comparative effects between bone marrow and mesenchymal stem cell transplantation in gdnf expression...

Comparative Effects between Bone Marrowand Mesenchymal Stem Cell Transplantation in GDNFExpression and Motor Function Recovery in a MotorneuronDegenerative Mouse Model

Diego Pastor & Mari Carmen Viso-León &

Jonathan Jones & Jesus Jaramillo-Merchán &

Juan José Toledo-Aral & Jose M. Moraleda &

Salvador Martínez

Published online: 30 June 2011# Springer Science+Business Media, LLC 2011

Abstract Motorneuron degenerative diseases, such asamyotrophic lateral sclerosis (ALS), are characterized bythe progressive and rapid loss of motor neurons in the brainand spinal cord, leading to paralysis and death. GDNF(glial cell line derived neurotrophic factor) has beenpreviously shown to be capable of protecting motor-neurons in ALS animal models although its delivery tothe spinal cord after systemic administration is blocked bythe blood brain barrier. Thus, it is necessary to develop newneurotrophic approaches to protect these motor neuronsfrom death. Bone marrow-derived stem cells have beenshown to be capable of improving a large variety of

neurodegenerative disorders through neurotrophic mediatedmechanisms. Here we analyzed the effect of transplantingwhole bone marrow or cultured mesenchymal stem cellsinto the spinal cord of a motor neuron degenerative mousemodel. Motor functions were analyzed using variousbehavior tests for several weeks after transplantation. Weobserved that bone marrow, and to a lesser degreemesenchymal stem cell, treated mice improved significantlyin the motor tests performed, coinciding with a higherGDNF immunoreactivity in the grafted spinal cord. Inseveral cases, the treated spinal cords were extracted, theengrafted bone marrow cells isolated and cultured, andfinally re-transplanted into the spleen of immunodeficientmice. Re-grafted cells were detected in the host spleen,bloodstream and bone marrow, demonstrating a phenotypicstability. Thus, bone marrow cells do not suffer significantphenotypic modifications and is an efficient procedure toameliorate motor-neuron degeneration, making it a possibletherapeutic approach.

Introduction

There are many motor neuron degenerative diseases, beingamyotrophic lateral sclerosis (ALS) one of the most wellknown. This disorder, also known as Lou Gehrig’s disease,mainly affects the spinal cord, brainstem and motor cortex[1]. The loss of motor neurons and their muscle synapseseventually leads to spasticity, hyperreflexia (when uppermotor neurons are exclusively affected), generalized weak-ness, muscle atrophy, dysphagia, and premature death due

D. PastorSports Research Center, University Miguel Hernández,Elche, Spain

M. C. Viso-León : J. Jones : J. Jaramillo-Merchán :S. Martínez (*)Neuroscience Institute UMH-CSIC,University Miguel Hernández,San Juan, Alicante, Spaine-mail: [email protected]

J. J. Toledo-AralBiomedicine Institute of Seville,University Hospital Virgen del Rocío/CSIC/University of Seville,Seville, Spain

J. M. MoraledaHospital Universitario Virgen de la Arrixaca,Universidad deMurcia, Hematology, BMTandCellular therapyUnit,Crta. Madrid Cartagena,30107, El Palmar, Murcia, Spaine-mail: [email protected]

Stem Cell Rev and Rep (2012) 8:445–458DOI 10.1007/s12015-011-9295-x

to respiratory failure [2, 3]. The cause of this degeneration isunknown in the majority of cases, and the clinical course ishighly variable, suggesting that multiple factors are involvedin the development of the disease, although severalhypotheses have been proposed, including mitochondrialdysfunction, protein aggregate formation, excitotoxicity,axonal transport malfunction, mutant-derived oxidativedamage, lack of growth factors, and inflammation [4].

As ALS, other related disorders such as spinalmuscular atrophy (SMA) and progressive muscularatrophy (PMA) present motor neuron degeneration [5,6]. Many authors have studied the use of stem cells to treatSMA [7] or ALS [8], by transplanting bone marrow-derivedstem cells in transgenic animal models, in order to try toameliorate the symptoms or delay the onset of the disease,with varied results [8–10].

Recent data has shown that transplanting bone marrowcells from healthy animals increases motor neuronsurvival [11]. In our lab, we have previously shown thathematopoietic stem cells, when transplanted into thelumbar regions of the spinal cord, are capable of protectingmotor neurons from entering apoptosis due to a trophic effectby the expression of GDNF [12]. This neurotrophic factor,known to improve neuronal survival, is not capable ofpassing through the blood-brain barrier [13], thus itssystemic administration is severely limited. Other methodsfor GDNF incorporation, such as intraventricular adminis-tration, could have significant collateral effects, affectingdifferent aspects of the blood-brain barrier’s function [14].

Several authors have shown that bone marrow cells cannot only ameliorate the symptoms but also support striatedmuscle regeneration [11] as well as improve electromyo-graphic signalling [12]. Also, mesenchymal stem cells arecapable of reducing the reactive gliosis and microgliaactivation that is typical in rodent models of ALS, resultingin improved motor skills as detected through variousbehavior tests [15]. Unfortunately, the methods by whichthese cell transplants avoid motor neuron death are not wellunderstood. Thus, the objectives in this work are to analyzeif whole bone marrow transplants or pure mesenchymalstem cells are capable of improving the motor functions ofan ALS mouse model, and characterize the process bywhich these cells induce this improvement.

Materials and Methods

Animals

All animal experiments have been performed in compliancewith the Spanish and European Union laws on animal carein experimentation (Council Directive 86/609/EEC), andhave been analyzed and approved by the Animal

Experimentation Committee of our university. Mice withthe muscle deficient/osteocondrodystrophy mutation (mdf/ocd) were obtained from the Pasteur Institute (Paris,France), as a gift from J.L. Guenet. These mice sufferfrom a currently unidentified autosomal-recessive spontaneousmutation, located on chromosome 19. Homozygous mdfmutant mice begin to present motor dysfunctions in theirhindlimbs at 3 weeks of age, which progresses to hindlimbparalysis and forelimb weakness [16]. Our mdf/ocd modelwas first described by Poirier [17] and has recently beenshown to be affected by a loss of function in Scyl1 [18], thusthe mouse model presents, in addition to motor neurondegeneration, cerebellar atrophy, purkinje cell loss and opticnerve atrophy. In our animal facilities, we have seen that theypresent a lifespan similar to wild-type mice. For the experi-ments, a total of 14 animals were used for post-graft spinalcord cultures (10 symptomatic and 4 non-symptomatic), 20mice for ELISA measurements and 21 mice for immunohis-tochemistry and behavior tests. As a control for ELISAmeasurements we used 6 C57BL/6 mice, the originalbackground for the mdf mutation [16].

The green fluorescent protein (GFP) mutant mice(C57Bl/6-Tg(ACTB-EGFP)1Osb/J) used were bred in ouranimal facilities. These mice ubiquitously express the GFPprotein under the control of the beta-actin promoter [19],which facilitates their detection when transplanted into ahost that does not present the GFP protein. Immunodefi-cient mice (Athymic Nude-Foxn1nu) were purchased fromHarlan Laboratories, Switzerland. GDNF/LacZ knockoutmice were bred in our animal facilities. Homozygous micedie at birth, while heterozygotes are viable and used tomaintain the colony [20].

Isolation of Bone Marrow Cells

Femurs were dissected from 6- to 8-week old GFP+ mice,sacrificed by cervical dislocation. Bone marrow wasextracted, and single-cell suspensions were obtained bymechanical dissociation. The cells were then counted andresuspended in standard basal medium (D-MEM, Invitrogen)to obtain a concentration of 1×106 cells in 5 μl for thetransplantation of whole bone marrow cells (BMC).

Isolation and Culture of BoneMarrow-Derived MesenchymalStem Cells

Unless stated otherwise, all the materials and substanceswere purchased from Sigma-Aldrich. Bone marrow wasextracted as previously commented, obtaining a cellsuspension. The cells were washed and centrifuged, andthe pellet was then resuspended in D-MEM (Invitrogen)supplemented with 15% FBS (Biochrom AG, Berlin),100 U/ml penicillin/streptomycin, 1 mM sodium pyru-

446 Stem Cell Rev and Rep (2012) 8:445–458

vate, and 1% non-essential amino acids. These cellswere placed in culture flasks and the plastic-adherentpopulation was isolated and allowed to proliferate for3–4 weeks, changing the media twice a week. Approx-imately every 7 days, the cells were tripsinized withtripsine/EDTA 0.25% and replated. After 3–4 weeks, allthe cells in culture expressed CD90 and CD44 and werenegative for CD34 and CD45, confirming that the cellpopulation was mainly composed of mesenchymal stemcells (MSC, data not shown) [21].

Isolation of Bone Marrow and Spleen Cells from GDNFKnockout Mice

Pregnant heterozygous GDNF knockout mice were sacrificed20 dpc and fetuses separated. The fetuses that did not presentkidneys, indicating that they are homozygous mutants, wereselected [20]. Femur bone and spleen were extracted,minced, centrifuged and resuspended in DMEM fortransplantation. Control fetuses were isolated in parallelto demonstrate that newborn GFP bone marrow cellsproduce the same effects as adult bone marrow.

Behavior Assays

A total of 3 tests were performed on the mdf mutant mice:footprint analysis to study stride length, accelerating rota-rodanalysis, and maximum speed in a treadmill. These tests wereperformed on a weekly basis for 3 weeks before thetransplantation process. After the transplant, the mice wereallowed to recover for 1 week, and then submitted again to theweekly behavior tests during at least another 3 weeks. Micewhere only standard culture medium was injected were usedas controls, which were submitted to the same tests as thoseperformed by the experimental mice.

Footprint (FP)

The hindlimb paws of the mice were painted with non-toxic,washable paints purchased from an ordinary paint shop. Then,they were placed in an opaque methacrylate tube with a black-colored cardboard box attached to one of its ends. The tubeand box were placed over a 60 cm long by 10 cmwide strip ofpaper, where the mouse was expected to walk towards thedarkness of the cardboard box, leaving its footprints on thepaper. In this manner, the stride length could bemeasured, as ithas been previously shown that mice with affected motorfunctions have a decreased stride length [22].

Rota-Rod (RR)

The rota-rod test was performed on an 8500 Rota-rod(Leica Scientific Instruments, Barcelona, Spain). The lane

is 500 mm wide, and the rod has a diameter of 30 mm.Before analysis, the mice were trained on the rod for 4 daysuntil they presented an adequate walking pace. Afterwards,the mice were tested weekly, placing each mouse 8 times onthe accelerating rota-rod cylinder, which uniformly increasedspeed from 4 to 40 rpm over a 5 min time span, until themouse fell. Both the time of fall and speed reached wererecorded and averaged per session [12, 23].

Treadmill (TM)

An LE 8700 treadmill model was used (Leica ScientificInstrument). The lane is 41 mm long and 10 mm wide. Therunway, when in movement, pushes the animal to the shockgrid, set at 0.4 mA. In this manner, it was possible tomeasure the maximum possible speed each mouse mayattain. The treadmill accelerated at a constant rate until themouse could not maintain the speed, and the maximumspeed reached was noted. Before analysis the mice weretrained for 2 days with the apparatus. Then, each mousewas placed 8 times in the treadmill per session and theaverage maximum speed attained was calculated. Otherauthors have analyzed the maximum speed attained inneurodegenerative animal models, however this is the firstreport to use the treadmill for this end [24].

Cell Transplantation

The mice were injected intraperitoneally with 0.1 mg ofbuprenorphine (Buprex, Schering-Plough, Madrid, Spain)per kilogram of mouse previous to the surgical intervention.Then, they were anesthetized with isofluorane gas (EsteveVeterinary, Milan, Italy) and placed in a stereotaxicapparatus (Stoelting, Wheat Lane Wood Dale, Indiana). Alaminectomy was performed at the L5–S1 level, and thecells were injected under a surgical microscope with aHamilton syringe (Hamilton, Reno, Nevada) into theanterior horn of the spinal cord through the posteriorfuniculus at 0.2 mm of the dorsal sulcus. Depending on thecase, either one million bone marrow-derived cells (BMC)(either from GFP or GDNF-KO mice), 500,000 mesenchymalstem cells (MSC), or fresh culture medium were injected, allin a maximum volume of 5 μl.

After the injection, the incision was sutured and the micewere monitored until they recovered from the anaesthesiaand during the following days in order to ensure that theoperation did not cause any additional motor damage, suchas paralysis or tremors in the hindlimbs.

Spinal Cord Culture

In order to analyze the viability and in vitro phenotype ofthe transplanted cells, BMC [6] and MSC [4]-treated mice

Stem Cell Rev and Rep (2012) 8:445–458 447

were sacrificed and their spinal cords removed for cellculture. Specifically, sections L4 to S2 were extracted anddissociated mechanically and enzymatically with 3 mg/mlcollagenase I during 15 min at 37°C. The resulting cellsuspension was washed, centrifuged and resuspended inMEM medium supplemented with 10% Calf serum, 2 mMglutamine, 100 U/ml penicillin/streptomycin, 1 mM sodiumpyruvate and 1% non-essential amino acids.

Transplantation of Previously Grafted Bone Marrow Cellsinto Nude Mice

For this experiment, the spinal cord cultures isolated fromBMC-treated mice were used. The cells were unattached withtripsine/EDTA 0.25%, washed and resuspended in 50 μlMEM, which was injected into the spleen of immunodeficientnude mice (Athymic Nude-Foxn1nu, Harlan Laboratories,Switzerland). The surgical procedure was similar to thespinal cord transplants, only in this case the mouse wasplaced over its back to reach the spleen. After 15 days, thenude mice were sacrificed, blood extracted and its bonemarrow isolated and cultured as previously commented. Thespleen, liver, lung, and brain were fixed in 4% paraformal-dehyde for 24 h before their histological analysis.

Tissue Preparation and Immunohistochemistry

Three to four weeks after the injection/grafting, the micewere anesthetized with isofluorane and fixed by intra-cardiac perfusion with 4% paraformaldehyde (PFA) inphosphate buffer (pH 7.4). The spinal cord was carefullyexcised and kept in 4%PFA overnight. After fixation, the spinalcords were cryoprotected by increasing concentrations ofsucrose and finally embedded in Tissue-Tek (Electron Micros-copy Science, Hatfield, PA) and frozen. Transverse sections of20 μm were obtained with a cryostat and mounted on slides.The sections were incubated for 60 min at room temperature in0.1% lysine in PBS to block non-specific binding. Thefollowing primary antibodies were used: anti-GFP (rabbit ormouse, 1:200, Molecular Probes, Eugene, OR), anti-GDNF(rabbit, 1:200, Santa Cruz Biotechnology, Santa Cruz, CA),anti-CD34 (rat, 1:250, eBioscience, San Diego, CA), anti-CD44 (rat, 1:250, BD Biosciences Pharmingen, San DiegoCA), anti-CD45 (rat, 1:250, BD Biosciences Pharmingen),anti-CD90 (rat, 1:250, BD Biosciences Pharmingen), Tuj-1(mouse, 1:1000, Covance, Berkeley, CA), anti-active caspase-3(rabbit, 1:500, BD Pharmingen), and beta-galactosidase anti-body (chicken, 1:400, Abcam). The sections were incubatedovernight at room temperature with the primary antibody.Afterwards, in all cases except for GFP staining, biotinylatedsecondary antibodies (1:200, Vector Laboratories, Burming-ham, California) or rabbit polyclonal secondary antibody tochicken IgY (1:200, Abcam) were used, followed by an

incubation with streptavidin conjugated either with Cy3(1:500, GE Healthcare, Buckinghamshire, UK) or Alexa Fluor488 in the case of GFP (mouse or rabbit, 1:500, MolecularProbes). Histological samples were observed in a fluorescencemicroscope (Leica DMR, Leica Microsystems) and micro-graphs taken with a confocal microscope (Leica DMR).

Enzyme-Linked Immunosorbent Assay (ELISA)

A group of 26 animals was used to study the possibleneuroprotective effects of glial cell line derived neurotrophicfactor (GDNF) levels in the grafted spinal cord area. Seven daysafter surgery, the mice were anesthetized with an overdose ofsodium pentobarbital (100 mg/kg) and decapitated. Differentgroups of animals were used: wild-type mice C57 (n=6), non-treated mdf/ocd clinically affected animals (n=6), BMCtransplanted mdf/ocd affected animals (n=7) and MSC trans-planted mdf/ocd affected animals (n=7). The lumbosacral (L4-S1 segments) enlargement of the spinal cords was removedwithin 3–5 min after decapitation. Tissue was stored at −80°C.All samples were used within 15 days after freezing. The tissuelevels of GDNF were measured with an ELISA kit (GDNFEmax ImmunoAssay System; Promega, Madison, WI),according to the protocol of the supplier.

Statistical Analysis

Statistical analysis was performed using SigmaPlot 10.0software. Both paired t-student tests (for behavior assay)and t-student (for protein quantification) were used.

Results

Whole Bone Marrow Transplants Significantly Improvethe Motor Functions of the Mice

A total of 21 symptomatic mdf/ocd mice were used for thisexperiment, placed in three groups, depending on the cellsource transplanted: 9 transplanted with whole bonemarrow (BMC), 6 with mesenchymal stem cells (MSC),and 6 controls where only culture medium was injected(SHAM). The mice were submitted on a weekly basis tothree behavior tests before and after the transplantation:treadmill, rota-rod and footprint (Fig. 1). The degenerationin mutant mice were found to progress in two differentmanners: either presenting acute, fast degeneration detectedby the decreasing scores obtained in the various teststhroughout the pre-grafting period (footprint n=7, rotarodn=6, treadmill n=5) (Fig. 1a represents the rotarodresults), or a slow degeneration (footprint n=10, rotarodn=7, treadmill n=12), obtaining low yet somewhat stablescores (Fig. 1b represents the rotarod results).


The results of the behavior tests varied significantlydepending on the experimental treatment (Fig. 1c). In thecase of the BMCmice, 2 out of 9 improved in one of the threetests, while 3 mice improved in two tests. As for the MSC-treated mice, only 2 of 6 mice improved significantly in one ofthe tests, with no mice improving in more than one test. Noneof the controls improved in any of the behavior tests.

By studying each of the behavior tests independently, wedetected that none of the MSC mice improved in thefootprint test (Fig. 1c). On the other hand, one third of theBMC mice (3 of 9) improved in this test.

As for the rotarod test, symptomatic mdf mice graduallyattained lower values on the accelerating rotarod,corresponding to the degeneration of the hindlimb motorneurons as well as the cerebellar degeneration due to theScyl1 mutation. Since Schmidt et al. [18] reported that at274 days there was still 42% of Purkinje cell remaining inthe mdf mice, and no functional data was presented in theirwork, we think that our early functional phenotype(analyzed at 45–90 postnatal days) was mainly due to

motorneuron degeneration. In the transplanted mice thesignificant improvements were evaluated with the pairedt-Student test. In this case, 22% (2 of 9) of the BMCpresented a significant improvement in the rotarod,compared to 16% of the MSC (1 of 6).

In the treadmill tests, the mice gradually improvedtheir motor performance, obtaining increasingly highervalues. As in the other cases, the non-transplantedsymptomatic mice gradually presented lower values,probably due to their neuronal degeneration. A third (3of 9) of the BMC mice and 16% (1 of 6) of the MSCimproved in this test.

Thus, as it can be seen, the two cellular treatmentshave different results on the mice. Specifically, bonemarrow transplantation significantly improved the motorskills in 55% of the treated mice (33% in more thanone test), compared to 33% of the mesenchymal stemcell-treated mice (with none improving in more than onetest). No MSC-treated mice improved in the footprinttest.

Fig. 1 Results in the behavior tests performed in the treated micecompared to controls. a and b Examples of the evolution of two micein a behavior test (rotarod) before and after whole bone marrowtransplantation. In (a), the mouse presented a rapid degeneration,indicated by the increasingly lower scores each week beforetransplantation. In (b), the mouse had a slow degeneration, where itwould maintain low scores throughout the pre-transplantation period.The dark blue bars indicate the results obtained after being submitted

to the treatment. Y axis indicates the time in seconds the mouse canwalk in the rotarod at increasing speeds (4–40 rpm) for each weekanalyzed. c Histogram depicting the number of mice that improvedsignificantly in each of the behavior tests. TM represents treadmill,RR is rotarod, FP footprint. BMC is whole bone marrow transplan-tation, MSC is mesenchymal stem cells, and Sham are mice treatedwith culture medium


Fig. 2 Immunohistochemical analysis of the treated mice. a Diagramsdepicting the location of the stem cell injection (red arrow) in thespinal cord. In the left image, D is dorsal, V is ventral, E is ependyma,DH is dorsal horn, VH is ventral horn. The green circles represent theBMCs and motor neurons are depicted in blue. Right image depictsthe location of the stem cell injection (red arrow) in the spinal cord. Inthis image, D is dorsal, V is ventral, A is anterior, P is posterior. Thegreen circles represent BMCs and the dashed line indicates the areawhere the BMCs were detected, dispersed in approximately a 1 cmarea. (b–i) Images taken of the spinal cords treated with whole bonemarrow cells. b Longitudinal section of the spinal cord (100×magnification). The bone marrow cells can be detected migratingthroughout the spinal cord (total length of the section=1 cm). cMagnification of the image in (B), the grafted cells present an

elongated morphology (400× magnification). d Some cells were alsospherical-shaped and forming small clusters (630× magnification). eClose-up of a grafted cell, with small prolongations (630× magnifi-ciation). (f–i) In the mice with whole bone marrow transplants, it ispossible to detect cells that express CD45 (f), CD34 (g), CD44 (h),and CD90 (i), seen in red. All images are taken at 630× magnificationexcept for (h) (400×). (j–l) Spinal cords of mice treated withmesenchymal stem cells. Cells are seen scattered in small clusters oras individual cells (j) (200× magnification). (k) Some can be seen incontact with blood vessels of the spinal cord (400× magnification). (l)These cells express CD90 (in red) (400× magnification). In all cases,blue is DAPI, green is GFP. Red staining is explained in each of theimages


Bone Marrow Grafted Cells Maintain Their Phenotypeand Do Not Undergo Transdifferentiation

The cells were injected into the anterior horn of the spinalcord, as depicted in Fig. 2a. After the experimental phase,which lasted for 3 weeks, the mice were perfused and spinalcords analyzed by immunohistochemistry. In BMC graftedmice, round-shaped GFP+ cells were detected forming acluster at the injection site, with many dispersed cellssurrounding it. Also, a number of cells could be detected asfar as 0.5 cm away from the central region of the graftedarea in both directions: rostrally and caudally (Fig. 2b). Thegrafted cells were either spherical-shaped or presented shortprolongations (Fig. 2c–e). The majority (90%) of the cellsexpressed CD45 and CD34 (hematopoietic markers), and afew (10% approximately) were CD44 or CD90 positive(mesenchymal markers) (Fig. 2f–i, respectively). Eightypercent of the BMC transplanted mice presented GFP+

cells; however, only 30% of the MSC grafts presented anequivalent distribution of live GFP+ cells as those foundin the BMC grafts, in the rest of cases the cells appearedlargely disseminated and in very low numbers. In thiscase, the cells were mainly small, round-shaped, andfound either in clusters or as individual cells surroundingthe cluster (Fig. 2j). Also, in the mice with the mostnumerous GFP+ cells, these were found many times inblood vessels (Fig. 2k), possibly contributing and promot-ing angiogenesis, as it has been previously described [25].These cells maintain the expression of the surface markerCD90 (Fig. 2l). In both cases, the regular morphology ofthe cell body, and the absence of cells with more than onenucleus indicated that there was no sign of transdiffer-entiation or cell fusion.

Increased Levels of GDNF in the BMC Mice

GDNF expression was analyzed in the treated (BMC andMSC) and non-treated mice (Fig. 3). GDNF immunore-activity was detected in the area where BMC cells weretransplanted and its expression decreased when weobserved areas farther away from the transplant (Fig. 3a).As for the MSC-treated mice as well as in the controlgroup, there was no significant expression of this neuro-trophic factor (Fig. 3b–c, respectively). In areas where thegrafted cells were detected near the motorneurons, theseneurons were GDNF positive, which was not detected inother motor neurons that were far from the grafted cells(Fig. 3d and e, respectively; p=0.028).

Quantification of GDNF expression in the grafted areaof the spinal cord was performed using the ELISA test(Fig. 3f). In the BMC grafted group, there was asignificantly higher concentration of GDNF compared tothe non-treated control group (39.5.1±12.8 pg/mg and

27.8±9.7 pg/mg, respectively). In the MSC grafted group,GDNF concentration was similar (40.7±20.9 pg/mg) toBMC grafted mice (39.5.1±12.8 pg/mg), however due toits high variability, it was not significantly higher than theNT group (p=0.094).

Bone Marrow-Derived Cells Maintain ProliferativeCapabilities in the Post-Graft Culture, WhereasMesenchymal Stem Cells Enter Apoptosis

To analyze the viability of the grafted cells and theirphenotypic characteristics, active caspase-3 was studied inthe spinal cord (Fig. 4). In the BMC grafted group, 35% ofthe cells were active caspase-3 positive (Fig. 4a). However,in the MSC group, 80% of the cells expressed activecaspase-3 (Fig. 4b), indicating that the mesenchymal stemcells do not survive well in the spinal cord, and thus enterapoptosis.

Three weeks after the transplantation, 6 BMC and 4MSC mice were sacrificed, and their spinal cord removedfor cell culture. In the MSC graft cultures, no GFP+ cellswere detected (data not shown). In BMC graft cultures,donor cells were detected in all the spinal cord cultures.Several GFP positive cells were found as early as 1 dayafter culture, many of which were in contact with hostcells (Fig. 5a). The majority of the host cells were Tuj1positive, indicating a neuronal phenotype. BMC cellswere also transplanted into non-symptomatic mice, main-tained for 3 weeks and spinal cords removed for cellculture (n=4). In this case, there was 80% more GFP+cells in culture compared to the cultures of BMCsymptomatic mice (n=4, Fig. 5b and c, respectively).We believe this was due to the neurodegenerativeenvironment of the spinal cord in the mutant mice, whichprobably created a different niche that modified in someway the culture properties of grafted cells.

Nevertheless, the majority of the surviving cells inculture were CD45 positive, indicating that they stillmaintained their bone marrow phenotype, and were ofhematopoietic origin (Fig. 5d–f).

Post-graft Cultured Bone Marrow-Derived Cells,When Transplanted into the Spleen of Nude Mice, Enterthe Bloodstream and Migrate to the host’s Bone Marrow

The bone marrow-derived cells that were cultured in thepostgraft protocol were re-transplanted into the spleen ofnude mice. After 3 weeks, the mice were sacrificed andseveral organs were analyzed by immunohistochemistry,including the spleen, lungs, liver, brain as well as the bonemarrow and peripheral blood. GFP positive cells weredetected in the spleen (Fig. 6a), as well as in the bonemarrow of the nude mice (Fig. 6b–c). These cells were


initially round-shaped, but soon adopted an elongatedmorphology. Thus, the cells, once re-transplanted into thespleen, travelled through the bloodstream and populated thehost’s bone marrow, its natural niche.

Donor GDNF Expression is Necessary to Improve MotorBehavior in Mdf Mice

To confirm that donor GDNF expression is necessaryfor motor-neuron survival, we transplanted bone mar-row and spleen cells of GDNF knockout (GDNF KO)mice in the spinal cord of mdf mice (n=4). Since thismouse model dies at p1 [20], it was necessary to isolatecells from the bone marrow and spleen at p0. Ascontrols we used newborn beta-actin GFP mice (n=6).As a result, no significant improvements were detected inthe behavior tests of GDNF KO bone marrow-treatedmice, while the mice treated with GFP-bone marrow

presented similar results as with adult bone marrow(Fig. 7a). The grafted cells of both GDNF KO andGFP cells were detected in the anterior horn of thelumbar region (Fig. 7b and c, respectively). In severalcases, the animals were sacrificed and analyzed for GDNFquantification (GDNF KO n=3, GFP n=4) by ELISA.GDNF KO bone marrow cells did not increase the GDNFcontent of the spinal cords, as opposed to the GFP cells,corroborating the importance of GDNF expression of thedonor cells (Fig. 7d).

Discussion

In this work we have shown that transplantation of wholebone marrow (BMC), and to a lesser degree mesenchymalstem cells (MSC), is capable of ameliorating the symptomsof a motor neuron degenerative mouse model. Over half of

Fig. 3 GDNF expression in spinal cords of treated mice. a Spinalcord of a mouse treated with whole bone marrow, where GDNFexpression (in red) can be detected around the grafted area. b Spinalcord of a mouse treated with mesenchymal stem cells, where little orno GDNF expression was detected. c Spinal cord of a control mousestained with GDNF (in red). d GFP+ cells were detected nearsurviving motor neurons, which expressed GDNF. e Image taken inthe anterior horn of the contralateral side of the same spinal cord as in

(d). In this case, little or no GDNF expression was detected. In (a–c),the images were taken at 100× magnification, whereas (d–e) were at400× magnification. f Histogram depicting GDNF values (in pg/mg)obtained in each 0.5 cm fraction (L4–S1) of the spinal cord of bonemarrow cells (BMC 39.59±4.0), mesenchymal stem cells (MSC 40.15±6.6), non-treated (NT 27.80±2.9), and wild-type (WT 33.57±3.8)treated mice. *P<0.05, Student t-test comparing each condition witheach other (BMC-NT n=0.027)


the BMC grafted mice improved in at least one of thebehavior tests, and 33% improved in two. Interestingly, inthe case of mesenchymal stem cell grafts, one thirdsignificantly improved in one behavior test, with noneimproving in more than one. In our experiments, BMC

presented better results than MSC when comparing themotor function of the treated mice. Also, BMC had a highersurvival rate in the spinal cord, where 80% of the BMCgrafted mice presented donor cells at the moment ofsacrifice compared to 30% of MSC grafted mice, which

Fig. 4 Apoptosis in presence ofthe transplanted cells. Wholebone marrow transplantspresented few active caspase-3positive cells (a), whereasalmost all the mesenchymalstem cells were active caspase-3positive (b). Images were takenat 200×, green staining is GFP,red is active caspase-3, and blueis DAPI


can explain the difference in the results obtained with bothtreatments, as well as the great variability observed inGDNF content of the MSC group.

As opposed to our previous work [12], where hemato-poietic stem cells were isolated and transplanted, here wegrafted whole bone marrow as well as mesenchymal stemcells for several reasons. First of all, bone marrow cells arerelatively easy to obtain, requires minimum manipulation,and can be extracted and rapidly re-implanted into the

region of interest. As for the mesenchymal stem cells, theirneuroprotective effects are well known and can be easilyexpanded in culture. Bone marrow contains hematopoieticstem cells, endothelial cell progenitors, mesenchymal stemcells, as well as other cell types that may differentiate intoimmature or mature nonhematopoietic cells of multipletissues [26], which possibly produce or induce the observedtrophic effect. Interestingly this effect was not observed inthe transplants using mesenchymal stem cells. In the BMC-

Fig. 6 Re-grafted total bone marrow cells in nude mice. a Presence ofcells in the spleen, where they were grafted. b Culture obtained fromnude mice´s bone marrow, where GFP+ cells were detected. Image

taken 1 day after extraction. c Image taken 5 days after extracting thebone marrow, where the cells adopt an elongated morphology. Imageswere taken at 200× magnification, green staining is GFP

Fig. 5 Spinal cord cultures of mice with total bone marrowtransplants and mesenchymal stem cells. a Spinal cord culture of asymptomatic mouse after 1–2 days of culture, where several GFP+cells could be detected growing over donor cells attached to the

culture dish. Many donor cells were Tuj1 positive (in red). b–c Spinalcord culture of non-symptomatic and symptomatic mice after 10 days(b and c, respectively). d–f The majority of the GFP+ cells in cultureexpress CD45 (in red). Images were taken at 200× magnification


treated mice, it is possible that the grafted cells, includingthe mesenchymal stem cells, cooperate by unknown cellcommunity effects and are better protected, resulting in a

higher possibility of survival as detected in the analyzedhistological slices and thus work together to generate theobserved neurotrophic effect. Indeed, mesenchymal stem

Fig. 7 Transplantation of bonemarrow isolated from GDNF-KO mice. a Histogram depictingthe number of mice that im-proved significantly in each ofthe behavior tests. TM repre-sents treadmill, RR is rotarod,FP footprint. P1 are animalsgrafted with newborn GFP+cells, GDNF −/− are grafts withGDNF KO cells, and Sham aremice treated with culture medi-um. b Composition of variousimmunohistochemical images(at 100× magnification) of thespinal cord of mdf mice 4 weeksafter stem cell injection isolatedfrom newborn GDNF KO mice.Blue is DAPI staining and ingreen β-galactosidase staining.In the image, VH is ventralhorn, DH is dorsal horn, E isependyma, D is dorsal, V isventral, R is rostral and C iscaudal. c Composition of images(at 200× magnification) of thespinal cord of mdf mice 4 weeksafter newborn GFP-bonemarrow injection. Blue indicatesDAPI staining and green is GFP.d Histogram comparing GDNFconcentration between differenttreatments. BMC are adult GFPbone marrow treatments (39.59±4.0), MSC are adult GFPmesenchymal stem cells (40.15±6.6), NT indicates non-treatedmice (27.80±2.9), WT iswild-type mice (33.57±3.8),GDNF −/− are animals graftedwith newborn GDNF-KO bonemarrow cells (11.38±2.2), andP1 are newborn GFP bonemarrow transplants (55.56±5.5),p<0.001 between P1 and NT


cells seem to require other bone marrow cell types in orderto survive in the spinal cord of this motor neurondegenerative model.

The majority of the transplanted cells maintained theirspherical morphology as well as their original bone marrowmarkers, although we could easily find cells with severalprolongations that may possibly be contacting adjacentcells. These cell contacts could underlay the production ofneurotrophic factors such as GDNF by these cells. Ourexperiments with GDNF-KO mice show that the graftedcells produce this neurotrophic factor, thereby increasingGDNF expression in the spinal cord, resulting in improvedmotor coordination. Also, our experiments with newbornGFP bone marrow obtained similar results to those seenwith adult bone marrow, indicating that the results obtainedwith GDNF-KO bone marrow is due to the lack of thistrophic factor.

In the BMC-treated mice, the majority of the graftedcells were clustered in the injection site; however cellscould be detected as far as 0.5 mm away from thetransplanted area. Interestingly, many cells were alsodetected surrounding blood vessels. In BMC transplants,the majority of the cells expressed CD45 or CD34,indicating a hematopoietic lineage, although mesenchy-mal markers such as CD44 and CD90 were alsodetected. In both cases (BMC and MSC), the cells didnot express any of the neuronal markers analyzed, thusthere was no indication of cell fusion nor transdiffer-entiation, and no tumor formations were observed, asother authors have shown [27–30]. Mesenchymal stemcells stop proliferating due to contact inhibition [31],which is an important property to consider in cell therapyin order to avoid tumor formation. This phenotypicstability is of great importance, since any externalmodification of the grafted cells, even towards normalcharacterization, would increase the risk of collateraleffects after cellular therapy.

We have used the mdf/ocdmouse model for this experimentbecause we believe it to be an interesting rodent model formotor neuron degeneration. Many authors use SOD1 mutantmice and rats, which possess traits found in only 2–3% of thehuman ALS patients [32, 33]. The striking pathological andclinical similarity between familial and sporadic disease hassparked enthusiasm that the animal models based on mutantSOD1 might provide insight into mechanisms of bothsporadic and familial disease. However, to date, there is nodirect evidence validating this assumption [2]. Mdf is anautosomal-recessive mutation located in mouse chromosome19 [16, 17]. These mice present successive motor neurondeath that result in a posterior waddling at 4–8 weeks of age.This waddling is progressive until reaching hindlimb paral-ysis and forelimb weakness, with the mouse ultimately dying

generally due to respiratory failure. In the skeletal muscles itis possible to observe neurogenic atrophy, as well as axonaldegeneration in the peripheral nerves. In the spinal cord, theneurons degenerate and there is an astrogliosis restricted tothe ventral horn of the spinal cord. These traits make the mdf/ocd mouse a highly appropriate model for the study of motorneuron degeneration, presenting a longer lifespan comparedto SOD1 mutant mice. Recently, this mouse model has beenshown to be affected by a loss of function of the Scyl1 gene[18], and includes cerebellar degeneration. The Scyl1 genemutation in humans produces Scyl1-encoded protein thatinteracts with the survival motor neuron and Purkinje cells, apossible cause for spinal muscular atrophy and spinocerbellarataxia.

There are several methods by which bone marrow cells arecapable of improving the symptoms of diseased models [12,34]. Besides cell fusion and transdifferentiation, bonemarrow cells possess immunomodulatory properties [15],and can release trophic factors. In the latter case, bonemarrow cells produce and secrete a large variety of trophicfactors, such as VEGF, NGF, BDNF, and GDNF, amongothers [21, 35–41]. Several of these trophic factors have beenshown to ameliorate the degeneration of motor neurons invarious animal models [42–45], which has resulted in thedevelopment of various clinical trials for ALS [10, 45]. Ofthese factors, GDNF has been proven to protect thedegenerating motor neurons [46]. This was achieved bytransplanting human neural progenitors, genetically modifiedto secrete GDNF, in the spinal cord of an ALS rat model. Asa result, the dying motor neurons were rescued, but theiraxonal projections to the muscle were lost. Our results showthat bone marrow stem cells are capable of not onlyprotecting the motor neurons but also their projections,increasing significantly their motor functions. Also, we haveshown that the cells are capable of inducing GDNFexpression, thus whole bone marrow can be used as avehicle to increase GDNF expression in the damaged spinalcords, avoiding motor degeneration. This increased GDNF isrestricted to the area the cells are transplanted, thus it ispossible that the factors released by the grafted cells to theadjacent motor neurons increase their probabilities ofsurvival and thus avoid the loss of motor functions, asopposed to a possible recovery of lost neurons. In addition,retrograde trophic effects may act through the activation ofGDNF axonal receptors in cortical motor neurons, thus alsoactivating neurotrophic mechanisms in the motor cortex.

In conclusion, whole bone marrow transplantation is arelatively simple method to consider as a potentialtreatment of motor neuron degenerative diseases, capableof ameliorating its symptoms by reducing motor neurondeath, inducing GDNF production, ultimately resulting inan improvement of the host’s motor functions.


Acknowledgements We appreciate the help of M. Rodenas, C.Redondo, O. Bahamonde, A. Torregrosa, and A. Estirado for theirtechnical assistance. This work has been financed by EUCOMMTOOLS,Science and Innovation Ministry (MICINN BFU-2008-00588,CONSOLIDERCSD2007-00023), Valencian government (PROMETEO/2009/028), Cell Therapy Network-Carlos III Health Institute (RD06/0010/0023 and RD07/0010/2012), Alicia Koplowitz Foundation, 5P-Syndrome Foundation, and Diógenes Foundation/Elche (CATEDRAELA).

Disclosures The authors indicate no potential conflicts of interest.

References

1. Vucic, S., & Kiernan, M. C. (2006). Axonal excitability propertiesin amyotrophic lateral sclerosis. Clinical Neurophysiology, 117,1458–1466.

2. Bruijn, L. I., Miller, T. M., & Cleveland, D. W. (2004). Unravelingthe mechanisms involved in motor neuron degeneration in ALS.Annual Review of Neuroscience, 27, 723–749.

3. Rowland, L. P. (1998). Diagnosis of amyotrophic lateral sclerosis.Journal of Neurological Sciences, 160(Suppl 1), S6–S24.

4. Boillee, S., Yamanaka, K., Lobsiger, C. S., et al. (2006). Onset andprogression in inherited ALS determined by motor neurons andmicroglia. Science, 312, 1389–1392.

5. Stavarachi, M., Apostol, P., Toma, M., Cimponeriu, D., &Gavrila, L. (2010). Spinal muscular atrophy disease: a literaturereview for therapeutic strategies. Journal of Medicine and Life,3, 3–9.

6. Visser, J., van den Berg-Vos, R. M., Franssen, H., et al. (2007).Disease course and prognostic factors of progressive muscularatrophy. Archives of Neurology, 64, 522–528.

7. Lorson, C. L., Rindt, H., & Shababi, M. (2010). Spinal muscularatrophy: mechanisms and therapeutic strategies. Human MolecularGenetics, 19, R111–R118.

8. Silani, V., Cova, L., Corbo, M., Ciammola, A., & Polli, E. (2004).Stem-cell therapy for amyotrophic lateral sclerosis. Lancet, 364,200–202.

9. Svendsen, C. N., & Langston, J. W. (2004). Stem cells forParkinson disease and ALS: replacement or protection? NatureMedicine, 10, 224–225.

10. Mazzini, L., Fagioli, F., Boccaletti, R., et al. (2003). Stem celltherapy in amyotrophic lateral sclerosis: a methodologicalapproach in humans. Amyotrophic Lateral Sclerosis and OtherMotor Neuron Disorders, 4, 158–161.

11. Corti, S., Locatelli, F., Donadoni, C., et al. (2004). Wild-type bonemarrow cells ameliorate the phenotype of SOD1-G93A ALS miceand contribute to CNS, heart and skeletal muscle tissues. Brain,127, 2518–2532.

12. Cabanes, C., Bonilla, S., Tabares, L., & Martinez, S. (2007).Neuroprotective effect of adult hematopoietic stem cells in amouse model of motoneuron degeneration. Neurobiology ofDisease, 26, 408–418.

13. Kastin, A. J., Akerstrom, V., & Pan, W. (2003). Glial cell line-derived neurotrophic factor does not enter normal mouse brain.Neuroscience Letters, 340, 239–241.

14. Corti, S., Locatelli, F., Papadimitriou, D., Strazzer, S., &Comi, G. P. (2004). Somatic stem cell research for neuralrepair: current evidence and emerging perspectives. Journal ofCellular and Molecular Medicine, 8, 329–337.

15. Vercelli, A., Mereuta, O. M., Garbossa, D., et al. (2008). Humanmesenchymal stem cell transplantation extends survival, improvesmotor performance and decreases neuroinflammation in mouse

model of amyotrophic lateral sclerosis. Neurobiology of Disease,31, 395–405.

16. Blot, S., Poirier, C., & Dreyfus, P. A. (1995). The mouse mutationmuscle deficient (mdf) is characterized by a progressivemotoneuron disease. Journal of Neuropathology and ExperimentalNeurology, 54, 812–825.

17. Poirier, C., Blot, S., Fernandes, M., et al. (1998). A high-resolution genetic map of mouse chromosome 19 encompassingthe muscle-deficient osteochondrodystrophy (mdf-ocd) region.Mammalian Genome, 9, 390–391.

18. Schmidt, W. M., Kraus, C., Hoger, H., et al. (2007). Mutation inthe Scyl1 gene encoding amino-terminal kinase-like proteincauses a recessive form of spinocerebellar neurodegeneration.EMBO Reports, 8, 691–697.

19. Okabe, M., Ikawa, M., Kominami, K., Nakanishi, T., &Nishimune, Y. (1997). 'Green mice' as a source of ubiquitousgreen cells. FEBS Letters, 407, 313–319.

20. Sanchez, M. P., Silos-Santiago, I., Frisen, J., He, B., Lira, S.A., & Barbacid, M. (1996). Renal agenesis and the absenceof enteric neurons in mice lacking GDNF. Nature, 382, 70–73.

21. Jones, J., Jaramillo-Merchan, J., Bueno, C., Pastor, D., Viso-Leon, M., & Martinez, S. (2010). Mesenchymal stem cellsrescue Purkinje cells and improve motor functions in a mousemodel of cerebellar ataxia. Neurobiology of Disease, 40, 415–423.

22. McGavern, D. B., Zoecklein, L., Drescher, K. M., & Rodriguez,M. (1999). Quantitative assessment of neurologic deficits in achronic progressive murine model of CNS demyelination.Experimental Neurology, 158, 171–181.

23. Kaspar, B. K., Frost, L. M., Christian, L., Umapathi, P., &Gage, F. H. (2005). Synergy of insulin-like growth factor-1 andexercise in amyotrophic lateral sclerosis. Annals of Neurology,57, 649–655.

24. Kayatekin, B. M., Gonenc, S., Acikgoz, O., Uysal, N., & Dayi, A.(2002). Effects of sprint exercise on oxidative stress in skeletalmuscle and liver. European Journal of Applied Physiology, 87,141–144.

25. He, Q., Wan, C., & Li, G. (2007). Concise review: multipotentmesenchymal stromal cells in blood. Stem Cells, 25, 69–77.

26. Herzog, E. L., Chai, L., & Krause, D. S. (2003). Plasticity ofmarrow-derived stem cells. Blood, 102, 3483–3493.

27. Giordano, A., Galderisi, U., & Marino, I. R. (2007). From thelaboratory bench to the patient's bedside: an update on clinicaltrials with mesenchymal stem cells. Journal of Cellular Physiology,211, 27–35.

28. Vitry, S., Bertrand, J. Y., Cumano, A., & Dubois-Dalcq, M. (2003).Primordial hematopoietic stem cells generate microglia but notmyelin-forming cells in a neural environment. Journal ofNeuroscience, 23, 10724–10731.

29. Alvarez-Dolado, M., Pardal, R., Garcia-Verdugo, J. M., et al.(2003). Fusion of bone-marrow-derived cells with Purkinjeneurons, cardiomyocytes and hepatocytes. Nature, 425, 968–973.

30. Corti, S., Locatelli, F., Donadoni, C., et al. (2002). Neuro-ectodermal and microglial differentiation of bone marrow cellsin the mouse spinal cord and sensory ganglia. Journal ofNeuroscience Research, 70, 721–733.

31. Javazon, E. H., Beggs, K. J., & Flake, A. W. (2004). Mesenchymalstem cells: paradoxes of passaging. Experimental Hematology, 32,414–425.

32. Hall, E. D., Oostveen, J. A., & Gurney, M. E. (1998). Relationshipof microglial and astrocytic activation to disease onset andprogression in a transgenic model of familial ALS. Glia, 23,249–256.


33. Howland, D. S., Liu, J., She, Y., et al. (2002). Focal loss of theglutamate transporter EAAT2 in a transgenic rat model of SOD1mutant-mediated amyotrophic lateral sclerosis (ALS). Proceedingsof the National Academy of Sciences of the United States ofAmerica, 99, 1604–1609.

34. Zhao, C. P., Zhang, C., Zhou, S. N., et al. (2007). Humanmesenchymal stromal cells ameliorate the phenotype of SOD1-G93A ALS mice. Cytotherapy, 9, 414–426.

35. Caplan, A. I., & Dennis, J. E. (2006). Mesenchymal stem cells astrophic mediators. Journal of Cellular Biochemistry, 98, 1076–1084.

36. Auffray, I., Chevalier, S., Froger, J., et al. (1996). Nerve growthfactor is involved in the supportive effect by bone marrow–derived stromal cells of the factor-dependent human cell line UT-7. Blood, 88, 1608–1618.

37. Labouyrie, E., Dubus, P., Groppi, A., et al. (1999). Expression ofneurotrophins and their receptors in human bone marrow.American Journal of Pathology, 154, 405–415.

38. Chen, X., Katakowski, M., Li, Y., et al. (2002). Human bonemarrow stromal cell cultures conditioned by traumatic brain tissueextracts: growth factor production. Journal of NeuroscienceResearch, 69, 687–691.

39. Crigler, L., Robey, R. C., Asawachaicharn, A., Gaupp, D., & Phinney,D. G. (2006). Humanmesenchymal stem cell subpopulations expressa variety of neuro-regulatory molecules and promote neuronal cellsurvival and neuritogenesis. Experimental Neurology, 198, 54–64.

40. Araki, T., & Milbrandt, J. (2000). Ninjurin2, a novel homophilicadhesion molecule, is expressed in mature sensory and enteric

neurons and promotes neurite outgrowth. Journal of Neuroscience,20, 187–195.

41. Garcia, R., Aguiar, J., Alberti, E., de la Cuetara, K., & Pavon,N. (2004). Bone marrow stromal cells produce nerve growthfactor and glial cell line-derived neurotrophic factors.Biochemical and Biophysical Research Communications, 316,753–754.

42. Nagano, I., Shiote, M., Murakami, T., et al. (2005). Beneficialeffects of intrathecal IGF-1 administration in patients withamyotrophic lateral sclerosis. Neurological Research, 27, 768–772.

43. Storkebaum, E., Lambrechts, D., Dewerchin, M., et al. (2005).Treatment of motoneuron degeneration by intracerebroventriculardelivery of VEGF in a rat model of ALS. Nature Neuroscience, 8,85–92.

44. Wang, Y., Mao, X. O., Xie, L., et al. (2007). Vascular endothelialgrowth factor overexpression delays neurodegeneration and pro-longs survival in amyotrophic lateral sclerosis mice. Journal ofNeuroscience, 27, 304–307.

45. Blanquer, M., Perez Espejo, M. A., Iniesta, F., et al. (2010). Bonemarrow stem cell transplantation in amyotrophic lateral sclerosis:technical aspects and preliminary results from a clinical trial.Methods and Findings in Experimental and Clinical Pharmacology,32(Suppl A), 31–37.

46. Suzuki,M.,McHugh, J., Tork, C., et al. (2007). GDNF secreting humanneural progenitor cells protect dying motor neurons, but not theirprojection tomuscle, in a rat model of familial ALS.PloS One, 2, e689.


comparative effects between bone marrow and mesenchymal stem cell transplantation in gdnf expression...

Documents