Mouse precision-cut liver slices as an ex vivo model to study drug-induced cholestasis.

Drugs are often withdrawn from the market due to the manifestation of drug-induced liver injury (DILI) in patients. Drug-induced cholestasis (DIC), defined as obstruction of hepatic bile flow due to medication, is one form of DILI. Because DILI is idiosyncratic, and the resulting cholestasis complex, there is no suitable in vitro model for early DIC detection during drug development. Our goal was to develop a mouse precision-cut liver slice (mPCLS) model to study DIC and to assess cholestasis development using conventional molecular biology and analytical chemistry methods. Cholestasis was induced in mPCLS through a 48-h-incubation with three drugs known to induce cholestasis in humans, namely chlorpromazine (15, 20, and 30 µM), cyclosporin A (1, 3, and 6 µM) or glibenclamide (25, 50, and 65 µM). A bile-acid mixture (16 µM) that is physiologically representative of the human bile-acid pool was added to the incubation medium with drug, and results were compared to incubations with no added bile acids. Treatment of PCLS with cholestatic drugs increased the intracellular bile-acid concentration of deoxycholic acid and modulated bile-transporter genes. Chlorpromazine led to the most pronounced cholestasis in 48 h, observed as increased toxicity; decreased protein and gene expression of the bile salt export pump; increased gene expression of multidrug resistance-associated protein 4; and accumulation of intracellular bile acids. Moreover, chlorpromazine-induced cholestasis exhibited some transition into fibrosis, evidenced by increased gene expression of collagen 1A1 and heatshock protein 47. In conclusion, we demonstrate that mPCLS can be used to study human DIC onset and progression in a 48 h period. We thus propose this model is suited for other similar studies of human DIC.

Potential impact of sitagliptin on collagen-derived dipeptides in diabetic osteoporosis.

It is known that diabetes coincides with an increased risk of osteoporosis. While a disturbed collagen metabolism is proposed as a possible cause, much remains unknown about the enzymes involved and changes in the collagen-derived dipeptides and amino acids. Therefore, we sought to study this intricate pathway and the effect of dipeptidyl peptidase 4 (DPP4) inhibitors. Control and streptozotocin-nicotinamide-induced diabetic rats were treated for 12 weeks with vehicle or sitagliptin, a DPP4 inhibitor (Con/VH, Con/SG, DM/VH and DM/SG). The activities of four key enzymes involved in collagen breakdown were determined in serum (DPP4, matrix metalloproteinase 2 and 9 and prolidase). Dipeptide (Ala-Pro, Gly-Pro, Pro-Pro and Pro-Hyp) and amino acid (Pro and Hyp) concentrations were measured by liquid chromatography coupled to mass spectrometry. We found three-fold higher MMP9 activities in DM/VH than in controls, while in DM/SG this rise was attenuated. MMP2 and prolidase did not differ in the investigated groups. Furthermore, we are the first to report on two-fold higher Ala-Pro and Pro-Pro levels in diabetes compared to controls. In contrast, Pro-Hyp concentrations were lower in diabetes (DM/VH and DM/SG). DPP4 inhibition does not seem to have a direct influence on the collagen metabolism in streptozotocin-nicotinamide-induced diabetic rats. Instead, it probably acts through its effect on osteoprotective substrates. In diabetes, increased MMP9 activities seem to favour the production of Ala-Pro and Pro-Pro containing collagen fragments. The high Pro-Hyp levels in untreated controls might have a bone-stimulating effect. Nevertheless, the biological significance of these dipeptides is not yet clear and should be further investigated.