Reconstructing vascular networks promotes the repair of skeletal muscle following volumetric muscle loss by pre-vascularized tissue constructs

Preparation of acellular collagen and methacrylated collagen (CMA) hydrogels

After removing the fat layer, porcine skin tissues were cut into small pieces (1 cm × 1 cm) and further minced using a mixer grinder. To remove non-collagenous proteins, minced tissues were immersed into 0.1 M NaOH at 4°C for 3 days, rinsed with distilled water, and immersed into 10% butanol at 4°C for 24 h to remove the residual fat and debris. Collagen was extracted using 0.5 M acetic acid at 4°C for 3 days with continuous stirring. The supernatants of the extracted solutions were collected by centrifugation (7000 rpm), desalted by adding NaCl to a final concentration of 3 M, and centrifuged again at 4°C. The precipitate was collected and re-dissolved in 0.5 M acetic acid, dialyzed with 0.5 M acetic acid for 1 day and distilled water for 1 day to remove salts, and lyophilized. Freeze-dried collagen was dissolved in 0.02 M acetic acid by stirring at 4°C at the indicated concentration. 1 mL of collagen solution was mixed with 0.1 mL of cooled calcium- and magnesium-free Dulbecco’s phosphate-buffered saline (DPBS) and 0.025 mL of (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer solution in an ice bath; the pH was adjusted to 7.4–7.6 using 1 M NaOH, and the solution was loaded into the PDMS-made mold to form disk-shaped Collagen hydrogels with 1 mm thick and incubated at 37°C for 30 min to initiate a self-assembled gelation process for the formation of a collagen hydrogel with 1 mm thick (acellular Collagen).

The CMA conjugate was synthesized by conjugating methacrylic anhydride to collagen molecules. The details were modified from the published literature.^22,23 Briefly, Freeze-dried collagen was dissolved at a 0.0625% (weight/volume (w/v)) collagen solution in 0.1 M HCl aqueous solution at 4°C overnight. Morpholinoethanesulfonic acid (MES; Sigma-Aldrich, USA) was added to collagen solutions to reach a concentration of 50 mM and the pH was subsequently adjusted to 6 by adding 0.1 N NaOH solution at 4°C. Next, 0.459 g of methacrylic anhydride (Sigma-Aldrich, USA), 0.337 g of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC; Sigma-Aldrich, USA), and 0.08 g of N-hydroxysulfosuccinimide (NHS; Sigma-Aldrich, USA) were added to 100 mL of collagen solution, and the mixture was placed at 4°C for 24 h. Next, the solution was dialyzed with 0.005 M HCl for 14 days and freeze-dried to obtain the powder form of CMA conjugates. To prepare acellular CMA hydrogels, a CMA prepolymer solution was prepared by mixing the freeze-dried CMA (0.3% (w/v) final) and the photo-initiator (Irgacure 2959) (0.125% (w/v); CIBA Chemicals, USA) dissolved in PBS at 4°C. The glass spacers, 1 mm thick, were added on both sides of the glass slide to define the thickness of the hydrogel. 100 μL of CMA prepolymer solution was dropped in the middle of the glass slide and covered with polydimethylsiloxane (PDMS)-modified slides. After light exposure, a disk-shaped CMA gel was formed by removing the PDMS-modified slide.

Photocrosslinking was achieved by exposing the prepolymer to 11.42 mW/cm² UV light (365 nm; M&R Nano Technology, Taiwan) to 15 s (acellular CMA-15) and 20 s (acellular CMA-20). The intensity of UV light was monitored using an ultraviolet (UV) intensity meter (Model 202; G&R Labs, Taiwan) at room temperature.

Characterization of hydrogels

Proton nuclear magnetic resonance (1H NMR) spectra were recorded by the NMR spectrometer (AV-400 (400 MHz); Bruker, USA) at 4°C to characterize the conjugation of the methacrylic group in the CMA conjugates (1 mg per mL in D₂O). For the measurement of circular dichroism (CD) (J-815 CD; Jasco, USA), all samples were adjusted to a concentration of 0.03% (w/v). The solution was scanned at a wavelength range of 190–270 nm at 4°C. Rheological measurements of hydrogel formation were performed using a strain-controlled parallel-plate rheometer (AR-G2; TA Instruments, USA) at room temperature as previously described.^24,25 Analysis of three-dimensional (3D) microstructure of the hydrogel in an aqueous environment was performed under the multi-photon confocal microscopy (TCS-SP5-X; Leica, Germany). Atto 550 NHS ester (Sigma-Aldrich, USA) was conjugated to collagen-based hydrogels, according to the instructions provided by the manufacturer, through incubation overnight at room temperature. Fluorescence-labeled hydrogels were imaged using a multi-photon confocal microscope system (TCS-SP5-X; Leica, Germany). To analyze the microstructure of freeze-dried hydrogels, the hydrogels were immersed in liquid nitrogen for 1 min and placed in a freeze-dryer for 1 day. The microstructure of the hydrogels was investigated through scanning electron microscopy (SEM) (S-4300; Hitachi, Japan). To understand its drug release profile, the preformed hydrogel was immersed in 2 mg/mL of fluorescein isothiocyanate-dextran (70 kDa) isothiocyanate-dextran (FITC-dextran; Sigma-Aldrich, USA) solution. After 24 h of absorption, the hydrogels were washed three times with PBS and soaked in PBS to measure their cumulative release profiles for FITC-dextran at 5, 10, 30, 60 min and 24 h. This was achieved by measuring the FITC intensity with a UV-visible spectrophotometer (SpectraMax Plus 384; Molecular Devices, USA) at an excitation wavelength of 495 nm.

Cell culture

HUVECs (Lonza, USA) cultured for 4–7 passages in endothelial cell growth medium (EGM-2; Lonza, USA) containing 2% (volume/volume (v/v)) fetal bovine serum (FBS) and 1% (v/v) penicillin-streptomycin (PS) were used. MSCs were isolated from human normal deidentified discarded white adipose tissue obtained during a clinically-indicated procedure following an Institute Review Board-approved protocol from Dr. Juan M. Melero-Martin at Harvard Medical School (Boston, MA, USA), as previous. ²⁶ MSCs were cultured for 8–9 passages in MSC growth medium (MSCGM; Lonza, USA) containing 20% (v/v) FBS, 10 ng/mL basic fibroblast growth factor and 1% (v/v) PS and subsequently used. NIH-3T3 fibroblasts were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% FBS and 1% (v/v) PS.

Two-dimensional (2D) cell viability assays

Cells (2000 cells per well) were cultured in 96-well plates in regular culture medium. After 1 day of culture, photo-initiator at various concentrations (0.0625-1%) was added to the culture medium, and the wells were exposed to UV light for 0–60 s. Furthermore, cells were cultured in Collagen or CMA hydrogels in regular medium. After an additional 2 days, cell viability was calculated as the number of live cells and expressed as a percentage of the control using a live/dead viability kit (Invitrogen, USA) according to the instructions provided by the manufacturer.

3D cell encapsulation

Prepolymer Collagen and CMA solutions were gently mixed with MSCs alone (1 × 106 cells/mL) or a mixture of MSCs and HUVECs (1:1 ratio; 1 × 106 cells/mL). The 100 μL of cell-laden Collagen hydrogels were formed at 37°C for 30 min into the PDMS-made mold to form disk-shaped cell-laden Collagen hydrogels with 1 mm thick. The cell-laden CMA gel formation occurred under UV irradiation for 15 (CMA-15) or 20 (CMA-20) seconds. In this process, a Petri dish (Corning Inc., USA) was used as the substrate. The glass spacers, 1 mm thick, were added on both sides of the glass slide to define the thickness of the hydrogel. 100 μL of cell-laden prepolymer solution was dropped in the middle of the glass slide and covered with polydimethylsiloxane (PDMS)-modified slides. After light exposure, a disk-shaped cell-laden CMA gel was formed by removing the PDMS-modified slide.

The cell-laden gels were cultured in 24-well dishes for 2 days. After staining with a live/dead viability kit, the spreading area of live cells (green-labeled cells) was imaged using a multiphoton confocal microscope system (TCS-SP5-X; Leica, Germany). Serial optical sections were captured in the central depth plane with respect to the surface of the hydrogel structures, and the resultant z-stacks were merged into an image using ImageJ (National Institutes of Health (NIH), USA) to analyze the cell spreading area inside the hydrogel. Cell spreading was quantified by measuring the area of cells from z-stacks confocal images (n = 3 to 5 for each group).

In vivo vasculogenic assay

All animal experiments were approved by the institutional animal care and use committee of National Tsing-Hua University, Hsinchu, Taiwan (Protocol number: 108030). Male BALB/cAnN.Cg-Foxnlnu/CrlNarl mice (aged 5–7 weeks) were purchased from the National Laboratory Animal Center, Taiwan. The formation of vascular networks in vivo was evaluated using a xenograft transplantation model in immunodeficient mice.^24,27 HUVECs alone or MSCs alone cannot support vessel formation in collagen hydrogels (Figure S1). Therefore, MSCs and HUVECs were directly co-cultured at a 1:1 cell ratio in the cell density of 1 × 107 cells/mL, a suitable combination to generate vascular networks.^26,28–30 For the cell-laden Collagen group, the collagen prepolymer solutions (100 μL) were mixed with MSCs and HUVECs (1:1 ratio; 1 × 107 cells/mL) before gelation. The cell-laden CMA mixture (100 μL) was pipetted onto a polydimethylsiloxane-coated glass slide, with aliquots separated using a 1 mm spacer following 15 and 20 s of UV light exposure to form the cell-laden CMA-15 and CMA-20 groups, respectively. Two hydrogels were surgically implanted into the subcutaneous sites of a mouse. Each experimental condition was performed in 5–7 replicates. On day 7 after implantation, the mice were sacrificed and the hydrogel constructs were removed from the mice, and a 3D laser scanning confocal microscope (VK-X1000; Keyence, Taiwan) was used to scan and determine the 3D surface and topography of tissue constructs. This allowed the measurement of the thickness, horizontal area, and volume of the structures prior to tissue processing for histological analysis. The microvessels were quantified by evaluating five randomly selected fields under 100× magnification of H&E-stained horizontal sections taken from the middle part of the implants. The microvessels were identified as luminal structures containing red blood cells. The values reported for each experimental condition are the mean values ± the standard deviation, obtained from five to seven individual hydrogel for each group.

VML injury model and implantation of cell-laden hydrogels pre-formed at the subcutaneous space for 7 days or freshly prepared cell-laden hydrogels

BALB/cAnN.Cg-Foxnlnu/CrlNarl mice (aged 7–8 weeks) were anesthetized through inhalation of 1% isoflurane. Hair was removed from the incision site on the hindlimb, and the skin was disinfected using ethanol and iodine solution. An incision was performed to expose the quadriceps muscle on the left foot of mice. Using a biopsy punch tool (diameter: 4 mm) (VWR, USA) to make consistent VML full-thickness defects on the rectus femoris and vastus intermedius of the quadricep ³¹ 2 mm above the knee (Figure S6(b)–(c)). The muscle (4 mm in diameter and 3 mm thick) was removed from the bone to have the VML defect (Figure S6(c)–(d)). Cell-laden tissue constructs pre-formed at the subcutaneous site of mice for 7 days, or freshly prepared cell-laden hydrogels were implanted into the defect sites. Each mouse received one implant for cell-laden or acellular hydrogel groups, and the VML was left untreated in the non-treated empty group. After the injury, the skin was closed using sutures.

Histological and immunofluorescence analyses

Tissue constructs were fixed in 4% paraformaldehyde overnight. Staining of tissue sections using hematoxylin and eosin (H&E) and Masson’s trichrome was performed to investigate the formation of the vascular network and repair of muscle tissue. For immunofluorescence staining, tissue constructs were permeabilized with 0.5% Triton X-100 in DPBS for 25 min. The hydrogels were subsequently blocked with 5% bovine serum albumin (Sigma-Aldrich, USA) in DPBS for 2 h and incubated overnight with rabbit anti-alpha-smooth muscle actin (anti-α-SMA; 1:100 dilution; Clone 1A4; Sigma-Aldrich, USA), mouse anti-human-CD31 (1:100 dilution; Clone JC70A; Dako Cytomation, USA), rabbit anti-laminin (1:200 dilution; Sigma-Aldrich, USA), mouse anti-myosin heavy chain (anti-MHC; 1:100 dilution; DSHB, USA), mouse anti-human nuclear (1:100 dilution; Clone 235-1, Abcam, USA), anti-mouse Ly6 G antibodies (1:50 dilution; Clone1A8, eBioscience, USA), anti-mouse F4/80 antibodies (1:50 dilution; Clone CI A3-1, BD Pharmingen, USA), anti-mouse CD45 antibodies (1:50 dilution; Clone 30-F11, BD Pharmingen, USA), and mouse anti-β-Tubulin III (1:100 dilution; Abcam, USA) antibodies at 4°C. After washing three times with DPBS, the cell-laden hydrogels were incubated with Alexa Fluor 488-conjugated or Fluor 647-conjugated secondary antibodies (1:200 dilution; Invitrogen, USA) at room temperature for 2 h. Finally, sections were stained with Hoechst solution (Invitrogen, USA) and examined using a confocal microscope (TCS-SP5-X; Leica, Germany). For human CD31 immunohistochemical analysis, sections (thickness: 7 mm) were deparaffinized; this was followed by antigen retrieval, and the sections were blocked for 30 min at room temperature. After incubation with mouse anti-human CD31 antibodies (1:100 dilution; Clone JC70A; Dako Cytomation, USA) for 1 h, the sections were incubated with horseradish peroxidase-conjugated secondary antibodies (1:100 dilution in blocking buffer; Vector Laboratories, USA) for 1 h at room temperature. The bound antibody was detected using an ImmPACT DAB substrate kit (Vector Laboratories, USA). Finally, the sample was counterstained with hematoxylin and mounted. After 1- and 2-month, repaired tissue constructs were gently excised from the site of the muscle defect for assessment of muscle revascularization and repair. To investigate the types of tissues grown in tissue constructs, parafilm tissue sections (thickness: 5 μm) in front- and side-view were prepared through H&E and Masson’s trichrome staining. To quantify myofiber regeneration within the constructs, tissue sections were immune-fluorescently double-stained with anti-laminin (1:200 dilution; Sigma-Aldrich, USA) and anti-MHC (1:100 dilution; DSHB, USA) antibodies, stained with Hoechst solution, and mounted with mounting media (Thermo Fisher Scientific, USA); z-stacked fluorescent images were analyzed through confocal microscopy (TCS-SP5-X; Leica, Germany). The total number, thickness, and area of each muscle bundle that dually expressed MHC-positive and laminin-positive myofibers in the tissue constructs were counted using the multipoint tool in ImageJ. The uninjured group at the same age of mice were used as positive control and set as 100% of the tissue area. To quantify staining within the defect regions, the defect region was identified using DAPI staining to manually demarcate an area within the muscle section. A percentage of positive tubulin III in the sections was calculated by the numbers of positive β-tubulin III cells divided by the total numbers of cells into the repaired muscle region.

Statistical analysis

Three independent experiments were performed for each experimental group, with 3–5 replicates per experiment. The results were analyzed using Origin Pro 8.5 (OriginLab, Northampton, MA, USA) and presented as the mean ± standard deviation. Statistical analysis was performed using one-way variance (ANOVA) followed by Tukey’s honest significant difference test for equal variance and Tamhane T2 test for groups with unequal variance. Statistical significance was assumed at p-values < 0.05, 0.01, and 0.001 (indicated as *, **, and ***, respectively).

Synthesis and characterization of acellular collagen and CMA hydrogels

In this study, high yield of acid-soluble collagen was obtained from porcine skin with high purity and bioactivity of type I collagen molecules according to our optimal conditions (Figure S2). To attain a photocurable collagen hydrogel, CMA conjugates were synthesized via direct reaction between the free lysin amino groups in collagen molecules (Collagen) and methacrylic anhydride through nucleophilic substitution using low-concentration (0.0625%) collagen aqueous solutions at pH 6 and 4°C. The characterization of CMA using 1 H-NMR spectroscopy (Figure 1(a)) displayed the presence of a peak at δ = 1.4, 1.9, and 2.8 ppm characteristic of the methyl group of methacrylates,^23,31,32 implying the successful synthesis of CMA conjugates. The structural integrity of CMA was verified using CD analysis. The CD spectra of unmodified collagen (Collagen) (Figure 1(b)) showed negative absorption in the region of 190–200 nm. We prepared Collagen hydrogels that are physically crosslinked via hydrogen bonds between native collagen molecules with concentrations of 0.3% and 0.6%, and CMA hydrogels that are chemically crosslinked via covalent bonds between modified MA groups through free-radical polymerization reaction with exposure to light for various periods. These preparations were conducted to understand the tunability of the properties of these collagen-based hydrogels (i.e. acellular Collagen and CMA hydrogels). The mechanical properties of hydrogels were evaluated through rheological measurements. With increasing concentration of collagen molecules from 0.3% to 0.6%, the gelation time of the Collagen hydrogel was markedly decreased from 710 s to 180 s, whereas its storage modulus (G’) (Figure 1(c)) was increased by 50% from 48 Pa to 72 Pa. Furthermore, the prepolymer of Collagen with a concentration >0.6% was excessively gelling, thereby impairing the preparation of uniform cell-laden gels. Under 0.3% collagen concentration, the G’ of CMA hydrogels (Figure 1(d)) increased by approximately four-fold (from 47 to 187 Pa) through an increase in the time of exposure to light from 10 s (CMA-10) to 20 s (CMA-20). Collectively, the adjustability of the gelation time and storage modulus of Collagen hydrogels was limited and highly dependent on the concentration of collagen molecules. In contrast, CMA hydrogels were more tunable in terms of stiffness, which is independent of collagen concentration compared to the collagen hydrogels. We examined the effects of hydrogels on nutrient transport within the hydrogel by monitoring the diffusion rate of 70 kDa FITC-dextran in an aqueous solution (Figure 1(e)). The results showed that the cumulative release rate of 50% of the dextran from the 0.3% Collagen hydrogel was approximately 2.5-fold faster than that noted from the 0.6% Collagen and CMA-15 hydrogel, and four-fold faster than that of CMA-20 hydrogel. Interestingly, the permeability of the hydrogel was highly dependent on the stiffness of the hydrogel instead of the crosslinking methods in these soft gels (G’ <200 Pa) (Pearson coefficients of correlation, r = 0.792) The microstructure of collagen-based hydrogels was investigated by confocal microscopy (Figure 1(f)) and scanning electron microscopy (Figure 1(g)). The results showed numerous low-density and interconnected fibrils in Collagen, and a densely packed honeycomb-like structure in CMA hydrogels, which possessed distinct microstructures. An increase in collagen concentration significantly increased collagen fibrils’ density in Collagen hydrogels. The CMA-10 is too soft to encapsulate cells to form a cell-laden CMA-10 hydrogel, so was not used for the following studies. In CMA-15 and CMA-20 hydrogels, the pore size and density decreased with increasing time of exposure to light. In CMA-20 hydrogels, pores were barely visible. In this study, we successfully synthesized Collagen and photocurable CMA hydrogels precisely without substantially altering the native secondary structure and bioactivity of collagen. Both resultant hydrogels exhibited tunable physio-biochemical properties over the mechanical properties, microstructure, and permeability. To successfully create vascularized tissues with tunable vessel density in vivo, we selected the 0.3% Collagen, CMA-15 (exposure to UV: 15 s), and CMA-20 (exposure to UV: 20 s) hydrogels with distinctly different levels of stiffness and permeability for the subsequent experiments.

OPEN IN VIEWER

Figure 1.

(a) ¹H NMR spectra of Collagen and CMA macromers. Successful conjugation of MA groups was labeled with blue arrows. (b) CD spectra of Collagen and CMA prepolymer solutions at 4°C. (c and d) The mechanical strength of (c) acellular Collagen and (d) acellular CMA hydrogels can be tailored by altering the protein concentrations and the time of exposure to UV, respectively. (n = 3) (e) Cumulative release profiles of 70 kDa FITC-dextran from acellular Collagen and CMA hydrogels with various crosslinked conditions (n = 3–4). The microstructures of acellular hydrogels were studied using (f) laser confocal microscopy and (g) scanning electron microscopy (SEM).

Hydrogel-mediated encapsulated cell behavior in vitro

HUVECs and MSCs commonly used in engineering three-dimensional (3D) vascularized constructs were used as model cells to investigate cellular responses inside collagen-based hydrogel scaffolds. Crosslinkers were always considered harmful to cells, so the conditions used for the photo-crosslinking process (0.125% (w/v) photoinitiator (PI) and 0–60 s of UV irradiation) were first confirmed nontoxic for both HUVECs and MSCs ²⁹ (viability >90%; Figure S3(a)–(d)). To evaluate the biocompatibility to of our collagen-based hydrogels to cells, NIH3T3, HUVECs and MSCs cells were seeded on the top of preformed Collagen and CMA-15 gels and cultured in a two-dimensional environment. After 2 days, there is no difference in cell proliferation rates between the Collagen and CMA-15 groups among three kinds of cells (proliferation rate >90%; Figure S3(e)) compared to that on the standard tissue culture plates. To explore the 3D cell behavior and capability of vasculogenesis in different collagen-based hydrogels, HUVECs, and MSCs were comixed with Collagen and CMA prepolymer solutions to form Collagen, CMA-15, and CMA-20 cell-laden hydrogels. After 2 days of culture (Figure 2(a) and (b)), HUVECs (hCD31 + (green) and/or α-SMA+ (red), yellow and transparent yellow arrows in Figure 2(b)) significantly elongated and well-spread along MSCs (α-SMA+ (red), white arrows in Figure 2(b)) into Collagen and CMA-15 hydrogels, while both cells in CMA-20 hydrogels were kept round in both bright field and fluorescence-stained images. To verify the cell behavior, we acquired the fluorescent images by a confocal microscope to quantify the area for HUVEC and MSCs using ImageJ software. There is no significant difference in the spreading area of the embedded HUVEC cells (Figure 2(c)) between the Collagen and CMA-15 hydrogels, but it significantly decreased in the CMA-20 gel exhibiting only about 60% of the cell spreading area observed in the Collagen gel. The same trend was also observed on embedded MSCs (Figure 2(d)) in hydrogels, probably due to the stiffer, lower permeability in CMA-20 hydrogel. No lumen structures are observed in any cell-laden hydrogels after 2 days of co-culture. The soft hydrogels no matter the crosslinked methods, that is, Collagen and CMA-15, with better oxygen permeability can support cell spread in vitro. Here, collagen-based hydrogels provided a permissible environment to support cell viability and spreading in 2D and 3D culture systems.

OPEN IN VIEWER

Figure 2.

Representative (a) optical and (b) confocal fluorescence images of co-encapsulated HUVECs and MSCs in the Collagen, CMA-15, and CMA-20 hydrogels. The cells were immunostained with hCD31 (green) and α-SMA (red), and counterstained with Hoechst for nuclei (blue). The labeling of HUVECs is as follows: HUVECs exhibiting both hCD31+ (green) and α-SMA+ (red) staining are denoted by yellow arrows, while HUVECs solely displaying hCD31+ (green) staining are indicated by transparent yellow arrows. MSCs, on the other hand, are identified by α-SMA+ (red) staining alone, and these cells are pointed out by white arrows. Cell spreading area of the co-culture of (c) HUVECs and (d) MSCs inside the indicated hydrogel group. Data are presented as the mean ± SD. *** p < 0.001 indicates significant differences versus the Collagen group at the same time point (n = 3–5).

Control of perfused vascular formation in the cell-laden hydrogels at a subcutaneous site through tailoring of the physical properties of hydrogels

The timely and adequate supply of blood is critical to tissue repair and regeneration. Therefore, the formation of functional blood vessels in the large-sized construct within a week is crucial for increasing the survival rate of transplanted tissues or regeneration rate at the site of injury. To examine the effect of acellular hydrogel on the formation of a vascular network, we implanted acellular Collagen, CMA-15, and CMA-20 hydrogels (diameter: 10 mm; thickness: 2 mm) into a subcutaneous site in mice. After 7 days, no vessel ingrowth or formation inside the implanted gels was observed in any of the groups (Figure S4).

Simultaneous improvement of vasculogenesis and angiogenesis is the strategy for shortening the necessary time for the formation of functional vascular networks inside constructs. Thus, HUVECs and MSCs were co-encapsulated into Collagen, CMA-15, and CMA-20 hydrogels, and implanted cell-laden hydrogels subcutaneously into mice to engineer tissue-engineered functional blood vessels. After 7 days, cell-laden hydrogels were removed to analyze their volume and formation of stable vascular networks. In Figure 3(a), the thicknesses of the cell-laden Collagen and CMA-20 hydrogel were similar (Collagen: 1.55 mm; CMA-20: 1.61 mm), and only ~55 % of that observed in CMA-15 structures (2.86 mm). The horizontal-sectional area of the cell-laden Collagen hydrogels was the smallest (~57 mm²), while those of the other hydrogels were similar (~81 mm²). The volumes of explants were: cell-laden Collagen (~40 mm³), CMA-20 (~68 mm³), and CMA-15 (~103 mm³) (Figure 3(e)). The chemical-crosslinked covalent bonds inside CMA hydrogels allowed 1.5-2.5-fold better retention of the thickness of hydrogels by compared with the physical-crosslinked Collagen hydrogels. Moreover, the spacer allowed two-fold better retention of the thickness of the Collagen hydrogels versus that of others produced without using the spacer previously reported in the literature.^24,33–35

OPEN IN VIEWER

Figure 3.

Cell-laden hydrogels formed in the subcutaneous space of mice for 7 days. (a) Macroscopic images (inset) and surface topography scanned using laser scanning confocal microscopy of the cell-laden hydrogels. (b) Representative horizontal-sectional H&E images of the cell-laden hydrogel revealing distribution of perfused blood vessels (blue arrows) in the structures. (c) Representative images of sections stained with human CD31+HUVECs identified by immunohistochemistry. (d) Mature perfused human vessels were lined exclusively with hCD31-expressing HUVECs (green) and surrounded by αSMA-single positive expressing MSCs (red). Nuclei (blue) were labeled with Hoechst. The extent of vascular network formation was quantified by measuring the entire explant volume, and counting the numbers and areas of erythrocyte-filled lumens. (e) Explant volume, (f) vessel density, and (g) vessel area in the engineered cell-laden tissue constructs removed 7 days after implantation. *p < 0.1 and *** p < 0.001 indicate significant differences versus the Collagen group. Hash symbols (## p < 0.01 and ### p < 0.001) indicate significant differences versus the CMA-15 group (n = 5–7).

Histological examination via H&E staining of both horizontal (Figures 3(b) and S5) and cross sections (Figure S6) revealed numerous mature and functional blood vessels containing erythrocytes in Collagen and CMA-15 explants, which were evenly distributed throughout the whole explants (Figures 3(b)–(d) and S5). Fewer perfused blood vessels were observed in the CMA-20 hydrogels; these vessels were negative for human CD31 and mainly distributed on the periphery of the constructs, showing the ingrowth of vasculature in mice through angiogenesis. In Collagen and CMA-15 explants, most micro-vessels were stained positive for human CD31 (Figure 3(c) and (d)), confirming that the vessels were lined with implanted HUVECs. A close examination of the human micro-vessels revealed the presence of murine erythrocytes (Figure 3(c)), indicating that functional anastomoses with the existing murine vasculature had occurred. These results demonstrated that the formation of microvascular vessels within the hydrogels was induced by the implanted cells, and that this formation was not merely due to invasion by host blood vessels. Furthermore, most human micro-vessels were completely covered by α-SMA-single positive perivascular cells (Figure 3(d)), indicating vessel maturation. Quantitative evaluation revealed that the density of perfused blood vessels was: Collagen hydrogel (51.5 ± 21.1 vessels/mm²), CMA-15 hydrogel (14.2 ± 2.0 vessels/mm²), and CMA-20 hydrogel (1 ± 1.8 vessels/mm²) (Figure 3(b) and (f)). Importantly, the area covered by the perfused blood vessels was larger in the Collagen hydrogel, followed by CMA-15 and CMA-20 hydrogels (Figure 3(b) and (g)). In addition, the density of the perfused blood vessels in the horizontal sections was 2-threefold larger than that noted in the cross sections. This indicated that the growth of the engineered vascular network within the constructs occurred parallel to the skin (Figures 3(f) and S6). The present results showed that pre-vascularized constructs with three degrees of functional vascular networks were successfully created into collagen-based hydrogels within 1 week through precise control of the stiffness and permeability of hydrogels, corresponding to low (CMA-20), intermediate (CMA-15), and high (Collagen) density of vascular networks in vascularized soft tissues. These constructs can be used to investigate the effects of vascularized tissue graft on repair of VML defects.

Acellular hydrogels possess limited capability for repair of VML in muscles

To investigate the muscle repair and regeneration on the critical defect size, we established a VML model at the left rectus femoris in mice through the creation of a non-recovery muscle defect (Figure S7). Acellular CMA-20, CMA-15, and Collagen grafts with no microvessels pre-formed at the subcutaneous site for 7 days (Figure S4) were implanted into the muscle’s injury site. The average area of tissues in the uninjured group was set as 100%. After 1 month, the void created by surgery at the defect site remained visible from gross appearance in all groups (left panel in Figure 4(a)–(c)). There are no remaining hydrogels at the injury site in all groups. In all acellular groups, there was <60% of whole reconstructed tissues and <40% muscle tissues compared with those observed in the uninjured group. Furthermore, there was no significant difference in the area of repaired connective and muscle tissues between the groups, as shown by collagen staining with Masson’s trichrome stain (middle and right panels in Figure 4(d)–(f)). These results demonstrated that acellular Collagen, CMA-15, and CMA-20 hydrogels possess the limited capability to repair VML injured muscle.

OPEN IN VIEWER

Figure 4.

Volumetric skeletal muscle tissue reconstruction by implanting the acellular CMA-20, CMA-15, and Collagen hydrogels in the injury VML model, at 4 weeks after implantation. (a-c) Macro-images with different perspectives at the defect site after repair are presented in the left column (yellow-dashed line). Representative Masson’s trichrome staining images of the repaired construct in the front-view and side-view sections are shown in the middle and right columns. The repaired muscle (M) and connective tissue (C) were labeled. Quantification of (d) whole reconstructed area, (e) muscle, and (f) whole tissue area in the reconstruction at the defect site (the uninjured group was set as 100%). *p < 0.1 indicate significant differences versus the acellular Collagen group. (n = 4–5).

Muscle construction at the critical defect in VML injured muscles by various degrees of vascularized cell-laden hydrogel

Next, to investigate the effects of various degrees of pre-vascularized tissue constructs on restoring VML, pre-formed perfused vascularized soft tissue grafts of various degrees (i.e. low, intermediate, and high) were implanted into the muscle defect site of mice. The non-treated and uninjured groups served as the negative and positive control, respectively; the uninjured group was set as 100% for calculating the repair of tissues (Figure S8). After 1 month, in the non-treated group, the void at the defect area created by surgery remained visible from gross appearance (Figure 5(a)). And, there are no remaining hydrogels at the injury site in all cell-laden hydrogel groups. In the low (Figure 5(b)) and intermediate groups (Figure 5(c)), the atrophic appearance of the muscle partially attenuated compared with that observed in the non-treated group. Surprisingly, in the high pre-vascularization group, it was found that the created void was filled, and the surface gloss of the muscle was also restored (Figure 5(d)). To investigate the histological changes at the defect site 1-month post-injury, the front- and side-view sections of the repaired tissue were stained with Masson’s trichrome staining. Consistent with the observation of gross appearance, in the non-treated empty and low pre-vascularization group (right panel in Figure 5(b)), VML defect could still be seen and there was little tissue repair (right panel in Figure 5(a)). In the intermediate pre-vascularization group (right panel in Figure 5(c)), the repair diameters with more regenerated muscle and less dense connective tissues were larger by 1.5 times than that in the non-treated and low pre-vascularization groups. A significant increase in the volume of reconstructed tissue was observed in the high pre-vascularization group (reconstruction tissue and muscle ~ 90%). This treatment exerted a greater effect on muscle regeneration and reduction of the adipose and fibrosis connective tissues. (right panel in Figure 5(d)–(g)).

OPEN IN VIEWER

Figure 5.

Volumetric skeletal muscle tissue reconstruction by implanting the vascularized cell-laden hydrogels pre-formed at the subcutaneous space for 7 days in the injury VML model at 4 weeks after implantation. (a) The non-treated empty group served as the negative control. According to the density of the engineered functional vascular networks inside, the pre-vascularized cell-laden hydrogels were classified as (b) CMA-20 (Low), (c) CMA-15 (Intermediate), and (d) Collagen (High) groups. Macro-images with different perspectives at the defect site after repair are presented (yellow-dashed line in the left column). Representative Masson’s trichrome staining images of the repaired construct in the front-view and side-view sections are shown in the middle and right columns. The inset was the enlarged region of connective tissue in the cell-laden hydrogel. The repaired muscle (M) and connective tissue (C) were labeled. Quantification of (e) whole reconstructed area, (f) muscle tissues area, and (g) distribution of connective and muscle tissues in the whole reconstructed area compared with those determined in the uninjured group at the defect site (the tissue areas in the uninjured group were set as 100%). (n = 4–5) *p <0.05, ** p < 0.01, and *** p < 0.001 indicate significant differences versus the CMA-20 (Low) and non-treated group. Hash symbols (#p < 0.05) indicate significant differences versus the CMA-15 (Intermediate) group. $p < 0.05 and $$p < 0.001 indicate significant differences versus their respective acellular groups.

Preformed perfused and well-connected vascular networks supported the formation of mature muscle fibers

Histological analyses were performed by double-staining, namely laminin (red) and MHC (green), to quantitatively calculate the distribution and average size of repaired muscle fiber bundles after 1 month (Figure 6(a)). According to the front-view section of the repaired muscle region in construct (Figure 6(b)), the non-treated group presented significantly higher numbers (1100 #/mm²) of myofibers, but with reduced size (<500 μm² each), while the non-injury group showed fewer numbers of the bundle (~530 #/mm²) with the largest bundle area (>1500 μm² each). The low and high pre-vascularization groups showed similar numbers of myofiber bundles (~900 #/mm²), while only ~600 #/mm² myofiber bundles were noted in the intermediate group. The intermediate and high pre-vascularization groups demonstrated similar size distribution of myofiber bundles, which was larger than that recorded in the low pre-vascularization group. Based on the side-view section of the repaired construct (Figure 6(c)), there was no significant difference in the number and size of myofiber bundles between the groups; notably, the number in the low pre-vascularization group was slightly lower than in the other groups. Taken together with the results on repair muscle size (Figures 5 and 6), these results indicate improvement of muscle repair in the high pre-vascularization group and impairment in muscle regeneration in the non-treated and low pre-vascularization group.

OPEN IN VIEWER

Figure 6.

Muscle regeneration at the repaired muscle region in VML defects at 4 weeks after implantation. (a) Representative laminin (red)/MHC (green)/nuclei (blue) staining images of the repaired muscle in VML defect in the front view (top row) and side view (second row) implanted by the non-treated, cell-laden CMA-20(Low), cell-laden CMA-15 (Intermediate), cell-laden Collagen (High), and non-injury groups. The non-injury group served as the positive control. Scale bar: 50 μm. (b and c) Quantification of muscle fiber bundle areas and numbers in the (b) front-view and (c) side-view sections of the repaired tissue construct. ***p < 0.1 indicate significant differences versus the non-treated group (n = 4–5).

Freshly prepared cell-laden hydrogels impeded muscle repair at the VML injury site

We sought to determine whether the improvement in regenerative capacity achieved using the high-degree perfused vascularized cell-laden hydrogel s for the treatment of muscle defects was due to a connection between the preformed well-connected vascular network and host vasculature. For this purpose, instead of subcutaneously implanting pre-vascularized tissue grafts, freshly prepared cell-laden Collagen (Fresh cell-laden Collagen) and CMA-15 (Fresh cell-laden CMA-15) hydrogels were directly implanted into the injury site after the creation of VML (serving as non-pre-vascularized groups). After 1 month, the size and area of the repaired connective and muscle tissues formed at the injury site and myofibers were analyzed (Figure 7). And, there are no remaining hydrogels at the injury site in both fresh prepared cell-laden hydrogel groups. Direct implantation of freshly prepared cell-laden Collagen hydrogels (Fresh cell-laden Collagen group) significantly impeded the repair in connective and muscle tissues compared with the cell-laden Collagen group. Moreover, the same effect was observed in the Fresh cell-laden CMA-15 groups, although it did not reach statistical significance (Figure 7(c) and (d)). This may be attributed to insufficient muscle repair in the cell-laden CMA-15 group. Furthermore, no significant difference was found in the area and number of repaired myofibers between the groups (Figure 7(e) and (f)).

OPEN IN VIEWER

Figure 7.

Volumetric skeletal muscle tissue reconstruction by directly implanting the freshly prepared cell-laden Collagen (Fresh Collagen) and CMA-15 (Fresh CMA-15) hydrogel in the injury defect site at 1 month after implantation. (a and b) Macro-images with different perspectives at the defect site after repair are presented (yellow-dashed line in the left column). Representative Masson’s trichrome staining images of the repaired construct in the front-view and side-view sections are shown. The repaired muscle (M) and connective tissue (C) were labeled. Representative laminin (red)/MHC (green)/nuclei (blue) staining images of implants in the front view (in the right column). (Scale bar: 50 μm) Quantification of (c) whole reconstructed area, (d) muscle, and muscle fiber bundles areas and numbers in the (e) front-view and (f) side-view sections of the repaired tissue construct (the uninjured group was set as 100%). *** p < 0.001 indicate significant differences versus the Collagen group (n = 3–4).

Fast anastomosis of vascularized cell-laden hydrogel with host micro-vessels at the site of injury to improve muscle repair

To further investigate the promotive effect of vascularized cell-laden hydrogel on muscle repair, the pre-vascularized cell-laden collagen hydrogels and the non-pre-vascularized freshly prepared cell-laden collagen groups were analyzed at different time points immediately after surgery. The findings revealed the progression of vasculogenesis, angiogenesis, and anastomosis at the injury site during the repair process (Figure 8).

OPEN IN VIEWER

Figure 8.

Pre-vascularized cell-laden hydrogel construct formed at the subcutaneous site supported volumetric skeletal muscle tissue reconstruction through the successful connection between engineered vascular networks and host vasculature. (a) The representative H&E and hCD31+ images of the cell-laden Collagen group (Collagen) pre-formed at the subcutaneous site for 7 days before (pre-D7) and after 3 days of implantation in the muscle defect site. (b) The representative H&E and hCD31+ images of the freshly prepared cell-laden Collagen group (Fresh Collagen) at different time points after direct implantation into the muscle defect site (n = 3). Representative immunofluorescence staining of (c) Ly6G+(green) CD45+(red) neutrophils and (d) F4/80+ (green) CD45+(red) macrophages. Nuclei are stained with DAPI. (e) Quantification of neutrophils and macrophages in the Collagen group and the Fresh Collagen on day 3 (D3) after implantation.

In the pre-vascularized cell-laden collagen hydrogels group (Collagen group; Figures 8(a) and S9(b)), on day 3 after implantation (D3), all lumen structures were perfused and uniformly distributed in the constructs (~45 vessels/mm²). This indicates that the lumen structures were unchanged on day 3 compared to before implantation (51 vessels/mm² on Day 0). This implied that angiogenic sprouting from the host vasculature and anastomosis with the engineered perfused lumens had occurred within 3 days after implantation. In the freshly prepared cell-laden Collagen group (Fresh cell-laden Collagen group; Figure 8(b)), on day 3 (D3), there were some perfused blood vessels (~20 lumens/mm²) at the peripheral site and some non-perfused lumen structures in the central region of the constructs. By day 5 (D5), the lumen density was at a similar level; however, a larger number of perfused vessels was observed in the central region. These findings demonstrate that the processes of vasculogenesis and angiogenesis were initiated together, whereas anastomosis from the edges of the constructs began later. By day 7 (D7), almost all the lumen structures located in the central regions had been perfused, thus increasing the density of perfused vessels (~53 vessels/mm²); these mature vessels were stabilized by day 10 after implantation (55 vessels/mm²; Figure 8(b) and S9(a)). To understand whether host myeloid cells, neutrophils, and macrophages, participate in the revascularization of vascularized cell-laden hydrogel here, immunofluorescent staining was performed at the injury site on day 3 after implantation (D3). More Ly6G+ CD45+ neutrophils were observed into the Collagen group compared to that in the Fresh Collagen group, while the densities of F4/80+ CD45+ macrophages were similar in both groups (Figure 8(c)–(e)). And, the area of perfused micro-vessels in the prevascularizd cell-laden collagen hydrogels (cell-laden Collagen PreD7+D3, pre-formed in the subcutaneous space of mice for 7 days, and exercised and implanted into the other mice for 3 days, Figure S9(b)) was three-fold larger than that observed in the non-prevascularized group (Fresh cell-laden Collagen D10, Figure S9(a)) for 10 days in animals. Moreover, the total density of perfused lumens in the freshly prepared cell-laden collagen group after 1-week (Fresh cell-laden Collagen D7: ~53 vessels/mm²) was similar to that recorded for the vascularized group at the subcutaneous site (Collagen PreD7: ~45 vessels/mm²); nevertheless, the average area of micro-vessels remained significantly smaller. In the freshly prepared cell-laden group, there was a time delay in forming blood vessels and the perfusion process. To clarify whether this time delay significantly affected muscle repair, we extended the observation time to 2 months after surgery (Figure S10). Results showed in the prevascularization group (cell-laden Collagen group; Figure S11(a)–(c)), the total area of reconstruction tissue remained constant, but the muscle area increased by 1.5 times from 2 to 3 mm² after 2 months. In the non-prevascularization group (Fresh cell-laden Collagen group; Figure S11(a)–(c), the total area of reconstruction tissue significantly increased from 0.5 to ~ 3 mm², and the area of muscle significantly increased to 3 mm², which is similar to that in the prevascularization cell-laden Collagen group.

Neuromuscular regeneration after critical VML

To characterize the innervation of motor neurons and repaired myofibers at the injury site, we performed immunofluorescent staining sections for β-Tubulin III on the repaired tissue construct to visualize the presence of neurons (Figure 9(a) and (b)). The healthy uninjured group of the same age of mice was used as the positive control. At 1 month after implantation, similar percentages of β-Tubulin III-positive neurons surrounded by repaired myofibers were observed in the uninjured group and pre-vascularized cell-laden Collagen group, which are more than that in the freshly prepared cell-laden Collagen group (Figure 9(b)). Besides, the total repaired muscle tissue in the pre-vascularized cell-laden Collagen group is 4 times larger than that in the Fresh cell-laden Collagen group (Figure 7(d)). To confirm these differences in muscle repair, the walking behaviors of mice treated with the uninjured healthy, non-treated group, cell-laden Collagen group and Fresh Collagen groups were recorded after 1 month of treatment (Video S1). In the Fresh cell-laden Collagen group, mice displayed abnormalities in exploratory behavior, such as insufficient strength in the back-left leg. The results were in line with the quantification data on motor neuron innervation and the amount of repaired muscle volume observed at the injury site 1-month post-surgery. These experiments suggested that the cell-laden Collagen group pre-formed at the subcutaneous site for 7 days compared to the freshly prepared cell-laden Collagen group, promoted functional recovery resulting from accelerated tissue regeneration.

OPEN IN VIEWER

Figure 9.

Pre-vascularized cell-laden hydrogel promoted the repair of damaged skeletal muscle and the recovery of motor function in mice. (a) Representative immunofluorescent staining for innervation (β-Tubulin III, green) in the front-view sections of the non-injury group, the pre-vascularized cell-laden Collagen group (Collagen), and the freshly prepared cell-laden Collagen group (Fresh Collagen) at 1-month after implantation. (b) The quantitation analysis of the β-Tubulin III-positive neurons in repaired muscle regions for 1-month post-surgery. White arrows indicate β-Tubulin III-positive neurons. (c) Motion functional deficits were assessed by the balance beam and inclined plane test (n = 6). (*p < 0.05, **p < 0.01, and ***p < 0.001 as compared to the non-injury group. #p < 0.05 and ##p < 0.01 as compared to the Collagen group. N.S. means no significant difference).

These results demonstrated that the time delay of muscle repair was observed in the freshly cell-laden group compared to that in the vascularized tissue graft for muscle repair. Moreover, uninjured muscle tissues at the quadriceps muscle on the right foot of injured mice were also investigated to understand the injured effects on the other uninjured foot (Figure S11(d)–(e)). Results showed that the muscle area on the right uninjured quadriceps significantly increased with time after injury (Figure S10(e)), implying that the VML injury caused extra weight loading on the other uninjured foot to increase its strength during recovery, like rehabilitative therapies. The use of well-connected perfused vascular networks flap at the defect site immediately after injury plays an important role in the following muscle remodeling and repair process. These findings demonstrate that our pre-engineered human vascular networks connected and revascularized with host vasculature within 3 days after implantation and further supported muscle tissue repair and regeneration. In summary, these results show that the prevascularized cell-laden hydrogels provided a permissive environment for the formation of blood vessels, which further supported muscle repair at the damage site.