Conceived and designed the experiments: JCD CC SP. Performed the experiments: DG MM YX SP. Analyzed the data: DG JCD CC SP. Contributed reagents/materials/analysis tools: ER. Wrote the paper: DG JCD CC.
The authors have declared that no competing interests exist.
Progression of chronic kidney disease (CKD) is a major health issue due to persistent accumulation of extracellular matrix in the injured kidney. However, our current understanding of fibrosis is limited, therapeutic options are lacking, and progressive degradation of renal function prevails in CKD patients. Uncovering novel therapeutic targets is therefore necessary.
We have previously demonstrated reversal of renal fibrosis with losartan in experimental hypertensive nephropathy. Reversal was achieved provided that the drug was administered before late stages of nephropathy, thereby determining a non-return point of CKD progression. In the present study, to identify factors critically involved in the progression of renal fibrosis, we introduced losartan at the non-return point in L-NAME treated Sprague Dawley rats. Our results showed either reversal or progression of renal disease with losartan, defining 2 groups according to the opposite evolution of renal function. We took advantage of these experimental conditions to perform a transcriptomic screening to identify novel factors potentially implicated in the mechanisms of CKD progression. A secondary analysis of selected markers was thereafter performed. Among the targets identified, periostin, an extracellular matrix protein, presented a significant 3.3-fold higher mRNA expression in progression compared to reversal group. Furthermore, independent of blood pressure, periostin was strongly correlated with plasma creatinine, proteinuria and renal blood flow, hallmarks of hypertensive renal disease severity. Periostin staining was predominant in the injured regions, both in experimental hypertensive and human nephropathy.
These results identify periostin as a previously unrecognized marker associated with disease progression and regression in hypertensive nephropathy and suggest measuring periostin may be a sensitive tool to evaluate severity, progression and response to therapy in human kidney disease associated to hypertension.
In the kidney, sustained insult commonly leads to an increased synthesis of extracellular matrix, which surrounds and eventually replaces the injured structures. In chronic kidney diseases, this fibrotic process spontaneously autoaggravates and contributes to a progressive reduction in the number of functioning nephrons, irrespective of the initial cause of the disease
We have previously demonstrated the possibility of therapeutic reversal of renal fibrosis in experimental hypertensive nephropathy, especially with losartan, an angiotensin II receptor antagonist
In the present study we hypothesized that actors crucially involved in the orientation of disease at the non-return point may play an important role in the pathophysiology of renal fibrosis, and may consequently be useful biomarkers of ongoing injury and promising therapeutic targets. To identify candidate proteins we performed a transcriptomic analysis of factors associated with the progression of chronic kidney disease. Thereafter, we further characterized selected targets at different stages of hypertensive nephropathy, including progression and reversal of renal disease after introduction of losartan. We concluded that periostin expression more than indices of endothelial or tubular dysfunction was strongly related to the progression and the regression of experimental hypertensive nephropathy, independently of changes in systolic blood pressure.
After initiation of L-NAME treatment, rats rapidly developed severe persistent hypertension (MAP = 211±7 mmHg, and 212±5 mmHg at 6 and 10 weeks treatment respectively) (
Original magnification ×20.
Original magnification ×20.
C | LOS | LN 6w | LN 10w | LN reg | LN no reg | |
|
||||||
MAP (mmHg) | 128±4 | 118±9 | 211±7 |
212±5 |
167±5 |
202±5 |
P/C (g/mmol) | 0.16±0.03 | 0.17±0.02 | 1.31±0.24 |
0.95±0.13 |
0.30±0.02 | 0.45±0.06 |
pCr (µmol/l) | 43±2 | 43±2 | 64±8 | 100±14 |
50±2 | 76±3 |
RBF (ml/min) | 10.1±0.3 | 10.5±0.4 | 4.4±0.5 |
3.3±0.4 |
8.8±0.4 | 6.4±0.5 |
|
||||||
Interstitial Fibrosis | 0.3±0.2 | 0.2±0.1 | 0.6±0.3 | 2.1±0.4 |
0.7±0.2 | 1.1±0.2 |
Vascular Fibrosis | 0.2±0.1 | 0.2±0.2 | 0.6±0.2 | 1.6±0.3 |
0.4±0.2 | 1.3±0.2 |
Glomerulosclerosis | 0.1±0.1 | 0.2±0.3 | 0.6±0.2 | 1.4±0.4 |
0.3±0.1 | 1.0±0.2 |
Tubular Lesions | 0.1±0.1 | 0.1±0.2 | 0.5±0.2 | 1.8±0.4 |
0.4±0.1 | 0.8±0.2 |
Inflammation | 0.4±0.2 | 0.3±0.3 | 1.4±0.5 |
2.2±0.4 |
0.6±0.1 | 1.9±0.2 |
CD3+ cells/mm2 | 6±1 | 6±2 | 53±6 |
100±14 |
37±6 |
67±12 |
C: Control, LOS: Losartan, LN: L-NAME, LN Reg: L-NAME Regression, LN No Reg: L-NAME No Regression.
MAP: Mean Arterial Pressure, P/C: Proteinuria/Creatininuria, pCr: Plasma Creatinine, RBF: Renal Blood Flow.
N for each group is shown in
To study determinants of renal disease progression/regression, hypertensive rats displaying standardized mild nephropathy (proteinuria/creatininuria ∼1 g/mmol) were treated with losartan, together with the continuation of L-NAME. Important heterogeneity in the evolution of renal disease was observed.
Rats were thereafter separated in two groups according to the median value of plasma creatinine after 4 weeks losartan+L-NAME treatment (
Transciptomic analysis was performed in a limited number (n = 4) of animals from the No Reg and Reg groups comparing the expressions of 180 genes involved in fibrosis, extracellular matrix regulation and inflammation.
As shown by RT-qPCR, periostin was expressed in the rat kidney at baseline and was 13- and 18-fold up-regulated after 6 and 10 weeks L-NAME, respectively (
(A): Real-time quantitative PCR for periostin renal mRNA expression. Results are expressed as arbitrary units. Error bars represent SEM. *
Immunostainings revealed a weak expression of periostin in the normal rat kidney, within the media of arteries and arterioles. Hypertensive nephropathy was characterized by a strong increase in periostin staining in the media and the adventitia of renal vessels. Interestingly periostin also exhibited a focal
Original magnification ×20.
To evaluate the importance of periostin as a marker and/or an actor of kidney disease progression, we performed regression analyses with hallmarks of hypertensive nephropathy severity as dependent variables. We found a very strong association between periostin mRNA expression and plasma creatinine (r = 0.68,
Plasma Creatinine | Proteinuria/Creatininuria | Renal Blood Flow | |
Periostin | 0.68 |
0.71 |
−0.64 |
Col1A2 | 0.41 |
0.57 |
−0.45 |
Col3A1 | 0.42 |
0.36 |
−0.41 |
MMP2 | 0.52 |
0.32 | −0.32 |
Vimentin | 0.63 |
0.67 |
−0.59 |
MCP-1 | 0.22 | 0.22 | −0.22 |
VCAM-1 | 0.38 |
0.20 | −0.30 |
ESel | 0.40 |
0.47 |
−0.33 |
ET-1 | 0.38 |
0.51 |
−0.61 |
All variables were log-transformed, except from Renal Blood Flow which exhibited normal distribution.
Values are r coefficients.
Since endothelial injury is present in hypertension and may play a central role in the pathophysiology of renal vascular diseases, we investigated the transcriptional regulation of ET-1 propeptide and E-selectin in this model (
Real-time quantitative PCR for endothelin-1 (ET1, A) and E-selectin (ESel, B) Col3A1 (C) and vimentin (D). Results are expressed as arbitrary units. Error bars represent SEM. *
Although significant fibrosis was detected on week 10, but not on week 6 (
To determine whether the latter findings may translate to human disease, we further analyzed periostin immunostaining on human kidney biopsy specimens. Periostin was not expressed in glomeruli or tubules in normal renal tissue. Only a weak staining could be found in small vessels (
Original magnification ×40. Periostin (E) and vimentin (F) staining on serial sections in chronic allograft nephropathy. Original magnification ×20.
The renin-angiotensin system is a central player in multiple mechanisms responsible for the progression of renal fibrosis
In this study we analyzed the reversal of L-NAME hypertensive nephropathy with losartan, and focused on the identification of markers of a non-return point of renal disease reversal. Our results show that losartan ameliorated renal hemodynamic alterations, proteinuria, and vimentin expression in all groups treated, compared to L-NAME 6w and 10w groups. Since these variables are classical determinants of renal disease progression, it could have been expected that the latter effects would be associated with a global protection against the functional and structural alterations ultimately induced by NO-deficiency hypertension. Instead, in spite of these beneficial effects, the No Reg group presented severe functional and structural disease, characterized by elevated plasma creatinine and vascular fibrosis similar to LN 10w group. Important additional factors implicated in renal disease progression and influenced by losartan are therefore responsible for the differential evolution between the Reg and No Reg groups. The reason why these important factors were different between the two groups is uncertain. In spite of the standardized cut-off point chosen to introduce losartan, we cannot exclude that the heterogeneity observed between Reg and No Reg groups may be due to subtle differences in the evolution of the renal disease around the non-return point, when losartan was introduced. Alternatively, preexisting heterogeneity in the regulation of pro- or anti-fibrotic genes between the Sprague Dawley rats, an outbred strain in which genome is not totally identical between animals, could account for the differences between the Reg and No Reg groups. We took advantage of these original experimental conditions to perform a transcriptomic analysis of selected markers associated with the progression of hypertensive nephropathy.
In the Reg group, rats presented reduced renal fibrosis compared to the No Reg group. Renal fibrosis is characterized by the accumulation of extracellular matrix, including fibrillar collagen. Since collagen III production, evaluated by Col3A1 RT-qPCR, was not different between the Reg and No Reg groups, the histological differences observed may be due to increased degradation of fibrillar collagen, as previously described in our laboratory
L-NAME hypertension is associated with features of endothelial dysfunction
Periostin is an extracellular matrix protein first identified in the periosteum and the periodontal ligament
Our results identified a progressive induction of periostin in the kidney with the progression of the hypertensive nephropathy. Regression analyses found a strong association between periostin mRNA expression and robust functional markers of kidney disease. Importantly, these associations held true when systolic blood pressure was added as a covariate, which suggests that periostin is correlated to renal injury independently of the degree of hypertension. Immunohistochemical analyses revealed that the localization of periostin was predominantly perivascular, in areas where important deposits of extracellular matrix occur in this model. We also observed an intense predominantly extracellular staining for periostin in the injured fibrotic tubulo-interstitial regions of chronic allograft nephropathy, which further demonstrates overexpression of periostin in human kidney disease. Identification of the main cells responsible for the interstitial accumulation of periostin requires further investigation. Data from previous studies suggests that fibroblasts, smooth muscle cells and tubular epithelial cells may be involved in periostin expression in this setting
We found that after the onset of hypertensive renal disease, curative treatment with losartan was associated with diminished periostin expression in Reg, not in No Reg group, which suggests that the reduction of periostin may be implicated in the mechanisms of angiotensin II-related disease progression and that reduction of periostin expression may be a critical determinant of disease reversal. Interestingly, the analysis of histological fibrosis scores shows that the difference between Reg and No Reg groups was most evident for perivascular fibrosis, in accordance with periostin distribution in experimental hypertensive nephropathy. These original data suggest that periostin may be related to the pathophysiology of extracellular matrix accumulation at the site of renal injury.
Together, the results of this work identify periostin as a previously unrecognized marker associated with hypertensive nephropathy. Further research is necessary to precise the potential of renal, plasma and urine periostin as prognostic biomarkers to monitor the progression and therapeutic control of human chronic kidney diseases.
Male Sprague-Dawley rats, weighing 250 g, were maintained on a normal-salt diet and had free access to chow and tap water. NO synthesis was inhibited by L-NAME (orally, 15 mg.kg−1.day−1). We have previously found that this dose produced a gradual elevation of blood pressure accompanied by the progression of renal disease. As shown in
After anesthesia by pentobarbital sodium (50–60 mg/kg intraperitoneally, Nembutal, Abbott, Chicago, IL), animals were placed on a servo-controlled table kept at 37°C and the trachea was cannulated to facilitate respiration. The left femoral artery was catheterized for measurement of invasive arterial pressure, and a femoral venous catheter was used for infusion of volume replacement. An ultrasound transit-time flow probe (1RB, Transonic, Ithaca, NY) was placed around the left renal artery. Bovine serum albumin (4.7 g/dl of saline solution) was infused initially at 50 µl/min to replace surgical losses, and then at 10 µl/min for maintenance. Arterial pressure was measured via a pressure transducer (Statham P23 DB); RBF was measured by a flowmeter (T 420, low-pass filter, 40 Hz, Transonic). RBF values were controlled for zero offset determined at the end of an experiment after cardiac arrest. Data were recorded, stored, and analyzed using a DataTranslation analog-to-digital converter and the IOX software (EMKA Technologies, Paris, France).
Morning urine samples were collected over a standardized 4-h period. Urinary protein concentration was normalized to creatinine concentration, and values were expressed as g protein per mmol creatinine. Central venous blood samples were withdrawn on the last day of the study, and plasma creatinine (µmol/l) was measured by automated Jaffe's method.
Kidneys were stained with Masson's trichrome solution. Four-µm sections of kidneys were examined on a blinded basis by two investigators independently to estimate inflammation, tubular lesions, interstitial fibrosis, vascular fibrosis, glomerular sclerosis and vascular necrosis using a 0 to 4 injury scale as described previously
Four-micrometer-thick sections of paraffin-embedded kidneys were dewaxed, heated in citric acid solution (pH 6) at 98°C for 30 min, and incubated first with a polyclonal goat anti-rat CD3 antibody recognizing T lymphocytes (Santa Cruz Biotechnology, Santa Cruz, CA) for two hours at 37°C and then incubated for 30 min at room temperature with a second antibody from N-Histofine kit (Nichirei Biochemicals, Japan). Staining was revealed by applying AEC (Dako), counterstained with hematoxylin QS (Vector, Burlingame, CA), and finalized with Permanent Aqueous Mounting Media (Innovex). Quantification of CD3-positive cells was performed using Olympus analysis software.
On rat tissue, 4-µm thick frozen sections were incubated with polyclonal rabbit anti-rat periostin antibody (Abcam, Cambridge, UK) overnight at 4°C and then incubated for 30 min at room temperature with a second antibody from N-Histofine kit (Nichirei Biochemicals, Japan). Staining was revealed by applying AEC (Dako), counterstained with hematoxylin QS (Vector, Burlingame, CA), and finalized with Permanent Aqueous Mounting Media (Innovex).
Renal biopsies from patients (1 normal kidney, 1 chronic allograft nephropathy) were retrospectively analyzed. Informed consent was given for use of part of the biopsy for scientific purpose. All procedures and use of tissue were performed according to the national ethical guidelines and were in accordance with the declaration of Helsinki. Paraffin-embedded sections were dewaxed and hydrated. The antigens were retrieved by 20-min boiling in 10 mM citric acid solution (pH 6). The sections were incubated overnight with 1/4000 anti-periostin polyclonal rabbit antibody (Biovendor, France) or 1/1000 anti-vimentin mouse monoclonal antibody V9 (Zymed, Invitrogen, Cergy Pontoise, France). The sections were then incubated with anti-rabbit or anti-mouse antibody conjugated with peroxydase-labeled polymer (Dako, Trappes, France). Immunoreactive proteins were visualized with a 3-amino-9-ethylcarbazole-containing peroxydase kit (Dako) and counterstained with hematoxylin. For negative controls, the primary antibody was replaced by an equal concentration of rabbit or mouse IgG.
We extracted RNA from the renal cortex and the abdominal aorta using TRIzol solution (Life Technologies BRL, Gaithersburg, MD). RNA quality was checked by measuring the ratio of optical densities at 260 and 280 nm and residual genomic DNA was removed by DNase I treatment for 30 min at 37°C (Fermentas). We used reverse transcription with Revert Aid H minus First Strand DNA Synthesis kit (Fermentas) to convert 1 µg RNA into cDNA. Transcriptomic analyses were performed with RT2Profiler PCR Array (Superarray, Bioscience Corp, Tebu Bio, Le Perray en Yvelines, France). cDNA was amplified by PCR using a LightCycler 480 (Roche Diagnostic) using SYBR Green (Fast Start DNA Master SYBRGreen I; Roche Applied Science, Roche Diagnostic), specific primers for selected mRNA and hypoxanthine-guanine phosphoribosyltransferase (HPRT) as housekeeping gene under the following conditions: 95°C for 5 min, and 45 cycles at 95°C for 15 s and 60°C for 15 s, then 72°C for 15 s. Specific primers were designed by Universal Probe Library system (UPL, Roche Applied Science), sequences are shown in
mRNA | Strand | Sequence |
Periostin | Sense |
|
Antisense |
|
|
Col3A1 | Sense |
|
Antisense |
|
|
Col1A2 | Sense |
|
Antisense |
|
|
MCP-1 | Sense |
|
Antisense |
|
|
VCAM-1 | Sense |
|
Antisense |
|
|
MMP2 | Sense |
|
Antisense |
|
|
ESel | Sense |
|
Antisense |
|
|
Vimentin | Sense |
|
Antisense |
|
|
TGF-beta1 | Sense |
|
Antisense |
|
|
ET-1 | Sense |
|
Antisense |
|
|
HPRT | Sense |
|
Antisense |
|
Statistical analyses for the
(PPT)
(PPT)