Role of CX 3 C-chemokine CX 3 C-L/fractalkine expression in a model of slowly progressive renal failure

Background. The chemokine/chemokine receptor pair CX 3 C-L/CX 3 C-R is suspected to play a role in renal fibrogenesis. The aim of this study was to investigate their function in an animal model of slowly progressive chronic renal failure. Methods. Functional data were analysed in folic acid nephropathy (FAN) at different time points (up to day 142 after induction). Immunostaining for CX 3 C-L, CD3, S100A4, collagen type I, fibronectin, alpha-smooth muscle actin, Tamm-horsfall protein, aquaporin 1 and 2 as well as quantitative real-time PCR (qRT – PCR) for CX 3 C-L, CX 3 C-R and fibroblast-specific protein 1 (FSP-1) were performed. Additionally, regulatory mechanisms and functional activity of CX 3 C-L in murine proximal and distal tubular epithelial cells as well as in fibroblasts were investigated. 3.4-fold at day 7 further increasing up to 7.1-fold at day 106. The expression of mRNA CX 3 C-L correlated well with CX 3 C-R ( R 2 = 0.96), the number of infiltrating CD3+ cells ( R 2 = 0.60) and the degree of tubulointerstitial fibrosis ( R 2 = 0.56) and moderately with FSP-1 ( R 2 = 0.33). Interleukin-1 β , tumour necrosis factor- α , transforming growth factor- β as well as the reactive oxygen species (ROS) H 2 O 2 were identified by qRT – PCR as inductors of CX 3 C-L/fractalkine (FKN) in tubular epithelial cells. Functionally, CX 3 C-L/FKN chemoattracts peripheral blood mononuclear cells, activates several aspects of fibrogenesis and induces the mitogen-activated protein kinases in renal fibroblasts. Conclusions. In FAN, there is a good correlation between the expression of CX 3 C-L with markers of interstitial inflammation and fibrosis which may result from upregulation by pro-inflammatory and pro-fibrotic cytokines as well as by ROS in tubular epithelial cells. The FKN system may promote renal inflammation and renal fibrogenesis. Quantitative real-time PCR from human kidney biopsies The tubulointerstitial mRNA expression of CX 3 C-L/FKN was examined after the microdissection of kidney biopsy specimens from patients with tubulointerstitial fibrosis (fibrosis; n = 11) using quantitative real-time PCR and biopsy material from the European Renal cDNA Bank-Kroener Fresenius Biopsiebank and non-fibrotic, non-inflammatory biopsy speci- mens dedicated to variable primary or secondary nephropathies (no fibrosis; n = 9) as described previously [7]. Moreover, the acquired data were correlated with proteinuria. was found upregulated within tubular epithelial cells in phases characterized by interstitial inflammation (D) as well as in the phase of tubulointerstitial fibrosis (E and F). In addition, increased expression could be found within the mesangium, glomerular capillaries, BC and within the interstitium. G – 3 show the negative control using goat/rabbit IgG as the primary antibody. In FAN, we found a co-expression of CX 3 C-L/FKN (H1 – K1, red staining signal) with AQ1 (H, small arrow) but also tubular segments with exclusive CX 3 C-L/FKN expression and, to a lesser extent, a co-expression within AQ2-expressing tubules (I3, small arrow). Moreover, some tubular segments sparsely co-expressed CX 3 C-L/FKN and THP (J3), whereas others showed positivity for CX 3 C-L/FKN alone (small In K and L , we found a co-expression of α -sm actin or S100A4, respectively, and CX 3 C-L/FKN in selected tubular epithelial cells (small interstitial cells (K After immunohistochemical staining of kidneys' expression of tubular CX 3 C-L/FKN ( was evaluated semiquantitatively as described in the AbstractBackground. The role of the kallikrein – kinin system in diabetic nephropathy remains controversial. Methods and Results. High-glucose (HG) super-induced interleukin (IL)-6, CCL-2, transforming growth factor (TGF)- β , vascular endothelial growth factor (VEGF) and B 2 K receptor (B 2 KR) mRNA in cultured proximal tubular epithelial cells (PTEC), whereas bradykinin (BK) upregulated IL-6, CCL-2 and TGF- β mRNA. HG activated mitogen-activated protein kinase (MAPK) p42/p44 and protein kinase C (PKC) signals, whereas BK only activated MAPK. Tubular expression of these mediators and tissue kallikrein 1 (KLK1) was confirmed in human diabetic kidney biopsies. Inhibition of MAPK p42/p44 by PD98059 partially reduced HG and BK induction of IL-6, CCL-2 and TGF- β , whereas inhibition of PKC by staurosporine partially reduced HG- but not BK-induced overexpression of these cytokines and that of VEGF. Staurosporine and PD98059 synergistically reduced the effect of HG on IL-6, CCL-2 and TGF- β expression. The B 2 KR blocker, icatibant, downregulated BK- and HG-induced MAPK p42/ p44 but not HG-induced PKC activation and partially reduced both HG- and BK-induced IL-6, CCL-2 and TGF- β secretion. HG stimulated expression of KLK1 and low-molecular-weight kininogen (LMWK) and its downstream effects were attenuated by aprotinin (tissue kallikrein inhibitor). The peroxisome proliferator-activated receptor- γ (PPAR- γ ) agonist, rosiglitazone, attenuated HG-induced PKC but not HG- or BK- induced MAPK p42/44 activation and reduced HG-stimulated VEGF, along with IL-6, CCL-2 and TGF- β secretion. Rosiglitazone plus icatibant further reduced these effects of HG. Conclusions. In conclusion, HG stimulates tubular proinflammatory, profibrotic and angiogenic signals, which is partly mediated through BK via MAPK signalling and partly through PKC independent of BK. The potential ther-apeutic role of complementary B 2 KR blockade and PPAR- γ activation deserves clinical investigation.


Introduction
Tubulointerstitial fibrosis is a key component of the final common pathway in progressive chronic kidney disease and its extent is closely associated with the loss of renal function [1,2]. Whereas the early phases are potentially reversible if the underlying cause can be controlled, the later phases are often characterized by autonomous progression despite apparent resolution of the underlying disease [3,4].
In the past, there have been several reports on the role of chemokines and their receptors in progressive renal failure [5]. For example, in the induction phase of interstitial inflammation, chemokines regulate the influx of inflammatory cells into the tubulointerstitial space. However, chemokines do not only mediate interstitial inflammation but may have direct pro-fibrotic effects through receptors on parenchymal cells [6].
Thus, we chose to investigate the time course of CX 3 C-L/FKN expression in the chronic folic acid nephropathy (FAN) mouse model, analysing at the same time the expression of markers of cellular infiltration or tubulointerstitial f ibrosis. Folic acid induces dose-dependent nephrotoxicity in mice and rats, with the rapid appearance of folic acid crystals within renal tubules and subsequent acute tubular necrosis, followed by epithelial regeneration. Progressive renal cortical scarring leading to end-stage renal failure has been described in selected mice [30,31]. Furthermore, we wanted to identify the regulation and functional activity of CX 3 C-L/FKN expression in tubular epithelial cells and renal fibroblasts.

Experimental model
Animal protocols were approved by the local animals committee (#509.42505/01-04.99) and were conducted in conformity with the Guiding Principles in the Care and Use of Animals. Forty-day-old female CD1 mice were administered folic acid (200 mg/kg body weight) in vehicle (150 mM NaHCO 3 ; pH 7.4) or vehicle-only by a single intraperitoneal injection. This folic acid dose reliably induces severe nephrotoxicity, as assessed by histological finding of grossly flattened renal epithelia after 72 h [31]. The mice were killed under CO-narcosis and kidneys were harvested at days 3,5,7,14,21,56,84,112 and 142 with three to five FAN (the experimental groups) and two vehicle (the sham groups) animals at each time point. Mice were sacrificed by neck fracture and kidneys were removed within minutes. Left kidneys were used for routine histology and immunohistochemistry (IH)/immunofluorescence; right kidneys were used for real-time RT-PCR. In the latter case, organs were snap-frozen in liquid nitrogen. Proteinuria and blood urea nitrogen (BUN) were measured by standard routine methods.

Cell culture
The mouse proximal (MCT) and distal (NP-1) tubular epithelial cells, the renal fibroblast cell lines (Tk173, Tk188) as well as primary renal fibroblasts (Tk461) have been described previously [7,32,33]. For experiments, cells were FCS-starved for 24 h before they were stimulated as indicated.
Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood of untreated mice by density gradient centrifugation using Ficoll Histopaque as described previously [35].
Immunohistochemical/immunofluorescence staining for CX 3 C-L/FKN, CD3, Col I, FN, α-sm actin, S100A4, AQ1/2 and THP CX 3 C-L/FKN, CD3 and S100A4 were detected in paraformaldehydefixed, paraffin-embedded tissue and Col I and FN in cryostat sections using a modified sandwich technique as described previously [7,34]. Citric acid/sodium citrate microwave pretreatment was used for antigen unmasking in paraffin sections. As primary antibodies, polyclonal goat antibody to mouse CX 3 C-L/FKN, polyclonal rabbit antibodies to CD3, S100A4, FN or Col I, respectively, or goat/rabbit IgG as a negative control were used. As secondary antibodies, HRP-conjugated secondary rabbit anti-goat antibody (CX 3 C-L/FKN, S100A4) or Dako Envision anti-rabbit ® (CD3) were used. As tertiary antibody, Dako Envision anti-rabbit ® (CX 3 C-L/FKN) was applied. AEC was used as the chromogenic substrate in paraffin sections and DAB chromogen in cryostat sections. Nuclear counterstaining was performed using haemotoxylin. Slides were mounted on glycerol-gelatin and viewed in a non-inverted microscope.
Double immunofluorescence (DIF) staining was performed according to previously published protocol [7] with double labelling of CX 3 C-L/ FKN with PE (red) on one hand and α-sm actin, AQ1, AQ2 or THP on the other hand with FITC (green).

Evaluation of sections
The relative interstitial volume was evaluated by morphometric analysis using a 10-mm 2 graticule fitted into the microscope after Masson's trichrome staining as described previously [7]. Five randomly selected cortical areas, which included glomeruli, were evaluated for each animal [35,36].
Intensity and distribution of immunohistochemical stainings for CX 3 C-L/FKN, Col I and FN were compared in consecutive sections of biopsies and compared with the mean number of interstitial CD3-positive cells of four different visual fields using a 10-mm 2 graticule fitted into the microscope. For CX 3 C-L/FKN, the different renal compartments, interstitium, endothelial cells, tubular epithelial cells, epithelial cells of Bowman's capsule (BC), mesangial cells as well as infiltrating leukocytes were evaluated separately and scored semiquantitatively (0 to 3+) as described previously [7]. Interstitial deposition of Col I and FN was determined in a similar fashion.

RNA isolation and quantitative real-time PCR
Total RNA from whole kidneys or cells was extracted as described previously [7] using the phenol-guanidine isothiocyanate reagent RNA-Bee (Tel-Test, Friendswood, TX, USA). Oligo(dT)-primed reverse transcription was performed at 42°C for 50min after denaturation of the RNA at 70°C for 10min. CX 3 C-L/FKN, fibroblast-specific protein 1 (FSP-1) and GAPDH PCR of reverse-transcribed RNA (3μg) was performed using the primer set and annealing temperature indicated in Table 1  Quantitative real-time PCR from human kidney biopsies The tubulointerstitial mRNA expression of CX 3 C-L/FKN was examined after the microdissection of kidney biopsy specimens from patients with tubulointerstitial fibrosis (fibrosis; n = 11) using quantitative real-time PCR and biopsy material from the European Renal cDNA Bank-Kroener Fresenius Biopsiebank and non-fibrotic, non-inflammatory biopsy specimens dedicated to variable primary or secondary nephropathies (no fibrosis; n = 9) as described previously [7]. Moreover, the acquired data were correlated with proteinuria.
Cell motility assay FCS-starved MCT or NP-1 were left untreated (co) or stimulated with either IL-1β, TNF-α, TGF-β (each 10ng/mL) or 0.1mM H 2 O 2 for 48 h. Thereafter, medium was replaced and, after another 24 h, supernatants were stored. Cell motility assay was assessed by a migration assay as described previously [7]. In the upper chamber, 4 × 10 4 PBMCs per well were added in IMDM and the supernatants from the cell culture experiments were used as chemotactic agent in the presence or absence of CX 3 C-L/FKN neutralizing antibody (1:200) in the lower chamber.

Analysis of aspects of renal fibrogenesis
Renal fibroblasts were stimulated with CX 3 C-L/FKN and several aspects of renal fibrosis were investigated. This included analyses of cell viability by cell count, bromodeoxyuridine (BrdU) incorporation, FACS analysis for annexin V and propidium iodide, zymographic detection for matrix metalloproteinase 2 (MMP-2) and 9 (MMP-9), ELISA for FN and Col I and PCR for α-sm actin. These methods were described in detail in our previous study [4].

Statistical analysis
Values are shown as the mean ± standard deviation (SD). Statistical analysis was carried out as indicated in the text using the Statistica programme version 7.1 (StatSoft, Tulsa, USA). Results with levels of P < 0.05 were considered significant.

Time course of renal function, proteinuria and tubulointerstitial fibrosis in FAN
Functional data of BUN, proteinuria and the degree of renal fibrosis in sham-treated mice and FAN are summarized in Figure 1.
FSP-1 is an S100A4 protein constitutively expressed in the cytoplasm of tissue fibroblasts [37]. It identifies fibroblasts and tubular epithelium undergoing epithelial-mesenchymal transition (EMT) and is critically related to the progression of renal diseases [38]. Thus, we investigated the time course of this protein in our model. In sham-treated animals, the mRNA quotient of FSP-1/GAPDH was 12.8 ± 4.5 (× 10e − 8 copies). In FAN, the ratio was increased, peaking at day 14 [44.3 ± 30.1 (× 10e − 8 copies); 3.6-fold compared to sham]. Thereafter, FSP-1/GAPDH decreased, but was slightly elevated at the end of the observation period [day 142 = 27.0 ± 28.1 (× 10e − 8 copies); 2.0-fold compared to sham]. Figure 2C summarizes the time course of the FSP-1/GAPDH ratios.

Expression of CX 3 C-L/FKN in human kidney-derived fibrotic and non-fibrotic nephropathies
In order to further corroborate the relevance of the data obtained in the animal model, we additionally analysed the tubulointerstitial CX 3 C-L/FKN expression in human biopsies. The total expression of CX 3 C-L/FKN mRNA was significantly upregulated in biopsies from fibrotic kidneys compared to non-fibrotic nephropathies (1.6-fold) by quantitative real-time PCR (P = 0.041 comparing no fibro- sis vs fibrosis) ( Figure 3). There was only a very minor correlation between the mRNA quotient CX 3 C-L/GAPDH and the daily amount of proteinuria (R 2 = 0.3483).

Immunohistochemical expression and distribution of CX 3 C-L/FKN
In sham-treated animals, CX 3 C-L/FKN was constitutively expressed in the endothelial cell layer of arterioles and arteries ( Figure 4A). No additional cellular expression of CX 3 C-L/FKN could be detected. To confirm the specificity of CX 3 C-L/FKN immunostaining, we used a rabbit isotype control antibody, which did not show any immunoreactivity ( Figure 4B).   In FAN animals, we found a de novo expression of CX 3 C-L/FKN within the peritubular as well as glomerular capillaries, in mesangial cells and predominantly in tubular epithelial cells. In endothelial cells of arteries or arterioles, CX 3 C-L/FKN was upregulated, peaking at day 14, although a low-degree constitutive expression was detectable even at day 3. In tubular epithelial cells, CX 3 C-L/FKN upregulation persisted to the end of the observation period. Figure 4C-F displays the representative expression of CX 3 C-L/FKN at different stages of FAN. In the interstitium, CX 3 C-L/FKN was also discretely upregulated throughout the study period, whereas the cellular source could not be further defined.
In DIF staining, we found co-expression of CX 3 C-L/ FKN with AQ1 ( Figure 4H) and, to a lesser extent, AQ2 ( Figure 4I), but sparsely with THP ( Figure 4J). Moreover, in progressive stages, we found a co-expression of CX 3 C-L/FKN and α-sm actin or S100A4 in selected tubular ep- . CX 3 C-L/FKN was analysed by IH. B displays the negative control. In FAN, de novo expression of CX 3 C-L/FKN within peritubular capillaries was detectable even at day 3 (C) lasting throughout the phase of tubulointerstitial fibrosis (F). CX 3 C-L/FKN was found upregulated within tubular epithelial cells in phases characterized by interstitial inflammation (D) as well as in the phase of tubulointerstitial fibrosis (E and F). In addition, increased expression could be found within the mesangium, glomerular capillaries, BC and within the interstitium. G1-3 show the negative control using goat/rabbit IgG as the primary antibody. In FAN, we found a co-expression of CX 3 C-L/FKN (H1-K1, red staining signal) with AQ1 (H, small arrow) but also tubular segments with exclusive CX 3 C-L/FKN expression and, to a lesser extent, a co-expression within AQ2-expressing tubules (I3, small arrow). Moreover, some tubular segments sparsely co-expressed CX 3 C-L/FKN and THP (J3), whereas others showed positivity for CX 3 C-L/FKN alone (small arrow). In K and L, we found a co-expression of α-sm actin or S100A4, respectively, and CX 3  ithelial cells, respectively ( Figure 4K and L). In tubular epithelial cells, we found robustly increased expression peaking twice at day 5 as well as at day 84, remaining elevated until the end of the observation period. Figure 4M summarizes the semiquantitative analysis of the time course within the tubules.
Immunohistochemical expression of CD3, S100A4, FN and Col I The extent of renal interstitial inflammation was studied by staining for CD3 as an exemplary marker for CX 3 C-R-expressing inflammatory cells. In sham-treated animals, no relevant infiltration with CD3+ cells could be detected.
In FAN, we found increased amounts of CD3+ cells starting at day 3 with increasing numbers until day 21 and remaining elevated throughout the observation period. A representative example is shown in Figure 5A and B. S100A4 belongs to the S100 family of proteins and has been implicated in the progression of fibrosis. Several of the S100A4 effects described resemble the processes that occur during EMT [39]. In sham-treated animals, no relevant S100A4 expression was found ( Figure 5C). In FAN, we detected an upregulation within tubular epithelial and interstitial cells ( Figure 5D).
Col I and FN have been shown to be two predominating components that are deposited within the interstitial extracellular matrix during renal fibrosis [3]. In sham-injected animals, there were only negligible amounts of FN or Col I detectable within the interstitium. Conversely, in FAN animals, we found an increase of FN within the interstitial space beginning as early at day 3 with increasing amounts through the observation period. Col I deposition was upregulated first at day 5, peaking at day 84 and remaining elevated until day 142. However, although FN and Col I expression tended to increase in FAN, statistical significance (P < 0.05 compared to sham) was achieved for Col I only on day 84. Figure 5E-H shows representative examples and Figure 6A and B summarizes the results of semiquantitative analyses after IH in sham and FAN animals.
Expression of CX 3 C-L/FKN, CX 3 C-R or FSP-1 correlates poorly with proteinuria, but highly with tubulointerstitial fibrosis and each other Next, we correlated the number of copies of mRNA ratios of CX 3 C-L/GAPDH and CX 3 C-R/GAPDH. Not unexpected, the expression highly correlated with R 2 = 0.9627. The number of mRNA copies of CX 3 C-L/GAPDH correlated well with the degree of interstitial fibrosis (R 2 = 0.5576) and modestly with the mRNA FSP-1/GAPDH (R 2 = 0.3317). We found a good correlation with an R 2 -value of 0.5986 after comparison of mRNA CX 3 C-L/GAPDH and the sum of CD3 positively stained cells.
Comparing the amount of proteinuria with the copies of mRNA CX 3 C-L/GAPDH demonstrated a poor correlation with R 2 = 0.101. CX 3 C-L/FKN is upregulated in response to pro-inflammatory and pro-fibrotic cytokines as well as the reactive oxygen species H 2 O 2 To examine whether CX 3 C-L/FKN could be upregulated by pro-inflammatory or pro-fibrotic cytokines in tubular epithelial cells, FCS-starved proximal (MCT) or distal (NP-1) tubular epithelial cells were left untreated (co) or treated for 24 h with 10 ng/mL of either the proinflammatory cytokines IL-1β or TNF-α or the pro-fibrotic cytokines PDGF, epidermal growth factor (EGF), basic FGF-2 or TGF-β. Additionally, H 2 O 2 (0.1 mM) was used as a reactive oxygen species (ROS) stimulus under the same conditions.

CX 3 C-L/FKN upregulates CX 3 C-L/FKN expression and has modest effects on renal fibrogenesis in renal fibroblasts
To further analyse the relevance of the CX 3 C-L/CX 3 C-R pair, we investigated the effects of CX 3 C-L/FKN on CX 3 C-R-expressing renal fibroblasts [7] regarding potential autoinduction and aspects of fibrogenesis. CX 3 C-L/FKN-stimulated Tk173 and 188 fibroblasts dose-dependently increased the CX 3 C-L/GAPDH mRNA levels with a maximum of 312.6 ± 61.8% in Tk173 and 335.8 ± 130.4% in Tk188 (both n = 3, P < 0.01) ( Figure 9A).
In a first set of functional experiments, we observed a dose-dependent effect of increasing doses of CX 3 C-L/ FKN on cell growth. Counted cells increased up to 243 ± 59% of controls after 48 h stimulation with 20 ng/mL CX 3 C-L/FKN in Tk461, but only modestly with 125 ± 5% in Tk173 and 127 ± 11% in Tk188 fibroblasts (all n = 3, P < 0.05; not shown). We further analysed whether this effect is due to changes in cell proliferation, apoptosis or necrosis rates. The proliferation rate increased dosedependently up to a maximum of 160 ± 28% (n = 3, P < 0.01) in Tk461, 231 ± 92 in Tk173 and 224 ± 30% in Tk188 fibroblasts after stimulation with CX 3 C-L/FKN 20 ng/mL (all n = 3, P < 0.01; Figure 9B). Conversely, the apoptosis rate was decreased to 59 ± 38% in Tk461, 71 ± 8% in Tk173 and unchanged with 120 ± 35% in Tk188 (all n = 5; P < 0.05 for Tk461 and Tk173, not shown) as evaluated by FACS analysis for annexin V. Although necrosis rates tended to be lower, it failed to reach significance in all three cell types (n = 5; n.s.).
To further characterize the effects of CX 3 C-L/FKN on renal fibroblasts, Tk461 were stimulated with increasing, physiologically relevant doses of CX 3 C-L/FKN and activity of MMP-2 and MMP-9 were determined by zymography. MMP-9, but not MMP-2, was significantly induced by CX 3 C-L/FKN up to 157 ± 41% of the control (n = 3, P < 0.05). Determination of FN and Col I synthesis in CX 3 C-L/FKN-stimulated fibroblasts revealed no significant change (not shown).
Activation of fibroblasts to myofibroblasts is accompanied by the expression of α-sm actin. To analyse the influence of CX 3 C-L/FKN on this aspect, we stimulated Tk461 dose-dependently with CX 3 C-L/FKN as described above, and α-sm actin expression was detected by RT-PCR. However, we found no effects in any investigated dose ( Figure 9C).

Discussion
After the inhibition of CX 3 C-L/FKN or its corresponding receptors, respectively, a significant reduction of infiltrating sites and a reduction of irreversible, chronic tubulointerstitial damage has been achieved in several models of progressive glomerular and tubulointerstitial diseases [8,13,28]. However, no data exist on the time course of CX 3 C-L/FKN expression in progressive tubulointerstitial disease and the role of tubular CX 3 C-L/FKN expression. We detected the expression of CX 3 C-L/FKN, CD3, S100A4, Col I, FN, AQ1, AQ2, THP and α-sm actin by immunostaining in FAN as well as CX 3 C-L/FKN, CX 3 C-R and FSP-1 by qRT-PCR. Our follow-up investigations lasted from day 3 to 142 after induction. However, our study had some limitations as we were not able to detect monocytes/macrophages and CX 3 C-R by immunostaining despite using different sample processing methods, staining protocols, antigens or antibodies. Moreover, the earliest stadium investigated was day 3 after induction as we focused our investigations on the progression phase of tubulointerstitial fibrosis.
In IH, we found that CX 3 C-L/FKN was constitutively expressed within the endothelial cell layer of arteries and arterioles in a low amount, but not within the other compartments. This is in concordance with a previously published mouse model in which anti-CX 3 C-L/FKN immunoreactivity was detected mainly on endothelial cells throughout the kidney in sham-operated mice [8]. In FAN, we found a rapid de novo CX 3 C-L/FKN expression within the microvascular endothelial cell layer, suggesting a role in attracting inflammatory cells into the damaged tissue [40]. This is in accordance with previous in vitro [41] and in vivo data on renal inflammation [9,29] and further supported by the fact that CX 3 C-R-expressing lymphocytes [7] were found to correlate very well with the amount of mRNA CX 3 C-L/FKN in our model. Next, we found an upregulation within the tubules, peaking first at day 5 which lasted until the end of the observation period at day 142. So far, previous animal studies focused on early changes in expression with the longest observation period of 14 days after induction [8]. In DIF, we found a double labelling with AQ1 and, to a lesser extent, AQ2, but only sparsely with THP, suggesting a CX 3 C-L/FKN expression predominantly in proximal tubules. Several studies demonstrated an upregulation of CX 3 C-L/FKN within the tubules in inflammatory diseases, e.g. acute tubulointerstitial rejection in humans [14,42]. Otherwise, and in accordance with our observation, an upregulation has also been observed in slow-declining progressive fibrotic and sclerotic renal and non-renal diseases, e.g. chronic allograft nephropathy [43], prolonged mesangial proliferative glomerulonephritis [44], chronic pancreatitis [45], pulmonary hypertension [46], chronic liver injury [47], systemic sclerosis [40] and atherosclerosis [24]. Corroborating the hypothesis of an involvement of the CX 3 C-L/CX 3 C-R system in renal fibrosis, we found an expression of CX 3 C-L/FKN in tubular epithelial cells putatively undergoing tubular epithelial transition, demonstrated by a double labelling with the mesenchymal marker α-sm actin or S100A4, respectively, and, moreover, by our findings in human fibrotic and non-fibrotic nephropathies. In previous studies, tubular upregulation of CX 3 C-L/FKN was induced by protein overload [15], TNF-α [48] or IFNγ stimulation [49]. However, we found no good correlation with proteinuria in FAN and only a minor correlation in humans. Otherwise, our in vitro analyses identified the pro-inflammatory cytokines IL-1β and TNF-α, the pro-fibrotic cytokine TGF-β as well as the ROS H 2 O 2 as the main inductors of CX 3 C-L/FKN in murine proximal and distal tubular epithelial cells in vitro. These data suggest a role of mediators of the early inflammation, but also of mediators expressed during the progression phase of renal scarring. This was not unsuspected, since ROS are known to mediate an upregulation of several chemokines in tubular epithelial cells [50,51], although that had not been described for CX 3 C-L/ FKN. In rat mesangial cells, an upregulation of CX 3 C-L/ FKN in response to the pro-fibrotic cytokines FGF-2 or PDGF has been described [20], but not to TGF-β. However, these differences may depend on the kind of investigated cells and species. In our previous study, we have shown an upregulation of CX 3 C-R in human tubular epithelial cells as well as a de novo expression in renal fibroblasts in progressive renal diseases mediated by oxidative stress. Thus, we chose this fibroblast model to further investigate potential profibrotic effects downstream [7]. Stimulation of CX 3 C-Rexpressing renal fibroblasts with CX 3 C-L/FKN enhanced their migration in vitro [7], which might be provided by an upregulation of matrix metalloproteinase. Moreover, we found effects on cell viability and an MEK-1-/ERK-1/2dependent autoinduction, but not on matrix synthesis or transformation to myofibroblasts. Furthermore, tubular CX 3 C-L/FKN caused an enhanced migration of PBMCs in the migration assay. Taken together with this new data, we speculate that the CX 3 C-L/CX 3 C-R system may be an autoinductive system upregulated in progressive renal diseases which may enhance fibrosis by an increased invasion of mononuclear cells and also by an increased migration and proliferation of fibroblasts to the place of injury. This hypothesis is corroborated by several previous findings: (1) Lu and co-workers have demonstrated that blockade of CX 3 C-R failed to block serum creatinine increase in a cisplatin-induced acute renal failure model within the early phase (day 3 after induction) but without investigation of later stadiums [52], (2) Furuichi and colleagues have shown that CX 3 C-L/FKN-dependent fibrosis was not an early but a late event during an ischaemia-reperfusion injury model of mouse kidneys [8], (3) Feng et al. have demonstrated that CX 3 C-L/FKN inhibition was less effective at the early stage than at the late stage in crescentic glomerulonephritis [28] as it ameliorated the progression of lupus nephritis in MRL/lpr mice [29]. the national meeting of the German Society of Nephrology in September 2006 in Essen, Germany. The authors thank Mrs. S. Woock, A. Krüger and F. Agdas for the excellent technical assistance and cand. med. A. Imling for the IH staining for CX 3 C-L and CD3 and trichrome staining. This work was supported in part by the grant "Expression, Funktion und Regulation des Chemokins CX 3 C-L und seines Rezeptors CX 3 C-R bei Entzündungsprozessen der Niere" from the Georg-August-University to M.J.K. (Forschungsförderungsprogramm).
Conflict of interest statement. None declared.