Dynamic control of contractile resistance to iPSC-derived micro-heart muscle arrays

Abstract

Many types of cardiovascular disease are linked to the mechanical forces placed on the heart. However, our understanding of how mechanical forces exactly affect the cellular biology of the heart remains incomplete. In vitro models based on cardiomyocytes derived from human induced pluripotent stem cells (iPSC-CM) enable researchers to develop medium to high-throughput systems to study cardiac mechanobiology at the cellular level. Previous models have been developed to enable the study of mechanical forces, such as cardiac afterload. However, most of these models require exogenous extracellular matrix (ECM) to form cardiac tissues. Recently, a system was developed to simulate changes in afterload by grafting ECM-free micro-heart muscle arrays to elastomeric substrates of discrete stiffnesses. In the present study, we extended this system by combining the elastomer-grafted tissue arrays with a magnetorheological elastomeric substrate. This system allows iPSC-CM based micro-heart muscle arrays to experience dynamic changes in contractile resistance to mimic dynamically altered afterload. Acute changes in substrate stiffness led to acute changes in the calcium dynamics and contractile forces, illustrating the system’s ability to dynamically elicit changes in tissue mechanics by dynamically changing contractile resistance.

1 |. INTRODUCTION

Cardiovascular disease is the leading cause of death in the United States, accounting for nearly 1 in 5 deaths in 2020.¹ Many types of heart disease are linked to the mechanical forces on the cells in the heart.² These mechanical forces generally fall into two categories; preload, the strain applied to the cardiac tissue as the heart fills with blood, and afterload, the resistance to the heart’s ability to contract against systemic blood pressure.^3,4 These forces directly affect heart muscle biology, illustrated by cardiac hypertrophy often being a secondary response to hypertension-induced pressure overload in patients.⁵ Changes in preload and afterload have been linked to the prognosis of genetically inherited cardiomyopathies. For example, patients with hypertrophic cardiomyopathy (HCM) have a much worse prognosis if they are hypertensive and/or obese.⁶ Despite this implication of a potential interaction between genetics and mechanical loading in disease pathology,⁷ the link between mechanical forces, genetic mutations, and cardiac disease needs to be more clearly elucidated.

Due to the challenge of accurately and non-invasively measuring forces on the heart in clinical subjects, other systems must be used to study heart muscle mechanobiology. Early animal models were created to evaluate how increased blood pressure led to a cardiac response, even when the hormonal feedback systems are disconnected from the heart.⁸ This pioneering work showed that the heart was mechanosensitive. More recently, animal models have been developed to test how mechanical loads on the heart, typically caused by surgical maneuvers, impact physiology of wild type animals or in animal models of inherited diseases.^9–14 However, while indispensable for our understanding of heart muscle mechanobiology, animal models present several fundamental obstacles. First, it is difficult to precisely control cardiac mechanical loading via surgical maneuvers, and distinguishing the effects of mechanical changes alone versus surgically induced inflammation is challenging, particularly at early time points.¹⁵ Second, critical differences in physiology between species limit the translational potential of animal models.¹⁶ Finally, throughput is an inherently limiting factor for in vivo studies.

These limitations associated with in vivo models have led to the development of various in vitro models as alternatives.^17–20 Many such approaches are variations of 3D engineered heart tissues.^21–24 These in vitro 3D models have been further developed to analyze how different levels of mechanical forces simulating preload or after-load can affect heart tissue.^25–29 As the heart in vivo experiences dynamic changes in mechanical loading, there has been a push to devise systems that allow dynamic changes in mechanical forces applied to engineered tissues to simulate changes of both preload^30,31 and afterload.^32–34

A limitation of these prior approaches to dynamically modulate mechanical forces on 3D engineered tissues is that they all require the cells to be encapsulated in gels made from exogenous extracellular matrix (ECM). These gels, typically comprised of pure collagen²² or fibrin,³⁵ are different in composition than the native ECM of the heart,³⁶ and this may directly affect the biology of cells within.³⁷ Further, the use of exogenous ECM for cell encapsulation presents additional experimental challenges because the short “working time” before these ECM gels crosslink limits experimental throughput. Further technical challenges include working with small volumes of viscous pre-gel polymers to produce micro-scale tissues.

Based on these limitations, approaches to dynamically modulate mechanical forces on in vitro engineered cardiac tissues made without exogenous ECM are desirable. Magnetorheological elastomer (MRE) substrates were developed to allow reversible tuning of material stiffness.³⁸ These materials have been applied on iPSC derived cardiac fibroblasts and cardiomyocytes (iPSC-CM) cultured as single cells and/or confluent monolayers in 2D.^39,40 MRE substrates exhibit dramatic, rapid, and reversible shifts in stiffness (e.g., Young’s modulus) in the presence of magnetic fields. In turn, these stiffness changes modulate cellular structure, physiology, and transcriptomes. Further developments of these MRE materials have been reported to enable even larger ranges of Young’s Modulus tunability.⁴¹ In the present study, we demonstrate the combination of our previously described elastomer-grafted 3D micro-heart muscle (μHM) system to simulate cardiac afterload^28,29 with MRE substrates to enable dynamic control of afterload. By devising a new approach to image these μHM in real-time immediately following substrate stiffness changes, we observed the time-course of how the physiology of 3D engineered cardiac micro-tissues evolves over time in response to acute changes in substrate stiffness.

2 |. MATERIALS AND METHODS

2.1 |. Magnetorheological elastomer fabrication

Magnetorheological elastomers were fabricated with polydimethylsiloxane (PDMS) Sylgard 527 Silicone Dielectric Gel (Dow Corning, Midland, MI) and carbonyl iron powder (ChemicalStore.com, Clifton, NJ; 3–10 μm Diameter Zero Valent Microspheres) using a modified version of the method previously described by Corbin et al.³⁹ Briefly, Sylgard 527 part A, Sylgard 527 part B, and Fe microspheres were combined at a 1:1:2 weight ratio, such that the final mixture was 50% Sylgard 527 and 50% Fe by weight. This Fe/PDMS mixture was then homogenized vigorously by hand for 10 min. The mixture was then placed into a Branson 1510 Ultrasonic Bath Cleaner (Branson Ultrasonics, Brookfield, CT) and sonicated for 3 min to break up Fe aggregates and further homogenize. Next, the Fe/PDMS mixture was poured into acrylic molds to generate materials of defined geometry. The acrylic molds were pre-coated with Rain-X to decrease their adhesion of the elastomer for easier removal after curing. The poured mixture was then degassed in a vacuum desiccator chamber for 10 min to remove any air bubbles. Finally, the Fe/PDMS mixture was then placed into a 60°C oven to cure overnight to form the MRE samples. All control samples of pure 527 were fabricated using this method but omitted the iron particles such that Sylgard 527 parts A and B were mixed at a 1:1 ratio. Using this method, the final MRE samples fabricated were cylinders with a thickness of 2.95 ± 0.05 mm and a diameter of 15 mm.

2.2 |. Magnetic field characterization

To modulate MRE stiffness, various neodymium permanent magnets were placed at varying distances from the sample, altering the magnetic field around the materials. The magnets used in this study were ⅛-inch thick, 1-inch diameter, N48 grade, neodymium rare earth disc magnets (Element Magnets Inc, Centennial, CO, 172 mT surface field strength) and 1/4-inch thick, 1-inch diameter, N52 grade, neodymium rare earth disc magnets (Element Magnets Inc, Centennial, CO, 335 mT surface field strength). To test the magnetic field at different axial and radial distances a GM1-ST DC Gaussmeter (AlphaLab, Inc., Salt Lake City, UT) was used by holding the magnetic field sensor probe at different axial distances from the surface of the magnet, as measured with 8-inch digital calipers (General Tools and Instruments, New York, NY), at the center and the radial edge of the magnet. For the two types of magnets, each location was tested on three different magnets. The mean magnetic field strengths for the magnets at the different distances were calculated.

2.3 |. Shear modulus measurement

As the primary mode by which cells within μHM deflect their substrate is shear, the shear modulus⁴² was measured. Standard rheometers are not suitable to be used with the magnetic field induced by neodymium magnets, so a custom setup was devised to measure shear modulus (Figure 1A). Briefly, nylon fishing wire was attached via superglue to the center of a 25 mm × 25 mm square glass coverslip and the other side of the coverslip was attached to the sample via the sample’s adhesive properties. The exposed side of the sample was then stuck to the wall of a custom 3D-printed poly(lactic) acid (PLA) stand using the adhesive properties of the sample. A downward shear force was then applied to the surface sample by attaching a weight of known mass to the end of the fishing wire. A photo was taken of the sample before and after adding the weight and the distance the sample was displaced in the shear direction was measured using ImageJ (Figure 1B).

FIGURE 1.

Shear modulus of MRE materials can be dynamically varied. (A) Example set up of the custom shear test platform with a MRE sample in place and a binder clip loaded as a weight from the front view (left) and side view (right). (B) Representative image of a sample before (left) and after (right) adding the weight, and the lines drawn in ImageJ to calculate the displacement distance. (C) Rheologic testing of untreated MRE samples outside of an applied magnetic field reported consistent shear storage and loss modulus values up to 10% oscillation strain when measured at 1 rad/s. The storage modulus was measured to be 0.39 kPa and the loss modulus was measured to be 0.11 kPa. (D) Unpaired t-test comparison of the shear modulus measured using the custom shear setup and the storage modulus measured using a rheometer on unmagnetized (0 mT) untreated MRE samples indicated that the difference in measured values was not statistically significant (p = .3111). (E) There was no statistical change in shear modulus of autoclaved Sylgard 527 when absent of Fe particles (G = 0.61 kPa), regardless of the applied magnetic field, and when compared to the autoclaved MRE outside of a magnetic field (G = 0.59 kPa, p = .9046). When applying a magnetic field, there was statistically significant increase in the shear modulus of the autoclaved MRE, with a shear modulus of 2.60 kPa at ~90 mT and 5.26 kPa at ~180 mT (p < .0001). ****indicates significant difference p < .0001. Error bars: standard deviation (SD), n ≥ 4 (C–E).

The Young’s modulus was also then be calculated using the relationship between shear modulus and Young’s modulus,⁴² as shown in Equation 1, where E is the Young’s modulus and ν is the Poisson’s ratio. As the Poisson’s ratio of PDMS has previously been shown to be approximately equal to 0.50,⁴³ the Young’s modulus was assumed to be three times the shear modulus (Equation 1).

$$E=2G(1+\nu ).$$ (1)

To investigate how the shear modulus was affected by changing the magnetic flux through MRE samples, this same protocol was applied while changing the magnetic flux by placing a permanent magnet a defined distance from the MRE via a spacer on the PLA stand. Sylgard 527 and MRE samples fabricated as described were tested without a magnetic field, as well as with the N48 and N52 neodymium magnets 2 mm from the near side of the sample (~5 mm from the far side of the sample) to induce a defined magnetic field.

As the samples would be autoclaved (121°C for 10 min) for sterilization prior to cell culture experiments, we tested whether exposing the samples to these high temperature conditions might alter their mechanical and magneto-responsive properties. To test this, the custom shear test method was performed with the same samples untreated and after being autoclaved. To see if more extreme heat treatment would affect MRE baseline stiffness and response to magnetic fields, the samples were baked again after autoclaving at 200°C for 24 h. The samples were then again tested using the previously described protocol for the custom shear test method.

The custom shear test was validated by measuring the shear storage modulus (G′) and loss modulus (G″) using a TA Instruments HR-20 rheometer (TA Instruments, New Castle, DE) fitted with an 8 mm diameter circular parallel plate and conducting strain sweeps from 0% to 10% at 1 rad/s on untreated MRE samples absent of an external magnetic field. The MRE samples were tested after first measuring them on the custom shear setup and then punching them with an 8 mm biopsy punch.

2.4 |. iPSC-cardiomyocyte differentiation

All iPSC studies were performed with the Wild Type C (WTC) human hiPSC cell line previously engineered to express a single-copy of CAG-driven GCaMP6f, which is knocked into the first Exon of the AAVS1 “safe harbor” locus.⁴⁴ The parent cell line (WTC) was reprogrammed from fibroblasts derived from a healthy 30-year-old male with a normal electrocardiogram and no known family history of heart disease (Coriell Repository # GM25256). To differentiate iPSC into cardiomyocytes, iPSC were first passaged at least three times in Essential 8 medium on tissue culture substrates coated with growth factor reduced Geltrex^™ (Thermo Fisher Scientific, Waltham, MA) diluted at a 1:100 ratio in KnockOut^™ DMEM (Thermo Fisher Scientific). For differentiation, iPSC were plated in Essential 8 media with 10 μM Y27632 (Biogems, Westlake Village, CA) at 25,000 cells/cm² onto 1:100 Geltrex^™. iPSC were then guided to a cardiac lineage using small molecule Wnt signaling manipulation.⁴⁵ Briefly, when iPSC were confluent (day 0 of differentiation), GSK-3β inhibition was performed using 6 μM CHIR99021 (Biogems). 48 h later, Wnt signaling was inhibited with 5 μM IWP-2 (Biogems). From days 0 to 4 of differentiation, cells were cultured in RPMI 1640 (Thermo Fisher Scientific) with B27 supplement without insulin, supplemented with 150 μg/mL L-Ascorbic acid (Sigma, Millipore-Sigma, St. Louis, MO). From differentiation day 6 onward, cells were cultured in RPMI 1640 with B27 containing insulin (B27c). Spontaneously beating cells were typically observed before differentiation day 10.

2.5 |. iPSC-CM cell adhesion assay

A cell adhesion assay was performed to assess whether the Fe doped substrates hindered cell activity, potentially due to cytotoxicity. MRE samples were fabricated as previously described. Sylgard 527 substrates free of Fe particles were used as a positive control. The samples were attached to glass coverslips using Sylgard 184 as a glue, and autoclaved (121°C for 10 min) to sterilize. Sterile fibronectin (FN; Sigma, 20 μg/mL) was next physisorbed onto the surface of the substrates for 1 h at 22°C followed by three rinses in Dulbecco’s Phosphate Buffered Saline (dPBS).

Cardiomyocytes at day 28 following initiation of cardiomyocyte differentiation from iPSC were dissociated using 10× TrypLE Select Enzyme (Thermo Fisher Scientific) and resuspended into embryoid body 20 medium (EB20; knockout DMEM with 20% Fetal Bovine Serum) supplemented with 10 μM Y27632, 150 μg/mL L-ascorbic acid, 4 μg/mL vitamin B12, and 3.2 μg/mL penicillin. Cells were plated onto substrates at a density of 2 × 10⁴ cells/cm². After 24 h incubation at 37°C with 5% CO₂, the cells were fixed using 4% paraformaldehyde for 15 min and then washed three times with dPBS. To differentiate myocytes from other cell populations, the cells were permeabilized for 15 min with 0.1% Triton-X-100, blocked for 45 min with 3% goat serum and 3% bovine serum albumin in 0.1% Triton-X-100, and then incubated with ACTN2 primary antibody (Sigma, mAB Clone EA-53, 1:1000 dilution of stock to a working concentration of ~1 μg/mL) overnight at 4°C, secondary antibody for 2 h at room temperature.

To quantify cell adhesion and spreading, the cells were then stained for F-actin using phalloidin (Alexa Fluor^™ 488 Phalloidin; Thermo Fisher Scientific) for 1 h at room temperature. Cell nuclei were counterstained for 10 min at room temperature with Hoechst 33342 (Thermo Fisher Scientific). Fluorescent micrographs were then taken for 4–5 fields for each sample using an epifluorescence microscope and individual projected cells areas were automatically quanti-fied using ImageJ. The automatic quantification was done by applying a binary threshold to the image and using the ImageJ built-in “Analyze Particles” feature to automatically measure the cell sizes. Particles below 500 μm, above 10,000 μm and any obvious non-cellular debris were excluded from the analysis.

2.6 |. Characterization of MRE photothermal response

It was hypothesized that Fe-doped substrates would absorb the high intensity light during fluorescence imaging, leading to photothermal heating that could damage cells.⁴⁶ The photothermal properties of the MRE substrates were tested using the 475 ± 28 nm LED light from an epifluorescence microscope (Eclipse Ts2R, Nikon; Tokyo, Japan) equipped with an Aura Light Engine (Lumencor, Beaverton, OR). Sylgard 527 samples free of Fe particles were used as a positive control. In some cases, a thin layer (~5 μm) of pure Sylgard 527 was applied atop the MRE to determine if this could provide insulation. The top layer of PDMS without carbonyl iron was applied to the MRE using a spin coater (Laurell Technologies, North Wales, PA) at 3000 rpm for 30 s and curing again for >2 h at 60°C.

During tests of MRE photothermal response, samples were suspended in both air and RPMI 1640 media to see if the surrounding environments would affect the surface heating. LED light was focused through a 10× objective with a numerical aperture of 0.3 on the samples for 30 s at varying intensities. This objective was chosen because it is typically used for assessing μHM physiology. After 30 s of continuous illumination, the temperature of the point where the light was focused on the samples was measured using a Digital Temperature Probe (Traceable; Webster, TX) as well as at distances of 2 mm and 10 mm outside of the area of the incident light. The radiant power at the sample surface for each of the intensities was measured using a Newport 1918-R Power Meter equipped with an 818-SL photodetector (Newport Corp., Irvine, CA). The intensity of the light was then calculated by dividing the radiant power by the light spot size at the focal plane.

The temporal temperature dynamics of the MRE and Sylgard 527 samples were similarly tested by suspending the samples in air and then focusing light of varying intensities on the surface through a 10× objective. To measure the heating profile, the light was left on the sample for 120 s, and temperature was recorded every 10 s. The light was then turned off and temperature was recorded every 10 s for another 80 s to measure the cooling profile.

2.7 |. Stencil fabrication

To create μHM, “dog bone” shaped stencils were fabricated using the previously described Hydrogel Assisted STereolithographic Elastomer (HASTE) prototyping method.⁴⁷ This method allows replication of high resolution, stereolithographic 3D prints into PDMS elastomer. Briefly, initial prints were designed using Autodesk Inventor Professional to have multiple dog bone shaped extrusions with knobs on each end having 1 mm × 1 mm square dimensions connected by a 1 mm × 200 μm rectangular “shaft” on a 3 mm thick base plate. The prints were then printed in ClearResin^™ on a Form 3 SLA 3D Printer (Formlabs Inc., Somerville, MA). The resultant printed resins were first replica molded into 1.5% wt/v agar, which was then used as a negative to produce positive, PDMS replicas of the original 3D printed resin.

To generate the PDMS stencils with dog bone-shaped through-holes from the PDMS replica, the replica was oxidized in an oxygen plasma chamber (Harrick Plasma, Ithaca, NY; pressure/flow rate: 580–680 μTorr) for 90 s at 30 W and then immediately treated with trichloro(1H,1H,2H,2H-perfluorooctyl)silane (Sigma) via vapor deposition to enable PDMS-off-PDMS molding for final stencil creation. Stencils of ~1 mm thickness with the dog bone shaped through-holes were created by pouring Sylgard 184 onto the PDMS replicas and clamping them between two acrylic plates and curing overnight at 60°C. Finally, the thin stencil sheets were then removed from the PDMS replica and cut out into individual groups of 3 dog bones.

2.8 |. Preparation of MRE substrates for μHM formation

MRE samples of 50% wt/wt Fe in Sylgard 527 were formed as described above (Figure 2A). Following MRE crosslinking, a drop of Sylgard 184, mixed at a 10:1 base to curing agent ratio, was placed onto a 22 mm × 22 mm glass coverslip to glue down the MRE by curing the Sylgard 184 in between the coverslip and MRE at 60°C for at least 4 h. The substrates were then autoclaved to remove potential contaminants deep within the thick PDMS materials that may not be removed during ethanol disinfection. The devices were brought back into a non-sterile environment after this point under the assumption that contaminants would not be able to reach the insides of the samples.

FIGURE 2.

Overview of experimental design to acutely modulate μHM contractile resistance. (A) Schematic illustrating preparation of MRE substrate for seeding of iPSC-CMs to generate μHM. (B) Schematic detailing experimental procedure to form iPSC-CM μHM tissues and magnetically stiffen and soften the underlying MRE substrate. (C) Schematic detailing substrate stiffness from experimental day 0 to day 4 and time scales at which different mechanisms will cause changes in tissue physiology.

To enable force measurements, a thin layer of fluorescent beads was next applied on top of the MRE.^28,29,48 Briefly, a 10 μL aliquot of fluorescent bead solution (Fluorescein; 1.06 μm diameter; Poly-sciences Inc., Warrington, PA) was desiccated in a vacuum chamber overnight. Dried beads were then thoroughly mixed with 1.33 g of Sylgard 527 and 1.33 g of Fe powder, and then degassed in a vacuum desiccator to remove bubbles. The beads mixture was then spun down using a spin coater onto the PDMS coverslips at 3000 rpm for 30 second and cured overnight at 60°C to generate ~5 μm thick layers of bead-laden PDMS atop the MRE substrates.

Following generation of the beads layer, FN was covalently grafted to the surface of the substrate.²⁸ PDMS substrates were first oxidized in an oxygen plasma chamber for 90 s at 30 W. Immediately after plasma treatment, oxidized substrates were placed into 5% (3-aminopropyl)triethoxysilane (APTES; Sigma, 5% v/v in methanol) for 1 h on an orbital shaker to graft the surfaces with amine groups. The substrates were then washed 3 times for 5 min with methanol and left in dPBS overnight to reduce autofluorescence.²⁹ Following the dPBS wash, substrates were reacted with 2.5% glutaraldehyde (Thermo Fisher Scientific) in dPBS for 2 h on an orbital shaker followed by 3 dPBS washes. 20 μg/mL FN was placed on the substrates for at least 12 h to graft the ECM protein to the surface via glutaraldehyde groups, followed by 3 dPBS washes and 3 DI water washes. Finally, substrates were reacted with 2.5% w/v ethanolamine hydrochloride (Alfa Aesar, Haverhill, MA) in dPBS to quench any unreacted aldehyde groups, followed by 3 dPBS washes.

Substrates were next moved to 6 well tissue culture plates and disinfected by soaking with sterile 70% ethanol for 48 h and then removing the ethanol and allowing the remaining to evaporate off for 1 h while substrates were exposed to the 30 W UV source of the biosafety hood. The substrates were then rinsed with 3 dPBS washes. Autoclaved dog bone shaped stencils were then soaked in sterile Pluronic F68 (Sigma, 1% w/v in dPBS) for 1 h and rinsed 3 times with dPBS. Following substrate disinfection, the stencils were bonded to the substrates with 0.2 mm filter-sterilized methanol for 1–2 h at 60°C.

2.9 |. μHM formation

10 days before MRE experiments, iPSC-CM monolayers at days 18–22 following initiation of cardiomyocyte differentiation with CHIR99021 were dissociated using 10× TrypLE Select and resuspended into EB20 supplemented with 10 μM Y27632, 150 μg/mL L-ascorbic acid, 4 μg/mL vitamin B12, and 3.2 μg/mL penicillin, to a density of 5 × 10⁷ cells/mL. Dog bone-shaped microwells were next pre-wetted by centrifugation at 300 rcf, for 3 min, with 3 mL of cell-free resuspension media. Next, 3 μL of this cell suspension was seeded into each of the dog bones. After cell seeding, a small volume of resuspension media was applied to the periphery of plate wells, to maintain a humid environment as the tissues formed. The devices were incubated at 37°C for 2.5 h to allow cells to agglomerate to form tissue before filling the entire well to cover tissues with the resuspension media. 2 days after seeding cells, media was changed to RPMI 1640 supplemented with B27, 4 μg/mL vitamin B12, and 3.2 μg/mL penicillin.

2.10 |. μHM calcium transient analysis

Calcium dynamics were visualized using the encoded GcaMP6f indicator.⁴⁴ 10 days following μHM formation (Supplemental Figure 1), the tissues were visualized at 4× magnification and the LED at 0.077 W/cm² intensity and recorded at 100 fps using an epifluorescence microscope. This intensity level was chosen to minimize any potential artifacts caused by heating of the MRE substrates through photothermal effects. Importantly, whereas MRE photothermal response was calculated with an illumination time of 30 s during which light is applied continuously, for dynamic imaging, physiology of micro-tissues was obtained by imaging samples for 5–7 s, with 10 ms pulses. This condition was observed to cause no detectable photo-thermal heating.

Calcium transients of the entire μHM were obtained by averaging fluorescence intensity across all pixels for each frame. An open-source MATLAB pipeline⁴⁹ was used to analyze the epifluorescence videos, producing data including, but not limited to, upstroke duration (UPD), signal decay from peak to 25% above baseline (τ₇₅), maximum calcium amplitude (measured as the change in intensity divided by the baseline intensity), and integrated calcium intake per beat (total area under the curve).

2.11 |. μHM traction force microscopy (TFM)

The previously fabricated fluorescent beads layer was used to conduct TFM analysis in order to measure the substrate deflection and contractile force generated from the spontaneously beating μHM. Following formation of the μHM, deflection of the fluorescent beads could be observed and were imaged using an epifluorescence microscope. Images were taken at 100 fps with 10× magnification and the 0.174 W/cm² LED intensity at the “knob” of the tissue as this is where the tissues fully attach to the substrate, such that the majority of contractile deflection transfers to the substrate there. Contractility of the spontaneously beating μHM against the substrates was quanti-fied using a modified version of an open source TFM software.⁵⁰

2.12 |. μHM edge deflection

To determine whether bead deflections measured using TFM were directly coupled to μHM contractile motion, we manually tracked deflection of the edges of each “knob,” of the tissues during peak contraction. To do this, videos of beating tissues pre-magnetization, 5 min after stiffening, and 2 h after stiffening were loaded into ImageJ. The location of the edge of the tissue was obtained when the tissue was at rest and during peak contraction to obtain μHM contractile motion.

2.13 |. Modified tissue culture setup to allow acute MRE magnetization while imaging

Four days after seeding the cells, the devices were inverted onto PDMS inversion frames (Supplemental Figure 2A) that were previously sterilized using 70% ethanol sterilization and overnight exposure to the 30 W biosafety cabinet UV, and the plate lid was changed out with a custom PLA plate lid (Supplemental Figure 2B) to enable higher levels of magnetic flux through the substrates. The tissues were visualized on an epifluorescence microscope to confirm coherent tissues had formed. μHM were imaged 10 days after cell seeding to generate baselines of the calcium transients and traction forces using the previously described methods (Figure 2B).

A non-motorized version of the magnetic actuation platform described by Stottlemire et al., was fabricated and modified to dynamically control the magnetic field experienced by the MRE array.⁵¹ Briefly, a 3D-printed magnetic holder attachment (Supplemental Figure 2C), was loaded with 1/4-inch thick, 1-inch diameter, N52 grade, neodymium disc magnets, to allow the magnets to be placed precisely 2 mm away from the inverted cell cultured substrates, or ~5 mm from the cell culture surface.

Ten days after cell seeding (termed experimental day 0) the calcium transients had stabilized and the magnet holder stage was moved to place magnets ~5 mm away from the culture surface, increasing the magnetic flux through the MRE substrate and thus increasing stiffness. The tissues were imaged before adding the magnet and immediately after adding the magnet, as well as 5 min, 2 h, 24 h, and 48 h later. This spread of timepoints was chosen to allow us to sample physiologic responses to stiffening that rely on rapid adaptation mechanisms (e.g., stretch sensitive ion channels) as well as longer-term, response mechanisms such as transcription (Figure 2C).⁵²

In a subset of studies, after the 48-h time point, the magnet was again removed, and the tissues were imaged immediately, as well as 5 min, 2 h, 24 h, and 48 h later, to determine how μHM respond to a subsequent acute decrease in substrate stiffness. The control samples were imaged at all of the same time points except for immediately after changing substrate stiffness and 5 min later as tissue dynamics are expected to be stable over this time frame. The effects of the changing substrates stiffness on the tissue calcium handling and traction forces were analyzed using the previously described imaging methods and open-source MATLAB software packages.

2.14 |. Statistical analysis

GraphPad Prism 9.5.1 was used for statistical analysis in all the results shown throughout this thesis. One-way or two-way analysis of variance (ANOVA) followed by multiple comparison between groups was performed using a post hoc Holm Sidak’s test. For comparison between two groups, the unpaired t-test with Welch’s correction was used. p < .05 was considered a statistically significant difference.

3 |. RESULTS

3.1 |. Measured magnetic field strengths

The mean magnetic flux densities measured for the two types of magnets showed that the N52 magnets have magnetic flux densities twice as strong as that of the N48 magnets, with a peak magnetic flux density at the center of the magnet of 305 ± 10 mT for the N52 compared to 140 ± 5 mT for the N48 (Supplemental Figure 3). The range of values measured between the center and edge of the surface of the magnets encompassed the manufacturer reported values, validating our measurements. All magnets used during this study had 1-inch (25.4 mm) diameters, nearly 70% larger diameter than the MRE samples. As such, it is unlikely that magnetic boundary flux effects would affect tissues formed at the center of the MRE substrates. The magnetic flux density 2 mm from the surface of the N48 magnet was measured to be 130 ± 5 mT at all radial distances within the diameter of the magnet. The magnetic flux density measured 5 mm from the N48 magnet was measured to be 102 ± 5 mT at the center and 70 ± 5 mT at the radial edge, with an overall average of ~90 mT. The magnetic flux density measured 2 mm away from the N52 magnet was found to be 260 ± 10 mT at the center of the magnet and 230 ± 10 mT at the radial edge of the magnet, with an overall average of ~250 mT. The magnetic flux 5 mm from the N52 magnet was measured to be 193 ± 10 mT at center and 132 ± 5 mt at the radial edge, with an overall average of ~180 mT.

3.2 |. MRE shear modulus

Rheological testing showed that untreated MRE samples had a shear storage modulus of 0.39 ± 0.02 kPa and a shear loss modulus of 0.11 ± 0.01 kPa (Figure 1C). Importantly, when using the same samples there was no significant difference (p = .3111) between the static shear modulus (G = 0.42 ± 0.03 kPa) measured using the custom setup and the dynamic shear storage modulus (G’ = 0. 39 ± 0.02 kPa) measured using the rheometer (Figure 1D).

Using the custom shear system, we found that absent external magnetic fields, autoclaved MRE samples and autoclaved Sylgard 527 control samples had similar shear modulus values of 0.59 ± 0.09 kPa and 0.61 ± 0.08 kPa, respectively (Figure 1E). When magnets were placed ~5 mm from the far side of the MRE samples, the shear modulus increased to 2.60 ± 0.35 kPa when inducing a magnetic field of ~90 mT via the N48 and to 5.26 ± 0.71 kPa when inducing a magnetic field of ~180 mT via the N52. The shear modulus of the Sylgard 527 did not change when placed in the same magnetic fields. The Young’s modulus of the autoclaved Sylgard 527 samples was calculated to be 1.83 ± 0.24 kPa at all levels of magnetic field strength while the autoclaved MRE samples’ Young’s modulus was 1.76 ± 0.28, 7.81 ± 1.06, and 15.77 ± 2.14 kPa in magnetic fields of 0, ~90, and ~180 mT, respectively.

Importantly, autoclaving did not induce significant changes in MRE shear modulus at 0 mT (p = .6078), 90 mT (p = .6108) or 180 mT (p = .3670) (Supplemental Figure 4). In contrast, MRE samples that were heated to 200°C for 24 h, exhibited a significant increase in baseline shear modulus (p = .0024). Further, MRE samples treated at this high temperature for 24 h exhibited a markedly lower degree of stiffening in the presence of a 180 mT magnetic field.

3.3 |. iPSC-CM adhesion to MRE substrates

Although we observed a slight decrease in iPSC-CM spreading area on MRE substrates (Figure 3A) compared to on pure Sylgard 527 substrates (Figure 3B), the difference was not statistically significant (p = .0728), and the difference appeared to stem from a small sub-population of the cells (Figure 3C). Moreover, Z-disk staining via antibodies against sarcomeric α-actinin (Figure 3D) verified that iPSC-CM that attached and spread on the FN-coated MRE substrates were grossly normal in their structure.

FIGURE 3.

MRE substrates are not toxic to iPSC cardiomyocytes. (A) Representative images of iPSC-CM attachment (F-actin; green) on (i) MRE (50% Sylgard 527, 50% Fe particles, w/w%) and (ii) Sylgard 527 without Fe particles. (B) iPSC-CM cell spreading area on different materials coated with 20 μg/mL FN. No significant change in cell area was measured between the different materials (p = .0728). (C) Representative image of iPSC-CM adhered to MRE substrate, showing the presence of contractile mechanisms (ACTN2, orange). Scale bars: (A) 400 μm, (C) 20 μm. Error bars: SD, n > 190 cells.

3.4 |. MRE photothermal properties

MRE samples, which started at the ambient room temperature of ~22°C, did not show significant amounts of photothermal heating when exposed to the 0.174 W/cm² intensity light used during the in vitro studies whether the surrounding environment was air (Figure 4A) or media (Figure 4B), heating to only about 23.58 ± 0.75°C and 22.42 ± 0.55°C, respectively. When exposing the MRE samples to 2.230 W/cm² intensity light, the surface temperature significantly (p < .0001) increased to 36.28 ± 1.80°C in air and 35.28 ± 2.14°C in media. When testing the surface heating of MRE samples in air with intermediate intensities of light, it was found that with 0.291 W/cm², the lowest intensity tested above 0.174 W/cm², there was a significant amount of surface heating (p = .0023) (Supplemental Figure 5A). In contrast, no temperature increases were measured when exposing pure Sylgard 527 samples to any of the intensities of light, staying at ~22°C. Adding a ~5 μm layer of Sylgard 527 on top of the MRE surface to act as an insulator did not have any significant effect on reducing the photothermal surface heating of the MRE. Although surface heating penetrated through the thin layer of Sylgard 527, it was shown that outside of the light there was significantly less heating of the MRE surface than the area of the MRE near the incident light when exposed to 30 s of 2.230 W/cm² of light in air (Supplemental Figure 5B). As previously mentioned, inside the light at these conditions the surface temperature was measured to be 36.28 ± 1.80°C. In comparison, the surface temperature of the MRE 2 mm outside the area of the light was measured to be 27.04 ± 1.76°C, and 10 mm outside the area of the light the temperature remained at baseline (22.38 ± 0.41°C).

FIGURE 4.

MRE samples had a significant photothermal response to LED light used for epifluorescence imaging. (A) When exposing the MRE samples in air to 2.230 W/cm² LED light (blue, 475 ± 28 nm) for 30 s, the MRE surface temperature rose significantly (p < .0001) from 22.24 ± 0.45°C when there was no light to 36.28 ± 1.80°C. In contrast, when using 0.174 W/cm² there was no significant (p = .1266) change in temperature measuring at 23.58 ± 0.75°C. (B) When instead exposing the MRE samples to the light while in media there was little change in the heating effects with the MRE still significantly (p < .0001) heating from 21.38 ± 0.37°C to 35.28 ± 2.14°C due to 30 s of 2.230 W/cm² of light. There was also still no significant surface heating (p = .2780) when exposed to 30 s 0.174 W/cm² of light with temperatures measured to be 22.42 ± 055°C. (C) The heating profile shows that when using 2.230 W/cm², the surface temperature begins heating very fast, becoming noticeably, albeit not significantly (p = .2442), warmer at 25.50 ± 2.69°C after only 10 s of exposure, and significantly (p = .0048) warmer to 27.80 ± 0.14°C after 20 s. After the entire 120 s of 2.230 W/cm² intensity light the surface temperature reaches 52.15 ± 2.05°C (p < .0001). When using 0.494 W/cm² intensity light, the surface temperature began to become noticeably different around 50 s reaching a temperature of 25.05 ± 0.35°C (p = .8565), although it did not become significantly warmer until 110 s (p = .0067) when the temperature reached 27.75 ± 1.20°C, reaching 28.65 ± 1.63°C after 120 s (p = .0007). In contrast when the MRE was exposed to 0.174 W/cm² intensity light for 120 s there was never any significant heating (p > .9999 at all points) and only became slightly warmer after 120 s reaching a temperature of 24.2 ± 1.70°C. The pure Sylgard 527 did not experience a temperature change within the 120 s when exposed to 2.230 W/cm², staying at 22.50 ± 0.15°C for the duration. (D) The cooling profile shows that the MRE returned back to room temperature after only 10 s following 120 s of exposure to 0.174 W/cm² intensity light. In contrast, when exposed to 0.494 W/cm² or 2.230 W/cm² it took 40 and 80 s, respectively, to return to room temperature.

The effects of different levels of light intensity was further illustrated by the time dependency of the surface heating. With all conditions starting at ~22.5°C, the MRE became noticeably, albeit not statistically significantly (p = .2442), warmer at 25.50 ± 2.69°C after only 10 s of exposure to 2.230 W/cm² of light (Figure 4C). Within 20 s this became a significant difference (p = .0048) with the temperature rising to 27.80 ± 0.14°C. After 120 s of exposure to 2.230 W/cm² of light the surface of the MRE was even more significantly heated (p < .0001) reaching a temperature of 52.15 ± 2.05°C. In comparison, it took 110 s of exposure at light at 0.494 W/cm², to induce a statistically significant change in MRE surface temperature (p = .0067). Importantly, photothermal heating did not cause a statistically significant change in MRE surface temperature with an incident light flux of 0.174 W/cm² intensity light applied for over 120 s (Figure 4D). Following the 120 s of exposure to 0.494 W/cm² and 2.230 W/cm² intensity light it took 40 s and 80 s for the MRE samples to return to room temperature, respectively. As expected, pure Sylgard 517 samples did not experience any heating effects during the 120 s duration while exposed to 2.230 W/cm² intensity light.

3.5 |. Effects of acute increases in substrate stiffness on μHM

When stiffening the substrates which μHM contract against, we observed significant changes in the tissues’ physiology. Immediately upon increasing the magnetic flux density through the MRE substrates, the spontaneous beat rate of the tissues significantly increased (Supplemental Figure 6). However, within tens of seconds of substrate stiffening, the spontaneous beat rate had returned to baseline, and remained constant for the remainder of the study. There were large variations in the magnitude of this instantaneous beat rate increase, as shown by the four example changes in spontaneous beat rates. Due to the rapid time course and large variability of the spontaneous beat rate adaption period, we subsequently focused on the dynamic response of μHM to stiffness increase at longer timepoints (e.g., ≥5 min after applying the magnetic field).

Over a 2-h time period, we observed significant shifts in Ca²⁺ dynamics (Figure 5). For example, although there was no change within 5 min of stiffening the substrates, integrated Ca²⁺ flux (Figure 5A), background corrected peak Ca²⁺ amplitude (Figure 5B), and Ca²⁺ τ₇₅ (Figure 5C), all exhibited significant decreases over a 2-h period. These changes in the Ca²⁺ dynamics were sustained for as long as the substrate was in the stiffened state. Alternatively, the Ca²⁺ UPD stayed constant within the first 2 h when acutely increasing the substrate stiffness before diverging from the control tissues (Figure 5D). However, this change was largely due to changes in the UPD in a subset of the control samples rather than changes in the experimental samples, which stayed near the baseline value for all time points. There were no changes in any of the Ca²⁺ dynamics for the control samples, other than the subset that showed increased UPD at the later time points.

FIGURE 5.

μHM response to acute increases in substrate stiffness over a 48-h time course. (A) Integrated calcium flux, (B) Calcium amplitude, and (C) Calcium Decay 75 all decreased within 2 h of increasing substrate stiffness and remained significantly decreased compared to the control samples, which did not significantly shift from baseline, for as long as the substrate stayed stiffened (p < .002 for all three behaviors at the 48-h time point). (D) Calcium upstroke duration did not significantly change in the experimental samples when increasing substrate stiffness. Significant difference was seen between the experimental and control samples after 24 h (p = .0045) and 48 h (p < .0001) due to an increase in the upstroke duration of a subset of the control samples. (E) Bead displacements did not significantly change over the 48-h time course as a function of either time or application of the magnetic field (p = .3417, 2-way ANOVA). (F) The traction forces increased significantly within seconds-to-minutes of stiffening the substrates, but significance compared to the control group cannot be reported due to lack of control data at the 5-min time point. The contractile force remained elevated in the experimental group for the remaining 48 h that the substrate was stiffened and remained significantly larger than the contractile forces generated by the control tissues (p < .0001 at 48-h time point). ****, ***, ** and * indicates significant difference p < .0001, .001, .01 and .05, respectively. Error bars: SD.

These changes in calcium handling were concurrent with changes in the contraction forces and contraction kinetics of the μHM. Strikingly, μHM contraction induced displacements of the substrates stayed constant at the baseline value with no significant change (Figure 5E). There was also no significant change in substrate displacements caused by contraction forces of the control tissues. As such, because twitch-force induced contractile displacement of the substrates did not change while substrate stiffness increased (experimental condition), the μHM exhibited marked increases in their peak contraction forces (Figure 5F). This was concurrent with a decrease in total contraction time (Supplemental Figure 7).

Interestingly, while peak contraction-induced substrate deflection did not change, motion of the tissue edge transiently increased (at the 5-min time-point) after the magnet was applied to stiffen the MRE (Supplemental Figure 8). However, by 2 h after applying the magnet, peak displacement of the free tissue edge decreased back down to the level observed in the substrates before the magnet had been applied. Like the transient increases in Ca²⁺ τ₇₅ and upstroke duration (Figure 5), this may reflect that immediately upon experiencing after-load, μHM adapt by increasing calcium intake to trigger an overall larger contraction motion, without all of the motion going toward mechanically productive work (e.g., substrate deflection). Addressing this short-term adaptation will be an interesting topic for future studies.

3.6 |. Reversibility of changes in μHM physiology induced by acute changes in substrate stiffness

To determine whether μHM reacted to in situ softening of the substrate, we continued to image a subset of μHM on MRE substrates after removing the magnets following the first 48-h magnetization. In μHM where substrate stiffness was returned back to the softened state, the changes in the calcium dynamics originally induced by stiffening generally rebounded rapidly to their pre-stiffened baseline values within 5 min (Figure 6A), illustrated by the changes in the Ca²⁺ τ₇₅ (Figure 6B), although there was variability as some μHM took up to 24 h to return to baseline. Additionally, displacements (Figure 6C) and traction forces (Figure 6D) reverted to values similar to the baseline values within minutes and were no longer significantly different from the control values at the same time points. However, while the integrated calcium increased back toward the baseline value, as shown by the traces in Figure 6A, it did not reach the baseline value within the 48-h time course, except for an initial rebound in the first 5 min of softening the substrate when there was a large amount of variation in the integrated calcium measurements (Supplemental Figure 9).

FIGURE 6.

The μHM response to heighted substrate stiffness for 48 h can be reversed by reverting the substrate stiffness to the softened state. (A) Example calcium transients of a tissue at baseline, 2 h, and 48 h after stiffening the substrate, and 2 and 48 h after the substrate is returned to the softened state. The width and amplitude of the transients decreased upon increasing the substrate stiffness. When returning the substrate stiffness to the original softened state, the transient width return to the pre-stiffened baseline width within 2 h, while the transient amplitude increased back toward the baseline value but did not reach the baseline value within the 48-h time course. (B) Calcium decay initially (≤5 min) went above the baseline value when returning the substrate to the softened state before remaining at the pre-stiffened baseline value. After 48 h of the substrate being returned to the softened state, there was a slightly-significant difference between the experimental and control samples due to an increase in the decay time in the control samples (p = .0189). (C) Peak traction forces returned to the pre-stiffened baseline values within 5 min of softening the substrate and remained at baseline 48 h later with no significant difference in comparison to the control traction forces (p = .9466). (D) Bead displacements did not significantly change over the 48-h time course as a function of either time or application of the magnetic field (p = .4083, 2-way ANOVA). **** and * indicates significant difference p < .0001 and .05, respectively. Error bars: SD.

4 |. DISCUSSION

We created a system to dynamically control μHM contractile resistance via MRE substrate elasticity to simulate afterload changes in vitro. The custom plate lid and magnet holders enabled the substrate stiffness to be manipulated during the in vitro experiments by changing the distance of the magnet to the MRE substrate. In future iterations, a linear actuator could be applied to the system to allow slower and more controllable changes of substrate stiffness, as has been done in similar MRE systems.⁵¹ Additionally, previous uses of MRE substrates for studies of cardiac mechanobiology required the use of an upright fluorescence microscope.^32,39,40 Our system, with the use of an inversion frame with controlled thickness, allowed for the use of inverted epifluorescence imaging; this in turn offers the benefit of shorter working distance between the objectives and the sample, and eliminates a potential need for using dipping lenses for high-resolution imaging. However, because the substrate is opaque, brightfield imaging is not feasible, making applications like automated tissue motion tracking and strain calculations difficult. Techniques like fluorescent membrane labeling such that the tissue intensity remains constant could be used in combination with computational analyses like the one described by Das et al. and/or the “virtual blebbistatin” approach recently described by our group to perform high-throughput measurements of tissue motion and contractile strains.^53,54

At low strains, Sylgard 527 has previously been shown to be linear elastic.⁵⁵ Additionally, the rheological testing (1 rad/s = 0.1592 Hz) and the spontaneously beating μHM (≤1 Hz) both occurred at low frequencies, over which there is little frequency dependence in the shear modulus of Sylgard 527.⁵⁶ As such, the dynamic and static modulus should be similar. This expectation is consistent with our results, which suggested no significant difference between the shear modulus measured using this platform and shear storage modulus measured using the rheometer. Because we tested dynamic rheology at ~0.15 Hz, we performed our static shear measurement within a few seconds of applying the load. However, long-term creep behavior of the PDMS may be of interest, which could be tested with timelapse imaging of samples in our testing setup in the future.

Both Sylgard 527 mechanical properties and the magnetic response of MRE are known to have temperature dependence.^57,58 However, we observed that an additional short exposure to 121°C for 10 min while autoclaving did not affect the baseline MRE stiffness or the magneto-responsiveness of our MRE. This is likely because of the short duration of this heating. However, consistent with these previous findings, prolonged exposure to higher temperatures (200°C for 24 h) both increased the baseline stiffness and reduced the magneto-responsiveness of the MRE materials.

Using the validated custom shear system, we were able to measure a significant change in the shear modulus from 0.59 up to 5.26 kPa when increasing the magnetic flux from 0 to ~180 mT. When converted into Young’s modulus the stiffness was calculated to increase from 1.76 to 15.77 kPa, a range at which physiological changes have been shown to be visible in elastomer grafted μHM on substrates of discrete stiffnesses.^28,29 Thus, it was determined that we had fabricated substrates with stiffnesses that could be dynamically controlled to induce physiological changes in μHM. Although these stiffness changes were shown to be sufficient to illustrate this system’s ability to cause changes in the tissue response, larger shifts in modulus will be required to mimic the transition from fetal, to physiologic, to pathologic tissue stiffness.⁵⁹ Recent work by Clark et al. provides a framework for achieving this stiffness range using MRE materials.⁴¹

Due to the photothermal response of the opaque Fe-doped materials, imaging parameters had to be carefully monitored. Beating frequency of iPSC-CMs have previously been shown to increase when environmental media temperature increase, up to around 42°C before beating quickly becomes too faint to be analyzed, signaling possible decreases in contractility and calcium flux.⁶⁰ However, the response seen by Laurila et al. when heating the media was mimicked using local photothermal heating of iPSC-CMs treated with polydopamine nanoparticles, a material with a large photothermal response.⁴⁶ In these studies, the authors found that low light intensities increased spontaneous beat rate, but high light intensities led beating to stop entirely before slowly recovering when removed from the light.

The phenomenon seen by Laurila et al. and Gholami Derami et al. was observed during the studies described in this paper when exposing the μHM on MRE substrates to high intensities (>0.3 W/cm²) of blue light (data not shown). In contrast, we observed no such effects when we minimized the energy of LED light (≤0.2 W/cm²). Future iterations of this approach should use minimal light intensity, and ideally, higher wavelengths of light. Further, imaging on a single area should be kept to short sessions. Different areas of the same sample can be safely sampled continuously as the heat does not appear to be particularly radiative, as shown by the reduction of temperature 2 mm and 10 mm outside the area of the light. However, our results also show that layers of PDMS that are thin enough (~20 μm⁶¹) to allow cells to sense stiffness changes in the underlying MRE are not sufficient to insulate the surface of the composite material from photothermal heating.

In the system described in this study, changes in calcium and contractile dynamics showed that μHM could sense the changes in substrate stiffness when altering the magnetic field through MRE substrates. The non-sustained change in spontaneous beat rate over the first few seconds of stiffening the substrate may stem from rapid adaptive mechanisms such as force-responsive ion channels.⁶² In comparison, the sustained changes in the calcium dynamics that started to occur between 5 min and 2 h after substrate stiffening are likely due to other mechanisms, such as changes in the phosphorylation of proteins like phospholamban, which regulates the cardiac cycle; calcium contraction coupling; and/or transcriptional changes.^63,64 These changes returned to their pre-stiffened baseline values over a similar time course when returning the substrates to the original softened state, suggesting against “mechanical memory” effects that may require longer times to be spent at a stiffened state, but further studies of the biomolecular changes, through the use of techniques like qPCR and other assays, are required to make more definitive conclusions.⁶⁵

Counterintuitive to the notion that calcium intake is proportional to cardiac twitch force,⁶⁶ while the calcium intake decreased when stiffening the substrates, the contractile force increased. One possible explanation for this divergence could be mechano-sensing thick filaments, which have been shown to be able to regulate force generation independently of calcium by increasing the recruitment of myosin motors.^67–69 Future experiments should target the cause of this phenomena, such as through the use of calcium sensitizers and/or calcium-contraction uncoupling drugs. Regardless of the mechanism of force generation, the increase or decrease of μHM contractile force within seconds-to-minutes of increased or decreased contractile resistance is consistent with original observations described by von Anrep, which have been reproduced in numerous ex vivo studies.^8,70 Interestingly, despite these changes in calcium dynamics and contractile forces, the μHM kept near constant displacements when increasing the substrate stiffness. However, 24 h after returning the substrates to a softened state, the displacements exhibited a trend toward being increased compared to the control group, albeit not statistically significant. This potentially points to maturation from mechanical conditioning. The inability to recapture the original calcium flux following the return of the substrate to its original stiffness could similarly be a sign of maturation and increased force-calcium efficiency. Future studies may identify the molecular underpinnings of these phenomena.

In addition to investigating the molecular mechanisms for mechano-adaptation, future studies could also center around rapidly changing stiffness on the time-scale of seconds, so that afterload increases are only experienced by the tissues during systole, as occurs in vivo. Likewise, future studies could assess the impact of genomic variants associated with cardiomyopathies like HCM on mechano-adaptation.^71,72 HCM patients who have hypertension suffer from increased mortality.^73,74 Consistent with the notion that this relates to altered mechano-sensation and/or mechano-adaptation in HCM, mice harboring an HCM genotype that reach adulthood without treatment do not respond to blood pressure reducing agents, whereas wild type mice suffering from surgically-induced hypertension show a reversion of hypertrophic remodeling when receiving the same drugs.^75,76 Using the system described here would allow for the investigation of how changes in contractile resistance differentially affect these genetic variations.

5 |. CONCLUSION

We have developed an MRE substrate based μHM system to simulate cardiac afterload in vitro. The MRE fabricated was shown to be able to dynamically control stiffness, with shear moduli values that could be modulated from 0.59 kPa to at least 5.25 kPa. When culturing μHM on these MRE substrates, acute, yet reversible, changes in calcium dynamics and contractile forces were elicited. This platform may be useful in accelerating iPSC cardiomyocyte maturation via mechanical conditioning. Additionally, the use of this system in chronic studies will allow for careful examination of the mechano-induced signals and whether cardiomyocyte genotype affects mechanical adaptation.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are openly available in Digital Commons@Becker at https://digitalcommons.wustl.edu/data/.