An Atypical NF-κB-Regulated Pathway Mediates Phorbol Ester-Dependent Heme Oxygenase-1 Gene Activation in Monocytes¹

Heme oxygenase (HO)³-1 is the first and the rate-limiting enzyme of heme degradation (1). The catalytic cleavage of the prooxidant heme by HO produces iron, biliverdin, and carbon monoxide (2). Biliverdin is converted into the potent antioxidant bilirubin (3) via biliverdin reductase (4), and HO-derived carbon monoxide plays an important physiological role as a signaling gas (5, 6). HO-1 is highly inducible by a variety of oxidative stress stimuli and has been known for many years to provide antioxidant cellular protection (6). More recently, HO-1 knockout mice and a human case of genetic HO-1 deficiency have been shown to exhibit phenotypical alterations of chronic inflammation (7, 8). Furthermore, HO-1^−/− mice were highly susceptible to the toxicity of the proinflammatory mediator LPS (7, 9), and induction of HO-1 expression, either by gene transfer or by pharmacological stimulation, has emerged to be of potential therapeutic use for the treatment of inflammatory diseases in animal models (10, 11, 12, 13, 14, 15).

HO-1 is regulated primarily at the level of transcription (6, 16). An array of cis-acting regulatory elements (RE), which are targeted by transcription factors (TF) such as NF-E2-related factor 2 (Nrf2), AP-1, or USF-2, have been identified in the promoter regions of avian and mammalian HO-1 genes and are involved in HO-1 regulation (6, 17). Although the TF NF-κB, which provides cytoprotection against oxidative stress (18), has been shown to be activated by various stimuli that are also known to up-regulate HO-1 gene expression such as curcumin (19), LPS (20), or dietary polyphenols (21), the regulatory role of NF-κB for HO-1 gene regulation is discussed controversially (17, 22). Moreover, a functional κB site of the HO-1 promoter, which is the direct target of this TF, has not been identified to date. Thus, the goal of the present study was to investigate the regulation of HO-1 by the phorbol ester PMA, which is a prototypical activator of NF-κB in monocytic cells and a potent inducer of protein kinase C (PKC) (23).

In this article it is reported that PMA induces HO-1 gene expression in monocytes. This up-regulation is mediated via a newly identified κB element of the rat HO-1 proximal promoter that is a target of the NF-κB subunit p65/RelA. An atypical IκB kinase (IKK)-independent NF-κB pathway, which requires the activation of p38 MAPK and CK2, is involved in PMA-dependent induction of HO-1 gene expression in monocytes.

DMEM, RPMI 1640, and MEM were obtained from PAA Laboratories, FBS was from Biochrom, Ficoll-Paque was from Pharmacia, CD14⁺ immunomagnetic microbeads were from Miltenyi Biotec, and polyvinylidene difluoride membranes were from Millipore. All other chemicals were purchased from Sigma-Aldrich and Roche Applied Science unless otherwise indicated.

Liver tissue macrophages (LTM), peritoneal macrophages (24), rat hepatocytes (25) and human PBMC were isolated and cell culture was maintained in culture as described previously (26). RAW264.7 cells were from American Type Culture Collection, mouse embryonic fibroblasts (MEF) from p65^−/− mice were from Dr. H. Nakano (Department of Immunology, Jutendo University School of Medicine, Tokyo, Japan) (27) and were grown in DMEM supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. All cell cultures were kept under air/CO₂ (19:1) at 100% humidity. Treatment of cells with PMA (0.5 μM) and LPS (Escherichia coli 0111:B4; 1 μg/ml) was performed with serum-free medium. Specific inhibitors of CK2, apigenin, 5,6-dichloro-1-β-d-ribofuranosylbenzimidazole (RFBD) (Calbiochem), p38 MAPK inhibitor SB202190 (Calbiochem), IKK2 inhibitor SC-514 (Calbiochem), and N-acetylcysteine (NAC) were added to the culture medium 30 min to 1 h before treatment with PMA, as indicated.

Total RNA isolation and Northern blot analysis were essentially performed as described previously and, as a probe for hybridization, the cDNA of rat HO-1 and a 28S ribosomal RNA oligonucleotide was applied (28).

Cells were washed with 0.9% NaCl and thereafter lysis was performed as described (29). The lysate was centrifuged for 5 min at 13,000 × g at 4°C and the protein concentration in the supernatant was determined by the BCA (bicinchoninic acid) protein assay kit (Pierce). Fifty micrograms of total protein was separated on a 12% SDS-polyacrylamide gel and electroblotted onto polyvinylidene difluoride membranes. Membranes were blocked with Tris-buffered saline containing 5% skim milk or 5% BSA, 50 mM Tris-HCl (pH 7.6), 150 mM NaCl, and 0.1% Tween 20 for 1 h at room temperature. The primary Abs against HO-1 (Stressgen), and GAPDH (Hytest) were used at 1/1000 dilutions. The primary Abs for the detection of phosphorylated IκBα serine 32 (Cell Signaling) and tyrosine 42 (ECM Biosciences) were applied at the concentrations recommended by the manufacturers. Secondary Abs were goat anti-rabbit IgG HRP and anti-mouse IgG HRP (DPC Biermann) and were used at 1/20,000 and 1/100,000, respectively. The ECL chemiluminescent detection system (Amersham Biosciences) was used for detection according to the manufacturer’s instructions.

The luciferase reporter gene constructs pHO-1338, pHO-754, and pHO-347 have been described previously (28), pTNF-585 was from Dr. G. Duff (University of Sheffield, Sheffield, U.K.), and pNF-κB was from Dr. L. Schmitz (University of Giessen, Giessen, Germany). The expression vector for dominant negative IκBα was from Dr. R. Gaynor (University of Texas Southwestern Medical Center, Houston, TX), the expression plasmid pRSV-NF-κB (p65/RelA) was from Dr. D. Schmoll (Sanofi Aventis Pharma) (30), and the expression vector for CK2α was from Dr. R. Kemler (Max-Planck-Institute for Immunobiology, Freiburg, Germany) (31). The plasmid pFA-CHOP with the transactivation domain of the TF CHOP fused with the DNA-binding domain of yeast Gal4 and the empty control vector pFC2-dbd were purchased from Stratagene. The reporter gene construct pHO-347 κBmut, which contains targeted mutations within the rat HO-1 proximal κB site, was generated with the template pHO-347 and the oligonucleotides mutκBfor (5′-GAATTGTCTCCTAGTTCTCCTACCTTGGAGATTCC TGAGAGGGC-3′) as forward primer and mutκBrev (5′-GCCTCTCAGGAATCTCCAAGGTAGGAGAACTAGGAGACAATTC-3′) as reverse primer. The reporter construct pHO-347 MTE was generated with the oligonucleotides (5′-GAGCTTGCCAGAGCTATACAATTTATCCCCATAC-3′) as forward primer and (5′-GTATGGGGATAAATTGTATAGCTGTGGCAAGCTC-3′) as reverse primer and pHO-347 as template. Both plasmids were generated with the QuikChange XL site-directed mutagenesis kit (Stratagene) according to the manufacturer’s instructions. Plasmid pHO-20 was generated from synthetic oligonucleotides containing the rat HO-1 region from −20 to + 71 bp and XhoI or KpnI restriction sites at their ends (MWG-Biotech). Annealed dsDNA was ligated into XhoI/KpnI sites of pGL3basic (Promega) by using standard molecular cloning methods. All constructs were verified by DNA sequencing in both directions.

After growth for 24 h, transfection of plasmid DNA into RAW264.7 cells and MEF was performed by using FuGENE (Roche Applied Science) as described previously (29). Unless otherwise mentioned, cells were transfected with 0.5–1 μg of the reporter plasmid and, in cotransfection experiments, with 0.1–1 μg of the indicated expression vectors. Transfection efficiency was controlled using 0.1 μg of Renilla luciferase expression vector pRL-SV40 (Promega) as described previously (29). Cells were lysed with luciferase lysis reagent (Promega) and luciferase activity was determined with a commercial Dual-Luciferase reporter assay system (Promega) according to the manufacturer’s instructions. Cells were either harvested 24 h after transfection or treated for the time points indicated with PMA or other reagents, as indicated. Relative light units of Firefly luciferase activity were normalized with Renilla luciferase activity.

NE were prepared as described previously (29). The sequences of the biotin-labeled oligonucleotides (MWG-Biotech) used for the EMSA are as follows: HO-κB-B, (5′-CCTAGTTCTGGAACCTTCCAGATTCCTGA-3′), HO-κBmutant (HO-κB-Bmut), (5′-CCTAGTTCTTTAACCGTTAAGATTCCTGA-3′), and NF-κB consensus oligonucleotide with sequence (5′-AGTTGAGGGGACTTTCCCAGGC-3′) with respective oligonucleotides of the noncoding strand. For competition assays, an excess of unlabeled oligonucleotide was added as indicated. After preincubation for 10 min at room temperature, the biotin-labeled probe was added and incubation was continued for another 20 min. For supershift analysis, 3 μl of an Ab directed against the NF-κB p65 subunit (Cell Signaling) was added to the EMSA reaction. The reaction mixture was loaded on a 6% native polyacrylamide gel in 0.5% Tris-borate-EDTA and blotted onto nylon membranes (Pierce). After UV-cross-linking, the LightShift chemiluminescent EMSA kit (Pierce) was used to detect interaction between the biotin end-labeled DNA and the protein with a streptavidin-HRP conjugate and the chemiluminescent substrate.

To investigate the regulation of HO-1 gene expression by the phorbol ester PMA in cell cultures of the monocytic cell line RAW264.7, we determined mRNA levels of HO-1 after exposure to PMA. For a comparison, the effect of PMA on HO-1 mRNA expression was also determined in primary rat LTM, peritoneal macrophages, and hepatocytes. PMA induced HO-1 mRNA levels in RAW264.7 cells and to a similar extent also in primary LTM (Fig. 1,A) and peritoneal macrophages (data not shown), but not in hepatocytes (Fig. 1,A). Subsequently, we examined the effect of PMA on the regulation of endogenous HO-1 protein expression in cell cultures of RAW264.7 cells and LTM. Similarly as for the regulation of HO-1 mRNA levels in response to PMA, HO-1 protein expression was markedly induced by this treatment (Fig. 1,B). It is also remarkable, that PMA-dependent induction of HO-1 gene expression was observed in human PBMC (Fig. 1 B). The data indicate that PMA induced HO-1 gene expression in the monocytic cell line RAW264.7 and in various primary monocytic cells, but not in hepatocytes.

The proinflammatory mediator LPS has previously been shown to be a potent inducer of HO-1 (32). To further investigate the regulatory mechanism(s) of PMA-dependent induction of HO-1 in monocytic cells, we compared the time course of HO-1 protein expression by PMA with that elicited by LPS. Treatment with PMA induced HO-1 gene expression in RAW264.7 cells in a time-dependent manner with a maximum after 6 h that persisted up to 18 h, whereas LPS-dependent induction of HO-1 was retarded (Fig. 1,C). Moreover, we have also determined the combined effect of PMA and LPS on HO-1 gene expression. Simultaneous treatment with PMA and LPS induced HO-1 gene expression in an additive manner (Fig. 1 D). The data suggest that various mechanisms regulate HO-1 gene expression in response to PMA and LPS in RAW264.7 cells.

PMA is not only known to induce HO-1 gene expression (33), but it also activates the NF-κB pathway in monocytes (34). To further study whether HO-1 induction by PMA could be mediated via NF-κB, we searched for potential κB elements in the proximal promoter region (positions −1338 to + 1) of the rat HO-1 gene. Two putative κB elements, HO-κB-A (−1002 to −993) and HO-κB-B (−284 to −275), were identified. Both elements matched the consensus sequence of the prototypical κB element in eight or nine of 10 nucleotides, respectively (Fig. 2,A). In addition, a macrophage-specific 12-O-tetradecanoyl-phorbol-13-acetate-responsive element (MTE), which has previously been shown to mediate PMA-dependent induction of the human HO-1 gene (33), was identified in the proximal 5′-flanking sequence of the rat HO-1 gene promoter (−140 to −131)(Fig. 2 A).

To further characterize the HO-1 gene activation in response to PMA, reporter gene constructs with serially 5′-deleted HO-1 promoter sequences were transiently transfected into RAW264.7 cells. Loss of the distal HO-κB-A element in pHO-754 and pHO-347 did not cause a major reduction of the PMA-induced luciferase activity. By contrast, deletion of the proximal HO-κB-B site and the MTE in pHO-20 abrogated the PMA-dependent induction, indicating a possible regulatory function of these two elements (Fig. 2 B).

To assess the regulatory capacity of the HO-κB-B element and the MTE, pHO-347 reporter gene constructs, either with or without mutations of the κB-B site and the MTE, were transfected into RAW264.7 cells. Targeted mutations of the HO-κB-B site led to a marked reduction of PMA-dependent induction. By contrast, point mutations of the MTE sequence did not have a major effect on PMA-induced luciferase activity when compared with the wild-type construct (Fig. 2 C). It is also important to note that targeted mutations within the HO-κB-A site of pHO-1338 did not affect PMA-dependent up-regulation of luciferase activity (data not shown). Taken together, the data suggest that the HO-κB-B site, but not the MTE, mediate PMA-dependent induction of rat HO-1 gene expression.

In EMSA studies we examined the binding activity of nuclear proteins from RAW264.7 cells that were treated with either PMA or LPS to an oligonucleotide with the HO-κB-B site. NE from PMA-treated cells showed markedly stronger DNA-binding activity to the HO-κB-B oligonucleotide when compared with NE from control cells. By contrast, NE from cells that were treated with LPS only exhibited minor inducible DNA binding to the HO-κB-B site (Fig. 3,A). The intensity of the band formed by the DNA-protein complex of HO-κB-B with NE from PMA-treated cells was decreased by an excess of unlabeled HO-κB-B oligonucleotide in a dose-dependent manner (Fig. 3,B). Binding of NE to the HO-κB-B site was abolished by an excess of unlabeled oligonucleotides for HO-κB-B and NF-κB, respectively, but not by an excess of an oligonucleotide with a targeted mutation in the HO-κB-B site (Fig. 3,C). Moreover, incubation of the binding reaction with an Ab against the NF-κB subunit p65 caused a reduction of DNA-protein complex formation (Fig. 3 C), suggesting that the HO-κB-B site is a nuclear target for p65.

The NF-κB subunit p65, which is also termed RelA, is a member of the Rel family of proteins and is activated in response to a variety of stimuli (18). To investigate the functional regulatory role of p65 on HO-1 promoter activity, RAW264.7 cells were cotransfected with HO-1 reporter gene constructs and an expression vector for p65. Basal luciferase activity of the reporter gene constructs pHO-1338 and pHO-347, but not that of pHO-347κBmut with a targeted mutation of the HO-κB-B site, was markedly augmented by overexpressed p65. As a control, luciferase activity of a reporter gene plasmid with three copies of the prototypical κB site (pNF-κB) was induced by cotransfected p65 to a similar extent when compared with pHO-347 (Fig. 4). The data indicate that p65-dependent HO-1 activation is mediated via the proximal HO-κB-B site of the rat HO-1 gene promoter.

To substantiate the involvement of NF-κB and its subunit p65 in PMA-dependent HO-1 induction, we examined the HO-1 gene expression in p65^−/− and p65^+/+ MEF. We found that PMA was not able to induce HO-1 expression in p65^−/− MEF that lack p65, whereas PMA induced HO-1 expression in wild-type p65^+/+ MEF (Fig. 5,A, left panel). Importantly, in p65^−/− MEF in which p65 was reconstituted by stable transfection with an expression vector for p65 (27), PMA-dependent induction of HO-1 gene expression was similar to that in wild-type cells (Fig. 5,A). We also determined PMA-dependent regulation of HO-1 promoter activity in p65^−/− and p65^+/+ MEF. Luciferase activity of pHO-1338 was induced in p65^+/+ but not in p65^−/− MEF. Similar to the observations on endogenous HO-1 gene regulation, PMA-dependent induction of HO-1 promoter activity was restored in p65-reconstituted p65^−/− MEF (Fig. 5 B). Taken together, the data confirm that p65 mediates the transcriptional induction of PMA-dependent HO-1 gene expression.

The classical activation pathway of NF-κB by proinflammatory stimuli such as LPS and TNF-α is mediated via IKK-dependent phosphorylation of IκBα at serine 32 and serine 36. Phosphorylation of these regulatory serine residues leads to proteolysis of cytosolic IκBα via the proteasome, after which p65/RelA is translocated into the nucleus (18, 35). To investigate whether IκBα may be involved in PMA-dependent induction of HO-1, we determined the effect of overexpressed dominant negative IκBα on the level of PMA-dependent HO-1 promoter induction. As demonstrated in Fig. 6 A, dominant negative IκBα markedly inhibited up-regulation of HO-1 promoter activity by PMA. Moreover, PMA-dependent induction of the control plasmid pTNF-585, which is known to be regulated via functional κB elements, was inhibited by dominant negative IκBα to a similar extent.

To investigate the potential role of IKK2 for PMA-dependent induction of HO-1 gene expression, we also determined the effect of an overexpressed dominant negative mutant of IKK2 on HO-1 promoter activity. As shown in Fig. 6 B, dominant negative IKK2 did not have an inhibitory effect on PMA-dependent up-regulation of luciferase activity of pHO-1338, but markedly reduced PMA-dependent induction of the control reporter gene construct pTNF-585. No regulatory effect on PMA-dependent HO-1 promoter regulation was observed for the specific pharmacological IKK2 inhibitor SC-514 (data not shown).

To determine the PMA-dependent activation of IκBα, RAW264.7 cells were not only treated with PMA but, for a comparison, also with LPS for various lengths of time, and cell extracts were analyzed for the phosphorylation of IκBα at serine 32. Treatment with LPS caused a marked increase of IκBα phosphorylation at serine 32 for up to 8 h. In contrast, PMA induced IκBα phosphorylation at serine 32 only to a minor extent (Fig. 6,C, upper panel). Because it has previously been shown that IκBα can also be phosphorylated at tyrosine 42 by oxidative stress (36, 37), we also determined the phosphorylation of IκBα at this regulatory residue in response to PMA and LPS. Treatment with PMA caused a rapid and transient IκBα phosphorylation at tyrosine 42 after 15 min (second panel from top). In contrast, treatment with LPS caused a stronger and more persistent IκBα phosphorylation at tyrosine 42 as compared with PMA (Fig. 6 C, second panel from top).

Taken together, the data suggest that PMA-dependent activation of HO-1 is mediated via a nonclassical NF-κB pathway that is independent of IKK2 activity.

PMA has previously been shown to up-regulate the generation of reactive oxygen species (ROS) in monocytes (38). To determine whether ROS as potential secondary messengers would be involved in HO-1 gene induction in our cell culture model of RAW264.7 cells, we examined the effect of the antioxidant NAC on PMA-dependent induction of HO-1. Pretreatment with NAC decreased PMA-dependent up-regulation of HO-1 in a dose-dependent manner (Fig. 7,A). Moreover, we also determined the effect of NAC on the regulation of HO-1 promoter activity by PMA in RAW264.7 cells. Pretreatment with NAC significantly lowered PMA-induced promoter activity of the pHO-1338 reporter gene construct (Fig. 7 B), suggesting that the induction of HO-1 gene expression by PMA is mediated via ROS.

A major target of ROS in monocytes is p38 MAPK (39). Accordingly, phosphorylation of p38 was markedly induced by PMA in our model of RAW264.7 cells (Ref. 40 and data not shown). To investigate the potential role of p38 MAPK for PMA-dependent up-regulation of HO-1 gene expression, we determined the effect of the pharmacological p38 inhibitor SB202190 on PMA-dependent induction of endogenous HO-1 gene expression and promoter activity. Pretreatment with SB202190 markedly decreased PMA-dependent up-regulation of HO-1 gene expression (Fig. 8,A). We also examined the influence of this inhibitor on PMA-dependent HO-1 promoter activity. Pretreatment with SB202190 significantly attenuated the PMA-induced activity of the pHO-1338 reporter gene construct (Fig. 8 B). Taken together, the data suggest that p38 MAPK is involved in PMA-dependent induction of HO-1 gene expression.

CK2 is a stress-activated serine/threonine protein kinase (41, 42) that has previously been shown to be involved in IKK2-independent activation of NF-κB (43). To investigate the potential regulatory role of CK2 for PMA-dependent up-regulation of HO-1 gene expression, we determined the effect of two specific CK2 inhibitors, RFBD and apigenin, on PMA-dependent induction of endogenous HO-1 and promoter activity. Pretreatment with RFBD and apigenin markedly reduced the PMA-dependent up-regulation of HO-1 gene expression in a dose-dependent manner (Fig. 9, A and B). Similar to the observations for endogenous HO-1 gene regulation, pretreatment with RFBD or apigenin attenuated PMA-dependent induction of HO-1 promoter activity (Fig. 9,C). Finally, we also evaluated the effect of cotransfection of an expression vector for CK2α on basal HO-1 promoter activity. Overexpressed CK2α markedly augmented the activity of the HO-1 promoter construct pHO-1338 and that of the control reporter gene plasmid pNF-κB (Fig. 9 D). Thus, the data indicate that CK2 is involved in the regulation of HO-1 gene induction by PMA.

Activation of CK2 by various stress stimuli such as UV light has previously been shown to be regulated via p38 MAPK (43). To determine whether p38 MAPK is required for PMA-dependent CK2 activation, p38 MAPK activity was determined with a fusion plasmid containing the transactivation domain of the transcription factor CHOP and the DNA-binding domain of yeast Gal4 (pFA-CHOP). Transactivation via pFA-CHOP is specifically controlled by p38-dependent phosphorylation of two adjacent regulatory serine residues of the CHOP transactivation domain (44). Treatment with PMA strongly induced pFA-CHOP activity, and pretreatment with the p38 MAPK inhibitor SB202190 lowered PMA-dependent pFA-CHOP-mediated luciferase activity. By contrast, pretreatment with CK2 inhibitors had no effect on PMA-dependent, pFA-CHOP-mediated luciferase activity (Fig. 10), suggesting that p38 MAPK is an upstream kinase of PMA-dependent CK2 activation.

Expression of HO-1 is up-regulated by multiple stress stimuli, and the enzymatic products of this reaction not only have antioxidant cytoprotective effects but also anti-inflammatory functions in various animal models (10, 11, 12, 13, 14, 15). The present study demonstrates the following: 1) HO-1 is induced by PMA in monocytic cells with a regulatory pattern different from that by LPS; 2) HO-1 induction by PMA occurs on the transcriptional level and is mediated via a proximal κB site of the rat HO-1 promoter that is a nuclear target of p65/RelA; and 3) an IKK-independent, atypical NF-κB pathway mediates PMA-dependent induction of HO-1 via activation of p38 MAPK and CK2.

In the present report it is shown that HO-1 gene expression is induced by PMA in monocytic cells, but not in hepatocytes (Fig. 1,A). These findings correspond with a previous report demonstrating that PMA induced HO-1 gene expression in a monocyte-specific manner in human myelomonocytic cells (33). The regulatory mechanism(s) that mediate(s) HO-1 gene induction by PMA appear(s) to be different from that by LPS, because HO-1 gene expression was up-regulated with a different time course by these two compounds (Fig. 1,C). This assumption is also supported by the observation that simultaneous treatment with PMA and LPS induced HO-1 gene expression in an additive manner (Fig. 1,D). Distinct kinetics of gene induction by PMA and LPS were also reported for cyclooxygenase-2 gene expression in monocytes (26, 45). The different kinetics of HO-1 gene induction by PMA and LPS may also correspond with the distinct binding of nuclear proteins to the HO-κB-B site (Fig. 3).

Activation of the TF NF-κB is a major pathway for mediating cell survival during oxidative stress (18). Because multiple stress stimuli that are known to induce HO-1 gene expression also activate NF-κB (19, 20, 21), we hypothesized that the HO-1 promoter may be targeted by this TF. Two potential κB sites within the proximal rat HO-1 gene promoter have been identified that share high sequence identity with the prototypical NF-κB consensus sequence (5′-GGGRNNYYCC-3′) (Fig. 2,A). Although two potential candidate κB elements were found, we demonstrate that PMA-dependent induction of HO-1 promoter activity was only mediated via the proximal HO-κB-B site. In addition, it is shown that this HO-κB-B element was a nuclear target for the NF-κB subunit p65/RelA (Figs. 2–4). Although NF-κB has been implicated in the transcriptional regulation of HO-1 gene expression (17), to our knowledge the rat HO-κB-B site is the first functional RE of the HO-1 gene that is directly targeted by NF-κB. In an earlier report on the human HO-1 gene promoter, a putative κB site has been shown to exhibit in vitro DNA-binding with the recombinant NF-κB subunit p50, but the functionality of this RE has not been examined (46). In other reports, the in vitro binding activity of nuclear proteins to synthetic NF-κB oligonucleotides, which were not necessarily found in the HO-1 gene promoter, have been correlated with the induction of HO-1 gene expression by various identical stimuli (17, 21, 47). It is conceivable that a functional HO-1 κB element may have been overlooked in earlier studies, because human and mouse HO-1 gene promoter regions, which have previously been studied in more detail, exhibit significant sequence differences when compared with the rat HO-1 gene promoter. Sequence alignment of the first 1338 bp of the promoter 5′-flanking region of the rat, mouse, and human HO-1 genes revealed only 47% (rat vs human), 49% (mouse vs human), and 69% (rat vs mouse) sequence similarity, respectively. Remarkably, the murine sequence corresponding to the functional rat HO-κB-B element did not contain a homologous κB sequence, which may suggest species-specific functionality of the HO-κB-B site. Independently, the rat sequence that corresponds with a previously identified PMA-responsive MTE of the human HO-1 promoter (33) was not functional in the context of the rat HO-1 gene promoter (Fig. 2). In conclusion, discrepancies of the promoter structure may explain species-specific differences of HO-1 gene regulation that have also been observed for HO-1 induction by hypoxia or heat shock (17, 48). The present study, however, does not exclude the possibility that PMA-dependent induction of HO-1 is regulated by TF other than NF-κB. In fact, the TF AP-1 and Nrf2 have also been shown to mediate PMA-dependent induction of the mouse HO-1 gene (49, 50).

The classical NF-κB pathway is regulated via IKK-dependent phosphorylation of serine 32 and 36 in the N-terminal region of IκBα in response to a variety of stimuli such as LPS and TNF-α (18, 35). Accordingly, inhibition of PMA-dependent HO-1 promoter activation by overexpressed dominant negative IκBα, but not by dominant negative IKK2 (Fig. 6), indicated that the classical NF-κB pathway does not play a major role for this regulation. This conclusion is consistent with the observation that IκBα is phosphorylated to a minor extent at serine 32 in response to PMA rather than in response to LPS (Fig. 6 C).

Furthermore, it has been proposed that ROS could be involved in the activation of atypical NF-κB regulatory pathways, because PMA could increase intracellular levels of ROS in mononuclear phagocytes (38) and affect the phosphorylation of IκBα at tyrosine 42 (36, 37). In line with this proposal, we showed in the present study that the action of PMA on HO-1 gene expression was abolished by treatment with the antioxidant NAC (Fig. 7), which indicates the involvement of ROS. We also examined whether PMA would affect phosphorylation of IκBα at tyrosine 42 and found a minor and transient level of IκBα phosphorylation at tyrosine 42 after treatment with PMA (Fig. 6,C). This minor up-regulation of IκBα phosphorylation at tyrosine 42 did not correlate with the marked PMA-dependent induction of HO-1 gene expression (Fig. 1) and the inducible DNA binding of nuclear extracts to the HO-κB site in response to PMA (Fig. 3). Therefore, IκBα phosphorylation at tyrosine 42 does not seem to play a major role for PMA-dependent induction of HO-1 gene expression. These latter observations may correspond with a report, in which H₂O₂ has been demonstrated to stimulate NF-κB, but phosphorylation of IκBα at tyrosine 42 per se was not sufficient for NF-κB activation (36). Moreover, PMA as a prototypical activator of PKC, which mimics the intracellular effects of the endogenous mediator diacylglycerol (23), has recently been found to activate PKC in a p38 MAPK-dependent manner (51). This assumption would partially correspond with our data on PMA-dependent induction of HO-1 gene expression by NF-κB via ROS and a p38 MAPK/CK2-dependent pathway (Figs. 7–10). Thus, the present study strongly suggests that this signaling cascade is mediated via an atypical NF-κB pathway that involves phosphorylation of regulatory sites in the C-terminal domain of IκBα. A similar regulatory signaling cascade has previously been shown for the activation of NF-κB by UV light in HeLa cells (43).

The findings of our present study, along with those of a previous report (33) that demonstrate a monocyte-specific induction of HO-1 gene expression by PMA, suggest an important physiological role of PKC-dependent HO-1 up-regulation in monocyte differentiation and/or activation. This assumption would correspond with a recent study in which it has been shown that the activation of PKC is essential for the differentiation of CD14⁺ monocytes into macrophages or dendritic cells (52). Moreover, the proinflammatory mediator LPS has recently been shown to exert its cellular effect via activation of PKC and NF-κB in macrophages, which may also involve the induction of HO-1 gene expression (53).

Inflammatory processes play a major role in the pathogenesis of cancer and cardiovascular disease. Evidence has accumulated that HO-1 has potent anti-inflammatory functions, because genetic HO-1 deficiency causes a chronic inflammatory phenotype and high vulnerability to LPS (7, 9). Anti-inflammatory protection via the induction of HO-1 has initially been described in a model of acute complement-dependent pleurisy (54). More recently, the potential clinical relevance of HO-1 has also been shown in various animal models of inflammatory diseases and organ transplantation, in which targeted overexpression of HO-1 provided efficient protection (10, 11, 12, 13, 14, 15, 55). Finally, it is remarkable that HO-1 gene expression is not only regulated via NF-κB as demonstrated in the present report, but that HO-1 can modulate the activity of NF-κB in various cell types (56, 57, 58).

In conclusion, the present study defines a new regulatory mechanism of HO-1 gene expression by PMA in monocytic cells and may help to further understand the complexity of the gene regulation of this protective gene.

We thank Silke Werth and Annette Zeyer for excellent technical assistance. We also thank Dr. Gordon Duff, Dr. Richard Gaynor, Dr. Rolf Kemler, and Dr. Dieter Schmoll for supply of plasmids and Dr. Hiroyasu Nakano for providing p65-deficient fibroblasts.

The authors have no financial conflict of interest.