INTRODUCTIONAmyotrophic lateral sclerosis (ALS) is

an adult-onset neurodegenerative disor-der affecting upper and lower motorneurons and leading to muscular atro-phy and paralysis. The pathological hall-marks of ALS are the atrophy of dyingmotor neurons, which are surrounded byreactive astrocytes and microglia (1).Riluzole is the only drug approved by

the U.S. Food and Drug Administrationfor the treatment of ALS patients. How-ever, its effect is both minimal and con-troversial (2). In the emerging field of re-generative medicine, mesenchymal stemcell (MSC) transplantation holds thepromise of treating a variety of degener-ative diseases where no specific or effec-tive treatment is currently available (3).MSCs can differentiate into multiple

mesoderm-type and non–mesoderm-typecells, including neuronal-like cells (4–6).

Indeed, a previous study in our labora-tory demonstrated that human adultskeletal muscle–derived stem cells (SkmSCs) may also differentiate intoneurogenic cell lineages (7). In this study,we demonstrated that human fetal skele-tal muscle may contain MSC-like cellsendowed with neurogenic differentiationcapabilities in vitro.

Moreover, considering the antiinflam-matory and immunomodulatory proper-ties and hypoimmunogenic nature ofMSCs (8), we hypothesized that musclemay be a tissue source for the isolation ofMSCs for development of cell-basedtherapies for human neurodegenerativediseases with multifactorial features suchas ALS. The potential of stem cell ther-apy has already been demonstrated in a

M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2 | C A N Z I E T A L . | 4 0 1

Human Skeletal Muscle Stem Cell Antiinflammatory ActivityAmeliorates Clinical Outcome in Amyotrophic LateralSclerosis Models

Laura Canzi,1,2 Valeria Castellaneta,3 Stefania Navone,1 Sara Nava,1 Marta Dossena,1 Ileana Zucca,4

Tiziana Mennini,3 Paolo Bigini,3* and Eugenio A Parati1*

1Department of Cerebrovascular Disease, IRCCS Foundation, Neurological Institute “C. Besta,” Milan, Italy; 2Department ofBiotechnology and Biosciences, University of Milano-Bicocca, Milan, Italy; 3Department of Molecular Biochemistry andPharmacology, Mario Negri Institute for Pharmacological Research, Milan, Italy; and 4Science Direction Unit, IRCCS Foundation,Neurological Institute “C. Besta,” Milan, Italy

Mesenchymal stem cell (MSC) therapy is considered one of the most promising approaches for treating different neurode-generative disorders, including amyotrophic lateral sclerosis (ALS). We previously characterized a subpopulation of human skele-tal muscle–derived stem cells (SkmSCs) with MSC-like characteristics that differentiate into the neurogenic lineage in vitro. In thepresent study, we evaluated the SkmSC therapeutic effects in the most characterized model of spontaneous motor neuron de-generation, the Wobbler (Wr) mouse. Before evaluating the therapeutic efficacy in the Wr mouse, we followed the route of Skm-SCs at different times after intracerebroventricular injection. Two exogenous tracers, superparamagnetic iron oxide (SPIO)nanoparticles and Hoechst 33258, were used for the in vivo and ex vivo tracking of SkmSCs. We found that the loading of bothHoechst and SPIO was not toxic and efficiently labeled SkmSCs. The magnetic resonance imaging (MRI) system 7 Tesla allowedus to localize transplanted SkmSCs along the whole ventricular system up to 18 wks after injection. The ex vivo Hoechst 33258 vi-sualization confirmed the in vivo results obtained by MRI analyses. Behavioral observations revealed a fast and sustained im-provement of motor efficacy in SkmSC-treated Wr mice associated with a relevant protection of functional neuromuscular junc-tions. Moreover, we found that in SkmSC-treated Wr mice, a significant increase of important human antiinflammatory cytokinesoccurred. This evidence is in accordance with previous findings showing the bystander effect of stem cell transplantation in neu-rodegenerative disorders and further strengthens the hypothesis of the possible link between inflammation, cytotoxicity and ALS.Online address: http://www.molmed.orgdoi: 10.2119/molmed.2011.00123

*PB and EAP contributed equally to this work.

Address correspondence to Laura Canzi, IRCCS, Neurological Institute “C. Besta,” Via

Celoria, 1 20133 Milan, Italy. Phone: +39-02-23942272; Fax: +39-02-23942272; E-mail:

laura.canzi@istituto-besta.it.

Submitted April 1, 2011; Accepted for publication November 3, 2011; Epub

(www.molmed.org) ahead of print November 4, 2011.

familial model of ALS—the transgenicmutant superoxide dismutase 1 (SOD1)mouse (9).

Thus, in the present study, the possi-ble therapeutic effects of human SkmSCtransplantation by a single bilateral in-tracerebroventricular administration wasevaluated in a well-characterized murinemodel of spontaneous motor neuron de-generation, the Wobbler (Wr) mouse(10). The Wr mouse carries a homozy-gous missense mutation (L967Q) in thevacuolar protein sorting 54 (Vps54) genethat codifies for a protein involved inthe retrograde transport from late endo-somes to the Golgi apparatus (11). Al-though this mutation is not yet associ-ated with any case of ALS, it leads toearly motor neuron disorder that sharesthe many neuropathological and clinicalfeatures found in patients with ALS.Moreover, the evidence that chronictreatment with Riluzole improves motorbehavior, prevents biceps muscle atro-phy and decreases the amount of motorneuron loss in Wr mice makes it an in-teresting and reliable model to investi-gate the pathological mechanisms in-volved in sporadic ALS and the effect ofa wide range of therapeutic approaches(12,13).

Our results show that treatment of Wrmice with human SkmSCs during thepresymptomatic phase can delay diseasesymptoms of Wr mice, possibly via im-munomodulation and inhibition of in-flammatory response. We hope that agreater understanding of human SkmSCtherapeutic action may create new per-spectives in the pharmacological treat-ment of ALS.

MATERIALS AND METHODS

SkmSC Isolation, Culture andDifferentiation

SkmSCs were derived from muscles of12- to 16-wk-old fetuses obtained follow-ing the ethical guidelines of the Euro-pean Network for Transplantation (NECTAR) and were processed as previ-ously reported (7). Single cells are platedin the presence of 20 ng/mL EGF and

10 ng/mL basic fibroblast growth factor(bFGF) in a serum-free stem-cell mediumoptimized for neural stem cell growthsupplemented with Noggin (100 ng/mL;R&D Systems, Minneapolis, MN, USA)(growth medium) (14). The experimentalprotocol was approved by the ethicscommittee of the Foundation IRCCSNeurological Institute “C. Besta,” Milan,Italy, and the Foundation IRCCS Policlinico-Mangiagalli-Regina Elena,Milan Italy.

Fluorescence-Activated Cell Sorter(FACS) Analysis

The phenotypic characterization ofSkmSCs was evaluated by flow cytome-try (fluorescence-activated cell sorter[FACS]) as described previously [7]) withCD44, CD166, CD73 (BD Pharmingen,San Jose, CA, USA), CD56, CD133/2,CD34 (Miltenyi Biotec, Bisley, Surrey,UK), CD105 (ABDSerotec, Raleigh, NC,USA) and CD90 (Millipore, Temecula,CA, USA).

ImmunocytochemistryFor differentiation assay, SkmSCs were

plated as previously reported (7). Cellcomposition was analyzed by means ofimmunostaining with lineage-specific antibodies: glial fibrillary acidic protein(GFAP) (rabbit, 1:500); NG2 (new glue 2;CSPG4, chondroitin sulfate proteogly-can 4) (rabbit, 1:200); class III β tubulin(mouse, 1:100) (Chemicon, Temecula,CA, USA) and Nestin (mouse 1:100; R&DSystems).

Magnetic Labeling of Human SkmSCsSuperparamagnetic iron oxide (SPIO)

labeling (Endorem AMI-25; Guerbet; par-ticle size, 80–150 nm; stock solution, 11.2mg Fe/mL) was performed as previouslyreported (15).

Hoechst 33258 Dye Staining of SPIO-Labeled Human SkmSCs

SPIO-labeled SkmSC cultures in 12-wellplates were incubated with 2 μmol/LHoechst 33258 reagent (MolecularProbes, Eugene, OR, USA) for 1 h at 37°Cin the dark. After incubation, the cells

were plated on matrigel-coated glass-chamber slides (Nunc, Naperville, IL,USA) in serum-free growth medium for2 d and fixed in 4% paraformaldehyde(PFA). After that, SPIO efficiency was an-alyzed by means of immunostainingwith anti-dextran antibody (STEMCELLTechnologies Inc., Vancouver, BC, Canada).

AnimalsProcedures involving animals and

their care were conducted in conformitywith the institutional guidelines that arein compliance with national and interna-tional laws and policies (16–18) and theNational Institutes of Health (NIH) Guidefor the Care and Use of Laboratory Animals(19). The protocol for the use of labora-tory animals was approved by the ItalianMinistry of Health and by an internalethical committee.

Wr mice were originally obtained fromNIH genetics and then bred at CharlesRiver Italy, Calco (LC), Italy. Heterozy-gous founders were designed by geno-typing (20).

Experimental ScheduleThe recruitment of Wr mice for our

studies was carried out after clear diag-nosis of the disease on the basis of phe-notype analysis (fourth week of life).Age-matched healthy littermates, ho-mozygous for the wild-type form of theVps54 gene, were selected after genotyp-ing (12).

For each experimental group, thesame number of males and femaleswere recruited, since no sex-related dif-ference in the evolution of the Wr dis-ease was reported. A schematic timelineof the experimental plan is shown inSupplementary data 1: ExperimentalPlan Timeline.

Human SkmSC Transplantation in ALSAnimal Models

For injection of SkmSCs, mice wereanesthetized using isoflurane and placedin a stereotaxic frame. The head wasshaved and a ~1-cm incision was madeto expose the skull. By use of a micro-

4 0 2 | C A N Z I E T A L . | M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2

S k m S C I M M U N O M O D U L A T O R Y E F F E C T I N T H E A L S M O D E L

drill, two 1- to 2-mm holes were pro-duced to allow two intracerebroventricu-lar injections (stereotaxic coordinates of4-wk-old control mice ± 1.0 mm lateral tobregma, ± 0 mm rostro-caudal to bregmaand 3.0 mm deep from the skull surface;stereotaxic coordinates of 4-wk-old Wrmice ± 0.8 mm lateral to bregma, + 0.5 mmrostro-caudal to bregma and 2.5 mmdeep from the skull surface). As shamcontrols, animals were injected with 5 μLphosphate-buffered saline (PBS) withoutcells. All injections were performed witha Hamilton syringe. A total of 5 × 105

cells with a total volume of microliterswas injected per animal. No immuno-suppression was given, since humanfetal cells with low immunogenicity wereused.

Magnetic Resonance ImagingExperiments

Magnetic resonance imaging (MRI)scanning was performed by a 7 Tesla and30-cm bore Bruker spectrometer (BioSpin70/30 USR) equipped with a surfacemouse brain radio-frequency (RF) coil. Athree-orthogonal plane gradient echoTriPilot scan (TriPilot protocol of ParaVision 4.0 software; Bruker BioSpinMRI GmbH, Ettlingen, Germany) wasused as a geometric reference to locatethe olfactory bulbs, and the sections werechosen from this scan. T2-weighted im-ages (axial, sagittal, coronal; rapid acqui-sition with refocusing echoes (RARE) fac-tor = 8, time of repetition (TR) = 3.3 s, timeto echo (TE) = 54 ms, interecho time = 13.5 ms, field-of-view (FOV) = 2.2 × 2.2 cm2,data matrix 256 × 256, slice thickness = 1 mm) were acquired to visualize anatom-ical details. T2*-weighted images (axial;FLASH Flip Angle = 30 degrees, TR = 1 s,TE = 7 ms, FOV = 1.8 × 1.8 cm2, data ma-trix 128 × 128, slice thickness = 0.6 mm)were acquired to visualize the suscepti-bility artifacts due to the presence of iron.To acquire the image of the spinal cordwith a gray-white matter contrast, T1-weighted images (axial; spin echo, TR =300 ms, TE = 11.7 ms, FOV = 2.7 × 2.7cm2, data matrix 256 × 256, slice thick-ness = 0.5 mm) were acquired.

Disease Progression AssessmentBehavioral trials were performed ac-

cording to the procedure described previously (21).

Motor Neuron and NeuromuscularJunction Counting

Choline acetyl transferase (ChAT)-pos-itive neuron counting was carried out inthe cervical region corresponding toC2–C5 in 7- and 9-wk-old Wr and age-matched controls (n = 5 for each group),according to the procedure previouslydescribed (22).

Neuromuscular junction counting wasperformed by employing bromo indoxylacetate dye–staining methods as de-scribed (23).

ImmunohistochemistryThe mouse astroglial markers were re-

vealed by antibodies against GFAP(mouse monoclonal antibody; Immuno-logical Sciences, Rome, Italy; dilution1:1,000), CD11b (rat monoclonal antibodyOX 42; Serotec, Oxford, UK; dilution1:300). Sections were observed with anOlympus Fluoview microscope BX61with the confocal system FV500. Pseudo-colors were used, and the signals fromthe three different channels were auto-matically merged by Olympus Fluoviewsoftware.

Real-Time Quantitative PolymeraseChain Reaction

RNA extractions were performed onfour total brains and cervical spinal cordsof 6-wk-old SkmSC-transplanted Wr mice(2 wks after intracerebroventricular trans-plant with SkmSCs), on four total brainsand cervical spinal cords of 6-wk-old Wrmice and on controls of SkmSCs culturedas described above. Real-time polymerasechain reaction (PCR) was performedusing the Human Inflammatory Cy-tokines & Receptors RT2 Profiler™ PCRArray (SuperArray Bioscience Corpora-tion, Frederick, MD, USA).

Pathway-focused gene expressionanalysis was performed with the PCRArray System and the PCR Array DataAnalysis Web Portal. RT2 Profiler™ PCR

Arrays were done on separate cDNAs atleast three times.

Bio-Plex AssayProtein extractions were performed on

four total brains and cervical spinal cordsof 6-wk-old SkmSC-transplanted Wr mice(2 wks after intracerebroventricular trans-plant with SkmSCs), on four total brainsand cervical spinal cords of 6-wk-old Wrmice and on controls of SkmSCs culturedas described above, with Bio-Plex celllyses kit (Bio-Rad, Hercules, CA, USA)following the manufacturer’s instruc-tions. Two human cytokines (hIL-13 Bio-Plex human group I region 56, hIL-10Bio-Plex human group I region 51) wereestimated by using Bio-Plex ProteinArray System (Bio-Rad). Samples werediluted in lyses buffer plus bovine serumalbumin 0.5% at a final concentration of0.5 and 10 mg/mL. The assay was per-formed according to the manufacturer’sinstructions. Data have been normalizedwith background fluorescence intensity.

Statistical AnalysisNumerical values are expressed as

mean ± standard deviation (SD). Com-parisons of parameters were performedwith a Student t test. Comparisons of pa-rameters among three groups were madewith a one-way analysis of variance test.P < 0.05 was considered significant.

All supplementary materials are availableonline at www.molmed.org.

RESULTS

In Vitro Characterization of HumanSkmSCs

We characterized five human SkmSClines to verify their proliferative and dif-ferentiative potential (7). Under serum-free culture conditions, we were able toisolate a homogeneous SkmSC popula-tion with human MSC-like morphologythat grew as adherent cultures. As Skm-SCs gradually continued to grow andreached confluency, they were constantlycollected to count the cell number withTrypan Blue Assay. As a result, we ob-

R E S E A R C H A R T I C L E

M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2 | C A N Z I E T A L . | 4 0 3

4 0 4 | C A N Z I E T A L . | M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2

S k m S C I M M U N O M O D U L A T O R Y E F F E C T I N T H E A L S M O D E L

Figure 1. In vitro characterization of SkmSCs. (A) Growth rate of SkmSCs: growth rate = (n + 1) – n/n values; n = number of passage. (Band B′) FACS analysis of SkmSCs. (C) Differentiated SkmSCs stained positively for β-tubulin-III (red) and GFAP (green). (D) Immunostainingfor Nestin (red). DIV, days in vitro. All nuclei were counterstained with 4′-6-diamidino-2-phenylindole (DAPI) (blue). Results are representa-tive of five different SkmSC lines. Results are displayed as means ± SD. (C, D) Scale = 50 μm.

served that SkmSCs kept in culture forseveral weeks (up to 20 passages in vitro[P20]) maintain a stem cell–like prolifera-tion profile (Figure 1A). The immunophe-notypic analysis by FACS of this cell pop-ulation showed that a high percentage ofthe SkmSCs (P8–10) expressed character-istic cluster of differentiation (CD) mark-ers of MSCs, for example, CD105 (89.18 ±4.44%), CD73 (98.27 ± 2.57%), CD90(96.59 ± 2.57%), CD44 (98.9 ± 0.75%),CD166 (69.25 ± 5.95%) and CD56 (69.7 ±11.6%) (Figures 1B, B′) (24).

However, hematopoietic stem cell CDmarkers seemed to be present in a lowerfraction in human SkmSCs (P8–10), forexample, CD34 (44.8 ± 27.8) and CD133(33.5 ± 19.1) (see Figures 1B, B′). This re-sult suggested a common precursor be-tween the hematopoietic and mesenchy-mal lineages (25).

Human SkmSCs, when induced withneural differentiation media for 7 d,started showing morphological changessuggestive of cells of astrocytic/neuronallineages by d 4 and appear highly posi-tive for the neural stem cell markerNestin, for example (99.34 ± 0.2%) (Fig-ure 1D). At 7 d after neural differentia-tion conditions, the distribution of im-munoreactivity against GFAP andβ-tubulin-III was both high and intense,confirming the neurogenic potential ofthe human SkmSCs.

A small number of human SkmSC β-tubulin-III–positive cells also appearedto be positively immunoreactive withGFAP, for example, β-tubulin-III (44.62 ± 7.4%) and GFAP (87.97 ± 9.5%)(Figure 1C). At this time, a subset of cellscoexpressed neuronal (β-tubulin-III) andglial marker GFAP, indicating that thesehuman SkmSC samples might be in theprogenitor stage (or glioblast) of neuro-genic differentiation (β-tubulin-III/GFAP39.9 ± 6.8%) (see Figure 1C) (26).

In Vitro Labeling of SkmSCs with SPIOand Poly-L-Lysine Transfect Agent

We observed that SPIO labeling ofSkmSCs was feasible and did not impaircell viability or biological characteristicsin vitro (Figures 2A, A′).

In Vivo–Ex Vivo Correlation of SkmSCsTracking Hoechst/SPIO DoubleLabeling

The double labeling of SkmSCs withfluorescent (Hoechst 33258) and SPIOtracers allowed us to evaluate the possi-ble anatomical correspondence betweenthe hypointense signal associated withthe presence of iron oxide and the fluo-rescence related to the living nuclear dye.MRI analyses, performed 1 d after celltransplantation, showed hypointenseareas corresponding to the ventricularsystem. These results were similar to re-sult observed in animals only receivingthe SPIO-labeled human SkmSCs (Fig-ures 3A, B) and were observed for alltransplanted Wr mice.

Figures 2B and B′′′ show four differentrepresentative images of axial MRI slices,1 wk after cell transplantation, corre-sponding to the anterior body of lateralventricles (see Figure 2B), the body oflateral ventricles close to the hippocam-pus (see Figure 2B′) and the periaque-ductal gray and cistern magna in thecerebellum and brainstem (see Figure 2B’’).Interestingly, the observation of spinalcord slices revealed that a relevantamount of Hoechst-positive nuclei waslocalized in the external surface. This re-sult could likely be associated with men-ingeal layers (see Figure 2B′′′). The MRIanalysis confirmed the pattern of SPIOdistribution observed 1 d after cell trans-plantation. In addition, highly magnified

R E S E A R C H A R T I C L E

M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2 | C A N Z I E T A L . | 4 0 5

Figure 2. SkmSC SPIO internalization and in vivo MRI tracing. (A, A′) SPIO-labeled SkmSCsstained with an anti-dextran antibody (red) and counterstained with a viable cells dyeHoechst 33258 (blue) for quantitative and qualitative analysis of SPIO labeling efficiency;scale = 100 μm. (B–B′′′) In vivo MRI of Wr mice brain postgrafted with Hoechst/SPIO-doublelabeled SkmSCs 1 wk after injection.

pictures showed the presence of Hoechst33258–positive nuclei corresponding toventricular compartments characterizedby a strong decay of signal intensity (seeFigure 2B, B′, lower-left insert). Althoughseveral spinal cord sections analyzedfrom Wr mice 1 and 6 d after transplan-tation displayed a relevant number ofHoechst 33258–positive nuclei, we didnot find a migration of fluorescent cellsto the gray matter and, in particular, inthe ventral horns (data not show).

In spite of a small decay of SPIO- induced hypointense signal detected byMRI, a drastic reduction in both the in-

tensity and the overall number ofHoechst 33258–positive nuclei was ob-served in the ventricular system of Wrmice 8 wks after human SkmSC trans-plantation (data not show). This apparentdichotomy among the two different ap-proaches of cell tracking could likely bedue to the bleaching of fluorescent dye orthe detachment of bysbenzimide mole-cules from the minor groove of DNA.

The immunostaining results of Wrbrains 18 wks after human SkmSC trans-plantation showed only a modest humanSkmSC integration in the brainparenchyma, as confirmed by the pres-

ence of cells immunoreactive for thehuman-nuclei and carboxy-dextran (datanot show).

Clinical OutcomeFrom the fourth to the eighth week of

life, the body weight increase was similarbetween the Wr mice transplanted withhuman SkmSCs and the group receivingPBS alone. In the last week of clinical ob-servation, the group of treated miceshowed a higher trend than that of vehicle-treated mice, but this differencedid not statistically differ among the twogroups (Figure 4A). The rapid and pro-gressive alteration of forepaw adductionobserved in Wr mice receiving PBS in-tracerebroventricularly was greatly atten-uated by human SkmSC transplantation(Figure 4B). The beneficial effect was al-ready found at the first week after trans-plantation and was maintained for thewhole duration of clinical observation.The attenuation of the degree of forelimbmuscle atrophy was accompanied by asignificant improvement of muscularstrength in Wr mice transplanted withhuman SkmSCs (Figure 4C). The signifi-cant difference in terms of muscularstrength between the two experimentalgroups occurred at the sixth week of lifeand was protracted until the end of thestudy.

Histopathology of Human SkmSCsTransplanted Wr Mice

Representative images of ChAT- positive motor neurons in the cervicalspinal cord from 9-wk-old healthy mice(Figure 4D) and from age-matched Wrmice (Figure 4E) clearly show themarked process of degeneration occur-ring in affected mice. A reduction of~60% of ChAT- positive neurons wasfound in the cervical spinal cord Wr micereceiving PBS alone (12.6 ± 1.32 meanmotor neurons (MNs) ± standard error ofthe mean [SE] for each section) comparedwith age-matched healthy mice (32.6 ±2.7). The transplantation with SkmSCsdid not lead to a reduction in the rate ofmotor neuron death at the cervical spinalcord level of 9-wk-old Wr mice (12.4 ±

4 0 6 | C A N Z I E T A L . | M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2

S k m S C I M M U N O M O D U L A T O R Y E F F E C T I N T H E A L S M O D E L

Figure 3. Long-term monitoring of SPIO-labeled SkmSCs. (A, B) Representative MRI axialsections of T2-weighted images mouse brains (A) and T1-weighted images cervical spinalcord (B). (C) Photograph of Wr CNS. The immunostaining of Wr whole-mount spinal cordcervical region with anti–human nuclei (AHN) staining (mouse monoclonal antibody[mAb] anti–human nuclei; Chemicon, Temecula, CA, USA; 1:250) (red) 18 wks after trans-plant. All nuclei were counterstained with DAPI. T0 = MRI acquisition of pretransplantedmouse brain (4 wks of age); T1 = 0 h posttransplant (4 wks of age); T2 = 2 wks posttrans-plant (6 wks of age); T3 = 4 wks posttransplant (8 wks of age); T4 = 8 wks posttransplant(12 wks of age); T5 = 14 wks posttransplant (18 wks of age).

0.4). The counting of ChAT-positivemotor neurons, performed at an earlierstage of symptom progression (at theseventh week of age), did not show a sig-nificant difference between Wr micetransplanted with SkmSCs (14.9 ± 1.3)and PBS-treated Wr (17.4 ± 1.9) and alsorevealed a trend of protection in mice re-ceiving cells (P = 0.0715).

Figures 4D′ and E′ show a representa-tive view of biceps muscle fibers in bothhealthy (see Figure 4D′) and sympto-matic Wr mice (see Figure 4E′). Thecomparison between the two images re-veals a marked alteration in the in-tegrity of muscle fibers and a hypotro-phy of neuromuscular junctions. In

accordance with this observational re-sult, we found a drastic reduction onthe percentage of active neuromuscularjunctions in 7-wk-old Wr mice com-pared with their age-matched controls(the percentage of healthy mice wasnormalized and expressed as 100).However, a little but significant preser-vation was found in mice treated withSkmSCs (56.3 ± 1.9) compared with PBS-treated age-matched Wr mice (39.9 ±6.0) (*P = 0.221). We also examined themigration of transplanted SkmSCs. His-tograms showing the values of ChAT-positive motor neurons and active neu-romuscular junctions are shown inFigures 4F–H.

Molecular Characterization of SkmSCRegenerative Proprieties

To characterize the molecular factorsinvolved in therapeutic proprieties ofhuman SkmSCs in the Wr mice, we per-formed a human gene expression pro-file of the regulation of key genes in-volved in the neurogenesis processesand in the inflammatory response intotal brains and spinal cords of 6-wk-old Wr mice (2 wks after human SkmSCtransplant). We focused our analysis onchanges in gene expression higher thanfourfold (Table 1). We also performed ahuman protein expression analysis toconfirm the production of human inter-leukin (IL)-10 and IL-13 proteins in

R E S E A R C H A R T I C L E

M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2 | C A N Z I E T A L . | 4 0 7

Figure 4. Clinical outcome and histopathology of human SkmSC-trasplanted Wr mice. (A–C) Behavioral scores of Wr mice that receivedhuman SkmSCs or vehicle. Each point represents the mean ± SD of 15 animals per group. —●—, Wobbler PBS; ––▲––, Wobbler SkmSCs. (D, E)Representative immunohistochemistry of ChAT-positive motor neurons in cervical spinal cord from 9-wk-old healthy mice (D) and from age-matched Wr mice (E); scale = 100 μm. (D′ and E′) Representative view of biceps muscle fibers in both healthy (D′) and Wr mice (E′); scale =25 μm. Bromo indoxyl acetate (blue). (F–H) Histograms showing the values of ChAT-positive motor neurons and active neuromuscular junc-tions. Data represent mean ± SD of 14 replications. *P < 0.05; **P < 0.01; ***P < 0.0001; ^^^P < 0.0001. Q36 . CTR, Control healthy mice.

human SkmSC transplant Wr mice(Table 1).

Immunofluorescent experiments at thecervical spinal cord level of both 7-wk-old Wr mice treated with SkmSCs (Fig-ures 5B, B′′′) and age-matched PBS-treated Wr mice (Figures 5C, C′′′)confirmed the relevant process of as-

trogliosis occurring during disease pro-gression. However, although SkmSCsdid not arrest the process of hypertrophyof astrocytes (Figures 5B′, C′), a mild at-tenuation of microglia activation (moni-tored by the expression of CD11b) seemsto be produced by cell transplantation(Figures 5C, C′).

DISCUSSIONIn this study, a total of 500,000 human

SkmSCs were injected into the intracere-broventricular space of 4-wk-old Wrmice to assess their therapeutic value inreducing the pathology in the Wr mousemodel of ALS. Treated Wr mice did ex-hibit significant improvements in motorfunction, although none of these miceapproached the functional capabilities ofwild-type mice. The choice of SkmSCswith MSC distinctiveness as a possiblecell therapy in motor neuron disorders isclosely related to previous experimentalevidence on MSC treatments for centralnervous system (CNS) diseases (27).

Recent studies have revealed that thehypothesized mechanism of MSC actionseems to be related to the release of pro-tective factors, even far from the injuredsite, rather than the actual replacementof degenerating neurons. Such a thera-peutic effect may be provided by differ-ent classes of molecules, includingtrophic factors, antiinflammatory cy-tokines and immunomodulatorychemokines released from transplantedcells (28).

Four main results emerged from thepresent study: (a) SkmSCs spread in thewhole ventricular system and survive in-side the ventricles for a long time, andthis survival is more prolonged in Wrmice than in controls; (b) in both Wr andhealthy mice, transplanted SkmSCs donot significantly integrate in the brainparenchyma; (c) despite the lack of evi-dence of SkmSC migration to the affectedWr cervical spinal cord parenchyma, celltransplantation significantly improvedmotor activity; (d) the therapeutic effectsof SkmSCs, which persisted until the endof the behavioral observation, are not re-lated to a marked motor neuroprotectionbut are likely associated with an im-munomodulatory effect.

The efficacy of the intracerebroventric-ular administration of stem cells in themurine models of neurodegenerative dis-ease has been documented (29). The fluxof cerebrospinal fluid in the different dis-tricts of the CNS represents one of themain advantages of intracerebroventricu-

4 0 8 | C A N Z I E T A L . | M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2

S k m S C I M M U N O M O D U L A T O R Y E F F E C T I N T H E A L S M O D E L

Figure 5. Neuropathology of human SkmSC-transplanted Wr mice: astrogliosis and mi-croglial activation. Immunostaining for GFAP (red) and CD11b (green) at the cervicalspinal cord level of both 7-wk-old Wr mice treated with SkmSCs (B–B′′′) and age-matchedPBS-treated Wr (A–A′′′) are shown. All nuclei were counterstained with DAPI (blue). Scale =100 μm.

Table 1. Gene and protein expression profile of human SkmSC-transplanted Wr CNS.

Human antiinflammatory cytokines and receptors

Gene Fold difference P

Decreased gene expressionCCL2 –5.5996 0.000189CXCL2 –255.7635 0.004607CXCL6 –447.3727 0.003429IL-1R1 –563.6543 0.00016

Increased gene expressionIL-10 4812.82 0.000001IL-10RA 950.582 0.015IL-13 21658.8 0.015

Bio-Plex cytokine assay

Average Protein fluorescence intensity SD P

IL-13 9.82 0.606 0.038611IL-10 0.56 0.737 0.18

Gene expression profile: Results are expressed in fold change signal between in vitroSkmSCs, vehicle-treated Wr CNS and SkmSC-transplanted Wr CNS. Values are expressed asmean ± SD. P < 0.05 for human SkmSC-transplanted Wr CNS versus human SkmSCs culturedin vitro. Protein expression profile: Results are expressed in average fluorescence intensitysignal between SkmSC-transplanted Wr CNS and vehicle-treated Wr CNS. Values areexpressed as mean ± SD. P < 0.05 for human SkmSC-transplanted Wr CNS versus vehicle-treated Wr CNS.

lar transplantation. On the other hand,this approach requires a detailed investi-gation on the interaction between trans-planted cells and host tissue. The inter-nalization of SPIO has already beendemonstrated to be an efficient, reliable,safe and long-lasting procedure to visu-alize stem cells at different times aftertransplantation in the CNS of rodentsand to follow their migration or disap-pearance. In the present study, MRI anal-ysis, weekly performed in both Wr andhealthy littermates after transplantation,showed that human SkmSCs rapidlyspread along the ventricular system(third ventricle, body and posteriorhorns of lateral ventricles, cisternamagna, fourth ventricle, meningeal lay-ers) in both groups (see Figure 3A). Thelongitudinal observation revealed anoverall decrease of the signal for bothgroups (see Figure 3A). However, in thebrain of Wr mice, the decay was delayedand lower compared with healthy mice(see Figure 3A). The reason why trans-planted SPIO-labeled SkmSCs remain indiseased mice for a longer time is farfrom clear. The hypothesis is thatchemoattractant factors, released frominjured neurons or activated glial cells,may somehow favor the permanence oftransplanted cells. However, this intrigu-ing hypothesis seems to be in part miti-gated by the evidence of the almost com-plete lack of a selective migration ofSkmSCs in the affected ventral horns ofcervical spinal cord in Wr mice. To local-ize the engrafted human SkmSCs with ahigher level of resolution, the vital nu-clear intercalant (Hoechst 33258) wasused to label the cells before the trans-plantation in the Wr brain ventricles. Inboth Wr and healthy littermates, a rele-vant amount of Hoechst 33258–positivecells were found to be confined to theventricular system. A large number ofHoechst 33258 fluorescent nuclei wasfound at the cervical region 1 wk afterSkmSC transplantation. The nuclei weremainly localized close to the dorsal whitematter or attached to the meningeallayer. Their presence in cervical spinalcord sections of Wr mice until the sixth

week after transplantation may favor apossible cross-talk between SkmSCs andaffected Wr regions. The overlapping be-tween the SPIO-related signal, MRIslices, the presence of Hoechst 33258 nu-clei and histology confirmed the almostexclusive localization of SkmSCs insidethe ventricle layers.

Our in vitro studies have shown thatSkmSCs express some important markersof neural lineage (see Figures 1C, D).However, the lack of apparent cell migra-tion to the side of injury (ventral horn ofthe Wr cervical spinal cord) and the ab-sence of their differentiation towardmotor neurons did not allow us to specu-late that there is a cell replacement bySkmSCs. Whereas other MSC transplan-tation studies in ALS animal modelsdemonstrate a motor efficacy improve-ment without grafts replacement, wehave chosen to focus our investigationon the evaluation of the SkmSC pharma-cological effects (9).

Here we first report the reliability ofthis dual modality of fluo-paramagneticlabeling of stem cells for in vivo and exvivo tracking. Although the loading ofSPIO and Hoechst 33258 is not reportedto be associated with mechanisms of cy-totoxicity, the transplantation of SkmSCsin Wr mice for clinical estimation wasperformed by avoiding their prelabelingwith exogenous tracers. We opted forthis approach for two main reasons: apractical reason (we suppose that thepossible future application of cell ther-apy in humans will not consider labeledcells) and an experimental reason (wepreviously demonstrated the reliabilityof ex vivo tracking with antibodies di-rected against endogenous markers). Itwas recently reported that the recruit-ment of a relatively low number of ani-mals is considered one of the main fac-tors responsible for the “lost intranslation” between the results obtainedfrom trials in the SOD1 mouse and inhuman ALS (30). Generally, to avoid therejection of human cells, cyclosporine isgiven from the day of transplantation tothe end of symptoms. However, the inhi-bition of immune response by cy-

closporine A may play a therapeutic roleby acting as an antiinflammatory agentor by reducing mitochondrial impair-ment and cell death in neurons (31). Be-cause of the low immunogenicity of thefetal SkmSCs, the present study was car-ried out in the absence of immunomodu-lant agents so we could avoid any possi-ble bias due to interaction betweencyclosporine A and SkmSC transplant effects.

In our study of SkmSC tracking, whichlasted for over 3 months, after cell trans-plantation for both Wr mice and healthycontrols, the lack of cyclosporine A didnot lead to acute toxicity and/or highermortality compared with untreated littermates.

In our study, Wr mice receiving Skm-SCs rapidly and significantly showed re-duced progression of foreleg atrophyand, therefore, as a consequence, suf-fered less loss of forelimb muscularstrength. The transplantation of SkmSCsseemed to produce a faster and strongeroutcome in comparison with our previ-ous results obtained by intraperitoneallyinjecting Riluzole in Wr mice (32).

On the other hand, in contrast to theresults obtained by Riluzole treatment,SkmSC transplantation did not signifi-cantly preserve the rate of motor neuronloss at the end stage of our study (seeFigures 4F, H).

CONCLUSIONThe beneficial effect emerging from

this study, therefore, may be explainedby the production and secretion of a va-riety of growth factors, cytokines andchemokines that could play a beneficialrole in maintaining a functionality in thesurvived motorneurons.

The link between the immunomodulantand therapeutic effects of MSCs on ALSanimal models has already been hypothe-sized (9). In particular, the analysis of thelevels of cytokines measured in vehicle-treated ALS mice and in ALS mice receiv-ing cells suggested that transplantationmay favor the maintenance of Th-2 in-nate adaptive immune response by delay-ing the switch to the Th1-mediated pro -

R E S E A R C H A R T I C L E

M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2 | C A N Z I E T A L . | 4 0 9

inflammatory response and therefore at-tenuating the clinical progression (9).

The role of proinflammatory cytokinesin exacerbating pathological and clinicalfeatures of ALS has already been de-scribed in patients, animals and neuronalcultures (32). Here, we first report the ex-pression levels of a panel of human spe-cific anti–cytokines-chemokines, ex-pressed by SkmSCs 2 wks afterintracerebroventricular transplantation inlateral ventricles of Wr mice.

The RT and protein profiler analysis ofhuman SkmSC-transplanted Wr in brainrevealed that several factors involved inthe antiinflammatory response, such ashuman IL-10 and its receptor IL-10RA(IL-10 receptor α) (33), are highly ex-pressed (Table 1) and that their expressionis increased in vivo compared with in vitro.The IL-10 gene encodes for the antiinflam-matory cytokine of type 2 helper T cells(34). IL-10 is a potent inhibitor of inflam-mation because of its ability to reduceT-cell activity and subsequent inflamma-tory cytokine production. In the past 10years, the beneficial effects of IL-10 wasobserved in several neuroinflammatorydisease models (35). Interestingly, an in-crease of IL-10 and a shift from Th1 toTh2 response has been similarly describedin SOD1G93A mice receiving human um-bilical cord blood (hUCB) cellsand show-ing a therapeutic benefit. IL-13 is also in-volved in the antiinflammatory processby means of similar mechanisms de-scribed above for IL-10 (33).

In the brainstem and spinal cord ofboth patients with ALS and Wr mousemodels, increasingly, many activated mi-croglia and astrocytes, IgG and T lym-phocytes may be capable of secreting thenumerous cytokines that play key rolesin the inflammatory reaction (32). The increase of macrophage-cytokines suchas IL-1α, IL-1β and IL-1RA and the up-regulation of the tumor necrosis factor(TNF) gene make them likely candidatesfor inflammatory effectors (32). IL-10 andIL-13 inhibit the production of severalproinflammatory cytokines includingTNF and IL-1β in monocytes andmacrophages (33). It is therefore impor-

tant to underline that, in Wr mice, thelevels of TNF-α at the cervical spinalcord region rapidly increase at the earlysymptomatic phase and are accompaniedby a parallel increase of two importantdownstream stress kinases (p-JNK and p-P38) having potential neurotoxicity(36). It has also been reported that thechronic treatment with a soluble receptorfor the tumor necrosis factor receptor 1(TNFR1) (Onercept), which had alreadyshown antiinflammatory activity in pe-ripheral districts, drastically reduced thesymptom progression in Wr mice (36).

The main indication emerging fromour study is that SkmSCs have a by-stander effect by releasing some impor-tant antiinflammatory cytokines far fromthe pathological areas. Thus, we hypoth-esize that a deeper comprehension of themechanisms involved in the beneficial ef-fects observed after human SkmSC treat-ment could open a new prospective onthe formulation of a practical, safe andpotential therapeutic strategy for patientswith ALS.

ACKNOWLEDGMENTSWe are grateful to Andrea Smith for

the helpful review of the English in themanuscript. We also thank GiulioAlessandri and Giorgio Battaglia for crit-ically reading this scientific article. Thiswork was supported by IRCCS Founda-tion Neurological Institute “C. Besta”(LR8) and by the Italian Ministry ofHealth (RF-INN-2007-644440).

DISCLOSUREThe authors declare that they have no

competing interests as defined by Molec-ular Medicine, or other interests thatmight be perceived to influence the re-sults and discussion reported in thispaper.

REFERENCES1. Boillee S, Cleveland DW. (2008) Revisiting oxida-

tive damage in ALS: microglia, Nox, and mutantSOD1. J. Clin. Invest. 118:474–8.

2. Wokke J. (1996) Riluzole. Lancet. 348:795–9.3. Lodi D, Iannitti T, Palmieri B. Stem cells in clini-

cal practice: applications and warnings. J. Exp.Clin. Cancer Res. 30:9.

4. Justesen J, Stenderup K, Eriksen EF, Kassem M.(2002) Maintenance of osteoblastic andadipocytic differentiation potential with age andosteoporosis in human marrow stromal cell cul-tures. Calcif. Tissue Int. 71:36–44.

5. Jiang Y, et al. (2002) Pluripotency of mesenchy-mal stem cells derived from adult marrow. Na-ture. 418:41–9.

6. Sanchez-Ramos J, et al. (2000) Adult bone mar-row stromal cells differentiate into neural cells invitro. Exp. Neurol. 164:247–56.

7. Alessandri G, et al. (2004) Isolation and culture ofhuman muscle-derived stem cells able to differ-entiate into myogenic and neurogenic cell line-ages. Lancet. 364:1872–83.

8. Gotherstrom C, et al. (2004) Immunologic proper-ties of human fetal mesenchymal stem cells. Am.J. Obstet. Gynecol. 190:239–45.

9. Garbuzova-Davis S, et al. (2008) Human umbilicalcord blood treatment in a mouse model of ALS:optimization of cell dose. PLoS One. 3:e2494.

10. Falconer DS. (1956). Mouse News Letter. 15:23.11. Schmitt-John T, et al. (2005) Mutation of Vps54

causes motor neuron disease and defectivespermiogenesis in the wobbler mouse. Nat.Genet. 37:1213–5.

12. Fumagalli E, Bigini P, Barbera S, De Paola M,Mennini T. (2006) Riluzole, unlike the AMPA an-tagonist RPR119990, reduces motor impairmentand partially prevents motoneuron death in thewobbler mouse, a model of neurodegenerativedisease. Exp. Neurol. 198:114–28.

13. Palmisano R, et al. (2011) Endosomal accumula-tion of APP in wobbler motor neurons reflectsimpaired vesicle trafficking: implications forhuman motor neuron disease. BMC Neurosci.12:24.

14. Wada MR, Inagawa-Ogashiwa M, Shimizu S, Yasumoto S, Hashimoto N. (2002) Generation ofdifferent fates from multipotent muscle stemcells. Development. 129:2987–95.

15. Neri M, et al. (2008) Efficient in vitro labeling ofhuman neural precursor cells with superpara-magnetic iron oxide particles: relevance for invivo cell tracking. Stem Cells. 26:505–16.

16. D.L. no. 116, G.U. suppl. 40, 18 February 1992.17. Circolare no. 8, G.U., 14 Luglio 1994.18. EEC Council Directive 86/609, OJ L 358, 1, Dec.

12, 1987.19. Institute of Laboratory Animal Resources; Com-

mission on Life Sciences; National ResearchCouncil. (1996) Guide for the Care and Use of Labo-ratory Animals. Washington (DC): National Acad-emy Press. Available from: http://www. nap.edu/openbook.php?record_id=5140

20. Rathke-Hartlieb S, Schmidt VC, Jockusch H,Schmitt-John T, Bartsch JW. (1999) Spatiotempo-ral progression of neurodegeneration and glia ac-tivation in the wobbler neuropathy of the mouse.Neuroreport. 10:3411–6.

21. Mennini T, et al. (2006) Nonhematopoietic erythro-poietin derivatives prevent motoneuron degenera-tion in vitro and in vivo. Mol. Med. 12:153–60.

4 1 0 | C A N Z I E T A L . | M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2

S k m S C I M M U N O M O D U L A T O R Y E F F E C T I N T H E A L S M O D E L

22. Bigini P, Mennini T. (2004) Immunohistochemicallocalization of TNFalpha and its receptors in therodent central nervous system. Methods Mol.Med. 98:73–80.

23. Bigini P, et al. (2010) Neuropathologic and bio-chemical changes during disease progression inliver X receptor beta-/- mice, a model of adultneuron disease. J. Neuropathol. Exp. Neurol.69:593–605.

24. Dominici M, et al. (2006) Minimal criteria fordefining multipotent mesenchymal stromal cells.The International Society for Cellular Therapyposition statement. Cytotherapy. 8:315–7.

25. Yin AH, et al. (1997) AC133, a novel marker forhuman hematopoietic stem and progenitor cells.Blood. 90:5002–12.

26. Ogawa Y, et al. (2002) Transplantation of in vitro-expanded fetal neural progenitor cells results inneurogenesis and functional recovery after spinalcord contusion injury in adult rats. J. Neurosci.Res 69:925–33.

27. Cristofanilli M, Harris VK, Zigelbaum A, et al.(2011) Mesenchymal stem cells enhance the en-graftment and myelinating ability of allogeneicoligodendrocyte progenitors in dysmyelinatedmice. Stem Cells Dev. 20:2065–76.

28. Mazzini L, et al. (2010) Mesenchymal stem celltransplantation in amyotrophic lateral sclerosis: aphase I clinical trial. Exp. Neurol. 223:229–37.

29. Rehni AK, Singh N, Jaggi AS, Singh M. (2007)Amniotic fluid derived stem cells amelioratefocal cerebral ischaemia-reperfusion injury in-duced behavioural deficits in mice. Behav. BrainRes. 183:95–100.

30. Benatar M. (2007) Lost in translation: treatmenttrials in the SOD1 mouse and in human ALS.Neurobiol. Dis. 26:1–13.

31. Keep M, Elmer E, Fong KS, Csiszar K. (2001) In-trathecal cyclosporin prolongs survival of late-stage ALS mice. Brain Res. 894:327–31.

32. Mosley RL, Gendelman HE. Control of neuroin-flammation as a therapeutic strategy for amy-otrophic lateral sclerosis and other neurodegen-erative disorders. Exp. Neurol. 222:1–5.

33. Opal SM, Keith JC, Palardy JE, Parejo N. (2000)Recombinant human interleukin-11 has anti- inflammatory actions yet does not exacerbatesystemic Listeria infection. J. Infect. Dis.181:754–6.

34. Hoffmann KF, Cheever AW, Wynn TA. (2000)IL-10 and the dangers of immune polarization:excessive type 1 and type 2 cytokine responsesinduce distinct forms of lethal immunopathologyin murine schistosomiasis. J. Immunol.164:6406–16.

35. Qiu F, et al. (2006) [Effect of ligustrazine on cellproliferation in subventricular zone of lateralcerebral ventricle after adult rat suffering fromfocal cerebral ischemia]. Sichuan Da Xue Xue BaoYi Xue Ban 37:726–9, 780.

36. Bigini P, et al. (2007) Lack of caspase-dependentapoptosis in spinal motor neurons of the wobblermouse. Neurosci. Lett. 426:106–10.

R E S E A R C H A R T I C L E

M O L M E D 1 8 : 4 0 1 - 4 1 1 , 2 0 1 2 | C A N Z I E T A L . | 4 1 1