Abstract
Soluble epoxide hydrolase (sEH) hydrolyzes a wide variety of endogenous and exogenous epoxides. Many of these epoxides are believed to be formed by cytochrome P450 epoxygenases. Here we report the distribution of sEH and cytochrome P450 epoxygenases 2C8, 2C9, and 2J2 by immunohistochemistry. A large number of different tissues from different organs were evaluated using high-throughput tissue microarrays. sEH was found in the liver, kidney, and in many other organs, including adrenals, pancreatic islets, pituitary gland, lymphoid tissues, muscles, certain vascular smooth muscles, and epithelial cells in the skin, prostatic ducts, and the gastrointestinal tract. Immunolabeling for sEH was highly specific for particular tissues and individual cell types. CYP2C9 was also found in almost all of these organs and tissues, suggesting that 2C9 and sEH are very similar in their tissue-specific patterns of expression. CYP2C8 and 2J2 were also widely distributed in human tissues but were less frequently associated with sEH. The results suggest potentially distinct pathways of endogenous fatty acid epoxide production and hydrolysis in a variety of human tissues.
S
Since the documentation of the existence of soluble epoxide hydrolase as a different enzyme from the microsomal epoxide hydrolase (Ota and Hammock 1980), its exact physiological substrate(s) and role(s) have been the focus of considerable research. With the many suggested biological roles for sEH, possibly in different tissues, our rationale was to screen whether sEH and P450 epoxygenases exist in these tissues. Many of the biological activities for EETs are re-giospecific (Peri et al. 1997; Node et al. 1999; Capdevila et al. 2000). P450s in the arachidonate epoxygenase pathway epoxidize arachidonic acid with distinct regioselectivity (Capdevila et al. 2000). Hence, the tissue- or organ-specific distribution of these enzymes may be useful in understanding their biological roles in different tissues.
A previous study investigated the distribution of sEH in human tissues by measuring enzyme activity in tissue homogenates, and it has been shown that the liver and kidney have the highest levels of enzyme activity (Pacifici et al. 1988). In this study, using immunohistochemistry (IHC), we provide a comprehensive evaluation of sEH protein localization in an array of normal human tissues and compare distribution and expression levels with those of CYP2C8, 2C9, and 2J2. The results suggest that sEH has a much broader range of expression in humans than has been previously thought, and that CYP2C9 and sEH tissue-specific expression patterns are similar.
Materials and Methods
Tissues, Cell Lines, and Antibodies
Normal tissue samples were obtained from the Cooperative Human Tissue Network (CHTN), International Bioresearch Solutions (IBS) (Tucson, AZ), and normal tissue microarrays from Zymed Laboratories (S. San Francisco, CA), and Ambion Incorporation (Austin, TX). In most cases, samples were provided with background information (e.g., age, sex) about individuals from which the samples were obtained. For SDS-PAGE and immunoblotting, samples were kept at −80C before use.
For immunohistochemical (IHC) procedures, 4-μm sections of paraffin-embedded tissues were used. Sections were histologically evaluated with no pathological findings. The tissue microarrays provided a high-throughput method for evaluation of a large number of tissues, and ensured identical experimental conditions and reagent concentrations for all tissues on the same slide.
Polyclonal anti-human sEH rabbit serum specificity was determined by immunoblotting. Polyclonal anti-human CYP2C8 rabbit IgG (Puracyp; Carlsbad, CA) and polyclonal anti-human CYP2C9 rabbit serum (Gentest; Woburn, MA) immunoblotting and specificity were evaluated by the manufacturers. Polyclonal anti-human CYP2J2 rabbit serum (Dr. Zeldin, Laboratory of Pulmonary Pathobiology; NIEHS, National Institutes of Health) specificity has been previously demonstrated by immunoblotting (Wu et al. 1996). Liver sections were used as a positive control for all four antibodies (Pacifici et al. 1988; Wu et al. 1996; Lapple et al. 2003). In addition, the specificity of the antibodies was confirmed by pre-absorption of each primary antibody against the corresponding antigen before incubation with liver sections. sEH was obtained from Baculovirus-infected Sf-21 cells expressing recombinant human sEH [published genomic EPHX2 sequence (Sandberg and Meijer 1996)], CYP2J2 was a gift from Dr. Zeldin, and CYP2C8 and CYP2C9 were purchased from Gentest.
Electrophoresis and Immunoblotting
Adrenal (175 mg wet weight) and liver (167 mg wet weight) samples were homogenized in 2 ml cold 0.1 M sodium phosphate (pH 7.4), 250 mM sucrose, and 1 mM EDTA and were centrifuged at 13,000 × g (Beckman TL-100 ultracentrifuge) for 25 min. The supernatant fraction from the liver samples was diluted 10-fold. Twenty μl of diluted liver or undiluted adrenal supernatant fraction was added per lane for SDS-PAGE and Western blotting. Protein concentrations were measured by BCA reagent from Pierce (Rockford, IL) as recommended. SDS-PAGE was performed using 5% stacking gels and 10% resolving gels. Proteins were transferred to 0.2-μm PVDF transfer membrane (Millipore; Bedford, MA) overnight at 30V in Tris-glycine buffer with 0.037% SDS and 20% methanol. The membrane was blocked for 1 hr in 5% (w/v) non-fat dry milk plus 0.25% Tween-20 in PBS (pH 7.4), and then incubated with polyclonal rabbit anti-human sEH (1:3000) for 2 hr. The secondary antibody, goat anti-rabbit-IgG-peroxidase conjugate (1:10,000; Sigma, St Louis, MO) was incubated for 3 hr. Bound secondary antibodies were detected with supersignal chemiluminescent substrate as recommended (Pierce). Bands were visualized with a Kodak image station 440CF, with Kodak 1D software. Baculovirus (Sf-21) cells expressing recombinant human sEH were used as positive controls.
Immunohistochemistry
The Vectastain Elite ABC avidin-biotin-alkaline phosphatase kit (Vector Labs; Burlingame, CA) was used. The staining procedure was essentially that provided by the manufacturer. Slides were deparaffinized and hydrated by passage through a series of xylene, ethanol, and distilled water washes. Retrieve-All (Signet; Dedham, MA) was used to unmask antigenic sites as recommended. The sections were then washed in Tris buffer (0.096 M Tris-HCl, 0.029 M Tris base, 0.347 M NaCl with 0.025% Triton X-100, pH 7.6) for 10 min. The sections were incubated with 5% goat blocking serum in Tris buffer for 30 min, and then with primary antibodies (1:25 dilution for sEH, 2C9, and 2J2 antibodies and 1:50 for 2C8 antibody) for 45 min. Primary antibody dilution and incubation period were determined on the basis of titration and optimization experiments. The sections were incubated for 30 min with biotinylated goat anti-rabbit secondary antibody followed by alkaline phosphatase reagent for 30 min. To detect bound antibodies Vector red chromogen was used as substrate. Tris buffer was used for washing between different steps. The sections were then counterstained with hematoxylin, washed in distilled water, and passed through a series of rapid dips in ethanol and xylene, then coverslipped. Negative control studies were done using preimmune rabbit IgG for 2C8 antibody and preimmune rabbit serum for the other primary antibodies, at equivalent concentrations and incubation periods. All experiments were run in replicate (n = 2-5 replicates per experiment) and experiments were repeated on different occasions to confirm constancy of staining in each sample. The expression level of each enzyme in different tissues was evaluated as intensity in stained cells by three independent observers (AE, RF, and MT) and described as absent (−), low (+), moderate (+ +), or high (+ + +). Samples from different individuals were assessed for frequency of expression and were expressed as the number of positively stained samples out of the total number of samples for any particular tissue from different individuals.
Results
We used IHC to determine the distribution and expression of sEH, CYP2C8, 2C9, and 2J2 in an array of normal human tissues. Interestingly, sEH and all three CYPs were expressed widely. However, different distributions and levels of expression were found in different organs examined. Variations observed in staining intensity and frequency of expression for any particular tissue or organ among different individuals from whom the samples were obtained are shown in Table 1. Western blotting analysis for the sEH antibody suggests limited crossreactivity with other proteins (Figure 1A). Negative controls included pre-immune serum, IgG (Figures 1B and 1C) and pre-absorption with the corresponding antigen before incubation with tissue sections (not shown). Because the liver has been previously shown to express all four enzymes (Pacifici et al. 1988; Wu et al. 1996; Lapple et al. 2003), liver sections were used as positive controls (Figures 1D-1G).
(
Semiquantitative evaluation of staining intensity in different tissues and frequency of staining among different individuals for any particular tissue
F, frequency of expression of each enzyme among different individuals in different tissues, expressed as number of positive staining/total number per tissue from different individuals.
I, intensity of staining observed as absent (−), low (+), moderate (+ +), or high (+ + +).
N/A, not applicable. Staining of blood vessels was dependent on the vascular bed and varied among different individuals. Therefore, the frequency could not be assessed. However, staining was more common in small to medium blood vessels for most cases.
Both sEH and CYP2C9 show similar tissue distribution patterns. The distribution of sEH and the three CYPs appeared to follow three main patterns of distribution: sEH+2C9 + 2C8 + 2J2 (Figure 2), sEH+2C9+2J2 (Figure 3), and sEH+2C9 (Figure 4). In the liver, sEH and CYP2C9 showed diffuse distribution in hepatocytes with high (+ + +) staining intensity (Figures 1D and 1F) in most samples examined. Interestingly, in liver samples from a 14-year-old girl we detected a patchy periportal and centrilobular low (+) staining intensity (not shown) for both enzymes. The level of expression of CYP2J2 in the renal cortex (Figure 2J) was different from that of the other enzymes (Figures 2A, 2D, and 2G), being higher in the distal tubules than in the proximal tubules.
(
(
(
In the endocrine system, some interesting cell type-specific patterns of staining were detected. In the pancreatic islets of Langerhans, both sEH and CYP2J2 were localized in the peripheral cells (Figures 3A and 3I), and in the anterior pituitary CYP2J2 expression was cell-specific (Figure 3K). However, identification of these particular cell types requires further investigation using hormone markers. In lymphoid tissues, sEH staining was detected in the germinal centers of the lymphoid follicles, whereas CYP2C9 and 2C8 were detected only in the parafollicular and mononuclear cells (not shown).
Immunostaining for sEH and 2C9 was detected in blood vessels from many organs and tissues, including cervix (Figure 4D and 4H), uterus, spleen, small intestine, and colon. CYP2J2 (Figure 2K) and sEH (Figure 2B) were also detected in blood vessels of the adrenal gland.
Discussion
The precise biological function of sEH is still largely unknown. sEH is reported to be different from and complementary to microsomal epoxide hydrolase, and many exogenous and endogenous substrates have been identified (Meijer and DePierre 1988). sEH has been suggested to be involved in xenobiotic metabolism (Meijer and DePierre 1988) and in regulating the levels of biologically active fatty acid epoxides produced in the CYP450 epoxygenase pathway (Capdevila et al. 2000). To the best of our knowledge, this is the first report to describe the distribution and localization patterns of sEH together with three major CYP450 epoxygenases in a large array of normal human tissues using IHC. Examining the localization of sEH, we conclude that there is a widespread distribution pattern in several tissues and organs. However, we detected differential levels of expression in different tissues and organs and among tissues and organs obtained from different individuals. We also identified three main patterns of distribution with regard to CYP2C8, 2C9, and 2J2, suggestive of potentially distinct pathways of endogenous fatty acid epoxide production and hydrolysis in different human tissues. Our finding of lower level of expression of sEH in the liver of a 14-year-old girl is consistent with the previously reported age and gender differences found for sEH in mice (Pinot et al. 1995).
A comprehensive knowledge of tissue- or cell type-specific patterns of expression of sEH is essential for evaluation of its functional significance, especially with the growing interest in its possible role in many biological activities, including regulation of blood pressure (Sinal et al. 2000), hypertension (Imig et al. 2002), and vascular inflammation (Node et al. 1999). A potential role in blood pressure regulation, hypertension, and/or vascular inflammation is supported by our findings demonstrating the expression of sEH and these CYP450s in the kidney, adrenal gland, and blood vessels.
The cell-specific pattern of distribution of sEH and 2J2 in the pancreatic islets is consistent with previous findings for 2J2 (Zeldin et al. 1997), for which it was suggested that EETs regulate the levels of insulin and glucagon (Falck et al. 1983; Zeldin et al. 1997). CYP2J2, 2C9, and sEH expression in the pituitary gland is intriguing in light of previous data suggesting the influence of EETs on the release of pituitary hormones such as growth hormone (Snyder et al. 1989) and prolactin (Cashman et al. 1987).
Interestingly, different EET regioisomers have differential vasodilator or vasoconstrictor roles (Takahashi et al. 1990; Carroll et al. 1992; Imig et al. 1996) as well as differential effects in inflammation (Peri et al. 1997; Node et al. 1999; Kozak et al. 2003). These differential effects may be regulated in different tissues by the patterns of CYP450 epoxygenases, which have been shown to produce different ratios of EET regioisomers (Capdevila et al. 2000). Furthermore, the level of expression and the distribution pattern of sEH and CYP450 epoxygenases in any particular tissue may alter the availability of arachidonic acid for other pathways in the arachidonic acid metabolic cascade, thus indirectly influencing production of other vasoactive metabolites or proinflammatory metabolites such as prostaglandins.
Finally, sEH has been proposed to play a role in vascular smooth muscle cell and epithelial cell proliferation (Chen et al. 1998; Yu et al. 2000; Davis et al. 2002), which is again supported by the widespread distribution pattern in epithelial cells and blood vessels.
In conclusion, as a xenobiotic-metabolizing enzyme, sEH has a widespread pattern of distribution in normal human tissues and may have a particularly important role in xenobiotic metabolism in some tissues such as the pituitary gland, which lacks a blood-brain barrier and has been previously shown to have higher xenobiotic metabolizing enzyme activities than other parts of the brain (Ghersi-Egea et al. 1992). The detection of sEH and P450 epoxygenases in a variety of normal human tissues supports the different suggested biological roles in these tissues and organs. CYP450-generated epoxides and their corresponding sEH metabolites would be suggested to act as mediators, the levels of which are regulated by expression levels of these enzymes in different tissues. The three main patterns of distribution identified may determine the differential effects of different epoxide regioisomers in any particular tissue. Moreover, these patterns may have functional significance by altering the availability of arachidonic acid for other pathways. Finally, among the P450s studied here, sEH and 2C9 seem to have the greatest similarities in distribution patterns. Further investigation will be needed to determine the biological significance of these patterns of sEH and CYP450 distribution and the differential effects of metabolites produced by these pathways in different tissues.
Footnotes
Acknowledgements
Supported by NIH Grants ES011630 and GM56708.
We are grateful to Dr B. Hammock (University of California, Davis) for providing the polyclonal anti-human soluble epoxide hydrolase rabbit serum and to Dr D. Zeldin (Laboratory of Pulmonary Pathobiology; NIEHS, National Institutes of Health) for providing the polyclonal anti-human CYP2J2 rabbit serum. We also thank Ione Jackman and Lynn Howlett (Histology Laboratory, University of Connecticut) for advice and assistance during the completion of this work.