To determine the potential role of PBI-4050 on diabetic kidney disease, we utilized an accelerated model of type 2 diabetic nephropathy, eNOS–/– db/db mice. In our initial studies, we treated mice with either vehicle or PBI-4050 (100 mg/kg) by daily gavage from 8–20 weeks of age. eNOS–/– db/db mice are moderately hypertensive (14, 15), and PBI-4050 had no effect on blood pressure (Supplemental Figure 1A; supplemental material available online with this article; https://doi.org/10.1172/jci.insight.120365DS1). In addition, there was no effect on body weight with PBI-4050 administration compared with vehicle-treated mice (Supplemental Figure 1B). By 8 weeks of age, the eNOS–/– db/db mice had already developed significant albuminuria, which progressively increased until 20 weeks of age in vehicle-treated mice (Figure 1A). In contrast, PBI-4050 prevented any further increase in albuminuria over the treatment period. Furthermore, compared with vehicle-treated mice, which exhibited a decrease in glomerular filtration rate (GFR) from 8–20 weeks, PBI-4050 preserved GFR (Figure 1B). Histological analysis indicated marked decreases in glomerulosclerosis (Figure 2A) and interstitial fibrosis (Figure 2B) in the PBI-4050–treated mice compared with vehicle-treated mice. Podocyte number per glomerulus section markedly decreased from 8–20 weeks of age in vehicle-treated mice, while PBI-4050 treatment significantly slowed this loss (Figure 2B).
To determine the effect of PBI-4050 in established diabetic nephropathy, we performed further studies in which we did not initiate therapy in eNOS–/– db/db mice until 16 weeks of age, a time at which the mice had already developed substantial glomerulopathy. As a comparator, we treated a subset of mice with the angiotensin-converting enzyme (ACE) inhibitor captopril, with or without concomitant administration of PBI-4050. As with earlier initiation of treatment, administration of PBI-4050 at 16 weeks of age did not affect body weight (Supplemental Figure 2). However, it did prevent further increases in albuminuria, while captopril led to a gradual decline in albuminuria. Of interest, the combination of PBI-4050 and captopril induced a marked early and sustained decrease in albuminuria (Figure 3A). Even with the delayed initiation of treatment with PBI-4050, there was a clear decrease in glomerulosclerosis and tubule injury at 24 weeks compared with vehicle-treated mice (Figure 3B), with reduced collagen deposition, as indicated by Masson’s trichrome and Picrosirius red staining (Figure 3, C and D). A decrease in collagens I an IV expression was also observed by IHC, along with reductions in the myofibroblast marker smooth muscle actin α (α-SMA) and the profibrotic marker connective tissue growth factor (CTGF) (Figure 4A). Furthermore, phosphorylation of SMAD3, which results from activation of TGF-β, was significantly reduced by PBI-4050 (Figure 4B). There was also decreased renal macrophage infiltration with PBI-4050 (Figure 5A) and evidence for reduced oxidative stress, as indicated by decreased nitrotyrosine staining (Figure 5B).
As noted above, PBI-4050 acts as an agonist for GPR40 and as an antagonist for GPR84. In addition to its expression in the kidney, GPR40 is highly expressed in the pancreatic islets and has been reported to stimulate insulin release (16). Hence, we examined blood glucose in eNOS–/– db/db mice treated with PBI-4050 from 8–20 weeks. Compared with vehicle-treated mice, hyperglycemia was markedly decreased with PBI-4050 (Figure 6A), and glucose tolerance was markedly improved (Figure 6B). In mice in which PBI-4050 treatment was not initiated until 16 weeks, blood glucose still decreased compared with vehicle-treated mice (Figure 6C).
The db/db mice on the C57BLKS/J (BKS) background have been reported to develop progressive pancreatic isletitis and decreased insulin production (17). We confirmed progressive decrease in plasma insulin levels in vehicle-treated eNOS–/– db/db mice (Figure 7A). In contrast, when the mice were treated with PBI-4050 beginning at 8 weeks of age, there was preservation of plasma insulin levels (Figure 7B), while initiation of treatment at 16 weeks of age reverted plasma insulin levels toward levels seen at 8 weeks (Figure 7C). In these mice, PBI-4050 treatment resulted in a significant preservation of insulin expression in pancreatic islets at 24 weeks compared with vehicle-treated mice (Figure 7D), along with markedly decreased pancreatic fibrosis (Figure 8A). Since PBI-4050 is also an antagonist or inverse agonist of the proinflammatory GPR84, we examined its effects on islet inflammation and found a marked decrease in infiltration of both macrophages and CD8+ T cells (Figure 8, B and C). There was decreased expression of the ER stress marker CHOP (Supplemental Figure 3), as well as an increase of autophagic activity in islets from PBI-4050–treated mice, as indicated by decreased expression of p62 (Figure 9). Finally, in addition to preservation of pancreatic islets, PBI-4050 markedly decreased hepatic steatosis in the eNOS–/– db/db mice (Supplemental Figure 4).
eNOS–/– db/db mice have a reduced lifespan and normally die between 24 and 28 weeks of age (18). We studied a subset of eNOS–/– db/db mice in which treatment with PBI-4050 or vehicle was initiated at 16 weeks. Compared with the 8 vehicle-treated mice, who had all died within 10 weeks, 6 of 8 PBI-4050–treated mice were still alive and appeared healthy after 14 weeks, when the experiment was terminated (Figure 10).
We have previously described a model of spontaneous tubulointerstitial fibrosis resulting from the selective overexpression of the EGFR ligand, HB-EGF, in renal proximal tubules (hHB-EGFTg/Tg) (19). These mice have augmented interstitial myofibroblast accumulation and progressive increase in interstitial extracellular matrix deposition. Because there are no apparent systemic metabolic abnormalities in these mice during the development of tubulointerstitial fibrosis, we tested whether PBI-4050 could affect the inexorable progression of the tubulointerstitial fibrosis in this nondiabetic model.
Our previous studies indicated an important role for EGFR-mediated ERK activation in induction of tubulointerstitial fibrosis (20), and treatment of hHB-EGFTg/Tg mice with a MEK inhibitor to inhibit ERK activation attenuated development of the tubulointerstitial fibrosis (19). In the present studies, we found that PBI-4050 significantly decreased phospho-ERK overexpression seen in proximal tubules of the vehicle-treated mice (Figure 11A).
Accumulation of extracellular matrix is the pathological hallmark of tubulointerstitial fibrosis. Sustained EGFR activation in the proximal tubule in the vehicle-treated hHB-EGFTg/Tg mice induced significant collagen deposition by 15 weeks, as indicated by Picrosirius red staining (Figure 11B). In contrast, PBI-4050 administration markedly decreased tubulointerstitial fibrosis. PBI-4050 treatment also resulted in decreased mRNA expression of collagen I, the myofibroblast marker; α-SMA; the profibrotic cytokines TGF-β and CTGF; and the extracellular matrix component, fibronectin (Figure 11C). Immunoblotting and immunostaining confirmed PBI-4050–mediated decreases in α-SMA and fibronectin protein levels, as well as decreased phosphorylation of SMAD3 (Figure 11, A and C).
There is increasing evidence that tubular dysfunction contributes to the initiation and progression of fibrosis, and that when renal epithelial cells revert to a dedifferentiated state, they produce profibrotic and proinflammatory cytokines. A number of markers of epithelial dedifferentiation have been described, including decreased expression of the adherens junction protein E-cadherin and increased expression of the transcription factors Snail1 and Snail2. In this regard, compared with vehicle-treated hHB-EGFTg/Tg mice, PBI-4050–treated mice had relative preservation of E-cadherin expression and inhibition of Snail2 expression (Figure 12, A and B). Furthermore, there was a marked decrease in renal interstitial macrophage infiltration in the PBI-4050–treated mice (Figure 12C), in accordance with results seen in the eNOS–/– db/dbmice (Figure 5A).