Effects of tumor necrosis factor-α inhibition on kidney fibrosis and inflammation in a mouse model of aristolochic acid nephropathy

Tumor necrosis factor (TNF)-α is a potent mediator of inflammation and is involved in the pathophysiology of chronic kidney disease (CKD). However, the effects of TNF-α inhibition on the progression of kidney fibrosis have not been fully elucidated. We examined the effects of TNF-α inhibition by etanercept (ETN) on kidney inflammation and fibrosis in mice with aristolochic acid (AA) nephropathy as a model of kidney fibrosis. C57BL/6 J mice were administered AA for 4 weeks, followed by a 4-week remodeling period. The mice exhibited kidney fibrosis, functional decline, and albuminuria concomitant with increases in renal mRNA expression of inflammation- and fibrosis-related genes. The 8-week ETN treatment partially but significantly attenuated kidney fibrosis and ameliorated albuminuria without affecting kidney function. These findings were accompanied by significant suppression of interleukin (IL)-1β, IL-6, and collagen types I and III mRNA expression. Moreover, ETN tended to reduce the AA-induced increase in interstitial TUNEL-positive cells with a significant reduction in Bax mRNA expression. Renal phosphorylated p38 MAPK was significantly upregulated by AA but was normalized by ETN. These findings indicate a substantial role for the TNF-α pathway in the pathogenesis of kidney fibrosis and suggest that TNF-α inhibition could become an adjunct therapeutic strategy for CKD with fibrosis.

www.nature.com/scientificreports/ and renal levels of TNF-α have been reported in human CKD and experimental kidney disease, including unilateral ureteral obstruction (UUO), ischemia-reperfusion injury (IRI), and cisplatin-induced nephropathy models 13,[15][16][17] . Furthermore, high serum levels of TNF-α are positively correlated with the severity of kidney injury 15 . Pharmacological and genetic inhibition of TNF-α ameliorates kidney injury in diabetic nephropathy and IRI models in which inflammation plays a critical role in the pathogenesis [18][19][20] . On the other hand, to elucidate the pathogenesis of the TNF-α pathway in the development of kidney fibrosis, several previous studies have employed the UUO model which is the most widely used mouse model of kidney fibrosis. In these studies, TNF-α inhibition using the pegylated form of soluble TNF receptor type 1 or genetic ablation of TNF receptors successfully had an anti-fibrotic effect against kidney fibrosis induced by the UUO model, indicating that the TNF-α pathway is a promising therapeutic target for kidney fibrosis [21][22][23][24][25] . However, the UUO model has potential limitations, such as the inability to assess changes in kidney function and the rapid progression to kidney fibrosis caused by hydronephrosis, which differs from the process of kidney fibrosis typically observed in human CKD 26 .
In this study, we employed aristolochic acid (AA)-induced nephropathy (AAN) as another model of kidney fibrosis to corroborate the findings reported for the UUO model. AAN is caused by direct cytotoxicity to renal tubular cells and is characterized by progressive tubular atrophy accompanied by chronic interstitial inflammation and fibrosis [27][28][29][30][31][32][33] . To investigate whether TNF-α inhibition prevents the development of kidney fibrosis, and to elucidate the underlying molecular mechanisms, we evaluated the effects of etanercept (ETN), which is a fusion protein competitively acting as a "TNF-α decoy receptor" to inhibit the binding of TNF-α to its cell surface receptor 34 , on kidney functional decline, inflammation, and fibrosis in a murine model of AAN.

Results
Etanercept does not affect body weight, tissue weight, systolic blood pressure, or heart rate. The changes in body weight (BW), tissue weight, systolic blood pressure (BP), and heart rate (HR) are shown in Fig. 1b-d. At the start of the experimental period, there were no significant differences in BW, systolic BP, or HR among control, AA and AA + ETN groups. BW in the AA and AA + ETN groups decreased during the 4 weeks of AA administration and increased after AA discontinuation ( Fig. 1b). At the end of the experimental period, BW was significantly lower in the AA and AA + ETN groups than in the control group (P < 0.001 and P < 0.001, respectively). There were no significant differences in heart to BW ratio among the three groups (Fig. 1c), but the kidney to BW ratio was significantly decreased in the AA and AA + ETN groups compared to the control group (both P < 0.001). There were no significant differences in BP and HR among the three groups ( Fig. 1d-e).
Etanercept does not affect plasma creatinine, blood urea nitrogen, or creatinine clearance but reduces albuminuria. Parameters of kidney function are shown in Fig. 2a. Plasma creatinine (Cr) and blood urea nitrogen (BUN) levels were significantly increased in the AA and AA + ETN groups compared to the control group (Cr, P < 0.001 and P < 0.001; BUN, P < 0.001 and P < 0.001, respectively). In addition, creatinine clearance levels were significantly decreased in the AA and AA + ETN groups compared to the control group (both P < 0.001). There were no differences between the AA and AA + ETN groups in these parameters. Urinary albumin excretion in the AA and AA + ETN groups was significantly increased compared to the control group (both P < 0.001); however, urinary albumin excretion in the AA + ETN group was significantly decreased compared to the AA group (P = 0.047). In the pathological evaluation, the glomerular area in the AA group was significantly decreased, and that in the AA + ETN group tended to be decreased compared to the control group (P = 0.001 and P = 0.095, respectively) ( Fig. 2b).
Etanercept partially reduces kidney fibrosis without affecting TGF-β expression. Renal mRNA expression levels of type I and III collagen, major components of the extracellular matrix in the kidneys, were significantly increased in the AA and AA + ETN groups compared to the control group (both P < 0.001) (Fig. 3a). However, the increased expression of type I and III collagen was partially but significantly attenuated by ETN (P = 0.011 and P = 0.012, respectively). Renal mRNA expression of TGF-β1, a pivotal profibrotic cytokine 9 , was significantly and similarly increased in the AA and AA + ETN groups compared to the control group (P < 0.001 and P < 0.001, respectively). In the pathological evaluation, AA administration caused severe kidney fibrosis, indicated by an increased area of picrosirius red (PSR) staining; however, ETN partially but significantly reduced kidney fibrosis (Fig. 3b). The quantitative assessment showed a significant decrease in the area of kidney fibrosis in the AA + ETN group compared to the AA group (P = 0.030). In a preliminary experiment in which we administered ETN at a lower dose (1 mg/kg, twice per week), the expression of type I and III collagen and pathological kidney fibrosis were comparable between the AA and AA + ETN groups (see Supplementary Fig. S1 online).

Etanercept reduces proinflammatory cytokine expression and phosphorylated p38-MAPK.
To determine how etanercept suppressed kidney fibrosis induced by AA, we examined the mRNA expression of genes involved in inflammation ( Fig. 4a-b). AA administration elicited remarkable increases in renal TNF-α, interleukin (IL)-6, and IL-1β expression (all P < 0.001). Although there were no significant differences in renal TNF-α expression between the AA and AA + ETN groups, renal IL-6 and IL-1β expression was significantly decreased in the AA + ETN group compared to the AA group (P = 0.026 and P = 0.016, respectively) (Fig. 4a). Regarding macrophage infiltration, renal MCP-1, F4/80, and CD68 expression were markedly and similarly increased in the AA and AA + ETN groups (both P < 0.001). Immunohistostaining of F4/80 revealed obvious renal interstitial macrophage infiltration in the AA and AA + ETN groups (Fig. 4c). Because MAPK pathways are downstream of the TNF-α pathway and are involved in the pathogenesis of kidney fibrosis 14 www.nature.com/scientificreports/ elevated in the AA group compared to the control group (P = 0.011) but was normalized by ETN (P = 0.034) (Fig. 4d).

Etanercept tends to suppress renal interstitial cell apoptosis, and reduces Bax expression.
To investigate the mechanism of the anti-fibrotic effect of TNF-α inhibition, we performed a TdT-mediated dUTP nick-end labeling (TUNEL) assay and evaluated apoptosis-related factors. Although the number of TUNELpositive cells was significantly increased in the AA and AA + ETN groups compared to the control group (both P < 0.001) ( Fig. 5a-b), ETN tended to reduce the number of TUNEL-positive cells but this trend did not reach statistical significance (P = 0.080). Renal expression of Bax, a pro-apoptotic factor, was significantly increased in the AA group compared to the control group (P < 0.001) but was normalized by ETN (P < 0.001) (Fig. 5c). The anti-apoptotic factor Bcl-2, activated by AA, was not affected by ETN.

Discussion
We demonstrated that TNF-α inhibition by ETN attenuated kidney inflammation, fibrosis, and albuminuria in a mouse model of AAN. TNF-α inhibition suppressed the AA-induced increase in renal expression of inflammation-and fibrosis-related genes, including IL-1β, IL-6, and type I and III collagen. Moreover, TNF-α inhibition tended to reduce the AA-induced increase in renal interstitial TUNEL-positive cells and significantly reduced the AA-induced increase in renal Bax mRNA expression, a pro-apoptotic factor. These findings imply that the  www.nature.com/scientificreports/ TNF-α pathway plays a substantial role in the development of kidney fibrosis, and that TNF-α inhibition may exert its anti-fibrotic effect by suppressing kidney inflammation and renal interstitial cell apoptosis. p38 MAPK is downstream of the TNF-α signaling pathway, and its activation promotes apoptosis and the production of proinflammatory cytokines, including TNF-α, IL-1β, and IL-6 13,14,36 . In this study, the AA-induced increase in renal phosphorylated p38 MAPK was normalized by ETN, suggesting that ETN attenuated the upregulated renal TNF-α pathway in the AAN model. This finding is consistent with the result that the AA-induced increase in renal expression of proinflammatory cytokines was significantly attenuated by ETN. Nonetheless, the renoprotective effect of ETN was limited to partial improvement of fibrosis and albuminuria. These results may be associated with the TGF-β pathway, macrophage infiltration, or the method of inhibiting the TNF-α pathway.
In this study, the AA-induced increase in renal TGF-β expression was not affected by ETN. The TGF-β pathway is a master regulator of fibrosis that promotes the epithelial-mesenchymal transition and production of extracellular matrix protein via the Smad-signaling pathway 9,10 . Therefore, the lack of a preventive effect of ETN on upregulation of the renal TGF-β pathway may explain its limited anti-fibrotic effect in the AAN model. Notably, previous studies on the roles of the TNF-α and p38 MAPK pathways in the development of kidney injury also demonstrated that inhibiting these pathways does not attenuate renal TGF-β expression 22,35,37,38 , indicating that the TGF-β pathway is less influenced by inhibiting the TNF-α pathway. No study has evaluated the effects of inhibiting both TNF-α and TGF-β signaling on kidney fibrosis, and future studies addressing this issue might make the anti-fibrotic effect of TNF-α inhibition shown in this study more prominent. www.nature.com/scientificreports/ Macrophages play a pathogenic role in the development of kidney inflammation and fibrosis 11,39 . Various strategies, including depletion or repletion of macrophages and blockade of chemokines or their receptors, have demonstrated a pivotal role for macrophages in kidney diseases [40][41][42][43][44][45][46] . However, in the present study, ETN did not affect renal interstitial macrophage infiltration induced by AA, and the renal gene expression levels related to macrophage infiltration, including MCP-1, F4/80, and CD68, were not affected by TNF-α inhibition. Previous studies on the effects of TNF-α inhibition in kidney injury models have also reported a lack of improvement of macrophage infiltration. For example, Meldrum et al. showed that administering PEG-TNFR1 has no effect on macrophage infiltration in a rat model of UUO 25 . Similarly, Misaki et al. reported that administering ETN reduces interstitial mononuclear cell infiltration 3 days after, but not 7 and 14 days after UUO 47 . The reason that ETN did not affect renal interstitial macrophage infiltration in the AAN model is unclear, but insufficient suppression of macrophage infiltration may explain the limited anti-fibrotic effects of ETN.
Genetic ablation of TNF-α receptors has been employed in several studies as another way to inhibit the TNF-α pathway 21,22 . In these studies, genetic ablation of TNF-α receptors had a certain anti-fibrotic effect (20-30% reduction) against UUO-induced kidney fibrosis, which seemed to be more prominent than the results of our study. Therefore, the limited renoprotective effect in this study might be associated with an insufficient ability of ETN to inhibit the TNF-α pathway.
In this study, ETN partially but significantly attenuated the AA-induced increase in albuminuria, whereas it did not affect structural changes in the glomerulus. Because AA causes proximal tubular epithelial cell death without affecting the glomerulus, the AAN model exhibits mild albuminuria, which is considered tubular proteinuria 29 . Given that ETN had its renoprotective effect mainly in the renal interstitium, the partial reduction  www.nature.com/scientificreports/ in AA-induced albuminuria may reflect the degree of improvement in renal interstitial damage by TNF-α inhibition. This study has limitations. First, an ETN group in which mice received an i.p. injection of DMSO and an s.c. injection of ETN was lacking. Therefore, it is unknown whether TNF-α inhibition has any effect on normal kidneys. Second, because ETN did not improve the AA-induced reduction in creatinine clearance, it is unclear whether TNF-α inhibition has a renoprotective effect on kidney functional decline in the AAN model. This absence of an effect may be attributed to the experimental design such as the experimental period, doses of ETN, or the kidney injury model. Studies using longer and/or different doses of ETN and other kidney injury models, such as adenine-infusion and 5/6 nephrectomy, are needed to address this issue. Third, although the present study demonstrates the anti-fibrotic effect of ETN by suppressing inflammation and apoptosis and inhibiting p38 MAPK signaling, it is unclear which factor contributed more prominently.
In conclusion, our findings show that TNF-α inhibition by ETN partially attenuates the development of kidney fibrosis and albuminuria by suppressing kidney inflammation and interstitial cell apoptosis in a mouse model of AAN. Our results support evidence of an anti-fibrotic effect of TNF-α inhibition on kidney fibrosis that has mainly been established by the UUO model. Collectively, these findings indicate a substantial role for the TNF-α pathway in the development of kidney fibrosis and imply that TNF-α inhibition could become an adjunct therapeutic strategy for CKD with fibrosis.

Methods
Animals. This study was performed following the National Institutes of Health guidelines for the use of experimental animals and was reviewed and approved by the Animal Studies Ethics Committee of Yokohama City University. All experiments were performed according to the ARRIVE guidelines. Male C57BL/6 J wild- www.nature.com/scientificreports/ type mice were purchased from Oriental Yeast Co, Ltd. The mice were housed in a controlled environment with a 12:12 light/dark cycle and ambient temperature (25℃) and humidity. The mice were allowed free access to food and water. They were fed a standard diet (0.5% NaCl, 3.6 kcal/g, and 13.3% energy as fat; Oriental MF, Oriental Yeast Co, Ltd.).

Experiment protocol and administration of aristolochic acid and etanercept. After 2 weeks
of acclimatization, mice (9-11 weeks old) were randomly assigned to either the control group, AA group, or AA + ETN group (n = 8 mice per group) (Fig. 1a). The control group received an intraperitoneal (i.p.) injection of dimethyl sulfoxide (DMSO) and a subcutaneous (s.c.) injection of normal saline (NS). The AA group received an i.p. injection of AA (3 mg/kg) and an s.c. injection of NS 32,48 . The AA + ETN group received an i.p. injection of AA (3 mg/kg) and an s.c. injection of ETN (5 mg/kg) [49][50][51] . The ETN dose was determined from previous studies and our preliminary experiment ( Supplementary Fig. S1 online). In the preliminary experiment, ETN at a lower dose (1 mg/kg, twice per week) did not exert any anti-fibrotic effect, thus we employed the higher dose (5 mg/ kg, twice per week) for this study. Figure 1a shows the time course of the experiment. AA or DMSO was administered twice per week for 4 weeks. ETN or NS was administered twice per week for 8 weeks. Based on previous studies 32, 48 , we adopted a protocol of a 4-week AA administration followed by a 4-week remodeling period, to evaluate kidney fibrosis in a more chronic phase.
Blood pressure and heart rate measurements. Systolic BP and HR were measured via a noninvasive procedure using a computerized tail-cuff plethysmograph (MK-2000 BP monitor; Muromachi Kikai, Tokyo, Japan), as described previously 52 . All measurements were performed at the same time between 10:00 and 13:00 to avoid diurnal variation in BP. At least 10 BP and HR measurements were analyzed. Systolic BP and HR were measured 0, 4, and 8 weeks after the intervention (Fig. 1a).

Metabolic cage analysis.
A metabolic cage analysis was performed at the end of the experimental period, as described previously 53 . Mice were housed in metabolic cages for 2 consecutive days and given free access to food and water. A 24 h urine collection procedure was performed on day 2.
Biochemical assays. Blood samples were collected by cardiac puncture when the mice were euthanized in the fed state. Whole-blood samples were centrifuged at 3,000 rpm at 4 °C for 10 min to separate plasma, which was snap-frozen and stored at -80 °C until use. Plasma Cr, BUN, and urinary Cr were measured using an autoanalyzer (Hitachi 7180; Hitachi, Tokyo, Japan). Urinary albumin was measured by an immunoturbidimetric assay (Fujifilm Wako Pure Chemical Corp., Tokyo, Japan).
Histological and immunohistochemical analyses. Histological analyses were performed as described previously 54,55 . Briefly, mouse kidneys were fixed in 4% paraformaldehyde in PBS, incubated overnight at 4 °C, and embedded in paraffin. Sections (4 μm thick) were stained with periodic acid-Schiff (PAS) and PSR. The glomerular area was measured by tracing the outline of the glomerular tuft of at least 50 glomeruli in the cortical fields of PAS-stained specimens. Fibrotic areas were measured digitally using a fluorescence microscope (BZ-X800; Keyence, Osaka, Japan) in the cortical fields of PSR-stained specimens. Immunohistochemistry was performed as described previously 56 . Sections were incubated with anti-F4/80 antibodies (1:100; ab111101; Abcam, Cambridge, MA, USA). The TUNEL assay was conducted using an in situ apoptosis detection kit (MK-500; Takara Biomedicals, Tokyo, Japan). Interstitial TUNEL-positive cells were counted in 10 randomly selected cortical fields (magnification: × 200). All measurements were blinded.

Real-time quantitative reverse transcription polymerase chain reaction (PCR) analysis. Total
RNA was extracted from kidney tissues using ISOGEN (Nippon Gene, Tokyo, Japan), and cDNA was synthesized using the SuperScript III First-Strand System (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's protocol. Real-time quantitative reverse transcription PCR analysis was performed using an ABI PRISM 7000 Sequence Detection System by incubating the reverse transcription products with the TaqMan PCR Master Mix and designed TaqMan probes (Applied Biosystems, Foster City, CA, USA Immunoblot analysis. Western blotting analysis was performed as described previously 56 . Briefly, kidney tissues were mechanically homogenized in SDS-containing sample buffer with the complete protease inhibitor cocktail (Roche, Basel, Switzerland), and were denatured by heating at 95 °C for 5 min. Proteins were quantified using the RC DC protein assay kit (Bio-Rad, Hercules, CA, USA). Equal amounts of protein extract were separated by 5-20% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene difluoride membrane using the iBlot Dry Blotting System (Invitrogen, Paisley, UK). The membranes were blocked with 5% skim milk for 1 h at room temperature and probed with specific primary antibodies to p38 MAPK (