Self-aligned myofibers in 3D bioprinted extracellular matrix-based construct accelerate skeletal muscle function restoration

Abstract

To achieve rapid skeletal muscle function restoration, many attempts have been made to bioengineer functional muscle constructs by employing physical, biochemical, or biological cues. Here, we develop a self-aligned skeletal muscle construct by printing a photo-crosslinkable skeletal muscle extracellular matrix-derived bioink together with poly(vinyl alcohol) that contains human muscle progenitor cells. To induce the self-alignment of human muscle progenitor cells, in situ uniaxially aligned micro-topographical structure in the printed constructs is created by a fibrillation/leaching of poly(vinyl alcohol) after the printing process. The in vitro results demonstrate that the synergistic effect of tissue-specific biochemical signals (obtained from the skeletal muscle extracellular matrix-derived bioink) and topographical cues [obtained from the poly(vinyl alcohol) fibrillation] improves the myogenic differentiation of the printed human muscle progenitor cells with cellular alignment. Moreover, this self-aligned muscle construct shows the accelerated integration with neural networks and vascular ingrowth in vivo, resulting in rapid restoration of muscle function. We demonstrate that combined biochemical and topographic cues on the 3D bioprinted skeletal muscle constructs can effectively reconstruct the extensive muscle defect injuries.

I. INTRODUCTION

The demand for bioengineered skeletal muscle tissues has risen rapidly for reconstructive procedures of trauma or tumor ablation. More than 20% of the original muscle loss results in functional impairment due to the limited regenerative capacity.¹ The skeletal muscle is composed of uniaxially aligned muscle fibers developed through the fusion of mononucleated muscle cells, which are critical for the contractility of skeletal muscle and force generation for movement.² To recapitulate this structural organization, various muscle-specific biochemical and topographical cues have been applied to bioengineer functional muscle regeneration.^3,4

To obtain the spatially cellular orientation, several strategies have been tested.^5,6 In particular, topographical cues between micro- and nano-scales play a key role in cellular alignment. Furthermore, the micro- or nano-structure stimulated the myoblasts to produce extracellular matrix (ECM) components by regulating transmembrane molecules such as integrins and rearranging intracellular components.⁷ Conventionally, the techniques employed for topographical cell stimulation are restricted to 2D platforms, such as soft lithography,⁸ micro-/nano-patterning,⁹ and electrospinning;¹⁰ these approaches, however, allow single-layered myofiber bundle or micrometer-scaled tissue construction that are insufficient for the treatment of extensive muscle defect injuries.¹¹ Several strategies using molding,¹² microfluidics,¹³ and rolling¹⁴ techniques have also been applied to acquire high degrees of muscle cell alignment in 3D volumetric tissue constructs. Although these strategies demonstrated structural maturation and force-generating capacity of the bioengineering muscle constructs, they are still deficient to build clinically relevant sized volumetric muscle constructs that are composed of highly oriented multi-layered myofiber bundles.

Three-dimensional bioprinting technologies have been previously utilized to create millimeter- to centimeter- scaled muscle constructs with structural integrity.¹⁵ These muscle constructs were successfully implanted in a rat muscle defect injury model, and, more importantly, they were integrated with host nerve and vascular systems, resulting in the restoration of muscle function. Even though we have successfully created large-scaled muscle constructs, polymeric pillars have been used to induce the compaction phenomenon for cell alignment in the 3D printed cell-laden constructs.^15,16 With the polymeric pillars, the printed muscle cells were able to stretch along the longitudinal axis of the constructs, and the constructs underwent compaction, keeping the fibers taut during myofiber formation, whereas the printed cells without pillar support did not show cellular alignment. The use of the polymeric pillars, however, might be a limitation to fabricate a clinically relevant sized muscle construct because it requests a multiple material printing and limits only one direction of cellular orientation. Thus, we previously applied the topographical cues directly to the printed cell-laden constructs by the poly(vinyl alcohol) (PVA) fibrillation/leaching method to induce the self-alignment of the printed cells without the polymeric pillar system.¹⁷ The printed C2C12 myoblasts were fully oriented and effectively differentiated with a high degree of myofiber formation without polymeric pillars.

The hydrogel-based bioink system is a major component of cell-based bioprinting applications.^18,19 This bioink system should provide not only the printability and structural integrity but also the biological microenvironment.²⁰ To fulfill these prerequisites, decellularized tissue-derived ECMs (dECMs) have been introduced as hydrogel bioinks owing to the chemical and structural components of the dECMs.^17,21,22 Indeed, we previously developed a photo-crosslinkable skeletal muscle-specific dECM methacrylate (dECM-MA)-based bioink that could provide a proper microenvironment for efficient myofiber formation in the 3D bioprinted tissue constructs.^4,17 Based on the aforementioned principles on skeletal muscle bioprinting, we were able to fabricate the biochemically and topographically mimicked muscle constructs combined by the muscle-specific dECM-MA bioink and the PVA fibrillation/leaching method.¹⁷ These biomimetic muscle constructs showed the accelerated myogenic differentiation by the dECM components, while the self-aligned myofibers were achieved by the topographical cues. In this study, we further investigate the feasibility of using the bioprinted self-aligned skeletal muscle dECM constructs containing clinically relevant human primary muscle progenitor cells (hMPCs) for functional restoration of a muscle defect injury in rats.

II. RESULTS

A. Skeletal muscle-specific bioink preparation and its rheological properties

A photo-crosslinkable dECM-MA was prepared by the methacrylation of dECM obtained from biopsied porcine skeletal muscle tissue [Fig. 1(a)]. The degree of methacrylation of the dECM-MA was approximately 80%, as measured by amino group substitution degree with 2,4,6-trinitrobenzene-sulfonic acid solution (TNBS). The amount of the residual DNA in dECM and dECM-MA was noticeably decreased (less than 50 ng/mg of dry tissue sample²³), whereas major ECM components were preserved in dECM and dECM-MA [Fig. 1(b) and Fig. S1 in the supplementary material]. To induce topographical cues in the bioprinted dECM-MA constructs, PVA was used as an additive/sacrificing component. After dECM-MA/PVA printing, PVA was washed out and formed micro-scaled fibrillation in the bioprinted construct. The rheological properties, storage modulus (G′), and complex viscosity (η*), of the dECM-MA with testing modes (temperature and frequency sweep) are shown in Figs. S2(a) and S2(b) in the supplementary material. The dECM-MA showed near 37 °C of the physical gelation point because the collagen is the main component of the dECM, and the rheological properties of the dECM-MA sensitively responded to the UV exposure (300 mW/cm² for 60 s) [Fig. S2(c) in the supplementary material]. Based on the rheological results, the printability (P_r = L²/16A, where A and L represent the area and perimeter of the pore),^18,24 which was determined by the printed lattice structure of the dECM-MA bioink, was improved by the application of temperature (37 °C) in the printing plate and UV cross-linking [Figs. S2(d) and S2(e) in the supplementary material].

FIG. 1.

Preparation and characterization of dECM-MA bioinks. (a) Schematics of preparation of a photo-crosslinkable skeletal muscle-derived dECM bioink. (b) DNA content (n = 5) and ECM components; collagen (n = 4), elastin (n = 5), and glycosaminoglycans (GAGs, n = 6) in native tissue, dECM, and dECM-MA (**p < 0.01 and ***p < 0.001). (c) Rheological properties of the dECM-MA bioinks with different molecular weights of PVAs (hPVA, m-PVA, and l-PVA) containing hMPCs (2 × 10⁷ cells/ml) tested in stress sweep (n = 3; ***p < 0.001). (d) Comparison of storage modulus (G′) of the dECM-MA bioinks before and after treatment with gelation temperature (37 °C) and UV exposure (300 mW/cm² for 2 min) at 1 Hz measured using frequency sweep (n = 3). All data are represented as mean $\pm$ SD. The p-values by one-way ANOVA followed by Tukey's test are indicated.

B. Micro-topographical cues induced by bioprinting and PVA fibrillation/leaching

The different molecular weights of PVA with the dECM-MA bioink could influence the micro-topographical pattern of the printed dECM-MA constructs due to the flow behavior in a microscale nozzle during the printing process. To optimize the microfibrillation, the effect of molecular weight of PVA (l-PVA: 31–50 kg/mol, m-PVA: 89–98 kg/mol, and h-PVA: 146–186 kg/mol) with the dECM-MA bioinks containing hMPCs was examined by the rheological measurement. The rheological properties (G′ and yield stress, τ_y) of the dECM-MA bioink were significantly decreased with l-PVA or m-PVA, while an increase in G′ and yield stress was observed upon addition of h-PVA [Fig. 1(c)]. Additionally, the G′ of all dECM-MA bioinks was significantly enhanced after applying temperature (37 ± 2 °C) at the printing stage and simultaneous UV exposure (300 mW/cm² for 60 s) [Fig. 1(d) and Fig. S3(a) in the supplementary material]. After PVA leaching, the printed dECM-MA constructs presented aligned fibrous structure with m-PVA and h-PVA, while l-PVA induced randomly oriented fibrous structure [Figs. S3(b) and S3(c) in the supplementary material]. However, the physical properties of the dECM-MA constructs after PVA leaching were not significantly changed with different molecular weights of PVAs [Fig. S4 in the supplementary material].

To evaluate the cellular activities by the PVA molecular weight, the dECM-MA bioinks with different molecular weights of PVAs (l-, m-, and h-PVA) were printed with hMPCs (printing plate temperature: 37 ± 2 °C and moving speed of the nozzle = 10 mm/s). To attain the same volume flow rate (∼0.74 μL/s) of the dispensed bioinks, different pneumatic pressures were applied on the three dECM-MA bioinks (75 kPa for l-PVA, 250 kPa for m-PVA, and 320 kPa for h-PVA) (Table S1 in the supplementary material). The dispensing bioinks were crosslinked by in situ UV exposure (300 mW/cm²). The Pr values indicated the m-PVA and h-PVA dECM-MA bioinks presented more stable lattice structures compared with l-PVA [Fig. 2(a)]. The printed cell-laden dECM-MA struts with l-PVA and m-PVA showed significantly higher cell viability at 1 day after printing when compared with the bioink with h-PVA [Figs. 2(b) and 2(c)]. The orientation factor [f = (90 − ϕ_o)/90, where ϕ_o is the full width at half maximum] of the F-actin of hMPCs, fusion index [the ratio of the number of nuclei in myosin heavy chain (MHC)-positive myofibers to the total number of nuclei], and maturation rate (the ratio of the myofibers containing five or more nuclei to the total number of myofibers) were measured to examine the alignment of F-actin and myofiber formation [Figs. 2(d) and 2(e)]. The myofibers formed by hMPCs in the dECM-MA bioink with m-PVA were well aligned and matured, while the printed struts of the bioink with l-PVA showed a significantly low orientation factor of the myofibers. In addition, a few myofibers were observed in the bioink with h-PVA. Thus, the m-PVA was selected to fabricate the cell-laden dECM-MA constructs to induce the self-alignment of myofibers.

FIG. 2.

Optimization of PVA fibrillation/leaching. (a) Pr values of the printed dECM-MA constructs with h-PVA, m-PVA, and h-PVA. (b) LIVE (green)/DEAD (red), DAPI (blue)/phalloidin (red), and DAPI (blue)/MHC (green) stained images of the printed constructs. (c) Cell viability of hMPCs in the bioinks and printed structures obtained from LIVE/DEAD^TM stained images (n = 5; ***p < 0.001). (d) Orientation factor (f) of the cytoskeletons and myofibers of the hMPCs in the printed dECM-MA constructs by phalloidin and MHC stained images (n = 3; ***p < 0.001). (e) Fusion index and maturation rate of the myofibers in the printed constructs quantitatively estimated using MHC stained images (n = 5; **p < 0.01 and ***p < 0.001). (f) Processing diagram of the dECM-MA/m-PVA bioink containing hMPCs at various pneumatic pressures (50–400 kPa), showing the strut diameter and F-actin alignment (orientation factor, f) of the bioprinted constructs (n > 50). (G) Optical, LIVE/DEAD, and DAPI/phalloidin stained images of the printed struts using three different pneumatic pressures (100, 250, and 350 kPa). All data are represented as mean $\pm$ SD. The p-values by one-way ANOVA followed by Tukey's test are indicated.

The optimal pneumatic pressure applied in the nozzle during the printing was determined. The volume flow rate derived by the pneumatic pressure significantly influenced the printing outcomes as well as the structural alignment [Figs. S5(a)–S5(c) in the supplementary material]. Figure 2(f) shows the diameter of the printed strut and cell alignment, determined by the orientation factor of F-actin, of the printed hMPC-laden dECM-MA/m-PVA construct. The pneumatic pressure (or volume flow rate) ranged from 50 to 400 kPa significantly affected the cellular alignment [Fig. 2(f) and Fig. S5(d) in the supplementary material]. The F-actin in the printed struts was uniaxially aligned with the orientation factor (f > 0.6) when the applied pneumatic pressure ranged from 200 to 300 kPa during the printing, while the cell orientation was randomly distributed or partially oriented (f < 0.6) under lower than 200 kPa and higher than 300 kPa. Notably, the printed struts at a pressure lower than 200 kPa showed non-continuous structures, and the cell viability was significantly decreased at higher pressure (>300 kPa) [Figs. 2(f) and 2(g) and Fig. S5(e) in the supplementary material].

C. In vitro self-aligned myofiber formation in 3D bioprinted dECM constructs

To evaluate the effects of self-alignment and tissue-specific dECM on in vitro myogenic activities, uniaxially oriented cell-laden constructs (15 × 7 × 3 mm³) without a supporting polymeric frame structure were printed using (i) Gelatin methacrylate (Gel-MA)/PVA, (ii) dECM-MA, and (iii) dECM-MA/PVA bioinks [Fig. 3(a)]. Gel-MA was used as a control, as it is the most commonly used hydrogel for tissue engineering and bioprinting.²⁵ The LIVE/DEAD™ stained images revealed high cell viability in the printed constructs (>95%) after 3 days in culture [Fig. 3(b)], and cell proliferation was significantly higher in dECM-MA/PVA than in Gel-MA/PVA and dECM-MA constructs, as determined by MTT assay [Fig. 3(c)]. At 14 and 21 days in culture, the multinucleated myofibers were uniaxially oriented along with the printing direction in the Gel-MA/PVA and dECM-MA/PVA constructs, while aligned myofibers were not observed in the dECM-MA construct without PVA [Fig. 3(d)]. In addition, the aspect ratio of nuclei, F-actin alignment, and fusion and maturation index of the myofibers were significantly higher in the dECM-MA/PVA constructs [Fig. 3(e)–3(g)]. Regarding the myogenic gene expression, the expression levels of pre-expressed genes (paired box protein pax-7; Pax7), fast-expressed genes (myoblast determination protein 1; Myod1, myogenin; Myog, and myosin heavy chain 7; Myh7), and slow-expressed genes (myosin heavy chain 4; Myh4 and myosin heavy chain 2; Myh2) in hMPCs in the dECM-MA/PVA constructs were also significantly higher than those in the controls (Gel-MA/PVA and dECM-MA) [Fig. 3(h)]. Especially, the expression of the maturation-state myogenic genes (Myh4 and Myh2) was dramatically induced in the dECM-MA/PVA constructs compared to that of determination- and differentiation-state myogenic genes. Additionally, 2D myogenic differentiation of hMPCs on the dECM-MA–coated plates also demonstrated the dECM-MA could enhance cellular activities, including cell proliferation and differentiation [Fig. S5 in the supplementary material].

FIG. 3.

In vitro myogenic differentiation of hMPCs in 3D bioprinted constructs. (a) Schematics and optical images of the printed structures constructed using Gel-MA/PVA (GM-P), dECM-MA (dEM), and dECM-MA/PVA (dEM-P) bioinks. (b) LIVE/DEAD^TM stained images at 3 days in culture. (c) Cell proliferation of hMPCs in the printed constructs at 1, 3, and 7 days, as confirmed by MTT assay (n = 6; *p < 0.05, **p < 0.01, and ***p < 0.001). (d) DAPI/MHC at 14 and 21 days and SEM at 21 days of the printed constructs. (e) Nuclei aspect ratio (max. diam./min. diam.) (n > 30; ***p < 0.001) and (f) myofiber orientation factor (f) (n = 3; **p < 0.01 and ***p < 0.001) at 14 and 21 days, as obtained from DAPI/MHC stained images. (g) Fusion index and maturation rate of the myofibers in the printed constructs at 14 and 21 days, as obtained from MHC stained images (n = 5; ***p < 0.001). (H) Myogenic gene expression, including determination-state (Pax7), differentiation-state (Myod1, Myog, and Myh7), and maturation-state (Myh4 and Myh2) genes (normalized by GAPDH) in hMPCs of the printed GM-P, dEM, and dEM-P constructs (n = 3; *p < 0.05, **p < 0.01, and ***p < 0.001). All data are represented as mean $\pm$ SD. The p-values by one-way ANOVA followed by Tukey's test are indicated.

D. In vivo functional restoration by bioprinted self-aligned muscle construct

To validate the bioprinted muscle constructs, an extensive muscle defect injury model was created by removing approximately 40% of the tibialis anterior (TA) muscle mass (10 × 7 × 3 mm³) followed by extensor digitorum longus (EDL) and extensor hallucis longus (EHL) muscle ablation in immunodeficient Rowett nude rat (RNU) rats.^15,26 The bioprinted cellular constructs printed by Gel-MA/PVA, dECM-MA, and dECM-MA/PVA were implanted into the region of defected TA muscle [Fig. 4(a)]. After 4 and 8 weeks of implantation, the TA muscles implanted with the bioprinted constructs were harvested for further analyses. Age-matching (uninjured) and non-treated (injured) animals were used as controls (a total of 40 animals, n = 4 per group at 4- and 8-week time points).

FIG. 4.

Rat TA muscle defect injury model and histological examination. (a) Gross appearance of construct implantation after the creation of TA muscle defect (40% muscle mass) in rats. (b) The retrieved TA muscles at 4 and 8 weeks after implantation. (c) TA muscle weight (% of contralateral normal TA muscle) (n = 4) and (d) tetanic force (N·mm/kg) (n = 4) at 4 and 8 weeks after implantation. All data are represented as mean $\pm$ SD. The p-values by one-way ANOVA followed by Tukey's test are indicated. (e) H&E (purple: nuclei, pink: cytoplasm) and MTC (red: muscle fiber, blue: collagen) stained images of the retrieved TA muscles at 4 and 8 weeks after implantation.

Grossly, the bioprinted muscle constructs in dECM-MA and dECM-MA/PVA groups restored their original muscle volume at 8 weeks after implantation, as compared with the age-matching control group [Fig. 4(b)]. In contrast, the muscle injury in the non-treated group showed severe muscular atrophy at 4 and 8 weeks after implantation. Indeed, the bioprinted dECM-MA/PVA constructs demonstrated a significant increase in muscle weight (82.8 ± 3.5% and 88.7 ± 8.4%) when compared to non-treated (50.5 ± 3.5% and 61.9 ± 3.4%) and Gel-MA/PVA (67.7 ± 4.4% and 75.1 ± 3.9%) constructs at 4 and 8 weeks after implantation, respectively [Fig. 4(c)].

For in vivo functional analysis, the tetanic muscle force of the injured leg in response to electrical stimulation of peroneal nerve was measured at 4 and 8 weeks after implantation [Fig. 4(d)]. The bioprinted dECM-MA/PVA constructs also exhibited a significant functional recovery when compared with other groups at 4 weeks (62.4 ± 1.2 N·mm/kg in dECM-MA/PVA, 32.7 ± 1.7 N·mm/kg in non-treated, 35.6 ± 0.8 N·mm/kg in Gel-MA/PVA, and 42.2 ± 0.6 N·mm/kg in dECM-MA) and 8 weeks (80.7 ± 0.5 N·mm/kg in dECM-MA/PVA, 38.4 ± 2.6 N•mm/kg in non-treated, 58.7 ± 1.9 N•mm/kg in Gel-MA/PVA, and 72.8 ± 2.4 N•mm/kg in dECM-MA) after implantation.

E. In vivo skeletal muscle regeneration by bioprinted self-aligned muscle construct

To examine skeletal muscle regeneration by the bioprinted muscle constructs, histological and immunofluorescent analyses were performed. The defected TA muscles treated with the bioprinted constructs were harvested and cut longitudinally to examine microscopic structure at 4 and 8 weeks after implantation. The general morphology and collagenous fibrotic tissue formation in the defected muscle region were observed using hematoxylin and eosin (H&E) and Masson's trichrome (MTC) staining, respectively [Fig. 4(e)]. The histological results revealed that the bioprinted constructs in the Gel-MA/PVA, dECM-MA, and dECM-MA/PVA groups accommodated newly formed myofibers in the defected TA muscles. Among those, highly organized dense myofibers were found in the dECM-MA/PVA constructs. In the non-treated group, there was no sign of newly formed myofibers in the defect region. Furthermore, the fibrotic tissue area was quantified based on the MTC-stained images of the defect region [Fig. S7 in the supplementary material]. The bioprinted dECM-MA/PVA constructs had significantly less fibrotic tissue area compared with other groups.

To further evaluate the skeletal muscle regeneration, immunofluorescence for MHC was performed. The expression of MHC in the defected TA muscle indicated newly formed myofibers in the Gel-MA/PVA, dECM-MA, and dECM-MA/PVA groups [Fig. 5(a) and Fig. S8(A) in the supplementary material]. Quantitatively, the dECM-MA/PVA group showed significantly higher ability to form MHC⁺ myofibers when compared with the bioprinted constructs in the Gel-MA/PVA and dECM-MA groups, as confirmed by the percentage area of MHC⁺ myofibers [Fig. 5(b)] and diameter [Fig. 5(c)] and length [Fig. 5(d)] of myofibers. In order to track the printed hMPCs, anti-human laminin (HL), anti-human leukocyte antigen (HLA), and anti-mitochondrial ribosomal protein L12 (MRPL 12) were used [Fig. S9 in the supplementary material). The results indicated that the hMPCs in the bioprinted constructs were differentiated and formed myofibers (MHC⁺/MRPL 12⁺) at 4 and 8 weeks after implantation [Fig. 5(e) and Fig. S8(b) in the supplementary material], while cells in the age-matching control and non-treated groups did not express MRPL 12. Interestingly, higher contribution of hMPCs (77.1 ± 2.7% in the dECM-MA and 86.2 ± 2.2% in the dECM-MA/PVA) in both dECM-MA groups was detected, while the cells in the Gel-MA/PVA group showed 16.2% MHC⁺/MRPL 12⁺ myofibers at 8 weeks after implantation [Fig. 5(f)]. The results indicated that the combined effects from topographical cues and tissue-specific dECM components improved the printed hMPCs' contribution to the skeletal muscle regeneration.

FIG. 5.

Newly formed myofibers in bioprinted muscle constructs. (a) Immunofluorescent images of the retrieved TA muscles for MHC (red) at 4 and 8 weeks after implantation. (b) Quantification of the myofiber area per HFP (%, ×400) (n = 3; *p < 0.05). (c) Diameter of myofibers (μm²) (n = 20, *p < 0.05). (d) Myofiber length (μm²) (n = 20, *p < 0.05). (e) Double-immunofluorescent images of the retrieved TA muscles for MRPL12 (green)/MHC (red) at 4 and 8 weeks after implantation. (f) Quantification of myofiber area (MRPL12⁺) per HPF (%, × 400) (n = 3; *p < 0.05). All data are represented as mean $\pm$ SD. The p-values by one-way ANOVA followed by Tukey's test are indicated.

F. Host neural integration and vascularization of the bioprinted muscle constructs

The development of neuromuscular junction (NMJ) of the implanted constructs was determined for the functional restoration. Double-immunofluorescence was performed for neurofilament (NF) and alpha-bungarotoxin (α-BTX). Co-localization of NF+ axons of peripheral nerve and α-BTX⁺ acetylcholine receptor (AChR) in the implanted region indicated NMJ formation (NF⁺/α-BTX⁺) between the host nerve and newly formed myofibers in the bioprinted constructs [Fig. 6(a)]. Both dECM-MA and dECM-MA/PVA constructs had mature NMJs and neuronal contact, whereas the Gel-MA/PVA constructs had insufficiently developed NMJs. Quantitatively, more NMJs per field were found in the dECM-MA/PVA group than in the dECM-MA and Gel-MA/PVA groups at 4 and 8 weeks after implantation [Fig. 6(b)].

FIG. 6.

Innervation and vascularization of bioprinted muscle constructs. (a) Double-immunofluorescent images of the retrieved TA muscle for NF (green)/$\alpha$-BTX (red) at 4 and 8 weeks after implantation. (b) Quantification of NMJs (NF⁺/$\alpha$-BTX⁺) per HPF (n = 3; *p < 0.05). (c) Double-immunofluorescent images of the retrieved TA muscles for vWF (green)/α-SMA (red) at 4 and 8 weeks after implantation. (d) Quantification of vessels per HPF (n = 4) and (e) area of vessels per HPF (μm², ×400) (n = 4). All data are represented as mean $\pm$ SD. The p-values by one-way ANOVA followed by Tukey's test are indicated.

Host vascular ingrowth into the implanted bioprinted constructs was evaluated by double-immunofluorescence for Willebrand factor (vWF) and α-smooth muscle actin (α-SMA). At 4 weeks after implantation in the dECM-MA and dECM-MA/PVA groups, the bioprinted constructs were highly vascularized when compared with the Gel-MA/PVA group [Fig. 6(c)]. However, the vascularization in both dECM-MA and dECM-MA/PVA groups was dramatically declined at 8 weeks after implantation, while the bioprinted Gel-MA/PVA constructs showed the gradual buildup of vascularization [Figs. 6(d) and 6(e)].

III. DISCUSSION

It is well known that the biochemical and topographical cues notably impact the cell function, orientation, and differentiation on skeletal muscle regeneration.^7,27 In our previous study, we successfully fabricated the biomimetic tissue constructs by skeletal muscle-derived dECM-MA bioink together with the PVA fibrillation/leaching method.¹⁷ These biomimetic constructs represented the biochemical and topographical features of skeletal muscle that accelerated the myogenic differentiation by dECM components and simultaneously induced the self-alignment of the printed cells by topographical cues in vitro. Our in vitro results strongly suggested that the combined biochemical and topographical cues could enhance the cellular alignment and differentiation of hMPCs in the bioprinted constructs.

In this present study, we applied this strategy to investigate the clinical feasibility of using the clinically relevant human cell source in the rat muscle defect injury model. The hMPCs used in this study are clinically relevant cell sources. The safety and efficacy of using these cells have already been demonstrated in several clinical trials.²⁸ The acellular or cell-free approaches may provide off-the-shelf products with minimal translational hurdles.²⁹ However, the acellular approaches often require controlled delivery systems of bioactive factors such as growth factors or chemokines to stimulate the host tissue microenvironment.³⁰ These approaches will be limited if there is no endogenous cell source available from the adjacent tissue due to the extensive tissue damage. Although ex vivo cellular manipulation is required for the cell-based approaches, this strategy could apply more extensive tissue regeneration with high regenerative capability.^29,31 Thus, treatment options should be carefully chosen for the specific patient and specific condition being treated.

In the bioink formulation, the PVA component provided the topographical cues directly to the cell-laden constructs because the fibrillated PVA was aligned alongside a distribution of shear stress in a microscale nozzle during the printing process.^17,32 To evaluate the effect of the molecular weight of PVA on the self-alignment of the printed hMPCs, we selected three different molecular weights of PVA. With l-PVA (Mw: 31–50 kg/mol), the cell alignment and printing outcome were not sufficiently achieved due to the relatively low elongational force of PVA chemical chains in the shear flow of the bioink.³³ Dissimilarly, h-PVA (Mw = 146–186 kg/mol) provided a good printing outcome (P_r ≈ 1.0) but the cell viability (∼42.4%) was significantly decreased due to the greater wall shear stress caused by the high viscosity during the shear flow in the printing nozzle. Thus, we selected m-PVA (Mw = 89–98 kg/mol) that provided high cell viability (>90%) and effective topographical cues for the self-alignment as well as a proper printing outcome (0.9 < P_r < 1.0).

Process parameters, including extrusion pressure, nozzle diameter, feed rate, and temperature, influence the printing outcomes. Among the process parameters, the extrusion pressure can be the most considerable parameter affecting the flowing behavior of the bioink within the microsized printing nozzle, which can directly disturb the cell viability and alignment due to the wall shear stresses within the nozzle.³⁴ When proper wall shear stress was developed in the nozzle, the PVA and dECM-MA molecules were uniaxially oriented along with the printing direction. However, a deficient wall shear stress by lower pressure or a large amount of extruded bioink by higher pressure did not induce the aligned structures. Thus, the effect of the extrusion pressure was examined on the cell viability and cellular alignment of the bioprinted hMPCs. By assessing the cell viability and anisotropic cell-cytoskeleton of the bioprinted struts, 250 kPa of the extrusion pressure was selected. Along with the extrusion pressure, other processing parameters (37 °C of printing plate temperature, 10 mm/s of moving speed, and 300 mW/cm² of UV intensity) were also optimized for the bioprinted dECM-MA/PVA containing hMPCs, resulting in more than 90% initial cell viability and a high degree of cellular alignment. This optimal printing process presented a reasonable printing outcome (0.9 < P_r < 1.0), enabling us to create a cell-laden volumetric muscle construct (15 × 7 × 3 mm³) of anatomically relevant size for in vivo animal study.

Along with the topographical cues, skeletal muscle-derived dECM bioink provided tissue-specific biochemical cues, which include various growth factors, cytokines, and structural/adhesive molecules.³⁵ Previously, we chemically modified the dECM-based materials as a photo-crosslinkable hydrogel by methacrylation to improve the structural integrity.^17,22 Additionally, this chemical modification of dECM might offer a more stable microenvironment to the cells. The bioactive molecules after the solubilization will have a short effective half-life, low stability, and rapid inactivation by enzymes under physiological conditions. It has been reported that the covalent immobilization of growth factors could maintain their stability and protect the rapid inactivation by enzymes, resulting in prolonged bioactivities.^22,36 In our previous study, several growth factors and cytokines were detected from skeletal muscle-derived dECM-MA.¹⁷ Especially, high amounts of transforming growth factor β1 (TGF-β1), vascular endothelial growth factor (VEGF), growth hormone 1 (GH1), insulin-like growth factor 1 (IGF-1), and bone morphogenic protein 7 (BMP-7) were maintained in the dECM-MA.

In this study, the combined biochemical and topographic cues exhibited the synergistic effect that induces highly efficient cell-to-cell/cell-to-matrix interactions. Gene expression in the dECM-MA/PVA group was significantly higher in various stages of gene expression (determination, differentiation, and maturation) compared with the Gel-MA/PVA (no biochemical cues) and dECM-MA (no topographic cues) groups. This might have been caused by the strong influence of cell polarity through the topographical pattern³⁷ known as “contact guidance” and the various tissue-specific bioactive molecules preserved in the dECM-MA bioink. According to our in vitro evaluations, we concluded that the dECM-MA/PVA constructs efficiently induced the self-alignment and matured myofiber formation of hMPCs.

To investigate the clinical feasibility using this self-organized bioprinted human skeletal muscle construct, we used the well-established immunodeficient rat model of the extensive muscle loss injury.^15,26 This defect injury model without treatment shows the irreversible structural and functional deficits with 62% and 38% recovery in muscle weight and tetanic force, respectively, following 8 weeks after injury, while the muscle force was recovered by 81% of normal muscle force with 89% muscle weight recovery in the dECM-MA/PVA group. Using this critically defected injury model, the results indicated that the bioprinted muscle constructs could facilitate effective skeletal muscle regeneration, resulting in functional restoration in vivo. Histologically, the bioprinted dECM-MA/PVA constructs were well integrated with host native skeletal muscle with minimum fibrotic tissue formation. More interestingly, 86% and 77% of MHC⁺ myofibers formed in the implanted constructs were originated from the printed hMPCs in the dECM-MA/PVA and dECM-MA, respectively, as confirmed by double-immunofluorescence for MHC and MRPL 12 at 8 weeks after implantation. However, only 16% of MHC⁺ myofibers were found in the Gel-MA/PVA constructs. This indicates that the dECM-MA provides a more proper microenvironment to accommodate hMPCs in the bioprinted constructs.

The in vivo outcomes also showed effective host neural integration and vascular ingrowth into the bioprinted dECM-MA and dECM-MA/PVA constructs. Effective innervation is essential to successfully bioengineer a functional muscle construct in vivo. The established contacts of bioengineered muscle constructs with host neural network are critically important following implantation, as improper innervation will lead to atrophy of muscle tissue and loss of contractile function.³⁸ Following the interconnection between the host nerve and AChR clusters on the myofibers, a higher number of NMJs were formed in the dECM-MA/PVA constructs at 4 and 8 weeks after implantation. Also, the vascular ingrowth should be rapidly occurred in the implanted bioengineered tissue constructs to avoid necrosis of the transplanted cells in the constructs due to the limited oxygen and nutrient supply.^5,39 In our results, both dECM-MA and dECM-MA/PVA groups showed a higher number and area of vessel formation compared with the age-matching control after 4 weeks of implantation, resulting in accelerated skeletal muscle regeneration. After maturation of the developed myofibers, the microvessels formed in both dECM-MA constructs were stabilized at 8 weeks after implantation, resembling the age-matching control. This could be explained by angiogenic factors (e.g., VEGF) in the dECM-MA. These results are also in agreement with previous studies showing that skeletal muscle-derived ECM scaffolds can induce a strong angiogenic response,⁴⁰ and a highly oriented topology can promote vascularization.⁴¹ These results demonstrated that our dECM-MA/PVA constructs are capable of strong innervation and vascularization capacity after implantation.

IV. CONCLUSIONS

We develop the human skeletal muscle constructs using combined biochemical and topographical cues. The photo-crosslinkable dECM-MA bioink provides skeletal muscle-specific biochemical cues, and the self-alignment of the printed hMPCs is achieved by the PVA fibrillation/leaching and in situ printing process. The self-aligned myofibers in the bioprinted dECM-MA constructs show a rapid restoration of muscle function in the rat TA muscle defect injury model. This advanced bioprinting approach for bioengineering functional skeletal muscle constructs may be an effective therapeutic option for treating extensive muscle defect injuries with accelerated innervation and vascularization.

V. EXPERIMENTAL METHODS

A. Skeletal muscle-derived ECM bioink preparation

Tissues were biopsied from the lower limbs of female Yorkshire pigs (10–15 months old) under the Wake Forest University Institutional Animal Care and Use Committee (IACUC) guidelines. Following the removal of the epimysium, the tissue was chopped into a small cube (smaller than 8 × 8 × 3 mm³). The chopped tissues were then washed trice using Dulbecco's phosphate-buffered saline (DPBS; HyClone^TM, Cytiva, Marlborough, MA) and treated with 1 w/v% sodium dodecyl sulfate (SDS) solution for 3 days in a sterilized glass bottle until the tissues became transparent. Next, the tissues were treated sequentially with Triton X-100 (1% in DPBS; 2 days), antibiotic/antimycotic (1% in DPBS; 1 h), and DNase I solution (2 h). The decellularized tissues (dECM) were rinsed with DPBS and de-ionized water, followed by lyophilization in a freeze-dryer (Labconco, Kansas City, MO).

Pepsin solution (0.1 w/v% in 0.5 M acetic acid) was used to digest the lyophilized dECM (10 mg dECM per 1 ml of pepsin solution) at room temperature for 2 days, followed by precipitation with sodium chloride in the digested solution. The precipitated dECM was dialyzed using a dialysis tube (1000 kDa molecular cutoff; Spectrum Chemical Manufacturing Corp., New Brunswick, NJ) at 4 °C for 2 days and lyophilized.

The lyophilized dECM was dissolved in acetic acid solution (0.5 M) at a ratio of 3.75 mg dECM per 1 ml acetic acid solution, and the dECM solution was adjusted to pH to 8.5 by adding 1 M NaOH at 4 °C under continuous stirring. To obtain a photo-crosslinkable dECM, methacrylic anhydride (MAA) was added for 2–3 days at a ratio of 621 mg MAA per 600 mg dECM at 4 °C. After chemical modification, the dECM methacrylate (dECM-MA) was dialyzed and lyophilized. The lyophilized dECM and dECM-MA were stored at −80 °C until the use. All chemicals were purchased from Millipore Sigma (Saint Louis, MO), unless stated otherwise. The degree of methacrylate was measured by TNBS assay kit (Thermo Fisher Scientific), as reported previously.^17,22 The following equation was used to calculate the degree of methacrylate (DM):

$$DM=\left[1-\frac{{OD}_{\mathit{dECM}-MA}}{{OD}_{\mathit{dECM}}}\right]\times 100.$$ (1)

B. Characterizations of dECM-MA bioink

DNeasy Blood & Tissue kit (Qiagen Inc., Hilden, Germany) and Quant-iT™ PicoGreen™ dsDNA Assay kit (Thermo Fisher Scientific, Waltham, MA) were used to extract and quantify the DNA content, respectively. DNA quantification was performed according to the manufacturer's instructions.

ECM components were examined by solubilizing the lyophilized dECM, as follows: pepsin solution (0.1 mg pepsin per 1 ml of 0.5 M acetic acid) for collagen assay; extraction solution containing papain for the glycosaminoglycans (GAGs) assay; and oxalic acid solution (0.25 M) for the elastin assay. Sircol^TM Soluble Collagen, Blyscan^TM Sulfated Glycosaminoglycans, and Fastin^TM Elastin assay kits (Biocolor Life Sciences Assays, Carrickfergus, UK) were used for quantifying collagen, GAGs, and elastin, respectively, according to the manufacturer's instructions.

To examine the rheological properties, storage modulus (G′), complex viscosity (η*), and yield stress (σ_y), of the dECM-MA bioinks with PVA, a cone-and-plate geometry (4° angle, 40 mm diam., and 150 μm gap) supplemented rotational rheometer (BohlinGemini HR Nano; Malvern Instruments, Malvern, UK) was used. A frequency sweep (0.1–10 Hz) was conducted with 1% strain on the 5 wt. % dECM-MA with different molecular weights of 20 wt. % PVAs at 25 °C without UV and 37 °C with UV cross-linking. To evaluate the τ_y, G′ was examined under a stress sweep (0.1–500 Pa) with 1 Hz at 25 °C without UV exposure.

C. Primary human muscle progenitor cell culture

Human musculus gracilis muscles (from 51- and 64-year-old women, de-identified) were biopsied following the approved protocol by the Wake Forest University Institutional Review Board (IRB), as previously reported.^15,26 Briefly, minced biopsies were digested in DMEM containing collagenase type I (0.2 w/v%; Worthington Biochemical, Lakewood, NJ) and dispase (0.4 w/v%) for 2 h at 37 °C. The digested tissue fibers were gently pipetted, filtered using a 100 μm pore strainer, and centrifuged at 1500 rpm for 5 min. The obtained pellet was transferred to culture plates coated with collagen type I (1 mg/ml, BD Biosciences, San Jose, CA) and dissolved in DMEM/F12 containing fetal bovine serum (FBS; 18%), dexamethasone (0.4 μg/ml), human insulin (10 μg/ml), human basic fibroblast growth factor (hbFGF; 1 ng/ml), and human epidermal growth factor (hEGF; 10 ng/ml) overnight at 37 °C in a humidified atmosphere with 5% CO₂. The non-adhered cells contained in the supernatant were transferred into other plates coated with collagen. After 80% confluence about 8–10 days in culture, the cells were sub-cultured in growth medium consisting of DMEM/high glucose (HyClone^TM), 20% FBS (Gemini Bio-Products, West Sacramento, CA), 2% chicken embryo extract (Gemini Bio-Products), and 1% penicillin/streptomycin (Thermo Fisher Scientific) and expanded to passage 4–5 for the experiments. The medium was changed every 2 days.

D. Bioprinting process and PVA fibrillation/leaching

After dissolving the dECM-MA (200 mg/ml) in 0.1 M acetic acid overnight, it was mixed with 10× DMEM at a 1:1 volume ratio. Then, the PVA (200 mg/ml) in DPBS was added to the dECM-MA solution (100 mg/ml) at a 1:1 volume ratio. For the complete mixing, the dECM-MA and PVA were loaded in two syringes and a three-way stopcock, and the solution was pumped back and forth through the stopcock for 5 min. The final concentrations of dECM-MA and PVA were 5 wt/vol % (50 mg/ml) and 10 wt/vol % (100 mg/ml), respectively. The gelatin methacrylate (Gel-MA) as a control was synthesized based on our previous work.⁴² The prepared dECM-MA/PVA, dECM-MA, and Gel-MA/PVA solutions were mixed with 2 × 10⁷ cells/ml of hMPCs and 3 mg/ml of 2-hydroxy-4'-(2-hydroxyethoxy)-2-methylpropiophenone (Irgacure 2959). For the printability (P_r = L²/16A), a two-layer crosshatch pattern (9 square-shaped pores arranged in a 3 × 3 pattern) was printed. The Pr values were measured for each pore.²⁴

The bioinks containing hMPCs were loaded into a sterile syringe with a 30 G single nozzle (inner diam. = 150 μm), and the syringe was aseptically inserted into a dispensing module (AD-3000C; Ugin-tech, Seoul, South Korea) that was connected to a printing system (DTR3–2210 T-SG; DASA Robot, South Korea) in a closed humidified chamber. The cell-laden bioinks were printed at 50–400 kPa of pneumatic pressure onto a 12 × 12 mm² sterilized poly(ethylene terephthalate) (PET) film on a working plate. During the printing process, the cell-laden bioink was crosslinked in situ using UV light (300 mW/cm²) [Fig. S2(c) in the supplementary material]. After removing the PET film, the printed muscle constructs (15 × 7 × 3 and 10 × 7 × 3 mm³) were placed in a 12-well plate with the growth medium at 37 °C for 1 day, and then washed thrice using pre-warmed DPBS (37 °C) to leach out PVA from the construct. The printed conditions based on the bioink formulations are listed in Table S1 in the supplementary material.

1. In Vitro Evaluations of Bioprinted dECM-MA Constructs

The bioprinted constructs containing hMPCs were fixed in neutral-buffered formalin (Leica Biosystem Inc., Buffalo Grove, IL), blocked using a serum-free blocking agent (Agilent), and permeabilized with Triton X-100. The cytoskeleton of the hMPCs in the bioprinted constructs was visualized by immunofluorescence for Alexa Fluor 594-conjugated phalloidin (green; 1:100 in DPBS; Invitrogen), and nuclei were stained using 4′,6-diamidino-2-phenylindole (DAPI; blue; 1:100 in DPBS; Invitrogen, Carlsbad, VA). Samples were imaged using a confocal microscope (Carl Zeiss). The nuclei aspect ratio and cytoskeleton coverage and alignment were analyzed using ImageJ software.

For MHC immunofluorescence, the samples were treated with anti-MF20 primary antibody (5 μg/ml; Developmental Studies Hybridoma Bank, Iowa City, IA) and incubated overnight at 4 °C. Subsequently, the samples were washed with DPBS and incubated with secondary antibodies (1:50 in DPBS; Invitrogen) conjugated with Alexa Fluor 488 for 1 h. Finally, the samples were counterstained with DAPI. Immunofluorescent images were captured using a confocal microscope (Leica DM4000B; Leica, Wetzlar, Germany) with z-stack mode to visualize combined images, and the maturation and orientation of the myofibers were analyzed by ImageJ software. The quantification of myofiber maturity (fusion index and maturation rate) was carried out after normalizing the values to the number of DAPI.

The viability of cells was examined using LIVE/DEAD staining assay kit (Life Technologies, Carlsbad, CA). The cells were stained using 0.15 mM calcein AM and 2 mM ethidium homodimer-1 at 37 °C for 1 h. The stained cells were visualized using a Carl Zeiss confocal microscope (LSM 700; Carl Zeiss, Oberkochen, Germany). ImageJ software (National Institutes of Health, Bethesda, MD) was used to calculate cell viability using LIVE/DEAD-stained images.

Cell proliferation was determined by MTT assay (cell proliferation kit I; Boehringer Mannheim, Mannheim, Germany). The samples were incubated in the growth medium containing 0.5 mg/ml MTT at 37 °C for 4 h and dissolved using 10% SDS in 0.01 M HCl solution. Optical density (OD) was measured at 570 nm wavelength using a microplate reader (EL800; Bio-Tek Instruments, Winooski, VT). The OD values at 7 days were normalized to those on 1 day to calculate the proliferation rate.

The samples were fixed using 10% neutral-buffered formalin and dehydrated using ethanol series (50%, 60%, 70%, 80%, 90%, and 100%), followed by lyophilization. The longitudinal cross-sectional morphology of the freeze-dried tissues was visualized using scanning electron microscopy (SEM; Flex SEM 1000, Hitachi, Tokyo, Japan).

Using quantitative RT-PCR, the expression levels of myogenesis-related genes, including determination-state [paired box protein pax-7 (Pax-7)], differentiation-state [myogenic differentiation 1 (Myod1), myogenin (Myog), and myosin heavy chain 7 (Myh7)], and maturation-state [myosin heavy chain 4 (Myh4) and myosin heavy chain 2 (Myh2)], were examined after 14 and 21 days of differentiation. Total RNA was isolated from cells using a TRI reagent. Thereafter, RNA concentration and purity were determined using a spectrophotometer (SpectraMax M5; Molecular Devices). For complementary DNA (cDNA) synthesis, total RNA was first treated with RNase-free DNase. Next, a QuantiTect Reverse Transcription Kit (Qiagen) was used, followed by amplification of the cDNA using a QuantiStudio^TM 3 Real-Time PCR System (Applied Biosystems, Foster City, CA). cDNA samples were used for quantitative RT-PCR in triplicate using Power SYBR^® Green PCR Master Mix (Quiagen) according to the manufacturer's instructions. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the housekeeping gene. The bioprinted Gel-MA/PVA constructs at 14 days in culture were used to determine baseline reads. The primer sequences are listed in Table S2 in the supplementary material.

E. Rat TA muscle defect injury model

The skeletal muscle defect model was created in RNU rats (male, 10–12 weeks old, Charles River Laboratory, Wilmington, MA) and all animal procedures were performed according to protocols approved by the IACUC at Wake Forest University. Animals were anesthetized with 3% isoflurane, the skin of the lower left leg was incised 3–5 cm, and the muscle was separated from the fascia. The EDL and EHL muscles were removed to prohibit compensatory hypertrophy during muscle regeneration, and approximately 40% of the TA muscles were excised and weighed. The TA muscle weight of each animal was calculated with the following equation: y(g) = 0.0017 × body weight (g) – 0.0716.⁴³ The bioprinted constructs were implanted in the defect regions and covered with fascia. The fascia was closed using absorbable sutures, and the skin was sutured with surgical staples. Five groups were performed at 4- and 8-week time points (40 animals in total, n = 4 per group and time point): (1) Age-matching control (uninjured), (2) non-treated (defect only), (3) Gel-MA/PVA, (4) dECM-MA, and (5) dECM-MA/PVA.

1. Tetanic Muscle Force Measurement

To evaluate the muscle function restoration, the tetanic force of TA muscles was measured using a dual-mode muscle lever system (Aurora Scientific, Inc., Mod, 305 b, MI-RAT 2.74, Aurora, Canada) in a blinded manner (n = 4 per group and time point, three repeated measurements per sample). Animals were anesthetized, and the posterior left injured foot was fixed to the footplate. After stabilizing the knees and ankles at a 90° angle, two sterile platinum needle electrodes were placed in the posterior compartment of the lower leg along either side of the peroneal nerve, and the nerve was stimulated using a Grass stimulator (S88) at 100 Hz and 10 V with a pulse width of 0.1 ms. Muscle force (N·mm/kg) in response to electrical stimulation was calculated (peak isometric torque × foot length) per body weight. After functional evaluation, the TA muscles of each lower leg were collected and weighed, and the percentage of TA muscle weight (percent of contralateral) was calculated (n = 4 per group and time point).

F. Histological and immunofluorescent staining

For the histological evaluation, the harvested TA muscle samples were rapidly frozen in liquid nitrogen and frozen-sectioned into 7-μm thick slices. The sliced samples were fixed with 4% paraformaldehyde for 10 min, and H&E staining were performed. The MTC-stained images were analyzed by ImageJ software to quantify the fibrotic area (n = 3 per sample and time point, four random fields in each sample).

For immunofluorescence, sectioned samples were fixed with 4% paraformaldehyde for 15 min, treated with serum-free blockers for 1 h at room temperature, and incubated with primary antibody overnight at 4 °C. The samples were washed three times with TBS buffer and incubated with either Alexa 488-conjugated anti-rabbit IgG (1:200 dilution, Invitrogen) or Alexa 594-conjugated anti-mouse IgG (1:200 dilution, Invitrogen) at room temperature for 40 min. The samples were mounted with mounting medium and DAPI. To examine the area, density, diameter, and length of the newly formed muscle fibers in the implantation regions, the samples were double-immunostained with mouse anti-MF-20 and MRPL12 (rabbit polyclonal anti-mitochondrial ribosomal protein L11, species reactivity: human, 1:1000 dilution, Abcam, Cambridge, UK). The area of myofibers per field (%), density, diameter, and length of myofibers were measured with immunofluorescent images for MHC/DAPI (×400 magnification) in a blinded fashion (n = 3 per sample and time point, four random fields in each sample). NMJ formation was visualized with double-immunofluorescence using rabbit anti-NF/rat anti-AChR antibodies. The number of NMJ per field (%) was obtained using double-immunofluorescence for NF/AChR (×400 magnification; n = 3 per group and time point, four random fields per sample). To evaluate the vascularization of the implanted constructs, the samples were stained with rabbit anti-vWF (1:400 dilution, Dako, Agilent, Santa Clara, CA) and mouse anti-α-SMA (1:50 dilution, Santa Cruz Biotechnology, Santa Cruz, CA). The number and area of vessel per field (μm²) were measured with immunofluorescence images for vWF/α-SMA (×400 magnification, n = 4 per sample and time point, four random fields per sample). All images were analyzed using Image J software.

G. Statistics

One-way analysis of variance (ANOVA) and post hoc analysis (Tukey's HSD test) were conducted using SPSS™ software (SPSS Inc., Chicago, IL). Variables are expressed as a mean ± standard deviation (SD), and differences between experimental groups were considered statistically significant at *P < 0.05, **P < 0.01, and ***P < 0.001.

SUPPLEMENTARY MATERIAL

See the supplementary material for the technical details of immunofluorescence for collagen type I, elastin, and laminin, and characterization methods for the printed dECM-MA constructs. It entails immunofluorescent images of native and decellularized tissues, rheological and printability outcomes, SEM images, and physical properties of the printed dECM constructs with different molecular weights of PVAs. It also presents quantitative analyses of cellular orientation and activities. It contains in vivo outcomes, including quantification of the fibrotic area (%), immunofluorescent images for MHC and MRP11/MHC, and various human antigens. It provides tables for printing conditions based on bioink formulations and human primer sequences used for quantitative RT-PCR analyses.

AUTHORS' CONTRIBUTIONS

H.L. and W.K. contributed equally to this work. All authors reviewed the final manuscript.

DATA AVAILABILITY

The data that support the findings of this study are available within the article and its supplementary material.