The MEK Inhibitor Trametinib Ameliorates Kidney Fibrosis by Suppressing ERK1/2 and mTORC1 Signaling

Materials

Antibodies to vimentin, pAKT, AKT, pERK1/2, ERK1/2, pP70S6K, pS6, S6, p4E-BP1, pSTAT3, STAT3, pP65, and P65 were obtained from Cell Signaling Technology. Antibodies to β-actin, αSMA, and tubulin were obtained from Sigma. The antibody for NADPH oxidase-4 (Nox4) was obtained from Santa Cruz Biotechnology, the antibody for Ki-67 was obtained from Abcam, the F4/F80 antibody was obtained from Serotec, and TGF-β1 was obtained from R&D Systems. Trametinib was obtained from Selleck Chemicals. All other chemicals were obtained from Sigma, unless otherwise stated.

Animal Procedures and Trametinib Treatment

All animal experiments were conducted in accordance with the United Kingdom Home Office Animals 1986 Scientific Procedures, with local ethical committee approval (Project License 70/8356).

Male C57BL/6J mice (Charles River UK Ltd.) weighing between 20 and 25 g at 6–8 weeks of age were used. UUO was established according to published methods.^29,30 Briefly, mice were anesthetized with isoflurane and the abdominal cavity was exposed using midline laparotomy. Subsequently, the right ureter was isolated and tied off 0.5 cm from the pelvis, using a sterile 5–0 silk-braided suture (Pearsalls). The left ureter was left unclamped and served as the sham-operated control. All incisions were closed using 5–0 Proline suture (Ethicon). The following day and for 6 days in total, trametinib (3 mg/kg) was administered once daily by oral gavage (100 μl per animal). The final trametinib dose was given 2 hours before euthanizing. Trametinib was initially dissolved in DMSO and diluted into an aqueous pooled dose containing a final concentration of 0.5% hypromellose (Sigma) and 2% Tween-80 (Sigma), according to published methods.³¹ Control animals received the vehicle. Final DMSO concentration was 0.2% v/v, and trametinib or vehicle were administered in a blinded fashion. In a parallel study, trametinib (3 mg/kg) or vehicle were administered 4 days after UUO once daily for 6 days. At day 10 the kidneys were harvested. Both the kidneys from each animal (UUO and sham) were cut in half longitudinally. One half of each kidney was snap-frozen in liquid nitrogen and subsequently stored at −80°C for Western blotting. The other half was fixed with 10% Formalin (Sigma) for 16 hours at 4°C, and then transferred to a 70% v/v ethanol solution for a further 24 hours before embedding in paraffin for immunohistochemistry.

In a second set of experiments, mice were fed a rodent diet (824534 RM1) containing 0.15%w/w adenine (SDS diets; LBS-Biotech, Hookwood, UK) for 2 months. Excessive adenine is enzymatically modified and eventually precipitates in the tubules, resulting in progressive tubular injury culminating in necrosis, accompanied by macrophage infiltration, kidney fibrosis, and diminished renal function.³² We used 0.15% w/w adenine-containing diet because this was found to result in progressive kidney failure over time (up to 4 months) without the excessive weight loss observed with higher adenine-enriched diets (J. Kieswich, personal communication). Animals were administered trametinib (0.3 mg/kg) or vehicle once daily for 2 months by oral gavage, and kidneys were harvested as described for the UUO model.

Cell Culture and Protein Extraction

Primary human adult kidney fibroblasts from a single donor were obtained from Cambridge Bioscience and cultured in I-GRO medium containing growth supplements (Cambridge Bioscience). For acute (up to 30 minutes) TGF-β1 stimulation, cells (approximately 100,000) were serum-starved for 1 hour in a physiologic serum-free buffer as previously described.^33,34 For longer-term (24 hours) stimulation, fibroblasts were in complete medium including serum and growth supplements. Fibroblasts were preincubated with trametinib (10 or 20 nM) or vehicle for 30 minutes before TGF-β1 stimulation (5 ng/ml). At the end of the experiment fibroblasts were lysed and stored at −80°C until further use, as previously described.^33,35

Western Blot

Immunoblot analysis of kidney samples or fibroblast cellular lysates were conducted using the NuPAGE electrophoresis and buffers system (Invitrogen), as previously described.^33,35 Proteins were visualized with ECL Prime (GE Healthcare). Optical densities of bands of interest were determined using ImageJ 1.46r (National Institutes of Health) and normalized against the appropriate loading controls. The value of the normalized control sample was arbitrarily set to 1. Membranes were stripped using the Restore Plus reagent (Fisher Scientific), and reprobed for tubulin, β-actin, total ERK1/2, total STAT3, or total P65. For technical reasons, total S6 and total AKT were estimated from parallel Western blots.

Cell Proliferation Experiments

Human renal fibroblasts were seeded at a density 3×10³ cells per well, in 96-well plates, (Nunc) in complete medium. The following day, fresh medium was added along with the indicated concentrations of trametinib or vehicle. After 30 minutes of incubation, 5 ng/ml TGF-β1 was added to the corresponding wells. After 24 or 48 hours of culture, the number of viable cells in each well was quantified with the MTT assay (Promega) according to the manufacturer’s instructions. Each condition was assayed in triplicate and the experiment was repeated three times. All data were normalized against the absorbance of the unstimulated (no TGF-β1) control sample at 24 hours.

Immunohistochemical Staining

Formalin-fixed kidneys were embedded in paraffin and 4-μm sections were cut by the Barts Cancer Institute Pathology Unit. Staining with Sirius red or hematoxylin and eosin were performed as previously described.³⁶ Immunostaining was performed with the DISCOVERY XT (Ventana) automated slide processing instrument using the OmniMap reagents (Ventana), according to the manufacturer’s recommendations. Images were captured at ×20 magnification, unless otherwise stated, using a Zeiss AxioPhot microscope with an AxioCam HRc camera. Five images of kidney cortex were captured per mouse and staining was quantified as percentage of total area, using ImageJ 1.46r.³⁷ For Ki67 staining, positive nuclei per field of view were counted instead.

Urine and Blood Biochemistry

Twenty four hours before the end of the adenine-diet experiment, animals were placed in metabolic cages for urine collection. Blood was taken by cardiac puncture into heparinized syringes, decanted into Eppendorf tubes, and placed on ice. Subsequently, the blood was centrifuged at 9900×g for 3 minutes to separate serum. Creatinine clearance (C_CL, an indicator of GFR) was calculated using the following formula: C_CL (ml/min)=[C_U (µmol/L)×urine flow (ml/min)]/C_S (µmol/L), where C_U is the concentration of creatinine in the urine and C_S is the creatinine concentration in the serum. Fractional excretion of sodium (FE_NA), an indicator of tubular function, was calculated with the following formula: FE_Na (%)=100×[Na_U (µmol/L)×urine flow (ml/min)]/[C_CL (ml/min)×Na_S (µmol/L)], where Na_U is the sodium concentration in the urine and Na_S is the sodium concentration in the serum. All biochemical parameters in serum and urine were measured by a commercial veterinary testing laboratory (IDEXX Bioresearch, Ludwigsburg, Germany).

Statistical Analyses

Data are expressed as the means±SEM. Statistical significance was determined by one- or two-way analysis of variance and Tukey post hoc test, using GraphPad Prism v5, as appropriate. Values of P<0.05 were deemed statistically significant. Where possible, experimenters were blinded. No animals were excluded from the analysis.

Trametinib Ameliorated Kidney Fibrosis after UUO

We investigated whether ERK1/2 pathway inhibition can hinder renal fibrosis progression by subjecting mice to UUO. As reported previously,^29,30 UUO kidneys exhibited severe morphologic lesions characterized by tubular dilation and tubulointerstitial expansion (Figure 1A). Conversely, the kidneys of animals treated with trametinib displayed considerably less morphologic abnormalities (Figure 1A). To measure collagen deposition in the interstitium, a hallmark of fibrosis,^2,3 kidneys were stained with Sirius red (Figure 1B). UUO significantly increased Sirius red staining, whereas staining was dramatically decreased by 92% in the kidneys of trametinib-treated animals (Figure 1C).

Next, the effect of trametinib on the expression of the myofibroblast marker αSMA^8–10 was investigated. UUO resulted in considerably increased αSMA-positive staining, primarily in the tubulointerstitium (Figure 1D). In animals that received trametinib, αSMA staining was reduced by 75% (Figure 1, D and E). Western blotting also confirmed this, revealing 62% reduction of αSMA expression in trametinib-treated UUO kidneys (Figure 1, F and G).

The cytoskeletal protein vimentin is upregulated in myofibroblasts⁵ and its expression can be modulated by ERK1/2.³⁸ Thus, we investigated whether ERK1/2 inhibition could suppress vimentin upregulation. No vimentin bands were detected in sham-operated kidneys by Western blotting, whereas dramatic upregulation of vimentin was observed in UUO samples (Figure 1F). Treatment with trametinib significantly attenuated vimentin expression by 68% (Figure 1H).

Trametinib Suppressed ERK1/2 and AKT Activation and Myofibroblast Expansion in UUO

As previously reported,^20,26 phospho-ERK1/2 was significantly upregulated in UUO kidneys, as determined by Western blot (Figure 2A) and immunostaining (Figure 2D). Trametinib administration completely suppressed ERK1/2 phosphorylation in both sham- and UUO-operated animals (Figure 2, A and B), as confirmed by immunohistochemistry (Figure 2, D and E). AKT is phosphorylated during UUO and contributes to injury.²⁶ Western blotting revealed augmented AKT phosphorylation after UUO and trametinib application reduced phospho-AKT levels by 46% (Figure 2, A and C).

Myofibroblast proliferation is a major driver of fibrosis^2,3 and ERK1/2 modulates renal fibroblast expansion.³⁹ Accordingly, we tested the effect of trametinib on myofibroblast proliferation by staining for Ki67, a marker of proliferating cells. Hardly any Ki67-positive nuclei were detected in sham kidneys (Figure 2F). UUO resulted in a dramatic increase in the detected Ki67-positive nuclei primarily in the interstitium, whereas trametinib significantly reduced the number of proliferating cells by 65% (Figure 2G).

Trametinib Blocked UUO-Induced mTORC1 Activation and Nox4 Overexpression

mTORC1 activation downstream of TGF-β1 promotes kidney fibrosis¹⁹ and contributes to injury by upregulating Nox4.^40–42 As previously reported,¹⁹ significant phosphorylation of the key downstream mTORC1 effectors p70S6 kinase (P70S6K) and eIF4E binding protein-1 (4E-BP1)⁴³ was observed upon UUO (Figure 3, A–C). In contrast, in trametinib-treated animals mTORC1 activation was significantly suppressed, as exemplified by the reduction of phospho-P70S6K and phospho-4E-BP1 levels by 80% and 81%, respectively (Figure 3, A–C). Increased phosphorylation of the ribosomal protein S6, a direct target of P70S6K⁴³ (Figure 3, A and D), was also attenuated by 85% by trametinib administration (Figure 3D). This was further corroborated by decreased phospho-S6 immunostaining observed in the kidneys of trametinib-treated animals (Figure 3, E and 3F). In agreement with previous studies reporting Nox4 upregulation downstream of mTORC1 and ERK1/2,^40–42 Western blotting revealed marked induction of Nox4 protein in UUO kidneys (Figure 3G). Trametinib significantly blunted Nox4 upregulation by 51% (Figure 3, G and H).

Trametinib Attenuated UUO-Induced Phosphorylation of STAT3, NF-κB, and Smad2/3 and Macrophage Infiltration

Renal fibroblast differentiation requires STAT3 phosphorylation and phospho-STAT3 is upregulated in myofibroblasts during UUO.⁴⁴ Moreover, STAT3 is activated downstream of mTORC1.⁴⁵ Negligible phospho-STAT3 was observed in sham-operated kidneys (Figure 4A), whereas UUO resulted in a striking increase of phospho-STAT3, which was significantly attenuated by trametinib administration by 80% (Figure 4, A and B).

Genes transcribed downstream of NF-κB prominently contribute to myofibroblast expansion and differentiation.²¹ As shown in Figure 4A, phosphorylation of the NF-κB subunit P65, was barely detectable in sham-operated kidneys, whereas considerable upregulation of phospho-P65 was observed upon UUO. Trametinib significantly suppressed phospho-P65 by 51% when normalized by total P65 levels (Figure 4, A and C). Total P65 was also significantly upregulated by 83% in the UUO kidneys when compared with controls, whereas trametinib administration blunted P65 expression, although the difference was still significant (Figure 4, A and D).

Phosphorylation of the main effectors of canonical TGF-β1 signaling, Smad2/3,¹⁴ was also enhanced in control UUO kidneys, and treatment with trametinib suppressed phospho-Smad2/3 levels by 65% (Figure 4, A and E).

Persistent inflammation is a major contributor to kidney fibrosis and infiltration of macrophages in UUO kidneys exacerbates injury.⁴⁶ Thus, we tested whether trametinib treatment reduced macrophage infiltration during UUO. A dramatic increase in F4/F80 staining (a marker of macrophages) was observed in the interstitium of UUO control animals, whereas significantly less staining (75% reduction) was detected in UUO kidneys from trametinib-treated animals (Figure 4, F and G).

Delayed Administration of Trametinib Suppressed Kidney Fibrosis and Profibrotic Signaling upon UUO

We determined whether ERK1/2 pathway inhibition could slow kidney fibrosis subsequent to injury by administering trametinib or vehicle 4 days post-UUO surgery for 6 days. Significant myofibroblast activation and phospho-ERK1/2 upregulation was observed in kidneys after 4 days of UUO (Supplemental Figure 1). Considerable Sirius red staining was observed in the tubulointerstitium after 10 days of UUO (Figure 5, A and B). Treatment with trametinib 4 days postsurgery decreased Sirius red staining by 74% (Figure 5, A and B). Western blotting revealed that in animals that received trametinib, αSMA and vimentin expression were significantly reduced by 60% and 77%, respectively (Figure 5, C–E). Decreased αSMA-positive immunostaining was also detected in the obstructed kidneys of these animals (Figure 5, F and G).

Significant upregulation of ERK1/2 and AKT phosphorylation was observed in kidneys 10 days post-UUO (Supplemental Figure 2, A–C). However, in animals that received trametinib 4 days postinjury, ERK1/2 and AKT activation were suppressed by 92% and 52%, respectively (Supplemental Figure 2, A–C). Immunohistochemical analysis confirmed almost complete phospho-ERK1/2 suppression in trametinib-treated animals (Supplemental Figure 2, D and E).

Immunoblotting revealed that delayed administration of trametinib resulted in significant reduction in mTORC1 pathway activation as judged by the reduction of the phosphorylated levels of P70S6K, 4E-BP1 and S6 by 68%, 88%, and 33%, respectively (Supplemental Figure 2, F–I).

Immunoblotting showed significant kidney Nox4 overexpression upon 10 days of UUO, which was blunted by trametinib administration (Supplemental Figure 3, A and B).

Substantial activation of STAT3, NF-κB, and Smad2/3 and upregulation of total P65 was also detected in UUO kidneys (Supplemental Figure 3, C–G). As shown in Supplemental Figure 3, D–G, trametinib administration significantly ameliorated STAT3 and Smad2/3 phosphorylation by 73% and 83%, respectively. Phospho-P65 was reduced significantly by 33% even when corrected for total P65, which was upregulated in UUO. Additionally, trametinib decreased macrophage infiltration by 74%, as demonstrated by immunostaining with F4/F80 (Supplemental Figure 3, H and I).

Trametinib Reduced Kidney Fibrosis in a Chronic Model of CKD Induced by Adenine-Rich Diet

We next evaluated whether trametinib could ameliorate renal fibrosis in a chronic model of CKD. Mice fed a diet containing 0.15% w/w adenine for 2 months³² received trametinib (0.3 mg/kg), a dose that results in approximately 75% ERK1/2 pathway inhibition,³⁷ or vehicle for the duration of the experiment.

As previously reported,³² gross morphologic lessons were observed in the kidneys of adenine-fed animals. Unlike UUO, trametinib did not appear to reduce tubular dilation (Figure 6A). Considerable Sirius red staining, an indicator of collagen deposition, was observed in the tubulointerstitium of adenine-diet fed animals. Treatment with trametinib (0.3 mg/kg) significantly reduced staining by 63% (Figure 6, B and C). Decreased αSMA immunostaining (by 66%) was also detected in the kidneys of adenine-fed animals that received trametinib (Figure 6, D and E). Western blotting revealed that trametinib treatment suppressed αSMA and vimentin expression by 41% and 57%, respectively (Figure 6, F–H).

In the kidneys of adenine-fed animals significant upregulation of ERK1/2 and AKT phosphorylation was observed (Supplemental Figure 4, A–C). ERK1/2 and AKT activation was suppressed by 84% and 56%, respectively, in the trametinib-treated animals (Supplemental Figure 4, A–C). Immunohistochemical analysis showed phospho-ERK1/2 suppression by 80% (Supplemental Figure 4, D and E), accompanied with a reduction of the number of proliferating, Ki67-positive cells (Supplemental Figure 4, F and G) in the kidneys of animals that received trametinib.

Increased mTORC1 pathway activation was observed by immunoblotting in the kidneys of adenine-fed animals and this was blunted by trametinib administration, as exemplified by the reduction of phospho-P70S6K, phospho-4E-BP1, and phospho-S6 by 64%, 73%, and 85%, respectively (Supplemental Figure 5, A–D). Additionally, considerable macrophage infiltration was observed in the kidneys of adenine-fed animals. Conversely, trametinib reduced F4/F80-positive immunostaining by 78% (Supplemental Figure 6, A and B).

Mice that were fed an adenine-rich diet exhibited a significant reduction in kidney function, as exemplified by a four-fold increase in plasma creatinine and urea, a two-fold increase in urine protein, and a significantly increased fractional excretion of sodium and creatinine clearance ratio (Table 1). Trametinib treatment, despite significantly reducing kidney fibrosis, did not improve kidney function in adenine-fed mice (Table 1).

Adenine-fed animals exhibited a 5% reduction in body weight after 2 months. Trametinib-treated mice displayed a modest, albeit significant, reduction in weight by further 7% (Supplemental Figure 7). This was not due to reduced food consumption because once weight differences became apparent (after 5–6 weeks), we measured food consumption for the remainder of the experiment and no significance differences were found (data not shown). Trametinib-treated animals did not display any signs of discomfort or reduced physical activity. A similar trend was observed in the control arm of the study. Given that ERK1/2 is known to regulate metabolism in mice,³¹ it is probable that the observed weight differences were due to altered basal metabolism. No excessive weight loss has been reported in humans that received trametinib.^27,28

Trametinib Ameliorated TGF-β1–Induced Renal Fibroblast Activation

We next investigated the effect of trametinib on renal fibroblasts expansion and signaling by stimulating primary human fibroblasts with TGF-β1, the predominant profibrotic cytokine.^13,15

Acute application of TGF-β1 resulted in significantly increased mTROC1 and ERK1/2 pathways activation, as previously reported^19,39 (Figure 7A). Trametinib (10 nM) completely inhibited ERK1/2 activation at all time points tested (Figure 7, A and B). Concurrently, at 5 minutes poststimulation, phosphorylation of P70S6K and S6 was attenuated by 88% and 75%, respectively (Figure 7, A, C, and D). Trametinib also significantly blunted mTORC1 pathway activation for all of the other time points tested (Figure 7A). Conversely, Smad2/3 phosphorylation was not significantly affected by trametinib, suggesting no direct interference with TGF-β1 receptor signaling (Figure 7, A and E).

Likewise, trametinib suppressed TGF-β1–induced ERK1/2 and mTORC1 activation in the presence of serum over 24 hours. Fibroblasts preincubated with 20 nM trametinib exhibited reduced phospho-ERK1/2 levels by 58% at 24 hours after TGF-β1 treatment (Figure 7, F and G). Phospho-P70S6K and phospho-S6 were also significantly suppressed by 67% and 91%, respectively, in these cells (Figure 7, F, H, and I). In contrast to short-term stimulation, trametinib also ameliorated TGF-β1–induced Smad2/3 phosphorylation after 24 hours (Figure 7, F and J). TGF-β1 significantly increased αSMA expression by 1.57-fold and trametinib (20 nM) resulted in 51% reduction in αSMA expression (Figure 7, K and L). Additionally, trametinib dose-dependently suppressed both TGF-induced and basal fibroblast proliferation, with a more prominent effect for 20 nM after 48 hours in culture (Figure 7, L and M). Specifically, at 48 hours, 20 nM trametinib reduced fibroblast expansion at basal or TGF-β1–stimulated conditions by 39% and 41%, respectively.

Considerable αSMA expression was detected in proliferating fibroblasts at basal conditions (no TGF-β1), suggesting that in the presence of serum, human fibroblasts acquired some myofibroblast properties. This is in agreement with studies using murine fibroblasts.^39,44 Nonetheless, TGF-β1 had an additive effect, significantly upregulating αSMA and increasing fibroblast proliferation (Figure 7, K and N).