Copper overload impairs hematopoietic stem and progenitor cell proliferation via prompting HSF1/SP1 aggregation and the subsequently downregulating FOXM1-Cytoskeleton axis

Summary Unbalanced Cu homeostasis has been suggested to be associated with hematopoietic disease, but the roles of Cu overload in the hematopoietic system and the potential mechanisms are obscure. Here, we report a novel association and the novel potential pathways for Cu overload to induce proliferation defects in zebrafish embryonic hematopoietic stem and progenitor cells (HSPCs) via down-regulating expression of foxm1-cytoskeleton axis, which is conserved from fish to mammals. Mechanistically, we show the direct binding of Cu to transcriptional factors HSF1 and SP1 and that Cu overload induces the cytoplasmic aggregation of proteins HSF1 and SP1. These result in the reduced transcriptional activities of HSF1 and SP1 on their downstream FOXM1 as well as the FOXM1 transcriptional activities on cytoskeletons in HSPCs, which leads to ultimately cell proliferation impairment. These findings unveil the novel linkage of Cu overload with specific signaling transduction as well as the subsequent HSPC proliferation defects.


INTRODUCTION
The unbalanced homeostasis of copper (Cu), an essential trace element in living organisms, can lead to developmental defects and various diseases. Cu deficiency causes the blockage of hematopoietic cell differentiation, leading to anemia, neutropenia and thrombocytopenia. [1][2][3] Cu overload in the body can cause a variety of hematopoietic diseases, such as acute hemolytic anemia and leukemia, 4 and higher serum Cu is observed in children with leukemia, [4][5][6] suggesting unbalanced Cu homeostasis is strictly related to hematopoietic diseases although the potential differences and mechanisms are complicate and unknown.
The transcriptional and signaling orchestration is usually common in the development of both hematopoietic stem and progenitor cells (HSPCs) and leukemia, and the orchestration of the development of HSPCs is highly conserved from zebrafish to mammals. 7 The HSPC emerge in zebrafish embryos initiates around 26-28 hours post fertilization (hpf) and is termed the definitive hematopoiesis, differentiating from flk1 positive hemogenic endothelium located in aorta-gonad mesonephros (AGM), followed by migration to caudal hematopoietic tissue (CHT) and proliferation. The embryonic HSPCs can be transplantable, 8 marked by genes cmyb and runx1. 9 Recently, studies have unveiled that Cu overload induces embryonic skeletal myofibrillogenesis 10 and neural system myelin defects, 11 and impairs embryonic angiogenesis and lymphangiogenesis. 12 Meanwhile, it is reported that Cu overload has a strong negative effect on zebrafish hematopoietic system via reducing hematopoiesis potential of head kidney in common carp, 13 suggesting the potential linkage of Cu overload with hematopoiesis potential of HSPCs. In this study, we will study the roles of Cu overload in the emergence and proliferation of HSPCs.
A clear picture of the mechanisms underlying Cu overload-induced developmental defects during vertebrate embryogenesis has not yet emerged, although Cu overload impairs angiogenesis and lymphangiogenesis via blocking the migration rather than the proliferation of endothelial cells, 12 induces myofibrillogenesis 10 and myelination defects 11 via epigenetic regulation, and induces retinal and intestinal development defects via triggering endoplasmic reticulum (ER) stress and reactive oxygen species The collected runx1 promoter driving GFP + cells (runx1GFP + cells) showed abundant expression of HSPC genes runx1, cmyb, and gata2b, while little expression of neural genes (olig2, mbp), liver gene (fabp2), and muscle gene (myod), suggesting the HSPC identity of the runx1GFP + cells (Figures 1E and S1B) as reported recently. 27 After FACS analysis of runx1GFP + cells (HSPCs) labeled by Cu ion probe, the proportion of PE + GFP + cells labeled with Cu probe ( Figure 1A) and the Cu content in individual runx1GFP + cells (HSPCs) ( Figure 1B) were found to be significantly increased (p < 0.05) in the Cu-stressed embryos. Additionally, red fluorescence was observed in some cells but not in the runx1GFP + cells of the whole embryos labeled with Cu ion probe in low magnification ( Figure S1C). HSPCs showed a significant reduction in the number and percentage in Cu-stressed embryos at both 33 hpf ( Figure 1C) and 58 hpf ( Figure 1D), with significantly reduced expression of the HSPC genes runx1, cmyb ( Figures 1E and 1F), and gata2b (Figure 1E), suggesting the intracellular Cu accumulation and the subsequently impaired the production of HSPCs.
Additionally, posterior lateral mesoderm gene pax2a and trunk mesoderm gene myoD exhibited no obviously changed expression in Cu-stressed zebrafish embryos at 14 hpf ( Figures S2C1 and S2C2), and upregulated expression of scl with no changed expression of gata1a, mpx, fli1 and flk1 existed in Cu-stressed zebrafish embryos ( Figures S2C3-S3C12), suggesting Cu overload exerted little effects on primitive hematopoiesis in zebrafish embryos, and the intracellular Cu accumulation induced impaired production of HSPCs was specific.
Effect of copper chelator tetrathimolybdate and copper ionophores elesclomol on hematopoietic stem and progenitor cell development in zebrafish embryos Copper transporter 1 (Ctr1) is a membrane Cu transporter that plays essential roles in Cu acquisition and for zebrafish development. 28 In this study, gene ctr1 showed identical expression in runx1GFP + cells and in runx1GFP À cells, while Cu stresses significantly suppressed ctr1 transcriptional expression in runx1GFP + cells at 33 hpf ( Figure S2B1) and in the whole embryos at 33 hpf and 72 hpf (Figures S2B2-S2B5). iScience Article Cu chelators remove copper ions from the body, but Cu ionophores are small molecules binding copper ions and help transport copper ions into the cell. 29,30 To further identify the roles of intracellular Cu accumulation in HSPC development, we treated the embryos with TTM and elesclomol respectively and tested their HSPC development. TTM-treated embryos exhibited hypo-pigmentation at 33 hpf ( Figures 2A3 and  2A9), and the embryos exhibited a slightly developmental delay with shorten body and enlarged yolk sac at 72 hpf ( Figures 2A15 and 2A21). Meanwhile, the embryos in Cu 2+ and elesclomol co-treated group exhibited obvious trunk abnormalities, enlarged yolk sac, and thoracic enlargement at 33 hpf and 72 hpf ( Figures 2A6, 2A12, 2A18, and 2A24).
Compared with the Cu concentration in the control embryos at 33 hpf, it was significantly decreased (p < 0.05) to 0.07 G 0.02 mg/g in TTM-treated embryos, was significantly increased (p < 0.05) to (E) The expression of HSPC genes runx1, cmyb, and gata2b in runx1GFP À cells, runx1GFP + cells, and Cu-stressed runx1GFP + cells, respectively. (F) The expression of HSPC genes runx1 and cmyb in embryos at 33 hpf and 72 hpf. F9, F10, calculation of runx1 and cmyb expression level in the Cu-stressed and control embryos, respectively. Each experiment was repeated three times, and a representative result is shown. N changed /N total in the right bottom corner of each panel indicates embryos with changed expression/total tested embryos, and N in calculation panels indicates the number of embryos with changed expression in each group. The same for the numbers in the following figures. F1-F8, lateral view, anterior to the left, and dorsal to the up. Data are mean G SD. t-test, *p < 0.05, **p < 0.01, ***p < 0.001. Scale bars, 20 mm (D1-D2), 50 mm (F1-F4), and 200 mm (F5-F8).
12.29 G 0.23 mg/g in Cu-stressed embryos, and was further significantly increased (p < 0.05) to 17.53 G 1.63 mg/g in Cu and elesclomol co-treated embryos ( Figure 2B). Consistently, the variation tends of Cu concentration in individual runx1GFP + cells (HSPCs) at different treatments were similar to that of the total Cu concentration in zebrafish embryos from different groups ( Figure 2C). The percentage of HSPCs and expression of the HSPC genes showed a slight reduction in TTM-treated embryos at 33 hpf, and their expression decreased sharply in Cu and elesclomol co-treated embryos ( Figures 2D and 2E). Meanwhile, in Cu and elesclomol co-treated embryos, more concentrated Cu and more reduced HSPCs were observed compared to that in Cu single-treated embryos ( Figures 2D and 2E).

Cu overload inhibits hematopoietic stem and progenitor cell proliferation in zebrafish embryos
Cell cycle damage was potentially underlying mechanisms for reduced cell emergence during embryogenesis, 31,32 so runx1GFP + cells (HSPCs) were sorted at 33 hpf and used first for cell cycle detection in this study ( Figure 3A1). The proportion of HSPCs was reduced significantly (p < 0.05) at G2/M stage from 16.57 G 1.92% to 9.37 G 2.63%, while increased (p < 0.05) at the G1 stage from 77.44 G 2.45% to 85.38 G 1.12% in Cu-stressed embryos at 33 hpf (Figures 3A2 and  Cytoskeleton regulates copper overload-induced hematopoietic stem and progenitor cell deficiency Runx1GFP + cells (HSPCs) were sorted at 33 hpf for RNA deep sequencing (RNA-Seq) analysis ( Figure 4A and Tables S11, S12, S13, S14, S15, S16, S17, and S18). RNA-Seq unveiled the elevated erythrocyte differentiation and myeloid cell homeostasis ( Figures S3A3, S3A4 and Tables S11, and S12) as reported recently. 33,34 Differentially expressed cytoskeleton genes exhibited significantly reduced expression and enrichment in Cu overload HSPCs ( Figures 4B and S3A2), which was further confirmed by qRT-PCR (Figure S3B) and WISH assays ( Figure S3C). Additionally, damage to the cytoskeleton of runx1GFP + cells (HSPCs) and reduction in the cytoskeleton positive immunofluorescence signaling were observed ( Figure 4C).
Four cytoskeletal genes (tubala, stmn4, tmsb2, tubb5) downregulated responding to Cu overload in HSPCs were selected for further analysis. The four morphants exhibited little developmental defects independent of p53 (Co-injection of p53-MO is an effective strategy to test the specificity or off-target effects of MOs 35-37 ( Figure S4A). Individual gene knockdown further confirmed that impaired HSPC emergence and proliferation occurred in the four morphants independent of functional p53 ( Figures S4B-S4D). A striking HSPC-restricted phenotype that a remarkable reduction in runx1GFP + cell (HSPCs) numbers (Figure S5B) and in G2/M stage HSPCs ( Figure S5C), was found in the four cytoskeleton genes together Figure 2. Cu chelator TTM and ionophore elesclomol affect embryonic and HSPC development during zebrafish embryogenesis (A) Phenotypes of zebrafish embryos in the control, DMSO exposed, TTM exposed, Cu exposed, Cu and DMSO co-exposed, or Cu and elesclomol co-exposed group, at 33 hpf and 72 hpf, respectively. (B) Cu concentration in zebrafish embryos from the control, DMSO exposed, TTM exposed, Cu exposed, Cu and DMSO co-exposed, or Cu and elesclomol co-exposed group at 33 hpf, respectively. (C) Cu content in individual runx1GFP + cells (HSPCs) of zebrafish embryos in control, DMSO exposed, TTM exposed, Cu exposed, Cu and DMSO co-exposed, or Cu and elesclomol co-exposed group at 33 hpf, respectively. iScience Article morphants. Moreover, the cytoskeleton genes together morphants showed a similar spindle malformation of HSPCs in metaphase as that of HSPCs in Cu-stressed embryos ( Figure 4D), with severely damaged cytoskeleton protein distribution and the down-regulated cytoskeleton positive fluorescence in HSPCs in embryos with knockdown of either cytoskeleton gene tuba1a, stmn4, tmsb2, or tubb5 ( Figure S5D). These findings are consistent with Cu's distinctive connection to cytoskeleton-mediated HSPC cell cycle impairments.
Foxm1 acts upstream of cytoskeleton in mediating copper overload-induced hematopoietic stem and progenitor cell deficiency We further checked the RNA-Seq data of both Cu overload embryos 38 and HSPCs (this study) (Tables S19, S20, and S21), and found the down-regulated expression of foxm1 (Tables S20 and S21) [which is essential for proper cell cycle progression by regulating the G1/S and G2/M transition and the execution of the mitotic program, 39 is also required for maintaining HSPC pool in mammal adult 40 ] ( Figure S7A), and its reduced expression was further confirmed by qRT-PCR ( Figure S7B1) and WISH ( Figures S7B2-S7B4). Its knockdown induced a little developmental defect in zebrafish independent of iScience Article gene p53 ( Figure S7C) as well as a significant reduction in the expression of HSPC markers ( Figure S7D) and the number of flk1 + runx1 + cells ( Figure S7E). Additionally, foxm1 knockdown led to a significant reduction (p < 0.05) of HSPCs at G2/M phase ( Figure S7F), and showed disorganized spindle microtubule organization of HSPCs in metaphase ( Figure 5A)  iScience Article and 5C) exhibited almost normal-like phenotype ( Figure 5D), but with the significant drop of HSPCs cells at both 33 hpf and 72 hpf ( Figure 5E). Overexpression of foxm1 rescued the HSPC phenotype in foxm1 À/À mutant ( Figure S8C).

Copper directly binds transcriptional factors HSF1/SP1 and induces their cytoplasmic aggregation
The transcriptional activities of human FOXM1 promoter in cells were significantly decreased after Cu stresses ( Figure 8A2). 5 truncated promoters of gene FOXM1 (À2100/+28, À1086/+28, À755/+28, À316/+28, À106/+28) were constructed to unveil the Cu targeted domain on FOXM1 promoter ( Figure 8A1) based on previous report. 43 Cu overload was shown to significantly suppress the transcriptional activities of À2100/+28 and À1086/+28 of FOXM1 promoter rather than À755/+28, À316/+28, À106/+28 ( Figure 8A3). HSF1 has been unveiled to effect on the À2100/-1086 region and SP1 on the À1086/-755 region on gene FOXM1 promoter 43 ( Figure 8A1), and the transcriptional activities of FOXM1 promoter were found to be significantly up-regulated by ectopic expression of both zebrafish and human HSF1 or SP1, but significantly inhibited after Cu stresses ( Figures 8B and S10B). Meanwhile, ATOX1 is a Cu-dependent transcriptional factor in the activation of platelet-derived growth factor (PDGF), SOD3, and cyclinD1. 15,16 In this study, ectopic expression of human ATOX1 inhibited the transcriptional activity of FOXM1, and its transcriptional activity was significantly inhibited iScience Article after Cu stresses (Figures S10C1 and S10C3). However, ectopic expression of zebrafish ATOX1 did not change the transcriptional activity of FOXM1 although its transcriptional activity was significantly inhibited after Cu stresses (Figures S10C2 and S10C4).
The protein levels remained unchanged in SP1 and HSF1 after Cu stresses ( Figures 8C1 and S10D). The mRNA levels of hsf1 and sp1 showed identical in runx1GFP + cells and in runx1GFP À cells, and maintained stable in runx1GFP + cells after Cu stresses ( Figure S10E). However, SP1, MYC-SP1, or MYC-HSF1 ChIP-qPCR analyses showed that SP1 or HSF1 signals were less enriched at the promoter of FOXM1 in Custressed group compared with the control (Figures 8C2-8C4), and Cu overload significantly up-regulated expression of HSF1 negative target BAX 44,45 while significantly down-regulated expressions of HSF1 positive target CDC6 and CKS1B 46 and SP1 positive targets CDC25A 47 and MYCA 48 ( Figure 8D).
The experiments described above establish a connection between Cu and the expression of HSF1/SP1 but do not establish a direct mechanistic link. We hypothesized that Cu might directly bind to HSF1/SP1 iScience Article proteins. To test this hypothesis, we purified HSF1/SP1 via Cu 2+ -charged resins from total K562 cell lysates as studies performed recently, 29 and found that SP1 proteins bound to Cu 2+ -charged resin but not to Fe 3+ resins, while HSF1 proteins bind to both Cu 2+ and Fe 3+ charged resins (Figures 8E and S10F). Meanwhile, Cu overload caused pronounced induction of Cu-HSF1/SP1 foci in the cytoplasm in K562 cells ( Figure 8F), and Cu-SP1 foci were significantly more increased in K562 cells co-treated with Cu and elesclomol, but significantly decreased in K562 cells treated with TTM ( Figure S10G).

DISCUSSION
It is well-known that the unbalanced Cu homeostasis in individuals is associated with hematopoietic system diseases even leukemia, 4 and Cu ionophores (increasing bioavailable Cu, inducing Cu overload) have been suggested as agents in treating cancer which characterized with active cell proliferation, 50,51 but little is known about the related molecular characteristics and underlying mechanisms. Here, we unveiled for the first time that bioavailable Cu at a certain concentration range could impair the cell cycle in both zebrafish HSPCs and mammalian cells, and show the effects of intracellular Cu on the development of zebrafish HSPCs. Overload Cu directly interacts with HSF1/SP1 and induces the cytoplasmic aggregation of the proteins. This results in the reduced transcriptional activities of the two proteins on FOXM1 as well as FOXM1 transcriptional activities on cytoskeleton genes, which leads to the ultimately HSPC cell cycle and proliferation impairments.
Increased Cu content in HSPCs with a significant reduction in the number of nascent and fetal HSPCs as well as G2/M phase BrdU + are observed in Cu overload embryonic HSPCs, strengthening that Cu induces HSPC developmental defects is primarily dependent on intracellular Cu accumulation. Moreover, we observe down-regulated expression of cytoskeleton genes and proteins as well as impaired cytoskeleton protein distribution in Cu overload HSPCs, which is consistent with reports that orderly and accurate rearrangement of cytoskeleton helps push the cell cycle forward, 52,53 while their integrity disruptions lead to cell-cycle arrest. 52,54 In this study, we demonstrate that Cu at a certain concentration range (nearly 2-fold increase) damages cell cycle, suggesting Cu might inhibit cell proliferation via inhibiting cell cycle, at the concentration which also induces cell senescence in zebrafish embryonic blood vessels 12 and in mammalian cells. 55 Meanwhile, the down-regulated expression of foxm1 in Cu overload HSPCs, the spatial distribution disorder of cytoskeleton in HSPCs caused by foxm1 knockdown/knockout, with the observation that the conservative roles of FOXM1 in directly regulating cytoskeleton genes and cell cycle from fish to mammalian cells are shown in this study, consistently with reports that FOXM1 regulate the cell cycle by regulating the structure of chromosomal bodies, 56 activates the transcription of CDC25B requiring for G2/M transition, 57,58 and affects both G1/S and G2/M transition during cell cycle. 57 Consistently, this study observes the significantly reduced G2/M stage in the cell cycle in both Cu overload and in FOXM1 knockdown embryonic Cu ionophores induced Cu overload is reported to reduce tumor growth and possess anticancer activities, [62][63][64] and induce a distinct form of regulated cell death termed cuproptosis in a 15-to 60-fold increase intracellularly. 29 Meanwhile, some studies have reported that Cu chelators (reducing bioavailable Cu) functioned effectively in the expansion or maintenance of progenitor cells in vitro 65 as well as in patients with lung cancer (a high serum Cu level, >23.6 mM) 66 and with breast cancers (average: 51.2 mM), 67 suggesting Cu bioavailable or not might be the potential differences of system Cu overload in cell proliferation, in cell senescence, or in cell cuproptosis. 50,51 In this study, we observe that Cu ionophore elesclomal co-treated with Cu induces more Cu probe positive fluorescence and more accumulated Cu in HSPCs and the resulted in more reduced percentage of HSPCs in embryos compared with that in Cu single-treated embryos. The observations in this study are not only consistent with reports that Cu ionophores induced Cu overload possess anticancer (anti-cell proliferation) activities, but also suggest that Cu overload impairs HSPC development is concentration-dependent. However, Cu chelator TTM not only induces reduced Cu probe positive fluorescence and little accumulated Cu in HSPCs compared with that in WT, but also lead to reduced percentage of HSPCs in embryos, suggesting Cu homeostasis is strictly for the emergence, maintenance, or amplification of HSPCs, Cu overload or Cu deficiency, both induce reduced HSPCs during zebrafish embryogenesis.
Cu significantly suppressed the transcriptional activities of FOXM1 promoter in both À2100 to +28 and À1086 to +28 rather than in À775 to +28, À316 to +28, and À106 to +28, where transcriptional factors SP1(-891), HSF1(-1792/-1767), and so forth are pivotal in regulating FOXM1 transcription. 43 Consistently, we show that the ectopic expression of either SP1 or HSF1 could significantly up-regulate the FOXM1 transcriptional activities, while the up-regulated transcriptional activities of both SP1 and HSF1 could be significantly inhibited by Cu overload. In this study, we unveil that Cu binds directly with transcriptional factors iScience Article SP1 or HSF1, and Cu overload significantly reduces the binding enrichment of either SP1 or HSF1 in the gene FOXM1 promoter, and induces the aggregation of proteins SP1 and HSF1 in the cytoplasm in HSPCs. These findings support a model that the down-regulated expression of FOXM1 in HSPCs is mediated at least in part by the aberrant oligomerization of both SP1 and HSF1 proteins in the cytoplasm and their subsequently reduced functional transcriptional activities on FXOM1 and other targeting genes BAX, CDC6, CKS1B, CDC25A, and MYCA. Consistently, a recent study also reports that Cu-induced cell death is mediated at least in part by the aberrant oligomerization of lipoylated TCA cycle proteins. 29 ATOX1 is a well-known Cu-dependent transcriptional factor, however, this study unveils that Cu regulates FOXM1 transcription independent of ATOX1. Additionally, different from ATOX1 dependent on Cu in the activation of PDGF, SOD3, and cyclin D1, 15 this study unveils that Cu suppressed transcriptional activities of SP1 or HSF1 on gene FOXM1, suggesting the double-sided effects of Cu on nuclear transcriptional proteins depending on the Cu spatial distribution. Moreover, whether other Cu binding transcriptional factors in nuclear, such as MTF1/Nrf2, MT1/2, and ect., 22 are involved in Cu overload-induced HSPC proliferation defects, needs to be studied in the further days.
HSF1 has been unveiled to be a master regulator in the heat shock response 68,69 and in responses to environmental conditions such as metals and so forth 70 promotes hematopoietic stem cell fitness and proteostasis in response to ex vivo culture stress, 71 activates beige fat metabolism against obesity. 72 SP1, which belongs to the SP-like family, is a transcription factor that controls diverse cell functions and behaviors, 73 and SP1 has been shown to be required for the transcriptional regulation of Cu transporter hCtr1 in response to the cellular Cu level 74 and to be a pivotal transcriptional factor in early embryonic development. 75 In this study, we show that Cu directly interacts with HSF1 and SP1 proteins and induces the cytoplasmic aggregation of the two transcriptional proteins. These findings expand the novel roles of both SP1 and HSF1 in linkage with Cu in the development of HSPCs, and widen the Cu binding transcriptional factors from MTF1/Nrf2, ATOX1, and ect., to MTF1/Nrf2, ATOX1, SP1, and HSF1, and ect..
In the case of genetic disorders of Cu homeostasis (atp7b À/À for Wilson's disease, atp7a À/À for Menke's disease, cox17 À/À ), our observations that no obvious change is observed in the expression of HSPC genes and foxm1 in atp7a À/À mutants with or without Cu stresses, and reduced expression of HSPC genes in atp7b À/À and both foxm1 and HSPC genes still exhibits down-regulated after Cu stresses in the mutants, not only strengthening the down-regulated expression of foxm1 and HSPC proliferation impairment are primarily dependent on intracellular Cu accumulation because atp7a À/À mutants lack the ability to pump Cu into the portal circulation and exhibit Cu overload in intestines while deficiency in other tissues, 76 while atp7b À/À mutants lack the ability to excrete excess cellular Cu, 77 but also indicating hematopoiesis potential defects might be another contributor for anemia occurred in Wilson patients. Meanwhile, expression of gene HSPC genes and foxm1 in cox17 À/À mutants do not change obviously before Cu stresses but still exhibits down-regulated after Cu stresses, excluding the effects of cox17-deletion-induced genetic disorders of Cu homeostasis on Cu overload-induced HSPC proliferation impairment.
In summary, this study shows that intracellular Cu overload effects negatively on the proliferation of zebrafish HSPCs, and confirms that Cu directly interacts with SP1 and HSF1 transcriptional proteins and Cu overload induces the aggregation of the two proteins in HSPCs, which leads to the reduced SP1 and HSF1 transcriptional activities on gene FOXM1 and the down-regulation of cytoskeleton genes and the ultimately HSPC proliferation impairment. All the integrated data from this study might enrich the theoretical basis for the research of embryonic HSPC/blood cancer cell proliferation in a certain concentration range of trace element Cu, and provide some novel hints for the underlying mechanism of hematopoietic diseases in humans with unbalanced Cu homeostasis.

Limitations of the study
In this study, we show that excess Cu directly binds transcriptional factors HSF1/SP1 and induces their cytoplasmic aggregation in the cytoplasm in cells, and whether it is a general effect for excess Cu on transcriptional factors in metal signaling transduction, we need more solid data to convince it in the future days.

STAR+METHODS
Detailed methods are provided in the online version of this paper and include the following:

DECLARATION OF INTERESTS
The authors have declared that no conflict of interest exists.  iScience Article

METHOD DETAILS
The full names and their abbreviations for genes mentioned in this study are listed in Table S1.

Morpholinos and Cas9/gRNA
In this study, the CRISPR/Cas9 genome edition system was used to construct tubulin alpha 1a (tuba1a), stathmin-like 4 (stmn4), and forkhead box M1 (foxm1) zebrafish mutants. The guide RNAs (gRNAs) were designed to target the first exon of the aforementioned genes by ZiFiT Targeter Version 4.2 (http://zifit. partners.org/ZiFiT/CSquare9Nuclease.aspx). Sequences of gRNAs are listed in Table S2. The genotyping assays of tuba1a, stmn4 and foxm1 heterozygote and homozygous mutants were performed in the F1 and F2 generation from the fundamental fish with genome edition at targeted loci in germline as reported previously, 25,81 and the genotyping primers are listed in Table S3. The phenotypes of mutants at different developmental stages were observed and photographed. Tuba1a knockout homozygous mutants were embryonic lethal from 6 dpf (days post fertilization), all died aroud 8-9 dpf, and tuba1a À/À mutants were one-by-one genotyped for dead fish and in whole-mount in situ hybridization (WISH) assays in this study. The morpholinos (MOs), including tuba1a-MO, stmn4-MO, tubb5 (tubulin, beta 5)-MO, tmsb2 (thymosin beta 2)-MO, foxm1-MO, and p53-MO were purchased from Gene Tools, LLC (Philomath, Oregon, USA) and their sequences are listed in Table S4. In all experiments, the MOs were injected into one-cell stage embryos, with the MO doses of tuba1a, stmn4, tubb5, tmsb2 and foxm1 at 0.9 mM, 0.9 mM, 0.9 mM, 0.9 mM and 0.6 mM, respectively, and MO dose of p53 at 1.0 mM.
Studies shown that the off-target effects of MOs are mediated by P53 activation, 35-37 and co-injection of p53 MO is a well-known standard to estimate whether the tested morphology or molecular defects are derived from the ameliorate apoptosis induced by MO off-targets, from P53 activation, or from the down-regulation of gene expression.

ShFOXM1 cell lines construction
FOXM1 shRNA (Table S5) was synthesized to target the different loci of gene FOXM1. Lentiviruses for different shFOXM1 or shGFP controls were generated by transfecting HEK293T cells with a transducing vector as well as packaging vectors pMD.2 and pspAX. 81 After transfection for 48 h, viral particles in the medium were harvested and transduced into target cells (HUVECs). The virus infected cells were selected with puromycin (1 mg/mL) and verified by cell-directed qRT-PCR to check the FOXM1 expression. The cells infected with specific shFOXM1 lentiviruses and exhibiting significantly reduced expression of FOXM1 were expanded and maintained as stable shFOXM1-knockdown cell lines. 12 Drug exposure and phenotype observation The Cu ion (Cu) exposure solution was prepared as we performed recently. 82 Briefly, embryos were exposed to Cu 2+ (CuSO 4 $5H 2 O) (Sigma, Cat#61245) before sphere stage at 3.9 mM Cu concentration. 38 Elesclomol (Selleck, Cat#S1052) and was dissolved in Dimethyl Sulphoxide (DMSO) (Biosharp, Cat#BS087) at 10 mM for stock solutions, and tetrathiomolybdate (TTM) (Sigma, Cat#323446) were dissolved in DMSO at 20 mM for stock solutions. 100 nM elesclomol was added to the Cu treated groups, 29,83 and 200 mM TTM added to the control groups. 29 The embryos were collected at the indicated stages. Embryos from the control and the treated groups were observed and photographed using a light microscope (Leica M205FA) to examine their embryonic morphology. The embryos were collected at the indicated stages for experiments as indicated in Schema 1.

Cellular Cu ion level assays
Tg(runx1:GFP) embryos before sphere stage were stressed with 3.9 mM Cu to 33 hpf. After treatment, the embryos were homogenized, followed by cell filtration. Cells of embryos were collected with ice-cold PBS.
Next, 500 mL ice-cold PBS and 0.5 mL Cu ion probe (Rhodamine type Cu ion probe, labelling intracellular Cu ion, kindly provided by Pro. Gao, Hebei University, Baoding 071002, P. R. China) was added to each sample, followed by incubation in the dark for 2 min at 28 C. Analysis of the cellular Cu ion in Cu stressed runx1GFP + cells using FACS (CytoFLEX S, Beckman Coulter, USA).

mRNA synthesis and injection
For mRNA preparation, capped mRNAs were synthesized using the mMessage mMachine kit (Ambion, Cat#AM1344) as instructed by the manufacturer. The synthesized mRNAs were diluted into different concentrations and injected into one-cell stage embryos as reported previously, 85 and the mRNA concentrations of foxm1, tuba1a, and stmn4 both at 200 ng/mL. Full-length zebrafish foxm1, tuba1a, and stmn4 were amplified from cDNA pools by using the appropriate set of primers shown in Table S8.

Fluorescence microscopy observation
Embryos were anesthetized with a low dose of tricaine and mounted on dishes with 1% low-melting agarose for observing the flk1 + runx1 + double positive cells in different treated Tg(runx1:GFP/flk1:mcherry) embryos. Confocal images were acquired with a Leica TCS SP8 confocal laser microscope (Wetzlar, Germany). The number of HSPCs was counted based on yellow particles obtained by the overlap of the red and green fluorescence in the cells, and the fluorescence intensity in the positive cells was analyzed by Image J software.

Cell cycle analysis
The Cu stressed Tg(runx1:GFP) embryos and the control embryos were homogenized separately and filtered, followed by sorting the runx1GFP + cells by FACS (BD FacsAria SORP 650110M3 BioDot, American), and the collected cells were fixed in the pre-cooled 75% absolute ethanol for over 2 h. Then, after washing with PBS, 100 mL of RNase A was added to the cells for 30 min at 28 C. Next, the DNA was stained with 500 mL DNA Prep stain (Propidium iodide solution, PI) (Wanleibio, Cat#WLA010a) and incubated in the dark for 30 min at 28 C. Finally, the cellular DNA content was analyzed using the CytoFLEX Flow Cytometer (Beckman Coulter, USA), followed by estimating the percentage of cells in the G1 phase, S phase, and G2/M phase.
Additionally, Tg(runx1:GFP) embryos at 33 hpf and 72 hpf were fixed with 4% PFA overnight at 4 C, followed by dehydration with methanol, incubation at -20 C overnight, and then treatment with 1 mg/mL collagenase (Life-iLab Biotech, Cat#AC15L141) for 45 min at RT. After washing with PBST, the embryos were blocked in 4% BSA for 1 h, followed by incubation separately with the primary antibodies of a-Tubulin

Plasmids Construction
Briefly, genomic DNA was extracted from $ 50-60 embryos or HEK293T cells using the ammonium acetate method and quantified using a Nanodrop spectrophotometer (Thermo Fisher) for amplification of gene promoter fragments. The primers in Table S9 were used for amplification of 5 0 unidirectional human TUBA1A promoter, including À1950; human STMN4 promoter, including À1959; human FOXM1 promoter, including À2176, À1086, À755, À317 and À106; zebrafish tuba1a promoter, including À1950; zebrafish stmn4 promoter, including À2024; zebrafish foxm1 promoter, including À2113, followed by cloning them separately into the pGL3 vector. Full-length human ATOX1, FOXM1, SP1, HSF1 and zebrafish atox1, foxm1, sp1, hsf1 were amplified from cDNA pools using the appropriate sets of primers (Table S9). Human ATOX1, FOXM1, SP1, HSF1 and zebrafish atox1, foxm1, sp1, hsf1 were sub-cloned into the pCGN-HAM and pCMV-Myc vectors, respectively. All plasmids were verified by sequencing. iScience Article re-suspended in sterile water and used as a template for qPCR. The tested genes and their primers used for ChIP-qPCR are listed in Table S10, and qPCR and data analysis as described above.

Luciferase reporter assay
In this study, the reporter vectors of FOXM1, STMN4, TUBA1A, and different truncated mutants of FOXM1 were used for luciferase reporter assays as described previously. 10,11,81 All plasmids were transiently transfected into HEK293T cells using lipofectamineä 2000 (Thermo Fisher Scientific, Cat#11668-019) following the manufacture's protocol. After harvesting the transfected cells, the luciferase activity assays were performed using the Dual-Luciferase Reporter Assay System (Promega, Cat#E1910), and the luciferase assay data were analyzed using GraphPad Prism 8.0.

Statistical analysis
A sample size larger than 10 embryos (n > 10) was used for different experiments in each group, with 2-3 biological replicates for each test. Percentage analysis of the results among different groups was performed using hypergeometric distribution in the R-console software. 86 For WISH results in different experimental groups, the number in the figure was shown as N changed /N total , where N changed indicates the number of embryos exhibiting reduced or increased expression, and N total indicates the total number of embryos in a group; the number in the figure for the control groups was shown as N normal /N total , where N normal indicates the number of embryos exhibiting normal expression and N total indicates the total number of embryos in a group. The signal area of representative images was calculated in each representative embryo (N R 3) in WISH by Image J software (NIH, Bethesda, Maryland) (Firstly, the WISH's picture was converted to type 8-bit, then the area was adjusted by Image Adjust Brightness/Contrast, then Set Scale, and finally the measure) and shown as scatter plots, with each dot indicating the signal level of a representative image in an individual embryo in each group, and the data were analyzed using t-test by GraphPad Prism 8.00 software as we performed recently. 10,11 Additionally, the percentage of the WISH results was determined by hypergeometric distribution analysis using the R-console software.
The number of HSPCs and HSPC proliferation were quantified from the confocal images of the AGM (470 mm3470 mm) and CHT (470 mm3470 mm) with z-stacks spanning the entire trunk thickness, and the number of flk1 + runx1 + and anti-GFP + (runx1 + ) BrdU + cells were manually counted using ImageJ software (NIH). At least ten randomly selected units were analyzed for each control group and experimental group. The number of PH3 + HSPCs was quantified in a similar way as described above. Also, the fluorescence intensity of cytoskeleton in a single cell was also quantified from the confocal images of the positive b-Tubulin signals in runx1GFP + cell (22.15 mm322.15 mm) with z-stacks spanning the entire cell thickness, and the fluorescence intensity of cytoskeleton was measured using ImageJ. The statistical data of the signal area and fluorescence level in different samples were analyzed using t-test by GraphPad Prism 8 software, with each dot representing the signal level in an individual embryo in each group.
Statistical data of cell cycle were processed by GraphPad Prism 8 software. Each dot represents the cell percentage in each cell cycle phase. The qRT-PCR data were analyzed by one-way analysis of variance (ANOVA) and post hoc Tukey's test in the Statistic Package for Social Science (SPSS) 19.0 software, with each dot representing one repeat. The statistical analysis for luciferase reporter assay results was performed using GraphPad Prism 8 software (unpaired t-test) (GraphPad Software Inc). Data were presented as mean G SD, *P < 0.05, **P < 0.01, ***P < 0.001.