From Fiber Architecture to Functional Attachment: A Clinically Relevant, Mechanically Tunable Cardiac Patch

ABSTRACT

Contractile engineered cardiac patches hold great potential for treating myocardial infarction, serving as biological ventricular assist devices (BioVADs). However, optimal design and attachment of cardiac patches remain insufficiently explored, although both are essential for the mechanical support of damaged hearts. This study presents a platform for personalized macroscale patches with a multi‐zonal microarchitecture combining a regenerative zone for cell alignment, a stiff force transmission zone for load transfer, and an elastic attachment zone enabling integration. Based on computational modeling, the design is implemented using a custom G‐code generator for melt electrowriting (MEW). Digital image correlation reveals up to a 2.6‐fold strain difference between scaffold zones under physiological deformation, confirming zonal interplay. Biaxial testing with preconditioning shows scaffold mechanics replicating native myocardium properties up to 10% strain. For epicardial suture attachment, a reinforced outline enables shape‐morphing and increases suture retention 2.16‐fold. Dynamic BioVAD cultivation with fibrin‐embedded cardiomyocytes significantly (p = 0.01) improves cell alignment versus controls. Finally, in a porcine myocardial infarction model, the BioVAD achieves complete epicardial attachment and vascular ingrowth within 7 days, compared to partial attachment in controls. This study highlights MEW as a versatile platform for tailoring cardiac scaffold mechanics to support tissue integration and cardiac function.

A personalizable platform featuring a multi‐zonal cardiac patch design at clinical scale enables precise customization of fiber architecture at micro‐ and macroscale via melt electrowriting. Distinct scaffold zones support cell alignment with mechanical force transfer and epicardial attachment. In a porcine myocardial infarction model, the patch integrates within 7 days with capillary ingrowth, demonstrating strong potential to restore heart function.

1. Introduction

Cardiovascular diseases, particularly myocardial infarctions (MI) and ischemic heart disease (IHD), are the leading cause of death worldwide, accounting for approximately one‐third of all fatalities [1, 2]. The resulting lack of oxygen and nutrients leads to cardiomyocyte necrosis and a consequent decline in contractile function. This is followed by the disruption of the organized, anisotropic microenvironment and the formation of fibrotic scar tissue, impairing coordinated myocardial contraction. Progressive adverse remodeling, including ventricular dilation and reduced ejection fraction, ultimately results in heart failure (HF) [3]. Current treatment options, including pharmacological and mechanical interventions, mainly manage symptoms and remain inadequate with only a 50% survival rate after 5 years for HF [4]. Heart transplants represent the only viable option for end‐stage HF, despite the risks of rejection and its limitation due to donor shortage of compatible organs. Traditional Ventricular Assist Devices (VAD) provide mechanical circulatory support and serve either as bridge to heart transplantation or as destination therapy [5, 6]. Although VAD technology continues to evolve, there are significant risks and drawbacks, including the need for anticoagulation therapy, increased risk of bleeding, and the risk of thromboembolic events [7]. Furthermore, its inability to adequately adapt to varying demands, to the growth of the patient, or simply to regenerate or repair damaged cardiac tissue highlights the need for alternative, regenerative approaches.

As cardiomyocytes are terminally differentiated and incapable of significant mitotic proliferation, novel tissue engineering approaches have emerged since the early 2000s [8]. Over the years, this field progressed toward the creation of functional tissue in vitro to be implanted onto the damaged myocardium. These engineered heart tissues (EHT) evolved into patches to provide contractile function to the damaged heart [9]. With the advent of human induced pluripotent stem cell‐derived cardiomyocytes (hiPSC‐CM), the EHT have gained increasing interest by eliminating both the need for human donor cardiomyocytes and the risk of xenogeneic immune response [10]. However, the initial enthusiasm has been tempered by major challenges in clinical translation [11]. As reviewed by Madonna et al., these include low cell retention and survival rates due to the hostile environment along with poor structural and functional integration into host tissue [12]. Despite these challenges, the group of Zimmermann et al. recently published encouraging preclinical data that enabled approval for a first‐in‐human clinical trial and showed promising initial results [13]. However, the ideal architectural patch design, including its mechanical properties and attachment strategy, remains largely unexplored. Notably, the mechanical composition and the attachment to the native myocardium are critical to overcoming the key challenge of tissue integration with an effective force transfer to the damaged heart, thereby enabling cardiac support.

Due to the emerging (bio)fabrication technologies, novel tissue engineering strategies offer promising solutions to overcome these structural challenges of cardiac patches. The primary fabrication methodologies can be categorized into molding, 3D (bio)printing, and electrohydrodynamic (EHD) processing [14]. The 3D (bio)printing approaches enable the production of more intricate structures with higher resolution compared to molding [15, 16]. Nevertheless, the resulting constructs often exhibit mechanical properties that are insufficient for the intended application. To overcome this limitation, polycaprolactone (PCL) scaffolds have been incorporated as mechanical reinforcements [17].

Processes based on EHD principles enable the fabrication of biomimetic matrices. Electrospinning, for example, holds the potential to create elastic meshes that mimic the collagen fiber nanoarchitecture of cardiac tissue with controlled pore size [18, 19]. While Melt Electrowriting (MEW) operates on a similar principle, it offers the distinct advantage of controlled and precise fiber deposition at the micrometer scale. Due to an electric field and computer‐controlled axes of the collector plate, a molten polymer jet can be positioned in a wide range of patterns. Over the years, PCL has become the polymer of choice for biomedical applications of MEW due to its excellent processability, biodegradability, biocompatibility, and Food and Drug Administration (FDA) clearance for certain applications [20]. As a layer‐by‐layer approach, MEW is also classified as a 3D printing technology that generates a porous scaffold [21]. Due to the highly versatile scaffold architectures that mimic the extracellular matrix, MEW has been proven as an outstanding technology for soft tissue engineering. This has been demonstrated in a wide range of applications, including but not limited to osteochondral tissue [22], ligaments [23], skin [24], tympanic membrane [25], blood vessels [26], heart valves [27, 28, 29], and cardiac tissue [30]. For heart valves, Vernon et al. focused on reproducing the interface of soft tissue mechanics and collagen fiber organization in vitro. While this shows great potential, scaffolds of 1 × 4 cm were manufactured only on a microstructural level of 5‐layers, which will limit the current clinical translatability [27]. Moreover, MEW revealed great potential for myocardial tissue formation and has the ability for electroconductivity by polypyrrole (PPy) coating [31, 32]. A hexagonal microstructure of the fibers increased the cellular beating rate, enhanced cell alignment, and promoted cardiac maturation via enhanced presence of related marker expression as compared to a rectangular‐pore conventional scaffold [33]. In a follow‐up study, these structures were integrated with an extrusion‐based bioprinting process to pattern a pre‐vascular network, facilitating the future maturation of thick scaffolds [34]. However, current MEW scaffolds remain too small for the treatment of clinically relevant myocardial tissue defects and have only focused on the microstructural features, while the macroscopic design and mechanics for force transmission and attachment to the native myocardium have been largely overlooked.

Here, we present a platform that goes beyond the current state of the art, which reproduces the interface of soft tissue mechanics on a microstructural level. Our BioVAD scaffolds have been upscaled for application and combine the microstructural fiber architecture within a clinically relevant macroscale design, comprising 28 layers spanning up to 11cm × 11 cm. Further, they have been successfully implemented in a large‐animal model, highlighting their potential as a therapy for myocardial infarction. Its key novelty lies in the unique multi‐zonal design, which addresses the critical gap in cardiac patch attachment by employing the tailorable micro‐ and macroarchitecture that serves two main purposes: An inner regenerative zone (RZ) with a microstructure that promotes the alignment and maturation of contracting hiPSC‐CMs that is surrounded by an outer attachment zone (AZ) and force transmission zone (FTZ) to enable integration and effectively transfer mechanical circulatory support to the heart.

The inner fiber architecture follows a hexagonal fiber pattern, which has been previously found to be advantageous for hiPSC‐CMs organization and scaffold integrity [33]. The outer scaffold zones are designed with a combination of stiffer and more elastic structural elements to enable both tissue integration and functional force transfer, whilst avoiding impeding ventricular filling. This macroscopic BioVAD design is based on input from computational modeling [35, 36, 37]. Personalization and integration of different fiber architectural elements are achieved using a Python‐based G‐Code generator. Individual zones of the BioVAD scaffold and their interfaces are mechanically analyzed. Subsequently, an in‐depth characterization of the entire BioVAD scaffold is conducted, including biaxial testing with comparison to native cardiac tissue, and assessment of its attachment to myocardial tissue. The dynamic culture of the BioVAD with hiPSC‐CMs results in alignment harmonized with the macroscopic design. Finally, the BioVAD scaffold is tested in a custom chamber using digital image correlation (DIC) to analyze the zonal interplay and is implanted in pigs after myocardial infarction (MI) to assess its attachment and integration in vivo.

2. Results

2.1. BioVAD Design Framework Guided by Computational Modeling

The overall design of the BioVAD (Figure 2A) must fulfill a range of requirements. On a microscopic level, the most critical need is achieving homogeneous cell alignment to enable the formation of a functional syncytium in the RZ. From a macroscopic perspective, the force generated by this RZ must be effectively transmitted to the native heart via the FTZ to provide adequate contractile support. The additional AZ aims for a smooth interface of the BioVAD to the epicardium without micromovements to enable an effective transplantation and tissue integration.

FIGURE 2.

Zonal characteristics and interface. (A) Schematic illustration of the zonal composition of the BioVAD scaffold, with stereomicroscopic images showing a hexagonal fiber pattern in the regenerative zone (RZ), a rectangular fiber pattern in the force transmission zone (FTZ), and a sinusoidal fiber pattern in the attachment zone (AZ). In reference to the RZ, the mechanical properties were tailored in both directions: increased stiffness along the force transmission direction (B), and enhanced elasticity along the attachment direction (D). Scaffold zones are compared by (i) stress–strain curves, (ii) E‐Modulus, and (iii) UTS. Images of (C) and (E) depict tensile testing at the interface (i) between the regenerative zone and the attachment/force transmission zones, confirming the adapted differences in scaffold stiffness and strain (ii) along both directions.

To give insight into the basic principles behind BioVAD function, a simplified 1D model was formulated, representing the central part of the BioVAD along FTZ‐axis, the direction of force generation (Figure 1A, left panel). It consists of a central RZ, connected on either side to an FTZ. The model excludes the AZ, as it is outside the principal direction of load transfer and does not affect the mechanical behavior of the RZ and the FTZ. Material properties of the RZ consist of a passive component and an active component. The passive component is modeled by a nonlinear stress‐strain relation according to experimental data acquired from mice cardiac muscle (Figure 1A, right panel) [38]. Active material properties are modeled to depend on strain (Figure 1A, right panel), time, and shortening velocity according to experiments performed by Janssen et al. [39]. The ability of activated RZ tissue to generate more stress at higher stretch is crucial for the interpretation of the model results. Material properties of the FTZ were assumed passive, similar to those of the RZ.

FIGURE 1.

1D computational simulations to support BioVAD design principles. (A) Model representing the central part of the BioVAD in the FTZ direction, including the RZ and the FTZ (left). The BioVAD is subjected to an experiment consisting of a passive stretching phase I (mimicking LV filling), an isometric contraction phase II (isovolumic contraction), an isotonic contraction phase III (LV ejection), and an isometric relaxation phase IV (isovolumic relaxation) (middle panel). Active and passive material stress‐strain relations are shown in the right panel. (B) Stress–strain results from a reference model for the RZ (solid red lines) and the FTZ (solid blue lines). The area enclosed in RZ loop represents work density. Dotted lines indicate results with either a stiffer FTZ (left panel), a shorter FTZ (middle panel), or a stiffer RZ (right panel). (C) Work density in RZ with variation of either RZ and FTZ stiffness (left panel), or FTZ stiffness and length (middle panel), or RZ stiffness and FTZ length (right).

To investigate design aspects of the BioVAD, a numerical experiment was performed to mimic the mechanical load experienced by the BioVAD when placed over the epicardium. The results were evaluated in terms of stress, defined as the force per unit of cross‐sectional area in the initial unloaded state, and strain, defined as the change in length divided by the length in the unloaded state. Throughout the experiment, the condition of mechanical equilibrium must be satisfied, implying that the force (and thus our stress measure) in the RZ and the FTZ must be equal at any moment. The experiment consists of a passive loading (phase I, mimicking passive left ventricular (LV) filling), an isometric contraction (phase II, mimicking LV isovolumic contraction), an isotonic contraction (phase III, mimicking LV ejection), and an isovolumic relaxation (phase IV, mimicking LV isovolumic relaxation) as shown in Figure 1A (middle).

A reference situation was defined in which the length of the two FTZ combined equaled 30% of the total length of the BioVAD. The mechanical behavior of this model is represented by stress‐strain plots in Figure 1B, for the RZ (solid red) and the FTZ (solid blue). During phase I, an increasing preload stress is applied to both ends of the BioVAD. Since the RZ and FTZ have identical passive material properties, they stretch by the same amount until state 1 in Figure 1B. During phase II, the RZ is activated while the total BioVAD length is kept constant. The increasing stress in the RZ must be balanced by the stress in the FTZ. This can only be achieved by lengthening the FTZ, causing the RZ to shorten in view of the constant BioVAD length. Phase III starts once the afterload stress is reached (at state 2 in Figure 1B). During this phase, stress is constant. Therefore, FTZ strain remains constant, but the RZ shortens further. As soon as the RZ cannot maintain the afterload stress anymore (at state 3 in Figure 1B), phase IV starts, during which BioVAD length is constant once again. The decline in active stress leads to shortening of the FTZ and lengthening of the RZ. The area enclosed within the stress–strain loop of the RZ represents the work density, that is, done by the tissue per unit of tissue volume. It is equivalent to the area enclosed by an LV pressure‐volume loop, where it represents stroke work. In this case, it is equal to 0.41 mJ/mL.

To investigate the relation between BioVAD design and function, the experiment was repeated with different settings. First, the stiffness of the FTZ was increased tenfold. During phase I, this leads to a reduced lengthening of the FTZ while lengthening of the RZ remains unchanged. During the isometric phase II, further lengthening of the stiffer FTZ is reduced, leading to less shortening of the RZ. This increases the ability of the RZ to generate active stress (Figure 1A, right panel). During the remainder of the experiment, this causes an increased area within the stress‐strain loop, corresponding to an increase in tissue work density. Second, the FTZ was shortened threefold, covering only 10% of the total BioVAD length. Since FTZ tissue properties remained unchanged, the stress‐strain relation was unchanged, as indicated by the blue dotted curve in the middle panel of Figure 1B, which overlaps the original solid blue curve. During phase I, stress and strain in the RZ and FTZ proceed as in the original experiment. However, during phase II, the increase in strain in the FTZ translates into a reduced increase in length, since the unloaded length is shorter. Thus, shortening of the RZ is reduced, eventually causing a similar increase in work density as in the case of the stiffer FTZ. Third, the stiffness of the RZ was increased threefold. During phase I, this causes a reduction of stretch of the RZ. This reduced RZ stretch continues throughout the experiment, causing a reduction of the ability to generate active stress and eventually causing a reduction of work density (Figure 1B, right panel).

Figure 1C shows work density in the RZ for further variations with respect to the reference model: variations of RZ and FTZ stiffness (Figure 1C, left panel), of FTZ stiffness and length (Figure 1C, middle panel), and of RZ stiffness and FTZ length (right panel). These results confirm the detailed results in Figure 1B. In summary, the model suggests that BioVAD function benefits from a compliant RZ, placing the tissue in a favorable position for force generation. Further, a stiffer and shorter FTZ helps to effectively transfer the contractile force generated in the RZ to the endogenous muscle.

2.2. Zonal Configuration and Characterization of the BioVAD

The in silico modelling suggests that effective contractile support requires several functionally differentiated zones (Figure 2A). The cellularized RZ needs to be pliable to generate contractile forces and to be coupled on both ends with the FTZ for the functional input. This FTZ will be aligned with the CM and, according to the 1D simulations, should provide a high stiffness and short length for maximal efficiency. However, attaching a stiff zone to the epicardium, which experiences high strains, is highly likely to disrupt native physiology and cause micromovements of the BioVAD that hinder tissue integration. Consequently, the additional AZ is highly elastic to absorb epicardial strains and smoothly stretches the BioVAD over the epicardial curvature, also adapting to the change in volume by ventricular filling. Thus, this AZ ensures a continuous fit to the epicardium, promoting both attachment and integration. The AZ is therefore oriented perpendicular to the force axis, facilitating a tight BioVAD‐tissue interface. The RZ is positioned centrally between the force transfer and attachment zones.

For the optimal BioVAD design, the mechanical properties of individual zones are tailored to their specific requirements. MEW has proven to be an exceptional technology for customizing the mechanical properties of soft tissue [26, 27, 28, 29]. Referring to the literature, a hexagonal fiber architecture represents the current standard for EHT, including CM alignment [33]. Therefore, a hexagonal scaffold (HEX, 10‐layer with 700 µm fiber spacing 400 µm side length) was taken as a reference. Based on this given standard for the central region, the surrounding scaffold properties were tailored along the axis of the AZ direction and FTZ direction (Figure 2A). To enable a force transfer to the myocardium along the FTZ direction, a stiffer fiber architecture with a rectangular fiber pattern was selected (blue arrow in Figure 2A). By varying the fiber spacing (FS, 400, 700, and 1000 µm) of the 10‐layer scaffolds, the E‐modulus can be tailored to specific mechanical requirements (Figure 2Bi‐iii) and was found to be 11.76 MPa (9.97–13.04 MPa), 7.48 MPa (6.95–7.87 MPa), and 5.01 MPa (4.63–5.59 MPa), respectively (data is presented as median with interquartile range: Q1–Q3) All fiber spacings exhibited significantly higher stiffness compared to the that of RZ (E = 0.88 MPa [0.77–1.28 MPa]), with the 400 µm fiber scaffold showing an 13.36‐fold increase in E‐modulus. Additionally, the ultimate tensile stress (UTS) varied by a factor of 1.97 (400 µm), 1.21 (700 µm), and 0.89 (1000 µm) relative to the RZ (UTS = 0.2838 MPa [0.2797–0.3106 MPa]). For further experiments, a fiber spacing of 700 µm was selected as it provides structural consistency between the zones. Image analysis of tensile testing of the interface confirmed the findings (Figure 2Ci‐ii). During tensile testing, the FTZ largely preserved its geometry with a strain of 63.43% (37.53–70.68%), while the RZ experienced a significantly higher strain of 236% (229–263%), resulting in stretching to failure with a 3.72‐fold strain increase.

Along the axis of the AZ direction (green arrow in Figure 2A), the aim was to provide a higher elasticity than the RZ. Therefore, a sinusoidal fiber pattern was selected for the AZ. Uncoiling of fiber curvature allows an elastic stretching that mimics the extracellular matrix (ECM) of soft tissue. Hence, these structures have the ability to recapitulate nonlinear and anisotropic mechanical behavior, particularly of cardiovascular connective tissue [40]. Typically, the stress‐strain graphs exhibit a J‐shaped curve representing the gradual stretch and alignment of wavy collagen fibrils. By altering the amplitude, the initial stretch of the J‐shaped curve, also known as the toe region, can be adjusted to the required elasticity (Figure 2Di). In the G‐code, three distinct amplitude magnitudes were implemented via the input parameter “AZ curvature offset” in the graphical user interface, resulting in small (312 µm [310–315 µm]), medium (340 µm [336–342 µm]), and large (444 µm [443–447 µm]) amplitudes, as characterized in the supporting information (Figure S1A). The small amplitude of 312 µm was chosen as the lower bound, as it showed similar initial‐phase stretch and no significant difference in elastic modulus compared to the reference RZ (E  =  0.57 MPa [0.52–0.64 MPa] versus 0.65 MPa [0.63–0.67 MPa] for the hexagonal pattern). The upper amplitude of 444 µm corresponds to a toe region of 25% with a maximum load‐bearing capacity of ∼0.1 N (≈0.125 MPa). The intermediate amplitude of 340 µm was defined by the input factor between the lower and upper bounds. With increasing amplitude from medium to large, the scaffolds exhibited progressively increased stretch and significantly reduced E‐modulus values of 0.32 MPa (0.29–0.34 MPa) and 0.31 MPa (0.28–0.32 MPa), respectively. This allows tuning the stiffness of the AZ up to 2.10‐fold lower than that of the RZ (Figure 2Dii). As the epicardial surface is typically exposed to strains up to 25%, the medium amplitude of 340 µm was selected for further investigations, providing sufficient elasticity to accommodate physiological strain while maintaining structural integrity [41]. All sinusoidal scaffolds showed significantly lower UTS than the hexagonal architecture (Figure 2Diii). So even under extreme conditions, the AZ will be torn apart and keep the BioVAD on the epicardial surface, while the RZ mostly retains its required structure. Again, those results were verified in a tensile interface testing (Figure 2E). The image analysis revealed that the AZ was stretched to failure, exhibiting significantly higher strain of 280% (272–284%) compared to the RZ with 19% (16–28%). This corresponds to a 14.74‐fold increase in strain, while the RZ retained its structural integrity (Figure 2Ei,ii). In summary, it is hypothesized that the RZ, located centrally between the AZ and FTZ, represents an optimal scaffold design for functional BioVAD attachment. While the elastic AZ facilitates continuous tissue contact to enable integration, the stiff FTZ provides load transfer to mechanically support the damaged myocardium.

2.3. Full BioVAD Implementation and Characterization

2.3.1. Versatile G‐Code Generator for Multi‐Scale Personalization

Unlike other additive manufacturing technologies, MEW creates scaffolds through continuous fiber deposition, preventing the extrusion from starting and stopping at will. As a very specialized technology, no commercial software is available for G‐code generation or 3D object slicing. Hence, G‐codes are often programmed manually, highly limiting print complexity. Thus, we developed a G‐code generator platform for cardiac patches to overcome these limitations (Figure 3A). By processing multiple input parameters, the generator produces a single G‐code print path, enabling seamless merging of scaffold zones and customization of their geometry and architecture at both micro‐ and macroscopic scales. The previously described design principles were thereby implemented.

FIGURE 3.

Implementation of full BioVAD scaffold. (A) The graphical user interface of the Python‐based G‐Code generator allows input of various parameters to customize the mechanics and geometry of the BioVAD at both micro‐ and macroscopic scales. (B) Three BioVAD scaffolds with varying AZ‐lengths demonstrate customization. (C) Selection of SEM images of the rectangular FTZ (i), the sinusoidal AZ (ii), the hexagonal RZ (iii), and a continuous fiber interface including the reinforced outline (iv).

The patch geometry is designed with an oval shape to minimize stress peaks. BioVAD scaffolds can be fabricated with varying layer numbers and fiber spacings (pore sizes), while individual zones can be adjusted in printing speed and overall dimensions, including curvature radii. Additionally, the AZ features tunable sinusoidal amplitude and fiber orientation, aligning fibers along the attachment direction toward the central axis via an indentation input. Optionally, the G‐code generator also allows reinforcement of the BioVAD outline by decelerating below the critical translation speed (CTS), causing the jet to coil in a semi‐controlled manner and form fiber accumulation lines. This reinforced outline facilitates BioVAD suturing onto cardiac tissue. The FTZ and AZ enclose the RZ and allow for the incorporation of an additional rectangular extension zone for mounting on a custom dynamic culturing device. Before implantation, and after maturation, this extension is cut off along the reinforced suture outline.

Figure 3B (center) illustrates the resulting BioVAD fabricated with the previously selected parameters, while scaffolds on the left and right demonstrate variations in AZ‐length, highlighting customization options. Scanning electron microscopy (SEM) images (Figure 3C) depict the described scaffold zones: the FTZ with a rectangular fiber architecture, the RZ with a hexagonal pattern, the AZ with a sinusoidal fiber arrangement, and a continuous scaffold interface with a reinforced outline. The scaffold exhibits an average fiber diameter of around 10 µm in all zones (Figure S1B, supporting information) and is fabricated with 28 layers, measures 0.55 mm in thickness, and covers a total area of 11 cm × 11 cm. After removal of the culturing extension scaffold, the implantable BioVAD sizes 5 cm × 6.5 cm, providing a RZ sufficient to cover myocardial tissue defects of up to 4 cm × 3 cm. Unlike previously reported MEW cardiac patches, which cover 4 cm × 1 cm, this is the first BioVAD designed for implantation to fully cover the clinical‐size MI area [33].

2.3.2. Revealing the Zonal Interplay by Digital Image Correlation

The three previously selected geometries with varying AZ‐lengths were compared in terms of their zonal behavior to identify the optimal shape. The selected BioVAD scaffolds were mounted in a custom culture device and subjected to dynamic mechanical stimulation designed to mimic physiological conditions. For this purpose, the scaffold extension was embedded with silicone into a circular fixation ring to ensure stable mounting. A linear actuator applied a central, cyclic deformation of 7.5 mm in the Z‐direction for 10 cycles, resulting in a convex, membrane‐like deformation of the BioVAD. To localize the strain behavior and validate the BioVAD design concept, digital image correlation (DIC) was applied during scaffold deformation (supplementary Figure S2). To the best of our knowledge, this is the first application of DIC to MEW scaffolds, allowing for localized analysis of complex scaffold structures.

Figure 4A (top row) visualizes the total strain distribution of the three samples with varying AZ size, at the peak deformation using a heat map. Videos of the entire loading cycles, including the heat map can be found in the supporting information (Video S1–S3), which present the AZ‐length in ascending order from small (S1) to medium (S2) to large (S3). For all samples, the overall design principle proves effective. The bottom row of Figure 4A illustrates the BioVAD surface strain vectors, demonstrating a homogeneous and directional strain pattern in the RZ that aligns with the intended cardiomyocyte orientation toward the FTZ. In both illustrations, the RZ and FTZ show predominantly green coloration, indicating mild total strain values, whereas the AZ appears fully dark red, emphasizing high deformation. To further quantify this behavior, strains were recorded at selected points in each zone (9 within the RZ, and 3 each in the FTZ and AZ). Figure 4B plots the total strain values of the individual zones for all three samples, with each data point representing the mean total strain across 10 peak deformation cycles. Through this quantification, a significantly higher total strain in all AZs becomes evident, while the FTZ and RZ within each group show no significant differences. For example, the AZ of the medium‐length scaffold exhibited mean total strain of 7.70% (7.06–8.67%) compared to 2.63% (2.22–3.71%) in the RZ and 2.99% (2.04–3.84%) in the FTZ. Generally, it can be said that with increasing AZ‐length, the local strain in this region also rises, with a significant difference between the small and large AZ. This further highlights the capabilities to alter the mechanics of the BioVAD with our customizable and personalizable design concept. During cyclic loading, the AZ functions as a shock absorber, providing tunable elasticity while maintaining constant deformation in the RZ.

FIGURE 4.

DIC of BioVAD scaffolds in a custom culturing device. (A) BioVAD scaffolds with increasing AZ‐length (from left to right) at peak deformation with strain visualized using a heat map (top) and strain vectors (bottom). Highlighted points in black were used for quantitative data acquisition. (B) Strain representation for each scaffold zone at different AZ‐lengths, where each data point corresponds to the mean value of a tracking point calculated from the 10 peak deformation measurements. (C) Mean strain curves of the medium AZ‐length scaffold plotted over 10 cycles, showing strain cycles for each zone over time.

The BioVAD with the small AZ exhibits artifacts of elevated strains within the RZ, indicating insufficient elasticity. Conversely, the scaffold with the large AZ has a total implantation size of 6.5 cm × 5 cm, which is at the upper implantation limit for a porcine heart. The BioVAD with the medium AZ‐length demonstrated the most homogeneous strain distribution in the RZ, representing the optimal geometry, and was therefore selected for further experiments. Figure 4C presents the average strain of the medium size BioVAD for each zone, elucidating the zonal characteristic and the static strain response over the cycles.

The performance of the small and large AZ scaffolds aligns with this principle, and the corresponding graphs are presented in the supporting information (Figure S3). In summary, DIC combined with physiologic mechanical stimulation has validated our design hypothesis and demonstrated the ability to fine‐tune the mechanical response. Finally, we highlight the potential of DIC for MEW scaffold analysis, particularly as recent advancements enable the fabrication of larger, more complex scaffolds that still lack adequate characterization methods.

2.3.3. Biaxial Characterization and the Potential of Preconditioning

A biaxial characterization of the complete BioVAD geometry was performed, focusing on the macroscopic mechanical properties that could influence the cardiac tissue response after implantation. For a future application, the BioVAD will undergo mechanical stimulation and maturation in a bioreactor prior to implantation. Thus, the aim was to analyze the effect of such preconditioning and establish a stable mechanical response. As the DIC experiments confirmed sufficient elastic stretch for the medium AZ‐length, this geometry was selected for further testing. Since myocardial strain in healthy tissue typically ranges from 15% to 20% [41], five samples underwent five consecutive strain levels (ranging from 5% to 25% in 5% increments), with three cycles per strain level. Figure 5A,B visualizes the BioVAD during the loading at the initial 0% strain and the final 25% strain with its corresponding video in the supporting information (Video S4, shown at 20 × speed). Figure 5C illustrates the evolution of the mechanical response for a representative specimen along the AZ direction, which is also consistent with the FTZ direction. The BioVAD scaffold achieves a stable response within three cycles, with the second (red) and third cycle (blue) exhibiting nearly identical responses. Furthermore, significant differences were noted between the initial and preconditioned responses, including substantial energy dissipation and an increase in the toe region in the preconditioned curves. This results in lower stiffness during the early deformation stages. The extent of preconditioning was strain‐dependent, with greater differences observed at higher deformations.

FIGURE 5.

Biaxial characterization summary. (A,B) BioVAD is undergoing biaxial testing during the initial cycle at 0% strain and the final cycle at 25% strain. (C) Cyclic preconditioning results of a representative specimen in FTZ direction; solid lines represent loading stages and dotted lines unloading stages. C1, C2, and C3 stand for cycles 1 to 3, respectively. (D) Preconditioned response at different strain levels, solid lines correspond to FTZ direction and dotted lines to AZ direction. (E) Magnified view of the cyclic stage highlighting the preconditioned response up to 30% strain. (F) Comparison of the device unconditioned mechanical response with the native cardiac tissue from [37] and [43]. Evolution of the response at (G) 10% and (H) 20% strain preconditioning. In F‐H, for BioVAD plots (blue) solid line corresponds to FTZ direction and the dotted line to AZ direction; for cardiac tissue (red and black) solid line corresponds to main‐fiber direction and dotted line to the cross‐fiber direction. Lines correspond to mean values, and the shaded area to the standard deviation.

Considering the preconditioned state as the first loading step of each strain level, five distinct responses were identified, ranging from the initial, unconditioned state (0%) to preconditioning at 20% strain (Figure 5D,E). Following cyclic loading, a final monotonic extension was applied until mechanical failure occurred (Figure 5D). Around 30% strain, a constant‐stress extension phase was observed, likely due to mechanical failure mechanisms such as plastic deformation or delamination. Given that the maximum expected strain after implantation is below 25%, mechanical failure of the structure is not expected to be a concern. Once again, experimental results support our in silico‐based design.

Figure 5E provides a magnified view of the different preconditioning levels, focusing solely on the cyclic phase. Once again, preconditioning effects on mechanical behavior are evident. Controlled cyclic loading enables modulation of the BioVAD scaffold's mechanical response within a defined strain range, resulting in a wider toe region and, consequently, reduced stiffness in that range. Additionally, a distinctly anisotropic response was observed, with the FTZ direction exhibiting greater stiffness than the AZ direction. At 20% strain, the averaged anisotropy ratio (AR = PFT/PA) across all samples was calculated to be 1.41. These trends were consistently observed across all tested samples (Figure 5F‐H). However, it should be noted that the BioVAD scaffold was tested with a reinforced outline, and all zones were analyzed in superposition, which reduces the anisotropy effect and limits the elastic toe region of the native AZ.

Finally, it is essential to compare the mechanical response of the BioVAD scaffold with that of native cardiac tissue. For this purpose, the device's mechanical profile was compared with data from recent studies on the biaxial characterization of porcine myocardial tissue from the medial free wall of the left ventricle, both under physiological conditions [42]. and 6 weeks after transmural myocardial infarction [43]. These studies employed a biaxial testing protocol similar to the one used here. Native cardiac tissue exhibits an exponential stress‐strain response, where the principal fiber direction (solid line in Figure 5F‐H) is stiffer than the transverse direction [42]. Additionally, infarcted tissue shows significant stiffening, with a more pronounced exponential response due to structural remodeling [43].

The BioVAD demonstrates greater stiffness compared to native tissue, particularly at low strain levels, but remains within the mechanical range of infarcted tissue. With increasing preconditioning, the BioVAD's mechanical response gradually converges with that of native tissue due to the previously observed widening of the toe region. At 20% strain preconditioning, the BioVAD's mechanical behavior closely resembles infarcted tissue (Figure 5H) and approximates the physiological tissue response up to approximately 10% strain. These findings suggest that precycling enhances the compliance of the scaffold at low strain levels, converging its mechanical behavior more similar to that of native tissue. Thus, an initial phase of biaxial precycling may improve cell integration within the structure and better mimic physiological tissue mechanics, ultimately enhancing implantation outcomes.

2.4. Reinforced Suturing as a Novel Approach for Cardiac Patch Attachment

Although the attachment strategy of patches to the myocardium is crucial for enabling tissue integration and functional support, it remains a significant challenge [14]. Various strategies have been explored in preclinical models, including physical approaches such as (micro)needle‐based designs and chemical adhesives based on functional groups like catechol or dopamine [44, 45, 46]. However, these approaches have not been translated to clinically relevant, large‐animal or human applications. In clinical practice, surgeons rely on conventional suturing or fibrin glue to attach cardiac patches. In our study, we focused on comparing these clinically relevant strategies and additionally investigated an innovative reinforced outline in the BioVAD design to improve suture retention with force distribution and shape‐morphing, thereby enabling functional attachment while maintaining scaffold integrity. To benchmark these strategies, we also evaluated a medical‐grade cyanoacrylate, which is known for its high adhesive strength, but has limited use in the cardiovascular field due to its toxicity and stiffness. A custom‐built setup was used to mount porcine epicardial tissue and evaluate the adhesive strength of the interface with a rectangular scaffold during shear testing (Figure 6Ai). As expected, the cyanoacrylate demonstrated a strong and reproducible adhesive strength (Figure 6Aii,iii), leading to a plastic scaffold deformation with an ultimate tensile force (UTF) of 1.12 N (1.08–1.14 N). However, it was only used as a reference, and further testing was discontinued due to its toxicity. Fibrin glue, commonly used in clinical applications, exhibited very poor adhesive strength of 0.46 N (0.27–0.73 N) with low reproducibility and also creates a possible barrier for functional integration [47]. Suturing the scaffold to the epicardium resulted in a strong and reproducible interface with an UTF of 0.84 N (0.83–0.90 N). However, this connection showed lower stiffness compared to cyanoacrylate due to localized, point‐based loading and scaffold deformation. This remains the major drawback of suture connections.

FIGURE 6.

BioVAD attachment. (A) Custom‐built set‐up to mount epicardial tissue (i) for adhesive shear testing of attached scaffolds. Force‐strain (ii) and UTS (iii) comparison of scaffolds attached to epicardium by cyanoacrylate‐, fibrin glue, and suturing. (B) Visualization of suture retention of scaffolds with and without outline (i), highlighting the benefits of force distribution enabled by reinforcement through the force‐strain curve (ii), UTF, and failure angle (iii). (C) Uniaxial stretching of the complete BioVAD with reinforced outline along the AZ direction (i) results in a characteristic out‐of‐plane convex curvature (ii). A surgical planning experiment (iii) on a contractile rubber heart demonstrates that suturing and stretching in AZ direction induces a shape‐morphing response, conforming to the epicardial contour, thereby providing a continuous tissue interface.

To overcome this limitation, a reinforced outline was integrated into the scaffold, as described in Section 2.3.1. This outline was further investigated in a suture retention test and compared to scaffolds without reinforcement. As visible in Figure 6Bi scaffolds without reinforcement experienced localized loading, leading to premature suture retention. In contrast, the reinforced outline enabled a homogeneous force distribution across all scaffold fibers, preventing failure. This was further confirmed by the force‐strain diagram, which showed that the outline contributed to increased stiffness and a rise in UTF from 0.84 N (0.77–0.88 N) to 1.81 N (1.72–1.82 N), which corresponds to a 2.16‐fold suture retention (Figure 6Bii). An additional readout parameter is the failure angle, which enables the quantification of the force distribution of the suture‐scaffold loading. This failure angle significantly increased from 19.8° (19.22–20.76°) to 72.79° (71.91–73.31°) with reinforcement, further supporting our hypothesis (Figure 6Biii). In conclusion, the reinforcement line, created by printhead translation below the CTS, represents a simple yet effective innovation that enables suturing of MEW scaffolds.

Based on the clear improvement in suture retention, the reinforcement was implemented as a continuous outline around the entire BioVAD scaffold. To fully exploit the BioVAD design during surgical attachment, the scaffold is intended to be stretched along the direction of the elastic AZ to conform to the curvature of the heart. To analyze the influence of the highly stiff reinforced outline on this procedure, the complete construct was subjected to uniaxial deformation along the AZ direction (Figure 6Ci). Upon tensile loading, the BioVAD scaffold exhibited a characteristic convex curvature predominantly oriented along the tensile axis (Figure 6Cii; fully visible in Video S5 in the supporting information). This behavior arises from the mechanical mismatch between the reinforced outline and the more compliant inner scaffold. The outline constrains in‐plane deformation, while the interior accommodates strain via out‐of‐plane deformation.

This concept was further evaluated in a surgical planning experiment using a contractile rubber heart. During this test, the BioVAD scaffold was sutured with two stitches in each AZ and thereby stretched over the epicardial surface. This stretching resulted in the described convex curvature, demonstrating a shape‐morphing response that closely followed the epicardial contour (Figure 6Ciii; fully visible in Video S6 in the supporting information). For complete attachment, two additional sutures were placed in each FTZ. Importantly, this experiment further highlighted that the AZ absorbs contractile deformations, generating a curvature that gently presses against the epicardial tissue, supporting intimate contact and facilitating early tissue integration.

2.5. BioVAD Architecture Enhances hiPSC‐CM Alignment under Mechanical Stimulation

To investigate the impact of BioVAD architecture on hiPSC‐CM orientation, we used a mixture of hiPSC‐derived CMs and cardiac fibroblasts (CFs) embedded in a fibrin hydrogel in a 9:1 ratio and cast on BioVAD scaffolds. A total of 20 million cells were seeded into the BioVADs in different areas distributed throughout the RZ. A recently published protocol has further demonstrated the scalability toward clinically relevant cell numbers [48]. To reproduce the native mechanical deformation of a contracting heart, 14‐day‐old cellularized BioVADs were introduced into a custom maturation chamber and subjected to mechanical stimulation for up to 7 days. Within the dynamic culture, the contractions became synchronized and developed a strong contractile force, resulting in visible macroscopic scaffold deformation (Video S7 of the BioVAD scaffold at day 14 and Video S8 at day 21 are available in the supporting information). Cellularized scaffolds from the RZ, only consisting of a hexagonal structure, were cultured as a control under static conditions without mechanical stimulation for an equivalent period of time. The control group without dynamic culture developed less synchronized contractions with lower force, barely causing any macroscopic scaffold deformation (Video S9 of the control scaffold at day 14 and Video S10 at day 21 are available in the supporting information). After culture, both cardiac constructs were characterized by histological analysis, and hiPSC‐CMs orientation was examined as described in the supporting information (Figure S4). Tissues exhibited a homogeneous cellular distribution throughout the scaffolds. Human iPSC‐CM strongly expressed cardiac α‐actinin contractile protein, displaying well‐arranged sarcomeres (Figure 7A). Importantly, cells in BioVADs showed a significantly different directionality, with sarcomeres aligned along the direction of the FTZ (p = 0.0101, Figure 7B). Quantification of sarcomeric filament alignment in BioVADs resulted in a distribution of 29.52 ± 3.18% sarcomeres aligned in the range of 0–40° with respect to the direction of the FTZ, with a higher peak between 30 and 40° (10.94%), compared with 21.99 ± 0.57% in control samples, demonstrating that the geometry of the FTZ provides lower stresses to guide the orientation of the hiPSC‐CM. In contrast, control scaffolds resulted in wider frequency distributions resulting in a predominantly orientation of 30.66 ± 1.59% between 110 and 150° versus 19.34 ± 0.86% in BioVADs. These results indicate that optimal, application‐based design of advanced MEW printed structures combining different architectures can provide the mechanical requirements and the alignment cues for contracting hiPSC‐CMs, enhancing patch contractile functionality.

FIGURE 7.

Morphology and organization of human cellularized BioVADs: (A) Confocal immunofluorescence analysis of hiPSC‐derived cells seeded on BioVAD scaffolds, showing representative images of the hexagonal structure fully filled with α‐actinin + hiPSC‐CMs and VIM + hiPSC‐CFs. White lines indicate PCL fibers. Scale bars: 100 µm; higher magnification (right), 20 µm. (B) Polar plot showing the distribution of filament directions (*p = 0.0101). N = 3 independent experiments.

2.6. The BioVAD Demonstrates Strong Epicardial Attachment and Integration in a Large‐Animal Myocardial Infarction Model

The BioVAD scaffold shown in Figure 8A was compared with a control patch having a standard hexagonal scaffold architecture [33], with both groups (n = 2) tested in a clinically relevant porcine model. The surgical application of the standard patch on a non‐infarcted heart was suboptimal, which prompted us to improve the design prior to proceeding with the application on the infarcted porcine heart. In our myocardial infarction model using an ischemia‐reperfusion protocol, the objective was then to evaluate the novel design's impact on attachment and integration, demonstrating its safety and feasibility as an epicardial device. Myocardial infarction was induced four weeks prior to implantation of the BioVAD to reflect the clinical circumstances when additional support is required.

FIGURE 8.

In vivo comparison of the novel BioVAD design with a standard patch. (A) Demonstrates the enhanced attachment performance of BioVAD during surgery, as well as the improved interface observed in (B) MRI imaging at 1 week post‐surgery, showing firm attachment of the clips attached onto the BioVAD scaffold (thin red arrows) but presence of air caused by detachment in the control patch (thick red arrows). (C) Postmortem analysis reveals a robust macroscopic attachment of the BioVAD to the epicardium of the porcine heart, compared to an incomplete attachment of the standard patch after 1 week. (D) Visualizes the cross‐sectioning process for histological analysis (E), which shows cellular infiltration around the scaffold (HE stain), in part consisting of CD45‐positive inflammatory cells as well as host‐derived capillaries (red arrows, lectin‐stained) and arterioles (red arrows, smooth muscle actin‐stained) growing into the BioVAD (red stars), indicating biological integration of the scaffold with the epicardial surface.

The BioVADs were implanted via thoracotomy, positioned on the epicardium with the FTZ aligned obliquely toward the apex from the transversal plane. This orientation ensured optimal FTZ positioning relative to myocardial stress fibers [35]. The attachment was performed according to the procedure described in Figure 6Ciii with four sutures at the AZ, allowing the BioVAD to stretch over the curved surface, followed by four additional sutures at the FTZ to ensure a tight fit and effective energy transmission. Already during implantation, the novel design demonstrated superior performance compared to the standard patch, which was similarly attached using 4–6 sutures. Stretching the AZ resulted in a snug and uniform epicardial interface, whereas air pockets accumulated between the standard patch and the epicardium.

Following implantation, the pigs were monitored for seven days. Electrocardiography (ECG) confirmed the absence of major arrhythmias. Magnetic resonance imaging (MRI) scans were performed on day 7 to track the position and confirm the correct attachment of the scaffolds by ligation clips (Figure 8B). The initial observation of the attachment was confirmed by post‐mortem examinations on day 7, which revealed full and firm adherence to the epicardial surface. In contrast, the standard scaffold showed incomplete attachment on postmortem examination of the normal porcine heart (Figure 8C). Cross‐sectioning with histological analysis further supported the biocompatibility of these results and demonstrated robust host‐derived capillary (lectin‐stained) and arteriolar (smooth muscle actin‐stained) ingrowth into the BioVAD scaffold (Figure 8D,E). The cellular infiltrate at the interface between BioVAD and myocardium likely represents a foreign body host reaction, in part mediated by CD45 ‐positive cells in non‐immunosuppressed pigs. Vascular ingrowth observed after just 7 days provides strong evidence of successful biological integration and further supports our hypothesis that the BioVAD does not require additional vascularization but is sufficiently supplied through diffusion and native vascular ingrowth. As PCL is known for its comparatively slow degradation kinetics, this characteristic represents a distinct advantage, enabling the scaffold to sustain mechanical support and directional guidance during ongoing tissue integration and maturation [49, 50].

Taken together, these results demonstrate the safety, feasibility, and biocompatibility of large clinical‐scale BioVADs, validating our superior design approach to overcoming current challenges in patch attachment and integration. These findings provide a solid foundation for further preclinical studies using innovative biological therapeutic strategies. Long‐term studies in a larger cohort are currently ongoing to evaluate the effect of cellular input on cardiac function by MRI. Future work will focus on mechanical and electrical maturation to enhance the patches' force generation already before implantation and further enhance their potential to improve their circulatory support.

3. Conclusive Discussion and Outlook

To contextualize the BioVAD platform within the recent landscape of cardiac patch technologies, we performed a comparative analysis against representative state‐of‐the‐art strategies (Table 1). Each of these technologies has advanced specific aspects of myocardial repair, such as promoting cellular alignment (fiber‐based constructs [33]), enabling conformal adhesion (Janus hydrogels [46]), or enhancing electrical coupling (conductive hydrogels or PPy‐coated MEW fibers [32, 51]). However, they remain limited in one or more key translational dimensions. Hydrogel‐ and Janus‐based patches, while biocompatible and adaptable, typically lack controlled anisotropy and mechanical robustness, restricting their ability to transmit contractile forces effectively to the native myocardium. Conductive hydrogels improve electrical conductivity and adhesion but often do not incorporate aligned cardiomyocytes to support macroscopic force transmission at clinically relevant scales. Fiber‐based scaffolds promote cardiomyocyte alignment but lack macroscopic coupling to the native myocardium and have so far been demonstrated primarily at small, non‐clinical scales.

TABLE 1.

Comparative overview (sorted by year) of state‐of‐the‐art cardiac patch technologies and the BioVAD platform. Legend: −: Not present / o: Limited / +: Present / ++: Extensive.

Reference	Fabrication	Macro‐architecture	Mech. characterization	Customizability	CM alignment	Size	In vivo model
BioVAD	MEW & Fibrin, hiPSC‐CM/CF	++ Multi‐zonal design for functional attachment	++ DIC, biaxial, uniaxial and attachment testing	+ Micro/macro scale via G‐code	++ Fiber guidance & dynamic culture	++ Clinical‐scale	+ Porcine MI model
[52]	Bilayer Natural hydrogel	+ Janus	+ Attachment, uniaxial, friction	− Not reported	− Not reported	− Small patches	o Rat
[13]	Collagen & hiPSC‐CMs	− Isotrope design	− Limited characterization	o Patch layering	− Not reported	++ Clinical‐scale	++ Primates
[51]	Conductive Adhesive Hydrogel	o Paintable	+ Attachment	+ Individual painted	No CM incorporated	+ Medium	o Rat
[46]	Janus Hydrogel + hiPSC‐CMs	+ Janus structure	+ Attachment testing	− Not reported	− Not reported	o Small patches	o Rat
[53]	Stereolithography of Hydrogel & hiPSC	+ 4D self‐morphing	+ Uniaxial	− Not reported	+ Fiber guidance	− Small patches	o Rat
[32]	Conductive PPy coated MEW	Single zone auxetic	+ Cyclic and Poisson's Ratio investigation	− Not reported	− Not reported	o Small patches	− Not reported
[54]	Bioprinting ECM hydrogel	+ Thick, perfusable	− Limited characterization	+ personalized hydrogel & macro‐geometry	− Not reported	+ Medium patches	− Not reported
[33]	MEW & Hydrogel, hiPSC‐CMs	o Anisotropic single zone	+ Uniaxial	− Not reported	+ Fiber guidance	o Small patches	+ Porcine
[55]	Fibrin & hiPSC	− Isotrope disk like	− Limited characterization	− Not reported	− Not reported	+ Medium patches	+ Porcine MI model

Collectively, existing technologies either optimize the microscale, focusing on cellular alignment and local bioactivity or the macroscale, focusing on adhesion and macroscopic force transfer, but rarely address both simultaneously. These limitations illustrate that most current patch concepts remain insufficiently scaled, mechanically optimized, or functionally integrated for clinical translation, thereby constraining their overall translational potential.

In contrast to existing technologies, the BioVAD platform represents the first clinically scaled cardiac patch with a multi‐zonal scaffold that can be customized at both the fiber‐level microarchitecture and overall geometric design. It is therefore capable of simultaneously meeting microscale requirements to guide cardiomyocyte alignment during dynamic cultivation, while integrating these features within a macroscale architecture that promotes tissue integration and efficiently transmits contractile forces in a large‐animal porcine model. By bridging micro‐ and macroscales, the platform directly addresses remaining translational bottlenecks and positions the BioVAD as a next‐generation, clinically relevant cardiac patch. For this, the multi‐architectural BioVAD scaffold design, consisting of three distinctive zones, was based on insights from computational modeling experiments. The RZ is based on hexagonal fiber architecture, which was previously shown as state‐of‐the‐art for CM culture and cell alignment. Extending this alignment, a stiff FTZ of rectangular fibers facilitates efficient energy transmission to the myocardium. The AZ adds elastic stretchability, enabling the BioVAD to conform seamlessly to the cardiac curvature, promoting continuous interface formation and tissue integration.

The implementation of the multi‐architectural MEW scaffolds was enabled by a Python‐based G‐code generator that allows customization of the BioVAD on both a microscopic and macroscopic level to tailor mechanical and geometric properties. Mechanical testing of the individual scaffold zones validated the foundation of the design concept. The full clinical‐sized patch was then mounted into the custom set‐up, and DIC was applied to MEW scaffolds for the first time. This analysis visualized the design principle and the zonal interplay under in vivo‐like mechanical stimulation. The AZ was able to absorb high strains, while the RZ and FTZ experienced significantly lower strains for force transmission. Additionally, biaxial testing investigated the impact of mechanical stimulation and emphasized the capabilities and importance of such preconditioning. Performing these precycles allows further fine‐tuning of the mechanical response and its convergence to the properties of the native myocardium.

During scaffold attachment testing on cardiac tissue, a reinforced scaffold outline demonstrated improved force distribution at the suturing interface, addressing the major limitation of punctual suture loading. Engineered tissue generation of the regenerative zone with hiPSC‐CMs and hiPSC‐CFs under mechanical stimulation confirmed that BioVAD scaffolds provide an excellent environment for guiding CM alignment. Finally, the advanced BioVAD scaffold design was compared with a standard scaffold control in a porcine myocardial infarction model. In contrast to the control, the novel BioVAD showed a remarkable epicardial attachment within 7 days, and even facilitated capillary ingrowth, highlighting ongoing integration. To demonstrate the therapeutic use of the BioVADs, long‐term studies on pigs are currently ongoing in a larger cohort. These studies are particularly important to assess improvements in cardiac function via MRI, as well as histological evaluation of long‐term tissue formation and maturation within the scaffold, while simultaneously monitoring material degradation behavior, providing essential data to pave the way for clinical translation. To further advance the BioVAD function prior to implantation, additional maturation studies are ongoing. Taken together, our novel design strategy advances the functionality of cardiac patches by improving epicardial attachment, thereby providing a promising platform for regenerative mechanical cardiac support.

4. Experimental Section

4.1. G‐Code Generator

The G‐code generator was developed in Python (Python Software Foundation, USA) to enable the continuous fiber deposition interface of multiple regions. To achieve this, the requirements for BioVAD scaffolds were defined, and the necessary input parameters were derived accordingly. The variable parameters for the entire scaffold include the number of layers, fiber spacing, and reinforced outlines, while individual zones can be customized with dimensions, radius, printing speed, and sinusoidal amplitude. A script featuring a graphical user interface (GUI) allows users to input these parameters, which are then processed in subsequent scripts for architectural calculations. Finally, the results are passed to a script that generates the continuous G‐code and outputs it along with a graphical visualization.

4.2. Melt Electrowriting

A custom‐built MEW device consisting of a heated pressurized polymer reservoir with a high voltage source and a computer‐controlled 3‐axis collector was used to manufacture all scaffolds. A 3 mL plastic cartridge with nozzle (22 G precision tips, Nordson Corporation, USA) was inserted into the printhead and PCL (PURASORS PC12, Corbion) heated to 85°C. A pressure of 0.2 bar was applied to extrude the melt, while a high‐voltage source (HCP 14–20000, FuG Elektronik GmbH, Germany) was used to generate a 5.5 kV potential between the nozzle tip and the collector. A central control unit (Bosch Rexroth AG, Germany) was used to implement the G‐code and actuate the axes for the laydown pattern on a 1 mm glass slide. The distance between nozzle tip and collector was maintained at 4,6 mm. To ensure print path accuracy across layers, the following power function was applied to the translation speed for print path correction [56]

$$\mathit{V}=\left({\mathit{V}}_0\right)\times {\mathit{L}}^{-\mathit{\alpha }}$$

The initial translation speed V ₀ was adjusted over the number of layers L by the correction factor α  =  0.05. For the RZ V ₀ set to 375 mm/min, whereas it was increased to 400 mm/min for the attachment zone AZ and the FTZ.

For the single‐zonal characterization, only 10 layers (corresponding to a total height of 0.1 mm) of the specific scaffolds were printed, with the architecture varying as follows: rectangular scaffolds with fiber spacings of 400, 700, and 1000 µm, as well as sinusoidal scaffolds with a fiber spacing of 700 µm and amplitudes of 312, 340, and 444 µm. For the interface characterization, scaffolds combining two distinct architectures were investigated, both with a fiber spacing of 700 µm.

The full BioVAD scaffolds and the scaffolds used for attachment tests were all printed with 700 µm fiber spacing and 28 layers corresponding to a total height of 0.55 mm. Two reinforced outlines were printed with a translation of 25 mm/min side by side with an offset of 0.5 mm, resulting in one dense reinforcement line. The hexagonal RZ measured 40 × 30 mm with a radius of 50° and extended into the rectangular FTZ with a length of 5 mm. With an indentation of 10 mm the sinusoidal AZ exhibited a width of 30 mm, an outside radius of 30°. For DIC the length of the AZ was varied between 5, 10 and 15 mm, with further experiments being conducted at a standardized medium length of 10 mm.

4.3. Mechanical Testing

4.3.1. Uniaxial Tensile Testing

For the single‐zonal characterization and the interface testing, at least n = 6 scaffolds per group were cut by a laser cutter (Reyjet 50, Trotec Laser GmbH, Austria) into specimens measuring 8 mm × 20 mm. The cut scaffolds were glued onto paperboard for stabilization and mounted in the test machine (ElectroForce 5500, Bose Corporation, USA). This resulted in an active testing area of 8 × 8 mm which was subjected to a uniaxial tensile test with a translation speed of 5 mm/min to a maximum displacement of 12 mm. The stress was calculated based on the dimensions of the cross‐sectional area. This procedure disregards the high porosity; however, for the final BioVAD, these pores are filled with fibrin, and it is assumed that this will not significantly alter the tensile properties. Consequently, for each sample, the stress–strain curves were generated, and the area within the error bands was shaded grey. The E‐Modulus was calculated from the slope of the linear region and the maximum stress determined as the ultimate tensile strength. During the test of the scaffold interface, n = 4 videos were captured, and the strain in each region was measured with ImageJ (National Institutes of Health, USA).

Scaffolds undergoing the attachment test to myocardial tissue were cut into specimens (n = 4 scaffolds per group) measuring 16 × 20 mm, with a 16×8 mm area designated for gluing and tensile testing. One side of the scaffold, measuring 4 mm, was mounted onto a paper frame, while the other side was attached to the epicardial surface. The scaffold was either attached with a surgical knot (5‐0 Prolene, ETHICON Inc., USA), or the interface covered with the Fibrin (TISSEEL, Baxter Inc., USA) or Cyanoacrylate glue (TRUglue, Trusetal GmbH, Germany). The interface was then exposed to shear testing, employing the same protocol as for the zonal characterization. Force‐strain graphs were generated, and the ultimate tensile strength was determined as previously described.

In the suture retention test (n = 6 per group), the suture was stitched through a scaffold pore and mounted directly to the testing machine. The scaffolds measured a width of 20 mm, representing the distance between two sutures, and had a height of 16 mm. Of this, 4 mm on each side was used for suture or paper frame attachment, resulting in an active testing length of 8 mm. One group included an additional reinforcement line, while the other consisted of standard scaffolds without reinforcement. The previously described testing protocol was performed at a translation speed of 15 mm/min, and a video was captured. Force‐strain graphs were created, and ImageJ was used to measure the failure angle of the scaffold deformation.

4.3.2. Biaxial Testing

Cyclic equibiaxial tests were performed using a true biaxial testing equipment (Instron Planar Biaxial Soft Tissue Test System, Figure 5A) with a 10 N loading cell. Five different stretch levels were consecutively applied (5–25% in 5% increments), with 2 preconditioning cycles and 1 measuring cycle at each stretch to get a steady response. These deformation levels were selected based on the work of Nagata et al., who reported myocardial longitudinal strain generally in the 15–20% range [41]. After these consecutive cyclic loadings, a final monotonic stretch until mechanical failure was applied. A 2 mN equibiaxial preloading was applied in all samples. Samples were clamped right after the external reinforcement. Tests were performed under quasistatic conditions (5 mm/min), maintaining room temperature. Engineering stress (P) was obtained for the evaluation of the mechanical response as: Pi = Fi/(t·Lj), where Fi corresponds to the normal forces in each direction; t is the specimen thickness in the unloaded reference configuration (which was taken as 0.55 mm in all specimens); and Lj are the transverse‐to‐the‐force lengths of the sample sides in the undeformed state. Global device dimensions were considered for stress calculation, which correspond to 60 mm in the attachment direction and 50 mm in the force‐transmitting direction. A total of five samples were employed.

4.3.3. Digital Image Correlation

The BioVAD, composed of a PCL scaffold, underwent cyclic deformation administered by an actuator (Moog Inc., NY, USA) on a corresponding test bench (supplementary Figure S2). For digital image correlation (DIC) and 3D test data acquisition, GOM Correlate Pro software systems (Zeiss AG, Oberkochen, Germany) were employed, utilizing advanced point‐tracking algorithms with ARAMIS software. The bioVADs were mounted by their extension zone into the custom‐built dynamic device. Testing parameters comprised a real amplitude (displacement) of 7.5 mm, a duration of 10 cycles, and a frequency of 0.2 Hz. For data acquisition and analysis of the strain distribution through the samples, several points were selected strategically. Boxplot graphs were generated, and the different zones were statistically compared. Additionally, a strain‐time graph was plotted for each region, with each data point in the boxplot representing the mean of all 10 maximum strain peaks.

4.4. hiPSC Differentiation, BioVAD Cellularization, and Filament Orientation Analysis

4.4.1. hiPSC Cell Culture and Differentiation

Human pluripotent stem cells (hiPSC) (ESi007‐A; cell line registered in the Human Pluripotent Stem Cell Registry (hPSCreg)) were cultured on growth factor reduced matrigel (GFR‐MG) coated dishes (1:80 dilution) with mTeSR1 medium (Stem Cell Technologies) and passaged 1:20 by incubation with 0.5 mM EDTA (Invitrogen) every 4–5 days. Cells were differentiated into CMs in monolayer culture using a biphasic Wnt modulation protocol as previously described [57]. Briefly, when hiPSC achieved confluence, the medium was changed to RPMI supplemented with B27 minus insulin (RPMI B27‐) and 10 µM GSK3 inhibitor CHIR99021 (CHIR, Axon Medchem) for 24 h. After 24 h, the medium was changed to RPMI B27− for another 2 days and then replaced with RPMI B27− and 5 µM Wnt inhibitor C59 (Axon Medchem) for 48 h. After 2 days in RPMI B27−, differentiations were maintained in RPMI supplemented with B27 (RPMI B27). Beating usually started from days 9 to 10. At this point, to purify iPSC‐CM, cells were subjected to 2 cycles of 72 h in RPMI without glucose supplemented with 4 mM lactate. Finally, cells were reseeded by incubating them for 7–10 min with TrypLE (Gibco), onto matrigel‐coated 12‐well plate (1:80 dilution) at a density of 1.0 million cells per well in RPMI B27 supplemented with 10% Knock‐out serum replacement (KSR, Gibco), and 10 µM of Y‐27632 (Y27, Tocris). After 24 h, the media was changed into RPMI B27 with frequent media change every 2 days.

To differentiate hiPSC into cardiac fibroblasts (CFs), the method described by Zhang et al. was used [58]. Briefly, cells at 90% confluency were incubated with 10 µM CHIR on RPMI B27− for 48 h. After 48 h, the medium was changed to RPMI B27−, and at day 3, cells were then incubated with 5 µM C59 on RPMI B27− for 48 h. At day 5, cells were reseeded on matrigel‐coated 12‐well plate (1:80 dilution) in RPMI B27− supplemented with 5 µM CHIR, 2 µM retinoic acid (Sigma–Aldrich), 10% KSR, and 10 µM Y27. The following day, the media was changed to RPMI B27− supplemented with 2% KSR. At day 8, the media was changed to Advanced Dulbecco´s Modified Eagle´s Medium (ADMEM, Gibco) supplemented with 2% KSR, and kept for 72 h. At day 11, cells were replated on 0.1% gelatin (Merck Milipore) coated plates, and switched to Fibroblast Growth Medium 3 (FGM3, PromoCell) supplemented with 10 µM SB431542 (Sigma–Aldrich), 10 ng/mL of FGF2, and 10 µM Y27. Cells were cultured in FGM3 medium supplemented with 10 µM SB431542, changed every 2 days until their use.

4.4.2. BioVAD Cellularization and Culture

BioVAD scaffolds were treated with O₂ plasma (Diener Electronic) for 5 min in order to increase their hydrophilicity and facilitate cell infiltration. The scaffolds were then sterilized in 70% ethanol for 30 min, washed thrice in sterile distilled water, and left to dry. Human iPSC‐CM and hiPSC‐CF were detached using TrypLE and seeded in a 9:1 ratio. Cells were seeded in different areas (with a diameter of 8 mm and 0.5 cm apart) along the RZ of the BioVAD scaffold (supporting information, Figure S5). For control tissues, scaffold discs from the RZ with a similar diameter were obtained using an 8 mm biopsy punch. A total of 2 × 10⁶ cells in 35 µL of hydrogel, consisting of RPMI B27 media supplemented with 10% KSR, 1 P/S, 10 µM of Y27, and 6 mg/mL bovine fibrinogen (Sigma–Aldrich) were used for each seeding area. The volume containing fibrinogen with cells was pipetted into a Teflon mold where the BioVAD scaffold was placed, followed by the addition of 1.8 µL thrombin (100 U/mL (Biopur)). The scaffold was then incubated for 1 h at 37°C to induce fibrin crosslinking. Afterward, the BioVAD scaffolds were transferred to 10 cm diameter culture dishes with RPMI B27 supplemented with 10% KSR, 1% P/S, 10 µM of Y27, and aprotinin (0.1% (wt/vol); 33 µg/mL) (Sigma–Aldrich) to avoid fibrin degradation. On the next day, the media was refreshed with RPMI B27 supplemented with 1% P/S and aprotinin, changed every other day from then on. BioVADs were maintained in culture 14 days and then were transferred into the bioreactor chamber and cultured under mechanical stimulation for 7 days, while control tissues were cultured on 12‐well plates without mechanical stimulation for 21 days. For the in vivo application of BioVAD scaffolds on the infarcted myocardium, the preparation of the scaffolds was identical, but no cells were mixed into the hydrogel.

4.4.3. Immunofluorescence and Quantification of Filament Alignment

For staining, BioVAD scaffolds were fixed with zinc‐buffered formalin (Thermo Scientific) for 30 min at RT. After three washes in PBS, scaffolds were incubated overnight at 4°C with anti‐cardiac α‐actinin (1:400, Sigma–Aldrich, A7811) and anti‐vimentin (1:500, Abcam, ab92547) primary antibodies diluted in 1% BSA (Sigma–Aldrich) after 20 min of permeabilization in 0.2% Triton‐X100 (Sigma–Aldrich). Alexa Fluor 488‐ and 594‐conjugated secondary antibodies (1:500, Invitrogen) were employed and incubated in 1% BSA for 1 h at RT. Composite images of 9 images were taken at 8–10 different focal distances (interplane separation of 10 µm) at 40× magnification with 0.5× zoom/crop using a LSM 800 Zeiss Confocal microscope, and processed using ImageJ software (National Institutes of Health, USA). Cardiac α‐actinin confocal stacks from BioVAD scaffolds were used for quantifying sarcomere filament alignment. A custom macro was developed based on ImageJ (available in supporting material). A minimum of 10 and a maximum of 25 sarcomeres were analyzed per plane. Specifically, at least 150 sarcomeres were quantified (n ≥ 150) for each independent experiment. The user was blinded to the group of origin to avoid any bias in the analysis.

4.5. In Vivo Study

The implantation procedure was developed in a surgical planning experiment using a rubber heart equipped with a beating controller of the Chamberlain Group. Here, the BioVAD was sutured with two stitches in each AZ and thereby stretched over the epicardial surface, followed by two stitches in each FTZ. For the in vivo study, a posterolateral wall infarction (MI) in pigs was induced by 90 min balloon inflation in the circumflex artery. Four weeks after MI, the BioVAD was attached according to the established implantation procedure. Cine MRI imaging was performed to visualize the location and motion of vessel clips. Postmortem histology was performed to evaluate capillaries and arteriolar ingrowth in the scaffold region using lectin and smooth muscle actin stains.

4.6. Statistics

Boxplots display the full data range, with the box edges representing the 25th and 75th percentiles (interquartile range) and the central line indicating the median. Whiskers extend to the minimum and maximum values, and individual data points are shown as dots. Normality was assessed using the Shapiro–Wilk test, followed by a one‐way analysis of variance (ANOVA) to determine statistical significance. Statistical significance was defined as p < 0.05. Where applicable, p‐values were indicated on the graphs using asterisks to denote their magnitude: ns > 0.1234, * p < 0.0332, ** p < 0.0021, *** p < 0.0002, and **** p < 0.0001. To maintain visual clarity in graphs with multiple comparisons, statistical significance for groups with very strong differences (p < 0.0001) was not always labeled with asterisks when the difference was visually obvious.

4.7. Ethics

This investigation conforms to the Belgian National Institute of Health guidelines for care and use of laboratory animals, and the protocol of this study was approved by the local Ethics Committee for Animal Experimentation at KU Leuven (EC approval number P114/2020).

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File 1: adma72298‐sup‐0001‐SuppMat.docx.

Supporting File 2: adma72298‐sup‐0002‐Data.zip.

Supporting File 3: adma72298‐sup‐0003‐VideoS1.mp4.

Supporting File 4: adma72298‐sup‐0004‐VideoS2.mp4.

Supporting File 5: adma72298‐sup‐0005‐VideoS3.mp4.

Supporting File 6: adma72298‐sup‐0006‐VideoS4.mp4.

Supporting File 7: adma72298‐sup‐0007‐VideoS5.mp4.

Supporting File 8: adma72298‐sup‐0008‐VideoS6.mp4.

Supporting File 9: adma72298‐sup‐0009‐VideoS7.mp4.

Supporting File 10: adma72298‐sup‐0010‐VideoS8.mp4.

Supporting File 11: adma72298‐sup‐0011‐VideoS9.mp4.

Supporting File 12: adma72298‐sup‐0012‐VideoS10.mp4.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.