中国科技论文在线 http://www.paper.edu.cn

Toxicology Letters 147 (2004) 109–119

Inhibitory effect of tellimagrandin I on chemically induced differentiation of human leukemia K562 cells

Zongchun Yi a,b, Zhao Wang a,∗, Haixia Li a, Mingjie Liu a

a Department of Biological Sciences and Biotechnology, Tsinghua University, Beijing 100084, China b Department of Biological Engineering, Beijing University of Aeronautics and Astronautics, Beijing 100083, China Received 12 September 2003; received in revised form 12 September 2003; accepted 12 October 2003

Abstract

Tellimagrandin I is a compound widely present in plants. In this study, the effect of tellimagrandin I on chemically induced erythroid and megakaryocytic differentiation was investigated using K562 cells as differentiation model. It was found that tellimagrandin I not only inhibited the hemoglobin synthesis in butyric acid (BA)- and hemin-induced K562 cells with IC50 of 3 and 40 ␮M, respectively, but also inhibited other erythroid differentiation marker including acetylcholinesterase (AChE) and glycophorin A (GPA) in BA-induced K562 cells. Tellimagrandin I also inhibited 12-O-tetradecanoylphorbol-13- acetate (TPA)-induced expression of CD61 protein, a megakaryocytic marker. RT-PCR analysis showed that tellimagrandin I decreased the expression of erythroid genes (␥-globin and porphobilinogen deaminase (PBGD)) and related transcription factors (GATA-1 and NF-E2) in BA-induced K562 cells, whereas tellimagrandin I induced the overexpresison of GATA-2 transcription factor that played negative regulation on erythroid differentiation. These results indicated that tellimagrandin I had inhibitory effects on erythroid and megakaryocytic differentiation, which suggested that tannins like tellimagrandin I might influence the anti-tumor efficiency of some drugs and the hematopoiesis processes. © 2004 Elsevier Ireland Ltd. All rights reserved.

Keywords: Erythroid differentiation; GATA-1; GATA-2; ␥-Globin; Megakaryocytic differentiation; NF-E2; Porphobilinogen deaminase; Tellimagrandin I

1. Introduction Tannins have been showed to possess a variety of pharmacological activities such as antiviral, antimi- Tannins are present in varied plants utilized as crobial, antioxidant, antimutagenic, and anti-tumor foods and medicinal herbs (Chung et al., 1998a). In activities (Chung et al., 1998b). The adverse effects addition, tannins have been used as food additives. of tannins, including hepatotoxic, antinutritional, and It was estimated that people in the US ingested each carcinogenic activities, have been paid attention. In day 1 g of tannic acid (a tannin) (Sanyal et al., 1997). addition, several toxicity studies had been performed to evaluate the health safety of and propyl ∗ gallate. As early as 1948, Orten and his colleagues Corresponding author. Tel.: +86-10-62772241; fax: +86-10-62772240. observed reduced food intake, growth inhibition and E-mail address: [email protected] (Z. Wang). at dietary dose levels of 11,700 mg of propyl gallate

0378-4274/$ – see front matter © 2004 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.toxlet.2003.12.008 转载 中国科技论文在线 http://www.paper.edu.cn 110 Z. Yi et al. / Toxicology Letters 147 (2004) 109–119

(Orten et al., 1948). In another study, propyl gallate and doxorubicin (Jeannesson et al., 1997), while at the dose of 10,000 mg/kg feed produced severe 12-O-tetradecanoylphorbol-13-acetate (TPA) induced anemia and growth retardation (Heijden et al., 1986). K562 cells to differentiate towards the megakary- Recently, a subchronic toxicity study of gallic acid ocytic lineage (Villeval et al., 1983). Out of these on F344 rats showed that toxic effects following ad- inducers, anthracyclines, Ara-C, and BA are widely ministration of 0.6% or more in males and 5% in used in cancer therapy through their inducing differ- females included reduction of hemoglobin concentra- entiation activities, suggesting that the inhibition of tion, hematocrit, and red blood cell counts, suggest- their inducing differentiation activities would result in ing the development of anemia (Niho et al., 2001). the loss of anti-tumor function. In this study, the effect In a subacute study, oral of gallic acid at a dose of of tellimagrandin I on chemically induced erythroid 1000 mg/kg body weight (BW) for 28 days reduced and megakaryocytic differentiation was investigated hemoglobin level in mice (Rajalakshmi et al., 2001). in K562 cells. During erythroid differentiation, tel- All these studies demonstrated that gallic acid could limagrandin I not only inhibited the hemoglobin syn- induce the development of anemia. It was proposed thesis in BA- and hemin-treated K562 cells, but also that this hydrolysate (gallic acid) of hydrolysable tan- inhibited other erythroid differentiation marker in- nins and hydrolysable tannins per se could interfere cluding acetylcholinesterase (AChE) and glycophorin erythropoiesis processes. A (GPA) in BA-induced K562 cells. When K562 cells Tellimagrandin I is a hydrolysable tannin com- were simultaneously treated with TPA and tellima- pound widely present in plants such as Punica grana- grandin I, the TPA-induced expression of megakary- tum, Myrtaceae, and Elaeagnaceae, which possesses ocytic surface marker CD61 was also inhibited by various biological activities, such as inhibitory ef- tellimagrandin I. These results showed that tellima- fect against carbonic anhydrase (Satomi et al., 1993), grandin I inhibited K562 cell differentiation, which antibacterial activity against Helicobacter pylori suggested that hydrolysable tannins such as tellima- (Yoshida et al., 2000), inhibitory activity on the syn- grandin I might influence the efficiency of some cytia formation (Kim et al., 2001), toxicities towards anti-tumor agents and the hematopoiesis processes. the nematode and the brine shrimp (Yamasaki et al., 2002), restoration of effectiveness of beta-lactams and tetracycline on methicillin-resistant Staphylococ- 2. Materials and methods cus aureus (Shiota et al., 2000), anti-tumor activities against sarcoma-180 in mice (Miyamoto et al., 1993b) 2.1. Materials and induction of IL-1␤ production from human peripheral macrophages (Miyamoto et al., 1993a). Tellimagrandin I (purity > 95%) was kindly pro- Nevertheless, the effect of tellimagrandin I on hema- vided by Prof. Yanze Liu at Henan College of Tradi- tological differentiation and hematopoiesis processes tional Chinese Medicine, Zhengzhou, China. has not been characterized. It has been extensively demonstrated that K562 2.2. Cell culture cells can be induced to differentiate towards erythroid and megakaryocytic lineages by various differenti- K562 cells were grown in RPMI 1640 medium ation inducers (Sutherland et al., 1986). K562 cells (GIBCO) supplemented with 10% (v/v) fetal bovine have been used as a model to screen anti-tumor serum (HyClone), 100 units/ml penicillin, and drugs inducing differentiation and to research hema- 100 ␮g/ml streptomycin (Sigma) in 5% (v/v) CO2 tological cell differentiation (Koeffler and Golde, humidified atmosphere at 37 ◦C. For the experiment, 1980). Erythroid differentiation of K562 cells could exponentially growing K562 cells were collected and be achieved by exposure to several pharmacolog- re-suspended in fresh culture medium. After 24 h, ical agents, including hemin (Dean et al., 1981), the cells was added hemin (Sigma) at a final con- arabinofuranosyl cytosine (Ara-C) (Watanabe et al., centration of 40 ␮M, BA (Sigma) at 0.5 mM, TPA 1985), butyric acid (BA) (Lozzio et al., 1979; Chénais (Sigma) at 50 ng/ml, and tellimagrandin I at 10 ␮M et al., 1997), and anthracyclines such as aclarubicin, or at indicated concentrations. 中国科技论文在线 http://www.paper.edu.cn Z. Yi et al. / Toxicology Letters 147 (2004) 109–119 111

2.3. Benzidine staining and determination (Biometra). The following specific primers sets were used for PCR. ␥-Globin: sense strand primer, The percentage of cells straining for hemoglobin 5-ACAAGCCTGTGGGGCAA-3, antisense strand   was estimated by staining with benzidine/H2O2 es- primer, 5 -GCCATGTGCCTTGACTTT-3 ; porpho- sentially as previously described (Ngo-Nyoung et al., bilinogen deaminase (PBGD): sense strand primer, 1994). The hemoglobin-positive cells were stained 5-GGTCCTACTATCGCCTCCCTC-3, antisense str- with blue by benzidine. The benzidine-positive cells and primer, 5-CCAGCCTCTGTCCCCTCCAGC-3; were counted in a hemacytometer on a microscope GATA-1: sense strand primer, 5-CAGTCTTTCAGG- from 500 cells for each sample. Then the percentage TGTACCC-3, antisense strand primer, 5-GAGTG- of benzidine-positive cells was calculated. ATGAAGGCAGTGCAG-3; NF-E2: sense strand primer, 5-ATTTGAGCCCCAAGCCCCAGC-3, anti- 2.4. AChE activity assays sense strand primer, 5-CCAGCCTCTGTCCCCTCC- AGC-3; GATA-2: sense strand primer, 5-ATCAAGC- The method used to determine AChE activity was CCAAGCGAAGACTG-3, antisense strand primer, modified from that of Ellman et al.’s (1961). Briefly, 5-ACATTGTGCAGCTTGTAGTAGAGGC-3; ␤- the enzyme reaction was undergone at 37 ◦Cina actin: sense strand primer, 5-TGGACTTCGAGCAA- 1-ml reaction system containing 0.5×106 K562 cells, GAGATGG-3, antisense strand primer, 5-ATCTCCT- 0.1 M potassium phosphate buffer (pH 8.0), 0.6 mM TCTGCATCCTGTCG-3. The amplification re- 5,5-dithiobis-(2-nitrobenzoic acid) and 0.75 mM actions were initiated by a denaturation step for acetylthiocholine iodide substrate. After 20 min, each 5 min at 95 ◦C and then subjected to 30 cycles sample was centrifuged at 4 ◦C. The absorbance of of 95 ◦C for 1 min, 60 ◦C for 45 s, and 72 ◦C for the supernatant was measured spectrophotometrically 45 s. ␤-Actin was used as a control. PCR prod- at 405 nm. ucts were analyzed on 2% agarose gel, stained with ethidium bromide and photographed with Image- 2.5. Flow cytometric analysis Master Video Documentation System (Pharmacia Biotech). The expression of erythroid and megakaryocytic antigens was determined by direct immunofluores- 2.7. Data analysis cence using the following conjugated antibodies: flu- orescein isothiocyanate (FITC)-conjugated anti-CD61 All data were presented as mean ± S.D. Student’s antibodies (Becton–Dickinson), and phycoerythrin t-test was used to determine the statistical signifi- (PE)-conjugated anti-GPA antibodies (Caltag). Mouse cance. P<0.05 were considered statistically signi- isotype IgG1-FITC (Becton–Dickinson) and IgG1-PE ficant. (Caltag) antibodies served as controls, respectively. Flow cytometric analysis was performed using the FACScan instruments (Coulter, Epics Elite, USA). 3. Results

2.6. Isolation of total RNA and RT-PCR 3.1. Inhibition of hemoglobin synthesis and AChE activity in K562 cell by tellimagrandin I Total RNA Minipreps Classic Kit (Sangon, Shang- hai, China) was used to extract total RNA from cul- Firstly, we determined induced hemoglobin syn- tured cells. The RNA quantity was calculated from thesis in K562 cells using benzidine staining. K562 the absorbance at 260 nm. First strand cDNA was syn- cells were treated with 40 ␮M of hemin or 0.5 mM of thesized from 25 ␮g total RNA by M-MuLV reverse BA combining with different concentration of tellima- transcriptase (Sangon, Shanghai, China) in a 100-␮l grandin I for 72 h. In hemin-induced cells, the treat- reaction system. ment with tellimagrandin I decreased the percentage PCR was carried out using 10 ␮l of cDNA of benzidine-positive cells in concentration-dependant ina50␮l reaction with UNOII Thermocycler manner with approximately IC50 of 40 ␮M(Fig. 1A). 中国科技论文在线 http://www.paper.edu.cn 112 Z. Yi et al. / Toxicology Letters 147 (2004) 109–119

(A) 4 - Tellimagrandin I 80 + Tellimagrandin I cells) 6

3 60

** ** 40 2

20 1

0 Inhibition of benzidine-positive cells (%) cells of benzidine-positive Inhibition 0 AchE activity (nmol/min/10 AchE 0 10 20 30 40 50 Control BA Tellimagrandin I (µM) Fig. 2. The effect of tellimagrandin I on AChE activity in K562 cells. The cells were treated with 0.5 mM BA and 10 ␮M tel- (B) limagrandin I for 72 h. Data represent the mean ± S.D. from five 100 independent experiments.

80 showed that tellimagrandin I inhibited the AChE 60 activity in K562 cells either treated or not treated with BA. 40 3.2. Inhibition of expression of erythroid and 20 megakaryocytic surface markers by tellimagrandin I

Inhibition of benzidine-positive cells (%) 0 We further observed whether tellimagrandin I could 0246810 change the expression of erythroid surface marker Tellimagrandin I (µM) GPA on the surfaces of BA-induced K562 cells. After ␮ Fig. 1. Inhibition of tellimagrandin I on K562 cell hemoglobin K562 cells were treated with 0.5 mM of BA and 10 M synthesis induced by hemin (A) and BA (B). The K562 cells were of tellimagrandin I for 72 h, expression of GPA protein simultaneously treated with tellimagrandin I at different concentra- was analyzed by flow cytometry using PE-conjugated tion and differentiation inducer (40 ␮M hemin or 0.5 mM BA) for anti-GPA antibodies. As shown in Fig. 3, tellima- 72 h. After treatment, the cells were stained by benzidine/H2O2, grandin I inhibited GPA expression in the K562 cells, and benzidine-positive cells were counted out from 500 cells per sample. The percentage of benzidine-positive cell in control K562 whereas tellimagrandin I reduced the GPA-positive cells varied from 2% to 8%. Data represent the means from three percentage of BA-treated cells obviously. independent experiments. In addition, effect of tellimagrandin I on TPA- induced megakaryocytic differentiation of K562 cells was investigated using CD61 as megakaryocytic Compared with hemin-induced cell, BA-treated cells marker (Fig. 4). When K562 cells were only treated were more sensitive. It was only at the concentration with 10 ␮M of tellimagrandin I for 72 h, the expression of 3 ␮M that tellimagrandin I reduced the differen- CD61 remained at the control level. After the treat- tiation of BA-treated cells by 50%. Furthermore, an ment of 50 ng/ml of TPA for 72 h, the CD61-positive almost complete inhibition of benzidine-positive cells percentage markedly increased. However, when the was obtained in BA-treated cells in presence of 10 ␮M cells were simultaneously treated with TPA and tel- of tellimagrandin I (Fig. 1B). limagrandin I, the TPA-induced CD61 expression We also evaluated the effect of tellimagrandin I on was inhibited. These results suggested tellimagrandin AChE activity of K562 cells treated with 0.5 mM of I inhibited the TPA-induced megakaryocytic differen- BA and 10 ␮M of tellimagrandin I for 72 h. Fig. 2 tiation of K562 cells. 中国科技论文在线 http://www.paper.edu.cn Z. Yi et al. / Toxicology Letters 147 (2004) 109–119 113

Fig. 3. Expression of erythroid surface markers GPA protein in control K562 cells (A) and the cells after treatment with 10 ␮M tellimagrandin I (B), 0.5 mM BA (C), or 0.5 mM BA in combination with 10 ␮M tellimagrandin I (D) for 72 h. Data from a typical experiment are representative of two. 中国科技论文在线 http://www.paper.edu.cn 114 Z. Yi et al. / Toxicology Letters 147 (2004) 109–119

Fig. 4. Expression of megakaryocytic surface marker CD61 protein in control K562 cells (A) and the cells after incubation with 10 ␮M tellimagrandin I (B), 50 ng/ml TPA (C), or 50 ng/ml TPA in combination with 10 ␮M tellimagrandin I (D) for 72 h. Data from a typical experiment are representative of two. 中国科技论文在线 http://www.paper.edu.cn Z. Yi et al. / Toxicology Letters 147 (2004) 109–119 115

3.3. Effects of tellimagrandin I on erythroid and transcriptional level in response to BA. Therefore, megakaryocytic mRNA expression we investigated the influence of tellimagrandin I on the transcriptional activation of ␥-globin and PBGD In order to define the effect of tellimagrandin I genes in BA-induced cells. Fig. 5A and B shows that on cell differentiation at molecular level, we further the mRNA levels of both genes were up-regulated in determined the mRNA expression of erythroid and k562 cells induced by 0.5 mM of BA for 72 h as ex- megakaryocytic related genes in the K562 cells which pected. However, the mRNA levels for ␥-globin and were treated with tellimagrandin I and differentiation PBGD in BA-treated cells were obviously reduced in inducers. presence of 10 ␮M tellimagrandin I. Moreover, basal The transcriptional activation of the ␥-globin gene, mRNA levels of these two genes were also reduced in which encoding one of hemoglobin subunits, has been K562 cells only treated with tellimagrandin I. The in- demonstrated in BA-induced K562 cells (Chénais duced expression of GATA-1 and NF-E2 gene, which et al., 1997). The heme synthesis pathway enzymes are two transcription factors playing critical roles in PBGD was also shown to be up-regulated at the erythroid differentiation, has been demonstrated to

Fig. 5. Effects of tellimagrandin I on mRNA expression of erythroid and megakaryocytic specific genes in K562 cell. After the K562 cells were treated with 10 ␮M tellimagrandin I in combination with 0.5 mM BA or 50 ng/ml ATP for 72 h, the total RNA was extracted form the cells, and RT-PCR was used to analyze mRNA expression level of genes. ␤-Actin serves as a control. Data from a typical experiment are representative of two. 中国科技论文在线 http://www.paper.edu.cn 116 Z. Yi et al. / Toxicology Letters 147 (2004) 109–119 involve in BA-induced erythroid differentiation of common feature of hydrolysable tannins, which could K562 cells (Chénais et al., 1997; Chénais, 1998). In be due to the gallic acid unit. Tannins are present percent study, the basal mRNA level of NF-E2 in in a variety of plants utilized as food including food K562 cells was markedly reduced in the presence of grains (such as sorghum, millet, barley, beans, and 10 ␮M tellimagrandin I, while the GATA-1 mRNA peas), and un-ripe fruits (such as apples, bananas, level slightly declined in these cells. When tellima- dates, grapes, plums, and strawberries), although they grandin I was present, the BA-induced expression might not be and tellimagrandin I of both GATA-1 and NF-E2 genes was distinctly (Chung et al., 1998a). In addition, there are various reduced. Chinese medicine herbs containing tannins such as GATA-2 transcription factor is required for pomegranates and eucalyptus leaf. Therefore, when erythroid differentiation at early step, but its over- we consume these tannin-rich food grains and fruits, expression inhibited erythroid differentiation and or use these medicine herbs, we should consider the promotes megakaryocytic differentiation (Ikonomi potential hazard of tannins on erythropoiesis pro- et al., 2000a,b). To examine whether GATA-2 was in- cesses. Furthermore, there are many anticancer drugs volved in the inhibitory effects of tellimagrandin I on that exert effects on cancer cell through inducing dif- BA-induced erythroid and TPA-induced megakary- ferentiation. For example, arabinofuranosyl cytosine, ocytic differentiation, we also analyzed GATA-2 anthracyclines, BA and doxorubicin can induce ery- mRNA levels. As shown in Fig. 5A and C, tellima- throid differentiation of K562 cell. Tellimagrandin grandin I induced the overexpression of GATA-2 in I is likely to disturb the anticancer effects of these K562 cells without treatment of any inducers and agents. Therefore, when a lot of tannin-rich food was BA-induced K562 cells, while BA only induced the ingested, the adverse effect on erythropoiesis should moderate up-regulation of this gene transcription. been monitored, and hemoglobin levels and AChE TPA induced the mRNA expression of GATA-2 gene activity of blood cells could be used as monitoring while the presence of tellimagrandin I promoted the biomarkers. inducing effect of TPA. The transcriptional activation of the ␥-globin gene has been demonstrated in BA-induced K562 cells (Chénais et al., 1997). In addition, heme biosynthetic 4. Discussion enzyme genes such as PBGD (Chénais et al., 1997), ␦-aminolevulinate dehydratase (Chang and Sassa, In the present study, we demonstrated that tellima- 1995), and ␦-aminolevulinate synthase (Kawasaki grandin I inhibited BA-induced erythroid differen- et al., 1996) have been also shown to be up-regulated tiation of K562 cells, as shown with the decrease at the transcriptional level in response to BA. The of hemoglobin synthesis, GPA protein expression down-regulation of mRNA expression for ␥-globin and AChE activity. In addition, tellimagrandin I in- and PBGD genes induced by tellimagrandin I in hibited hemin-induced erythroid differentiation and BA-induced K562 cells indicated that tellimagrandin I TPA-induced megakaryocytic differentiation in K562 inhibited heme and globin synthesis at transcriptional cells. level, leading to inhibition of hemoglobin synthesis. We previously also found that another tannin The NF-E2 transcription factor plays an important compound chebulinic acid has similar inhibitory ef- role in transcriptional activation of globin genes and fect on erythroid differentiation of K562 cells. Both heme biosynthetic enzyme genes such as PBGD and chebulinic acid and tellimagrandin I belong to hy- ferrochelatase (Jarman et al., 1991; Taketani et al., drolysable tannins composed of and gallic 1992; Mignotte et al., 1989; Gong et al., 1996). acid in chemical structure. The administration of gal- Numerous GATA protein binding sites have been lic acid or propyl gallate has been observed to result observed in the promoters, enhancers, and locus con- in the decrease of blood hemoglobin levels in vivo trol regions of globin genes and most other genes (Orten et al., 1948; Heijden et al., 1986; Niho et al., expressed in erythroid cells (Ikonomi et al., 2000a). 2001; Rajalakshmi et al., 2001). It was presumed that GATA-1 transcription factor is essential for sur- inhibition of erythroid differentiation could be the vival and terminal maturation of erythroid precursors 中国科技论文在线 http://www.paper.edu.cn Z. Yi et al. / Toxicology Letters 147 (2004) 109–119 117

(Weiss et al., 1997; Shivdasani and Orkin, 1996; and expression of GATA-1 and NF-E2 in K562 cells, Orkin and Weiss, 1999). Previous evidences have and H2O2 induced erythroid differentiation in K562 been shown that GATA-1 and NF-E2 are required in cells, suggesting that reactive oxygen species (ROS) BA-induced erythroid differentiation process (Chénais production was involved in BA-induced erythroid et al., 1997; Chénais, 1998). The transcriptional ac- differentiation process. ROS seem to act early via the tivation of GATA-1 and NF-E2 was also inhibited by specific transcription factors GATA-1 and NF-E2 as tellimagrandin I, which suggested that the decrease of indicated by the inhibition observed with an antioxi- ␥-globin and PBGD mRNA level induced by tellima- dant or via other factors able to induce overexpression grandin I was due to the expression down-regulation of ␥-globin and PBGD genes (Chénais et al., 2000). of these transcription factors. It was suggested that the antioxidant activity was GATA-1 and GATA-2 transcription factors are both the initial reason for the inhibitory effect of tellima- expressed in erythroid cells yet with distinct patterns grandin I on BA-induced differentiation. ROS are of expression. During erythropoiesis, GATA proteins implicated in numerous processes such as carcino- switch from GATA-2 in early proliferating progeni- genesis or inflammation. These compounds are also tors to GATA-1 in terminally differentiating erythroid produced at physiological levels and are involved in cells (Pevny et al., 1991; Leonard et al., 1993; Cheng normal processes such as cell division and probably et al., 1996). GATA-2 can determine in part differen- differentiation. In fact, an hypothesis has been pro- tiation along the erythroid/megakaryocyte pathway, posed, assuming that part of the OH free radicals and the overexpression of GATA-2 would inhibit derived from H2O2 by iron-induced heterolysis may the erythroid differentiation (Persons et al., 1999; play a physiologically important role in cell growth Ikonomi et al., 2000a,b). Here we found that tel- and maturation (Zs-Nagy, 1992; Sauer et al., 2001). limagrandin I induced the overexpression of GATA-2 It has been fully demonstrated that iron and hypoxia, gene in K562 cells. Consequently, GATA-2 occupied which are critical roles in normal erythropoiesis, can the GATA binding sites in erythroid genes regulated induce the generation of ROS (Britton et al., 2002; by GATA-1, resulting in the block of expression of Paddenberg et al., 2003), suggesting that oxidative erythroid genes and the inhibition of erythroid differ- stress at physiological level involved in normal ery- entiation. It has been showed that GATA-1 expression thropoiesis, and tellimagrandin I could inhibit the in GATA-1-null cells induces erythroid differentia- normal erythropoiesis by its antioxidant activity. tion and represses GATA-2 (Crispino et al., 1999). Hemin was regarded as a nonoxidative stress-related This GATA-1-dependent transcriptional repression of and reversible inducer (Dean et al., 1981). Chénais GATA-2 is via disruption of positive autoregulation et al. (2000) found that among the antioxidants used, and domain-wide chromatin remodeling (Grass et al., PDTC and quercetin have no effect on hemin-induced 2003). Thus, the GATA-1-dependent transcriptional erythroid differentiation, and only NAC at a high dose repression of GATA-2 transcription is rescued by inhibited hemin-induced differentiation, indicating the down-regulation of GATA-1 expression induced that the mechanism of hemin is not linked to a rad- by tellimagrandin I, leading to overexpression of ical process. Hemin-induced differentiation did not GATA-2. involve in the up-regulation of GATA-1 and NF-E2 Tellimagrandin I belongs to polyphenol compounds transcription factors (Morceau et al., 1996). These with strong antioxidant activity (Bravo, 1998). Re- observations could interpret why the sensitivity of cently, BA-induced oxidative stress was evidenced by hemin-induced K562 cells to tellimagrandin I (IC50 changes in GSSG and GSH levels, the decrease in of 40 ␮M) was lower than that of BA-induced cells cellular antioxidant enzyme activities and the oxida- (IC50 of 3 ␮M). tion of the dihydroethidium probe at the early step of We have also observed the inhibitory effect of erythroid differentiation in K562 cells (Chénais et al., tellimagrandin I on TPA-induced megakaryocytic dif- 2000). The addition of chemically unrelated antioxi- ferentiation in K562 cells. Tellimagrandin I blocked dants, including N-acetylcysteine (NAC), pyrrolidine the TPA-induced expression of megakaryocytic dithiocarbamate (PDTC), and quercetin, considerably marker CD61 protein. This inhibitory effect seem decreased the BA-induced erythroid differentiation to be specific for tellimagrandin I, because another 中国科技论文在线 http://www.paper.edu.cn 118 Z. Yi et al. / Toxicology Letters 147 (2004) 109–119 hydrolysable tannin compound chebulinic acid, with bilinogen deaminase and NF-E2 mRNA levels. Leukemia 11, similar inhibitory effect on erythroid differentiation, 1575–1579. did not change TPA-induced megakaryocytic differ- Chénais, B., Andriollo, M., Guiraud, P., Belhoussine, R., Jeannesson, P., 2000. Oxidative stress involvement in chemical- entiation in K562 cells. We further observed that tel- ly induced differentiation of k562 cells. Free Radic. Biol. Med. limagrandin I induced overexpression of GATA-2 in 28, 18–27. K562 cells, whereas CD61 protein expression did not Chung, K.T., Wong, T.Y., Wei, C.I., Huang, Y.W., Lin, Y., 1998a. up-regulate. Furthermore, tellimagrandin I promoted Tannins and human health: a review. Crit. Rev. Food Sci. Nutr. the up-regulation of GATA-2 mRNA level induced by 38, 421–464. Chung, K.T., Wei, C.I., Johnson, M.G., 1998b. Are tannins a TPA. Therefore, the inhibitory effect of tellimagrandin double-edged sword in biology and health. Trends Food Sci. I on megakaryocytic differentiation was not due to Technol. 9, 168–175. GATA-2 gene expression change, although GATA-2 Crispino, J.D., Lodish, M.B., MacKay, J.P., Orkin, S.H., 1999. Use plays critical role in megakaryocytic differentiation. of altered specificity mutants to probe a specific protein-protein In conclusion, we have demonstrated that tel- interaction in differentiation: the GATA-1: FOG complex. Mol. Cell 3, 219–228. limagrandin I inhibited both BA-induced erythroid Dean, A., Erard, F., Schneider, A.B., Schechter, A.N., 1981. differentiation and TPA-induced megakaryocytic dif- Induction of hemoglobin accumulation in human K562 cells ferentiation in K562 cells. The further studies are to by hemin is reversible. Science 212, 459–461. be preformed to establish whether the inhibitory ef- Ellman, G.L., Courtney, K.D., Andres, V., Featherstone, R.M., fects can disturb the efficiency of anticancer drugs and 1961. A new and rapid colorimetric determination of acetylcho- affect the hematopoiesis in vivo, and whether other linesterase activity. Biochem. Pharmacol. 7, 88–95. Gong, Q.H., McDowell, J.C., Dean, A., 1996. Essential role hydrolysable tannins possess the similar inhibitory of NF-E2 in remodeling of chromatin structure and trans- effect. criptional activation of the epsilon-globin gene in vivo by 5 hypersensitive site 2 of the beta-globin locus control region. Mol. Cell. Biol. 16, 6055–6064. Acknowledgements Grass, J.A., Boyer, M.E., Pal, S., Wu, J., Weiss, M.J., Bresnick, E.H., 2003. GATA-1-dependent transcriptional repression of This work was partly supported by the China GATA-2 via disruption of positive autoregulation and domain- wide chromatin remodeling. Proc. Natl. Acad. Sci. U.S.A. 100, Postdoctoral Science Foundation, the Basic Research 8811–8816. Foundation of Tsinghua University (JC2002046), and Heijden, C.A.V.D., Janssen, P.J., Strik, J.J., 1986. Toxicology of the National Natural Science Foundation of China gallates: a review and evaluation. Food Chem. Toxicol. 24, (project no. 30271593). 1067–1070. Ikonomi, P., Noguchi, C.T., Miller, W., Kassahun, H., Hardison, R., Schechter, A.N., 2000a. Levels of GATA-1/GATA-2 trans- cription factors modulate expression of embryonic and fetal References hemoglobins. Gene 261, 277–287. Ikonomi, P., Rivera, C.E., Riordan, M., Washington, G., Schechter, Bravo, L., 1998. Polyphenols: chemistry, dietary sources, meta- A.N., Noguchi, C.T., 2000b. Overexpression of GATA-2 bolism, and nutritional significance. Nutr. Rev. 56, 317–333. inhibits erythroid and promotes megakaryocyte differentiation. Britton, R.S., Leicester, K.L., Bacon, B.R., 2002. Iron toxicity and Exp. Hematol. 28, 1423–1431. chelation therapy. Int. J. Hematol. 76, 219–228. Jarman, A.P., Wood, W.G., Sharpe, J.A., Gourdon, G., Higgs, H., Chang, C.S., Sassa, S., 1995. ␦-Aminolevulinate dehydratase Higgs, D.R., 1991. Characterization of the major regulatory in human erythroleukemia cells: an immunologically distinct element upstream of the human ␣-globin genecluster. Mol. enzyme. Blood 65, 939–944. Cell. Biol. 11, 4679–4689. Cheng, T., Shen, H., Giokas, D., Gere, J., Tenen, D.G., Scadden, Jeannesson, P., Lahlil, R., Chénais, B., Devy, L., Gillet, R., D.T., 1996. Temporal mapping of gene expression levels during Aries, A., Morceau, F., Trentesaux, C., 1997. Anthracyclines the differentiation of individual primary hematopoietic cells. as tumor cell differentiating agents: effects on the regulation Proc. Natl. Acad. Sci. U.S.A. 93, 13158–13163. of erythroid gene expression. Leuk. Lymphoma 26, 575– Chénais, B., 1998. Requirement of GATA-1 and p45 NF-E2 587. expression in butyric acid-induced erythroid differentiation. Kawasaki, N., Morimoto, K., Tanimoto, T., Hayakawa, T., 1996. Biochem. Biophys. Res. Commun. 253, 883–886. Control of hemoglobin synthesis in erythroid differentiating Chénais, B., Molle, I., Trentesaux, C., Jeannesson, P., 1997. K562 cells. Arch. Biochem. Biophys. 328, 289–294. Timecourse of butyric acid-induced differentiation in human Kim, H.J., Lee, J.S., Woo, E.R., Kim, M.K., Yang, B.S., Yu, K562 leukemic cell line: rapid increase in ␥-globin, porpho- Y.G., Park, H., Lee, Y.S., 2001. Isolation of virus-cell fusion 中国科技论文在线 http://www.paper.edu.cn Z. Yi et al. / Toxicology Letters 147 (2004) 109–119 119

inhibitory components from Eugenia caryophyllata. Planta mutation in the gene for transcription factor GATA-1. Nature Med. 67, 277–279. 349, 257–260. Koeffler, H.P., Golde, D.W., 1980. Human myeloid leukemia cell Rajalakshmi, K., Devaraj, H., Devaraj, S.N., 2001. Assessment of lines: a review. Blood 56, 344–350. the no-observed-adverse-effect level (NOAEL) of gallic acid Leonard, M., Brice, M., Engel, J.D., Papayannopoulou, T., 1993. in mice. Food Chem. Toxicol. 39, 919–922. Dynamics of GATA transcription factor expression during Sanyal, R., Darroudi, F., Parzefall, W., Nagao, M., Knasmuller, S., erythroid differentiation. Blood 82, 1071–1079. 1997. Inhibition of the genotoxic effects of heterocyclic amines Lozzio, C.B., Lozzio, B.B., Machado, E.A., Fuhr, J.E., Lair, S.V., in human derived hepatoma cells by dietary bioantimutagens. Bamberger, E.G., 1979. Effects of sodium butyrate on human Mutagenesis 12, 297–303. chronic myelogenous leukaemia cell line K562. Nature 281, Satomi, H., Umemura, K., Ueno, A., Hatano, T., Okuda, T., Noro, 709–710. T., 1993. Carbonic anhydrase inhibitors from the pericarps of Mignotte, V., Wall, L., deBoer, E., Grosveld, F., Romeo, P.H., Punica granatum L. Biol. Pharm. Bull. 16, 787–790. 1989. Two tissue-specific factors bind the erythroid promoter Sauer, H., Wartenberg, M., Hescheler, J., 2001. Reactive oxygen of the human porphobilinogen deaminase gene. Nucleic Acids species as intracellular messengers during cell growth and Res. 17, 37–54. differentiation. Cell Physiol. Biochem. 11, 173–186. Miyamoto, K., Murayama, T., Nomura, M., Hatano, T., Yoshida, Shiota, S., Shimizu, M., Mizusima, T., Ito, H., Hatano, T., T., Furukawa, T., Koshiura, R., Okuda, T., 1993a. Anti-tumor Yoshida, T., Tsuchiya, T., 2000. Restoration of effectiveness activity and interleukin-1 induction by tannins. Anticancer Res. of beta-lactams on methicillin-resistant Staphylococcus aureus 13, 37–42. by tellimagrandin I from rose red. FEMS Microbiol. Lett. 185, Miyamoto, K., Nomura, M., Murayama, T., Furukawa, T., Hatano, 135–138. T., Yoshida, T., Koshiura, R., Okuda, T., 1993b. Anti-tumor Shivdasani, R.A., Orkin, S.H., 1996. The transcriptional control activities of against sarcoma-180 in mice. Biol. of hematopoiesis. Blood 87, 4025–4039. Pharm. Bull. 16, 379–387. Sutherland, J.A., Turner, A.R., Mannoni, P., McGann, L.E., Turc, Morceau, F., Aries, A., Lahlil, R., Devy, L., Jardillier, J.C., J.M., 1986. Differentiation of K562 leukemia cells along Jeannesson, P., Trentesaux, C., 1996. Evidence for distinct erythroid, macrophage and megalaryocyte linerages. J. Biol. regulation processes in the aclacinomycin and doxorubicin- Resp. Mod. 5, 250–262. mediated differentiation of human erythroleukemic cells. Taketani, S., Inazawa, J., Nakahashi, Y., Abe, T., Tokunaga, R., Biochem. Pharmacol. 51, 839–845. 1992. Structure of the human ferrochelatase gene. Exon/intron Ngo-Nyoung, M., Trentesaux, C., Aries, A., Carpentier, Y., gene organization and location of the gene to chromosome 18. Jardillier, J.C., Gorisse, M.C., Jeannesson, P., 1994. Effect of Eur. J. Biochem. 205, 217–222. aclacinomycin-doxorubicin association on differentiation and Villeval, J.L., Pelicci, P.G., Tabilio, A., Titeux, M., Henri, A., growth of human erythroleukemic K562 cells. Anticancer Res. Houesche, F., Thomopoulos, P., Vainchenker, W., Garbaz, 14, 1203–1208. M., Rochant, H., Breton-Gorius, J., Edwards, P.A., Testa, U., Niho, N., Shibutani, M., Tamura, T., Toyoda, K., Uneyama, C., 1983. Erythroid properties of K562 cells: effect of hemin, Takahashi, N., Hirose, M., 2001. Subchronic toxicity study of butyrate and TPA induction. Exp. Cell Res. 146, 428– gallic acid by oral administration in F344 rats. Food Chem. 438. Toxicol. 39, 1063–1070. Watanabe, T., Mitchell, T., Sariban, E., Sabbath, K., Griffin, Orkin, S.H., Weiss, M.J., 1999. Cutting red cell production. Nature J., Kufe, D., 1985. Effects of 1-␤-D-arabinofuranosylcytosine 401, 433–436. and phorbol ester on differentiation of human K562 Orten, J.M., Kuyper, A.C., Smith, A.H., 1948. Studies on the erythroleukemia cells. Mol. Pharmacol. 27, 683–688. toxicity of propyl gallate and of antioxidant mixtures containing Weiss, M.J., Yu, C., Orkin, S.H., 1997. Erythroid-cell-specific propyl gallate. Food Technology, Champaign 2, 308. properties of transcription factor GATA-1 revealed by Paddenberg, R., Goldenberg, A., Faulhammer, P., Braun-Dullaeus, phenotypic rescue of a genetargeted cell line. Mol. Cell. Biol. R.C., Kummer, W., 2003. Mitochondrial complex II is essential 17, 1642–1651. for hypoxia-induced ROS generation and vasoconstriction in Yamasaki, T., Sato, M., Mori, T., Mohamed, A.S., Fujii, K., the pulmonary vasculature. Adv. Exp. Med. Biol. 536, 163– Tsukioka, J., 2002. Toxicity of tannins towards the free-living 169. nematode Caenorhabditis elegans and the brine shrimp Artemia Persons, D.A., Allay, J.A., Allay, E.R., Ashmun, R.A., Orlic, D., salina. J. Nat. Toxins 11, 165–171. Jane, S.M., Cunningham, J.M., Nienhuis, A.W., 1999. Enforced Yoshida, T., Hatano, T., Ito, H., 2000. Chemistry and function of expression of the GATA-2 transcription factor blocks normal vegetable polyphenols with high molecular weights. Biofactors hematopoiesis. Blood 93, 488–499. 13., 121–125. Pevny, L., Simon, M.C., Robertson, E., Klein, W.H., Tsai, S.F., Zs-Nagy, I.A., 1992. proposal for reconsideration of the role of D’Agati, V., Orkin, S.H., Costantini, F., 1991. Erythroid oxygen free radicals in cell differentiation and aging. Ann. differentiation in chimaeric mice blocked by a targeted N.Y. Acad. Sci. 673, 142–150.