Expression of heme oxygenase-1 is repressed by interferon-c and induced by hypoxia in human retinal pigment epithelial cells Reiko Udono-Fujimori 1 , Kazuhiro Takahashi 1 , Kazuhisa Takeda 1 , Kazumichi Furuyama 1 , Kiriko Kaneko 1 , Shigeru Takahashi 2 , Makoto Tamai 3 and Shigeki Shibahara 1 1 Department of Molecular Biology and Applied Physiology, Tohoku University School of Medicine, Sendai; 2 Laboratory of Environmental Molecular Physiology, School of Life Science, Tokyo University of Pharmacy and Life Science, Hachioji; 3 Department of Ophthalmology, Tohoku University School of Medicine, Sendai, Japan The retinal pigment e pithelium ( RPE) is essential for main- tenance of photoreceptors and normally functions under conditions enriched with reactive oxygen species. RPE therefore expresses various defense e nzymes against oxida- tive stress, including heme oxygenase-1 (HO-1). HO-1 catalyzes heme breakdown to release iron, carbon monox- ide, and biliverdin, which is reduced to bilirubin, a potent radical s cavenger. HO-1 expression is i nduced by various environmental factors, which has b een established as a def- ense mechanism. To explore the hypothesis that the expression level of HO-1 is reduced in those RPE cells under certain conditions, we a nalyzed the effects of i nterferon-c and hypoxia, each of which represses the expression of HO-1 mRNA in other types of human cells. Expression levels of HO-1 mRNA were reduced by interferon- c in two human RPE cell lines, D407 and ARPE-19, which was consistently associated with the induction of mRNA for Bach1, a transcriptional repressor for the HO-1 gene. On the other hand, HO-1 and Bach1 mRNAs were induced by hypoxia in D407 cells but remained unchanged in ARPE-19 cells, suggesting that Bach1 is not a sole regulator for HO-1 expression. The hypoxia-mediated induction of HO-1 mRNA in D 407 cells depends on gene transcription and protein synthesis, as judged by the effects of their inhibitors. The half-life of HO-1 mRNA did not change during hyp- oxia. Thus, hypoxia may increase tran scription of the HO-1 gene through a certain protein factor in RPE cells. These results indicate that RPE cells maintain retinal homeostasis by repressing or inducing the expression of HO-1, depending on the microenvironment. Keywords: heme oxygenase-1; hypoxia; interferon-c;oxida- tive stre ss; retinal pigment epithelium. The retinal pigment epithelium ( RPE) forms a single cell layer located between the retinal photoreceptors and the vascular-rich choroids, thereby constituting the blood– retinal barrier. Thus, RPE normally functions under relatively high oxygen tensions in postnatal life. At the apical side, RPE contacts with outer segments of photo- receptors through i ts large numbers of villi, and is involved in phagocytosis of shed outer segments [1] and in uptake, processing, and transport of retinoids [2]. RPE also participates in absorption of light with its melanin granules. RPE is therefore essential f or visual function and s urvival of the photoreceptors. Conversely, dysfunction of RPE may lead to loss of photoreceptors or retinal degeneration [3], which accounts for a m ajor cause of aging-dependent visual impairment and blindness in the developed world. Heme oxygenase is a rate-limiting enzyme in heme catabolism and cleaves heme to form biliverdin I Xa, carbon monoxide, and iron [4]. Biliverdin IXa is immediately reduced to bilirubin IXa (bilirubin) during the last step of heme breakdown reaction [5]. There are two functional isozymes of heme oxygenase, heme oxygenase-1 (HO-1) and heme oxygenase-2 (HO-2) [6,7]. Expression of HO-1 mRNA is increased in human cells by the substrate heme [8], heavy metals [9–11], UV irradiation [10], and nitric oxide donors [12–14]. Because bilirubin functions as a natural radical scavenger [15,16], inductio n of HO-1 represents a protective response against oxidative stress. The physiolo- gical importance of HO-1 has been confirmed by the severe phenotypic alterations of the HO-1 deficient mice [17] and a patient with HO-1 deficiency [18]. In contrast, H O-2 is constitutively expressed and the expression levels of HO-2 mRNA are maintained within narrow ranges in human cells [12,13,19]. RPE expresses various enzymes that are important in protection against oxidative stress, including HO-1 [20,21], thereby coping with large amounts of oxygen radicals generated b y light exposure and during a ctive phagocytosis of shed outer segments. The pioneering study in the bovine ocular tissues has demonstrated the high activities of heme oxygenase and cytochrome P450-dependent monooxygen- ases in the ciliary body and the RPE [20]. Subsequent studies have shown that HO-1 is induced in rat retina by intense Correspondence to S. Shibahara, Department of M olecular Biology and Applied Physiology, Tohoku University School of Medicine, 2-1 Seiryo-machi, Aoba-ku, Sendai, Miyagi 980-8575, Japan. Fax: + 81 22 7178118, Tel.: + 81 22 7178117, E-mail: shibahar@mail.tains.tohoku.ac.jp Abbreviations: HO, heme oxygenase; HRE, hypoxia-responsive ele- ment; IFN, interferon; IL, interleukin; RPE, retinal pigment epithe- lium; TGF, transforming growth factor; TNF, tumor necrosis factor. (Received 3 1 March 2004, revised 21 May 2004, accepted 2 J une 2004) Eur. J. Biochem. 271, 3076–3084 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04241.x visible light [22] and in cultured human RPE cells by transforming growth factor-b1 (TGF-b1) [23] and oxidative stresses [21,24,25]. These results support the notion that expression of HO-1 is important in the s urvival and maintenance of RPE in the adult retina. On the other hand, we have shown that HO-1 expression is reduced in human cell lines by treatmen t with interferon-c [26,27] or by hypoxia in primary cultures of human umbilical vein endothelial cells, coronary artery endothelial cells, and astrocytes [28], as well as in various human cell lines [27]. These results raise the possibility that some degree of reduction in the HO-1 expression is an important defense mechanism in human cells under certain conditions [29]. To explore the above h ypothesis, we analyzed the effects of interferon-c and hypoxia on HO-1 expression in human RPE cells, both of which have been shown to repress HO-1 expression in other types of human cells. In addition, interferon-c induces the expression of major histocompat- ibility complex class II antigen on the cell surface of human RPE [30,31], and is involved in the pathophysiology of inflammatory ocular diseases, such as proliferative vitreo- retinopathy [ 32,33], c ytomegalovirus retinitis [34] and toxo- plasma chorioretinitis [35]. Hypoxia has been suggested as a risk factor for diabetic retinopathy [36], which a ccounts for a common cause of blindness. However, t here is no report concerning the effects of interferon-c or hypoxia on HO-1 expression in human RPE cells. Here we show the repression of HO-1 expression by interferon-c and the induction of HO-1 expression by hypoxia in human RPE cells. Experimental procedures Materials Recombinant human interferon-c (IFN-c-1a, Imunomax-c) was a gift from Shionogi Co. (Osaka, Japan). Human tumor necrosis factor-a (TNF-a) and interleukin-1b (IL-1b)were obtained from PeproTech EC Ltd (London, UK). Cell culture The human RPE cell lines, ARPE-19 [37] and D407 [38], were provided by L. M. Hjelmeland (Department of Biological Chemistry, University of California, Davis, CA, USA) and R. C. Hunt (Department of Microbiology, University of South Carolina Medical School, Columbia, SC, USA), respectively. ARPE-19 cells were cultured in a 1 : 1 mixture of DMEM and nutrient mixture F12 contain- ing 10% fetal bovine serum, 2 m ML -glutamine, a nd antibi- otics (100 UÆmL )1 penicillin and 0 .1 mgÆmL )1 streptomycin) [37]. D407 cells were cultured in DMEM containing 10% fetal bovin e s erum, 2 m ML -glutamine and antibiotics (at the same amounts as above) [38]. Cells were cultured at 37 °C under 5% CO 2 and 95% air, unless otherwise indicated. To examine the effects of cytokines on the expression of HO-1 mRNA, D407 a nd ARPE-19 cells were cultivated in fresh medium for 24 h and then exposed to the following three cytokines: interferon-c (100 UÆmL )1 ), TNF-a (20 ngÆmL )1 )andIL-1b (10 ngÆmL )1 ), as described previ- ously [39]. The cells were incubated at 37 °C for 24 h and were harvested for RNA extraction. For hypoxia experi- ments, human RPE cells were cultured in a chamber equilibrated with 5% CO 2 ,94%N 2 ,and1%O 2 [40]. The cells were cultivated under normoxia or hypoxia for indicated hours, and harvested for extraction o f RNA and protein. D407 cells were also incubated under normoxia o r hypoxia with or without actinomycin D (1 lgÆmL )1 )or cycloheximide (1 lgÆmL )1 ) for 12 or 24 h. The D407 cells were also incubated with the recombinant human TGF-b1 for the 12 h. Northern blot analysis Total RNA was extracted from cultured RPE cells and subjected to Northern blot analysis, as detailed previously [39,40]. The northern probes used for heme oxygenase mRNAs were the XhoI/XbaI fragment (nucleotide positions )64 to 923) derived from the human heme oxygenase-1 cDNA, pHHO1 [8], and the HinfI/HinfI fragment of human heme oxygenase-2 cDNA, pHHO2-1 [19]. The northern probe for human Bach1 mRNA was the Pst1 fragment of human Bach1 cDNA [41]. The human Nrf2 cDNA segment was prepared from total RNA of D407 cells by RT-PCR using a forward primer (5¢-AGTCAGAAACCAGT GGATCT-3¢) (GenBank accession number S74017; nucleo- tide positions 269–289) and a reverse primer (5¢-AGAT TCCACTGAGTGTTCTG-3¢) (nucleotide positions 1061– 1080). The human Nrf2 cDNA fragment of 810 basepairs was c loned into pBluescript KS, yielding a subclone, p BS- hNrf2. The EcoRI/NcoI fragment (vector/vector) of pBS- hNrf2 was used as a northern probe for Nrf2 mRNA. The expression of b-actin mRNA was examined as an internal control. The p robe for b-actin mRNA was the SmaI/ScaI fragment (nucleotides 124–1050) of a human b-actin cDNA provided by T. Yamamoto (Tohoku University, Sendai, Japan). These DNA fragments were labeled with [a- 32 P]dCTP (Amersham Biosciences) by the random priming method and were used as hybridization probes. Total RNA (15 lg per sample) was electrophoresed on 1% agarose gels containing 2 M formaldehyde, tran sferred to nylon membranes filter (Zeta-probe membrane; Bio-Rad), and fixed with a UV-linker (Stratalinker 1800; Stratagene). The RNA blot was hybridized w ith each 32 P-labeled probe, as described previously [39,40]. Radioactive signals were detected by exposing the filters to X-ray films (X-AR5; Kodak) or with a Bioimage A nalyzer (BAS1500; Fu ji Film Co. Ltd). The exposure time to X-ray films varied depend- ing on the experiments. The intensity of hybridization signals was determined by photo-stimulated luminescence with a Bioimage Analyzer. Northern blot analysis for HO-1, TGF-b1andb-actin was also performed with DIG Northern Starter Kit (Roche Diagnostics, Mannheim, Germany) according to the manu- facturer’s protocol. For preparing HO-1 RNA probe, the cDNA fragment (corresponding to nucleotides 81–878 of human HO-1 cDNA) (GenBank Accession Number X06985) was amplified by PCR using Pfu Turbo DNA polymerase (Stratagene, La Jolla, CA, USA), then cloned into pCR-bluntII-TOPO (Invitrogen, Carlsbad, CA, USA), and named pCR-hHO1. RT-PCR was performed for preparation of the human TGF-b1 cDNA. First strand cDNA was synthesized by Thermoscript TM reverse tran- scriptase (Invitrogen) using mRNA from a human eryth- roleukemia cell line [42]. Then, a part of human TGF-b1 Ó FEBS 2004 Dynamic regulation of HO-1 expression in RPE (Eur. J. Biochem. 271) 3077 cDNA (corresponding to nucleotides 395–2159 of GenBank accession number NM_000660) was amplified using FastStart DNA polymerase (Roche Diagnostics) with GC-RICH solution, amplified fragment was cloned into pGEM-Teasy vector (Promega, Madison, WI, USA), and named pGEM-hTGFb1. SP6 RNA polymerase was used for transcription of RNA probe from pCR-hHO1. Western bolt analysis D407 human RPE cells were lysed in triple detergent lysis buffer containing 50 m M Tris/HCl (pH 8.0), 150 m M NaCl, 0.02% sodium azide, 0.1% SDS, 100 lgÆL )1 phenyl- methylsulfonyl fluoride, 1 lgÆmL )1 aprotinin, 1 lgÆmL )1 Nonidet P40, and 0.5% sodium dexycholate. The cell lysates were centrifuged at 150 000 g for 10 min, and the superna- tant (10 lg of protein) was analyzed on a SDS-polyacryl- amide gel (10%). The proteins in the gel were treated with 20% methanol buffer containing 48 m M Tris, 39 m M glycine, and 0.037% SDS and electrophoretically trans- ferred to a poly(vinylidene difluoride) membrane (Immobi- lon-P, Millipore Corporation), which was pretreated with the same buffer. Expression of HO-1 was determined with anti-HO-1 Ig [19]. The specifi c immunocomplexes were detected with a Western blot kit (ECL Plus, Amersham Biosciences). Expression of a-tubulin was determined as an internal control with a-tubulin antibody (Neo Markers, Fremont, CA, USA). HO-1 mRNA stability assays D407 cells were incubated for 12 h in fresh medium under hypoxia or normoxia, followed by addition of actinomy- cin D (1 lgÆmL )1 ). The cells were further i ncubated for 1, 2.5, and 5 h after the addition of actinomycin D under hypoxia or normoxia and harvested at each time p oint for RNA extraction. Transient transfection assays D407 cells in six-well plates (1 · 10 6 cells per well) were transfected by the FuGENE TM 6 protocol (Boehringer Mannheim). Reporter plasmids, pSV40 promoter-Epo HRE-Luc containing four copies of hypoxia-responsive element (HRE) and pSV40 promoter-Luc lacking HRE [43], were used as a positive and a negative control for hypoxic induction. Four constructs, phHOLUC40, phHO- LUC ()1976), phHOLUC ()981) and phHOLUC ()58) were reported previously [44]. The numbers indicate the nucleotide positio ns of the 5¢-ends of the growth region. The 4.5-kb PstI/Xho1 f ragment of the human HO-1 gene [45,46] was inserted into the SmaIandXhoIsiteofpGL3-Basic vector (Promega), yielding phHOLUC45. pHHOLUC ()31), containing the f ragment ()31 to +24), was prepared by inserting synthetic double-stranded oligonucleotides into the SmaI/XhoIsiteofpGL3-Basicvectorandlacksthe putative HRE site, CACGTG sequence ()44 to )39). D407 cells were transfected with each plasmid DNA, followed by 24-h incubation and harvested. The amounts of DNA usedfor t ransfection were 1 lg of a test fusiongene and 20 ng of an internal control, pRL-TK, containing the herpes simplex virus thymidine kinase promoter region upstream of Renilla luciferase (Promega). The fusion genes contained the firefly luciferase gene a s a reporter under the control of the 5¢-flanking region of the human HO-1 gene. A promotorless construct (pGL3 Basic) was used as a control. Expression of reporter genes and pRL-TK was determined with the Dual-Luciferase TM Reporter Assay System (Promega). Results Repression of HO-1 expression by interferon-c in the RPE cell lines It has been reported t hat HO-1 expression was induced by oxidative stress in two human RPE cell lines, D407 and ARPE-19 [21,25], indicating that these RPE cells posse ss the de fense system involving HO-1. These RPE cell lines therefore provide a good system to explore the hypothesis that expression of HO-1 is re duced as a defense mechanism under certain conditions. We have focused on interferon-c that has been shown to reduce the expression of HO-1 in human glioblastoma cells [26]. For comparison, we also analyzed the e ffects of the pro-inflammatory cytokines, TNF-a and I L-1b, both of which are also involved in the pathogenesis of proliferative vitreoretinopathy [32,33]. The dose–response and time-course studies showed that the maximum reduction was obtained with interferon-c at thedoseof100UÆmL )1 and at 24 h of incubation (data not shown). T he expression levels of HO-1 mRNA were decreased by the treatment with interferon-c,whichwas consistently associated with the induction of Bach1 mRNA (Fig. 1). Bach1 has been shown to function as a repressor of the HO-1 gene in mice [47] and in cultured human lung cancer cells [27]. The expression levels of HO-1 and Bach1 mRNAs are inversely regulated by interferon-c in both RPE cell lines. I n contrast, T NF-a,IL-Ib, or their combination did not noticeably change the expression levels of HO-1 and Bach1 mRNAs in D407 cells, but each combination with interferon-c reduced the HO-1 expression and induced the Bach1 expression (Fig. 1A). In ARPE-19 cells, however, TNF-a or IL-Ib induced the expression of HO-1 mRNA but not Bach1 mRNA, and any combination with inter- feron-c decreased the HO-1 mRNA levels and increased Bach1 mRNA levels (Fig. 1B). Notably, the combination of TNF-a and IL-1b caused the concomitant induction of HO-1 and Bach1 mRNAs in ARPE-19 cells. The d ifferen- tial effects of TNF-a and IL-1b in the two RPE cell lines may reflect the hetero geneity of R PE cells [31,48]. Effects of hypoxia on the expression of HO-1 in human RPE cell lines In contrast to interferon-c, hypoxia indu ced the expression levels of HO-1 mRNA in D407 cells and exerted no or only marginal effects in ARPE-19 cells (Fig. 2). In D407 RPE cells, the levels of HO-1 mRNA were increase d by 6 h after exposure to hypoxia and reached the maximum by 12 h of hypoxia, in which b-actin mRNA levels remained unchanged (Fig. 2A). Likewise, hypoxia increased expres- sion of Bach1 mRNA, but decreased the mRNA levels of Nrf2, a transcriptional inducer of the HO-1 gene [49]. Thus, the expression levels of HO-1 and Bach1 mRNAs were concomitantly increased in D407 cells under hypoxia. In 3078 R. Udono-Fujimori et al. (Eur. J. Biochem. 271) Ó FEBS 2004 ARPE-19 cells, hypoxia caused no noticeable changes in the expression levels of Bach1 mRNA and rather reduced Nrf2 mRNA (Fig. 2B). Induction of HO-1 protein in D407 RPE cells by hypoxia HO-1 protein levels were increased by hypoxia in D407 RPE cells by 12 h and 24 h, as judged by Western blot analysis (Fig. 3). In contrast, hypoxia had no noticeable effects on the expression levels of a-tubulin, an internal control. Nature of the induction of HO-1 mRNA by hypoxia in D407 RPE cells The time course study of HO-1 mRNA induction in D407 cells showed the relatively late peak of the induction at 12 h (Fig. 2A), compared to the effects of c admium [46] or 12-O-tetradecanoylphorbol-13-acetate [50] with the maxi- mum induction by 3 h, e ach of which activates the transcription of the human HO-1 gene. Thus, the hypoxia- mediated indu ction of HO-1 expression may be a conse- quence of the production of a certain endogenous factor in D407 cells. We therefore studied the effects of actinomy- cin D, an inhibitor of transcription, and cycloheximide, an inhibitor of translation, on the hypoxia-mediated induction of HO-1 mRNA in D407 cells (Fig. 4). The expression levels of HO-1 and Bach1 mRNAs were reduced to the undetectable levels by the treatment with actinomycin D. Importantly, the hypoxia-mediated induction of HO-1 mRNA was prevented by the treatment with cycloheximide. Thus, RNA synthesis and new protein synthesis are Fig. 2. Differential effects of hypoxia on expression of HO-1, Bach1, and Nrf2 mRNAs in human RPE cells. D407 (A) and ARPE19 (B) human RPE cells were cultivated un der norm oxia (20% O 2 ;N)or hypoxia (1% O 2 ; H) for the indicated time (h) and then harvested for RNA preparation. Shown are representative examples of Northern blot analyses. The lane labe led with Ô0Õ contained RNA prepared from the untreated cells. The lower gel image in each panel s hows b-actin mRNA as an internal control. Fig. 3. Induction of HO-1 protein in D407 RPE cells by hypoxia. Western blot analysis of the HO-1 and a-tubulin protein. Each lane contained whole c ell extrac ts prepared fro m D407 hu man R PE ce lls exposed to norm oxia o r hypoxia for the indicated number of hours. See Fig. 2 legend for key. Fig. 1. Effects of cytokines on HO-1 mRNA expression in human RPE cells. Northern blot analyses of HO-1 mRNA and Bach1 mRNA in humanRPEcellstreatedwiththefoll owing c ytokines: interferon-c (IFNc), TNF-a and IL-1b. D407 RPE cells (A) and ARPE-19 cells (B) were treated with one of these three cytokines, or a combination of two or three cytokines for 2 4 h. Each lane contains 15 lgtotalRNA.The lower gel image in each panel shows b-actin mRNA as an internal control. Th e d ata shown are from one of t wo in dep endent experiments. Ó FEBS 2004 Dynamic regulation of HO-1 expression in RPE (Eur. J. Biochem. 271) 3079 required for the induction of HO-1 mRNA. Notably, the expression of Bach1 mRNA was increased by the treatment with cycloheximide under normoxia, and was further increased under hypoxia. These results suggest that the degradation of Bach1 mRNA may be enhanced by a protein factor with a short half-life, which is down-regulated by hypoxia. Stability of HO-1 mRNA under hypoxia We then analyzed the stability of HO-1 mRNA in D407 cells under hypoxia (Fig. 5), as it has been reported that hypoxia increased the stability of HO-1 mRNA in human dermal fibroblasts [51]. There was no significant difference in the half-life of HO-1 mRNA between under normoxia (mean ± SEM, 1.42 ± 0.12 h, n ¼ 3) and under hypoxia (1.63 ± 0.02 h). The short half-lives of HO-1 mRNA are consistent with the remarkable inhibitory effects of actino- mycin D (Fig. 4). These results suggest that the hypoxia- mediated induction of HO-1 mRNA expression is not due to the increased stability of HO-1 mRNA. It is also noteworthy that t he maximum levels o f HO-1 mRNA w ere significantly reduced after 5 h under hypoxia even in the absence of a ctinomycin D, indicating that the hypoxia- mediated induction of HO-1 expression represents an acute response in RPE cells. Functional analysis of the HO-1 gene promoter in D407 RPE cells All the data suggest that hypoxia may increase the transcription of the HO-1 gene as a consequence of the production of a certain endogenous factor in D407 cells. On the other hand, we have been interested in th e presence of a putative HRE sequence CACGTGA (positions )44 to )39) that overlaps the functional E-box in the human HO-1 gene promoter [11,45]. HRE is the binding site for h ypoxia- inducible factor-1. We therefore performed transient expres- sion assays to analyze the effects of hypoxia on the promoter activity of the HO-1 gene in D407 cells (Fig. 6). The basal promoter activity of phHOLUC45 was higher than that of phHOLUC40, which may be due to the presence of a Maf recognition element. Hypoxia did not influence t he expression o f any HO-1 constructs containing a putative HRE sequence CACGTGA, whereas the p uta- tive HRE (the E-box motif) appears to be required for the basal promoter activity of th e HO-1 gene. In contrast, hypoxia consistently increased the promoter activity of a construct, HRESV40, which contains four copies of HRE, but showed only marginal effects on the promoter activity of N-HRESV40, a negative control. Thus, the cis-acting elements located in the 5¢-upstream region of 4.5 kb are unable t o confer the hypoxia-mediated induction or repres- sion on a reporter gene in D407 cells. M oreover, treatment with interferon-c exerted no noticeable effects on the Fig. 4. Inhibitory effects of cycloheximide or actinomycin D o n expression of HO-1 mRNA in D407 RPE cells. D407 human RPE cells were incubated with out or with actinomycin D (1 lgÆmL )1 )or cycloheximide (1 lgÆmL )1 ) for the indicated time (h) under normoxia (N) o r hypoxia (H). Middle and lower gel images show Bach1 mRNA expression for comparison and b-actin mRNA as an internal control. Fig. 5. Effect of hypoxia on the stability of H O-1 mRNA. (A) N orthern blot analysis [conditions; normoxia + actinomycin D (N+AM-D), hypoxia (H), and hypo xia + actinomycin D (H+AM-D)]. D4 07 human RPE cells were incubated for 12 h under hypoxia or normoxia, and then further incubated for 1, 2.5, and 5 h after addition of actinomycin D (1 lgÆmL )1 ). Other conditions are the same as in Fig. 1. The data shown a re from one o f three indepe ndent exp eriments with similar results. (B) Relative expression levels of HO-1 mRNA. The intensities representing H O-1 mRNA at the time of addition of actinomycin D un der e ach co ndition were con sidered to be 100% . Th e data shown are mean ± S EM (n ¼ 3). 3080 R. Udono-Fujimori et al. (Eur. J. Biochem. 271) Ó FEBS 2004 promoter activity of phHOLUC45 (data not shown), despite the presence of a Maf recognition element that is bound by Bach1-small Maf complexes [52]. Discussion A number of studies have shown the beneficial role of HO-1 induction, as described above, but the red uced expression of HO-1 has been largely ignored. We have hypothesized that a certain degree of reduction in the HO-1 expression levels may be beneficial under pathological conditions, such as infectious diseases or cancers, because the reduced HO-1 expression may result in the restriction of iron supply to pathogens or tumor cells or the preservation of intracellular heme that is an essential component of some defense enzymes [29]. In the former context, it has been reported that human pancreatic tumor cells over-expressing HO-1 show the aggressive properties when inoculated in immu- nodeficient mice, i ncluding the enhanced growth, angiogen- esis and lung metastasis [53], and treatment of these mice with a HO-1 inhibitor reduced the occurrence of lung metastasis [53]. These results provide the in vivo evidence suggesting the beneficial role o f the reduced HO-1 activity. To explore the physiological significance of the reduced HO-1 expression, we analyzed the effect of cytokines on the expression of HO-1 in the RPE, which normally functions under the extreme conditions enriched with reactive oxygen species. We show that HO-1 expression is consistently reduced by interferon -c,eveninRPEcells,whichmay reflect an important mechanism for the maintenance of the retinal homeostasis. It should be noted that the reduced expression levels of HO-1 mRNA are maintained at the detectable levels, unlike the severe effect of actinomycin D (Fig. 4), as HO-1 is important in cell s urvival. The R PE is a target cell infected by cytomegalovirus [54] and by the intracellular parasite Toxoplasma gondii [34]. Toxoplasmosis is a common protozoal infection in the developed world, and chorioretinitis is a major complication of ocular toxoplasmosis in infants, patients with acquired immune deficiency syndrome, and organ transplant recip- ients [35]. Incidentally, interferon-c at the c oncentration of 100 UÆmL )1 used in this study was shown to completely inhibit Toxoplasma gondii replication in cultured human RPE cells [35]. I n this case , interferon-c has been shown to inhibit proliferation of parasites by inducin g the expression of indoleamine 2,3-dioxygenase, a heme-containing enzyme that converts tryptophan to kynurenine, thereby depleting cellular tryptophan [35]. Likewise, the protective role of the interferon-c-induced indoleamine 2,3-dioxygenase has been shown in cytomegalovirus retinitis [34]. T hese results raise the intriguing possibility that the reduced HO-1 expression may result in the preservation of intracellular heme that is an essential component of indoleamine 2,3-dioxygenase. Hypoxia is a potent stimulus for neovascularization that is normally prevented in the adult retina. Therefore, hypoxia may reflect severe pathologic conditions for the RPE, such as retinal vascular occlusive diseases and retinal detachment, which may be followed by a ngiogenesis as s een in d iabetic retinopathy and age-related macular degeneration. Here we show that hypoxia transiently induces the expression of HO-1 mRNA through an unknown p rotein factor in D407 cells but exerts no noticeable effects in ARPE-19 cells. Such differences in hypoxic responses may suggest the presence of the regulatory mechanism that maintains the HO-1 Fig. 6. Effect of hypoxia on the human HO-1 g ene promote r function. D407 human RPE cells were transfected with each reporter c onstruct and incubated under normoxia or hypoxia. A putative HRE in the HO-1 gene promoter is marked with Ô?Õ. The two constructs, shown n ear the b ottom, represent positive and negative controls for hypoxia. Relative luciferase activity under normoxia or hypoxia is shown as the ratio t o the normalized luciferase activity obtained with pG L3Basic under normoxia or hypoxia, respectively. T he data are means ± S EM of three independent experi- ments. Ó FEBS 2004 Dynamic regulation of HO-1 expression in RPE (Eur. J. Biochem. 271) 3081 expression levels within narrow ranges under hypoxia in RPE c ells. Alternative ly, the diffe ren ce may s imp ly r eflect the regional variations in the differentiation state or proliferative capacity of RPE cells [31,48]. In fact, we have reported that hypoxia increases production of endothelin-1 in D407 cells but not in ARPE-19 cells, whereas hypoxia induces expres- sion of adrenomedullin in bothRPE cell lines [40]. M oreover, two transcription factors, microphthalmia-associated tran- scription factor and OTX2, which are important in differ- entiation of RPE, are more abundantly expressed in ARPE- 19 cells than in D407 cells [55]. Thus, ARPE-19 cells may retain m ore differentiated properties, which is consistent with the previous report by other investigators [25]. Hypoxia up-regulates HO-1 expression in D407 RPE cells without affecting the half-life of HO-1 mRNA (Fig. 5), whereas other investigators have reported that hypoxia induces the expression of HO-1 in human dermal fibroblasts by increasing the half-life of HO-1 mRNA [51]. I n human dermal fibroblasts, the induction was inhibited by cyclo- heximide and peaked by 10 h [51], which is similar to the properties of the HO-1 induction observed in D407 cells. These results suggest that hypoxia may i nduce HO-1 expression through newly synthesized protein factors but the induction mechanisms are different between RPE cells and skin fibroblasts. In addition to the well-known inter- species variations [29], the present study has shown the intercell differences in the h ypoxic induction of human HO-1 gene expression. A question that remains to be answered is the i dentity of a protein factor that is responsible for the hypoxia-mediated induction of HO-1 expression in D407 RPE cells. Probably, many factors a re induced by hypoxia at their protein levels in RPE cells. For example, we have shown the increased production of endothelin-1 and adrenomedullin in D407 RPE cells [40]. Furthermore, it has been reported that HO-1 expression is induced by TGF-b1 in human RPE cells [23], human renal epithelial cells [56], a nd A549 human lung cancer cells [57]. Thus, TGF-b1 is a candidate that may mediate the hypoxic induction of HO-1 in RPE cells. In fact, hypoxia increased the expression of TGF-b1mRNAin D407 cells (data not shown), as reported in other human RPE cells [58]. However, hypoxia has been shown to reduce the production of TGF-b1 in human RPE cells [58]. F urther studies, such as DNA microarray analysis, may help us to find such a factor. Bach1 is a heme-regulated transcriptional repressor for the HO-1 gene and plays an important role in the feedback regulation of HO-1 expression [47,59]. However, the expression of HO-1 and Bach1 mRNAs w as concomitantly induced by the treatment with the combination of TNF-a and IL-1b in ARPE-19 cells, indicating that Bach1 is not a sole determinant for HO-1 mRNA expression in ARPE-19 cells. Likewise, hypoxia induced expression of HO-1 and Bach1 mRNAs in D407 cells, indicating that Bach1 may not be a key determinant for the hypoxia-mediated induction of HO-1 expression in D407 cells. In addition, hypoxia coordinately and rapidly induced expression of both HO-1 and Bach1 mRNAs in cultured rat and monkey cells, indicating that increased e xpression of Bach1 does not necessarily result in the inhibition of HO-1 transcription [27]. These results indicate that additional regulators, such as corepressors, may influence Bach1 activity. It is noteworthy that HO-1 expression is not reduced by hypoxia in th e two RPE cell lines, unlike other types of human cells [27,28]. To confi rm these findings, we wish to perform experiments using primary cultures of human RPE, but are unable to obtain sufficient numbers of original RPE cells from donors. Instead, using a hypoxic chamber (10% oxygen) [60], we are w orking on a rodent model of high altitude retinopathy, although the interspecies difference has been well known in th e regulation of HO-1 expression [29]. In summary, the present study has shown that interferon- c consistently reduces the expression of HO-1 mRNA in two t ypes of human RPE cell lines, in which HO-1 mRNA is induced or remains unchanged under hypoxia. Thus, human RPE cells up-regulate or down-regulate the HO-1 expression through different pathways in a dynamic manner to cope with the changes in the retinal microenvironment. Acknowledgements We thank Dr R.C. Hunt for D 407 RPE cells and Dr L.M. Hjelmeland for ARPE-19 RPE cells. We also thank Y. Fujii-Kuriyama and E. Ito for the HRE constructs and human Bach1 cDNA, respectively. This work was supported in part by Grants-in-Aid for Scientific Rese arch (B), Scientific Research (C), f or Exploratory Research, and for Priority Area from the Ministry of Education, Science, Sports, and Culture of Japan and by the 21st Century COE Program Special Resea rch Gra nt Ôthe Center for Innovative Therapeutic Developm ent for Common DiseasesÕ from the Ministry of Education, Science, Sports, and Culture of Jap an. This work was also sup portedinpartbythegrantsprovided by Uehara Memorial Foundation. References 1. Bok, D. (1988) Retinal photoreceptor disc shedding and pigment epithelium ph agocytosis. I n Re tinal D iseases: Medical and Sur gical Management (Rayan, S .J., Ogde n, T.E. & Schachat, A.P., e ds), p p. 69–81. C V. Mosby Co, St Louis, MO. 2. Bok, D. (1993) The retinal pigment epithelium: a versatile partner in vision. J. Cell Sci. Suppl. 17, 189–195. 3. Gao, H. & Hollyfield, J.G. (1992) Aging of the human retina. Differential loss o f neuro ns and r etinal pigme nt epithe lial cells. Invest. Ophthalmol. Vis. Sci. 33, 1–17. 4. Tenhunen, R., Marver, H.S. & Schmid, R. (1968) The enzymatic conversion of heme to bilirubin by microsomal heme oxygenase. Proc.NatlAcad.Sci.USA61, 748–755. 5. Yoshida, T. & Kikuchi, G. (1978) Features of the reaction of heme degradation catalyzed by the reconstituted microsomal heme oxygenase system. J. Biol. Chem. 253, 4230–4236. 6. Shibahara, S., Muller, R., Taguchi, H. & Yoshida, T. (1985) Cloning and expression of cDNA for rat heme oxygenase. Proc. NatlAcad.Sci.USA82, 7865–7869. 7. Cruse, I. & Maines, M.D. (1988) Evidence suggesting that t he two forms of h eme oxygenase and prod ucts o f d ifferent genes. J. Biol. Chem. 263, 3348–3353. 8. Yoshida, T., Biro, P., Cohen, T., Muller, R.M. & Shibahara, S. (1988) Human heme oxygenase cDNA and induction of its mRNA by hemin. Eur. J. Biochem. 171, 457–461. 9. Taketani, S., Ko hno, H ., Y oshinaga, T. & T okunaga, R. (1989) The human 32-kDa stress p rotein induced by exposure to arsenite and cadmium ions is heme oxygenase. FEBS Lett. 245, 173–176. 10. Keyse, S.M. & Tyrrell, R.M. (1989) Heme oxygenase is the major 32-kDa stress protein induced in hum an skin fibrob last by UVA radiation, hydrogen peroxide, and sodium arsenite. Proc. Natl Acad. Sci. USA 86, 99–103. 3082 R. Udono-Fujimori et al. (Eur. J. Biochem. 271) Ó FEBS 2004 11. Sato, M., Ishizawa, S., Yoshida, T. & Shibahara, S. (1990) Interaction of upstream stimulatory factor with the human heme oxygenase gene promoter. Eur. J. Biochem. 188, 231–237. 12. Takahashi, K., Hara, E., Suzuki, H. & Shibahara, S. (1996) Expression of heme oxygenase isozyme mRNA in the human brain and induction of heme oxgenase-1 by nitric oxide donors. J. Neurochem. 67, 482–489. 13. Takahashi, K., Hara, E., Ogawa, K., Kimura, D., Fujita, H. & Shibahara, S. (1997) Possible implications of the induction of human heme oxygenase-1 by nitric oxide donors. J. Biochem. (Tokyo) 121, 1162–1168. 14. Hara, E., Takahashi, K., Takeda, K., Nakayama, M., Yoshizawa, M.,Fujita,H.,Shirato,K.&Shibahara, S. (1999) Induction o f heme oxygenase-1 g ene as a response in sensing the signals evoked by distinct nitric oxide donors. Biochem. Ph ar macol. 58, 227–236. 15. Stocker, R., Yamamoto, Y., McDonagh, A.F., Glazer, A.N. & Ames, B.N. (1987) Bilirubin is an antioxidant of possible physio- logical i mportance. Science 23 5, 1043–1046. 16. Yamaguchi, T., Shioji, I., Sugimoto, A., Komoda, Y. & Nakaji- ma, H. (1994) Chemical structure of a new family of bile pigments from human urine. J. Biochem. (Tokyo) 116 , 298–303. 17. Poss, K.D. & Tonegawa, S. (1997) Heme oxygenase 1 is required for mammalian iron reutilization. Proc. Natl Acad. Sci. USA 94, 10919–10924. 18. Yachie, A., Niida, Y., W ada, T., Igarashi, N., Kaneda, H., Toma, T., Ohta, K., Kasahara, Y. & Koizumi, S. (1999) Oxidative stress causes enhanced endothelial cell injury in human heme oxygenase- 1 deficiency. J. Clin. Invest. 103, 129–135. 19. Shibahara, S., Yoshiz awa, M., Suzuki, H., Takada, K., Megu ro, K. & Endo, K. (1993) Functional analysis of cDNAs for two types of human heme oxygenase and evidence for their separate reg- ulation. J. Biochem. (Tokyo ) 113, 214–218. 20. Schwartzman, M.L., Masferrer, J., Dunn, M.W., McGiff, J.C. & Abraham, N.G. (1987) Cytochrome P450, drug metabolizing en- zymes and arachidonic acid me tabolism in bovine ocular tissues. Curr. Eye Res. 6, 623–630. 21. Hunt, R.C., Hunt, D.M., Gaur, N. & Smith, A. (1996) Hemo- pexin in the human retina: protection of th e retina against heme- mediated toxicity. J. Cell. Physiol. 168, 71–80. 22. Kutty, R .K., Kutty, G., Wiggert, B., Chader, G.J., Darrow, R.M. & Organisciak, D.T. (1995) Induction of heme oxygenase 1 i n the retina by intense visible light: suppression by the antioxidant dimethylthiourea. Proc.NatlAcad.Sci.USA92, 1177–1181. 23. Kutty, R.K., Nagineni, C.N., Kutty,G.,Hooks,J.J., Chader, G.J. & Wiggert, B. (1994) Increased expression of heme oxygenase-1 in human retinal pigment epith elial cells by transforming growth factor-beta. J. Cell. P hysiol. 159, 371–378. 24. Honda, S., Hjelmeland, L.M. & Handa, J.T. (2000) T he use o f hyperoxia to induce chronic mild oxidative stress in RPE cells in vitro. Mol. Vis. 7, 63–70. 25. Alizadeh, M., Wada, M., Gelfman,C.M.,Handa,J.T.&Hjel- meland, L.M. (2001) Downregulation of differentiation specific gene expression by oxidative stress in ARPE-19 cells. Invest. Ophthalmol. Vis. Sci. 42, 2 706–2713. 26. Takahashi, K., Nakayama, M., Takeda, K., Fujita, H. & S hiba- hara, S. (1999) Suppression of heme oxygenase-1 mRNA expres- sion by interferon-c in human glioblastoma cells. J. Neurochem. 72, 2356–2361. 27. Kitamuro, T., T akaha shi, K., Ogawa, K., Udono-Fu jimori, R., Takeda, K., Furuyam a, K., Naka yama, M ., Sun, J., Fujita, H., Hida, W., Hattori, T ., Shirato, K., Igarashi, K. & Shibahara, S. (2003) Bach1 function s as a hypoxia-inducible repressor for the heme oxygenase-1 gene in human cells. J. Biol. Chem. 278, 9125– 9133. 28. Nakayama, M., Takahashi, K., Kitamuro, T., Yasumoto, K., Katayose, D., Shirato, K., Fujii-Kuriyama, Y. & Shibahara, S. (2000) Repression of heme oxygenase-1 by hypoxia in vascular endothelial cells. Biochem. Biophys. Res. Commun. 271, 665–671. 29. Shibahara, S. (2003) The heme oxygenase dilemma in cellular homeostasis: new insights for the feedback regulatio n of heme catabolism. Tohoku J. Exp. Med. 200, 1 67–186. 30. Detrick, B., Newsome, D.A., Percopo, C.M. & Hooks, J.J. ( 1985) Class II antigen expression and gamma interferon modulatio n of monocytes and retinal pigment epithelial cells from patients with retinitis pigmentosa. Clin. Immunol. Immunopathol. 36, 201–211. 31. Casella, A.M., Taba, K.E., Kimura, H., Spee, C., Cardillo, J.A., Ryan, S.J. & Hinton, D.R. (1999) Retinal pigment epithelial cells are heterogeneous in their expression of MHC-II after stimulation with interferon-gamma. Exp. Eye Res. 68, 423–430. 32. Limb, G.A., Little, B.C., Meager, A., Ogilvie, J .A., Wolstencroft, R.A., Franks, W.A., Chignell, A.H. & Dumonde, D.C. (1991) Cytokines in proliferative vitreo retinopathy. Eye 5, 686–693. 33. Limb,G.A.,Alam,A.,Earley,O., Green, W., Chignell, A.H. & Dumonde. D.C. (1994) Distribution of cytokine proteins within epiretinal membranes in proliferative vitreoretinopathy. Curr. Eye Res. 13, 791–798. 34. Bodaghi, B., Goureau, O., Zipeto, D., Laurent, L., Virelizier, J.L. & Michelson, S. ( 1999) Role of IFN-c-induced indoleamine 2,3- dioxygenase and inducible nitric oxide synthase in the re plication of human cytomegalovirus in retinal pigment epithelial cells. J. Immunol. 162, 9 57–964. 35. Nagineni, C.N., Pardhasaradhi, K., Martins, M .C., Detrick, B. & Hooks, J.J. (1996) Mec hanisms of in terfer on-induce d inhibition o f Toxoplasma gondii replication in human retinal pigment epithelial cells. Infect. Immun. 64, 4188–4196. 36. Arden, G.B. (2001) The absence of diabetic retinopathy in patients with retinitis pigmentosa: implications for pathophysiology and possible treatment. Br. J. O phthalmol. 85, 366–370. 37. Dunn, K.C., Aotaki-Keen, A.E., Putkey, F.R. & Hjelmeland, L.M. (1996) ARPE-19, a human retinal pigment epithelial cell line with differentiated p roperties. Exp. Eye Res. 62, 155–169. 38. Davis, A.A., Bernstein, P.S., Bok, D., Turner, J., Nachtigal, M. & Hunt, R.C. (1995) A human retinal p igm ent epithelial ce ll line that retains epithelial characteristics after prolonged cu lture. Invest. Ophthalmol. Vis. Sci. 36, 955–964. 39. Udono, T., Takahashi, K., Nakayama, M., Murakami, O ., Durlu, Y.K., Tamai, M. & Shibahara, S. (2000) Adrenomedullin in cul- tured human retinal pigment ep ithelial cells. Invest. Ophthalmol. Vis. Sci. 41 , 1962–1970. 40. Udono, T., Takahashi, K., Nakayama, M., Yoshinoya, A., Totsune, K., Murakami, O., Durlu, Y.K., Tamai, M. & Shiba- hara, S. (2001) Induction of ad renomedu llin by hypo xia in c ul- tured retinal pigment epithelial cells. Invest. Ophthalmol. Vis. Sci. 42, 1080–1086. 41. Kanezaki, R., Toki, T., Yokoyam, M., Yomogida, K., Sugiyama, K.,Yamamoto,M.,Igarashi,K.&Ito,E.(2001)Transcription factor BACH1 is recruited to the nucleus by its novel alternative spliced isoform. J. Biol. Chem. 276, 7278–7284. 42. Endo, K., Harigae, H., Nagai, T., Fujie, H., Meguro, K., Watanabe, N., Furuyama, K., Kameoka, J., Okuda, M., Hayashi, N., Yamamoto, M. & Abe, K. (1993) Two chronic myelogenous leukaemia cell lines which represent different stages of erythroid differentiation. Br. J. Haematol. 85 , 653–662. 43. Ema, M., Taya, S., Yokotani, N., Sogawa, K., Matsuda, Y. & Fujii-Kuriyama, Y. (1997) A novel bHLH-PAS factor with close sequence similarity to hypoxia-inducible factor 1alpha regulates the VEGF expression and is potentially involved in lung and vascular develop ment. Proc.NatlAcad.Sci.USA94, 4273–4278. 44. Takahashi, S., Takahashi, Y., Ito, K., Nagano, T., Shibahara, S. & M iura, T. (1999) Positive and negative regulation of the human heme oxygenase-1 gene expression in cultured cells. Biochim. Biophys. Acta 1447, 231–235. Ó FEBS 2004 Dynamic regulation of HO-1 expression in RPE (Eur. J. Biochem. 271) 3083 45. Shibahara, S., Sato, M., Muller, R.M. & Yoshida, T. (1989) Structural organization of the human heme oxygenase gene and the function of its promoter. Eur. J. Biochem. 179, 557–563. 46. Takeda, K., Ishizawa, S., Sato, M., Yoshida, T. & Shibahara, S. (1994) Identification of a cis-acting element that is responsible for cadmium-mediated induction of the human heme oxygenase. J. Biol. Chem. 269, 22858–22867. 47. Sun, J., Hoshino, H., Takaku, K., Nakajima, O., Muto, A., Suzuki, H., Tashiro, S., Takahashi, S., Shibahara, S., Alam, J., Taketo, M .M., Yamamoto, M. & Igarashi, K . (2002) Hemopro- tein Bach1 regulates enhancer availability of heme oxygenase-1 gene. EMBO J. 21, 5216–5224. 48. McKay, B.S. & Burke, J.M. (1994) Separation of phenotypically distinct subpopulations of cultured human retinal pigment epi- thelial cells. Exp. Cell Res. 213, 85–92. 49. Alam,J.,Stewart,D.,Touchard,C.,Boinapally,S.,Choi,A.M.& Cook, J.L. (2000) Nrf2, a Cap’n’Collar transcription factor, regulates induction of the h eme oxygenase-1 gene. J. Biol. Chem. 274, 26071–26078. 50. Muraosa, Y. & Shibahara, S. (1993) Identification of a cis-acting element and putative trans-acting factors responsible for 12-O- tetradecanoylphorbol-13-acetate (T PA)-mediated induction of heme oxygenase expression in myelomonocytic cell lines. Mol. Cell. Biol. 13, 7881–7891. 51. Panchenko, M.V., Farber, H.W. & Korn, J.H. (2000) Induction of heme oxygenase-1 by hypoxia and free radicals in human dermal fibroblasts. Am.J.Physiol.278, C92–C101. 52. Oyake, T., Itoh, K., Motohashi, H., Hayashi, N., Hoshino, H ., Nishizawa, M., Yamamoto, M. & Igarashi, K. (1996) Bach proteins belong to a novel family of BTB-basic leucine zipper transcription factors that interact with MafK and regulate tran- scription through the NF-E2 site. Mol. Cell. Biol. 16, 6083–6095. 53. Sunamura, M., Duda, D.G., Ghattas, M.H., Lozonschi, L., Motoi, F., Yamauchi, J., Matsuno, S., Shibahara, S. & Abraham, N.G. (2003) Heme oxygenase-1 accelerates tumor angiogenesis of human pancreatic cancer. Angiogenesis 6, 15–24. 54. Miceli, M.V., Newsome, D.A., Novak, L.C. & Beuerman, R.W. (1989) Cytomegalovirus replication in cultured human retinal pigment epithelial cells. Curr. Eye Res. 8, 835–839. 55. Takeda, K., Yokoyama, S., Yasumoto, K., Saito, H., Udono, T., Takahashi, K. & Shibahara, S. (2003) OTX2 regulates expression of DOPAchrome tautomerase in h uman retinal pigment epithe- lium. Biochem. Biophys. Res. C ommun. 300, 9 08–914. 56. Hill-Kapturczak, N., Truong, L ., Thamilselvan, V., Visner, G.A., Nick, H.S. & Agarwal, A. (2000) Smad7-dependent regulation of hemeoxygenase-1bytransforming growth factor-beta in human renal epithelial cells. J. Biol. Chem. 275, 40904–40909. 57. Ning, W., Song, R., Li, C., Park, E., Mohsenin, A., Choi, A.M. & Choi, M.E. (2002) TGF-beta1 stimulates HO-1 via the p 38 mitogen-activated protein kinase in A549 pulmonary epithelial cells. Am. J. Physiol. Lung Cell Mol. Physiol. 283, L1094–L1102. 58. Khaliq, A., Patel, B., J arvis-Evans, J., Moriarty, P., McLeod, D. & Boulton, M. (1995) Oxygen modulates production of bFGF and TGF-beta by retinal cells in vitro. Exp. Eye Res. 60, 415–423. 59. Ogawa, K., Sun, J., Taketani, S., Nakajima, O., Nishitani, C., Sassa, S., Hayashi, N., Yamamoto, M., Shibahara, S., Fujita, H. & Igarashi, K. (2001) Heme mediates de-repression of Maf recognition element through direct binding to transcription repressor Bach1. EMBO J. 20, 2835–2843. 60. Katayose, D., Isoyama, S., Fujita, H. & Shibahara, S. (1993) Separate regulation of heme oxygenase and heat shock protein 70 mRNA expression in the rat he art by hemodynamic stress. Biochem. Biophys. Res. Commun. 191, 587–594. 3084 R. Udono-Fujimori et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . Expression of heme oxygenase-1 is repressed by interferon-c and induced by hypoxia in human retinal pigment epithelial cells Reiko Udono-Fujimori 1 ,. protein factor in RPE cells. These results indicate that RPE cells maintain retinal homeostasis by repressing or inducing the expression of HO-1, depending on