Effects of cyclosporin and FK-506 on glomerular ... - Springer Link

2 downloads 0 Views 748KB Size Report
Effects of cyclosporin and FK-506 on glomerular mesangial cells. Evidence for ...... (1991) Cyclosporine and FK 506 induced inhibition of renal epi- thelial cell ...
Eur J Clin Pharmacol (1993) 44 [Suppl 1]: S 11~S16

BQ[s: eqerleg @Springer-Verlag 1993

Effects of cyclosporin and FK-506 on glomerular mesangial cells Evidence for direct inhibition of thromboxane synthase by low cyclosporin concentrations H. H. Radeke, S. Kuster, V. Kaever, and K. Resch Institut for Molekularpharmakologie, Medizinische HochschuleHannover, Hannover, Germany

Summary. The cellular sources or molecular mechanisms responsible for the derangement of vasoactive prostanoid levels during immunosuppressive cyclosporin (CSA) therapy have not been defined. Using cultured rat glomerular mesangial cells (MC), the cytostatic, cytotoxic and prostanoid synthesis modulating effects of CSA and FK506 have been measured and compared with the immunosuppressive action of these drugs. Both, CSA and FK-506 inhibited proliferation of MC at similar doses (ICs0 -~ 1 gg- ml l). Lymphoproliferation was suppressed with ICs0s of 50 ng. ml -~ and < 1 ng .m1-1, respectively. In contrast, and unlike FK-506, CSA caused mesangiolysis (ICs0 = 4.5 pg. ml- 1) and concentration dependently inhibited the interleukin-lfl (IL-lfl) stimulated mesangial cell release of TXB2 at nanomolar doses (IC50 = 50ng.ml-1). In kinetic experiments (6-48h), CSA i ng- ml- ~ partially and 1 gg. ml- ~ completely abolished the IL-lfl augmented mesangial secretion TXB2 at all the time points tested. Both, low and high doses of CSA reduced PGE2 release by only 20-40 % and then not until at least 24 h of incubation. Measuring enzymatic capacity of membrane fractions of MC to generate TXB2 or PGE; from added arachidonic acid (10 5 M), CSA (0.11000 ng. ml-1) caused a dose dependent reduction in cyclooxygenase (COX)/thromboxane synthase activity up to 76%, while PGE2 synthesis (COX/prostaglandin synthase) was decreased by 34 %. Immunoblots with a specific COX-1 antiserum revealed that COX-lprotein expression of MC was not affected by CSA. The in vitro data suggest that, unlike FK-506, CSA has a specific and direct effects on the cytochrome P-450 III enzyme, thromboxane synthase, at concentrations in the range of these producing immunosuppressive effects. Key words: Mesangial cell, Cyclosporin, FK-506; thromboxane, prostaglandin synthase, cyclooxygenase Major investigations towards the mode of action of cyclosporin (CSA) have concentrated on its immunosuppressive effects [1, 2]. Considerable progress in this area has been made with the definition of a CSA binding protein,

cyclophilin, and its characterisation is cis-trans prolin isomerase [3]. A current hypothesis claims that CSA inhibits cytokine transcription via D N A enhancer/promoter sequences following binding to cytosolic cyclophi!in (and in a second step to calcineurin) [1-3]. A similar mechanism has also been proposed for the activity of another new immunosuppressant, FK-506, whereas rapamycin acts slightly differently [4]. In the elucidation of signalling mechanisms involved in lymphocyte activation CSA, FK506 and rapamycin have proved to be very useful tools. However, serious adverse effects on kidney function associated with CSA therapy [5, 6] have not been explained by these investigations. There is no indication that the effect of CSA on the transcription of cytokines might be relevant for the appearance of renal adverse effects. With respect to the pathological changes in the kidney, such as decrease in GFR, tubular or endothelial lesions or interstitial fibrosis different concepts of toxic CSA actions have been discussed [5]: enhancement of sympathetic neurotransmitter release, induction of endothelin [7], calcium dependent potentiation of vasoconstriction [8, 9], and a direct effect on mitochondrial function [10]. In addition to these investigations, the majority of studies have been focussed on alterations in the vasoactive eicosanoid balance. Increased urinary and glomerular secretion of thromboxane has been associated with CSA toxicity [11, 12]. Subsequently, a combination of thromboxane synthase inhibitors and receptor antagonists prevented the adverse renal effects of CSA [12]. On the other hand it has been demonstrated in several investigations that the administration of vasodilator prostaglandins resulted in an improvement in CSA related renal dysfunction [13]. Despite the apparent evidence of involvement of vasoconstrictive thromboxane in the course of CSA toxicity, some authors have failed to show a correlation between renal toxicity related to CSA and urinary eicosanoid secretion in vivo [14,15,16]. In the current investigation we have examined the effects of CSA and FK-506 on thromboxane B2 and prostaglandin E2 formation by cultured glomerular mesangial cells. In contrast to studies in vivo or with isolated glomeruli, we found a very significant inhibition of TXB2 release from cultured MC by low CSA concentrations, while FK-

$12 506 h a d n o effect. F u r t h e r e x p e r i m e n t s s u g g e s t e d that C S A was d i r e c t l y i n h i b i t i n g the activity o f t h e c y t o c h r o m e P-450 e n z y m e , t h r o m b o x a n e synthase.

any crossreaction with other prostaglandins or arachidonic acid, with the exception that the anti-TXB2 antibody recognise to 8.6 % the TXB2 metabolite 2,3-dinor-TXB2 [22].

Material and methods

Enzymatic activity of the enzyme combinations cyclooxygenase/prostaglandin E2 synthase and cyclooxygenase/thromboxane synthase

Materials Cyclosporin (CSA) was a kind gift from Dr. B. Ryffel, Sandoz AG, Basle, Switzerland. The stock solution of 10 mg/ml was kept in the dark at 4°C in dehydrated dimethylsulfoxide (DMSO). FK-506 (10 mg/ml in DMSO) was kindly donated by Dr. U. Christians, Institute of General Pharmacology, Medical School, Hannover, Germany. All cell culture reagents came from GIBCO, Wiesbaden, FRG, if not otherwise stated. The sources of the antisera for immunocytochemical characterisation of MC have been described in [17]. A polyclonal rabbit antiserum direct against sheep-cyclooxygenase (COX-l; crossreacting with human COX-1 [18]) was kindly donated by David DeWitt (Dep. Biochemistry, Michigan State University, East Lansing, USA).

Mesangial cell culture MC were prepared from Sprague Dawley rat kidneys as described elsewhere [17, 19]. The cells were characterized by immunofluorescence as showing a positive reaction for myosin, vimentin, fibronectin, desmin and collagen type IV, and they were negative with keratin and factor VIII antisera. The culture medium used for the experiments consisted of DMEM, supplemented with 2 retool.1 1 glutamine, 100 U.m1-1 penicillin, 100 gg. ml I streptomycin, 1% non-essential amino acids and 5 % fetal calf serum (Gibco). MC cultures were routinely shown to be negative for mycoplasma contamination by DAPI (4'6-diamino-2-phenylindole, Sigma, Deisenhofen, FRG) staining.

Determination of antiproliferative and cytotoxic effects DNA synthesis in subconfluent MC exposed for 48 h to the appropriate dilutions of CSA or FK-506 in DMSO, or DMSO controls, were performed as described in [19, 20]. Control experiments to determine immunosuppressive effects on human peripheral blood mononuclear cells (PBL) were performed as described [19, 21]: PBL (buffy coat cells) from 200 ml heparinised blood (freshly obtained from healthy volunteers (Blood transfusion Unit, Medical School Hannover) were isolated on a standard Ficoll-hypaque gradient with centrifugation. Cells seeded at a density of 2 x 106/ml were incubated either with no stimulus (control) or phytohemagglutinin (PHA, 5 gg. ml-1) in the presence of serial dilutions of the immunosuppressants for 20 h, and a further 4 h with [3H]-thymidine to determine DNA synthesis. Cytotoxic effects of CSA or FK-506 on adherent MC were measured in analogy to [19], using the DNA-stain crystal violet (0.1%, pH 5.6). Absorption was determined at 590 nm after dye solubilisation with 10 % CH3COOH. The intensity of staining was directly correlated with the number of viable cells over the range 103 2 x 105 MC/well in microtiter plates.

Thromboxane B 2 (TXBO and prostaglandin E2 (PGE2) determination TXB2 and PGEz concentrations in supernatants taken from incubations of MC in triplicate, in 24-well plates, or from the cyclooxygenase assay reaction mixtures (see below) were determined by the appropriate enzyme-linked, immunoabsorbant assays (ELISA), performed according to Reinke et al. [22] for TXB2, or as adapted for PGE2 by Kaever et al. in our laboratory [23]. For both eicosanoids highly specific monoclonal antibodies were used: anti-TXB2 antibody (clone 4E-TBR1) and anti-PGE2 antibody (clone 6B-E2R2), both donated by Dr. K. Brune, Pharmacology Dpt., University of Erlangen, Germany. As previously described neither antibody showed

Crude plasma membrane fractions of mesangial cells for the determination of arachidonic acid metabolising activity and cyclooxygenase (COX-l) protein content (see below: COX-1 Western blot) were prepared according to [24]. After the relevant stimulation protocol ( _+IL-lfl; + CSA/FK-506 dilutions) in 80 cm2culture flasks (three/point) the enzymatic capacity of MC membrane fractions to produce either prostaglandin Ez or thromboxane B2 from arachidonic acid was determined using the conditions described for the "Cyclooxygenase assay" in [25]. Protein fractions from crude membrane preparations adjusted to 20gg/sample in 50rnM Hepes/140 mM KC1, pH 8.0 were preincubated in the assay mixture (8 gM bovine haemoglobin (type H2625, Sigma), 320 mM DL-tryptophan (Sigma), 3.75 mM glutathione (Boehringer Mannheim, Germany)) for 10 rain at 37°C. In the final experiments CSA was not added to riving MC, but directly to the membrane fraction 10 rain before starting the COX assay. The assay was performed in duplicate with one additional sample containing the specific cyclooxygenase inhibitor diclofenac. The enzymatic reaction was started by the addition of 10 5 M arachidonic acid (Sigma) and was allowed to proceed for 30 min at 37°C. After stopping the reaction, the supernatants were collected for the determination of TXB2 or PGE2 by highly specificELISAs;theresultsareexpressedasng-mg ~protein x 30rain.

Determination of cyclooxygenase-1 protein expression by western blots Western blots were performed essentially as described [24]. The appropriate amounts of protein were subjected to SDS-PAGE (10% acryl-bisacrylamide,4 % precast, 1% SDS) after being denatured by heating for 5 min, at 95 °C, in L~mmli buffer. After electrophoresis, proteins were transferred to an Immobilon pR nylon membrane (Millipore) by semi-dry blotting, and after blocking, they were incubated with a 1/200 dilution of the polyclonal rabbit anti-sheep cyclooxygenase antiserum. Bound antibodies were incubated in two steps with biotinylated donkey anti-rabbit IgG antibody and a streptavidin/biotinalkaline phosphatase complex. Colour development was performed with NBT and phenanzine methosulphate.

Statistical analysis Control experiments to determine the cytostatic and cytotoxic effects on MC or the antiproliferative effects of CSA/FK-506 on human PBL were performed at least three times in 96-well plates, each with 6 replicates. The induction of mesangial cell TXB2 or PGF~ release by IL-lfl, and the actions of the immunosuppressants, were measured in 4-6 separate experiments, in 24-well plates in triplicate, with determination of the eicosanoid/protein formation in subsequent assays in additional duplicate incubations. Membrane enzymatic activities representing either the combined cyclooxygenase/thromboxane synthase (TXB2 produced) or cycloocygenase/prostaglandin E2 activity (PGEz produced) were measured in 3-7 separate experiments. The significance of differences was calculated by Student's t-test as mentioned in the Results.

Results

Antiproliferative and cytotoxic effects of CSA and FK-506 W e first e x c l u d e d n o n s p e c i f i c i n t e r f e r e n c e b y c y c l o s p o r i n with t h e t h r o m b o x a n e assay s y s t e m ( n o t shown). Subs e q u e n t e x p e r i m e n t s w e r e p e r f o r m e d to d e t e r m i n e t h e

S13 mesangiolysis (75 % D N A staining compared to 100% control), whereas CSA, as seen by microscopy, led to significant lysis of MC at concentrations above 2 gg. ml-1 (ICs0 = 4.5 gg. ml- 1), starting with vacuolisation of the cytoplasm, membrane bleb formation and disintegration of the nuclei.

"-'csA

"~ E

.

I1~11

o. o ._1 ¢¢o

"

\l

Inhibition of release of glomerular mesangial cell TXB2 by CSA and FK-506 in comparison to their immunosuppressive activities

1.0,

o o

"



_c u •,=

FKSO6

e,

0.s

The known immunsuppressive activities of CSA and FK506 [20, 21] on agglutinin stimulated human lymphocytes are shown in Fig. 1A; FK-506 was 100-fold more potent than CSA. These experiments were performed in parallel to the determination of the inhibitory effect of the immunosuppressants on the interleukin-lfl (200 U.ml-1) stimulated glomerular mesangial cell release of TXB2 using identical serial dilutions (Fig. 1 B). This experimental setting permitted a direct comparison of the specific action of CSA and FK-506 on lymphocytes with the "nonspecific" effect on mesangial cell thromboxane formation. FK-506 inhibited PHA-stimulated lymphocyte proliferation with an ICs0 < 1 ng-ml 1, whereas it did not inhibit mesangial cell TXB2 release at concentrations up to 5 gg. ml-1. In contrast, CSA inhibited both lymphocyte activation and stimulated mesangial cell TXB2 release in comparable concentration ranges (ICs0 = 50 ng/ml).

v

~B'~- •'lIra ~ i-eii

-

ixii._i 0.0 contr. PHA

0.01

0.1

A

1.0

10

I0 L

+ CSA/FK506[ng/ml] 400

m~m

T

"\

oID



-



200

I--

!

• - - I I FK506 •--•

100

Time-dependent effect of CSA on stimulated glomerular mesangial cell release of TXB2 and PGE2 1~

CSA

.L contr.

B

IL-lp

)4 + CSA/F'K506[ng/ml]

Fig. 1. A Specificimmunosuppressive effects of CSA and FK-506on human peripheral blood lymphocytes (PBL). White blood cells, isolated as described in "Methods", were adjusted to a density of i x 10/ml, 6 and were incubated in 10 % FCS medium alone or stimulated with 5 gg. ml- 1phytohaemagglutinin (PHA). Serial dilutions of CSA or FK-506 were added to the stimulated samples in microtiter plates and [3H]-thymidineincorporation was determined after 24 h. B Inhibition of glomerular mesangial cell thromboxane B2 (TXB2) release by CSA and FK-506. Simultaneously with the lymphocyte tests (A), the serial dilutions of CSA and FK-506 were added to confluent mesangial cells cultured in 24-wellplates. The results in A and B represent the means of five separate experiments, with 6 replicates for the PBL's, and triplicates ( + duplicates in the TXB2 ELISA) for the mesangial cell assay

dose range of the cytostatic and cytotoxic actions of CSA or FK-506 on glomerular mesangial cells (not shown). As in previous studies [19], the ICs0 of CSA-induced inhibition of D N A synthesis was about 1.0 gg. ml-1. FK-506 inhibited mesangial cell proliferation in a comparable concentration. On the other hand, FK-506, even at the concentration of 10 gg. ml- 1, produced almost no visible

To further define the specificity of the effect of CSA on mesangial cell TXB2 release, we examined the time dependency of the CSA action on TXB2 production (Fig.2A) in comparison with mesangial PGEz release (Fig. 2B). The data from one representative (in triplicate) of four similar experiments showed that IL-lfl (200 U. ml- 1) caused 2-3 fold stimulation of MC secretion of TXB2 between 6-48 h. This stimulated release was significantly inhibited by i ng. m l - 1CSA (P < 0.05 for 24, 36, 48 h), while 1 gg. ml- 1 completely abolished thromboxane sythesis already after 6 h (P < 0.01). PGE2 secretion by MC gradually appeared with time after IL-lfl (200 U. ml 1) stimulation, reaching a plateau after 24 h, with a 20-fold increase compared to basal PGE2 secretion (Fig. 2 B). CSA only marginally inhibited mesangial PGE2 release at low or high concentrations.

Inhibition of the mesangial cell membrane cyclooxygenase/thromboxane synthase or cyclooxygenase/prostaglandin E2 synthase activity by CSA After a 24 h stimulation with IL-lflin the presence or absence of CSA, crude membrane fractions were isolated and their arachidonic acid metabolising activity determined by measuring either TXB2 or PGE2 (not shown). As in previous experiments evaluating the amount of eicosanoids secreted by intact MC (Figs. 1, 2), IL-lflin opt•-

$14 kDa

kDa

IL-111 + C S A :

- ._ n ,m,2 +r,

97 - 66 - -

,..-,600 c IL-1B

.

...97 .66

45 - -

~45

31 - -

~31

/

Q. ~

~)400

vL ~

~11

25 T2.5 microsomo[

2.5

0.5

p r o t e i n (lag)

+

+ + +

control

~X200

T

12

A

r

r

r

24 time [h]

36

T

I

48

40 IL-11"1 + CSA: -m

[1 ng/ml]4~.[1 pg/ml] I L - 1 B ~ ' *

+ +

medium 1L-1-13 CsA (t ng/m[)

Fig.& Influence of CSA on the basal and interleukin-lfl(IL-lfl) induced expression of glomerular mesangial cell cyclooxygenase-1 protein determined in immunoblots with a specific rabbit anti-sheep COX-1 antibody (see "Materials and Method"-section). Mesangial cells were treated with IL-lfl with or without 1 ng- ml- 1CSA in culture flasks for 24 h. The left blot shows that staining of the 68 kDa COX-1 protein band was dependent on the protein concentration. Based on these results the right immunoblot was performed with 12.5 gg mesangial membrane protein initially loaded onto each lane in order to detect differences in the amount of COX-1 protein expressed by MC. Neither the 24 hour IL-lfl stimulation nor the presence of CSA significantlychanged the expression of cyclooxygenase

~'3o

Q. O~ E 20

Direct inhibition of mesangial cell thromboxane synthase by CSA

c

According to the results of the COX/TXS and COX/PGS enzyme assays, the putative common enyzmes involved in the synthesis of both products, namely phospholipase A2 and cyclooxygenase, seemed not to contribute to the observed effects. In a first approach to prove this point we subjected the isolated membrane protein fractions of MC to immunoblotting with a specific cyclooxygenase antibody. Crude membranes were prepared after a 24 h incubation of MC in medium (control) or with IL-lflin the absence or presence of i ng- ml- ~ CSA. The Fig. 3 shows in the left blot that a single 68 kDa band was detected, and second, that the intensity of the bands stained with the COX-1 antiserum was dependent on the amount of membrane protein loaded onto the gel (0.5 to 25 gg). This result represents a suitable prerequisite in order to define quantitative differences in COX-1 protein expression in MC. However, as can be seen in the right immunoblot, neither significant induction of COX-1 protein by IL-lfl nor any reduction in basal or stimulated expression by CSA was found. In particular, based on the very rapid inhibition of TXB2 release by CSA already apparent after 6 h in the kinetic experiments (Fig.2A), we assumed that CSA might act as a direct inhibitor of the P-450 cytochrome enzyme, thromboxane synthase. In a final experimental series (n = 5), we added CSA (0.1, 1.0, 1000 rig-ml ~-1)directly to the cyclooxygenase/thromboxane synthase assay. Fig. 4 demonstrates that the TXB2-producing activity of membranes isolated from mesangial cells stimulated for 24 h with IL-lfl (200 U . ml- 1) was immediately reduced by CSA. Even a very low CSA concentration (0.1 ng. m1-1) caused a 63.7 % decrease in TXB2 synthesis from exogenous arachidonic acid.

u2 a. lO-

+," o

I

°o=, I

I

i

24 36 48 B time [hi Fig.2A, B. Time-dependent inhibition of (A) stimulated glomerular mesangial cell thromboxane B2 and (B) prostaglandin E2 release by CSA (CSA): confluent rat MC in triplicates in 24-wellplates were exposed for the appropriate times either to 1 rig. ml- 1or 1 gg. mlCSA in the presence of IL-1/~ [200 U.ml-~]. Data represent the mean of triplicates of on representative experiment out of a series of four similar tests. The standard deviation did not exceed 15 % and has been omitted for reasons of clarity o

12

mal concentrations enhanced the TXB2-producing capacity of MC membranes by 2-3-times, while 500 U. ml- 1 of this cytokine raised the PGE2 production by the COX/PGE2 synthase enzyme complex 13-fold. CSA (0.1, 1.0 and 1000 ng. m1-1) present during the 24 h incubation reduced the IL-lfl stimulated COX/thromboxane synthase activity dose dependently by up to 76 % (diclofenac inhibitable control: 0.60 (0.14); IL-lp: 2.57 (0.56); IL-1/~ + 0.1 ng.m1-1 CSA: 1.47 (0.07); IL-lfl + i ng.m1-1 CSA: 1.08 (0.02) ng .ml -~ protein x 30 min, respectively). The capacity of membranes from CSA pretreated MC to synthesise PGE2 was only inhibited to a minor extent (control: 1.54 (0.29); IL-lfl: 22.0 (3.0); IL-lfl + 1 gg.ml -~ CSA: 15.0 (1.0) ng. m!- 1protein x 30 • i n , respectively).

$15 10.0

E 8.0Q X C

T

'~ 6.0 el

4.0£ 2.0

0.0 -

control

11-113

~

0.1 1.0 1000 + cyclosporin [ng/ml]

Fig.4. Immediate inhibition of intefleukin-lfl (IL-lfl) stimulated cyclooxygenase/thromboxanesynthase activityby CSA added directlyto the enzymaticassaysamples. In contrastto the experiments described so far, where CSA was present throughoutthe incubation period prior to the membrane protein preparation, in this experimental series CSA was added 10 min before the enzymaticreaction withisolatedMC membraneswas startedby the additionof 10 5 M arachidonicacid

Discussion

The most significant finding in the experimental series presented here is that CSA inhibited glomerular mesangial cell release of thromboxane at concentrations comparable to those required to produce its immunosuppressive activities. Based on initial experiments CSA seemed to exert this effect by a specific and direct interaction with thromboxane synthase. Several lines of experimental evidence were accumulated in support of this suggestion. First neither nonspecific cytotoxic nor cytostatic effects of CSA could account for this activity. Although CSA was definitely more cytotoxic than FK-506 (albeit having similar cytostatic activity), the micromolar concentration (ICs0 = 4.5 ~tg. ml- 1) of CSA leading to mesangiolysis was two orders of magnitude above that inhibiting thromboxane production (ICs0 = 50 ng. ml-1). In addition, parallel lymphocyte proliferation assay not only confirmed that serial dilutions of CSA and FK-506 inhibited lymphocyte activation in the well known concentration range (ICs0 = 50 ng. ml-~ and 1 ng. ml 1, respectively), but clearly established that TXB2 release was affected by CSA in comparable concentrations. Similarly, in the subsequent experiments the enzymatic capacity of stimulated MC membranes to synthesise TXB2 was significantly reduced by even lower doses of CSA. Again specificity of the CSA effect was observed, as PGE2 synthesis was only marginally attenuated. In agreement with our proposition that CSA might directly affect specific "distal enzymes" of the arachidonic acid metabolisation cascade are the profound quantitative differences in the IL-lfl effect on TXB2 compared with PGE2 secretion and synthetic capacity (2-4-fold vs 10-20-fold, respectively). Finally, the immediate and almost complete inhibition of

TXB2 formation by the cyclooxygenase/thromboxane synthase complex produced by CSA directly added to the enzyme assay, together with the lack of effect on COX-1 protein expression, strongly supports our assumption. Several recent studies have described effects of CSA on MC [8, 19, 26] and other cells [27]. Investigations using micromolar doses of CSA [14] are not really comparable to the present study as our data showed that at such concentrations CSA started to be cytotoxic. Additional experiments not presented here clearly showed that CSA in doses higher than 2 gg. ml- l led to a toxic influx of calcium into MC. Working closer to the present concentration range Fan et al. [28] found that CSA 100 ng. ml- 1reduced prostacyclin production by rat peritoneal macrophages by 25 %, albeit significance was reached only at 300 ng/ml CSA. It should be mentioned that prostacyclin synthase represents a P-450 enzyme like thromboxane synthase [29]. In partial agreement with our data, in short-term incubations the generation of basal or stimulated PGE2/PGFla by mesangial cells was decreased by borderline to high doses of CSA in two other studies [30, 31]. The authors found evidence of inhibition of phospholipase A2 by CSA, thus reducing the availability of arachidonic acid. We did not examine the effect of CSA on phospholipase A2. However, in our enzyme assay system, the requirement of a substrate delivering phospholipase was avoided by the addition of exogenous arachidonic acid. Under these conditions there was still specific reduction in TXB2 synthesis. Recent publications have suggested the presence of a second "inducible" cyclooxygenase (represented by a 4.1 kB RNA) [32], which, as an alternative to the protein product of the 2.8 kB cyclooxygenase RNA detected by D. L. DeWitt's antibody in our immunoblots, might be the target of the CSA action. Again, although we could not rule out an effect of CSA on a putative COX-2 form, the observed differential effects of CSA on TXB2 and PGE2 synthesis made this idea less likely. In the end, we are left with our initial hypothesis that CSA might preferentially interfere with cytochrome P-450 enzymes, such as prostacyclin or thromboxane synthase. In this respect in should be noted that both the microsomal P-450 oxygenases mainly responsible for CSA metabolism [33, 34] and the thromboxane synthase seem to belong to a common P-450 IIIA subfamily [29]. However, recent data reveal that FK506 is also metabolized by this class of P-450 III cytochromes [34]. Therefore, a similar correlation to that drawn for CSA, namely that the inhibitory activity is due to specific binding to P-450 III cytochromes, should also apply to FK-506. But, as shown by our data, FK-506 had almost no effect on TXB2 formation. Possibly related to this discrepancy, a recent study described how, unlike FK506 or an analogue (FK-520), long-term CSA treatment resulted in a 67 % increase in the total renal P-450 cytochrome content [35]. However, there is no indication that thromboxane synthase is also inducible by CSA. In conclusion, we obtained evidence of specific inhibition of glomerular mesangial cell TXB2 release by CSA, which seemed to be dependent on an interaction of CSA with the homologous P-450 III enzyme, thromboxane synthase. Although possibly not relevant to CSA-associated nephrotoxicity in vivo, these findings stress the possibility that CSA in very low concentrations, e. g. in the range re-

$16

quired for its immunosuppressive activity, might have a specific effect on arachidonic acid metabolising P-450 cytochromes. Further studies are necessary directly to prove the interaction of CSA with such arachidonic acid metabolising P-450 enyzmes.

Acknowledgements. This work was supported by Deutsche Forschungsgemeinschaft grant SFB 244/B1. We appreciate the excellent technical assistance of Mrs. J. vonder Ohe.

References 1. Brabletz T, Pietrowski I, Serfling E (1991) The immunosuppressives FK 506 and cyclosporin A inhibit the generation of protein factors binding to the two purine boxes of the interleukin 2 enhancer. Nucleic Acids Res 19:61.67 2. Flanagan WM, Corth6sy B, Bram RJ, Crabtree GR (].991) Nuclear association of a T-cell transcription factor blocked by FK506 and cyclosporin A. Nature 352:803-807 3. Liu J, Farmer JD, Jr., Lane WS, Friedman J, Weissman I, Schreiber SL (199I) Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell 66: 807815 4. Kimball PM, Kerman RH, Kahan BD (1991) Production of synergistic but nonidentical mechanisms ofimmunosuppression by rapamycin and cyelosporine. Transplantation 51:486M90 5. Sigal NH, Dumont F, Durette R Siekierka JJ, Peterson L, Rich DH, Dunlap BE, Staruch M J, Melino MR, Koprak SL, Williams D, Witzel B, Pisano JM (1991) Is cyclophilin involved in the immunosuppressive and nephrotoxic mechanism of action of cyclosporin A. J Exp Med 173:61%628 6. Devarajan R Kaskel FJ, Arbeit LA, Moore LC (1989) Cyclosporine nephrotoxicity: blood volume, sodium conservation, and renal hemodynamics. Am J Phsio1256:FTl-F78 7. Kon V, Sugiura M, Inagami ~I, Harvie BR, Ichikawa I, Hoover RL (1990) Role of endothelin in cyclosporine-induced glomerular dysfunction. Kidney Int 37:1487-1491 8. Pfeilschifter J (1988) Cyclosporin A augments vasoconstrictorinduced rise in intracellular free calcium in rat renal mesangial cells. Biochem Pharmaco137:4205-4210 9. Locher R, Huss R, Vetter W (1991) Potentiation of vascular smooth muscle cell activity by cyclosporin A. Eur J Clin Pharmacol 41:297-301 10. Griffiths EJ, Halestrap AP (1991) Further evidence that cyclosporin A protects mitochondria from calcium overload by inhibiting a matrix peptidyl-prolyl cis-trans isomerase. Implications for the immunosuppressive and toxic effects of cyclosporin. Biochem J 274:611-614 11. Perico N, Pasini M, Gaspari F, Abbate M, Remuzzi G (1991) Coparticipation of thromboxane A2 and leukotriene C4 and D4 in mediating cyclosporine-induced acute renal failure. Transplantation 52:873-878 12. Grieve EM, Hawksworth GM, Simpson JG, Whiting PH (1990) Effect of thromboxane synthetase inhibition and angiotensin converting enzyme inhibition on acute cyclosporin A nephrotoxicity. Biochem Pharmaco140(10): 2323-2329 13. Makowka L. Lopatin W, Gilas T, Falk J, Phillips MJ, Falk R (1986) Prevention of cyclosporine (CyA) nephrotoxicity by synthetic prostaglandins. Clin Nephro125 [Suppl 1]: $89-$94 14. McCauley J, Studer R, Craven R Murray S (1991) The effects of cyclosporine A, cyclosporine G, and FK 506 upon prostaglandin production in renal mesangial ceils in culture. Transplant Proc 23:3141-3142 15. Heering R Strobach H, Schr6r K, Grabensee B (1992) The role of thromboxane and prostacyclin in ciclosporin-induced nephrotoxicity. Nephron 61:26-31 16. Schnabel FR, Wait RB, Kahng KU (1991) The relationship of urinary thromboxane excretion to cyclosporine nepbrotoxicity. Transplantation 51:686-689 17. Radeke HH, Meier B, Topley N, F16ge J, Habermehl GG, Resch K (1990) Interleukin 1-a and tumor necrosis factor-a induce

oxygen radical production in mesangial cells. Kidney Int 37: 767775 18. DeWitt DL, EI-Harith EA, Kraemer SA, Andrews M J, Yao EF, Armstrong RL, Smith WL (1990) The aspirin and heme-binding sites of ovine and murine prostaglandin endoperoxide synthases. J Biol Chem 265(9): 5192-5198 19. Radeke HH, Christians U, Bleck JS, Sewing K-F, Resch K (1991) Additive and synergistic effects of cyclosporine metabolites on glomerular mesangial cells. Kidney Int 39:1255-1266 20. Radeke HH, Christians U, Sewing K-F, Resch K (1992) The synergistic immunosuppressive potential of cyclosporin metabolite combinations. Int J Immunopharmac 14(4): 595-604 21. Christians U, Radeke HH, Kownatzki R, Schiebel HM, Schottmann R, Sewing K-F (1991) Isolation of an immunosuppressant metabolite of FK 506 generated by human microsome preparations. Clin Biochem 24:271-275 22. Reinke M, Piller M, Brune K (1989) Development of an enzymelinked immunosorbent assay of thromboxane B2 using a monoclonal antibody. Prostaglandins 37(5): 577-586 23. Kaever V, Goppelt-Strabe M, Resch K (1988) Enhancement of eicosanoid synthesis in mouse peritoneal macrophages by organic mercury compound thimerosal. Prostaglandins 35:885-902 24. Radeke HH, Cross AR, Hancock JT, Jones OTG, Nakamura M, Kaever V, Resch K (1991) Functional expression of NADPH oxidase components (a- and fi-subunits of cytochrome bs58and 45-kDa flavoprotein) by intrinsic human glomerular mesangial cells. J Biol Chem 266:21025-21029 25. Topley N, Floege J, Wessel K, Hass R, Radeke HH, Kaeyer V, Resch K (1989) Prostaglandin E2 production is synergistically increased in cultured human glomerular mesangial cells by combinations of IL-1 and tumor necrosis factor-a. J Immunol 143: 198%1995 26. Rodriguez-Puyol D, Lamas S, Olivera A, L6pez-Farr6 A, Ortega G, Hernando L, Ldpez-Novoa JM (1989) Actions of cyclosporin A on cultured rat mesangial cells. Kidney Int 35:632-637 27. McCauley J, Farkus Z, Prasad SJ, Plummer HA, Murray SA (1991) Cyclosporine and FK 506 induced inhibition of renal epithelial cell proliferation. Transplant Proc 23:2829-2830 28. Fan T-PD, Lewis GP (1985) Mechanism of cyclosporin A-induced inhibition of prostacyclin synthesis by macrophages. Prostaglandins 30(5): 735-747 29. Yokoyama C, Miyata A, Ihara H, Ullrich V, Tanabe T (1991) Molecular cloning of human platelet thromboxane A synthase. Biochem Biophys Res Commun 178:1479-1484 30. Walker R J, Lazzaro VA, Duggin GG, Horvath JS, Tiller DJ (1990) Structure-activity relationships of cyclosporines: Inhibition of angiotensin II-stimulated prostaglandin release in cultured rat mesangial cells. Transplantation 50:343-345 31. Bunke M, Wilder L, Martin A (1991) The effect of cyclosporine on agonist-stimulated glomerular and mesangial cell vasodilatory prostaglandin production. Transplantation 52:718-722 32. Xie W, Chipman JG, Robertson DL, Erikson RL, Simmons DL (1991) Expression of a mitogen-responsive gene encoding prostaglandin synthase is regulated by mRNA splicing. Proc Natl Acad Sci USA 88:2692-2696 33. Lucey MR, Kolars JC, Merion RM, Campbell DA, Aldrich M, Watkins PB (1990) Cyclosporin toxicity of therapeutic blood levels and cytochrome P-450 IIIA. Lancet 335:11-15 34. Sattler M, Guengerich FR Yun C-H, Christians U, Sewing K-F (1992) Human and rat liver microsomal cytochrome P~s03A4is involved in biotransformation of FK 506 and rapamycin. Biochem Pharmacol (in press) 35. Vincent SH, Wang RW, Karanam BV, Klimko M, Alvaro R, Lee Chin S-H (1991) Effects of the immunosuppressant FK-506 and its analog FK-520 on hepatic and renal cytochrome P450 mixedfunction oxidase. Biochem Pharmaco141(9): 1325-1330 Dr. H.H. Radeke Institut ffir Molekularpharmakologie Medizinische Hochschule OE5320 Konstanty-Gutschow-Strasse 8 W-3000 Hannover 61, Germany