204 Biochimica et Biophysica Acta, 1157 (1993) 204-~208 © 1993 Elsevier Science Publishers B.V. All rights reserved 0304-4165/93/$06.00

BBAGEN 23811

Localisation of a , /z and zr class glutathione S-transferases in kidney: comparison with CuZn superoxide dismutase

Simon J. Davies a, Richard D'Sousa a, Helen Philips a, Derek Mattey a, Christopher Hiley a, John D. Hayes b, Geoffrey M. Aber a and Richard C. Strange a

a School of Postgraduate Medicine, University of Keele, North Staffordshire Hospital Centre, Stoke-on-Trent, Staffordshire (UK) and b University Department of Clinical Biochemistry, Royal Infirmary, Edinburgh (UK)

(Received 15 December 1992)

Key words: Glutathione S-transferase; CuZn superoxide dismutase; Kidney; Mesangial cell; Epithelial cell

We describe studies in whole kidney, cortical and medullary homogenates and, glomerular cells in culture to determine the relative levels of expression of a (Ya, Yc, Yk), /x (Ybl/Yb2), ~- (Yf) glutathione S-transferases (GST) and CuZn superoxide dismutase (CuZn SOD) in different regions of the nephron. Immunoblotting and immunohistochemistry were used to demonstrate relatively weak expression of a , /z GST and, CuZn SOD in the glomerulus compared to that in particularly distal tubules. Whilst expression of Ya was found within glomerular cells, Yc, Yk and Yf were not detected. Immunofluorescence showed that Ya and Ybl/Yb2 but not Yf were expressed in cultured epithelial and mesangial cells studied between passages 1 and 3. While Ya was distibuted in cytosol, Ybl/Yb2 was primarily located in nuclei.

Introduction

The major cytosolic isoenzymes of the glutathione S-transferase (GST) supergene family can be grouped into the a , / z , 0 and ~- classes [1-3]. In the rat, a genes encode the Ya, Yc and Yk subunits (subunits 1, 2 and 8, respectively), /z genes encode Ybt, Yb 2 (subunits 3 and 4) and the ~- gene encodes Yf (subunit 7) [3]. Endogenous substrates appear to include the products of reactive oxygen species (ROS) attack on lipid and DNA. For example, a isoenzymes catalyse the reduc- tion of linoleic acid hydroperoxide and 4-hydroxy non- 2-enal whilst/x enzymes catalyse the reduction of DNA hydroperoxides and 5'-hydroxymethyluracil [4]. These data suggest certain GST act in concert with antioxi- dant enzymes such as the superoxide dismutases (SOD) to protect cells from ROS. This view is supported by studies in developing human tissues showing similar levels, cellular location and time-specific changes in expression of GST and CuZn SOD [5] as well as the finding of an antioxidant responsive element in the flanking region of the Ya gene that is activated by H 2 0 2 and other oxidants [6].

Correspondence to: R. Strange, School of Postgraduate Medicine, University of Keele, North Staffordshire Hospital Centre, Stoke-on- Trent, Staffordshire, ST4 7QB, UK.

Oxidant stress is a feature of many forms of glomerular injury that are characterised by mesangial cell prol iferat ion/hypertrophy and resultant glomeru- losclerosis [7,8] and, it is likely that the ability of cells to detoxicate ROS will determine their response to renal injury. ROS have complex effects; low concentra- tions have some mitogenic effects while the higher concentrations associated with pathological events such as reperfusion-hypoxia or, infiltration with inflamma- tory macrophages and leucocytes can cause cell dam- age and death [8].

The sensitivity of different cell types within the kidney to ROS varies; injury following reperfusion- ischaemia is most marked in the outer medullary strip and juxta-medullary cortex [7] while tubular damage is found after treatment with cisplatin, an agent that causes a reduction in expression of several relevant enzymes including glutathione peroxidase, glutathione reductase, GST and cytochrome P450 [9]. The reasons for these differences in sensitivity may relate to differ- ent levels of antioxidant protection. Thus, studies using light microscopy have shown that glomerular expres- sion of the superoxide dismutases and a and tz GST is weak compared with the proximal tubule [5]. Low power microscopy however, does not allow identifica- tion of the different intrinsic cell types and it is unclear whether endothelial, mesangial and epithelial cells ex- hibit uniformly low levels of expression or, expression is concentrated within one cell type.

We describe studies using light microscopy, immu- noblotting and immunofluorescence designed to com- pare expression of GST isoforms and CuZn SOD in different parts of the kidney, the glomerulus and tubule and, in cultured glomerular epithelial and mesangial cells.

Mater ia l s and M e t h o d s

Tissue studied Kidneys were obtained from male WKY rats (20-25

wk). Half of each kidney was immediately fixed in buffered formalin for immunohistochemistry. Ho- mogenates were also prepared from whole kidneys and, after macroscopic dissection, from cortex and medulla.

Glomerular preparations Glomeruli were isolated using a mechanical seiving

technique [10]. Cortices were passed through 250/zm, 100 /~m and 75 /~m nylon sieves and glomeruli col- lected from the 75/~m sieve. Purity of the glomerular suspensions, assessed by light microscopy, was greater than 90%.

Mesangial and epithelial cell cultures Isolated glomeruli were disaggregated by initial in-

cubation (10 min) with collagenase (1 mg/ml). Trypsin (0.5%, w/v) and EDTA (0.02%, w/v) were then added and the incubation continued for a further 10 min.

Cultures of epithelial cells were obtained by plating disaggregated glomeruli onto alcohol-flamed coverslips and growing to confluence (8-10 days) in RPMI 1640 containing 10% FCS and insulin/transferrin/sodium selenite (Sigma). During this time mesangial cells were not evident.

To obtain mesangial cells, disaggregated glomeruli were seeded into culture flasks (25 cm 2) containing RPMI 1640 supplemented with 17% FCS and insulin/transferrin/sodium selenite. Mesangial cells appeared after about 7 days and grew to confluence after 21-28 days [11]. Cells, studied between passages 1-3, were seeded onto cover slips and, after 24 h, used in immunofluorescent studies.

The identity of epithelial cells was confirmed by their typical morphology under phase-contrast mi- croscopy and characteristic immunofluorescence stain- ing pattern for cytokeratin. Mesangial cells also demonstrated typical morphology and were positive for a-actin, desmin, vimentin, but negative for cytokeratin, Factor VIII and leucocyte common antigen CD45. Immunofluorescence identification of these proteins was performed as described below using antibodies obtained from Sigma Chemical Co., Dorset, UK.

205

Antibodies Antisera to a (Ya, Yc, Yk), /z (Ybl, Yb2) and ~r

(Yf) GST were used [3]. Antilaodies to Ybl cross-re- acted with Yb2 but antisera to'Yb2 did not cross-react with Ybl. Ya and Yf cross-reacted with antisera to human a and ~" GST respectively [5]. Antisera to human CuZn SOD (monomer mol wt 18000 Da) [12], demonstrated cross-reactivity with the rat isoform (monomer mol wt 16 500 Da) on immunoblots.

Immunoblotting Homogenates of cortex, medulla and isolated

glomeruli were adjusted to a total protein concentra- tion of 2 mg/ml and subjected to SDS/PAGE in 12.5% resolving gels and blotted onto nitrocellulose membranes [12,13]. Enzyme standards (about 40 ng per gel) were included on each gel. Antigens were detected using an alkaline phosphatase method [12] and quanti- fied by scanning each antigen band using a Confocal 2000 Image Analyser (Confocal Technologies, Liver- pool, UK). A background reading, obtained from the mean of six random measurements (coefficient of vari- ation < 2%) of each immunoblot, was subtracted from a similar mean of the greyness of each antigen band. Results from four experiments, each comprising three kidneys, were compared.

Immunohistochemistry Half kidneys were fixed in buffered formalin, pro-

cessed to paraffin wax, cut and stained with antibodies or appropriate controls using an avidin-biotin complex and peroxidase method [14,15]. Antibodies to human a and ~- GST were used to cross check the immunohisto- chemical data with identical results. Tissue sections

TABLE I

Localisation of GST and CuZn SOD within glomerul~ renal cortex and medulla determined by immunoblotting

Samples of cortex, medulla and glomeruli were examined for expres- sion of GST and CuZnSOD by immunoblotting. Antigen bands were quantified by image analysis and compared with the appropriate enzyme standard (40 ng). Results were expressed as greyness of sample/greyness of standard × 100. Data shows mean + SEM of four experiments.

Glomeruli Cortex Medulla a GST

Ya 58+ 10 350+61 110+43 Yc 2+ 0.2 20+ 4 125:2 Yk 25:0.3 17+ 3 65:1

/z GST YB1/Yb2 75:0.9 9+ 1 145:1.2

• r GST Yf 0 255:3 3+ 0.3

CuZn SOD 28 5:5 91 + 12 25 5:3

206

TABLE II

Relative expression of GST and CuZn SOD in glomeruli, proximal and distal tubules and medulla determined immunohistochemically

Glomeruli Tubules Medulla

proximal distal

a GST Ya + + + + + + + + Yc - + + + + + + + Yk - + + + + + + +

/z GST Ybl + + + + + + Yb2 + + + + + +

~" GST yf - _ + + + + +

CuZn SOD + + + + + + + +

Score:- = no positive cells, + = occasional positive, + + = majority cells positive, + + + = all cells positive.

were viewed bl ind and a scoring system appl ied to each of four areas: glomeruli , medulla , proximal and distal tubules (Table 2); Score: 0 = no positive cel ls ,+ = occasional positive, + + = majori ty cells positive, + + + = all cells positive.

lmmunofluorescence Epithel ia l and mesangia l cells grown on cover slips

were washed in PBS, fixed (30 min) with 3.5% formal- dehyde and permeabi l i sed with 100% ethanol ( - 2 0 ° C , 2 min). Af ter blocking (1% BSA), cells were incuba ted (30 min) with pr imary ant ibody (1:150), washed and incubated (30 min, 20°C) with a secondary F I T C la- bel led ant ibody (1:40). Pr imary ant ibody was omited in controls. Cul tu red h u m a n fibroblasts were used as C u Z n SOD positive control.

Results

Immunoblott ing The results of immunoblo t t ing glomerular , cortical

and medul la ry homogena tes are summar ised in Table I. All the enzymes were expressed in whole cortical

homogena tes bu t appeared to be p redominan t ly lo-

calised to cortical tubules as expression of a (Ya), /z G S T ( Y b l / Y b 2 ) and C u Z n SOD in isolated glomeru- lar p repara t ions was relatively weaker. A lpha class Ya conta in ing isoforms were p r edominan t in each of the areas of the kidney studied. Expression of Yk and Yc was relatively much weaker. Expression of ~" class Yf was also weak and this m o n o m e r was not detec ted in isolated glomeruli (Table I).

All the enzymes examined were detec ted in medul lary homogena tes bu t expression was usually weaker than in cortex.

Fig. 1. a. Glomerular mesangial cells (passage 3) cultured on glass cover slips examined 24h after seeding. The cells demonstrate im- munofluorescence localisation of/z class GST (Ybl/Yb2) predomi- nantly to nuclei. Staining is evident in two nuclei. Some weaker cytosolic staining is also evident, b. Cultured epithelial cells (not passaged) examined 7-8 days after seeding whole glomeruli on glass cover slips. The cells demonstrate localisation of /z class GST (Ybl/Yb2) within nuclei, c. Cultured glomerular mesangial cells (as Fig. la) demonstrating localisation of a class GST (Ya) within

cytosol with little nuclear staining.

Immunohistochemistry Kidneys from 4 animals were examined and results

(Table II), were identical in each case. All enzymes studied showed intense expression within distal tubules throughout the cortex, a and /z GST and CuZn SOD were, in addition, localised to proximal tubules where expression was strongest in the deep juxta-meduIlary cortex. Expression of Ya and Yb l /Yb2 while weaker than in tubular cells, was Iocalised to glomerular cells. CuZn SOD appeared faintly on capillary loops.

Medullary staining, positive for all enzymes exam- ined, was localised predominantly to collecting ducts. The greater amount of negatively staining interstitial tissue in medulla compared with cortex, may account for the comparatively fainter bands found on immuno- blotting of medullary homogenates.

Irnmunofluorescence Immunofluorescence was used to examine the sub-

cellular localisation of Ya, Yb l /Yb2 and Yf GST and CuZn SOD in cultured mesangial and epithelial cells. Both cell types were negative for CuZn SOD and Yf although weak positivity for CuZn SOD was obtained after immunoblotting cultured cells. In contrast, stain- ing for/~ (Ybl /Yb2) and a (Ya) GST was consistently positive in cells obtained from 5 separate cultures. Subcellular distribution differed; predominantly nu- clear for Yb l /Yb2 (Fig. la,b) and cytosolic for Ya (Fig. lc) in both cell types.

Discussion

We have described the localisation and relative lev- els of expression of cytosolic GST and CuZn SOD in different regions of the kidney and, the distribution of Ya and Yb l /Yb2 within cultured glomerular epithelial and mesangial cells. Results from cultured ceils were supportive of those obtained in tissue preparations. This agreement may reflect the study of cultured cells at low passage numbers since expression of xenobiotic-metabolising enzymes can alter in culture.

The study complements that of Guthenberg et al. [16] who showed expression of YaYa, YaYc, YcYc, Yb2Yb2 and Yf in rat kidney cytosols. Both studies show that Ya is strongly expressed in kidney, the Ybl monomer is weakly expressed. Our immunoblotting studies demonstrated that the GST and CuZn SOD were present principally in the cortex. Immunohisto- chemical data also demonstrated localisation to tubules, with only sparse glomerular staining. We have previ- ously reported the predominance of tubular compared to glomerular expression for GST and CuZn SOD in the human, although confined our observations to im- munohistochemical studies [14,15]. A similar pattern for CuZn SOD has been described in the Syrian ham- ster where this enzyme is co-expressed with Se-glutath- ione peroxidase and catalase [17].

207

Immunoperoxidase staining showed localisation of a and /.~ GST and CuZn SOD to deep proximal and distal tubules and medullary collecting ducts. This pat- tern of ot GST expression differs from that in humans where a is confined to proximal tubules [14]. The significance of this species difference is unclear. Inter- estingly, a and/~ GST and CuZn SOD, enzymes with a presumed antioxidant function, are expressed in seg- ments of the nephron with high metabolic activity [7] and, known to be vulnerable to oxidant damage follow- ing events such as ischaemia-reperfusion [18]. Presum- ably constitutive expression of antioxidant enzymes is inadequate when faced with sudden increases in ROS. Pi class GST have relatively low activities towards oxidised lipid and DNA [4] and were confined to distal tubules and collecting ducts.

The detection by immunoblotting of a (Ya) and /x (Ybl /Yb2) GST in isolated glomeruli was supported by the finding of positive staining in occasional glomerular cells viewed at high power. Only very weak expression of Yc and Yk was detected by either of these approaches. YaYa and YkYk have markedly different activities towards various lipid substrates; YaYa catalyses the formation of prostaglandins (PG) E 2 and F2~ from PGH 2 and, the reduction of linoleic acid hydroperoxide. YkYk has no activity towards these substrates but has high activity towards 4-hydroxy non- 2-enal [4,19]. The significance of these different pat- terns of expression may relate to the recognised impor- tance of glomerular prostaglandin production. Further- more, tubular cells are metabolically active and may undergo high rates of lipid peroxidation and/or be particularly sensitive to the genotoxic effects of alke- nals.

Both mesangial and epithelial cells expressed Ya and Ybl /Yb2 although immunofluorescence studies showed their subcellular distributions differed; Ya was distributed in cytosol while Yb l /Yb2 was predomi- nantly perinuclear. These data support reports of nu- clear staining for GST [20] and the view that /~ GST catalyse the reduction of peroxidised DNA [4].

Immunofiuorescence staining of early passage cul- tured cells for CuZn SOD was negative although weak positivity was detected on immunoblots indicating the low level of expression of this enzyme. The enzyme was detected in isolated glomeruli by immunoblotting and, faint glomerular loop positivity was found using im- munoperoxidase staining. We interpret these data as indicating the majority of glomerular CuZn SOD is located in endothelial ceils. This suggestion is sup- ported by observations of CuZn SOD activity in iso- lated glomeruli and cultured bovine glomerular en- dothelial cells [21,22]. The presence of CuZn SOD within glomerular endothelial cells indicates these cells constitute a protective barrier to ROS derived from the circulation.

208

The relatively low levels of expression of a and /x GST and CuZn SOD within glomeruli implies that filtration represents less metabolic work than tubular reabsorption, and/or supports the suggested physio- logical role for ROS, particularly H20 2 within the glomerulus. It also suggests glomeruli are particularly vulnerable to oxidant injury, especially when the en- dothelium is disrupted allowing infiltration of inflam- matory cells which release ROS, a phenomenon char- acteristic of many forms of glomerulonephritis.

Acknowledgements

We thank Dr. Clive Stonier for help with image analysis and the National Kidney Research Fund, ICI Pharmaceuticals and the North Staffordshire Medical Institute for support. RD'S is a holder of the North Staffordshire Medical Institute Research Fellowship.

References

1 Mannervik, B., Alin, P., Guthenberg, C., Jensson, H., Tahir, M.K., Warholm, M. and Jornvall, H. (1985) Proc. Natl. Acad. Sci. USA 82, 7202-7206.

2 Pickett, C.B. and Lu, A.Y.H. (1989) Annu. Rev. Biochem. 58, 743-764.

3 Hayes, J.D. (1990) in: Hayes, J.D., Pickett, C.B. and Mantle, T.J., eds. Glutathione S-transferases and drug resistance, pp. 17-33. Taylor and Francis, London.

4 Ketterer, B. and Meyer, D.J. (1989) Murat. Res. 214, 33-40. 5 Strange, R.C., Fryer, A.A., Hiley, C., Bell, J., Cossar, D. and

Hume, R. (1990) in: Hayes, J.D., Pickett, C.B. and Mantle, T.J.,

eds. Glutathione S-transferases and drug resistance, pp262-271. Taylor and Francis, London.

6 Rushmore, T.H., Morton, M.R. and Pickett, C.B. (1991) J. Biol. Chem. 266, 11632-11639.

7 Canavese, C., Stratta, P. and Vercellone, A. (1988) Nephron 49, 9-15.

8 Shah, S.V. (1989) Kidney Int. 35, 1093-1106. 9 Bompart, G., Orfila, C. and Manuel, Y. (1991) Nephron 58,

68-74. 10 Messenger, E.A., Stonier, C. and Aber, G.M. (1988) Clin. Sci. 75,

191-196. 11 Camazine, S.M., Ryan, G.B., Unanue, E.R. and Karnovsky, M.J.

(1976) Lab. Invest. 35, 315-326. 12 Strange R.C., Cotton, W., Fryer, A.A., Jones, P., Bell, J. and

Hume, R. (1990) J. Lab. Clin. Med. 116, 666-673. 13 Towbin, H., Staehelin, T. and Gordon, J. (1979) Proc. Natl. Acad.

Sci. USA 76, 4350-4354. 14 Hiley, C., Bell, J., Hume, R. and Strange, R.C. (1989) Biochim.

Biophys. Acta 990, 321-324. 15 Strange, R.C., Hiley, R.C., Roberts, C., Jones, P., Bell, J. and

Hume, R. (1989) Free Rad. Res. Commun. 7, 105-112. 16 Guthenberg, C., Jensson, H., Nystrom, L., Osterlund, E., Tahir,

M.K. and Mannervik, B. (1985) Biochem. J. 230, 609-615. 17 Oberley, T.D., Oberley, L.W., Slattery, A.F., Lauchner, L.J. and

Elwell, J.H. (1990) Am. J. Pathol. 137, 199-214. 18 Mason, J. and Torhorst, J. (1984) Kidney Int. 126, 283-293. 19 Ujihara, M., Tsuchida, S., Satoh, K., Sato, K. and Urade, K.

(1988) Arch. Biochem. Biophys. 264, 428-433. 20 McCusker, F.M., Phillips, M.F., Boyce, S.J. and Mantle, T.J.

(1990) in: Hayes, J.D., Pickett, C.B. and Mantle, T.J., eds. Glu- tathione S-transferases and drug resistance, pp. 262-271. Taylor and Francis, London.

21 Yoshioka, T., Bills, T., Moore-Jarrett, T., Greene, H.L., Burr, I.M. and Ichikawa,I. (1990) Kidney Int. 38, 282-288.

22 Yoshioka, T., Kawamura, T., Yoshida, H., Beckman, J.K., Hoover, R.L., Magnuson, M.A., Mayrick, B.O. and Ichikawa, I. (1991) J. Am. Soc. Nephrol. 2569.