Background

Interstitial fibroblasts play a critical role in renal fibrogenesis, and autocrine proliferation of these cells may account for continuous matrix synthesis. Basic fibroblast growth factor (FGF-2) is mitogenic for most cells and exerts intracrine, autocrine, and paracrine effects on epithelial and mesenchymal cells. The aims of the present studies were to localize and quantitate the expression of FGF-2 in normal and pathologic human kidneys and to study the in vitro effects of FGF-2 on proliferation, differentiation, and matrix production of isolated cortical kidney fibroblasts.

Methods

FGF-2 protein expression was localized by immunofluoresence double labelings in normal and fibrotic human kidneys. Subsequently, interstitial FGF-2 labeling was determined semiquantitatively in 8 normal kidneys and 39 kidneys with variable degrees of interstitial fibrosis and was correlated with the morphometrically determined interstitial cortical volume. In addition, FGF-2 expression was quantitated by immunoblot analysis in three normal and six fibrotic kidneys. FGF-2 mRNA was localized by in situ hybridizations. Seven primary cortical fibroblast lines were established, and expression of FGF-2 and FGF receptor-1 (FGFR-1) were examined. The effects of FGF-2 on cell proliferation were determined by bromodeoxyuridine incorporation and cell counts, those on differentiation into myofibroblasts by staining for -smooth muscle actin, and those on matrix synthesis by enzyme-linked immunosorbent assay for collagen type I and fibronectin. Finally, proliferative activity in vivo was evaluated by expression of MIB-1 (Ki-67 antigen).

Results

In normal kidneys, FGF-2 expression was confined to glomerular, vascular, and a few tubular as well as interstitial fibroblast-like cells. The expression of FGF-2 protein was increased in human kidneys, with tubulointerstitial scarring correlating with the degree of interstitial fibrosis (r = 0.84, P < 0.01). Immunoblot analyses confirmed a significant increase in FGF-2 protein expression in kidneys with interstitial scarring. In situ hybridization studies demonstrated low-level detection of FGF-2 mRNA in normal kidneys. However, FGF-2 mRNA expression was robustly up-regulated in interstitial and tubular cells in end-stage kidneys, indicating that these cells are the source of excess FGF-2 protein. Primary cortical fibroblasts express FGF-2 and FGFR-1 in vitro. FGF-2 induced a robust growth response in these cells that could be blocked specifically by a neutralizing FGF-2 antibody. Interestingly, the addition of the neutralizing antibody alone did reduce basal proliferation up to 31.5%. In addition, FGF-2 induced expression of -smooth muscle actin up to 1.6-fold, but no significant effect was observed on the synthesis of collagen type I and fibronectin. Finally, staining for MIB-1 revealed a good correlation of interstitial FGF-2 positivity with interstitial and tubular proliferative activity (r = 0.71, P < 0.01 for interstitial proliferation, N = 30).

Conclusions

Interstitial FGF-2 protein and mRNA expression correlate with interstitial scarring. FGF-2 is a strong mitogen for cortical kidney fibroblasts and may promote autocrine fibroblast growth. Expression of FGF-2 correlates with interstitial and tubular proliferation in vivo.

METHODS

Materials

Rabbit polyclonal antibody to FGF-2 (Ab-2) and mouse monoclonal antibody to FGF receptor-1 (FGFR-1) were purchased from Calbiochem (La Jolla, CA, USA). Human recombinant FGF-1 and FGF-2 were from R&D Systems (Minneapolis, MN, USA), as was the neutralizing goat polyclonal antibody to FGF-2. Mouse monoclonal antibody to collagen type I was obtained from Southern Biotechnology (Birmingham, AL, USA). Polyclonal rabbit antibody to fibronectin antibody was purchased from Sigma (St. Louis, MO, USA), and mouse monoclonal antibody to heparan sulfate was from Boehringer Mannheim (Mannheim, Germany). For cell characterization, the following additional mouse monoclonal antibodies were used: anticytokeratin (Dako, Carpinteria, CA, USA), antivimentin (Boehringer Mannheim), anti-factor VIII (Dako), anti-human lymphocyte antigen-DR (HLA-DR) (Dako), anti–-smooth muscle actin (Paesel+Lorei, Wiesbaden, Germany), anticollagen type III (Southern Biotechnology), anti-CD 44 (Pharmingen, San Diego, CA, USA), anti-CD 54, and anti-CD 68 (Dako). The mouse monoclonal antibody to collagen type IV was a generous gift from Steven Gay (University of Zürich, Zürich, Switzerland). Trypsin-ethylenediaminetetraacetic acid (EDTA), Dulbecco's modified Eagle's medium (DMEM), Iscove's modified Dulbecco's medium, and fetal calf serum (FCS) were obtained from GIBCO BRL Ltd. (Paisley, Scotland). Cell culture dishes were from Becton Dickinson (Franklin Lakes, NJ, USA).

Patient material

Kidney tissues from 39 patients with primary and secondary renal diseases and a variable degree of tubulointerstitial fibrosis were used in this study. Tissues included 32 biopsies and 7 nephrectomy specimens. The nephrectomies were performed in patients with end-stage renal disease caused by intractable hypertension or recurrent urinary tract infections. The use of parts of nephrectomy specimens or kidney biopsies for research purposes was approved by the ethics committee of the Georg-August-University, and written consent was obtained from all patients prior to nephrectomy or kidney biopsy. Normal-appearing kidney tissue from kidneys that proved either unsuitable for transplantation or were explanted because of tumor nephrectomies were used as controls (N = 8).

Cell culture

Human medullary fibroblast cell lines Tk 173 (obtained from a normal kidney) and Tk 188 (from a kidney with tubulointerstitial fibrosis) have been characterized previously²⁶ and were cultured in DMEM supplemented with 10% FCS, penicillin (100 U/mL), and streptomycin (100 U/mL). For the isolation of primary human renal fibroblasts, cortical parts of renal biopsy cylinders or nephrectomy specimens were cut and emerged in DMEM medium with 20% FCS and penicillin/streptomycin in the previously mentioned concentrations. When cells had grown to confluence, they were characterized by immunofluorescence for cytokeratin, vimentin, -smooth muscle actin, collagen types I, III, IV, factor VIII, CD 68, and HLA-DR. Cells that were positive for vimentin and at least collagen types I and III, and negative for cytokeratin, factor VIII and HLA class II were considered to be fibroblasts and were used for induction assays in passages 2 through 8. For detection of heparan sulfate proteoglycane expression, fluorescence activated cell sorting (FACS) analysis was performed as described previously²⁶.

Specificity of the basic fibroblast growth factor antibody

To determine the specificity of the applied antibody, FGF-2 expression in Tk 173 and Tk 188 fibroblasts (which are easier to culture than primary fibroblasts) was first analyzed by immunocytochemistry as described previously, using the anti–FGF-2 and the secondary FITC-antirabbit antibodies at concentrations of 1:20²⁷. For Northern blot analyses, total cellular RNA was extracted from cultured cells using RNA-clean™ (AGS, Heidelberg, Germany) according to the manufacturer's instructions. Concentration of RNA was determined by absorbance at 260 nm, and samples were stored at -80°C prior to use. Northern blot analysis was performed as described²⁸. Briefly, 40 g of total RNA were electrophoresed on a 1.0% agarose gel containing 2.2 mol/L formaldehyde using 1 MOPS, pH 7.0, as the running buffer. Photographs of ethidium-bromide–stained gels were taken under ultraviolet illumination. RNA was subsequently transferred to a nylon membrane (Hybond N; Amersham, Arlington Heights, IL, USA) by capillary transfer for 12 hours with 20 standard saline citrate (SSC) as the transfer buffer. Blots were baked at 80°C and prehybridized for one hour at 68°C in prehybridization solution containing 0.5 mol/L Na₂HPO₄ buffer (pH 7.2), 0.5 mol/L EDTA (pH 8.0), 25% sodium dodecyl sulfate (SDS), and 1.5% blocking reagent. FGF-2 specific oligonucleotides (R&D Systems) were labeled with digoxigenin using a kit (Boehringer Mannheim) according to the manufacturer's instructions. Hybridizations were performed overnight at 58°C using the prehybridization solution. After washing, anti-DIG alkaline phosphate and disodium 3-(4-methoxyspiro{1,2-dioxetane-3,2'-(5'-chloro) tricyclo[3.3.1.1^3,7]decan}-4-yl)phenyl phosphate (CSPD) substrate (both Boehringer Mannheim) were added, and autoradiograms were obtained.

Subsequently, immunoblots were performed as described previously²⁷. Briefly, lysates from Tk 173 and Tk 188 fibroblasts were obtained by lysis with a detergent-based buffer containing 0.4% sodium deoxycholate, 1% NP-40, 1.9% EGTA, and 10 mmol/L Tris (pH 7.4). One hundred micrograms of total cellular protein and 50 ng FGF-2 control protein were run on an 18% SDS-polyacrylamide electrophoresis gel and transferred to a nitrocellulose membran (Hybond™ ECL™; Amersham) by electroblotting. Membranes were stained with Ponceau red to control for adequate transfer and equally loaded amounts. After blocking, the membrane was incubated with the respective antibody (anti–FGF-2 in a concentration of 1:40) followed by the secondary antibody (donkey-antirabbit, horseradish peroxidase-linked, concentration 1:3000; Amersham). Positive reaction products were identified by enhanced chemiluminescence (ECL; Amersham) according to the manufacturer's protocol.

Immunhistochemical staining for basic fibroblast growth factor and morphometric evaluation of interstitial volume

Tissue from normal human kidneys, renal biopsies, and end-stage kidneys were snap frozen in liquid nitrogen. Sections of 4 to 6 m were cut, mounted on microscope slides, and fixed in ice-cold acetone for 10 minutes. Sections were then allowed to dry, and unspecific binding was blocked by incubation with goat serum. Anti–FGF-2 antibody was added in a concentration of 1:20. After incubation for 60 minutes and several washes in phosphate-buffered saline (PBS), FITC-antirabbit was added in a concentration of 1:20. Positivity was evaluated with an Axiophot S100 microscope (Zeiss, Jena, Germany). Interstitial cortical staining for FGF-2 was rated by two blinded investigators as in normal kidneys (+), increased (++), robustly increased (+++), or absent (-). Negative controls consisted in substitution of the primary antibody with an irrelevant rabbit polyclonal antibody or PBS.

In selected sections, double labelings were performed to characterize FGF-2–positive interstitial cells. In that case, an addition of the anti–FGF-2 antibody was followed by staining for -smooth muscle actin, vimentin, cytokeratin, HLA-DR, CD 44, CD 54, and CD 68. FITC-antirabbit and the appropriate rhodamin-labeled antibody were added sequentially at a concentration of 1:20. Photographs were taken either by double or by sequential exposure using Zeiss camera and software (MC 200 Chip).

Selected sections from normal and fibrotic kidneys were also paraffin embedded after fixation in formaldehyde using a standard avidin-biotin complex (ABC) method, as described previously²⁷.

For morphometric evaluation, all tissue specimens were fixed in 4% buffered formalin solution (pH 7.4) and embedded in paraffin. Tissue sections were stained for trichrome or PAS, and point counting was performed as described by Bohle et al²⁹. Only those specimens containing at least five glomeruli were evaluated. Five projection fields of the cortex were examined per kidney using a lattice of 1 cm at a magnification of 250, discriminating between interstitium, tubular epithelia, glomeruli, and vasculature. Big blood vessels were ignored. The value for the interstitium was expressed as a percentage of the analyzed cortex.

Quantitation of basic fibroblast growth factor expression

Expression of FGF-2 protein was quantitated by immunoblot analysis. Briefly, lysates were obtained from six kidneys with considerable interstitial scarring and three control kidneys. Immunoblots were performed as described earlier in this article using 100 g of protein and 50 ng of FGF-2 recombinant protein as positive control. After staining for Ponceau red, blots were analyzed densitometrically for later normalization. Positive reaction products were again identified by chemiluminescence using a Fluor-S™ Multiimager (Bio-Rad, Hercules, CA, USA), and densitometric analysis was performed with Multi-Analyst™ software (Bio-Rad).

Vector assembly and in situ hybridization for basic fibroblast growth factor

In situ hybridizations were performed on 6 m sections of snap-frozen tissue from freshly explanted kidneys using a digoxigenin-labeled cRNA probe. FGF-2 cDNA was purchased from R&D Systems, digested with Hind III and BamHI, and cloned into pBS SK II (Stratagene, La Jolla, CA, USA) as described³⁰. The new vector was named FGF-2-pBS. Correct ligation was verified by sequence analysis using Alf-Express (Pharmacia, Piscataway, NJ, USA). The vector was linearized with the appropriate enzyme, and FGF-2 sense/antisense riboprobes were synthesized and labeled with digoxigenin by in vitro transcription with the DIG RNA labeling kit (Boehringer Mannheim) as described³¹. Tissue sections were mounted on Sectionlock™ slides (Polysciences, Eppelheim, Germany), briefly heated to 50°C to ensure fixation, dried for 30 minutes, and fixed in 3% paraformaldehyde for 7 minutes. Washing in PBS and 2 SSC was followed by prehybridization for 60 minutes at 43°C in hybridization buffer (Amersham) containing 50% formamide (Sigma). Hybridization was then performed in 40 to 60 L hybridization buffer per section containing the cRNA probe at a concentration of 200 ng/mL, followed by washing in 2 SSC for 5 minutes at 37°C and twice for 15 minutes in 0.1 SSC, SDS at 68°C. After a final wash in 100 mmol/L Tris-HCl and 150 mmol/L NaCl for five minutes at room temperature, 1% blocking buffer and 10% goat serum (Dako) were added, and the sections were incubated with a 1:500 dilution of alkaline-phosphatase conjugated anti-DIG antibody for two hours at room temperature. Detection was performed after several washes in 100 mmol/L Tris-HCl, 150 mmol/L NaCl with detection buffer containing 0.18 mg/mL 5-bromo-4-chloro-3-indolyl-phosphate (BCIP), 0.34 mg/mL nitro blue tetrazolium (NBT), and 240 g/mL levamisole. After 12 hours, color reaction was stopped by washing in 10 mmol/L Tris (pH 8) and 1 mmol/L EDTA for five minutes, and coverslips were mounted. Sections from human synovia, which is known to express FGF-2 abundantly, served as positive control.

Proliferation assay

Proliferation studies in primary cortical fibroblasts were performed using nonradioactive bromodeoxyuridine incorporation assays (Amersham) based on the method by Gratzner³². Four 10³ cells were plated per well in 96-well microtiter plates containing DMEM medium with the usual supplements and incubated overnight to allow cell attachment. Cells were then made quiescent by replacement of DMEM medium with Iscove's medium and subsequent incubation for 24 hours. In preliminary experiments, cytokines were added, and proliferation was measured after 24, 48, 72, and 96 hours. FGF-2 was used in concentrations of 0.1, 1, 10, and 100 ng/mL (equivalent to 5.5 10^-12 mol/L, 5.5 10^-11 mol/L, 5.5 10^-10 mol/L, and 5.5 10^-9 mol/L, respectively). In a second set of experiments, 1 ng/mL FGF-2 was added, and bromodeoxyuridine incorporation was measured after 72 hours. The specificity of the response to FGF-2 was determined by addition of a neutralizing FGF-2 antibody in a concentration of 10 g/mL (which had been determined in titration experiments with Tk 173 fibroblasts) or a nonrelevant IgG. Both antibodies were also added without FGF-2 to study their intrinsic effects on proliferation. Medium was changed to Iscove's containing BrdU/FdU in a dilution of 1:500 four hours prior to measurements. Cells were then washed three times in PBS and fixed in methanol containing 2% hydrogen peroxide followed by denaturation in 1N HCl for 10 minutes. After a further three times of washing in PBS containing 0.1% Tween and blocking for 30 minutes in PBS (0.1% Tween, 3% BSA), 50 L of anti-BrdU antibody were added and incubated for 45 minutes. Cells were again washed three times in PBS/Tween, and peroxidase substrate was added (100 L/well). Optical densities were subsequently determined photometrically at 405 nm (Dynatech MR 4000; Dynatech, Denkendorf, Germany). All experiments were performed in triplicate and repeated five times. Results are shown in percentage of negative control (Iscove's medium = 100%).

In a third set of experiments, confluent monolayers of fibroblasts were trypsinized, and cells were seeded at 4 10³ cells per well in 96-well culture plates. After 12 hours, DMEM medium was replaced by Iscove's medium. After a further 24 hours, cells were incubated with various factors (FGF-2 at concentrations of 0.1, 1, 10, and 100 ng/mL, FGF-2 (1 ng/mL) + FGF-neutralizing antibody (10 g/mL), FGF-2 (10 ng/mL) + irrelevant IgG, and FGF-2 (1 ng/mL) + heparin (10 g/mL) for 72 hours before cell counts were performed. PDGF-BB at 10 ng/mL and heparin alone (10 g/mL) served as controls. Again, all experiments were performed in triplicate and repeated five times.

Immunofluorescence staining of -smooth muscle actin

The effects of FGF-2 on the differentiation of kidney fibroblasts into myofibroblasts were assessed by immunofluorescent staining of -smooth muscle actin after continuous exposure to FGF-2 for 48 hours. Cells were seeded in microchamber slides and incubated in regular medium for 12 hours. After 24 hours of incubation in Iscove's medium for quiescence, FGF-2 was added in concentrations of 0.1, 1.0, and 10 ng/mL. At the end of incubation, cells were washed twice with PBS and fixed with ethanol/acetic acid (50:50, vol/vol) for 20 minutes at 4°C. The cells were subsequently washed again and incubated for two hours at room temperature with a 1:200 dilution of a monoclonal mouse antibody to human -actin isoform of smooth muscle cells. Cells were then washed again with PBS and incubated for 60 minutes at room temperature with the FITC-conjugated rabbit-antimouse antibody. Control cultures were incubated without application of the primary antibody. Incubation with TGF-1 (1 ng/mL) served as a positive control. Cells with the typical -actin structure were considered to be -actin positive. Additional staining with the Hoechst dye H33258 to identify nuclei was performed to facilitate cell counting. For each group, 500 cells were determined microscopically using an Axiophot S100 microscope (Zeiss) by a blinded investigator, and the percentage of -smooth muscle actin-positive cells was calculated. Each experiment was repeated three times independently.

Matrix synthesis

Enzyme-linked immunosorbent assay (ELISA) of supernatants were performed as described³³, with some modifications, including the use of chemiluminescence to enhance sensitivity³⁴. Eight 10³ cells were plated per well, and cells were again made quiescent by incubation in serum-free medium for 24 hours. Next, FGF-2 was added in concentrations of 0.1, 1.0, and 10 ng/mL followed by incubation for 48 hours. Human recombinant TGF-1 (R&D Systems) in a concentration of 1 ng/mL served as a positive control. Fifty g/mL ascorbic acid and 50 g/mL propionitrile (both from Sigma) were added in experiments evaluating collagen synthesis. Supernatants were transferred to a Maxisorp™ plate and incubated overnight at room temperature. Plates were subsequently dried for two hours and blocked with 3% dried milk. Incubation with 50 L of the primary antibody (anticollagen type I was used in a concentration of 1:300 and antifibronectin at 1:5000) was followed by washing two times with PBS/0.1% Tween and incubation with the secondary antibody [antimouse-IgG-AP and antirabbit-IgG-AP (both Boehringer Mannheim, Germany), both used at a concentration of 1:1000]. After two additional wash steps, 100 L of CSPD-Ready-to-Use substrate (Boehringer Mannheim) were added, and quantitation was performed in a luminometer (Mikrolumat CB 96P; Berthold, Bad Wildbad, Germany) using MikroWin™ software (Mikrotek, Overath, Germany). Nonspecific binding was determined by incubation with the secondary antibody only. The value was normally less than 5% of the total chemiluminescence and was subtracted from each assay. All assays were performed in triplicate and repeated five times. Standardization was obtained for the collagen type I ELISA using human type I collagen (Becton Dickinson, Bedford, MA, USA). The limit of detection was 1 ng/mL, with a working range of 10 to 1000 ng/mL. Human fibronectin (Sigma) served as standard for the fibronectin ELISA. The working range of that ELISA was 0.1 to 10 g/mL, with a limit of detection of 10 ng/mL. All values were corrected for cell counts (10³ cells), which were performed after transfer of supernatants.

Evaluation of cell proliferation in vivo

Immunostaining with MIB-1 (Dako) was used to evaluate cell proliferation in vivo. Five-micrometer thick sections were cut, dewaxed, and rehydrated according to routine procedures. Sections were pretreated in a microwave oven for 5 minutes and then incubated for 60 minutes at room temperature with the MIB-1 antibody, which was used in a concentration of 1:50. The streptavidin-biotin/alkaline phosphatase complex technique (Dako) was performed, and Fast Red™ was used as the chromogen. The sections were counterstained with hematoxylin. All stained nuclei were rated as positive. Proliferative activity within the interstitium and tubular epithelium was quantitated separately by counting MIB-1–positive cells in four fields at 250 magnification using an integration grid.

Statistical analysis

All values are expressed as mean SEM unless specified. Analysis of variance (ANOVA) was used to determine statistical differences between growth factor-treated groups and controls using Sigma-Stat™ software (Jandel Scientific, San Rafael, CA, USA). Further analysis was carried out using Bonferroni's test to control for multiple testing. For nonnormally distributed data (MIB-1 staining), the Mann–Whitney rank sum test was applied. A correlation between semiquantitative interstitial FGF-2 staining, morphometric volume, and MIB-1 positivity was evaluated by Spearman's rank-order correlation coefficient. P values < 0.05 were considered significant.

RESULTS

Specificity of the anti-basic fibroblast growth factor antibody

Since the exact localization of FGF-2 in the kidney is still controversial, we first determined the specificity of the antibody used for our studies. First, immunofluorescence labeling demonstrated that the two fibroblast lines, Tk 173 and Tk 188, were positive for FGF-2 using the Ab-2 antibody. These results were corroborated by Northern and Western blot analyses Figure 1. Northern blot analysis confirmed that both fibroblast lines express FGF-2 mRNA, with a main band at 7.0 kb Figure 1a. In addition, the immunoblot demonstrated that the antibody detects a specific 18 kd band in the two fibroblast lines corresponding to the size of the recombinant protein Figure 1b. The two additional bands found mainly in Tk 173 fibroblasts probably correspond to multiple translation initiation sites³⁵. There was no cross-reactivity of the antibody with FGF-1 (acidic fibroblast growth factor; data not shown).

Figure 1.

Specificity of the anti–basic fibroblast growth factor (anti-FGF-2) antibody. The two kidney fibroblast cell lines Tk 173 and Tk 188 express FGF-2 as determined by Northern blot (A) and immunoblot analyses (B). Northern blot gave a main band at 7.0 kb, as expected for FGF-2 (A). Immunoblot analysis demonstrated a main band at 18 kD corresponding to the size of the recombinant protein. Additional bands at 22.5 and 24 kD correspond to multiple transplantation initiation sites (B). One hundred micrograms of total cellular protein were loaded onto each lane, and 50 ng of recombinant FGF-2 control protein (rFGF-2) served as positive control.

Full figure and legend (22K)

Basic fibroblast growth factor expression is expressed in interstitial fibroblast-like cells and increased in kidneys with tubulointerstitial fibrosis

Basic fibroblast growth factor expression was first analyzed in frozen sections of normal kidneys by indirect immunofluorescence. Immunolabeling of normal kidneys demonstrated FGF-2 protein within glomeruli, tunica media of blood vessels, few tubular (mainly distal by morphology), and many, albeit not all, interstitial cells Figure 2. The number of positive interstitial cells was considerably smaller in paraffin sections, probably because of differences in antigen stability (data not shown). To further characterize the FGF-2–positive cells within the interstitium, immunofluorescent double labelings were performed in normal kidneys and biopsies from kidneys with minimal interstitial involvement (interstitium morphometrically <25%). The results are depicted in Figure 2. Most of the FGF-2–positive cells within the interstitium stained also positive for CD 54 (intercellular adhesion molecule-1; Figure 2e) and CD 44 Figure 2f, although many FGF-2–positive cells did not label for both. In addition, almost all FGF-2–positive cells did stain for vimentin as well (data not shown). There was a considerable overlap in interstitial staining for FGF-2 and -smooth muscle actin Figure 2c; however, there were some cells that expressed -smooth muscle actin but not FGF-2. There was some coexpression of FGF-2 and HLA-DR Figure 2a, but most FGF-2–positive interstitial cells were negative for the HLA class II molecule, which is expressed on dendritic cells and macrophages. Only occasionally was coexpression of FGF-2 and the macrophage marker CD 68 observed Figure 2b, indicating that only very few FGF-2–positive cells are macrophages. Judging by the difference in staining for CD 68 and HLA-DR, some of the cells that stained for FGF-2 may be dendritic cells, but the majority clearly are not. CD 44, CD 54, and -smooth muscle actin have all been propagated as potential markers of fibroblasts^36,37. Thus, it seems reasonable to assume that most FGF-2–positive cells within the interstitium are probably fibroblasts or fibroblast-like cells.

Figure 2.

Localization of FGF-2–positive cells in cortical tissue of a normal kidney by immunofluorescence. Indirect immunofluorescence double labelings for FGF-2 (FITC labeled) and HLA-DR (A), CD 68 (B), -smooth muscle actin (C), cytokeratin (D), CD 54 (E), and CD 44 (F) (rhodamine labeled) were performed in a human kidney without interstitial scarring. There is a close correlation in staining within the interstitium for FGF-2 and -smooth muscle actin, CD 54, and CD 44. However, some interstitial cells stain exclusively for -smooth muscle actin (C), some exclusively for FGF-2 and not for CD 54 (E), or CD 44 (F). A number of cells coexpress FGF-2 and HLA-DR (A), although most interstitial cells stain only for FGF-2. There are very few coexpressions between FGF-2 and CD 68 (B). Interstitial cells do not express cytokeratin; however, some tubular cells display coexpression of FGF-2 and cytokeratin (D). Original magnification 200.

Full figure and legend (148K)

We next compared the sections from eight normal renal tissues (from 3 female and 5 male patients, mean age 56.9 years) and 39 kidneys (from 12 female and 27 male patients, mean age 45.0 years), each with varying degrees of tubulointerstitial fibrosis. Since we were particularly interested in the correlation between FGF-2 expression and the degree of interstitial scarring, a variety of renal diseases was examined. Table 1 displays the characteristics of patients and tissue types. The relative percentage of the cortical interstitium was determined by morphometric analysis. In normal kidneys, the relative percentage of the interstitium varied between 15 and 25%. Conversely, in the nephrectomy specimens from patients with renal disease, the interstitium occupied between 49 and 72%. No change of FGF-2 expression was observed in normal human kidneys. However, robust up-regulation of FGF-2 expression was noted in kidneys with interstitial scarring. Table 1 summarizes the results. Figure 3 depicts the staining results in a kidney with marked interstitial fibrosis. FGF-2–positive cells Figure 3 a, c and e also stained positive for vimentin Figure 3b, -smooth muscle actin Figure 3d, and CD 44 Figure 3f. Again, there was only little co-staining for the dendritic cell and macrophage marker HLA-DR and only occasional co-staining for the macrophage marker CD 68 (data not shown), excluding prominent expression in these cells. In paraffin-embedded sections, there was an accentuation of extracellular matrix staining, including tubular basement membrane, whereas fewer cellular elements were labeled.

Figure 3.

Staining for FGF-2 in a fibrotic kidney. Indirect immunofluorescence for FGF-2 (A, C, E) and vimentin (B), -smooth muscle actin (D), and CD 44 (F) in a human kidney with marked interstitial fibrosis. The staining for all four molecules is increased compared with a normal kidney. There is marked coexpression for FGF-2 and all three markers of interstitial fibroblasts; however, again, the overlap in staining is not complete. Some interstitial cells label exclusively for FGF-2 and some exclusively for -smooth muscle actin (C and D). Original magnification 400 (A and B) and 200 (C, D, E and F).

Full figure and legend (113K)

Table 1 - Characteristics of the kidney specimens.

Full table

When the expression of FGF-2 was correlated with the cortical interstitial volume, we found a good correlation (r = 0.64, P < 0.01). However, four of the biopsy specimens displayed an enlargement of the interstitial volume caused exclusively by interstitial infiltration without significant interstitial scarring (numbers 14, 15, 33, and 34). No increase in FGF-2 expression was observed in any of the four specimens. Thus, when FGF-2 protein expression was correlated with the interstitial volume only in those kidneys in which widening of the interstitial space was due to scarring, we found an even better correlation (r = 0.84, P < 0.01). Figure 4 illustrates the correlation graphically.

Figure 4.

Correlation of staining for FGF-2 and morphometric volume. There was a positive correlation between semiquantitative FGF-2 labeling and morphometrically determined volume. Four cases of acute tubulointerstitial nephritis have been omitted for this analysis (discussed in the Results section). P < 0.01; r = 0.84, by Spearman's rank order coefficient.

Full figure and legend (5K)

When the specimens from patients with glomerulonephritis (N = 20) were analyzed exclusively, the correlation was still significant (r = 0.65, r < 0.01). The number of individual diseases was too small to analyze them separately. However, looking at the specimens with IgA nephritis, there may be a biphasic expression pattern of FGF-2 with a +++ expression at interstitial volumes of 37 to 46% and a ++ expression at volumes of 55 to 71%. There was no correlation between interstitial FGF-2 staining and interstitial volume in the specimens from patients with tubulointerstitial nephritis.

FGF-2 protein expression is increased in cortical tissue from kidneys with interstitial fibrosis

Although our immunofluorescence labeling studies had shown a good correlation between interstitial volume and interstitial FGF-2 labeling, we wanted to confirm this finding by immunoblot analysis. Thus, protein was isolated from cortical tissue from three normal kidneys and six kidneys with marked interstitial scarring [from patients 10 (that is, FK1), 11 (FK2), 13 (FK3), 18 (FK4), 19 (FK5), and 24 (FK6)]. With the exception of FK1, all fibrotic kidneys contained more FGF-2 than their normal counterparts Figure 5a. Figure 5b summarizes the result after densitometric analysis (P = 0.048).

Figure 5.

Expression of FGF-2 protein in kidneys with and without interstitial scarring. FGF-2 protein expression was examined by immunoblot analysis in three normal (NK1 through NK3) and six kidneys with at least 10% interstitial fibrosis (FK1 through FK6). With the exception of FK1, all fibrotic kidneys express more FGF-2 than their nonfibrotic counterparts (A). (B) The densitometric analysis. *P < 0.05.

Full figure and legend (12K)

In situ hybridizations reveal increased FGF-2 mRNA expression within the tubulointerstitium in end-stage kidneys

In situ hybridizations were performed to localize FGF-2–synthesizing cells in three normal and three end-stage kidneys with interstitial fibrosis, tubular atrophy, and dilation. As expected, hybridization with the sense probe gave no signal Figure 6a. Conversely, hybridization with the antisense probe demonstrated weak staining within glomeruli and blood vessels, as well as occasional tubular and interstitial cells in normal kidneys Figure 6b, indicating that these cells synthesize FGF-2 themselves. Furthermore, similar to what was observed by immunofluorescent staining, in situ hybridizations of fibrotic end-stage kidneys revealed a robust increase in signal for FGF-2 mRNA within interstitial cells Figure 6 c, d. The majority of interstitial cells displayed FGF-2 mRNA positivity. There was also increased staining in tubular cells, even in some tubules that were dilated and atrophic. Conversely, there was no increased FGF-2 mRNA synthesis within the glomeruli of the analyzed kidneys, indicating that most FGF-2 is, in fact, synthesized in nonglomerular cells. Thus, we conclude that the results of the in situ hybridizations parallel the protein stainings, demonstrating that increased FGF-2 mRNA synthesis takes place in interstitial cells within the kidney and that increased protein detection is not due to mere accumulation of extrarenally synthesized FGF-2.

Figure 6.

In situ hybridizations for FGF-2 in normal control kidney and end-stage kidneys. There is no staining in a normal kidney hybridized with the FGF-2 sense riboprobe (A). Hybridization of a normal kidney with the FGF-2 antisense riboprobe revealed faint staining of some glomerular, tubular, vascular, and interstitial (arrows) cells. The hybridization signal was robustly up-regulated in kidneys from patients with end-stage renal disease (C and D). Staining for FGF-2 mRNA was concentrated within the interstitium (C) and in some tubular cells (D), even in some tubules that appear atrophic (C). Conversely, there was no increased staining for FGF-2 mRNA within glomeruli. Original magnification 200 (A, B, and D) and 100 (C).

Full figure and legend (114K)

Detection of FGF-2 expression and its low- and high-affinity receptors in human primary cortical fibroblasts

To evaluate the possible role of FGF-2 in human renal fibrogenesis, the expression of FGF-2 and its high- and low-affinity receptors were examined in seven primary cortical fibroblast lines. All cell lines were characterized as fibroblasts and synthesized extracellular matrix components, including collagen types I and III, as well as fibronectin. Expression of FGFR-1, one of four high-affinity receptors, was analyzed by immunoblot analysis. The immunoblot analysis demonstrated a double band at approximately 120 kD in all seven fibroblast lines Figure 7a. The most robust expression of FGFR-1 was observed in fibroblasts isolated from kidneys 25, 31, and 39, with interstitial volumes of 41, 74, and 79%, respectively. In addition, FACS analyses confirmed the expression of heparan sulfate proteoglycane, which serves as a low-affinity receptor for FGF-2 in all seven cell lines. There was no difference in the level of expression (data not shown).

Figure 7.

Expression of FGF-2 and FGFR-1 in seven primary cortical fibroblast lines. (A) The expression of FGFR-1. (B) The expression of FGF-2 by immunoblot analyses. F33 and F34 were isolated from kidney biopsies without interstitial scarring, and the other five fibroblast lines were from kidneys with moderate (F25 and F26) and severe interstitial scarring (F24, F31, and F39). There was a good correlation between the degree of interstitial scarring and the expression of FGF-2, however, no such correlation could be found regarding the expression of FGFR-1 (discussed in the Results section).

Full figure and legend (17K)

Basic fibroblast growth factor protein itself could also be detected in all seven fibroblast lines studied Figure 7b. Interestingly, the level of FGF-2 protein expression correlated with the degree of interstitial scarring of the kidney from which the cells were isolated. Whereas F33 and F34, isolated from patients 33 and 34 with tubulointerstitial nephritis and no interstitial scarring, displayed relatively moderate positivity for FGF-2, the level of expression was much higher in the fibroblasts isolated from kidneys with interstitial scarring. In addition, there was a good correlation between the relative level of FGF-2 protein expression in the seven fibroblast lines and the semiquantitative FGF-2 immunofluorescence labeling in the kidneys from which the cells were isolated (r = 0.76, P = 0.038).

FGF-2 is a strong mitogen for renal fibroblasts, induces differentiation into myofibroblasts, and has no major effect on matrix protein synthesis

To evaluate the function of FGF-2, proliferation, differentiation into myofibroblasts, and matrix protein synthesis of human renal fibroblasts were examined in five different primary cortical fibroblast lines. In a first set of experiments, performed in F25 fibroblasts, FGF-2 caused a robust induction of proliferation, which was time and dose dependent. The increase of bromodeoxyuridine incorporation was significant after 48 hours (187.8 8.9%) at a concentration of 1 ng/mL. The maximum increase was seen with a concentration of 10 ng/mL after 72 hours (267.3 24.9%). Figure 8 summarizes the findings obtained with bromodeoxyuridine incorporation after stimulation with FGF-2 at a concentration of 1 ng/mL for 72 hours in three of the five fibroblast lines. F25 and F26 were isolated from biopsies of patients 25 and 26, displaying increased interstitial volumes because of fibrosis of 41 and 42%, respectively. Conversely, F34 originated from patient 34 with tubulointerstitial nephritis and no interstitial scarring. Bromodeoxyuridine incorporation was increased between 166.3 3.2% in F34 cells and 247.0 15.4% in F26 fibroblasts. This increase could be specifically inhibited by addition of a neutralizing antibody to FGF-2, whereas a nonrelevant IgG did not have an effect on FGF-2–induced proliferation. Interestingly, the addition of the FGF-2 neutralizing antibody alone did decrease bromodeoxyuridine incorporation in all five fibroblast lines. However, because of the low level of basal proliferative activity, that decrease was statistically significant in only three cell lines (maximum inhibition 31.5% in F25 cells). Conversely, the addition of the nonrelevant IgG resulted in either no change or an increase in bromodeoxyuridine uptake. In a third set of proliferation experiments, bromodeoxyuridine incorporation essays were confirmed by cell counting. In these experiments, cell numbers increased dose dependently to a maximal 234.3 15.8% (P < 0.01) in F25 and to 197.4 18.6% (P < 0.01) in F34 fibroblasts after 96 hours of incubation at a concentration of 10 ng/mL. Again, the higher concentration (100 ng/mL) did not have any superior effect to the 10 ng/mL concentration. There was no additional effect of heparin, which by itself caused a slight reduction in proliferation (maximum inhibition of 18.9 4.5%; NS).

Figure 8.

Effects of FGF-2 on proliferation in primary cortical fibroblasts. FGF-2 was added in a concentration 1.0 ng/mL to F26 (A), F25 (B), and F34 (C) fibroblasts and proliferation was determined after 72 hours by bromodeoxyuridine incorporation. Specificity was determined by the addition of a neutralizing FGF-2 antibody and a nonspecific immunoglobulin (IgG). FGF-2 induced a robust proliferative response in all three fibroblast lines, which could be blocked specifically by the neutralizing FGF-2 antibody. The addition of the neutralizing FGF-2 antibody alone resulted in a significant inhibition of proliferation in two of the three tested cell lines, whereas the nonspecific IgG resulted rather in a promitogenic response. Values are given in percentage of negative control (Iscove's medium) and are a mean of five independent experiments. All experiments were conducted in triplicates. *P < 0.05 vs. control; **P < 0.01 vs. control.

Full figure and legend (15K)

Activation of fibroblasts into so-called myofibroblasts represents a key step in renal fibrogenesis. We evaluated the effects of FGF-2 on the expression of -smooth muscle actin, which serves as a marker for myofibroblast formation. In two primary cortical fibroblast lines, F6 isolated from a kidney with mild interstitial scarring (interstitial volume 38%) and F39 from a kidney with an interstitial volume of 71%, FGF-2 resulted in a time- and dose-dependent induction of -smooth muscle actin expression. Figure 9 summarizes the results observed after 72 hours. The increase was similar to that obtained with TGF-1 (increase of -smooth muscle actin-positive cells from 52.3 3.3 to 82.3 1.9% after 72 hours of incubation with 1 ng/mL FGF-2; P < 0.01). Of course, we do not know if these changes were due to de novo expression of -smooth muscle actin or due to clonal selection of positive cells caused by cell population heterogeneity.

Figure 9.

Expression of -smooth muscle actin in primary cortical fibroblasts. The number of -smooth muscle actin-positive cells was determined after incubation with three different concentrations of FGF-2 (0.1, 1.0, or 10 ng/mL) for 48 hours. Iscove's medium served as a negative control, and TGF-1 (1 ng/mL) served as a positive control. FGF-2 induced the expression of -smooth muscle actin in both cell lines significantly. **P < 0.01 vs. control.

Full figure and legend (18K)

Finally, the influence of FGF-2 on extracellular matrix synthesis was evaluated by ELISAs for collagen type I and fibronectin in F6 and F39 fibroblasts. Whereas TGF-1 (at a concentration of 1 ng/mL) increased secretion of collagen type I (from 101.0 9.8 ng/10³ cells to 1018.9 98.5 ng/10³ cells in F39 fibroblasts after stimulation for 48 h) and fibronectin (from 31.8 4.5 g/10³ cells to 48.7 3.5 g/10³ cells, again after 48 h of stimulation) compared with controls in both cell lines, no significant effect could be observed using three different concentrations of FGF-2, although there was a tendency for a decrease in secretion for collagen type I in both cell lines (data not shown).

Increased FGF-2 expression correlates with increased interstitial and tubular MIB-1 staining

After having established that FGF-2 is a mitogen for human cortical kidney fibroblasts in vitro, we tried to determine whether increased FGF-2 expression was correlated with increased proliferative activity in vivo. Thus, we stained for expression of the Ki-67 antigen, which is associated with cell proliferation using the MIB-1 antibody. We were able to perform stainings for MIB-1 in cortical tissue from 24 patients and 8 normal controls. Fifteen patient specimens could not be analyzed because of the lack of tissue that had been used up for other purposes (specimens from patients 5, 7, 14 through 17, 21, 23, 27, 29, 30, 32, and 35 through 37). In normal kidneys, only few interstitial and tubular cells displayed MIB-1 positivity (median 1.5 in both; Figure 10a), whereas the number of MIB-1–positive cells was considerably higher in patient specimens (median 13.5, P = 0.001 in interstitial, median 9.0, P = 0.008 in tubular cells; Figure 10 b, c and d). The highest number of MIB-1–positive cells within the interstitium was observed in the two biopsies from patients with tubulointerstitial nephritis (numbers 33 and 34), which is not surprising given the high proliferative activity of infiltrating monocytes/macrophages and lymphocytes. When the number of MIB-1–positive cells was correlated with the semiquantitative FGF-2 staining, there was a positive correlation with the number of interstitial (r = 0.48, P < 0.01) and tubular (r = 0.53, P < 0.01) positive cells. However, the correlation was much more robust when we omitted the two patients with acute interstitial nephritis, who had a high proliferative activity of infiltrating mononuclear cells but a low level interstitial FGF-2 staining. In that case (N = 30), correlation of FGF-2 staining with MIB-1–positive interstitial cells (r = 0.71, P < 0.01) and tubular cells (r = 0.66, P < 0.01) was much better. The results are displayed in Figure 11. Thus, the results of the MIB-1 staining demonstrate that FGF-2 positivity is correlated positively with the number of MIB-1–positive cells within the interstitium and in tubular epithelia in scarring processes. This was even true when glomerular diseases were analyzed separately with a positive correlation between FGF-2 staining and interstitial MIB-1 positivity (r = 0.64, P < 0.05), although the number of analyzed tissues was small (N = 11). There does not seem to be a correlation of FGF-2 staining with acute inflammatory processes, although the number of tissues studied is too limited in that regard.

Figure 10.

Immunohistochemical staining for MIB-1–positive cells. Whereas normal kidneys displayed either no or only occasional MIB-1–positive cells (A), staining of kidneys with variable degrees of interstitial scarring demonstrate positive tubular and interstitial cells (highlighted by arrows). (B) Taken from patient 4 (IgA nephritis, interstitial space 46%, FGF-2 +++). (C) Taken from patient 36 (SLE, interstitial space 42%, FGF-2 ++). (D) Taken from patient 10 (end-stage renal disease caused by amyloidosis, interstitial space 68%, FGF-2 +++). Original magnification 400.

Full figure and legend (263K)

Figure 11.

Correlation of interstitial FGF-2 labeling with MIB-1–positive interstitial (A) and tubular epithelial (B) cells. There was a positive correlation between semiquantitative FGF-2 labeling and the number of proliferating interstitial (A; P < 0.01, r = 0.71), as well as tubular epithelial (B; P < 0.01, r = 0.66) cells. Two cases of acute tubulointerstitial nephritis have been omitted for this analysis (see RESULTS for details). The r depicts the value for Spearman's rank order coefficient.

Full figure and legend (9K)

DISCUSSION

Virtually all forms of chronic progressive renal disease are accompanied by changes in the tubulointerstitium, irrespective of the primary renal disease³⁸. This is not surprising since renal tubules, interstitium, and vasculature comprise more than 80% of the renal cortex³⁹. Tubulointerstitial changes consist of interstitial inflammation followed by fibrosis, tubular atrophy, and dilation. Along with glomerular scarring, interstitial fibrosis is an accurate predictor of renal function, even for primary glomerular diseases²⁹. One of the most perplexing features of renal fibrogenesis is the fact that fibrosis often progresses despite resolution of the primary inflammatory stimulus⁴⁰. Thus, we and others have propagated the existence of autocrine cytokine loops, which may represent one of the key differences between regular wound healing and scarring^41,42,43.

The current studies that combined descriptional and functional analyses were undertaken to elucidate the role of FGF-2 in renal fibrogenesis. We have carefully analyzed the usefulness of the antibody used in our study given the controversial information regarding the detection of FGF-2. Different splice variants may be the reason that a variety of studies came to different conclusions regarding the localization of FGF-2. As was shown in the medullary Tk 173 and Tk 188 fibroblasts, the Ab-2 antibody mainly detects the 18 kD band, but also the 22.5 and 24 kD bands described previously³⁵. These higher molecular weight forms may target intracellular FGF-2 to a nuclear localization⁴⁴; however, their exact role is currently unclear. The higher molecular weight forms could not be detected in our primary cortical fibroblasts. Thus, the 18 kD band seems to be the dominating isoform in kidney fibroblasts.

Our study demonstrates that FGF-2 protein can be detected in glomerular, vascular, a few (mainly distal) tubular, as well as interstitial fibroblast-like cells in normal human kidneys. In cryosections, more interstitial cells were labeled than in paraffin-embedded tissue sections, which may explain the differences to the study by Floege et al, who described only occasional interstitial cell positivity¹⁸. The in situ hybridizations confirmed that the same cells that stain positive for FGF-2 in cyrosections also synthesize FGF-2 mRNA. In addition, we have demonstrated for the first time that the expression of FGF-2 protein and mRNA is increased in kidneys with tubulointerstitial scarring, irrespective of the primary disease. Furthermore, our study has shown that primary cortical fibroblasts express FGF-2 and one of its high-affinity receptors. FGF-2 causes a robust induction in proliferation in these cells and induces -smooth muscle actin expression as marker for myofibroblast formation but has no major effect on synthesis of collagen type I and fibronectin. Proliferative activity in tubular epithelial and interstitial cells is increased in diseased human kidneys and correlates with staining for FGF-2. Finally, basal proliferation in cortical fibroblasts can be reduced by the addition of a neutralizing antibody to FGF-2, pointing to a possible autocrine mechanism of fibroblast proliferation. Thus, our study provides evidence for the existence of an autocrine loop in human renal fibrogenesis, which is characterized by constitutive expression of FGF-2 and its receptors, up-regulation in disease state, induction of proliferation, and decreased proliferation after neutralization of the cytokine.

The increased expression of FGF-2 in renal fibrogenesis, although demonstrated for the first time, is not surprising. The association of FGF-2 expression with proliferative fibrogenesis was first demonstrated by Gonzalez et al in patients with Dupuytren's contracture⁴⁵. They noted already intense staining for FGF-2 associated with increased synthesis for FGF-2 and FGFR-1 mRNA in fibroblasts. Charlotte et al found increased expression of FGF-2 in the carbon tetrachloride model of rat liver fibrosis⁴⁶. In that study, FGF-2 expression was localized to myofibroblastic liver cells. Interestingly, FGF-2 knockout mice display delayed wound healing⁴⁷, further pointing to a critical role for the cytokine in matrix formation. Unlike most other polypeptide growth factors, FGF-2 lacks a conventional leader sequence for secretion, and the mechanisms for its release are still unclear¹¹. Chronically injured cells, however, release FGF-2 efficiently, and the mitogenic effect of the cytokine is potentiated by tissue injury⁴⁸. In addition, as was originally shown by Mignatti, Morimoto, and Rifkin, FGF-2 may be secreted by viable cells⁴⁹.

What is the role of proliferation in renal fibrogenesis? Proliferation was thought to play only a little role, if at all, in the progression of renal disease. However, as was demonstrated by Nadasdy et al in a study of 28 human end-stage kidneys, even scarred kidneys display high proliferative activity in tubuli and interstitial cells⁵⁰. Similar results were recently described by Thomas et al in the subtotal nephrectomy model of chronic renal failure⁵¹. These findings are corroborated by our study. It is intriguing to speculate that fibroblasts and tubular epithelial cells stimulate each other in the process of renal fibrogenesis, as indicated by the paralleled proliferation. We and others have recently demonstrated that tubular epithelial cells may acquire certain characteristics of fibroblasts^27,52. Although we did not examine this aspect in the current study, FGF-2 may be a candidate cytokine that could play a role in that process. Clearly, the interaction between tubular epithelial cells and fibroblasts plays a role not only in the initiation of interstitial inflammation, but also in progressive renal scarring. In the study by Ray et al in their transgenic mouse model of HIV-associated nephropathy, FGF-2 detection was associated with interstitial areas that contained both cell injury and proliferation²⁵. Thus, proliferation does play a critical role in the progression of renal disease, and autocrine/paracrine secretion of FGF-2 may be important in the perpetuation of the process. Autocrine stimulation of cell proliferation caused by FGF-2 has been described in a number of cells, including tubular epithelial⁴³, endothelial⁵³, and vascular smooth muscle⁵⁴ cells. In these cell types, as in cortical fibroblasts, the efficacy of the neutralizing antibody indicates that the protein does get secreted before binding to its receptors.

Basic fibroblast growth factor exerts its effects via low- and four high-affinity receptors. The high-affinity receptors consist of four transmembrane tyrosine kinases (FGFR-1 to -4)¹¹. Our study has demonstrated that primary cortical fibroblasts express FGFR-1, also known as flg, and heparan sulfate proteoglycan, which potentiates the binding to the high-affinity receptors⁵⁵. Floege et al noted occasional interstitial staining for flg¹⁸. In our study, only three of the five examined fibroblast lines from kidneys with interstitial fibrosis displayed an up-regulated expression for FGFR-1. One possible explanation is the redundancy of the fibroblast receptor system, including multiple splice variants. In a recent study by Ford et al, all seven receptor isoforms assayed by the group were expressed in whole cortex by reverse transcription-polymerase chain reaction assay⁵⁶. Thus, increased expression of one of the other high-affinity receptors may explain the lack of up-regulation of FGFR-1 in two kidneys with interstitial scarring. Alternatively, down-regulation of FGFR-1 during a late stage of fibrogenesis may protect the kidney from uncontrolled proliferation. The mechanism of overexpression of FGF-2 and FGFR-1 has been described in certain tumor cells, including human hepatoma cells⁵⁷.

Heparan sulfate proteoglycan is a low-affinity receptor for FGF-2. Its expression in all seven primary cortical fibroblasts may explain why heparin (which otherwise is known to increase the proliferative effects of FGF-2) did not have any additional effects on proliferation. Conversely, the slight reduction in proliferation caused by heparin itself may be explained by competition with this low-affinity receptor⁵⁸.

Our findings of FGF-2 and FGFR-1 detection in human kidney fibroblasts or fibroblast-like cells are in concordance with observations by other groups in dermal⁵⁹, mammary gland⁶⁰, and human liver⁶¹ fibroblasts. However, we extend those findings, demonstrating that renal fibroblasts (or fibroblast-like cells) do express FGF-2 in vivo as well. The interstitial cells that express FGF-2 are predominantly renal fibroblast-like cells as shown by double labelings. Vimentin, -smooth muscle actin, CD 44, and CD 54 are all markers for fibroblasts in the kidney^36,37. However, though almost all CD 44- and CD 54-positive cells within the interstitium expressed FGF-2, some interstitial cells displayed exclusive FGF-2 positivity. In addition, a number of -smooth muscle actin-positive cells were negative for FGF-2. Renal interstitial fibroblasts are a heterogeneous population⁶², and studies on these cells are still hampered by the lack of a definite fibroblast marker¹. Most of these cells seem to express FGF-2 in vivo, although some subpopulations may not. Conversely, some nonfibroblast cells such as monocytes/macrophages or dendritic cells express FGF-2, although the majority of these cells do not under normal conditions.

Previous studies suggested that extrarenal sources may contribute to FGF-2 deposition in the kidney²⁵. Although we cannot completely exclude that possibility, it seems unlikely given the marked up-regulation of FGF-2 mRNA synthesis in fibrotic kidneys, as demonstrated by the in situ hybridization studies. Increased staining for FGF-2 in renal disease has been described before in various animal models^14,25 and in patients with focal and segmental glomerulosclerosis and IgA nephropathy⁶³. In that latter study, FGF-2 could be detected in glomeruli and in areas of tubulointerstitial damage, correlating with the degree of interstitial injury. Morita et al examined the expression of FGF-2 binding domains in human kidneys. FGF-2 binding heparan sulfate domains are present in the basement membrane proteoglycan perlecan, but are absent from other glycosaminoglycans such as fibroglycan and glypican⁶⁴. Heparan sulfate was detected in sections with glomerular sclerosis and interstitial fibrosis; however, only the fibrotic sections contained FGF-2 binding⁶⁵. Interestingly, a small population of macrophage-like parenchymal cells was identified in that study which expressed the FGF-2–binding domain on the cell surface, which is in concordance with the results of our study.

No significant effect of FGF-2 on synthesis of matrix components was observed in our study, although there was a tendency for decreased synthesis of collagen type I. Decreased synthesis of collagen type I in cells of mesenchymal origin after incubation with FGF-2 has been described before. Tan et al, for example, demonstrated a down-regulation of collagen gene expression in keloid fibroblasts⁶⁶. Thus, unlike TGF-1, a role of FGF-2 in the accumulation of extracellular matrix is not supported by our study nor by the literature. However, FGF-2 does induce expression of -smooth muscle actin as an indicator of myofibroblast formation. In that regard, it is similar to TGF-1, which is the prototype of a profibrogenic cytokine in the kidney. Expression of -smooth muscle actin has been correlated with interstitial matrix formation and even with prognosis in a number of studies in the kidney⁶⁷ and other organs⁶⁸. It certainly represents a critical step in fibrogenesis⁶⁹.

In summary, we have demonstrated that FGF-2 is expressed in interstitial fibroblast-like cells and that it is robustly up-regulated within these and tubular epithelial cells in human kidney fibrosis. Whereas FGF-2 does not seem to play a role in interstitial matrix production, it may have a critical role in fibroblast proliferation and myofibroblast formation. The continuity of fibroblast proliferation is probably one of the major differences between fibrosis and restitutional healing. Our study points to an important role of FGF-2 in that process. Clearly, further studies are necessary that delineate the differences between regular wound healing with no or minor scar formation and interstitial fibrosis resulting in progressive loss of organ function.

References

1. Eddy, AA: Molecular insights into renal interstitial fibrosis. J Am Soc Nephrol 1996 7:2495–2508,  | PubMed | ISI | ChemPort |
2. Adzick, NS & Lorenz, HP: Cells, matrix, growth factors and the surgeon: The biology of scarless fetal wound repair. Ann Surg 1994 220:10–18,  | PubMed | ISI | ChemPort |
3. Whitby, DJ & Ferguson, MWJ: The extracellular matrix of lip wounds in fetal, neonatal, and adult mice. Development 1991 112:651–668,  | PubMed | ISI | ChemPort |
4. Whitby, DJ & Ferguson, MWJ: Immunohistochemical localization of growth factors in fetal wound healing. Dev Biol 1991 147:207–215,  | Article | PubMed | ISI | ChemPort |
5. El Nahas, AM, Muchaneta-Kubara, EC, Essawy, M, Soylemezoglu, O: Renal fibrosis: Insights into pathogenesis and treatment. Int J Biochem Cell Biol 1997 29:55–62,  | Article | PubMed | ChemPort |
6. Bohle, A, Müller, GA, Wehrmann, M, Mackensen-Haen, S, Xiao, JC: The pathogenesis of chronic renal failure in the primary glomerulopathies, renal vasculopathies, and chronic interstitial nephritides. Kidney Int 1996 49 (Suppl 54): S2–S9,  | ISI |
7. Knecht, A, Fine, LG, Kleinman, KS, Rodemann, HP, Müller, GA, Woo, DDL, Norman, JT: Fibroblasts of rabbit kidney in culture. II. Paracrine stimulation of papillary fibroblasts by PDGF. Am J Physiol 1991 261:F292–F299,  | PubMed | ISI | ChemPort |
8. Tang, WW, Ulich, TR, Lacey, DL, Hill, DC, Qi, M, Kaufman, SA, Yee, J: Platelet derived growth factor-BB induces renal tubulointerstitial myofibroblast formation and tubulointerstitial fibrosis. Am J Pathol 1996 148:1169–1180,  | PubMed | ISI | ChemPort |
9. Sporn, MB & Roberts, AB: Transforming growth factor-: Recent progress and new challenges. J Cell Biol 1992 119:1017–1021,  | Article | PubMed | ISI | ChemPort |
10. Border, WA & Noble, NA: Mechanisms of disease: Transforming growth factor in tissue fibrosis. N Engl J Med 1994 331:1286–1292,  | Article | PubMed | ISI | ChemPort |
11. Friesel, RE & Maciag, T: Molecular mechanisms of angiogenesis: Fibroblast growth factor signal transduction. FASEB J 1995 9:919–925,  | PubMed | ISI | ChemPort |
12. Shinozaki, M, Kawara, S, Hayashi, N, Kakinuma, T, Igarashi, A, Takehara, K: Induction of subcutaneous tissue fibrosis in newborn mice by transforming growth factor : Simultaneous application with basic fibroblast growth factor causes persistent fibrosis. Biochem Biophys Res Commun 1997 237:292–296,  | Article | PubMed | ISI | ChemPort |
13. Baird, A, Esch, F, Böhlen, P, Ling, N, Gospodarowicz, D: Isolation and partial characterization of an endothelial cell growth factor from the bovine kidney: Homology with basic fibroblast growth factor. Regul Pept 1985 12:201–213,  | Article | PubMed | ISI | ChemPort |
14. Floege, J, Eng, E, Lindner, V, Young, BA, Reidy, MA, Johnson, RJ: Rat glomerular mesangial cells synthesize basic fibroblast growth factor: Release, upregulated synthesis, and mitogenecity in mesangial proliferative glomerulonephritis. J Clin Invest 1992 90:2362–2369,  | PubMed | ISI | ChemPort |
15. Takeuchi, A, Yoshizawa, N, Yamamoto, M, Sawasaki, Y, Oda, T, Senoo, A, Niwa, H, Fuse, Y: Basic fibroblast growth factor promotes proliferation of rat glomerular visceral epithelial cells in vitro. Am J Pathol 1992 141:107–116,  | PubMed | ISI | ChemPort |
16. Hughes, SE & Hall, PA: Immunolocalization of fibroblast growth factor receptor 1 and its ligands in human tissues. Lab Invest 1993 69:173–182,  | PubMed | ISI | ChemPort |
17. Cauchi, J, Alcorn, D, Cancilla, B, Key, B, Berka, JL, Nurcumbe, V, Ryyan, GB, Bertram, JF: Light-microscopic immunolocalization of fibroblast growth factor-1 and -2 in adult rat kidney. Cell Tissue Res 1996 285:179–187,  | Article | PubMed | ISI | ChemPort |
18. Floege, J, Hudkins, KL, Eitner, F, Cui, Y, Morrison, RS, Schelling, MA, Alpers, CE: Localization of fibroblast growth factor-2 (basic FGF) and FGF receptor-1 in adult human kidney. Kidney Int 1999 56:883–897,  | Article | PubMed | ISI | ChemPort |
19. Silver, BJ, Jaffer, FE, Abboud, HE: Platelet-derived growth factor synthesis in mesangial cells: induction by multiple peptide mitogens. Proc Natl Acad Sci USA 1988 86:1056–1060,
20. Kanda, S, Nomata, K, Saha, PK, Nishimura, N, Yamada, J, Kanatake, H, Saito, Y: Growth factor regulation of renal cortical tubular cells by epidermal growth factor, insulin-like growth factor-1, acidic and basic fibroblast growth factor, and transforming growth factor-. Cell Biol Int Rep 1989 13:687–699,  | PubMed | ISI | ChemPort |
21. Ballermann, BJ: Regulation of bovine glomerular endothelial cell growth in vitro. Am J Physiol 1989 256:C182–C189,  | PubMed | ISI | ChemPort |
22. Floege, J, Eng, E, Young, BA, Alpers, CE, Barrett, TB, Bowen-Pope, DF, Johnson, RJ: Infusion of platelet-derived growth factor or basic fibroblast growth factor induces selective glomerular mesangial cell proliferation and matrix accumulation in rats. J Clin Invest 1993 92:2952–2962,  | PubMed | ISI | ChemPort |
23. Floege, J, Kriz, W, Schulze, M, Susani, M, Kerjaschki, D, Mooney, A, Couser, WG, Koch, KM: Basic fibroblast growth factor augments podocyte injury and induces glomerulosclerosis in rats with experimental membranous nephropathy. J Clin Invest 1995 96:2809–2819,  | PubMed | ISI | ChemPort |
24. Kriz, W, Hähnel, B, Rösener, S, Elger, M: Long-term treatment of rats with FGF-2 results in focal segmental glomerulosclerosis. Kidney Int 1995 48:1435–1450,  | PubMed | ISI | ChemPort |
25. Ray, PE, Bruggeman, LA, Weeks, BS, Kopp, JB, Bryant, JL, Owens, JW, Notkins, AL, Klotman, PE: bFGF and its low affinity receptors in the pathogenesis of HIV-associated nephropathy in transgenic mice. Kidney Int 1994 46:759–772,  | PubMed | ISI | ChemPort |
26. Müller, GA, Frank, J, Rodemann, HP, Engler-Blum, G: Human renal fibroblast cell lines (tFKIF and tNKF) are new tools to investigate pathophysiologic mechanisms of renal interstitial fibrosis. Exp Nephrol 1995 3:127–133,  | PubMed | ISI | ChemPort |
27. Strutz, F, Okada, H, Lo, CW, Danoff, T, Carone, B, Tomaszewski, J, Neilson, EG: Identification and characterization of fibroblast-specific protein 1 (FSP1). J Cell Biol 1995 130:393–405,  | Article | PubMed | ISI | ChemPort |
28. Engler-Blum, G, Meier, M, Frank, J, Müller, GA: Reduction of background problems in nonradioactive northern and southern blot analyses enables higher sensitivity than 32P-based hybridizations. Anal Biochem 1993 210:235–244,  | Article | PubMed | ChemPort |
29. Bohle, A, Grund, KE, Mackensen, S, Tolon, M: Correlations between renal interstitium and level of serum creatinine: Morphometric investigations of biopsies in perimembranous glomerulonephritis. Virchows Arch A Pathol Anat Histol 1977 373:15–22,  | Article | PubMed | ChemPort |
30. Albert, SE, Strutz, F, Shelton, K, Haverty, T, Sun, MJ, Li, S, Denham, A, Maki, RA, Neilson, EG: Characterization of a cis-acting regulatory element which silences expression of the class II-A gene in epithelium. J Exp Med 1994 180:233–240,  | Article | PubMed | ISI | ChemPort |
31. Dijkman, HB, Mentzel, S, de Jong, AS, Assmann, KJ: RNA in situ hybridization using digoxigenin labeled cRNA probes. Biochemica 1995 2:21–25,
32. Gratzner, HG: Monoclonal antibody to 5-bromo- and 5-iododeoxyuridine: A new reagent for detection of DNA replication. Science 1982 218:474–475,  | PubMed | ISI | ChemPort |
33. Jacot, TA, Striker, GE, Stetler-Stevenson, M, Striker, LJ: Mesangial cells from transgenic mice with progressive glomerulosclerosis exhibit stable, phenotypic changes including undetectable MMP-9 and increased type IV collagen. Lab Invest 1996 75:791–799,  | PubMed | ISI | ChemPort |
34. Kricka, LJ: Labels, labeling, analytical strategies, and applications, in Nonisotopic Probing, Blotting, and Sequencing, 1995, edited by Kricka LJ, San Diego, New York, Boston, London, Sydney, Tokyo, Toronto, Academic Press, pp 3–40
35. Florkiewicz, RZ & Sommer, A: Human basic fibroblast growth factor gene encodes four polypeptides: Three initiate translation from non-AUG codons. Proc Natl Acad Sci USA 1989 86:3978–3981,  | PubMed | ChemPort |
36. Clayton, A, Steadman, R, Williams, JD: Cells isolated from the human cortical interstitium resemble myofibroblasts and bind neutrophils in an ICAM-1-dependent manner. J Am Soc Nephrol 1997 8:604–615,  | PubMed | ISI | ChemPort |
37. Alpers, CE, Hudkins, KL, Floege, J, Johnson, RJ: Human renal cortical interstitial cells with some features of smooth muscle cells participate in tubulointerstitial and crescentic glomerular injury. J Am Soc Nephrol 1994 5:201–210,  | PubMed | ISI | ChemPort |
38. Nath, KA: Tubulointerstitial changes as a major determinant in the progression of renal damage. Am J Kidney Dis 1992 20:1–17,  | PubMed | ISI | ChemPort |
39. Palmer, BF: The renal tubule in the progression of chronic renal failure. J Invest Med 1997 45:346–361,  | ISI | ChemPort |
40. Fogo, AB & Kon, V: Pathophysiology of progressive renal disease,, in Immunologic Renal Diseases, 1997, edited by Neilson EG,Couser WG, Philadelphia, Lippincott-Raven, pp 683–704
41. Lonnemann, G, Shapiro, L, Engler-Blum, G, Müller, GA, Koch, KM, Dinarello, CA: Cytokines in human renal interstitial fibrosis. I. IL-1 is an autocrine growth factor for cultured fibrosis-derived kidney fibroblasts. Kidney Int 1995 47:837–844,  | PubMed | ChemPort |
42. Phillips, AO, Topley, N, Morrisey, K, Williams, JD, Steadman, R: Basic fibroblast growth factor stimulates the release of preformed transforming growth factor 1 from human proximal tubular cells in the absence of de novo gene transcription or mRNA translation. Lab Invest 1997 76:591–600,  | PubMed | ISI | ChemPort |
43. Anderson, RJ & Ray, CJ: Potential autocrine and paracrine mechanisms of recovery from mechanical injury of renal tubular epithelial cells. Am J Physiol 1998 274:F463–F472,  | PubMed | ISI | ChemPort |
44. Mason, IJ: The ins and outs of fibroblast growth factors. Cell 1994 78:547–552,  | Article | PubMed | ISI | ChemPort |
45. Gonzalez, A-M, Buscaglia, M, Fox, R, Isacchi, A, Sarmientos, P, Farris, J, Ong, M, Martineau, D, Lappi, DA, Baird, A: Basic fibroblast growth factor in Dupuytren's contracture. Am J Pathol 1992 141:661–671,  | PubMed | ISI | ChemPort |
46. Charlotte, F, Win, KM, Préaux, AM, Dhumeaux, D, Zafrani, ES, Mavier, P: Immunolocalization of heparin-binding growth factors (HBGF) types 1 and 2 in rat liver: Selective hyperexpression of HBGF-2 in carbon tetrachloride-induced fibrosis. J Pathol 1993 169:471–476,  | Article | PubMed | ISI | ChemPort |
47. Ortega, S, Ittmann, M, Tsang, SH, Ehrlich, M, Basilico, C: Neuronal defects and delayed wound healing in mice lacking fibroblast growth factor 2. Proc Natl Acad Sci USA 1998 95:5672–5677,  | Article | PubMed | ChemPort |
48. Muthukrishnan, L, Warder, E, McNeil, PL: Basic fibroblast growth factor is efficiently released from a cytosolic storage site through plasma membrane disruptions of endothelial cells. J Cell Physiol 1991 148:1–16,  | Article | PubMed | ISI | ChemPort |
49. Mignatti, P, Morimoto, T, Rifkin, DB: Basic fibroblast growth factor released by single, isolated cells stimulates their migration in an autocrine manner. Proc Natl Acad Sci USA 1991 88:11007–11011,  | PubMed | ChemPort |
50. Nadasdy, T, Laszik, Z, Blick, KE, Johnson, DL, Silva, FG: Tubular atrophy in the end-stage kidney: A lectin and immunohistochemical study. Hum Pathol 1994 25:22–28,  | Article | PubMed | ISI | ChemPort |
51. Thomas, GL, Yang, B, Wagner, BE, Savill, J, Elnahas, AM: Cellular apoptosis and proliferation in experimental renal fibrosis. Nephrol Dial Transplant 1998 13:2216–2226,  | Article | PubMed | ISI | ChemPort |
52. Ng, YY, Huang, TP, Yang, WC, Chen, ZP, Yang, AH, Mu, W, Nikolic-Paterson, DJ, Atkins, RC, Lan, HY: Tubular epithelial-myofibroblast transdifferentiation in progressive tubulointerstitial fibrosis in 5/6 nephrectomized rats. Kidney Int 1998 54:864–876,  | Article | PubMed | ISI | ChemPort |
53. Schweigerer, L, Neufeld, G, Friedman, J, Abraham, JA, Fiddes, JC, Gospodarowicz, D: Capillary endothelial cells express basic fibroblast growth factor, a mitogen that promotes their own growth. Nature 1987 325:257–259,  | Article | PubMed | ISI | ChemPort |
54. Davis, MG, Zhou, M, Ali, S, Coffin, JD, Doetschman, T, Dorn, GW: Intracrine and autocrine effects of basic fibroblast growth factor in vascular smooth muscle cells. J Mol Cell Cardiol 1997 29:1061–1072,  | Article | PubMed | ISI | ChemPort |
55. Yayon, A, Klagsbrun, M, Esko, JD, Leder, P, Ornitz, DM: Cell surface heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptors. Cell 1991 64:841–848,  | Article | PubMed | ISI | ChemPort |
56. Ford, MD, Cauchi, J, Greferath, U, Bertram, JF: Expression of fibroblast growth factors and their receptors in rat glomeruli. Kidney Int 1997 51:1729–1738,  | PubMed | ISI | ChemPort |
57. Kin, M, Sata, M, Ueno, T, Torimura, T, Inuzuka, S, Tsuji, R, Sujaku, K, Sakamoto, M, Sugawara, H, Tamaki, S, Tanikawa, K: Basic fibroblast growth factor regulates proliferation and motility of human hepatoma cells by an autocrine mechanism. J Hepatol 1997 27:677–687,  | Article | PubMed | ISI | ChemPort |
58. Rieck, P, Oliver, L, Engelmann, K, Fuhrmann, G, Hartmann, C, Courtois, Y: The role of exogenous/endogenous basic fibroblast growth factor (FGF2) and transforming growth factor (TGF-1) on human corneal endothelial cells proliferation in vitro. Exp Cell Res 1995 220:36–46,  | Article | PubMed | ISI | ChemPort |
59. Wang, Y, Rao, U, Mascari, R, Richards, TJ, Panson, AJ, Edington, HD, Shipe-Spotloe, JM, Donnelly, SS, Kirkwood, JM, Becker, D: Molecular analysis of melanoma precursor cells. Cell Growth Differ 1996 7:1733–1740,  | PubMed | ISI | ChemPort |
60. Barraclough, R, Fernig, DG, Rudland, PS, Smith, JA: Synthesis of basic fibroblast growth factor upon differentiation of rat mammary epithelial to myoepithelial-like cells in culture. J Cell Physiol 1990 144:333–344,  | Article | PubMed | ISI | ChemPort |
61. Rosenbaum, J, Blazejewski, S, Preaux, A-M, Mallat, A, Dhumeaux, D, Mavier, P: Fibroblast growth factor 2 and transforming growth factor 1 interactions in human liver fibroblasts. Gastroenterology 1995 109:1986–1996,  | Article | PubMed | ISI | ChemPort |
62. Müller, GA & Strutz, F: Renal fibroblast heterogeneity. Kidney Int 1995 48 (Suppl 50): S33–S36,
63. Stein-Oakley, AN, Maguire, JA, Dowling, J, Perry, G, Thomson, NM: Altered expression of fibrogenic growth factors in IgA nephropathy and focal and segmental glomerulosclerosis. Kidney Int 1997 51:195–204,  | PubMed | ChemPort |
64. Aviezer, D, Hecht, D, Safran, M, Eisinger, M, David, G, Yayon, A: Perlecan, basal lamina proteoglycan, promotes basic fibroblast growth factor-receptor binding, mitogenesis and angiogenesis. Cell 1994 79:1005–1013,  | Article | PubMed | ISI | ChemPort |
65. Morita, H, Shinzato, T, David, G, Mizutani, A, Habuchi, H, Fujita, Y, Ito, M, Asai, J, Maeda, K, Kimata, K: Basic fibroblast growth factor-binding domain of heparan sulfate in the human glomerulosclerosis and renal tubulointerstitial fibrosis. Lab Invest 1994 71:528–535,  | PubMed | ISI | ChemPort |
66. Tan, EML, Rouda, S, Greenbaum, SS, Moore, JH, Fox, JW, Sollberg, S: Acidic and basic fibroblast growth factors down-regulate collagen gene expression in keloid fibroblasts. Am J Pathol 1993 142:463–470,  | PubMed | ISI | ChemPort |
67. Diamond, JR, Vangoor, H, Ding, G, Engelmyer, E: Myofibroblasts in experimental hydronephrosis. Am J Pathol 1995 146:121–129,  | PubMed | ISI | ChemPort |
68. Zhang, K, Rekter, MD, Gordon, D, Phan, SH: Myofibroblasts and their role in lung collagen gene expression during pulmonary fibrosis. Am J Pathol 1994 145:114–125,  | PubMed | ISI | ChemPort |
69. Strutz, F & Müller, GA: Interstitial pathomechanisms underlying progressive tubulointerstitial damage. Kidney Blood Press Res 1999 22:67–75,

Acknowledgments

This work was supported in part by grants from the Deutsche Forschungsgemeinschaft (DFG) to F. Strutz (DFG Str 388/3-1) and G.A. Müller (DFG Mü 523/7-1). M. Zeisberg was awarded a graduate grant from the DFG. The authors wish to acknowledge the excellent technical assistance of M. Dietrich and A. Renziehausen. Parts of this work were presented in abstract form at the national meetings of the American Society of Nephrology in November 1996 in New Orleans, LA, USA (J Am Soc Nephrol 7:A2596, 1996) and in November 1997 in San Antonio, TX, USA (J Am Soc Nephrol 8:A2456, 1997).