Restoration of Anal Sphincter Function after Myoblast Cell Therapy in Incontinent Rats

Abstract

Fecal incontinence (FI) remains a socially isolating condition with profound impact on quality of life for which autologous myoblast cell therapy represents an attractive treatment option. We developed an animal model of FI and investigated the possibility of improving sphincter function by intrasphincteric injection of syngeneic myoblasts. Several types of anal cryoinjuries were evaluated on anesthetized Fischer rats receiving analgesics. The minimal lesion yielding sustainable anal sphincter deficiency was a 90° cryoinjury of the sphincter, repeated after a 24-h interval. Anal sphincter pressure was evaluated longitudinally by anorectal manometry under local electrostimulation. Myoblasts were prepared using a protocol mimicking a clinical-grade process and further transduced with a GFP-encoding lentiviral vector before intrasphincteric injection. Experimental groups were uninjured controls, cryoinjured + PBS, and cryoinjured + myoblasts (different doses or injection site). Myoblast injection was well tolerated. Transferred myoblasts expressing GFP integrated into the sphincter and differentiated in situ into dystrophin-positive mature myofibers. Posttreatment sphincter pressures increased over time. At day 60, pressures in the treated group were significantly higher than those of PBS-injected controls and not significantly different from those of normal rats. Longitudinal follow-up showed stability of the therapeutic effect on sphincter function over a period of 6 months. Intrasphincteric myoblast injections at the lesion borders were equally as effective as intralesion administration, but an injection opposite to the lesion was not. These results provide proof of principle for myoblast cell therapy to treat FI in a rat model. This strategy is currently being evaluated in humans in a randomized double-blind placebo-controlled clinical trial.

Introduction

Fecal incontinence (FI) is a frequent chronic debilitating disease with a profound impact on quality of life and is recognized as an economic and public health issue (26,30). FI causes may be underlying neurological disorders or degradation of the anal sphincter, for example, postsurgical or postdelivery (5,31). Current therapy remains insufficient despite proposed therapies such as biofeedback treatment (28), sacral nerve stimulation (18), injection of bulking agents (34), or very invasive surgery for artificial anal sphincter (22). Therefore, cell therapy using autologous myoblasts may represent a new attractive option in FI. This would allow regeneration of the damaged striated sphincter, possibly allowing reinnervation of newly formed myofibers and providing sustained recovery of sphincter function. Cell therapy using myoblasts is technically feasible because these adult stem cells can be obtained from muscle biopsies, as already performed in early human clinical trials in an attempt to treat muscular dystrophies (10,14,33) or later for heart failure (20,21).

Regarding sphincter deficiency, the interest of developing cell therapy in FI is supported by the results of myoblast injections in treating urinary incontinence. Indeed, there is now significant experimental evidence that myoblast injection may restore damaged urinary sphincter function in rats (4,36), pigs (24), and dogs (7), which has led to the development of clinical trials (3,6, 23,29).

To establish the proof of principle of myoblast therapy in FI, we developed an animal model in which sphincteric function could be quantitated in vivo and followed up longitudinally. We then investigated whether intrasphincteric injection of myoblasts prepared using a protocol mimicking a clinical-grade process could durably improve sphincteric function.

Materials and Methods

Animals

All experiments were conducted using normal 8- to 12-week-old female Fischer rats (Charles River Laboratories, L'arbresle, France) in compliance with animal care and welfare regulations. This protocol was approved by an institutional ethics committee (CENOMEXA N/03–10–09/19/10–12).

Preparation of Myoblasts

Cell preparation was adapted from a human clinicalgrade myoblast preparation protocol previously used in clinical trials (29). Briefly, muscle was obtained from gastrocnemius of syngeneic rats, minced, and digested. Then, isolated cells were cultured in a selection medium (Celogos, Paris, France) for approximately 7 days. At this stage, a step of transduction with a green fluorescent protein (GFP)-encoding lentiviral vector [self-inactivating phosphoglycerate kinase-enhanced GFP Woodchuck hepatitis virus vector (pSIN.PGK.EGFP.WHV), kind gift of D. Klatzmann, Pitié-Salpêtrière Hospital/INSERM, Paris, France] was added to the protocol. Briefly, at day 1, myoblasts were seeded at 100,000 cells/well into six-well plates (BD Falcon, Franklin Lakes, NJ, USA) in expansion medium (Celogos). The next day (day 0), the transduction medium was prepared as expansion medium supplemented with protamine sulfate at 2 μl/ml (Sanofi Aventis, Madrid, Spain) and GFP-encoding lentiviral vector suspension (multiplicity of infection, MOI = 20). After addition of 2 ml/well of freshly prepared transduction medium, plates were centrifuged at 450 × g for 2 h at 32°C and further incubated for 2 h 30 min at 37°C. After medium removal, 2 ml of transduction medium was added anew and incubated at 37°C overnight. Using this procedure, transduction efficiency was 52% to 88%. Myoblasts were then amplified for two rounds in an expansion medium (2 weeks), harvested, and resuspended in phosphate-buffered saline (PBS; Life Technologies, Saint-Aubin, France) at a concentration of 3 × 10⁶ to 3 × 10⁸ cells/ml depending on the cell dose to be injected.

Flow Cytometry

Transduced myoblasts were stained with a rabbit anticluster of differentiation 56 (CD56) antibody (AB5032, Millipore, Molsheim, France) at 1/200. After washing, cells were incubated with 1/100 secondary phycoerythrin-labeled donkey anti-rabbit antibody (12–4739, eBioscience, Paris, France). Cells were analyzed on a FC500 flow cytometer (Beckman Coulter, Villepinte, France). The mean frequency of CD56-positive cells was 88%.

RT-PCR Analysis

Total RNA was extracted from myoblasts with TRIzol^® (Life Technologies). cDNA was prepared by reverse transcription of 1 μg of total RNA. Primers used were desmin, 5′-ACC TGC GAG ATT GAT GCT CT-3′ F and 5′-AGG CCA TCT TCA CAT TGA GC-3′ R (199 bp); myogenic factor 5 (Myf5), 5′-AAG CTT TCG AGA CGC TCA AG-3′ F and 5′-CCA TCA GAG CAG TTG GAG GT-3′ R (196 bp); myogenic differentiation (myoD), 5′-TAC CCA AGG TGG AGA TCC TG-3′ F and 5′-CAT CAT GCC ATC AGA GCA GT-3′ R (200 bp); glyceraldehyde 3-phosphate dehydrogenase (GAPDH), 5′-AGA ACA TCA TCC CTG CAT CC-3′ F and 5′-GTC CTC AGT GTA GCC CAG GA-3′ R (227 bp); GFP, 5′-AGA ACG GCA TCA AGG TGA AC-3′ F and 5′-GAA CTC CAG CAG GAC CAT GT-3′ R (197 bp). PCR was performed using 35 cycles (94°C–60°C–72°C) on a PCR apparatus (Eppendorf, Hamburg, Germany). GAPDH was used as positive control.

Anal Sphincter Cryoinjury and Myoblast Injection

Cryoinjury of the anal sphincter was performed using a 3-mm-diameter aluminum rod chilled with liquid nitrogen (single application for 30 s). After testing several conditions, two consecutive applications at a 24-h interval on a single 90° sector were chosen for the study. After the second cryoapplication, a quantity of myoblasts, varying from 3 × 10⁵ to 3 × 10⁷ cells, suspended in PBS, was injected into the anal sphincter using a 25-gauge syringe (BD, Columbus, OH, USA) under binocular microscopic guidance (Leica Microsystems, Wetzlar, Germany). Depending on the experiment, the cell dose was administered as three individual injections of 1 × 10⁵ (n = 6 rats), 1 × 10⁶ (n = 8), or 1 × 10⁷ (n = 6) myoblasts, each at equidistant sites within the rat anal sphincter, two of which at the borders of the cryoinjured site (total volume injected 100 μl). Alternatively, a single dose of 1 × 10⁶ (n = 6) myoblasts was injected in a unique site, within or opposite to the lesion. PBS alone was used as a control.

Histological Analysis

Animals were euthanized by injection of a lethal dose of ketamine (Imalgene; Merial, Lyon, France) and xylazine (Rompun; Bayer Health Care, Puteaux, France). The anorectal region was dissected immediately, snap frozen in liquid nitrogen, cut into 8-μm sections, and processed for histological analysis. Staining for morphological assessment was by standard hematoxylin (MHS32; Sigma-Aldrich, Saint-Quentin Fallavier, France) and eosin (861006; Sigma-Aldrich) method. Alternatively, frozen sections were preincubated with 1% fetal calf serum (Life Technologies) to block nonspecific binding and stained with one of the following primary antibodies: anti-desmin (1/100, D33; DakoCytomation, Les Ulis, France), anti-dystrophin (1/20, VP-D505; Vector Laboratories, Peterborough, UK), anti-GFP (1/200, sc-9996; Santa-Cruz, Heidelberg, Germany), anti-neurofilament (1/100, 2F11; DakoCytomation), antimyosin heavy chain fast (1/20, AbC10-M072; Eurobio, Montpellier, France), anti-myosin heavy chain slow (1/20, AbC10-M074; Eurobio), or anti-myosin heavy chain embryonic (1/20, F1.652; DSHB, Iowa City, IA, USA) before visualizing with an appropriate Alexa 647-labeled secondary antibody (1/2,000); or with α-bungarotoxin conjugates with Alexa 680 (1/400, B35452; Life Technologies), and counterstaining with Hoechst 33258 (1/50,000, 861405; Sigma-Aldrich). Control for staining specificity was omission of the primary antibody and was always negative.

Sphincter Functional Evaluation

Animals were anesthetized intraperitoneally with ketamine (Imalgene^® 1000; 100 mg/kg) and xylazine (Rompun^® 2%; 10 mg/kg). Anal sphincter pressure was monitored by anorectal manometry under local electrostimulation. The probe was inserted into the anus and perfused at a constant flow (30 ml/h). Sphincter contraction was elicited by electrical stimulation at different voltages. The signal from the probe was amplified and recorded using the software Spike 2 version 6.0 (Cambridge Electronic Design, Cambridge, UK). The reference range of manometric pressures (6.35 V, 5 s) was defined as the interval between the 10th and the 90th percentiles of the distribution obtained with normal rats (n = 11).

Statistical Analysis

Manometry data were expressed as median ± 25th percentile. The Mann–Whitney test was used for comparing two samples and the Kruskal–Wallis test for comparing more than two samples. A value of p < 0.05 was considered to be statistically significant.

Results

Model of Fecal Incontinence

Sphincteric function was evaluated in vivo by water perfusion anorectal manometry under local noninvasive electrostimulation, which induced a voltage-dependent increase in contraction force (Fig. 1A, B). FI was obtained by performing local cryoinjury on anesthetized Fischer rats receiving analgesics and antibiotics. After testing different lesion types, the selected injury was a duplicate cryoapplication on a 90° sector, which led to significant and sustained sphincteric deficiency (Fig. 1C). This procedure provoked sectorized sphincter destruction that was accompanied by transient cellular infiltration (Fig. 2). While inflammation disappeared after a few days together with reepithelialization, the striated sphincter rupture was durable (Fig. 2).

Figure 1.

Model of fecal incontinence. Standard recording of anal sphincteric contractility in control rats. (A) Example of anal sphincteric pressure curve under 5-V electrical pulses in control anesthetized animals. (B) Anal sphincteric pressure at different voltages. The values shown are area under curve of peak pressures. Data are expressed as median ± 25th percentile (n = 11 rats). Data show a voltage-dependent increase in contraction force whose maximal effect was observed with a 5-s pulse at 6.35 V. This condition was thus chosen as the standard signal for subsequent analyses. (C) Validation of the model of fecal incontinence. Topography of the different sphincter lesions evaluated (left). The condition selected as a model of fecal incontinence is boxed. Anal sphincter manometric analysis at day 30 after different types of cryoinjury (right). Data are expressed as median ± 25th percentile (n = 3–5).

Figure 2.

Model of fecal incontinence: histological analysis of anal sphincter cross sections. Hematoxylin–eosin staining of sphincter from normal (left) or untreated rats at day 1 (D1) (middle) and D30 (right) after 90° 2× cryoinjury. Location of the external (EAS) and internal anal sphincter (IAS) and lumen (L) is depicted (upper left). Uninjured sphincter is outlined by dotted ellipse (upper middle and upper right).

Characterization of GFP-Transduced Rat Myoblasts

Originating from muscle satellite cells, myoblasts can be expanded in vitro from a biopsy (32). Here, rat myoblasts were prepared by adapting a clinical-grade protocol that had been previously used to prepare human myoblasts for treating urinary incontinence in a human therapeutic trial (29). Cell culture included an additional step of lentiviral transduction intended to permit follow-up of donor-type engraftment (Fig. 3A, B). The expanded cell population showed typical features of myoblasts, including gene transcription of myogenic factors such as myoD and myf5 (Fig. 3B) and CD56 and desmin expression (Fig. 3C, D).

Figure 3.

Characterization of GFP-transduced rat myoblasts. (A) Flow cytometrical analysis of GFP expression. Percentage indicates the frequency of GFP-transduced cells. (B) RT-PCR analysis of desmin and myogenic factor gene expression. Myf5, myogenic factor 5; myoD, myogenic differentiation 1; GAPDH, glyceraldehyde 3-phosphate dehydrogenase. (C) Flow cytometric analysis of surface expression of cluster of differentiation 56 (CD56; negative control is overlaid, dotted line). (D) Fluorescence analysis after anti-desmin staining (red). GFP fluorescence is displayed in green. Hoechst nuclear staining in blue in overlay image.

Intrasphincteric Myoblast Injection Restores Anal Function

Syngeneic GFP-transduced myoblasts (1 × 10⁶; three injections) were injected into the incontinent anal sphincter at three equidistant sites (Fig. 4A). Functional evaluation was performed at different time points from day 0 (graft) to day 60. The control group received PBS instead of myoblasts. In the grafted group, sphincteric pressures significantly increased compared to controls at day 30 (p < 0.001) and were not statistically different from those of normal rats at day 60 (Fig. 4A). Histological analysis demonstrated reconstitution of the external anal sphincter at day 60 by regeneration of striated muscle in the myoblast-treated group, whereas a persistent striated sphincter defect was observed in the control group receiving PBS (Fig. 4B). Fluorescence microscopy analysis revealed that the transferred GFP-expressing myoblasts had differentiated in situ into mature myofibers expressing GFP (Fig. 4C), dystrophin (Fig. 4D), and fast myosin (Fig. 4E). Consistent with this late time point after muscle regeneration, embryonic and slow myosin was negative (not shown). GFP⁺ fibers represented approximately 20% of the sphincter surface on transversal sections, and no GFP⁺ was observed outside the injured area at day 60 despite three equidistant injections (data not shown). Interestingly, neurofilaments contacting GFP⁺ fibers were observed (Fig. 4F), and acetylcholine receptors binding α-bungarotoxin were expressed on GFP⁺ fibers (Fig. 4G), suggesting that neuromuscular endplates had formed on contact with newly differentiated myofibers. Functional recovery was maintained over a period of at least 6 months (Fig. 5A). There was no significant benefit in augmenting the delivered cell dose to 10 × 10⁶ myoblasts per injection (Fig. 5B). A lower dose of 0.1 × 10⁶ myoblasts per injection was not effective (difference not significant compared to PBS-injected rats, Kruskal–Wallis test) although GFP⁺ fibers were present (Fig. 5C). There were no clinically detectable side effects or extrasphincteric dissemination of grafted cells as attested by PCR examination of multiple organs (data not shown). Hence, syngeneic myoblasts allow durable recovery of sphincter function after local injection.

Figure 4.

Intrasphincteric myoblast injection restores anal function in incontinent rats. (A) Topography of the sites of injection (upper). 1 × 10⁶ GFP-transduced syngeneic myoblasts (squares, n = 8) were injected in the cryoinjured anal sphincter at three different sites, two of which were located at the lesion borders. This condition is referred to as 1 × 10⁶ (three injections). Controls received PBS in place of myoblasts (rhombuses, n = 8). Pressures under electrostimulation (6.35 V, 5 s) were measured by water perfusion manometry (lower). Data are presented as median ± 25th percentile. Reference range (shaded area) is 10th to 90th percentile of the distribution of sphincteric pressures from normal rats (n = 11). ***p < 0.001. (B) Hematoxylin–eosin staining of sphincter from 1 × 10⁶ (three injections) myoblast-treated (left) or PBS-control rats (right). Fluorescence analysis of anal sphincter from 1 × 10⁶ (three injections) myoblast-treated rats at day 60 after injection: (C) anti-GFP staining (green), (D) anti-dystrophin (red) and GFP (green), (E) anti-fast myosin heavy chain (red) and GFP (green), (F) anti-neurofilament (red) and GFP (green), (G) α-bungarotoxin (red) and GFP (green). Nuclei are counterstained with Hoescht 33342 (blue).

Figure 5.

Efficacy of myoblast therapy for fecal incontinence. (A) Long-term functional restoration after intrasphincteric injection of 1 × 10⁶ (three injections) myoblasts (n = 5). (B) Analysis at day 60 after intrasphincteric injection of different cell doses from 0.1 × 10⁶ (three injections) to 10 × 10⁶ (three injections) myoblasts (n = 6–8 per group). (C) Fluorescence analysis of anal sphincter at day 60 after injection of 0.1 × 10⁶ (three injections) (left) or 10 × 10⁶ (three injections) (right) myoblasts: anti-dystrophin (red) and GFP (green). Nuclei are counterstained with Hoescht 33342 (blue).

We also evaluated the effect of a single injection of myoblasts within or opposite to the lesion. Myoblasts injected opposite to the site of cryoinjury were unable to restore anal sphincter pressures (Fig. 6). In contrast, a single intralesion injection of 1 × 10⁶ myoblasts yielded the same functional result (Fig. 6) as when injecting 1 × 10⁶ myoblasts at three equidistant sites, two of which were at the borders of lesion site (sphincter pressure was not statistically different at day 60 between the group receiving three injections of 1 × 10⁶ cells and the group with single injection of 1 × 10⁶ cells within the lesion, Mann–Whitney test).

Figure 6.

Influence of site of injection on efficacy of myoblast cell therapy. Topography of the injection site within or opposite to the cryolesion (upper). 1 × 10⁶ myoblasts were injected within (black circles, n = 6) or opposite to the lesion site (white circles, n = 6). Pressures under electrostimulation (6.35 V, 5 s) were measured at day 60. Data are expressed as median ± 25th percentile.

Discussion

Together, these results provide proof of principle for myoblast therapy as a candidate treatment in FI. Cryoapplication provoked a sphincter lesion with fast epithelial reconstitution and without durable inflammation. We observed spontaneous functional amelioration at day 30 after a single 90° or 180° sphincter lesion (Fig. 1C). Hence, to maintain sustained sphincter deficiency, we preferred a double 90° cryoapplication (90° 2×), which was thereafter chosen as a model of FI. This enabled achievement of histologically proven destruction of one sector of the anal sphincter, provoking a significant decrease in contractile force. We preferred this approach to sphincterotomy (1,12,19,35), since it is not only more convenient to perform but also less traumatic for the rat.

In the present study, the capacity of myoblasts to restore anal sphincter function was evaluated by anorectal manometry as it represents a reliable method for providing robust measurement of anal pressures in humans (28). A first attempt by Kang et al. to evaluate the efficacy of myoblasts in rodents failed to show a statistically significant difference in sphincter contractility after treatment; however, the method used to determine function was ex vivo analysis of explanted sphincters after chemical stimulation (13). Anorectal manometry was preferred because it is non-traumatic and can be repeated over time. Hence, among the few experimental studies of myoblast anal sphincter injection, this is the first report of a longitudinal functional study in individual animals (12,13,35). To achieve myoblast tracking, injected cells were transduced with a GFP-encoding lentiviral vector. This method allowed long-term follow-up of donor-type myoblasts and their progeny without risk of dye dilution such as with PKH-26 used in other studies (1,12,13).

We show herein that myoblast cell therapy allows functional sphincter regeneration after 2 months and permits long-term recovery of sphincter contractility. This is secondary to donor-type tissue reconstitution, consistent with the encouraging results obtained in rabbits after grafting myoblasts prepared using a fiber explant protocol (12). After staining dystrophin as a marker of muscle fiber, double positive GFP⁺ dystrophin⁺ fibers were predominant at the lesion site, indicating that muscle regeneration occurred by formation of new muscle fibers from implanted myoblasts, even if we cannot exclude the formation of some hybrid fibers with recipient cells. Indeed, the added myoblasts may participate in fusing with host satellite cells and/or with host myotubes or integrating the host remnants of muscle fibers. Consequently, the presence of GFP in one fiber indicates that one donor cell, at least, is participating in muscle regeneration but does not guarantee that the whole muscle fiber originates only from the injected cells. Hence, fibers expressing GFP may not be completely new ones of exogenous origin, and the frequency of GFP-positive fibers on a sphincter section may overestimate the extent of extrinsic regeneration.

Also, the present results suggest that motor endplates can contact donor-type newly formed anal sphincter fibers, as already shown for the urinary sphincter (15,17,36) and that treated sphincter recovery may thus remain under physiological voluntary control. Yet, the presence of neurofilaments and acetylcholine receptor on muscle fibers, 2 months after cryoinjury, is not a complete demonstration that new fibers are innervated de novo. Indeed, the sites of innervations remain in place after degeneration, leaving time for colonization of the empty tube delineated by its basement membranes. The presence of these neuromuscular junctions should rather be considered as an indication of functionality than a definitive proof of an establishment de novo.

Three injections of 1 × 10⁶ myoblasts each, two of which were performed at the borders of the injured area, led to similar recovery as a single administration directly into the lesion. In contrast, a single injection opposite to the lesion was ineffective. We did not find GFP⁺ fibers outside the lesion site at day 60 in rats having received three injections, suggesting that their survival and differentiation were restricted to the injured area. This indicates that myoblasts have a limited capacity to migrate within the sphincter, as already seen in other striated muscles (27). Also, despite their capacity to secrete exosomes (16), the paracrine effect of myoblasts in our conditions remains limited in terms of distance of action. Mesenchymal stem cells may exert paracrine effects but poorly differentiate into muscle (8). To our knowledge, nothing has been published on the paracrine effect of AC133 (CD133) cells or mesoangioblasts. Regarding induced pluripotent stem cells and embryonic stem cells, it remains very difficult to achieve myogenic differentiation without MyoD or paired box 7 (Pax7) gene transduction, which would render these cells unsuited for clinical use in this indication at this stage. Side population cells promote muscle regeneration most probably by a paracrine effect, since they do not differentiate into the myogenic lineage (25). However, their limited number hampers their use in autologous cell therapy. Hence, we consider that performing several equidistant injections of myoblasts into the sphincter represents the best treatment option in clinical conditions to date.

Regarding tolerance of the treatment, we observed no evident side effects, and no rat developed tumors, including the group followed for 6 months. These myoblasts were produced following a process that mimicked one used in clinical trials in humans that has been extensively tested for safety (2). Absence of lentivirus-transduced cells in extrasphincteric sites, as attested by PCR using GFP and vector long terminal repeat-specific sequences, confirmed that the injected myoblasts lacked the capacity to disseminate.

Here, myoblasts were injected rapidly after performing the lesion, which represents a limitation of the model, since human patients would be treated for chronic rather than acute incontinence. A suitable model for chronic incontinence could be to induce severe sphincter degeneration, then to wait until possible spontaneous regeneration has stabilized, and then to inject the myogenic cells. Yet, as urinary incontinence has now moved to clinical evaluation (11,23,29), a first open study in chronic FI has recently shown good tolerance and suggested possible clinical benefit (9). Based on these clinical data and on the present proof of concept in an animal model, we are now evaluating the clinical efficacy of this mini-invasive therapeutic strategy in a randomized placebo-controlled clinical trial in patients with refractory FI (ClinicalTrials.gov NCT01523522).