Substrate stiffness affects the functional maturation of neonatal rat ventricular myocytes

Abstract

Cardiac cells mature in the first postnatal week, concurrent with altered extracellular mechanical properties. To investigate the effects of extracellular stiffness on cardiomyocyte maturation, we plated neonatal rat ventricular myocytes for 7 days on collagen-coated polyacrylamide gels with varying elastic moduli. Cells on 10 kPa substrates developed aligned sarcomeres, whereas cells on stiffer substrates had unaligned sarcomeres and stress fibers, which are not observed in vivo. We found that cells generated greater mechanical force on gels with stiffness similar to the native myocardium, 10 kPa, than on stiffer or softer substrates. Cardiomyocytes on 10 kPa gels also had the largest calcium transients, sarcoplasmic calcium stores, and sarcoplasmic/endoplasmic reticular calcium ATPase2a expression, but no difference in contractile protein. We hypothesized that inhibition of stress fiber formation might allow myocyte maturation on stiffer substrates. Treatment of maturing cardiomyocytes with hydroxyfasudil, an inhibitor of RhoA kinase and stress fiber-formation, resulted in enhanced force generation on the stiffest gels. We conclude that extracellular stiffness near that of native myocardium significantly enhances neonatal rat ventricular myocytes maturation. Deviations from ideal stiffness result in lower expression of sarcoplasmic/endoplasmic reticular calcium ATPase, less stored calcium, smaller calcium transients, and lower force. On very stiff substrates, this adaptation seems to involve RhoA kinase.

FIGURE 1

The Young's modulus of elasticity of polyacrylamide gels varies over a physiologic range of 1–50 kPa with varying monomer from 2% to 7% and a 1:20 monomer/cross-linker ratio.

FIGURE 2

NRVMs on polyacrylamide gels and labeled for α-actin have poorly defined striations on soft 1 kPa substrates (A), well defined and aligned striations on 10 kPa substrates (B), and unaligned striations with long, large stress fibers on stiff 50 kPa gels (C). NRVMs plated on polyacrylamide gels and labeled with phalloidin (green) and DiI (red) show an axially-aligned cytoskeleton throughout the cell on 1 kPa (D) and 10 kPa (E) gels but F-actin concentrated on the periphery and nucleus and no axial alignment on 50 kPa gels (F). Zoomed-in confocal images of NRVMs on 10 kPa gels (G) and 50 kPa gels (H) better show differences in sarcomeric structure and alignment. Scale bars = 10 μm.

FIGURE 3

Cell area (A) and aspect ratio (B) are not significantly different across the gel stiffnesses (ANOVA, Area p = .13). Circularity index (C) is significantly different between cells on 10 kPa and cells on 50 kPa gels (p < 0.05). Data points represent 15 cells/point. Error bars represent SE.

FIGURE 4

Traction force microscopy provides a measurement of cell contractile force without disturbing cell attachment. Cells were stimulated at 0.5 Hz for 0.8 ms and the movement of fluorescent beads embedded in the substrate was imaged. Bead displacement between images was tracked using a cross-correlation algorithm (A). Scale bar = 10 μm, scale arrow is 10 μm displacement. These displacements were converted to shear stresses on the gel surface (B). Scale bar = 10 μm, scale arrow is 1 kPa (1000 nN/μm²). The stresses were projected onto the long and short axis of the cell and integrated over the cell area to calculate an axial and longitudinal force over time (C).

FIGURE 5

Axial contraction force (A) peaks in cells on 10 kPa gels and decreases in cells on stiffer or softer gels. The contraction force was calculated through dynamic traction force microscopy. Contraction force is significantly different across all gels per ANOVA and force was significantly greater on 10 kPa gels compared to 1 kPa and 5 kPa gels (p < 0.05). The velocity of shortening of the major axis (B) trends downward as stiffness increases though individual velocities are not significantly different per ANOVA. The resting axial force increases as the substrate stiffness increase (C) though individual points are not different per ANOVA. All points are averages of 15 cells. Error bars represent SE.

FIGURE 6

The relationship between measures of calcium handling in NRVMs and substrate stiffness mirrors the force relationship. The magnitude of calcium transients (A), measured as peak fluorescence divided by baseline fluorescence, in Fura-2 or Fluo-4 labeled NRVMs on 10 kPa gels was significantly greater than transients on 1 kPa and 50 kPa gels (p < 0.05) (A). n = 6, 7, 10, 7, and 9 for 1, 5, 10, 25, and 50 kPa substrates, respectively. The magnitude of sarcoplasmic calcium stores (B), measured as the plateau of Fluo-4 fluorescence after stimulation with caffeine divided by baseline fluorescence, was significantly different per ANOVA (p < 0.05) and the calcium release of the cells cultured on 50 kPa gels was significantly lower than on 1 kPa and 10 kPa gels (p < 0.05). n = 13, 8, 12, 4, and 4 for 1, 5, 10, 25, and 50 kPa substrates, respectively. Representative traces of calcium transients (C) and of caffeine-induced calcium release transients (D) are also shown.

FIGURE 7

Results of Western blots of cell lysates for cardiac α-actin and MHC (A) show no differences in expression (B and C). Amounts of α-actin are significantly different per ANOVA and cells cultured on 10 kPa gels have significantly more actin than 0-day controls. Amount of myosin is greater in all cases but individual results were not significant. Western blot for SERCA2a (D) shows differences in expression that mirrors sarcoplasmic calcium stores, calcium transients, and force (E). Expression of SERCA2a is significantly different across gels per ANOVA (p < 0.05) though individual differences are not significant. Data presented are the average of protein analysis from four separate lysates, normalized to the average expression of the protein from day-0 unplated NRVMs.

FIGURE 8

Inhibition of ROCK with hydroxyfasudil (A) significantly inhibits the decrease in force observed by NRVMs on stiff gels of 25 kPa and 50 kPa elastic modulus (p < 0.05) and force continually increases with increasing stiffness in fasudil-treated cells. Similarly, inhibition of RhoA with C3 toxin (B) also inhibits the force reduction on the stiffest gels (p < 0.05). Results are normalized to average for control cells at 10 kPa elastic modulus. These results suggest that the adaptation of NRVMs to stiff gels requires the RhoA/ROCK pathway. In addition, force increases directly with elastic modulus of the gel. Fasudil-treated n = 8, 9, 9, 10, and 7 for 1, 5, 10, 25, and 50 kPa substrates, respectively. Control n = 5 in all cases. Error bars represent SE.