Gata2 Is a Rheostat for Mesenchymal Stem Cell Fate in Male ... · Gata2 Is a Rheostat for...

Gata2 Is a Rheostat for Mesenchymal Stem Cell Fatein Male Mice

Xiaoxiao Li, HoangDinh Huynh, Hao Zuo, Marjo Salminen, and Yihong Wan

Department of Pharmacology (X.L., H.H., H.Z., Y.W.), The University of Texas Southwestern MedicalCenter, Dallas, Texas 75390; and Department of Veterinary Biosciences (M.S.), University of Helsinki,Helsinki 00014, Finland

Gata2 is a zinc finger transcription factor that is important in hematopoiesis and neuronal devel-opment. However, the roles of Gata2 in the mesenchymal lineages are poorly understood. In vitrostudies suggest that Gata2 modulates adipocyte differentiation and mesenchymal stem cell (MSC)proliferation. To systematically determine the in vivo functions of Gata2 in the MSC lineage com-mitment and development, we have generated three mouse models in which Gata2 is specificallydeleted in MSCs, adipocytes, or osteoblasts. During the MSC expansion stage, Gata2 promotesproliferation and attenuates differentiation; thereby Gata2 loss in MSCs results in enhanced dif-ferentiation of both adipocytes and osteoblasts. During the differentiation stage, Gata2 also playsMSC-independent roles to impede lineage commitment; hence, Gata2 loss in adipocyte or osteo-blast lineages also augments adipogenesis and osteoblastogenesis, respectively. These findingsreveal Gata2 as a crucial rheostat of MSC fate to control osteoblast and adipocyte lineagedevelopment. (Endocrinology 157: 1021–1028, 2016)

Mesenchymal stem cells (MSCs) are multipotent stro-mal cells that have the potential to differentiate into

several lineages such as osteoblasts and adipocytes (1).There are several key transcription factors that commitand guide MSCs to specific fate: in the osteoblast lineage,runt-related transcription factor 2 (Runx2) orchestratesthe activation of downstream genes that control the earlystage of osteoblastogenesis; in the adipocyte lineage,CCAAT/enhancer binding proteins (CEBPs) are indis-pensable for triggering the initiation of adipogenesis,whereas peroxisome proliferator-activated receptor-�(PPAR�) activation promotes adipocyte maturation (2).MSCs can be isolated from the bone marrow, expanded,and differentiated into different lineages ex vivo. Thus,they present exciting potential for stem cell therapy torestore and replace damaged mesenchymal tissues (3, 4).

Gata2 is a zinc finger transcription factor that playscrucial roles in hematopoietic development. Gata2 ho-mozygous null embryos die during embryogenesis due to

a failure in blood cell generation (5), and Gata2 haploin-sufficiency resulted in altered integrity of the definitivehematopoietic stem cell (HSC) compartment, leading to asignificant reduction in HSC number (6). Gata2 deficiencyhas been implicated in many human diseases. For exam-ple, 84% of GATA2-deficient patients develop myelodys-plastic syndrome, in which the ability of bone marrow tomake blood cells is compromised. In addition, these pa-tients are also more susceptible to acute myeloid leukemia,chronic myelomonocytic leukemia, immunodeficiency,and vascular/lymphatic dysfunction (7).

Although Gata2 regulation of HSCs has been well stud-ied, whether Gata2 plays a functional role in MSCs or itssublineages is largely unknown. Most studies to date havebeen done in adipocyte lineage using cell lines such as3T3-F442A or TBR343 preadipocytic cell lines. Tong et al(8) found that Gata2 inhibits preadipocyte-to-adipocytetransition by suppressing PPAR� and interacting withCEBP family transcription factors (9). In humans, it was

ISSN Print 0013-7227 ISSN Online 1945-7170Printed in USACopyright © 2016 by the Endocrine SocietyReceived September 25, 2015. Accepted January 20, 2016.First Published Online January 26, 2016

Abbreviations: ALP, alkaline phosphatase; BAT, brown adipose tissue; BrdU, 5-bromo-2�-deoxyuridine; CEBP, CCAAT/enhancer binding protein; �CT, microcomputed tomogra-phy; FOP-flash, mutant T cell factor negative control reporter; HSC, hematopoietic stemcell; MSC, mesenchymal stem cell; ORO, Oil Red O; P1NP, amino-terminal propeptide oftype I collagen; PPAR�, peroxisome proliferator-activated receptor-�; Prx1, paired relatedhomeobox 1; Runx2, runt-related transcription factor 2; TOP-flash, T cell factor reporter;UCP1, uncoupling protein-1; WAT, white adipose tissue.

O R I G I N A L R E S E A R C H

doi: 10.1210/en.2015-1827 Endocrinology, March 2016, 157(3):1021–1028 press.endocrine.org/journal/endo 1021

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 11 March 2016. at 13:38 For personal use only. No other uses without permission. . All rights reserved.

reported that patients with aplastic anemia express lowerGata2 and higher PPAR� in bone marrow cells, whichcould explain the fatty marrow symptom (10). Recentlyanother in vitro study showed that Gata2 knockdown de-creased cell proliferation and increased the number of cellsin G1/G0 phase (11). However, very little is known abouthow Gata2 regulates MSCs and their potential to differ-entiate into sublineages such as osteoblasts and adipocytesin a physiological context. In particular, it is essentiallyunknown whether and how Gata2 regulates osteoblasto-genesis. Furthermore, no in vivo model has been estab-lished to study whether Gata2 is functionally significantfor MSC renewal and differentiation. In this study, wecreated mouse models in which Gata2 is specifically de-leted in MSCs, adipocytes, or osteoblasts to dissect itsphysiological roles in mesenchymal lineage development.

Materials and Methods

MiceGata2 flox mice have been previously described (12). Trans-

genic mice for paired related homeobox 1 (Prx1)-Cre (13), adi-ponectin-Cre (14), and osteocalcin-Cre (15) were from the Jack-son Laboratory. Prx1 is a transcription coactivator expressed inearly limb bud mesenchyme and in a subset of craniofacial mes-enchyme. Prx1-Cre mice are widely used to target mesenchymalstem cells to study mesenchymal specific gene functions. Micewere fed standard chow containing 4% fat ad libitum unlessstated otherwise. All studies were conducted with 2-month-oldmale littermates. Sample size estimate was based on power anal-yses performed using SAS 9.3 TS X64_7PRO platform at theUniversity of Texas Southwestern Medical Center BiostatisticsCore. With the observed group differences and the relativelysmall variation of the in vivo measurements, n � 4 and n � 3 willprovide more than 90% and more than 80% power at the typeI error rate of 0.05 (two sided test), respectively. All animal ex-periments were approved by the Institutional Animal Care andUse Committee of the University of Texas Southwestern MedicalCenter.

Bone analysesHistomorphometry of femoral sections was conducted using

the BIOQUANT software (Bioquant) as previously described(16, 17). Alkaline phosphatase (ALP) staining of osteoblasts wasperformed using an ALP staining kit containing fast blue RR salt(Sigma). Oil Red O (ORO) staining of adipocytes was performedas described (18). An ELISA of a serum bone formation markeramino-terminal propeptide of type I collagen (P1NP) and a mi-crocomputed tomography (�CT) were performed as previouslydescribed (16, 17).

Ex vivo bone marrow MSC expansion anddifferentiation

For the MSC expansion, mouse bone marrow cells were pu-rified with 100 �m cell strainer (19, 20) and then cultured for 3–7days in MSC media (mouse MesenCult proliferation kit; Stem-

Cell Technologies Inc). Osteoblasts were differentiated frommarrow MSCs as described (21). Briefly, MSCs were cultured in�-MEM containing 10% fetal bovine serum, 5 mM �-glycero-phosphate, and 100 �g/mL ascorbic acid (mineralization me-dium) for 9 days. Mature osteoblasts were identified as ALP�

cells using fast red violet LB salt (Sigma). For adipocyte differ-entiation, MSCs were cultured in adipogenesis medium (Mes-enCult basal medium � mouse MesenCult adipogenic stimula-tory supplement; StemCell Technologies) for 7 days. Matureadipocytes were identified as ORO� cells. Differentiation wasquantified by the mRNA expression of marker genes using RT-quantitative PCR.

5-Bromo-2�-deoxyuridine (BrdU) cell proliferationassay

Bone marrow cells were cultured for 4 days in MSC medium.The cells were serum starved overnight and then restimulatedwith serum for 6 hours to induce the S phase, during which BrdUwas provided in the culture medium. Cell proliferation wasquantified as BrdU incorporation using the BrdU ELISA assay inthe BrdU cell proliferation assay kit (GE Healthcare LifeSciences).

TOP-flash/FOP-flash reporter assayTo quantify �-catenin activity, human embryonic kidney-293

cells were transfected with a T cell factor (TOP-flash) luciferasereporter or a mutant T cell factor (FOP-flash)-negative controlreporter together with a cytomegalovirus-�gal reporter (as in-ternal control for transfection efficiency) as well as an expressionplasmid for a constitutively active �-catenin (�-catenin-�ex3) orvector control as we previously described (22). All transfectionswere performed using FuGENE HD (Roche) (n � 6) and re-peated for at least three times. Reporter assays were conducted48 hours after transfection. Data are presented as TOP to FOPratio.

RNA expression analysesRNA expression was analyzed by RT-quantitative PCR.

RNA was extracted with TRIZOL (Invitrogen), reverse tran-scribed into cDNA using an ABI high capacity cDNA RT kit(Invitrogen), and then analyzed using real-time quantitative PCR(SYBR Greener; Invitrogen) with gene-specific primers (Table 1)in triplicate. All RNA expression was normalized by mitochon-drial ribosomal protein L19, one of the most stable and reliableendogenous control genes (23).

Statistical analysesStatistical analyses were performed with a Student’s t test and

represented as mean � SD (*, P � .05; **, P � .01; ***, P � .005;****, P � .001; n.s., nonsignificant).

Results

Gata2 expression is differentially regulated duringMSC proliferation and differentiation

To study how Gata2 expression changes during MSCproliferation, we collected bone marrow MSCs and ex-

1022 Li et al Gata2 Controls Mesenchymal Lineages Endocrinology, March 2016, 157(3):1021–1028


panded them in culture. We found that Gata2 expressiondecreased as MSCs proliferated (Figure 1A). To study howGata2 expression changes during differentiation, wetreated the expanded MSCs with cocktails that triggeradipogenesis or osteoblastogenesis. The results show thatthe Gata2 level steadily increased during both adipocyteand osteoblast differentiation (Figure 1, B and C). Theseresults suggest that Gata2 is likely important for bothMSC proliferation and differentiation into different lin-eages such as adipocytes and osteoblasts. Furthermore,given Gata2’s opposite expression pattern during prolif-eration vs differentiation, it could play distinct and inde-pendent roles in MSC expansion and lineage commitment.

Gata2 deletion in early mesenchymal lineagedecreases MSC proliferation

To study how Gata2 affects early mesenchymal lineagein vivo, we generated MSC Gata2 conditional knockout(MSC-gata2-KO) mice by crossing Gata2 flox mice (12)with the previously established Prx1-Cre transgenic mice(13). By comparing MSC-gata2-KO mice with littermatecontrol mice, we found that Prx1-Cre successfully de-pleted Gata2 expression in all three mesenchymal popu-lations: MSCs, adipocytes, and osteoblasts (Figure 2,

A–C). As a result, MSC proliferation was markedly de-creased, measured by BrdU incorporation (Figure 2D).This is consistent with the in vitro observations by Kamataet al (11) that Gata2 small interfering RNA knockdown inhuman MSCs reduced cell proliferation and down-regu-lated genes in cell cycle and DNA replication. Moreover,we found that Gata2 loss does not affect MSC apoptosisbecause the expression of proapoptotic genes such as Fas,Bad, and Bax or antiapoptotic genes such as Bcl-2 andBcl-6 was unaltered (Figure 2E). These findings indicatethat Gata2 promotes MSC proliferation.

Gata2 deletion in early mesenchymal lineageenhances differentiation

We next examined how Gata2 loss in MSCs affectsadipocyte and osteoblast differentiation. There are twopossibilities: MSC Gata2 deletion may inhibit differenti-ation by reducing the number of progenitors due to de-creased proliferation; alternatively, MSC Gata2 deletionmay promote differentiation by dampening proliferation,thereby facilitating the proliferation-to-differentiationswitch. Our results supported the second hypothesis andshowed that Gata2 loss in MSCs increased the differenti-ation of both adipocytes and osteoblasts. In ex vivo ex-

Table 1. Mouse Primers for RT-PCR

Gene Forward Primer Reverse Primer

Gata2 CCTGGCGCATAACTACATGGA GAGTCGAGATGGTTGAAGAAGACAFas GCAGACATGCTGTGGATCTG AGTTTCATGAACCCGCCTCBad GAGCAACATTCATCAGCAGG TAAGCTCCTCCTCCATCCCTBax TTGCTGATGGCAACTTCAAC GAGGAAGTCCAGTGTCCAGCBcl-2 AGTACCTGAACCGGCATCTG GCTGAGCAGGGTCTTCAGAGBcl-6 AGTTTCTAGGAAAGGCCGGA ACTAGCGTGCCGGGTAAACTAdiponectin CAGTGGATCTGACGACACCAA GAACAGGAGAGCTTGCAACAGTPPARg2 GCTGATGCACTGCCTATGAGC CGGAGAGGTCCACAGAGCTGUCP1 AAGCTGTGCGATGTCCATGT AAGCCACAAACCCTTTGAAAAOsteocalcin TGACAAAGCCTTCATGTCCA ATAGCTCGTCACAAGCAGGGCol1a1 CGTCTGGTTTGGAGAGAGCAT GGTCAGCTGGATAGCGACATCRunx2 GCTCACGTCGCTCATCTTG ACACCGTGTCAGCAAAGCL19 GGTCTGGTTGGATCCCAATG CCCATCCTTGATCAGCTTCCT

Abbreviations: L19, mitochondrial ribosomal protein L19.

A B C

Figure 1. Gata2 decreases during MSC expansion but increases during differentiation. A, Gata2 expression decreased during MSC proliferation(n � 3). B, Gata2 expression increased during adipocyte differentiation (n � 3). C, Gata2 expression increased during osteoblast differentiation(n � 3). Error bars, SD.

doi: 10.1210/en.2015-1827 press.endocrine.org/journal/endo 1023


periments, adipocyte differentiationcultures from MSC-gata2-KO bonemarrow exhibited stronger OROstaining (Figure 2F) as well as higherexpression of adipocyte markerssuch as adiponectin and PPAR�2(Figure 2G) compared with controldifferentiation cultures. Consistentwith this observation, ORO stainingof mouse femur sections showedmore marrow fat in vivo (Figure 2H).Similarly, MSC Gata2 deletion alsoenhanced osteoblastogenesis shownby the stronger ALP staining (Figure2I) as well as elevated osteoblastmarkers such as osteocalcin and Col-lagen, type I, alpha 1 (Figure 2J) in exvivo differentiation cultures. In ac-cordance with this finding, ALPstaining of mouse femur sectionsshowed increased osteoblast numberand surface in vivo (Figure 2K);ELISA showed an increased serumbone formation marker P1NP (Fig-ure 2L); �CT showed higher bonevolume and trabecular number witha lower trabecular separation (Fig-ure 2M). Together these data sup-port that at the MSC level, Gata2 isa physiologically significant regula-tor that promotes proliferation andattenuates differentiation.

Gata2 deletion in adipocytelineage promotes adipocytedifferentiation

Because the expression of Gata2increases during adipocyte differen-tiation whereas it decreases duringMSC proliferation, we decided to in-vestigate whether Gata2 plays a syn-ergistic or opposite role in theadipose lineage that may be indepen-dent of those in MSC proliferation.To achieve this goal, we created adi-pocyte-specific Gata2 conditionalknockout (Ad-gata2-KO) micedriven by the previously establishedadiponectin-Cre (14). The adiponec-tin-Cre driver significantly reducedGata2 expression in the adipocytelineage (Figure 3A), whereas it did

A B C

D E

F G H

I J K

L M

Figure 2. Gata2 in MSC promotes proliferation to prevent differentiation. A–C, Gata2 level wassignificantly decreased in MSC (A), adipocyte (B), and osteoblast (C) ex vivo cultures derived fromMSC-gata2-KO bone marrow compared with littermate controls (n � 3). D, MSCs from MSC-gata2-KO mice (n � 5) exhibited reduced proliferation compared with controls (n � 10),measured by BrdU incorporation. E, MSCs from MSC-gata2-KO mice displayed unalteredexpression of proapoptotic and antiapoptotic genes (n � 3). F and G, Adipocyte differentiationwas enhanced for MSCs from MSC-gata2-KO mice, shown by the increased ORO staining (F) andthe higher expression of adipogenic markers such as adiponectin and PPAR�2 (G) (n � 3). H,Femurs from MSC-gata2-KO mice exhibited more marrow fat, shown by the higher adipocytenumber/bone area (Ad.N/B.Ar) and adipocyte area/bone area (Ad.Ar/B.Ar) (n � 6). I and J,Osteoblast differentiation was enhanced for MSCs from MSC-gata2-KO mice, shown by theincreased ALP staining (I) and the higher expression of osteoblastogenic markers such asosteocalcin and Col1a1 (J) (n � 3). K, Femurs from MSC-gata2-KO mice displayed higherosteoblast number/bone area (Ob.N/B.Ar) and osteoblast surface/bone surface (Ob.S/B.S) (n � 6).L, ELISA analysis showed that serum bone formation marker P1NP was higher in MSC-gata2-KOmice (n � 6). M, �CT analysis of the trabecular bone in proximal tibiae showed that MSC-gata2-KO mice exhibited higher bone mass (n � 4). BV/TV, bone volume/tissue volume ratio; Ctrl,control; Tb.N, trabecular number; Tb.Sp, trabecular separation. Error bars, SD.



not affect Gata2 expression in MSCs (Figure 3B). As aresult, ex vivo adipogenesis was augmented in the Ad-gata2-KO cultures compared with controls, shown by theincreased expression of adipoctye markers such as adi-ponectin and CEBP� (Figure 3C) as well as more ORO�

cells (Figure 3D).

In vivo analysis by ORO stainingrevealed that femur sections fromAd-gata2-KO mice exhibited moreabundant marrow fat comparedwith littermate control mice (Figure3E). In addition to marrow adi-pocytes, there are other types of ad-ipose tissue such as brown adiposetissue (BAT) and peripheral whiteadipose tissue (WAT) that displaydifferent gene expression profile andfunctions in the body. Thus, we de-cided to examine how Gata2 dele-tion affects peripheral WAT andBAT development in vivo. Hematox-ylin and eosin staining showed thatboth sc WAT and visceral WAT fromAd-gata2-KO mice contained moremature adipocytes compared withthe littermate control mice, evi-denced by the larger size, indicatingan enhanced WAT adipogenesis(Figure 3E). Furthermore, BAT fromAd-gata2-KO mice exhibited lesslipid droplets, indicating an aug-mented BAT development (Figure3E). In support of the histology ob-servations, gene expression analysisshowed thatboth scWATandvisceralWAT showed significantly higher ex-pression of adipose markers such asadiponectin and PPAR�2 (Figure 3, Fand G). Moreover, BAT from Ad-gata2-KO mice also expressed ahigher level of uncoupling protein-1(UCP1), a marker for thermogenesis(Figure 3H). Subcutaneous WATmass was significantly higher in theAd-gata2-KOmicecomparedwiththecontrols, whereas visceral WAT mass,BAT mass, or body weight did notshow significant differences (Figure 3,I and J). These in vivo observationsfurther support the notion that Gata2loss in the adipocyte lineage enhancesadipogenesis.

Given that the adiponectin-Cre driver specifically de-leted Gata2 in adipocyte but not in MSCs (Figure 3A),these findings indicate that Gata2 has a secondary func-tion in the adipocyte lineage to further suppress adipo-genesis. Our genetic and in vivo observations are in linewith the report by Tong et al (8, 9), which shows that

A B C D

E

F G H

I J

Figure 3. Gata2 in adipocyte lineage suppresses adipogenesis. A and B, Gata2 expression wassignificantly diminished in adipocytes (A) but not MSCs (B) derived from the bone marrow of Ad-gata2-KO mice (n � 3). C and D, Ex vivo bone marrow adipocyte differentiation was enhancedfor Ad-gata2-KO mice compared with littermate controls, shown by the higher expression ofadipogenic markers such as adiponectin and CEBP� (C), and more ORO staining in thedifferentiation cultures (D) (n � 3). E, Ad-Gata2-KO mice exhibited enhanced adipogenesis invivo, shown by the more abundant marrow adipocytes and larger adipocytes in subcutaneousand visceral WAT as well as more differentiated BAT. Representative images for ORO orhematoxylin and eosin staining are shown. Scale bars, 25 �m. F and G, Expression of WATmarkers such as adiponectin and PPAR�2 was higher in the sc WAT (F) and visceral WAT (G) fromAd-gata2-KO mice compared with controls (n � 3). H, Expression of BAT markers such as UCP1was higher in the BAT from Ad-gata2-KO mice compared with controls (n � 3). I, SubcutaneousWAT from Ad-gata2-KO mice was heavier compared with controls, whereas visceral WAT andBAT weights were similar (n � 3). J, Body weight was not significantly altered in Ad-gata2-KOmice (n � 3). Ctrl, control. Error bars, SD.



Gata2 inhibits preadipocyte to adipocyte transition invitro by suppressing PPAR� and CEBPs, as well as thereport by Tsai et al (24), which shows that Gata2 alsoinhibits BAT differentiation in vitro by suppressing bothUCP1 and peroxisomal proliferator-activated receptor-�coactivator 1� promoters.

Gata2 deletion in osteoblastlineage promotes osteoblastdifferentiation

We have uncovered a novel func-tion of Gata2 to impede osteoblas-togenesis. This is partially attributedto the ability of Gata2 to sustainMSC proliferation and thereby pre-vent differentiation. Because the ex-pression of Gata2 increases duringosteoblast differentiation whereas itdecreases during MSC proliferation,we decided to further investigatewhether Gata2 plays a synergistic oropposite role in the osteoblast lin-eage that may be independent ofthose in MSCs. To achieve this goal,we generated osteoblast-specificconditional knockout (Ob-gata2-KO) mice driven by the previouslyestablished osteocalcin-Cre (15).Our results showed that the osteo-calcin-Cre driver successfully low-ered Gata2 expression in osteoblastwithout affecting Gata2 levels inMSCs (Figure 4A). Ex vivo bonemarrow differentiation assays re-vealed that Ob-gata2-KO culturesdisplayed enhanced osteoblastogen-esis, shown by the significantlyhigher expression of osteoblastmarker genes such as Col1a1 and os-teocalcin (Figure 4B). In vivo analy-sis also showed that osteoblast num-ber and surface were significantlygreater in the femurs sections fromOb-gata2-KO mice compared withcontrols (Figure 4C), leading to anincreased bone formation (Figure4D) and a higher bone mass (Figure4E). Together these findings sug-gest that Gata2 loss in the osteo-blast lineage facilitates osteoblastdifferentiation.

We next sought to explore themechanism of how Gata2 in the os-

teoblast lineage inhibits osteoclast differentiation. Runx2is a key transcription factor that promotes osteoblasto-genesis. We found that both human and mouse Runx2promoters contain multiple GATA binding motifs, such asthe examples shown in Figure 4F, suggesting that Runx2may be a Gata2 target gene. Indeed, we found that Gata2

A B C

D E F

G I

H

Figure 4. Gata2 in osteoblast lineage suppresses osteoblastogenesis. A, Gata2 expression wassignificantly diminished in osteoblasts (left) but not MSCs (right) derived from the bone marrowof Ob-gata2-KO mice (n � 3). B, Ex vivo bone marrow osteoblast differentiation was enhancedfor Ob-gata2-KO mice compared with littermate controls, shown by the higher expression ofosteoblast markers such as osteocalcin and Col1a1 (n � 3). C, Femurs from Ob-gata2-KO micedisplayed higher osteoblast number/bone area (Ob.N/B.Ar) and osteoblast surface/bone surface(Ob.S/B.S) (n � 6). D, ELISA analysis showed that serum bone formation marker P1NP was higherin Ob-gata2-KO mice (n � 6). E, �CT analysis of the trabecular bone in proximal tibiae showedthat Ob-gata2-KO mice exhibited higher bone mass (n � 4). BV/TV, bone volume/tissue volumeratio. F, Multiple GATA binding motifs were identified in the Runx2 promoter in human, mouse,and zebrafish. Examples are shown. G, Runx2 expression was increased in osteoblastsdifferentiated from the bone marrow of Ob-gata2-KO mice (n � 3). H, Gata2 dose dependentlysuppressed �-catenin activity in a transient transfection and TOP/FOP reporter assay (n � 6). I, Aworking model for how Gata2 regulates mesenchymal cell fate. Gata2 in the early mesenchymallineage promotes MSC proliferation and suppresses terminal differentiation. Gata2 in thecommitted osteoblast and adipocyte lineages also inhibit differentiation by suppressing keytranscription factors: in the osteoblast lineage, Gata2 down-regulates �-catenin and Runx2; inthe adipocyte lineage, Gata2 attenuates PPAR�2 and CEBPs. Ctrl, control.



loss led to an increased Runx2 mRNA expression in os-teoblasts (Figure 4G), indicating that Gata2 inhibition ofosteoblastogenesis is partially mediated by repressingRunx2 transcription. In addition to Runx2, Wnt/�-catenin signaling also boosts osteoblast differentiation(25). To determine the effects of Gata2 on �-catenin ac-tivity, we performed transient transfection and a TOP/FOP �-catenin reporter assay. Compared with vector con-trol, Gata2 significantly decreased �-catenin activity in adose-dependent manner (Figure 4H), indicating thatGata2 inhibition of osteoblastogenesis also may be con-tributed by suppressing �-catenin signaling.

Discussion

In this study, we have provided genetic and in vivo evi-dence that Gata2 is a crucial and physiologically signifi-cant rheostat of MSC fate and mesenchymal lineage de-velopment. Moreover, we have uncovered multiplecellular and molecular mechanisms that confer Gata2 reg-ulation. In MSCs, Gata2 sustains stem cell proliferation toprevent differentiation (Figure 4J). In lineage-committedcells, Gata2 further inhibits differentiation by impedingkey transcription factors and signaling pathways: in adi-pocytes, Gata2 suppresses PPAR� and CEBPs; in osteo-blasts, Gata2 suppresses Runx2 and �-catenin (Figure 4I).The notion that Gata2 targets Runx2 expression is furthersupported by our analysis of the Encyclopedia of DNAElements chromatin immunoprecipitation-sequencing da-tabase (26), which shows that Gata2 can bind to multiplesites at the human Runx2 promoter and enhancer region.Together these findings provide important new insights toMSC function and cellular differentiation, with potentialimplications in tissue regeneration and degenerativediseases.

Gata2’s expression pattern is controlled by its upstreamregulators, whereas its roles in proliferation and differen-tiation are related to its positive or negative effects ondownstream targets. These two readouts do not necessar-ily correlate with or contradict each other. Our observa-tion that Gata2 decreases during MSC expansion yet pro-motes proliferation suggests that Gata2 expression inearly MSCs is required for MSC proliferation, but its ex-pression may be down-regulated by a negative feedbackloop at the late stage of MSC expansion. Similarly, ourobservation that Gata2 increases during differentiationyet suppresses adipogenesis and osteoblastogenesis is con-sistent with our findings that Gata2 is a negative regulatorthat inhibits positive transcription factors such as PPAR�,CEBPs, Runx2, and �-catenin, as our data and model haveshown; therefore, the Gata2 level must be low at the be-

ginning of the differentiation to derepress these positivetranscription factors and trigger the differentiation pro-gram, but Gata2 expression gradually bounces back totimely resolve the signaling and cease the differentiationpossibly due to a negative feedback loop at late stage ofdifferentiation; as a result, Gata2 loss enhances differen-tiation. This dynamic regulation has been shown in manyother cell types. For example, the level of the master proos-teoclastogenic transcription factor Nuclear Factor of Ac-tivated T-Cells, Cytoplasmic, Calcineurin-Dependent 1 ishigh at the early stage of osteoclast differentiation, yet itsexpression decreases during the late stage of osteoclastdifferentiation (27–29). In addition, we have recentlyshown that the expression of the nuclear receptor Nur77increases during osteoclastogenesis, yet it suppresses os-teoclast differentiation by triggering Nuclear Factor of Ac-tivated T-Cells, Cytoplasmic, Calcineurin-Dependent 1degradation (28).

Gata2 functions in MSCs may exert an impact beyondmesenchymal lineages. MSCs, osteoblasts, and adipocyteshave all been shown to help to form and support the HSCniche (30, 31). Therefore, Gata2 in MSC and mesenchy-mal cell types may indirectly regulate HSCs and hemato-poietic lineages. Gata2 loss in mesenchymal lineages maydisrupt the HSC niche, further worsening the dysregula-tion of blood cells caused by Gata2 loss in hematopoieticlineages. Furthermore, marrow fat infiltrations are com-mon features in both myeloplastic syndrome and aplasticanemia, both of which have been related to Gata2 defi-ciency. In these cases, Gata2 loss in the mesenchymal cellcompartment could exacerbate the deficiency in bloodcells by up-regulating marrow adipocyte differentiation.

In addition to adipocyte and osteoblast, other lineagessuch as chondrocyte and myocyte can also be differenti-ated from MSCs. A recent study suggests that Gata2 dys-regulation may be a marker for skeletal muscle hypertro-phy (32). In future studies, it will be interesting to seewhether and how Gata2 can also regulate these other mes-enchymal lineages. Knowledge of how Gata2 controlsMSC expansion, lineage commitment, and cellular differ-entiation will educate us to design better strategies for stemcell therapy and regenerative medicine. To this end,K-7174 and K-11706 are two compounds that can inhibitGata2 activity (33), which may hold exciting potential asnovel bone anabolic agents for the treatment of skeletaldegenerative diseases such as osteoporosis.

Acknowledgments

Y.W. is a Virginia Murchison Linthicum Scholar in MedicalResearch.



Address all correspondence and requests for reprints to: Yi-hong Wan, PhD, Department of Pharmacology, The Universityof Texas Southwestern Medical Center, 6001 Forest Park Road,Dallas, TX 75390. E-mail: [email protected].

This work was supported in part by National Institutes ofHealth Grant R01 DK089113 (to Y.W.), Cancer Prevention andResearch Institute of Texas Grant RP130145 (to Y.W.), Depart-ment of Defense Breast Cancer Research Program Idea AwardBC122877 (to Y.W.), March of Dimes Grant 6-FY13-137 (toY.W.), The Welch Foundation Grant I-1751 (to Y.W.), and theUniversity of Texas Southwestern Endowed Scholar StartupFund (to Y.W.).

Disclosure Summary: The authors have nothing to disclose.

References

1. Prockop DJ. Marrow stromal cells as stem cells for nonhematopoi-etic tissues. Science. 1997;276(5309):71–74.

2. James AW. Review of signaling pathways governing MSC osteo-genic and adipogenic differentiation. Scientifica. 2013;2013:684736.

3. Bornes TD, Adesida AB, Jomha NM. Mesenchymal stem cells in thetreatment of traumatic articular cartilage defects: a comprehensivereview. Arthritis Res Ther. 2014;16(5):432.

4. Perdisa F, Gostynska N, Roffi A, Filardo G, Marcacci M, Kon E.Adipose-derived mesenchymal stem cells for the treatment of artic-ular cartilage: a systematic review on preclinical and clinical evi-dence. Stem Cells Int. 2015;2015:597652.

5. Johnson KD, Hsu AP, Ryu MJ, et al. Cis-element mutated inGATA2-dependent immunodeficiency governs hematopoiesis andvascular integrity. J Clin Invest. 2012;122(10):3692–3704.

6. Rodrigues NP, Janzen V, Forkert R, et al. Haploinsufficiency ofGATA-2 perturbs adult hematopoietic stem-cell homeostasis.Blood. 2005;106(2):477–484.

7. Spinner MA, Sanchez LA, Hsu AP, et al. GATA2 deficiency: a pro-tean disorder of hematopoiesis, lymphatics, and immunity. Blood.2014;123(6):809–821.

8. Tong Q, Dalgin G, Xu H, Ting CN, Leiden JM, Hotamisligil GS.Function of GATA transcription factors in preadipocyte-adipocytetransition. Science. 2000;290(5489):134–138.

9. Tong Q, Tsai J, Tan G, Dalgin G, Hotamisligil GS. Interactionbetween GATA and the C/EBP family of transcription factors iscritical in GATA-mediated suppression of adipocyte differentiation.Mol Cell Biol. 2005;25(2):706–715.

10. Xu Y, Takahashi Y, Wang Y, et al. Downregulation of GATA-2 andoverexpression of adipogenic gene-PPAR� in mesenchymal stemcells from patients with aplastic anemia. Exp Hematol. 2009;37(12):1393–1399.

11. Kamata M, Okitsu Y, Fujiwara T, et al. GATA2 regulates differ-entiation of bone marrow-derived mesenchymal stem cells. Haema-tologica. 2014;99(11):1686–1696.

12. Haugas M, Lillevali K, Hakanen J, Salminen M. Gata2 is requiredfor the development of inner ear semicircular ducts and the sur-rounding perilymphatic space. Dev Dyn. 2010;239(9):2452–2469.

13. Logan M, Martin JF, Nagy A, Lobe C, Olson EN, Tabin CJ. Ex-pression of Cre recombinase in the developing mouse limb buddriven by a Prxl enhancer. Genesis. 2002;33(2):77–80.

14. Eguchi J, Wang X, Yu S, et al. Transcriptional control of adiposelipid handling by IRF4. Cell Metab. 2011;13(3):249–259.

15. Zhang M, Xuan S, Bouxsein ML, et al. Osteoblast-specific knockoutof the insulin-like growth factor (IGF) receptor gene reveals an es-sential role of IGF signaling in bone matrix mineralization. J BiolChem. 2002;277(46):44005–44012.

16. Krzeszinski JY, Wei W, Huynh H, et al. miR-34a blocks osteopo-rosis and bone metastasis by inhibiting osteoclastogenesis and Tgif2.Nature. 2014;512(7515):431–435.

17. Wei W, Dutchak PA, Wang X, et al. Fibroblast growth factor 21promotes bone loss by potentiating the effects of peroxisome pro-liferator-activated receptor gamma. Proc Natl Acad Sci USA. 2012;109(8):3143–3148.

18. Wan Y, Saghatelian A, Chong LW, Zhang CL, Cravatt BF, EvansRM. Maternal PPAR � protects nursing neonates by suppressing theproduction of inflammatory milk. Genes Dev. 2007;21(15):1895–1908.

19. Wei W, Zeve D, Suh JM, et al. Biphasic and dosage-dependent reg-ulation of osteoclastogenesis by �-catenin. Mol Cell Biol. 2011;31(23):4706–4719.

20. Wei W, Zeve D, Wang X, et al. Osteoclast progenitors reside in theperoxisome proliferator-activated receptor �-expressing bone mar-row cell population. Mol Cell Biol. 2011;31(23):4692–4705.

21. Krum SA, Miranda-Carboni GA, Hauschka PV, et al. Estrogen pro-tects bone by inducing Fas ligand in osteoblasts to regulate osteoclastsurvival. EMBO J. 2008;27(3):535–545.

22. Jin Z, Wei W, Dechow PC, Wan Y. HDAC7 inhibits osteoclasto-genesis by reversing RANKL-triggered �-catenin switch. Mol En-docrinol. 2013;27(2):325–335.

23. McNeill RE, Miller N, Kerin MJ. Evaluation and validation of can-didate endogenous control genes for real-time quantitative PCRstudies of breast cancer. BMC Mol Biol. 2007;8:107.

24. Tsai J, Tong Q, Tan G, Chang AN, Orkin SH, Hotamisligil GS. Thetranscription factor GATA2 regulates differentiation of brown adi-pocytes. EMBO Rep. 2005;6(9):879–884.

25. Case N, Rubin J. �-catenin—a supporting role in the skeleton. J CellBiochem. 2010;110(3):545–553.

26. Gerstein MB, Kundaje A, Hariharan M, et al. Architecture of thehuman regulatory network derived from ENCODE data. Nature.2012;489(7414):91–100.

27. Kim JH, Kim K, Jin HM, et al. Negative feedback control of oste-oclast formation through ubiquitin-mediated down-regulation ofNFATc1. J Biol Chem. 2010;285(8):5224–5231.

28. Li X, Wei W, Huynh H, Zuo H, Wang X, Wan Y. Nur77 preventsexcessive osteoclastogenesis by inducing ubiquitin ligase Cbl-b tomediate NFATc1 self-limitation. eLife. 2015;4:e07217.

29. Takayanagi H, Kim S, Koga T, et al. Induction and activation of thetranscription factor NFATc1 (NFAT2) integrate RANKL signalingin terminal differentiation of osteoclasts. Dev Cell. 2002;3(6):889–901.

30. Bianco P. Bone and the hematopoietic niche: a tale of two stem cells.Blood. 2011;117(20):5281–5288.

31. Valtieri M, Sorrentino A. The mesenchymal stromal cell contribu-tion to homeostasis. J Cell Physiol. 2008;217(2):296–300.

32. Musaro A, McCullagh KJ, Naya FJ, Olson EN, Rosenthal N. IGF-1induces skeletal myocyte hypertrophy through calcineurin in asso-ciation with GATA-2 and NF-ATc1. Nature. 1999;400(6744):581–585.

33. Nakano Y, Imagawa S, Matsumoto K, et al. Oral administration ofK-11706 inhibits GATA binding activity, enhances hypoxia-induc-ible factor 1 binding activity, and restores indicators in an in vivomouse model of anemia of chronic disease. Blood. 2004;104(13):4300–4307.



mailto:[email protected]

Gata2 Is a Rheostat for Mesenchymal Stem Cell Fate in Male ... · Gata2 Is a Rheostat for...

Documents

Transcript of Gata2 Is a Rheostat for Mesenchymal Stem Cell Fate in Male ... · Gata2 Is a Rheostat for...