Physiologic biomechanics enhance reproducible contractile development in a stem cell derived cardiac muscle platform

Figures

• 

  Fig. 1. Micron-scale two-dimensional cardiac muscle bundles (2DMBs) exhibit enhanced myofibrillar organization. a iPSCs were differentiated through wnt modulation (CHIR, IWP4) with retinoic signaling inhibition (RI) on days 3–7. During lactate purification, CDM3 was formulated with RPMI without glucose and without glutamine to improve purification. b Day-25 hPSC-CMs were replated to 8 kPa PDMS micropatterned with fibronectin in 7:1 aspect ratio, 308 × 45 μm rectangles, each separated by a buffering space to minimize mechanical interactions among adjacent 2DMBs. The overall pipeline improves efficiency by requiring only 45 k hPSC-CMs per substrate, each yielding many high-quality, matured 2DMBs per device for structure and function studies. The approach reduces the overall time required for contractile and myofibrillar structural analyses, including both imaging and running automated image-processing algorithms. c hPSC-CMs avidly adhere and conform to PDMS micropatterns but not to surrounding PDMS to generate 2DMBs (image from brightfield microscopy with ×40 objective). d Myofibrils (marked by MyBP-C) develop along the 2DMB long-axis traversing continuously across cell junctions (marked…

• 

  Fig. 2. Mechanical uncoupling of adjacent micron-scale two-dimensional cardiac muscle bundles (2DMBs) improves contractile function and quantification reproducibility. a Modeling analysis demonstrated that 2DMBs are mechanically uncoupled from adjacent 2DMBs (see also Supplemental Fig. 3). 3D representation of an 2DMB on 70-μm thick PDMS at 11% fractional shortening shown with heatmap depicting local displacements. Pink lines demarcate the inner 50% length of the 2DMBs where regional muscle bundle deformations are most stable for accurate measurements of fractional shortening. b Brightfield image of 2DMB with regions of interest (ROIs) annotated for displacement tracking using cross-correlation in MATLAB with the ContractQuant algorithm. c Representative output from ContractQuant for an 2DMB showing (top) concurrent normalized displacements and velocities for the full time-series capture (second row) fractional shortening of merged contractions, (third row) merged absolute values of contractile velocity, and (fourth row) absolute values for acceleration. In rows 2–4, the solid line indicates the mean and shaded regions the 95% confidence intervals. d Representative output from…

• 

  Fig. 3. Maturation of calcium handling, electrophysiology, and gap junction development is dependent on the biomechanical and contractile environment. a RNA expression of ion channel and calcium handling genes by RNA-seq in micron-scale two-dimensional cardiac muscle bundles (2DMBs) was compared to standard hPSC-CMs using normalized RNA-seq reads (data points from three biologic replicates per condition from five pooled batches of iPSC-CMs, two-sided P value shown from unpaired t test; mean ± standard deviation shown by error bars). b Representative action potential from an 2DMB by patch clamping. c–f Patch clamping of 2DMBs showed reproducible resting membrane potentials (c), action potential amplitude (d), action potential duration (e), and maximum dV/dt during phase 0 depolarization (f). Resting membrane potentials were measured from 18 2DMBs from five batches and triggered action potentials were recorded for 11 2DMBs from four batches (mean ± standard deviation shown by error bars except in (e) where the standard error of the mean is shown). g Resting fura-2 ratios were lower in 2DMBs with reduced variability across samples (N = 56 from three batches for each condition; two-sided P…

• 

  Fig. 4. Micron-scale two-dimensional cardiac muscle bundles (2DMBs) increase the precision of pharmacologic testing in hPSC-CMs. a, b Maximal fractional shortening with mavacamten in standard hPSC-CMs (a) and 2DMBs (b) on 8 kPa PDMS (N = 22 and N = 39, respectively). c, d Maximal fractional shortening with verapamil in standard hPSC-CMs (c) and 2DMBs (d) on 8 kPa PDMS (N = 29 and N = 37, respectively). e Concentration-response of increasing calcium concentration on maximal fractional shortening in 2DMBs (N = 19). f Maximal fractional shortening with mavacamten in 2DMBs on 8 kPa PDMS from a different iPSC line (Allen Institute AICS-0053 cl.16, N = 22). g, h Calcium-transient amplitudes with verapamil in standard hPSC-CMs (g, N = 48) and 2DMBs (h, N = 55). P values are shown for paired t tests for consecutive time points for all experiments. The mean ± standard deviation is shown by error bars for all experiments. Source data are provided as a Source Data file.

• 

  Fig. 5. Physiologic biomechanics drive transcriptional development in hPSC-CMs. a Relative expression (log-2 fold change) of genes associated with maturation is shown by heatmap across 2D maturation conditions. Genes expected to be downregulated during maturation are indicated with a downward yellow arrowhead at the right of each row. b Gene ontology analysis of maturation conditions versus standard monolayer hPSC-CMs demonstrated overrepresentation of differentially expressed genes among several summative categories: (1) tissue connectivity (e.g., multicellular organism development, extracellular matrix organization, response to mechanical stimulus, cell–cell junction organization shown left–right), (2) cardiac development (e.g., cardiac muscle cell differentiation, actin cytoskeleton organization, (3) cardiac muscle hypertrophy, cardiac ventricle development shown left–right), (4) electrophysiology (e.g., regulation of membrane potential, cardiac muscle cell action potential, membrane repolarization, cellular calcium ion homeostasis shown left–right), (5) mitochondrial respiration (e.g., cellular respiration, oxidative phosphorylation, mitochondrion organization, respiratory electron…

• 

  Fig. 6. Metabolic and hormonal inputs increase myofilament maturation and adrenergic response in micron-scale 2D cardiac muscle bundles (2DMBs) distinct from the biomechanical environment. a In standard hPSC-CM monolayers, switching to a media that requires oxidative phosphorylation (OxPhos) induces expression of cTnI compared to ssTnI (measured by mass spectrometry, each data point indicates the mean of N = 6 biologic replicates at each time point, two each for three different hPSC-CM lines) beginning in the first 24 h after media change (mean ± standard deviation shown by error bars). b Relative TNNI3 mRNA expression (as a proportion to total of TNNI3 + TNNI1 mRNA) is increased in standard hPSC-CMs in OxPhos media synergistically with T3 after 5 days (N = 6, triplicates for two different hPSC-CM lines). c Representative images of cTnI expression in 2DMBs in regular (top) and OxPhosT3 media (bottom). More hPSC-CMs in regular media exhibited minimal myofibrillar expression of cTnI (lower cells in top image). d Quantitative immunofluorescence of cTnI in hPSC-CMs in 2DMBs in regular glucose-containing media or OxPhosT3 media (N = 197 and N = 125, respectively). e Contractile quantification…

• 

  Fig. 7. Contraction force and calcium concentration are critical for myofilament and calcium handling development in 2DMBs. a qRT-PCR analysis of micron-scale two-dimensional cardiac muscle bundles (2DMBs) cultured for 8 days with or without the cardiac myosin-specific contractile antagonist, mavacamten (500 nM), demonstrates greater myofilament growth-related gene expression in the control condition. In contrast, genes associated with electrophysiologic maturation are unchanged. Each biologic replicate was obtained from 2 2DMB substrates comprised of >8000 total 2DMBs (DF19-9-11 line). *P < 0.05 for two-sided t test (HOPX P = 0.01, MYH7 P = 0.02, MYL2 P = 0.01, SCN5A P = 0.32, KCNJ2 P = 0.75, GJA1 P = 0.44). b Immunofluorescence quantification of MLC-2v confirmed increased protein abundance in the absence of mavacamten (N = 14 control; N = 17 mavacamten). c Representative immunofluorescence images of 2DMBs cultured with and without mavacamten for 8 days stained for alpha-actinin and MLC-2v (DF19-9-11 line). d 2DMBs cultured for 8 days in near-physiologic calcium (1.45 mM, N = 17) exhibited higher alpha-actinin abundance by immunofluorescence quantification than either 2DMBs in standard…

• 

  Fig. 8. Controlling the biomechanical environment improves the precision of contractile kinetic investigation for cardiomyopathy disease modeling. a Representative example of a cardiomyopathy 2DMB (MYBPC3pr−/−) showing normally aligned myofibrils (labeled with SiR-actin). b Maximal fractional shortening in control vs. MYBPC3pr−/− 2DMBs (N = 209 and N = 42). c–e Normalized contraction velocity in control vs. MYBPC3pr−/− 2DMBs at early (20% of max shortening), mid (50% of max shortening), and late (80% of max shortening) contraction. f Contraction deceleration time in control vs. MYBPC3pr−/− 2DMBs (defined as time from deviation from linear contractile slope to peak contraction). g–i Normalized relaxation velocity in control vs. MYBPC3pr−/− 2DMBs at early (80% of max shortening), mid (50% of max shortening), and late (20% of max shortening) relaxation. j Summary of relative differences in contraction and relaxation velocities in MYBPC3pr−/− 2DMBs shown on a representative fractional shortening curve from an MYBPC3pr−/− 2DMB. The mean ± standard deviation is shown by error bars and statistical significance tested using two-sided t tests for all graphs. Source data are provided as a Source…