Cardiac fibroblast-mediated ECM remodeling regulates maturation in an in vitro 3D engineered cardiac tissue

Abstract

Cardiac fibroblasts play an important role in heart homeostasis, regeneration, and disease by producing extracellular matrix (ECM) proteins and remodeling enzymes. Under normal conditions, fibroblasts exist in a quiescent state and maintain homeostasis, such as tissue structure and ECM turnover. However, if they become activated upon stimuli, such as injury, aging, or mechanical stress, which can lead to disease through excessive cell proliferation and ECM production. In addition to their role in disease progression, it remains unclear how cardiac fibroblasts contribute to cardiac maturation during development and whether the mechanism driving cytokine and ECM production during development aligns with those observed in pathological conditions. In this study, we investigated the functional and structural maturation of engineered cardiac tissue by modulating fibroblast activity within a three-dimensional (3D) in vitro model. In this model, human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs) and human primary cardiac fibroblasts (FBs) were co-cultured in a fibrin gel and their morphology, beating characteristics, beating force, and mRNA expression profiles were analyzed. The results demonstrate that functional and structural maturation were enhanced by fibroblast-driven tissue contraction and collagen deposition, while inhibition of ECM remodeling impaired both processes. However, excessive collagen accumulation reduced functional maturation by limiting contractile efficiency. Our data suggest that ECM remodeling by cardiac fibroblasts is essential for cardiac tissue maintenance and maturation. Additionally, the regulation of collagen deposition by fibroblast activity will be a key focus of future research, as it may critically influence both cardiac development and the progression of heart disease.

Introduction

The microenvironment of heart tissue consists of biochemical factors such as growth factors and cytokines by cell-cell interactions and biomechanical factors such as ECM (extracellular matrix) properties and geometric guidance by cell-ECM interactions, which play a crucial role in the development of cardiac structure and function.^1–3 During heart development or disease progression, the composition of the cells and ECM changes dynamically. The differences between development and disease are determined by the relationship between the major cells of the heart, such as cardiomyocytes, fibroblasts, and endothelial cells in a three-dimensional (3D) environment.^4,5 For example, endothelial cells regulate cardiac trabeculation and endocardial cushion through angiocrine signaling pathways and fibroblasts can regulate cardiac function through paracrine communication with cardiomyocytes.^6,7 Therefore, efforts to elucidate the mechanisms of heart development and disease progression using key components of the heart highlight the importance of non-cardiomyocytes as well as cardiomyocytes.^8–10

Cardiac fibroblasts are known as important cells in the formation of a microenvironment by secreting various cytokines and ECM proteins in the myocardial tissue.^11,12 Cardiac fibroblasts normally exist in a quiescent state and contribute to the structural, mechanical, and electrical physiology of the myocardium, ¹³ however, in response to injury or disease, they undergo a transition from a quiescent state to a myofibroblast phenotype, which involves increased fibroblast proliferation, decreased tissue contractility, and enhanced ECM remodeling, such as excessive collagen secretion and tissue compaction. These processes ultimately contribute to the progression of cardiac fibrosis.^14,15 To understand the mechanism of disease progression by cardiac fibroblasts, cardiac fibrosis has been studied and although several factors are involved in orchestrating the fibrotic response, the predominant mediator is TGF-β. ¹⁶ A cardiac fibrosis model formed by treatment of co-cultured human cardiac fibroblasts and pluripotent stem cell-derived cardiomyocytes with TGF-β exhibits the classic features of fibrosis-induced heart failure, including high collagen deposition, increased tissue stiffness, BNP secretion, and passive tension. ¹⁷ To further understand the role of fibroblasts in the heart, studying the differences in their role in development and disease progression is necessary, which requires the development of three-dimensional models of in vitro cardiac tissue using cardiomyocytes and cardiac fibroblasts.

Cardiac fibroblasts significantly contribute to the formation of the microenvironment by ECM-related protein secretion as well as interactions with cardiomyocytes during heart development.^18–20 Significant differences in the proportion of fibroblasts in the atria and ventricles with different structures and functions demonstrate that fibroblasts contribute to the structure and function of the heart during developmental stages. ²¹ In addition, the heart shows dynamic changes in the major ECM components, such as collagen type I and III, fibronectin, and laminin, depending on the developmental stage.^22,23 Along with increased collagen deposition, particularly during early heart development, tissue stiffness and expression of cardiac maturation markers are also increased. ²⁴ These alterations in the microenvironment, including the ECM, induce changes in mechanical forces, such as tension in the surrounding tissues, and are related not only to structural maturation, but also to functional maturation.^25–27

To elucidate mechanisms of heart development and disease progression, various platforms have been developed, such as cardiac spheroids, organoids, and 3D engineered cardiac tissue, to mimic the function and structure of the heart. However, both platforms, spheroids and organoids, are not suitable for studying structure development due to the lack of cell-ECM interactions and mechanical forces, as there is little ECM in the tissue. Compared with spheroids and organoids, the 3D engineered cardiac tissue model, that can form while controlling cells, and the ECM not only has mechanical forces such as tension, but also can measure those forces.^28–31 Based on these advantages, the engineered cardiac tissue model may be used in myocardial tissue maturation models by mechanical stimulation, disease models, such as fibrosis and afterload, and genetic disease models using genetically modified cells.^31–34 However, significantly fewer models have studied the role of fibroblasts in heart development compared with cardiac disease models.

To address this gap, we investigate how fibroblast-mediated ECM remodeling influences the structural and functional maturation of cardiac tissue during early development. We established a 3D engineered cardiac tissue model using human induced pluripotent stem cell-derived cardiomyocytes (hiPSC-CMs), human primary cardiac fibroblasts (FBs) and fibrin gel. We hypothesized that fibroblasts promote cardiac tissue maturation through ECM remodeling processes such as collagen deposition and tissue compaction, and that dysregulation of this remodeling-either by inhibition or excessive activation-would disrupt tissue development. Consistent with this, our model exhibited both structural and functional maturation, characterized by fibroblast-mediated tissue contraction and collagen deposition, whereas inhibition of ECM remodeling significantly impaired these maturation processes. In particular, sustained collagen deposition by activated fibroblasts increased the expression of genes associated with structural maturation, but led to a decline in functional properties, such as contractility. These findings suggest that the regulation of collagen secretion via fibroblast activation plays an important role in myocardial tissue maturation and may contribute to pathological fibrosis. Our results are expected to be applied to new drug development and developmental research aimed at controlling fibroblast activity or the structural development of tissues to inhibit fibrosis.

Results

Role of fibroblasts in fibrin-based 3D engineered cardiac tissues

A 3D engineered cardiac tissue was generated using human cardiac cells, fibrin gel, and aprotinin, as previously described.^35,36 Aprotinin was applied until Day 7 to stabilize the tissue, after which the constructs were detached from the mold and allowed to undergo natural ECM remodeling without further aprotinin treatment until Day 14.^37,38 To investigate the role of fibroblasts in 3D engineered cardiac tissue formation, two groups were established: a cardiomyocyte-only (CM only) group and a cardiomyocyte-fibroblast (CMFB) co-culture group, consisting of 80% hiPSC-CMs and 20% human cardiac fibroblasts, reflecting the cellular composition of human cardiac tissue (Figure 1).^21,39,40 As a result, the CM only group did not maintain the 3D tissue structure, leading to complete degradation (Figure 1(b) and Supplemental Figure S1). In contrast, the CMFB group maintained its rod-shaped structure, though a gradual reduction in tissue width over time (Figure 1(c)). Given the prominent role of fibroblasts in collagen deposition during early cardiac development, we analyzed collagen expression differences in fibrin-based 3D cardiac tissues between groups with and without fibroblasts. ²⁴ As a result, COL1A1 expression, indicative of collagen type I production, was significantly higher in the CMFB group than in the CM only group (Figure 1(d)). To further examine the role of fibroblasts in structure maintenance beyond collagen secretion, 3D engineered cardiac tissues were fabricated using a collagen gel instead of a fibrin gel. The results showed that the rod-shaped structure was maintained regardless of the presence of fibroblasts in 3D engineered cardiac tissues with collagen gel (Figure 1(e)). However, although the structure was preserved in the CM-only (collagen) group, the absence of width reduction due to tissue compaction resulted in irregular beating patterns (Figure 1(f) and (g) and Supplemental Figure S1 and Supplemental Video S1). In the CMFB group, width reduction and tissue compaction resulted in synchronized beating (Figure 1(f) and (g) and Supplemental Figure S1 and Supplemental Video S2). These results suggest that fibroblasts may significantly contribute to the formation and maintenance of 3D engineered cardiac tissue by remodeling the ECM through collagen type I production and tissue compaction.

Figure 1.

The presence of cardiac fibroblasts affects 3D engineered cardiac tissue formation. (a) Schematic diagram of the 3D engineered cardiac tissue formation process. (b) Morphological change in the CM only group with fibrin gel on Day 1, 7, and 14 (rod-shaped structure fully degraded at Day 14). Scale bars: 2 mm. Magnification: 500 μm. (c) Morphological change of the CMFB (co-culture) group with fibrin gel on Day 1, 7, and 14 (rod-shaped structure maintained until Day 14). Scale bars: 2 mm. Magnification: 500 μm. (d) mRNA expression of ECM marker (COL1A1) in the CM only and CMFB group on Day 7 and 14 (n = 3 for all groups). A significant difference was observed among groups (F(3, 8) = 1538). (e) CM only and CMFB rod-shaped structure with a collagen type I gel. Scale bars: 500 μm. (f) Analysis of width changes in 3D engineered cardiac tissue structure depending on the presence of fibroblasts and the type of gel over time (n = 4 or 5 for each group). (g) Width deformation rate at Day 14 across groups. Kruskal-Wallis test revealed significant group differences (p = 0.0127). Pairwise Mann-Whitney U test indicated Significant differences between CM-only and CMFB (fibrin), and between CMFB (collagen) and CMFB (fibrin). Data are shown as the mean ± SD. Statistical analysis: one-way ANOVA for (d); Kruskal-Wallis test and Mann-Whitney U tests for (g).

*p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Role of fibroblasts in structural maturation of 3D engineered cardiac tissue

The lack of fibroblasts in 3D engineered cardiac tissue formation failed to maintain structural integrity, indicating fibroblasts are essential for ECM remodeling during tissue formation. To investigate the effect of ECM remodeling by fibroblasts over time, the structural maturation markers of the CMFB group were analyzed in 3D engineered cardiac tissues formed with fibrin gels. Unlike collagen gels, where intense tissue compaction occurred within a single day, Fibrin gels, exhibiting gradual tissue compaction over time, allowed specific evaluation of fibroblast-mediated ECM remodeling, and were thus selected as the base material for this study. Collagen type I deposition, a key component of the heart ECM was examined pre- and post-remodeling on Day 7 and Day 14 using second harmonic generation (SHG), which enables label-free visualization of fibrillar collagen. ⁴¹ Our results showed that collagen deposition and fiber alignment were significantly higher in the Day 14 group (post-remodeling) compared to Day 7 group (pre-remodeling), as shown by the directionality histogram. Dispersion and collagen area measurement indicated enhanced ECM anisotropically organization with ECM remodeling (Figure 2(a–c)). Tissue stiffness, measured via wet-atomic force microscopy (AFM), showed a shift in Gaussian peak distribution between Day 7 and Day 14 and increased from Day 7 (99.5 %) to Day 14 (123.4 %) relative to Gel only group, reflecting collagen-induced structural reinforcement (Figure 2(d) and (e) and Supplemental Figure S2). To further evaluate cellular organization, F-actin (red) staining was employed to visualize overall cell distribution, while sarcomeric α-actinin (SαA, green) staining highlighted cardiomyocyte arrangement and sarcomere formation; F-actin signals without α-actinin colocalization were interpreted as fibroblasts (Figure 2(f)). At Day 7, the cells exhibited a dispersed distribution with isotropic orientation, whereas by Day 14, the cells were densely packed and demonstrated clear anisotropic alignment, indicative of structural organization and tissue compaction. Quantification of the cell-cell distance, measured by nuclear distance, showed reduced distance on Day 14 compared to Day 7, which appears to result from tissue compaction rather than fibroblast proliferation, as indicated by significant width reduction between Day 7 and Day 14 (Figure 2(g)), and the predominant colocalization of α-actinin and F-actin signals (Figure 2(f)). Additionally, the sarcomere length, a marker of cardiomyocyte structural maturation, significantly increased between Day 7 and Day 14, further indicating advanced tissue maturation (Figure 2(h)). In mRNA expression analysis via qPCR revealed no significant difference in TNNT2 (cardiac troponin T), a general cardiomyocyte marker, between Day 7 and Day 14. However, the levels of structure-related maturation markers, including TNNI3 (cardiac troponin I), MYL2 (ventricular myosin light chain-2), and TTN (titin), were notably elevated on Day 14 (Figure 2(i)). Interestingly, the expression of COL1A1 (collagen type I), ACTA2 (smooth muscle α actin) and TCF21 decreased on Day 14. These markers of fibroblast activity suggested a natural regulation of fibroblast activity as structural maturation advanced (Figure 2(i)). These findings imply that as ECM remodeling progresses—encompassing fibrin gel degradation, tissue compaction, and collagen type I deposition—tissue stiffness and structural maturation of the 3D engineered cardiac tissue are enhanced, with fibroblast activity dynamically adapting to support this developmental stage.

Figure 2.

Fibroblast-mediated ECM remodeling enhances structural maturation in 3D engineered cardiac tissue. (a) Analysis of collagen type I formation using second harmonic generation (SHG) and directionality in 3D engineered cardiac tissues on Day 7 and 14. Left panel: 3D reconstruction of collagen type I distribution. Right panel: Corresponding Z-projection of collagen type I image shown in the left panel. Scale bars: 100 μm. (b) Quantitation of collagen deposition area in 3D engineered cardiac tissues on Day 7 and 14 (n = 3). (c) Quantitation of collagen type I dispersion in Day 7 and 14. (dande) Histograms and Young’s modulus measurements of gel only, Day 7 and 14 3D engineered cardiac tissues, obtained using wet atomic force microscopy (AFM) (N = 25 for Gel only, N = 189 for Day 7 and N = 161 for Day 14, across three biological replicates). (f) Immunofluorescence images of sarcomeric α-actinin (green), F-actin (red), and DAPI (blue) in the 3D engineered cardiac tissues on Day 7 and 14. Scale bars: 100 μm. (g and h) Measurement of cell-to-cell distance and sarcomeric length in 3D engineered cardiac tissues (n = 3 for Day 7 and n = 5 for Day 14). (i) mRNA expression of structural maturation markers (TNNT2, TNNI3, MYL2, and TTN), ECM markers (COL1A1 and POSTN) and fibroblast activity markers (ACTA2 and TCF21) on Day 7 and 14 in the 3D engineered cardiac tissue group (n = 3). Data are shown as the mean ± SD. Statistical analysis was performed using unpaired t-test for (b, c, i), one-way ANOVA for (e) and Mann-Whitney U test for (g, h).

NS: no significance.

*p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Impact of ECM remodeling on contractile force in 3D engineered cardiac tissue

To examine the functional maturation of 3D engineered cardiac tissue influenced by ECM remodeling, we measured contractile force generated by the tissue over time. Contractile force was evaluated in three components: twitch force (the force generated by spontaneous beating), resting tension (the baseline force generated by passive tension), and net force (the combined total of twitch force and resting tension) (Figure 3(a) and Supplemental Videos S3 and S4). The result showed that both twitch force and net force increased as ECM remodeling progressed (Figure 3(c) and (d) and Supplemental Figure S3). In particular, twitch force significantly increased after day 11, whereas resting tension initially increased and then slightly decreased after day 11 (Figure 3(b) and (c) and Supplemental Figure S3). By displaying the twitch force graph alongside tissue width graph, we observed a correlation in which contractile force increased as ECM remodeling progressed, accompanied by a reduction in width, indicating a potential link between tissue compaction and increased contractile force (Figure 3(e) and Supplemental Figure S3). These findings suggest that, as ECM remodeling advances, the contractile force generated by the 3D engineered cardiac tissue escalates. The observed tissue width reduction can help establish a correlation between ECM remodeling and observed enhanced contractile force.

Figure 3.

As ECM remodeling decreases the width of the 3D engineered cardiac tissue, its contractile force increases. (a) Schematic diagram of the measured forces within 3D engineered cardiac tissue. (b–d) Measurement of resting tension, twitch force, and net force changes from day 7, 9, 11, 13, 14 (each day, n = 4) (e) Analysis of width and twitch force changes over time. Data are shown as the mean ± SD.

NS: no significance.

*p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Establishment of a 3D engineered cardiac tissue model to study ECM remodeling

To assess beating properties, we measured intensity changes in ROIs during 10 s of spontaneous beating, which was used to capture intrinsic variability, as field stimulation induces synchronized contractions (Figure 4(ai)). As a result, irregular beating was observed on Day 7 (Figure 4(aii) and Supplemental Video S5), progressing synchronized beating on Day 14 (Figure 4(aiii) and Supplemental Video S6). In addition, we quantified peak-to-peak duration, contraction-relaxation duration, and beats per minute (BPM) using a beating graph, as mature cardiomyocytes are characterized by slower spontaneous beating rates and prolonged contraction-relaxation cycles ⁴² (Figure 4(b)). Result showed increased peak-to-peak and contraction-relaxation durations on Day 14, alongside a decrease in BPM compared to Day 7, indicating more advanced functional maturation at Day 14 (Figure 4(c)). mRNA expression analysis via qPCR revealed significant upregulation of beating-related genes, including ion channel (ATP2A2, RYR2, and KCNA4) and gap junction (GJA1) on Day 14 relative to Day 7 (Figure 4(d)). These results suggested that functional maturation, as evidenced by enhanced contractile force, synchronization and gene expression, progresses in this 3D engineered cardiac tissue as ECM remodeling advances.

Figure 4.

Structurally mature 3D engineered cardiac tissues exhibit mature beating characteristics. (a) Analysis of the beating intensity changes within the region of interest (ROI) (i) on Day 7 (ii) and Day 14 (iii) in the 3D engineered cardiac tissue. Scale bar: 500 μm. (b) Schematic diagram of the beating properties within the beating graph. (c) Comparisons of peak-to-peak (i), contraction-relaxation duration (ii) and beating rate per minute (iii) in 3D engineered cardiac tissue on Day 7 and 14 (n ⩾ 15 from five distinct ROIs in a representative sample). (d) mRNA expression of ion channel markers (ATP2A2, RYR2, and KCNA4) and gap junction marker (GJA1) in the Day 7 and 14 3D engineered cardiac tissue group (n = 3). Data are shown as the mean ± SD. Statistical analysis was performed using unpaired t-test for (c, d).

*p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Effects of ECM remodeling modulation on morphological and functional changes in 3D engineered cardiac tissue

Through in vitro 3D engineered cardiac tissue, we demonstrated that ECM remodeling by fibroblasts plays a significant role in promoting both structural organization and functional enhancement in cardiac tissues. To further assess the effects of ECM remodeling inhibition and fibroblast activation on tissue formation, we included two additional experimental groups alongside the REMDL (control) group. The REMDL group refers to cardiac tissues that received aprotinin treatment for the first 7 days to transiently inhibit ECM remodeling, followed by natural remodeling from Day 7 to 14 without aprotinin—corresponding to the Day 14 group presented previously. In contrast, the i-REMDL (ECM remodeling inhibition) group indicates cardiac tissues treated with aprotinin until Day 14 to evaluate the effects of continuous ECM remodeling inhibition compared to natural remodeling. This treatment did not affect the morphology or mRNA expression of genes associated with cardiomyocyte or fibroblast function (Supplemental Figure S4). In the Act-FB group, fibroblasts were pre-treated with TGF-β for 4 days before tissue formation to induce fibroblast activation (Figure 5A and Supplemental Figure S5). The results indicated that almost no width reducing and tissue compaction occurred in the i-REMDL group, while the Act-FB group developed 3D engineered cardiac tissue structure resembling the REMDL group (Figure 5(b–d) and Supplemental Figure S6). In addition, beating analysis at Day 14 showed that the Act-FB group displayed synchronized beating similar to the REMDL group (Supplemental Video S7), whereas the i-REMDL group exhibited irregular beating (Figure 5(e) and Supplemental Video S8). Both the i-REMDL and Act-FB groups demonstrated shorter contraction-relaxation durations and higher beating frequencies than the REMDL group, suggesting less mature beating properties, as indicated by peak-to-peak duration and BPM analysis (Figure 5(f)). Contractile force measurement revealed that the i-REMDL group generated significantly lower force overall, whereas Act-FB exhibited the highest net force. However, the force in the Act-FB group was largely attributed to tissue compaction, resulting in increased resting tension and reduced twitch force (Figure 5(g–i) and Supplemental Figure S7). These results suggest that the i-REMDL group, in which ECM remodeling was inhibited, exhibited minimal width change and insufficient tissue compaction, resulting in asynchronous and immature beating properties. This highlights the essential role of ECM remodeling in cardiac tissue development. In contrast, although the Act-FB group showed a similar degree of width reduction to the REMDL group, it exhibited immature beating properties, including significantly higher resting tension relative to twitch force, suggesting that excessive fibroblast activation induced a fundamentally different remodeling process. Therefore, both inhibition of ECM remodeling and excessive fibroblast activation adversely affect the functional maturation of 3D engineered cardiac tissues.

Figure 5.

Modulation of ECM remodeling interferes with the functional development of 3D engineered cardiac tissues. (a) Schematic diagram of the 3D engineered cardiac tissue formation process by ECM remodeling inhibition or activated fibroblast. (b) Morphological change of the ECM remodeling inhibition group (i-REMDL) and activated fibroblast (Act-FB) co-culture group at Day 14. Scale bars: 2 mm in the first row and 500 μm in the second row. (c) Width changes of the 3D engineered cardiac tissue structure over time (n = 4). (d) Quantification of the width deformation rate (n = 3 for i-REMDL group, n = 5 for REMDL and Act-FB group). (e) Analysis of the beating intensity in the region of interest (ROI) for i-REMDL (i) and Act-FB (ii) group, with each graph representing three distinct ROIs. (f) Comparisons of peak-to-peak (i), contraction-relaxation duration (ii) and beating per minute (iii) in the REMDL, i-REMDL, and Act-FB group at Day 14 (n ⩾ 15 from three distinct ROIs). Measurement of twitch force (g), resting tension (h), and net force (i) changes from Day 7 to 14 depending on the ECM remodeling or fibroblast activation (n = 3 for i-REMDL and Act-FB group, n = 4 for REMDL group). Data are shown as the mean ± SD. Statistical analysis was performed using one-way ANOVA for (d, f).

*p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Structural maturation of 3D engineered cardiac tissues based on the degree of ECM remodeling

To investigate the structural differences underlying the observed functional variations among the groups, we measured collagen deposition, tissue alignment, and gene expression in 3D engineered cardiac tissue. In the i-REMDL group, collagen deposition and fiber alignment were significantly reduced, indicating minimal ECM remodeling. In contrast, the Act-FB group, displayed enhanced collagen fiber alignment and increased collagen deposition compared to the REMDL group (Figures 6(a) and (b) and 2(a); REMDL group is represented in Figure 2 as the Day 14 group). The full width at half maximum (FWHM) derived from Gaussian fitting of the directionality analysis serves as a quantitative metric for assessing fiber alignment. Lower FWHM values indicate a higher degree of alignment in a singular direction, whereas higher FWHM values reflect greater dispersion in fiber orientation. In this study, the i-REMDL group exhibited significantly elevated FWHM values compared to other experimental groups, indicating a markedly reduced degree of fiber alignment (Figure 6(c)). Analysis of mRNA expression revealed that structural maturation markers, including cTNI and MLC2v, as well as activated fibroblast markers, such as COL1A1 and POSTN (periostin) were significantly increased in the Act-FB group (Figure 6(d)). Conversely, the expression of ion channels affecting cardiac function, such as RYR2, ATP2A2, and KCNA4, were downregulated in the Act-FB group, consistent with the immature beating and reduced twitch force seen in Figures 5 and 6(d). These results suggest that inhibiting ECM remodeling limits structural maturation of 3D engineered cardiac tissue, while excessive ECM remodeling via fibroblast activation promotes structural maturation, albeit with potential trade-offs in functional ion channel expression.

Figure 6.

Activated fibroblasts induce the structural development of 3D engineered cardiac tissues at Day 14. (a) Analysis of collagen type I formation using second harmonic generation (SHG) and directionality in the i-REMDL and Act-FB group. Scale bars: 100 μm. (b) Quantitation of collagen deposition area depending on ECM remodeling and fibroblast activation (n = 3 for REMDL group, n = 6 for i-REMDL group, n = 5 for Act-FB group). (c) Comparison of full-width half maximum (FWHM) in directionality histogram based on ECM remodeling and fibroblast activation (n = 3). (d) mRNA expression of structural maturation markers (TNNT2, TNNI3, MYL2, and TTN), ECM markers (COL1A1 and POSTN), ion channel markers (RYR2, ATP2A2, and KCNA4), Gap junction marker (GJA1) and fibroblast activity markers (ACTA2 and TCF21) depending on ECM remodeling and fibroblast activation (n = 3). Data are shown as the mean ± SD. Statistical analysis was performed using one-way ANOVA for (b–d).

*p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Internal structural changes in 3D engineered cardiac tissue based on the degree of ECM remodeling

To visualize and examine the internal structure of 3D engineered cardiac tissue based on ECM remodeling, we conducted immunofluorescent staining for α-actinin and F-actin (Figure 7(a) and Supplemental Figure S8). Analysis of sarcomere length using α-actinin showed comparable developmental patterns in the REMDL and Act-FB groups, while the i-REMDL group showed underdeveloped sarcomeres (Figure 7(b)). In addition, the i-REMDL group showed reduced tissue contraction due to the inhibited ECM remodeling, resulting in increased cell-cell distances and a more isotropic alignment of the cardiomyocytes, indicating limited structural organization (Figure 7(c–e)). In the Act-FB group, cardiomyocytes and fibroblasts were anisotropically aligned, suggesting organized structural formation; however, the cell-cell distance remained larger compared to the REMDL group. The lower ratio of sarcomeric α-actinin expression to F-actin in the Act-FB group indicated a higher relative presence of fibroblasts within the cardiac tissue, likely contributing to the observed tissue compaction and higher resting tension (Figures 5(h) and 7). These results support our interpretation that excessive ECM remodeling in the Act-FB group—marked by increased collagen deposition and a higher fibroblast-to-cardiomyocyte ratio—results in reduced functional maturation. The increased spacing between cardiomyocytes and the presence of excessive ECM components likely disrupt cell-cell coupling and mechanical signaling, both of which are crucial for proper functional integration.

Figure 7.

Inhibition of ECM remodeling hinders sarcomeric development of cardiomyocytes in 3D engineered cardiac tissues at Day 14. (a) Immunofluorescence images of sarcomeric α-actinin (green), F-actin (red) and DAPI (blue) in the REMDL, i-REMDL, and Act-FB group. Scale bars: 100 μm. (b and c) Measurement of sarcomeric length and cell-to-cell distance of the REMDL, i-REMDL, and Act-FB group (n = 5 for REMDL group and Act-FB group, n = 4 and 3 for i-REMDL group). (d) Quantitation of expression of sarcomeric alpha actinin/F-actin depending on ECM remodeling and fibroblast activation (n = 4 for REMDL group and Act-FB group, n = 3 for i-REMDL group). (e) Quantification of directionality in the rod-shaped structure of the REMDL, i-REMDL, and Act-FB group (n = 3). Data are shown as the mean ± SD. Statistical analysis was performed using Kruskal-Wallis test for (b), one-way ANOVA for (c–e).

*p < 0.05. **p < 0.01. ***p < 0.001. ****p < 0.0001.

Overall, these findings suggest that while ECM remodeling is essential for both structural development and functional performance of cardiac tissues. These results indicate that insufficient ECM remodeling is associated with underdeveloped tissues and reduced functional properties, whereas excessive ECM remodeling may lead to increased tissue stiffness, resembling characteristics observed in fibrotic tissues. These observations highlight the potential importance of maintaining a balanced ECM remodeling process to promote both structural and functional maturation in 3D engineered cardiac tissues, although further studies with larger sample sizes will be needed to fully validate these trends and their implications for healthy and diseased cardiac models.

Discussion

In vitro 3D engineered cardiac tissue models offer a powerful platform for mimicking and analyzing the structural and functional dynamics of heart development and disease. These models enable the investigation of critical processes such as ECM remodeling, tissue assembly, and force generation, providing insights into normal development and pathological conditions. ²⁸ Fibroblasts play a central role in these 3D engineered cardiac tissue by modulating ECM composition and interacting with cardiomyocytes^18–20; however, their specific contributions to 3D cardiac tissue development and disease progression have not been elucidated. In this study, we developed a 3D engineered cardiac tissue by incorporating cardiomyocytes and fibroblasts in hydrogels, recapitulating key features of cardiac development: collagen deposition, increased tissue stiffness, enhanced contractile force, and cardiomyocyte maturation. ²⁴ In this model, we aimed to investigate fibroblast-mediated ECM remodeling and its impact on cardiac tissue development during the formation process by either activating fibroblasts prior to tissue formation or inhibiting hydrogel degradation.

Fibroblasts are known to play a crucial role within cardiac tissue by secreting ECM, thereby altering the properties and physical characteristics of the tissue, particularly facilitating structural changes during development. Moreover, they promote wound healing and contribute to tissue regeneration; however, in pathological states, myofibroblasts can form fibrotic scars leading to the promotion of disease. Therefore, the regulation of fibroblast activity is important to myocardial tissue maturation or disease progression. ⁴³ Previous studies have demonstrated that treating cardiac tissues, composed of cardiomyocytes, fibroblasts and hydrogels, with TGF-β resulted in increased collagen deposition, fibrous tissue formation, irregular beating, and arrhythmias during cardiac remodeling.^27,44,45 Furthermore, a higher fibroblast-to-cardiomyocyte ratio has been shown to promote the formation of fibrotic tissue. ⁴⁶ While fibroblasts are commonly utilized in the formation of 3D cardiac tissues, research on cardiac development has predominantly focused on cardiomyocytes, with fibroblast-specific studies primarily investigating their roles in disease modeling.^31,47 As a result, the positive contributions of fibroblasts to cardiac development remain relatively unexplored. In this study, we investigated the role of fibroblasts in myocardial tissue development by analyzing three key aspects: 1) the impact of fibroblast presence or absence on myocardial tissue formation, 2) the effect of pre-activated fibroblasts on the myocardial tissue formation process, and 3) the influence of ECM remodeling inhibition on cardiac tissue maturation. Our results suggest that cardiomyocytes alone are insufficient to form structured 3D engineered cardiac tissues (Figure 1). Furthermore, we observed that activated fibroblasts with TGF-β enhanced structural maturation through increased deposition of collagen and improved cell alignment. However, this activation negatively impacted functional maturation, as evidenced by irregular beating properties (Figures 5 and 6). This discrepancy may be attributed by two factors: (1) TGF-β-induced fibroblast proliferation may have increased the fibroblast-to-cardiomyocyte ratio, thereby reducing the relative expression of cardiomyocyte-specific ion channel genes supported by Figure 7(c) and (d); and (2) excessive collagen deposition (Figure 6(a) and (b)) may have elevated resting tension (Figure 5(g)), limiting contractility and diminishing mechanical cues essential for electrophysiological maturation—an effect commonly observed in cardiac fibrosis. ⁴⁸ These observations highlight the importance of tightly regulating fibroblast activity to promote proper myocardial development.

ECM (extracellular matrix) remodeling during the development of cardiac muscle structure is a complex process that includes cell-cell interactions and cell-matrix interplay. A biomimetic matrix showed that the control of cell adhesion by adhesive peptide on the matrix and the modulation of the matrix degradation rate by protease-sensitive peptide as a crosslinker result in mature tissue formation during tissue regeneration and remodeling.^49–52 In addition, the attenuation of remodeling also leads to immature organ development and malfunction during development in vivo. When matrix remodeling is disrupted by the knockout of angiopoietin-1 (Angpt1), which is involved in dynamic ECM structure remodeling during cardiac development, it results in the impaired degradation of cardiac jelly, leading to malformations in the development of the atrium and ventricular structures. ⁵³ In our 3D engineered cardiac tissue model, inhibiting ECM remodeling with aprotinin reduced collagen deposition, disrupted cell alignment, and resulted in immature tissue structures. Conversely, excessive ECM remodeling by activated fibroblasts increased collagen deposition and enhanced structural maturation, but impaired functional maturation (Figures 2 and 5–7). This result is consistent with previous studies showing that the rate of remodeling of the matrix is critical during tissue regeneration.^51,52 These findings suggest that balanced fibroblast-mediated ECM remodeling appears to be important for supporting proper myocardial tissue development and minimizing potential abnormalities associated with either insufficient or excessive remodeling.

While our study provides valuable insights, additional analyses are needed to further characterize cardiac tissue maturation. We measured certain aspects of myocardial tissue maturation, including mRNA expression, contractile force, and sarcomere length through staining; however, these provide only a partial understanding of the maturation process. Assessments such as electrophysiology, calcium handling, and the effects of external mechanical forces could offer a more comprehensive understanding of the interplay between ECM remodeling and myocardial function. ⁵⁴ Despite these limitations, our findings suggest the importance of fibroblast-mediated ECM remodeling in heart development and underscore the role of matrix dynamics as a key factor in tissue engineering. In conclusion, this study highlights the central role of fibroblast activity and ECM remodeling in 3D engineered cardiac tissue formation, underscoring their importance in achieving structural and functional maturation. These insights have implications for both developmental biology and the development of regenerative therapies, as understanding and controlling ECM remodeling may enhance heart repair and regeneration.

Conclusions

Our study highlights the essential role of fibroblast-mediated ECM remodeling in the structural and functional maturation of 3D engineered cardiac tissues. Fibroblasts drive collagen deposition, tissue compaction, and cell alignment, which are critical for forming mature cardiac structures. While fibroblast activation enhances structural organization, it impairs functional synchronization, underscoring the need for balanced ECM remodeling. Conversely, inhibiting ECM remodeling disrupts collagen organization and alignment, leading to immature tissues. These findings emphasize the importance of regulated ECM dynamics in myocardial development and provide insights to advance cardiac tissue engineering and regenerative therapies.

Methods

PDMS mold preparation

A pre-polymer mixture containing PDMS (Sylgards^® 184, Dow Chemical Co., Midland, MI, USA) precursor and a curing agent at a 10:1 ratio was added to a 24-well plate. Then, Teflon spacers (EHT technologies GmbH, Hamburg, Germany) were placed upon it as a mold. After polymerization at 60°C for 2 h, the Teflon spacers were removed to obtain a PDMS mold.

Cell culture

Human induced pluripotent stem cell (hiPSC)-CMs (Cardiosight-S, NEXEL Co., Ltd., Seoul, South Korea) were stored in a liquid nitrogen tank, thawed in a 37°C water bath for 1 min, and used immediately in the experiment. We utilized multiple hiPSC-CM batches from NEXEL, exceeding three distinct batches. Cardiac fibroblasts (NHCF-V, Normal Human Ventricular Cardiac Fibroblast, LONZA, MD, USA) were cultured in FBM™ basal medium with FGM™-3 SingleQuot Supplements (Lonza) in a 37°C, 5% CO₂ humidified cell culture incubator. A passage below 8 was used and 10 ng/ml of Transforming Growth Factor-β (TGF-β, Human, Sigma-Aldrich) was added for 4 days to activate fibroblasts.

3D engineered cardiac tissue fabrication and culture

A mixture to generate a fibrin-based 3D engineered cardiac tissue was prepared on ice as follows: Final concentration: 1 × 10⁶ cells/100 μl (2 × 10⁵ cells of fibroblasts and 8 × 10⁵ cells of cardiomyocytes), 12.5 mg/mL bovine fibrinogen (stock solution: 15 mg/mL fibrinogen in 0.9% saline solution with 20 mM HEPES buffer, Sigma-Aldrich, Saint Louis, MO, USA) and 4 U/ml of thrombin (stock solution: 25 U/ml thrombin, Sigma-Aldrich). Unlike conventional EHT fabrication methods that commonly use agarose molds, we utilized PDMS molds to enable reusability and allow for rigorous sterilization without the need for antibiotics. PDMS molds were incubated with a 1% F-127 solution for 2 h to minimize surface adhesion between the PDMS molds and the cell-containing gel. After washing the PDMS mold with PBS, 84 μl of fibrinogen and cell mixture were mixed with 16 μl of thrombin (25 U/mL, Sigma-Aldrich) and loaded into the PDMS mold. The silicone post rack (EHT technologies) on the plate according to the shape of the PDMS mold. For fibrinogen polymerization, the constructs were placed in a 37°C, 5% CO₂ humidified cell culture incubator for 20 min. Then, 1 ml of CM plating medium (Cardiosight-S plating medium, Nexel) with 0.02 mg/ml of aprotinin (Sigma-Aldrich) was added per well and the constructs were maintained in a 37°C, 5% CO₂ humidified cell culture incubator. The 3D engineered cardiac tissues were cultured until Day 7 in CM maintenance medium (Cardiosight-S maintenance medium, Nexel) supplemented with 0.02 mg/ml of aprotinin (Sigma-Aldrich) and 213 μg/ml of L-ascorbic acid (Sigma-Aldrich). On Day 7, the 3D engineered cardiac tissues were detached from the PDMS mold, transferred to a fresh 24-well plate, and cultured in CM maintenance medium without aprotinin. During the demolding process, approximately 10–20% tissue loss may occur, depending on the relative strength of PDMS-tissue adhesion versus tissue-post rack adhesion. 3D engineered cardiac tissues were cultured for 14 days and media was changed once every 2 days.

For the collagen-based 3D engineered cardiac tissue, the solution for gelation (collagen concentration, 2 mg/ml) was prepared by mixing collagen type I (rat tail, BD Biosciences) with phosphate-buffered saline (PBS) supplemented with phenol red, and the pH was adjusted to 7.4 using 0.5 N NaOH. The solution (100 μl) was mixed with cells, 1 × 10⁶ cells/100 μl (2 × 10⁵ cells of fibroblasts and 8 × 10⁵ cells of cardiomyocytes), and loaded into the PDMS mold. The silicone post-rack (EHT technologies) was placed on the plate based on the shape of the PDMS mold. The constructs were placed in a 37°C, 5% CO₂ humidified cell culture incubator for 40 min. Then, 1 ml of CM plating medium was added per well and the constructs were maintained in a 37°C, 5% CO₂ humidified cell culture incubator. The 3D engineered cardiac tissue was cultured in CM maintenance medium supplemented with 213 μg/ml of L-ascorbic acid (Sigma-Aldrich). The media was changed once every 2 days.

Morphology analysis

Bright-field images were captured by microscopy (Nikon, Ti2-E, Tokyo, Japan) at 4× magnification, and full images of the 3D engineered cardiac tissue were taken at the same magnification and stitched to a 2 × 5 image. Width and post-to-post length were measured by ImageJ software.

Second harmonic generation for detecting collagen fibers

The SHG microscope system is based on an Olympus BX-51 microscope using a mode-locked Ti: sapphire laser (Chameleon Ultra II, Coherent) as a light source with a central wavelength of 810 nm (time pulse width = 140 fs). Two photomultiplier tubes (PMT; Hamamatsu) were used in photon counting mode for signal detection. Images were captured using an objective lens (20×, 0.5 NA, Olympus) with a pixel resolution of 1024 × 1024 × 40 voxels using a galvanometer mirror-based 2D point scanning system (GVS002, Thorlabs, USA).

Atomic force microscopy (AFM) nanoindentation

We measured the Young’s modulus (E, Pa) of each sample using atomic force microscopy (AFM)-based nanoindentation testing.^55,56 Silicon nitride cantilevers were used with approximately 0.08 N/m of the spring constant (CP-PNP-PS, NanoAndMore, USA) with 10.8 μm polystyrene attached to the end of the cantilevers. The cantilever was loaded to AFM (NX10, Park System, South Korea) to perform nanoindentation in a liquid phase (i.e., PBS buffer). Before each nanoindentation measurement, we arranged the x- and y-axis of the AFM sample stage using Smart Scan software (Park Systems, South Korea) and calibrated the spring constant of the cantilever following the manufacturer’s instructions at room temperature. The approach and retraction velocities of the cantilever were set at 1 μm/s and 5 nN of threshold in a closed z-loop. Nanoindentation was performed by the movement of the cantilever with a depth of 4 μm. The force-distance (F-D) curve was fitted by XEI software (Park Systems, South Korea), and the E was calculated by the Hertz model using equation (1).

$$F=\frac{4}{3}\frac{E}{1-{\mu }^2}\sqrt{R{\delta }^3}$$ (1)

Where F indicates the applied force, R is the radius of the sphere, δ is an indentation depth, and μ is a Poisson ratio (~0.5) of hydrogels that are assumed to be a linearly elastic isotropic network. We performed the nanoindentation tests within the maximum duration (~4 h) to minimize the negative effects of buffer evaporation. Based on this process, we measured the E of the sample with the non-destructive method. Tissue stiffness was evaluated by collecting multiple measurements from spatially distinct regions within each sample, across three biological replicates.

Beating analysis and force measurement

The contractile properties of the 3D engineered cardiac tissues were estimated by analyzing movies taken with a microscope (Nikon, Ti2-E, Tokyo, Japan) at 4× magnification. Movies captured images at 50 frames per second and the movies were analyzed by NIS software (Nikon), which detects the variation in light intensity in a selected region for beating characterization. Spontaneous beating was intentionally analyzed instead of field-stimulated contractions to preserve natural heterogeneity within the cardiac tissues, as field stimulation can artificially synchronize contractions. Since spontaneous beating can introduce greater variability across samples, a representative cardiac tissue exhibiting typical spontaneous beating behavior was selected. To focus on intrinsic regional heterogeneity and minimize sample-to-sample variability, five independent regions of interest (ROIs) within the same tissue were analyzed. Continuous 10-s segments were consistently selected across all conditions. For force measurement, movies were analyzed in MATLAB (MathWorks, Inc., Natick, MA, USA) using a custom source code to quantify the distance traveled by the post and variations in light intensity within 60 s (Supplemental Videos 3 and 4). The force measurement is obtained by a formula based on the distance δ traveled by a post with radius R, length L, and modulus of elasticity E (2).

$$F=\frac{3\pi E{R}^4\delta }{4L^3}$$ (2)

The post of the silicone post-rack is 1 mm in diameter and 12 mm in length, with a modulus of elasticity of 1.7 MPa.

Immunofluorescence staining

The 3D engineered cardiac tissues were washed with PBS (Gibco) and fixed with 4% paraformaldehyde (PFA) solution (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at room temperature. Permeabilization was performed with 0.4% Triton X-100 for 40 min, and blocking was done using 5% bovine serum albumin (BSA; Sigma-Aldrich) for 1 h at room temperature. After an overnight incubation at 4°C with primary antibodies, which included mouse sarcomere anti-alpha-actinin (Abcam, Cambridge, UK) at a dilution of 1:400 in 5% BSA solution. Samples were washed five times with 0.01% PBST for 15 min each. Alexa Fluor 488 goat anti-mouse IgG (H + L) was used as a secondary antibody at a dilution of 1:400 in a 5% BSA solution. Then, Alexa Fluor 594 phalloidin (Thermo Scientific) against anti-F-actin and 4’,6-diamidino-2-phenylindole, di-hydrochloride (DAPI; 1:2000; Molecular Probes, Carlsbad, CA, USA) for nuclear staining were added at dilutions of 1:400 and 1:2000 in 5% BSA solution, respectively, and incubated for 3 h at room temperature. After staining, the samples were washed five times with 0.01% PBST for 15 min each. The samples were visualized using an LSM 800 (Carl Zeiss, Oberkochen, Germany) confocal fluorescence microscope. Sarcomere length and cell-to-cell distance were measured by analyzing multiple regions within each sample, across three independently prepared 3D engineered cardiac tissues (biological replicates).

Quantitative PCR

Total RNA was extracted from each sample using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) with slight modifications of the manufacturer’s recommendations. Briefly, each sample was homogenized by pipetting 1 ml of TRIzol reagent for 5 min and then mixed with 0.2 ml of chloroform (Sigma), followed by vortexing for 15 s. The resulting mixture was centrifuged at 15,000 rpm for 15 min at 4°C and the colorless supernatant was transferred to a new tube, which was gently mixed with the same amount of isopropanol (Sigma). The supernatant mixture was centrifuged at 15,000 rpm for 10 min at 4°C, yielding an RNA pellet. After removing the supernatant, the RNA pellet was washed with 500 μl of 75% ethanol and centrifuged at 15,000 rpm for 10 min at 4°C. This step was repeated twice. To remove the ethanol completely, the pellet was dried in a dry oven (56°C) for 5 min. The dried RNA pellet was dissolved in nuclease-free duplex buffer (Integrated DNA Technologies, Coralville, IA, USA). The concentration of the extracted mRNA was measured using a NanoDrop 2000c (Thermo Fisher Scientific). Extracted RNA (500 ng) was reverse-transcribed into complementary DNA (cDNA) using the PrimeScript™ cDNA synthesis kit (Takara, Shiga, Japan). PCR was performed with iQSYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA, USA) using a MyiQ 2 real-time PCR detection system (Bio-Rad). The gene-specific primers are listed in Table 1 where β-actin served as the reference gene. The expression of each gene was normalized to that of the control group using the ddCt method.

Table 1.

List of primers used for quantitative PCR analysis.

Gene	Forward primer	Reverse primer	Company
β-actin	GGACCTGACTGACTACCTCAT	CGTAGCACAGCTTCTCCTTAAT	Integrated DNA Technologies (USA)
TNNT2	CGATGGATTCCAGTTCGAGTATG	CTTGCAGTGGTAGGTGATGTT	Integrated DNA Technologies (USA)
TNNI3	GACAAGGTGGATGAAGAGAGATAC	CTTGCCTCGAAGGTCAAAGA	Integrated DNA Technologies (USA)
MYL2	CGGAGAAGAGAAGGACTAGGA	ACAGACAAGGTAGGGACAGA	Integrated DNA Technologies (USA)
TTN	GATGACAGTGGAACCTACCG	GTCAGCTCAGGGAAAACAGA	Integrated DNA Technologies (USA)
RYR2	TGGACAGAGTTCGCACAGTA	TGTACTCGGTTCCACCTGAT	Integrated DNA Technologies (USA)
ATP2A2	TTTGGCTTGGTTTGAAGAAG	CGATACACTTTGCCCATTTC	Integrated DNA Technologies (USA)
KCNA4	CCCACCCAGGATCATTCTT	TCATGCAGAAGAAGCACTTCAC	Integrated DNA Technologies (USA)
GJA1	TGGGTCCTGCAGATCATATT	TCGCATTTTCACCTTACCAT	Integrated DNA Technologies (USA)
COL1A1	CCTGTCTGCTTCCTGTAAACTC	GTTCAGTTTGGGTTGCTTGTC	Integrated DNA Technologies (USA)
POSTN	CTAATGGGGTTGTCACTGTT	GTTTCTCAAAAGCCTCATTG	Integrated DNA Technologies (USA)
ACTA2	TGACCCTGAAGTACCCGATA	CGTCCAGAGGCATAGAGAGA	Integrated DNA Technologies (USA)
TCF21	TCCTGGCTAACGACAAATACGA	TTTCCCGGCCACCATAAAGG	Integrated DNA Technologies (USA)
TGFβ1	TTGCTTCAGCTCCACAGAGA	TGGTTGTAGAGGGCAAGGAC	Integrated DNA Technologies (USA)

Statistical analysis

Statistical analyses were performed using GraphPad Prism software (GraphPad Software Inc., CA, USA). For comparison between two groups, either the Student’s t-test or the Mann-Whitney U test was used, depending on whether the data satisfied assumptions of normality evaluated by the Shapiro-Wilk test. For comparisons among multiple groups, one-way ANOVA followed by Turkey’s post-test or the Kruskal-Wallis test was used accordingly, as indicated in each figure legend. Data are presented as the mean value ± standard deviation. Experiments were performed in triplicate. A significant difference was defined as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.