NEOPLASIA

SALL4 is a key regulator of survival and apoptosis in human leukemic cellsJianchang Yang,1 Li Chai,2 Chong Gao,2 Taylor C. Fowles,1 Zaida Alipio,1 Hien Dang,1 Dan Xu,1 Louis M. Fink,1

David C. Ward,1 and Yupo Ma1

1Division of Laboratory Medicine, Nevada Cancer Institute, Las Vegas; and 2Department of Pathology, Joint Program in Transfusion Medicine, Brigham andWomen’s Hospital/Children’s Hospital Boston, Harvard Medical School, Boston, MA

Increasing studies suggest that SALL4may play vital roles in leukemogenesisand stem cell phenotypes. We havemapped the global gene targets of SALL4using chromatin immunoprecipitationfollowed by microarray hybridizationand identified more than 2000 high-confidence, SALL4-binding genes in thehuman acute promyelocytic leukemic cellline, NB4. Analysis of SALL4-binding sitesreveals that genes involved in cell death,cancer, DNA replication/repair, and cellcycle were highly enriched (P < .05).

These genes include 38 importantapoptosis-inducing genes (TNF, TP53,PTEN, CARD9, CARD11, CYCS, LTA) andapoptosis-inhibiting genes (Bmi-1, BCL2,XIAP, DAD1, TEGT). Real-time polymer-ase chain reaction has shown that expres-sion levels of these genes changed signifi-cantly after SALL4 knockdown, whichubiquitously led to cell apoptosis. Flowcytometry revealed that reduction ofSALL4 expression in NB4 and other leuke-mia cell lines dramatically increasedcaspase-3, annexin V, and DNA fragmen-

tation activity. Bromodeoxyuridine-incorporation assays showed decreasednumbers of S-phase cells and increasednumbers of G1- and G2-phase cells indi-cating reduced DNA synthesis, consis-tent with results from cell proliferationassays. In addition, NB4 cells that ex-press low levels of SALL4 have signifi-cantly decreased tumorigenecity in immu-nodeficient mice. Our studies provide afoundation in the development of leuke-mia stem cell–specific therapy by target-ing SALL4. (Blood. 2008;112:805-813)

Introduction

SALL4 is a zinc-finger transcription factor and a member of theSALL gene family, which was originally cloned based on sequencehomology to Drosophila spalt (sal).1-3 In Drosophila, sal is ahomeotic gene and essential in the development of posterior headand anterior tail segments.4

Human SALL4 mutations are associated with the Duane-radialray syndrome (DRRS; also called Okihiro syndrome), a humanautosomal-dominant syndrome involving multiple organ de-fects.3,5,6 Our group and others have shown that murine SALL4plays an essential role in maintaining the properties of embryonicstem (ES) cells and governing the fate of the primitive inner cellmass by interacting with Oct4 and Nanog.7-11 Interestingly, SALL4protein expression is correlated with stem and progenitor cellpopulations in various organ systems including bone marrow.8,12,13

We have demonstrated previously that SALL4 is expressed consti-tutively in human leukemia cell lines and primary acute myeloidleukemia (AML) cells derived from patient samples.12 Transgenicmice that overexpress SALL4B, one of the SALL4 isoforms,exhibit myelodysplastic syndrome (MDS)–like symptoms andsubsequently develop AML that is transplantable.12

The polycomb gene Bmi-1 has been identified recently as amajor downstream target of SALL4.14 Bmi-1 is an apoptosisregulator that has been well studied in many models includingleukemia, the epidermis, neural lineages, and recently prostatecancer.15-21 Bmi-1 plays an essential role in regulating adult,self-renewing, hematopoietic stem cells (HSCs) and leukemic stemcells (LSCs).18,22-27 Bmi-1 is highly expressed in purified HSCs,and this expression declines with differentiation.23 Knockout of

Bmi-1 in mice results in a progressive loss of all hematopoieticlineages. This loss results from the inability of the Bmi-1�/� stemcells to self-renew.18 The presence of Bmi-1 appears to beimportant in the maintenance of LSC identity. Recently, Bmi-1expression has been used as an important marker for predictingprogression of MDS to AML.28

Despite the prominent roles of SALL4 in ES cells and leukemiccells, the mechanisms controlling cell growth, proliferation, andapoptosis through the downstream targets of SALL4 remain largelyunexplored. We undertook a genome-wide analysis of gene promot-ers directly or indirectly bound by SALL4 using chromatinimmunoprecipitation followed by microarray hybridization (ChIP-chip) to determine potential mechanism whereby SALL4 regulatesproliferation and apoptosis. Our results indicate that SALL4 acts asa key regulator of cell growth and apoptosis in leukemic cells.Massive apoptosis and significant growth arrest occurs whenSALL4 expression is down-regulated in various leukemic cells. Inaddition, reduction of SALL4 markedly diminishes tumorigenicityof leukemic cells in immunodeficient mice.

Methods

Mice

Nonobese diabetic–severe combined immunodeficient (NOD.SCID)/NCrmice (6-8 weeks old) were maintained under specific pathogen-freeconditions. All animal experiments were preapproved by the Office ofLaboratory Animal Welfare, Institutional Animal Care and Use Committee.

Submitted November 30, 2007; accepted April 19, 2008. Prepublished online asBlood First Edition paper, May 16, 2008; DOI 10.1182/blood-2007-11-126326.

The online version of this article contains a data supplement.

The publication costs of this article were defrayed in part by page chargepayment. Therefore, and solely to indicate this fact, this article is herebymarked ‘‘advertisement’’ in accordance with 18 USC section 1734.

© 2008 by The American Society of Hematology

805BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

Cell culture and ChIP

Human acute promyelocytic leukemia NB4 cells were maintained inRPMI-1640 supplemented with 10% (vol/vol) fetal bovine serum (FBS),and ChIP assays were performed as previously described.14

ChIP-chip

A complete protocol was provided by NimbleGen Systems (Madison, WI). Inbrief, NB4 cells were grown, cross-linked with formaldehyde, and sheared bysonication.An anti-SALL4 antibody and nonimmune rabbit serum were used forimmunoprecipitation.12,29 ChIP-purified DNAwas blunt-ended using T4 polymer-ase, ligated to linkers, and subjected to low-cycle polymerase chain reaction(PCR) amplification. Promoter tiling arrays were produced by NimbleGenSystems. The RefSeq human promoter array is a single array containing 2.7 kbsequence with 2.2 kb representative of the proximal promoter region and 500 bpof the 5� terminal coding sequence (from National Center for BiotechnologyInformation [NCBI, Bethesda, MD] Build 36; HG18). The promoter region iscovered by 50- to 75-mer probes at roughly 100-bp spacing depending on thesequence composition of the region. The arrays were hybridized, and the datawere extracted according to standard procedures by NimbleGen Systems.

Data extraction was done using NimbleScan by NimbleGen Systems.NimbleScan searches for 4 or more probes above a cutoff value ranging from90% to 15% using a 500-bp sliding window. The cutoff value is a percentage ofthe hypothetical maximum determined using the mean plus 6 standard devia-tions, and under default scanning conditions decreases in 1% increments from90% to 15%. The data are then randomized 20 times to evaluate the possibility offalse positives, and each peak is assigned a false discovery rate (FDR) based onthis randomization. Predicted binding sites were confirmed using quantitativereal-time (Q-RT)–PCR. Negative control primers were designed flanking pre-dicted SALL4-binding sites.

Retrovirus production and SALL4 knockdown

Three short-hairpin RNA-expressing plasmids, 1 control (pRS) and2 SALL4 specific (no. 7410, no. 7412; all 3 from Origen, Rockville, MD),were transfected into Phoenix packaging cells (Orbigen, San Diego, CA)using Lipofectamine 2000 (Invitrogen, Frederick, MD). Shed virus washarvested 48 hours after transfection, and control or stable SALL4knockdown NB4 clones were obtained under puromycin (1.2 �g/mL)selection after 7 days.

Quantitative real-time PCR

Total RNA extraction and Q-RT-PCR assays were performed as previouslydescribed.12

Microarray-based mRNA profiling

A catalog human microarray (NCBI Build 36; HG18) was manufactured byNimbleGen Systems. The human genes in this design (n � 47 633) were selectedfrom the Homo sapiens entries in the RefSeq collection, full-length humanmRNAs, and experimental gene models from NCBI. Nine 60-mer probes werepositioned 1.5-kb upstream from the 3� end of each target. Probes were spacedevenly over the length of the target region (� 1.5 kb) where possible so that theexact spacing is dependent on the length of the target sequence. Redundant geneswere removed by identifying identical target probe sets and placing only oneexemplar sequence on the array for each probe set. Signal intensities werecalculated as the mean 3 � 3-grid pixel comprising each feature. Gene expres-sion analysis was done using robust multichip average.

Expression microarrays (NimbleGen Systems) were used to examine geneexpression patterns before and after RNA interference in NB4 cells. After virusinfection, up to 75% of endogenous SALL4 expression was reduced as observedusing Q-RT-PCR analysis. The experiment was repeated 3 times, and the totalRNA isolated from each experiment was pooled (both control and knockdowngroups) for expression assays. Ingenuity Systems (version 2.01; Redwood City,CA) software was used to analyze the hybridization results, and only geneswhose expression levels were altered more than 2-fold were considered

differentially expressed and singled out for further analysis. All microarray datafor this paper have been deposited in the Gene Expression Omnibus (GEO) underaccession number GSE10734.30

Apoptosis, cell proliferation, and cell-cycle arrest assays

Cultured cells (approximately 106 cells) were treated with control or SALL4shRNA retroviruses. Cell viability was evaluated and quantified by cell countingfollowing trypan blue exclusion. To quantify apoptosis, cells were harvested forflow cytometry-based caspase-3, annexin V/propidium iodide (Pharmingen, BDBiosciences, San Diego, CA), and DNA fragmentation assays (ApoAlert DNAFragmentation Assay Kit; Clontech, Palo Alto, CA) through flow cytometry. ForDNA synthesis and cell-cycle arrest assays, a portion of the harvested cells wereused to incorporate BrdU (Pharmingen, BD Biosciences) following the manufac-turer’s instructions.

Statistics

Ingenuity Pathway Analysis (Ingenuity Systems, http://www.ingenuity.com) tools were used for ChIP-chip data. Preprocessed data files from Nimble-Gen Systems were filtered to remove duplicate genes according to the NCBIEntrez Gene ID setting.31 The list of genes was then inputted into IngenuityPathwayAnalysis with the NCBI Entrez Gene ID as the observed gene identifier.A more stringent FDR less than 0.10 was used for pathway analysis. Genes wereanalyzed according to relevant biologic functions and pathways.

Preprocessed gene expression data from NimbleGen Systems was usedto convert to the log2 ratio. Two controls (C) and 2 treated (T) data sets thenwere used to generate 4 ratios according to the following expression: log2

(T/C) � log2 (T) � log2 (C). It is necessary to do this to obtain more valuesfor this sample t test. A one-sample t test was performed using TigrMultiExperiment Viewer (TIGR) with P less than .05.32 The significantgene list was inputted into the Ingenuity Pathway Analysis tool.

Results

SALL4 regulatory networks in leukemic cells

Previously we have demonstrated that SALL4 plays an important role inleukemogenesis. To identify the molecular pathway(s) that regulateSALL4-mediated leukemogenesis, we used ChIP-chip as a screeningtool in the AML-derived cell line NB4, which expresses relatively highlevels of SALL4. Using highly specific anti-SALL4 antibodies, weidentified more than 2500 target genes bound by SALL4 withinpromoter regions spanning 2.2-kb upstream and 500-bp downstream ofthe transcription start site (Table S1, available on the Blood website; seethe Supplemental Materials link at the top of the online article). Wechose a promoter array for the ChIP-chip because it has been shown tocontain approximately 90% of transcription factor–binding sites.33 Toverify that the SALL4-binding sites identified with the RefSeq arraywere authentic, we performed independent ChIP experiments coupledwith a quantitative real-time PCR (ChIP-Q-RT-PCR) in NB4 cells.Through this strategy, 81 promoter regions were subjected to ChIP-enrichment validation. We successfully validated the binding of SALL4to 39 randomly chosen promoter regions (39/43) in addition to35 apoptosis genes (35/38, 91% of the selected targets total; Figure 1;Figure S2). Thus, using ChIP-chip assays, roughly 91% of identifiedpromoters were considered genuine SALL4-binding loci in NB4 cells.

We next categorized the signaling pathways associated with theSALL4-binding sites identified through ChIP-chip assay. UsingIngenuity Pathway Analysis, we determined that SALL4-bindinggenes are involved in more than 30 different signaling pathways.The most prominent of these pathways was the Wnt, apoptosis,PTEN, NF-�B, and p53 signaling pathways, each of which playimportant roles in apoptosis (Figure 2A). We also assessedpathway distributions of ChIP-chip data for SALL4-binding

806 YANG et al BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

genes in primary AML (French-American-British classification:M0) and chronic acute leukemic transformation leukemic cells(data not shown). These results demonstrated similar pathwaydistributions as the NB4 cells. These observations suggest thattargets of SALL4 identified in NB4 cells will likely regulate abroader spectrum of AML cell types and is consistent with ourprevious suggestion that SALL4 plays a role in the transforma-tion of different types of AML.12

We further classified the SALL4-binding genes based on theirmolecular and cellular functions. Interestingly, cell death and

cancer genes were among the highly enriched groups with P valuesless than 10�11. Other classifications that were enriched includedcell cycle, and hematologic-related genes (Figure 2B; Table S2).These functional distributions suggest that SALL4 is tightlyassociated with leukemic cell growth.

SALL4 directly regulates expression of apoptosis genes

Substantial numbers of genes that are regulators of apoptosis werefound in the top tier of genes bound by SALL4 in the ChIP-chip

Figure 1. Validation of ChIP-chip–identified SALL4-binding sites by Q-RT-PCR. Chromatin immunoprecipitation experiments were performed in NB4 cells with antibodiesraised against SALL4. Forty-three randomly selected promoters identified as SALL4 targets with ChIP-chip arrays were analyzed by Q-RT-PCR. ChIP-Q-RT-PCR identified39 targets having more than 2-fold enrichment versus the input control. The fold change of SALL4 immunoprecipitated DNA over the input is presented on a log2 scale.Negative control primers were designed based on the genomic regions surrounding the peak as shown in Figure S2.

Figure 2. Classification of SALL4-bound genes inNB4 leukemia cells. SALL4 target genes were classi-fied based on annotations important to cell growth andcell death extracted from Ingenuity Pathway Analysis. Itshould be noted that specific genes may be representedin more than one classification group. (A) Classificationof SALL4 target genes is based on signaling pathwaysassociated apoptosis and cell growth properties accord-ing to Ingenuity Pathway Analysis knowledge base.P values for this analysis are not presented due to thelow number of genes within each pathway. The first andsecond percentages represent the number of genesbound within each pathway relevant to the total SALL4binding and the total number of genes within eachpathway relevant to the genes on the array, respectively.(B) Classification of molecular functions associated withSALL4 target genes. The P values presented for eachcategory are calculated using Fisher exact test againstthe genes represented on the promoter tiling array. Thefirst and second percentages are the total number ofgenes bound by SALL4 in each function relevant to thenumber of genes bound by SALL4 and the number ofgenes on the array that are classified in each functioncompared with the total number of genes representedon the array, respectively. Unannotated genes wereremoved from this analysis, and the analysis was doneusing the top tier of SALL4-bound genes. The genesbound by SALL4 and the genes within each classifica-tion are presented in Table S2.

SALL4 IS A KEY REGULATOR OF SURVIVAL AND APOPTOSIS 807BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

assays. These genes include apoptosis-inducing genes, such asTNF, LTA, TNFRSF10D, TP53, PTEN, CARD9, CARD11, CASP6,and ATF3; apoptosis-inhibiting genes, such as BCL2, TEGT,Bmi-1, BIRC4, and BIRC7; and apoptotic process related genessuch as CYCS and BAT3 (Figure 3). A detailed list of these genes isprovided in Table 1. These genes were of primary importance inregulating cell survival and may be responsible for the role thatSALL4 plays in leukemogenesis. For this reason, we sought toconfirm SALL4 binding to promoter regions of these genes using areplicate preparation of ChIP DNA. Primers were designed basedon locations of peaks from the ChIP-chip data and were used forQ-RT-PCR verification of immunoprecipitation. The resulting dataestablished that 35 of 38 apoptosis genes were indeed targets ofSALL4 in NB4 cells (Figure 3A). We also sought to determinewhether these apoptosis genes were targets of SALL4 in anothercell line. The CD34� cell population derived from human primarybone marrow cells expresses high levels of SALL4. Using this cellpopulation, we demonstrate that 28 of 37 apoptosis genes areindeed targets of SALL4 in the CD34� fraction (Figure 3B). Thesedata strongly suggest that SALL4 binding to the promoters of asubset of apoptosis genes may account for the malignant phenotypein NB4 cells (as well as other leukemic cell types) and that aberrantexpression of these apoptosis genes may lead to this phenotype.

We next focused on interrogating the transcriptional effects ofSALL4 on these genes in leukemic cells. We analyzed “directtarget” genes of SALL4 associated with apoptosis using retrovirus-induced SALL4 reduction combined with Q-RT-PCR. The term“direct target” includes all promoters that are bound by SALL4,either by direct interaction with the DNA or indirectly as acomponent in DNA-protein–binding complexes. Interestingly,17 apoptosis-inducing genes, including TNF, TP53, PTEN, CARD9,CARD11, ATF3, and LTA were consistently up-regulated in NB4cells in which SALL4 expression was reduced by approximately75% compared with control NB4 cells (Figure 4B). By contrast,10 apoptosis-inhibiting genes, including Bmi-1, BCL2, DAD1,

TEGT, BIRC7, and BIRC4 (XIAP) were consistently down-regulated in the same experiments. Notably, SALL4 appears tobind and regulate multiple target genes in all major apoptoticpathways (Figure 4C). The one notable outlier was the apoptosisinhibition gene MDM4, whose expression was up-regulated. How-ever, the functional role of MDM4 acting in apoptosis inhibition iscontroversial.35 Nevertheless, these expression studies indicate thatSALL4 is a key regulator of apoptosis and cell growth byup-regulating genes necessary for apoptosis and down-regulatinggenes that inhibit apoptosis.

Down-regulated SALL4 induces apoptosis and cell-cycle arrestin NB4 cells

To further test our hypothesis that SALL4 acts as a key regulator inleukemic cell apoptosis and cell growth and determine possiblyphenotypic consequences of this regulation, we adopted a RNAinference strategy to reduce SALL4 expression that was followedby flow cytometric assays in NB4 cells. Two short-hairpin (shRNA)retroviral constructs that target different regions of the SALL4mRNA, termed no. 7410 and no. 7412, were shown to reduceSALL4 mRNA in NB4 cells by Q-RT-PCR (Figure S1A). TheshRNA retroviral construct, no. 7412, was more effective insuppressing SALL4 expression and it was used in all furtherexperiments. We first performed flow cytometric assays to monitorthe activity of caspase-3 in different cell lines. Caspase-3 is one ofthe key protein markers for early stages of apoptosis. For virus-treated NB4 cells that express approximately 25% of wild-type(WT) SALL4 levels, there was a 21-fold increase in caspase-3activity. The number of cells expressing caspase-3 went from 4.6%in WT cells to 98.3% in SALL4-deficient cells as measured by flowcytometry (Figure 5A,B). The apoptotic effect was further verifiedusing data based on expression of the later-stage apoptosis markerannexin V (Figure S3). Annexin V revealed that approximately75% of the SALL4-reduced cells were apoptotic, whereas only

Figure 3. ChIP-Q-RT-PCR of apoptosis genes asso-ciated with SALL4 binding. Thirty-eight apoptosisgenes identified by ChIP-chip were subjected toChIP-Q-RT-PCR to confirm SALL4 binding by an alterna-tive method. These genes are candidates for SALL4involvement in leukemogenesis. (A) In NB4 cells, 35 ofthe 38 genes identified by ChIP-chip were confirmed asbound by SALL4 using ChIP-PCR. (B) Identification ofSALL4 binding to apoptosis genes in CD34� cells sortedfrom human bone marrow samples. CD34� cells arelikely progenitor cells to the promyelocytic cell line NB4within these lineages. Identified SALL4 target genes hadmore than 2-fold enrichment compared with the inputcontrol. The position of negative control primers is pre-sented in Figure S1. In Figure 2B, primer neg-2 failed toamplify the experimental DNA but did amplify the inputDNA, indicating the absence of DNA template in theexperimental group. This is represented by a fold changeof less than 0.001.

808 YANG et al BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

12% of the controls cells were apoptotic (more than a 6-foldchange). DNA fragmentation assays revealed an increase infragmenting DNA from 1.8% in the control cell population to 39%in the SALL4-reduced population (more than a 21-fold increase[Figure S4]). Another shRNA (no. 7410) against the SALL4mRNA was also able to produce the same apoptotic phenotype thatwas proportional to the degree of SALL4 reduction, suggesting thatthis apoptosis was not due to off-target effects (Figure 5; FigureS1B,C). Similar major increases in apoptosis were observed inother cell lines including KBM5 and NTERA2 when SALL4expression was decreased (data not shown). This observationstrongly suggests that the specific reduction of SALL4 in leukemiccells induces massive cell apoptosis.

We further monitored cell-cycle changes and cellular DNAsynthesis in SALL4-reduced and control NB4 cells throughbromodeoxyuridine (BrdU)–incorporation assays and flow cytom-etry. The NB4 cells that expressed roughly 25% of WT SALL4showed approximately a 4-fold decrease in S-phase cells and asignificant increase in the G1- and G2-phase cells (6- and 50-foldincreases, respectively), which paralleled the decrease in DNAsynthesis as measured by BrdU incorporation (Figure 5D,E). Weapplied these same studies on other cancerous human cells, such as

the embryonic carcinoma cell line NTERA2 and chronic myeloidleukemia cell line KBM5 (both purchased from ATCC, Manassas,VA), and similar results were observed (data not shown). Studies ofcell-cycle changes and cellular DNA synthesis were also consistentwith results from cell proliferation assays. As shown in Figure 5G,NB4 cells transduced with SALL4-specific shRNA retrovirusesgrew poorly during 5 days of virus infection. By comparison,growth of NB4 cells was not affected by infection with controlviruses. Taken together, these data clearly demonstrate that SALL4reduction can induce dramatic apoptosis and significant growtharrest in NB4 leukemia cells.

SALL4 knockdown-induced cell-cycle arrest and apoptosisreversed by Bmi-1

We have shown previously that Bmi-1 is a major downstream targetof SALL4 in leukemic cells.14 Bmi-1�/� cells display alteredexpression of the cell-cycle inhibitor gene p16INK4a, resulting in thepromotion of cell-cycle arrest and apoptosis. Given the apoptoticeffect caused by a reduction in SALL4, overexpression of Bmi-1, amajor SALL4 downstream target, may be able to rescue SALL4-induced apoptosis and cell-cycle arrest. SALL4-deficient NB4 cellswere transfected with an expression vector containing Bmi-1 andthe levels of caspase-3 activity, changes in cell cycle, and cellularDNA synthesis were measured by flow cytometry. As shown inFigure 5C, SALL4-induced caspase-3 activity was restored to anear-normal level by overexpression of Bmi-1 (compared withFigure 5A). Furthermore, reintroduction of Bmi-1 in SALL4-deficient NB4 cells resulted in an increase in the S-phase popula-tion and a decrease in the G1- and G2-phase populations (Figure5F). SALL4-deficient cells, where Bmi-1 was restored to arelatively normal level, incorporated BrdU in a manner similar tothe control NB4 cells. Interestingly, restoration of SALL4 func-tions occurred when the Bmi-1 expression level was corrected toapproximately 75% of the normal level (data not shown). This alsosuggests that cell apoptosis is due exclusively to the specificreduction of SALL4 mRNA. Overexpression of Bmi-1 had littleeffect on caspase-3 activity, cell-cycle changes, and DNA synthesisin control NB4 cells (data not shown). This may be due to thesecells having a high level of endogenous Bmi-1 expression.

Reduction of SALL4 diminishes tumorigenicity in humanleukemia cells

Decreased SALL4 expression in NB4 cells leading to massivecell apoptosis implies an essential role for SALL4 in thedevelopment of leukemia. To test this hypothesis in an animalmodel, we examined the effect of SALL4 knockdown on theability of NB4 AML cells to induce tumors in immunodeficientmice. NB4 AML cells infected with either control (pRS) orSALL4 shRNA-expressing retroviruses (no. 7412) were in-jected subcutaneously into the flanks of nonobese diabetic/severe combined immunodeficient (NOD/SCID) mice. Injec-tions were conducted using 3 � 106 or 1.3 � 106 cells. Miceinjected with control cells (pRS) at both doses developed largetumor masses after 3 weeks. Two of 4 mice injected withSALL4-reduced cells at the lower dose developed tumors,whereas mice injected with the higher dose developed tumors inall 4 animals. Both dosages showed significantly diminishedtumorigenic activity, producing tumors that were on averageapproximately 5 times smaller than tumors produced by thecontrol cells (Figure 6; P � .01). The difference in tumor size isquite large; however, it is not proportional to the degree of

Table 1. SALL4-bound genes related to apoptosis

Name Description

ABL1 v-abl Abelson murine leukemia viral oncogene homolog 1

ATF3 Activating transcription factor 3

BAG4 BCL2-associated athanogene 4

BAT3 HLA-B associated transcript 3

BCL2 B-cell CLL/lymphoma 2

BIRC4 Baculoviral IAP repeat-containing 4

BIRC7 Baculoviral IAP repeat-containing 7 (livin)

BMI1 B lymphoma Mo-MLV insertion region (mouse)

CARD11 Caspase recruitment domain family, member 11

CARD9 Caspase recruitment domain family, member 9

CISH Cytokine inducible SH2-containing protein

CUL3 Cullin 3

CYCS Cytochrome c, somatic

DAD1 Defender against cell death 1

DAPK3 Death-associated protein kinase 3

DAXX Death-associated protein 6

ENDOGL1 Endonuclease G-like 1

FASTK Fas-activated serine/threonine kinase

FGFRL1 Fibroblast growth factor receptor-like 1

FKBP8 FK506 binding protein 8, 38kDa

GDNF Glial cell derived neurotrophic factor

HIP1 Huntingtin interacting protein 1

HIP1R Huntingtin interacting protein 1 related

HNRPK Heterogeneous nuclear ribonucleoprotein K

LTA Lymphotoxin alpha (TNF superfamily, member 1)

MAGED1 Melanoma antigen family D, 1

MAP3K12 Mitogen-activated protein kinase kinase kinase 12

MDM4 Mdm4 transformed 3T3 cell double minute 4, p53 binding

protein (mouse)

PCBP4 Poly(rC) binding protein 4

PDCD4 Programmed cell death 4 (neoplastic transformation inhibitor)

PTEN Phosphatase and tensin homolog (mutated in multiple

advanced cancers 1)

SPHK2 Sphingosine kinase 2

STK17A Serine/threonine kinase 17a (apoptosis-inducing)

TEGT Testis enhanced gene transcript (BAX inhibitor 1)

TNF Tumor necrosis factor (TNF superfamily, member 2)

TNFRSF10D Tumor necrosis factor receptor superfamily, member 10d

TP53 Tumor protein p53 (Li-Fraumeni syndrome)

TRO Trophinin

SALL4 IS A KEY REGULATOR OF SURVIVAL AND APOPTOSIS 809BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

apoptosis seen in SALL4-reduced cells. We hypothesize thatbecause retrovirus infection is variable, tumors seen in miceinjected with SALL4-reduced cells may have derived from asubpopulation of NB4 cells that express relatively normal levelsof SALL4. To test this, Q-RT-PCR was performed to determinethe status of SALL4 expression on tumors from cells infectedwith SALL4 shRNA retroviruses. Interestingly, tumor cells of

animals injected with SALL4-reduced NB4 cells expressed anormal level of SALL4, similar to the level of SALL4 in thecontrol cell tumors (data not shown). This study indicates thatthe tumors derived from NB4 cells that were not derived fromthe SALL4-reduced cells that were originally injected. Takentogether, our results demonstrate that SALL4 is required fortumorigenicity in AML cells.

Figure 4. SALL4 binds and transcriptionally regu-lates genes involving various apoptotic pathways.(A) SALL4 binds to apoptosis-inducing genes, apoptosis-inhibiting genes, as well as genes involved in otherapoptosis processes based on the PANTHER classifica-tion system.34 (B) Q-RT-PCR shows differentially regu-lated expression levels of apoptotic genes in SALL4knockdown NB4 cells. Error bars represent standarddeviations for duplicate measurements (quantificationreplicas) normalized to GAPDH levels. Lowering levelsof SALL4 expression in NB4 cells induces gene expres-sion that favors apoptosis. (C) When comparing theSALL4 target gene set with the total analyzed gene pool,we found that SALL4 binds to genes involving variousapoptotic pathways, including p53, BCL2, TNF, andPTEN. Data were processed using Ingenuity PathwayAnalysis.

810 YANG et al BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

Discussion

One interesting finding of this study is the striking level of genesthat bind SALL4 within NB4 cells. Many reported transcriptionfactors do not have as extensive binding in genome-wide analyses,

although transcription factors NANOG and E2F1 have beendemonstrated to bind an equivalent number of gene promoters.36,37

In addition, ChIP-chip experiments in human and mouse ES cellshave shown that SALL4 also binds to multiple thousands of genepromoters (Y.M., unpublished data, June 2007). In recent studies,SALL4 has been used as a part of a gene signature and marker forsomatic cell reprogramming due to its vital role in ES cells.38,39

Inherent in the quality of ChIP-chip experiments is the quality of theantibody. We have used this antibody extensively for the previouslydescribed experiments and rigorously characterized it for use in ChIP. Itis currently used for immunohistochemistry,12,29 and flow cytometrywhere it uniquely and specifically identifies the SALL4 protein. Inaddition, we used a commercially available anti-HA antibody tocharacterize the specificity of the SALL4 antibody. Using mouse EScells, ChIP-enriched regions were identified using both ChIP-chip andChIP-PCR with the anti-SALL4 antibody. Then, in SALL4�/� mouseES cells transfected with exogenous SALL4-HA, we used the HAantibody to pull down ChIP-enriched fragments. The correlation be-tween these 2 sets of ChIP DNA was greater than 88%. Finally,ChIP-chip data have been collected for multiple leukemia cell lines andreflect similar, but complex, binding patterns (data not shown).

The functional categories of SALL4 target genes in NB4 leukemiccells revealed a wide variety of cellular processes. Most striking, SALL4

Figure 5. Flow cytometric assays of caspase-3 activ-ity, cell-cycle changes, and cell proliferation assaysin SALL4 knockdown NB4 cells. Two shRNA retroviralconstructs that targeted different regions of the SALL4were made, and their ability to reduce SALL4 mRNA inNB4 cells was confirmed by Q-RT-PCR (Figure S1). NB4cells with approximately 75% reduction of SALL4 expres-sion levels (via shRNA no. 7412) were used in theseexperiments. (A,D) NB4 cells transduced with the controlretrovirus. (B,E) NB4 cells transduced with SALL4 shRNAretroviruses. (C,F) Bmi-1–expressing plasmid was trans-fected into SALL4 knockdown NB4 cells. Caspase-3activity analysis shows shRNA-mediated reduction ofSALL4 induces apoptosis in NB4 cells (A,B). SALL4knockdown of NB4 cells can be rescued from apoptosiswith ectopic expression of Bmi-1 (C). Cell-cycle changesand cellular DNA synthesis in control NB4 cells andSALL4 knockdown NB4 cells were monitored by BrdU-incorporation assay and analyzed by flow cytometry (3%background debris were excluded); data show knock-down of SALL4 induces cell-cycle arrest and decreasedDNA synthesis (D,E). By ectopically expressing Bmi-1,SALL4 knockdown cells can be rescued from cell-cycleand DNA synthesis arrest (F). (G) Proliferation assayshows that SALL4 knockdown affects the growth rate ofNB4 cells. Cells (5 � 104) from stable clones expressinga control or SALL4 shRNA retrovirus (experiment) wereplated in RPMI with 10% FBS. This assay was performedin duplicate and cell numbers were counted daily using atrypan blue–based assay.

Figure 6. Knockdown of SALL4 inhibits the tumorigenicity of NB4 leukemiacells. Tumor growth in NOD/SCID mice injected with control NB4 cells or SALL4reduction cells was estimated with caliper measurements. Cells were injected at2 different dosages, and tumor size was determined on day 21 or 30 after injection.Error bars represent the standard deviation. *P � .05; **P � .01.

SALL4 IS A KEY REGULATOR OF SURVIVAL AND APOPTOSIS 811BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

frequently binds to target genes involving various apoptotic pathways,including p53, BCL2, TNF, and PTEN, when comparing the SALL4target genes with those in the total analyzed gene pool. All of theseapoptotic pathways experience a dramatic effect by the down-regulationof SALL4 and result in apoptosis. Similar results were observed in othercell lines including NTERA2 and KBM5 (unpublished data). Theseobservations suggest that down-regulation of SALL4 could modulatethe various apoptotic pathways disseminating apoptotic signals. Insupport of this interpretation, we have observed that reduction in SALL4levels by 75% triggers massive apoptosis in the human leukemic cellline NB4.

Our conclusion that SALL4 acts as a key regulator in leukemic cellswas further strengthened by mRNA microarray profiling (NimbleGenSystems). Global mRNA changes in NB4 cells after SALL4 knock-down were analyzed, and affected genes (either significantly up-regulated or down-regulated) were classified using Ingenuity PathwayAnalysis. As shown in Figure 7, SALL4 not only binds genes involvedin important cell growth pathways such as the Wnt, apoptosis, PTEN,and NF-kB signaling pathways, but gene expression levels in thesepathways are also affected by reduction of SALL4. Correlating pro-moter binding and affected pathway distributions (Pearson correlationcoefficient of 0.7) provides insight into the mechanisms wherebySALL4 may regulate leukemic cell growth. Alternately, functionalanalyses of expression data consistently demonstrated that cancer-related genes are regulated by SALL4. These expression microarraydata were considered reliable since approximately 50% to 75% SALL4

reduction was observed in duplicate experiments, which is consistentwith Q-RT-PCR results.

Previously, we have demonstrated that Bmi-1 is a downstream targetfor SALL4 in leukemic cells. Here we demonstrate that inducedapoptosis and cell-cycle arrest in leukemic cells expressing reducedlevels of SALL4 can be reversed by ectopically expressing Bmi-1. Thisraises an interesting question: How does Bmi-1 do this? Perhaps SALL4and Bmi-1 modulate common target genes in apoptotic and cell-cyclepathways that are interrelated. Therefore, forced expression of Bmi-1would be capable of reversing the deleterious effects associated withSALL4 down-regulation in leukemic cells.

Recently, we have completed similar experiments on cancer stemcells and normal ES cells of both mouse and human origin (Y.M.,unpublished data, June 2007). Preliminary results showed that reductionof SALL4 led to significant apoptosis in NTERA2, a human embryonalcarcinoma cell line, as measured by caspase-3 activity (10-fold increase)and through morphologic examination. In contrast, no significant celldeath or increased caspase-3 activity was observed in ES cells express-ing normal levels of SALL4 (data not shown). In addition, in ourconventional SALL4 knockout animal model, SALL4�/� bone mar-rows displayed little cell death compared with WT controls, andimmature hematopoietic stem cell and hematopoietic progenitor cells inSALL4�/� mice were only mildly decreased compared with WT (datanot shown). The molecular mechanism of these phenotypes in cancer-ous and normal cells is unknown; however, data presented here suggestSALL4 binds to some different target genes in the NB4 cell line, such as

Figure 7. SALL4 binds and transcriptionally regulatesgenes involving various pathways and biologic func-tions. Gene-expression changes derived from NB4 cellmRNA expression profiling following SALL4 reductionwere classified and compared with ChIP-chip data.

represent the number of SALL4-bound promotersidentified from the ChIP-chip assay; represent thenumber of statistically significant (P � .05) altered mRNAtranscripts in SALL4 knockdown NB4 cells. Their distribu-tion in each signaling pathway (A) or biologic functionalcategory (B) was determined by Ingenuity Pathway Anal-ysis. The P values presented for each category are basedon overrepresentation and calculated against the knowl-edge base provided by Ingenuity Pathway Analysis.

812 YANG et al BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3

DAXX, BIRC4, and p53, compared with the CD34� cell population ofwhole bone marrows. Since the reduction of SALL4 has dramatic effecton the survival of leukemic cells but not normal ES cells, it is temptingto speculate that the SALL4/Bmi-1 regulatory pathway might be anattractive target for therapeutic intervention by inducing cancer stemcells to undergo apoptosis.

Acknowledgments

This work was supported in part by National Institutes of Health(NIH, Bethesda, MD) grants K08 CA097185 (Y.M.), NIHR01HL087948 (Y.M.), P20 RR016464 (Y.M.), and K08 DK063220(L.C.) and by the National Blood Foundation (Bethesda, MD;L.C.).

Authorship

Contribution: Y.M., J.Y., and L.C. designed research; J.Y., C.G.,and Z.A. performed research; L.M.F., D.C.W., H.D., J.Y., D.X.,Y.M., and T.C.F. analyzed data; and L.M.F., D.C.W., Y.M, J.Y.,L.C., and T.C.F. wrote the paper.

Conflict-of-interest disclosure: The authors declare no compet-ing financial interests.

Correspondence: Y. Ma, Division of Laboratory Medicine, NevadaCancer Institute, One Breakthrough Way, Las Vegas, NV 89135; e-mail:yma@nvcancer.org; or L. Chai, Department of Pathology, Joint Programin Transfusion Medicine, Brigham and Women’s Hospital/Children’sHospital Boston, Harvard Medical School, 75 Francis St, Boston, MA02115; e-mail: lchai@partners.org.

References

1. Kohlhase J, Schuh R, Dowe G, et al. Isolation,characterization and organ-specific expression oftwo novel human zinc finger genes related to theDrosophila gene spalt. Genomics. 1996;38:291-298.

2. Kohlhase J, Hausmann S, Stojmenovic G, et al.SALL3, a new member of the human spalt-likegene family, maps to 18q23. Genomics. 1999;62:216-222.

3. Al-Baradie R, Yamada K, St Hilaire C, et al.Duane radial ray syndrome (Okihiro syndrome)maps to 20q13 and results from mutations inSALL4, a new member of the SAL family. Am JHum Genet. 2002;71:1195-1199.

4. Kuhnlein RP, Schuh R. Dual function of the re-gion-specific homeotic gene spalt during Dro-sophila tracheal system development. Develop-ment. 1996;122:2215-2223.

5. Kohlhase J, Heinrich M, Schubert L, et al. Okihirosyndrome is caused by SALL4 mutations. HumMol Genet. 2002;11:2979-2987.

6. Borozdin W, Wright MJ, Hennekam RC, et al.Novel mutations in the gene SALL4 provide fur-ther evidence for acro-renal-ocular and Okihirosyndromes being allelic entities, and extend thephenotypic spectrum. J Med Genet (http://www.jmedgenet.com/cgi/content/full/41/8/e102).2004;41:e102.

7. Zhang J, Tam WL, Tong GQ, et al. Sall4 modu-lates embryonic stem cell pluripotency and earlyembryonic development by the transcriptionalregulation of Pou5f1. Nat Cell Biol. 2006;8:1114-1123.

8. Elling U, Klasen C, Eisenberger T, Anlag K, TreierM. Murine inner cell mass-derived lineages de-pend on Sall4 function. Proc Natl Acad Sci U S A.2006;103:16319-16324.

9. Wu Q, Chen X, Zhang J, et al. Sall4 interacts withNanog and co-occupies Nanog genomic sites inembryonic stem cells. J Biol Chem. 2006;281:24090-24094.

10. Zhou Q, Chipperfield H, Melton DA, Wong WH.A gene regulatory network in mouse embryonicstem cells. Proc Natl Acad Sci U S A. 2007;104:16438-16443.

11. Li SS, Liu YH, Tseng CN, Chung TL, Lee TY,Singh S. Characterization and gene expressionprofiling of five new human embryonic stem celllines derived in Taiwan. Stem Cells Dev. 2006;15:532-555.

12. Ma Y, Cui W, Yang J, et al. SALL4, a novel onco-gene, is constitutively expressed in human acutemyeloid leukemia (AML) and induces AML intransgenic mice. Blood. 2006;108:2726-2735.

13. Takahashi K, Tanabe K, Ohnuki M, et al. Induc-tion of pluripotent stem cells from adult humanfibroblasts by defined factors. Cell. 2007;131:861-872.

14. Yang J, Chai L, Liu F, et al. Bmi-1 is a target genefor SALL4 in hematopoietic and leukemic cells.Proc Natl Acad Sci U S A. 2007;104:10494-10499.

15. Molofsky AV, Pardal R, Iwashita T, Park IK, ClarkeMF, Morrison SJ. Bmi-1 dependence distin-guishes neural stem cell self-renewal from pro-genitor proliferation. Nature. 2003;425:962-967.

16. Cui H, Hu B, Li T, et al. Bmi-1 is essential for thetumorigenicity of neuroblastoma cells. Am JPathol. 2007;170:1370-1378.

17. Lee K, Adhikary G, Balasubramanian S, et al. Ex-pression of Bmi-1 in epidermis enhances cell sur-vival by altering cell cycle regulatory protein ex-pression and inhibiting apoptosis. J InvestDermatol. 2008;128:9-17.

18. Lessard J, Sauvageau G. Bmi-1 determines theproliferative capacity of normal and leukaemicstem cells. Nature. 2003;423:255-260.

19. Jacobs JJ, Kieboom K, Marino S, DePinho RA,van Lohuizen M. The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation andsenescence through the ink4a locus. Nature.1999;397:164-168.

20. Jacobs JJ, Scheijen B, Voncken JW, Kieboom K,Berns A, van Lohuizen M. Bmi-1 collaborates withc-Myc in tumorigenesis by inhibiting c-Myc-induced apoptosis via INK4a/ARF. Genes Dev.1999;13:2678-2690.

21. Berezovska OP, Glinskii AB, Yang Z, Li XM,Hoffman RM, Glinsky GV. Essential role for acti-vation of the Polycomb group (PcG) protein chro-matin silencing pathway in metastatic prostatecancer. Cell Cycle. 2006;5:1886-1901.

22. Raaphorst FM. Self-renewal of hematopoieticand leukemic stem cells: a central role for thePolycomb-group gene Bmi-1. Trends Immunol.2003;24:522-524.

23. Park IK, Qian D, Kiel M, et al. Bmi-1 is requiredfor maintenance of adult self-renewing haemato-poietic stem cells. Nature. 2003;423:302-305.

24. Dick JE. Stem cells: Self-renewal writ in blood.Nature. 2003;423:231-233.

25. Ohta H, Sawada A, Kim JY, et al. Polycomb groupgene rae28 is required for sustaining activity ofhematopoietic stem cells. J Exp Med. 2002;195:759-770.

26. van der Lugt NM, Domen J, Linders K, et al. Pos-terior transformation, neurological abnormalities,

and severe hematopoietic defects in mice with atargeted deletion of the bmi-1 proto-oncogene.Genes Dev. 1994;8:757-769.

27. Iwama A, Oguro H, Negishi M, et al. Enhancedself-renewal of hematopoietic stem cells medi-ated by the polycomb gene product Bmi-1. Immu-nity. 2004;21:843-851.

28. Mihara K, Chowdhury M, Nakaju N, et al. Bmi-1 isuseful as a novel molecular marker for predictingprogression of myelodysplastic syndrome andpatient prognosis. Blood. 2006;107:305-308.

29. Cui W, Kong NR, Ma Y, Amin HM, Lai R, Chai L.Differential expression of the novel oncogene,SALL4, in lymphoma, plasma cell myeloma, andacute lymphoblastic leukemia. Mod Pathol. 2006;19:1585-1592.

30. National Center for Biotechnology Information.Gene Expression Omnibus. http://www.ncbi.nlm.nih.gov/geo/. Accessed May 1, 2008.

31. National Center for Biotechnology Information.Entrez Gene. http://www.ncbi.nlm.nih.gov/sites/entrez?db�gene. Accessed May 4, 2007.

32. Pan W. A comparative review of statistical meth-ods for discovering differentially expressed genesin replicated microarray experiments. Bioinfor-matics. 2002;18:546-554.

33. Boyer LA, Lee TI, Cole MF, et al. Core transcrip-tional regulatory circuitry in human embryonicstem cells. Cell. 2005;122:947-956.

34. SRI International. Protein ANalysis THrough Evo-lutionary Relationships (PANTHER). http://www.pantherdb.org/. Accessed November 20, 2006.

35. Mancini F, Gentiletti F, D’Angelo M, et al. MDM4(MDMX) overexpression enhances stabilizationof stress-induced p53 and promotes apoptosis.J Biol Chem. 2004;279:8169-8180.

36. Loh YH, Wu Q, Chew JL, et al. The Oct4 andNanog transcription network regulates pluripo-tency in mouse embryonic stem cells. Nat Genet.2006;38:431-440.

37. Bieda M, Xu X, Singer MA, Green R, FarnhamPJ. Unbiased location analysis of E2F1-bindingsites suggests a widespread role for E2F1 in thehuman genome. Genome Res. 2006;16:595-605.

38. Lowry WE, Richter L, Yachechko R, et al. Gen-eration of human induced pluripotent stem cellsfrom dermal fibroblasts. Proc Natl Acad SciU S A. 2008;105:2883-2888.

39. Takahashi K, Tanabe K, Ohnuki M, et al. Induc-tion of pluripotent stem cells from adult humanfibroblasts by defined factors. Cell. 2007;131:861-872.

SALL4 IS A KEY REGULATOR OF SURVIVAL AND APOPTOSIS 813BLOOD, 1 AUGUST 2008 � VOLUME 112, NUMBER 3