Preservation of Hematopoietic Properties inTransplanted Bone Marrow Cells in the Brain

Kenji Ono,1,2 Ken Yoshihara,1,2 Hiromi Suzuki,2 Kenji F. Tanaka,2 Takemasa Takii,1

Kikuo Onozaki,1 and Makoto Sawada2*1Department of Molecular Health Sciences, Graduate School of Pharmaceutical Sciences, Nagoya CityUniversity, Mizuho, Nagoya, Aichi, Japan2Joint Research Division for Therapies Against Intractable Diseases, Institute for Comprehensive MedicalScience, Fujita Health University, Toyoake, Aichi, Japan

Recent studies have described the possible transdiffer-entiation of bone marrow cells (BMC) into neurons andglia when they migrate to the brain. However, we havereported that some immature BMC migrating into thebrain parenchyma after bone marrow transplantation ex-press early hematopoietic markers but not neural or glialmarkers. The present study further characterizes trans-planted BMC that migrate to the brain. Double immuno-labeling confirmed that BMC migrating to the brain ex-pressed hematopoietic but not neural markers, such asnestin, microtubule-associated protein-2 and glial fibril-lary acidic protein, even 4 and 18 weeks after bonemarrow transplantation. BMC that expressed green flu-orescent protein also expressed hematopoietic but notneural markers when cultured with mixed brain cells ac-cording to double immunolabeling and single-cell dis-section using a laser. Analysis of the DNA content indi-cated that most of the migrated BMC were arrested atthe G0/G1 phase, and aneuploidy or tetraploidy wasundetectable. Thus, BMC that migrate to the brain prob-ably have preserved hematopoietic properties underphysiological conditions. © 2003 Wiley-Liss, Inc.

Key words: bone marrow transplantation; transdifferen-tiation; ER-MP12; green fluorescent protein; microglia

Recent studies have demonstrated that transplantedbone marrow stem cells are not limited to the formation ofblood cells but can also generate neurons and glia (Brazel-ton et al., 2000; Mezey et al., 2000; Nakano et al., 2001).Moreover, transplanted adult bone marrow cells (BMC)have generated unexpected phenotypes in vivo, includingmuscle (Gussoni et al., 1999), liver (Lagasse et al., 2000),and other types of cells, even brain cells (Orlic et al., 2001;Hess et al., 2002). Thus, adult bone marrow might containpluripotent stem cells or stem cells that could becomepluripotent under the appropriate conditions. On theother hand, we have reported that some immature BMCmigrate into the brain parenchyma after bone marrowtransplantation (BMT) and that most of them express earlyhematopoietic and mature monocytic markers (Ono et al.,1999). This implies that BMC would retain a hematopoi-etic phenotype in the host brain and not transdifferentiate

into neurons or glia. A recent report has indicated that thephenotype of transplanted bone marrow stem cells is al-tered not by the direct conversion of blood cells to braincells but rather through the spontaneous generation ofhybrids between these two cell types (Terada et al., 2002).Therefore, we further characterized transplanted BMCthat migrate to the brain.

MATERIALS AND METHODS

Green fluorescent protein (GFP) transgenic mice[TgN(�-act-EGFP)Osb; Okabe et al., 1997] were killed, andmuscle was removed from both femora. The ends of the boneswere cut off, and then marrow cells from the femora wereflushed out with RPMI 1640 medium containing 10% fetal calfserum delivered through 21-gauge needles. The cells werewashed, resuspended in medium, and transplanted. Freshly iso-lated BMC were confirmed for expression of CD34 and Sca-1in the preparations by fluorescence-activated cell sorting (FACS)analysis. Adult B6 mice (18 g–48 g body weight) were treatedwith 5-fuorouracil (5-FU), and then 1 � 107 cells were injectedvia the tail vein as described elsewhere (Ono et al., 1999). At 7days and 4 and 18 weeks after BMT, the hearts of anesthetizedmice were perfused with about 100 ml isotonic saline.

The brain and peripheral tissues were isolated, frozen inliquid nitrogen, and embedded in OCT compound (Tissue Tek;Miles, Elkhart, IN). Frozen sections (7 �m) were serially cutinto seven slices using a microtome (Leica), then transferred togelatin-coated slides and air dried. The sections were fixed with4% paraformaldehyde in PBS at 4°C for 15 min and doublylabeled with polyclonal antibodies against either GFP (MBL,Nagoya, Japan) or glial fibrillary acidic protein (GFAP; a markerof astrocytes; Immunon, Pittsburgh, PA) and with monoclonalantibodies against GFP (Clontech, Palo Alto, CA), ER-MP12 (amurine early-stage myeloid precursor marker; BMA Switzer-land), Mac-1 (a murine complement component C3bi receptor

*Correspondence to: Dr. Makoto Sawada, Joint Research Division forTherapies Against Intractable Diseases, Institute for Comprehensive Med-ical Science, Fujita Health University, Toyoake, Aichi 470-1192, Japan.E-mail: msawada@fujita-hu.ac.jp

Received 24 September 2002; Revised 10 January 2003; Accepted 15January 2003

Journal of Neuroscience Research 72:503–507 (2003)

© 2003 Wiley-Liss, Inc.

marker; BMA), CD45 (a leukocyte common antigen, Ly-5;PharMingen Japan), CD34 (a hematopoietic stem and progen-itor marker; PharMingen), nestin (a marker for neural stem andprogenitors; Chemicon, Temecula, CA), or microtubule-associated protein-2 (MAP-2; a stringent marker for neurons;Roche, Nutley, NJ) as described elsewhere (Ono et al., 1999).The reaction was visualized with R-phycoerythrin-conjugatedgoat F(ab�)2 anti-rat IgG (Caltag, Burlingame, CA) at a dilutionof 1:200, rhodamine-conjugated goat anti-rat IgG (H�L chains;Biomeda Corp., Foster City, CA) at a dilution of 1:200, fluo-rescein isothiocyanate (FITC)-conjugated goat F(ab�)2 anti-rabbit IgG (Caltag) at a dilution of 1:200, Cy3-conjugated goatanti-rabbit IgG (Jackson Immunoresearch, West Grove, PA),the M.O.M. Immunodetection kit (Vector, Burlingame, CA),and UltraAvidin-Fluorescein (Leinco Technologies, St. Louis,MO) or Cy3-conjugated egg-white avidin (Jackson Immunore-search), then photographed. Replacement of the BMC withcells isolated from the femora of a recipient mouse was moni-tored using a FACSCalibur cell sorter (Becton-Dickinson, SanJose, CA) equipped with a 530-nm filter (bandwidth �15 nm)and a 585-nm filter (bandwidth �21 nm) and Cell Questsoftware (Becton-Dickinson).

Mixed glial cultures (MGC) were prepared from neonatalB6 mice as described elsewhere (Sawada et al., 1989). MGCwere incubated for 14 days in vitro, and then 1 � 105 freshBMC isolated from GFP-transgenic mice were added. After 7 or14 days, the cocultures were fixed with 4% paraformaldehyde inphosphate-buffered saline (PBS) at 4°C for 15 min and labeledwith a monoclonal or polyclonal antibody against one of celllineage-specific differentiation markers described above. Over3,000 GFP-positive cells from the 14-day cocultures were alsomicrodissected using a laser (LM200; Olympus, Tokyo, Japan)with a 7.5-�m beam diameter, and then mRNA expression wasanalyzed.

RNA extracted from about 3,000 dissected cells using amodification of the acid phenol-guanidine method (Sawada et

al., 1992) was used as a template for first-strand cDNA synthesisas follows. A random primer (0.1 �g) was incubated at 95°C for10 min with the RNA (1 �g) in a volume of 30 �l, then placedon ice for 5 min. This mixture was then incubated at 37°C for90 min with a mixture of 100 U M-MLV reverse transcriptase(Gibco BRL, Grand Island, NY), 1� reverse transcriptionbuffer, 10 mM dithiothreitol, 40 U RNasin, and 0.56 mM eachof dATP, dGTP, dCTP, and dTTP in a volume of 50 �l, thenat 95°C for 10 min. The cDNA was amplified with Taq DNApolymerase (Takara, Tokyo, Japan) using primer pairs specific toenhanced GFP (EGFP; sense primer: TCG TGA CCA CCCTGA CCT; antisense primer: GTG ATC GCG CTT CTCGTT), PU.1 (sense primer: TCT GAT GGA GAA GCT GATGG; antisense primer: CTT GAC TTT CTT CAC CTC GC),erythropoietin receptor (EpoR; sense primer: CGG TTC TCCTCG CTA TCA C; antisense primer: CCC TCA AAC TCGCTC TCT G), or neurofilament medium (NF-M; sense primer:AAG TGG GAA ATG GCT CGT CA; antisense primer: GCGATG GCT GTG AGG GTT TC) for 35 cycles (94°C for1 min, 55°C for 1 min, and 72°C for 2 min), and GFAP (senseprimer: GCC AAG GAG CCC ACC AAA CT; antisenseprimer: ATC TCC TCC TCC AGC GAT TCA AC) orGAPDH (sense primer: GGA GAT TGT TGC CAT CAACGA C; antisense primer: ATG AGC CCT TCC ACA ATGCCA AAG) for 30 cycles. The 473-bp (EGFP), 300-bp (PU.1),501-bp (EpoR), 314-bp (NF-M), 275-bp (GFAP), and 441-bp

Fig. 1. Immunohistochemical analy-sis of cells expressing GFP in brain 7days after BMT. Sections werestained with antibodies against ER-MP12 (A,B), Mac-1 (C,D), nestin(E,F), GFAP (G,H), and MAP-2(I,J). A, C, E, G, and I indicateGFP fluorescence. A and B, C andD, E and F, G and H, and I and Jare photographs of the same sec-tions. Arrows indicate cells express-ing GFP. Scale bars � 40 �m.

TABLE I. Characterization of Cells Expressing GFP in Brain 7Days After Bone Marrow Transplantation

Hematopoietic marker Neural marker

CD45 CD34 ER-MP12 Mac-1 Nestin GFAP MAP-2

Ratioa 27/27 0/25 36/37 39/42 0/32 0/32 0/25aRatio indicates total number of stained cells among those expressing GFP.

504 Ono et al.

(GAPDH) polymerase chain reaction (PCR) products wereresolved by electrophoresis in 2% agarose gels, stained withethidium bromide, and photographed.

A portion of the mouse tissues after BMT was minced andenzymatically dispersed. Cell suspensions were digested with 0.5mg/ml RNase A and stained with propidium iodide (PI). TheDNA content of cells expressing GFP was analyzed using aFACS method and ModFit software (Becton-Dickinson).

RESULTSAt 7 days after BMT, GFP-positive cells formed clusters

in the brain as described elsewhere (Ono et al., 1999). Tocharacterize the cells immunologically, brain sections werestained with antibodies against several cell lineage differenti-ation markers and an anti-GFP antibody. Figure 1 and TableI show that all of the GFP-positive clusters examined werealso CD45 positive, and most of them reacted with ER-MP12 and Mac-1 antibodies, whereas they were negative forCD34, nestin, GFAP, and MAP-2.

At 4 and 18 weeks after BMT, brain sections weresimilarly stained with several antibodies (Fig. 2). Almost allGFP-positive cells examined were hematopoietic markerpositive, but none of them was neural marker positive.

A mixed glial culture is one of the most suitable invitro models of the brain cell network. To investigate theeffects of brain cells on immature BMC differentiation, wecocultivated BMC that expressed GFP with MGC for 7and 14 days and then analyzed the surface antigens onthese cells. Figure 3 and Table II show that most of thecells expressing GFP were positive for CD45 and Mac-1and that about 10% of them reacted with ER-MP12antibodies, whereas they were negative for CD34, nestin,GFAP, and MAP-2. Cells expressing GFP on MGC werecaptured by the laser microdissection system, and theexpression of cell lineage-specific mRNA was analyzed byreverse transcription (RT)-PCR. We detected the expres-

sion of mRNA specific to hematopoietic cells, such asPU.1 and EpoR, but not of that specific to neural cells,such as NF-M and GFAP.

We analyzed the DNA content in cells expressingGFP in the brain and peripheral tissues 7 days after BMT(Fig. 4). Flow cytometric analysis of PI-stained GFP-positive cells did not detect any extraordinary ploidy (an-euploidy nor tetraploidy), suggesting the absence of cellfusion. We observed that 15.57% � 2.46% and 34.45% �5.36% of cells expressing GFP in the bone marrow andspleen, respectively, were in the S phase of the cell cycle.In that 10.95% � 1.60%, of cells in S phase of the originalBMC expressing GFP migrated, those cells that migratedto the bone marrow and spleen seemed to be more pro-liferative than the original cells. On the other hand, insofaras 7.75% � 1.56% of the GFP-positive BMC in the brainwere in S phase, these cells seemed to be less proliferativethan the original cells.

DISCUSSIONSome studies have shown that cells expressing exog-

enous markers migrate into the brain of both normal anddefective mice after BMT. Most of these studies indicatedthat cells derived from donor cells in the brain expressedantigens against microglia and macrophages, such asMac-1 or F4/80. However, others have reported that

TABLE II. Characterization of BMC Expressing GFP in MixedGlial Cultures

Hematopoietic marker Neural marker

CD45 CD34 ER-MP12 Mac-1 Nestin GFAP MAP-2

Ratioa 249/251 0/218 22/265 270/278 0/268 0/272 0/250aRatio indicates total number of stained cells among cells expressing GFP.

Fig. 2. Immunohistochemical anal-ysis of cells expressing GFP inbrain 4 and 18 weeks after BMT.Sections were stained with anti-bodies against ER-MP12 (A,B,G,H), CD45 (C,D,I,J), andGFAP (E,F,K,L). A–F are sec-tions 4 weeks and G–L are sec-tions 18 weeks after BMT. A, C,E, G, I, and K indicate GFP flu-orescence. A and B, C and D, Eand F, G and H, I and J, and Kand L are photographs of the samesections. Arrows indicate cells ex-pressing GFP. Scale bars � 40�m.

Hematopoietic Properties of Transplanted BMC in Brain 505

some donor-derived cells in the brain express neural mark-ers, such as GFAP (Eglitis and Mezey, 1997); specificmarker for astrocytes; and the neuronal markers NeuN,neuron-specific enolase (NSE), and neurofilament (Bra-zelton et al., 2000; Mezey et al., 2000). However, thepresent study found that GFP-positive cells in the brainexpressed immature and mature myeloid-lineage antigensbut not neural markers such as GFAP, MAP-2, and nestin(Table I), even at 4 and 18 weeks after BMT, when manygroups have evaluated transdifferentiation. This may bedue to a difference in the experimental protocol. A recentstudy pointed out that irradiation creates significant tissuedamage for a long period, including damage of the cellswith which the blood–brain barrier is consistent. We used5-FU to reduce the numbers of recipient hematopoieticcells, whereas others have applied irradiation. We con-firmed that no breakage of the blood–brain barrier oc-curred with our experimental protocol using Evans bluedye (data not shown). We also examined the actual con-

tribution of circulating donor-derived cells soon after theinjection to determine the difference in properties ofdonor-derived cells in brain and other peripheral tissues.We found that BMT cells directly migrated into the brainwithout, homing in to the host bone marrow.

Recent reports have suggested that the altered pheno-type of transplanted bone marrow stem cells appears to arisenot through the direct conversion of immature blood cellsinto brain cells but rather through the spontaneous genera-tion of hybrids between these two cell types (Terada et al.,2002). The incidence of such cell fusion increases whenimmature cells are irradiated (Terzoudi and Pantelias, 1997).However, most of the papers reporting transdifferentiationdid not exclude this possibility. In this experiment, we ex-amined the DNA content of donor-derived cells in the brainand found that most of the migrated BMC were arrested atthe G0/G1 phase and that aneuploidy or tetraploidy wasundetectable. This means that our results can exclude resultsfrom accidental cell fusion.

Fig. 3. Immunocytochemical and RT-PCR analyses of BMC express-ing GFP in coculture with MGC. A–H: Cocultures stained withantibodies against ER-MP12 (A,B), Mac-1 (C,D), GFAP (E,F), andMAP-2 (G,H). A, C, E, and G indicate GFP fluorescence. A and B, Cand D, E and F, and G and H are photographs of the same areas. Arrows

indicate cells expressing GFP. I: Total RNA from dissected cellsexpressing GFP was analyzed by EGFP-, PU.1-, EpoR-, GFAP-, orNF-M-specific RT-PCR. BMC, bone marrow cells; Mph, peritonealmacrophages; MG, primary purified microglia; MGC-MG, MGC afterpurification of microglia. Scale bars � 40 �m.

506 Ono et al.

Bone marrow stromal cells also differentiate into notonly multiple mesenchymal cells, such as chondrocytesand adipocytes (Pittenger et al., 1999), but also neuronalphenotypic cells that exhibit NSE or NF-M in vitro(Woodbury et al., 2000). Stromal cells stereotaxically in-jected into the brain differentiate into astrocytes and cellswith a neuronal phenotype (Kopen et al., 1999). It ispossible that the donor-derived cells in the brain with aneuronal phenotype after BMT are derived from cells ofthe mesenchymal lineage, such as bone marrow stromalcells. In conclusion, we have indicated that some of theBMC that migrate to the brain directly from circulationafter an injection probably do not undergo transdifferen-tiation under nondamaged, nontoxic, and physiologicalconditions.

ACKNOWLEDGMENTSWe thank Drs. Masaru Okabe and Masahito Ikawa

for providing GFP-transgenic mice.

REFERENCESBrazelton TR, Rossi FM, Keshet GI, Blau HM. 2000. From marrow to

brain: expression of neuronal phenotypes in adult mice. Science 290:1775–1779.

Eglitis MA, Mezey E. 1997. Hematopoietic cells differentiate into bothmicroglia and macroglia in the brains of adult mice. Proc Natl Acad SciUSA 94:4080–4085.

Gussoni E, Soneoka Y, Strickland CD, Buzney EA, Khan MK, Flint AF,Kunkel LM, Mulligan RC. 1999. Dystrophin expression in the mdxmouse restored by stem cell transplantation. Nature 401:390–394.

Hess DC, Hill WD, Martin-Studdard A, Carroll J, Brailer J, Carothers J.2002. Bone marrow as a source of endothelial cells and NeuN-expressingcells after stroke. Stroke 33:1362–1368.

Kopen GC, Prockop DJ, Phinney DG. 1999. Marrow stromal cells migratethroughout forebrain and cerebellum, and they differentiate into astro-cytes after injection into neonatal mouse brains. Proc Natl Acad Sci USA96:10711–10716.

Lagasse E, Connors H, Al-Dhalimy M, Reitsma M, Dohse M, Osborne L,Wang X, Finegold M, Weissman IL, Grompe M. 2000. Purified hema-topoietic stem cells can differentiate into hepatocytes in vivo. Nat Med6:1229–1234.

Mezey E, Chandross KJ, Harta G, Maki RA, McKercher SR. 2000.Turning blood into brain: cells bearing neuronal antigens generated invivo from bone marrow. Science 290:1779–1782.

Nakano K, Migita M, Mochizuki H, Shimada T. 2001. Differentiation oftransplanted bone marrow cells in the adult mouse brain. Transplantation71:1735–1740.

Okabe M, Ikawa M, Kominami K, Nakanishi T, Nishimune Y. 1997.‘Green mice’ as a source of ubiquitous green cells: FEBS Lett 407:313–319.

Ono K, Takii T, Onozaki K, Ikawa M, Okabe M, Sawada M.1999. Migration of exogenous immature hematopoietic cells into adultmouse brain parenchyma under GFP-expressing bone marrow chimera.Biochem Biophys Res Commun 262:610–614.

Orlic D, Kajstura J, Chimenti S, Jakoniuk I, Anderson SM, Li B, PickelJ, McKay R, Nadal-Ginard B, Bodine DM, Leri A, Anversa P. 2001.Bone marrow cells regenerate infarcted myocardium. Nature 410:701–705.

Pittenger MF, Mackay AM, Beck SC, Jaiswal RK, Douglas R, Mosca JD,Moorman MA, Simonetti DW, Craig S, Marshak DR. 1999. Multilineagepotential of adult human mesenchymal stem cells. Science 284:143–147.

Sawada M, Kondo N, Suzumura A, Marunouchi T. 1989. Production oftumor necrosis factor-alpha by microglia and astrocytes in culture. BrainRes 491:394–397.

Sawada M, Suzumura A, Kondo N, Marunouchi T. 1992. Induction ofcytokines in glial cells by trans activator of human T-cell lymphotropicvirus type I. FEBS Lett 313:47–50.

Terada N, Hamazaki T, Oka M, Hoki M, Mastalerz DM, Nakano Y,Meyer EM, Morel L, Petersen BE, Scott EW. 2002. Bone marrow cellsadopt the phenotype of other cells by spontaneous cell fusion. Nature416:542–545.

Terzoudi GI, Pantelias GE. 1997. Conversion of DNA damage into chro-mosome damage in response to cell cycle regulation of chromatin con-densation after irradiation. Mutagenesis 12:271–276.

Woodbury D, Schwarz EJ, Prockop DJ, Black IB. 2000. Adult rat andhuman bone marrow stromal cells differentiate into neurons. J NeurosciRes 61:364–370.

Fig. 4. DNA content of GFP-expressing cells analyzed in tissues afterBMT. A–D indicate flow cytometric histograms of cells expressingGFP in donor BMC (A), BMC (B), spleen (C), and brain (D) 7 daysafter BMT.

Hematopoietic Properties of Transplanted BMC in Brain 507