Fractalkine expression on human renal tubular epithelial cells: potential role in mononuclear cell adhesion

Abstract

Fractalkine (CX₃CL1) is a transmembrane molecule with a CX₃C chemokine domain attached to an extracellular mucin stalk which can induce both adhesion and migration of leucocytes. Mononuclear cell infiltration at renal tubular sites and associated tubular epithelial cell damage are key events during acute renal inflammation following renal allograft transplantation. Using northern and Western blot analysis, we have demonstrated the expression of fractalkine message and protein by renal tubular epithelial cells in vitro. The expression was up-regulated by TNF-α, a key proinflammatory cytokine in acute rejection. Investigation of surface expression of fractalkine on cultured proximal tubular epithelial cells revealed only a subpopulation of positively staining cells. Immunohistochemistry revealed that only a proportion of tubules in renal allograft biopsies showed induction of fractalkine expression. Studies using a static model of adhesion demonstrated CX₃CR1/fractalkine interactions accounted for 26% of monocytic THP-1 cell and 17% of peripheral blood natural killer cell adhesion to tubular epithelial cells, suggesting that fractalkine may have a functional role in leucocyte adhesion and retention, at selected tubular sites in acute renal inflammation. Thus, fractalkine blockade strategies could reduce mononuclear cell mediated tubular damage and improve graft survival following kidney transplantation.

INTRODUCTION

Fractalkine (FKN, CX₃CL1), the only described CX₃C chemokine, was originally detected in human umbilical vein endothelial cells (HUVEC) [1] and more recently in lung microvascular endothelial cells (LMVEC) [2]. A more ubiquitous expression is now recognized, including neuronal astrocytes [3], cardiac myocytes [4], intestinal epithelial cells [5] and dendritic cells [6,7]. Low levels of fractalkine mRNA have been detected in a wide range of normal human tissues, including kidney [1,8].

Fractalkine is a glycoprotein which can exist either anchored to the cell membrane or as a soluble form generated by proteolytic cleavage from the membrane [1]. It can function as both a potent chemoattractant molecule (soluble form) [1][9–11], and an adhesion molecule (membrane-anchored form) for monocytes, NK cells and subsets of CD8+ T cells [12–16]. Thus, it not only promotes chemotaxis of target leucocytes but also has the potential to function as an adhesion molecule to mediate capture and firm adhesion of circulating leucocytes [14,17,18]. In addition, unlike other chemokines identified to date, fractalkine binds to a specific receptor, CX₃CR1, that is not shared by any other chemokine [12].

In vitro, TNF-α, IL-1 and LPS stimulate fractalkine expression on human endothelial cells including HUVEC [1,11,15–17] and epithelial cells derived from the lung vasculature [2] and the intestine [5]. Fractalkine is inducible in other cell types such as neuronal astrocytes [3], intestinal epithelial cells [5] and bone marrow derived dendritic cells [7]. In all cases, constitutive expression of fractalkine is low or absent.

Proximal tubular epithelial cells (PTEC) are a prolific source of a variety of inflammatory mediators, including cytokines and chemokines. Here, we provide evidence to demonstrate that fractalkine message and protein expression by PTEC is modulated by TNF-α, a key pro-inflammatory cytokine in renal inflammatory conditions [19,20]. Further, in vitro adhesion studies demonstrates the ability of surface expressed fractalkine, on PTEC, to support the adhesion of human peripheral blood NK cells and THP-1 cells of the monocyte lineage, which express the receptor for fractalkine. Thus, our observations suggest proximal tubule-derived fractalkine may have a role in the recruitment and retention of leucocytes to the interstitium during renal inflammatory responses.

MATERIALS AND METHODS

Antibodies and reagents

Tissue-expressed fractalkine was determined using a polyclonal goat antihuman fractalkine antibody (R & D Systems Europe Ltd, Oxon, UK), a biotinylated rabbit antigoat secondary antibody (R & D Systems Europe Ltd) and a streptavidin ABC/HRP complex (DAKO Ltd, Ely, UK). The polyclonal antifractalkine antibody was also used for Western blot analysis. The fractalkine receptor was detected using a fractalkine-SEAP fusion protein (a gift from Millennium Pharmaceuticals Inc., Boston, USA), a polyclonal goat anti-SEAP secondary antibody (Binding Site, Birmingham, UK) and a FITC conjugated antigoat detection antibody (Binding Site). Surface expression of fractalkine, ICAM-1 and VCAM-1 on PTEC was detected using FACs analysis with goat antihuman fractalkine (R & D Systems Europe Ltd), mouse antihuman ICAM-1 (CD54) (DAKO Ltd) and mouse antihuman VCAM-1 (CD106) (DAKO Ltd) antibodies, respectively. Adhesion blocking studies were done using recombinant fractalkine (R & D Systems Europe Ltd), mouse antihuman VLA-4 (CD49d) (Max68; gift from Martyn Robinson, Celltech, Slough, UK) and mouse antihuman LFA-1 (CD11a, CD18) (DAKO Ltd). CD56-positive NK cells were isolated using a mouse antihuman CD56 antibody (DAKO Ltd) and Dynabeads sheep antihuman IgG kit (Dynal Biotech Ltd, Bromborough, UK).

Immunohistochemistry

Cryostat sections of renal biopsy specimens from patients with acute cellular allograft rejection (scoring 4(IA) to 4(IIA) on the 1997 Banff diagnostic categorization system [21] were used in these studies. Control tissue comprised cortical fragments from the unaffected pole of kidneys removed for renal cell carcinoma. Endogenous peroxidase activity was blocked with tris buffered saline (TBS pH 7·4) containing 0·3% H₂O₂ and 0·1% NaN₃, for 10 min This was followed by sequential treatment with 0·1% avidin and 0·01% biotin to block endogenous biotin and 10% rabbit serum. Three stage indirect immunostaining was performed with a primary goat polyclonal antihuman fractalkine antibody at 15 μg/ml. The specificity of this antibody was confirmed and the optimal working concentration was determined previously [22]. On control sections the primary antibody was substituted with preimmune serum. This was followed sequentially by a biotinylated rabbit antigoat IgG at 1:400 and then HRP conjugated streptavidin ABC complex. Binding was visualized by the addition of 3,3′-diaminobenzidine (DAB) (Vector Laboratories Ltd, Peterborough, UK), the sections counterstained with haematoxylin and mounted in dibutyl polystyrene xylene (DPX; Merck Ltd, Lutterworth, UK).

Cells and cell culture

All primary PTEC cultures were grown in serum-free PTEC growth medium which consisted of DMEM:Nutrient Ham's F₁₂ 1:1 medium supplemented with 5μg/ml insulin, 5μg/ml transferrin, 5μg/ml sodium selenite, 36ng/ml hydrocortisone, 4pg/ml tri-iodothyronine, 10ng/ml epidermal growth factor, 2 mm l-glutamine, 1000U/ml penicillin and 1000μg/ml streptomycin as described previously [23]. PTEC used in all further in vitro studies were maintained in PTEC growth medium and were from passages two to five. THP-1 cells, a monocytic cell line (a gift from Anthony Lammas, University of Birmingham) were cultured in RPMI-1640 medium (Sigma-Aldrich Co Ltd, Gillingham, UK) containing 10% foetal calf serum (FCS). CD56 positive natural killer (NK) cells were isolated from blood as previously described [24] by positive selection using immunomagnetic beads.

Positive selection of CD56-positive NK cells

Peripheral blood mononuclear cells (PBMC) were isolated from heparinized blood collected from normal volunteers by centrifugation over Ficoll/Hypaque gradient. PBMC were suspended in phosphate buffered saline (PBS; pH 7·4) containing 1% FCS at a density of 2 × 10⁸ cells per ml and incubated with 15 μg/ml mouse antihuman CD56 at 4°C for 45min The cells were subsequently washed in ice-cold PBS/1% FCS, resuspended at 2 × 10⁷ cells per ml and incubated with Dynabeads pan mouse IgG (Dynal Biotech Ltd) at a concentration of 2 × 10⁷ beads per ml at 4°C for 30min The rosetted cells were recovered using a Dynal MPC magnet. The CD56-positive NK cells were recovered by detaching the beads from the cells using a releasing buffer (Dynal Biotech Ltd) as per manufacturer's instructions. Purity of CD56-positive cells were verified by flow cytometry.

Northern blotting

Confluent monolayers of PTEC in 75 cm² tissue culture flasks were incubated for 3, 6 and 9h in the absence and presence of 10ng/ml tumour necrosis factor-α (TNF-α) (Sigma-Aldrich Co Ltd); in PTEC growth medium. It was necessary to maintain PTEC under optimal growth conditions during TNF-α stimulation since these cells do not survive for extended lengths of time in the absence of growth supplements. These conditions were in line with other studies of chemokine production by tubular epithelial cells in vitro[25–28].

Peak serum TNF-α levels of 2·5ng/ml have been reported during acute cellular renal allograft rejection [29], however, tissue levels of TNF-α are likely to be much greater. The TNF-α concentration of 10 ng/ml was optimized during earlier studies investigating the induction of adhesion molecules and chemokines in PTEC. Total RNA was isolated from PTEC using RNAzol (Biogenesis Ltd, Poole, UK) according to manufacturer's instructions. 10 μg of RNA was denatured, electrophoresed on a 1·2% denaturing agarose gel and transferred to Genescreen plus nylon membrane (NEN Life Science Products, Boston, USA) and the RNA immobilized to the membrane by exposure to UV light.

A 443 base pair fractalkine cDNA insert was isolated from a PGEM-7Zf + vector cloned in our laboratory [22]. Similarly, a 1078 base pair cDNA fragment coding for the human 18S ribosomal RNA was isolated from a pBS-KS + based vector (gift from Dr Yuen-Ling Chan, University of Chicago, USA) [30] as a template for the control probe. ³²P-labelled probe was synthesized using a random primed DNA labelling kit (Boehringer Mannheim, Lewes, UK), according to manufacturer's instructions. The probes used in our studies typically had counts of 1 × 10⁹ cpm per μg.

Following standard prehybridization treatment, the RNA blot was incubated with the ³²P-labelled probe at a final concentration of 6 × 10⁵ cpm per ml and incubated with the membrane at 65°C for 16 h. The membranes were washed and exposed to a phosphorimager screen. A 3·5 kilobase fractalkine and a 1·9 kilobase 18S transcript were visualized using a phosphorimager (Molecular Dynamics, Sunnyvale, CA, USA). Owing to the differences in the relative sizes and amounts of the two transcripts there was no need to strip the membrane providing the membrane was probed first for fractalkine and then for 18S ribosomal RNA.

Western blot analysis of fractalkine protein expression in PTEC

PTEC grown to confluence were treated with TNF-α (10ng/ml) as described above for 6, 12 and 18 h. Following the specified incubation period, the cells were detached by incubation in cell dissociation solution (Sigma-Aldrich Co Ltd) at 37°C for 30 min The detached cells were solubilized in cell lysis buffer containing 20 mm Hepes pH 7·2, 2·5 mm EGTA, 2 mm EDTA, 1 mm NaCl, 5 mm MgCl² and 1% Triton X-100. The protein concentration was estimated using the Bio-Rad protein assay reagent (Bio-Ra-Laboratories, Hemel Hempstead, UK). The whole-cell lysates were mixed with an equal volume of Laemmli sample buffer containing 200 mm dithiothreitol and 4 mm phenylmethylsulphonyl fluoride and boiled for 10 min. Next 20 μg protein from each sample was loaded on a 12% sodium dodecyl sulphate/polyacrylamide gel, and 25 ng recombinant fractalkine and rainbow marker (Amersham Pharmacia Biotech UK, Little Chalfont, UK) were loaded in parallel. The proteins were transferred to a Hybond nitrocellulose membrane (Amersham Pharmacia Biotech UK). The nitrocellulose membrane was blocked with TBS pH 7·4 containing 1% BSA and 0·1% tween 20 and sequentially probed with a goat antihuman fractalkine antibody (same as for immunostaining) at 0·2 μg/ml overnight at room temperature followed by a horseradish peroxidase conjugated donkey antisheep/goat IgG (Binding site) at a 1/11000 dilution for one hour. Positive bands were visualized using enhanced chemiluminescence (Amersham Pharmacia Biotech UK) and autoradiography. The specificity of the antifractalkine antibody was demonstrated by absorption of the antifractalkine antibody using 1 μg/ml recombinant fractalkine (R & D Systems Europe) and our observations were verified by probing with a second rabbit polyclonal antifractalkine antibody (Millennium Pharmaceuticals) and also a mouse monoclonal antifractalkine antibody (Biosource International Inc, Nivelles, Belgium) in separate experiments.

Flow cytometry

CX₃CR1, fractalkine receptor expression on CD56-positive NK cells and THP-1 cells was determined by FACS analysis. The cells were incubated with 4 nm fractalkine-SEAP fusion protein (FKN-SEAP) or control SEAP (C-SEAP) in medium containing 1% FCS and 0·1% sodium azide (NaN₃) at room temperature for 45 min. The cells were washed wih ice cold PBS containing 1% FCS and 0·1% NaN₃ and incubated with goat anti-SEAP at a 1:10 dilution for 45 min at 4°C. Following a further wash the cells were incubated with a FITC conjugated antigoat IgG at a 1:10 dilution for 45 min at 4°C. After a final wash the FITC labelled cells were fixed with 2% paraformaldehyde and analysed. As reported in previous studies 70% of both THP-1 cells and NK cells used in our adhesion studies expressed the receptor for fractalkine [12,15].

Mouse monoclonal anti-ICAM-1 (R & D Systems Europe) and anti-VCAM-1 (R & D Systems Europe) primary antibodies followed by a FITC conjugated mouse IgG was used to determine ICAM-1 and VCAM-1 expression on the surfaces of PTEC. Similarly, mouse monoclonal anti-LFA-1 (DAKO) and anti-VLA-4 were used to determine LFA-1 and VLA-4 expression, respectively, on THP-1 cells.

The purity of the NK cells was verified using mouse anti-CD56 and FITC conjugated mouse IgG. CD56 enriched NK cell populations used in the adhesion studies were around 70% pure.

Surface expression of fractalkine on PTEC was detected using a goat antifractalkine primary antibody (same as for immunohistochemistry) and a FITC conjugated antigoat IgG.

Adhesion Studies

Adhesion of THP-1 cells and freshly isolated CD56-positive NK cells to PTEC was assessed based on methods described previously [16,31,32]. The cells suspended at 1 × 10⁶ cells per ml were incubated with 200 μCi of Chromium-51 (⁵¹Cr) (Amersham Pharmacia Biotech UK) in RPMI containing 10% FCS for 2h at 37°C. The ⁵¹Cr-labelled cells were washed and resuspended at 2 × 10⁵ cells per ml and incubated in control medium or medium containing blocking reagents for 30 min at 4°C. In the absence of any well characterized fractalkine receptor blocking antibodies, recombinant fractalkine has previously been used by various groups to block CX3CR1 [12,16] to demonstrate fractalkine mediated interactions. Initial studies were performed to determine the optimal inhibitory concentration of recombinant fractalkine required to block fractalkine mediated adhesion in our model. Optimal levels of inhibition was achieved using 1 μg/ml recombinant fractalkine, inhibition was not enhanced further using 5–10 μg/ml recombinant fractalkine. In blocking studies, THP-1 cells were pretreated with recombinant fractalkine at 1 μg/ml to block the fractalkine receptor which was completely down-regulated by soluble fractalkine at this concentration. A combination of anti-VLA-4 [33] and anti-LFA-1 [34] each at 20μg/ml were used to block leucocyte surface β1 and β2 integrins, for 30 min at 4°C. THP-1 cells were incubated with 1 μg/ml pertussis toxin for 1h at 37°C and subsequently washed prior to the adhesion assay, to block any G-protein dependent adhesion processes. Irrelevant mouse IgG was used as control.

PTEC grown to confluence in 48-well tissue culture plates were incubated in PTEC growth medium alone or supplemented with 10 ng/ml TNF-α for 18h at 37°C. The ⁵¹Cr-labelled cells were added to each well at 2 × 10⁴ cells in a volume of 100 μl and incubated for 30 min at 37°C. The nonadherent cells were removed by washing with PBS and the adherent cells lysed using 200 μl 1% Triton X-100 and the supernatant from each well was counted on a Multi-gamma II gamma counter (LKB, Turku, Finland).

The adhesion of THP-1 cells, pretreated with recombinant fractalkine at 1 μg/ml for 30 min at 4°C, to ICAM-1 and VCAM-1 was evaluated in additional studies. Briefly, 96-well tissue culture plates were coated with 100 ng/ml ICAM-1 (R & D Systems) [15] and 600 ng/ml VCAM-1 (R & D Systems Europe) [35] overnight at 4°C. The recombinant ICAM-1 and VCAM-1 were diluted in PBS supplemented with 0·1% bovine serum albumin (BSA). Control wells were coated with 0·1% BSA. The ⁵¹Cr-labelled cells were added to each well at 5 × 10⁴ cells in a volume of 100 μl [35] and incubated for 30 min at 37°C. The nonadherent cells were removed by washing with PBS and the adherent cells lysed using 1% Triton X-100 and the supernatant from each well was counted.

Data obtained as counts per minute (cpm), from quadruplicate wells from 4 different experiments were analysed and the number of adherent cells per well (± standard error of mean) were calculated as follows:

Significance of differences in adhesion between control cells and cells treated with blocking reagents was determined using Mann–Whitney one-tailed two-sample tests. P-values of 0·05 or less were taken as significant inhibition of adhesion.

RESULTS

Immunohistochemistry

Immunohistochemical analysis of biopsy material revealed only negligible diffuse fractalkine positivity in the control kidney (Fig. 1A). In contrast, analysis of biopsy material from patients with histological evidence of acute renal allograft rejection demonstrated fractalkine protein expression on tubular epithelial cells, with positive staining localizing to both basolateral and luminal membranes of the epithelial cells (Fig. 1B). We observed that fractalkine expression was only present in a proportion of tubules. In previous studies it was found that the pattern of fractalkine protein expression in the interstitium correlated with mRNA levels detected by in situ hybridization and infiltrates of monocytes in serial sections [22].

Fig. 1.

Immunohistochemical localization of fractalkine (A) in normal kidney and (B) on tubular epithelial cells in acute allograft rejection. Magnification, ×400.

Fractalkine mRNA expression in PTEC in vitro

To confirm that PTEC can express fractalkine mRNA, levels of mRNA in PTEC cultured in growth medium alone or medium supplemented with 10 ng/ml TNF-α, were monitored over a period of 9 h. Fractalkine mRNA levels in the absence of exogenous cytokine treatment were minimal. However, TNF-α induced fractalkine expression within 3h of cytokine treatment which was reduced slightly by 6h and nearly at baseline by 9h (Fig. 2A,B).

Fig. 2.

Northern blot analysis of fractalkine RNA and 18S ribosomal RNA expression by PTEC, untreated and treated with TNF-α for 3, 6 and 9 h. (A) Blot represents one of three separate experiments and (B) densitometric analysis of fractalkine/18S band density ratio from 3 experiments.

Fractalkine protein expression in PTEC in vitro

Fractalkine protein expression by PTEC and its inducibility by TNF-α was tested next. It was detectable in PTEC after 12h in the absence of cytokine treatment and remained detectable after 18h in culture. All treatments were performed in PTEC growth medium, one or more components of which may be capable of inducing basal fractalkine expression despite the absence of any cytokine treatment. However, TNF-α induced expression within 6h and further augmented expression after 12 h. By 12 h, levels were three to four fold greater after TNF-α treatment compared to untreated cells. At 18h after TNF-α treatment, expression levels remained high (Fig. 3A,B).

Fig. 3.

Western blot analysis of lysates from PTEC, untreated and treated with TNF-α for 6, 12 and 18 h, using goat polyclonal antifractalkine antibody. Relative migration of recombinant fractalkine is shown in parallel lanes. (A) Blot represents one of three separate experiments and (B) densitometric analysis of fractalkine band density from 3 experiments.

Thus, immunohistochemistry demonstrated fractalkine expression by tubular epithelial cells, localizing primarily to the cell membrane, in biopsy material obtained from patients undergoing episodes of acute rejection. The in vitro northern and Western blotting studies verified that exposure of PTEC to an inflammatory stimulus such as TNF-α, was capable of inducing enhanced levels of fractalkine message and protein. Acute renal allograft rejection is characterized by dense infiltrates in the interstitium composed of T cells and monocytes [36,37]. In addition, NK cells have been implicated in the damage of graft tubular epithelial cells [37,38]. Thus, we proceeded to investigate whether fractalkine expressed by PTEC in vitro was capable of supporting the adhesion of fractalkine receptor expressing leucocytes. To this end we examined the adhesion of THP-1 cells which represent cells of the monocyte cell lineage and freshly isolated peripheral blood NK cells, both of which strongly express the fractalkine receptor.

Adhesion Studies

Initial optimization studies indicated that treatment of PTEC with 10 ng/ml TNF-α for a time period between 12 and 18h allowed maximal levels of leucocyte adhesion to PTEC. Maximal induction of fractalkine was also observed by Western blotting (Fig. 3) during this time period. Thus, in all further adhesion studies we used untreated PTEC and PTEC treated with TNF-α for 18 h. We also examined fractalkine expression on the surface of PTEC at this time point since it is the membrane bound form of fractalkine that is involved in mediating leucocyte adhesion.

Fractalkine expression on PTEC surface

Flow cytometric analysis of PTEC revealed surface expression of fractalkine on 7%± 4·2% of total cells in untreated PTEC (mean percentage positive cells ± standard error of mean) which increased to 15%± 3·5% of total cells in TNF-α treated cells (Fig. 4A,D). Expression of fractalkine seemed to be confined to a subpopulation of TNF-α treated PTEC as had been suggested by the immunohistochemistry studies on renal allograft biopsy sections. The expression was significantly lower than the level of ICAM-1 expression (Fig. 4B,E), which was present on 99% of cells in both untreated and TNF-α treated PTEC. VCAM-1, another key adhesion molecule expressed in acute renal inflammation, was nominal in untreated PTEC and increased to 28% of cells in TNF-α treated PTEC (Fig. 4C,F).

Fig. 4.

FACS analysis of fractalkine (A, D), ICAM-1 (B, E) and VCAM-1 (C, F) expression on the surface of untreated (A-C) and TNF-α treated (D-F) PTEC. Histograms illustrate negative control (shaded) and positively stained cells (unshaded).

Since ICAM-1 and VCAM-1 are prominent adhesion molecules expressed on PTEC in acute rejection, we compared the effect of blocking their main ligands LFA-1 and VLA-4 simultaneously using blocking antibodies, to the effect of fractalkine receptor blockade using recombinant fractalkine.

THP-1 cell adhesion to PTEC

THP-1 cells adhered to both untreated and TNF-α treated PTEC. We examined the role of fractalkine in leucocyte adhesion in the presence of participation from conventional adhesion molecules of the immunoglobulin superfamily such as ICAM-1 and VCAM-1 which are also up-regulated on PTEC by TNF-α. We used a combination of anti-LFA-1 and anti-VLA-4 antibodies to inhibit ICAM-1/LFA-1 and VCAM-1/VLA–4 mediated interactions. The results from 6 separate experiments performed in quadruplicate are summarized in Fig. 5. Adhesion to TNF-α treated PTEC was significantly inhibited by up to 26% (P < 0·05) (Fig. 5B) by blocking the fractalkine receptor, CX₃CR1, using 1 μg/ml recombinant fractalkine (optimal inhibitory concentration determined in preliminary experiments). Adhesion was blocked more efficiently by a combination of anti-LFA-1 and anti-VLA-4 antibodies reaching a maximum of 57% (P < 0·005) reduction of THP-1 cell adhesion to TNF-α treated PTEC (Fig. 5B). Blockade of all three receptors on THP-1 cells, however, did not produce any further decrease in adhesion to PTEC (Fig. 5A,B). It is possible that the inhibitory effect of the recombinant fractalkine was reversed in the presence of anti-integrin antibodies. Indeed, it has been shown that under some conditions recombinant fractalkine can enhance integrin-mediated adhesion [15]. In further studies THP-1 cells were treated with 1 μg/ml pertussis toxin prior to the adhesion assay. Although 60% of THP-1 cell adhesion was inhibited following pertussis toxin treatment a significant proportion of THP-1 cell adhesion was maintained. This observation supports a role for G-protein independent pertussis toxin insensitive adhesion mechanisms such as fractalkine/CX₃CR1 interactions.

Fig. 5.

Inhibition of THP-1 cell adhesion to (A) untreated and (B) TNF-α treated PTEC using recombinant fractalkine, anti-LFA-1 mAb and anti-VLA-4 mAb, as described in Methods.

THP-1 cell adhesion to immobilized ICAM-1 and VCAM-1

In order to determine the influence of pretreating THP-1 cells with 1 μg/ml recombinant fractalkine for 30 min at 4°C on integrin function, the adhesion of THP-1 cells to immobilized ICAM-1 and VCAM-1, ligands for THP-1 cell expressed integrins LFA-1 and VLA-4, was examined. Following pretreatment with recombinant fractalkine, THP-1 cell adhesion to ICAM-1 was significantly up-regulated by 38% (P < 0·05), indicating fractalkine-mediated modulation of the β2 integrin, LFA-1 function. However, no such augmentation of THP-1 cell adhesion to VCAM-1 was observed, following treatment with recombinant fractalkine (Fig. 6).

Fig. 6.

Adhesion of untreated and 1 μg/ml recombinant fractalkine-treated THP-1 cells to immobilized ICAM-1 and VCAM-1, as described in Methods.

To determine if fractalkine altered expression levels of ICAM-1, VCAM-1 or fractalkine on the surface of PTEC, or expression of their corresponding receptors LFA-1, VLA-4 and CX3CR1 on THP-1 cells, each cell type was treated separately with 1 μg/ml recombinant fractalkine for 30 min at 4°C and analysed by FACS. There was no detectable change in the levels of ICAM-1, VCAM-1 or fractalkine on PTEC in response to recombinant fractalkine (data not shown). Similarly there was no effect of recombinant fractalkine on the levels of LFA-1 or VLA-4 on THP-1 cells (data not shown). There was, however, a marked down-regulation of CX3CR1 on THP-1 cells following treatment with recombinant fractalkine (Fig. 7).

Fig. 7.

FACS analysis of fractalkine receptor, CX3CR1 expression on the surface of untreated (A) and 1 μg/ml fractalkine-treated (B) THP-1 cells. Histograms illustrate negative control (shaded) and positively stained cells (unshaded).

NK cell adhesion to PTEC

As seen with THP-1 cells, NK cells also adhered to both untreated and TNF-α treated PTEC in vitro. Blockade of the fractalkine receptor on NK cells using 1 μg/ml recombinant fractalkine did not produce significant levels of inhibition of adhesion on untreated PTEC, reaching 11% (from 6225 ± 416–5524 ± 466 cells per well). A slightly greater and significant degree of inhibition of 17% (from 6359 ± 385–5282 ± 441 cells per well; P < 0·05) was observed on TNF-α treated PTEC. Data were obtained from quadruplicate wells from four separate experiments. Adhesion to untreated and TNF-α treated PTEC was significantly blocked using the cocktail of anti-integrin antibodies by up to 26% (P < 0·01) which was marginally greater than that achieved with fractalkine receptor blockade. When CX₃CR1, LFA-1 and VLA-4 were blocked simultaneously, no further significant increase in the inhibition of NK cell adhesion was observed.

Thus the adhesion studies using the monocytic THP-1 cells and CD56+ NK cells demonstrated that fractalkine induced on the surface of renal tubular epithelial cells in vitro was capable of supporting CX₃CR1/fractalkine mediated adhesion of specific leucocyte subsets. However, this interaction accounted for modest levels of adhesion compared to LFA-1/ICAM-1 and VLA-4/VCAM–1 interactions, which may reflect the relatively greater levels of ICAM-1 and VCAM-1 expression on PTEC.

DISCUSSION

The starting point for these studies was the observation that the CX₃C chemokine, fractalkine, was expressed by renal tubular epithelial cells during episodes of acute renal allograft rejection. This prompted in vitro studies on the cytokine inducibility of fractalkine message and protein by primary cultures of human PTEC. Northern and Western blotting studies confirmed that PTEC in culture express both fractalkine message and protein. Further, levels could be induced by the proinflammatory cytokine, TNF-α.

TNF-α has been identified in acute renal inflammation and has previously been shown to induce the expression of chemokines such as MCP-1 and RANTES either alone or in combination with IFN-γ[25], both of which have been implicated in renal inflammation. Initial in vitro studies demonstrated a role for TNF-α in the induction of fractalkine mRNA and protein in endothelial cells [1]. These studies have been complemented by more recent reports of induction of fractalkine by TNF-α and IL-1 in neuronal and epithelium-derived cells [3,5].

Immunohistochemical staining of biopsy material showed fractalkine expression on both the basal and luminal membranes of tubular epithelial cells and in some cases fractalkine was localized to epithelial cell junctions. In addition to our Western blotting studies which showed fractalkine protein expression in whole cell lysates, FACS analysis demonstrated a membrane bound form of fractalkine on the surface of the PTEC.

As acute renal allograft rejection is characterized by inflammatory cell infiltrates at tubular, perivascular, periglomerular regions of the interstitium, we wished to determine whether membrane bound fractalkine may play a role in adhesion and retention of leucocytes at tubular sites, relevant to development of acute rejection.

Thus, a static adhesion assay system [15,16,31,32] was used to investigate the ability of PTEC-expressed fractalkine to mediate leucocyte adhesion. Previously, static adhesion assay systems have demonstrated adhesion of CX₃CR1-expressing cells to immobilized fractalkine, fractalkine-transfected endothelial cells or TNF-α treated human umbilical vein endothelial cells [12–16].

In initial studies we used flow cytometry to confirm that THP-1 cells, which represent cells of the monocyte lineage and human peripheral blood NK cells, can both express the receptor for fractalkine, CX₃CR1. We also used flow cytometry to check surface expression of fractalkine on PTEC. Fractalkine expression was lower than expected and appeared to be confined to a subpopulation of TNF-α treated PTEC. This might accord with the observation that only a proportion of tubules in any one renal biopsy appeared to be fractalkine positive during acute rejection. However, at the present time we cannot say whether this reflects heterogeneity in the ability of tubular cells, either in vivo or in vitro, to respond to inflammatory cytokines. Nonetheless, the low level of surface expression of fractalkine on tubular epithelial cells may account for the modest levels on a percentage basis of fractalkine/CX₃CR1 dependent adhesion of peripheral blood NK cells and THP-1 cells that were observed to cytokine activated PTEC.

Compared to LFA-1/ICAM-1 and VLA-4/VCAM-1 dependent adhesion, the fractalkine/CX₃CR1 system appeared to confer little extra benefit. However, since fractalkine receptor blockade alone was capable of inhibiting both THP-1 cell and NK cell adhesion to TNF-α activated PTEC and, given the possible patchy nature of PTEC fractalkine expression, it is possible that the fractalkine/CX₃CR1 system mediates biologically significant recruitment or retention of leucocytes at selected tubular sites. THP-1 cells and NK cells express a range of adhesion molecules including MAC-1 [39,40] which is a ligand for ICAM-1, and VLA-6 on THP-1 cells [41] and VLA-5 on NK cells [40] which are ligands for extracellular matrix components laminin and fibronectin, respectively. These adhesion interactions may account for the inability to achieve total inhibition of adhesion by blocking CX3CR1, LFA-1 and VLA-4.

Further, no additive effect was observed with the combined blocking strategy which suggested an effect of fractalkine on integrin functionality and expression. In additional studies, treatment of THP-1 cells with recombinant fractalkine induced a marked and significant enhancement of adhesion to immobilized ICAM-1, although cell adhesion to VCAM-1 remained unchanged. This marked increase in LFA-1/ICAM–1 interactions is also reflected in the absence of an additive inhibitory effect with a combination of recombinant fractalkine and anti-integrin antibodies and in the modest inhibition of cell adhesion seen following treatment with recombinant fractalkine alone. There was however, no change in the expression of LFA-1 and VLA-4 on THP-1 cells or their PTEC-expressed ligands ICAM-1 and VCAM-1, respectively, following exposure to soluble fractalkine. Similarly there was no change in fractalkine expression on PTEC, while fractalkine receptor expression was efficiently down-regulated following treatment with soluble fractalkine.

The chemotactic function of soluble, secreted fractalkine from renal tubular epithelial cells was also evaluated in our laboratory. These studies demonstrated that cytokine-treated PTEC supernatant induced migration of IL-2 activated peripheral blood lymphocytes in an in vitro chemotaxis assay. This migration was significantly inhibited by 29% (P < 0·001) by blocking lymphocyte expressed CX3CR1 using recombinant fractalkine [42]. Although fractalkine played a significant role in mediating chemotaxis, it does not appear to be a major player in leucocyte chemotaxis compared to other tubular epithelial cell-derived chemokines.

Studies using an antibody against the fractalkine receptor, in a rat model of anti-GBM glomerulonephritis, demonstrated inhibition of leucocyte trafficking and crescent formation suggesting an important role for fractalkine in renal inflammation [43], while fractalkine expression in the renal interstitium of patients with crescentic glomerulonephritis correlated with NK cell infiltrates at these sites [44]. Further, in a mouse model of cardiac transplant, an increase in graft survival was reported in mice treated with antibodies to CX₃CR1 [45]. Whilst these studies, taken together with the data presented here, provide strong evidence for a role of fractalkine in renal inflammation further studies are required to elucidate the contribution of fractalkine during an inflammatory response in the kidney in vivo.