NAD(P)H Oxidase Mediates TGF-β1–Induced Activation of Kidney Myofibroblasts

Abstract

TGF-β1 expression closely associates with activation and conversion of fibroblasts to a myofibroblast phenotype and synthesis of an alternatively spliced cellular fibronectin variant, Fn-ED-A. Reactive oxygen species (ROS), such as superoxide, which is a product of NAD(P)H oxidase, also promote the transition of fibroblasts to myofibroblasts, but whether these two pathways are interrelated is unknown. Here, we examined a role for NAD(P)H oxidase–derived ROS in TGF-β1–induced activation of rat kidney fibroblasts and expression of α-smooth muscle actin (α-SMA) and Fn-ED-A. In vitro, TGF-β1 stimulated formation of abundant stress fibers and increased expression of both α-SMA and Fn-ED-A. In addition, TGF-β1 increased both the activity of NADPH oxidase and expression of Nox2 and Nox4, homologs of the NAD(P)H oxidase family, indicating that this growth factor induces production of ROS. Small interfering RNA targeted against Nox4 markedly inhibited TGF-β1–induced stimulation of NADPH oxidase activity and reduced α-SMA and Fn-ED-A expression. Inhibition of TGF-β1 receptor 1 blocked Smad3 phosphorylation; reduced TGF-β1–enhanced NADPH oxidase activity; and decreased expression of Nox4, α-SMA, and Fn-ED-A. Diphenyleneiodonium, an inhibitor of flavin-containing enzymes such as the Nox oxidases, had no effect on TGF-β1–induced Smad3 but reduced both α-SMA and Fn-ED-A protein expression. The Smad3 inhibitor SIS3 reduced NADPH oxidase activity, Nox4 expression, and blocked α-SMA and Fn-ED-A, indicating that stimulation of myofibroblast activation by ROS is downstream of Smad3. In addition, TGF-β1 stimulated phosphorylation of extracellular signal–regulated kinase (ERK1/2), and this was inhibited by blocking TGF-β1 receptor 1, Smad3, or the Nox oxidases; ERK1/2 activation increased α-SMA and Fn-ED-A. Taken together, these results suggest that TGF-β1–induced conversion of fibroblasts to a myofibroblast phenotype involves a signaling cascade through Smad3, NAD(P)H oxidase, and ERK1/2.

Progression of renal fibrosis involves expansion of interstitial myofibroblasts and extracellular matrix accumulation, resulting in the loss of function and ultimately renal failure.^1,2 The origin of myofibroblasts is under extensive investigation, and evidence indicates the cells may be derived from several sources, including an expansion of activated resident fibroblasts, perivascular adventitial cells, blood-borne stem cells that migrate into the glomerular mesangial or interstitial compartment, or tubular epithelial-to-mesenchymal transition and migration into the peritubular interstitial space. Regardless of their origin, there is common agreement that the myofibroblast is the cell most responsible for interstitial expansion and matrix accumulation during the course of renal fibrosis. TGF-β1 is the predominant growth factor responsible for matrix synthesis by mesenchymal cells such as fibroblasts in vitro and during renal fibrosis.^3,4 Indeed, there is a close correlation in the cellular expression of TGF-β1, a fibroblast transition to an activated, α-smooth muscle actin (α-SMA)-positive myofibroblast phenotype, and synthesis of an alternatively spliced isoform of fibronectin, Fn-ED-A.⁵ TGF-β1 differentially regulates the expression of Fn-ED-A in fibroblasts^6–8 and induces expression of α-SMA in a variety of mesenchymal cells in culture.^9,10 Indeed, a functional ED-A domain is mandatory for α-SMA induction by TGF-β1.^7,8,10 Moreover, TGF-β1 is frequently associated with a myofibroblast phenotype in liver, lung, and kidney disease,^1,11–13 and all three proteins frequently co-localize in these disease settings. In addition, a co-localization of α-SMA and Fn-ED-A is frequently observed in fibrotic disease as well as in glomerular and interstitial lesions in kidney diseases previously investigated in our laboratory.^14–17

Accumulating evidence also indicates that reactive oxygen species (ROS), mainly in the form of superoxide, play a significant role in the initiation and progression of cardiovascular^18,19 and renal^20–25 disease. ROS are involved in distinct cell functions, including hypertrophy, migration, proliferation, apoptosis, and regulation of extracellular matrix.^25–28 More specific, the NAD(P)H oxidases of the Nox family have gained heightened attention as mediators of injury associated with vascular diseases, including hypertension, atherosclerosis, heart disease, and diabetes.^18,19,29,30 NAD(P)H oxidase generation of superoxide is recognized as an important mediator of cell proliferation in glomerulonephritis²² and matrix accumulation in diabetic nephropathy^25,31–33 and fibrosis.^21,24 Adventitial fibroblasts are also a major source of superoxide in the aorta,^19,34–36 therefore being highly relevant to renal disease. This is because the renal perivascular space is noticeably reactive and is the site where myofibroblasts may first appear during the course of renal disease and fibrosis.^17,37–39

The observations that both TGF-β1 and ROS induce fibroblasts to α-SMA–positive myofibroblast phenotype^40–42 suggest that these two pathways are interrelated and may share signaling pathways in kidney disease. TGF-β signaling occurs through a well-established process involving two downstream pathways: Smad and extracellular signal–regulated kinase (ERK).^43–45 TGF-β/Smad signaling (Smad2 and Smad3) is tightly controlled by mitogen-activated protein kinase (MAPK; ras/MEK/ERK) signaling cascades.⁴⁶ A regulatory role for ROS in PDGF and angiotensin II–induced signal transduction has gained recognition^47,48; however, a role for ROS in TGF-β signaling is less well understood. It is also unknown whether kidney myofibroblasts express NAD(P)H oxidase homologs or generate ROS in response to TGF-β1. Given TGF-β1–induced myofibroblast activation and matrix synthesis during renal disease may be linked to ROS, we examined a role for NAD(P)H oxidase in TGF-β1–induced Smad3 and ERK signaling as well as kidney myofibroblast activation, as assessed by a switch to an α-SMA–positive phenotype and expression of Fn-ED-A expression in vitro.

Results

TGF-β1 Activates a Myofibroblast Phenotype (α-SMA Expression) and Fn-ED-A Expression via ROS

α-SMA and Fn-ED-A are expressed in negligible amounts in resting fibroblasts as assessed by immunohistochemistry and immunoblotting (Figure 1). TGF-β1 activated a fibroblast-to-myofibroblast transition characterized by the acquisition of an α-SMA–positive phenotype showing increased protein in well-defined actin-positive stress fibers by immunofluorescence histochemistry (Figure 1A) and protein level by Western blotting (Figure 1B). In addition, TGF-β1 induced an increase in Fn-ED-A expression in these cells as detected by both methods. To examine whether stimulation of kidney fibroblasts by TGF-β1 works through enhanced ROS generation, we exposed the cells to diphenyleneiodonium (DPI) before addition of TGF-β1 and examined expression levels of α-SMA and Fn-ED-A. The results suggest that ROS plays a role in myofibroblast activation as indicated by DPI inhibition of TGF-β1–induced expression of both proteins in immunohistochemistry and immunoblotting procedures (Figure 1). Immunoblots show approximately 60% reduction in both proteins after DPI plus TGF-β1 relative to TGF-β1 alone.

Figure 1.

TGF-β1 induces a ROS-dependent kidney fibroblast conversion to a myofibroblast phenotype. (A and B) TGF-β1 induces a myofibroblast phenotype as indicated by an increase in α-SMA and Fn-ED-A protein expression by immunofluorescence microscopy (A) and Western blotting (B). DPI inhibits TGF-β1–enhanced expression of both proteins, implicating a role for ROS in kidney myofibroblast activation. Data are means ± SEM from three independent experiments. *P < 0.05 versus control; #P < 0.05 versus TGF-β1; §P < 0.05 versus TGF-β1+inhibitor.

TGF-β1 Stimulates a Rapid Increase in ROS Generation and NADPH Oxidase Activity

TGF-β1 stimulated a rapid increase in fibroblast generation of ROS as revealed by increased 2′,7′-dichlorodihydrofluorescein fluorescence from 15 min to 4 h after addition relative to diluent controls as detected by confocal microscopy (Figure 2, A and B). TGF-β1 also markedly enhanced NADPH oxidase activity as assayed by superoxide excitation of lucigenin using NADPH as a substrate. The results showed a near doubling of NADPH oxidase activity within 30 min after addition of TGF-β1 with sustained increases in activity during a 4-h period (Figure 2D). TGF-β1–induced elevations in NADPH oxidase activity were completely blocked by the DPI (Figure 2, C and D), indicating a substantial input to ROS generation from this enzyme family. These findings were verified in primary fibroblasts (A2F11), showing TGF-β1–stimulated increase in NADPH oxidase activity (Figure 2F) and a reduction in α-SMA and Fn-ED-A expression after treatment with DPI (Figure 2G).

Figure 2.

Effect of TGF-β1 on production of ROS in kidney fibroblasts is shown. Fibroblasts were serum-starved and treated with 1 ng/ml TGF-β1 for the indicated periods. 2′,7′-dichlorodihydrofluorescein (DCF) fluorescence reflecting the relative levels of ROS was imaged with a confocal laser scanning fluorescence microscope. (A through C) Representative photomicrographs show DCF fluorescence in kidney fibroblasts under basal conditions (A), 15 min after addition of TGF-β1 (B), and in cells pretreated with DPI (10 μM) before TGF-β1 (C). (D) TGF-β1 induced a sustained increase in the rate of NADPH oxidase activity (relative luminescent units per minute per milligram of protein) in these cells, peaking at 30 min after exposure as illustrated in a representative graph of three separate experiments. DPI inhibits both ROS detection by DCF (C) and NADPH oxidase activity (D). (E) Western analysis indicates that kidney fibroblasts express Nox2 and Nox4 as early as 5 min after TGF-β1 treatment, which progressively increases over time. Kidney fibroblasts do not express Nox1 or Nox3 homologs. *Nox4 seems to be the predominant homolog expressed by kidney fibroblasts given the amount of lysate loaded onto gels was one fourth that for all other homologs. (HEK: HEK293 cell lysates were used as a positive control for Nox1 and Nox3). (F and G) NADPH oxidase activation was verified in primary kidney fibroblast (A2F11) showing TGF-β1 induction of NADPH oxidase activity in a representative graph (F) and inhibition of growth factor–stimulated α-SMA and Fn-ED-A protein expression by DPI (G).

Kidney Fibroblasts Express Nox2 and Nox4

Western blot analysis of cell lysates from NRK-49F kidney fibroblasts grown under normal growth conditions showed no detectable expression of Nox1 or Nox3 homologs (Figure 2E); however, a weak basal expression of Nox2 and Nox4 proteins was observed in nonstimulated cells (Figure 2E). TGF-β1 had no effect on Nox1 or Nox3 expression; however, both Nox2 and Nox4 proteins were enhanced with progressive increases from 5 min to 24 h after addition of growth factor (Figure 2E). Signal for Nox4 expression began earlier and was more intense than Nox2 in controls and TGF-β1–stimulated cells considering that the amount of protein lysate loaded onto gels was one fourth the amount for detection of all other Nox homologs. These studies were verified by immunofluorescence microscopy (data not shown).

Inhibition of Nox4 Suppresses NADPH Oxidase Activity and Expression of α-SMA and Fn-ED-A

We used a small interfering RNA (siRNA) knockdown strategy to examine the contribution of Nox4 oxidase components to TGF-β1–induced enhancements of α-SMA and Fn-ED-A expression. The effect of siRNA in kidney fibroblasts was confirmed by the observation that siNox4 decreased Nox4 protein expression relative to scrambled (Scr) siRNA (Figure 3A). siNox4 also reduced TGF-β1–induced increments in NADPH oxidase activity (Figure 3B) and α-SMA and Fn-ED-A protein expression levels (Figure 3, C through E) relative to scrambled control (Scr siRNA). Scr siRNA increased the TGF-β1 response of α-SMA and Fn-ED-A expression. A cause for this finding is unclear, but these results were not significantly different from TGF-β1 alone and may be attributed to greater variability in expression levels in this treatment group. Overall, these data suggest that Nox4 is an important mediator of kidney fibroblast activation and matrix synthesis as assessed by α-SMA and Fn-ED-A expression through NAD(P)H oxidase–derived ROS.

Figure 3.

Nox4 mediates kidney myofibroblast activation in response to TGF-β1. siRNA complementary to Nox4 (siNox4) and scrambled siRNA (Scr) were examined in TGF-β1–induced myofibroblast activation. (A) The effect of siRNA in kidney fibroblasts was confirmed by the observation that siNox4 decreased basal Nox4 protein expression relative to Scr siRNA. (B through E) siNox4 also reduced TGF-β1–induced increments in NADPH oxidase activity (B) as well as α-SMA (C and D) and Fn-ED-A (C and E) protein expression, indicating a role for superoxide generation by this homolog in kidney myofibroblast activation. Data are means ± SEM from three independent experiments. *P < 0.05 versus control; #P < 0.05 versus TGF-β1; §P < 0.05 versus TGF-β1+Scr.

Inhibition of TGF-βR Abolishes Myofibroblast Activation, NADPH Oxidase Activity, and Nox4 Expression

TGF-βR1 kinase activity inhibitor SB431542 (SB) abolished α-SMA and Fn-ED-A expression (Figure 4A) and NADPH oxidase activity (Figure 4B) and markedly reduced Nox4 protein expression (Figure 4C), indicating that TGF-β1 regulates myofibroblast activation through its receptor and NAD(P)H oxidase–derived ROS generation.

Figure 4.

TGF-β1 stimulates kidney myofibroblast α-SMA and Fn-ED-A expression, NADPH oxidase activity, and Nox4 via TGFBR. To determine a role for NAD(P)H oxidase–derived ROS in TGF-β1 signaling of kidney myofibroblast activation, we used experiments inhibiting TGFR1 with SB and determining expression levels of α-SMA and Fn-ED-A, NADPH oxidase activity, and Nox4 protein. (A through C) SB inhibited TGF-β1–induced increases in α-SMA and Fn-ED-A expression (A) and substantially reduced NADPH oxidase activity shown in a representative graph (B) as well as Nox4 protein expression (C). Data are means ± SEM from three independent experiments. *P < 0.05 versus control; #P < 0.05 versus TGF-β1; §P < 0.05 versus TGF-β1+inhibitor.

Role of NADPH Oxidase in Smad and ERK1/2 Signaling

To determine a role for Smad3 and ROS signaling of kidney myofibroblast activation, we examined the effect of SIS3, an inhibitor of Smad3, on NADPH oxidase activity and Nox4, α-SMA, and Fn-ED-A protein expression. In time-course studies, TGF-β1 induced a rapid and progressive increase in p-Smad3 from 5 min to 1 h (Figure 5A). SB abolished p-Smad3 (Figure 5B), verifying that myofibroblast activation and matrix expression is TGF-β1 dependent and involves downstream Smad3 activation; however, DPI had no effect on p-Smad3 expression (Figure 5C), indicating that ROS may occupy a regulatory role for myofibroblast activation below Smad3. SIS3 reduced TGF-β1–induced increases in NADPH oxidase activity and Nox4 protein expression to basal levels (Figure 5, D and E), indicating that Smad3 regulates ROS generation. In addition, SIS3 abolished TGF-β1–induced increases in α-SMA and Fn-ED-A expression (Figure 5, F and G). These findings indicate that Smad3 regulates TGF-β1–stimulated NADPH oxidase and myofibroblast activation.

Figure 5.

TGF-β1–induced myofibroblast activation involves Smad3 regulation of NAD(P)H oxidase. To assess the NAD(P)H oxidase–derived ROS relative to p-Smad3, we treated kidney fibroblasts with SB, DPI, or an inhibitor of Smad3 (SIS3) before stimulation with TGF-β1. Smad3 phosphorylation and expression levels of α-SMA, Fn-ED-A, Nox4, and NADPH oxidase activity were examined. (A) TGF-β1 activates pSmad3 by 5 min progressively increasing over 1 h. (B and C) TGF-β1 activation of p-Smad3 is inhibited by SB (B); however, DPI had no effect on TGF-β1–induced Smad3 phosphorylation (C). (D and E) Inhibition of p-Smad3 by SIS3 blocked TGF-β1–induced stimulation of NADPH oxidase activity (D) and Nox4 (E), indicating that NAD(P)H oxidase is downstream of Smad3. (F and G) Similarly, SIS3 inhibited TGF-β1–induced stimulation of α-SMA (F) and Fn-ED-A (G) expression. Data are means ± SEM from three independent experiments. *P < 0.05 versus control; #P < 0.05 versus TGF-β1.

To determine a role for ERK1/2 in TGF-β1–induced ROS signaling and kidney myofibroblast activation, we conducted studies using the inhibitors to TGF receptor, Smad3, and ROS generation on ERK1/2 phosphorylation. In time-course studies, TGF-β1 induced a rapid spike in pERK1/2 at 5 min, which returned to basal levels at later time points up to 24 h after exposure to the growth factor (Figure 6A). UO126, an inhibitor of MEK and upstream of ERK1/2, inhibited TGF-β1–induced phosphorylation of this signaling protein (Figure 6B) as well as reduced the expression of α-SMA and Fn-ED-A (Figure 6C). To examine the function of pERK1/2, Smad3, and ROS, we treated the cells with SB, SIS3, or DPI, then stimulated them with TGF-β1 and assessed pERK1/2 protein expression. The results showed that all three inhibitors significantly reduced TGF-β1–induced pERK1/2 expression (Figure 6, D through F). UO126 did not affect TGF-β1–induced stimulation of NADPH oxidase activity (data not shown). Collectively, these results support a cascade of signaling events involving Smad3, ROS, and ERK1/2 in TGF-β1 induction of myofibroblast activation and matrix synthesis, as depicted in Figure 7.

Figure 6.

ERK1/2 has a central role in Smad3 and NAD(P)H oxidase signaling of TGF-β1–induced myofibroblast activation. (A) TGF-β1 induced a rapid expression of p-ERK1/2 at 5 min, disappearing thereafter, and was absent up to 24 h after stimulation by growth factor; therefore, all subsequent studies were performed at 5 min after stimulation with TGF-β1. (B and C) UO126, an inhibitor of MEK, completely eliminates TGF-β1–induced ERK1/2 phosphorylation (B) and substantially reduces α-SMA and Fn-ED-A protein expression (C). (D) In addition, inhibition of TGFBR with SB blocks p-ERK expression, indicating that TGF-β1–induced myofibroblast activation is partially regulated through this signal transduction molecule. (E) Similarly, DPI inhibits p-ERK, indicating that this signal protein is downstream of TGFBR and ROS. (F) SIS3 inhibition of Smad3 also inhibited p-ERK1/2 expression, supporting a TGFBR/Smad3/ROS/ERK1/2 signaling cascade in TGF-β1 stimulation of kidney myofibroblast activation and matrix synthesis (see Figure 7). Data are means ± SEM from three independent experiments. *P < 0.05 versus control; #P < 0.05 versus TGF-β1; §P < 0.05 versus TGF-β1+inhibitor.

Figure 7.

Proposed signaling cascade involving Smad3, ROS, and ERK1/2 in TGF-β1–induced kidney myofibroblast activation and matrix synthesis is shown.

Discussion

This study shows that kidney myofibroblast activation, characterized by acquisition of an α-SMA phenotype and increased expression of Fn-ED-A, is regulated through NAD(P)H oxidase–generated ROS. By using chemical inhibitors of key signaling proteins, we demonstrated that TGF-β1–induced myofibroblast activation works through a TGF receptor 1 (TGFR1)/Smad/ROS/ERK1/2 signaling cascade (Figure 7). TGF-β1 enhances the expression of Nox2 and Nox4 homologs in the NAD(P)H oxidase family. In addition, Nox4 seems to be the primary oxidase in kidney myofibroblast activation because of its robust expression level and ability of siRNA to reduce TGF-β1–induced increases in NADPH oxidase activity, α-SMA, and Fn-ED-A expression.

Roles for NAD(P)H oxidase–generated ROS in cell function are not entirely clear and may be regulated by multiple factors depending on the cell type and variant of Nox homolog(s) present in the cell.²⁹ Similarly, there seems to be considerable heterogeneity in the expression of Nox homologs in different cell types.^29,30 For example, Nox2 contributes to NAD(P)H oxidase activity in aortic fibroblasts and smooth muscle cells from resistance vessels but is minimally detectable in smooth muscle cells from conduit vessels.^29,49 Nox1 is primarily expressed in smooth muscle cells, whereas Nox4 seems to be present in all cells of the vascular wall, including adventitial fibroblasts. Moreover, a lung fibroblast line expresses Nox4 but not Nox2.⁵⁰ Cardiac fibroblasts also express Nox4.⁵¹ Aortic fibroblasts express Nox1, Nox2, and Nox4 homologs.³⁶

The significance of the heterogeneity of Nox homologs in different cell types is not known and may represent important differences in cellular function. NAD(P)H oxidases have been linked to a variety of cellular functions, including proliferation, migration, hypertrophy, regulation of metalloproteinases, matrix synthesis, cell senescence, and contraction.^{28,29,36,52–54} Indeed, differential roles of Nox1 and Nox4 in vascular smooth muscle cells have been correlated with differential compartmentalization in specific signaling domains in the membrane and focal adhesions.⁵⁵ Such observations combined with the differential expression of multiple Nox proteins in vascular smooth muscle in vitro and during restenosis⁵⁶ suggest distinct functions for these proteins.

This study is the first report to describe expression of both Nox4 and Nox2 homologs in kidney fibroblasts. Moreover, it is important to note that both homologs are upregulated by TGF-β1 in these cells, implicating an important role for agonist-induced NAD(P)H oxidase–derived ROS generation in kidney myofibroblast activation. These studies also agree with previous studies indicating a close relationship of NAD(P)H oxidase–generated ROS and α-SMA expression and a myofibroblast phenotype in adventitial fibroblast and vascular smooth muscle cells.^40–42 Our results show that expression levels of Nox4 in kidney fibroblasts greatly exceed Nox2 and that siRNA inhibition of Nox4 substantially inhibits α-SMA and Fn-ED-A expression, providing evidence that this homolog plays a pivotal role in promoting kidney myofibroblast activation and expression of the extracellular matrix protein Fn-ED-A.

It is well established that ROS can function as classic second-messenger molecules.^57,58 The realization that ROS have regulatory roles in signal transduction is based on the observations that exogenous oxidants (H₂O₂) activate growth factor–mediated signaling pathways.^59–62 Also, a number of growth factors elicit ROS and activate MAPK, ERK1/2, phosphatidylinositol-3-kinase/Akt, p38-MAPK, and c-Jun N-terminal kinase signaling pathways in a wide variety of cell types.^{28,40,47,48,59,62–65} Conversely, inhibitors of ROS block signaling and cell function, linking ROS with downstream pathways. A role for ROS in activation of TGF-β signaling is less understood. TGF-β binds to a receptor on the cell surface, forming a complex of subunits known as TGFR1 and TGFR2. Both TGFR1 and TGFR2 activate serine/threonine kinases that subsequently signal through the Smad family of transcriptional activators.^44,45 In mesangial cells and fibroblasts, TGF-β/Smad signaling (Smad 2/3) is tightly controlled by MAPK (i.e., ras/MEK/ERK) signaling cascades.⁴⁶ Studies showed that TGF-β1 stimulates not only a Smad pathway but also a parallel H₂O₂-mediated ERK pathway⁶¹ in human keratinocytes. Also, TGF-β1 stimulates production of ROS and proliferation of human pulmonary arterial smooth muscle cells⁶⁶ as well as conversion of cardiac fibroblasts to myofibroblasts,⁴⁰ both effects regulated by Nox4. In cardiac fibroblasts, depletion of Nox4 was shown to inhibit baseline levels as well as TGF-β1 stimulation of Smad 2/3 phosphorylation.⁴⁰ Our studies also indicate that NAD(P)H oxidase plays a key role in TGF-β1 activation of kidney myofibroblast and Fn-ED-A expression through Smad3 and ERK signaling pathways; however, NAD(P)H oxidase–generated ROS appears downstream of Smad3 and upstream of ERK1/2 in activated kidney fibroblast, unlike cardiac fibroblast, where Nox4 seems to regulate Smad2/3. The molecular mechanism of TGF-β1 stimulation of Smad3, NAD(P)H oxidase, and ERK1/2 in these experiments has not been defined. Stimulation of Smad3, Nox4, and ERK1/2 began as early as 5 min after addition of TGF-β1. Such a rapid phenomenon could be interpreted as a translational regulation of early signaling events, perhaps representing a novel pathway in early response to injury.

In vivo, the roles that inflammatory, tubular, and vascular cells play in renal fibrosis are complex, but they are believed to involve an induction phase characterized by the release of profibrogenic cytokines, growth factors, and ROS followed by activation of fibroblasts to myofibroblasts and extracellular matrix synthesis.^1,2,4 These in vitro studies indicate that ROS play an important role involving Smad3, ERK1/2, and NAD(P)H oxidase in kidney myofibroblast activation by TGF-β1. A similar role for such a pathway in kidney myofibroblast activation during renal fibrosis in vivo has not been determined; however, the association of TGF-β1 with myofibroblast activation in the pathogenesis of renal interstitial fibrosis is well documented.^1–3 Smad3 mediates TGF-β–induced fibronectin in mesangial cells, and the Smad pathway is activated in the diabetic mouse kidney.⁶⁷ Moreover, mice lacking Smad3 are protected against streptozotocin-induced diabetic glomerulopathy.⁶⁸ Mesangial cells are resident pericyte-like cells located in the renal glomerulus and, like renal fibroblast, acquire α-SMA and Fn-ED-A during renal disease.¹⁶ Also, NAD(P)H oxidase–generated ROS is important in mesangial cell hypertrophy and matrix synthesis during diabetic nephropathy²⁵ and signals through ERK1/2.⁶⁵ Furthermore, interstitial fibroblasts express ERK1/2 in kidney fibrosis after unilateral ureteral obstruction that can be ameliorated by inhibitors of ERK1/2.⁶⁹ Kidney myofibroblast during renal fibrosis originate, at least in part, from perivascular adventitial cells.^17,37–39 This observation is consistent with the known role of NAD(P)H oxidase–generated ROS in adventitial fibroblast activation in aortic perivascular disease.^19,36 Indeed, antioxidant therapy was shown to ameliorate renal disease in several models^21–25; however, a specific role for NAD(P)H oxidase–derived ROS in kidney fibroblast activation during renal interstitial fibrosis is plausible but awaits further investigation.

Concise Methods

Kidney Fibroblast Cell Culture

Rat kidney fibroblasts (NRK-49F) were obtained from American Type Culture Collection (Rockville, MD). The cells are a fibroblastic clone of NRK (mixed culture of normal rat kidney) cells commonly used for in vitro assay of profibrotic cellular events. In addition, a primary kidney fibroblast cell line established in our laboratory was used in a few key experiments to verify TGF-β1 induction of NADPH oxidase and α-SMA and Fn-ED-A expression. For these studies, cells were grown out of diced (<0.5 mm) tissue collected from the outer stripe of the outer medulla of rats with interstitial fibrosis 2 wk after Habu venom plus angiotensin II infusion.¹⁷ The cells were grown in RPMI medium supplemented with 10% FCS and propagated using standard cell culture methods.

Regulation studies were performed by growing the cells to 80% confluence and rendering them quiescent by replacing the complete medium with serum-free medium for 24 h, and then incubating them with incremental concentrations of TGF-β1 from 0.1 to 10.0 ng/ml. Diluent was used as a control in all experiments. Twenty-four hours later, the cells were examined for expression of two proteins selectively induced by TGF-β1: The cytoskeletal protein α-SMA⁹ and the matrix protein Fn-ED-A⁶ as outcome measurements for myofibroblast activation. Concentration of TGF-β1 (1 ng/ml) was determined to be optimal and used for subsequent studies to examine the effect of this substance on signaling through TGF-β1 receptor, ROS, and Smad3, as well as ERK1/2, using immunofluorescence microscopy and Western blot analysis (see the Immunofluorescence Microscopy section). Inhibitors of TGF-β1 receptor, NAD(P)H oxidase, Smad3, and ERK1/2 were used to interrupt signaling. All experiments were repeated at least three times.

Detection of Endogenous ROS Generation

Kidney fibroblasts were cultured and made quiescent as already described. For examination of whether TGF-β1 induces endogenous ROS, cells were stimulated with TGF-β1, and the peroxide-sensitive fluorescence probe 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate, acetyl ester (CM-H₂DCFDA; Molecular Probes, Eugene, OR) was used to assess the generation of intracellular ROS as described previously.^25,65 Images were obtained at 15 and 30 min after addition of TGF-β1 using an Olympus inverted microscope with a 20A planfluor objective, 2× zoom, and an Olympus FluoView confocal laser scanning attachment. 2′,7′dichlorodihydrofluorescein fluorescence was detected with an excitation wavelength of 488 nm of light, and its emission was detected using a 510- to 550-nm band-pass filter.

NADPH Oxidase Assay

NADPH oxidase activity was measured by the lucigenin-enhanced chemiluminescence method previously described.^25,64 Kidney fibroblasts were stimulated with TGF-β1 or diluent, as already described, and at incremental time points from 30 min to 4 h. The cells were washed and homogenized in lysis buffer containing 20 mM KH₂PO₄ (pH 7.0), 1 mM EGTA, 1 mM PMSF, 10 μg/ml aprotinin, and 0.5 μg/ml leupeptin. A total of 100 μl of homogenate was immediately added to 900 μl of 50 mM phosphate buffer (pH 7.0) containing 1 mM EGTA, 150 mM sucrose, 5 μM lucigenin as the electron acceptor, and 100 μM NADPH as an electron donor. Photon emission expressed as relative light units was measured every 30 s for 5 min in a luminometer. There was no measurable activity in the absence of NADPH. A buffer blank (<5% of the cell signal) was subtracted from each reading before calculation of the data. Superoxide production was expressed as the rate of relative chemiluminescence (light) units per minute per milligram of protein. Protein content was measured using the Bio-Rad protein assay reagent. NADPH oxidase activity experiments were also repeated three times, each showing the same trends. Data are illustrated in representative graphs.

Expression and Regulation of NAD(P)H Oxidase Homologs

To date, seven Nox homologs have been described.³⁰ Of these, Nox1 and Nox2 are present in the vasculature; Nox 3 is present in fetal kidney; Nox4 is present in the fetal and adult kidney; and Nox 5 is present in extrarenal organs including the spleen, mammary glands, and cerebrum. Two additional homologs, duox 1 and 2, have dual oxidase and peroxidase activities and are present in extrarenal organs including thyroid, cerebellum, lungs, pancreas, and prostate.³⁰ The profile of NAD(P)H oxidase homologs in kidney fibroblasts has not been determined; therefore, kidney fibroblasts were cultured in serum-free conditions, then incubated with TGF-β1 or diluent as already described. Nox1, Nox2, Nox3, and Nox4 protein expression was examined by Western analysis and immunofluorescence histochemistry. Nox5 or the duox homologs were not examined because of the tissue specificity in organs other than the kidney.

siRNA Inhibition of Nox4

We observed that TGF-β1–induced Nox4 protein expression markedly exceeded that of Nox2 (see the Results section), suggesting that Nox4 is the predominant NAD(P)H oxidase homolog involved in signaling of myofibroblast activation. Studies were therefore conducted to examine whether myofibroblast activation and expression of α-SMA and Fn-ED-A was Nox4 dependent using siRNA. Kidney fibroblasts were transfected with SiControl Nontargeting siRNA pool or siGenome Smart pool rat Nox4 designed by Dharmacon Research (Lafayette, CO) using X-tremeGene siRNA transfection reagent (Roche Diagnostics Corp., Indianapolis, IN) with minor modifications to the manufacturers' instructions. Briefly, kidney fibroblasts were grown to 80% confluence in Costar six-well cell culture plates (Corning, Corning, NY), and fresh medium was provided 30 min before transfection. The medium was then replaced with serum/antibiotic/antimycotic-free RPMI containing nontargeting and Nox4 siRNAs at a final concentration 400 nM. After 6 h of exposure, the medium was replaced with RPMI 1640 containing serum without antibiotic or antimycotic, and the cells were allowed to grow for an additional 48 h. The transfected cells were then made quiescent as already described, and TGF-β1 was added achieving a final concentration of 1 ng/ml. For verification of siNox4 knockdown of NAD(P)H oxidase, the transfected cells were examined for Nox4 protein expression and TGF-β1–induced stimulation of NADPH oxidase activity as outlined already. Transfected cells were also exposed to TGF-β1 for 24 h, then lysed in RIPA buffer containing 1% NP-40 and frozen for subsequent Western analysis for α-SMA and Fn-ED-A.

Western Analysis

Immunoblotting was performed as described previously.¹⁷ At termination of each experiment, the kidney fibroblasts were lysed in 0.5 ml of RIPA buffer (50 mM Tris-HCl [pH 7.5], 1 mM EGTA, 140 mM NaCl, and 1.0% NP-40) containing 1 μg/ml leupeptin and aprotinin, 50 mM sodium fluoride, 1 mM sodium orthovanadate, and 1.0 mM PMSF. Insoluble proteins were removed by high-speed centrifugation at 10,000 × g, then protein concentrations were determined using the Bio-Rad DC protein assay (Bio-Rad Laboratories, Hercules, CA). Protein lysates were boiled in sample buffer for 10 min, then equal amounts of sample were loaded onto 7.5% SDS-PAGE gels and electrophoretically separated. The proteins were transferred to polyvinylidene difluoride membranes using a Bio-Rad Trans-Blot cell followed by blocking with 5% nonfat dry milk in PBS containing 0.1% Tween 20 and incubated overnight in primary antibody diluted in ECL advance blocking agent (Amersham Pharmacia Biotech, Piscataway, NJ). The antigens are detected and identified by enhanced chemiluminescence using standard ECL techniques as recommended by the manufacturer (Amersham). Immunoblotting was used for Nox homologue identification (as already described) and analysis of Fn-ED-A expression levels. Antibodies for immunoblotting included anti-Nox1 (Mox1; Santa Cruz Biotechnology, Santa Cruz, CA) Nox2 (Upstate Biotechnology, Lake Placid, NY), and Nox3 (Santa Cruz Biotechnology). In addition, antibodies to Nox4 were produced in our laboratory by immunization of New Zealand white rabbits with a recombinant GST-Nox4 fragment (peptides 299 through 515) representing the N-terminus of the protein²⁵ that works well in Western blotting and immunohistochemistry. In addition, the following antibodies were used for Western blotting: α-SMA (Sigma Chemical Co., St. Louis, MO), Fn-ED-A (Abcam, Cambridge, MA), pERK, ERK1/2, and pSmad3 (Cell Signaling, Beverly, MA). Western blot data were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH; anti-GAPDH; Novus, Littleton, CO) and represented by column graphs as average intensity percentage of TGF-β1.

Immunofluorescence Microscopy

Kidney fibroblasts were grown in multiwell plastic Lab-Tek chamber microscope slides (Nalge Nunc Int., Naperville, IL), incubated with TGF-β1, and examined for expression of Nox homologs and α-SMA and Fn-ED-A protein. After addition of experimental conditions, the cells were washed in PBS and fixed in cold −20°C methanol for 5 min, then briefly rinsed with 0.02 M PBS (pH 7.4). The slides were blocked with PBS containing 0.1% BSA, and then the specific protein of interest was detected by indirect immunofluorescence using primary antibodies described already followed by a FITC-labeled mouse anti-mouse or rabbit IgG second antibody (Millipore/Chemicon, Billerica, MA). The stained cells were washed with PBS containing 0.1% BSA, mounted, and viewed by epifluorescence and photographed using an Olympus AX70 Research microscope equipped with a DP-70 digital camera (Melville, NY).

ROS Signaling in Kidney Myofibroblast Activation

Studies were conducted to examine the relative position of NAD(P)H oxidase–derived ROS in TGF-β1 signaling of kidney myofibroblast activation (α-SMA and Fn-ED-A expression). To begin, expression levels of proteins known to have major roles in TGF-β1 signal transduction including Smad3 and ERK were examined in time-course studies by Western analysis. Quiescent cells were treated with TGF-β1 and incubated at multiple time points from 5 min to 24 h. The cells were lysed, and phosphorylated forms of Smad3 and ERK1/2 proteins were examined by Western blot analysis. ERK1/2 or GAPDH proteins served as loading controls. For elucidation of the relative expression of NAD(P)H oxidase–derived ROS in myofibroblast signaling, cells were preincubated with inhibitors of TGFR1 (SB431542, 10 μg/ml, 30 min; Sigma), NAD(P)H oxidase (DPI, 10 μM, 30 min; Sigma), Smad3 (SIS3, 5 μM,1 h; Calbiochem, Gibbstown, NJ), and ERK1/2 (UO126, 10 μM, 2 h; Cell Signaling Technologies, Danvers, MA), then incubated with TGF-β1. NADPH oxidase activity, pSmad3, and p-ERK1/2 expression was examined by Western blotting. Kidney fibroblasts treated with diluent in place of TGF-β1 or inhibitor served as basal controls. In separate experiments, 24 h after addition of inhibitors and TGF-β1, the cells were examined for expression of α-SMA and Fn-ED-A protein levels.

Statistical Analysis

Experiments were analyzed by one-way ANOVA, followed by Newman-Keuls multiple comparison post hoc test performed by GraphPad Prism 5 software (GraphPad, San Diego, CA). Significance was achieved at P < 0.05.

Disclosures

None.