OBJECTIVE—Studies in rodent models have suggested that reduction in renal transforming growth factor (TGF)-β1 may underlie the renoprotective effects of the renin-angiotensin system (RAS) blockade. However, the role of the RAS blockade in abrogating TGF-β in human disease is unknown. Accordingly, we sought to examine TGF-β gene expression and biological activity in human renal biopsies, before and after ACE inhibition.
RESEARCH DESIGN AND METHODS—RNA was extracted from renal biopsies taken from participants in the Diabiopsies study, a randomized controlled 2-year trial of 4 mg/day perindopril versus placebo that reported a reduction in proteinuria and cortical matrix expansion in type 2 diabetic nephropathy. Biopsies taken at study entry and at 2 years were obtained in 12 patients (6 placebo and 6 taking perindopril). TGF-β1 and its receptor mRNA were quantified by real-time PCR, and its biological activity was assessed by examining the activation of its intracellular signaling pathway (phosphorylated Smad2) and the expression TGF-β–inducible gene H3 (βig-H3).
RESULTS—At baseline, TGF-β1 expression was similar in both placebo- and perindopril-treated groups and was unchanged over a 2-year period in biopsies of placebo-treated subjects. In contrast, perindopril treatment led to a substantial diminution in TGF-β1 mRNA (mean 83% reduction, P < 0.05). Phosphorylated Smad2 immunolabeling and βig-H3 mRNA were similarly reduced with ACE inhibition (P < 0.05) but unchanged in the placebo group. No differences were noted in the gene expression of TGF-β receptor II in biopsies of either placebo- or perindopril-treated subjects.
CONCLUSIONS—This study demonstrates that over a 2-year period, treatment with perindopril in patients with type 2 diabetes and nephropathy leads to a reduction in both renal TGF-β1 gene expression and its downstream activation.
Blockade of the renin-angiotensin system (RAS) reduces the rate of progression of renal dysfunction in patients with both type 1 and type 2 diabetes (1,2). While the renoprotective effects of ACE inhibitors and angiotensin receptor blockers have been attributed to their hemodynamic effects, recent evidence suggests that these agents may also exert beneficial effects by reducing the production of locally active growth factors. In particular, angiotensin II potently induces synthesis of transforming growth factor (TGF)-β (3), a profibrotic growth factor consistently implicated in the pathogenesis of diabetic nephropathy (4).
While the glomerulus and, in particular, the mesangium have been the focus of intense investigation in diabetes, tubulointerstitial injury also is a major feature of diabetic nephropathy and an important predictor of renal dysfunction (5). For instance, in a longitudinal intervention study of patients with type 2 diabetic nephropathy treated with ACE inhibitor therapy, histological improvement only was seen in the tubulointerstitium after 2 years of treatment (6). However, while these beneficial effects may be hypothesized to be due to a reduction in intrarenal TGF-β expression, to date studies examining the interaction between the RAS and TGF-β within the diabetic kidney have been confined to cell culture, rodent models, and the measurement of TGF-β in urine (7) and serum (8). Accordingly, we sought to examine the effects of ACE inhibition on TGF-β1 expression and activity in human kidney tissue.
RESEARCH DESIGN AND METHODS—
Renal tissue was obtained from biopsy specimens stored as part of the Diabiopsies study, the structural and functional aspects of which have been reported (6). In brief, this prospective, biopsy-based study examined the effects of the ACE inhibitor perindopril (4 mg/day) in patients with type 2 diabetes, proteinuria (range 70–4,210 mg/day), relatively well-preserved renal function (creatinine clearance >60 ml/min), and evidence of diabetic glomerulosclerosis on initial biopsy before entry into the trial (Table 1). Of 19 participants who completed the original study, tissue obtained at study entry and follow-up (24 months) was available from 12 participants (6 taking perindopril and 6 placebo) for use in the current substudy. At biopsy, specimens were divided and either snap frozen, immersed in Dubosq-Brazil fixative and paraffin embedded, or fixed in 2.5% glutaraldehyde before embedding in Epon. Snap-frozen material subsequently was cut by cryostat, placed on gelatin-coated sides, and stored at −80°C. Control tissue was obtained from eight patients who underwent nephrectomy for the treatment of a tumor. All control patients had normal renal function, were normotensive, and had no evidence of urinary protein. Sections of kidney remote to the tumor were selected and snap frozen at −80°C and normal histology microscopically confirmed. Estimation of the percentage of globally sclerosed glomeruli was made as previously described (6) on paraffin-embedded tissue. Percentage of cortical interstitial fibrosis was also determined, as previously described (6), using computer-assisted stereologic analysis of samples stained with Masson’s trichrome of the whole available cortical tissue.
RNA extraction and cDNA synthesis
RNA was extracted from snap-frozen study tissue stored at −80°C using previously established methods (9). In brief, 200 μl of water-saturated phenol and chloroform were added to the RNA sample, mixed well, and centrifuged for 2 min at 4°C. Following removal of the upper aqueous phase, the phenol/chloroform extraction was repeated, followed by a further chloroform extraction. Samples were then precipitated with 180 μl isopropanol, 18 μl 3M sodium acetate (pH 5.2), and 1 μl of 20 mg/ml of carrier glycogen at 20°C. After centrifuging for 15 min at 4°C, the supernatant was discarded and the RNA pellet washed with 1 ml of cold 70% ethanol containing diethyl pyrocarbonate–treated water, centrifuged for 10 min at 4°C, and resuspended in 10 μl RNase-free (diethyl pyrocarbonate) water. One microliter of RNA extracted from frozen tissue sections was reverse transcribed in a final volume of 20 μl using M-MLV reverse transcriptase (Life Technologies-BRL) in the manufacturer’s buffer containing 1 mmol/l dNTPs, 40 units RNase inhibitor (Amersham Pharmacia Biotech, Freiburg, Germany), 300 ng of random hexamers (Pharmacia), and 7 μl RNA. The reactions took place at 42°C for 60 min, followed by 95°C for 5 min and 4°C for 5 min.
Real-time PCR
Gene expression of TGF-β1, its type II receptor, and TGF-β–inducible gene H3 (βig-H3) were quantified using the GeneAmp 5700 Sequence Detector (PE Biosystems, Foster City, CA) according to the manufacturer’s instructions. Primers and probes were obtained from Sigma-Aldrich and PE Biosystems, as listed in Table 2. The labeled probes (Applied Biosystems, Foster City, CA) comprised a fluorescence reporter (6-carboxyfluorescein) at the 5′-end and a fluorescent quencher (6-carboxytetramethylrhodamine [TAMRA]) at the 3′-end (10). A commercial, predeveloped 18S control kit labeled with the fluorescent reporter dye (VIC) on the 5′-end and the quencher (TAMRA) on the 3′-end (PE Biosystems) was used as the endogenous control. Similar amplification efficiencies of all gene targets and 18S were validated, permitting semiquantitative measurement of gene expression.
The 25-μl PCR mixture contained 12.5 μl of Taqman Universal PCR Master Mix, 500 nmol/l of primers (forward and reverse), 100 nmol/l of labeled probe, and 1 μl of cDNA template. PCR was performed at 50°C for 2 min, 95°C for 10 min, and then run for 50 cycles at 95°C for 15 s and 60°C for 1 min. Comparative cycle threshold (Ct) calculations were all relative to the control group. 18S Ct values were subtracted from the gene of interest Ct values to give a ΔCt value. ΔΔCt values were achieved by subtracting the average control ΔCt value for each treatment group, and the target gene expression relative to control was derived using the equation 2t−ΔΔC. The derived normalized values were the mean of four runs.
TGF-β receptor activation
Activation of the TGF-β receptor was assessed by quantifying the nuclear expression of phosphorylated Smad2 in kidney tissue, as previously described (7), using a rabbit anti–phospho-Smad2 antibody (Cell Signaling Technology, Boston, MA) that detects endogenous Smad2 only when dually phosphorylated at Ser463 and Ser465 (11). Four-micron-thick sections were cut from paraffin-embedded tissue and immunostained, as previously described (10), according to the manufacturers instructions. Sections incubated with 1:10 normal goat serum, instead of the primary antiserum, served as the negative control.
Quantitation of matrix deposition and phospho-Smad2 expression
The accumulation of matrix and the extent of immunostaining phospho-Smad2 were quantified using computer-assisted image analysis, as previously reported (12,13). In brief, images from the entire field of each biopsy were captured and digitized using a BX50 microscope attached to a Fujix HC5000 digital camera. Digital images were then loaded onto a Pentium III IBM computer. An area of blue on trichrome-stained sections (for matrix) or brown on immunostained sections (for phospho-Smad2) were selected for their color ranges, and the proportional area of tissue with their respective ranges of color were then quantified. Calculation of the proportional area stained blue (matrix) or brown (phospho-Smad2) was then determined using image analysis (Analytical Imaging Station version 6.0; AIS, Ontario, Canada).
Statistical analysis
Data are expressed as means ± SE, unless otherwise stated. Results from gene expression studies demonstrated a skew distribution, such that data were log transformed before analysis. Statistical significance of change in gene expression over the trial period was determined by a two-way ANOVA with a Fisher’s post hoc comparison. Analyses were performed using Statview II+Graphics package (Abacus Concepts, Berkeley, CA) on an Apple Macintosh G4 computer (Apple Computers, Cupertino, CA). A P value <0.05 was regarded as statistically significant.
RESULTS
Clinical and histological parameters
As reported in the Diabiopsies study cohort (6), the subgroup examined in the present study also showed no differences in blood pressure control, serum creatinine, creatinine clearance, or glycated hemoglobin between perindopril- and placebo-treated groups either at baseline or over the period of the study (Table 1). Urinary protein excretion increased in the placebo group and was unchanged in the perindopril group over the trial period (Table 1).
Histological parameters
Extensive glomerulosclerosis was noted in all biopsies from diabetic subjects. However, as previously reported for the entire study group, no change in the extent of glomerulosclerosis in either the perindopril or placebo groups was noted in the present substudy (Table 1). Moreover, as previously reported in the initial study, mean cortical interstitial fractional volume significantly increased in placebo-treated patients but was unchanged in the perindopril-treated patients (Table 1).
Gene expression
Analysis of paired biopsies from trial entry and 2 years for each patient demonstrated a significant reduction in TGF-β1 gene expression in the perindopril-treated biopsies. In contrast, no change in TGF-β1 mRNA was present in placebo-treated patients. This represented a significantly greater decline in TGF-β1 gene transcription in the ACE inhibitor group compared with placebo (P < 0.05; Fig. 1A). Similarly, βig-H3 mRNA decreased by 60 ± 16% (P < 0.05) after 2 years in biopsies from perindopril-treated subjects, contrasting a 47 ± 11% increase (NS) in placebo-treated subjects. No difference in TGF-β type II receptor mRNA was noted in either group.
Phospho-Smad2
Minimal nuclear phospho-Smad2 was detected in normal nephrectomy tissue (Fig. 2A). In contrast, kidneys from diabetic subjects showed abundant nuclear immunolabeling, localized predominantly to tubular epithelial cells (Fig. 2B–E). No change in the proportional area of tissue showing nuclear phospho-Smad2 staining was noted in placebo-treated patients, after 2 years. In contrast, perindopril was associated with a significant reduction in phospho-Smad2 immunolabeling during the study interval (P < 0.05; Fig. 1B).
CONCLUSIONS—
To our knowledge, the present study provides the first data showing that blockade of the RAS with ACE inhibitor therapy leads to a reduction in intrarenal TGF-β1 expression and activity in human diabetic nephropathy. Although previous human studies in diabetic nephropathy have examined the effects of the interrupting the RAS, these have been confined to its analysis in serum and urine. For instance, in a subgroup of participants from the Collaborative Study Group Captopril Trial, ACE inhibition led to a diminution in serum TGF-β over a 6-month period (8). Similarly, in a prospective study of patients with type 2 diabetes and microalbuminuria, the angiotensin receptor blocker losartan reduced urinary TGF-β over a 4-week period (7).
Over the past several years, experimental evidence consistently has suggested a key role for TGF-β in the pathogenesis of the extracellular matrix accumulation that characterizes diabetic nephropathy (4) and closely correlates with declining renal function (14). The mechanisms for this fibrogenic or prosclerotic action of TGF-β are multiple and include both stimulation of extracellular matrix synthesis and inhibition of its degradation (15). Several studies in experimental animals have documented that not only is TGF-β expression increased in the diabetic kidney but that it is also present in a biologically active form (16,17). Moreover, inhibition of TGF-β action has been shown to attenuate mesangial expansion and interstitial fibrosis in experimental diabetic nephropathy (18).
The pivotal clinical findings of major intervention trials (1,2,19) with ACE inhibitors and angiotensin receptor blockers in diabetic nephropathy have been accompanied by cell culture data showing that angiotensin II, the effector molecule of RAS, stimulates extracellular matrix protein synthesis through induction of TGF-β in mesangial cells (20), renal interstitial fibroblasts (21), and proximal tubular epithelial cells (22). While also expressed in the glomerulus, studies in the streptozotocin-induced diabetic rat have found the tubulointerstitium to be the predominant site of TGF-β overexpression and that this increase could be attenuated by ACE inhibition (17). The findings of the present study are consistent with previous preclinical studies and show that as in the rodent model, ACE inhibition leads to a significant diminution in TGF-β expression, tubulointerstitial fibrosis, and proteinuria (17).
TGF-β1 is synthesized as a 391–amino acid precursor molecule with little biological activity, requiring cleavage of its NH2-terminal latency-associated peptide to give rise to its active form (23). In addition, its biological effects also may be modified by the presence of the proteoglycan decorin (24) and the scavenging protein α 2-macroglobulin (25), such that increased TGF-β1 mRNA or protein may not accurately reflect TGF-β1 activity. In the present study, we assessed the biological effects of TGF-β1 by examining one of its specific intracellular actions, the phosphorylation of the TGF-β receptor–activated protein Smad2 (26,27). While prominent nuclear staining of phosphorylated Smad2 was present in all diabetic biopsies at baseline, a significant diminution only was noted in kidney tissue from those patients who had received perindopril over the succeeding 2 years. In addition, gene expression of a TGF-β–inducible matrix protein βig-H3 that has been previously used to indirectly assess TGF-β activity in a range of clinical and experimental settings including diabetic nephropathy (28,29). In contrast to the reduction in TGF-β1 and indexes of its activity, expression of the TGF-β type II receptor, in the present study, was unaffected by ACE inhibitor therapy.
As the glomerular fraction accounts for <10% of total cortical volume, we consider the analysis of mRNA from the biopsy material, using the methods described in this study, to be largely indicative of changes in the tubulointerstitium of the kidney. However, the findings of this current study do not exclude the glomerulus as a further major source of TGF-β gene transcription. Indeed, previous studies using microdissected glomeruli from patients with diabetic nephropathy showed significant increases in TGF-β expression (28).
While the present study focuses on diabetic nephropathy, an observational report in IgA nephropathy noted significantly less TGF-β mRNA in the biopsies of patients who were receiving ACE inhibitor therapy (30). Together, these findings suggest that a reduction in TGF-β transcription and consequent fibrosis may contribute to the renoprotective effects of ACE inhibition in humans.
A: Fold change in TGF-β1 mRNA in human renal biopsies at 2 years compared with baseline (*P < 0.05). Compared with baseline, perindopril treatment led to an eightfold decrease in TGF-β mRNA. In contrast, biopsies from patients treated with placebo showed no detectable difference in TGF-β gene expression. B: Fold change in percentage area of phospho-Smad2 immunostaining at 2 years compared with baseline (*P < 0.05). Compared with baseline, perindopril treatment led to a significant decrease in phospho-Smad2 immunostaining. In contrast, biopsies from patients treated with placebo showed no difference in immunostaining. C: Fold change in βig-H3 mRNA in human renal biopsies at 2 years compared with baseline (*P < 0.05). Compared with baseline, perindopril treatment led to a threefold decrease in βig-H3 mRNA. In contrast, biopsies from patients treated with placebo showed no detectable difference in βig-H3 gene expression.
A: Fold change in TGF-β1 mRNA in human renal biopsies at 2 years compared with baseline (*P < 0.05). Compared with baseline, perindopril treatment led to an eightfold decrease in TGF-β mRNA. In contrast, biopsies from patients treated with placebo showed no detectable difference in TGF-β gene expression. B: Fold change in percentage area of phospho-Smad2 immunostaining at 2 years compared with baseline (*P < 0.05). Compared with baseline, perindopril treatment led to a significant decrease in phospho-Smad2 immunostaining. In contrast, biopsies from patients treated with placebo showed no difference in immunostaining. C: Fold change in βig-H3 mRNA in human renal biopsies at 2 years compared with baseline (*P < 0.05). Compared with baseline, perindopril treatment led to a threefold decrease in βig-H3 mRNA. In contrast, biopsies from patients treated with placebo showed no detectable difference in βig-H3 gene expression.
Representative photomicrograph of phosphorylated Smad2 immunostaining of tubulo-interstitial areas of control (A) and diabetic nephropathy (B–E), demonstrating a marked increase in immunostaining of nuclei in diabetic biopsies (arrow). Compared with the trial entry biopsy, there was a reduction in phopsho-Smad2 immunostaining over 2 years in the perindopril-treated patient (B and C) compared with no change in the placebo-treated patient (D and E). Magnification ×320.
Representative photomicrograph of phosphorylated Smad2 immunostaining of tubulo-interstitial areas of control (A) and diabetic nephropathy (B–E), demonstrating a marked increase in immunostaining of nuclei in diabetic biopsies (arrow). Compared with the trial entry biopsy, there was a reduction in phopsho-Smad2 immunostaining over 2 years in the perindopril-treated patient (B and C) compared with no change in the placebo-treated patient (D and E). Magnification ×320.
Clinical and histological data of the study participants demonstrating a measurable increase in urinary protein and percentage cortical fibrosis in the placebo-treated patients over the study period that was not evident in the perindopril-treated patients
. | Study time point (months) . | . | . | . | |||
---|---|---|---|---|---|---|---|
. | Placebo . | . | Perindopril . | . | |||
. | 0 . | 24 . | 0 . | 24 . | |||
Age (years) | 48 ± 3 | 42 ± 5 | |||||
Sex (male/female) | 3/3 | 5/1 | |||||
Mean arterial blood pressure (mmHg) | 108 ± 6 | 110 ± 3 | 101 ± 3 | 103 ± 4 | |||
HbA1c (%) | 6.1 ± 0.6 | 8.53 ± 1.1 | 6.63 ± 0.6 | 7.77 ± 0.4 | |||
Creatinine clearance (ml/min) | 105 ± 10 | 85 ± 19 | 142 ± 24 | 123 ± 18 | |||
Glomerulosclerosis (%) | 15.3 ± 7.5 | 16.8 ± 9.3 | 13.5 ± 8.5 | 12.2 ± 7.0 | |||
Urinary protein (mg/24 h) | 1,292 (1,905–208) | 3,786 (4,677–426)* | 512 (741–204) | 275 (380–138) | |||
Cortical interstitial fibrosis (%) | 33.46 ± 2.5 | 41.49 ± 5.0* | 32.8 ± 1.8 | 32.8 ± 2.2 |
. | Study time point (months) . | . | . | . | |||
---|---|---|---|---|---|---|---|
. | Placebo . | . | Perindopril . | . | |||
. | 0 . | 24 . | 0 . | 24 . | |||
Age (years) | 48 ± 3 | 42 ± 5 | |||||
Sex (male/female) | 3/3 | 5/1 | |||||
Mean arterial blood pressure (mmHg) | 108 ± 6 | 110 ± 3 | 101 ± 3 | 103 ± 4 | |||
HbA1c (%) | 6.1 ± 0.6 | 8.53 ± 1.1 | 6.63 ± 0.6 | 7.77 ± 0.4 | |||
Creatinine clearance (ml/min) | 105 ± 10 | 85 ± 19 | 142 ± 24 | 123 ± 18 | |||
Glomerulosclerosis (%) | 15.3 ± 7.5 | 16.8 ± 9.3 | 13.5 ± 8.5 | 12.2 ± 7.0 | |||
Urinary protein (mg/24 h) | 1,292 (1,905–208) | 3,786 (4,677–426)* | 512 (741–204) | 275 (380–138) | |||
Cortical interstitial fibrosis (%) | 33.46 ± 2.5 | 41.49 ± 5.0* | 32.8 ± 1.8 | 32.8 ± 2.2 |
Data are means ± SE or geometric mean (95% CI).
P < 0.05.
Sequences of primers and probes
. | Forward primer . | Reverse primer . | Probe . |
---|---|---|---|
TGF-β1 | CACCCGCGTGCTAATGG | ATGCTGTGTGTACTCTGCTTGAACT | CCACAACGAAATCTA |
TGF-β receptor II | TCCTGTGGACGCGTATCG | TGTCAGTGACTATCATGTCGTTATTAACC | CGATCCCACCGCACG |
βig-H3 | GTTTATCGTAATAGCCTCTGCATTGA | TCCCGTACCTCCCCCTCTT | AGTCCACTTTAAATCCTT* |
. | Forward primer . | Reverse primer . | Probe . |
---|---|---|---|
TGF-β1 | CACCCGCGTGCTAATGG | ATGCTGTGTGTACTCTGCTTGAACT | CCACAACGAAATCTA |
TGF-β receptor II | TCCTGTGGACGCGTATCG | TGTCAGTGACTATCATGTCGTTATTAACC | CGATCCCACCGCACG |
βig-H3 | GTTTATCGTAATAGCCTCTGCATTGA | TCCCGTACCTCCCCCTCTT | AGTCCACTTTAAATCCTT* |
TAMRA, all others MGB.
Article Information
This work was supported in part by grants from the National Health and Medical Research Council of Australia and the Juvenile Diabetes Research Foundation (JDRF). D.J.K. is a recipient of a JDRF career development award.
References
A table elsewhere in this issue shows conventional and Système International (SI) units and conversion factors for many substances.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.