Matrix accumulation in the renal tubulointerstitium is predictive of a progressive decline in renal function. Transforming growth factor-β1 (TGF-β1) and, more recently, connective tissue growth factor (CTGF) are recognized to play key roles in mediating the fibrogenic response, independently of the primary renal insult. Further definition of the independent and interrelated effects of CTGF and TGF-β1 is critical for the development of effective antifibrotic strategies. CTGF (20 ng/ml) induced fibronectin and collagen IV secretion in primary cultures of human proximal tubule cells (PTC) and cortical fibroblasts (CF) compared with control values (P < 0.005 in all cases). This effect was inhibited by neutralizing antibodies to either TGF-β or to the TGF-β type II receptor (TβRII). TGF-β1 induced a greater increase in fibronectin and collagen IV secretion in both PTC (P < 0.01) and CF (P < 0.01) compared with that observed with CTGF alone. The combination of TGF-β1 and CTGF was additive in their effects on both PTC and CF fibronectin and collagen IV secretion. TGF-β1 (2 ng/ml) stimulated CTGF mRNA expression within 30 min, which was sustained for up to 24 h, with a consequent increase in CTGF protein (P < 0.05), whereas CTGF had no effect on TGF-β1 mRNA or protein expression. TGF-β1 (2 ng/ml) induced phosphorylated (p)Smad-2 within 15 min, which was sustained for up to 24 h. CTGF had a delayed effect on increasing pSmad-2 expression, which was evident at 24 h. In conclusion, this study has demonstrated the key dependence of the fibrogenic actions of CTGF on TGF-β. It has further uniquely demonstrated that CTGF requires TGF-β, signaling through the TβRII in both PTCs and CFs, to exert its fibrogenic response in this in vitro model.
- proximal tubule cells
- cortical fibroblasts
- collagen IV
tubulointerstitial pathology, characterized by expanded extracellular matrix (ECM), tubular atrophy, and a variable inflammatory response, is the key structural hallmark of progressive renal disease independent of the primary pathology (3, 19). The cellular elements primarily involved in the pathological changes in human renal fibrosis are the cells that predominate in the cortical tubulointerstitium, namely, proximal tubule cells (PTC) and cortical fibroblasts (CF) (16, 19).
Transforming growth factor-β1 (TGF-β1) is the cytokine most causatively associated with disorders characterized by inflammation and fibrosis. It is clear that both PTC and CF produce a significant amount of TGF-β1 under basal conditions (10, 12). This significant amount of TGF-β1 is then amplified in disease states and correlates with clinical and histological markers of progressive pathology (2). Three different TGF-β receptors have been identified. However, it is recognized that the TGF-β type II receptor (TβRII) plays a key role in TGF-β binding and activation, leading to signal transduction mediated by a number of TGF-β receptor-responsive proteins (4). TGF-β is known to mediate its fibrotic effects by activating the receptor-associated Smads (Smad-2 and -3). Phosphorylated Smad-2 (pSmad-2) and Smad-3 associate to form a heteromultimer with Smad-4 (Co-Smad). This complex is then translocated to the nucleus, where it can regulate target gene expression (17).
Emerging evidence suggests that connective tissue growth factor (CTGF) may be an important downstream mediator of the profibrotic effects of TGF-β. TGF-β- and Smad-responsive elements in the CTGF promoter have been identified (7, 9), and it has been recently demonstrated that CTGF may mediate the altered α5β1-integrin expression induced by TGF-β1 (20). A receptor specific to CTGF has to date not been identified, although CTGF has been shown to bind to low-density lipoprotein receptor-related protein and integrins (18). Furthermore, the specific molecular mechanisms by which CTGF exerts its profibrotic effects in the human renal interstitium remain to be elucidated. A recent report by Abreu et al. (1) showed that CTGF may function as a chaperone to modify the conformation or solubility of TGF-β1 and thus facilitate TGF-β1 binding to the TβRII. However, the mechanisms have not been defined. Given that CTGF may have effects both independent from and dependent on TGF-β1, it is of prime importance to delineate the independent and convergent roles of CTGF and TGF-β1 in renal fibrosis and progressive renal disease to identify novel therapeutic targets.
Fibronectin and collagen IV are major ECM proteins that serve as a scaffold for the deposition of other proteins. Furthermore, fibronectin functions as a fibroblast chemoattractant and promotes their differentiation, which may be a crucial phenomenon in the pathogenesis of tubulointerstitial fibrosis (21).
The present study was therefore performed to determine the extent to which the effects of CTGF on renal fibrogenesis in vitro are dependent on TGF-β and TGF-β receptor signaling.
MATERIALS AND METHODS
Human PTC and CF cell culture.
Segments of macroscopically and histologically normal renal cortex were obtained under aseptic conditions from patients undergoing nephrectomy for small (<6 cm) tumors (11). Patients were accepted for inclusion in the study if there was no history of renal or systemic disease known to be associated with tubulointerstitial pathology, which was confirmed by subsequent histopathology. Written informed consent was obtained from each patient before surgery, and ethics approval for the study was obtained from the Royal North Shore Hospital Human Research Ethics Committee. The methods for isolation of primary culture of both PTC and CF are described in detail elsewhere (10, 11). In brief, the kidney cortex was dissected from the medulla, finely minced, and then digested by collagen (383 U/mg, Worthington) for 30 min at 37°C and passed through 100-μm mesh. The filtrate was resuspended in 50 ml of 45% Percoll (Pharmacia, Uppsala, Sweden) and centrifuged at 20,000 rpm at 4°C for 30 min. The bottom-most tissue band containing highly purified PTC and the top band containing CF were carefully removed and washed. The PTC fragment pellet was resuspended in serum-free hormonally defined media consisting of 1:1 (vol/vol) DMEM and Hams F-12 (DMEM/F-12; Trace) supplemented with 10 ng/ml (1.64 nM) epidermal growth factor, 5 mg/ml human transferrin, 5 mg/ml (0.87 mM) bovine insulin, 0.05 mM hydrocortisone, 50 mM prostaglandin E1, 50 nM selenium, and 5 pM triiodothyronine (all from Sigma, St. Louis, MO). The CF pellet was resuspended in DMEM/F-12 containing 10% fetal calf serum (Invitrogen). Passage 2 cells were used for all experiments. The ultrastructure, growth, and immunohistochemistry of both PTC and CF have been well characterized in our laboratory and shown to reproducibly reflect the biology and physiology of their in vivo counterparts (10, 11).
TGF-β1 was purchased from Sigma, and rhCTGF was a generous gift from FibroGen. TGF-β-neutralizing antibody, TβIIR-neutralizing antibody, negative control rabbit IgG, and goat IgG were all purchased from R&D Systems (Minneapolis, MN). Mouse monoclonal fibronectin, rabbit polyclonal collagen IV, goat polyclonal β-actin, and pSmad-2 antibodies were from Neomarkers, Abcam (Cambridge, UK), Santa Cruz Biotechnology (Santa Cruz, CA), and Cell Signaling, respectively. The CTGF polyclonal antibody used was raised in a New Zealand White rabbit against full-length purified recombinant human CTGF protein as previously described (14).
The cell culture supernatant and cell lysate were collected after PTC and CF were treated with CTGF for 48 h. The supernatant was centrifuged at 4°C for 5 min to eliminate cell debris and stored at −80°C. Cells were lysed in cell lysis buffer (Cell Signaling) with 1 mM PMSF, and samples were centrifuged at 4°C for 5 min. The supernatant was removed and stored at −80°C. TGF-β1 was measured using an ELISA (Promega) according to the manufacturer's instructions. Results were corrected for cell protein and expressed as nanograms per milligram protein. Data shown are standardized to control values, represented as 100%.
Cells were grown to ∼80% confluence and then made quiescent by incubation for 24 h in basic media (DMEM/F-12 containing 5 ng/ml human transferrin and 5 mM glucose). Subsequently, the cells were exposed to the various experimental conditions described below for a further 48 h, with all experiments being performed in medium containing 5 mM glucose plus 0.01% BSA. Unless otherwise stated, in all experiments the concentration of added reagents were as follows: 20 ng/ml CTGF and 2 ng/ml TGF-β1.
After achieving subconfluence, cells were exposed to CTGF, TGF-β1, or a combination of CTGF and TGF-β1 for 48 h. In the combination experiments, cells were exposed to 0.5 ng/ml TGF-β1 to induce a submaximal response. Fibronectin and collagen IV, measured in the supernatant by Western blot analysis, were used as markers of the profibrotic response. A neutralizing antibody strategy was used to determine the extent to which TGF-β mediated the profibrotic effects of CTGF. Cells were exposed for 48 h in separate wells to CTGF, CTGF plus TGF-β-neutralizing antibody (5 μg/ml, preincubated for 30 min before the addition of CTGF), CTGF plus normal rabbit IgG (5 μg/ml, also preincubated for 30 min before the addition of CTGF), or TGF-β-neutralizing antibody (5 μg/ml) alone, and fibronectin and collagen IV were measured in the supernatant. A similar neutralizing antibody strategy was used to determine whether the effects of CTGF were mediated via the TβRII. Cells were exposed to CTGF, CTGF plus TβRII-neutralizing antibody (20 μg/ml, preincubated for 30 min before the addition of CTGF), CTGF plus normal goat IgG (20 μg/ml, also also preincubated for 30 min before the addition of CTGF), or TβRII-neutralizing antibody (20 μg/ml) alone, and fibronectin and collagen IV were measured. A time course of the TGF-β blockade in response to CTGF using TGF-β-neutralizing antibody (5 μg/ml) was performed. Cells were exposed for 48 h to CTGF, CTGF plus TGF-β-neutralizing antibody (preincubated for 30 min before the addition of CTGF), CTGF plus TGF-β-neutralizing antibody (TGF-β-neutralizing antibody was added after cells were treated with CTGF for 6 and 24 h, respectively). Conditioned media were then collected for the measurement of fibronectin and collagen IV.
To detect changes in the levels of protein for fibronectin, collagen IV, CTGF, β-actin, and pSmad-2, we used standard Western immunoblotting techniques. Western blot analysis was performed on Triton X-100-soluble fractions of cells for detection of CTGF, β-actin, and pSmad-2 and on conditioned media for fibronectin and collagen IV. Samples were subjected to SDS-PAGE under reducing conditions. Proteins were then transferred to Hybond ECL nitrocellulose membrane (Amersham Pharmacia Biotech). Nonspecific binding sites were blocked for 1 h (5% nonfat milk and 0.1% Tween 20 in PBS), after which the membranes were exposed to the respective primary antibodies overnight at 4°C, followed by being washed four times, after which they were incubated with peroxidase-labeled secondary antibodies (Amersham Pharmacia Biotech) for 1 h and again washed four times. The blots were then detected using ECL (Amersham Pharmacia Biotech). The bands corresponding to fibronectin (220 kDa), collagen IV (180 kDa), CTGF (38 kDa), β-actin (47 kDa), and pSmad-2 (58 kDa) were quantitated using National Institutes of Health Image software, version 1.60. Coomassie brilliant blue staining or β-actin was used to confirm that an equal amount of protein was loaded in each lane.
RT-PCR and real-time RT-PCR.
Semiquantitative RT-PCR and quantitative real-time PCR following reverse transcription were used to assess transcript levels of CTGF and TGF-β1. To ensure equal loading of cDNA samples, concurrent PCR reactions for β-actin (in RT-PCR) and 18S (in real-time PCR) were performed. Both water blank and non-reverse transcribed RNA samples were used as negative controls. The number of amplification cycles in the semiquantitative PCR was determined from the linear portion of the PCR cycle, and amplification was performed with increasing numbers of cycles for CTGF, TGFβ1, and β-actin. Briefly, total RNA was extracted using an RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. Total RNA (2 μg) was treated with DNase I (Invitrogen), and then cDNA was synthesed using reverse transcriptase Superscript II RT (Invitrogen). In RT-PCR, sequence-specific primers for human CTGF (GenBank accession no. NM_001901), TGF-β1 (accession no. NM_000660), and β-actin (accession no. NM_001101) were synthesed by Sigma Genosys (Sydney, Australia). CTGF primers were forward 5′-CGAGCTAAATTCTGTGGAGT-3′ and reverse 5′-CCATGTCTCCGTACATCTTC-3′; TGF-β1 primers were forward 5′-CCATGTCTCCGTACATCTTC-3′ and reverse 5′-CGCCCGGGTTATGCTGGTTGT-3′; and β-actin primers were forward 5′-GCTCGTCGTCGACAACGGCTC-3′ and reverse 5′-CAAACATGATCTGGGTCATCTTCTC-3′. The sizes of the PCR products for CTGF, TGF-β1, and β-actin were expected as 208, 288, and 353 bp, respectively. Amplification products were electrophoresed through 1.5% (wt/vol) agarose gels and visualized by ethidium bromide staining. To further confirm the RT-PCR data, real-time PCR was used to measure CTGF and TGF-β1 mRNA. Specific primers for the use of SYBR green were designed as follows: CTGF, forward 5′-GGCTTACCGACTGGAAGAC-3′ and reverse 5′-AGGAGGCGTTGTCATTGG-3′; TGF-β1, forward 5′-GCAACAATTCCTGGCGATACC-3′ and reverse 5′-CTCCACGGCTCAACCACTG-3′; 18S served as an internal control: forward 5′-CGGCTACCACATCCAAGGAA-3′ and reverse 5′-GCTGGAATTACCGCGGCT-3′. Primer specificity in real-time PCR reactions was confirmed using RT-PCR. Twenty-five microliters of the real-time PCR reaction included Brilliant SYBR Green QRT-PCR Master Mix as per the manufacturer's instructions (Stratagene). Real-time quantitations were performed with the Bio-Rad iCycler iQ system (Bio-Rad). The fluorescence threshold value was calculated using the iCycle iQ system software. The calculation of relative change in mRNA was performed using the delta-delta method (15), with corrections for the housekeeping gene 18S.
All results are expressed as a percentage of the control value (100%) with the exception of real-time PCR results, which are expressed as a fold-change compared with the control value. Experiments were performed in at least three different culture preparations, and at least three data points for each experimental condition were measured in each preparation. Results are expressed as means ± SE, with n reflecting the number of culture preparations. Statistical comparisons between groups were made by ANOVA, with pairwise multiple comparisons made by Fisher's protected least-significant difference test. Analyses were performed using the software package Statview, version 4.5 (Abacus Concepts, Berkeley, CA). P values <0.05 were considered significant.
Pan-specific TGF-β- and TβRII-neutralizing antibody effect on CTGF-induced fibronectin and collagen IV production.
When recombinant human (rh)CTGF was added to PTC, it induced fibronectin and collagen IV protein secretion to 148 ± 8 (P < 0.005) and 140 ± 10% (P < 0.005) of control, respectively (Fig. 1, A and B). In CF, rhCTGF induced fibronectin and collagen IV protein secretion to 144 ± 16 (P < 0.005) and 169 ± 12% (P < 0.005) of control, respectively (Fig. 1, C and D). The addition of TGF-β-neutralizing antibody prevented the CTGF-induced fibronectin and collagen IV secretion in both PTC and CF. Rabbit IgG and TGF-β-neutralizing antibody alone had no effect on either fibronectin or collagen IV secretion.
The CTGF-induced fibronectin and collagen IV protein secretion in both PTC and CF was blocked by TGF-β-neutralizing antibody when added to the cells 30 min before the addition of CTGF. The inhibitory effect by TGF-β-neutralizing antibody was not seen on the cells treated with CTGF for 6 or 24 h before the addition of TGF-β-neutralizing antibody (data not shown).
Similar to the effect seen with the TGF-β-neutralizing antibody, the addition of TβRII-neutralizing antibody blocked CTGF-induced fibronectin and collagen IV secretion to control levels in both PTC and CF, respectively (Fig. 2). Goat IgG and TβRII-neutralizing antibody alone had no effect on either fibronectin or collagen IV secretion.
Collectively, these data suggest that CTGF-induced fibronectin and collagen IV secretion requires endogenous TGF-β bioactivity and that the signaling of CTGF is mediated through the TβRII in both PTC and CF.
CTGF does not stimulate TGF-β1 expression in either PTC or CF.
As CTGF action was shown to be dependent on TGF-β signaling in both PTC and CF, we considered it possible that CTGF mediated its effects through stimulation of production of endogenous TGF-β1. To determine whether rhCTGF regulated gene expression of TGF-β1, steady-state mRNA levels of TGF-β1 were measured after exposure of PTC and CF to 20 ng/ml CTGF for 30 min, 6 h, and 24 h. This treatment did not cause a significant increase in TGF-β1 mRNA expression compared with the control values in either PTC or CF measured by either semiquantitative RT-PCR (Fig. 3, A and B) or real-time PCR (Fig. 3, C and D). To determine whether rhCTGF regulated TGF-β1 protein expression, TGF-β1 in both the conditioned media and cell layer was measured after exposure of PTC and CF to 20 ng/ml CTGF for 48 h. Treatment did not cause a significant increase in TGF-β1 protein expression compared with the control values (Fig. 3, E and F). The amount of total TGF-β1 secreted by PTC ranged from 2.5–3.3 ng/mg protein (in conditioned media) and 1.55–3.77 ng/mg protein (in the cell lysate) over a period of 24 h. In CF, the amount of total TGF-β1 secreted ranged from 1.85 to 2.4 ng/mg protein (in conditioned media) and 0.9 to 1.3 ng/mg protein (in the cell lysate) over a period of 24 h. Thus, although TGF-β1 is required by CTGF to induce fibronectin and collagen IV secretion, its expression is not stimulated by CTGF beyond the levels of basal production. This suggests that the known basal secretion of TGF-β by both PTC and CF is sufficient to facilitate the action of CTGF.
TGF-β1 stimulates CTGF expression.
In contrast to a lack of effect of rhCTGF on TGF-β1 gene expression, exposure of PTC and CF to 2 ng/ml TGF-β1 for 30 min, 6 h, and 24 h increased CTGF mRNA levels. This was observed at 30 min and was sustained at 24 h in both PTC and CF (P < 0.05) measured by semiquantitative RT-PCR (Fig. 4, A and B) or real-time PCR (Fig. 4, C and D). CTGF protein expression was measured by Western immunoblotting using CTGF antibody. In parallel, CTGF protein increased to 186 ± 21.3 and 136 ± 7.5 (P < 0.05), respectively, after PTC and CF were exposed to 2 ng/ml TGF-β1 for 48 h (Fig. 4, E and F).
Combined effect of CTGF and TGF-β1 on PTC and CF.
We further explored whether the combination of CTGF and TGF-β1 had an additive or synergistic effect on matrix production. Exposure of PTC to CTGF or TGF-β1 (0.5 ng/ml) increased fibronectin to 197 ± 31 (P < 0.05) and 468 ± 24% (P < 0.005), respectively, compared with control. Combined CTGF and TGF-β1 treatment further increased fibronectin by 752 ± 252%, a level significantly greater than that observed with TGF-β1 alone (P < 0.05) (Fig. 5A). In CF, exposure to CTGF or TGF-β increased fibronectin expression to 232 ± 18 (P < 0.0001) and 644 ± 135% (P < 0.005), respectively, compared with control. Combined CTGF and TGF-β1 further increased fibronectin by 356 ± 25%, again to a level significantly greater than that observed with TGF-β1 alone (P < 0.05) (Fig. 5C). Collagen IV secretion was measured in the above experimental conditions, and similar results as seen for fibronectin were observed in both PTC and CF. In PTC, combined CTGF and TGF-β1 treatment further increased collagen IV by 1,145 ± 99%, a level significantly greater than that observed with TGF-β1 alone (P < 0.05) (Fig. 5B). In CF, combined CTGF and TGF-β1 further increased collagen IV by 2,218 ± 163%, again to a level significantly greater than that observed with TGF-β1 alone (P < 0.05) (Fig. 5D). The observed additive effects of CTGF and TGF-β1 suggest that CTGF magnifies TGF-β-induced fibronectin and collagen IV secretion.
CTGF is unlikely to directly induce the pSmad-2.
PTC and CF were treated with TGF-β1 (2 ng/ml) or 20 ng/ml CTGF for 0, 0.25 (15 min), 0.5 (30 min), 2, 6, and 24 h. TGF-β1 induced pSmad-2 in both PTC (Fig. 6A) and CF (Fig. 6B) within 15 min. Although CTGF induced pSmad-2 after 24 h in both PTC (Fig. 6A) and CF (Fig. 6B), CTGF did not detectably induce pSmad-2 at earlier time points. These data could be explained by the basal secretion of TGF-β1 interacting progressively with CTGF to induce pSmad-2 over a delayed time course.
This study provides important information concerning the integrated actions of TGF-β and CTGF in inducing renal fibrosis. We have demonstrated, using in vitro models in human renal PTC and CF, that CTGF-induced fibronectin and collagen IV expression is dependent on TGF-β. Although CTGF is known to bind to low-density lipoprotein receptor-related protein and integrins (13, 18, 20), a receptor specific to CTGF and the mechanism whereby CTGF promotes a profibrotic response have to date not been elucidated. It has been suggested that because of the multimodular structure of CTGF, its effects may depend on interactions with other proteins and their associated signaling pathways. The present study has shown that the profibrotic effect of CTGF was completely abrogated in the presence of an antibody to TβRII, which prevents phosphorylation of the TGF-β receptor type I and thus inhibits signal transduction. We considered it possible that CTGF may signal through the TβRII. However, the effects of CTGF were also abrogated in the presence of a pan-specific TGF-β antibody, which prevents TGF-β binding to TβRII. This provides evidence that the key role of CTGF may be to facilitate the action of TGF-β1 at TβRII, rather than independently signal via TβRII. CTGF had a delayed effect on the induction of pSmad-2, suggesting that it facilitates the action of the endogenous TGF-β1 over the 24-h period of the study. This concept of CTGF facilitating TGF-β is consistent with that of Abreu et al. (1), who showed that CTGF may function as a chaperone to modify the conformation or solubility of TGF-β1 and thus facilitate TGF-β1 binding to TβRII.
It is well known that CTGF is an important downstream mediator in TGF-β-induced ECM in the kidney, but there are overlapping and distinct fibrogenic effects between TGF-β and CTGF in human renal cells (6). Our results also confirm previous observations that TGF-β1 increases CTGF expression (8) and further suggest that the increase in CTGF induced by TGF-β1 provides an autocrine mechanism to enhance the bioactivity of TGF-β1. The mechanism whereby TGF-β1 increases CTGF mRNA has not been defined by the current study. However, TGF-β- and Smad-responsive elements in the CTGF promoter have been identified (7, 9), and it has recently been demonstrated that CTGF may mediate the altered α5β1-integrin expression induced by TGF-β1 (20), which has been implicated in the development of fibrosis. Conversely, our data do not suggest that CTGF induces autocrine TGF-β1 production. Despite this lack of induction of TGF-β1, our study strongly suggests that CTGF cannot act independently of TGF-β to induce profibrotic effects. Given the absence of specific upregulation of TGF-β1 mRNA by CTGF, it is possible that the effects of CTGF are mediated through an interaction between an alternative isoform of TGF-β and TβRII.
That the maximal effective dose of CTGF on the cells (20 ng/ml) when added alone was less potent than the maximal dose of exogenously added (recombinant) TGF-β1 is consistent with the observation that added CTGF requires endogenous TGF-β1 to exert its effect. Under physiological conditions, only a small fraction of total TGF-β1 produced by the cells is in the active form. The level of total TGF-β1 produced by the cells under basal conditions is 2.52 ng/mg protein on average over a 24-h period (10), and <5% of this total amount is in the active form (5). In contrast, all of the added recombinant TGF-β1 used in the present studies is in the active form. Hence the studies in the current work showing the interdependence of CTGF and TGF-β1 indicate that CTGF has a greater potency in the presence of increased active TGF-β1 in the cell system and that the amount of active TGF-β1 present limits the maximal potency of CTGF added to these cells.
In summary, the integrated action of both TGF-β1 and CTGF is required to induce in vitro ECM in our model of human primary cell cultures, and by implication, in renal fibrosis. TGF-β induces autocrine production of CTGF, which then serves to amplify the effect of TGF-β1 in both human PTC and CF. It remains to be determined whether CTGF can be targeted in vivo to specifically prevent the profibrotic responses of TGF-β1, while leaving unaffected the antiproliferative and immunosuppressive effects induced by TGF-β1.
These studies were supported by the National Health and Medical Research Council of Australia and the Juvenile Diabetes Research Foundation. W. Qi was supported by a National Health and Medical Research Council Scholarship.
CTGF was a generous gift from FibroGen, Inc. (San Francisco, CA).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society