Reconstitution of the complete rupture in musculotendinous junction using skeletal muscle-derived multipotent stem cell sheet-pellets as a “bio-bond”

Animals

Green fluorescent protein transgenic mice (GFP-Tg mice; C57BL/6 TgN[act EGFP]Osb Y01, provided by Dr. M Okabe, Osaka University, Osaka, Japan) (Okabe et al., 1997) were used as donor mice for the cell transplantation experiments (male, 4–8 week old, n = 6), and wild-type mice (C57BL/6N) were used as recipients (male, 8–12 week old, n = 26). All experimental procedures were approved by the Tokai University School of Medicine Committee on Animal Care and Use (153015).

Cell purification and preparation of stem cell sheet-pellets

Sk-MSC sheet-pellets (Tamaki et al., 2013) were investigated for their effects on the regeneration of completely ruptured MTJ. Sheet-pellets generally showed gel-like characteristics (Fig. 1A), and they could be lifted using forceps (Fig. 1B). The thigh and lower leg muscles (tibialis anterior, extensor digitorum longus, soleus, plantaris, gastrocnemius, and quadriceps femoris) of GFP-Tg mice were removed and used in subsequent experiments. Muscle sampling was performed under an overdose of pentobarbital (60 mg/kg, Schering–Plough, combined with butorphanol tartrate 2 mg/kg, Meiji Seika, Tokyo, Japan, i.p.). Average total muscle mass removed during the procedure was 512 ± 67 mg/GFP-Tg mouse (mean ± SE). Muscles were not minced, and were subsequently treated with 0.1% collagenase type IA (Sigma–Aldrich, Tokyo, Japan) in Dulbecco’s modified Eagle’s medium (DMEM, Wako, Osaka, Japan) containing 7.5% fetal calf serum (FCS, Equitech Bio, TX, USA) with gentle agitation for 30 min at 37 °C. Following a short digestion, whole muscles were divided into fiber-bundles, which were washed with culture medium (Iscove’s modified Dulbecco’s medium; IMDM, Wako, Osaka, Japan) containing 10% FCS, and cultured in IMDM/20% FCS with 100 units/ml penicillin G, 100 µg/ml streptomycin sulfate (Wako, Osaka, Japan), 10 µg/ml gentamycin sulfate (Schering–Plough, Osaka, Japan), and 0.1 mM β-mercaptoethanol (Wako, Osaka, Japan) for 3 days. Cultured fiber-bundles associated with expanded cells were treated with trypsin-EDTA (0.05% trypsin, 0.53 mM EDTA; Life Technologies, Tokyo, Japan), in order to dissociate individual cells. Single-cell suspension was filtered through 70-, 40- and 20-µm nylon filters in order to remove muscle fibers and other debris, and, after washing, isolated cells were re-cultured in IMDM/20% FCS for 2–3 days, until they reached confluence. In total, cells were expanded in culture for 5–6 days, which represents a reduced culture time compared with previous studies (Soeda et al., 2013; Tamaki et al., 2013), thus preserving the myogenic potential of sheet-pellets. After they reached confluence, the cells were gently detached from culture dishes using 2 mM EDTA solution. This step mainly affected Ca²⁺-dependent cell adhesion (e.g., cadherins), while cell-to-cell contact was maintained, sheet-like cell aggregations were collected and centrifuged, and stem cell sheet-pellets were obtained. Total sheet-pellet mass was 120 ± 17 mg/mouse (mean ± SE), which means that over 100 mg of sheet-pellets were obtained from about 500 mg of skeletal muscle tissue. Throughout cell isolation, 7.5–10% FCS was added to the collagenase and washing solutions in order to minimize contaminating protease activity and to protect isolated cells as much as possible.

RT-PCR and protein analysis of the sheet-pellets

Quality, differentiation potential and relative expressions of cytokines of the cells forming sheet-pellets was confirmed by RT-PCR and antibody array analysis immediately before the transplantation. Cells composed sheet-pellet was prepared for the RT-PCR, and their supernatant was prepared for antibody array analysis. For the RT-PCR analysis, specific primers and the analyzed materials are summarized in Table S1. Cells were lysed and total RNA was purified using a QIAGEN RNeasy Micro Kit (Hilden, Germany). First-strand cDNA synthesis was performed with Invitrogen SuperScript III system using dT30-containing primer (see Table S1), and specific PCR (35 cycles of 30 s at 94 °C, 30 s at 60–65 °C and 2 min at 72 °C) was performed in a 15-µl volume containing Ex-Taq buffer, 0.8 U of ExTaq-HS-polymerase, 0.7 µM specific sense and antisense primers, 0.2 mM dNTPs, and 0.5 µl of cDNA. Relative expression was normalized to the expression of a housekeeping control (HPRT). Details of this analysis were described previously (Tamaki et al., 2013).

Concurrently, several cytokines, which are related to the muscular and vascular regeneration, were also analyzed by antibody array kit (Proteome Profiler, ARY013, R & D Systems, Minneapolis, MN), as a protein level. Cell culture supernatant of sheet-pellet just before the transplantation was obtained after removal of particulates by centrifugation, and 500-µl of supernates was prepared for the analysis. Culture medium containing 20% FCS was also prepared for the same analysis in order to check the background effects. The relative expression levels of several cytokines; such as interleukin-6 (IL-6), insulin-like growth factor-1 and 2 (IGF-1 and 2) and their related proteins IGF binding protein-1, 2, 5, and 6 (IGFBP-1, 2, 5 and 6), which are the critical regulator of myogenesis; a regulator/enhancer of macrophage such as chemokine (C–C motif) ligand 2 (MCP-1) and colony stimulating factor 1 (M-CSF); a regulator/activator of wide range of cell types such as fibroblast growth factor-21 (FGF-21) and tissue inhibitor of metalloproteinase 1 (TIMP-1); and vascular endothelial growth factor (VEGF) were determined.

Complete rupture of MTJ model and application of sheet-pellets

In order to prepare the experimental complete rupture model, we manually detached muscle fibers from the distal tendons of the right tibialis anterior (TA) muscles of recipient mice (n = 19). All surgical preparations were performed under the inhalation anesthesia (Isoflurane; Abbot, Osaka, Japan). A summary of the procedure is shown in Fig. 1. First, TA muscle was exposed (C, step 1), and muscle fibers were subsequently detached from the distal tendon and peeled off using cotton swabs for one half of the entire TA length (D, step 2), and everted necrotic fiber portions were removed (step 3). The average removed muscle mass was 21.5 ± 1.8 g, and this represented about 40% of total muscle mass. Sheet-pellets were then adhered to the open region of completely ruptured MTJ (Fig. 1E, step 4; Sheet-pellet, SP group, n = 14), and skin was sutured (E, white arrows). The left side was preserved as a contralateral control. The non-transplanted control group (C group, n = 10) underwent the same surgery, and the same amount of cell-free culture medium was administered. The mice were allowed full freedom of movement after surgery.

Functional assessment of the regenerated TA muscles

After the transplantation, the animals were divided into two groups as short-term (4–8 weeks) and long-term (14–18 weeks) recoveries, and prepared for the functional assessment. The first assessment was begun at 4 weeks using each one mouse in SP and C group, and tetanic tension outputs of regenerated TA muscles were measured in both left (non-operated control side) and right (operated side) legs. The tetanic tension output was considered as the total functional recovery of the operated TA muscle, and recovery ratio was determined based on the contralateral non-operated control side. Following assessments were performed every week (0.5–1.5 week interval) during each group terms using one by one mouse, then, SP and C groups were compared. Measurements were performed in situ under inhalation anesthesia (Isoflurane; Abbot, Osaka, Japan), and body (rectal) temperature was maintained at 36 ± 1 °C with a radiant heat light throughout the measurement. Details of this measurement were as described previously (Tamaki et al., 2005). Briefly, the distal tendon of TA muscle and sciatic nerve (about 10 mm) on both sides were carefully exposed, and tissues were coated with mineral oil to prevent tissue drying and to minimize electrical noise interference. A bipolar silver (Ag/Ag) electrode (inter-electrode distance: 2 mm) was placed under the sciatic nerve. A stainless steel hook was attached to the distal tendon of each TA muscle using a silk ligature. The animal was transferred to a custom-made operating table that allowed stabilization of the head and limbs in a supine position using surgical tape. A stainless steel hook was attached to a force-distance transducer (FD-Pickup, TB-611T; Nihon Kohden, Tokyo, Japan) connected to the carrier amplifier (AP-621G; Nihon Kohden). This enabled the measurements of the muscle contraction force and its distance to be conducted. A bipolar silver electrode (inter-electrode distance: 5 mm/1 mm diameter) was attached to the surface of the reference muscle as well, in order to obtain an evoked electrical myogram, as a confirmation of stable muscle contractions. We have taken care to avoid interference of the reference muscle and nerves with the normal blood supply. Afterward, twitches were elicited by single pulse (1 ms duration, 0.5 Hz) electrical stimulation via the sciatic nerve, at a voltage above the threshold for a maximum response (1.5–4.0 V). Subsequently, peak tetanic tension was determined using stimulation frequencies of 10, 20, 40, 60, 80, 100, 120, and 140 Hz of 0.5 s duration at 15 s intervals. The frequency that produced the highest tetanic tension was considered the optimal stimulation for tetanus. All mechanical and electrical measurements were recorded on a Linearcorder (Mark VII, WR3101; Graphtec, Tokyo, Japan) as analog data.

Macroscopic observation and immunostaining

Following the functional measurements above, recipient mice (including the animals used for morphological analysis only, 5–10 weeks after transplantation, n = 5) were given an overdose of pentobarbital (60 mg/kg, combined with butorphanol tartrate 2 mg/kg, i.p.), and the engraftment of donor-derived GFP⁺ cells into the damaged portion of TA muscle was confirmed by fluorescence stereomicroscopy (SZX12; Olympus, Tokyo, Japan, Fig. 4). Recipient mice were perfused with warm 0.01 M phosphate-buffered saline (PBS, Wako, Osaka, Japan) through the left ventricle, followed by fixation with 4% paraformaldehyde/0.1 M phosphate buffer (4% PFA/PB, Wako, Osaka, Japan). Muscles were removed and fixed overnight in 4% PFA/PB, washed with graded sucrose (0–25%, Wako, Osaka, Japan)/0.01 M PBS series, and quick frozen with isopentane (Wako, Osaka, Japan) pre-cooled by liquid nitrogen, followed by storage at −80 °C. Subsequently, 7 µm cross-sections were obtained. Skeletal muscle fibers were stained with anti-skeletal muscle actin (αSkMA; dilution, 1:200; incubation, room temperature for 2 h; Abcam, Cambridge, UK). Nerve fiber localization (axons) was detected by rabbit polyclonal anti-Neurofilament 200 (N-200, dilution, 1:1000; incubation, room temperature for 1 h; Sigma, St. Louis, MO, USA). Schwann cells were detected using anti-p75 (rabbit polyclonal, 1:400, 4 °C overnight; CST, Boston, MA, USA). Blood vessels were detected with rat anti-mouse CD31 (1:500, 4 °C overnight; BD Pharmingen, San Diego, CA, USA) monoclonal antibody, which is a known vascular endothelial cell marker, and mouse monoclonal α-smooth muscle actin (αSMA, Cy3-conjugated; 1:1500; room temperature for 1 h; Sigma, St. Louis, MO, USA). Dystrophin formation in the skeletal muscle fibers was detected using goat anti-dystrophin polyclonal antibody (1:50, 4 °C, overnight; Santa Cruz Biotechnology, Dallas, TX, USA). Neuromuscular junctions were detected by α-bungarotoxin (Alexa Fluor 594 conjugated, 1:100, room temperature,1 h; Molecular Probes, Eugene, OR, USA). Reactions were visualized using Alexa Fluor-594-conjugated goat anti-rabbit and anti-rat antibodies (1:500, room temperature, 2 h; Molecular Probes, Eugene, OR, USA). Nuclei were counter-stained with DAPI (4,6-diamino-2-phenylindole).

Statistical analysis

Differences between two groups (short and long term recovery group) were tested using Student’s t test, and the significance level was set at p < 0.05. Values are expressed as mean ± SE.

Quality and therapeutic potential of Sk-MSC sheet-pellets

Quality, differentiation, and putative therapeutic potential of the sheet-pellet were first confirmed by RT-PCR and protein array analysis (Fig. 2). The expression of specific myogenic, neurotrophic, and vasculogenic factor mRNAs in the sheet-pellets immediately before the transplantation is shown in Fig. 2A. The Sk-MSC sheet-pellets showed expressions of various myogenic factors (MyoD, Myf5, Pax7, Myogenin, c-met, Mcad, MyH, Desmin, and IGF-1), neurotrophic factors (NGF, BDNF, GDNF, CNTF, LIF, Ninjurin, Galectin, Nestin, and Sox10) and vascular growth factors (VEGF, HGF, PDGF, TGF-β, EGF, and FGFb), except Pax3. These results agree with a previous report (Tamaki et al., 2013), which showed good quality of this method of sheet-pellet preparation.

In addition, relative increase in the expressions of proteins (cytokines) was also detected in the same sheet-pellet culture supernatant (Fig. 2B). An increase of myokine (Munoz-Canoves et al., 2013; Pedersen, 2012), which is a critical regulator of muscle regeneration, such as IL-6 and IGF-1 and 2 associate with their relating IGFBP-2, 3, 5 and 6, the chemokine, which is an up-regulator of monocyte/macrophage (MCP-1 and M-CSF) in the tissue regeneration (Pantsulaia et al., 2005; Shiba et al., 2007), and a regulator/activator of a wide range of cell types (FGF-21 and TIMP-1) (Mas et al., 2007; Wan, 2013) was detected, showing a paracrine capacity of the sheet-pellet.

Functional recovery

The results of functional assessment are summarized in Table 1, and group differences between the short-term and the long-term recovery group, with these composed individual plots at each measurement point are also shown in Fig. 3. The SP short term recovery group showed about 64% recovery of muscle mass, and about 28% recovery of tetanic tension output relative to the contralateral side (Figs. 3A and 3B). Then, the SP long term recovery group showed increased recovery of average muscle mass (77%), and significant recovery of tetanic tension output (49%, Figs. 3A and 3B). Importantly, the media control group showed significantly (P < 0.05) lower recovery at 4 week (Fig. 3A) with strong atrophy coalescence into the tibial bone (see next Figs. 4I–4K), and incomplete recover of the muscle–tendon unit continuity. This atrophy apparently progressed toward 6 weeks (compare I and J). Thus, it was clear that this model is irreversible spontaneously. Therefore, we were unable to continue measuring muscle mass and tension output in the media Control group subsequently. Average recovery ratio of muscle mass gradually increased in the long-term group, and the tetanic tension showed significantly higher recovery (Figs. 3A and 3B), showing that a development of recoveries continued/progressed over 14 weeks after operation.

When the recoveries were analyzed individually in each group, a linear relationship between the term and the manner of recoveries was observed to be relatively stronger in the long-term group than that of the short term group. This may be representing the effect of increase in standard body activities of mice in the cages, because we observed that general activities (walking around, hanging and upside-down walking on the cage lid) clearly occurred more frequently in the long-term recovery group than in the short term group.

Macroscopic examination

Typical fluorescence macroscopic features at 4–10 weeks after the engraftment of transplanted GFP⁺ sheet-pellets are shown in Fig. 4. In situ observation revealed that a large volume of GFP⁺ tissues was engrafted in the damaged TA muscle portion after transplantation, and these tissues showed 5 different patterns. Pattern 1 was the most common one (5/13 samples), showing broad and thick engraftment through the tendon to the mid portion of the muscle (Figs. 4A–4C). GFP⁺ engrafted tissue was observed in blood vessel networks (arrows in Fig. 4B), and the continuity was clearly maintained (Fig. 4C). This pattern was mainly composed of connective tissue with few muscle fibers. Pattern 2 was rarely observed (1/13 samples), and it mostly comprised of the muscle fibers (Fig. 4D) that extended from the mid portion of MTJ to the upper portion, and with few connective tissues. Pattern 3 (1/13 samples) in contrast, mainly comprised of the connective tissue at the distal portion of MTJ, with few muscle fibers (Fig. 4E). Patterns 4 (3/13 samples) and 5 (3/13 samples) were mixed types, showing both muscle fibers and connective tissues (Figs. 4F–4H), but GFP was relatively sparse, and the amount of connective tissues was small in Pattern 4 (Fig. 4F), while Pattern 5 showed even distribution of both tissues (Fig. 4G). Active blood vessels in and around GFP⁺ tissues were equally observed in all Patterns (arrows in Figs. 4B, and 4D–4H). Severe atrophy of TA muscle was generally observed in the C group (Figs. 4I and 4J), and this pattern was observed in 7/7 control mice. On the lateral side of the tibia (black arrows in Figs. 4I and 4J), a portion of the TA muscle was clearly hollowed (white arrows in Fig. 4I), and there were no muscle fibers present. Apparent adipose tissue formation (red arrows in Fig. 4J) was also observed, suggesting lack or very weak continuity of the TA muscle–tendon unit (Fig. 4J). An apparent muscle deficit of distal-half of operated TA with a meager tendon was confirmed in the isolated TA muscle of C group (K), suggesting the difficulty of force generation. Progressive muscle atrophy of the remaining proximal-half portion of TA was also apparent (compare I and J in dotted circles).

Immunohistochemical analysis of the engrafted cells

At 4–10 weeks following the surgery, the behavior of engrafted GFP⁺ cells was analyzed in cross-sections. Figure 5 shows the result of Pattern 1. Thick tissues composed of GFP⁺ cells closely adhered to skeletal muscle fibers (Fig. 5A). GFP⁺ cells surrounded the tendon, and van Gieson elastic fiber staining showed connective tissue networks (Fig. 5B). This suggests that the engrafted GFP⁺ tissue formed connective tissue networks, which connected both the tendon and the muscle fibers, displaying a “bio-bond” role. Additionally, invagination of GFP⁺ cell-derived connective tissue was observed in the tendon (yellow arrows in Fig. 5C), and showed differentiation into vascular endothelial cells (CD31 + GFP, white arrows in 5C). A similar trend was also observed on the muscle fiber side, and the migration of GFP⁺ cells could be seen between muscle fibers (Fig. 5D, below the dotted line). A close relationship (not double staining) of N200⁺ nerve axons and GFP⁺ cells, was observed on both the muscle side (below the dotted line) and the connective tissue side (Fig. 5D, upper-side of dotted line). GFP⁺ cells in the connective tissue were positive for p75 (Fig. 5E, right side of dotted line) and therefore, they were considered Schwann cells. These results support a close relationship between GFP⁺ cells and nerve axons, which is detectable in Fig. 5D. The relationship among engrafted GFP⁺ connective tissue, tendon, and muscle fibers is more apparent in Fig. 4F. GFP⁺ connective tissue bridging tendon (T) and muscle fibers (right side of the panel) expressed dystrophin (red), which suggests a strong relationship between connective tissue and muscle fibers through dystrophin complexes.

Similarly, longitudinal profiles obtained from pattern 5 are shown in Fig. 6. Engrafted GFP⁺ cell-derived connective tissues closely adhered to the muscle fibers (Fig. 6A), and GFP⁺ muscle fibers were also observed (Fig. 6B). Differentiation of GFP⁺ cells into vascular smooth muscle cells was seen in the connective tissue network (Fig. 6C, arrows, double staining with GFP + SMA), contributing to a relatively large blood vessel formation. The involvement in the peripheral nerve reconstitution was indicated by a close relationship between GFP⁺ cells and axons (Fig. 6D, arrows, close distributions of GFP⁺ cells and N200⁺ axons). This relationship was supported by the differentiation of GFP⁺ cells into Schwann cells positive for p75 (Fig. 6E, arrows, double staining of GFP + p75). Similarly, a close relationship of GFP⁺ cells and muscle fibers, nerve axons, and the neuromuscular junctions was evident (Fig. 6F, in the dotted line circle as α-bungarotoxin⁺), confirming that GFP⁺ cells contributed to peripheral nerve extensions, reaching to the end of a motor nerve. These results indicate that the transplanted GFP⁺ Sk-MSC sheet-pellets mainly form connective tissue networks together with a certain amount of muscle fibers, and that they physically bind the tendon and muscle fibers, contributing to the peripheral nerve-blood vessel formation.