Induction of Anoikis Following Myoblast Transplantation into SCID Mouse Muscles Requires the Bit1 and FADD Pathways

Authors


* Corresponding author: Jacques P. Tremblay, Jacques-P.Tremblay@crchul.ulaval.ca

Abstract

Seventy-five percent of the myoblasts transplanted in the mouse muscle die during the first 4 days following transplantation. The purpose of this study was to determine if anoikis plays a role in this phenomenon. Survival and proliferation of myoblasts in vitro were determined by Hoescht-PI labeling and cell counts respectively. In vivo cell survival and proliferation were quantified by injecting human male myoblasts labeled with 14C-thymidine in SCID mouse muscles. Survival and proliferation of the transplanted myoblasts were evaluated by scintigraphy and quantitative PCR of human Y chromosomal DNA. Inclusion of the extracellular matrix protein fibronectin enhanced transplanted myoblast survival by 1.7-fold while vitronectin improved their proliferation by 1.8-fold. Reductions in FADD and Bit1 expression reduced anoikis in vitro and improved the injected myoblast survival in vivo. Ectopic expression of the anti-apoptotic protein Bcl-2 completely abolished myoblast anoikis in vitro and enhanced cell survival by 3.1-fold in vivo. Cell death following transplantation appears to me mediated in part by anoikis. Inclusion of extracellular matrix proteins enhanced both survival and proliferation. Reduced expression of the proapoptotic proteins Bit1 and FADD or overexpression of Bcl-2 improved myoblast survival.

Introduction

Transplantation of muscle precursor cells (MPC) or myoblasts is considered as a potential approach to repair damaged skeletal muscles in a number of human muscle and myocardial pathologies (1–3). However, the utility of myoblast transplantation therapies is limited by the death of the majority of the cells within 4 days following their injection (4,5).

One possible reason for the massive loss of cells following transplantation is that the injected MPCs form pockets that reduce their contact with extracellular matrix components which in turn leads to an apoptotic death known as anoikis. Anoikis is caused by the detachment of the cells from their substrate or by their inability to adhere to a matrix (6). Blockade of adhesion molecules such as integrins and cadherins can rapidly induce apoptosis in cultured cells and the loss of cell-cell contacts results in anoikis in vivo in many different cell types (7–11). The survival signals induced by cell-substrate interactions are mediated by the anchorage of the cells to the extracellular matrix (ECM) via integrins, which stimulate the autophoshorylation of the focal adhesion kinase FAK (12). This in turn results in the activation of PI3K and the Akt (13) and inhibition of apoptosis by repressing BAD, Bid and caspase 9 (14). The activation of PI3K also induces the overexpression of FLIP, an inhibitor of Fas-associated death domain (FADD) protein and reduces death receptor CD95 (Fas)-induced apoptosis. Cell-cell adhesion occurs via cadherins and involves proteins from the catenin family, which form complexes with the actin cytoskeleton (15). Loss of cell-cell and cell-substrate interactions triggers cell death by recruiting apoptotic factors and/or mitochondrial proteins. In fact, the dephosphorylation of FAK and the repression of Akt activity (6) results in reduced expression of Bcl-2. BAD, Bid and BAX are translocated from the cytoplasm to the mitochondrion leading to cytochrome c release, caspase activation and FLIP repression (16). Anoikis is also enhanced via the P53-dependant release of Fas-L. This signaling molecule binds to and activates the CD95 death receptor, which then promotes caspase-8 activation and apoptosis signal amplification (17). A recent publication also suggests that a mitochondrial protein Bit1 is released into the cytosol following loss of α5β1 and αvβ3 integrins. Cytoplasmic Bit1 forms a complex with the transcriptional co-regulator AES (amino-terminal enhancer of split) and induces anoikis (18).

In present study we have tested the hypotheses that activation of anoikis is responsible for the loss of transplanted myoblasts. Skeletal myoblasts express αVβ3, αVβ5 and α5β1 integrins, which bind to the extracellular matrix (ECM) proteins (19–21), and participate in the regulation of skeletal muscle differentiation (22). αvβ3 integrin was classified as vitronectin receptor but it is known to bind to fibronectin as well (23). α5β1 and αVβ5 integrins have been reported to be major receptors for fibronectin (24). We tested the ability of ectopic fibronectin and vitronectin to enhance myoblast survival. In parallel series of experiments, we evaluated the roles of the anoikis signaling molecules Bit1, FADD and Bcl-2 both in vitro and in vivo. Our data strongly support the hypothesis that anoikis mediates the death of human myoblasts under conditions where cell-substrate adhesion is restricted and suggest that treatments designed to inhibit anoikis will enhance the utility of myoblast transplantation therapies.

Material and Methods

This work was authorized and supervised by Laval University Animal Care Committee. Manipulations were conducted according to guidelines set by the Canadian Council of Animal Care.

Reagents

Fetal bovine serum (FBS), Dulbecco's modified Eagle medium (DMEM), OptiMem medium, Trizol and penicillin/streptomycin were obtained from Invitrogen (Burlington, Canada). Hank's balanced salt solution (HBSS), vitronectin, fibronectin and collagenase were purchased from Sigma Aldrich (St. Louis, MD). Agarose was obtained from JT Baker (Phillipsburg, NJ). Rabbit anti-human FADD and mouse anti-human Bcl-2 antibodies were purchased from NeoMarkers (Fremont, CA). Rabbit anti-human Bit1 was obtained from Prosci (Poway, CA). Alkaline phosphatase conjugated goat anti-rabbit antibody was obtained from Caltag (Burlington, CA). Peroxydase conjugated goat anti-mouse and goat anti-rabbit antibodies were purchased from Dako (Mississauga, Canada). T7 Ribomax Express Large Scale RNA Production System was obtained from Promega (Madison, WI). The LightCycler Fast Start DNA Master SYBRGreen was purchased from Roche (Indianapolis, IN). The siPORTAmine was purchased from Ambion (Austin, TX). The BCIP-NBT kit for Western blot detection was purchased from BD Biosciences Pharmingen (San Diego, CA). The chemo-luminescence reagent for Western blot detection and the [methyl-14C] -thymidine were purchased from PerkinElmer (Boston, MA). QIAamp DNA extraction minikit was obtained from QIAGEN (Mississauga, Canada).

Cell culture

Myoblasts used for in vitro and in vivo experiments were obtained after enzymatic digestion of human muscle biopsies with collagenase (2%) in HBSS for 45 minutes at 37.5°C. The cells were then cultured at 37°C in a humidified atmosphere with 5% CO2 in proliferation medium (DMEM-high glucose) supplemented with 20% FBS and 1% penicillin–streptomycin. In some experiments, myoblasts were cultured in serum free DMEM-HG in presence or absence of vitronectin (25 μg/mL) and/or fibronectin (50 μg/mL).

In vitro anoikis assay

To induce anoikis in order to perform in vitro cell death assay, the ability of myoblasts to adhere to the culture vessel walls was blocked using the method described by Schwartz et al. (25). Basically, T25 flasks were first coated with heat denatured BSA (10 mg/mL) in PBS for 5 minutes at room temperature followed by 2 mL of melted 2% agarose in PBS. Adhesion blockade of cells cultured under these conditions was determined visually using an optical microscope. The technique completely blocked human myoblast adhesion.

In vitro proliferation assay

In order to evaluate the effect of ectopic molecules or siRNA transfection on human myoblast proliferation, cells were plated in 6 well plates at low confluence under normal and nonadherent conditions. Cultured cells were then harvested at different time points and counted using Thomas hematocytometer.

In vitro cell death assay

This test was performed to quantify human myoblast survival in vitro. Cells viability was evaluated by using the Hoescht-Propidium iodide (PI) labelling technique as previously reported (26). Briefly, cells were harvested, washed with HBSS and resuspended in 50 μL of PI (20 μg/mL) for 30 min at 4°C in the dark. The suspension was fixed with 25% ethanol solution for 5 min and then 25 μL of a Hoechst 33342 solution (112 μg/mL) were added for 1 hour at 4°C in the dark. The labeling of the cells was determined with by fluorescence activated cell sorting (FACS).

siRNA production and transfection

Synthetic siRNA against human FADD and Bit1 were produced to transfect human myoblasts and to transiently inhibit FADD and Bit1 expression. siRNAs were created using the T7 Ribomax Express Large Scale RNA Production System according to the manufacturer's recommendations. The target sequence for Bit1 siRNA was CTTGGTTATGGAATATTTG (18) and the target sequence for FADD was AATGCGTTCTCCTTCTCTGTGCCTGTC (27). The target sequences in the inverted orientation were used to produce nonspecific siRNA controls which were CTGTCCGTGTCTCTTCCTCTTGCGTAA for FADD and GTTTATAAGGTATTGGTTC for Bit1.

FADD and Bit1 siRNA transfections were performed using OptiMem medium according to the manufacturer reverse transfection protocol. Briefly, adherent cells were trypsinized and diluted in DMEM-HG growth medium. Three million myoblasts were then cultured in 6 well plates at a density of 500,000 cell per well. RNA (10%) and siPort (3%) were diluted in OPTIMEM medium and incubated 10 minutes at room temperature. The mix was then added to the cells in suspension at 37°C until the time of assay.

Western blot analysis

FADD, Bit1 and Bcl-2 protein expression for each experiment was determined by Western blot. Myoblasts were harvested and resuspended in a lysis buffer. The suspension was sonicated and proteins (30 μg) were then fractionated in a 12% polyacrylamide SDS gel and then transferred onto a nitrocellulose membrane (Biorad, Canada), which was blocked in saline-Tween-Blotto (5%). The membrane was then incubated with a primary antibody (1:100 dilution of rabbit anti-human FADD, 1:500 dilution of rabbit anti-human Bit1 or 1:250 dilution of mouse anti-human FADD) for 90 minutes at room temperature. The blots were then incubated for one hour with their respective secondary antibody (a 1:2000 dilution of phosphatase alkaline-conjugated rabbit anti-mouse immunoglobulin G antibody for FADD and a 1:10,000 dilution of peroxydase-conjugated goat anti-rabbit antibody for Bit1 and Bcl-2). Blots were visualized using the BCIP/NBT revelation kit for FADD and using the chemoluminescence kit for Bcl-2 and Bit1. To confirm that equal amounts of protein were loaded into each well, gels were stained with Coomassie blue.

Bcl-2 expression construct, virus production and human myoblast infection

This experiment was performed to evaluate the effect of Bcl-2 overexpression on human myoblast survival in vitro under anoikis conditions and in vivo following their transplantation in SCID mouse TA muscles. The day before the transfection, 3 million PA317 cells were plated in 10 cm culture dish in DMEM-HG medium supplemented with 10 FBS. Cells were transfected with 15 μg of the a pBabe retroviral vector encoding human Bcl-2 and a puromycin-resistance gene using the PEI transfection reagent (Medicorp, Quebec) mixed with 0.15 M NaCl and incubated at 37°C overnight. PA317 transfection medium was then removed and a freshly prepared DMEM-HG medium containing 6 μg/μL of puromycin was added to the cells for 6 days. Puromycin-resistant cells were then harvested, plated in T75 flasks and incubated to reach 70% of confluency. The medium containing the Bcl-2 encoding retrovirus was then harvested, frozen in liquid nitrogen and stored at −80°C. Human myoblast infection was performed by adding the medium containing Bcl-2 retrovirus to cells for 24 hours. Two rounds of 4-day selection on human myoblasts were performed using 6 μg/μL of puromycin and the surviving cells then cultured for ten days.

Myoblast transplantation

Four-week-old SCID female mice were used as hosts for these experiments. Both TA of recipient mice were used for the implantation of 1 × 106 human male myoblasts in each muscle. The myoblasts were either transfected with the siRNAs or cultured in the presence or absence of fibronectin (50 μg/mL) and/or vitronectin (25 μg/mL). On the day of transplantation the cells were detached using 0.125% trypsin and washed in HBSS. Pellets were obtained by centrifugation at 6500 rpm for 5 min. The skin was opened, and the myoblast pellets, resuspended in 20 μL of HBSS, were slowly injected throughout the TA muscles. At various times, the animals were then killed, and the injected muscles were removed and subjected to DNA extraction.

DNA extraction from muscle tissue

To evaluate the proportion of remaining radioactivity and also to quantify the amount of human Y chromosome in each condition, DNA was extracted from transplanted Tibialis anterior (TA) muscles using the QIAamp DNA minikit according to the manufacturer protocol.

In vivo cell death assay

To quantify the in vivo cell mortality, myoblasts were radiolabeled by culturing them 24 hours in growth medium containing 0.25 μCi/ml [methyl-14C] thymidine (50 mCi/mmol). Radiolabeled cells (1 × 106) were injected into 12 sites of the TA muscle using a glass micro-pipette (Drummond Scientific Co., Broomall, PA). The right TA was injected with myoblasts resuspended in saline solution containing fibronectin and/or vitronectin while the left TA muscle was used as control (transplanted with myoblasts resuspended in saline solution). Muscles were then removed at the indicated time points, snap-frozen in liquid nitrogen, and stored at −80°C. The amount of radio-label within each TA muscle was measured on DNA extracts using liquid scintillation counter (Mod. Wallac 1409, Woodbridge, Ontario, Canada)

In vivo quantification of myoblast proliferation by real-time PCR

To quantify myoblast proliferation in vivo, we first created a human Y chromosome standard by transplanting 250.000, 500.000, 1.000.000 and 2.000.000 human male myoblasts in SCID mouse TA muscles (n = 4). Muscles where then harvested and DNA was extracted. 100 ng of total DNA extract were used to produce the human Y chromosome standard by qPCR using a Corbett Rotorgene thermocycles.

Human male myoblasts (1 × 106) either transfected with the siRNAs or cultured in the presence or absence of and/or vitronectin were then transplanted into SCID mouse TA muscle. Muscles were then removed at various time points and their DNA was extracted and 100 ng of each DNA extract was used to evaluate the amount of human Y chromosome. At the end of the qPCR, the Rotorgene software (version 6.0.1) estimated the number of human myoblasts by transposing the Y chromosome amplification in each extract to the standard curve.

Primers used to amplify human Y chromosome were: 5′-CACCTACTGTGCCAGACAATGTG-3′ and 5′-CCCATGCCATGTTTGTCATACT-3′. The PCR profile for the human Y chromosome repeated for 45 cycles was 45 sec at 94°C followed by 1 min at 60°C and 45 sec at 72°C.

Statistical analysis

Differences among groups were statistically analyzed using analysis of variance test. A value of p < 0.05 was considered significant. Data are represented as means ± SD.

Results

Human myoblasts survival and proliferation require both substrate adhesion and growth factors in vitro

As a first step in evaluating the potential role of anoikis in myoblast death, human primary myoblasts were cultured in flasks under normal (adherent) conditions and those under which adhesion was blocked (nonadherent) as shown in Figure 1A. The basal level of cell death measured in the control cultures was the same at time 0 and 72 hours later. In contrast, loss of substrate adhesion resulted in a 4-fold increase in cell death following 72 hours of culture. These data suggest that human myoblasts undergo anoikis when they are prevented from attaching to the substrate. It is also known that myoblasts are dependent on growth factors for survival in vitro (28,29). Removal of serum from the medium increased the cell death by 10-fold after 72 hours in vitro. To determine if the combined loss of serum and adhesion exacerbated the tendency of cells to undergo apoptosis, we removed both of these stimuli and examined cells 72 hours later. The combined loss of both serum and substrate lead to a profound loss of cells, with almost 50% the cells having died by the time of assay (Figure 1A). Under normal growth conditions used to culture these human myoblasts, we observed that the cells doubled approximately every 36 hours, so that by 72 hours, there was a 3.7-fold increase in the number of myoblasts (Figure 1B). Removal of serum from the cultures reduced proliferation but it was still 2.3-fold greater than at time 0. In contrast, loss of substrate adhesion blocked proliferation in either the presence or absence of serum. Therefore while serum was dispensable for proliferation, substrate contact is required.

Figure 1.

Anoikis in human primary myoblasts in vitro. Human myoblasts were cultured in either serum-free medium or medium supplemented with 20% FBS under adherent and nonadherent conditions and examined at time 0 (To, black bars) and 72 hours later (To + 72, grey bars). (A) The percentage of apoptotic cells was determined by FACScan analysis. (B) Myoblasts were stained by trypan blue dye and counted using a Thomas hemacytometer at the indicated time points (*, τ p < 0,05; mean ± SD, n = 5).

Fibronectin reduced in vitro anoikis of human myoblasts cultured under nonadhesion conditions but did not alter proliferation

Human myoblasts were cultured under adherent and nonadherent conditions in presence or absence of fibronectin (50 μg/mL) in serum-free medium. The addition of fibronectin to cells cultured in growth medium conferred a slight survival benefit (6.9%) after 72 hours of culture (Figure 2A). Fibronectin reduced apoptosis by 2.3-fold and 1.7-fold, respectively, in adhesion and nonadhesion conditions in comparison to the control without fibronectin. This figure also suggests that fibronectin reduced apoptosis due to serum deprivation. The provision of ectopic fibronectin slightly enhanced human myoblast proliferation in vitro, however the effect was not statistically significant (Figure 2B).

Figure 2.

The effects of co-treatment of human primary myoblasts from in vitro with fibronectin. Human myoblasts were cultured in serum-free medium under adherent and nonadherent conditions in the presence or absence of ectopic fibronectin (50 μg/mL). Cells were examined at time 0 (To, black bars) and 72 hours later (To + 72, grey bars). (A) The percentage of apoptotic cells was determined by FACScan analysis. (B) Myoblasts were stained by trypan blue dye and counted using a Thomas hemacytometer at the indicated time points (*p < 0,05; mean ± SD, n = 5).

Vitronectin enhanced human myoblast proliferation in vitro but did not reduce anoikis

We next tested the effects of vitronectin (25 μg/mL) on the survival and proliferation of human myoblasts cultured in serum-free medium under adherent and nonadherent conditions. The addition of vitronectin did not confer any discernable survival benefit to the myoblasts (Figure 3A). However, its presence did enhance proliferation about 1.8-fold, independent of cells ability to bind to the substrate (Figure 3B).

Figure 3.

The effects of treatment of human primary myoblasts fromin vitrowith vitronectin. Human myoblasts were cultured in serum-free medium under adherent and nonadherent conditions in the presence or absence of ectopic vitronectin (25 μg/mL). Cells were examined at time 0 (To, black bars) and 72 hours later (To + 72, grey bars). (A) The percentage of apoptotic cells was determined by FACScan analysis. (B) Myoblasts were stained by trypan blue dye and counted using a Thomas hemacytometer at the indicated time points (*p < 0.05; mean ± SD, n = 5).

Co-administration of vitronectin and fibronectin reduced anoikis and enhanced proliferation of human myoblast

The aim of this experiment was to determine if co-administration of vitronectin and fibronectin improved human myoblast survival and proliferation under nonadherent conditions better than either one alone. Myoblasts were co-cultured under adhesion and nonadhesion conditions in presence or absence of vitronectin (25 μg/mL) and/or fibronectin (50 μg/mL) in serum-free medium. Co-administration of these two ECM proteins had no effect on cell survival under conditions where the cells were adherent to the substrate (Figure 4A). However, there was a 1.7-fold reduction in apoptosis when the cells were plated in medium containing fibronectin and vitronectin. Indeed, under nonadherent conditions, 42% of the myoblasts were apoptotic in the absence of both vitronectin and fibronectin at 72 hours, whereas only 25% of the cells were apoptotic in presence of these ligands. The combination of fibronectin and vitronectin thus had an effect similar to that of fibronectin alone. The results presented in Figure 4B demonstrate that the co-administration of vitronectin and fibronectin improved enhanced proliferation 1.6-fold under nonadherent conditions relative to controls. As with anoikis, this effect was similar to the one observed with vitronectin alone.

Figure 4.

The effects of co-treatment of human primary myoblasts from in vitro with fibronectin and vitronectin. Human myoblasts were cultured in serum-free medium under adherent and nonadherent conditions in the presence or absence of ectopic vitronectin (25 μg/ml) and fibronectin (50 μg/mL). Cells were examined at time 0 (To, black bars) and 72 hours later (To + 72, grey bars). (A) The percentage of apoptotic cells was determined by FACScan analysis. (B) Myoblasts were stained by trypan blue dye and counted using a Thomas hemacytometer at the indicated time points (*p < 0.05; mean ± SD, n = 5).

The inclusion of fibronectin and/or vitronectin enhanced the survival and proliferation of human myoblasts transplanted into mouse

To determine if fibronectin and/or vitronectin can enhance the survival or proliferation of myoblasts in vivo, we injected 1 million human myoblasts that had been prelabeled with 14C-thymidine into TA muscle of SCID mice. The right TA was injected with myoblasts resuspended in fibronectin or/and vitronectin while the left TA muscle served as a control and only received myoblasts resuspended in saline solution. Muscles were then removed To, To+2 days and To+4 days, and subjected to DNA extraction. There was no statistically significant treatment effect at day two post-injection, but by day 4, the effects were clear. Administration of fibronectin enhanced the survival of transplanted myoblasts by 1.3-fold. The combination of fibronectin and vitronectin had an effect similar to that of fibronectin alone. Vitronectin had no effect on myoblast survival in vivo.

To evaluate the effect of vitronectin and/or fibronectin on human myoblast proliferation in vivo, and to quantify the number of transplanted cells or their progeny, we performed qPCR analysis for human Y chromosome sequences using the DNA extracted from the animals described above. The addition of vitronectin did not statistically alter the number of surviving cells although there was a trend in the direction of a benefit. To determine the proliferation index for the control and ECM-treated myoblasts, we took the ratio of surviving cells to the total number present in the tissue at 4 days (Figure 5C). Although fibronectin treatment had no discernable effect, treatment with vitronectin resulted in a 1.4-fold increase in the number of transplanted cells or their progeny within the host muscle. We have also showed that inclusion of ectopic fibronectin enhanced Bcl-2 expression. Vitronectin also enhanced Bcl-2 expression but less than fibronectin. Effect of combination of both fibronectin and vitronectin on Bcl-2 expression was slightly greater than the one observed with fibronectin alone. Moreover, fibronectin and vitronectin treatment did not change FADD and Bit1 expression (Figure 5D).

Figure 5.

The effects of vitronectin and/or fibronectin treatment on myoblast survival and proliferation in vivo. Human male primary myoblasts were labeled in vitro with 14C-thymidine and then injected into TA muscles of SCID mice in presence or absence of vitronectin and/or fibronectin. Transplanted muscles were then harvested and their DNA was extracted. (A) Radio-labeled DNA was quantified by liquid scintillation counting to determine the relative amount of transplanted myoblast DNA persisting in the mouse muscles. (B) The relative amount of human Y chromosome DNA was determined by qPCR. (C) Proliferation index was calculated by comparing the proportion of Y chromosome to the percentage of remaining radioactivity. (D) Western blots were performed to quantify FADD, Bit1 and Bcl-2 expression following cell treatments with ectopic fibronectin and/or vitronectin (V: vitronectin; F: fibronectin; V+F: vitronectin and fibronectin) (*, τ p < 0,05; mean ± SD, n = 4).

FADD and Bit-1 play a role in myoblast anoikisboth in vivo and in vitro

Because Bit1 and FADD have been shown to mediate anoikis in several different lineages, we designed a series of experiments to evaluate their roles in the death of human primary myoblasts. We transfected cells with siRNAs directed against FADD and Bit1, either alone or together, and monitored the effects on targeted protein expression via Western blot analysis. The siRNA treatment for FADD reduced the expression of that protein by 60% in one experiment (Figure 6A) and by 65% in second study (Figure 8A), while the siRNA for Bit1 reduced Bit1 protein expression by 70% in one experiment (Figure 7A) and by 75% in a second (Figure 8A).

Figure 6.

Effect of siRNA on the FADD expression and anoikis in human primary myoblasts. Human myoblasts were transfected with synthetic siRNA against FADD and analyzed for the repression of FADD expression and anoikis. (A) Western blot analysis was performed in order to quantify FADD inhibition within transfected cells. (B) Anoikis was quantified in control (siFADD-) and FADD siRNA-treated (siFADD+) myoblasts in vitro. Cells were grown under adherent and nonadherent conditions and then harvested after 72 hours. The proportions of living or dead cells were evaluated using the Hoescht-PI labeling technique (*p < 0.05; mean ± SD, n = 5).

Figure 8.

Effects of co-repression of FADD and Bit1 on anoikis in human primary myoblasts in vitro and in vivo. Human myoblasts were transfected with synthetic siRNAs against both FADD and Bit1 and analyzed for protein expression and anoikis. (A) Western blot analysis was performed in order to quantify the repression of both FADD and Bit1 in transfected cells. (B) Anoikis was quantified in control (siBit1/siFADD -) and FADD/Bit1 siRNA-treated (siBit1/FADD+) myoblasts in vitro. Cells were grown under adherent and nonadherent conditions and then harvested after 72 hours. The proportion of living or dead cells was evaluated using the Hoescht-PI labeling technique (*p < 0.05; mean ± SD, n = 5). (C) In vivo, the effect of the inhibition of Bit-1 and/or FADD on cell survival was evaluated by injecting 14C-thymidine labeled nontransfected or co-siRNA transfected human myoblasts into the TA muscles of SCID mice. Muscles were harvested at time of injection, 2 days and 5 days following the myoblast transplantation and the DNA isolated. The amount of radioactivity remaining within the muscle DNA was quantified. Data in (C) represent the percentage of surviving cells at 5 days posttransplantation (*, ω, ϕ p < 0,05; mean ± SD, n = 4).

Figure 7.

Effect of siRNA on the Bit1 expression and anoikis in human primary myoblasts. Human myoblasts were transfected with synthetic siRNA against FADD and analyzed for the repression of Bit1 expression and anoikis. (A) Western blot analysis was performed in order to quantify Bit1 inhibition within transfected cells. (B) Anoikis was quantified in control (siBit1 -) and Bit1 siRNA-treated (siBit1+) myoblasts in vitro. Cells were grown under adherent and nonadherent conditions and then harvested after 72 hours. The proportions of living or dead cells were evaluated using the Hoescht-PI labeling technique (*p < 0.05; mean ± SD, n = 5).

To determine physiological consequences of reducing the expression of FADD and Bit1, we repeated the various in vitro and in vivo anoikis and proliferation assays described above. Inhibition of FADD expression had no effect on myoblast survival under adherent growth conditions but reduced cell death 1.3-fold under nonadherent growth in vitro (Figure 6B). Reductions in Bit1 also had no effect on cell survival under adherent conditions but led to a 1.9-fold reduction in anoikis (Figure 7B). Similar results were obtained when FADD and Bit1 were co-repressed (Figure 8). Under nonadherent growth conditions, co-repression resulted in a 1.7-fold reduction in apoptosis, comparable to what was obtained with siRNA against Bit1 alone, suggesting that these treatments were not synergistic.

Repression of FADD and Bit1 expression alone, or in combination, were evaluated for their ability to enhance the survival of transplanted myoblasts. Cells were labeled with 14C-thymidine as before and injected into the TA muscles of host mice. At 4 days posttransplantation, cells treated with siRNAs against FADD and Bit1 displayed a 1.6-fold reduction in cell death (Figure 8C). Co-repression of FADD and Bit1 resulted in a 1.8-fold reduction in myoblast death (Figure 8C).

Bcl-2 overexpression reduced human myoblast death in vivo and in vitro

To provide additional support for the hypothesis that human primary myoblasts die by anoikis, we infected cells with a retroviral construct encoding a Bcl-2 construct that co-expresses the antibiotic puromycin. Cultures of stably infected cells displayed a 4-fold increase in Bcl-2 expression relative to control cells receiving the empty vector (Figure 9A). Cells expressing ectopic Bcl-2 failed to undergo anoikis when cultured under nonadherent conditions (Figure 9B). To determine if the survival benefit of ectopic Bcl-2 expression impacted the survival of engineered myoblasts, we injected these cells in to the TA muscles of mice and evaluated survival 72 hours later. Cells expressing ectopic Bcl-2 displayed a 3.1-fold increase in survival relative to control myoblasts carrying the empty vector (Figure 9C) by 4 days posttransplantation.

Figure 9.

Effect of ectopic Bcl-2 on anoikis in human primary myoblasts myoblasts both in vitro and in vivo. Myoblasts were infected with the SV40-Bcl-2-IRES-puromycin or the SV40-IRES- puromycin retroviral vectors and the puromycin-resistant cells were cultured for 10 days. (A) Proteins were extracted for Western blot analysis using an anti-Bcl-2 antibody. (B) Puromycin-resistant myoblasts were cultured in serum-free medium under adherent and nonadherent conditions to determine if ectopic Bcl-2 conferred protection from anoikis. Grey and black bars represent respectively the percentage of apoptotic myoblasts at time 0 (to) and 72 hours later (To + 72). (C) Myoblasts expressing ectopic Bcl-2 were labeled with 14C-thymidine, transplanted in the TA muscles of SCID mice and then harvested on day 0 and day 5 for scintillation counting. The percent persisting myoblast DNA was calculated (*p < 0.05; mean ± SD, n = 4).

Discussion

Myoblast transplantation therapy can be used to repair damaged tissues including skeletal and cardiac muscles (30,31). One of the primary problems is the observation that the majority of transplanted cells die shortly after transplantation. In this study we test the hypothesis that many of these transplanted cells die by anoikis. The transplantation process appears to predispose myoblasts to anoikis because they are introduced in vivo at a high concentration as a bolus that likely precludes the majority of cells from contacting the substrate. In the present study, we tested the hypothesis that human primary myoblasts undergo anoikis and that treatments designed to counter anoikis enhance the survival of these cells.

A number of extracellular matrix proteins and intracellular signaling molecules have been implicated in anoikis. We manipulated the expression of several key ECM and intracellular signaling molecules that have been strongly implicated in the control of anoikis. Our results demonstrate that human myoblasts are susceptible to anoikis in vitro. Myoblasts express several integrins including αVβ3, αVβ5 and α5β1 which can bind the ECM proteins fibronectin and vitronectin (19–21) which in turn have been shown to promote cell survival or proliferation in other cell lines (32). Inclusion of fibronectin enhanced the survival of myoblasts following detachment from the substrate and thus inhibited anoikis in vitro and in vivo. In contrast, vitronectin did not confer a survival benefit in either environment but did enhance proliferation both in culture and in animals. Activation of these integrins stimulates the autophoshorylation of the focal adhesion kinase (12), which in turn induces PI3K and Akt activation (13). Moreover, binding of αVβ3, αVβ5 and α5β1 integrins to extracellular matrix proteins promotes cell survival by repressing apoptotic proteins as BAD and caspase 9 (14).

Apoptosis is regulated by a series of signal transduction molecules that ultimately converge on the activation of initiator and executioner caspases (33,34). Many of the extracellular signals that regulate apoptosis signal through TNF-family death receptors such as Fas and TNFR. Both of these death receptors in turn utilize the Fas associated death domain protein FADD. Blockade of FADD expression or function is sufficient to block anoikis in several models and we have found that it plays a critical role in myoblast anoikis as well (17). Reductions in FADD expression greatly reduced myoblast death in vitro under nonadherent growth conditions and in vivo following transplantation. Our results also suggest that in vivo, FADD was implicated in apoptosis caused by other factor(s) than just the loss of cell to matrix anchorage. Indeed, our results demonstrated that in vitro myoblast apoptosis was reduced by 1.3-fold while in vivo, their survival was improved by 1.7-fold.

Bit1 was recently identified in a screen for anoikis-regulating proteins in CHO cells and functions downstream of fibronectin and integrin signaling (18). Detachment of cells leads to a downregulation of Akt signaling which results in the release of Bit1 from mitochondria. Bit1 then forms complex with the transcriptional co-regultor AES (amino-terminal enhancer of split) and induces caspase-independent-apoptosis. In our studies we have demonstrated that siRNA-dependent reductions in Bit1 expression protect myoblasts from anoikis both in vitro and in vivo.

The Bcl-2 protein has been shown to play a critical role in caspase activation and can block apoptosis induced by a wide variety of different cell types. In experimental systems, ectopic expression of Bcl-2 protects many cell types from apoptosis induced by such stimuli as: growth factor withdrawal, detachment, toxins, and DNA damage. It has been reported that Bcl-2 expression can be upregulated upon the activation of α5β1 and αVβ3 integrins in CHO cells (32), supporting a role in the control of anoikis. We have evaluated by Western blot the expression of Bcl-2, FADD and Bit1 to indicate how the fibronectin and/or vitronectin treatment changes the expression of these factors (Figure 5D). Our results suggest that the improvement of human myoblast survival following their treatment with these extracellular matrix proteins is due to an increased Bcl-2 expression that represses FADD and Bit1 proteins. We also infected human primary myoblasts with a Bcl-2 expressing retrovirus and evaluated its ability to protect cells from anoikis. We observed that a 4-fold increase in Bcl-2 expression was sufficient to completely abrogated anoikis in human myoblasts in vitro and reduced their death by 3.1-fold in vivo. These results suggest that apoptosis is implicated in the death of about 40% of transplanted human myoblasts and that anoikis is a key mediator of this effect death. In a previous report, we observed that less than 2% of injected myoblast were positive for the active caspase-3 during a 5-day follow-up (35). The apparent low abundance of apoptotic cells is likely due to the asynchrony of apoptotic process, the speed of the execution phase of apoptosis and the phagocytosis of apoptotic cells by macrophages and other cells. The active caspase-3 positive cells only include those cells that are actively dying at the time of the labeling. Our approach consisted in evaluating human myoblast apoptosis when a potent anti-apoptotic gene was overexpressed. Our results suggest that about 40% of dying cells are lost by apoptosis. The rest of the dying cells are presumably lost by necrosis or a Bcl-2 independent pathway.

Our data provide strong support for the hypothesis that anoikis plays a critical role in the loss of transplanted myoblasts and provide the first report of the role for ECM proteins in loss of these cells in vivo. In addition, they point to the regulatory roles of FADD and Bit1 in this process both in vitro and in vivo in human primary myoblasts. These results have significant implications for the use of myoblasts for therapeutic applications and suggest treatments that could be used to enhance the survival of transplanted cells. The use of RNAi against FADD or Bit1 could also provide a transient treatment that would allow cells to survive the initial period following injection and provide the cells more time to establish proper cell-substrate interactions. Ectopic Bcl-2, while a potent antiapoptotic manipulation, would unlikely to be of value in a clinical setting due to its ability to allow cells with DNA damage to evade apoptosis. Nevertheless, Bcl-2 provides insight into the maximal benefits that could be achieved by inhibiting anoikis.

Acknowledgments

We would like to thank Dr Roy PH for giving us the access to Corbett Rotorgen Q-PCR machine and Dr. Yanhui Hu for the pBabe Bcl-2 construct. This work was supported by A.F.M (Association Française contre les Myopathies).

Ancillary