11, 1E Kent Ridge Road, Singapore, 119288, Singapore.
2
Department of Surgery, Yong Loo Lin
School of Medicine, National University of Singapore, NUHS Tower Block, Level 8, IE Kent Ridge Road, Singapore,
119228, Singapore.
3
BioIonic Currents Electromagnetic Pulsing Systems Laboratory, BICEPS, National University
of Singapore, MD6, 14 medical Drive, #14-01, Singapore, 117599, Singapore.
4
Institute for Electromagnetic Fields,
Swiss Federal Institute of Technology (ETH), Rämistrasse 101, 8092, Zurich, Switzerland.
5
Tissue Engineering
Group (TEG), National Orthopaedic Centre of Excellence for Research and Learning (NOCERAL), Department of
Orthopaedic Surgery, Faculty of Medicine, University of Malaya, Pantai Valley, Kuala Lumpur, 50603, Malaysia.
6
Tissue Engineering Program, Life Sciences Institute, National University of Singapore, DSO (Kent Ridge) Building,
#04-01, 27 Medical Drive, Singapore, 117510, Singapore. Correspondence and requests for materials should be
addressed to A.F.-O. (email: suraf@nus.edu.sg) or J.H.P.H. (email: james_hui@nuhs.edu.sg) or Z.Y. (email: lsiyz@
nus.edu.sg)
Received: 13 April 2017
Accepted: 28 July 2017
Published: xx xx xxxx
OPEN
www.nature.com/scientificreports/ 2ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
methodologies commonly contain a mixture of =bro- and hyaline cartilage
6–8
that do not achieve the biochemical,
mechanical or functional properties of native cartilage.
MSCs can be di+erentiated along di+erent cell lineages of mesodermal origin including osteoblasts, chondro-
cytes, skeletal myocytes or visceral stromal cells
9
. Chondrogenic induction of MSCs entails proliferation, conden-
sation, di+erentiation and maturation
10
, necessitating endogenous transcriptional and developmental regulators,
cell-cell and cell-matrix interactions that, in turn, are modulated by environmental stimuli including mechanical
forces, temperature and oxygen levels
10–13
. A common objective is to recreate as closely as possible the in vivo
environmental conditions in vitro so that the rate and quality of chondrogenic development is enhanced and
the functionality of the repaired tissue improved. To this end, various environmental stimuli such as hypoxia,
mechanical, electric and electromagnetic stimulation are currently being explored
14–16
.
Mechanical stimulation can be applied in a semi-controlled manner with the use of bioreactors designed to
impart shear, compression, tension, or pressure on developing tissues. Appropriately applied mechanical stim-
ulation positively in×uences MSC-induced chondrogenic di+erentiation, ECM deposition and the mechanical
properties of the generated cartilage
17–20
. At the cellular level the transduction of mechanical signals (mech-
anotransduction) involves their conversion into biochemical responses, often with the assistance of mech-
anosensitive calcium channels
21–24
. Electromagnetic =eld (EMF)-stimulation has been shown to promote cell
di+erentiation via the modulation of extracellular calcium entry via plasma membrane-embedded cation chan-
nels
25–27
, raising the intriguing possibility that EMFs may be recruiting related pathways.
Studies examining time-variant or pulsing electromagnetic =elds (PEMFs) have alluded to a bene=t over
articular chondrocytes or cartilaginous tissue in vitro , particularly with reference to chondrocyte proliferation,
extracellular matrix (ECM) deposition, secretory activity and in×ammatory status
28–35
. Studies have also exam-
ined the e+ects of PEMF-treatment over the chondrogenic di+erentiation of stem cells derived from bone mar-
row
36–38
, adipose
35
, umbilical cord Wharton jelly
39
, synovial ×uid
40
or peripheral blood sources
40
. Re reported
consequences of PEMF-stimulation over chondrogenesis, however, are largely inconsistent. Some studies report
modest enhancements in the gene expression of Sox9, aggrecan, type II collagen (Col 2) as well as deposition
of sulfated glycosaminoglycan (sGAG), typically on the order of 2-folds
35, 36
, whereas other studies show little
to no e+ect
38–41
. On the extreme end of the spectrum, Wang et al .
37
reported inhibition of both Sox9 and Col 2
expression concomitant with induction of hypertrophy and mineralization in response to exposures of 3 h per
day at an amplitude of 1 mT. Obvious di+erences in stimulation protocols likely underlie reported discrepancies.
Existing EMF studies have typically employed exposure durations between 30 minutes to 8 h per day and more
consistently in the low milli Tesla amplitude range (3–5 mT). Empirical determination of the appropriate expo-
sure and signal parameters for a speci=c biological response and given tissue are essential as there are indications
that cell responses to magnetic =elds obey an electromagnetic eβcacy window de=ned by a speci=c combination
of frequency, amplitude and time of exposure that gives rise to optimum cell response
42, 43
. Here, we systemat-
ically characterized the e+ects of PEMF exposure over MSC chondrogenic di+erentiation by varying the =eld
amplitude, exposure duration and dosage with an emphasis on determining the briefest and lowest amplitude
electromagnetic exposure to render a developmental outcome. Given that both mechanical stimuli and calcium
entry
21, 22
in×uences chondrogenic di+erentiation, we investigated the ability of PEMF exposure to in×uence cal-
cium homeostasis during early induction of MSCs into the chondrogenic lineage, in particular that attributed to
the Transient Receptor Potential (TRP) family of cation-permeable channels, which has been broadly implicated
in cellular mechanotransduction
23, 44
. We show that brief and single exposures to low amplitude PEMFs were
most e+ective at stimulating MSC chondrogenesis. Our results also implicate the involvement of calcium in×ux
and the mechanosensitive TRP channels, TRPC1 and TRPV4, in the chondrogenic development stimulated by
targeted PEMF exposure.
Results
E�ect of PEMF intensities and exposure durations on MSC chondrogenesis. We =rst sought to
determine the magnetic =eld amplitude and duration of exposure at which MSCs undergoing chondrogenic
induction are most responsive using starting conditions preliminarily tested in MSCs for chondrogenic regen-
eration
39
. MSC pellets in chondrogenic di+erentiation medium were subjected to single exposures to PEMFs
of 10 min duration at intensities ranging in amplitudes between 0–4 mT (Fig. 1A), then subjected to exposure
durations between 5 and 60 min at 2 mT intensity (Fig. 1B), applied on the =rst day of chondrogenic induction.
RNA analysis monitoring MSC chondrogenic progression at 7 days post-induction showed greatest increases in
response to 10 min exposures applied at an amplitude of 2 mT as evidenced by enhancements in Sox9, aggrecan
and Col 2 mRNA expression. By contrast, lower (1 mT) or higher (>3 mT) amplitude of PEMFs (Fig. 1A), or
briefer (5 min) or longer (>20 min) durations of exposure (Fig. 1B), resulted in overall smaller e+ect sizes. ±e
same EMF e cacy window translated to the expression of cartilaginous ECM macromolecular proteins (Fig. 1C).
In response to 2 mT amplitude pulsing, a 3-fold increase in Col 2 protein was detected 21 days a er chondrogenic
induction, whereas no increase was detected with exposure to 3 mT. Moreover, a 2-fold increase in sGAG was
detected in response to exposure to 2 mT PEMFs, whereas 3 mT PEMFs produced a signi=cantly smaller increase.
±e relative ine+ectiveness of prolonged exposure to PEMFs was also corroborated at the protein level. Sixty min
exposures to 2 mT PEMFs did not elicit signi=cant increases in Col 2 formation than 10 min exposures (Fig. 1C).
With reference to sGAG production, PEMF amplitudes greater than 2 mT, or exposure durations of one hour,
produced inferior results to 10 min exposures at 2 mT.
Dosage e�ects of PEMFs over MSC chondrogenesis. We next investigated the e+ect of repetitive
exposures to PEMFs. MSCs were exposed to PEMFs at 2 mT for 10 min/day once, twice or thrice on days 1, 2
and 4 following chondrogenic induction (Fig. 2A). RNA analysis a>er 7 days of di+erentiation showed that a
single exposure produced the greatest and most consistent increase in the expression of chondrogenic markers
www.nature.com/scientificreports/ 3ScieNtific REPOrTs }� 9421 D �10.103~ s415~-01}-0~2-w
(Fig. 2A). In another series of experiments, MSCs pellets were exposed once on the =rst day of chondrogenic
induction, or weekly for 3 consecutive weeks (Fig. 2B). RNA analysis a>er 21 days of chondrogenic di+erentia-
tion showed that a single exposure to 2 mT PEMFs for 10 min given on day 1 of induction gave the greatest and
most consistent increase in expression of chondrogenic markers relative to no exposure (0 mT) (Fig. 2B). ±ree
weekly exposures either rendered no additional bene=t (Col 2) or gave similar results to control (0 mT) (Sox9 and
aggrecan). Moreover, the amount of ECM produced was inversely related to the total number of exposures. Single
exposures produced > 2-folds and ~1-fold increases in Col 2 and sGAG, respectively (Fig. 2C), whereas triple
weekly exposures for three weeks (9 total exposures) completely precluded an increase Col 2 and sGAG forma-
tion. ±e change in the amount of DNA across samples varied less than 0.2-fold, although reaching signi=cances
at 2 mT, indicating that cell proliferation was only modestly a+ected within our pellet culture system (Fig.−2C).
E�ect of PEMF treatment to deposition of ECM. ECM deposition in response to PEMF-exposure was
also analyzed using Safranin O staining for proteoglycan and immunohistochemical staining for type II collagen
(Fig. 3). Stained images of day 21 samples showed an enhanced deposition of proteoglycan and type II collagen in
samples exposed only once to 2 mT for 10 min as compared to control (0 mT). By contrast, MSC samples exposed
for longer (60 min), to greater amplitude (3 mT) or repeatedly (3x, 9x) yielded comparable, or inferior, ECM
deposition to control.
Ca
2+
entry pathways implicated in transducing the e�ects of PEMFs over MSC di�erentia-
tion.
To investigate whether PEMF-stimulated MSC chondrogenesis was depended on calcium in×ux, EGTA
(2 mM) was co-administered to the culture medium during PEMF−exposure and summarily replaced a>erwards
with age-matched chondrogenic control media. RNA analysis at day 7 showed that the inclusion of EGTA sig-
ni=cantly decreased the mRNA expression of Sox9, Col 2 and aggrecan in PEMF-treated samples (Fig.−4A),
Figure 1. E+ects of PEMF amplitude (A) and exposure duration (B) on MSC chondrogenesis. Real-time
PCR analysis of cartilaginous markers expression a>er 7 days of di+erentiation was normalized to GAPDH and presented as fold-changes relative to levels in undi+erentiated MSC. (C) Quanti=cation of cartilaginous extracellular matrix macromolecules (Col−2 and sGAG) generated a>er 21 days of chondrogenic di+erentiation of MSC subjected to distinct PEMF parameters. All data shown are mean ± SD, n = 6 from 2 independent
experiments. *Denotes signi=cant increase, or decrease, compared to non-PEMF control.
#
Denotes signi=cant
decrease compared to 2 mT (A), or 10 min (B) PEMF exposure.
www.nature.com/scientificreports/ 4ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
indicating that PEMF-exposure stimulates calcium in×ux. Conversely, transiently supplementing the di+eren-
tiation medium with elevated extracellular Ca
2+
(5 mM CaCl
2) enhanced the mRNA expression of Sox9, Col 2
and aggrecan in otherwise non-exposed samples, and moreover, accentuated chondrogenic gene expression in
PEMF-treated samples. ±ese results corroborate that calcium in×ux is part of the upstream signalling cascade
recruited by PEMFs contributing to chondrogenic induction.
To reveal the Ca
2+
in×ux pathway recruited by PEMFs, we pharmacologically dissected the contribution of
candidate channels utilizing 2-APB (100 µM) or Ruthenium Red (RR, 10 µM) as TRPC or TRPV cation chan-
nel antagonists, respectively, or Nifedipine (1 µM), as a dihydropyridine-sensitive, L-type voltage-gated calcium
channel (VGCC) antagonist. Calcium channel antagonists were included into the di+erentiation medium 10 min
Figure 2. Dosage e+ects of PEMFs over MSC chondrogenesis. (A) MSCs were exposed once (1x), twice (2x) or
thrice (3x) per week. (B) MSCs were subjected to either a single exposure on day 1 of chondrogenic induction (1x) or once per week for 3 weeks (3x). Real-time PCR analysis of cartilaginous markers expression at 7 (A) or 21 days (B) a>er the induction of di+erentiation was normalized to GAPDH and presented as fold-changes relative to level in undi+erentiated MSCs. (C) Quanti=cation of cartilaginous ECM macromolecules generated
during chondrogenic di+erentiation of MSCs in response to distinct PEMF dosing as indicated. MSC pellets
were subjected to either a single PEMF exposure given on day 1 of chondrogenic induction (1x) once per week
for 3 weeks (3x), or thrice weekly for 3 weeks (9x). Data represents the mean ± SD, n = 6 from 2 independent
experiments. *Denotes signi=cant increase compare to non-PEMF (0 mT) control.
#
Denotes signi=cant decrease
compared to single PEMF (1x) exposure.
www.nature.com/scientificreports/ 5ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
before exposure to PEMFs and removed immediately a>erwards with age-matched control chondrogenic media.
Both 2-APB and Ruthenium Red completely inhibited the PEMF-triggered up-regulation of chondrogenic
genes, whereas Nifedipine had no signi=cant inhibitory e+ect (Fig.−4B). Chondrogenic inhibition by 2-APB and
Ruthenium Red was also observed in non-exposed samples, indicating that TRPC- and TRPV-mediated calcium
entry are similarly involved in constitutive chondrogenesis upon induction. By contrast, VGCC-mediated Ca
2+
entry does not appear to play a predominant role in the early induction of chondrogenesis.
We next investigated the expression pro=les of TRP channels (TRPC1, TRPC6, TRPV1, TRPV4, TRPV6)
previously implicated in chondrogenesis and correlated these to our PEMF-induced chondrogenic responses.
Amongst the panel of candidate TRP channels, the expression of TRPC1 and TRPV4 most closely correlated with
our delineated magnetic e cacy window governing chondrogenesis with reference to PEMF amplitude, duration
Figure 3. Histological analysis of pellets exposed to PEMFs of distinct amplitude, duration and dosage. Pellets
were harvested at day 21, sectioned and subjected to Safranin O or type II collagen immunohistochemistry
staining. Images presented were represenation of n = 3, taken at 100× magni=cation.
Figure 4. Investigation of calcium entry pathways implicated in the PEMF-e+ect. (A) Involvement of Ca
2+
in×ux in mediating the e+ects of PEMF-induced MSC chondrogenic di+erentiation. MSCs were exposed for 10 min at 2 mT alone (control, white bars), or in the presence of 2 mM EGTA (dark grey bars) or 5 mM
CaCl
2 (hatched bars) transiently added to the culture media. EGTA and CaCl
2 were included to the bathing
media 10 min before exposure and replaced with age-matched media control cultures 10 min a>er exposure.
(B) Involvement of candidate calcium channels in mediating the e+ect of PEMs over MSC chondrogenic di+erentiation. Control MSC chondrogenic di+erentiation medium (white bars) was supplemented with Nifedipine (1 µM, light grey bars), Ruthenium Red (RR, 10 µM, black bars), or 2-APB (100 µM, dark grey bars)
10 min before exposure and replaced with age-matched media control cultures 10 min a>er exposure. Real-
time PCR analysis was performed on day 7 of di+erentiation. Data represent the means ± SD, n = 6 from 2
independent experiments. *Denotes signi=cant increase, or decrease, compared to non-PEMF (0 mT) control.
#
Denotes signi=cant decrease relativeto 2 mT PEMF treatment.
www.nature.com/scientificreports/ 6ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
and dosage (Fig. 5 and Suppl. Figures 1 and 2). ±ese results corroborate an involvement of TRPC1 and TRPV4
in the PEMF-induced enhancement of chondrogenic di+erentiation of MSC we observed.
E�ect of recurring calcium in�ux on MSCchondrogenesis. We next investigated whether calcium
entry, particularly that via TRPchannels underlies the inhibitory e+ect observed with repeated PEMF exposures.
MSCs were exposed once, twice or thrice to 2 mT PEMFs for 10 min or, alternatively, exposed for 10 min to
aged-matched control di+erentiation media containing elevated extracellular calcium (5 mM CaCl
2) in lieu of
PEMF exposure. RNA analysis at day 7 showed that MSCs treated once with PEMFs, or transiently administered
elevated calcium, on day 1 exhibited enhanced chondrogenesis to comparable levels. By contrast, subsequent
exposures to elevated calcium, on days 2 and 4 suppressed chondrogenesis mirrored the e+ect of multiple expo-
sures to PEMFs (Fig. 6A and Suppl. Fig. 3A).
Analogously, precluding calcium entry (with EGTA) also exhibited dichotomous e+ects if applied during
the =rst versus the second or third exposition to PEMFs, although in opposite direction to that observed with
calcium administration or PEMFs. Whereas EGTA added during the initial exposure to PEMFs (1x) prevented
PEMF-induced chondrogenesis, EGTA applied during the second or third exposure partially counteracted the
Figure 5. Expression pro=les of TRPC1 and TRPV4 in response to determined PEMF eβcacy window
regulating MSC chondrogenesis. Real-time PCR analysis of TRPC1 and V4 exposed to di+erent (A) intensities, (B) durations, and (C) dosages of PEMFs. Data represent the means ± SD, n = 6 from 2 independent
experiments. *Denotes signi=cant increase, or decrease, compared to non-PEMF (0 mT) control.
#
Denotes
signi=cant decrease relative to 2−mT (A), 10 min (B), or single (1x, C) PEMF treatment.
www.nature.com/scientificreports/ 7ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
inhibition of di+erentiation exerted by serial PEMF exposure (Fig.−6B and Suppl. Fig. 3B). Notably, impeding cal-
cium entry with transient application of EGTA during both the second and third PEMF exposure was capable of
almost completely reversing the inhibition of chondrogenesis observed with repeated PEMF exposures, suggest-
ing that PEMFs are activating disparately functioning calcium mechanisms at early (day 1) and later stages (> day
2) of chondrogenic-induction that confer opposite e+ects over chondrogenesis. In contrast to the bene=cial e+ect
of calcium in×ux induced by PEMF at the initial stage of chondrogenesis, subsequent induction of calcium in×ux
by repeated pulsing at later stages of chondrogenesis was suppressive of MSC chondrogenesis. ±e contribution of
TRPC- and TRPV-mediated calcium entry to the chondrogenic-inhibition observed with repeated PEMF expo-
sures was investigated by co-administering 2-APB (100 µM) or Ruthenium Red (RR, 10 µM), respectively, during
PEMF exposure. As observed with transient EGTA application, antagonism of TRPC1/V4-mediated calcium
entry during the =rst exposition to PEMFs was strongly inhibitory of PEMF-induced chondrogenesis, where-
asTRPC1/V4 antagonism during subsequent PEMF expositions was somewhat less protective than EGTA over
di+erentiation (Fig.−6B), implicating other yet to be determined calcium pathways in the later calcium-dependent
inhibitory phase of chondrogenic progression. Notably, Ruthenium Red (TRPV4 antagonist) was capable of
reverting the inhibition of di+erentiation and expression of TRPC1 expression in response to repeated PEMFing,
whereas 2-APB (TRPC antagonist) was unable to revert the inhibition of di+erentiation and TRPV4 expression
in response to repeated PEMFing, suggesting that TRPV4-mediated calcium entry antagonizes TRPC1 expres-
sion leading up to di+erentiation suppression. ±e dichotomous e+ects of precluding calcium entry by EGTA,
Ruthenium Red or 2-APB when applied during the =rst, or the second and third, exposition to PEMFs was
corroborated at the protein level. Preventing calcium entry during the initial exposure to PEMFs (1x) prevented
Figure 6. (A) MSC chondrogenic di+erentiation in response to multiple exposures to PEMFs or exogenously elevated calcium. MSCs were subjected to either PEMF stimulation alone (white bars) or with transient supplementation of CaCl
2 alone (5 mM; hatched bars), once (1x), twice (2x) or thrice (3x) in a week. Dotted
lines refers to expression level of non-treated controls. Real-time PCR analysis was performed on day 7 of chondrogenic di+erentiation. *Denotes signi=cant increase relative to non-PEMF (0 mT) control.
#
and
+
denote signi=cant di+erences relative to respective single (1x) exposure (white and hatched bars, respectively). P = PEMF treatment, Ca = CaCl
2 supplementation. (B) MSC chondrogenic di+erentiation in response to
multiple exposures to PEMFs alone (white bars) or in combination with calcium chelator (EGTA) or TRP channel antagonists. EGTA (2 mM; dark grey bars; “E”), Ruthenium Red (10 µM; RR, black bars; “R”) or 2-APB
(100 µM; light grey bars; �C�) was added to the MSC di erentiation medium during PEMF expoure applied
once (1x), twice (2) or thrice (3x) per week. EGTA, RR and 2-APB were included 10 min before exposure
and replaced with media harvested from age-matched chondrogenic control cultures 10 min a>er exposure.
*Denotes signi=cant increase compare to non-PEMF (0 mT) control.
#
Denotes signi=cant decrease compared to
single PEMF exposure (1x).
+
Denotes signi=cant di+erence compared to respective PEMF control (white bar).
P = PEMF treatment, E = EGTA, R = Ruthenium Red (RR), C = 2-APB. Data shown are means ± SD, n = 6 from
2 independent experiments.
www.nature.com/scientificreports/ 8ScieNtific REPOrTs }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
PEMF-induced cartilaginous Col 2 and sGAG formation, while blocking calcium entry at later exposures coun-
teracted the inhibition of di+erentiation exerted by serial PEMF exposure (Fig.−7). Voltage-gated L-type calcium
channels, on the other hand, do not appear to be strongly implicated in the response as the expression of its subu-
nits (CACNA1C and CACNA2D1) was not perturbed by PEMF or calcium treatment (Suppl. Fig. 3).
Discussion
Pulsed electromagnetic =elds (PEMFs) have been demonstrated to be in×uential in numerous biological functions
including progenitor cell fate determination and di+erentiation. PEMF-based therapies have been previously
shown to enhance chondrocyte and cartilage explant anabolism while also limiting the catabolic consequences of
in×ammatory cytokines
29–31, 33–35, 40
. PEMF exposure has been also reported to enhance the chondrogenic induc-
tion of stem cells
34–37, 45
. Nevertheless, inconsistent and con×icting results plague the scienti=c literature in this
area of study, with PEMF exposures typically being applied on the order of hours per day for several days or
weeks at a time. Here we report a high-e cacy of unprecedentedly brief (10 min applied once) PEMF expo-
sure at inducing MSC chondrogenesis. We consistently detected increases in Sox9, Col 2 and aggrecan mRNA
(>2-folds) in response to lone exposure to 2 mT PEMFs applied at the commencement of induction for only
10 min (Fig. 1). ±ese increments in mRNA later translated into increased chondrogenic ECM protein forma-
tion (> 2-fold) a>er 21 days of di+erentiation. By contrast, stimulation with greater amplitudes (> 3 mT), longer
exposures (> 20 min) or more frequently (> 2x/week) rendered no additional bene=t, or was even less e+ective
at promoting chondrogenesis at both the gene and protein levels. Although higher PEMF amplitudes and longer
duration exposures were capable of augmenting aggrecan mRNA expression and macromolecular sGAG forma-
tion, the levels achieved were no better than those from samples treated only once with 2 mT PEMFs for 10 min.
Col 2 expression was especially susceptible to overstimulation, being negatively impacted by exposures > 2 mT or
longer than 10 minutes. To the best of our knowledge, all published EMF studies examining chondrogenesis have
employed exposure durations between 30 min to 8 h
28, 35, 37, 38
. For instance, Mayer-Wagner et al.
38
using PEMF of
15 Hz, 5 mT, exposed MSCs undergoing chondrogenesis for 45 min every 8 h for a total of 21 days and observed
less than a 2-fold increase in type II collagen expression, with no detected e+ect on Sox9 or aggrecan expression.
Wang et al.
37
using 1, 2, and 5 mT PEMFs at the frequency of 75 Hz exposed MSCs for 3 h per day for 4 weeks
and instead observed a loss of cartilaginous phenotype associated with increased cartilage-speci=c extracellular
matrix degradation in the later stage of chondrogenic di+erentiation.
Given that most conventional PEMF exposure paradigms employ a multiple exposure strategy and have
reported positive chondrogenic outcome
35
, we sought to determine the minimal number of exposures necessary
to promote chondrogenesis (Fig. 2). We found that exposing MSCs once at the commencement of chondrogenic
di+erentiation (1x) was necessary and suβcient to induce chondrogenic gene expression (Fig.−2A), which was
sustainable for up to 21 days post chondrogenic-induction (Fig. 2B). ±e superior e+ect of a single pulse was
also con=rmed at the level of sGAG and Col 2 protein deposition (Figs−2C and 3 ). Indeed, in response to 3
Figure 7. Quanti=cation of cartilaginous ECM macromolecules generated by chondrogenically di+erentiated
MSC in response to single or three weekly PEMF exposures alone (white bars) or in combination with calcium chelator (EGTA) or TRP channel antagonists as indicated. EGTA (2 mM; dark grey bars; “E”), Ruthenium Red
(10 µM; RR, black bars; “R”) and 2-APB (100 µM; light grey bars; “C”) were included once during single PEMF
exposures, or twice during the second and third PEMF exposure. EGTA, RR and 2-APB were added 10 min
before exposure and replaced with media harvested from age-matched chondrogenic control cultures 10 min
a>er exposure. *Denotes signi=cant increase compare to non-PEMF (0 mT) control.
#
Denotes signi=cant
decrease compared to single PEMF exposure (1x).
+
Denotes signi=cant di+erence compared to respective
PEMF control (white bar). P = PEMF treatment, E = EGTA, R = Ruthenium Red (RR), C = 2-APB. Data
represent means ± SD, n = 3.
www.nature.com/scientificreports/ 9ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
exposures per week (10 min pulsing) for 3 consecutive weeks (9x treatments), ECM deposition was unchanged,
or even inhibited, relative to unexposed samples. Ours is likely the =rst report to demonstrate an e+ectiveness
of lone, 10 min, low amplitude PEMF exposures over MSC chondrogenesis, while concomitantly demonstrating
the counter productivity of prolonged or repeated exposures. ±e possibility that prolonged or repeated PEMF
exposures were merely cytotoxic, rather than truly inhibitory to chondrogenesis, was ruled out by our =nding
that total DNA content across all treatments was largely unchanged, despite lower Col 2 yield. In addition, the
amount of sGAG was either unchanged or higher than that in control non-pulsed samples, further indicating
that prolonged/repeated PEMF exposure did not adversely in×uence cell viability. Finally, the PEMF paradigm
demonstrated here to best promote chondrogenesis (2 mT applied once for 10 min) did not alter the expression of
osteogenic genes, Runx2 and ALP (Suppl. Fig. 4). Provocatively, osteogenic markers did increase following 20 min
exposure to 2 mT PEMFs, thereby substantiating our assertion that reduced chondrogenic expression is not a
re×ection of cell death, but likely deferred chondrogenesis towards osteogenesis. Our demonstration of the high
e cacy of brief and early PEMF exposure might thus help explain the existing inconsistencies and the relatively
weaker responses previously reported
28, 30, 32, 34–38
.
Chondrogenesis is known to be modulated by calcium signaling cascades of speci=c temporal sensitivity
24, 46
.
±e dependence of chondrogenesis on extracellular Ca
2+
was =rst alluded to with the demonstration that elevated
extracellular Ca
2+
promoted chondrogenic di+erentiation in chick limb bud-derived cultures
47, 48
. Moreover,
Sox9, the master transcription factor of chondrogenesis, is subject to Ca
2+
-calmodulin regulation
49
. Elevation
in cytoplasmic calcium downstream of calcium in×ux has been demonstrated in response to electric =eld (EF)
or EMF stimulation during MSC-derived osteogenesis or chondrogenesis
27, 50, 51
. We show that MSC chondro-
genesis depends on the presence of extracellular Ca
2+
, whereby a transient (10 min) elevation of extracellular
Ca
2+
or brief (10 min) exposure to PEMFs (Figs 4A and 6A) enhanced MSC chondrogenic di+erentiation in an
additive manner. Previous studies have also revealed that chondrogenesis is positively responsive to intracellular
Ca
2+
within a tightly controlled concentration window
24, 46
. A 1.25-fold increase in cytosolic Ca
2+
concentration
was shown to promote di+erentiation, whereas a moderately greater increase (1.5-fold) negatively in×uenced
in vitro chondrogenesis
48
. It is thus feasible that high amplitude or prolonged PEMF exposures elevate cyto-
plasmic calcium levels beyond the bene=cial threshold for MSC chondrogenic di+erentiation
46
. It is also well
documented that the spatial and temporal patterns of intracellular free Ca
2+
concentration play important roles
in the regulation of various cellular processes, governed not only by absolute Ca
2+
level, but also by periodic
oscillatory changes of cytosolic Ca
2+
concentration
48, 52
. MSCs undergoing chondrogenesis increase their fre-
quency of Ca
2+
oscillations (waves) in the early stages of di+erentiation
48
, coinciding with the initial period of
cellular condensation during the =rst 2�4 days
10
. Conversely, sustained elevations of extracellular calcium inhibit
chondrogenesis
48
, demonstrating a temporal requirement for calcium. Here we show that transient pulsing with
elevated calcium recapitulates the temporal characteristic of the inhibitory actions of repeated PEMF exposures
(Fig. 6A). Moreover, preventing calcium entry (with EGTA) during repeated PEMF exposure precludes the inhi-
bition (Figs 6B and 7), de=ning a developmental change in calcium-sensitivity following calcium-dependent ini-
tiation of chondrogenesis. In this respect, single brief exposition to PEMFs de=ned by a speci=ed electromagnetic
window applied during the early stages of MSC chondrogenesis may be su cient to provide the correct catalytic
rise in intracellular Ca
2+
to optimally promote the initiation of chondrogenesis. Conversely, higher exposure
intensities or multiple exposures could result in excessive or sustained calcium in×ux that may instead disrupts or
interrupts MSC-induced chondrogenesis, respectively.
Ca
2+
in×ux via membrane-associated cation channels is a key event in initiating chondrogenesis, that can be
potentially mediated by either TRP channels and/or voltage-gated calcium channels (VGCC)
24, 44, 46
. ±e transient
receptor potential (TRP) channels are a diverse and widely distributed family of cation channel broadly impli-
cated in cellular mechanotransduction
23, 53–57
. ±e TRPC and TRPV subfamilies have been broadly implicated in
calcium homeostasis, ascribed mechanically-mediated gating
44, 57–61
, as well as implicated in the developmental
programs of diverse mechanosensitive tissues
62, 63
. Previous studies have shown that blocking TRPV4 during
the initial stages of induction inhibited chondrogenesis
60, 64
. TRPV4-mediated Ca
2+
signaling is also a positive
regulator of Sox9 and as such, has been shown to promote chondrogenesis
65
and in transducing the mechanical
signals that support cartilage extracellular matrix maintenance and joint health
44, 59
. TRPC1 is expressed during
early chondrocyte expansion
66
, as well as being involved in the proliferation of mesenchymal stem cells
67
. We
detected time, intensity, and PEMF dosage-dependent up-regulations of both TRPV4 and TRPC1 that closely
correlated with the PEMF-induced expression pattern of chondrogenic markers (Fig. 5). Blocking TRPC1 and
TRPV4 channels with 2-APB
68
and Ruthenium Red
64
, respectively, in the early stage of di+erentiation e+ectively
inhibited chondrogenesis, implicating these TRP channels in the initiation of chondrogenesis, and indicating that
PEMFs recruit the activity of these channels to enhance chondrogenesis. Notably, blocking calcium-permeation
through TRPV4 channels reverses the inhibition on chondrogenic di+erentiation and TRPC1 expression during
repeated PEMF exposure, whereas blocking TRPC1 channels was unable to revert the inhibition on di+eren-
tiation and expression of TRPV4 in response to repeated PEMF exposure, suggesting that TRPV4-mediated
calcium entry antagonizes TRPC1 expression and is an essential step in initiating di+erentiation (Figs−6B and 7).
TRPV4-mediated calcium entry may thus increase a>er the induction of di+erentiation (> 2 days) serving to cur-
tail TRPC1 expression and thereby promote di+erentiation by inhibiting TRPC1-medited proliferation (Suppl.
Fig. 2).
An involvement of voltage-gated calcium channels was more difficult to establish. A predominant role
for L-type VGCCs (CACNA1, CACNA2D1) in transducing PEMF’s effects was not supported given that a
chondrogenically-e+ective dose of, Nifedipine, a L-type VGCC antagonist
53–55
, had no signi=cant e+ect on the
PEMF-induced upregulation of MSC chondrogenesis (Fig. 4B). Moreover, the expression level of the L-type chan-
nel was not correlated with changes in calcium (Suppl. Fig. 3). Ca
2+
in×ux via the low-threshold T-type VGCC
had been previously implicated in tracheal chondrogenesis
56
. ±e expression of T-type VGCC (CACNA1H) was
www.nature.com/scientificreports/ 10ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
induced by lone early exposure to PEMFs or transient calcium administration, and was suppressed by repeated
exposures to PEMF or extracellular calcium, mirroring the expression pattern of chondrogenic markers under
identical conditions (Suppl. Fig. 3A). ±e induction of the T-type calcium channel in response to PEMF/calcium
exposure more likely re×ects chondrogenic di+erentiation, rather than a fully determinant role in PEMF-induced
chondrogenesis, as its expression was in the majority of conditions unchanged (relative to control) by removal of
extracellular calcium during PEMF exposure (Suppl. Fig. 3B). Our strongest data support the interpretation that
TRPC1 and TRPV4 play a more predominant role, although not necessarily exclusive, in transducing the chon-
drogenic e+ects of PEMFs. Further work will require to fully disentangle the intricasy of calcium homeostasis
during the chondrogenic developmental process.
In summary, we have provided comprehensive characterization of the e+ects of PEMFs over MSC chondro-
genic di+erentiation. MSCs undergoing chondrogenic induction are preferentially responsive to a well-de=ned
window of PEMF stimulation of particular amplitude (2 mT), duration (10 min) and dosage (once on day 1 induc-
tion). By contrast, treatment with higher amplitude PEMFs, longer exposure durations or repeated expositions,
as are more common in the =eld, are generally counterproductive, helping explain the lack of resolution in the
=eld. Our results indicate that PEMFs mediate their e+ect by activating calcium in×ux through mechanosensitive
calcium TRP channels. ±e unprecedented eβcacy of our low amplitude, exceptionally brief and non-invasive
PEMF-exposure protocol over MSC chondrogenesis has broad clinical and practical implications for the ultimate
translation of related PEMF-based therapeutic strategies for stem cell-based cartilage regeneration.
Methods
Human bone marrow MSCs culture and chondrogenic di�erentiation. Primary human mesen-
chymal stem cells (MSCs) were purchased from RoosterBio Inc. (Frederick, MD), supplied at passage 3. ±e
MSCs was further expanded in MSC High Performance Media (RoosterBio Inc.) at 37 °C in 5% CO
2 atmosphere.
±e expanded MSCs were used at passage 5�6. Chondrogenic di+erentiation of MSCs was induced through 3D
pellet culture as previously described
16, 69
. Brie×y, 2.5 × 10
5
cells were centrifuged to form pellets and cultured in a
chondrogenic di+erentiation medium containing high glucose DMEM supplemented with 4 mM proline, 50 µg/
mL ascorbic acid, 1% ITS-Premix (Becton-Dickinson, San Jose, CA), 1 mM sodium pyruvate, and 10
−7
M dexa-
methasone (Sigma-Aldrich, St Louis, MO), in the absence of antibiotics, for up to 7 or 21 days in the presence of
10 ng/mL of transforming growth factor-β 3 (TGFβ 3; R&D Systems, Minneapolis, MN). To investigate an involve-
ment of calcium in×ux or of calcium channels in transmitting the e+ects of PEMFs, cells were pre-incubated in
chondrogenic media supplemented with elevated calcium, EGTA or particular calcium channel antagonist for
10 minutes prior to pulsing. Ten minutes a>er exposure to PEMFs the supplemented chondrogenic media was
replaced with age-matched chondrogenic media (0 mT) cultures. To attenuate extracellular calcium in×ux, 2 mM
ethylene-bis(oxyethylenenitrilo) tetraacetic acid (EGTA; Sigma) was added to the bathing media as noted. To pro-
mote extracellular Ca
2+
in×ux the bathing media was supplemented with 5 mM CaCl
2 (Sigma). To block calcium
permeation through dihydropyridine-sensitive, L-type voltage-gated calcium channel (VGCC), Nifedipine (1 µM,
Sigma) was added to the bathing media. 2-aminoethoxydiphenyl borate (2-APB, 100 µM, Sigma) and Ruthenium
Red (10 µM, Merck Millipore) were administered as indicated to block calcium entry via TRPC and TRPV chan-
nels, respectively. Aminoglycoside antibiotics such as streptomycin were excluded in all MSC expansion and
chondrogenic di+erentiation media to avoid interference with mechanosensitive ion channels
70
.
PEMF Exposure system. ±e ELF-PEMF (extremely low frequency � pulsed magnetic =eld) delivery sys-
tem has been described previously
43
. For the purposes of this study a barrage of magnetic pulses of 6 ms duration
was applied at a repetition rate of 15 Hz and at ×ux densities between 1�4 mT. Each 6 ms burst consisted of a
series of 20 consecutive asymmetric pulses of 150 µs on and o+ duration with an approximate rise time of 17 T/s.
±e background magnetic ×ux density measured in the chamber was below 1 µT between 0 Hz to 5 kHz. ±e coil
size, position and individual number of windings were numerically optimized by a CST low frequency solver
for low =eld non-uniformity over a wide frequency range taking into consideration the shielding capacity of the
µ-metallic chassis. ±e measured =eld non-uniformity did not exceed 4% within the uniform exposure region of
the coils.
PEMF treatment. To investigate the optimum dosage of PEMF, MSCs in a 3D pellet culture were exposed to
PEMFs of di+erent exposure durations, dosage and the magnetic ×ux amplitude. MSCs were subjected to PEMFs
of 1–4 mT amplitude with exposure times ranging between 5 to 60 min on the day of chondrogenic induction,
applied once or multiple times as indicated in the respective =gure legend. Cell pellets to be treated once with
PEMFs (1x) were exposed on =rst day of chondrogenic induction. Two scenarios of multiple exposures were
administrated (Fig. 2). Firstly, multiple exposures were administrated during the course of a week; double expo-
sures (2x) were applied on days 1 and 2; triple exposures (3x) on days 1, 2 and 4. Alternatively, multiple exposures
were applied on a once a week basis, for up to three week. Non-exposed (control) cells were placed within the
PEMF device without current ×ux to produce a magnetic =eld to ensure that all cells were subject to the same
climatic and mechanical conditions.
Real time PCR analysis. Chondrogenic cell pellets were digested in 0.25% Type II collagenase (Gibco,
Life Technologies) followed by centrifugation. Total RNA was extracted using the RNeasy
®
Mini Kit (Qiagen,
Germany). Reverse transcription was performed with 100 ng total RNA using iScript
™
cDNA synthesis kit
(Bio-Rad, USA). Real-time PCR was conducted using the SYBR
®
green assay on ABI 7500 Real-Time PCR System
(Applied Biosystems, Life Technologies, USA). Real-time PCR program was set at 95 °C for 10 min, followed by
40 cycles of ampli=cations, consisting of a 15 s denaturation at 95 °C and a 1 min extension step at 60 °C. Primer
sequences used in this study were according to previous publication
16
and presented as Supplementary Table 1.
www.nature.com/scientificreports/ 11ScieNtific RePOrtS }� 9421 D �10.103~ s415~-01}-0~2-w
Re level of expression of the target gene, normalized to GAPDH, was then calculated using the 2
−ΔΔCt
formula
with reference to the undi+erentiated MSC. Results were averaged from triplicate samples of two independent
experiments.
ECM and DNA quanti�cation. Samples harvested were digested with 10 mg/mL of pepsin in 0.05 M acetic
acid at 4 °C, followed by digestion with elastase (1 mg/mL). A Blyscan sulfated glycosaminoglycan (sGAG) assay
kit (Biocolor Ltd., Newtownabbey, Ireland) was used to quantify sGAG deposition according to manufactur-
er’s protocol. Absorbance was measured at 656 nm and sGAG concentration was extrapolated from a stand-
ard curve generated using a sGAG standard. Type II Collagen (Col 2) content was measured using a captured
enzyme-linked immunosorbent assay (Chondrex, Redmond, WA). Absorbance at 490 nm was measured and
the concentration of Col 2 was extrapolated from a standard curve generated using a Col 2 standard. Values for
sGAG and Col 2 content obtained were normalized to the total DNA content of respective samples, measured
using Picogreen dsDNA assay (Molecular Probes, OR, USA). Quadruplicates of each group were analyzed from
two independent experiments.
Histological and immunohistochemical evaluation. Samples were =xed in formalin, dehydrated, par-
a n embedded, and cut into sections of 5 µm. For Safranin-O staining, the sections were incubated in hema-
toxylin (Sigma-Aldrich), washed and stained with fast green (Sigma-Aldrich), before staining with Safranin-O
solution (AcrosOrganics). For immunohistochemistry, ultra-vision detection kit (±ermo scienti=c) was used.
Endogenous peroxidase in the sections was =rst blocked with hydrogen peroxide before pepsin treatment for
20 min. Samples were treated with monoclonal antibodies of collagen type II (Clone 6B3; Chemicon Inc.) fol-
lowed by incubation with biotinylated goat anti-mouse (Lab Vision Corporation). A mouse IgG isotype (Zymed
Laboratories Inc.) was used as control for immunohistochemistry studies.
Statistical analysis. All experiments were performed in biological replicates (n = 3 or 4) and results
reported as mean ± standard deviation (SD). Statistical analysis was carried out by Students t-test for comparison
between two groups using the Microso> Excel so>ware. ±e level of signi=cance was set at p < 0.05. All quantita-
tive data reported here were averaged from at least two independent experiments.
References
1. Poole, A. R. et al . Composition and structure of articular cartilage: a template for tissue repair. Clin Orthop elat es S26–33 (2001).
2. Goldring, M. B. & Goldring, S. R. Osteoarthritis. J Cell Physiol 213, 626–34 (2007).
3. Ge, Z. et al. Osteoarthritis and therapy. Arthritis heum 55, 493–500 (2006).
4. Buckwalter, J. A. & Mankin, H. J. Articular cartilage: degeneration and osteoarthritis, repair, regeneration, and transplantation. Instr
Course Lect 47, 487–504 (1998).
5. Vasara, A.I. et al . Indentation sti+ness of repair tissue a>er autologous chondrocyte transplantation. Clin Orthop elat es , 233–42
(2005).
6. Wakitani, S. et al . Mesenchymal cell-based repair of large, full-thickness defects of articular cartilage. J Bone Joint Surg Am 76,
579–92 (1994).
7. Wakitani, S. et al . Human autologous culture expanded bone marrow mesenchymal cell transplantation for repair of cartilage defects
in osteoarthritic knees. Osteoarthritis Cartilage 10, 199–206 (2002).
8. Kuroda, R. et al . Treatment of a full-thickness articular cartilage defect in the femoral condyle of an athlete with autologous bone-
marrow stromal cells. Osteoarthritis Cartilage 15, 226–31 (2007).
9. Pittenger, M. F. et al . Multilineage potential of adult human mesenchymal stem cells. Science 284, 143–7 (1999).
10. Goldring, M. B., Tsuchimochi, K. & Ijiri, K. e control of chondrogenesis. J Cell Biochem 97, 33–44 (2006).
11. Mahmoudifar, N. & Doran, P. M. Chondrogenesis and cartilage tissue engineering: the longer road to technology development.
Trends Biotechnol 30, 166–76 (2012).
12. Woods, A., Wang, G. & Beier, F. Δegulation of chondrocyte di+erentiation by the actin cytos<eleton and adhesive interactions. J Cell
Physiol 213, 1–8 (2007).
13. DeLise, A. M., Fischer, L. & Tuan, R. S. Cellular interactions and signaling in cartilage development. Osteoarthritis Cartilage 8,
309–34 (2000).
14. Toh, W. S., Spector, M., Lee, E. H. & Cao, T. Biomaterial-mediated delivery of microenvironmental cues for repair and regeneration
of articular cartilage. Mol Pharm 8, 994–1001 (2011).
15. Klein, T. J., Malda, J., Sah, R. L. & Hutmacher, D. W. Tissue engineering of articular cartilage with biomimetic zones. Tissue Eng Part
B ev 15, 143–57 (2009).
16. Raghothaman, D. et al. Engineering cell matrix interactions in assembled polyelectrolyte =ber hydrogels for mesenchymal stem cell
chondrogenesis. Biomaterials 35, 2607–16 (2014).
17. Responte, D. J., Lee, J. K., Hu, J. C. & Athanasiou, K. A. Biomechanics-driven chondrogenesis: from embryo to adult. FASEB J 26,
3614–24 (2012).
18. Mouw, J. K., Connelly, J. T., Wilson, C. G., Michael, K. E. & Levenston, M. E. Dynamic compression regulates the expression and
synthesis of chondrocyte-speci=c matrix molecules in bone marrow stromal cells. Stem Cells 25, 655–63 (2007).
19. Grad, S., Eglin, D., Alini, M. & Stoddart, M. J. Physical stimulation of chondrogenic cells in vitro : a review. Clin Orthop elat es 469,
2764–72 (2011).
20. Cheing, G. L., Li, X., Huang, L., Kwan, R. L. & Cheung, K. K. Pulsed electromagnetic elds (PEMF) promote early wound healing
and myo=broblast proliferation in diabetic rats. Bioelectromagnetics 35, 161–9 (2014).
21. Steward, A. J., Kelly, D. J. & Wagner, D. R. e role of calcium signalling in the chondrogenic response of mesenchymal stem cells to
hydrostatic pressure. Eur Cell Mater 28, 358–71 (2014).
22. Pingguan-Murphy, B., El-Azzeh, M., Bader, D. L. & Knight, M. M. Cyclic compression of chondrocytes modulates a purinergic
calcium signalling pathway in a strain rate- and frequency-dependent manner. J Cell Physiol 209, 389–97 (2006).
23. Eijkelkamp, N., Quick, K. & Wood, J. N. Transient receptor potential channels and mechanosensation. Annu ev Neurosci 36,
519–46 (2013).
24. Matta, C. & Zakany, R. Calcium signalling in chondrogenesis: implications for cartilage repair. Front Biosci (Schol Ed) 5, 305–24
(2013).
25. Pall, M. L. Electromagnetic =elds act via activation of voltage-gated calcium channels to produce bene=cial or adverse e+ects. J Cell
Mol Med 17, 958–65 (2013).
www.nature.com/scientificreports/ 12ScieNtific RePOrtS }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
26. Ross, C. L. et al . ±e e+ect of low-frequency electromagnetic =eld on human bone marrow stem/progenitor cell di+erentiation. Stem
Cell es 15, 96–108 (2015).
27. Petecchia, L. et al. Electro-magnetic =eld promotes osteogenic di+erentiation of BM-hMSCs through a selective action on Ca(2+)-
related mechanisms. Sci ep 5, 13856 (2015).
28. Schmidt-Rohlfing, B., Silny, J., Woodruff, S. & Gavenis, K. Effects of pulsed and sinusoid electromagnetic fields on human
chondrocytes cultivated in a collagen matrix. heumatol Int 28, 971–7 (2008).
29. Vincenzi, F. et al. Pulsed electromagnetic =elds increased the anti-in×ammatory e+ect of A(2)A and A(3) adenosine receptors in
human T/C-28a2 chondrocytes and hFOB 1.19 osteoblasts. PLoS One 8, e65561 (2013).
30. De Mattei, M. et al . E+ects of pulsed electromagnetic =elds on human articular chondrocyte proliferation. Connect Tissue es 42,
269–79 (2001).
31. Sakai, A., Suzuki, K., Nakamura, T., Norimura, T. & Tsuchiya, T. E ects of pulsing electromagnetic elds on cultured cartilage cells.
Int Orthop 15, 341–6 (1991).
32. Hilz, F. M. et al . In×uence of extremely low frequency, low energy electromagnetic =elds and combined mechanical stimulation on
chondrocytes in 3-D constructs for cartilage tissue engineering. Bioelectromagnetics 35, 116–28 (2014).
33. Boopalan, P. Δ., Arumugam, S., Livingston, A., Mohanty, M. & Chittaranjan, S. Pulsed electromagnetic =eld therapy results in
healing of full thickness articular cartilage defect. Int Orthop 35, 143–8 (2011).
34. Mueller, M. B. & Tuan, R. S. Functional characterization of hypertrophy in chondrogenesis of human mesenchymal stem cells.
Arthritis heum 58, 1377–88 (2008).
35. Chen, C. H. et al . Electromagnetic fields enhance chondrogenesis of human adipose-derived stem cells in a chondrogenic
microenvironment in vitro. J Appl Physiol (1985) 114, 647–55 (2013).
36. Lim, H. L. et al . Dynamic Electromechanical Hydrogel Matrices for Stem Cell Culture. Adv Funct Mater 21 (2011).
37. Wang, J. et al. Pulsed electromagnetic =eld may accelerate in vitro endochondral ossi=cation. Bioelectromagnetics 36, 35–44 (2015).
38. Mayer-Wagner, S. et al. E+ects of low frequency electromagnetic =elds on the chondrogenic di+erentiation of human mesenchymal
stem cells. Bioelectromagnetics 32, 283–90 (2011).
39. Esposito, M. et al. Di+erentiation of human umbilical cord-derived mesenchymal stem cells, WJ-MSCs, into chondrogenic cells in
the presence of pulsed electromagnetic =elds. In Vivo 27, 495–500 (2013).
40. Ongaro, A. et al. Electromagnetic =elds counteract IL-1beta activity during chondrogenesis of bovine mesenchymal stem cells. J
Tissue Eng egen Med 9, E229–38 (2015).
41. Spaas, J. H. et al . Chondrogenic Priming at Reduced Cell Density Enhances Cartilage Adhesion of Equine Allogeneic MSCs - a
Loading Sensitive Phenomenon in an Organ Culture Study with 180 Explants. Cell Physiol Biochem 37, 651–65 (2015).
42. Berg, H. Problems of wea< electromagnetic =eld e+ects in cell biology. Bioelectrochem Bioenerg 48, 355–60 (1999).
43. Crocetti, S. et al. Low intensity and frequency pulsed electromagnetic =elds selectively impair breast cancer cell viability. PLoS One
8, e72944 (2013).
44. O’Conor, C. J., Case, N. & Guilak, F. Mechanical regulation of chondrogenesis. Stem Cell es Rer 4, 61 (2013).
45. Sun, L. Y. et al . E+ect of pulsed electromagnetic =eld on the proliferation and di+erentiation potential of human bone marrow
mesenchymal stem cells. Bioelectromagnetics 30, 251–60 (2009).
46. Fodor, J. et al. Store-operated calcium entry and calcium in×ux via voltage-operated calcium channels regulate intracellular calcium
oscillations in chondrogenic cells. Cell Calcium 54, 1–16 (2013).
47. San Antonio, J. D. & Tuan, R. S. Chondrogenesis of limb bud mesenchyme in vitro : stimulation by cations. Dev Biol 115, 313–24
(1986).
48. Matta, C. et al. Cytosolic free Ca2+ concentration exhibits a characteristic temporal pattern during in vitro cartilage di+erentiation:
a possible regulatory role of calcineurin in Ca-signalling of chondrogenic cells. Cell Calcium 44, 310–23 (2008).
49. Argentaro, A. et al. A SOX9 defect of calmodulin-dependent nuclear import in campomelic dysplasia/autosomal sex reversal. J Biol
Chem 278, 33839–47 (2003).
50. McCullen, S. D. et al . Application of low-frequency alternating current electric =elds via interdigitated electrodes: e+ects on cellular
viability, cytoplasmic calcium, and osteogenic di+erentiation of human adipose-derived stem cells. Tissue Eng Part C Methods 16,
1377–86 (2010).
51. Xu, J., Wang, W., Clark, C. C. & Brighton, C. T. Signal transduction in electrically stimulated articular chondrocytes involves
translocation of extracellular calcium through voltage-gated channels. Osteoarthritis Cartilage 17, 397–405 (2009).
52. Kawano, S. et al. Characterization of Ca(2+ ) signaling pathways in human mesenchymal stem cells. Cell Calcium 32, 165–74 (2002).
53. Curtis, T. M. & Schol=eld, C. N. Nifedipine bloc<s Ca2+ store re=lling through a pathway not involving L-type Ca2+ channels in
rabbit arteriolar smooth muscle. J Physiol 532, 609–23 (2001).
54. Wen, L. et al. L-type calcium channels play a crucial role in the proliferation and osteogenic differentiation of bone marrow
mesenchymal stem cells. Biochem Biophys es Commun 424, 439–45 (2012).
55. Zhang, J., Li, M., Kang, E. T. & Neoh, K. G. Electrical stimulation of adipose-derived mesenchymal stem cells in conductive sca olds
and the roles of voltage-gated ion channels. Acta Biomater 32, 46–56 (2016).
56. Lin, S. S. et al . Cav3.2 T-type calcium channel is required for the NFAT-dependent Sox9 expression in tracheal cartilage. Proc Natl
Acad Sci USA 111, E1990–8 (2014).
57. Guila<, F., Leddy, H. A. & Liedt<e, W. Transient receptor potential vanilloid 4: ±e sixth sense of the musculos<eletal system? Ann N
Y Acad Sci 1192, 404–9 (2010).
58. Maroto, R. et al. TRPC1 forms the stretch-activated cation channel in vertebrate cells. Nat Cell Biol 7, 179–85 (2005).
59. Eleswarapu, S. V. & Athanasiou, K. A. TRPV4 channel activation improves the tensile properties of self-assembled articular cartilage
constructs. Acta Biomater 9, 5554–61 (2013).
60. Phan, M. N. et al . Functional characterization of TRPV4 as an osmotically sensitive ion channel in porcine articular chondrocytes.
Arthritis heum 60, 3028–37 (2009).
61. Hung, C. T. Transient receptor potential vanilloid 4 channel as an important modulator of chondrocyte mechanotransduction of
osmotic loading. Arthritis heum 62, 2850–1 (2010).
62. Kurth, F. et al. Transient receptor potential vanilloid 2-mediated shear-stress responses in C2C12 myoblasts are regulated by serum
and extracellular matrix. FASEB J 29, 4726–37 (2015).
63. Clapham, D. E., Δunnels, L. W. & Strubing, C. ±e TΔP ion channel family. Nat ev Neurosci 2, 387–96 (2001).
64. Somogyi, C. S. et al . Polymodal Transient Receptor Potential Vanilloid (TRPV) Ion Channels in Chondrogenic Cells. Int J Mol Sci
16, 18412–38 (2015).
65. Muramatsu, S. et al. Functional gene screening system identi=ed TΔPV4 as a regulator of chondrogenic di+erentiation. J Biol Chem
282, 32158–67 (2007).
66. Gavenis, K. et al. Expression of ion channels of the TRP family in articular chondrocytes from osteoarthritic patients: changes
between native and in vitro propagated chondrocytes. Mol Cell Biochem 321, 135–43 (2009).
67. Torossian, F., Bisson, A., Vannier, J. P., Boyer, O. & Lamacz, M. TRPC expression in mesenchymal stem cells. Cell Mol Biol Lett 15,
600–10 (2010).
68. Lievremont, J. P., Bird, G. S. & Putney, J. W. Jr. Mechanism of inhibition of TRPC cation channels by 2-aminoethoxydiphenylborane.
Mol Pharmacol 68, 758–62 (2005).
www.nature.com/scientificreports/ 13ScieNtific REPOrTs }ã 9421 �wv�wvy~ �zw{~�vw}�v~x� ™
69. Yang, Z. et al. Improved mesenchymal stem cells attachment and in vitro cartilage tissue formation on chitosan-modi ed poly(L-
lactide-co-epsilon-caprolactone) sca old. Tissue Eng Part A 18, 242–51 (2012).
70. Winegar, B. D., Haws, C. M. & Lansman, J. B. Subconductance block of single mechanosensitive ion channels in skeletal muscle
bers by aminoglycoside antibiotics. J Gen Physiol 107, 433–43 (1996).
Acknowledgements
Re study was supported by University of Malaya HIR-MoE Grant (Reference number � UM.C/625/1/HIR/
MOHE/MED/32 account number – H20001-E000071) and Singapore-MIT Alliance for Research and Technology
(SMART) Foundation (ING14085-BIO). Dinesh Parate was supported by NUS Research scholarship.
Author Contributions
D.P. performed experiments, analyzed data and drafted the manuscript. A.F.O., J.F. and C.B. provided
technological expertise and contributed to the fabrication the PEMF facility. A.A.A., T.K., J.H.P.H. and A.F.O.
provided funding and critical reading of the manuscript. A.F.O. and Z.Y. designed the study, analyzed data and
provided critical revision of the manuscript. All authors reviewed the manuscript.
Additional Information
Supplementary information accompanies this paper at doi:10.1038/s41598-017-09892-w
Competing Interests: Re authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and
institutional a liations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International
License, which permits use, sharing, adaptation, distribution and reproduction in any medium or
format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Cre- ative Commons license, and indicate if changes were made. Re images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the
material. If material is not included in the article’s Creative Commons license and your intended use is not per-
mitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the
copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.
� Re Author(s) 2017