MyoD expression restores defective myogenic differentiation of human mesoangioblasts from inclusion-body myositis muscle.

Abstract

Inflammatory myopathies (IM) are acquired diseases of skeletal muscle comprising dermatomyositis (DM), polymyositis (PM), and inclusion-body myositis (IBM). Immunosuppressive therapies, usually beneficial for DM and PM, are poorly effective in IBM. We report the isolation and characterization of mesoangioblasts, vessel-associated stem cells, from diagnostic muscle biopsies of IM. The number of cells isolated, proliferation rate and lifespan, markers expression, and ability to differentiate into smooth muscle do not differ among normal and IM mesoangioblasts. At variance with normal, DM and PM mesoangioblasts, cells isolated from IBM, fail to differentiate into skeletal myotubes. These data correlate with lack in connective tissue of IBM muscle of alkaline phosphatase (ALP)-positive cells, conversely dramatically increased in PM and DM. A myogenic inhibitory basic helix-loop-helix factor B3 is highly expressed in IBM mesoangioblasts. Indeed, silencing this gene or overexpressing MyoD rescues the myogenic defect of IBM mesoangioblasts, opening novel cell-based therapeutic strategies for this crippling disorder.

Full text

Page 1

MyoD expression restores defective myogenic
differentiation of human mesoangioblasts
from inclusion-body myositis muscle
Roberta Morosetti*†, Massimiliano Mirabella*‡§, Carla Gliubizzi*‡, Aldobrando Broccolini*, Luciana De Angelis¶,
Enrico Tagliafico?, Maurilio Sampaolesi**, Teresa Gidaro*, Manuela Papacci*, Enrica Roncaglia?, Sergio Rutella††,
Stefano Ferrari?, Pietro Attilio Tonali*‡, Enzo Ricci*‡, and Giulio Cossu**‡‡§§
*Department of Neurosciences and†Interdisciplinary Laboratory for Stem Cell Research and Cellular Therapy, Catholic University, Largo A. Gemelli 8, 00168
Rome, Italy;‡Fondazione Don Carlo Gnocchi, 00194 Rome, Italy;‡‡Institute of Cell Biology and Tissue Engineering, San Raffaele Biomedical Science Park,
00128 Rome, Italy;¶Department of Histology and Embriology, University ‘‘La Sapienza,’’ 00161 Rome, Italy;?Department of Biomedical Sciences, University
of Modena and Reggio Emilia, 41100 Modena, Italy; **Stem Cell Research Institute, San Raffaele Hospital, 20132 Milan, Italy;††Institute of Hematology,
Catholic University, 00168 Rome, Italy; and§§Department of Biology, University of Milan, 20133 Milan, Italy
Edited by Tullio Pozzan, University of Padua, Padua, Italy, and approved September 19, 2006 (received for review April 28, 2006)
Inflammatory myopathies (IM) are acquired diseases of skeletal
muscle comprising dermatomyositis (DM), polymyositis (PM), and
inclusion-body myositis (IBM). Immunosuppressive therapies, usu-
ally beneficial for DM and PM, are poorly effective in IBM. We
report the isolation and characterization of mesoangioblasts, ves-
sel-associated stem cells, from diagnostic muscle biopsies of IM.
The number of cells isolated, proliferation rate and lifespan,
markersexpression,andabilitytodifferentiateintosmoothmuscle
do not differ among normal and IM mesoangioblasts. At variance
with normal, DM and PM mesoangioblasts, cells isolated from IBM,
fail to differentiate into skeletal myotubes. These data correlate
with lack in connective tissue of IBM muscle of alkaline phospha-
tase (ALP)-positive cells, conversely dramatically increased in PM
and DM. A myogenic inhibitory basic helix–loop–helix factor B3 is
highly expressed in IBM mesoangioblasts. Indeed, silencing this
gene or overexpressing MyoD rescues the myogenic defect of IBM
mesoangioblasts, opening novel cell-based therapeutic strategies
for this crippling disorder.
T
group of acquired muscle diseases and encompass three major
forms: dermatomyositis (DM), polymyositis (PM), and inclusion-
body myositis (IBM) (1). Causes of DM, PM, and IBM are
unknown,butanautoimmunepathogenesisissupportedbymarked
up-regulationofcytokinesandadhesionmolecules,evidenceofaT
cell-mediated myocytotoxicity in PM and IBM and of a comple-
ment-mediated microangiopathy in DM (2). Current immunother-
apies are usually effective in DM and PM patients, whereas IBM,
the most frequent myopathy in elderly patients, responds poorly or
not at all to immunosuppressive therapies and its course steadily
progresses to severe disability. In IBM muscle, the presence of
degenerativefeatures,suchasvacuolatedfiberscontainingamyloid
and amyloid-related proteins (3), reflects a complex pathogenesis
involving misfolded and unfolded proteins and increased oxidative
stress in the context of a cellular ‘‘aged’’ milieu acting in concert
with chronic inflammation (4). Regeneration and repair of muscle
fibers are fundamental processes accounting for rebuilding muscle
integrity and gradual recovery of muscle strength in IM after
suppression of mononuclear cells infiltration. Satellite cell-
dependent regeneration occurs also in IBM muscle wherein mul-
tiple metabolic pathways normally involved in muscle development
are activated (5, 6). However, in IBM, despite the activation of
potentially repairing mechanisms, regeneration is inefficient.
Mesoangioblasts are vessel-associated stem cells, firstly isolated
from dorsal aorta of mouse embryos (7), able to differentiate into
avarietyofmesodermtissuesincludingskeletal,cardiacandsmooth
muscle (8, 9). When delivered intraarterially, mesoangioblasts
he idiopathic inflammatory myopathies (IM), characterized by
mononuclear cells infiltration of skeletal muscle, are the largest
restore to a significant extent muscle morphology and function in
a mouse model of muscular dystrophy (10).
Because mesoangioblasts express numerous receptors for in-
flammatory cytokines, we assumed that the human counterpart of
murine mesoangioblasts should be recruited in high numbers
during muscle inflammation.
Here,wedescribetheisolationandfunctionalcharacterizationof
pericyte-derived adult mesoangioblasts (herein simply called me-
soangioblasts) from diagnostic muscle biopsies of IM patients and
show that IBM mesoangioblasts fail to differentiate into skeletal
muscle. This differentiation block can be corrected in vitro by
transient expression of MyoD, making these cells potential attrac-
tive candidates for cellular therapy of this disabling disease.
Results
Mesoangioblasts Are Efficiently Isolated from IM Muscle Biopsies.
After 10–15 days of organ culture from biopsies of three normal
controls, three DM, three PM, and six IBM, we isolated a popu-
lation of cells morphologically different from satellite cells. Ap-
proximately 3–4 ? 104cells could be obtained from each biopsy.
From the first passage on, cells were characterized by a triangular,
adherent,refractiveshapeandbyafloating?looselyadherentround
component, particularly abundant in DM (Fig. 1A). Peculiar cell
morphology, phenotypic characteristics, and differentiation poten-
tialindicatedthatourcellswerehumanmesoangioblasts,asrecently
characterized (A. Dellavalle, M.S., R. Tonlorenzi, E.T., B. Sac-
chetti, L. Perani, B. G. Galvez, G. Messina, R.M., S. Li, G. Peretti,
J. S. Chamberlain, W. E. Wright, Y. Torrente, S.F., P. Bianco, and
G.C., unpublished data).
Cellswerekeptincultureupto25populationdoublings(PD)(25
forcontrolandDM,19forIBM,and20forPM)withaproliferation
rate comparable for DM, IBM, and PM and independent from
patients’age.Doublingtimefromallbiopsieswas33.5?2.38h(Fig.
1B). At both early and late passages, cells kept a diploid kariotype
(data not shown). There were no differences in the number of cells
isolatedfromfreshlydissectedorfresh-frozenmusclesatbothearly
Authorcontributions:R.M.andM.M.contributedequallytothiswork;R.M.,M.M.,andG.C.
designedresearch;R.M.,M.M.,C.G.,A.B.,L.D.A.,E.T.,M.S.,T.G.,M.P.,E.Roncaglia,andS.R.
performed research; R.M., M.M., A.B., E.T., S.F., P.A.T., E. Ricci, and G.C. analyzed data; and
R.M., M.M., and G.C. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS direct submission.
Abbreviations: ALP, alkaline phosphatase; IM, idiopathic inflammatory myopathies; IBM,
inclusion-body myositis; DM, dermatomyositis; PM, polymyositis; SDMC, satellite-derived
myogenic cells; mdx, mouse muscular dystrophy; bHLH, basic helix–loop–helix; BHLHB3,
bHLH domain containing class B3 transcription factor.
§To whom correspondence should be addressed. E-mail: mirabella@rm.unicatt.it.
© 2006 by The National Academy of Sciences of the USA
www.pnas.org?cgi?doi?10.1073?pnas.0603386103PNAS ?
November 7, 2006 ?
vol. 103 ?
no. 45 ?
16995–17000
NEUROSCIENCE

Page 2

and late passages, neither phenotypic characteristics of the 12 IM
studied were significantly different after 8 and 18 PD (Fig. 2A and
Fig. 7, which is published as supporting information on the PNAS
web site). Because we isolated on average of 3–4 ? 104cells from
a single biopsy, the estimated final number of cells after 25 PD is
50–120 ? 1010, and the real number that could be obtained before
the appearance of senescent cells in significant proportion is
between 10 and 20 ? 109cells. This number would be suitable for
intraarterial delivery to adult patients, based on a per kg compar-
ison with the mouse model used before (10).
Clonogenic Potential, Cell Cycle, and Phenotypic Characteristics Do
Not Differ Among IM Mesoangioblasts. We dissociated mesoangio-
blaststosinglecellsuspensionandclonedthembylimitingdilution:
clones appeared in 9.75 ? 3.9, 8.87 ? 3.1, and 10.5 ? 4.0 wells for
DM, PM, and IBM, respectively, all with the same double mor-
phologyoftheoriginalcells.Byreplatingtheclonesatclonaldensity
they were able to give rise to new clones.
The cell cycle distribution was similar for all mesoangioblasts of
12 IM in three separate experiments (each one conducted in
duplicate) (G0?G1, 65.5 ? 6.4%; S, 23.7 ? 4.3%; G2?M, 10.8 ?
2.9%) regardless of the IM type. The pattern of distribution was
significantly different (P ? 0.01) from that observed in control
proliferating satellite-derived myogenic cells (SDMC) (G0?G1,
87.58 ? 3.6%; S, 8.43 ? 3.0%; G2?M, 4 ? 1.8%) (Fig. 1C). Results
were always consistent throughout all experiments. Cells from all
IM were strongly positive for CD44 and CD13, positive for CD49b,
homogeneously negative for CD34, CD133, CD45 by FACS (Fig.
2A), consistently with what observed in normal human mesoan-
gioblasts (A. Dellavalle, M.S., R. Tonlorenzi, E.T., B. Sacchetti, L.
Perani, B. G. Galvez, G. Messina, R.M., S. Li, G. Peretti, J. S.
Chamberlain,W.E.Wright,Y.Torrente,S.F.,P.Bianco,andG.C.,
unpublished data). By immunocytochemistry and Western blot, all
of the cells were positive for vimentin, weakly positive for ?-SMA
and desmin, and did not express glial fibrillar acidic protein
(GFAP), nestin, ?III-tubulin, and MyoD (data not shown). Alka-
line phosphatase (ALP) staining was positive in all IM mesoan-
gioblasts, with the highest levels observed in PM and only a weak
labeling in IBM (Fig. 2B). Together, these markers identify human
adult mesoangioblasts as the in vitro progeny of pericytes.
The Ability to Differentiate into Smooth Muscle Cells (SMCs) and in
Osteoblasts Are Similar Among All IM Mesoangioblasts. Murine
mesoangioblastsdifferentiateintomatureSMCsuponTGF?treat-
ment (11, 12). Therefore, we exposed mesoangioblasts from all
patients to TGF?. Approximately 80% of cells from all biopsies
differentiated into strongly positive ?-SMA-positive SMC, with no
significant difference between the various IM (Fig. 8 which is
published as supporting information on the PNAS web site).
Similarly to murine mesoangioblasts (8), human cells responded
to BMP2 with a rather low percentage (?5%) differentiating into
strongly ALP-positive osteoblast-like cells expressing osteocalcin
andosteopontin(datanotshown).Incontrast,bothcontrolandIM
mesoangioblasts failed to differentiate into neurons or glia when
grown in neural stem cell differentiation media (data not shown).
Genome-Wide Gene Expression in IM Mesoangioblasts. Proliferating
mesoangioblasts from normal and IM muscle were analyzed for
gene expression by Affimetrix gene array. As expected, gene
expression profile was similar in all samples with only few genes
differentially expressed. Clustering results are shown in Fig. 2C.
Two main classes were defined: the first included mesoangioblasts
from normal muscle, whereas the second consisted of mesoangio-
blasts from DM and IBM. Interestingly, the clustering procedure
paired together DM and IBM replicates.
AsummaryoftheanalysisisshowninTable1,whichispublished
as supporting information on the PNAS web site. In particular,
mesoangioblasts from controls and IM (DM and IBM) did not
express myogenic factors such as MyoD, or Pax3, Pax7, MEF2C, or
MEF2D. As expected for mesoderm cells, mesoangioblasts did not
Fig.2.
profiling of IM mesoangioblasts. (A) More than 90% of cells from all samples
were strongly positive for CD44 and CD13 with high percentage of cells
CD49b-positive.Noneofthemarkerspositiveinmurinemesoangioblastswere
significantly expressed. Bars represent the mean ? SD of 36 samples from the
12 patients with IM (3 DM, 3 PM, and 6 IBM) (each performed in triplicate). (B)
IM mesoangioblasts in vitro are all ALP-positive. After simultaneous staining
inthesamecultureconditions,moreintenselylabeledcellscanbeobservedin
PM and to a lesser extent in DM, whereas IBM mesoangioblasts are only
weakly positive. (Scale bar: 20 ?m.) (C) Clustering results show two main
classes: mesoangioblasts from normal controls (lanes 5–8), mesoangioblasts
from DM (lanes 1 and 2), and IBM (lanes 3 and 4). Clustering procedure pairs
together DM and IBM replicates.
FACS,immunophenotyping,ALPhistochemistry,andgeneexpression
Fig. 1.
culture on, refractive triangular, adherent, and round loosely adherent?
floating cells were observed. (Scale bar: 40 ?m.) (B) Cell growth was assessed
after 24 h, 48 h, and 72 h. Viable cells were judged by trypan blue exclusion.
Results are expressed as absolute counts. Bars represent mean ? SD of tripli-
cate samples of one representative experiment of three. (C) Cell cycle distri-
butionofproliferatingnormalhumanSDMCandmesoangioblastsfromthree
controls, three DM, three PM, and six IBM (run in duplicate) after 24 h of
culture was assessed by propidium iodide and FACS. For each sample, per-
centage of cells in G0?G1, S, or G2?M phases of cell cycle is indicated. One
representative experiment of three is shown.
Cell morphology, growth curve, and cell cycle. (A) From the organ
16996 ?
www.pnas.org?cgi?doi?10.1073?pnas.0603386103Morosetti et al.

Page 3

express cytocheratins or neurofilaments; at variance with their
embryonic counterparts, human postnatal mesoangioblasts did not
express endothelial markers but rather pericyte markers such as
ALP, PDGF-receptor ?, or NG2 proteoglycan. Differently from
controls,IMmesoangioblastsexpressednumerousimmune-related
genes, such as IFN-induced proteins and, particularly in DM, high
levels of integrins, that greatly facilitate diffusion into muscle
interstitium. As recently shown, mesoangioblasts muscle homing
can be increased by exposure to stromal cell-derived factor 1?
(SDF-1?) and subsequent transient expression of ?-4 integrin or
L-selectin(13).Ofnote,increasedSDF-1?hasbeenobservedinIM
muscle (14). These data suggest a possible advantage of IM
mesoangioblasts over controls in migration through endothelium
and into muscle fibers that can be more effective during inflam-
mation and after intra-arterial delivery. Interestingly, only IBM
mesoangioblasts expressed high levels of genes known to inhibit
myogenesis such as TGF?-1, SFRP-2 (secreted frizzled-related
protein 2), and BHLHB3 (basic helix–loop–helix domain contain-
ing class B3 transcription factor) (15–17).
DM and PM Mesoangioblasts Efficiently Differentiate into Skeletal
Muscle Under a Variety of Conditions. Upon coculture with C2C12
myoblasts, approximately 10% of myosin-positive myotubes con-
tained nuclei expressing human lamin A?C, thus confirming
fusion of human and murine cells (Fig. 3A). To investigate
whether mesoangioblasts from normal or IM muscle were
capable of spontaneous myogenic differentiation in vitro, cells at
80% confluence were maintained in culture medium without
growth factors (18). Control, DM and PM mesoangioblasts fused
into multinucleated myosin-positive myotubes with a fusion
index (expressed as number of myonuclei?number of total
nuclei) of 0.15 ? 0.05 for three controls and 0.1 ? 0.015 for three
DM and three PM. Differences between controls and DM and
PM did not reach statistically significance (P ? 0.05). This
feature was never observed in mouse mesoangioblasts, which
differentiate only upon coculture (8, 9). Interestingly, DM, PM,
and control mesoangioblasts, cultured for 4 days in growth
medium conditioned by normal human SDMC, acquired homo-
geneous myoblast-like morphology and, after exposure to dif-
ferentiation medium for 7 days, fused into numerous multinu-
cleated myosin-positive myotubes (fusion index of 0.7 ? 0.1)
(Fig. 3 A and C) with a marked up-regulation of myogenin and
myosin by Western blot (Fig. 3C). This evidence indicates that
factor?s produced by satellite-derived cells is?are necessary for
the myogenic commitment of human mesoangioblasts.
IBM Mesoangioblasts Fail to Differentiate into Skeletal Muscle. In
none of the above mentioned culture conditions, IBM mesoangio-
blastswereabletodifferentiateintomultinucleatedmyosin-positive
myotubes. However, when previously exposed to SDMC-
conditioned growth medium, they became oriented with elongated
morphology displaying a strong desmin immunoreactivity. MyoD
proteinexpressionwasnotdetectableafterexposuretoconditioned
medium(Fig.3B).ByWesternblot,IBMcellsharvestedatthesame
timepointsofDMdidnotexpressmyosinormyogenin,butshowed
that activation of the p38 pathway, known to be involved in skeletal
myogenesis, was not disrupted (19, 20) (Fig. 3C). Of note, satellite
cells isolated from the same IBM biopsies (18) were able to
normally differentiate in myotubes (Fig. 3D).
Connective Tissue of IM Biopsies Contains ALP-Positive Cells Recruited
to Myogenic Fate.ResultsfrominvitrostudieshaveshownthatALP
is a marker of adult mesoangioblasts. By investigating the distribu-
tion of ALP-expressing cells, we found ALP reactivity only in small
arteries in normal muscle, whereas in all DM and PM biopsies a
very strong ALP staining was evident in perimysial and endomysial
connectivetissue(Fig.4A),asdescribed(21).Onthecontrary,IBM
muscle showed no ALP connective tissue staining, although blood
vessels were normally represented. We asked whether at least part
ofthecellsstronglyALP-positiveinconnectivetissueofDMorPM
muscle could be recruited to a myogenic fate by neighboring
regenerating fibers. Indeed both PM and DM showed numerous
stronglyALP-positiveroundcellsintheendomysiumwithaMyoD-
positivenucleus(Fig.4B).Inthesameareas,satellitecellsexpressed
Pax7, whereas ALP-positive cells in the interstitium were PAX7-
negative. Because ALP is not expressed in satellite cells and
myoblasts (22), the presence of cells expressing both ALP and
myogenic markers suggests recruitment of pericytes-derived cells
into the myogenic lineage. Neither ALP-positive nor double-
positive cells were detected in the IBM sections analyzed. Because,
atleastearlyinthedisease,bothMyoD-positivesatellitecellsaswell
as pericytes are not significantly reduced, the lack of MyoD-
Fig. 3.
myosin and human lamin A?C. Mesoangioblasts from DM efficiently fused with
C2C12 murine myoblasts into mature myosin-positive myotubes, as indicated by
the presence of nuclei expressing human lamin A?C. (Scale bar: 10 ?m.) (Middle
andBottom)DMmesoangioblastsexposed4daystoSDMC-conditionedgrowth
medium (CM) and subsequently cultured for 7 days in differentiation medium
(Dif.M)spontaneouslyfuseintodifferentiatedmyotubes.(Scalebar:40?m.)Each
experimentwasperformedinduplicateatleastthreetimes.(B)IBMmesoangio-
blasts do not constitutively express MyoD (RT-PCR) and when exposed 4 days to
CM and 7 days to Dif.M are negative for MyoD and myosin, but display a strong
desmin immunoreactivity. (Scale bar: 20 ?m.) (C) Mesoangioblasts from DM, but
not from IBM, differentiate into multinucleated myotubes (phase contrast).
(Scale bar: 40 ?m.) For Western blot analysis, cells were harvested at day 0, after
4daysinCMandafter3and7daysinDif.M.Amarkedup-regulationofmyogenin
and myosin is observed already at day 3 of differentiation for DM mesoangio-
blasts; no up-regulation of myosin and myogenin is visible in IBM cells. For both
DMandIBMmesoangioblasts,activationofp38(phospho-p38)isobserved.One
representativeexperimentofsixisshown.(D)SDMCfromprimarymusclecultures
from all six IBM biopsies normally differentiate in multinucleated myotubes
(phase contrast, Upper) and show strong myosin-immunoreactivity (Lower).
(Scalebar:20?m.)Tworepresentativeculturesareshown.NCM,nonconditioned
medium.
Skeletal muscle differentiation. (A) (Top) Immunofluorescence for
Morosetti et al.PNAS ?
November 7, 2006 ?
vol. 103 ?
no. 45 ?
16997
NEUROSCIENCE

Page 4

expressingpericytes-derivedcellsseemscorrelatedwiththespecific
defect in myogenic commitment of this cell population in IBM.
Transient Expression of MyoD Rescues the Myogenic Defect of IBM
Mesoangioblasts. IBM mesoangioblasts did not give rise to termi-
nally differentiated myotubes and did not constitutively express
MyoD mRNA, and MyoD could not be induced by our differen-
tiating conditions (Fig. 3 B and C); therefore, we transduced IBM
mesoangioblasts with an adenoviral vector expressing full-length
murine MyoD. Expression of MyoD caused myogenic differentia-
tion in ?60% (fusion index 0.6 ? 0.02) of cells, as defined by
activation of muscle-specific genes and myotubes formation
(Fig. 5A).
Myogenic potential of IM mesoangioblasts in vivo was tested by
intramuscular transplantation in irradiated scid?mdx immunodefi-
cient mice (Table 2, which is published as supporting information
onthePNASwebsite).Aftertwoconsecutiveinjections,thetibialis
anterior (TA) of mice treated with mesoangioblasts from DM
displayed, in the injected areas, numerous muscle fibers containing
human nuclei and expressing human dystrophin (average percent-
age of positive fibers was 60 ? 15% and 57 ? 10% in two animals
transplanted with DM-derived mesoangioblasts, respectively; Fig.
5B Right). In contrast, none of such fibers could be detected in the
TA of two mice injected with ‘‘wild-type’’ IBM mesoangioblasts.
These muscles showed the presence of isolated (1–2%) human
lamin A?C-positive nuclei in the interstitium and within some
muscle fibers, but dystrophin expression was negligible (Fig. 5B
Left). However, after transplantation of adenoMyoD-transduced
IBM mesoangioblasts, the TA of the two mice treated showed, in
the injected region, the presence of large areas reconstituted with
dystrophin-lamin A?C-positive fibers (33 ? 10% and 29 ? 8%,
respectively) (Fig. 5B Middle). Controls are shown in Fig. 9, which
is published as supporting information on the PNAS web site).
siRNA Inhibition of BHLHB3 Expression Restores a Normal Myogenesis
in IBM Mesoangioblasts. IBM mesoangioblasts had increased
mRNAlevelsofBHLHB3(Fig.10,whichispublishedassupporting
information on the PNAS web site), a group E subfamily member
of bHLH factors, which negatively regulates myogenesis by mod-
ulating the transcriptional activity of MyoD (17). Therefore, we
examined the effect of siRNA-mediated suppression of BHLHB3
in mesoangioblasts from three IBM patients. Interestingly, siRNA-
transfected cells were able to differentiate, giving rise to multinu-
cleated myosin-positive myotubes after 7 days in differentiation
medium (Fig. 6). To verify the siRNA specificity, we used a second
siRNA for BHLHB3 and found that also this siRNA had a similar
ability to rescue myogenesis (data not shown), unlike nonsilencing
control siRNA that was ineffective.
Discussion
Our study demonstrates that human adult mesoangioblasts can be
efficiently isolated from diagnostic muscle biopsies of patients with
IM. Antigenic and molecular characterization of these cells indi-
cated that mesoangioblasts represent a distinct type of mesoderm
progenitor cells, different from mesenchymal stem cells (E.T., and
S.F.,unpublishedobservations).InallIMpatients,thesecellsretain
the same proliferation ability of cells isolated from normal muscle,
and can be grown and expanded for as many as 25–30 passages,
althoughnotindefinitely.HerewehaveshownthatexposureofDM
Fig.4.
arteriesshowALP-positivestaining.IntenseALPstainingispresentinPMandDM
biopsies in perymisial and endomysial connective tissue, particularly strong in
roundcellsassociatedtovessels’walls.IMbiopsiescontainalsoavariablenumber
of small-sized regenerating muscle fibers with slight punctuate ALP positivity.
IBM muscle shows no ALP connective tissue staining. (Scale bar: 40 ?m.) (B)
Immunohistochemistry for MyoD or Pax7 with peroxidase-antiperoxidase (PAP)
on the same unfixed frozen sections of biopsies from PM, DM, IBM, and normal
musclefromwhichmesoangioblastswereisolated.RepresentativefieldsofaPM
biopsy showing a round vessel-associated cell (open arrowhead) strongly ALP-
positive containing a MyoD-positive nucleus. Several PAX7-positive cells (filled
arrowheads), identified as activated satellite cells, are associated with muscle
fibers showing regenerative aspects with slight ALP positivity, whereas the
roundish ALP-positive cells in the interstitium are PAX7-negative (arrow). CTRL,
control. (Scale bar: 10 ?m.)
ALPhistochemistryinIMmusclebiopsies.(A)Innormalmuscleonlysmall
Fig. 5.
blasts. (A) As shown by immunofluorescence and Western blot, AdenoMyoD-
transduced IBM mesoangioblasts after 7 days in differentiation medium, fused
intomultinucleatedmyosin-positivemyotubes.Approximately90%ofcellswere
MyoD-positive. (Scale bar: 10 ?m.) Efficiency of transduction was evaluated 48 h
after the infection by Western blot showing a marked up-regulation of MyoD.
?-actin was used as loading control. A representative experiment of three, each
performed in duplicate, is shown. (B) Double immunohistochemistry for human
dystrophinandlaminA?C.InvivomyogenicityofIMmesoangioblastswastested
by intramuscular transplantion in scid?mdx mice. After two consecutive trans-
plants, the TA of mice injected with IBM mesoangioblasts showed the presence
of occasional human nuclei (blue with AMCA-labeled human lamin A?C, arrow-
heads)withinmusclefibersandintheinterstitium,butdystrophinexpressionwas
undetectable (Left). On the contrary, after IBM adenoMyoD-infected mesoan-
gioblasts transplantation (Middle), large areas of injected muscle were reconsti-
tuted with fibers containing human lamin A?C-positive myonuclei (blue, arrow-
heads) and expressing dystrophin along the sarcolemma (red with Texas red-
labeled specific antibody). When DM mesoangioblasts were used (Right), the
majority of muscle fibers in the injected areas expressed dystrophin and con-
tained human nuclei (positive for lamin A?C, arrowheads). (Scale bars: 10 ?m.)
SYTOXgreenstainingonthesamesectionswasusedtoidentifybothhumanand
murine nuclei.
MyoDtransductionandinductionofdifferentiationinIBMmesoangio-
16998 ?
www.pnas.org?cgi?doi?10.1073?pnas.0603386103Morosetti et al.

Page 5

and PM mesoangioblasts to normal SDMC-conditioned medium is
greatly effective in inducing skeletal muscle differentiation, outlin-
ing the importance of muscle-secreted factors for myogenic mat-
uration of these stem cells.
IBM mesoangioblasts show the same proliferation ability ob-
servedinnormalmusclesuggestingthatthediseasehasnotreduced
their proliferation potency. Nevertheless, although they normally
differentiate into SMCs, their differentiation into skeletal muscle
seems markedly impaired, because no myotubes were observed
under the same conditions that promote massive skeletal muscle
differentiation of DM, PM and control mesoangioblasts. Interest-
ingly, cultures of satellite cell-derived progenitors (MyoD-positive)
obtained from the same IBM biopsies show normal myogenic
differentiation.Thesedataagreewithpreviousstudiesshowingthat
cultured IBM muscle cells proliferate and differentiate normally
and can be properly innervated (23), although the possibility exists
that satellite cells activation in vivo could be defective.
However, we show here that a progenitor cell, resident in a
perivascular niche of IBM muscle, is defective in myogenic deter-
mination and differentiation. No significant differences of age exist
between our IBM and DM?PM patients, excluding a consequence
ofmuscleagingandstronglysuggestingacausalcorrelationwiththe
specific pathophysiology of IBM. In fact, IBM presents an inade-
quate long-term regeneration despite a normal number of satellite
cells, a complete lack in muscle connective tissue of ALP-positive
cells, likely activated pericytes representing the cells from which
mesoangioblasts are established in vitro, and a failure of mesoan-
gioblasts to differentiate in vitro. This defect can be rescued by
transientMyoDexpression,astrategythat,independentlyfromthe
underlying molecular mechanism, immediately offers a therapeutic
opportunity. Skeletal myogenesis, a complex multistep process
mainlyregulatedbybHLHmyogenicfactors,dependsuponafinely
controlled equilibrium with repressors bHLH such as Id (24). IBM
mesoangioblasts highly express BHLHB3, which inhibits MyoD
function (17). Our data demonstrate that reestablishing a correct
balance between positive and negative bHLH factors, by MyoD
transienttransductionorsiRNAinhibitionofBHLHB3expression,
restores a normal myogenesis in IBM mesoangioblasts. As
BHLHB3 gene expression is regulated in a cell-type specific man-
nerbymultipleextracellularstimuliincludingcytokinesandgrowth
factors,itislikelythatcytokinesorotherfactorspresentintheIBM
muscle ‘‘milieu’’ but not in other IM muscle induce BHLHB3
expression. Like most bHLH, BHLHB3 probably transactivates its
own promoter, thus maintaining its inhibitory effect on myogenesis
also in vitro, away from the microenvironment that initially induced
its expression. Whereas several genes may control muscle differ-
entiation of mesoangioblasts, we identified BHLHB3 as one of the
important mediators of myogenic block in IBM. How much this
defect impacts on the pathogenesis of IBM remains to be investi-
gated; it is however intriguing that satellite cells are not directly
affected and yet IBM muscle fails to regenerate.
Currently it is unknown to what extent other progenitors such as
mesoangioblasts participate to regeneration in normal muscle after
inflammatory damage and in degenerative myopathies, either
directly or by feeding into the satellite cells pool, should this
reservoir be exhausted. Postnatal mesoangioblasts in vivo express
ALP, are located in perithelial position and can be considered as
part of the pericyte population (A. Dellavalle, M.S., R. Tonlorenzi,
E.T., B. Sacchetti, L. Perani, B. G. Galvez, G. Messina, R.M., S. Li,
G. Peretti, J. S. Chamberlain, W. E. Wright, Y. Torrente, S.F., P.
Bianco, and G.C., unpublished data). A characteristic histochem-
ical feature of PM and DM is a strong ALP positivity in perimysial
and endomysial connective tissue, as opposed to IBM as well as
other myopathies with increased connective tissue, usually display-
ing no or very little ALP-positivity (21). ALP staining correlates
more with regenerative properties of muscle rather than with
inflammatory changes and has been thought to reflect activated or
proliferating fibroblasts (21). Our study indicates that human
ALP-positive mesoangioblasts likely represent the cell population
activated in muscle connective tissue of IM, originating from the
perivascular niche and susceptible of myogenic determination in
vivo, as indicated by MyoD expression.
Mesoangioblasts present a distinct advantage over satellite cells,
i.e., the ability to cross the vessel wall which, together with easy
isolation and good myogenic potential, make them an ideal candi-
date for muscle reconstitution therapy by means of cell transplan-
tation. Their safety as therapeutic tool is demonstrated by limited
lifespan, absence of kariotype abnormalities or tumorigenesis in
nude mice (A. Dellavalle, M.S., R. Tonlorenzi, E.T., B. Sacchetti,
L. Perani, B. G. Galvez, G. Messina, R.M., S. Li, G. Peretti, J. S.
Chamberlain,W.E.Wright,Y.Torrente,S.F.,P.Bianco,andG.C.,
unpublisheddata).Moreovermesoangioblastscanbeeasilyisolated
withhighefficiencyevenfromfrozenmaterialandexpandedinvitro
tonumberofcellssuitableforapotentialinvivotreatment.Theuse
of frozen material is particularly relevant in IBM, often misdiag-
nosed as PM and suspected only after unsuccessful immune-
suppressive therapy, because it makes possible to go back to stored
muscle to expand mesoangioblasts. An intrinsic myogenic defect of
IBM mesoangioblasts can be corrected with viral transduction of
exogenous MyoD that overcomes BHLHB3 inhibition and, by act-
ivatingendogenousMyoD,irreversiblycommitcellstomyogenesis.
Thus, in vitro manipulation of these autologous cells induced to
effectively differentiate into skeletal muscle (without a stable viral
transduction as necessary in muscular dystrophies) could be par-
ticularly valuable to promote muscle repair and regeneration in a
disease such as IBM where relentless muscle involvement occurs
very selectively. The characteristic early and severe atrophy of
quadriceps, responsible for sudden falls and disabling gait impair-
ment, could be treated by infusing mesoangioblasts with selective
catheter-mediated delivery into the iliac arteries. The use of autol-
ogous cells eliminates the obligatory need of immune suppression,
although more selective anti-dysimmune drugs may eventually be
Fig. 6.
BHLHB3 siRNA. (A) BHLHB3 siRNA (Hs 1) was transfected into IBM mesoan-
gioblasts. Twenty-four hours after transfection, cells were harvested for RNA
extraction and RT-PCR analysis. (B) MAPK siRNA was used as positive control;
24 h after transfection, cells were harvested for Western blot. (C) BHLHB3
siRNA-transfected cells were shifted to differentiation medium after 24 h.
After 7-day culture, cells were stained with an anti-myosin antibody or har-
vested for Western analysis. (Scale bar: 40 ?m.) A representative culture and
immunoblot of one of three independent experiments (each one run in
duplicate) are shown.
Induction of myogenic differentiation of IBM mesoangioblasts by
Morosetti et al.PNAS ?
November 7, 2006 ?
vol. 103 ?
no. 45 ?
16999
NEUROSCIENCE

Page 6

combined with regenerative cell therapy, given the inflammatory
background of IBM muscle. Also rare cases of unresponsive PM
and DM would be potentially treatable as well by targeting muscle
groups essentials for motor and respiratory functions.
Methods
Patients. Diagnostic muscle biopsies were performed after in-
formed consent at the Neurology Department of Catholic Univer-
sity. We used fresh and fresh-frozen muscles from three normal
controls (one fresh, two frozen) (48–84 years of age, average 64 ?
18.33 years of age) and 12 patients with IM: 3 DM (2 fresh, 1
frozen),3PM(1fresh,2frozen)(33–75yearsofage,average52.5?
9.5 years of age), and 6 sporadic IBM (3 fresh, 3 frozen) (56–75
yearsofage,average67.4?18.1yearsofage).Diagnosiswasbased
on clinical evaluation and laboratory studies. None of the patients
receivedsteroidsorimmunosuppressivetherapybeforebiopsy.This
research was approved by the ethical committee of our institution.
Cell and Organ Cultures. Fragments of intramuscular vessels and
surrounding mesenchymal tissue were plated as described (8, 10).
Details are provided in Supporting Methods, which is published as
supporting information on the PNAS web site.
A fragment from the same muscle biopsy was also cultured to
obtain primary muscle cultures from satellite cells by using the
explantation reexplantation method (18).
Characterization of Human Mesoangioblasts from IM by FACS, Cell-
Cycle Analysis, and Growth Curve. Cells (5 ? 104) were incubated
with FITC-, PE-, or APC-conjugated mAbs directed against
AC133?1, CD34, VEGF-RII (KDR), CD45, CD49b, CD44, and
CD13. Details are provided in Supporting Methods.
In Vitro Differentiation. Skeletal muscle differentiation. Mesoangio-
blasts were (i) cultured under standard differentiating conditions
for SDMC (17); (ii) cocultured with a 4-fold excess of C2C12
myoblasts; (iii) cultured in normal human SDMC-conditioned
medium and then exposed to differentiation medium. At each time
point, cells were fixed or harvested for protein extraction. Differ-
entiation assays were performed in all IM samples studied and
repeatedatleastthreetimesforeachpatientwithconsistentresults.
Smooth muscle, osteoblasts, and neural differentiation. Differentiation
of mesoangioblasts into SMCs, osteoblasts, and neural cells was
tested as described (8, 12, 25).
Immunostainings. Immunostainings were performed as described
(5, 6). Details are available in Supporting Methods.
Double Immunohistochemistry-Histoenzymatic ALP Staining. Immu-
nocytochemistry for MyoD or Pax7 was performed with peroxi-
dase-antiperoxidase followed by histoenzimatic ALP staining on
the same unfixed frozen sections of the biopsies used for mesoan-
gioblast isolation.
Gene Expression Profiling and Data Analysis. Proliferating mesoan-
gioblasts from normal and IM muscles were analyzed for gene
expression by Affimetrix gene array (26, 27). Details are provided
in Supporting Methods.
Western Blot Analysis and RT-PCR. Protein expression was analyzed
by Western blot according to standard methods. Details on anti-
bodiesandprimersareavailableinSupportingMethods.Primersand
PCR conditions for BHLHB3 have been described (28).
Cell Transduction. Mesoangioblasts from three IBM patients were
adenoMyoD-transduced (29), cultured for 24 h in growth medium,
and then either shifted to differentiation medium for 7 days or
injected in vivo. Details are available in Supporting Methods.
Intramuscular Transplantation of DM, IBM, and AdenoMyoD-Trans-
duced IBM Mesoangioblasts into Irradiated scid?mdx Mice. Mesoan-
gioblasts from DM, IBM, and IBM adenoMyoD-transduced were
injectedintotherightorleftTAofsixmice(twopergroup).Details
are available in Supporting Methods.
BHLHB3 siRNA. Predesigned siRNA directed against human
BHLHB3 (Hs BHLHB3 1 and Hs BHLHB2 HP siRNA; Qiagen,
Valencia, CA) were transfected into IBM mesoangioblasts. Details
are available in Supporting Methods.
Statistical Analysis. All data were expressed as mean ? SD. One-
way ANOVA was used to compare differences between groups.
Statistical significance was set at P ? 0.05.
We thank Dr. Libera Berghella and Gabriella Proietti for assistance, Dr.
Marco Crescenzi (Istituto Superiore di Sanita `, Rome, Italy) for adeno-
viral MyoD vector, and Dr. Yvan Torrente (University of Milan, Milan,
Italy) for scid?mdx mice. This work was supported by grants from the
Italian Ministries of Health and of University and Scientific and Tech-
nological Research, from Telethon, from the Duchenne Parent Project,
and from the European Community.
1. Dalakas MC, Hohlfeld R (2003) Lancet 362:971–982.
2. MastagliaFL,GarleppMJ,PhillipsBA,ZilkoPJ(2003)MuscleNerve27:407–425.
3. Askanas V, Engel WK (1998) Am J Pathol 153:1673–1677.
4. Askanas V, Engel WK (2003) J Child Neurol 18:185–190.
5. Broccolini A, Ricci E, Pescatori M, Papacci M, Gliubizzi C, D’Amico A,
Servidei S, Tonali P, Mirabella M (2004) J Neuropathol Exp Neurol 63:650–659.
6. Broccolini A, Gidaro T, Morosetti R, Gliubizzi C, Servidei T, Pescatori M,
Tonali PA, Ricci E, Mirabella M (2006) J Neurochem 96:777–789.
7. De Angelis L, Berghella L, Coletta M, Lattanzi L, Zanchi M, Cusella-De
Angelis MG, Ponzetto C, Cossu G (1999) J Cell Biol 147:869–878.
8. Minasi MG, Riminucci M, De Angelis L, Borello U, Berarducci B, Innocenzi
A, Caprioli A, Sirabella D, Baiocchi M, De Maria R, et al. (2002) Development
(Cambridge, UK) 129:2773–2783.
9. Cossu G, Bianco P (2003) Curr Opin Genet Dev 13:537–542.
10. Sampaolesi M, Torrente Y, Innocenzi A, Tonlorenzi R, D’Antona G, Pelle-
grino MA, Barresi R, Bresolin N, De Angelis MG, Campbell KP, et al. (2003)
Science 301:487–492.
11. Brunelli S, Tagliafico E, De Angelis FG, Tonlorenzi R, Baesso S, Ferrari S,
Niinobe M, Yoshikawa K, Schwartz RJ, Bozzoni I, et al. (2004) Circ Res
94:1571–1578.
12. Tagliafico E, Brunelli S, Bergamaschi A, De Angelis L, Scardigli R, Galli D,
Battini R, Bianco P, Ferrari S, Cossu G, et al. (2004) J Cell Sci 117:4377–4388.
13. Galvez BG, Sampaolesi M, Brunelli S, Covarello D, Gavina M, Rossi B,
Costantin G, Torrente Y, Cossu G (2006) J Cell Biol 174:231–243.
14. De Paepe B, Schroder JM, Martin JJ, Racz GZ, De Bleecker JL (2004)
Neuromuscul Disord 14:265–273.
15. Massague J, Cheifetz S, Endo T, Nadal-Ginard B (1986) Proc Natl Acad Sci
USA 83:8206–8210.
16. Polesskaya A, Seale P, Rudnicki MA (2003) Cell 113:841–852.
17. Azmi S, Ozog A, Taneja R (2004) J Biol Chem 279:52643–52652.
18. Askanas V, Engel WK (1992) in Handbook of Clinical Neurology, eds Rowland
LP, Di Mauro S (Elsevier, Amsterdam), pp 85–116.
19. Puri PL, Wu Z, Zhang P, Wood LD, Bhakta KS, Han J, Feramisco JR, Karin
M, Wang JY (2000) Genes Dev 14:574–584.
20. Simone C, Forcales SV, Hill DA, Imbalzano AN, Latella L, Puri PL (2004) Nat
Genet 36:738–743.
21. Askanas V, Engel WK, Mirabella M (1994) Curr Opin Neurol 7:448–456.
22. Asakura A, Komaki M, Rudnicki M (2001) Differentiation (Berlin) 68:
245–253.
23. McFerrin J, Engel WK, Askanas V (1999) Neurology 53:2184–2187.
24. Norton JD (2000) J Cell Sci 113:3897–3905.
25. Gritti A, Frolichsthal-Schoeller P, Galli R, Parati EA, Cova L, Pagano SF,
Bjornson CR, Vescovi AL (1999) J Neurosci 19:3287–3297.
26. Liu WM, Mei R, Di X, Ryder TB, Hubbell E, Dee S, Webster TA, Harrington
CA, Ho MH, Baid J, et al. (2002) Bioinformatics 18:1593–1599.
27. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U,
Speed TP (2003) Biostatistics 4:249–264.
28. Li Y, Xie M, Song X, Gragen S, Sachdeva K, Wan Y, Yan B (2003) J Biol Chem
278:16899–16907.
29. Murry CE, Kay MA, Bartosek T, Hauschka SD, Schwartz SM (1996) J Clin
Invest 98:2209–2217.
17000 ?
www.pnas.org?cgi?doi?10.1073?pnas.0603386103Morosetti et al.