Cyclopentenone Prostaglandins Inhibit Cytokine-Induced NF-κB Activation and Chemokine Production by Human Mesangial Cells

Abstract

In the kidney an uncontrolled inflammatory response to an acute insult may lead to chronic inflammation, permanent tissue damage, and progressive renal insufficiency. Resolution of acute inflammation likely is dependent on endogenous regulatory mechanisms activated in parallel with mediators of renal inflammation. These mechanisms are postulated to attenuate the renal expression of proinflammatory cytokines, including the chemokines responsible for recruiting leukocytes to the kidney, thus facilitating the transition from inflammation to healing. To understand the regulation of the inflammatory response within the kidney, the effects of anti-inflammatory J series cyclopentenone prostaglandins on chemokine production by human mesangial cells were examined. Treatment of mesangial cells with prostaglandin J₂ and 15-deoxy-Δ^12,14-prostaglandin J₂ blocked interleukin-1β—induced monocyte chemoattractant protein-1 mRNA expression and protein production. This correlated with failure of the transcription factor nuclear factor-κB (NF-κB) to translocate to the nucleus and bind to its recognition motif, a step required for cytokine-induced monocyte chemoattractant protein-1 gene activation. NF-κB failed to translocate because the cyclopentenone prostaglandins attenuated degradation of the NF-κB inhibitor IκB-α. These data suggest that certain prostaglandins can limit the extent of renal chemokine expression and thus may have an important role in resolving renal inflammation.

Chemokines, such as monocyte chemoattractant protein-1 (MCP-1), are induced in the kidney as a response to various renal insults (¹). These chemokines are thought to play a critical role in the pathogenesis of renal inflammation by recruiting leukocytes to the kidney (²). We postulated that resolution of renal inflammation depends in part on intrinsic mechanisms that limit chemokine expression and thereby attenuate further infiltration of leukocytes into the kidney. The identities of such endogenous chemokine regulatory mechanisms are unknown.

Recent work suggests that certain J series prostaglandins derived from prostaglandin D₂ (PGD₂) may function as intrinsic regulators of inflammation. PGD₂ undergoes dehydration to 9-deoxy-Δ⁹-PGD₂ (PGJ₂) in aqueous environments and is dehydrated further to 9-deoxy-Δ⁹,Δ¹²,-dihydroprostaglandin D₂ and 15-deoxy-Δ^12,14-prostaglandin J₂ (15-d-PGJ₂) in the presence of protein (^3,4). These J series prostaglandins have a cyclopentenone ring that contains a reactive, electrophilic carbon. This ring can bind covalently to intracellular nucleophiles, such as free sulfhydryls and cysteines, and in so doing modify protein function (³). In addition, the cyclopentenone prostaglandins (cyPG) are taken up by cells and incorporated into the nucleus (⁵). The potential anti-inflammatory role of cyPG is illustrated clearly in a mouse model of pleural inflammation. In this model, cyclooxygenase (COX) inhibitors worsened inflammation when used beyond the acute phase of disease (⁶). This was associated with a decrease in the concentration of PGD₂ and 15-d-PGJ₂ in the pleural space (⁶). Replacement of these prostaglandins led to resolution of inflammation (⁶).

Three lines of evidence suggest that cyPG of the J series are reasonable endogenous candidates for modulating renal chemokine expression: (1) J series cyPG are found in the kidney; murine mesangial cells make PGJ₂ after treatment with interleukin-1β (IL-1) (⁷), and 9-deoxy-Δ⁹,Δ¹²,-dihydroprostaglandin D2 is present in normal human urine (⁸); (2) 15-d-PGJ₂ has been shown to block induction of nuclear factor-κB (NF-κB), a transcription factor necessary for chemokine gene transcription in most renal cell types (^9,10,11,12); (3) 15-d-PGJ₂ is a natural agonist of peroxisome proliferator-activated receptorγ (PPARγ) (¹³). PPARγ is a nuclear hormone receptor and transcription factor that has a role in regulating fatty acid homeostasis (^4,13) and a role in glucose disposition that was realized when it became clear that the thiazolidinedione class of antidiabetic drugs are specific PPARγ ligands (¹⁴). Relevant to the theme of this investigation, recent studies have shown that PPARγ also has significant anti-inflammatory activities that include attenuating monocyte cytokine production, promoting macrophage apoptosis, and inhibiting NF-κB activation (^15,16,17,18).

This investigation was undertaken to examine the effect of the cyPG 15-d-PGJ₂ on IL-1—induced production of MCP-1 by cultured human mesangial cells. In addition, because NF-κB is a key regulatory factor for mesangial MCP-1 expression, studies to determine whether 15-d-PGJ₂ modulates NF-κB activity were conducted. This work demonstrated 15-d-PGJ₂ to be a potent inhibitor of mesangial MCP-1 expression and NF-κB activation. 15-d-PGJ₂ acts at the level of IκB-α degradation but is not dependent on PPARγ for its antichemokine effects.

Materials and Methods

Human Mesangial Cell Culture and Treatment

Human mesangial cells were cultured from kidneys not suitable for transplantation. Mesangial cells from at least three different donors were isolated and characterized as we have previously described (¹⁹) and used between passages 4 and 7.

Mesangial cells were grown in RPMI-1640 plus 10% fetal calf serum (Bio Whittaker, Walkersville, MD) and changed to RPMI-1640 plus 0.25% bovine serum albumin at the time of use. The cells were treated for the indicated times with 1.1 ng/ml human recombinant IL-1β (R&D Systems, Minneapolis, MN) in the presence or absence of cyPG. All prostaglandins were purchased from Cayman Chemical (Ann Arbor, MI), as solutions in methyl acetate. Ciglitazone, a PPARγ agonist of the thiazolidinedione class, was obtained from Biomol (Plymouth Meeting, PA) and dissolved in ethanol. Bisphenol A diglycidyl ether (BADGE), a synthetic PPARγ antagonist (²⁰), was purchased from Fluka (Milwaukee, WI). In each experiment, the organic solvents were added to control or IL-1—treated cells at the same concentrations as cells treated with prostaglandins or ciglitazone. Mesangial cells were pretreated with prostaglandins or ciglitazone, as indicated in the individual experiments. Reagents used in this study were endotoxin free. The effects of the various treatments on mesangial cell viability were measured by trypan blue exclusion.

Measurement of Mesangial Cell MCP-1 Expression

MCP-1 mRNA was measured by Northern blotting of total mesangial cell RNA, as we have previously described (¹⁹). The MCP-1 probe was an XhoI fragment from phJE34 (American Type Culture Collection, Rockville, MD). The blots were reprobed for glyceraldehyde-3-phosphate dehydrogenase by use of a full-length human cDNA (Clontech, Palo Alto, CA), to confirm equal RNA loading.

Mesangial cell production of immunoreactive MCP-1 was determined by use of a modification (²¹) of the double-ligand enzyme-linked immunosorbent assay (ELISA) originally developed by Evanoff et al. (²²). The capture antibody was a mouse monoclonal anti-human MCP-1 (R&D Systems), and the upper antibody was a rabbit polyclonal anti-human MCP-1 (Pepro Tech, Rock Hill, NJ). A horseradish peroxidase—conjugated goat anti-rabbit antibody was used for detection. The standard curve was constructed with human recombinant MCP-1 (Pepro Tech). The ELISA was sensitive to 100 pg/ml. The prostaglandins did not interfere with the ELISA. Mesangial cells were grown to confluence in six-well plates before use. All MCP-1 concentrations were corrected for small variations in the number of mesangial cells present in individual tissue culture wells by use of total cellular protein as an index of cell density.

Electrophoretic Mobility Shift Analysis

Nuclear proteins were harvested from mesangial cells as described previously (¹⁰). Ten μg of nuclear protein were incubated with 0.2 ng of double-stranded oligonucleotide probe containing a tandem repeat of the consensus NF-κB motif (5′-GATCCAAGGGGACTTTCCGCTG-3′; Life Technologies BRL, Grand Island, NY). Probes were ³²P-end—labeled with T₄ polynucleotide kinase (Life Technologies BRL) plus [γ-³²P]ATP (3000 Ci/mmol; New England Nuclear, Boston, MA). Binding of the nuclear proteins to the NF-κB oligonucleotide was performed at 27°C in 25 μl of buffer containing 10 mM Tris (pH 7.5), 100 mM NaCl, 1 mM dithiothreitol, 5 mM ethylenediaminetetraacetate, 4% glycerol, and 2 μg poly(dI-dC) for 20 min. Oligonucleotide bound by NF-κB protein was separated from unbound oligonucleotide by electrophoresis on a 6% nondenaturing polyacrylamide gel in 0.25× Tris-borate-ethylenediaminetetraacetate buffer at 4°C. Transcription factors bound to the radiolabeled probes were identified by the appearance of bands with retarded mobility relative to unbound probe in autoradiographs of these gels. Specificity of the binding reaction was confirmed by use of excess (100-fold) unlabeled oligonucleotide to compete with labeled probes binding to nuclear proteins.

Western Blot Analyses

Prostaglandin-treated mesangial cells were lysed in 10% sodium dodecyl sulfate at specific times after addition of IL-1. Fifty μl of cell lysate were separated by electrophoresis on 15% polyacrylamide gels and transferred to nitrocellulose with the use of a semidry technique (²³). The nitrocellulose was dried and quenched in phosphate-buffered saline, 0.1% Tween-20 with 2% bovine serum albumin for 120 min. After washing, the nitrocellulose was blotted overnight at 4°C with a 1:1000 dilution of rabbit polyclonal antibodies that recognize either native IκB-α or IκB-α that had been phosphorylated on the serine residue at position 32 (New England Biolabs, Beverly, MA). Control blots (not shown) were probed with nonspecific rabbit IgG (Zymed, South San Francisco, CA) to confirm specificity of the band identified as IκB-α bands. The blots then were washed and incubated with biotinylated goat anti-rabbit antibody (Zymed) for 45 min. After further washing, the bands of interest were demonstrated by use of enhanced chemiluminescence (Amersham Life Sciences, Inc., Arlington Heights, IL).

Nuclear proteins isolated from mesangial cells were immunoblotted with a rabbit polyclonal anti-PPARγ antibody (Affinity BioReagents, Golden, CO). Nonimmune rabbit IgG served as a control. Blots were developed by use of enhanced chemiluminescence, as described above.

Statistical Analyses

Data are presented as mean ± SEM. Comparisons were made with the use of the paired t test. P < 0.05 was considered significant.

Results

15-d-PGJ₂ Inhibits Mesangial MCP-1 Expression

Induction of MCP-1 mRNA by IL-1 was measured in human mesangial cells treated with 15-d-PGJ₂. As shown in Figure 1, 15-d-PGJ₂ caused a dose-dependent inhibition of MCP-1 mRNA expression, with complete inhibition occurring between 25 and 50 μM 15-d-PGJ₂. Treatment of mesangial cells with the prostaglandin alone had no effect on baseline MCP-1 mRNA levels or the expression of the glyceraldehyde-3-phosphate dehydrogenase housekeeping gene (not shown). Consistent with blocking MCP-1 mRNA expression, 15-d-PGJ₂ also inhibited the secretion of immunoreactive MCP-1 by mesangial cells with an IC₅₀ of approximately 14.5 μM (Figure 2). Under these experimental conditions, 15-d-PGJ₂ had no effect on cell viability assessed by trypan blue exclusion.

Figure 1: 15-deoxy-Δ^12,14-prostaglandin J₂ (15-d-PGJ₂) attenuates interleukin-1 (IL-1)—induced monocyte chemoattractant protein-1 (MCP-1) mRNA expression. Mesangial cells were pretreated with the indicated concentrations of 15-d-PGJ₂ for 2 h, followed by IL-1 (1.1 ng/ml) for 3 h. Total cellular RNA was isolated and probed for MCP-1 mRNA by Northern blot analysis. Blots were reprobed for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) to ensure equal RNA loading. This blot is representative of five independent experiments.

Figure 2: 15-d-PGJ₂ attenuates IL-1—induced MCP-1 production. (A) Mesangial cells were treated with 10 μM 15-d-PGJ₂ for 2 h, followed by the addition of IL-1 (1.1 ng/ml) for 16 to 18 h. Cell-free supernatants were harvested, and MCP-1 content was determined by use of a specific enzyme-linked immunosorbent assay (ELISA). MCP-1 concentrations were corrected for variations in total cell protein. ^*, P < 0.002 versus IL-1 alone (n = 6). (B) Mesangial cells were treated with IL-1 plus the indicated concentrations of 15-d-PGJ₂, as described above. MCP-1 production was measured by ELISA. 15-d-PGJ₂ inhibits mesangial MCP-1 production in a dose-dependent manner.

15-d-PGJ₂ Inhibits NF-κB Activation in IL-1—Treated Mesangial Cells

In mesangial cells, the transcription factor NF-κB seems to be necessary for activation of the MCP-1 gene by IL-1, and most manipulations that downregulate mesangial MCP-1 also prevent NF-κB activation (^10,24,25). This was found to be true for 15-d-PGJ₂ as well. Mobility shift analysis of nuclear proteins from mesangial cells treated with IL-1 plus varying concentrations of 15-d-PGJ₂ demonstrated a dose-dependent decrease in nuclear protein binding to a consensus NF-κB recognition motif (Figure 3). This effect was apparent with 10 μM 15-d-PGJ₂, and 25 μM caused complete inhibition. The effects of 15-d-PGJ₂ were not limited to cells activated by IL-1; a similar inhibition of NF-κB induction after phorbol ester treatment also was observed (data not shown). 15-d-PGJ₂ alone had no effect on mesangial NF-κB activation.

Figure 3: 15d-PGJ₂ attenuates IL-1—induced nuclear factor-κB (NF-κB) activation. Mesangial cells were pretreated with the indicated concentrations of 15-d-PGJ₂ for 2 h. IL-1 (1.1 ng/ml) was then added for 1 h. Nuclear proteins were harvested, and binding to a radiolabeled oligonucleotide containing the consensus NF-κB binding motif was determined by use of electrophoretic mobility shift analysis. Specificity of binding was confirmed by use of a 100-fold excess of unlabeled oligonucleotide to compete for binding (com). Unbound oligonucleotide migrated to the bottom of the gel (arrow), whereas migration of oligonucleotide bound to NF-κB was retarded. Representative of five experiments.

To examine the mechanism of 15-d-PGJ₂-mediated NF-κB inhibition, we looked at the effects of this prostaglandin on the NF-κB inhibitory protein IκB-α. Degradation of IκB-α by the 26S proteasome complex is necessary for NF-κB nuclear translocation and occurs after phosphorylation of IκB-α serine residues at positions 32 and 36 and ubiquitination of the protein (²⁶). As shown in Figure 4A, IL-1 induced a rapid (within 2 to 5 min) phosphorylation of serine 32 of IκB-α, followed by a decline in the phospho—IκB-α level within 10 min that was progressive and consistent with degradation of this protein. In cells treated with IL-1 plus 15-d-PGJ₂, the phosphorylation of IκB-α was slower, occurring within 5 to 10 min, and there was no tendency for the level of phospho—IκB-α to decline after peak phosphorylation (Figure 4A). Levels were maintained for 20 min. To confirm an impairment in IκB-α degradation by 15-d-PGJ₂, we immunoblotted cells for total IκB-α (Figure 4B). Cells treated with IL-1 alone showed a near absence of IκB-α 10 to 20 min after IL-1 was added. IκB-α levels began to recover by 30 to 60 min (not shown), as we documented previously (¹⁰). In contrast, this rapid and profound decline in IκB-α levels was not observed in cells treated with IL-1 plus 15-d-PGJ₂, which showed only a mild decrease in IκB-α after 15 to 20 min of incubation (Figure 4B).

Figure 4: Time course of IκB-α phosphorylation and degradation in mesangial cells treated with IL-1 or IL-1 plus 15-d-PGJ₂. Mesangial cells were pretreated with 50 μM 15-d-PGJ₂ or vehicle for 2 h, followed by IL-1 for the indicated times (minutes). Cells were lysed immediately, and whole-cell lysates were used for immunoblotting. (A) Cells were immunoblotted for serine 32-phosphorylated IκB-α. Three independent experiments are shown, which demonstrated the rapid (2 to 5 min) appearance of phospho—IκB-α, followed by a fall in its level beginning at 10 min in cells treated with IL-1 alone. In contrast, cells treated with IL-1 plus 15-d-PGJ₂ showed a slower rise in the level of phospho—IκB-α that did not decline over 20 min. (B) Cells were immunoblotted for total IκB-α. IL-1 treatment caused a rapid (by 10 min) disappearance of IκB-α, which was sustained for 20 min. In contrast, cells treated with IL-1 plus 15-d-PGJ₂ showed only a mild decrease in the level of IκB-α, which occurred after approximately 15 min. Representative of four experiments.

Effect of Other Prostaglandins on Mesangial Cells

To determine whether the modulation of mesangial cell activation is specific to 15-d-PGJ₂, we incubated cells with other 15-deoxy-Δ^12,14-prostaglandin derivatives, or PGJ₂, the parent prostaglandin of 15-d-PGJ₂. Like 15-d-PGJ₂, 15-d-PGA₁ and 15-d-PGA₂ are cyPG, formed by the dehydration of PGE₁ and PGE₂, respectively. As shown in Figure 5, 15-d-PGA₁, 15-d-PGA₂, and 15-d-PGD₂ had little effect on IL-1—induced NF-κB activation or MCP-1 mRNA expression at concentrations that were effective for 15-d-PGJ₂. Consistent with these results, 15-d-PGA₁, 15-d-PGA₂, and 15-d-PGD₂ did not block MCP-1 protein production by mesangial cells (data not shown). PGJ₂ did, however, cause a dose-dependent inhibition of NF-κB activation similar to that of 15-d-PGJ₂ (Figure 6). PGJ₂ also blocked IL-1—induced MCP-1 mRNA expression (Figure 5), and incubation with 10 μM PGJ₂ overnight reduced mesangial MCP-1 secretion by 45% (32.7 ± 2.5 ng/ml versus 58.9 ± 14.9 from cells treated with IL-1 alone). Although PGD₂ is not a cyPG, it was predicted that it could inhibit NF-κB activation if it were metabolized to J series prostaglandins. Although the brief administration (2 h) of PGD₂ in concentrations up to 50 μM had little effect on IL-1—induced NF-κB activation, sustained exposure (up to 18 h) did inhibit NF-κB (Figure 6), consistent with a requirement for the metabolism of PGD₂. The 18-h treatment with PGD₂ had no effect on cell viability, assessed by trypan blue exclusion and MTT assay.

Figure 5: The effect of other 15-deoxy-Δ^12,14-prostaglandin derivatives on IL-1—induced NF-κB activation and MCP-1 expression. (A) Mesangial cells (n = 3) were pretreated for 2 h with 15-d-PGA₁, 15-d-PGA₂, or 15-d-PGD₂, followed by IL-1 (1.1 ng/ml) for 1 h. Nuclear proteins were analyzed for NF-κB by electrophoretic mobility shift analysis, as in Figure 3. (B) Mesangial cells (n = 3) were incubated with 25 μM PGJ₂ (J₂), PGD₂ (D₂), 15-d-PGA₁, 15-d-PGA₂, and 15-d-PGD₂ for 2 h, followed by IL-1 (1.1 ng/ml) for 3 h. Total mesangial RNA was isolated and probed for MCP-1 and GAPDH mRNA by Northern blotting, as in Figure 1.

Figure 6: The effect of PGJ₂ and PGD₂ on IL-1—induced NF-κB activation. Mesangial cells were pretreated with the indicated concentrations of PGJ₂ or PGD₂ for 2 h (A and B) or 10 μM PGD₂ (C) for the indicated times (hours). IL-1 (1.1 ng/ml) was added for 1 h before harvest of nuclear proteins. NF-κB binding was measured by electrophoretic mobility shift analysis, as in Figure 3. Representative of three experiments.

Effects of 15-d-PGJ₂ Are Not Mediated through PPARγ

15-d-PGJ₂ is a potent ligand and activator of PPARγ. In some cell types, such as monocytes, activation of PPARγ by 15-d-PGJ₂ seems to block the expression of NF-κB—dependent cytokine genes. Two approaches were used to determine whether the effects of 15-d-PGJ₂ on mesangial MCP-1 production were PPARγ dependent. First, mesangial cells were treated with IL-1 in the presence or absence of ciglitazone, a thiazolidinedione and specific PPARγ agonist (¹⁴). As shown in Figure 7, A and B, treatment of mesangial cells with ciglitazone in concentrations as high as 100 to 200 μM had no effect on IL-1—induced MCP-1 mRNA expression or NF-κB expression. This was not because of a lack of expression of PPARγ by human mesangial cells. PPARγ was found to be constitutively present in nuclear protein extracts of untreated mesangial cells and seemed to increase after ciglitazone (10 μM) treatment (Figure 7C). The apparent molecular mass of the protein identified by the anti-PPARγ antibody is 51.7 kD (n = 5), consistent with the molecular mass of the PPARγ1 isoform (^27,28,29).

Figure 7: Ciglitazone does not affect IL-1—induced MCP-1 mRNA expression or NF-κB activation. Mesangial cells were preincubated with the indicated concentrations of ciglitazone for 2 h, followed by IL-1 (1.1 ng/ml) for 3 h (Northern blot analysis) or 1 h (electrophoretic mobility shift analysis). Total cellular RNA or nuclear proteins were harvested. (A) MCP-1 mRNA was determined by Northern blotting, as described in Figure 1. Representative of three experiments. (B) NF-κB activation was determined by electrophoretic mobility shift analysis, as in Figure 3. Representative of four experiments. (C) Nuclear proteins were isolated from ethanol-treated cells (C; n = 5) or cells treated with ciglitazone (CIG; n = 2) for 18 h, separated by electrophoresis, and immunoblotted for peroxisome proliferator-activated receptorγ (PPARγ) by use of a specific rabbit polyclonal antibody.

In a second set of experiments, mesangial cells were pretreated with BADGE, a synthetic PPARγ antagonist (²⁰), to determine whether the inhibitory effect of 15-d-PGJ₂ could be reversed. BADGE is a competitive inhibitor of PPARγ, binding to but not activating this receptor (²⁰). As shown in Figure 8, BADGE was unable to reverse the effect of 15-d-PGJ₂ on NF-κB activation. Although it seems that NF-κB binding was reduced further in cells treated with both BADGE and 15-d-PGJ₂, compared with 15-d-PGJ₂ alone (Figure 8), BADGE had no effect on IL-1—induced NF-κB activation (data not shown).

Figure 8: The PPARγ antagonist bisphenol A diglycidyl ether (BADGE) does not reverse the inhibitory effects of 15-d-PGJ₂ on IL-1—induced NF-κB activation. Mesangial cells were pretreated with BADGE for 2 h, followed by the indicated concentrations of 15-d-PGJ₂ for 2 h. IL-1 (1.1 ng/ml) was added, and, after 1 h, nuclear proteins were isolated, and NF-κB binding activity was measured by electrophoretic mobility shift analysis. Representative of two experiments. In a third experiment (not shown), BADGE did not reverse the effect of 15-d-PGJ₂ even after pretreating for 24 h.

Discussion

Chemokines are potent proinflammatory agents that contribute to the development of renal inflammation after immune and nonimmune renal injury (¹). We postulated that to limit parenchymal damage from inflammation and prevent progressive renal insufficiency, intrinsic mechanisms are activated during the inflammatory response to modulate inflammation by attenuating chemokine expression. COX-2 products, including cyPG, are considered to have anti-inflammatory effects (^6,30). The data from the present study support a role for J series cyPG in the endogenous regulation of renal chemokine expression. This conclusion is based on our observation that 15-d-PGJ₂ and PGJ₂ block mesangial cell MCP-1 expression and NF-κB activation in response to proinflammatory stimuli, along with published reports indicating that J series prostaglandins can be produced by renal cells (⁷) and are found in human urine (⁸). That PGJ₂ and 15-d-PGJ₂ also prevent nuclear translocation of NF-κB suggests that these prostaglandins may downregulate not only MCP-1 expression but also the transcription of other chemokines and cytokines that require NF-κB.

15-d-PGJ₂ seems to interfere directly with IκB-α processing. Although IL-1—induced IκB-α phosphorylation was slower in 15-d-PGJ₂—treated cells, the main inhibitory effect was at the level of IκB-α degradation. Preventing the degradation of IκB-α provides a mechanistic explanation for 15-d-PGJ₂—mediated inhibition of NF-κB nuclear translocation. These data suggest further that 15-d-PGJ₂ interferes with either the process of IκB-α ubiquitination after its phosphorylation, a step necessary for proteasomal degradation, or the function of the 26S proteasome complex, which is responsible for digesting IκB-α (²⁶). Which of these novel activities is to be added to the growing list of ways that 15-d-PGJ₂ interferes with the NF-κB activation cascade remains to be determined. At present, it seems that, in some cells, the reactive cyclopentenone ring may alkylate cysteine residues in the p50 and p65 subunits of NF-κB and thus block DNA binding (¹²). Alternatively, a number of recent reports demonstrated that IκB-α phosphorylation is prevented by 15-d-PGJ₂, because 15-d-PGJ₂ directly blocks the activity of IκB kinases (^11,12,31). This presumably occurs via covalent modification of critical residues of IκB kinase by the cyclopentenone ring. It is conceivable that cyPG could modify the structure of the proteasome in this manner and alter its function. The exact mechanism(s) involved in NF-κB inhibition may depend on the specific cell type. Indeed, one investigation showed that despite clear inhibition of IκB kinase activity in a macrophage cell line, 15-d-PGJ₂ had no effect on IκB kinase activity in HeLa cells (¹²). Studies to evaluate proteasome function in 15-d-PGJ₂—treated mesangial cells are under way in our laboratory.

Although 15-d-PGJ₂ is a potent agonist of PPARγ and human mesangial cells express this receptor, this study found no evidence supporting a role for PPARγ in the modulation of MCP-1 or NF-κB by 15-d-PGJ₂ in mesangial cells. In contrast, there is a report that the thiazolidinedione troglitazone attenuates MCP-1 production in mesangial cells (³²). Many effects of troglitazone may, however, be drug specific rather than class specific and do not necessarily involve PPARγ (³³). In our system, the thiazolidinedione ciglitazone did not reproduce the effects of 15-d-PGJ₂, nor were the effects of 15-d-PGJ₂ inhibited by a PPARγ antagonist. Similar conclusions have been reached by other investigators (^11,31,34), although it seems clear that PPARγ can modulate NF-κB activity and cytokine expression in at least some types of cells (^17,18,35). This may depend on the level of PPARγ expression in a given cell type. This also raises the interesting possibility that therapeutically increasing PPARγ levels in a tissue compartment may augment the intrinsic anti-inflammatory effects of the cyPG.

The exact J series prostaglandins that are functional in vivo in the kidney remain to be determined. Although there is some evidence that 15-d-PGJ₂ is produced at sites of inflammation (⁶), other investigators have questioned whether PGJ₂ is metabolized to 15-d-PGJ₂in vivo (³⁶). In addition, although the effects of 15-d-PGJ₂ seem to be due to its reactive cyclopentenone ring (¹²), it is clear from our data that not all cyPG are equally active, at least in human mesangial cells. 15-d-PGA₁ and 15-d-PGA₂ were less effective inhibitors of IL-1—induced NF-κB activation and MCP-1 expression than 15-d-PGJ₂. It is interesting that other studies also showed J series cyPG to be more potent when compared with A series cyPG (^34,37).

Despite this selectivity, the concentration of 15-d-PGJ₂ required to inhibit cytokine expression seems to be cell specific. In some cells, cytokine expression and NF-κB activation were inhibited by 15-d-PGJ₂ with IC₅₀s of <5 μM (^11,17). In other cells, the concentration of 15-d-PGJ₂ for NF-κB and cytokine inhibition was found to be in the 5- to 30-μM range (^35,38), similar to that of the mesangial cell.

The data from this study may help explain the observation that COX inhibitors can worsen renal inflammation in experimental models of nephritis. In anti—glomerular basement membrane antibody nephritis and anti—Thy-1 nephritis, administration of nonselective COX inhibitors increased glomerular leukocyte infiltration and MCP-1 expression (^39,40). Timing of the administration of COX inhibitors may determine whether these agents treat or exacerbate inflammation. For example, in the model of pleural inflammation discussed previously (⁶), two peaks of COX-2 activity were observed. An early peak was associated with production of proinflammatory prostaglandins, whereas a delayed rise in COX-2 activity was associated with the production of PGD₂ and cyPG. Appropriate use of COX inhibition in humans may require a detailed analysis of the time course of appearance of specific prostaglandins.

Activation of cyPG production in the kidney thus may be an endogenous mechanism that contributes to the resolution of the renal inflammation by attenuating NF-κB—dependent proinflammatory gene expression. It is tempting to speculate that clinically observed variations in tissue injury in response to inflammatory stimuli are determined by the level of J series prostaglandins that an individual is able to produce. Supporting the feasibility of this idea, a defect in the production of PGJ₂ has been described in mesangial cells from lupus-prone mice (⁷). Such defects may lead to a more severe disease phenotype because of uncontrolled inflammation. Understanding how to augment the production of cyPG in human disease therefore may prove therapeutically useful in treating renal inflammation.

This work was supported in part by United States Public Health Service Grant DK46055 (to BHR) and a research grant from the National Kidney Foundation of Ohio (to BHR). AMC was the recipient of an Undergraduate Student Summer Research Fellowship Award from the American Heart Association, Ohio Valley Affiliate, Inc. Portions of this work were presented in abstract form at the 2000 American Society of Nephrology Meetings, Toronto, Ontario.