Macrophage-induced rat mesangial cell expression of the 24p3-like protein alpha-2-microglobulin-related protein

Biochimica et Biophysica Acta 1645 (2003) 218 – 227 www.bba-direct.com

Macrophage-induced rat mesangial cell expression of the 24p3-like protein alpha-2-microglobulin-related protein Izabella Z.A. Pawluczyk a,*, Peter N. Furness b, Kevin P.G. Harris a a

Department of Nephrology, Leicester General Hospital, Gwendolen Road, Leicester LE5 4PW, UK b Department of Renal Pathology, University of Leicester, Leicester, UK Received 21 March 2002; received in revised form 9 October 2002; accepted 18 November 2002

Abstract During screening of a murine macrophage cDNA repertoire for factors potentially able to modulate glomerular cell responses to injury, we identified a gene coding for the murine protein 24p3 lipocalin. Immunostaining of normal rat kidney sections showed positive 24p3-like staining in distal tubules/collecting ducts and small muscular arteries. Although most glomeruli were negative, some did exhibit small numbers of positively stained cells. Cultured rat glomeruli and glomerular mesangial cells secreted the 24p3-like protein in response to macrophage-conditioned medium (MPCM) and the cytokine IL-1h. MPCM derived from TGFh-pretreated macrophages enhanced mesangial cell 24p3 secretion. In contrast, addition of anti-IL-1h neutralising antibody to MPCM or IL-1h resulted in suppression of 24p3 secretion. Co-culture of mesangial cells with varying numbers of non-LPStreated macrophages resulted in dose-dependent secretion of 24p3 into culture supernatants. Archival sections from polyvinyl alcohol-treated and cholesterol-fed rats showed positive glomerular staining for 24p3 in and around glomerular foam cells. Nucleotide sequencing of rat mesangial cell-derived 24p3 cDNA revealed it to be identical to rat alpha-2-microglobulin-related protein (a2AGRP), the rat homologue of murine 24p3. These data provide the first description of rat a2AGRP in the context of mesangial cell pathophysiology. D 2002 Elsevier Science B.V. All rights reserved. Keywords: 24p3; a-2-microglobulin-related protein; Lipocalin; Macrophage; Mesangial cell; Interleukin-1h

1. Introduction A prominent infiltrate of macrophages is a characteristic feature of most forms of proliferative human and experimental glomerulonephritis. These cells are a source of a vast array of factors that not only inflict acute injury but also chronically alter the physiology and biology of adjacent nonimmune cell types. The macrophage has emerged as playing a significant role in mediating the transition from acute immune glomerular injury to chronic glomerular dysfunction and eventual sclerosis. During a screen of a murine macrophage (RAW 264.7 cell) cDNA repertoire for factors potentially capable of modulating glomerular cell responses to injury, we fortuitously identified a gene encoding a 24-kDa secreted acute

* Corresponding author. Tel.: +44-116-2584123; fax: +44-116-2734989. E-mail address: [email protected] (I.Z.A. Pawluczyk).

phase protein known as 24p3 lipocalin. 24p3 was originally discovered in the supernatants of Balb/c 3T3 fibroblasts stimulated by growth factors [1] and was cloned from an SV40-transformed quiescent mouse primary kidney cell culture infected with a polyoma A-2 wild-type virus [2]. Lipocalins are a family of small, secreted proteins, which have little amino acid sequence homology (20 – 30%) but share a common three-dimensional structure [3,4]. The mature protein folds to form a cone-shaped, protease-resistant structure containing a hydrophobic interior [5]. The precise role of lipocalins is not known but they are thought to bind and transport small hydrophobic ligands through hydrophilic body fluids to target cells [5]. Tissue distribution studies have revealed that 24p3 is mainly expressed in the liver during an acute phase response [6,7]. It has also been detected in spleen, lung and the uterus. In the latter location, its expression has been found to be coincident with parturition — a time of major tissue remodelling and inflammation [8 –10]. 24p3 has also been detected in the conditioned media of LPS-stimulated murine PU5.1.8 mac-

1570-9639/02/$ - see front matter D 2002 Elsevier Science B.V. All rights reserved. doi:10.1016/S1570-9639(02)00535-6

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

rophages and therefore suggested to function in defence against infectious agents [11]. As mesangial cell responses to macrophages are critical in the regulation of glomerular cell biology, we set out to establish whether this protein played a modulatory role within the kidney particularly in the context of macrophage-mediated injury or repair.

2. Materials and methods Unless otherwise stated, all reagents were purchased from Sigma-Aldrich Chemical Co. (Poole, Dorset, UK). 2.1. Glomerular culture Glomeruli were isolated by serial sieving of adult, female Wistar rat kidneys using standard techniques [12]. Briefly, the cortex was carefully removed from the rat kidneys and chopped into very small pieces. The pieces were minced through 500-, 250- and 75-Am mesh sieves using Hank’s balanced salt solution (HBSS, Life Technologies, Paisley, UK) to wash the tissue through the sieves. The glomeruli were collected off the surface of the 75-Am sieve and washed with HBSS. Approximately 2000 glomeruli/ml/well were plated into 24-well plates (ICN Flow, Oxford, UK) in RPMI 1640 (Life Technologies) containing 0.5% foetal calf serum (FCS) to render the glomeruli quiescent for 72 h before stimulation. 2.2. Culture of rat mesangial cells Glomeruli from approximately four kidneys per preparation were incubated with 750 U/ml collagenase type IV solution to remove the Bowman’s capsules. The resulting glomerular cores were washed three times in HBSS, resuspended and cultured in RPMI 1640 supplemented with 20% heat-inactivated FCS, 100 U/ml penicillin (Life Technologies), 100 Ag/ml streptomycin (Life Technologies), 5 Ag/ml bovine insulin and 2 mM glutamine (Life Technologies). Mesangial cells were characterised by their typical stellate fusiform morphology, positive staining for the Thy-1 antigen [13] and their resistance to the toxic effects of D-valine [14]. Mesangial cells of passages 2– 10 were cultured in 24well plates or 25 cm2 flasks (Costar-Corning, Buckinghamshire, UK), allowed to grow to confluence and then rendered quiescent in RPMI containing 0.5% FCS for 72 h before use. The day 3 medium from quiescing mesangial cells was retained and used as mesangial cell-conditioned medium (mccm) where appropriate. 2.3. Isolation of macrophages and preparation of macrophage-conditioned medium (MPCM) Peritoneal macrophages were obtained from female adult Wistar rats (Leicester University breeding colony) by

219

injecting 10 ml of 3% thioglycolate broth into the peritoneal cavity. After 5 days, the peritoneal cavity was lavaged with 20 ml cold HBSS (Life Technologies). Greater than 90% of the exudate cells obtained in this way were macrophages as judged by positive immunohistochemical staining for the rat monocyte/macrophage marker ED1 (Serotec, Oxford, UK). MPCM was prepared using a modified method of Kohan and Schreiner [15]. The exudate cells were purified by temporarily plating them at a cell density of 5  105 cell/ml in 25 cm2 tissue culture flasks. After 2 h of incubation at 37 jC in a humidified 5% CO2, 95% air atmosphere, nonadherent cells were removed by washing with HBSS. The macrophages were then stimulated with LPS (E. coli 026 B6) at a final concentration of 1 Ag/ml or with LPS plus 25 ng/ml hTGFh (natural, derived from platelets, R&D Systems, Abingdon, Oxon, UK) in serumfree RPMI for 16 h after which they were washed three times with HBSS and cultured for a further 48 h in serumfree RPMI to generate standard or TGFh-treated conditioned media (MPCM and MPCMTGF, respectively). The MPCMs were harvested and centrifuged for 10 min at 2000 rpm and frozen until required. MPCMRAW was prepared from a murine macrophage cell line RAW 267.4 (ECACC 91062702). RAW 267.4 cells were grown to confluence in 25 cm2 flasks. The cells were then stimulated with 1 Ag/ml LPS in serum-free RPMI for 16 h after which they were washed three times with HBSS and cultured for a further 48 h to generate the conditioned medium. 2.4. Culture of glomeruli or mesangial cells in the presence of MPCM or cytokines Quiescent glomeruli or mesangial cells were exposed to a 50% solution of MPCM, MPCMTGF, MPCMRAW, or 10 ng/ ml hTGFh, mTNFa, r/m IL-1h or hPDGF (R&D Systems) or to medium alone for 3 days after which the culture supernatants were retained and analysed for secretion of 24p3. For Northern blot analysis, mesangial cells were stimulated for 24 h with the aforementioned agents before extraction of RNA. In additional experiments, mesangial cells were exposed to MPCM or IL-1h in the presence of 7 Ag/ml rabbit anti-rat IL-1h (R&D Systems). Mesangial cells were also exposed to varying concentrations of LPS (1.0, 0.1, 0.01, 0.001 and 0.0001 Ag/ml). 2.5. Co-culture of mesangial cells with macrophages Confluent quiescent mesangial cells were co-cultured with varying numbers of non-LPS-stimulated rat peritoneal macrophages (25, 12.5, 6.25, 3.13 and 1.56  104 cells) (25  104 cells/well is the equivalent number of macrophages generating 50% MPCM) for 3 days. Supernatants were retained and analysed for 24p3 expression by Western blotting.

220

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

2.6. Culture of macrophages in the presence of mccm Peritoneal macrophages were plated into 24-well plates at a density of 2.5  105 cells per well (the equivalent number of macrophages in 50% MPCM). Varying concentrations of mccm were added to the macrophages. These cultures were maintained for 3 days after which the supernatants were analysed for 24p3 by Western blotting.

cytoplasmic antigen ED1 was detected using mouse antirat ED1 at 1/20 dilution in PBS for 2 h at room temperature. After washing, the primary antibody was detected using a horseradish peroxidase (HRP)-labelled rabbit anti-mouse Ig secondary antibody (Dako) for 1 h. After two further washes, the sections were developed with diaminobenzidine (DAB) SIGMA-FAST substrate. Sections were counterstained with haemotoxylin for 10 s. A mouse IgG1 preparation (Dako) was used as an isotype control.

2.7. Polyvinyl alcohol (PVA) study 2.9. Northern blotting Archival sections from a study in which PVA had been employed to induce glomerular macrophage infiltration were used. Briefly, 24 female Wistar rats were given twice weekly subcutaneous injections of either PVA (50 mg/100 g body weight) (given as a 5% solution in 0.9% saline) (n = 12) or saline alone (n = 12). The animals were pair fed a diet containing 24% casein protein or an identical diet supplemented with 4% cholesterol and 1% cholic acid to induce hypercholesterolaemia. At 24 weeks, the animals were sacrificed by terminal anaesthesia. The kidneys were removed, bisected, fixed in Carnoy’s fixative, washed in 70%, 95% and absolute alcohol and embedded in paraffin wax for histological examination. 2.8. Immunocytochemistry 2.8.1. 24p3 staining Four-micrometer sections were cut from paraffin wax blocks of normal or PVA study kidneys, and air-dried overnight. Paraffin wax was dissolved in xylene for 20 min and sections were then hydrated through decreasing concentrations of ethanol ending with a tap water rinse. Hydration was followed by a 2-min wash in TBS (pH 7.5). The sections were blocked with 2% normal goat serum for 30 min. 24p3 was detected using a polyclonal rabbit anti24p3 antibody (generous gift from Professor Marit NilsenHamilton, Iowa State University, Iowa, USA, prepared and characterised in her laboratory [16]) at a 1/100 dilution in PBS for 2 h at room temperature. After washing in TBS, the primary antibody was detected using an alkaline phosphatase-conjugated swine anti-rabbit Ig secondary antibody (Dako Ltd., Ely, UK) at a 1/20 dilution in PBS for 1 h. After two further washes in TBS, the sections were developed using 2-nitro-5-thiocyanobenzoic acid/5-bromo-4chloro-3-indolyl phosphate (NBT/BCIP) substrate plus 1 nM levamisole. Sections were counterstained with haematoxylin for 10 s. Normal rabbit serum instead of primary antibody was used as the negative control. 2.8.2. ED1 staining Following dewaxing as described above, sections were incubated in a blocking solution containing 0.3% hydrogen peroxide (v/v) for 30 min to block endogenous peroxidase activity. The sections were then washed and blocked with 2% normal rabbit serum for 30 min. The macrophage

Northern blot analysis was carried out using a method previously described [17]. Briefly, RNA was extracted from confluent, quiescent mesangial cells using TRIzol reagent (Life Technologies). Thirty-microgram aliquots of RNA were resolved on a 1% formaldehyde-agarose gel, blotted onto Hybond nylon membranes (Amersham-PharmaciaBiotech, UK) and hybridised with the [32P]-dCTP-24p3 cDNA labelled using a random primer labelling system (Prime-a-Gene, Promega, UK) in a solution containing 50% formamide, 1% SDS, 5  Denhardt’s, 5  SSPE and 200 Ag/ml denatured salmon sperm DNA at 42 jC. Membranes were exposed to X-Omat AR films (Kodak) with intensifier screens at 70 jC. The resulting autoradiographs were scanned using a Bio-Rad GS-700 imaging densitometer (Bio-Rad, Hertfordshire, UK). Cyclophilin (gift from SmithKline Beecham Pharmaceuticals) was used as the ‘housekeeping’ gene to normalise for RNA loading. 2.10. Reverse transcription – polymerase chain reaction (RT-PCR) Aliquots of total RNA (0.5 Ag) from whole kidney, glomeruli, mesangial cells, peritoneal macrophages, or RAW 264.7 cells were reverse transcribed using AMV Reverse Transcription System (Promega) according to the manufacturer’s instructions. The resulting first-strand cDNA was amplified using ReddyMixkPCR Mastermix (ABgene, Surrey, UK) using primers custom-made by Life Technologies: (sense) 5VATGGCCCTGAGTGTCATGTGT 3V(54 – 74) and (antisense) 5VGTTGTCAATGCATTGGTC 3V(2879 – 2896). Primers were designed on the basis of the published sequence [18]. The following thermocycling conditions were used : 1 cycle at 94 jC for 5 min, 35 cycles of 94 jC for 30 s, 55 jC for 1 min, 68 jC for 2 min, 1 cycle at 68 jC for 7 min, 10 min soak at 4 jC. The 597-base pair (bp) PCR product was visualised following electrophoresis on a 1% agarose – TAE gel. GAPDH was used as the housekeeping gene. 2.11. 24p3 probe 24p3 cDNA probe was generated by RT-PCR of RNA extracted from RAW 264.7 cells. The 597-bp PCR product

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

221

Denmark) for 1 h at room temperature. Following washing, the membranes were developed with DAB (SIGMAFASTk). 2.14. Statistics Results, where appropriate, are expressed as mean F S.E. employing an unpaired t-test to show statistical significance ( P < 0.05). All experiments were carried out on at least three

Fig. 1. Macrophage expression of 24p3. Northern blot, showing expression of 24p3 mRNA by peritoneal macrophages or RAW 264.7 cells F stimustimulation with LPS. RNA loading has been normalised using cyclophilin ‘housekeeping’ gene.

was isolated by electrophoresis on a 1% agarose –TAE gel and purified from the gel using Sephaglas Band Prep (Amersham-Parmacia-Biotech) according to the manufacturer’s instructions. 2.12. Nucleic acid sequencing The purified 24p3 probe and rat mesangial cell-derived 24p3 PCR product were sequenced to confirm or ascertain identity using ABI PRISMk dRhodamine Terminator Cycle Sequencing Ready Reaction System (PE Applied Biosystems, Perkin-Elmer Corporation) according to manufacturer’s instructions. 2.13. Western blotting MPCM, MPCMRAW, tissue culture supernatants from stimulated glomeruli and mesangial cells, and from macrophage/mesangial cell co-cultures were concentrated on a Savant DNA110 Speed Vac (Savant Instruments Farmingdale, New York, USA) (4-fold for culture supernatants and 40-fold for MPCMs). The concentrated supernatants were mixed 1:1 with reducing sample buffer and resolved on 12% SDS-polyacrylamide gels. The gels were blotted onto nitrocellulose membranes and immunostained with a polyclonal rabbit anti-24p3 antibody (generous gift from Professor Marit Nilson-Hamilton, Iowa State University). Briefly, the membranes were blocked with a 2% solution of BSA in TTBS (TBS containing 0.5% Tween 20) for 1 h. The membranes were then incubated with a 1/200 dilution of rabbit anti-24p3 antibody in TTBS for 2 h at room temperature. After washing four times with TTBS, the membranes were incubated with goat anti-rabbit Ig-HRP (Dako Ltd.,

Fig. 2. 24p3 immunohistochemistry in normal kidney. Section A shows positive staining in the macula densa, small muscular arteries and distal tubules/collecting ducts. In addition, occasional glomeruli exhibited some positively stained cells ( 20). Section B shows normal rabbit serum control. (C) RT-PCR was performed on RNA derived from normal whole rat kidney or normal rat glomeruli using specific primers for 24p3 or GAPDH.

222

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

MPCMs by Western blotting although only after 40-fold concentration (data not shown). 3.2. 24p3 expression in normal kidney

Fig. 3. Glomerular secretion of 24p3. Western blot of tissue culture supernatants from glomeruli exposed to medium alone (control), PDGF, TGFh, TNFa, IL-1h, or MPCM stained for 24p3.

different rat mesangial cell preparations. Each treatment was carried out in quadruplicate.

3. Results

To examine the expression/distribution of 24p3 in the normal kidney, tissue sections from normal rat kidneys were analysed by immunohistochemistry. We found positive staining in the macula densa, distal tubules/collecting ducts and small muscular arterioles. Although the majority of the glomeruli were negative, some did contain the occasional positively stained cell (Fig. 2A). Staining in the loops of Henle was also observed and was consistent with previous observations [19] (data not shown). RT-PCR confirmed that 24p3 was expressed in normal whole kidney and normal glomeruli (Fig. 2C). 3.3. Glomerular expression of 24p3 To further investigate glomerular expression of 24p3, isolated glomeruli were exposed to MPCM, or 10 ng/ml of r/mIL-1h, mTNFa, hTGFh or hPGDF, or to medium alone.

3.1. Macrophage expression of 24p3 Northern blotting using a 24p3 cDNA probe demonstrated that both LPS-stimulated rat peritoneal macrophages and LPS-stimulated RAW 264.7 cells expressed the 1.0-kb 24p3 transcript (Fig. 1). However, RAW 264.7 cells expressed this message constitutively, whereas rat peritoneal macrophages expressed the message only following LPS stimulation. Trace amounts of 24p3 protein were detected in the

Fig. 4. Mesangial cell expression of 24p3 mRNA levels in response to MPCM or cytokines. Northern blot showing levels of expression of 24p3 mRNA in mesangial cells in response to MPCM or medium alone (A), or to 10 ng/ml TGFh, PDGF, TNFa, or IL-1h (B).

Fig. 5. 24p3 secretion by mesangial cells in response to MPCM or cytokines. Western blot showing 24p3 staining of tissue culture supernatants from mesangial cells exposed to standard MPCM, control medium, 10 ng/ml TGFh, PDGF, TNFa, or IL-1h (A) or mesangial cells exposed to MPCMRAW (F LPS) (B).

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

223

After 3 days, tissue culture supernatants were assessed for the presence of 24p3 by Western blotting. The blots demonstrated that MPCM and r/mIL-1h stimulated strong glomerular expression of 24p3. Very low levels of protein were detected in response to TNFa whilst only trace amounts were detected in control glomeruli and following stimulation with TGFh and PDGF (Fig. 3). 3.4. Effect of MPCM and cytokines on mesangial cell 24p3 expression To ascertain whether glomerular mesangial cells may be responsible for glomerular secretion of 24p3, mesangial cells were exposed to MPCM or medium alone, or to 10 ng/ml TGFh, PDGF, TNFa or IL-h. Northern blotting demonstrated de novo expression of the 24p3 mRNA in response to MPCM and IL-1h. Unstimulated mesangial cells did not express the 24p3 transcript nor did those stimulated with TGFh or PDGF (Fig. 4). Cells stimulated

Fig. 7. Coincubation of mesangial cells with macrophages on 24p3 secretion. Confluent quiescent mesangial cells were exposed to varying numbers of macrophages (0.156 – 2.5  105 macrophages/ml/well). Tissue culture supernatants were Western blotted and stained for 24p3.

with TNFa were not generally seen to induce 24p3 expression although occasionally, trace levels were detected. 24p3 protein secretion, as assessed by Western blotting of tissue culture supernatants from stimulated mesangial cells, was consistent with the mRNA data showing a band of the order of 24 kDa following MPCM and IL-1h stimulation (Fig. 5A). Although RAW 264.7 cells constitutively expressed the 24p3 transcript (irrespective of prior LPS stimulation), only conditioned medium from LPS-stimulated RAW cells induced secretion of 24p3 in mesangial cells (Fig. 5B). 3.5. Effect of LPS on mesangial cell 24p3 secretion Because LPS is known to induce 24p3 secretion in macrophages, we wanted to eliminate the possibility that LPS contamination of MPCM could be inducing the secretion of 24p3 in mesangial cells. Mesangial cells were directly exposed to varying concentrations of LPS (1 Ag/ ml – 0.1 ng/ml) for 3 days and 24p3 protein secretion into the supernatants was assessed by Western blotting. 24p3 protein was not induced in mesangial cells following LPS stimulation (data not shown) thus confirming that macrophage-derived factors are directly responsible for stimulating mesangial cell 24p3 secretion. 3.6. Effect of anti-IL-1b neutralising antibody on IL-1b- and MPCM-induced 24p3 secretion

Fig. 6. Effect of anti-IL-1h neutralising antibody on IL-1h- and MPCMmediated 24p3 secretion. Western blot of supernatants from mesangial cells exposed to IL-1h (A) or MPCM (B) F anti-IL-1h neutralising antibody and stained for 24p3.

To further confirm the specificity of IL-1h-induced 24p3 secretion, mesangial cells were incubated with IL-1h in the presence or absence of a neutralising anti-IL-1h antibody. Western blotting of the supernatants showed that anti-IL-1h antibody inhibited IL-1h-induced 24p3 secretion (Fig. 6A).

224

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

To assess the contribution of IL-1h to MPCM-induced 24p3 expression, mesangial cells were exposed to MPCM in the presence or absence of anti-IL-1h neutralising antibody. Western blotting of the supernatants demonstrated that antiIL1h was able to inhibit most of the MPCM-mediated 24p3 secretion (Fig. 6B). 3.7. Effect of co-incubation of mesangial cells with macrophages on 24p3 secretion Because mesangial cells secrete 24p3 in response to a solution of macrophage-derived products, we investigated whether mesangial cells could respond to direct contact with non-LPS-stimulated macrophages.

Western blotting of culture supernatants from mesangial cell/macrophage co-culture experiments demonstrated dosedependent secretion of 24p3 with increasing numbers of macrophages (Fig. 7). This was confirmed by densitometric analysis of the Western blot bands from three independent experiments 25  104: 32.3 F 6.8, 12.5  104: 28.4 F 4.9, 6.25  104: 23.1 F 0.7, 3.13  104: 15.9 F 0.7, 1.56  104: 8.1 F 1.5 (macrophage cell number: arbitrary densitometric units). 3.8. Effect of mccm on macrophage 24p3 expression To assess whether mesangial cell-derived factors could stimulate macrophages to secrete 24p3, 2.5  105 macro-

Fig. 8. 24p3 immunohistochemistry in kidneys exhibiting a glomerular macrophage infiltrate. Kidneys from rats given PVA and a high-cholesterol diet, labelled with anti-24p3 antibody or anti-ED1 antibody as described in Materials and Methods. Section A shows 24p3 staining in areas of foam cell infiltration coincident with ED1 staining (B). Section C shows minimal to no 24p3 staining in areas that are clearly positive for ED1 (D). Section E shows normal rabbit serum control. Section F shows isotype control for ED1 antibody.

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

225

phages/well were cultured in the presence of varying dilutions of mccm. 24p3 protein was not detected in the resulting culture supernatants as determined by Western blotting (data not shown) confirming that mesangial cells secrete 24p3 in response to macrophages or their products. Mesangial cellderived products do not induce macrophages to secrete 24p3. 3.9. 24p3 expression in kidneys exhibiting a glomerular macrophage infiltrate To observe the effects of an infiltrate of macrophages in glomerular cells in vivo, we used the PVA model of induced glomerular foam cell infiltration. PVA is a synthetic macromolecular polysaccharide which, when administered subcutaneously, preferentially localises in the glomerular mesangium. This in turn induces an influx of activated blood-borne monocyte/macrophage cells into the mesangium [20]. Superimposition of a high-cholesterol diet is known to increase macrophage number and exacerbate the underlying glomerular injury [21]. Sections from kidneys where rats had been treated with PVA and received a high-cholesterol diet were stained for either 24p3 (using normal rabbit serum as control) or the cytoplasmic macrophage antigen ED1 (using isotypematched murine IgG1 as control). Positive 24p3 staining was observed in the glomeruli in discrete areas of foam cell formation (Fig. 8A). While ED1 staining was observed in all foam cells (Fig. 8B), 24p3 staining was found in most but not all foam cells (Fig. 8C). This was apparent where staining had been carried out on sections from the same kidney (Fig. 8C and D). Both the normal rabbit serum control (Fig. 8E) and the isotype control (Fig. 8F) were negative for any staining.

Fig. 10. Effect of TGFh-suppressed macrophages on 24p3 protein levels. Western blot showing 24p3 staining of tissue culture supernatants from mesangial cells exposed to medium alone, MPCMTGF, and MPCM.

3.10. Effect of TGFb-suppressed macrophages on mesangial cells 24p3 expression To shed further light on the possible role of mesangial cell 24p3 expression during a macrophage insult, we employed a model of TGFh-induced macrophage suppression. We have previously demonstrated that conditioned medium derived from TGFh-suppressed macrophages (MPCMTGF) has a reduced ability to induce pro-fibrotic responses in cultured rat mesangial cells compared to standard MPCM [22]. We therefore examined what effect TGFh suppression would have on MPCM-mediated 24p3 expression in mesangial cells. Northern analysis demonstrated that 24p3 mRNA levels in response to MPCMTGF were 2.7 F 0.6-fold ( P < 0.05) higher than levels induced by standard MPCM (Fig. 9). This increase in mRNA levels was further supported by increased secretion of 24p3 protein into culture supernatants as determined by Western blotting (Fig. 10). 3.11. Sequencing of rat mesangial cell 24p3 To assess the nucleotide sequence of the rat mesangial cell-derived 24p3 mRNA (as opposed to murine RAW cellderived), nucleotide sequencing of the purified mesangial cell-derived PCR product was carried out. A search of EMBL and GenBank databases revealed that the rat mesangial cellderived product was identical to rat alpha-2-microglobulinrelated protein (a2AGRP), the probable rat homologue of murine 24p3.

4. Discussion

Fig. 9. Effect of TGFh-suppressed macrophages on 24p3 mRNA expression. Northern blot showing levels of expression of 24p3 mRNA in mesangial cells in response to medium alone, MPCMTGF, and MPCM. MPCMTGF induced 2.7 F 0.7 fold more 24p3 mRNA than standard MPCM (P < 0.05, n = 4).

The current study provides the first description of 24p3 lipocalin or more correctly the 24p3-like rat homologue a2AGRP in mesangial cells in response to macrophagederived products. Rat a2AGRP [23,24] is a protein akin to, but distinct from, hepatic alpha-2-microglobulin, a 18.6-kDa protein known to accumulate in the proximal tubule as a

226

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

15.5-kDa cleavage product and thought to facilitate proximal tubule fatty acid oxidation [25]. Very little is reported in the literature regarding the expression of a2AGRP although its presumed mouse analogue 24p3 has previously been described in the uterus at the time of parturition. Because the uterus undergoes extensive tissue remodelling during pregnancy and undergoes stress and tissue damage around parturition with an accompanying immune infiltrate, it has been hypothesised that 24p3 may play a role in the local inflammatory response and in tissue repair. We therefore considered whether this protein may have an analogous role in renal injury. In our study, a2AGRP/24p3 was expressed by LPSstimulated rat macrophages. Meheus et al. [11] reported a similar finding in LPS-stimulated PU5.1.8 macrophages. More interestingly, however, the protein was synthesised de novo by mesangial cells in response to direct macrophage co-culture, exposure to the conditioned medium of LPSstimulated macrophages, and stimulation with the macrophage-derived pro-inflammatory cytokine IL-1h. a2AGRP/ 24p3 was not expressed by quiesced mesangial cells or mesangial cells exposed to TGFh or PDGF. The fact that only conditioned medium from LPS-stimulated macrophages induced mesangial cell secretion of this protein may suggest that stimulation of a2AGRP/24p3 secretion is dependent upon the concentration of a soluble factor which is produced in greater quantities in LPSstimulated macrophages, such as is found, for example, with IL-1h. The fact that anti-IL-1h antibody was able to inhibit both MPCM- and IL-1h-induced a2AGRP/24p3 secretion indicates that IL-1h is the major stimulator of this protein in our system. It also verifies that the observed increase in a2AGRP/24p3 secretion was not due to contamination by LPS of the MPCM or IL-1h preparations and it supports the idea of an LPS-induced increase in macrophage IL-1h concentration in MPCM being able to stimulate mesangial cell a2AGRP/24p3. a2AGRP/24p3 secretion as a result of direct contact between non LPS-stimulated macrophages and mesangial cells is probably modulated using an adhesion mechanism, the nature of which is beyond the scope of the current study. Nevertheless, the fact that macrophages did not secrete a2AGRP/24p3 in response to varying doses of mccm, coupled with the fact that even following LPS stimulation peritoneal macrophages secreted very little a2AGRP/24p3 protein, would tend to suggest that the a2AGRP/24p3 detected in the co-culture supernatants was derived largely from mesangial cells rather than the macrophages. Although little a2AGRP/24p3 staining was observed in normal glomeruli, immunohistochemical staining of archival sections from the PVA study demonstrated that the protein was dramatically expressed in diseased glomeruli in what resembled discrete areas of foam cell infiltration. Because not all of the ED1-positive cells stained for 24p3, this may suggest that there are phenotypic differences within the glomerular foam cell population.

In the current study, murine 24p3 cDNA hybridised with rat-derived mesangial cell RNA and the anti-murine 24p3 antibody cross-reacted with the rat mesangial cell protein, further supporting the theory that 24p3 and a2AGRP are the mouse and rat homologues of the same protein. A computerassisted homology search has previously revealed 80.3% identity between 24p3 and a2AGRP [23]. 24p3 also shares 71% mRNA sequence homology with a 25-kDa human neutrophil gelatinase-associated lipocalin (NGAL) [26], a protein thought to be the human homologue of 24p3. NGAL is found in the specific granules of neutrophils [27] where a minor portion of it is found covalently complexed with a 95kDa neutrophil-derived gelatinase [25,28]. NGAL expression has been located to epithelial cells in neoplastic and inflammatory disorders of colon and appendix [29], in trachea, bone marrow, lung and stomach [30]. Friedl et al. [19] carried out a comprehensive examination of NGAL expression on a wide spectrum of normal and neoplastic tissues by immunocytochemistry and found that NGAL was expressed by a variety of tissues. One interesting observation was that NGAL was highly expressed in cell types associated with microbial defence and has also been found in inflammatory conditions characterised by a prominent infiltrate of inflammatory cells including macrophages, lymphocytes, plasma cells and neutrophils. Since NGAL has been shown to bind the neutrophil and macrophage chemoattractant N-formylmethionyl-leucyl-phenylalanine [31,32], it was suggested that it might be a scavenger of bacterial products at sites of inflammation and thereby play some sort of immunomodulatory role. Friedl et al. [19] also observed expression of NGAL and its rat analogue in the normal kidney with a staining pattern comparable to that observed in our study. 24p3 has most recently been reported to be produced by a variety of cell types and its activation has been shown to be under the control of complex mechanisms. Devireddy et al. [33] reported that depriving FL5.12 pro-B cells of IL-3 activated 24p3 transcription leading to the synthesis and secretion of a protein which induced apoptosis in naı¨ve FL5.12 cells. Following on from this work, Persengieu et al. [34] found that 24p3 actually represses the expression of a pro-apoptotic protein. 24p3 expression has also been shown to be induced in IL-9-stimulated T-cell lymphoma lines [35] and in adipose tissue under conditions of hyperglycemia [36]. The widespread expression of a2AGRP/24p3 within the normal kidney and particularly in the medulla would tend to suggest that it is unlikely to play an injurious role. In view of the fact that TGFh-suppressed macrophages with their reduced ability to induce pro-fibrogenic responses could upregulate mesangial cell production of a2AGRP/24p3 in our study would tend to support the premise that a2AGRP/24p3 may play a protective role. TGFh is known to play a role in ‘glomerular self-defence’ whereby it allows the glomerulus to control and/or recover from macrophage-induced profibrotic effects which would otherwise lead to glomeruloscle-

I.Z.A. Pawluczyk et al. / Biochimica et Biophysica Acta 1645 (2003) 218–227

rosis [37]. Although TGFh does not play a direct role in a2AGRP/24p3 expression, the fact that TGFh-suppressed macrophages are able to up-regulate expression of a2AGRP/ 24p3 may suggest that this protein is up-regulated in mesangial cells in defence against the macrophage infiltrate. This hypothesis is supported by the studies of Orabona et al. [33] and Devireddy et al. [34] describing the role of 24p3 in leucocyte apoptosis. It is therefore, possible that 24p3 is secreted by mesangial cells to protect the glomerulus from infiltrating macrophages by inducing their apoptosis and thereby allowing the inherent glomerular cells to recover from an insult. Ryon et al. [38] have postulated a similar function for uterocalin in suppressing the inflammatory response during the process of uterine involution. Clearly, expression of 24p3 and its homologous molecules is found in a wide variety of settings where it plays a number of different roles. However, despite many years of investigation, its precise physiological function remains to be fully determined. The data presented in this study provide the first evidence for 24p3/a2-AGRP expression in the context of kidney pathophysiology and particularly macrophage/mesangial cell interactions, from which perspective one could intuitively suggest a protective function for this protein. Studies are underway to establish the precise physiological role for 24p3/a2-AGRP in the kidney, particularly in the context of macrophage-associated kidney disease.

Acknowledgements The authors acknowledge the generous gift of anti-24p3 antibody from Professor Marit Nilsen-Hamilton of Iowa state University. Thanks also go to Mr. Jez Brown for assistance with animal handling and section cutting. The work was supported, in part, by a Leicester General Hospital Research Award.

References [1] M. Nilsen-Hamilton, R.T. Hamilton, G.A. Adam, Biochem. Biophys. Res. Commun. 108 (1982) 158 – 166. [2] S. Hraba-Renevey, H. Turler, M. Kress, C. Salomon, R. Weil, Oncogene 4 (1989) 601 – 608. [3] D.R. Flower, A.C.T. North, T.K. Attwood, Biochem. Biophys. Res. Commun. 180 (1991) 69 – 74. [4] M.C. Peitsch, M.S. Boguski, Trends Biochem. Sci. 16 (1991) 363. [5] J. Godovac-Zimmerman, Trends Biochem. Sci. 13 (1988) 64 – 66.

227

[6] Q.S. Liu, M. Nilsen-Hamilton, J. Biol. Chem. 270 (1995) 22565 – 22570. [7] Q. Liu, M. Nilsen-Hamilton, FASEB J. 9 (1995) 709. [8] J.W. Kasik, E.J. Rice, Am. J. Obstet. Gynecol. 173 (1995) 613 – 617. [9] S.T. Chu, H.L. Huang, J.M. Chen, Y.H. Chen, Biochem. J. 316 (1996) 545 – 550. [10] Q.S. Liu, J. Ryon, M. Nilsen-Hamilton, Mol. Reprod. Dev. 46 (1997) 507 – 514. [11] L.A. Meheus, L.M. Fransen, J.G. Raymackers, H.A. Blockx, J.J.Van Beeumen, S.M. Van Bun, A. Van de Voorde, J. Immunol. 151 (1993) 1535 – 1547. [12] J.B. Foidart, C.A. Dechenne, P. Mahieu, C.E. Creutz, J. De Mey, Invest. Cell Pathol. 2 (1992) 15 – 26. [13] W.M. Bagchus, P. Hoedemaker, J. Rozing, W.W. Bakker, Lab. Invest. 55 (1986) 680 – 687. [14] S.F. Gilbert, B.R. Migeon, Cell 5 (1975) 11 – 17. [15] D.E. Kohan, G.F. Schreiner, Am. J. Physiol. 254 (1988) F879 – F886. [16] T.R. Davis, L. Tabatabai, K. Bruns, R.T. Hamilton, M. Nilsen-Hamilton, Biochim. Biophys. Acta 1095 (1991) 145 – 152. [17] I.Z.A. Pawluczyk, K.P.G. Harris, J. Am. Soc. Nephrol. 8 (1997) 1525 – 1536. [18] E. Garay-Rojas, M. Harper, S. Hraba-Renevey, M. Kress, Gene 170 (1996) 173 – 180. [19] A. Friedl, S.P. Stoesz, P. Buckley, M.N. Gould, Histochem. J. 31 (1999) 433 – 441. [20] W.S. Marcel, J.R. Hoyer, R.B. Sterzel, Lab. Invest. 49 (1983) 26 – 37. [21] J.R. Diamond, M.J. Karnovsky, Kidney Int. 32 (1987) 671 – 678. [22] I.Z.A. Pawluczyk, K.P.G. Harris, Kidney Int. 60 (2001) 533 – 542. [23] Y.-L. Chan, V. Paz, I.G. Wool, Nucleic Acids Res. 16 (1988) 11368. [24] S. Triebel, J. Blaser, H. Reinke, H. Tschesche, FEBS 314 (1992) 386 – 388. [25] Y. Wang, M.A. Shia, T.G. Christensen, S.C. Borkan, Kidney Int. 57 (2000) 1015 – 1026. [26] S. Bartsch, H. Tschesche, FEBS Lett. 357 (1995) 255 – 259. [27] L. Kjeldsen, D.F. Bainton, H. Sengelov, N. Borregaard, Blood 83 (1994) 799 – 807. [28] L. Kjelsden, A.H. Johnsen, H. Sengelov, F. Boulay, L. Kjeldsen, N. Borregaard, J. Biol. Chem. 268 (1993) 10425 – 10432. [29] B.S. Nielsen, N. Borregaard, J.R. Bundgaard, S. Timshel, M. Sehested, L. Kjeldsen, Gut 38 (1996) 414 – 420. [30] J.B. Cowland, N. Borregaard, Genomics 45 (1997) 17 – 23. [31] R.A. Allen, R.W. Erickson, A.J. Jesaitis, Biochim. Biophys. Acta 991 (1989) 123 – 133. [32] H. Sengelov, F. Boulay, L. Kjeldsen, N. Borregaard, Biochem. J. 299 (1994) 473 – 479. [33] L.R. Devireddy, J.G. Teodoro, F.A. Richard, M.R. Green, Science 293 (2001) 829 – 834. [34] S.P. Perengiev, L.R. Devireddy, M.R. Green, Genes Dev. 16 (2002) 1806 – 1814. [35] C. Orabona, L. Dumoutier, J.-C. Renauld, Eur. Cytokine Netw. 12 (2001) 154 – 161. [36] Y. Lin, M.W. Rajala, J.P. Berger, D.E. Moller, N. Barzilai, P.E. Scherer, J. Biol. Chem. 276 (2001) 42077 – 42083. [37] M. Kitamura, L.G. Fine, Kidney Int. 55 (1999) 1639 – 1671. [38] J. Ryon, L. Bendickson, M. Nilsen-Hamilton, Biochem. J. 367 (2002) 271 – 277.