A DEAB-Sensitive Aldehyde Dehydrogenase Regulates Hematopoietic Stem and Progenitor Cells Development During Primitive Hematopoiesis in Zebraﬁsh Embryos

A DEAB-sensitive aldehyde dehydrogenase regulates hematopoietic stem and progenitor cells development during primitive hematopoiesis in zebraﬁsh embryos

ACH Ma, MIS Chung, R Liang and AYH Leung

Division of Haematology and Bone Marrow Transplantation, Department of Medicine, The University of Hong Kong, Hong Kong, Hong Kong SAR

Although aldehyde dehydrogenase (ALDH) activity has become interfering RNA expanded HSPC population, reminiscent of a surrogate of hematopoietic stem and progenitor cells ALDH inhibition with DEAB in human study.6 However, (HSPCs), its function during hematopoiesis was unclear. Here, targeted deletion of aldh1al7 and aldh1a28 had no effects on we examined its role in zebrafish hematopoiesis based on pharmacological inhibition and morpholino (MO) knockdown. murine HSPC development and yolk-sac hematopoiesis, sug- Zebrafish embryos were treated with diethylaminobenzal- gesting that other members in the Aldh family may be involved. dehyde (DEAB, 1 lmol/l) between 0- and 48 hour-post-fertilization The zebrafish embryo has emerged as a model for the study (hpf). MOs targeting aldhs were injected between 1 and 4-cell of embryonic hematopoiesis.9 Like other vertebrates, hemato- stage. The effects on hematopoiesis were evaluated at different poiesis occurs in two successive waves, known as the primitive stages. DEAB treatment between 0 and 18 hpf increased gene and definitive hematopoiesis. In zebrafish, the transitory expression associated with HSPC (scl, lmo2), erythropoiesis (gata1, a-andb-eHb) and myelopoiesis (spi1)aswellasgfpþ primitive hematopoiesis occurs in the intermediate cell mass cells in dissociated Tg(gata1:gfp) embryos. The effects were (ICM), whereas the definitive wave originates from the ventral ameliorated by all-trans retinoic acid (1 nmol/l). Definitive wall of dorsal aorta, thence the kidney, where hematopoiesis hematopoiesis and the erythromyeloid precursors were occurs life-long. The transcription factors involved in HSPC unaffected. In all, 14 out of 15 zebrafish aldhs were detectable formation and differentiation during hematopoiesis are also by reverse transcription PCR in 18 hpf embryos, of which only conserved. The zebrafish genome contains at least 15 aldh aldh1a2 and aldh16a1 were expressed in sites pertinent to genes. Missense mutation of aldh1a2, known as the neckless10 hematopoiesis. Molecular targeting by MOs was demonstrated 11 for 15 aldhs, but none of them, even in combined aldh1a2 and or no-fin mutation, was associated with defects in the aldh1a3 knockdown, recapitulated the hematopoietic expan- hindbrain and brachial cartilage development, absence of sion in DEAB-treated embryos. In conclusion, DEAB expands pectoral fins and more recently expansion of cardiac progenitor HSPC population during primitive hematopoiesis through pool.12 These developmental effects could be explained by inhibition of aldh and retinoic acid synthesis. The specific aldh defective endogenous retinoic acid biosynthesis secondary to isoform(s) remains to be determined. 13 Leukemia (2010) 24, 2090–2099; doi:10.1038/leu.2010.206; aldh1a2 mutation. Interestingly, aldh1a1, the isoform thought published online 7 October 2010 to be associated with mammalian HPSC regulation, is lacking in Keywords: aldehyde dehydrogenase; zebrafish; hematopoiesis; all teleosts.14 It remains unclear if aldh may play a role during hematopoietic stem cells embryonic hematopoiesis. To address this issue, we examined the effect of aldh inhibition in zebrafish using both pharmacological inhibitor and anti-sense morpholino (MO) knockdown. We demonstrated that pharmacological inhibition of aldh with DEAB resulted in Introduction expansion of primitive hematopoiesis. Definitive hematopoiesis was not affected. Molecular targeting of 15 individual aldh Aldehyde dehydrogenase (ALDH) are a group of NAD(P) þ - isoforms did not increase hematopoiesis, implicating a complex dependent enzymes involved in the biosynthesis of retinoic acid mode of regulation by aldh. as well as the metabolism of amino acid, fatty aldehydes, ethanol and cyclophosphamide.1 It was first demonstrated in mouse and human hematopoietic stem and progenitor cells Materials and methods (HSPCs) almost 20 years ago based on their resistance to the cytotoxic effects of cyclophosphamide. In human, inhibition of Zebrafish and MOs ALDH with diethylaminobenzaldehyde (DEAB) in-vitro has Wild-type zebrafish (Danio rerio) were obtained from local been shown to delay differentiation and expand HSPC by 2 aquarium and were maintained and raised under standard inhibiting endogenous retinoic acid biosynthesis. Immuno- conditions at 28 1C. Transgenic Tg(gata1:gfp) embryos was a gift histochemical and gene expression studies have suggested from Dr Anming Meng (Tsinghua University, Beijing, China) ALDH1A1 as the predominant form associated with HSPC 15 3–5 were used to quantify the erythroid cell populations. activity. In mouse, knockdown of Aldh1a1 by small Transgenic Tg(fli1:gfp) embryos bought from ZIRC were used to track vascular development. Anti-sense MOs (Gene-Tools, Correspondence: Dr AYH Leung, Division of Haematology and Bone Philomath, OR, USA) targeting splice junction of scl16 together Marrow Transplantation, Department of Medicine, Queen Mary with 15 MOs newly designed to target the 50-UTR or start codon Hospital, The University of Hong Kong, Pok Fu Lam Road, of zebrafish aldhs and a random sequence MO for setting up Hong Kong, Hong Kong SAR. E-mail: [email protected] control were used (Supplementary Table 1). Procedures for Received 9 July 2010; accepted 30 July 2010; published online micro-injection and whole-mount in-situ hybridization (ISH) 7 October 2010 have been described previously.17,18 ALDH regulates primitive hematopoiesis in zebrafish embryos ACH Ma et al 2091 Treatment with ALDH inhibitor Microscopy and imaging DEAB was used to treat zebrafish embryos at 1 mmol/l. Unless Bight field images were taken under Nikon SMZ800 otherwise specified, the embryos were treated from 0 hour-post- (Nikon Corporation, Kawasaki, Japan) and P-Plan 1X objective fertilization (hpf) until examination. In each experiment, an in 3% methylcellulose with Nikon Coolpix E4500 (Nikon Hong equal volume of dimethyl sulfoxide was added to the control Kong Ltd., Hong Kong), processed with Adobe Photoshop embryos. Version 7.0. Fluorescent images were taken under Olympus IX70 (Olympus Corporation, Tokyo, Japan) and 10X/0.3 NA objective in 3% methylcellulose, with Olympus DP71 (Olympus Cloning of zebrafish aldhs and synthesis of sense and Corporation, Shinjuku-ku, Tokyo, Japan) and Olympus DP-BSW anti-sense RNA probes for ISH basic Software, processed with Adobe Photoshop Version 7.0. Full-length or partial sequences of zebrafish aldhs were PCR amplified with specific primers (Supplementary Table 2) and TA-cloned using pGEM-T Easy Vector System (Promega, Statistical analysis Madison, WI, USA). The sequences and orientation were Results were expressed as mean±s.e.m. unless otherwise stated. confirmed by DNA sequencing and both sense and anti-sense Comparisons between groups of data were evaluated by paired RNA probes were synthesized as described previously.17,18 t-test, Mann–Whitney U-test and Kruskal–Wallis test, where appropriate. P-value of o0.05 was considered statistically significant. Cloning and injection of gfp-fusion expressing vectors for checking MO efficiency To confirm molecular targeting of MOs, we designed primers Results (Supplementary Table 3) to clone the MO targeting region (50-UTR and the start codon) of each aldh isoform in-frame into DEAB treatment induced ICM expansion without pcDNA3.1/CT-GFP-TOPO vectors using the GFP Fusion TOPO perturbing vascular formation TA Expression Kits (Invitrogen, Carlsbad, CA, USA). We injected To examine the effects of aldh inhibition, we treated embryos each GFP vector with or without the respective MO at one-cell with DEAB (1 mmol/l) at 0 hpf. At 18 and 24 hpf (Figures 1a–d), stage. Embryos were examined at 18 hpf under fluorescent there was no morphological abnormality. At 30 hpf, the ICM was microscopy to check for gfp signal. expanded (Figures 1e–j). Thereafter, the ICM expansion gradually decreased and became undetectable at 48 hpf (Figures 1k and l). We also examined the effect of DEAB treatment on vascular Flow cytometry development using Tg(fli1:gfp) embryos. At 30 hpf, both vasculo- Tg(gata1:gfp) embryos at 18 hpf were dechorionated and genesis and angiogenesis were intact (Figures 1m–p) and digested with 0.05% Trypsin/EDTA solution (Invitrogen) for circulation was patent (Supplementary video SV1-4). Therefore, 15 min at 28 1C. The embryos were completely dissociated to the ICM expansion was unlikely due to the accumulation of single-cell suspension by pipetting. Trypsin digestion was blood cells as in embryos with defective circulation. terminated by CaCl2 (2 mmol/l) and the cellular suspension was filtered through a 40 mm cell strainer (BD Falcon, BD Biosciences Discovery Labware, Bedford, MA, USA). Cells were DEAB treatment upregulated genes associated with washed and harvested in phosphate-buffered saline with 2% primitive hematopoiesis fetal bovine serum and the percentage of gfp þ cells were The expansion of ICM in DEAB-treated embryos led us to enumerated by flow cytometry (Cytomics FC500, Beckman examine its hematopoietic gene expression. To focus on Coulter, Brea, CA, USA). For aldehyde dehydrogenase activity primitive hematopoiesis, we evaluated the embryos before the assay, wild-type embryos were dissociated and examined using onset of circulation. At 18 hpf, DEAB treatment resulted in Aldefluor assay (StemCell Technologies Inc., Vancouver, BC, significant increase in the expression of scl and lmo2 (HSPC), Canada) and flow cytometry. gata1, a-ehb and b-ehb (erythroid), as well as spi1 (myeloid progenitor; Figure 2a). The results were confirmed by Q-PCR (Table 1). The expression of genes associated with neural Real-time quantitative RT-PCR (Q-PCR) development (pax2a and egr2a) were not affected (Table 1). The RNA was extracted from zebrafish embryos at different increase in hematopoietic gene expression could be ameliorated developmental stages and was reverse transcribed. Q-PCR for by concomitant treatment with all-trans retinoic acid (ATRA) at hematopoietic genes was performed using the SYBR green PCR 1 nmol/l, although ATRA alone at this dosage had no effect on master mix system (Applied Biosystems, Foster City, CA, USA). hematopoietic gene expression (Table 1). The increase in Expression level was presented as fold change calculated using primitive hematopoiesis has led us to re-evaluate our results at 17,18 the comparative CT method as described with b-actin as the earlier time point, when specification of hematopoiesis begins. internal reference. Primer sequences used for Q-PCR were At 12 hpf, expression of scl and lmo2 in both the anterior and shown in Supplementary Table 4. posterior lateral plate mesoderm was significantly increased (Figure 2b). The results suggested that DEAB treatment increased hematopoietic specification at the level of HSPC. Blood sampling from zebrafish embryos and cytospin preparation Embryos at 48 hpf were anesthetized by 0.016% Tricaine DEAB treatment increased erythropoiesis, which could (3-amino benzoic acid ethylester) in 0.9 Â PBS and were be ameliorated by concomitant ATRA treatment punctured at the Duct of Curvier. Blood cells were aspirated and To further quantify the expansion of primitive hematopoiesis, pooled from 30 embryos and centrifuged at 800 r.p.m. for 5 min which is mainly erythropoiesis, we treated Tg(gata1:gfp) (7620 Cytopro Cytocentrifuge, Wescor Inc., Logan, UT, USA), embryos with DEAB until 18 hpf and erythropoiesis was followed by routine cytological staining and examination examined by flow cytometry after embryos dissociation.

Leukemia 2092 Leukemia Tg( rae ( treated S hwn h xrsinof expression the showing ISH erfihebysa 8hf( hpf 18 at embryos zebrafish 1 Figure fli1:gfp l erfihebysa 8hf ( hpf. 48 at embryos zebrafish ) mro ne ursetmicroscope. fluorescent under embryos ) feto EBtetet(1 treatment DEAB of Effect LHrgltspiiiehmtpissi erfihembryos zebrafish in hematopoiesis primitive regulates ALDH

a 30 hpf 30 hpf 24 hpf 18 hpf

, 30 hpf 30 hpf 48 hpf 30 hpf scl b gh ef cd ab kl ij mn op a ,2 p ( hpf 24 ), and gata1 m m o/)o erfihebys ( embryos. zebrafish on mol/l) – c p , aclrptenn nDS oto ( control DMSO in patterning Vascular ) d t3 p nDS oto ( control DMSO in hpf 30 at DMSO n 0hf( hpf 30 and ) 200 μ m e C Ma ACH , f 200 .Isrsi ( in Inserts ). 200 200 gata1 200 200 μ tal et m μ μ c scl scl m μ μ 200 200 m a m200 m – f g iehlsloie(MO oto ( control (DMSO) sulfoxide Dimethyl ) μ μ , e m m i , n EBtetd( DEAB-treated and ) f niaetemgie iwo h C ein ( region. ICM the of view magnified the indicate ) m , n n EBtetd( DEAB-treated and ) DEAB 200 h , μ j mro.( embryos. ) m 200 200 o 200 gata1 200 100 , 100 k a p , μ , mro t3 p ssonby shown as hpf 30 at embryos ) μ μ μ l m μ c MOcnrl( control DMSO ) μ m m m μ , m m m e n EBtetd( DEAB-treated and ) g – j Whole-mount ) k n DEAB- and ) b , d , f ) ALDH regulates primitive hematopoiesis in zebrafish embryos ACH Ma et al 2093 18 hpf DMSO DEAB DMSO DEAB scl lmo2 spi1 gata1 -ehb -ehb β α

200 μm

12 hpf DMSO DEAB

ALPM PLPM scl lmo2

200 μm

Figure 2 Effect of DEAB treatment (1 mmol/l) on the gene expression associated with primitive hematopoiesis in zebraﬁsh embryos. (a) Whole- mount ISH showing the expression of genes associated with primitive hematopoiesis in dimethyl sulfoxide (DMSO) control and DEAB-treated embryos at 18 hpf. (b) Flat-mount ISH showing the expression of genes associated with primitive HSC along the anterior lateral plate mesoderm (ALPM) and posterior lateral plate mesoderm (PLPM) in DMSO control and DEAB-treated embryos at 12 hpf.

Table 1 Fold change of the gene expression associated with primitive hematopoiesis and neurogenesis at 18 hpf in DEAB-treated embryos

Gene DMSO DEAB (1 mmol/l) DEAB (1 mmol/l)+ATRA (1 nmol/l) ATRA (1 nmol/l) scl 1.00 2.69±0.20*** 1.34±0.03* 0.95±0.07NS lmo2 1.00 1.95±0.09*** NA NA gata1 1.00 2.33±0.12*** 1.41±0.10* 1.07±0.11NS a-eHb 1.00 1.85±0.04*** 1.37±0.07* 0.93±0.14NS b-eHb 1.00 1.74±0.09*** NA NA spi1 1.00 2.89±0.27*** 1.29±0.08* 1.12±0.17NS l-plastin 1.00 1.03±0.04NS 0.96±0.14NS 1.09±0.05NS pax2a 1.00 1.07±0.05NS NA NA egr2a 1.00 1.16±0.06NS NA NA Abbreviations: ATRA, all-trans retinoic acid; DEAB, diethylaminobenzaldehyde; DMSO, dimethyl sulfoxide; NS, not signiﬁcant; NA, not available. Data are presented as relative fold change compared with DMSO control. P-values are presented as *(Po0.05); ***(Po0.005).

In DEAB-treated embryos, the percentage of gfp þ cell repre- biosynthesis, we treated Tg(gata1:gfp) embryos with DEAB senting erythroid population was signiﬁcantly increased (DEAB: (1 mmol/l) and ATRA (1 nmol/l) simultaneously. The increase 6.15±0.29%; control: 4.39±0.11%, P ¼ 0.001; Figures 3a and in erythropoiesis was ameliorated upon ATRA treatment b). To examine if increased erythropoiesis in DEAB-treated (4.46±0.12%; Figure 3c). The latter had no effect on basal embryos was due to inhibition of endogenous retinoic acid erythropoiesis (4.09±0.09%; Figure 3d).

Leukemia ALDH regulates primitive hematopoiesis in zebraﬁsh embryos ACH Ma et al 2094 abDMSO DEAB 4.25% 7.77% SLnSS Lin SS Lin SS Lin

+ GFP+ GFP cdDEAB + ATRA ATRA

4.11% 4.19% SS Lin

+ GFP+ GFP

efDMSO DEAB

orthochromic polychromatophilic orthochromic polychromatophilic

Figure 3 Effect of DEAB treatment (1 mmol/l) on primitive erythropoiesis. (a–d) Representative ﬂow cytometry records using dissociated Tg(gata1:gfp) embryos showing the percentage of gfp positive erythroid cells in dimethyl sulfoxide (DMSO) control (a), DEAB-treated (b), DEAB and ATRA concomitantly treated (c) and ATRA-treated (d) embryos. (e, f) Effect of DEAB treatment on the differentiation and proliferation status of primitive erythroid cells. Representative late orthochromic and early polychromatophilic erythocytes from cytospin of blood extracted from DMSO control (e) and DEAB-treated embryos (f) at 48 hpf.

DEAB treatment did not affect erythroid cell at 0–12, 0–18 and 6–12 hpf produced comparable increase differentiation in hematopoietic gene expression (Figure 4a). On the other The increase in primitive hematopoiesis led us to evaluate if hand, DEAB treatment at 0–6 or 12–18 hpf had no effect DEAB might affect erythroid differentiation. In 48 hpf control (Figure 4a). Concomitant knockdown of scl by MO ameliorated embryos, the relative abundance of circulating polychromato- the increase in hematopoietic gene in DEAB treated embryos philic (early) and orthochromic (late) erythroblasts were (Figure 4b). 31.6±1.90 and 68.4±1.90%, respectively. In DEAB-treated embryos, the corresponding percentage were 32.7±2.89 and 67.3±2.89% (Figures 3e and f), suggesting that DEAB had no DEAB treatment did not affect the initiation of definitive effect on erythropoietic differentiation. hematopoiesis To evaluate if aldh might have similar regulatory functions in definitive hematopoiesis, we examined DEAB-treated embryos Expansion of primitive hematopoiesis by DEAB was at 30 and 36 hpf by whole-mount ISH and Q-PCR. At 30 hpf, the achieved between 6 and 12 hpf, and ameliorated by erythromyeloid progenitors at caudal hematopoietic tissue,19 as concomitant knockdown of scl shown by spi1 expression, was not affected (Supplementary We further examined the developmental stage at which DEAB Figure S1A). At 36 hpf, the expression c-myb and runx1, expanded primitive hematopoiesis by treating the embryos transcription factors associated with the initiation of definitive between 0–6, 0–12, 6–12 and 12–18 hpf and analyzed the effect hematopoiesis in the ventral wall of dorsal aorta, were also on hematopoietic gene expression at 18 hpf. DEAB treatment unaffected (Supplementary Figure S1B, Table 2).

Leukemia ALDH regulates primitive hematopoiesis in zebraﬁsh embryos ACH Ma et al 2095 DMSO Control DEAB treatment (0-18 hpf) DEAB treatment (0-6 hpf) 3.5 DEAB treatment (0-12 hpf) *** *** DEAB treatment (6-12 hpf) 3 DEAB treatment (12-18 hpf) *** *** *** 2.5 *** *** ** *** *** *** *** 2 *** ** *** 1.5 * 1

0.5

0 scl lmo2 gata1 -eHb spi1 l-plastin

DMSO Control DEAB treatment DEAB treatment + scl MO 3.5 scl MO ** 3 ***

2.5 ** 2

1.5

1 * * * 0.5 ** * *

0 scl gata1 spi1

Figure 4 Effect of DEAB treatment (1 mmol/l) with different exposure time and with concomitant scl knockdown on the gene expression associated with primitive hematopoeisis as shown by Q-PCR. (a) Fold change of the expression of hematopoietic genes in DEAB-treated embryos at 18 hpf after differential DEAB treatment period compared with dimethyl sulfoxide (DMSO) control. (b) Fold change of the expression of hematopoietic gene at 18 hpf in DEAB-treated embryos after concomitant knockdown of scl by morpholino. P-values are presented as *(Po0.05); **(Po0.01); ***(Po0.005).

Table 2 Fold change of the gene expression associated with population (54.71±2.49%, n ¼ 5 experiments including more definitive hematopoiesis at 36 hpf in DEAB-treated embryos than 150 embryos) whose activity was completely abolished by DEAB at both 15 and 1 mmol/l. Gene DMSO DEAB (1 mmol/l) P-value c-myb 1.00 1.11±0.09 0.272 runx1 1.00 1.16±0.20 0.483 Gene expression profiling and knockdown of aldh isoforms Abbreviations: DEAB, diethylaminobenzaldehyde; DMSO, dimethyl To examine the gene expression profile of different zebrafish sulfoxide. aldh isoforms, we have cloned and performed whole-mount ISH Data are presented as relative fold change compared with DMSO control. for individual aldh. All, except aldh5a1, were successfully PCR amplified in zebrafish embryos at 18 hpf. aldh1a2 was expressed in the lateral plate mesoderm at 11 hpf and in the somites and branchial arch primordium at 18 hpf (Figure 5a). aldh1a3 was DEAB treatment inhibits Aldefluor signal in zebrafish expressed ubiquitously at 11 hpf and more restrictively in the embryos developing eye at 18 hpf (Figure 5b). aldh16a1 was expressed in In human and mouse HSPC, ALDH activity is numerated by the ICM at 18 hpf (Figure 5c). The other isoforms were either Aldefluor assay (StemCell Technologies Inc.). Aldefluor activity undetectable by ISH, expressed ubiquitously or in non- is correlated with human HSPC in bone marrow, umbilical cord hematopoietic tissues (data not shown). We further screened blood and mobilized peripheral blood stem cells20–23 and is the hematopoietic functions of 15 aldh isoforms using MO. thought to reflect ALDH1A1 activity.3–5 However, it has not Dose-response experiments were performed to identify the been tested in zebrafish, in which the aldh1a1 isoform is maximally tolerable dosages for each MO. Furthermore, we lacking.14 Therefore, we examined if aldh activity might be successfully cloned the 50-UTR of 14 aldh genes upstream of an detectable in zebrafish and inhibited by DEAB. At 18 hpf, in-frame gfp coding sequence at C-terminal in the pcDNA3.1/ dissociated zebrafish embryos consistently exhibited an aldh þ CT-GFP-TOPO vector. The constructs, driven by cytomegalovirus

Leukemia ALDH regulates primitive hematopoiesis in zebraﬁsh embryos ACH Ma et al 2096 11 hpf 18 hpf

LPM somites BA aldh1a2

LPM aldh1a3

sense probe

developing eye

ICM aldh16a1

sense probe 250 µm

Figure 5 Expression of aldh isoforms during embryonic development. (a) aldh1a2 was expressed in the lateral plate mesoderm (LPM) at 11 hpf and in the somites and branchial arch primordium (BA) at 18 hpf. (b) aldh1a3 was expressed ubiquitously at 11 hpf and more restrictively in the developing eye at 18 hpf. (c) aldh16a1 was expressed in the ICM at 18 hpf. For ISH at 11 hpf, lateral view and dorsal view are shown for aldh1a2 with speciﬁc expression in LPM. For aldh1a3 and aldh16a1, only lateral views were shown. Controls ISH with sense RNA probes were also shown.

promoter, were injected into zebrafish embryos at one-cell stage Discussion and they gave rise to green fluorescence in a mosaic pattern. Co-injection with the respective MO completely abolished gfp We examined the role of aldh in zebrafish embryonic expression in all cases, confirming the specificity of molecular hematopoiesis using ALDH inhibitor DEAB and anti-sense targeting for each MO (Supplementary Figure S2). MO for 15 zebrafish aldh genes. DEAB significantly expanded Despite the effective molecular targeting, none of the MO primitive, but not definitive hematopoiesis during zebrafish recapitulated the expansion of primitive hematopoiesis as seen development. The hematopoietic phenotypes could be reversed in DEAB treatment. Interestingly, knockdown of aldh16a1 by ATRA at physiological dose. DEAB acted within a narrow resulted in significant reduction in gene expression associated window between gastrulation and somitogenesis. These with primitive hematopoiesis (Table 3). Knockdown of aldh1a2 observations supported the proposition that aldh modulated resulted in the absence of pectoral fin (Figure 6), phenocopying the initiation of primitive hematopoiesis through endogenous the aldh1a2 mutants as reported previously by Grandel et al.11 retinoic acid biosynthesis. Dorsoventral patterning was However, there was no hematopoietic phenotype in those unaffected, as the embryos were morphologically normal and embryos. We further tested the effects of concomitant genes associated with neural development was unperturbed. aldh1a2 and aldh1a3 knockdown by injecting both aldh1a2 None of the 15 MOs recapitulated the hematopoietic expansion and aldh1a3 MO at 2 ng each. A higher dosage was asso- by DEAB, either singly injected or in combination as in ciated with significant toxicity. At 18 hpf, the expression of concomitant aldh1a2 and aldh1a3 knockdown, suggesting a primitive hematopoietic genes was not affected (Table 4). complex mode of regulation. These data have shed novel lights To examine if knockdown of individual aldh isoforms to the current understanding of primitive hematopoiesis and its might have an effect on Aldefluor signal, we performed regulation. Aldefluor assay on each morphant. Unlike DEAB treatment, First, this was one of the first evidence demonstrating specific injection of these MOs, including co-injection of aldh1a2 role of aldh in modulating primitive hematopoiesis during HSPC and aldh1a3 MOs had no significant effect on Aldefluor signal specification. Very recently, de Jong et al.24 showed that DEAB (data not shown). treatment induced gata1 expression in wild type and partially

Leukemia ALDH regulates primitive hematopoiesis in zebraﬁsh embryos ACH Ma et al 2097 Table 3 Fold change of the gene expression associated with primitive hematopoiesis at 18 hpf in embryos injected with morpholinos targeting each aldh isoforms

Targeted Gene (dosage) scl gata1 a-eHb spi1 l-plastin

Control 1.00 1.00 1.00 1.00 1.00 aldh1a2 (6 ng) 1.14±0.07NS 1.13±0.04NS 0.89±0.05NS 1.08±0.07NS 1.09±0.10NS aldh1a3 (6 ng) 1.05±0.05NS 1.05±0.06NS 1.12±0.06NS 1.04±0.06NS 0.96±0.06NS aldh2a (4.5 ng) 0.96±0.07NS 0.82±0.06NS 0.94±0.14NS 1.08±0.16NS 1.30±0.34NS aldh2b (6 ng) 0.94±0.13NS 0.94±0.10NS 1.05±0.14NS 1.14±0.04NS 0.96±0.16NS aldh3a2 (3 ng) 1.12±0.09NS 0.87±0.10NS 1.18±0.24NS 1.65±0.17* 1.77±0.12* aldh3d1 (4.5 ng) 1.02±0.06NS 0.98±0.20NS 0.86±0.08NS 1.54±0.06* 0.89±0.06NS aldh4a1 (4.5 ng) 1.10±0.13NS 0.99±0.11NS 1.15±0.13NS 1.03±0.08NS 0.97±0.15NS aldh5a1 (6 ng) 0.94±0.11NS 1.11±0.15NS 0.95±0.14NS 0.91±0.18NS 1.12±0.13NS aldh6a1 (6 ng) 1.03±0.15NS 1.04±0.06NS 1.09±0.04NS 1.05±0.07NS 1.43±0.08* aldh7a1 (9 ng) 1.03±0.16NS 1.09±0.13NS 1.03±0.15NS 0.85±0.14NS 1.050±0.09NS aldh8a1 (6 ng) 0.94±0.04NS 0.98±0.06NS 1.04±0.16NS 0.91±0.07* 0.97±0.04NS aldh9a1a (3 ng) 1.13±0.03NS 1.38±0.06** 1.22±0.14NS 1.85±0.13*** 1.19±0.09NS aldh9a1b (3 ng) 1.03±0.02NS 1.01±0.20NS 1.32±0.19NS 1.23±0.26NS 1.08±0.06NS aldh16a1 (3 ng) 0.95±0.07NS 0.66±0.05* 0.47±0.04* 1.01±0.09NS 0.94±0.13NS aldh18a1 (4.5 ng) 1.00±0.05NS 1.06±0.15NS 0.92±0.19NS 1.07±0.13NS 0.89±0.07NS Data are presented as relative fold change compared with control injected with random sequence morpholino. P-values are presented as *(Po0.05); **(Po0.01); ***(Po0.005); NS (not signiﬁcant).

Control aldh1a2 MO Table 4 Fold change of the gene expression associated with primitive hematopoiesis at 18 hpf in embryos co-injected with both morpholinos targeting aldh1a2 and aldh1a3

Gene Control aldh1a2+aldh1a3 morpholino P-value

scl 1.00 1.03±0.09 0.770 72 hpf gata1 1.00 0.95±0.11 0.642 a-eHb 1.00 0.97±0.05 0.613 250 μm 250 μm spi1 1.00 1.16±0.13 0.375 l-plastin 1.00 1.11±0.14 0.542 aldh1a2 and aldh1a3 morpholinos are injected at 2 ng each. Data are presented as relative fold change compared with control injected with random sequence morpholino at 4 ng. 5 dpf

ATRA alone had no effect on hematopoiesis. This had enabled 250 μm 250 μm us to address specifically the mechanism of action of aldh, Figure 6 Defects in pectoral fin bud development in zebrafish independent of the ATRA effects on dorsoventral patterning and embryos after knockdown of aldh1a2. Comparing with control vascular formation, which was often seen at pharmacological injected with random sequence morpholino (a, c), aldh1a2 morphants doses.24–26 (b, d) showed the absence of pectoral fin at both 72 hpf and 5 day- Second, the present study has suggested a complex mode of post-fertilization (dpf). Red arrows indicate normal developing regulation of primitive hematopoiesis by aldh. We performed pectoral fins in control embryos. extensive gene expression profiling and functional screening for all 15 known zebrafish aldh isoforms. In mammals, ALDH1A1 has been implicated in the regulation of HSPC functions.2,3,4,6 rescued gata1 expression in the cdx4 mutant embryos. However, this isoform is lacking in zebrafish.14 Mutation of Remarkably, the rescue manifested only when the embryos aldh1a2, the neckless gene, has been described, but the mutant were treated with DEAB between gastrulation and somito- embryos had no hematopoietic phenotype.10,11 Whole-mount genesis. Taken together, our data and theirs provided unequi- ISH showed that aldh1a2 was expressed in lateral and paraxial vocal evidence that aldh is an important modulator of primitive mesoderm at 11 hpf. The results corroborated with those hematopoiesis. In addition, we demonstrated quantitatively that of Begemann et al.10 and Grandel et al.11 Furthermore, we DEAB treatment up to 18 hpf induced an approximately two-fold demonstrated for the first time the expression of aldh16a1 in the increase in HSPC and erythroid, and a nearly three-fold increase ICM at 18 hpf. When individual aldh gene was knocked down in myeloid gene expression. The hematopoietic expansion was with MO, none of them recapitulated the hematopoietic preceded by an increase in scl and lmo2 expression in both the phenotype of DEAB. We have also knocked down both aldh1a2 anterior and posterior lateral plate mesoderm at 12 hpf and and aldh1a3 simultaneously to exclude functional redundancy could be abolished by concomitant scl knockdown, supporting among members of the aldh1 family. Despite the presence of the proposition that DEAB increased HSPC population at early pectoral fin defect, which have been reported in the no-fin specification. Furthermore, physiological dose (1 nmol/l) of mutant embryos,11 there was no increase in hematopoiesis. ATRA completely reversed the effects of DEAB. This dosage of The lack of hematopoietic phenotype in these experiments

Leukemia ALDH regulates primitive hematopoiesis in zebrafish embryos ACH Ma et al 2098 might result from compensatory upregulation of alternative or References perhaps multiple aldh isoforms. Alternatively, DEAB might inhibit a yet unidentified aldh isoform that modulate HSPC 1 Russo JE, Hilton J, Colvin OM. The role of aldehyde dehydro- specification. These possibilities remained to be distinguished. genase isozymes in cellular resistance to the alkylating agent Third, our data demonstrated for the first time DEAB-sensitive cyclophosphamide. Prog Clin Biol Res 1989; 290: 65–79. Aldefluor activity in zebrafish embryos. This assay has been 2 Chute JP, Muramoto GG, Whitesides J, Colvin M, Safi R, Chao NJ et al. Inhibition of aldehyde dehydrogenase and retinoid signaling optimized to enumerate human HSPC and is thought to be a 5,27 induces the expansion of human hematopoietic stem cells. surrogate of ALDH1A1 activity. However, its use in other Proc Natl Acad Sci USA 2006; 103: 11707–11712. vertebrates has not been clearly validated. In particular, targeted 3 He X, Gonzalez V, Tsang A, Thompson J, Tsang TC, Harris DT. deletion of Aldh1a1 gene in mouse bone marrow had no effect Differential gene expression profiling of CD34+ CD133+ umbilical on Aldefluor or HSPC activity,7 underscoring a complex cord blood hematopoietic stem progenitor cells. Stem Cells Dev relationship between Aldefluor, aldh and HSPC activity. In the 2005; 14: 188–198. 4 Forsberg EC, Prohaska SS, Katzman S, Heffner GC, Stuart JM, present study, only DEAB, but none of the MO inhibited Weissman IL. Differential expression of novel potential regulators Aldefluor signal, reminiscent of their respective effects on in hematopoietic stem cells. PLoS Genet 2008; 1: e28. hematopoiesis. This observation led to the hypothesis that 5 Cheung AM, Wan TS, Leung JC, Chan LY, Huang H, Kwong YL Aldefluor activity is associated with aldh relevant to the et al. Aldehyde dehydrogenase activity in leukemic blasts defines a regulation of primitive hematopoiesis. HSPC-specific Aldefluor subgroup of acute myeloid leukemia with adverse prognosis and activity in zebrafish embryos is being examined and will provide superior NOD/SCID engrafting potential. Leukemia 2007; 21: 1423–1430. us with novel insights to this link. 6 Muramoto GG, Russell JL, Safi R, Salter AB, Himburg HA, Daher P Finally, our data provided direct evidence for the distinctive et al. Inhibition of aldehyde dehydrogenase expands hemato- regulation of primitive and definitive hematopoiesis. Specifi- poietic stem cells with radioprotective capacity. Stem Cells 2010; cally, although DEAB treatment significantly expanded primitive 28: 523–534. hematopoiesis, it had no effects on the transitory erythromyeloid 7 Levi BP, Yilmaz OH, Duester G, Morrison SJ. Aldehyde precursors and definitive hematopoiesis, as shown by both dehydrogenase 1a1 is dispensable for stem cell function in the mouse hematopoietic and nervous systems. Blood 2009; 113: ISH and quantitatively by real-time PCR. Definitive hemato- 28 1670–1680. poiesis was distinctively regulated by prostaglandin E2, 8 Goldie LC, Lucitti JL, Dickinson ME, Hirschi KK. Cell signaling 29 30 eNOS, Notch signaling. On the other hand, our model directing the formation and function of hemogenic endothelium highlighted the specific regulation of primitive hematopoiesis by during murine embryogenesis. Blood 2008; 112: 3194–3204. aldh. DEAB treatment at low dosage (1 mmol/l) expanded 9 Davidson AJ, Zon LI. The ‘definitive’ (and ‘primitive’) guide to primitive HSC without affecting the dorsoventral patterning, zebrafish hematopoiesis. Oncogene 2004; 23: 7233–7246. 10 Begemann G, Schilling TF, Rauch GJ, Geisler R, Ingham PW. The suggesting that aldh and retinoic acid might have a direct effect zebrafish neckless mutation reveals a requirement for raldh2 in on HPSC. mesodermal signals that pattern the hindbrain. Development Our data have provided us with solid ground for future 2001; 128: 3081–3094. research. In particular, we demonstrated that aldh16a1 was 11 Grandel H, Lun K, Rauch GJ, Rhinn M, Piotrowski T, Houart C expressed specifically in the ICM and its knockdown by MO et al. Retinoic acid signalling in the zebrafish embryo is necessary significantly reduced erythroid, but not HSPC or myeloid gene during pre-segmentation stages to pattern the anterior-posterior expression, suggesting that aldh16a1 might have distinctive axis of the CNS and to induce a pectoral fin bud. Development 2002; 129: 2851–2865. function in the initiation and/or maintenance of primitive 12 Keegan BR, Feldman JL, Begemann G, Ingham PW, Yelon D. erythropoiesis. Its function in mammalian hematopoiesis is Retinoic acid signaling restricts the cardiac progenitor pool. completely unknown and is currently examined in our Science 2005; 307: 247–249. laboratory. In addition, we have collected at least 15 different 13 Perz-Edwards A, Hardison NL, Linney E. Retinoic acid-mediated aldh MOs and molecular targeting has been confirmed in 14 of gene expression in transgenic reporter zebrafish. Dev Biol 2001; them. At present, information about the developmental function 229: 89–101. 14 Pittlik S, Domingues S, Meyer A, Begemann G. Expression of of ALDH isoforms outside the ALDH1 family was scarce. In this zebrafish aldh1a3 (raldh3) and absence of aldh1a1 in teleosts. respect, our MO database may form the groundwork for Gene Expr Patterns 2008; 8: 141–147. targeting individual aldh in zebrafish embryos. 15 Long Q, Meng A, Wang H, Jessen JR, Farrell MJ, Lin S. GATA-1 expression pattern can be recapitulated in living transgenic zebrafish using GFP reporter gene. Development 1997; 124: 4105–4111. Conflict of interest 16 Dooley KA, Davidson AJ, Zon LI. Zebrafish scl functions independently in hematopoietic and endothelial development. The authors declare no conflict of interest. Dev Biol 2005; 277: 522–536. 17 Ma AC, Ward AC, Liang R, Leung AY. The role of jak2a in zebrafish hematopoiesis. Blood 2007; 110: 1824–1830. 18 Ma AC, Chung MI, Liang R, Leung AY. The role of survivin2 in Acknowledgements primitive hematopoiesis during zebrafish development. Leukemia 2009; 23: 712–720. 19 Bertrand JY, Kim AD, Violette EP, Stachura DL, Cisson JL, Traver D. We thank Dr Anming Meng (Tsinghua University, China) for the Definitive hematopoiesis initiates through a committed erythro- generous gift of the Tg(gata1:gfp) fish-lines. The works were myeloid progenitor in the zebrafish embryo. Development 2007; supported by the General Research Fund (HKU 752006M, 134: 4147–4156. 770308M and 769809M), the Seed Funding Programme for Basic 20 Hess DA, Meyerrose TE, Wirthlin L, Craft TP, Herrbrich PE, Research (200711159060), Small Project Funding to ACH Ma Creer MH et al. Functional characterization of highly puri- fied human hematopoietic repopulating cells isolated based (200907176072) and Seed Funding Programme for Basic on aldehyde dehydrogenase activity. Blood 2004; 104: Research to R Liang (200611159093) as well as a grant from the 1648–1655. strategy research theme of cancer stem cells in the University of 21 Pearce DJ, Taussig D, Simpson C, Allen K, Rohatiner AZ, Lister TA Hong Kong. et al. Characterization of cells with a high aldehyde dehydrogenase

Leukemia ALDH regulates primitive hematopoiesis in zebrafish embryos ACH Ma et al 2099 activity from cord blood and acute myeloid leukemia samples. Stem 26 Wingert RA, Selleck R, Yu J, Song HD, Chen Z, Song A et al. The Cells 2005; 23: 752–760. cdx genes and retinoic acid control the positioning and 22 Lioznov MV, Freiberger P, Kro¨ger N, Zander AR, Fehse B. segmentation of the zebrafish pronephros. PLoS Genet 2007; 3: Aldehyde dehydrogenase activity as a marker for the quality of 1922–1938. hematopoietic stem cell transplants. Bone Marrow Transplant 27 Ginestier C, Hur MH, Charafe-Jauffret E, Monville F, Dutcher J, 2005; 35: 909–914. Brown M et al. ALDH1 is a marker of normal and malignant 23 Fallon P, Gentry T, Balber AE, Boulware D, Janssen WE, Smilee R human mammary stem cells and a predictor of poor clinical et al. Mobilized peripheral blood SSCloALDHbr cells have the outcome. Cell Stem Cell 2007; 1: 555–567. phenotypic and functional properties of primitive haematopoietic 28 North TE, Goessling W, Walkley CR, Lengerke C, Kopani KR, Lord cells and their number correlates with engraftment following AM et al. Prostaglandin E2 regulates vertebrate haematopoietic autologous transplantation. Br J Haematol 2003; 122: 99–108. stem cell homeostasis. Nature 2007; 447: 1007–1011. 24 de Jong JL, Davidson AJ, Wang Y, Palis J, Opara P, Pugach E et al. 29 North TE, Goessling W, Peeters M, Li P, Ceol C, Lord AM et al. Interaction of retinoic acid and scl controls primitive blood Hematopoietic stem cell development is dependent on blood flow. development. Blood 2010; 116: 201–209. Cell 2009; 137: 736–748. 25 Waxman JS, Yelon D. Increase Hox activity mimics the teratogenic 30 Bertrand JY, Cisson JL, Stachura DL, Traver D. Notch signaling effects of excess retinoic acid signaling. Dev Dyn 2009; 238: distinguishes 2 waves of definitive hematopoiesis in the zebrafish 1207–1213. embryo. Blood 2010; 115: 2777–2783.

Supplementary Information accompanies the paper on the Leukemia website (http://www.nature.com/leu)

Leukemia