Antidiabetic actions of a non-agonist PPARγ ligand blocking Cdk5-mediated phosphorylation

Abstract

PPARγ is the functioning receptor for the thiazolidinedione (TZD) class of antidiabetes drugs including rosiglitazone and pioglitazone. These drugs are full classical agonists for this nuclear receptor, but recent data have shown that many PPARγ-based drugs have a separate biochemical activity, blocking the obesity-linked phosphorylation of PPARγ by Cdk5. Here we describe novel synthetic compounds that have a unique mode of binding to PPARγ, completely lack classical transcriptional agonism and block the Cdk5-mediated phosphorylation in cultured adipocytes and in insulin-resistant mice. Moreover, one such compound, SR1664, has potent antidiabetic activity while not causing the fluid retention and weight gain that are serious side effects of many of the PPARγ drugs. Unlike TZDs, SR1664 also does not interfere with bone formation in culture. These data illustrate that new classes of antidiabetes drugs can be developed by specifically targeting the Cdk5-mediated phosphorylation of PPARγ.

Figure 1. Novel PPARγ ligands lack classical agonism, block phosporylation at Ser273

a, Chemical structures of SR1664 and SR1824. b, Transcriptional activity of a PPAR-derived reporter gene in COS-1 cells following treatment with rosiglitazone, SR1664 or SR1824 (n=3). c and d, In vitro Cdk5 assay with rosiglitazone, SR1664 or SR1824 with PPARγ or Rb substrates. e, TNF-α-induced phosphorylation of PPARγ in differentiated PPARγ KO MEFs expressing PPARγ^WT treated with rosiglitazone, SR1664 or SR1824. Error bars are s.e.m.

Figure 2. Structural and in vitro functional analysis of SR1664

a, Lipid accumulation in differentiated 3T3-L1 cells treated with rosiglitazone or SR1664 following Oil-Red-O staining. b, Expression of adipocyte-enriched genes in these cells was analyzed by qPCR (n=3). c, Mineralization of MC3T3-E1 osteoblast cells as determined by Alizarin Red-S. Error bars are s.e.m.; *p<0.05, **p<0.01, ***p<0.001, n.s.; not significant. NT, no treatment. d, Overlay of differential HDX data onto the docking model of 2hfp bound to SR1664 (see Supplemental Fig. 3). This overlay depicts the difference in HDX between ligand-free and SR1664 bound PPARγ LBD. Perturbation data are color coded and plotted onto the backbone of the PDB file according to the key. Observed changes in HDX were statistically significant (p<0.05) in a two tailed t-test (n=3).

Figure 3. Anti-diabetic activity of SR1664 in high-fat diet (HFD) mice

a, Dose-dependent inhibition of phosphorylation of PPARγ by SR1664 in white adipose tissue (WAT). Quantification of PPARγ phosphorylation compared to total PPARγ (right). b, Ad libitum fed glucose (p=0.062 at 10mg/kg), insulin and HOMA-IR in HFD mice. c, Glucose infusion rate (GIR), suppression of hepatic glucose production (HGP), whole body glucose disposal and WAT 2-deoxyglucose tracer uptake during hyperinsulinemic-euglycemic clamps. d, Expression of a gene set regulated by PPARγ phosphorylation in WAT. e, Expression of an agonist gene set (see Methods) in WAT. Error bars are s.e.m.; *p<0.05, **p<0.01.

Figure 4. SR1664 has potent anti-diabetic activity and does not promote fluid retention in ob/ob mice

a, Phosphorylation of PPARγ in WAT (left). Quantification of PPARγ phosphorylation compared to total PPARγ (right). b and c, Fasting body weight, blood glucose and insulin levels prior to glucose-tolerance tests (GTT) in ob/ob mice treated with vehicle, rosiglitazone or SR1664 (n=8). Whole-body weight (d) and fat change (e) with continued drug administration following the GTT. f, Packed cell volume (PCV) in whole blood from ob/ob mice treated with vehicle, rosiglitazone or SR1664. Error bars are s.e.m.; *p<0.05, **p<0.01, ***p<0.001. n.s.; not significant.