0022-1114/92/$3.30 The Joumal of Histochemistry and Cytochemistry Copyright 0 1992 by The Histochemical Society, Inc
Vol. 40, No. 12, pp. 1817-1863, 1992 Prinred in KXA.
Original Article
I Immunohistochemical Localization of Carbonyl Reductase in Human Tissues’ HANSPETER WIRTH2 and BENDICHT WERMUTH3 Department of Clinical Chemispy, University of Berne, 3010 Berne, Switzerland Received for publication March 30, 1992 and in revised form June 29, 1992; accepted July 3 , 1992 (2A2640).
Carbonyl reductase, an NADPH-dependent oxidoreductase of broad specificity, is present in many human tissues. Its precise localization, however, has remained unclear, as well as its physiological and possible pathophysiological significance. The present study reports the immunohistochemical localizationof the enzyme in normal human tissues.Immunostaining was detectable in all organs investigated. The highest concentrations were found in the parenchymal cells of the liver, the epithelial cells of the stomach and small intestine, the epidermis, the proximal tubules of the kidney,
Introduction Carbonyl reductase (secondary alcohol:NADP+oxidoreductase,EC 1.1.1.184) is a cytosolic, monomeric oxidoreductase that catalyzes the NADPH-dependent reduction of a variety of endogenous and xenobiotic carbonyl compounds. Multiple molecular forms of the enzyme differing in charge and size but with apparently identical enzymatic properties have been purified from various human tissues, including placenta (Lin and Jarabak, 1978), brain (Wermuth, 1981), and liver (Felsted and Bachur, 1982), and their occurrence in other tissues has also been demonstrated (Wermuth, 1981). Nothing is known, however, about the enzyme’s histological localization within the various organs. Little is also known about its physiological role and possible pathophysiological significance. The best substrates are quinones, e.g., quinones derived from polycyclic aromatic hydrocarbons(Wermuth et al., 1986),and several xenobiotic aldehydes and ketones (Nakayama et al., 1985;Jarabak et al., 1983; Wermuth, 1981; Ahmed et al., 1978). Many of these compounds exert toxic effects in biologicalsystems, and we have suggested that carbonyl reductase may serve as a general catalyst in the detoxification of these compounds (Wermuth et al., 1986). In addition to xenobiotics, the enzyme catalyzes the reduction of endogenous,
Supported by a grant from the Swiss National Science Foundation (By). Present address: Innere Medizin Gastroenterologie, Universitatsspital. 8091 Zurich, Switzerland. Correspondence ta B. Wermuth, PhD, Chemisches Zentrallabor Inselspital, 3010 Berne, Switzerland.
neuronal and glial cells of the central nervous system, and certain cells of the anterior lobe of the pituitary gland. Consistently pronounced staining was also observed in smooth muscle fibers and the endothelium of blood vessels. The results are in agreement with a housekeeping function of carbonyl reductase in the elimination of reactive carbonyl compounds. (JHistOchemCyrdem 40:1857-1863,1992) KEY WORDS:Reductase;Tissue distribution; Carbonyl metabolism;
Detoxification; Immunohistochemistry; Human.
biologically active compounds, e.g., prostaglandins (Lin and Jarabak, 1978), pterins (Park et al., 1991), and steroids (Inazu et al., 1992). although proof of a physiologically significant participation of carbonyl reductase in the metabolism of any of these substrates is lacking. To obtain more information about the possible physiological and pathophysiological role(s) of this enzyme and to provide a basis for further studies, we carried out a survey of its distribution and histological localization in human tissues.
Materials and Methods Enzyme and Antibodies. Carbonylreductase was purified from human brain (Wermuth, 1981) and liver (Wermuth et al., 1986).Antibodies against the liver enzyme were raised in rabbits and further purified by ammonium sulfate precipitationand DEAE-Sephadexchromatography(Wirth and Wcrmuth, 1985). The effect of formaldehyde on the recognition of carbonyl reductase by the antibodies was tested by ELISA. The wells of a microtiter plate were coated with carbonyl reductase (2 pglml), followed by bovine serum albumin. Part of the wells were then treated with 4% formaldehyde for 15 min before the addition of antibodies.The formation of immunocomplexes was allowed to proceed in 20 mM sodium phosphate buffer, pH 7.2, containing 150 mM NaCl and 0.05% Tween 20 (Buffer A) at room temperature for at least 2 hr. Binding of the anti-carbonyl reductase antibodies was detected with protein A-peroxidase conjugate (Dr. Bommeli, Berne. Switzerland)and 0.04% o-phenylenediamine/O.01%H202 in 150 mM sodium acetate (pH 5.0) as substrates. Extensive washing at the end of each incubation period was carried out in Buffer A. Double Immunodifhsion and Crossed ImmunoisoelectricFocusing. Tissue (5-10 g) was homogenized in 1 part ice-cold 0.1 M sodium phosphate (pH 7) with a Potter-Elvehjem-type homogenizer, and the homogenate was centrifuged at 40,000 x g for 30 min. The supernatant was used 1857
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1858
WIRTH. WERMUTH
Figure 1. Speciicity ofanti-human carbonyl reductase antibodies. (A) Double immunodiffusion of antibodies (center) against purified carbonyl reductase from human brain (1) and brain extract (2). ( 6 )Double immunodiffusionof antibodies(center)against extracts from brain (1.4). liver (2, 5). and kidney (3,6). (C) Crossed immunoisoelectric focusing of human brain exlract. Visualitetion of precipitin lines was by (A) Coomassie brillianlblue, (6)light diffraction, and (C) enzyme activity.
A 1 a
.
I
I
I
I
8
7
6
5
for double immunodiffusion. Immunoprecipitation was allowed to proceed in 1% agar for at least 36 hr at 4'C. For crossed immunoisoelectricfocusing the tissue extract was desalted on Sephadcx G-25(Pharmacia; Uppsala. Sweden) equilibrated with 1% glycine. Froreins *re separated in the first dimension by isoelectricfocusing on thin-layer polyacrylamide gels (pH 3.5-9.5)accordingto the instructions ofthe manufacturer (Pharmacia). The extract (50-80PI) was applied
over the whole width of the gel. and electrofocusing was terminated 30 min after hemoglobin had focused to a sharp band. Gel strips containing the separated proteins were cut out and subjected to immunoelectrophoresis in the second dimension. followingthe general instructionsof Laurel1 (1966).Electrophoresiswas carried out for 20-24 hr. Precipitin bands were visualized by staining with Coomassic brilliant blue R-250.Enzyme activity was revealed by immersing the gel in 0.1 M sodium phosphate (pH 7.5)
Figure 2. Immunohistochemical localization of carbonyl reductase in human tissues. (A) Liver. Strongly stalnlng hepaIic cells of a liver lobule and portal area with bile duct (arrowhead) exhibiting little immunoreactivity. (E) Gallbladder. The columnar epithelium of the mucosfd folds and the endothelium of blood vessels (center) show marked immunostaining. (C) Epithelium of the oral cavity. Wlth the exception of the most superficial cell layer (top right), all epithelial cells show marked immunostaining. (D) Stomach. Transverse section of gastric glands (fundus) shows a gradient of immunoreactivity from strongly staining chief cells in the deeper parts (bottom)to faintly staining mucous neck cells in the upper part of the gastric wall. (E) Duodenum. Longitudinal section of villi with strongly Staining epithelium. Goblet cells have lost their content in the course of the dehydration procedure and are visible as white spots. No immunoreactive protein is detectable in the duodenal glands (bottom left). (FJ Kidney. Proximal tubules intensely stain for carbonyl reductase, whereas distal (arrowhead) and collecting (arrow) tubules as well as glomeruli (asterisk) show weak or no staining. (G) Fallopian lube. Branching folds of Ihe mucous membrane with intensely staining epithelium. (H) Cerebellum. Section Ihrough cortex with, from left to right, molecular layer, Purkinje cell layer with strongly staining Purkinje cell bodies (arrowhead), and granular layer. Dark spots, which are also visible in the control, are cell nuclei counterstained with hematoxylin. (I)Medulla oblongata. Seaion of the nucleus with clearly staining neuronal (arrowhead) and glial (arrows) cells. Controls (cf. Materials and Methods) are shown in the lower part of each panel. D.E.H counterstained with hematoxylin; other panels no counterslain. Bars I 100 pm.
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
CARBONYL. REDUCTASE IN HUMAN TISSUES
1
G
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1859
1860
WIRTH, WERMUTH
containing 0.1 mM NADPH, 0.2 mM menadione, and 0.6 mM nitrotetrazolium blue chloride (Wermuth, 1981). Immunohistochemistry. Human tissue was obtained from legal medical autopsies and, whenever possible, from surgical preparations. The tissue was fixed with formaldehyde (4% in 67 mM sodium phosphate, pH 7.2). dehydrated in a graded series of alcohollxylene, and embedded in paraplast. Sections 4-5 wm thick were mounted on glass slides, rehydrated, and subjected to digestion with trypsin (1 mglml in 50 mM Tris-HC1, 40 mM NaCI, pH 7.6) for 10 min at 37'C. Incubation with and detection of antibodies were carried out as described above for the ELISA, with the exception that a saturated solution of 3,3'-diaminobenzidinein 50 mM TrisHCI (pH 7.0) was used as electron donor. Controls using immunoglobulins from non-immunizedrabbits at the same concentration as the specific antibodies (based on Azso) were routinely carried through the whole procedure. Photographs were taken with a Wild-Leitz Diavert microscope using identical conditions for the antibody-treated sections and the controls.
Characterization of Antibodies A single fused precipitin line was obtained when the antibodies were diffused against purified human brain or liver enzyme and extracts from either organ (Figure 1A). Similarly, diffusion of the antibodies against extracts from brain, liver, and kidney yielded a single b e d precipitin line (Figure 1B). On the other hand, crossed immunoisoelectricfocusing of the same extracts indicated the presence of multiple antigenically related proteins, consistent with the known heterogeneity of the enzyme (Nakayama et al., 1985; Felsted and Bachur, 1982;Wermuth, 1981) (Figure IC). The identity of the precipitated antigen with carbonyl reductase was verified by double staining the gels for protein and enzymatic activity. In no case did we observe a precipitin band that did not stain for enzyme activity. An ELISA used to test the effect of formaldehyde on the antibody-antigen reaction revealed no significant difference between the formaldehyde-treated and native enzymes.
Localization of Carbonyl Reductase Liver. Hepatocytes and Kupffer cells stained strongly and uniformly throughout the liver lobule, clearly standing out from the non-parenchymal tissue (Figure 2A). Bile ducts showed weak staining of the epithelium, whereas in the gallbladder the epithelial cells stained heavily, containing most of the immunoreactive protein in this organ (Figure 2B). Gastrointestinal Tmct. Carbonyl reductase was detectable throughout the gastrointestinal tract. The most intense staining was observed in the stomach, duodenum, and jejunum, followed
by the more distal parts and the esophagus. In all parts the epithelium exhibited a more intense staining than the adjacent smooth muscle layer, including muscularis mucosae, muscularis externa, and smooth muscle cells in the villi. In the oral cavity and the esophagus, immunoreactiveprotein was localized to the more basal cell layers, whereas the most superficialepithelial cell layers stained only weakly or seemed devoid of carbonyl reductase (Figure 2C). Faint staining was observed in the alveoli and excretory ducts of the submandibular gland. In the stomach and the gut, including duodenum, jejunum, ileum, colon, and rectum, immunoreactive protein was detectable,with the exception of goblet cells, throughout the epithelial lining. In the stomach, the intensity of the staining increased from the outermost layer and the mucous neck cells to the deeper parts of the mucosa, including the gastric glands (Figure 2D). In contrast, in the small intestine the strongest staining was observed at the tip of the villi, and the glandular parts were essentially indistinguishable from the control (Figure 2E). Similarly, in the colon and rectum immunostaining decreased from the surface epithelium to the deeper parts of the crypts. Urogenital Tract. In the kidney, carbonyl reductase was localized primarily to the convolutedand straight segments of the proximal tubules (Figure 2F). Very faint immunostaining was detectable in the glomerulus and the capsule of Bowman. However, only autopsy material was available, in which the inner layer of the capsule and the endothelium of the glomerular capillaries were not discernible. The thin segments (loop of Henle) and the distal and collecting tubules exhibited faint or no immunoreactivity, as did the medullary interstitium. Carbonyl reductase was detectable in smooth muscle fibers and the fibromuscular stroma of the ureter, myometrium, and fallopian tube (Figure ZG), and in ductus deferens, seminal vesicles, and prostate gland. The epithelia of the uterus, the uterine glands, and the fallopian tube also showed strong immunoreactivity. In contrast to these latter tissues derived from the Mullerian ducts, epithelia derived from the Wolffian ducts (i.e., epididymis, vas deferens, seminal vesicle, and prostate gland) exhibited little or essentially no staining. Nervous System. Sections from all parts of the central nervous system exhibited diffuse background staining with both ncuronal and glial cell bodies (astrocytes, oligodendrocytes) conspicuously standing out as dark spots. In general, gray matter showed more intense stainingthan white matter. In particular, the following parts of the central nervous system were examined and contained immunoreactive protein: cerebrum including the visual cortex and pre- and postcentral areas, putamen, nucleus caudatus, hypothalamus, pons, cerebellum (outer stellate cells, Purkinjecells, and stellate cells of the granular layer) (Figure 2H), medulla oblongata (Figure
w Figure 3. Immunohistochemical localization of carbonyl reductase in human tissues. (A) Pituitary gland. Anterior lobe with strongly staining cells of unknown function. (B) Parathyroid with strongly staining oxyphilic cells and less intensely staining principal cells. (C) Pancreas. Weakly staining cells of pancreatic islet (asterisk) surrounded by more intensely staining acinar cells of the exocrine pancreas. (D) Ovary. Fetal tissue with strongly staining primordial oocytes. (E) Blood vessel. Cross-section of skin artery. Both endothelial cells and smooth muscle fibers of the tunica media stain strongly. (F) Heart. Myocardial fibers stain strongly in contrast to the interspersed connective tissue, containing a blood vessel with immunopositive endothelium. (G) Lung. Cross-section of alveoli. (H) Placenta. Chorionicvilli with immunopositive cell layer (syncytial trophoblast) and weakly staining connectivetissue. (I)Skin. Section through hair follicle (left) and sebaceous gland (center right) exhibiting marked immunostaining. Controls (d. Materials and Methods) are shown in the bottom part of each panel. No counterstain. Bars = 100um.
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1861
CARBONYL REDUCTASE IN HUMAN TISSUES
A
I
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1862
WIRTH, WERMUTH
21). and spinal cord, including cervical, thoracic, and lumbar regions. No immunoreactiveprotein was detectable in the granule cells of the cerebellum. Cells of neuronal origin outside the central nervous system, such as the ganglia of the dorsal root, the myenteric plexus of the small and large intestines, and the retina, also showed strong staining. In peripheral nerves, immunoreactivity was observed predominantly in the Schwann cells and fibrocytes, whereas the Schwann sheaths and axons showed only weak staining.
Endocrine Glands. Carbonyl reductase was detectable in all endocrine glands investigated, although the intensity of the stain varied markedly. The pituitary gland exhibited strong staining in both lobes and in the pars intermedia. Standing out from the rest of the tissue, certain cells and cell clusters in the anterior lobe exhibited very strong staining (Figure 3A). When stained with hematoxylinleosin, most but not all of these cells were acidophilic. On the other hand, not all acidophilic cells showed increased immunoreactivity, and an assignment to a specific cell type is at present not possible. The thyroid exhibited weak staining, which was localized in the follicle cells. In the parathyroid, the oxyphilic cells stained strongly, whereas the chief cells exhibited markedly weaker staining (Figure 3B). The adrenal gland showed marked and about equal staining of all three cortical zones, but only faint staining of the medulla. In the pancreas, immunoreactive protein was localized essentially to the acini of the exocrine part of the gland, whereas the cells of the Langerhans islets stained only slightly more than the control (Figure 3C). The testes showed diffuse staining of the cells within the seminiferous tubules, i.e., spermatogenic cells, Sertoli cells and, to a lesser extent, interstitial Leydig cells. Very weak staining was observed in the lamina propria. In the ovary, carbonyl reductase was detectablein oocytes of both adult and fetal tissue, follicular cells, and the stroma (Figure 3D). No staining was detectable in corpora albicantia and corpora lutea. Other Tissues. Throughout this study, the endothelium of arterial and venous vessels of all sizes, including capillaries,exhibited distinct immunostaining (Figure 3E). Similarly, smooth muscle fibers of all tissues investigated and cardiac muscle (Figure 3F) yielded significant immunostaining. In contrast, skeletal muscle showed only faint staining. In the lung, carbonyl reductase was detectable in the bronchial glands, respiratory epithelium, and the alveolar wall (Figure 3G). We could not distinguish, however, whether the staining of the alveolar wall reflected the presence of the enzyme in pneumocytes or resulted from its occurrence in the capillary endothelium. In the placenta, syncytial trophoblasts of the chorionic villi stained heavily, in contrast to the weakly staining connective tissue (Figure 3H). The (non-lactating) mammary gland showed strong immunostaining of the tubular cells, whereas the surrounding connective tissue was essentially free of carbonyl reductase. Marked immunostaining was also detectable in the epidermis of the skin, including hair follicles, sebaceous glands, and sweat glands (Figure 31).
Discussion We have carried out an immunohistochemical survey of the distribution and localization of carbonyl reductase in human tissues. Immunostaining was detectable in all organs investigated. Differences existed, however, with respect to the amount of immunoreactive protein in the various tissues as well as the different parts and cell typeswithin one organ. Estimates of the amount of immunoreactive protein were made on the basis of the time required for the color to develop. The highest concentrationswere found in the parenchymal cells of the liver, the epithelia of the stomach and small intestine, the epidermis, the proximal tubule of the kidney, neuronal and glial cells of the central nervous system, and certain cells of the anterior lobe of the pituitary gland. Consistently pronounced staining was also observed in smooth muscle fibers and the endothelium of blood vessels. Weak to equivocal staining was generally detectable in connective tissue and white matter of the central nervous system, whereas distinct tissues with weak or no detectable staining included the granule cells of the cerebellum,the most superficial layers of the epithelia of the oral cavity and the esophagus, the epithelia of the epididymis, vas deferens, seminal vesicle, and prostate gland, the glomerular capsule, the loop of Henle, the distal and collective tubules and the medullary interstitium of the kidney, the colloid of the thyroid gland, the islets of Langerhans, the corpora albicantia and lutea of the ovary, and skeletal muscle fibers. The physiological function of carbonyl reductase is unknown. Its partiality for quinones (e.g., quinones derived from polycyclic aromatic hydrocarbons) as well as for several other xenobiotic carbonyl compounds suggested a role in the detoxification of these reactive substances (Wermuth et al., 1986; Felsted and Bachur, 1982). In agreement with this hypothesis, the liver, the epithelia of the gastrointestinaltract, and the epidermis, which are in close contact with exogenous compounds, have the highest enzyme concentrations. Moreover, Forrest et al. (1990) have shown that the concentration of carbonyl reductase-specificmRNA is increased three- to fourfold in HepG2 and MCF7 tumor cells after treatment with Sudan 1, P-naphthoflavone, or t-butylhydroxyanisole,compounds known to induce the enzymes involved in detoxification. On the other hand, the present study has shown that carbonyl reductase is not restricted to tissues exposed to potentially harmful xenobiotic compounds. On the contrary, the brain, whose cells are shielded from exogenous compounds by the blood-brain barrier, is one of the richest sources of carbonyl reductase. The almost ubiquitous distribution of the enzyme in the human body therefore suggeststhat endogenous carbonyl compounds are also substrates. In fact, carbonyl reductase catalyzes the reduction of various biologically active carbonyl compounds, notably prostaglandins (Lin and Jarabak, 1978), pterins (Park et al., 1991) and steroids (Inazu et al., 1992). In addition, it catalyzes the oxidationofthe 15-hydroxy group of prostaglandins (fin and Jarabak, 1978), the first step in the biological inactivation of these compounds. The Km values for the endogenous substrates, however, are much higher than their cellular concentrations, and alternative enzymes that catalyze the same reactions more specifically have been demonstrated in human tissues (Hara et al., 1990; Hasler and Niederwieser, 1986;Jarabak et al., 1983). These findings have raised questions about the
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1863
CARBONYL REDUCTASE IN HUMAN TISSUES
physiologicalsignificance of carbonyl reductase in the metabolism of these biologicallyactive compounds. Nevertheless, the distribution of the “specific” enzymes in human tissues is largely unknown. In tissues where no alternative enzyme is present, carbonyl reductase might still be responsible for the metabolism of the carbonyl substrates. The present study provides the basis for further investigations of the physiological significance of carbonyl reductase in human tissues.
Literature Cited Ahmed NK,Felsted RL, Bachur NR (1978): Heterogeneity of anthracydine antibiotic carbonyl reductases in mammalian livers. Biochem Pharmacol 27:2713 Felsted RL,Bachur NR (1982): Human liver daunorubicin reductases. Prog Clin Biol Res 114:291 Forrest GL, Akman S, Kruuik S, Paxton RJ, Sparkes RS, DoroshowJ, Felsted RL, Glover Cj, Mohandas T,Bachur NR (1990): Induction of human carbonyl reductase gene located on chromosome 21. Biochim Biophys Acta 1048:149 Hara A, Eniguchi H, Nakayama T, Sawada H (1990): Purificationand properties of multiple forms of dihydrodiol dehydrogenase from human liver. J Biochem 108:250 Hasler T,Niederwieser A (1986): Tetrahydrobiopterin-producingenzyme activities in liver of animals and man. In Cooper BA, Whitehead VM,eds. Chemistry and biology of pteridines. Berlin, Walter de Gruyter, 319
Inazu N. Ruepp B, Wirth H, Wermuth B (1992): Carbonyl reductase from human testis: purification and comparison with carbonyl reductase from human brain and rat testis. Biochim Biophys Acta lll6:50 Jarabak J, Luncsford A, Berkowitz D (1983): Substrate specificity of three prostaglandin dehydrogenases. Prostaglandins 26:849 Laurel1 C-B (1966): Quantitative estimation of proteins by electrophoresis in agarose gel containing antibodies. Anal Biochem 1545 f i n Y-M, JarabakJ (1978): Isolation oftwo proteins with 9-ketoprostaglandin reductase and NADP-linked 15-hydroxy-prostaglandindehydrogenase activities and studies on their inhibition. Biochem Biophys Res Commun 81:1227
Nakayama T,Hara A, Yashiro K, Sawada H (1985):Reductases for carbonyl compounds in human liver. Biochem Pharmacol 34107 Park YS, Heizmann CW, Wermuth B, Levine RA, Steinerstauch P, Guzman J, Blau N (1991): Human carbonyl and aldose reductases: new catalytic functions in tetrahydrobiopterin biosynthesis. Biochem Biophys Res Commun 175:738 Wermuth B (1981): Purification and properties of an NADPH-dependent carbonyl reductasefrom human brain: relationship to prostaglandin 9-ketoreductase and xenobiotic ketone reductase. J Biol Chem 256:1206 Wermuth B, Platt KL, Seidel A, Oesch F (1986): Carbonyl reductase provides the enzymatic basis of quinone detoxificationin man. Biochem Pharmacol 35:1277 Wirth H, Wermuth B (1985): h ” c h e m i c a l characterization of aldo-keto reductases from human tissues. FEBS Lett 187:280
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
Vol. 40, No. 12, pp. 1817-1863, 1992 Prinred in KXA.
Original Article
I Immunohistochemical Localization of Carbonyl Reductase in Human Tissues’ HANSPETER WIRTH2 and BENDICHT WERMUTH3 Department of Clinical Chemispy, University of Berne, 3010 Berne, Switzerland Received for publication March 30, 1992 and in revised form June 29, 1992; accepted July 3 , 1992 (2A2640).
Carbonyl reductase, an NADPH-dependent oxidoreductase of broad specificity, is present in many human tissues. Its precise localization, however, has remained unclear, as well as its physiological and possible pathophysiological significance. The present study reports the immunohistochemical localizationof the enzyme in normal human tissues.Immunostaining was detectable in all organs investigated. The highest concentrations were found in the parenchymal cells of the liver, the epithelial cells of the stomach and small intestine, the epidermis, the proximal tubules of the kidney,
Introduction Carbonyl reductase (secondary alcohol:NADP+oxidoreductase,EC 1.1.1.184) is a cytosolic, monomeric oxidoreductase that catalyzes the NADPH-dependent reduction of a variety of endogenous and xenobiotic carbonyl compounds. Multiple molecular forms of the enzyme differing in charge and size but with apparently identical enzymatic properties have been purified from various human tissues, including placenta (Lin and Jarabak, 1978), brain (Wermuth, 1981), and liver (Felsted and Bachur, 1982), and their occurrence in other tissues has also been demonstrated (Wermuth, 1981). Nothing is known, however, about the enzyme’s histological localization within the various organs. Little is also known about its physiological role and possible pathophysiological significance. The best substrates are quinones, e.g., quinones derived from polycyclic aromatic hydrocarbons(Wermuth et al., 1986),and several xenobiotic aldehydes and ketones (Nakayama et al., 1985;Jarabak et al., 1983; Wermuth, 1981; Ahmed et al., 1978). Many of these compounds exert toxic effects in biologicalsystems, and we have suggested that carbonyl reductase may serve as a general catalyst in the detoxification of these compounds (Wermuth et al., 1986). In addition to xenobiotics, the enzyme catalyzes the reduction of endogenous,
Supported by a grant from the Swiss National Science Foundation (By). Present address: Innere Medizin Gastroenterologie, Universitatsspital. 8091 Zurich, Switzerland. Correspondence ta B. Wermuth, PhD, Chemisches Zentrallabor Inselspital, 3010 Berne, Switzerland.
neuronal and glial cells of the central nervous system, and certain cells of the anterior lobe of the pituitary gland. Consistently pronounced staining was also observed in smooth muscle fibers and the endothelium of blood vessels. The results are in agreement with a housekeeping function of carbonyl reductase in the elimination of reactive carbonyl compounds. (JHistOchemCyrdem 40:1857-1863,1992) KEY WORDS:Reductase;Tissue distribution; Carbonyl metabolism;
Detoxification; Immunohistochemistry; Human.
biologically active compounds, e.g., prostaglandins (Lin and Jarabak, 1978), pterins (Park et al., 1991), and steroids (Inazu et al., 1992). although proof of a physiologically significant participation of carbonyl reductase in the metabolism of any of these substrates is lacking. To obtain more information about the possible physiological and pathophysiological role(s) of this enzyme and to provide a basis for further studies, we carried out a survey of its distribution and histological localization in human tissues.
Materials and Methods Enzyme and Antibodies. Carbonylreductase was purified from human brain (Wermuth, 1981) and liver (Wermuth et al., 1986).Antibodies against the liver enzyme were raised in rabbits and further purified by ammonium sulfate precipitationand DEAE-Sephadexchromatography(Wirth and Wcrmuth, 1985). The effect of formaldehyde on the recognition of carbonyl reductase by the antibodies was tested by ELISA. The wells of a microtiter plate were coated with carbonyl reductase (2 pglml), followed by bovine serum albumin. Part of the wells were then treated with 4% formaldehyde for 15 min before the addition of antibodies.The formation of immunocomplexes was allowed to proceed in 20 mM sodium phosphate buffer, pH 7.2, containing 150 mM NaCl and 0.05% Tween 20 (Buffer A) at room temperature for at least 2 hr. Binding of the anti-carbonyl reductase antibodies was detected with protein A-peroxidase conjugate (Dr. Bommeli, Berne. Switzerland)and 0.04% o-phenylenediamine/O.01%H202 in 150 mM sodium acetate (pH 5.0) as substrates. Extensive washing at the end of each incubation period was carried out in Buffer A. Double Immunodifhsion and Crossed ImmunoisoelectricFocusing. Tissue (5-10 g) was homogenized in 1 part ice-cold 0.1 M sodium phosphate (pH 7) with a Potter-Elvehjem-type homogenizer, and the homogenate was centrifuged at 40,000 x g for 30 min. The supernatant was used 1857
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1858
WIRTH. WERMUTH
Figure 1. Speciicity ofanti-human carbonyl reductase antibodies. (A) Double immunodiffusion of antibodies (center) against purified carbonyl reductase from human brain (1) and brain extract (2). ( 6 )Double immunodiffusionof antibodies(center)against extracts from brain (1.4). liver (2, 5). and kidney (3,6). (C) Crossed immunoisoelectric focusing of human brain exlract. Visualitetion of precipitin lines was by (A) Coomassie brillianlblue, (6)light diffraction, and (C) enzyme activity.
A 1 a
.
I
I
I
I
8
7
6
5
for double immunodiffusion. Immunoprecipitation was allowed to proceed in 1% agar for at least 36 hr at 4'C. For crossed immunoisoelectricfocusing the tissue extract was desalted on Sephadcx G-25(Pharmacia; Uppsala. Sweden) equilibrated with 1% glycine. Froreins *re separated in the first dimension by isoelectricfocusing on thin-layer polyacrylamide gels (pH 3.5-9.5)accordingto the instructions ofthe manufacturer (Pharmacia). The extract (50-80PI) was applied
over the whole width of the gel. and electrofocusing was terminated 30 min after hemoglobin had focused to a sharp band. Gel strips containing the separated proteins were cut out and subjected to immunoelectrophoresis in the second dimension. followingthe general instructionsof Laurel1 (1966).Electrophoresiswas carried out for 20-24 hr. Precipitin bands were visualized by staining with Coomassic brilliant blue R-250.Enzyme activity was revealed by immersing the gel in 0.1 M sodium phosphate (pH 7.5)
Figure 2. Immunohistochemical localization of carbonyl reductase in human tissues. (A) Liver. Strongly stalnlng hepaIic cells of a liver lobule and portal area with bile duct (arrowhead) exhibiting little immunoreactivity. (E) Gallbladder. The columnar epithelium of the mucosfd folds and the endothelium of blood vessels (center) show marked immunostaining. (C) Epithelium of the oral cavity. Wlth the exception of the most superficial cell layer (top right), all epithelial cells show marked immunostaining. (D) Stomach. Transverse section of gastric glands (fundus) shows a gradient of immunoreactivity from strongly staining chief cells in the deeper parts (bottom)to faintly staining mucous neck cells in the upper part of the gastric wall. (E) Duodenum. Longitudinal section of villi with strongly Staining epithelium. Goblet cells have lost their content in the course of the dehydration procedure and are visible as white spots. No immunoreactive protein is detectable in the duodenal glands (bottom left). (FJ Kidney. Proximal tubules intensely stain for carbonyl reductase, whereas distal (arrowhead) and collecting (arrow) tubules as well as glomeruli (asterisk) show weak or no staining. (G) Fallopian lube. Branching folds of Ihe mucous membrane with intensely staining epithelium. (H) Cerebellum. Section Ihrough cortex with, from left to right, molecular layer, Purkinje cell layer with strongly staining Purkinje cell bodies (arrowhead), and granular layer. Dark spots, which are also visible in the control, are cell nuclei counterstained with hematoxylin. (I)Medulla oblongata. Seaion of the nucleus with clearly staining neuronal (arrowhead) and glial (arrows) cells. Controls (cf. Materials and Methods) are shown in the lower part of each panel. D.E.H counterstained with hematoxylin; other panels no counterslain. Bars I 100 pm.
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
CARBONYL. REDUCTASE IN HUMAN TISSUES
1
G
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1859
1860
WIRTH, WERMUTH
containing 0.1 mM NADPH, 0.2 mM menadione, and 0.6 mM nitrotetrazolium blue chloride (Wermuth, 1981). Immunohistochemistry. Human tissue was obtained from legal medical autopsies and, whenever possible, from surgical preparations. The tissue was fixed with formaldehyde (4% in 67 mM sodium phosphate, pH 7.2). dehydrated in a graded series of alcohollxylene, and embedded in paraplast. Sections 4-5 wm thick were mounted on glass slides, rehydrated, and subjected to digestion with trypsin (1 mglml in 50 mM Tris-HC1, 40 mM NaCI, pH 7.6) for 10 min at 37'C. Incubation with and detection of antibodies were carried out as described above for the ELISA, with the exception that a saturated solution of 3,3'-diaminobenzidinein 50 mM TrisHCI (pH 7.0) was used as electron donor. Controls using immunoglobulins from non-immunizedrabbits at the same concentration as the specific antibodies (based on Azso) were routinely carried through the whole procedure. Photographs were taken with a Wild-Leitz Diavert microscope using identical conditions for the antibody-treated sections and the controls.
Characterization of Antibodies A single fused precipitin line was obtained when the antibodies were diffused against purified human brain or liver enzyme and extracts from either organ (Figure 1A). Similarly, diffusion of the antibodies against extracts from brain, liver, and kidney yielded a single b e d precipitin line (Figure 1B). On the other hand, crossed immunoisoelectricfocusing of the same extracts indicated the presence of multiple antigenically related proteins, consistent with the known heterogeneity of the enzyme (Nakayama et al., 1985; Felsted and Bachur, 1982;Wermuth, 1981) (Figure IC). The identity of the precipitated antigen with carbonyl reductase was verified by double staining the gels for protein and enzymatic activity. In no case did we observe a precipitin band that did not stain for enzyme activity. An ELISA used to test the effect of formaldehyde on the antibody-antigen reaction revealed no significant difference between the formaldehyde-treated and native enzymes.
Localization of Carbonyl Reductase Liver. Hepatocytes and Kupffer cells stained strongly and uniformly throughout the liver lobule, clearly standing out from the non-parenchymal tissue (Figure 2A). Bile ducts showed weak staining of the epithelium, whereas in the gallbladder the epithelial cells stained heavily, containing most of the immunoreactive protein in this organ (Figure 2B). Gastrointestinal Tmct. Carbonyl reductase was detectable throughout the gastrointestinal tract. The most intense staining was observed in the stomach, duodenum, and jejunum, followed
by the more distal parts and the esophagus. In all parts the epithelium exhibited a more intense staining than the adjacent smooth muscle layer, including muscularis mucosae, muscularis externa, and smooth muscle cells in the villi. In the oral cavity and the esophagus, immunoreactiveprotein was localized to the more basal cell layers, whereas the most superficialepithelial cell layers stained only weakly or seemed devoid of carbonyl reductase (Figure 2C). Faint staining was observed in the alveoli and excretory ducts of the submandibular gland. In the stomach and the gut, including duodenum, jejunum, ileum, colon, and rectum, immunoreactive protein was detectable,with the exception of goblet cells, throughout the epithelial lining. In the stomach, the intensity of the staining increased from the outermost layer and the mucous neck cells to the deeper parts of the mucosa, including the gastric glands (Figure 2D). In contrast, in the small intestine the strongest staining was observed at the tip of the villi, and the glandular parts were essentially indistinguishable from the control (Figure 2E). Similarly, in the colon and rectum immunostaining decreased from the surface epithelium to the deeper parts of the crypts. Urogenital Tract. In the kidney, carbonyl reductase was localized primarily to the convolutedand straight segments of the proximal tubules (Figure 2F). Very faint immunostaining was detectable in the glomerulus and the capsule of Bowman. However, only autopsy material was available, in which the inner layer of the capsule and the endothelium of the glomerular capillaries were not discernible. The thin segments (loop of Henle) and the distal and collecting tubules exhibited faint or no immunoreactivity, as did the medullary interstitium. Carbonyl reductase was detectable in smooth muscle fibers and the fibromuscular stroma of the ureter, myometrium, and fallopian tube (Figure ZG), and in ductus deferens, seminal vesicles, and prostate gland. The epithelia of the uterus, the uterine glands, and the fallopian tube also showed strong immunoreactivity. In contrast to these latter tissues derived from the Mullerian ducts, epithelia derived from the Wolffian ducts (i.e., epididymis, vas deferens, seminal vesicle, and prostate gland) exhibited little or essentially no staining. Nervous System. Sections from all parts of the central nervous system exhibited diffuse background staining with both ncuronal and glial cell bodies (astrocytes, oligodendrocytes) conspicuously standing out as dark spots. In general, gray matter showed more intense stainingthan white matter. In particular, the following parts of the central nervous system were examined and contained immunoreactive protein: cerebrum including the visual cortex and pre- and postcentral areas, putamen, nucleus caudatus, hypothalamus, pons, cerebellum (outer stellate cells, Purkinjecells, and stellate cells of the granular layer) (Figure 2H), medulla oblongata (Figure
w Figure 3. Immunohistochemical localization of carbonyl reductase in human tissues. (A) Pituitary gland. Anterior lobe with strongly staining cells of unknown function. (B) Parathyroid with strongly staining oxyphilic cells and less intensely staining principal cells. (C) Pancreas. Weakly staining cells of pancreatic islet (asterisk) surrounded by more intensely staining acinar cells of the exocrine pancreas. (D) Ovary. Fetal tissue with strongly staining primordial oocytes. (E) Blood vessel. Cross-section of skin artery. Both endothelial cells and smooth muscle fibers of the tunica media stain strongly. (F) Heart. Myocardial fibers stain strongly in contrast to the interspersed connective tissue, containing a blood vessel with immunopositive endothelium. (G) Lung. Cross-section of alveoli. (H) Placenta. Chorionicvilli with immunopositive cell layer (syncytial trophoblast) and weakly staining connectivetissue. (I)Skin. Section through hair follicle (left) and sebaceous gland (center right) exhibiting marked immunostaining. Controls (d. Materials and Methods) are shown in the bottom part of each panel. No counterstain. Bars = 100um.
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1861
CARBONYL REDUCTASE IN HUMAN TISSUES
A
I
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1862
WIRTH, WERMUTH
21). and spinal cord, including cervical, thoracic, and lumbar regions. No immunoreactiveprotein was detectable in the granule cells of the cerebellum. Cells of neuronal origin outside the central nervous system, such as the ganglia of the dorsal root, the myenteric plexus of the small and large intestines, and the retina, also showed strong staining. In peripheral nerves, immunoreactivity was observed predominantly in the Schwann cells and fibrocytes, whereas the Schwann sheaths and axons showed only weak staining.
Endocrine Glands. Carbonyl reductase was detectable in all endocrine glands investigated, although the intensity of the stain varied markedly. The pituitary gland exhibited strong staining in both lobes and in the pars intermedia. Standing out from the rest of the tissue, certain cells and cell clusters in the anterior lobe exhibited very strong staining (Figure 3A). When stained with hematoxylinleosin, most but not all of these cells were acidophilic. On the other hand, not all acidophilic cells showed increased immunoreactivity, and an assignment to a specific cell type is at present not possible. The thyroid exhibited weak staining, which was localized in the follicle cells. In the parathyroid, the oxyphilic cells stained strongly, whereas the chief cells exhibited markedly weaker staining (Figure 3B). The adrenal gland showed marked and about equal staining of all three cortical zones, but only faint staining of the medulla. In the pancreas, immunoreactive protein was localized essentially to the acini of the exocrine part of the gland, whereas the cells of the Langerhans islets stained only slightly more than the control (Figure 3C). The testes showed diffuse staining of the cells within the seminiferous tubules, i.e., spermatogenic cells, Sertoli cells and, to a lesser extent, interstitial Leydig cells. Very weak staining was observed in the lamina propria. In the ovary, carbonyl reductase was detectablein oocytes of both adult and fetal tissue, follicular cells, and the stroma (Figure 3D). No staining was detectable in corpora albicantia and corpora lutea. Other Tissues. Throughout this study, the endothelium of arterial and venous vessels of all sizes, including capillaries,exhibited distinct immunostaining (Figure 3E). Similarly, smooth muscle fibers of all tissues investigated and cardiac muscle (Figure 3F) yielded significant immunostaining. In contrast, skeletal muscle showed only faint staining. In the lung, carbonyl reductase was detectable in the bronchial glands, respiratory epithelium, and the alveolar wall (Figure 3G). We could not distinguish, however, whether the staining of the alveolar wall reflected the presence of the enzyme in pneumocytes or resulted from its occurrence in the capillary endothelium. In the placenta, syncytial trophoblasts of the chorionic villi stained heavily, in contrast to the weakly staining connective tissue (Figure 3H). The (non-lactating) mammary gland showed strong immunostaining of the tubular cells, whereas the surrounding connective tissue was essentially free of carbonyl reductase. Marked immunostaining was also detectable in the epidermis of the skin, including hair follicles, sebaceous glands, and sweat glands (Figure 31).
Discussion We have carried out an immunohistochemical survey of the distribution and localization of carbonyl reductase in human tissues. Immunostaining was detectable in all organs investigated. Differences existed, however, with respect to the amount of immunoreactive protein in the various tissues as well as the different parts and cell typeswithin one organ. Estimates of the amount of immunoreactive protein were made on the basis of the time required for the color to develop. The highest concentrationswere found in the parenchymal cells of the liver, the epithelia of the stomach and small intestine, the epidermis, the proximal tubule of the kidney, neuronal and glial cells of the central nervous system, and certain cells of the anterior lobe of the pituitary gland. Consistently pronounced staining was also observed in smooth muscle fibers and the endothelium of blood vessels. Weak to equivocal staining was generally detectable in connective tissue and white matter of the central nervous system, whereas distinct tissues with weak or no detectable staining included the granule cells of the cerebellum,the most superficial layers of the epithelia of the oral cavity and the esophagus, the epithelia of the epididymis, vas deferens, seminal vesicle, and prostate gland, the glomerular capsule, the loop of Henle, the distal and collective tubules and the medullary interstitium of the kidney, the colloid of the thyroid gland, the islets of Langerhans, the corpora albicantia and lutea of the ovary, and skeletal muscle fibers. The physiological function of carbonyl reductase is unknown. Its partiality for quinones (e.g., quinones derived from polycyclic aromatic hydrocarbons) as well as for several other xenobiotic carbonyl compounds suggested a role in the detoxification of these reactive substances (Wermuth et al., 1986; Felsted and Bachur, 1982). In agreement with this hypothesis, the liver, the epithelia of the gastrointestinaltract, and the epidermis, which are in close contact with exogenous compounds, have the highest enzyme concentrations. Moreover, Forrest et al. (1990) have shown that the concentration of carbonyl reductase-specificmRNA is increased three- to fourfold in HepG2 and MCF7 tumor cells after treatment with Sudan 1, P-naphthoflavone, or t-butylhydroxyanisole,compounds known to induce the enzymes involved in detoxification. On the other hand, the present study has shown that carbonyl reductase is not restricted to tissues exposed to potentially harmful xenobiotic compounds. On the contrary, the brain, whose cells are shielded from exogenous compounds by the blood-brain barrier, is one of the richest sources of carbonyl reductase. The almost ubiquitous distribution of the enzyme in the human body therefore suggeststhat endogenous carbonyl compounds are also substrates. In fact, carbonyl reductase catalyzes the reduction of various biologically active carbonyl compounds, notably prostaglandins (Lin and Jarabak, 1978), pterins (Park et al., 1991) and steroids (Inazu et al., 1992). In addition, it catalyzes the oxidationofthe 15-hydroxy group of prostaglandins (fin and Jarabak, 1978), the first step in the biological inactivation of these compounds. The Km values for the endogenous substrates, however, are much higher than their cellular concentrations, and alternative enzymes that catalyze the same reactions more specifically have been demonstrated in human tissues (Hara et al., 1990; Hasler and Niederwieser, 1986;Jarabak et al., 1983). These findings have raised questions about the
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
1863
CARBONYL REDUCTASE IN HUMAN TISSUES
physiologicalsignificance of carbonyl reductase in the metabolism of these biologicallyactive compounds. Nevertheless, the distribution of the “specific” enzymes in human tissues is largely unknown. In tissues where no alternative enzyme is present, carbonyl reductase might still be responsible for the metabolism of the carbonyl substrates. The present study provides the basis for further investigations of the physiological significance of carbonyl reductase in human tissues.
Literature Cited Ahmed NK,Felsted RL, Bachur NR (1978): Heterogeneity of anthracydine antibiotic carbonyl reductases in mammalian livers. Biochem Pharmacol 27:2713 Felsted RL,Bachur NR (1982): Human liver daunorubicin reductases. Prog Clin Biol Res 114:291 Forrest GL, Akman S, Kruuik S, Paxton RJ, Sparkes RS, DoroshowJ, Felsted RL, Glover Cj, Mohandas T,Bachur NR (1990): Induction of human carbonyl reductase gene located on chromosome 21. Biochim Biophys Acta 1048:149 Hara A, Eniguchi H, Nakayama T, Sawada H (1990): Purificationand properties of multiple forms of dihydrodiol dehydrogenase from human liver. J Biochem 108:250 Hasler T,Niederwieser A (1986): Tetrahydrobiopterin-producingenzyme activities in liver of animals and man. In Cooper BA, Whitehead VM,eds. Chemistry and biology of pteridines. Berlin, Walter de Gruyter, 319
Inazu N. Ruepp B, Wirth H, Wermuth B (1992): Carbonyl reductase from human testis: purification and comparison with carbonyl reductase from human brain and rat testis. Biochim Biophys Acta lll6:50 Jarabak J, Luncsford A, Berkowitz D (1983): Substrate specificity of three prostaglandin dehydrogenases. Prostaglandins 26:849 Laurel1 C-B (1966): Quantitative estimation of proteins by electrophoresis in agarose gel containing antibodies. Anal Biochem 1545 f i n Y-M, JarabakJ (1978): Isolation oftwo proteins with 9-ketoprostaglandin reductase and NADP-linked 15-hydroxy-prostaglandindehydrogenase activities and studies on their inhibition. Biochem Biophys Res Commun 81:1227
Nakayama T,Hara A, Yashiro K, Sawada H (1985):Reductases for carbonyl compounds in human liver. Biochem Pharmacol 34107 Park YS, Heizmann CW, Wermuth B, Levine RA, Steinerstauch P, Guzman J, Blau N (1991): Human carbonyl and aldose reductases: new catalytic functions in tetrahydrobiopterin biosynthesis. Biochem Biophys Res Commun 175:738 Wermuth B (1981): Purification and properties of an NADPH-dependent carbonyl reductasefrom human brain: relationship to prostaglandin 9-ketoreductase and xenobiotic ketone reductase. J Biol Chem 256:1206 Wermuth B, Platt KL, Seidel A, Oesch F (1986): Carbonyl reductase provides the enzymatic basis of quinone detoxificationin man. Biochem Pharmacol 35:1277 Wirth H, Wermuth B (1985): h ” c h e m i c a l characterization of aldo-keto reductases from human tissues. FEBS Lett 187:280
Downloaded from jhc.sagepub.com at MICHIGAN STATE UNIV on April 11, 2015
