Full-length dystrophin deficiency leads to contractile and calcium transient defects in human engineered heart tissues

Stem cell culture and directed cardiac differentiation

A urine-derived hiPSC line from a healthy male donor (control) was established as previously described. ²² An isogenic diseased cell line harboring a dystrophin mutation (DMD 263delG) was generated from the control cell line via the deletion of a single base pair in exon 1 of the DMD gene with CRISPR-Cas9 as previously described. ³¹ Undifferentiated hiPSCs were maintained in mTeSR1 (Stemcell Technologies) on tissue culture plates coated with a 1:60 dilution of Matrigel (Corning). Cardiomyocyte directed differentiation was performed using a modified small molecule Wnt-modulating protocol optimized to each cell line as previously described. ⁴¹ Briefly, hiPSCs were seeded at 1.5–2.5 × 10⁵ cells/cm² on Matrigel-coated plates in mTeSR1 with 10 µM Y-27632 ROCK inhibitor (Tocris). Cells were maintained in mTeSR1 with daily media changes until they reached confluency, which was typically 2–3 days after initial seeding. The initiation of directed differentiation (day 0) was defined by changing the media to RPMI 1640 (Gibco) plus B-27 supplement minus insulin (Life Technologies) and 12–14 µM Chiron 99021 (Axon Medchem) to activate Wnt/β-catenin signaling via inhibition of glycogen synthase kinase-3β (GSK-3β). After 24 h (day 1), media was changed to RPMI with B-27 minus insulin. On day 3, media was changed to RPMI with B-27 minus insulin and 5 µM IWP-4 (Stemgent) to inhibit Wnt/β-catenin signaling. On day 5, media was changed to RPMI with B-27 minus insulin. On day 7, the media was changed to RPMI with B-27 supplement (with insulin) and media was subsequently changed every 2–3 days. All differentiation steps were performed in 12-well plates with media volumes of 2 mL per well. hiPSC-CMs were replated into 10 cm dishes on day 14 to prepare for lactate enrichment. On days 16 and 18, media was changed to DMEM without glucose (Gibco) supplemented with 4 mM sodium L-lactate (Sigma-Aldrich) to enrich for cardiomyocytes. Cells were returned to RPMI with B-27 supplement plus insulin and cultured until day 25–27. Prior to EHT casting, cardiomyocyte purity (> 90%) was evaluated by flow cytometry for cardiac troponin T (Supplemental Figure S1). All cell culture media was supplemented with 100 U/mL penicillin-streptomycin.

EHT platform and tissue casting

The EHT post platform was designed and fabricated as described previously, resulting in racks of six pairs of posts measuring 12 mm long and 1.5 mm in diameter with a cap.^32–34 Briefly, uncured polydimethylsiloxane (PDMS, Sylgard 184 mixed at a 10:1 ratio of base to curing agent) was poured into a four-part acrylic mold with a glass capillary tube (Drummond, Cat #1-000-0500) inserted into one post of each pair to make it rigid. Posts were left to cure at room temperature for 24 h then baked at 65 °C overnight to cure completely before being separated from the mold. Prior to EHT casting, PDMS posts were submerged in 70% ethanol for 10 min, rinsed with sterile deionized water for 10 min, then UV sterilized for 10 min. The PDMS had a modulus of elasticity of 2.5 MPa, so the bending stiffness of the flexible post in each pair was calculated to be 0.95 µN/µm, as previously described.^32,42

EHTs were cast on day 25–27 of cardiac differentiation as previously described.^33,34 Briefly, rectangular 2% agarose casting molds were prepared in 24-well tissue culture plates using 3D-printed spacers. Each rack of PDMS posts was positioned upside down with the tips of posts centered in the agarose wells. A solution of 100 µL volume consisting of 7.5 × 10⁵ hiPSC-CMs and 5 × 10⁴ HS27a human bone marrow stromal cells (ATCC) in RPMI with B-27 with 5 mg/mL bovine fibrinogen (Sigma-Aldrich) and 3 U/mL thrombin (Sigma-Aldrich) was pipetted into each agarose well. The mixture was incubated at 37 °C for 80 min to form EHTs suspended between pairs of posts and then transferred into a new 24-well plate with EHT media (RPMI with B-27 supplemented with 5 mg/mL aminocaproic acid, Sigma-Aldrich) for culture. EHT media was exchanged every 2–3 days for 3 weeks before subsequent analysis.

Contractile force measurement

EHTs were placed in a Tyrode’s buffer (1.8 mM CaCl₂, 1 mM MgCl₂, 5.4 mM KCl, 140 mM NaCl, 0.33 mM NaH₂PO₄, 5 mM glucose, pH 7.35) at 37 °C for contractile analysis. A custom pacing apparatus with carbon electrodes built to fit a 24-well plate was used with an electrical stimulator (Astro Med Grass Stimulator, Model S88X) to provide biphasic field stimulation at 1.5 Hz (5 V/cm for 10 ms duration) during imaging. ³³ Videos of EHT contraction were taken at 66.7 frames/s using an ORCA-Flash4.0 C13440 CMOS camera (Hamamatsu) on a Nikon TEi epi-fluorescent microscope with a 2× objective and 0.7× coupler, providing 4.64 µm/pixel resolution and a field of view of 9.5 mm × 9.5 mm, which was sufficient to visualize the full length of the EHTs and the tips of both posts. A custom MATLAB script was used to track the deflection of the flexible post relative to the rigid post and from this analysis, we calculated twitch force, shortening velocity, time to peak, time to 50% and 90% relaxation, active twitch power, and total twitch work, as previously described.^42,43 Briefly, twitch force was calculated by subtracting the tissue force minima from the following maxima and contraction kinetic metrics were calculated from the unnormalized tissue force traces. Shortening velocity was calculated as the maximum derivative of tissue force of each EHT contraction. EHT cross-sectional area was calculated assuming an elliptical cross section with a measured tissue width and conservatively estimated thickness of 500 µm. Specific force was then calculated by dividing the twitch force of each EHT by its cross-sectional area. For each EHT, analysis was performed on five consecutive contractions.

Beat interval irregularity measurement

Before initiation of electrical pacing, video recordings of spontaneously beating EHTs were taken to assess beat rate variability as previously described.^23,44 Briefly, EHTs were submerged in Tyrode’s buffer at 37 °C and a video of the flexible post was recorded for 60 s at 10 frames/s during spontaneous contraction. Post deflection was tracked and beat intervals (BI) were measured using a custom MATLAB script. BI were identified using a noise tolerant, peak finding algorithm that identifies local maxima that are greater than a threshold of one-fourth the amplitude of the data above the minimum. BI irregularity (ΔBI) was calculated as the difference between subsequent BI (ΔBI = BI_n+1 − BI_n).

Ca²⁺ transient measurement

Ca²⁺ transients in EHTs were assessed using the ratiometric Ca²⁺ indicator dye Fura-2 AM (Invitrogen), under 1.5 Hz electrical stimulation in Tyrode’s buffer at 37 °C. To load Fura-2, EHTs were incubated in EHT media with 5 µM Fura-2 AM for 1 h followed by a washout period of 30 min in EHT media before being transferred to Tyrode’s buffer. Two custom filter cubes (Chroma) were used to excite the dye at 365 nm or 380 nm and capture its emission at 510 nm. Excitation at 365 nm is near the isosbestic point of Fura-2 (360 nm) while emission intensity with 380 nm excitation decreases with increasing Ca²⁺ levels. Videos were taken at 50 frames/s on an ORCA-Flash4.0 C13440 CMOS camera (Hamamatsu) on a Nikon TEi epi-fluorescent microscope with a 2× objective and 0.7× coupler as described. To enable ratiometric Ca²⁺ assessment, a series of three videos of EHT contraction were taken sequentially at 365 nm, 380 nm, and again at 365 nm excitation. These videos were then analyzed with a custom MATLAB script to track the position of the flexible post and determine the average pixel intensity within the EHT area during contraction. The fluorescence intensity of the two 365 nm excitation recordings at the start and end of the experiments were averaged to account for any effects of photobleaching of the dye. The average was then used in calculating a ratio of baseline fluorescence intensity at 365 nm excitation over the dynamic fluorescence intensity at 380 nm excitation due to changes in the concentration of intracellular Ca²⁺. The 365 nm/380 nm ratio was then used to calculate the kinetics of the Ca²⁺ transient, including time to peak, time to 50% decay, and rate to 50% decay. For each EHT, analysis was performed on 5 consecutive contractions.

Immunoblot

Protein lysates were obtained from a subset of individual EHTs after force and Ca²⁺ measurements using an ice-cold RIPA buffer supplemented with 2% protease inhibitor cocktail (Sigma P8340). Individual EHTs were homogenized in 75–100 µL of buffer using a micro-homogenizer with a 5 mm probe (VWR). Samples were lysed on ice for 30 min, then spun down at 21,000g for 10 min at 4 °C to remove cellular and matrix debris. The supernatant was isolated and protein concentration was measured using the Pierce BCA Protein Assay Kit (Thermo Fisher Scientific) according to manufacturer’s instructions.

Samples were prepared for electrophoresis by adding 4× Laemmli Sample Buffer (Bio-Rad) and 2.5% β-mercaptoethanol, after which samples were denatured at 95 °C for 10 min. Protein was loaded into 4–15% Mini-PROTEAN TGX Stain-Free Gel at 50 µg per sample and run at 100 V for 80 min in 1× Tris/Glycine/SDS running buffer (Bio-Rad). Protein gels were transferred onto Immun-Blot LF PVDF membranes (Bio-Rad) overnight at 30 V and 4 °C in 1× Tris/Glycine buffer (Bio-Rad) with 10% methanol. Membranes were blocked in Blocker FL Fluorescent Blocking Buffer (Thermo Fisher Scientific) for 1 h at room temperature. Primary antibodies against dystrophin (Abcam 15277, 1:1000), β-dystroglycan (DSHB MANDAG2(D11), 1:1000), and GAPDH (Sigma SAB4300645, 1:1000) were diluted in blocking buffer and incubation was performed overnight at 4 °C with agitation. Membranes were washed three times for 5 min in TBS-T at room temperature, then incubated for 1 h at room temperature with species-matched AlexaFluor-conjugated secondary antibodies (Invitrogen, 1:1000) diluted in blocking buffer with agitation. Membranes were again washed three times for 5 min in TBS-T before imaging on a ChemiDoc MP imaging system (Bio-Rad). Before re-probing for other targets, membranes were stripped of antibodies with Restore Plus Western Stripping Buffer (Thermo Fisher Scientific) for 10 min, then washed 3 times for 10 min in TBS-T at room temperature. Densitometric quantification of western blot band intensity was performed using ImageJ and protein levels were normalized to the level of GAPDH in the sample.

Immunofluorescent imaging and quantitative sarcomere analysis

A subset of EHTs were processed for histology as previously described. ³³ Briefly, following force and Ca²⁺ assessment, EHTs were submerged in 140 mM KCl for 1 min to inhibit contraction and induce EHT relaxation followed by fixation in 4% paraformaldehyde in DPBS for 15 min. EHTs were then washed with DPBS and dehydrated in a 30% w/v sucrose solution overnight at 4 °C. EHTs were removed from the posts and embedded in O.C.T. compound. Longitudinal and transverse cryosections of 20 µm thickness were used for histology. Sections were blocked and permeabilized with 1% BSA and 0.1% Triton X-100 for 30 min at room temperature, followed by overnight incubation at 4 °C with mouse primary antibodies against α-actinin (Sigma-Aldrich A7811, 1:800) or dystrophin (Leica NCL-DYS1 (monoclonal, 1:30) or Abcam 15277 (polyclonal, 1:200). The following day, sections were incubated with donkey anti-mouse Alexa Fluor 488, donkey anti-rabbit Alexa Fluor 555, or goat anti-mouse Alexa Fluor 594 (Invitrogen, 1:200), and Alexa Fluor 488 phalloidin (Invitrogen A12379, 1:150), for 1 h at room temperature. Cover slides were mounted with ProLong™ Gold Antifade Mountant with DAPI (Invitrogen). Images were taken on a Leica SP8 confocal microscope. Laser strength and gain were kept constant between all samples and fields of view. Resting sarcomere length and alignment were calculated using α-actinin stained images and a scanning gradient Fourier transform program in MATLAB, as previously described. ⁴⁵ Z-disk width was calculated using a custom MATLAB script that thresholds and measures the major axis length of α-actinin stained images. Dystrophin stain intensity was quantified by applying a threshold mask to images to select cell area then averaging the pixel intensity within the mask.

RNA sequencing

Following force and Ca²⁺ assessments, a subset of EHTs were removed from posts and stored in RNAlater (Ambion) at −80°C. Lactate enriched, age-matched hiPSC-CMs (7 weeks post differentiation) were similarly stored in RNAlater at −80°C. Individual EHTs and hiPSC-CM samples were sent to BGI genomics for RNA extraction and 100 base pair paired-end transcriptome sequencing on the DNBseq platform with 50 million reads per sample. Reads were aligned to the human genome using Rsubread (Bioconductor) to GENCODE GRCh38.p13. ⁴⁶ Aligned reads were then counted using featureCounts (Bioconductor), excluding chimeras, multi-mapping genes, multi-overlap, and single end genes, resulting in roughly 20 million counts per individual EHT. ⁴⁶ Differential expression analysis was performed using edgeR (Bioconductor) and the following criteria: counts > 10, p < 0.05, |log(FC)| > 0.3, FDR < 0.05. ⁴⁷ Gene ontology (GO, 02/2021 release) analysis was performed with clusterProfiler (Bioconductor). ⁴⁸

Biological replicates and statistical analysis

The data shown herein represent EHTs pooled from multiple experimental replicates as indicated, where an independent experiment refers to a batch of EHTs made from an independent differentiation. The data encompasses up to three independent experiments with 6–12 tissues per batch, as indicated in figure legends. All data points shown in figures represent values for a single EHT and different symbols designate separate experiments. All values are reported as mean ± standard error of the mean (S.E.M.) unless indicated otherwise. Results were compared using an unpaired, two-tail t-test unless indicated otherwise and differences with a p-value < 0.05 were considered statistically significant as denoted with an asterisk.

Generation of EHTs from dystrophin-mutant hiPSC-CMs

Control hiPSCs were generated from urine derived cells from a healthy donor as previously described. ²² The control line was genetically edited using CRISPR-Cas9 to generate an isogenic match that harbored a dystrophin-truncating mutation (DMD 263delG).^31,40 Briefly, the DMD 263delG line has a single base pair deletion within the first exon of the DMD gene, resulting in the expression of a truncated isoform of dystrophin lacking a significant portion of the actin-binding domain (exons 1–6) (Figure 1(a)). N-terminal DMD mutations are known to cause more severe cases of Becker’s muscular dystrophy. ⁴⁹ Both control and DMD 263delG hiPSCs were subjected to directed cardiomyocyte differentiation and used to generate EHTs by casting hiPSC-CMs in a fibrin gel between two silicone posts, one flexible and one rigid (Figure 1(b)). Both control and DMD 263delG hiPSC-CMs generated EHTs that compacted in size between the posts and began to generate observable spontaneous contractions within 1 week of casting (Figure 1(c)). Both control and DMD hiPSC-CMs produced EHTs of similar resting length and cross-sectional area, indicating that EHT compaction, and therefore resting tissue force and stress, was consistent between the two lines (Supplemental Figure S2). Western blot analysis using a polyclonal antibody against the C-terminal domain of dystrophin showed similar expression levels of dystrophin protein in control and DMD 263delG EHTs (Figure 1(d) and (e), Supplemental Figure S4). Additionally, we observed lower-trending expression of β-dystroglycan, another dystroglycan complex protein, in DMD 263delG EHTs (Figure 1(d) and (e), Supplemental Figure S4), recapitulating the most recent demonstration of dystroglycan protein complex assembly in hiPSC-CMs. ⁵⁰ However, while immunocytochemical analysis of sarcomeric F-actin confirmed the presence of uniaxially aligned sarcomeres in control and DMD 263delG EHTs, an antibody targeting the rod domain of dystrophin (exons 26–30) indicated a high amount of the full-length protein in the control EHTs and only a trace amount of the truncated protein in the DMD 263delG EHTs (Figure 1(f), Supplemental Figure S3). Similarly, in transverse sections, we observed that dystrophin localized strongly to the cell membrane in control EHTs, while less dystrophin appeared at the cell membrane in DMD 263delG EHTs (Figure 1(g)). Together, these results suggest that while there was equivalent total expression of dystrophin in the DMD 263delG EHTs as seen by western blot, there was decreased localization of this truncated dystrophin to the cell membrane as seen by immunofluorescence. The observed decrease in β-dystroglycan expression further supports this hypothesis of impaired dystroglycan protein complex formation in the absence of full-length dystrophin.

Transcriptional phenotype of DMD hiPSC-CMs and EHTs

We performed RNA-sequencing on age-matched hiPSC-CMs and EHTs to investigate the transcriptomic effects of 3D EHT culture downstream of the DMD 263delG mutation (Figure 2(a)). We observed that 2D cell culture produced more differentially expressed genes (DEGs, |log(FC) > 0.3|, p < 0.05, FDR < 0.05, n = 3) between control and DMD 263delG samples than EHTs (4233 DEGs for 2D cell culture versus 1527 DEGs for EHTs) (Figure 2(b)). Interestingly, there appeared to be very little overlap between DEGs upregulated or downregulated as a result of the DMD 263delG mutation in 2D cell culture when compared to EHTs. To better demonstrate the differences between 2D and EHT culture, we compared the top 30 upregulated and downregulated gene ontology (GO) biological processes (BP) as ranked by p-value (Figure 2(c) and (d)). When comparing 2D control and DMD 263delG hiPSC-CMs, very little emerges that is relevant to dystrophin function or cardiac biology. Interestingly, while there was a downregulation of genes related to heart development and cardiac muscle cell action potential, the majority of BP terms relate to a large upregulation in mitochondrial function and ATP generation, with downregulation of BPs related to neural development, epithelial morphogenesis, and hemostasis (Figure 2(c)). Conversely, when comparing control and DMD 263delG EHTs, we observed dysregulation of many BPs relevant to cardiac function, including many related to heart and cardiac muscle development, cardiac action potential and muscle contraction, as well as extracellular matrix organization (Figure 2(d)). These results suggest that the DMD 263delG mutation does not have a profound effect on the cardiac-specific transcriptome in 2D culture, perhaps due to decreased hiPSC-CM maturity in this less physiologically-relevant culture platform, while the uniaxial, 3D culture provided by EHTs produces a disease-relevant transcriptomic response.

A closer analysis of the transcriptional changes that occurred in DMD 263delG EHTs reveals potential mechanisms driving the dystrophic phenotype presented in EHTs. Direct comparison of a subset of DEGs with the highest fold change revealed a hierarchical clustering that clearly delineates the two genotypes (Figure 2(e)). GO analysis indicated dysregulation of several key BPs including those related to cardiac muscle tissue development, heart contraction, regulation of membrane potential, calcium ion homeostasis, and extracellular matrix organization (Figure 2(f)). Further inspection of key genes identified by GO analysis indicates potential drivers of the dystrophic phenotype presented in EHTs (Supplemental Table S1). Interestingly, we observed increased expression of DMD mRNA in DMD 263delG EHTs, potentially indicating an attempted compensatory mechanism for the dysfunctional mutant dystrophin protein. Ultimately, we observed that 3D EHT culture produces transcriptional changes relevant to cardiac biology, suggesting multiple mechanisms driving the dystrophic phenotype produced as a result of a lack of full-length dystrophin.

Decreased contractility and delayed contractile kinetics in DMD EHTs

Given the central role of dystrophin in maintaining mechanical integrity of the cell, it was of interest to assess the extent to which dystrophic EHTs are able to produce contractile forces. After 3 weeks in culture, control and DMD 263delG EHTs were subjected to auxotonic contractility assessment with electrical field stimulation at 1.5 Hz via optical measurement of the deflection of the flexible post. It was observed that DMD 263delG EHTs had a similar baseline tension as control EHTs but produced significantly decreased twitch forces (Figure 3(a) and (b)). Accordingly, DMD 263delG EHTs produced lower specific forces than the control EHTs when estimated cross-sectional areas were accounted for (Figure 3(c)). DMD 263delG EHTs were also observed to have lower twitch power, twitch work, and shortening velocity, as compared to control EHTs, illustrating an overall decrease in contractile performance (Figure 3(d)–(f)). Normalized traces of tissue force (Figure 3(g)) reveal slower kinetics of contraction in DMD 263delG EHTs, which significantly increased the time to peak force and time to 50% relaxation (Figure 3(h)–(j)). All representative tissue traces are shown in Supplemental Figure S5. Overall, these results indicate that the dystrophin mutation results in impaired contractile function in DMD 263delG EHTs.

Altered sarcomere structure in DMD EHTs

To investigate the underlying causes of impaired contractility, we performed quantitative image analysis of myofibril structure. Specifically, control and DMD 263delG EHTs were stained for Z-disk protein α-actinin, and confocal images were analyzed with custom image analysis scripts. We observed that both the control and DMD 263delG EHTs generated cardiomyocytes that had sarcomeres aligned in the longitudinal direction (Figure 4(a)). Sarcomere length was quantified using a scanning-gradient Fourier transform. ⁴⁵ We found that control EHTs had significantly longer sarcomeres as compared to DMD 263delG EHTs (Figure 4(b)). However, we observed no significant difference in Z-disk width between control and DMD 263delG EHTs (Figure 4(c)). Additionally, it was observed that control and DMD 263delG EHTs had similar overall sarcomere alignment (Figure 4(d)). The observed shorter sarcomere spacing is consistent with the lower contractile forces observed in DMD 263delG EHTs.

Altered Ca²⁺ transients in DMD EHTs

It was of interest to investigate the Ca²⁺ transients of dystrophic EHTs, as previous studies have pointed to Ca²⁺ dysregulation as a key phenotype of DMD presented in vitro.^{22–25,28–31} To assess the extent of Ca²⁺ dysregulation in DMD 263delG EHTs, the ratiometric Ca²⁺ dye Fura-2 was used to assess both the kinetics of Ca²⁺ transients as well as the relative cytosolic Ca²⁺ levels under electrical field stimulation at 1.5 Hz. Comparing the ratiometric fluorescence intensities (Figure 5(a)) revealed significantly higher baseline levels of cytosolic Ca²⁺ (Figure 5(b)) as well as higher peak levels (Figure 5(c)) but with a decreased amplitude of Ca²⁺ flux in DMD 263delG EHTs (Figure 5(d)). Examining the kinetics of Ca²⁺ transients of control and DMD 263delG EHTs, normalized traces of cytosolic Ca²⁺ flux (Figure 5(e)) revealed slower time to Ca²⁺ transient peak in DMD 263delG EHTs (Figure 5(f)). We observed a decreased rate to 50% and 90% Ca²⁺ transient decay in DMD 263delG EHTs (Figure 5(g) and (h)). While we observed some batch-to-batch variability in measured 365 nm/380 nm values, the trends remained consistent. An examination of average force-Ca²⁺ loops highlighted the elevated Ca²⁺ level and decreased Ca²⁺ transient amplitude, as well as the decreased twitch force in DMD 263delG EHTs (Figure 5(i)). The significantly decreased force-Ca²⁺ loop area provided further evidence toward the lower Ca²⁺ flux as a significant contributor to the reduction of force seen in DMD 263delG EHTs (Figure 5(j)). All representative Ca²⁺ traces and images from Fura-2 measurement are shown in Supplemental Figure S6. These results point to elevated cytosolic Ca²⁺ and Ca²⁺ transient abnormalities as key contributors to the dystrophic phenotype presented in vitro.

Increased beat interval irregularity in DMD EHTs

To investigate consequences of Ca²⁺ transient abnormalities, we next investigated whether previously observed increased beat interval irregularity (ΔBI) was conserved in this model.^23,30 We found that DMD 263delG and control EHTs had similar average spontaneous beating frequencies (Figure 6(a)), but upon closer inspection of individual traces, intermittent irregular beats were observed to occur at a higher frequency in DMD 263delG EHTS (Figure 6(b)). We compared the difference between subsequent beat intervals (ΔBI = BI_n+1 – BI_n), which revealed a higher distribution of ΔBI in 263delG EHTs when distributions were compared with a nonparametric statistical test (Figure 6(c)). We noted that the distribution of ΔBI in the DMD 263delG EHTs spanned a greater time interval than the control EHTs. To highlight this difference, we selected a cutoff value of 1000 ms and compared the frequency with which ΔBI events exceeded this cutoff. This analysis revealed an increased occurrence of highly irregular events in DMD 263delG EHTs (Figure 6(d)). All collected traces are shown in Supplemental Figure S7. These results indicate a potential downstream consequence of the altered Ca²⁺ transients observed in DMD 263delG EHTs.