2D and 3D Self-Assembling Nanofiber Hydrogels for Cardiomyocyte Culture

Collagen is a widely used biomaterial in cardiac tissue engineering studies. However, as a natural material, it suffers from variability between batches that can complicate the standardization of culture conditions. In contrast, synthetic materials are modifiable, have well-defined structures and more homogeneous batches can be produced. In this study, several collagen-like synthetic self-assembling nanofiber hydrogels were examined for their suitability for cardiomyocyte culture in 2D and 3D. Six different nanofiber coatings were used in the 2D format with neonatal rat cardiomyocytes (NRCs) and human embryonic stem-cell-derived cardiomyocytes (hESC-CMs). The viability, growth, and functionality of the 2D-cultured cardiomyocytes were evaluated. The best-performing nanofiber coatings were selected for 3D experiments. Hydrophilic pH-sensitive nanofiber hydrogel coassembled with hyaluronic acid performed best with both NRCs and hESC-CMs. Hydrophilic non-pH-sensitive nanofiber hydrogels supported the growth of NRCs; however, their ability to promote attachment and growth of hESC-CMs was limited. NRCs also grew on hydrophobic nanofiber hydrogels; however, the cell-supporting capacity of these hydrogels was inferior to that of the hydrophilic hydrogel materials. This is the first study demonstrating that hydrophilic self-assembling nanofiber hydrogels support the culture of both NRCs and hESC-CMs, which suggests that these biomaterials hold promise for cardiac tissue engineering.


Introduction
Heart failure arising from myocardial loss is a leading cause of morbidity and mortality worldwide [1] for which stem cell therapy is an emerging treatment option. For treatment of heart failure, cells can potentially be delivered to the site of injury where they can repopulate the injured area, integrate into the host tissue, and restore functionality to the myocardium. However, clinical studies have shown that cells delivered by direct injection into the myocardium or by intracoronary injection are rapidly lost [2][3][4]. Modest improvements in myocardial function have been suggested to merely arise from paracrine effects [5][6][7]. To achieve the desired cellular effects in addition to the paracrine effects, biomaterial scaffolds can be used to maintain the cells at the site of injection.
An ideal biomaterial scaffold for cardiac repair would allow cardiomyocytes to be grown in vitro in a 3D structure that is optimal for application to the heart, allows cellular differentiation, and integrates well into the host tissue aer implantation. In addition to clinical cell therapy applications, 3D structures can also provide better cardiac tissue models for studying the pathophysiology of cardiac diseases and the function of diseased cardiomyocytes in vitro. Additionally, these 3D models may provide more meaningful physiological data in drug discovery and toxicology assays.
Beating cardiac constructs have been obtained using collagen patches [8], poly(N-isopropylacrylamide) sheets [9], collagen rings [10], collagen sponges [11], and injectable �brin gels [12]. Scaffold-free human cardiac tissue has also been generated [13]. In our previous studies, we compared different synthetic and natural biomaterials for their ability to support cardiomyocyte growth [14]. Of the biomaterials we tested, natural collagen supported cardiomyocyte growth the best but showed signi�cant batch-to-batch variation. us, we continued to search for synthetic alternatives to collagen that had similar bioactivity but less variability.
Previously, we tested the commercially available amphiphilic self-assembling peptide PuraMatrix (BD Biosciences). Although PuraMatrix was shown to be less effective than collagen at supporting the growth and survival of cardiomyocytes, it performed well enough to attract our interest in this class of materials [14]. We previously reported the development of self-assembling nano�ber systems with a modular architecture based on a 1,3,5-triamide cis,cis-cyclohexane core [15]. is core functions as a generic nano�ber-forming scaffold that can be easily functionalized and tuned [15][16][17]. We demonstrated that vesicles can be easily entrapped and immobilized in hydrogels generated from these materials [18], thereby providing a system to study the activity of membrane proteins at the single vesicle level [19]. Preliminary in vitro and in vivo experiments indicated that hydrogels formed from these compounds are biocompatible [15,20], which encouraged us to study their use as coatings and hydrogel scaffold materials for the growth of cardiomyocytes.
e aim of this study was to screen different self-assembling nano�ber hydrogels as 2D nano�ber coatings for cardiomyocyte attachment and growth and to then further evaluate the best candidates for use as 3D nano�ber hydrogels. To assess the performance of our nano�ber materials we used NRCs and hESC-CMs, and we considered that an optimal biomaterial should support routine cardiomyocyte attachment, growth, and function. NRCs were used because they can be obtained fairly easily in large numbers, which enabled us to perform large-scale comparison experiments [21] before testing hESC-CMs. It has also been shown that 3D heart tissue-like structures can be created with NRCs [10,22] and with hESC-CMs [22][23][24].
ree different types of self-assembling nano�ber hydrogels were examined: hydrophilic pH-sensitive, hydrophilic non-pH-sensitive, and very hydrophobic non-pH-sensitive, hydrogels. We also tested the effect of �ber thickness on cell adhesion for the hydrophilic nano�ber hydrogels. In addition, we tested the effects of pH-sensitive nano�ber hydrogels coassembled with hyaluronic acid (HyA) [25] on cardiomyocyte growth.
Differentiation was induced by coculturing the stem cells with END-2 cells [28] in 0% KO-SR hES medium (Knockout DMEM, 2 mM GlutaMax, 1% NEAA, 50 U/mL P/S, 0.1 mM 2-mercaptoethanol) which was changed on days 5, 8, and 12. On day 15, the medium was replaced with 10% KO-SR hES medium (Knockout DMEM, 10% SR, 2 mM GlutaMax, 1% NEAA, 50 U/mL P/S, 0.1 mM 2-mercaptoethanol) that was replenished every third day thereaer. e �rst beating areas normally appeared aer 14 days of differentiation. When the cells had been differentiated for 20 to 55 days, they were dissociated into single cells. Approximately the same amount (approximately 4,000) of cells was plated on each coating, on top of or inside the gels, and on control wells (0.1% gelatin (Sigma-Aldrich) coated commercial 24-or 48-well plates). e hESC-CMs were usually cultured for one week.

Cell Characterization.
Cells were plated at equal numbers onto nano�ber hydrogels and control wells. e cells were observed daily using a phase-contrast microscope (Nikon Eclipse TS100, Nikon, Japan) and several qualitative parameters were scored to determine the suitability of nano�ber hydrogels for supporting cardiomyocyte culture. ese parameters included cell attachment, spreading, morphology, viability, detachment, and beating. Staining for cardiomyocyte markers was used to evaluate the alignment and spreading of cardiomyocytes. Cell attachment was also evaluated quantitatively by counting troponin T positive cells. With NRCs, cell attachment was evaluated by measuring the con�uency of troponin-positive cells/well, whereas with hESC-CMs all troponin-positive cells from every replicate were calculated.

2.2.1.
Viability. e LIVE/DEAD Viability/Cytotoxicity Kit for mammalian cells (Molecular Probes, Inc., Invitrogen), which contains calcein AM to stain live cells green and ethidium homodimer-1 to stain dead cells red, was used to assess viability. e stained cells were observed using phase contrast and �uorescence microscopy (Olympus I�51, Olympus, Japan) and photographed using an Olympus DP30BW camera (Olympus, Japan).

Cryogenic Transmission Electron Microscopy.
Several microliters of the nano�ber suspensions were deposited on bare 700 lines/inch mesh copper grids. Aer excess liquid was blotted away, the grids were plunged quickly into liquid ethane. Frozen-hydrated specimens were mounted on a cryoholder (Gatan, model 626) and observed using a Philips CM 120 electron microscope operated at 120 kV. Micrographs were recorded under low-dose conditions using a slow-scan CCD camera (Gatan, model 794).

�reparation o� 2� �ano�ber Coatings o� p��Sensiti�e �elators ��ano�ber Coatings 1� 2�.
A solution of 10 mg of the HCl salt (Boom B.V., e Netherlands) of the gelator in 3 mL of mQ water was prepared by gentle heating. e solution was neutralized by addition of 1 mL of 100 mM HEPES, pH 8 (Sigma-Aldrich). Aliquots of the 200 L neutralized solution were transferred to wells of a 24-well plate (SPL Life Sciences, Inc., the Republic of Korea). e solvent was evaporated overnight under ambient conditions to yield transparent to translucent coatings. e �brous nature of the coatings was con�rmed by optical microscopy (Motic AE31, China). Before use, the plates were sterilized for 5 minutes with UV light irradiation in a laminar �ow cabinet. Nano�ber coating 1 was tested using both NRCs and hESC-CMs and nano�ber coating 2 was tested using hESC-CMs.

�reparation o� �y��Containing 2� �ano�ber
Coatings o� p��Sensiti�e �elators ��ano�ber Coatings 1 + HyA, 2 + HyA). A solution of 10 mg of the HCl salt of the gelator in 2.6 mL mQ water was prepared by gentle heating. To this solution we added 0.4 mL of a 0.5% (w/v) solution of hyaluronic acid (Sigma-Aldrich) in mQ water. e resulting solution was neutralized by the addition of 1 mL of 100 mM HEPES, pH 8. Aliquots of the 200 L neutralized solution were transferred to the wells of a 24-well plate. e solvent was evaporated overnight under ambient conditions to yield T 1: �roperties of the nano�ber hydrogels and the growth of cardiomyocytes on them. Cell attachment, spreading, morphology, �iability, detachment, and beating were e�aluated and the comparison was always done compared to control (NRCs: uncoated well plate, hESC-CMs: 0.1% gelatin-coated well plate). transparent to translucent coatings. e �brous nature of the coatings was con�rmed by optical microscopy. Before use, the plates were sterilized for 5 minutes with UV light irradiation in a laminar �ow cabinet. Nano�ber coatings 1 + HyA and 2 + HyA were tested using both NRCs and hESC-CMs.

Preparation of 2D
Coatings of Non-pH-Sensitive Gelators �Nano��er Coatings and 3, 4, 5, 6). A solution of 10 mg of the gelator in 16 mL of a 95 : 5 ethanol/water mixture was prepared by gentle heating. Aliquots of 750 L of the stock solution were transferred to the wells of a 24-well plate. e solvent was evaporated overnight under ambient conditions to yield transparent to translucent coatings. e �brous nature of the coatings was con�rmed by optical microscopy. Before use, plates were sterilized for 5 minutes with UV light irradiation in a laminar �ow cabinet. Nano�ber coatings 3, 5, and 6 were tested using both NRCs and hESC-CMs and nano�ber coating 4 was tested using hESC-CMs.

2.3.�. Preparation of 3D Nano��er Hydrogels.
Based on the results from the 2D experiments, the nano�ber coatings 1 + HyA and 4 were chosen for the 3D nano�ber hydrogel experiments. Nano�ber hydrogel 1 was included in the hESC-CM experiments to control for the effects of HyA addition. e 3D nano�ber hydrogels 1 + HyA and 4 were �rst studied using NRCs and then using hESC-CMs. e gelators were dissolved in DMSO (Sigma-Aldrich) (4: 100-197 mg/mL) or 0.21 M HCl (1: 130 mg/mL). When using HyA in addition to the nano�ber hydrogel, the HyA was �rst diluted in medium (5 mg/mL). e gel stock and HyA solutions were then sterilized with UV light for 5 minutes. Finally, the gel stock solutions were diluted in medium (either with or without the cells) in a ratio of 1 : 9. If the medium did not contain cells, the cells were subsequently plated on top of the nano�ber hydrogels.

Statistics.
Statistical signi�cance for cell attachment was analyzed using e Kruskal-Wallis and Mann-Whitney tests.

Results
3.1. Nano��ers. e properties of the nano�ber hydrogels have been previously described in detail [15,16]. Brie�y, the nano�ber hydrogels are thermoreversible and their stability can be adjusted by adding amino-acid-based substituents. e substituents also affect the responsiveness of the hydrogel to pH changes [15], which results in hydrogels that are pH sensitive or non-pH sensitive (Table 1). e pH-sensitive nano�ber hydrogels are positively charged and the non-pH-sensitive hydrogels are neutral. e positively charged pH-sensitive hydrogels can be coassembled with negatively charged HyA by electrostatic interaction.
Nano�ber hydrogels have �bers thicknesses that range from nanometers to micrometers (Table 1) and they form �brous gel networks [15,16]. e �ber surfaces are either cationic (positively charged) or protic (protons on the surfaces that exchange with water) ( Table 1). Whereas  gelators 1, 3, 4, and 5 all produced very similar �ber surfaces with terminal hydrophilic alcohol groups reminiscent of polyethylene glycol, the self-assembly process for these four compounds resulted in �bers with pronounced differences in morphology. us, this series of compounds was studied to investigate the effects of similar surface chemistry but different morphology on cellular attachment. Cryogenic transmission electron microscopy (Cryo-TEM) was used to characterize the differences in �ber morphology for these four compounds. Compound 1 was previously reported to self-assemble into tubular �bers with a diameter of approximately 4.2 nm and a very homogeneous distribution [16]. Figure 2 shows that compound 3 self-assembles into bundles of �bers with a diameter of approximately 13 nm. Compound 4 self-assembles into ribbons of uniform thickness that are approximately 50-200 nm wide. Compound 5 self-assembles into sheets of uniform thickness with widths of 100 nm to 3 m.

2 Nano��er
Coatings. e cells were cultured on the nano�ber coatings for seven days, a�er which they were stained with the LIVE/DEAD kit or �xed and stained with cardiac-speci�c antibodies (troponin T, MHC or MLC2v). e suitability of nano�ber coatings for cardiomyocyte culture was evaluated by observing cell attachment, spreading, morphology, viability, detachment, and beating in comparison to cells cultured on control surfaces (NRCs on untreated commercial well plates and hESC-CMs on 0.1% gelatincoated commercial well plates). e results for the evaluation criteria for each material are summarized in Table 1. In addition, cell attachment was quanti�ed (Figures 3 and 4), but no statistical signi�cance was detected.
e ratio between live and dead cells for both cell types was almost the same on every nano�ber coating. Approximately 70% of cells were alive and 30% were dead (Figures 5(f) and 5(l)).

Neonatal Rat
Cardiomyocytes. ere were no major differences in cell growth among the nano�ber coatings when culturing NRCs. Nano�ber coatings 1 ( , Figure 5(b)), 1 + HyA ( , Figure 5(c)), 3 ( , Figure 5(d)), and 5 ( ) supported the growth of the NRCs equally as well as the control surface ( , Figure 5(a)). e cells spread evenly and their morphology was the same as the cells in the control wells. Most of the cells remained attached throughout the entire culture period (Figure 3). e beating rate and strength were similar between nano�ber coatings and control wells.
Nano�ber coatings 6 ( ) and 2 + HyA ( , Figure 5(e)) did not support the growth of NRCs as well as the control surface and the cells did not attach properly on these coatings (Figure 3). Also the attached cells tended to detach over time. At day 7, most of the cells had detached and formed aggregates on the coatings ( Figure 5(e)). Additionally, cells were not as evenly spread on these nano�ber coatings as on the control surface ( Figure 5(a)).

Human Embryonic Stem-Cell-Derived Cardiomyocytes.
Hydrophilic nano�ber coatings 1 + HyA ( , Figure 5(h)) and 4 ( , Figure 5(i)) were best suited for supporting hESC-CMs. e cells attached well and they spread as evenly on the nano�ber surfaces as on the gelatin control surface ( , Figure 5(g)). Nano�ber coating 1 + HyA supported the growth of hESC-CMs throughout the entire culture period of seven days. Nano�ber coating 4 also supported the growth and survival of the cells, although initially cells needed a few days to adapt to this coating. Once adapted, the cells spread evenly and exhibited regular beating. F 4: Cell attachment was determined by calculating the amount of troponin-T-positive hESC-CMs on each material. However, cell attachment was not the only criteria for optimal material and thus despite good attachment some nano�ber coatings were not optimal supporters for cardiomyocyte culture (for details, see Section 3). Nano�ber coatings 5 ( , Figure 5(j)) and 2 + HyA ( ) modestly supported hESC-CM growth. On nano�ber coating 5 ( Figure 5(j)), there were fewer hESC-CMs attached and the attached cells were smaller than those on the control surface ( Figure 5(g)). Nano�ber coating 2 + HyA did not perform well initially, but towards the end of the culture period, the hESC-CMs adapted and spread well on this surface.
Nano�ber coatings 1 ( ), 2 ( ), 3 ( ), and 6 ( ) did not support the growth of hESC-CMs. e cells either did not attach at all or they detached shortly aer attachment. Some attached cells remained spherical and did not spread on the nano�ber coatings. On nano�ber coating 6, the cells surrounded the nano�ber particles rather than growing on top of them ( Figure 5(k)). ). �n nano�ber coating (k) 6 ( ), the cells did not grow on top of the coating, but they surrounded it. e round structure ( * ) is a particle of the coating. e ratio between living and dead cells was approximately the same on every coating with both (f) NRCs and (l) hESC-CMs. e same scale bar (200 m) applies to every image. Nano�ber coatings 1 + HyA and 4 were evaluated as the best ones due to the cell attachment calculations ( Figure  4) and other evaluated criteria (namely, cell spreading, morphology, viability, detachment, and beating; Table 1). Although nano�ber coating 2 had higher median cell attachment than 1 + HyA and 4, it was classi�ed as one of the least supportive coatings because the morphology of the cells was not the same as on the controls but spherical and they did not spread on the coating. In addition, there was high variation on cell attachment. Also nano�ber coatings 1, 3, and 6 were classi�ed as the least supportive materials, not according to the amount of cells attached (Figure 4) but according to other evaluation criteria listed above (Table 1).

3.4.
3 Nano�ber Hydro�els. Nano�ber coatings 1 + HyA and 4 were chosen for the nano�ber hydrogel experiments because of their superior performance in the 2D coating experiments with hESC-CMs. Nano�ber coating 1 was also investigated as a control for the nano�ber hydrogel 1 + HyA. e purpose of the 3D experiments was to see how well the 2D results translated to 3D. e same evaluation parameters were used as in the 2D coating experiments. One additional parameter was used, that is, degradation of the hydrogels. LIVE/DEAD staining could not be used because of high background.

Neonatal Rat Cardiomyocytes.
NRCs grew well both on top of and inside the nano�ber hydrogels. Nano�ber hydrogel 1 + HyA ( 3) performed well and was stable for the entire duration of the experiment (7 days). Nano�ber hydrogel 4 ( 3) supported cell growth well but appeared to be better suited for short-term studies or for other applications, as the gels started to degrade a few days aer the cells were plated. e degradation of nano�ber hydrogel 4 was remarkably faster than that of nano�ber hydrogel 1 + HyA.

Human Embryonic Stem-Cell-Derived Cardiomyocytes.
e three nano�ber hydrogels also supported the growth of hESC-CMs. All of the tested nano�ber hydrogels allowed the hESC-CMs to grow and beat on top of the gels as well as inside the gels; however, nano�ber hydrogel 4 ( 3) degraded too rapidly for long-term use (this was also observed for NRCs). Nano�ber hydrogel 1 + HyA ( ) performed better than non-coassembled nano�ber hydrogel 1 ( 3). In hydrogel 1 + HyA, the cells remained inside the gel, whereas, in non-coassembled nano�ber hydrogel 1, the cells tended to migrate through the hydrogel to the bottom of the wells. In addition, nano�ber hydrogel 1 + HyA was more robust, as it remained intact for the entire 30-day duration of the experiment. Furthermore, this hydrogel was able to maintain the beating capability of the cells (see supporting data (video) in supplementary material available at doi:10.1155/2012/285678).

Discussion
In this paper, we evaluated the suitability of six different selfassembling nano�ber hydrogels for attachment and growth of NRCs and hESC-CMs, �rst as 2D coatings and then as 3D gels. �e examined nano�ber hydrogels that were pH sensitive, non-pH sensitive, hydrophilic, and hydrophobic. Hyaluronic acid was coassembled with the pH-sensitive hydrogels. Two hydrophilic synthetic nano�ber hydrogels were found to support human and rat cardiomyocytes in both 2D and 3D culture, thus providing an alternative platform for in vitro cardiac modeling.
Collagen has been extensively studied as a biomaterial for cardiac tissue engineering [8,10,11]. Although natural polymers such as collagen may be bene�cial for cell attachment and differentiation, they oen do not have the proper mechanical strength. Batch-to-batch variations [29] and possible contamination with animal compounds also raise concerns, especially when considering future clinical applications. In addition, for proper gelation of collagen, animal-derived materials such as Matrigel, chick embryo extract, horse serum, or extracellular matrix (ECM) from decellularized porcine hearts are added [10,23]. In contrast, fully synthetic materials are homogenous, well de�ned, and have low batch-to-batch variation. Synthetic compounds can easily be modi�ed using amino-acid-based substituents [15] and are thus considered to be reliable and customizable materials for in vitro and clinical applications. Furthermore, synthetic compounds can be produced without any animal products, which is desirable for clinical applications.
In our previous study [14], we showed that the growth of NRCs was best supported by natural collagen. erefore, we wanted to continue our studies using a material with properties similar to those of natural collagen. Selfassembling nano�ber hydrogels are an emerging class of synthetic biomaterials that offer highly bioactive nanostructures that can be of interest for many biomedical applications [30,31]. Structurally, self-assembling nano�ber hydrogels have a strong resemblance to natural collagen and their biocompatibility has been demonstrated [20]. Hence, these materials were potentially suitable for supporting the growth of cardiomyocytes, which was indeed shown in this study.
In our 2D coating experiments, all nano�ber coatings supported the growth and survival of NRCs. However, two of the nano�ber coatings supported only limited attachment and growth of NRCs: one was the hydrophobic non-pH sensitive nano�ber 6 and the other was the hydrophilic pH-sensitive nano�ber 2. e best coatings for NRCs were the hydrophilic pH-sensitive nano�ber coating 1 with or without HyA and the hydrophilic non-pH sensitive nano�ber coatings 3 and 5. According to these results, hydrophilicity is more effective at promoting cell attachment and growth than pH sensitivity. However, pH-sensitive nano�ber hydrogels are positively charged and are therefore expected to provide greater cell attachment because cells have negatively charged surfaces. In these experiments, pH sensitivity seemed to work well when combined with hydrophilicity.
For hESC-CMs, the differences in cell attachment and survival among 2D nano�ber coatings were more pronounced. e nano�ber 1 hydrogel with hydrophilic �ber surface provided the best attachment and growth support also for human cardiomyocytes. e addition of HyA to nano�ber coating 1 further improved cell attachment and survival and also provided the best support for cells in 3D. Hyaluronic acid is one of the components of the ECM. erefore, it was presumed that HyA improved cell attachment [32] as well as cell proliferation and migration [33]. It would have been interesting to see how the addition of HyA to other nano�ber hydrogels affects cell attachment and survival, but it is possible to co-assemble negatively charged HyA only with positively charged pH-sensitive hydrogels.
e nano�ber hydrogel 2 (± HyA) did not support the growth of cardiomyocytes; only NRCs grew on this material to some extent. e �ber surfaces of this hydrogel are lysineterminated and thus similar to polylysine surfaces in terms of chemistry. Polylysine is considered favorable in terms of cell attachment. e attachment mechanism of cells to polylysine has not been fully elucidated, but it is usually considered to involve an electrostatic interaction between anionic cell surfaces and the cationic polylysine surface [34]. erefore, nano�ber hydrogel 2 surface was expected to enhance cell attachment more effectively than other surfaces, but this was not observed, which could be due to the more basic nature of nano�ber hydrogel 2 than polylysine. Nano�ber hydrogel 6 was also unsuitable for supporting cardiomyocytes; however, this hydrogel has a very hydrophobic �ber surface and thus these results were expected.
Nano�ber hydrogels 1, 3, 4, and 5 all have very similar �ber surfaces, but the �ber thicknesses vary, which allowed us to examine the effects of similar surface chemistries and different �ber morphologies on cellular attachment. �e varied the thickness of the �bers� 1 (4.2 nm), 3 (13 nm), 4 (50-200 nm), and 5 (100 nm-3 m). For comparison, collagen �bers are 1-20 m thick and elastin �bers are 0.2-1.5 m thick [35]. Nano�bers in hydrogel 5 formed sheets which thicknesses are similar to the �ber thicknesses of collagen and elastin. Nano�ber hydrogel 1 had the thinnest �bers. In our study, nano�ber hydrogel 1 (+ HyA) had the best performance. e second thickest nano�ber hydrogel (nano�ber hydrogel 4) had �ber sheet thicknesses of 50-200 nm and also performed well with hESC-CMs. Among nano�ber hydrogels with protic �ber surfaces, �ber thickness had less in�uence on NRC attachment than HyA addition. However, the �ber dimensions and the addition of HyA both affected hESC-CMs attachment.
Some differences in the results between NRCs and hESC-CMs were observed, especially in 2D experiments. It has been reviewed in previous experiments that hESC-CMs are more sensitive to surrounding biomaterial than other cell types [36]. NRCs are more robust and also in our experiments grew better on several nano�ber hydrogels than hESC-CMs. Shapira-Schweitzer et al. also showed that NRCs are easier to handle, maturate faster, and contract more effectively than hESC-CMs [22]. Nevertheless, NRCs can be used as a model for human cells for preliminary experiments, for example, to screen different biomaterials. is was also demonstrated in our study as hESC-CMs did not grow well on any of the materials where NRCs did not grow.
�hen evaluating nano�ber materials as in the present study, it is important to note that both NRC-and hESCderived beating areas contain both cardiomyocytes and noncardiomyocytes. Consequently, when the beating areas are dissociated, there is always a mixture of cell types. As a result, it is currently not possible to obtain pure populations of human cardiomyocytes for testing. However, the other cell types existing in the beating areas, such as �broblasts, have been shown to support cardiomyocyte growth and functionality [37], which suggests that the mixed population of cells in our cultures is bene�cial. Additionally, the presence of other cell types in cardiac gras (i.e., �broblasts and endothelial cells (ECs)), has been demonstrated to improve the vascularization and function of gras in vivo [38].
As cells grow in 3D in vivo, we wanted to test whether nano�bers in 3D hydrogels would support cell growth better than on 2D coatings. However, our results did not support this hypothesis, possibly because of problems with the gelation process. e optimal amount of cells needed for formation of 3D cell structures is also not known. In 3D hydrogel experiments, some variation from well to well was observed, mainly because gelation occurred rapidly resulting in heterogeneous hydrogels. Consequently, in soer parts of the hydrogels, some of the cells migrated to the bottom of the wells rather than staying attached to the hydrogel matrix. e cells that stayed attached to the hydrogels grew well and retained their beating capability. e heterogeneous nature of the hydrogels sometimes limited the visualization of the cells. ere were no major differences in the growth pattern of cells inside or on top of the different nano�ber hydrogels for either of the cell types.
Not many 3D biomaterial studies have been performed using hESC-CMs. �ur study is the �rst to demonstrate the growth of hESC-CMs in 3D self-assembling nano�ber hydrogels. e �rst 3D vascularized human cardiac tissues were created by combining hESC-CMs, ECs, and �broblasts with PLLA(50%)/PLGA(50%) biodegradable scaffolds [24]. Transplantation of hESC-CMs in alginate scaffolds into infarcted heart tissue has also been described; however, this treatment did not promote myogenic differentiation or organization of the implanted cells [39]. Furthermore, hESC-CM function and maturation within PEGylated �brinogen (PF) hydrogels has been shown. e responsiveness of these cells to cardiac drugs demonstrated the potential to use this system for in vitro drug screening [22]. e same group showed that codelivery of PF matrix with hESC-CMs into infarcted areas provided additional therapeutic effects and prevented unfavorable postinfarction cardiac remodeling [40]. Madden et al. showed that hESC-CMs seeded into microtemplated poly(2-hydroxyethyl methacrylate-comethacrylic acid) (pHEMA-co-MAA) hydrogels cultured for 2 weeks could reach adult heart density in vitro. Additionally, acellular scaffolds implanted in rats enhanced angiogenesis [41]. In one study, porcine heart ECM and collagen I were used to induce cardiac differentiation of hESCs. Hydrogels with different ratios of ECM and collagen were prepared and cardiomyocyte maturation and contraction were evaluated. Hydrogels with a higher ECM content promoted cardiac maturation [23]. Human engineered heart tissue (hEHT) has also been developed from hESC-CMs and �brinogen; it forms a dense network that responds to chronotropic compounds [42]. Based on our study, self-assembling nano�ber hydrogels can be modi�ed to obtain bene�cial features that support the growth of cardiomyocytes. To further improve the performance of these synthetic cell-supporting structures, the gelation procedure should be optimized to allow more homogeneous formation of the hydrogels. Another possibility that we are now investigating is decoration of the self-assembling nano�ber hydrogels with functional groups (e.g., the RGD peptide sequence) to improve cell adhesion. Finally, because a main application for 3D cardiac tissue is drug discovery and testing, the possibility of measuring signals from hESC-CMs in a 3D hydrogel using a microelectrode array (MEA) is under investigation. e suitability of MEA platforms for measuring drug responses of hESC-CMs has been shown in 2D culture [28,43,44].

Conclusions
In our previous study, we compared different natural and synthetic biomaterials for cardiomyocyte culture. Collagen type I best supported the growth of cardiomyocytes. However, as a natural material, collagen has batch-to-batch variations. We therefore decided to investigate a synthetic material similar to collagen. In this study, neonatal rat cardiomyocytes and human embryonic stem-cell-derived cardiomyocytes were grown on different synthetic selfassembling nano�ber hydrogels. e pH-sensitive nano�ber hydrogel with hydrophilic and protic �ber surfaces and coassembled with hyaluronic acid best supported the growth of rat and human cardiomyocytes. ese nano�ber hydrogels are promising materials for the development of future cardiac tissue models.

Con�ict of �nterests
e author declare no con�icts of interest.