Role of transforming growth factor-β1 in experimental chronic cyclosporine nephropathy

Kidney International, Vol. 49 (1996), pp. 1141—1151

Role of transforming growth factor—1 in experimental chronic cyclosporine nephropathy FUAD S. SHIHAB, TAKESHI F. ANDOH, AMIE M. TANNER, NANCY A. NOBLE, WAYNE A. BORDER, NORA FRANCESCHINI, and WILLIAM M. BENNETT Division of Nephrology, University of Utah and Department of Veterans AJJdrs Medical Center, Salt Lake City, Utah, and Division of Nephrology, Hypertension and Clinical Pharmacology, Oregon Health Sciences University, Portland, Oregon, USA

Role of transforming growth factor (TGF)-pl in experimental chronic cyclosporine nephropathy. The pathogenesis of fibrosis in chronic cyclosporine (CsA) nephropathy remains unknown. Since TGF-f31 plays a key

role in the fibrogenesis of a number of renal diseases, we studied a salt-depleted rat model of chronic CsA nephropathy which shows similarity to the structural and functional lesions described in patients. Pair fed rats were treated with either CsA (15 mg/kg/day s.c.) or an equivalent dose

of olive oil and sacrificed at 7 and 28 days. Characteristic histologic changes of proximal tubular injury, tubulointerstitial fibrosis and arteriolopathy developed in CsA-treated rats at day 28. They were accompanied by physiologic changes of increased serum creatinine, decreased creatinine clearance, increased enzymuria and decreased concentrating ability. CsA-

treated rats showed a progressive increase in mRNA expression of TGF-1 and matrix proteins at days 7 and 28. Most of the changes were in the tubulointerstitial and vascular compartments by immunofluorescence with a predominant involvement of the medulla as compared to cortex. The mRNA expression of plasminogen activator inhibitor, a protease inhibitor stimulated by TGF-pl, followed TGF-/31 and matrix proteins, suggesting that the fibrosis of chronic CsA nephropathy likely involves the dual action of TGF-p on matrix deposition and degradation.

poorly understood [11]. The observed striped interstitial fibrosis does not improve consistently with discontinuation of the drug and occurs even in patients treated with low doses of CsA [6]. Because the fibrotic lesion of chronic UsA nephrotoxicity has been difficult to reproduce in animals, the lack of an animal model of chronic CsA nephrotoxicity has hampered the study of its pathogenesis. Using the observation that sodium depletion exacerbates CsA nephrotoxicity, a reproducible animal model of chronic CsA nephrotoxicity was recently established [12]. In this model, CsA treatment in rats on low salt diet induced physiologic and histo-

logic features that resemble the human lesion described in patients on long-term CsA therapy [13]. To date, no study has looked at the components of the fibrotic

lesion of chronic UsA nephrotoxicity in vivo and none has attempted to correlate the renal fibrosis or the expression of ECM

components with any growth factor, Transforming growth factor-p (TGF-/3) has been implicated in the fibrosis of a number of chronic diseases of the kidney and other organs [14, 15]. TGF-f3 directly stimulates the synthesis of individual ECM components The introduction of cyclosporine A (CsA) into clinical practice [16—18]. It also blocks ECM degradation by decreasing the has resulted in a major improvement in the short term outcomes synthesis of proteases and stimulating protease inhibitors like of solid organ transplantation [1] and the treatment of autoim- plasminogen activator inhibitor-i (PAl-i) [19]. In addition to its mune diseases [2]. However, with increased use has come the effect on matrix accumulation, TGF-/3 also promotes immune recognition that CsA has severe deleterious effects on renal suppression [20, 21]. Since TGF-13 is a key fibrogenic cytokine, we hypothesized that structure and function making chronic nephrotoxicity a major limiting side effect [3]. Acute renal dysfunction due to CsA is TGF-13 overexpression might he responsible for the fibrosis of reversible since it involves renal hemodynamic dysfunction and is chronic CsA nephrotoxicity. To test this hypothesis, we utilized not associated with any permanent histologic changes [4]. On the this animal model with lesions similar to human chronic CsA other hand, chronic CsA nephrotoxicity may progress to an nephropathy. Our findings indicate that fibrosis is mostly exirreversible renal lesion characterized by striped interstitial fibro- pressed in the medulla, that all major components of the ECM are sis, tubular atrophy and hyalinosis of the afferent arterioles [5, 6]. elevated and that their expression correlates with an increased Chronic CsA nephrotoxicity has been described in kidneys of expression of TGF-p and PAl-i. These results suggest that the recipients of renal and other organ allografts, as well as in patients ECM accumulation observed with chronic CsA therapy likely treated with CsA for autoimmune diseases [7—9]. In addition, involves the dual action of TGF-p on ECM deposition and progression to end-stage renal disease with chronic CsA use is degradation. well documented [10]. However, the exact mechanism underlying the development of fibrosis in chronic CsA nephrotoxicity remains

Methods Experimental design

Received for publication August 23, 1995 and in revised form November 1, 1995 Accepted for publication November 2, 1995

© 1996 by the International Society of Nephrology

Adult, male Sprague-Dawley rats (Charles River, Wilmington, MA, USA), 225 to 250 g, were housed in individual cages in a temperature and light controlled environment, and received a low salt diet (0.05% sodium, Teklad, Premier, WI, USA), with water

1141

1142

Shihab et al: TGF-f31 in CsA nephropathy

ad libitum. After one week on the low salt diet, weight matched pairs of rats were randomly assigned to receive daily subcutaneous

Table I. Physiologic parameters in CsA and placebo (VH) treated rats at 7 and 28 days

injections of CsA 15 mg/kg (Sandimmune®, Sandoz Research Day 7 Institute, East Hanover, NJ, USA), or an identical volume of the CsA VH vehicle olive oil as a control. Control rats were fed the exact 0.37 0.02 0.48 0.02 amount of food consumed the day before by the CsA-treated rats. Sr mg/dl 0.55 0.02 0.41 0.02" Animals were divided into two groups of 16 CsA-treated and 16 Cr ml/min/100 g control rats. Eight rats from each group were studied at days 7 and 28. Body weight was recorded daily. At the end of each treatment

period, 24-hour quantitative urine samples were collected in metabolic cages (Nalge Co., Rochester, NY, USA) and systolic

blood pressure (BP) was measured by tail plethysmography (Narco Bio-systems, Houston, TX, USA). The following day

0.06

FENa %

Urine AAP IU/gCr Urine NAG IU/gCr UV ml/24 hr

animals were anesthetized with ketamine and a blood sample was U0sm mOsm/kg obtained. After opening the abdomen through a midline incision, TcW the abdominal aorta was cannulated retrogradely below the renal ml/24 hr

arteries with an 18-gauge needle. With the aorta occluded by ligation above the renal arteries and the renal veins opened by a small incision for outflow, the kidneys were perfused with 20 ml of

0.01

45 10

25 1

0.04 31

10.2

1.2

12.9

1,848

253

1,037

49±6

0.01

168 34 4 1.4

178"

28±5"

Day 28 VH

CsA

0.37 0.59

0.03 1.04 0.05 0.22

0.12" 0.03"

0.02

0.01 0.05

0.01

52 8

86 21

20

27

1

1

11.6

1.3

24.8

1,344

168

662 110"

34±2

3.2"

23±6

Data are mean SEM of eight rats. Abbreviations are: VH, placebo; Sr, serum creatinine; C, creatinine clearance; FENa, fractional excretion of sodium; AAP, alanine aminopeptidase; NAG, N-acetyl J3-D-glucosamini-

cold heparinized saline. The kidneys were removed, the cortex dase; UV, urinary volume; Uom, urinary osmolality; TcW, free-water was carefully dissected from the medulla, and the tissues were reabsorption. a P < 0.01 vs. VH processed for evaluation by light microscopy, RNA analysis and "P < 0.05 vs. VH immunohistochemistry.

Functional studies

biopsy by an observer masked to treatment groups using a color Urinary and plasma sodium were measured by flame photomanalyzer (Olympus CIA-102, Olympus Inc., Tokyo, Japan). eter (Instrumentation Laboratories, Lexington, MA, USA). Uri- image The following semiquantitative score was used to asses the

nary and plasma osmolality were measured by freezing point depression (Osmette A, Precision Systems Inc., Natick, MA, USA). Urinary and serum creatinine, urinary protein (Uprot), urinary alanine aminopeptidase (AAP) and urinary N-acetyl

extent of changes in each category. For tubular injury, the following score was utilized: 0 no tubular injury; 0.5 = <5% of

/3-D-glucosaminidase (NAG) were measured by a Cobas autoanalyzer (Roche Diagnostics, Div. Hoffman-La Roche Inc., Nutley, NJ, USA). CsA blood levels were determined in whole blood by a specific radioimmunoassay (Incstar Corp., Stillwater, MN, USA).

65% of tubules injured and 3 = > 65% of tubules injured.

tubules injured; 1 = 5 to 20% of tubules injured; 1.5 = 21 to 35% of tubules injured; 2 = 36 to 50% of tubules injured; 2.5 = 51 to

Tubulointerstitial fibrosis was estimated by counting the percent-

age of injured areas per field of cortex and medulla, and was

scored semiquantitatively using the following: 0 = normal interThe creatinine clearance (Ce,.), fractional excretion of sodium stitium; 0.5 = < 5% of areas injured; I = 5 to 20% of areas (FENa) and free water reabsorption (TcW) were calculated using injured; 1.5 = 21 to 35% of areas injured; 2 = 36 to 50% of standard formulae. areas injured; 2.5 = 51 to 65% of areas injured and 3 = > 65% Morphology

Renal tissue samples were fixed in 10% buffered formalin and embedded in paraffin. Two to four micron thick sections were stained with periodic acid-Schiff's reagent (PAS) and Trichrome. Horizontal and sagittal sections were both performed. The histologic findings were subdivided into three categories: proximal tubular injury, interstitial inflammation and scarring, and arteriolopathy. Findings ascribed to tubular injury included cellular and intercellular vacuolization, tubular collapse (unassociated with interstitial fibrosis or tubular membrane thickening) and tubular distention. Features of interstitial inflammation were mononuclear infiltrates, edema and vacuolization of interstitial cells. The findings of scarring were matrix-rich expansion of the interstitium

of areas injured. Hyalinosis of the afferent arterioles was semiquantitatively determined by counting the percentage of juxtaglomerular afferent arterioles available for examination with a minimum of 100 glomeruli per biopsy assessed: 0 = no arterioles injured; 0.5 = < 15% of arterioles injured; 1 = 15 to 30% of arterioles injured; 1.5 = 31 to 45% of arterioles injured;

2=

46

to 60% of arterioles injured; 2.5 =

61

to 75% of

arterioles injured and 3 = > 75% of arterioles injured. RNA analysis

After separating the cortex from the medulla, the tissue was finely minced with a razor blade on ice. Total RNA was prepared by lysis in 4 M guanidine isothiocyanate containing 1% 2-mercaptoethanol and 0.5% lauryl sarcosyl and ultracentrifugation of the

with distortion of the tubules and thickening of the tubular lysate on a cesium chloride cushion. After resuspending in basement membranes. Renal arteriolopathy in chronic CsA- Tris-EDTA buffer, RNA concentrations were determined using induced nephrotoxicity was characterized by hyalinization and destruction of the afferent arterioles. The hyalinosis consisted of hyalin deposition within the tunica media of afferent arterioles and terminal portion of interlobular arteries. A minimum of 20 fields at lOOx magnification were assessed and graded in each

spectrophotometric readings at Absorbance260. Thirty micrograms of RNA were electrophoresed in each lane in 0.9% agarose gels containing 2.2 M formaldehyde and 0.2 M Mops (pH 7.0) and transferred to a nylon membrane (ICN Biomedicals, Inc., Costa Mesa, CA, USA) overnight by capillary blotting. Nucleic acids

1143

Shihab et al: TGF-/31 in CsA nephropathy

a

.5-..

4

Fig. 1. Histologic changes in experimental chronic CsA nephrotoxicily. Micrograph showing the renal cortex of a salt-depleted rat given CsA, 15 mg/kg/day

during four weeks. Striped tubulointerstitial fibrosis and tubular atrophy are extensive in this animal (A). There is also focal smooth muscle injury of the glomerular afferent arteriole showing hypertrophied and granular eosinophilic transformation (B) (periodic acid-Schiff, magnification X200).

1144

Shihab et al: TGF-131 in CsA nephropathy

were crosslinked by ultraviolet irradiation (Stratagene, La Jolla, CA, USA). The membranes were prehybridized for two hours at 42°C with 50% formamide, 10% Denhardt's solution, 0.1% SDS,

5X standard saline citrate (SSC), and 200 j.tg/ml denatured

Table 2. Semiquantitative scoring by light microscopy of the extent of changes observed in the chronic CsA lesion for proximal tubular injury, tubulointerstitial fibrosis and arteriolopathy

Day 7

Day 28

salmon sperm DNA. They were then hybridized at 42°C for 18 VH VH CsA hours with cDNA probes labeled with 32P-dCTP by random Proximal tubular injury 0.1 0.1 0.6 0.2 0.1 0.1 oligonucleotide priming (Boehringer Mannheim Corp., Indianap- Tubulointerstitial fibrosis 0 0 0 0 0.1 0.1 olis, IN, USA). The blots were washed in 2 >< SSC, 0.1% SDS at Arteriolopathy 0 0 0 0 0.1 0.1 room temperature for 15 minutes and in 0.1 x SSC, 0.1% SDS at Data are mean SEM of eight rats. VH is control. 50°C for 15 minutes. Films were exposed at —70°C for different P < 0.01 vs. VH

CsA 1.5 1.6 1.6

0.2° 0.3° 0.2°

time periods to ensure linearity of densitometric values and

exposure time. Autoradiographs were scanned on a laser densitometer (Ultrascan XL; Pharmacia LKB Biotechnology, Inc.). The density of bands for the glyceraldehyde-3-phosphate dehy- Mann-Whitney test, as appropriate. The level of statistical signifdrogenase (GAPDH) mRNA was used to control for differences icance was chosen as P < 0.05. in the total amount of RNA loaded onto each gel line. For Results quantitative purposes, the values were divided by the density of Physiologic studies bands for GAPDH in the same lane. The following probes were Although weight gain was progressive in the two treatment used for Northern blot analysis. A mouse TGF-/31 cDNA probe (plasmid MUI5) was kindly groups, animals treated with CsA failed to gain as much weight as provided by R. Deiynck [22]. A rat PAl-I cDNA probe [plasmid those receiving vehicle (VH) with the difference achieving statispBluescript SK(—)] was obtained from T.D. Gelehrter [23]. tical significance (P < 0.01) at day 28. CsA-induced hypertension Plasmid P16, which contains a human biglycan insert, was a gift of was not observed in this model; in fact, in CsA-treated rats, mean L.W. Fisher [241. A rat decorin cDNA probe (plasmid pGEM4) systolic BP tended to be numerically lower than in VH-treated was provided generously by K.L. Dreher [25]. A rat procollagen rats, although the difference was not statistically significant. There al cDNA (plasmid paiRl) was obtained from D. Rowe [261. A was a progressive increase in CsA whole blood level in CsArat GAPDH cDNA probe (plasmid pBluescript KS II) was a gift treated animals: 2930 250 ng/ml at day 7 and 5827 301 ng/ml at day 28. from J.M. Blanchard [27]. CsA induced a significant increase in serum creatinine (Sr) Immunohistochemistiy only at day 28 (P < 0.01) (Table 1). On the other hand, the Cc1. Immunofluorescence microscopy was performed on tissues was already lower in the CsA-treated rats at day 7 (P < 0.05) and snap frozen in cold isopentane with a clyostat microtome (Miles was progressively lower at day 28 (P < 0.01) (Table 1). There was Scientific, Naperville, IL, USA) as previously described [28]. Four no significant proteinuria in the CsA-treated rats or the placebo micrometer cryostat sections were fixed in acetone and washed in animals. The fractional excretion of sodium (FENa) was significantly PBS, pH 7.4. The deposition of the following ECM components, decorin, tenascin and fibronectin-cellular containing extra domain decreased compared to normal in CsA-treated and placebo A (fibronectin EDA+) was performed. The primary antibody for animals at all time points, reflecting the low sodium diet (Table 1). decorin was a rabbit anti-human decorin antibody and was CsA caused an increase in the urinary excretion of AAP at days 7 provided by E. Ruoslahti (La Jolla, CA, USA) [29]. The other and 28 that was significant at day 7 (P < 0.01; Table 1). Urinary primary antibodies used were rabbit anti-human tenascin (Life NAG excretion was also higher at both time intervals, but did not Technologies, Gaithersburg, MD, USA) and mouse anti-human reach statistical significance (Table 1). Urinary concentrating fibronectin EDA+ (Sera-lab, Crowley Down, UK). The secondary ability was clearly impaired as indicated by the progressive and antibodies used were fluorescein isothiocyanate (FITC)-conju- significant increase in urinary volume by day 28 (P < 0.01) and gated F(ab')2 donkey anti-rabbit IgG (Jackson Immunoresearch, decrease in U0sm by day 7 (P < 0.05) and day 28 (P < 0.01) (Table West Grove, PA, USA) and FITC-conjugated F(ab')2 rat anti- 1). There was also a decrease in TcW at both time intervals, mouse IgG (Jackson Immunoresearch). A minimum of 20 randomly although it was statistically significant only at day 7 (P < 0.05; selected areas per sample were observed at X250 magnification. Table 1). The severity of staining for the glomerular, tubulointerstitial and Histologic changes vascular compartments was evaluated by an observer blinded to CsA-treated rats had characteristic morphologic findings simithe treatment groups using the following semiquantitative scale: 0 = diffuse, very weak or absent staining, 1 = staining involving lar to the chronic human lesion, mostly evident at day 28. There less than 25%, 2 = staining involving 25% to 50%, 3 = staining was a mononuclear cell infiltrate and an early tubular injury by day involving 50% to 75% and 4 = staining involving 75% to 100%. 7. However, the typical lesion of chronic CsA nephrotoxicity was Photographs were obtained at identical exposure and develop- observed by day 28 (Fig. 1). Findings ascribed to tubular injury ment time intervals. consisted of cellular and intercellular vacuolization, tubular col-

lapse and tubular dilatation. Striped interstitial fibrosis was Statistical analysis

present with progressive inflammatory cell infiltrates associated with thickening of the tubular basement membranes (TBM) and Bowman's capsule. There was focal disruption of the TBM in VH animals were done by two-tailed unpaired Student's t-test or some areas of interstitial infiltrate with loss of tubular definition.

Results are presented as mean

SEM. Multiple comparisons were done by analysis of variance. Comparisons between CsA and

1145

Shihab et al: TGF-131 in CsA nephropathy

D

N

<(I)

0

I> N

cc C'4


0

D a)

I>

-D N-

cJ

Li)

0

-o r;.

I>

a) c'j

<

0 ci)

-c cc

I> c'J

Kb

A

at

— 2.5

B

—3.1

C

—2.6

D

WeWe

—1.8

E — 5.7

—4.7

F

Fig. 2. Northern blotting of mRNA from cortex and medulla in GsA and VH-ireated rat kidneys. Total RNA was isolated from whole cortex and medulla at days 7 and 28 from rats treated with — 1 .3 CsA 15 mg/kg/day or VH control and was hybrid-

sees Cortex

ized with eDNA probes to (A) TGF-j31, (B) PAT-i, (C) biglycan, (D) decorin, (E) type I Medulla

The vascular lesion was characterized by arteriolar hyalinosis

collagen, and (F) GAPDH. Molecular size markers are shown on the right.

Expression of TGF-131 and PAl-i mRNA

consisting of hyalin deposition within the tunica media of afferent

arterioles and the terminal portions of the interlobular arterics. TGF-pl mRNA was progressively increased by day 7 (P < The extent of changes was graded (Table 2) using a 0 to 3 + 0.00 1) and more so at day 28 (P < 0.001) in the CsA-treated rats semiquantitative scale, with 3 + indicating extensive changes. In when compared to control VH rats (Figs. 2 and 3). The mRNA this model, while no early histologic changes were observed by day expression of PAl-i, a protease inhibitor that blocks ECM 7, changes of chronic CsA nephrotoxicity manifested as proximal degradation by the plasmin protease system and is directly tubular injury, tubulointerstitial fibrosis and arteriolopathy devel-

stimulated by TGF-f3, is shown in Figures 2 and 4. PAl-i mRNA,

oped by 28 days (P < 0.01) when compared to VH-treated rat

in parallel to TGF-f31 mRNA expression, was progressively

kidneys.

up-regulated in the CsA-treated rats at seven days (P < 0.001) and

1146

Shihab et al. TGF-pl in CsA nephropathy

1.2

A

**

a 0 C

C

0

**

0.8

**

1 .2

1.0

C (9

9 0.8 Cu

a,

0

0.6

Cu

a,

0 0 (0

a,

Q.Q4

0.6

a,

a,

-a, 0.4

C

2

C

z

0.2

E 0.2 0.0 Cortex

Medulla

0.0 -. Cortex

Fig. 3. 3. Quantitation Qua,:ritatio,:ofofmRNA ,nRNAexpression erpn-ssionofofTGF-/31 TGF-$I in in GsA CsA and and VIIVUFig.

treatedrat putkidneys kid,,n's relative relative to to GAPDH. GAPDH. Total treated Total KNA RNAwas wasisolated isolated from fromwhole whole cortex and and 28 28 and andwas washybridized hybridized with with aa eDNA cDNA Cortex and medulla medulla at at days days 77 and probe to to TGF-1. TGF-$l. Symbols Symbols are: arc: (U) (•) GsA VII 77days; CsA CsA77days; probe days; () VH days;(R)()CsA 28days, days.(D) (D) VII2828 dav. ror each eachgroup. group. "Pcompared 28 VH days. NN = 8 8 for **p < Oil 0.001 comparedto to Vii control. VH control.

1.0

**

Medulla

B

12 ,

Io0

1.0

C (9

Ia

0.8

C (5

0

0

0.6

to

a,

. 0.6

**

(0

a,

a,

0.4 -

C

C

a,

**

z

E 0.2

E

0.2

0.0 -

**

Cortex

0.0

Cortex

Medulla

Fig. 4. Quantitation of mRNA expression of PA 1-1 in GsA and VH-treated rat kidneys relative to GAPDH. Total RNA was isolated from whole Cortex and medulla at days 7 and 28 and was hybridized with a cDNA probe to PAT-i. Symbols are: (U) CsA 7 days; () VH 7 days; () CsA 28 days, () VH 28 days. N = 8 for each group. **p < 0.001 compared to VH control.

more so at 28 days (P <

0.001)

as compared to control VH rats

Medulla

}:jg 5. QuaruitunuitofofntRAl oft/ic pmteohrans. big/sean Fig. 5. Quantitation mRNAexpression expression of the proteoglycans, biglycan and decodpi. in ('sA Cs-I and andVH-treated Vt/-treated rat rut kidneys kidneys relative tv/at/se to to GAPDH. GAPDH. Total decorin, in Total RI RNA was isolated isolated from from whole was whole cortex cortexand and medulla medullaatatdays days 77 and and 2$ 28 and and was hybridized with with aa eDNA eDNA probe probetoto(A) fl I biglycan hi2lvcan and and(B) I Ridecorin. deenririSymbols S nib hybridized are: (U) CsA 7 days; () VH 7 days; () CsA 28 days, () VH 28 days. N = 8 for each group. *D < 0.005 and ** < 0.001 compared to VU

control.

elevated in this model suggesting active matrix synthesis (Figs. 2 and 5). Biglycan mRNA expression closely followed TGF-131 mRNA expression and was progressively elevated at seven days

(P < 0.005 for cortex and < 0.001 for medulla) and at 28 days (P < 0.001) when compared to control VH. The behavior of the Extracellular matrix deposition in the kidney other proteoglycan, deeorin, was somewhat different. At day 7, it Northern blot analysis. The proteoglycans biglycan and decorin was not statistically elevated when compared to VH control but are components of the ECM. Their mRNA expression was was definitely up-regulated by day 28 (P < 0.001). Collagen type (Fig. 4).

1147

early expression of fibronectin EDA+ at day 7 (P < 0.001) that

decreased by day 28 to control level. The deposition of decorin followed its mRNA expression and that of tenascin and fibronectin EDA+ (Figs. 7 and 8). Most of the changes were observed at day 7 in the vessels (P < 0.001) and tubulointerstitium (P < 0.001) with no elevated expression in the glomeruli when compared to VH controls.

a

mANA expression relative to GAPDH

a,

Shihab et at: TGF-pI in CsA nephropathy

0)

Differential expression in cortex and medulla

The differential expression in the cortex versus the medulla for TGF-/31, PAt-I and the ECM proteins was examined by Northern

I')

blots. As shown in Figure 9, the changes observed in mRNA

\\

C

-&

**

expression for TGF-/31, PAl-i, type I collagen and the proteoglycans, biglycan and decorin, although present in the cortex, was

Fig. 6. Quantitation of mRNA expression of type I collagen in CsA and VH-treated rat kidneys relative to GAPDH. Total RNA was isolated from whole Cortex and medulla at days 7 and 28 and was hybridized with a cDNA probe to type I collagen, N = 8 for each group. Symbols are: (•) CsA 7 days; () VH 7 days; () CsA 28 days, () VH 28 days. *P < 0.005 and ** < 0.001 Compared to VH control.

more dramatic and significant (P < 0.001) in the medulla as compared to the cortex at both days 7 and 28. Discussion This report describes an animal model of chronic CsA nephrotoxicity that mimics the functional and histopathological findings

observed in the human lesion of chronic-CsA induced injury. Salt-depletion has been shown to accelerate CsA nephropathy with the lesion developing by three to four weeks in the rat [12] instead of three months or longer [30]. In this model, we found that the histologic changes were accompanied by physiologic changes namely increased creatinine, decreased creatinine clearance, increased tubular enzymuria, and a loss of medullary concentrating ability. While enzymuria stabilizes, the other phys-

I is normally present in the interstitial compartment of the kidney and is directly increased by TGF-13. In this model, the mRNA expression of type I collagen followed TGF-131 and PAl-i mRNA expression and was progressively up-regulated at day 7 (P < 0.005 for cortex and < 0.001 for medulla) and day 28 (P < 0.001; Fig. 6).

Immunohistochemisity. Certain components of the ECM were also examined by immunotluorescence in order to assess whether the increased mRNA expression of matrix proteins was translated into increased protein synthesis and deposition into matrix. This allowed us to assess which compartments of the kidney (glomeruli, vessels or tubulointerstitium) were most extensively involved in matrix expansion. We examined the expression of tenascin, fi-

iologic markers of CsA toxicity progress with continued treatment. The matrix expansion observed in this model was most marked in the interstitial and vascular compartments of the kidney. This is

in accordance with the characteristic histology of chronic CsA nephropathy that is interstitial fibrosis and arteriolopathy [5]. Although present in the cortex, the observed changes were mostly in the medulla and were manifested by a dramatic increase in the

presence of TGF-pl, PAl-i and matrix proteins. These changes were associated with a decrease in medullary concentrating ability. In accordance with our observation, recent data have shown that the early changes of CsA nephropathy are first seen in the medullary compartment of the kidney [31, 32]. The common belief is that the chronic form of CsA injury is a

bronectin EDA+, and decorin. The amount of tenascin and consequence of CsA-induced renal vasoconstriction. According to fibronectin EDA+ was elevated in the kidneys of CsA-treated rats as compared to VH controls at both days 7 and 28 in all the kidney compartments (Figs. 7 and 8). The changes were most significant in the tubulointerstitium for both tenascin (P < 0.005 at days 7 and 28) and fibronectin EDA+

this notion, the arteriolar vasospasm with resultant impairment in

and significant increase in the expression of tenascin and fibronec-

tubular atrophy and interstitial fibrosis actually progressed [13]. In

tin EDA+ in the vascular compartment of the kidneys of CsAtreated rats. While it was significant for fibronectin EDA+ at day 7 (P < 0.005) and day 28 (P < 0.005), the expression of tenascin did not reach statistical significance until day 28 (P < 0.01) when compared to VH controls (Fig. 7). Most of the changes reached

our model, while acute tubular enzymuria stabilized, the associ-

statistical significance in the tubulointerstitium and vascular com-

proteoglycans, glycoproteins, and collagens.

renal blood flow leads to tubulointerstitial fibrosis. Data to

support the role of ischemia as a primary stimulus for tubulointerstitial fibrogenesis have been offered [33]. However, more recent studies have suggested that there can be a dissociation (P < 0.005 at day 7 and < 0.01 at day 28; Fig. 8). The arteriolar between the functional and histological injuries induced by CsA hyalinosis observed in this model correlated with a progressive [13]. CsA withdrawal led to improved GFR but the degree of

partments of the kidney, in accordance with the characteristic

ated histologic changes continued to progress during the follow-up period and were dominated by renal fibrosis. Our findings

also indicate that the observed fibrosis was associated with an increased deposition of the three major components of the ECM: Certain matrix components have been shown to be up-regu-

histology of chronic CsA nephrotoxicity (Fig. 8). In the glomeruli, lated with the administration of CsA. In cell cultures treated with

none of the changes were statistically significant except for an

CsA, collagen type I and IV mRNAs were elevated in murine

1148

Shihab et al: TGF-31 in CsA nephropathy

A

B

C

D

E

F

G

H

I

Fig. 7. Immunofluorescence micrographs of CsA and VH-treated rat kidneys. Kidney sections from CsA treated rats for seven days (A, D, and G), for 28

days (B, E, and H) and VH controls (C, F, and 1) were stained with an antibody to tenascin (A, B, and C), fibronectin EDA+ (D, E, and F) and decorin (G, H, and I). The photographs were taken under identical conditions with equal exposures of 60 seconds (magnification x250).

1149

Shihab et a!: TGF-/31 in CsA nephropathy

A

A

4

2

I00

C)

**

0

C),

0 a)

**

>

E2

a)

**

0

a)

(I)

E

C')

a)

0x a) z

C!,

0

E Tenascin

Fibro-EDA+

Decorin

B

0

5

TGF-131

PAl-i

BG

DEC

COL I

B

C)

5

**

(1)

I

><

E Ca

0 0 a)

0

Ca

a)

.01

C

0 U)

0

0 Tenascin

Fibro-EDA+

Decorin

C

**

a)

**

z

**

**

E

5

0 C)

4

TGF-131

PAl-i

BG

DEC

COL I

C

Fig. 9. Quantitation of mRNA expression of TGF-f31, PAl-i and matrix proteins in cortex versus medulla in CsA -treated rat kidneys. Total RNA was isolated from whole cortex () and medulla (E) at days 7 (A) and 28 (B),

i3 ><

and was hybridized whh cDNA probes to TGF-pl, PAl-i, biglycan,

Ca

decorin, and type I collagen. Abbreviations are: biglycan (BO), decorin (DEC), and type I collagen (COL I). *P < 0.001 compared to cortex.

E Si22 a) U)

Fig. 8. Quantitation of deposition of extracellular matrix components in CsA

0 Tenascin

Fibro-EDA+

Decorin

and control treated rat kidneys by immunofluorescence. (A) Glomerular staining, (B) tubulointerstitial staining, and (C) vessels staining scores at days 7 and 28. Symbols are: (•) CsA 7 days; (hI) VH 7 days; () CsA 28 days, () VH 28 days. Abbreviation Fibro-EDA+ is fibronectin EDA+. P < 0.005 and ** < 0.001 compared to VH control.

1150

Shihab et al: TGF-/31 in CsA nephropathy

tubular cells [31, while no change was observed in a variety of rat and human renal cells [35]. On the other hand, collagen type III

synthesis was increased in human mesangial cells and renal fibroblasts treated with CsA [351. In addition, procollagen sl (I)

mRNA was shown to be elevated in the renal cortex of CsAtreated rats [361. CsA was also recently shown to enhance TGF-p mRNA expression in normal human T cell culture [37]. Another cytokine, platelet-derived growth factor (PDGF), was shown to be present in the arteriolar walls of CsA-treated rats [38]. However, while most of the experiments done were in vitro, no study has

hyperplasia within the kidney in such individuals similar to other one-clip, one-kidney models of hypertension [50]. In conclusion, by using a model of chronic CsA nephropathy that mimics the human lesion, we have shown that matrix proteins are increased, mostly in the medulla, and more specifically in the

interstitial and vascular compartments of the kidney. We have linked the expression of TGF-/31 to matrix expansion and have shown that both TGF-131 and PAl-I are increased. We propose that TGF-/31 is a key fibrogenic cytokine involved in the development of chronic CsA nephropathy by enhancing ECM deposition and inhibiting its degradation. However, other factors may also be involved in the fibrosis of chronic CsA nephrotoxicity. Whether long-term immunosuppression with CsA inevitably leads to renal

looked at the components of the fibrotic lesion of chronic CsA nephrotoxicity in vivo, and none has attempted to correlate the renal fibrosis with any growth factor. Transforming growth factor-n (TGF-13) has been implicated in fibrosis is unknown and has not been excluded by any properly the fibrosis of a number of chronic diseases of the kidney and controlled clinical study. other organs [14, 15]. In addition, it is known to induce its own production [39] and its persistent expression following repeated Acknowledgments injuries can lead to a cycle of continued TGF-/3 production [14]. TGF-13 causes fibrosis by stimulating synthesis of individual ECM components and simultaneously blocking ECM degradation [16—

A preliminary report of this work was presented at the Annual Meeting of the American Society of Transplant Physicians, Chicago, IL in May 1995. This work was supported in part by the Ogden Research Foundation

19]. In this model, we have observed an elevated expression of (F.S.S.) and a grant from the Oregon Health Sciences University Founrepresentative members of the proteoglycans, glycoproteins and dation and the Clinical Research Group of Oregon (W.M.B.). collagen families. Their expression correlated with a parallel and significant increase in TGF-131. In addition, the protease inhibitor Reprint requests to Fuad S. Shihab, M.D., Division of Nephrology, 4R312 PAl-i was up-regulated and its expression followed closely the Medical Center, University of Utah, 50 N. Medical Drive, Salt Lake City, Utah

expression of TGF-pl and the ECM proteins studied. Since

84132, USA.

TGF-f31 directly stimulates PAl-i expression [19], our data suggest that the observed ECM accumulation in this model likely

involves the dual action of TGF-13 on ECM deposition and degradation. What causes the elevation of TGF-13 remains unclear. Since CsA directly stimulates TGF-p in vitro, it has been proposed that the immune suppressive activities and renal fibrosis due to CsA are both mediated by TGF-13 [37, 40, 411 and may explain the elevated expression of TGF-131 in our model. It is also possible

that renal ischemia produced by the vasoconstriction of CsA elevates TGF-/3 [42]. Since the current model is produced by salt-depletion, the renin-angiotensin system (RAS) is activated [32]. Interestingly, there is evidence to suggest that activation of the RAS up-regulates TGF-f3. The renal RAS is induced by CsA and plays an important role not only in the regulation of glomerular hemodynamics but also in glomerular growth and sclerosis [431. Angiotensin (Ang) II and CsA have been separately shown

to stimulate fibrogenesis by mechanisms independent of their vasoconstrictive action [33, 44]. Chronic infusion of Ang II in rats

produce considerable interstitial fibrosis [45]. Ang converting enzyme inhibitors and Ang II receptor antagonists attenuate interstitial fibrosis and progressive glomerulosclerosis in several disease models and slow disease progression in several human studies [46, 47]. Kagami et al have recently demonstrated a link between Ang II and TGF-/3 by showing that Ang II stimulation of

References 1. KAI-IAN BD: Cyclosporine. N EnglJ Med 321:1725—1738, 1989 2. FEUTREN G, MIHATSCH MJ: Risk factors for cyclosporine-induced

nephropathy in patients with autoimmune diseases. N EngI J Med 326:1654—1660, 1992 3. MYERS BD: Cyclosporine nephrotoxicity. Kidney mt 30:964—974, 1988 4. ENGLISH J, EVAN A, HouGilToN DC, BENNETT WM: Cyclosporinc-

induced acute renal dysfunction in the rat. Transplantation 44:135— 141, 1987 5. MYERS BD, Ross JC, NEWTON LD, LUETSCHER JA, PERLROTH MG:

Cyclosporine-associated chronic nephropathy. N EngI J Med 311:699— 705, 1984 6. MYERS BD, NEWTON L, BOSHKOS C, MACOVIAK JA, FRIST WH, DERBY GC, PERLROTI-I MG, SIBLEY RK: Chronic injury of human renal microvessels with low-dose cyclosporine therapy. Transplantation 46:694—703, 1988 7. PALESTINE AG, AUSTIN HA, BALOW JC, ANTONOVYCI-I TT, SABSIS SG,

PEEUSS HG, NUSSENBLAVF RB: Renal histopathologic alterations in patients with cyclosporine for uveitis. N Engl J Med 314:1293—1298, 1986 8. WHEATLY HC, DATZMAN M, WILLIAMS JW, MILES DE, HATCH FE: Long-term effects of cyclosporine on renal function in liver transplant recipients. Transplantation 43:641—647, 1987 9. NIZZE H, MIHATSCH MJ, ZOLLINGER HU, BROCHERIOU C, GOKEL JM,

HENERY K, SLOANE JP, STOVEN PG: Cyclosporine-associated ne-

phropathy in patients with heart and bone marrow transplants Clin Nephrol 30:248-260, 1988

ECM protein synthesis in rat glomerular mesangial cells is

10. MYERS BD, SIBLEY R, NEWTON L, TOMLANOVICH SJ, BOSI-IKOS C, STIN50N F, LUETSCHER JA, WHITNEY DJ, KRASNY D, COPLON NS,

mediated by both an increase in TGF-/3 synthesis and an increase in the conversion from latent to active TGF-p [48]. In addition,

11. Kopp JB, KLOTMAN PE: Cellular and molecular mechanisms of

CsA administration in salt-depleted rats caused a significant

PERLROTH MG: The long-term course of cyclosporine-associated chronic nephropathy. Kidney mt 33:590—600, 1988 cyclosporin nephrotoxicity. JAm Soc Nephrol 1:162—179, 1990

increase in juxtaglomerular (JG) apparatus renin, and blockage of 12. ROSEN 5, GREENFELD Z, BREZIS M: Chronic cyclosporine-induced nephropathy in the rat. Transplantation 49:445—452, 1990 the RAS with a competitive Ang II receptor antagonist, losartan,

reduced the formation of interstitial fibrosis [49]. While the peripheral renin activity is often normal or suppressed in CsAtreated renal transplants, there is abundant evidence of JG cell

13. ELZINGA LW, ROSEN 5, BENNETT WM: Dissociation of glomerular

filtration rate from tubulointerstitial fibrosis in experimental chronic cyclosporine nephropathy: Role of sodium intake. JAm Soc Nephrol 4:214—221, 1993

Shihab et al: TGF-J31 in ('sA nephropathy 14. BORDER WA, RuosLsHTI E: Transforming growth factor-[3 in disease: The dark side of tissue repair. J Clin Invest 90:1—7, 1992 15. BORDER WA, NOBLE NA: Transforming growth factor 13 in tissue fibrosis. N Engi J Med 331:1286—1292, 1994 16. NAKAMURA T, MILLER D, RUOSLAI-JTI E, BORDER WA: Production of

extracellular matrix by glomerular epithelial cells is regulated by transforming growth factor-/fl. Kidney In! 41:1213—1221, 1990 17. IGN0Tz RA, MASSAGUE J: Transforming growth factor beta stimulates the expression of fibronectin and collagen and their incorporation into the extracellular matrix. J Biol Chem 261:4337—4345, 1986 18. BALZA E, BoRsI L, ALLEMANNI G, ZARDI L: Transforming growth factor [3 regulates the levels of different fibronectin isoforms in normal human cultured fibroblasts. FEBS Lett 228:42—44, 1988 19. LAIHO M, SAKSELA 0, KESKI-OJA J: Transforming growth factor-[3

1151

33. TRUONG LD, FARHOOD A, TASBY J, Gui UM D: Experimental chronic

renal ischemia: Morphologic and immunologic studies. Kidney mt 41:1676—1689, 1992

34. WOLF G, KILLEN PD, NEILSON EG: Cyclosporine A stimulates transcription and procollagen secretion in tubulointerstitial fibroblasts and proximal tubular cells. JAm Soc Nephrol 1:918—922, 1990 35. GHIGGERI GM, ALTIERI P, OLEGGINI R, VALENTI F, GINEVA F,

PERFUMO F, GUSMAN R: yclosporine enhances the synthesis of

selected extracellular matrix proteins by renal cells "in culture". Transplantation 57:1382—1388, 1994 36. NAST CC, ADLER SG, ARTJSHEVSKY A, KRESSER CT, AHMED K,

ANDERSON PS: Cyclosporine induces elevated procollagen al (I) mRNA levels in the rat renal cortex. Kidney mt 39:631—638, 1991 37. KHANNA A, Li B, STENZEL KH, SUTHANTHIRAN M: Regulation of new

induction of type-I plasminogen activator inhibitor. J Biol Chem

DNA synthesis in mammalian cells by cyclosporine. Transplantation

262:17467—17474, 1987 20. RUSCETFI FW, PALLADINO MA: Transforming growth factor-/3 and the

57:577—582, 1994 38. SHEHATA M, EL NAHAS M, BARKWORTH E, COPE GH, RAFTERY AT:

immune system. Frog Growth Factor Res 3:159—175, 1991

Increased platelet-derived growth factor in the kidneys of cyclosporintreated rats. Kidney mt 46:726—732, 1994

21. SHELL MM, ORMSBY I, KIER AB, PAWLOSKI S, DIEBOLD RJ, YIN M, ALLEN R, SIDNIAN C, PROETZEL G, CALVIN D: Targeted disruption of

39. OBBERGHEN-SCI-IILLJNG EV, ROCHE NS, FLANDERS KC, SPORN MB,

the mouse transforming growth factor-[31 gene results in multifocal

ROBERTS AB: Transforming growth factor [31 positively regulates its

inflammatoiy disease. Nature 359:693—699, 1992 22. DERYNCK R, JARREIT JA, CIIEN EY, GOEDDEL DV: The murine transforming growth factor-[3 precursor. J Biol Chem 261:4377—4379, 1986 23. ZEHEB R, GELEHRTER TD: Cloning and sequencing of eDNA for the rat plasminogen activator inhibitor-I. Gene 73:459—468, 1988 24. FISHER LW, TERMINE JD, YOUNG MF: Deduced protein sequence of small proteoglycan I (biglycan) shows homology with proteoglycan TI

(decorin) and several nonconnective tissue proteins in a variety of species. J Biol Chem 264:4571—4576, 1989

25. ASUNDI VK, DREHER KL: Molecular characterization of vascular smooth muscle decorin: Deduced core protein structure and regulation of gene expression. Eur J Cell Biol 59:314—321, 1992 26. GENOVESE C, ROWE D, KREAM B: Construction of DNA sequences

complementary to rat al and a2 collagen mRNA and their use in studying the regulation of type I collagen synthesis by 1 ,25-dihydroxyvitamin D. Biochemistry 23:6210—6216, 1984 27. FORT P, MARTY L, PIECHACZYK M, EL SABOURY S, DANI C, JEANTEUR

P, BLANCHARD JM: Various rat adult tissues express only one major mRNA species from the glyceraldehyde-3-phosphate dehydrogenase multigenic family. NuclAcid Res 13:1431—1442, 1985 28. SHIHAB FS, YAMAMOTO T, NAST CC, COHEN AH, NOBLE NA, GOLL)

LI, BORDER WA: Transforming growth factor-[3 and matrix protein expression in acute and chronic rejection of human renal allografts. J Am Soc Nephrol 6:286—294, 1995 29. YAMAGLJCHI Y, MANN DM, RuosLAl-ITI E: Negative regulation of transforming growth factor-beta by the proteoglycan decorin. Nature 346:281—284, 1990 30. BERTANI T, PERICO N, ABBATE M, BATFAGLEA C, RErouzzi G: Renal

injury induced by long-term administration of cyclosporine A to rats. Am J Pathol 127:569—579, 1987 31. JACKSON NM, Hsu C, VISSCHER GE, VENKATACHALAM MA, 1-IUMES

I-ID: Alterations in renal structure and function in a rat model of cyclosporine nephrotoxicity. J Pharm Exp Ther 242:749—756, 1987 32. YOUNG BA, BURDMANN EA, JOHNSON RJ, ALPERS CE, GHIACHELLI

CM, ENG E, ANDOH T, BENNETT WM, COUSER WG: Cellular prolif-

eration and macrophage influx precede interstitial fibrosis in cyclosporine nephrotoxicity. Kidney mt 48:439—448, 1995

own expression in normal and transformed cells. J Biol Owm 263: 7741—7746, 1988 40. PRASHAR Y, KEANNA A, SEHAJPAL P, SHARMA VK, SUTHANTHIRAN M:

Stimulation of transforming growth factor-j31 transcription by cyclosporine. FEBS Let! 358:109—112, 1995

41. Li B, SEHAJPAL PK, KHANNA A, VLASSARA H, CERAMI A, STENZEL

KH, SUTHANTHIRAN M: Differential regulation of transforming growth factor [3 and interleukin 2 genes in human T cells: Demonstration by usage of novel competitor DNA constructs in the quantitative polymerase chain reaction. J Exp Med 174:1259—1262, 1991 42. GOES N, URMSON J, RAMASSAR V, HALLORAN PF: Ischemic acute

tubular necrosis induces an extensive local cytokine response. Transplantation 59:565—572, 1995 43. MASON J, MULLER-SCHWEINITZER E, SUPONT M, CASELLAS D, Mi-

HATSCH M, MOORE L, KASKEL F: Cyclosporine and the reninangiotensin system. Kidney mt 39:S28—32, 1991 44. WOLF G, NEILSON EG: Angiotensin II induces cellular hypertrophy in cultured murine proximal tubular cells. Am J Physiol 259:F768—F777, 1990 45. JOHNSON RJ, ALPERS CE, YOSHIMURA A, LOMBARDI D, PRITZL P, FLOEGE J, SCHWARTZ SM: Renal injury from angiotensin TI-mediated hypertension. Hypertension 19:464—474, 1992 46. LAFAYETtE RA, MAYER G, PARK SK, MEYER TW: Angiotensin II

receptor blockade limits glomerular injury in rats with reduced renal mass. J Clin invest 90:766—771, 1992 47. KEANE WF, ANDERSON 5, AURELL M, DC ZEEUW D, NARINS RG,

POVAR G: Angiotensin converting enzyme inhibitors and progressive renal insufficiency. Ann Intern Med 111:503—516, 1989 48. KAGAMI 5, BORDER WA, MIUER DE, NOBLE NA: Angiotensin II stimulated extracellular matrix protein synthesis through induction of transforming growth factor-[3 expression in rat glonierular mesangial cells. J Clin Invest 93:2431—2437, 1994 49. BURDMANN EA, YOUNG B, ANDOH TF, EVANS A, ALPERS CE, LINDSLEY J, JOHNSON RJ, COUSER W, BENNETT WM: Mechanisms of

eyclosporine-induced interstitial fibrosis. Transplant Proc 26:2588— 2589, 1994 50. BENNETT WM, PORTER GA: Cyclosporine-associated hypertension. Am JMed 85:131—133, 1988