Effect of Peroxisome Proliferator-Activated Receptor α Activation on Leukotriene B₄ Metabolism in Isolated Rat Hepatocytes

Experimental Procedures

Materials.

The following drugs and chemicals were obtained from the sources indicated: LTB₄ and PGB₂ (Cayman Chemical, Ann Arbor, MI); tritium-labeled LTB₄([5,6,8,9,11,12,14,15-³H]LTB₄, 195 Ci/mmol) (PerkinElmer Life Science Products, Boston, MA); PPARα agonist WY-14,643 (BIOMOL Research Laboratories, Plymouth, PA); 9-cis retinoic acid (Sigma Chemical, St. Louis, MO); Hanks' balanced salt solution (Invitrogen, Carlsbad, CA); modified Waymouth's medium (WMB-002, 0.12 mM ornithine, without phenol red) (Specialty Media, Lavalette, NJ); and Ultima-Flo M Scintillation Cocktail used for radioactivity detection (Packard Instrument Co., Meriden, CT). All solvents used were HPLC grade or higher and obtained from Fisher Scientific (Fair Lawn, NJ).

Hepatocyte Isolation and Incubation.

Rat hepatocytes were harvested from nonfasted male Sprague-Dawley rats (Harlan Bioproducts for Science, Indianapolis, IN) under pentobarbital sodium anesthesia by using a collagenase perfusion procedure as previously described (Berry and Friend, 1969). Cell viability was greater than 95% as determined by trypan blue exclusion. Isolated hepatocytes (3.2 × 10⁶ cells) were suspended in 3 ml of modified Waymouth's medium supplemented with 142 μM ascorbic acid, antibiotics (penicillin, 100 ng/ml and streptomycin, 100 U/ml), and 100 nM insulin, and plated on Matrigel-coated 60-mm Petri dishes prepared as previously described (Schuetz et al., 1988). Cells were placed in a 37°C, 5% CO₂ incubator and allowed to attach to the plates for 3 h after which the medium was aspirated and replaced with treatment medium consisting of the modified Waymouth's medium containing 1 nM insulin and WY-14,643 (100 μM), LTB₄ (10 μM), or their diluent (DMSO). Cells were incubated with the treatment media at 37°C for the specified times. For incubation periods greater than 24 h, treatment media were replaced with fresh media after 24 h. After the incubation period cells were processed as described below.

LTB₄ Metabolism.

For LTB₄metabolism studies hepatocytes were washed twice with Hanks' balanced salt solution and then incubated at 37°C with 1 ml of medium containing 10 μM LTB₄ and 0.5 μCi [³H]LTB₄ as tracer. After 30 min, cells were lysed by the addition of 4 ml of ice-cold ethanol. PGB₂ (1.3 nmol) was added to each sample as an internal standard. Samples were kept at 0°C overnight to precipitate proteins, which were removed by centrifugation. Supernatants were decanted and evaporated to near dryness under vacuum. Samples were reconstituted in 10% methanol and LTB₄metabolites were purified by solid phase extraction as previously described (Wheelan and Murphy, 1997). Purified samples in methanol were stored at −20°C until analyzed. Just prior to analysis, samples were dried and reconstituted in initial HPLC conditions 85% solvent A (water/0.05% acetic acid, pH 5.0, adjusted with ammonium hydroxide), 15% solvent B (acetonitrile/methanol, 65:35), and an aliquot injected onto an Ultremex column (2.0 × 150 mm, C₁₈; Phenomenex, Torrance, CA). A linear gradient from 15 to 100% B over 60 min (flow rate 200 μl/min) was used to elute metabolites from the column. UV absorbance was monitored on an HP1040A photodiode array detector (Hewlett Packard, Palo Alto, CA) at a wavelength of 270 nm. Eluent from the UV detector was either introduced into a radioactivity monitor (Flow One/Beta Radiomatic, Tampa, FL) for quantification of the radioactive LTB₄ metabolites or collected into 1-min fractions for off-line confirmation of identity of the metabolites by mass spectrometry. The quantity of each metabolite was calculated from the area of each radioactive peak and normalized to the area of the UV peak of the internal standard PGB₂.

Mass Spectrometric Analysis of LTB₄ Metabolites.

HPLC fractions were directly analyzed by negative ion electrospray mass spectrometry on a Sciex API III⁺ triple mass spectrometer quadrupole (PE-Sciex, Thornhill, ON, Canada). Mass spectra were obtained by scanning from m/z 100 to 700 (MS) or m/z 50 to 400 (MS/MS) by using an ion spray voltage of −3400 V, an orifice voltage of −60 V, and a scan rate of 3.04 s/scan. Collision-induced decomposition of the metabolites for MS/MS analysis was performed at a collision offset energy of 20 eV and a collision gas (argon) thickness of 200 × 10¹² molecules/cm².

RNA Isolation and Northern Blotting.

Total RNA was extracted from hepatocytes by using an SV Total RNA Isolation System (Promega, Madison, WI) as previously described (Shiao et al., 2000). The RNA was separated on a 1.2% denaturing agarose-formaldehyde gel in borate buffer at 140 V for 4 h. RNA was transferred overnight to Hybond-N+ nitrocellulose (Amersham Pharmacia Biotech, Piscataway, NJ) with high-efficiency transfer solution by capillary action. RNA was subsequently fixed to the membrane by UV crosslinking. The acyl-CoA oxidase cDNA probe (provided by Dr. S. Clarke, University of Texas, Austin, TX) and CYP 4A1 cDNA probe (provided by Dr. Gordon Gibson, University of Surrey, Guildford, UK) were labeled with [³²P]dCTP (Amersham Pharmacia Biotech) by using the DECprime II (Ambion, Austin, TX) DNA labeling kit. Membranes were hybridized overnight at 63°C and unbound probe was removed by two washes in 2% sodium chloride-sodium citrate-0.1% SDS followed by two washes in 0.1% sodium chloride-sodium citrate-0.1% SDS for 30 min each. Membranes were exposed to film with intensifying screen at −70°C overnight, and the autoradiograms of the blots were quantitated using an imaging densitometer (Bio-Rad, Hercules, CA). In addition, blots were probed with a labeled 18S cDNA probe (Ambion) prepared as described above. The relative density of acyl-CoA oxidase and CYP4A1 mRNAs was normalized to the 18S RNA.

Microsome Preparation and Western Blotting.

Hepatocytes from two culture plates were scraped into 1-ml matrisperse (Collaborative Biomedical Products, Bedford, MA) and incubated for 30 min on ice while mixing gently to extricate hepatocytes from Matrigel. After centrifugation of the suspension, the supernatant was aspirated, leaving the cell pellet. Microsomes were isolated from the hepatocytes after the method by Liddle et al. (1992). A small aliquot of the final microsome suspension was analyzed using the bicinchoninic acid protein assay (Pierce Chemical, Rockford, IL) to determine protein concentration. SDS-polyacrylamide gel electrophoresis was carried out as described by Laemmli (1970). Briefly, microsomal protein samples from the rat hepatocytes (3 μg/sample) were diluted 1:1 in 2× loading buffer, incubated at 37°C for 30 min, and separated on 7.5% gels. A sample of human liver microsomes (2 μg, provided by Dr. J. Lasker, Mount Sinai School of Medicine, New York, NY) was included as a standard. After electrophoresis, proteins were transferred to nitrocellulose membranes following the procedure of Towbin et al. (1979). Blots were probed with an antibody for rat hepatic CYP4F (provided by Dr. Lasker) overnight at 4°C. Blots were subsequently incubated with a biotinylated anti-rabbit antibody (Amersham Pharmacia Biotech), which was followed by incubation with streptavidin-coupled horseradish peroxidase (Amersham Pharmacia Biotech). Blots were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech). The autoradiograms were quantitated by densitometry with a Bio-Rad laser densitometer.

Statistical Analysis.

Values are expressed as means ± standard error of the mean. Means of several groups were compared by analysis of variance. Statistical probabilities are expressed as *p < 0.05, **p < 0.01, and ***p < 0.001.

Results

The effect of PPARα activation on LTB₄metabolism was assessed in isolated rat hepatocytes treated with LTB₄, the known PPARα activator WY-14,643, or their diluent (control, DMSO). After 24 h of treatment, the ability of the hepatocytes to metabolize LTB₄ was assessed by incubating the cells for 30 min with media containing [³H]LTB₄ as a tracer and measuring formation of radioactive LTB₄metabolites. Following the metabolism of the [³H]LTB₄ had the unique advantage of permitting a dissociation of these metabolites from those formed during the treatment period in experiments when LTB₄ was tested as a PPARα agent. In addition, monitoring radioactivity would allow for the detection of LTB₄ metabolites that lacked a UV chromophore.

HPLC separation and on-line radioactivity monitoring of the effluent revealed formation of several metabolites, which were identified as 16-COOH tetranor-LTB₃, 18-COOH dinor-LTB₄, 20-COOH-LTB₄, and 20-OH-LTB₄ by their retention time relative to the internal standard PGB₂ (Fig.1 A, inset) and the presence of the characteristic triene chromophore with an absorbance maximum at 271 nm and shoulders at 260 and 281 nm (data not shown). In addition, the identity of the metabolites was confirmed by mass spectrometry (data not shown). The peak eluting at 47 min was identified as starting material (i.e., unmetabolized LTB₄). Very polar metabolites with retention times of less than 5 min were also observed but were not further identified.

Overall, in control treated hepatocytes after the 30-min incubation period 71.8 ± 7.0% of the [³H]LTB₄ remained unmetabolized. The ω-oxidation products 20-OH-LTB₄, 20-COOH-LTB₄, 18-COOH-dinor-LTB₄, and 16-COOH-tetranor-LTB₃ accounted for 2.0 ± 0.5, 12.7 ± 1.7, 2.7 ± 1.2, and 1.6 ± 0.7% of the total detected radioactivity, respectively; 16.8 ± 3.8% of the radioactivity was attributed to metabolites more polar than 16-COOH tetranor-LTB₃ (Table1).

Treatment of hepatocytes with either LTB₄ or WY-14,643 resulted in an identical pattern of metabolites as seen in the control-treated cells, and no novel or unexpected metabolites were observed in these samples (Fig. 1, B and C). Furthermore, quantitation of radioactivity in the peaks yielded no significant differences between cells treated with LTB₄ or WY-14,643 and control-treated cells with respect to the distribution of radioactivity between the metabolite peaks nor the total amount of metabolites produced, suggesting that activation of PPARα had no effect on LTB₄ metabolism in these cells.

A time course experiment was also performed in which hepatocytes were treated with LTB₄ or WY-14,643 for 3, 18, or 48 h in addition to the 24-h time point, and LTB₄ metabolism was assessed at these times (Fig.2). More LTB₄ was metabolized during the 30-min incubation period by hepatocytes treated for 3 or 18 h compared with the 24- or 48-h-treated hepatocytes, indicating that the rate of LTB₄ metabolism in these cells decreased as time in culture increased. However, analogous to what was observed at the 24-h time point, there were no significant differences between control and PPARα activator-treated cells with respect to either the amount of ω- or β-oxidized metabolites produced (20-COOH-LTB₄ or 18-COOH-LTB₄) or the amount of LTB₄ that remained unmetabolized at any of the treatment time points that were tested.

The ability of LTB₄ and WY-14,643 to activate transcription of PPARα-responsive genes was assessed by measuring the levels of acyl-CoA oxidase mRNA in cells treated with LTB₄ and WY-14,643 by Northern Blot analysis (Fig. 3). By using this approach to detect PPARα activation, treatment of hepatocytes with WY-14,643 caused a significant increase of acyl-CoA oxidase mRNA levels over control levels after 18 and 24 h of incubation. On the other hand, in hepatocytes treated with LTB₄, acyl-CoA oxidase mRNA expression remained unchanged from control levels at all time points tested. LTB₄ treatment of hepatocytes under these culture conditions did not result in an increased transcription of PPARα-responsive genes.

Because acyl-CoA oxidase is an enzyme involved in β-oxidation of LTB₄ but not involved in the initial steps of LTB₄ metabolism in the hepatocyte, investigation of enzymes known to be involved in the initial rate-limiting steps of LTB₄ metabolism was carried out. Specifically, the effect of LTB₄ or WY-14,643 treatment on the expression of CYP4F, the LTB₄ ω-hydroxylase that catalyzes the conversion of LTB₄ into 20-OH-LTB₄ in rat hepatocytes, was assessed by Western blot analysis (Fig. 4). An antibody directed against CYP4F2, the isozyme cloned in human liver, was used for the SDS-polyacrylamide gel electrophoresis and Western blotting experiments to determine levels of CYP4F protein in microsomal preparations from hepatocytes after 48 h of treatment. The presence of two distinct bands in the lanes containing the rat hepatocytes samples (Fig. 4B, lanes 2–4) suggests that this antibody cross-reacted with two or more of the CYP4F isozymes found in rat liver (J. Lasker, personal communication). Treatment of rat hepatocytes with WY-14,643 increased the amount of immunodetectable CYP4F protein 1.4-fold over control levels. LTB₄ treatment did not change CYP4F expression. As mentioned above, there was no effect of WY-14,643 or LTB₄ treatment on LTB₄ metabolism after 48 h of treatment (Fig. 2).

We also determined the effect of WY-14,643 and LTB₄ on expression of CYP4A1, another P450 enzyme whose transcription is known to be induced by peroxisome proliferators in rat liver (Chen and Hardwick, 1993). Northern Blot analysis of the mRNA revealed similar results as those from the Western blot experiments of CYP4F; specifically, treatment with WY-14,643 increased CYP4A1 mRNA levels 2.0-fold over a 24-h time period, but there was no change in the LTB₄-treated cells (Fig.5).

Discussion

Leukotriene B₄ plays an important role in inflammation, largely because of its chemotactic activity for human polymorphonuclear leukocytes. Due to its potency, the need for regulation of LTB₄ levels becomes obvious. Feedback inhibition has evolved as an important control mechanism by which levels of signaling molecules in the body are regulated. Previous reports have shown that LTB₄ can regulate its own biosynthesis. For example, we have previously shown that LTB₄ can significantly inhibit the 5-lipoxygenase-dependent pathway of LTB₄biosynthesis initiated within the neutrophil during phagocytosis (Fiedler et al., 1998).

Despite detailed investigations of the metabolic pathways of LTB₄ in various tissues, surprisingly few studies have investigated the regulation of LTB₄metabolism as an essential mechanism for limiting its chemotactic activity. It is clear that metabolism plays a central role in regulating the biological half-life of LTB₄through conversion of LTB₄ into inactive metabolites. Thus, the biological activity of LTB₄ is a result not only of the extent of its biosynthesis but also of the subsequent metabolism.

It was in this venue that it was suggested that LTB₄ could enhance its own metabolism in the hepatocyte through the transcription factor PPARα (Devchand et al., 1996). Considerable interest was generated with this hypothesis as a novel pathway for controlling the action of this important eicosanoid and the duration of inflammation in general. The mechanism implied that the expression of enzymes within the peroxisome was induced by activation of PPARα by LTB₄ could enhance LTB₄ metabolism to inactive products. However, no direct measurements of LTB₄ were made to assess whether LTB₄ could induce its own metabolism in the hepatocyte or any other cell. To further investigate this hypothesis, the present study was devised to directly assess whether LTB₄ or a known activator of PPARα, WY-14,643, could alter LTB₄ metabolism in primary rat hepatocyte cultures.

It is generally believed that lipid mediators such as LTB₄ do not circulate in the bloodstream in biologically significant concentrations, and it is therefore debatable of what importance hepatic metabolism of LTB₄ is for its systemic clearance. Still, the hepatocyte was chosen as a model system in these studies because the liver has previously been shown to be one of the few tissues in rats to express PPARα in high levels (Braissant et al., 1996) and because the enzymatic pathways of LTB₄ metabolism have been well defined in these cells (Harper et al., 1986; Shirley and Murphy, 1990; Shirley et al., 1992).

When LTB₄ was incubated with rat hepatocytes by using tritium-labeled LTB₄ as tracer, it was possible to detect the formation of several LTB₄metabolites as previously reported (Shirley and Murphy, 1990). These metabolites all retained a UV chromophore, and a direct correlation was found for the elution of UV active metabolites during the reversed phase-HPLC separation and radioactive components present in the incubation media. The radioactivity eluting from the HPLC column at the solvent front had little or no lipophilicity and likely corresponded to incorporation of tritium tracer atoms derived from metabolism of LTB₄ at carbon atom 15 into nonlipid biochemical intermediates as previously suggested (Shirley and Murphy, 1990). Incubation of rat hepatocytes with LTB₄ resulted in a pattern of metabolites identified as 20-hydroxy-LTB₄, 20-carboxy-LTB₄, 18-carboxy-dinor-LTB₄, and 16-carboxy-tetranor-LTB₃, as well as the polar metabolites.

After treatment of hepatocytes with either LTB₄or WY-14,643, neither a decrease in LTB₄ levels nor an increase in the quantity of any LTB₄metabolites was observed. The results clearly indicated that neither LTB₄ nor WY-14,643 increased LTB₄ metabolism in these cells. Moreover, attempts to increase metabolism either with an altered time course of incubation or treatment of hepatocytes with agents such as retinoic acid, in addition to LTB₄ and WY-14,643, or clofibrate did not alter the outcome (data not shown). In conclusion, no evidence was found to support the hypothesis that treatment of hepatocytes with PPARα activators could induce known or novel pathways of LTB₄ metabolism.

Analysis of the mRNA for acyl-CoA oxidase present in isolated hepatocytes was also carried out in cells treated cells WY-14,643 or LTB₄. There was no effect of treatment at the 3-h time point; however, an increase in transcription of acyl-CoA oxidase was consistently observed after treatment of hepatocytes with WY-14,643 for 18 or 24 h. On the other hand, LTB₄ had no effect on acyl-CoA oxidase mRNA levels in these hepatocytes at any time point, These results are inconsistent with previous observations where an 2.5- and 2-fold increase in acyl-CoA oxidase transcription was reported after only 4 h of treatment with LTB₄ or WY-14,643, respectively (Devchand et al., 1996). There is no obvious explanation for this incongruity, because concentration of the agonists and treatment conditions used were apparently identical. One possible explanation for the discrepancies could be differences in the isolation and culturing of the isolated hepatocytes. For example, it has been previously shown that increasing the concentration of insulin in the culture medium of rat adipocytes to 1 μM induces phosphorylation of PPARα and significantly enhances responsiveness of monkey kidney (CV-1) cells transfected with hPPARα to PPARα agonists (Shalev et al., 1996). Although the concentration of insulin in the medium in the former report was not defined, the concentration of insulin in the culture medium in our experiments was 1 nM, which is 1000-fold lower than the concentration that was necessary to produce the enhancement of transcription, and the potential for artificial increases in mRNA levels was therefore minimal. Nevertheless, even when acyl-CoA oxidase mRNA levels were increased in our experiments, as in the WY-14,643-treated cells, it did not result in a functional consequence, because LTB₄metabolism was unaffected in these cells (Fig. 1).

Studies of the effects of LTB₄ or WY-14,643 on expression of enzymes involved in the immediate steps of LTB₄ metabolism in rat hepatocytes, specifically, members of the CYP4F subfamily, yielded similar results as those for acyl-CoA oxidase, in that LTB₄ treatment had no effect, but a small increase in CYP4F protein levels was observed in cells treated with WY-14,643. These results are seemingly at odds with the findings by Kawashima and Strobel (1995) who reported that hepatic expression of CYP4F was inhibited in rats by treatment with the peroxisome-proliferative drug clofibrate. However, substantial differences exist between our studies with isolated hepatocytes and their model, in which rats were injected with clofibrate, and expression of CYP4F was assessed in whole liver samples. Furthermore, their data also indicate that the down-regulation of P4504F subfamily is not necessarily a universal phenomenon because CYP4F levels in brain and kidney were not changed after clofibrate treatment.

Recently, evidence has accumulated to support that the expression of the CYP4F enzymes is regulated in a species- and tissue-, and even isoform-dependent manner. In mice, for example, clofibrate treatment reduced hepatic expression of CYP4F16, the ortholog of the rat CYP4F5 enzyme, but had no effect on CYP4F15, the ortholog of rat CYP4F4 (Cui et al., 2001). Moreover, clofibrate treatment actually induced renal expression of CYP4F15 (Cui et al., 2001).

Consistent with previous reports of an induction of CYP4A1 transcription by PPARα activators (Chen and Hardwick, 1993), WY-14,643 treatment increased levels of CYP4A1 mRNA. However, failure of LTB₄ treatment to induce transcription of either acyl-CoA oxidase or CYP4A1 further contradicts the previous assumption that LTB₄ can induce transcription by activating PPARα in isolated hepatocytes.

Most importantly, however, in our experiments neither the increase in acyl-CoA oxidase mRNA nor CYP4F proteins was accompanied by a concomitant increase in LTB₄ metabolism. Previous reports by Jedlitschky et al. (1991) had revealed increased metabolism of LTB₄ after chronic treatment of rats with clofibrate. However, as was pointed out by the authors of this article, the chronic treatment of animals with clofibrate in these experiments resulted in considerable hepatomegaly and a marked increase in the relative concentration of peroxisomal fatty acid-metabolizing enzymes within the hepatocytes. In comparison, in our experiments with a rather acute treatment model the increase of both acyl-CoA oxidase and CYP4F caused by WY-14,643 treatment was rather modest, and therefore it was not surprising that an increase in ω-oxidation of LTB₄ was not observed.

In summary, our data indicate that activation of PPARα does not result in the up-regulation of known or novel pathways of LTB₄ metabolism, and no evidence was found to support the suggestion that LTB₄ treatment of hepatocytes leads to activation of PPARα. Therefore, the previously proposed model (Devchand et al., 1996) in which a feedback mechanism exist within the hepatocyte by which LTB₄activates PPARα to induce its own metabolism could not be supported.