Nitric oxide suppresses the expression of Bcl-2 binding protein BNIP3 in hepatocytes.

Nitric oxide (NO) is not only an important signaling molecule, but it also regulates the expression of a number of genes in the liver. We have previously shown that apoptosis in hepatocytes exposed to tumor necrosis factor-alpha and actinomycin D is prevented by NO derived from the inducible nitric-oxide synthase (iNOS), by mechanisms that are both dependent on and independent of modulation of cyclic guanosine monophosphate (cGMP) subsequent to activation of soluble guanylyl cyclase (sGC). We hypothesize that one mechanism by which NO exerts these effects is by regulating the expression of genes involved in apoptosis. We used differential display-polymerase chain reaction to isolate NO-regulated genes in hepatocytes from iNOS knockout mice (to eliminate endogenous inducible NO production). Using this analysis, we identified a NO-suppressed gene fragment homologous with the pro-apoptotic Bcl-2 binding protein BNIP3. Northern analysis confirmed the NO-dependent suppression of BNIP3 in cultured cells. Similarly, the NO donor S-nitroso-N-acetyl-dl-penicillamine (1-1000 microm) down-regulated the expression of BNIP3 in both iNOS knockout and wild-type hepatocytes. This effect of NO was reversed by the sGC inhibitor 1H-(1,2,4)-oxadiazole[4,3-a]quinoxalon-1-one (ODQ),suggesting the involvement of the sGC/cGMP pathway in the modulation of BNIP3 by NO. We propose that suppression of BNIP3 expression is one sGC/cGMP-dependent mechanism by which NO might affect the process of hepatocyte apoptosis.

Despite major advances in the information available on the mediators involved in the septic response, relatively little is known about the cellular and molecular events that either prevent or promote organ failure. The liver is a primary response organ during sepsis and one of the last organs to fail in multiple organ failure (1). One emerging theory regarding the pathology of liver failure in sepsis implicates apoptosis secondary to release of inflammatory mediator (2,3). It is now clear that widespread up-regulation of the inducible nitric-oxide (NO)1 synthase (iNOS; NOS2) 1 is part of the sepsis syndrome.
Importantly, the expression of iNOS by Kupffer cells and hepatocytes in the liver appears to play a central role in the pathobiology of sepsis (4); recent studies from our laboratory have shown that iNOS exerts an anti-apoptotic effect on the liver during experimental endotoxemia (2,5,6). Several mechanisms contribute to the anti-apoptotic effect of NO on hepatocytes, including the suppression of activity of multiple caspases by nitrosylation (7,8), inhibition of Bcl-2 cleavage and cytochrome c release (9), decreased mitochondrial formation of reactive oxygen intermediates (10), and suppression of the induction of pro-apoptotic genes (11). Some of these actions are cGMP-dependent and cGMP analogues suppress caspase activation, loss of mitochondrial membrane potential, and cytochrome c release (12).
We hypothesized that some of these anti-apoptotic effects of NO are mediated by the modulation of gene expression in hepatocytes. Nitric oxide is an important regulator of a wide variety of cellular functions (13), and affects the expression of genes in multiple cell types (14,15). We identified genes regulated by NO in mouse hepatocytes using the technique of differential display polymerase chain reaction (DD-PCR) (16). Because iNOS may be induced during manipulation of hepatocytes (17,18), we decided to use cells that are incapable of expressing iNOS and thus have never been exposed to elevated concentrations of NO. To carry this out, we used hepatocytes derived from iNOS null (iNOSϪ/Ϫ) mice (19). Because iNOS is usually induced in hepatocytes by treatment with strong proinflammatory co-stimuli (interferon-␥, interleukin-1, and tumor necrosis factor-␣ (TNF-␣)) (20), stimuli that by themselves can modulate gene expression, we instead infected hepatocytes with an adenoviral vector expressing the human iNOS gene (Ad-iNOS) that has been developed and characterized in our laboratory (21).
Among the transcripts identified by this analysis was the pro-apoptotic gene BNIP3 (formerly known as Nip3 (22)), which belongs to the Bcl-2 family and has been involved in cell death. Our observation that BNIP3 expression is markedly suppressed following iNOS expression and NO production suggests that this could be another mechanism by which NO prevents apoptosis. was synthesized as described previously (23), stored as a powder in the dark, and checked for stoichiometric S-nitrosothiol content by the method of Saville (24). Stock solutions were prepared at 100 mM in ethanol, and further dilutions were made in culture medium. N-Iminoethyl-L-ornithine (L-NIO) was purchased from Alexis Co. (San Diego, CA). AmpliTaq DNA polymerase, dNTPs, and PCR reagents were from PerkinElmer Life Sciences (Foster City, CA). Mouse recombinant TNF-␣ was obtained from R&D Systems (Minneapolis, MN). Formaldehyde and formamide were from Fluka Chemica-Biochemica (Ronkonkoma, NY); plasmid DNA preparation columns were from Qiagen (Chatsworth, CA); and tissue culture dishes were from Corning Glass Works (Corning, NY). Unless indicated otherwise, all other chemicals and proteins were purchased from Sigma Chemical Co. (St. Louis, MO).
Hepatocyte Isolation and Culture-All procedures involving animals were approved by the Animal Care and Use Committee of the University of Pittsburgh. Hepatocytes were harvested from 6-week-old iNOS knockout (iNOSϪ/Ϫ) mice prepared and bred as described by MacMicking et al. (19). As another source of hepatocytes male Sprague-Dawley rats (Harlan Sprague-Dawley) were used. Hepatocytes were isolated by collagenase perfusion using the method of Seglen (25) and purified to Ͼ98% purity by repeated centrifugation at 50 ϫ g, followed by further purification over 30% Percoll. Viability at time of plating was checked by trypan blue exclusion. Highly purified hepatocytes (Ͼ98% purity and Ͼ98% viability by trypan blue exclusion) were suspended in Williams medium E containing 10% heat-inactivated calf serum supplemented with 15 mM HEPES (pH 7.4), 1 M insulin, 2 mM L-glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin. The cells were plated on cell culture dishes at a density of 3 ϫ 10 6 cells/60-mm dish and placed in incubator at 37°C in 5% CO 2 /95% air. After overnight preculture, cells were infected with an adenoviral vector containing the gene for LacZ or human iNOS (Ad-iNOS) (26) and allowed to express these genes for 6 and 20 h in the presence or absence of the NOS inhibitors L-NIO (1 mM) or L-NMMA (0.5 mM). For experiments with exogenous NO, after the preculture period, the old medium was removed and cells were further incubated with fresh media containing 5% calf serum with or without test compounds for 20 -24 h.
Adenoviral iNOS Gene Transfer-Modified adenoviral vectors carrying the human iNOS or bacterial ␤-galactosidase (LacZ) cDNA were prepared as described previously (26). After overnight preculture, iNOS knockout hepatocytes (3 ϫ 10 6 /6-cm plate) were washed with Hanks' buffered saline and transfected with adenoviral vector containing either the human iNOS (Ad-iNOS) or LacZ (Ad-LacZ) cDNA at multiplicity of infection of 1 in a volume of 2 ml of Opti-MEM. Following a 2-h infection, the medium was changed to fresh Williams medium E containing 5% calf serum, and the cells were allowed to express the transfected genes for 6 and 20 h in the presence or absence of the NOS inhibitors L-NIO (1 mM) or L-NMMA (0.5 mM) as indicated. The total RNA was isolated at 6 or 20 -24 h after transfection.
RNA Isolation-Total RNA was isolated from the cultured hepatocytes using an UltraSpec RNA isolation reagent from Biotecx Laboratories, Inc. (Houston, TX). Briefly, the cells were lysed directly in a culture dish by adding UltraSpec (1 ml/3.5-cm Petri dish) and passing the cell lysate several times through a pipette. After extraction with chloroform (0.2 ml/ml of UltraSpec), the total RNA was precipitated from the aqueous phase by addition of isopropanol, washed with 75% ethanol, and solubilized in diethyl pyrocarbonate (DEPO)-treated water.
Differential Display of mRNA from iNOSϪ/Ϫ Hepatocytes-Cells were infected with adenoviral vectors and cultured for 6 and 20 -24 h as described above. Differential Display (DD) reactions were performed using cultured hepatocytes total RNA with primers from the HIERO-GLYPH mRNA profile system (Genomyx Corp., Foster City, CA). Each of the HIEROGLYPH kits contains ten 3Ј-oligo(dT)-anchoring primers with a set of four kit-specific 5Ј arbitrary primers. HeLa control RNA for verification of reverse transcriptase and DD-PCR parameters is also included. Protocols for reverse transcriptase, DD-PCR, electrophoresis, cDNA fragment excision, reamplification, and sequencing are provided in the Hieroglyph manual from the manufacturer. All reactions were run in duplicate to verify reproducibility of cDNA fragment patterns. Following high resolution electrophoresis on a genomyxLR DNA sequencer, the cDNA bands differentially displayed were excised from the gel, extracted with elution buffer, and re-amplified by PCR using the same set of primers. The re-amplified PCR products were gel-purified and used as probes in Northern blots. Only those that continued to show differential expression patterns between control and infected cells were subcloned and sequenced prior to Northern analysis.
Northern Blot Analysis-Total RNA was isolated from the cultured hepatocytes as described above. The RNA (20 g/lane) was electrophoresed on 0.9% agarose gel containing 12.3 M formaldehyde and transferred to nylon membranes (GeneScreen, PerkinElmer Life Sciences) by vacuum blotting. The membranes were pre-hybridized for 3-4 h at 43°C and hybridized with [ 32 P]dCTP-labeled probe (10 6 cpm/ml) at 43°C. Membranes were then washed 3ϫ with SSC/SDS at 53°C before exposure. Membranes were then stripped and probed for 18 S RNA to assess RNA loading. Northern blot analysis of iNOS mRNA levels in cultured hepatocytes was carried out using a murine iNOS cDNA probe as described previously (27). (The probe to inducible nitric-oxide synthase used was a 2.7-kilobase cDNA obtained by NotI digestion from a mouse macrophage cDNA clone (27).) Northern blots were hybridized with labeled probes previously sequenced. Radioactive membranes were quantified with storage phosphor screens (PhosphorImager, Molecular Dynamics), and the relative amount of mRNA is presented as the ratio of mRNA to 18 S RNA. Regulated expression of the BNIP3 gene was examined by Northern analysis using a probe obtained as described above.
Western Blot Analysis-Cultured hepatocytes (5 ϫ 10 6 cells/100-mm Petri dish) were washed twice with ice-cold phosphate-buffered saline and resuspended in 1 ml of ice-cold lysis buffer containing 2 mM Tris-HCl buffer (pH 7.5), 15 mM NaCl, EDTA and EGTA (both 100 M), 0.1% Triton X-100, 250 M sodium pyrophosphate, 100 M ␤-glycerolphosphate, 100 M Na 3 VO 4 , and the protease inhibitors leupeptin (0.1 g/ ml) and phenylmethylsulfonyl fluoride (1 mM). After 5-min incubation on ice, the cells were scraped off the dish and transferred to microcentrifuge tubes. Following 20-to 30-min incubation at 4°C, cell debris was removed by centrifugation at 13,000 rpm for 10 -15 min, and the supernatant was used as cell lysate and stored at Ϫ80°C when necessary. An aliquot was used to determine protein concentration using the BCA Protein Assay kit from Pierce (Rockford, IL) with bovine serum albumin as standard. Protein samples (30 g) were separated on 15% SDSpolyacrylamide gels, and the gels were electroblotted onto nitrocellulose membranes. Immunodetection of BNIP3 protein was done using the ECFTM Western blotting kit (Amersham Pharmacia Biotech, Arlington Heights, IL) and a rabbit polyclonal antibody against human BNIP3 (1:7,500 dilution). Instructions for the kit were provided by the supplier. The rabbit polyclonal anti-BNIP3 antibody and the BNIP3-transfected MCF-7 lysate used as positive control were kindly provided by Drs. A. Greensberg and D. Dubik from Manitoba Institute of Cell Biology (Winnipeg, Canada).
Viability Assay-Cell viability was determined by the crystal violet method as described previously (12). Crystal violet is taken up by adherent cells on a culture dish, and color intensity is therefore proportional to cell number. Briefly, cells were stained with 0.5% crystal violet in 30% ethanol/3% formaldehyde for 10 min at room temperature. Plates were washed six times with tap water. After drying, cells were lysed with 1% SDS solution, and dye uptake was measured at 550 nm using a 96-well microplate reader. Cell viability was calculated from relative dye intensity and presented as percentages relative to control samples.
Nitrite Assay-Nitrite concentrations in medium were assessed using the Griess reaction by mixing equal volumes (100 l) of cell supernatants with the Griess reagent (1% sulfanilamide/0.1% naphthylethylenediamine dihydrochloride/2.5% H 3 PO 4 ). After incubation at room temperature for 10 min, absorbance was measured spectrophotometrically at 550 nm using a kinetic microplate reader (Vmax, Molecular Devices, Sunnyvale, CA). Nitrite concentration was determined from a calibration curve using sodium nitrite as standard.
Statistical Analysis-Data are presented as mean Ϯ S.E. of three separate experiments. Comparisons were performed using Student's t test or one-way analysis of variance followed by Dunnett's post-hoc test where appropriate. Differences were considered significant at the 95% confidence interval (p Ͻ 0.05).

RESULTS
Infection of Primary iNOSϪ/Ϫ Hepatocytes with Ad-iNOS-To expose NO-naïve cells to iNOS activity, hepatocytes from iNOSϪ/Ϫ mice were isolated for primary cell culture and infected with adenoviral vectors carrying either iNOS (Ad-iNOS) or LacZ (Ad-LacZ). After 20 h, NO 2 -levels in the culture medium of cells infected with Ad-iNOS were 80 Ϯ 9 M. No nitrite was detected in culture medium of cells infected with Ad-LacZ. Northern analysis confirmed the presence of human iNOS mRNA after infection with Ad-iNOS but not Ad-LacZ (Fig. 1).
Identification of Genes Regulated following Ad-iNOS Infection in iNOSϪ/Ϫ Hepatocytes by Differential Display-We sought to identify genes that are differentially expressed as a consequence of exposure to iNOS activity. To this end, we used the differential display technique to compare patterns of mRNA expression from control, Ad-LacZ-infected, and Ad-iNOS-infected cells. Messenger RNA was extracted from the cultured cells and used as the template for cDNA synthesis with oligo(dT)-primed reverse transcription and subsequent differential display analysis. This analysis was performed using a series of 20 arbitrary primer combinations, amplifying 10% of the total cDNA. Differential display analysis revealed TABLE I cDNA fragments corresponding to mRNAs differentially expressed by NO in iNOSϪ/Ϫ mouse hepatocytes transfected with human Ad-iNOS Effect of Ad-iNOS infection: "1" indicates up-regulation, "2" indicates down-regulation, "NB (yes/n.d.)" indicates whether Northern Blotting analysis was performed or not. the presence of several genes that were either induced or reduced by Ad-iNOS infection of mouse hepatocytes (Table I). Thirty-six cDNA fragments that reproducibly displayed a different abundance pattern between hepatocytes infected with Ad-LacZ and cells infected with Ad-iNOS were identified, isolated, cloned, and sequenced. Of these, 26 known genes were identified following a GenBank TM /BLAST search.
Identification of BNIP3 as a Transcript Differentially Expressed in Hepatocytes from iNOSϪ/Ϫ Mice Infected with Ad-iNOS-The primer combination T7-T12-GG and M13-GCTAG-CATGG produced a gene fragment suppressed by NO, which was sequenced and identified as the Bcl-2 binding protein BNIP3 (Fig. 2). The suppression of BNIP3 in Ad-iNOS-infected cells was specific to the presence of iNOS, because infection with Ad-LacZ did not affect the expression of this gene. The differential expression of other mRNAs, including the BNIP3 was confirmed using their cDNAs as probes in Northern analyses (data not shown, except for BNIP3).
Cytotoxic Effect of TNF-␣ plus Actinomycin D Is Associated with Changes in BNIP3 mRNA Expression-Treatment of wildtype hepatocytes with TNF-␣ plus actinomycin D, a stimulus combination that induces hepatocyte apoptosis (7,12), led to a concentration-dependent increase in BNIP3 mRNA. This elevated expression was associated with a decrease in cell viability (Fig. 3).
Down-regulation of BNIP3 mRNA Expression in Hepatocytes from iNOSϪ/Ϫ Mice Infected with Ad-iNOS-One of the principal limitations of the differential display technique is the presence of false positives and the post-differential display issue of discriminating between false positives and the truly differentially expressed mRNAs. The identification of a differ-

FIG. 2. Differential display analysis of mRNA from iNOS؊/؊ mouse hepatocytes.
Hepatocytes were infected either with Ad-LacZ (control) or with Ad-iNOS. After 6 and 24 h culture, the differential display reactions were performed using the total RNA as described under "Experimental Procedures." The figure shows a section of the high resolution DD-gel showing the cDNA bands differentially expressed that show homology with the BNIP3 gene.

FIG. 3. Effect of TNF-␣ plus actinomycin D on BNIP3 mRNA expression and cell viability in wild-type mouse hepatocytes.
Hepatocytes were cultured in medium containing TNF-␣ (2000 units/ ml) and actinomycin D (5-200 ng/ml) for 6 or 24 h. After 6 h, Northern blotting was performed as described under "Experimental Procedures" with total RNA (20 g per lane). The same membrane was re-hybridized with 18 S oligonucleotide probe as a loading control. Densitometric quantification of mRNA is shown in the graph. The ratio of signal intensity for BNIP3 to that of 18 S is defined as 1 in control cells. Viability was measured by the crystal violet method after 24 h. A representative Northern blot with subcloned re-amplified BNIP3 cDNA probe and densitometric analysis is shown below each bar. entially expressed band therefore requires confirmation by other methods such as Northern blotting or RNase protection analysis. Accordingly, hepatocytes from iNOSϪ/Ϫ mice were isolated for primary cell culture and were infected with either Ad-iNOS or Ad-LacZ for 6 or 20 h, in the presence or absence of the iNOS inhibitor L-NMMA. Northern blot analysis of the mRNA from iNOSϪ/Ϫ hepatocytes was performed using a probe obtained from the isolated differential display fragment corresponding to BNIP3 (Fig. 2). Northern analysis showed a reduction in BNIP3 mRNA expression induced by Ad-iNOS after 24 h (Fig. 4). The mRNA expression profiles between Ad-LacZ-and Ad-iNOS-infected cells match those obtained from differential display analysis, confirming the effect of iNOS on BNIP3 (see Fig. 2). This analysis further demonstrated that the suppression of BNIP3 required the production of NO, because treatment of iNOSϪ/Ϫ hepatocytes with Ad-iNOS in the presence of the NOS inhibitors L-NMMA (Fig. 4) or L-NIO (data not shown) significantly reduced the inhibition of BNIP3 expression.

Reduction of BNIP3 mRNA by Ad-iNOS Can Be Mimicked by the NO Donor SNAP and Requires NO Production-To further
demonstrate that iNOS-derived NO suppressed the expression of BNIP3, hepatocytes from iNOSϪ/Ϫ mice were isolated for primary cell culture and were treated with either a fresh solution of the NO donor S-nitroso-N-acetyl-DL-penicillamine (SNAP) or a solution of oxidized (decomposed) SNAP (oxSNAP) for 24 h. Treatment of iNOSϪ/Ϫ hepatocytes with SNAP showed a concentration-dependent reduction in BNIP3 mRNA expression (Fig. 5). Control treatment with sodium nitrite (Fig.  6) or with oxSNAP (not shown) had no significant effect on BNIP3 expression, indicating that NO was the effector molecule in the observed reduction in mRNA expression. Similarly, the nitrosothiol S-nitrosoglutathione had an inhibitory effect on BNIP3 mRNA expression under the same experimental conditions (Fig. 7), further confirming the effect of NO on BNIP3 mRNA expression.
Effect of SNAP on BNIP3 Protein Expression in iNOSϪ/Ϫ Hepatocytes-We next investigated the effect of NO on BNIP3 FIG. 4. Northern blot analysis of control and infected iNOS؊/؊ mouse hepatocytes with subcloned re-amplified BNIP3 cDNA probe. Cells were infected with either Ad-LacZ or Ad-iNOS. After 6 or 24 h, Northern blotting was performed as described under "Experimental Procedures" with total RNA (20 g per lane). The membranes were re-hybridized with 18 S oligonucleotide probe as a loading control. Densitometric quantification of mRNA is shown in the graph. In non-infected cells (control) the ratio of signal intensity for BNIP3 to that of 18 S is defined as 1 for densitometric analysis. This experiment is representative of three separate experiments yielding similar results.
protein expression by Western analysis. When iNOSϪ/Ϫ hepatocytes were treated with SNAP (100 M) for 3-72 h, a timedependent reduction in BNIP3 protein was observed using Western analysis, which suggests that NO is also capable of reducing BNIP3 protein synthesis (Fig. 8). One reported anomaly regarding BNIP3 is that, despite its calculated molecular mass of 21.54 kDa, when transiently expressed it migrates on FIG. 5. Effect of exogenous NO on BNIP3 mRNA expression in iNOS؊/؊ mouse hepatocytes. iNOSϪ/Ϫ hepatocytes were cultured in medium containing SNAP (0 -1000 M) for 20 -24 h. Northern blotting was performed using subcloned re-amplified BNIP3 cDNA as a probe as described under "Experimental Procedures" with total RNA (20 g per lane). The same membrane was re-hybridized with 18 S oligonucleotide probe as a loading control. A representative Northern blot, with each condition performed in duplicate as part of one experiment is shown, which was then subjected to densitometric analysis. The ratio of signal intensity for BNIP3 to that of 18 S is defined as 1 in control cells. Results are the mean Ϯ S.E. of three separate experiments (*, p Ͻ 0.05 versus control).
FIG. 6. Effect of sodium nitrite on BNIP3 mRNA expression in mouse hepatocytes. iNOSϪ/Ϫ hepatocytes were cultured in medium containing sodium nitrite (0 -10 M) for 20 -24 h. Northern blotting was performed using subcloned re-amplified BNIP3 cDNA as a probe as described under "Experimental Procedures" with total RNA (20 g per lane). The same membrane was re-hybridized with 18 S oligonucleotide probe as a loading control. Densitometric quantification of mRNA is shown in the graph. For the densitometric analysis of the blot the ratio of signal intensity for BNIP3 to that of 18 S is defined as 1 in control cells. This experiment is representative of three. FIG. 7. Effect of exogenous NO on BNIP3 mRNA expression in mouse hepatocytes. iNOSϪ/Ϫ hepatocytes were cultured in medium containing S-nitrosoglutathione (0 -500 M) for 20 -24 h. Northern blotting was performed using subcloned re-amplified BNIP3 cDNA as a probe as described under "Experimental Procedures" with total RNA (20 g per lane). The same membrane was re-hybridized with 18 S oligonucleotide probe as a loading control. Densitometric quantification of mRNA is shown in the graph. The ratio of signal intensity for BNIP3 to that of 18 S is defined as 1.
SDS-polyacrylamide gel electrophoresis as a major band of 60 kDa and a minor band of 30 kDa (29). In our studies, BNIP-3 also appeared as a 60-kDa protein.

SNAP Down-regulates BNIP3 mRNA Expression in Wildtype Hepatocytes in a Manner Dependent on Soluble Guanylate
Cyclase-We wished to confirm whether the effect of NO on BNIP3 can be generalized to wild-type hepatocytes. To investigate this issue, hepatocytes from wild-type (iNOSϩ/ϩ) mice were treated either with SNAP or with oxSNAP for 24 h. As in the case with the iNOSϪ/Ϫ cells, treatment of wild-type hepatocytes with SNAP, but not control treatment with sodium nitrite or oxSNAP (not shown) led to decreased BNIP3 mRNA expression (Fig. 9). We also examined whether the effect of NO is mediated by cGMP by comparing the BNIP3 expression in the presence or absence of the sGC inhibitor 1H-(1,2,4)-oxadiazole[4,3-a]quinoxalon-1-one (ODQ). The inhibition of the enzymatic activity of sGC with ODQ prevented the reduction of BNIP3 by SNAP (Fig. 8), which suggests that production of cGMP by NO is involved in the reduction of BNIP3. Control treatment with the inhibitor had no significant effect on BNIP3 mRNA expression. DISCUSSION Nitric oxide has an important regulatory role in the liver in infection and inflammation (2, 4, 30 -32). Among the several functions of NO in multiple cell types is the regulation of gene expression (14,15). As part of our analysis of iNOS-mediated changes in hepatocyte gene expression using differential display, we carried a detailed evaluation of one of the iNOSsuppressed genes. Thirty-six cDNA fragments that reproducibly displayed a different abundance pattern between hepatocytes infected with Ad-LacZ and cells infected with Ad-iNOS were identified, isolated, cloned, and sequenced. From this analysis, we have reported that infection with Ad-iNOS increases the expression of cytochrome P450 2E1 in iNOS-null hepatocytes in the absence of inflammatory stimuli (33). Among the other genes regulated by NO and detected by the same differential display analysis was BNIP3, a gene known to promote apoptosis (29); this was demonstrated for the first time in hepatocytes in the present study. We have shown that NO suppresses BNIP3 independently of any other inflammatory stimuli, an effect largely mediated by the cGMP pathway.
Infection with Ad-LacZ had no effect on BNIP3 expression, suggesting that the effect was specific to iNOS expression.
BNIP3 localizes to mitochondria and other cytoplasmic mem- Northern blotting was performed using subcloned re-amplified BNIP3 cDNA as a probe as described under "Experimental Procedures" with total RNA (20 g per lane). The same membrane was re-hybridized with 18 S oligonucleotide probe as a loading control. A representative Northern blot with subcloned re-amplified BNIP3 cDNA probe is shown. The lanes corresponding to SNAP and SNAP plus ODQ were performed in duplicate as part of one experiment. In control cells the ratio of signal intensity for BNIP3 to that of 18 S is defined as 1 for densitometric analysis. Results are the mean Ϯ S.E. of three separate experiments (*, p Ͻ 0.05 versus control; #, p Ͻ 0.001 versus SNAP; **, p Ͻ 0.001 versus control and versus SNAP ϩ ODQ). brane structures (34) and is found widely expressed in a large number of mouse and human tissues (29). In the liver, BNIP3 is expressed as a major transcript of 2.5 kb and as a minor transcript of 1.7 kb (29). BNIP3 appears to have an overall resemblance to several BH3-containing Bcl-2 family proteins such as BIK, BID, HRK, and BAD, in which the BH3 domain plays an important role in eliciting apoptosis (35). Indeed, BNIP3 sensitizes Rat-1 cells to apoptosis by Granzyme B and topoisomerase inhibitors and overcomes Bcl-2 suppression of apoptosis (29).
Although both apoptotic and necrotic cell death mechanisms have been attributed to BNIP3 (22,35,36), the exact pathway(s) by which BNIP3 expression induces cell damage remains unknown. The BNIP3 BH3 domain was as efficient as the Bax BH3 domain in eliciting apoptosis, probably through heterodimer formation with anti-apoptotic proteins (35). Expression of BNIP3 (both mRNA and protein) is induced and related to hypoxia-induced apoptosis in CHO-K1 (Chinese hamster ovary), CV-1 (monkey kidney), Rat-1 (rat fibroblast), PAM212 (human epithelial), HepG2 (human hepatocellular carcinoma), and ECV-304 (human bladder carcinoma) cell lines (37). BNIP3 and Nix (a BNIP3 homologue sharing both structural and functional similarity (38), also known as BNIP3L (39), BNIP3␣ (40), and B5 (41)) are the only members of the Bcl-2 family of apoptotic factors induced in response to hypoxia (37). However, BNIP-3 may also mediate cell death via pathways other than apoptosis. Apoptosis induced by BNIP3 was only partially suppressed in cell lines expressing Bcl-2 (29,35). Ray et al. (36) demonstrated recently that BNIP3 heterodimerizes with Bcl-2/Bcl-xl and induces cell death independent of a BH3 domain at both mitochondrial and non-mitochondrial sites. Furthermore, another recent study (22) reported that BNIP3 induces necrosis-like cell death following integration into the mitochondrial outer membrane and rapid mitochondrial PT pore opening. This mechanism was independent of caspases, and the Apaf-1/cytochrome c mitochondrial pathway is accompanied by the suppression of the proton electrochemical gradient and increased ROS production and occurs before the appearance of nuclear damage (22). In the present study, treatment with a pro-apoptotic stimulus (TNF-␣ in combination with actinomycin D) (7, 12) resulted in increased expression of BNIP3. Clearly, the mechanisms by which BNIP3 exerts its pro-apoptotic and/or necrotic activity warrant further investigation.
Many of the cellular actions of NO are mediated by activation of sGC followed by increase in cGMP production (42). Regulation of gene expression by NO can be both cGMP-dependent and independent. However, when it is dependent on the GC/ cGMP pathway it is not always clear which of several cGMP target proteins mediates the effects on gene expression (43). We have demonstrated previously that the anti-apoptotic effect of NO on hepatocytes occurs both via cGMP-dependent and -independent mechanisms (7,12). In this study, the comparison of the BNIP3 expression in the presence or absence of the sGC inhibitor ODQ suggested that cGMP production stimulated by NO leads to suppression of BNIP3 mRNA. This effect suggests that, in this case, iNOS-derived NO is mediating a direct effect via interaction with sGC, rather than an indirect effect mediated by reactive nitrogen oxide species (44).
The role of nitric oxide (NO) in apoptosis is both diverse and complex, because NO can be considered either a pro-apoptotic or an anti-apoptotic molecule depending on a number of factors (6,45). The effect of NO on BNIP-3 described herein appears to be part of a broad-spectrum anti-apoptotic effect of NO on hepatocytes (6). In a previous study using DNA microarray technology, 2 we found that the anti-apoptotic genes BAG-1 (46) and HSP70 (47,48) were induced by iNOS in mouse hepatocytes. We also observed a reduction in certain pro-apoptotic genes following infection with Ad-iNOS. These genes include sox-4 (49), the calpain gene (50), the R-ras gene (51), the ICElike cysteine protease (Lice) gene (28), and the RIP gene. 2 Further study is required to determine the relative contribution of these gene-modulatory effects of NO to the overall antiapoptotic effect of this free radical.
In conclusion, our results demonstrate that endogenous as well as exogenous NO down-regulate the expression of BNIP3 in both NO naïve and wild-type hepatocytes. The sGC/cGMP pathway largely mediates this inhibitory effect. Because BNIP3 is a pro-apoptotic protein, its suppression by NO may contribute to the anti-apoptotic activity of NO in hepatocytes. We are currently investigating in more detail the expression of BNIP3 in other cell types and its relationship to cell death induced by different apoptosis and cell death inducers.