Modulation of human mesenchymal stem cells by electrical stimulation using an enzymatic biofuel cell

Abstract


Results
The electrical stimulation of EBFC changes cell morphology through cytoskeleton re-arrangement.In particular, the results of whole transcriptome NGS showed that speci c gene clusters were up-or downregulated depending on the magnitude of applied electrical current of EBFC.

Conclusion
Taken together, the ndings in this study demonstrate that EBFC-generated electrical stimulation can in uence the morphological and gene expression properties of stem cells and such capabilities can be useful for regenerative medicine applications related to wearable sensors and devices.

Background
There is broad interest in developing enzymatic biofuel cells as components within wearable sensors and bioelectronic devices to regulate cell behavior [1][2][3][4][5].Electrical stimulation plays an important role in regulating the function of mammalian cells and tissues, including neurons, muscle, and cardiomyocytes.
Additionally, electrical stimulation can regulate cellular respiration in mitochondria and aid cellular signaling [6][7][8][9][10].A speci c range of electrical stimulation conditions also plays a biophysical role in wound healing and homeostasis by regulating stem cell behavior [11,12].Several studies have investigated how electrical stimulation regulates stem cell behavior like migration, proliferation, and differentiation in vitro [8, [13][14][15][16][17]. Previous reports demonstrate that cellular-level electrical stimulation affects axon outgrowth and neuronal function [18][19][20][21].Although electrical stimulation regulates cells, it is not easily or directly applicable as a therapeutic tool since most electrical stimulation systems are bulky electrical generators which are neither implantable nor biocompatible.In the medical eld, direct current electrical stimulation has been used to treat Parkinson's disease [22].Implanted electrical devices, such as deep brain stimulators [9,12], often cause surgical complications and in ammation caused by inorganic materials.Although direct implantation of inorganic fuel cells can be substituted by indirect therapy using electromagnetic systems, the systems are expensive and provide only localized stimulation [23].As such, there is broad interest in alternative device designs, especially ones that are compatible with a wearable format.
In contrast to the above shortcomings, enzymatic biofuel cells (EBFCs) have several advantages.Basic EBFC structures consist of two different enzyme components for a hydrogen/oxygen fuel cell, an anode, a cathode, and a separator [24].EBFCs are composed of biocompatible organic materials, with a simple system design, nano-scale electrical current control, user-de ned fuel cell formulation, and economical operation compared with cellular electrical stimulation using other power sources [25][26][27][28].Until now, EBFCs have been introduced and developed for over 45 years as bioelectrical tools for electrical machines, biosensors, and bioelectronics.Although real-life viable EBFC application has been demonstrated in living plants, snails [29,30], and small animals [25], EBFCs still have limitations and intrinsic issues for commercial clinical application in humans, such as limited reaction time, low power density, and insu cient voltage for device operation [26,31,32].However, EBFCs have enough capacity as direct electrical stimulation tools on the cellular level in a physiological medium [33].One promising area involves stem cell-based regenerative medicine, however, there are no studies using EBFCs to control mesenchymal stem cell differentiation with speci c electrical stimulation.
Mesenchymal stem cells (MSCs) hold great potential for damaged tissue replacement and regeneration [34,35].The mesenchymal stem cell niche, which consists of micro-environmental cues that control stem cell fate, is highly dependent on the physical condition of speci c tissues [36,37].Physiological stimulation, such as electric current, could activate stem cells to differentiate, thereby inducing regeneration and recovery [15,18,21].Electrophysiology has been used to study the therapeutic e cacy of electrical stimulation for mesenchymal stem cell regulation [15,18].However, research in this direction requires further investigation, especially with EBFCs which remain to be tested in this application scope.Therefore, we investigated how EBFC-generated nano-currents could control MSC behavior via electrical stimulation.Moreover, EBFC systems show excellent potential to serve as the basis for self-powered implantable devices, which can aid injury recovery and/or stimulate MSC activity for self-regeneration without requiring stem cell transplantation.
Herein, we investigated the effect of EBFC-generated electrical stimulation on hAD-MSC cell morphology and gene expression using next-generation RNA sequencing, which is a powerful tool for screening and identifying molecular mechanisms [38].Speci cally, mRNA-seq (whole-transcriptome shotgun sequencing) can pro le total gene expression and identify signaling network pathways activated by speci c environmental conditions such as the applied electrical current [39].Based on this integrated approach, we demonstrate that speci c electrical currents can elicit mesenchymal stem cell differentiation and can guide the future development of EBFC-based sensors and devices as cell regulators.
We measured oxidation and reduction ratios of GOx and BOD enzymes in cell culture media.We utilized a speci c colorimetric dye, WST-8 (2-(2-methoxy-4-nitrophenyl)-3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2Htetrazolium, monosodium salt), as the reporter.It is a formazan chemical dye that, upon bioreduction of the electron carrier, changes color depending on the electron concentration in the culture media.The colorimetric reaction of GOx and BOD was analyzed with a microplate reader (Epoch microplate spectrophotometer; BioTek Instruments, Inc. US) based on the absorbance intensity at 450 nm wavelength.
EBFC set-up for 2D cell culture The hAD-MSCs (1 x 10 5 cells/mL) were seeded and cultured in a 35 mm dish set up under three different enzyme concentrations as 0.01 mg/mL, 0.05 mg/mL, and 0.1 mg/mL and without enzyme as the negative control (printed carbon only without enzyme).Cell seeding on the culture dish was completed until adhesive on the dish was replaced by alpha MEM high glucose (4500 mg/ml) media supplemented with 10% FBS for electrical stimulation of EBFC.The enzyme-loaded cathode and anode electrodes were inserted into the EBFC system.We made the lm ring and cascade of both BOD and GOX enzymes.
Starting at the initial time point, the cells were cultured at predetermined time points for required assays.

Measurements of electrical currents
A CHI 660B potentiostat/galvanostat (Austin, TX, USA) was used to obtain linear sweep voltammetry (LSV) and amperometry (I-t curve) data.The GOx based working electrode was carried out for LSV with/without 25mM glucose in 1X PBS solution.The I-t curves were measured using different concentrations of GOx modi ed electrode by immersing the EBFC in cell culture media.A BOD-modi ed screen-printed carbon electrode (SPCE; 3.0 mm diameter) was used as the working electrode.GOxmodi ed SPCEs were used as the counter and reference electrode.The experiments were carried out under ambient air conditions at room temperature.
Cell survival, proliferation and cell morphology under electrical stimulation Cell proliferation was determined under EBFC electrical stimulation on days 3 and 6.On each experimental day, cells were harvested by trypsinization with a 0.25% trypsin/0.53mM EDTA solution.Cell number was determined using a hemocytometer, and dead cells were identi ed by trypan blue staining.Each sample was analyzed in triplicate for statistical analysis.
For cell morphology analysis, we performed immunocytochemistry with anti-rabbit focal adhesion kinase antibody in electrically stimulated hAD-MSCs.hAD-MSCs were xed with 4% PFA for 5 min at 24 o C. The samples were then treated with 0.1% Triton-X for antigen retrieval and were washed with PBS.A polyclonal rabbit anti-FAK antibody (A-17, Santa Cruz Biotechnology, US; 1:100 diluted with 0.1% BSA in PBS) was applied to samples, and incubated overnight at 4 °C.Thereafter, the samples were incubated with the anti-rabbit secondary FITC-conjugated antibody and mixed with phalloidin (A34055, Alexa Fluor 555 phalloidin, Invitrogen) for 2 h at room temperature.The nuclei were stained with 1 µg/mL Hoechst® 33342 nuclear stain (H1399; Thermo Fisher Life Technologies, CA, USA) for 5 min at room temperature.After washing thrice with PBS, the stained cells were mounting by anti-fade media (H-1000, Vectashield mouting media Vector Labs, US).Images of the stained cells were acquired using a confocal microscope (LSM700, Zeiss AG, Oberkochen, Germany), and were captured from all angles, and the z-stacks of 25-30 images were used for image analysis using ZEISS software (ZEN2008, ZEISS, Oberkochen, Germany).

mRNA-seq analysis
After electrical stimulation, hAD-MSCs were harvested by trypsinization, washed with PBS, and stored at -80 ºC until mRNA puri cation.All samples were processed with an mRNA puri cation kit (RNeasy Mini Kit; 74104, Qiagen US).Isolated mRNA was transferred to TheragenEtex (South Korea) for NGS analysis.mRNA quality for each sample was measured, and only samples with optimal mRNA quality were used for subsequent NGS.Each sample sheet was prepared on a HiSeq 2500 System and 150 bp paired-end reads (Illumina, US).We analyzed differentially-expressed genes depending on each condition within speci c gene clusters to identify hAD-MSC gene expression pro les and interaction networks due to EBFC nano-scale electrical stimulation.

Gene ontology analysis of transcriptome data
A gene ontology (GO) database was used to infer signi cantly-enriched terms on each gene set using DAVID bioinformatics resources.The particular genes, which are shown by expression differences in each electrical stimulation condition, were evaluated to classify the function of genes using the DAVID functional annotation tool v6.8 (https://david.ncifcrf.gov/)with the UniProtKB dataset.Cumulative hypergeometric p-values and enrichment factors were calculated and used for ltering.The remaining signi cant terms were hierarchically clustered based on Kappa-statistical similarities among gene groups.A 0.3 kappa score was used as the threshold to categorize the tree into term clusters.A webbased program called REVIGO was used to determine the biological function of differentially expressed genes (DEGs).We employed a multidimensional scaling procedure to initially place the terms using an eigenvalue-decomposed pairwise distance matrix, followed by a stress minimization step.

Reverse transcription-qPCR analysis
We veri ed early phase differentiatial gene expression in hAD-MSCs by using RT-qPCR.Gene expression was analyzed using glial brillary acidic protein (GFAP), neuro lament (NF), osteopontin (OPN), alkaline phosphate (ALP), myogenin, and MyoD.Electrically-stimulated mesenchymal stem cells were cultured for 6 days, harvested, and centrifuged.The resulting cell pellets were homogenized with a guanidine isothiocyanate-based cell lysis buffer (Qiagen Ltd.).Total RNA was extracted from the lysate by using a RNeasy Mini Kit (Qiagen Ltd.).Total RNA was converted to cDNA using a Superscript kit (Invitrogen) with random hexamers.qPCR was performed using Sensimix Plus SYBR master mix (Quantace) in a spectro uorometric thermal cycler (Rotor-Gene 3000; Corbett Research).RT-qPCR data were analyzed by the comparative threshold cycle (CT) method, with Glyceraldehyde-3-Phosphate Dehydrogenase (GAPDH) as the reference gene.Triplicate samples were tested.RT-qPCR primer sequences are presented in Table 2.

Statistical Analysis
Statistical analysis of whole transcriptome NGS data was performed by modi ed Fisher's exact test method.All other data are expressed as mean ± standard error of the mean (SEM) in triplicate.Statistical analyses were carried out using one-way ANOVA with the signi cance level set at *p < 0.05.Signi cant differences are indicated by asterisks (p < 0.05) in each gure.

Characterization of EBFCs used for cellular studies
We investigated human mesenchymal stem cell behavior with electrical stimulation delivered by EBFCs.We fabricated EBFCs of different designs using different enzyme concentrations (0.01, 0.05, 0.1 and 1 mg/mL), to investigate the behavior of hAD-MSCs after nano-scale electrical stimulation.The basic design is presented in Figure 1A.We measured electron rich or poor conditions with WST-8 dye as the colorimetric chemical dye.In Figure 1B, GOx exhibited a signi cantly increased electron oxidation (rich) environment, BOD showed a signi cantly decreased electron reduction (poor) environment depending on the enzyme concentration in culture media.These data support that the EBFCs are operational.In addition, we measured the electrochemical glucose response on the anode electrode, as shown in Figure 2A.There are GOx enzymes present, which can oxidize glucose (with/without 25mM).It means that the GOx enzyme is working well to operate as an EBFC.And BOD enzyme electrode was investigated by LSV in Figure S1A.The power density and polarization curves with 0mM (Figure S1B) and 25 mM (Figure S2C) glucose were obtained by LSVs of Figure 2A and Figure S1A.Also, the I-t curve was measured to determine how much current exists in different conditions as a function of the enzyme concentration (Figure 2B).The different EBFC designs are described in Table 1.A BOD-modi ed cathode electrode was used as the working electrode, while a GOx-modi ed anode electrode was used as a single reference and counter electrode, as shown in the inset of Figure 2B.

EBFC electrical current for determining cell viability
The EBFC system components, including enzyme solution and conductors, exhibit strong biocompatibility without damaging hAD-MSCs.We investigated cellular behavior under electrical stimulation by immunocytochemistry to elucidate the effect of high electrical current on hAD-MSCs.In the case of high electrical current, F-actin structure disassembly occurred in the 1870± 305nA/cm 2 current levels within 6 hrs (Figure 3A).However, hAD-MSCs survived and grew in the presence of electrical currents between 127 ± 9 nA/cm 2 and 598 ± 75 nA/cm 2 .The number of cells stabilized from 2 hrs to 6 hrs within the 127 and 598nA current range, but high current caused the cell number to immediately decrease compared to the low current group (Figure 3B).Thus, the EBFC system could modulate cell survival characteristics and did not cause additional cell proliferation (Figure 3C).

Morphological analysis of electrically-stimulated hAD-MSCs
Figures 4A show the change in cell morphology based on the F-actin arrangement by immunocytochemistry. Figure 4B presents numerical analysis of F-actin arrangement and shows a marked change in its arrangement within the hAD-MSCs depending on the electrical current.127 nA/cm 2 electrical current caused the disappearance of focal adhesion protein at the edge point in the cells and increased transverse and ventral stress ber type actin arrangements in the cells (about 85%).And 248 nA/cm 2 electrical stimulation caused dramatically different types of arrangements involving web-like structures such as cross linked dorsal-ventral stress bers with a perinuclear actin cap in the cells (90%) compared with controls.Furthermore, the 598 nA/cm 2 current caused strongly stressed ber type actin assembly formation in the cells (80%).

Transcriptomic analysis of electrically-stimulated hAD-MSCs
We next investigated mRNA expression pro les of MSCs that were subjected to electrical stimulation over time.We rst con rmed that the sequencing quality of the puri ed mRNA was acceptable with several quality measures (e.g., total number of reads and genome coverage; Supplemental Figure S2 and Table S1).Cluster analysis of the DEGs in hAD-MSCs affected by electrical stimulation was performed.Therefore, we screened signi cantly up-and down-regulated gene clusters ( ltering p < 0.05) (Figure S3A: 3 days, S3B: 6 days).
Figure S4 shows the gene expression pro les for electrically-stimulated hAD-MSCs.To determine transcriptomic differences between the control and EBFC groups, we analyzed RNA-seq data by Pearson correlation coe cient and hierarchical clustering.The results showed a dose-dependent increase in the distance of gene expression clusters from 127 nA/cm 2 , 248 nA/cm 2 , and 598 nA/cm 2 electrical stimulation (Supplementary Figure S3), which means the gene expression pro ling signi cantly changed depending on the electrical current.A total of 166 up-regulated genes showing gene expression signi cance compared to the control in each electrical condition were used for hierarchical clustering analysis, based on log 2 Fragment per Kilobase of transcript per Million mapped reads (FPKM) values.We found that the up-regulated genes were divided into three clusters according to the highest expression level at each current (127 nA/cm 2 : red present 41, 248 nA/cm 2 : blue present 68, 598 nA/cm 2 : green present 57 genes).In the hierarchical clustering graph, the EBFC-generated electrical current induced differential gene expression in hAD-MSCs (Figure 5A).When comparing the expression level of the upregulated genes, the three clusters showed a signi cantly higher expression by the electrical stimulation of EBFC (Figure 5B).To classify the functions of the 166 up-regulated genes in the EBFC group, we analyzed the genes using the UniProtKB database in the DAVID functional annotation tool v6.8 [40].
Among the 166 up-regulated genes, 98 genes were signi cantly enriched in UniProtKB functional categories such as ribosomal protein, alternative splicing, ribonucleoprotein, and amino-acid biosynthesis (Figures 5B, C).Metabolic regulation likely changed rapidly by promoting the tricarboxylic acid (TCA) cycle and subsequent oxidative phosphorylation in mitochondria.Up-regulation of genes related to amino-acid and nucleotide synthesis may have increased as a result of changes in TCA cycle genes [41].
GO analysis revealed that each gene set was involved in unique biological pathways according to the biological process, cellular component, and molecular function databases, depending on the applied electrical current (Table S1).Therefore, Table S1 shows the genes differentially expressed by different electrical currents.These genes identify pathways involved in differentiation.At 3 days, electricallystimulated hAD-MSCs showed a more neuron-, muscle-, and bone-like-related gene pro le (Table S1A).
After 6 days, electrically-stimulated hAD-MSCs showed up-regulation of gene clusters more associated with differentiation (Table S1B).

Transcriptome visualization of electrically stimulated hAD-MSCs by Reduce Visualize Gene Ontology (REVIGO) program
The REVIGO program was used to visualize the statistically-enriched GO terms in DEGs (>2-fold), using a clustering algorithm that relies on semantic similarity measures [42].We imaged the major GO data using scatter plots and interactive graphs for each functional gene cluster under each culture condition (Figure S3).In Figure S4A, 127 nA/cm 2 of electrical stimulation induced up-regulation genes regarding "alcohol metabolism", "detection of stimulus", "NIK/NF-κB signaling", and "epithelial cell proliferation" on day 3. Next, "metal ion transport", "inositol phosphate biosynthesis", "drug membrane transport" and "NIK/NF-κB signaling" gene was notably up-regulated with 248 nA/cm 2 electrical current caused by the EBFC (Figure S4B).On the other hand, 598 nA/cm 2 of electrical current induced the up-regulated gene groups classi ed as "multicellular organismal process", "cell surface receptor signaling pathway", and "detection of stimulus" on day 3 (Figure S4C).Figures S4D, E showed the up-and down-regulation of speci c gene groups on day 6.Speci cally, 127 nA/cm 2 and 248 nA/cm 2 electrical stimulation signi cantly upregulated "G-protein coupled receptor signaling pathway", "response to stimulus", "cell fate commitment", "sensory perception", and "glial cell differentiation" over 6 days(Figure S4E).In the case of 598nA/cm 2 , the electrical stimulation led to up-regulation of "response to stimulus" and "regulation of localization and transport" gene groups on day 6.However, it down-regulated to "negative regulation of skeletal muscle cell differentiation" on day 6 (Figure S4 F).

Early-phase differentiation-related gene expression of electrically-stimulated hAD-MSCs
We next performed real-time qPCR analysis of early-phase differentiation-related genes in hAD-MSCs under electrical stimulation at day 6 (Figure 6A, B and C).We analyzed neurogenesis-, osteogenesis-, and myogenesis-gene related expression after nano-electrical stimulation.Each electrical current increased early differentiation markers in a current-speci c manner after 6 days.The current densities of the three EBFC conditions are summarized in Figure S4.The 127 nA/cm 2 and 248 nA/cm 2 levels of EBFC electrical current ranges induced early neurogenesis genes such as GFAP and NF, after 6 days culture.ALP, which is an early osteogenesis marker, increased after culture with 248 nA/cm 2 electrical current of EBFC.Finally, both 127 nA/cm 2 and 598 nA/cm 2 of electrical current enhanced MyoD gene-like early myogenesis markers after 6 days.

Discussion
In this study, we found that electrical stimulation of EBFCs can control the property of MSCs.Based on the immunocytochemistry images and transcriptomic data, electrically-stimulated hAD-MSCs clearly exhibited speci c DEG clusters after electrical stimulation for days 3 and 6 (Figures S4A to F, Supplemental Tables S1 and S2).
Interestingly, immunocytochemistry of electrically stimulated hAD-MSC exhibited dramatically altered actin assembly formation depending on the electrical current.F-actin as an actin-cytoskeleton block is an important signaling transfer and responder cue to cell from the extracellular space [43].FAK is one of the regulatory factors at adhesion zones for in uencing cell behaviors like proliferation, migration, and differentiation [44].Generally, FAK can improve cell migration, proliferation, and the prevention of cell apoptosis through the integrin signaling pathway [45].Electrically stimulated hAD-MSCs show that FAK proteins move from focal adhesion point and the edges of cell focal adhesion points.Intracellular localization of FAK promoted cell proliferation and migration [46].Also, FAK and F-actin assembly linked for cell signaling from outside signal for regulation of cell behaviors [47,48].In particular, the FAK linked with actin cytoskeleton for response from physical cure like mechanotransduction of mechanical sensing [49][50][51][52][53][54].Indeed, environmental mechanosensing is a crucial element for stress ber organization and the fate determination of stem cells [55,56].Therefore, optimal electrical stimulation in uences the FAK linked actin-cytoskeleton to stem cell behavior.Additionally, our results indicate that gene expression pro ling and hAD-MSC behavior can be regulated by nano-scale electrical currents (Figures S2A and B).Based on bioinformatics analysis, we found several hundred different genes that underwent up-or down-regulation in response nano-levels of electrical stimulation.Based on the GO data analysis, each of the gene pro les was associated with speci c up-, stay-, and down-regulated genes by speci c electrical currents, such as NF for neurogenesis, myo bril assembly for myogenesis, and acid phosphatase activity for osteogenesis [57][58][59].The functional annotation enrichment data presented an effect of EBFC electrical stimulation up-regulated protein translation activity, including ribosomal proteins, which is essential for cell differentiation [60].Therefore, the data presents the potential of EBFC for regenerative medicine applications.
Here, we attempted to understand the mechanism underlying EBFC electrical stimulation and which gene clusters are up-and down-regulated during electrically-induced hAD-MSC differentiation.Several reports indicate that protein-based ligand binding forces are determined by electrical status and electrical constants [61,62].Therefore, we suggest that EBFC electrical stimulation could supply a speci c electrical status or constant condition depending on the EBFC-generated electrical current.Ligand binding interactions with receptors are primarily driven by an electrical constant.Each ligand-like growth factor has a different electrical constant controlling receptor binding [61].Therefore, an electrical constant mimicked by EBFCs could induce speci c signaling, gene expression, or regulation similar to the speci c ligand as a biophysical cue for regulation of MSC behavior.In the real time-qPCR data, speci c electrical current range of EBFC occurred for expression of speci c early differentiation genes in the hAD-MSC like GFAP, ALP and MyoD [57,63,64].Therefore, nano-ampere range electrical stimulation may have an application tool for directional differentiation of mesenchymal stem cells that could be useful for regenerative medicine applications.Indeed, EBFCs have excellent reported biocompatibility for in vitro and in vivo implantation, and could also be useful for wearable sensors and devices.Indeed, our ndings show that EBFCs have excellent potential to create self-powered devices for wearable sensor and medical implant applications related to bioelectrical stimulation of cells [1,26,65].This potential is further supported by past animal studies showing that implanted EBFCs can be self-powered intracorporeally [65,66].
Taken together, we identi ed gene sets related to various biological pathways associated with electricallystimulated hAD-MSCs based on an EBFC system.Nano-scale electrical stimulation induced speci c gene expression pro les under constant electrical currents.The speci c range of electrical currents induced differentiation depending on the culture time.In general, regeneration processes require cell growth for tissue repair and differentiation for speci c cell or tissue recovery processes.Electrical stimulation can be used for directional hAD-MSC differentiation into neurons, muscle cells, and bone/multi-organ processing between 248 and 598 nA/cm 2 .Our results could have potential applications using MSC differentiation for wound repair via direct EBFC electrical stimulation.Further, our data show the feasibility of MSC-based biomedical engineering using direct EBFC-generated electrical stimulation.

Conclusions
This study demonstrates that EBFCs can induce speci c MSC differentiation via electrically-stimulated neurogenesis, osteogenesis, and myogenesis.Moreover, our ndings show how electrochemical characterization of EBFC-based sensors can be integrated together with biological methods and nextgeneration RNA sequencing to unravel the effects of EBFCs on hAD-MSC cell morphology and gene expression levels.Thus, EBFCs may have increased potential for therapeutic applications in regenerative medicine and can be further considered as components within wearable sensors and bioelectronic devices.

(
A) Schematic illustrating the EBFC mechanism using electrodes with Os2+/3+ and PEVGEC.(B) Oxidation/reduction ratios of GOX (Anode) and BOD (Cathode) enzymes in the culture media.Open bar is GOX enzyme only as oxidation status, BOD enzyme only as reduction status and Full set is GOX and BOD enzyme in the dish as electrical current status.

Figure 5 Gene
Figure 5