Troglitazone suppresses the secretion of type I collagen by mesangial cells in vitro  Robert Earl Routh, John Hardwick Johnson, Kevin John McCarthy  Kidney International  Volume 61, Issue 4, Pages 1365-1376 (April 2002) DOI: 10.1046/j.1523-1755.2002.00277.x Copyright © 2002 International Society of Nephrology Terms and Conditions

Figure 1 Identification of peroxisomal proliferator activated receptor (PPAR)-α and PPAR-γ in mesangial cells by reverse transcription-polymerase chain reaction (RT-PCR). Total RNA from rat mesangial cells were harvested and RT-PCR was done with PPAR-α and -γ specific primers according to standard protocols (Methods section). Each primer set gave an amplicon of the predicted size. BP is the molecular marker, (lane 1) PPAR-γ amplicon (351 bp), (lane 2) PPAR-α amplicon (435 bp). Kidney International 2002 61, 1365-1376DOI: (10.1046/j.1523-1755.2002.00277.x) Copyright © 2002 International Society of Nephrology Terms and Conditions

Figure 2 (A) Direct action of troglitazone (TRO) and clofibrate (CLO) on mesangial cells through the changes in spread cell area. Mesangial cells were grown on collagen in 5 mmol/L glucose, 30 mmol/L glucose, or 5 mmol/L glucose with 25 mmol/L mannitol. Cells were then left untreated (UNT), or treated with TRO or CLO. Spread cell area (100 cells/treatment group) on digitized images of mesangial cells labeled with the fluorescent dye CMFDA was done using a planimetry subroutine within the image analysis software. No significant differences were found in spread cell area for cells grown in 5 mmol/L glucose in the UNT, TRO, or CLO groups. However, the UNT group grown in 30 mmol/L glucose had a significant increase in size as compared to cells grown under identical growth conditions treated with either TRO or CLO (*P < 0.0001 vs. all groups; ∞P < 0.0001 from all 30 mmol/L groups). Treatment with TRO or CLO corrected in part the changes in spread cell area induced by hyperglycemia. (B) Morphology of representative cells in the treatment groups outlined in panel A. Cells grown in 5 mmol/L glucose in the UNT, TRO, or CLO groups and those cells grown in 30 mmol/L glucose and treated with either TRO or CLO showed a compact, spindle shaped morphology. Note that the cells grown in the 30 mmol/L glucose with no treatment (UNT) appear as having a spread, flattened morphology that is significantly different from that of the rest of the treatment groups. Final magnification ×350. Kidney International 2002 61, 1365-1376DOI: (10.1046/j.1523-1755.2002.00277.x) Copyright © 2002 International Society of Nephrology Terms and Conditions

Figure 3 PPAR-α and PPAR-γ activation in mesangial cells treated with TRO, CLO, or UNT in 5 mmol/L or 30 mmol/L glucose conditions. Cells were grown in RPMI-1640 media containing 5 mmol/L glucose, 30 mmol/L glucose or 5 mmol/L glucose with 25 mmol/L mannitol. The cells were subsequently left untreated or treated with TRO or CLO. To determine the extent of PPAR-γ activation, the cells were simultaneously transfected with a PPRE luciferase reporter plasmid and a pRL-TK control plasmid (data normalization). The data show that PPAR-γ activation occurs with TRO in either 5 or 30 mmol/L glucose conditions, but PPAR-γ activation is significantly higher in cells grown 5 mmol/L glucose than 30 mmol/L glucose. There was no difference from UNT in the activation of PPAR α in cells grown in 5 mmol/L glucose and treated with CLO. However, there was a significant increase in PPAR-α activation in cells grown in 30 mmol/L glucose and treated with CLO. *P < 0.0001 from 5 mmol/L UNT and CLO; 30 mmol/L UNT; and mannitol. ∞P < 0.0009 from 5 mmol/L UNT and CLO. ¶P < 0.0001 from all groups except 5 mmol/L TRO. Kidney International 2002 61, 1365-1376DOI: (10.1046/j.1523-1755.2002.00277.x) Copyright © 2002 International Society of Nephrology Terms and Conditions

Figure 4 Total protein production (A) and laminin production (B) in mesangial cells grown in either 5 mmol/L glucose or 30 mmol/L glucose medium. Rat mesangial cells were grown in serum free conditions in RPMI 1640 medium containing 5 mmol/L glucose, 30 mmol/L glucose, or 5 mmol/L glucose with 25 mmol/L mannitol for 72 hours and either left untreated (UNT) or treated with TRO or CLO. At the end of 72 hours, the conditioned medium was collected and stored at -70°C until use. Total proteins in the medium (A) and the relative amounts of laminin were measured (B). Panel A shows that total secreted protein was significantly elevated in mesangial cells grown in 30 mmol/L glucose compared to those cells grown in 5 mmol/L glucose or the mannitol control (*P < 0.0001 from all 5 mmol/L groups and mannitol). Treatment with either TRO or CLO did not alter the level of total protein production by cells grown in either medium condition. Panel B shows the relative amounts of laminin produced by cells within each treatment group. Hyperglycemia did not increase the relative amounts of laminin production by mesangial cells compared to the laminin production in mesangial cells grown in 5 mmol/L glucose or osmotic control. CLO treatment did not have any effect on mesangial cell laminin production in cells grown in either 5 or 30 mmol/L glucose. In contrast, TRO treatment decreased laminin production in the normoglycemic and hyperglycemic conditions, however, the measurement did not reach statistical significance. Kidney International 2002 61, 1365-1376DOI: (10.1046/j.1523-1755.2002.00277.x) Copyright © 2002 International Society of Nephrology Terms and Conditions

Figure 5 Troglitazone treatment affects the biosynthesis of type I collagen, possibly via activation/binding to PPAR-γ. The relative amounts of type I collagen produced by mesangial cells grown in either 5 or 30 mmol/L glucose or osmotic control conditions are shown. Cells grown in 30 mmol/L glucose had a slight but not significant increase in the amount of type I collagen secreted into the medium by mesangial cells compared to that secreted by cells grown in 5 mmol/L glucose. Type I collagen production was significantly decreased by TRO treatment in cells grown in either 5 or 30 mmol/L glucose when compared to the UNT cells grown in the same glucose conditions. CLO treatment did not affect collagen production in cells grown in 5 mmol/L glucose, but did decrease collagen production at 30 mmol/L glucose. The effects of TRO on type I collagen production in cells grown in 30 mmol/L glucose appeared to be blunted when compared to the effect seen at 5 mmol/L glucose. BADGE, a PPAR-γ antagonist, was used to block the activation of PPAR-γ in mesangial cells. (Inset) A dose response study using a PPRE-luciferase reporter construct as a measure of PPAR-γ activation was done to determine the optimum amount of BADGE that will inhibit PPAR-γ in mesangial cells. The introduction of 50 μmol/L BADGE into the same series of cell culture assays blocks the TRO mediated decrease in type I collagen production in both the 5 and 30 mmol/L glucose conditions, suggesting that TRO's effect in this system may be mediated via binding to PPAR-γ. *P < 0.0001 from 5 mmol/L UNT, CLO, or 30 mmol/L UNT, CLO, or mannitol; or P < 0.0002 from 30 mmol/L TRO. ∞P < 0.0001 from 5 mmol/L UNT, TRO or CLO, and 30 mmol/L UNT or mannitol; P < 0.002 from 30 mmol/L CLO. ¶P < 0.0001 from 5 mmol/L TRO; P < 0.0004 from 30 mmol/L UNT. §P < 0.002 from all other BADGE groups, including mannitol/BADGE. Kidney International 2002 61, 1365-1376DOI: (10.1046/j.1523-1755.2002.00277.x) Copyright © 2002 International Society of Nephrology Terms and Conditions