Mice with chimeric livers are an improved model for human lipoprotein metabolism

Abstract

Objective: Rodents are poor model for human hyperlipidemias because total cholesterol and low density lipoprotein levels are very low on a normal diet. Lipoprotein metabolism is primarily regulated by hepatocytes and we therefore assessed whether chimeric mice extensively repopulated with human cells can model human lipid and bile acid metabolism.

Design: FRG [ F ah(-/-) R ag2(-/-)Il2r g (-/-)]) mice were repopulated with primary human hepatocytes. Serum lipoprotein lipid composition and distribution (VLDL, LDL, and HDL) was analyzed by size exclusion chromatography. Bile was analyzed by LC-MS or by GC-MS. RNA expression levels were measured by quantitative RT-PCR.

Results: Chimeric mice displayed increased LDL and VLDL fractions and a lower HDL fraction compared to wild type, thus significantly shifting the ratio of LDL/HDL towards a human profile. Bile acid analysis revealed a human-like pattern with high amounts of cholic acid and deoxycholic acid (DCA). Control mice had only taurine-conjugated bile acids as expcted, but highly repopulated mice had glycine-conjugated cholic acid as found in human bile. RNA levels of human genes involved in bile acid synthesis including CYP7A1, and CYP27A1 were significantly upregulated as compared to human control liver. However, administration of recombinant hFGF19 restored human CYP7A1 levels to normal.

Conclusion: Humanized-liver mice showed a typical human lipoprotein profile with LDL as the predominant lipoprotein fraction even on a normal diet. The bile acid profile confirmed presence of an intact enterohepatic circulation. Although bile acid synthesis was deregulated in this model, this could be fully normalized by FGF19 administration. Taken together these data indicate that chimeric FRG-mice are a useful new model for human lipoprotein and bile-acid metabolism.

Figure 1. Lipoproteins in mouse serum.

A, Serum cholesterol lipoprotein profiles measured by size exclusion chromatography of wild type mice, human, high (90%) and low (45%) levels of repopulation in humanized FRG mice. Panel B showing percentage of different lipoprotein fractions, as well as ratio of LDL/HDL in wild type mice, human controls, repopulated FRG mice and FRG controls. C, Western blot analysis of human (h) and mouse (m) Apolipoprotein E in serum samples of human and mouse control samples, 1–6. Humanized FRG with different levels of repopulation are shown in lane 7–9.

Figure 2. Bile acid composition and expression of enzymes of the bile acid synthesis pathway.

A, Ratio of deoxycholic acid (DCA) over beta-muricholic acid (BMCA) in high (<80%), moderate (50–80%) or low (30–50%) population compared to non-transplanted FRG mice. Statistics were performed by a 1-way ANOVA on log-transformed data followed by Dunett’s test. The overall significance was p = 0.023 (all different vs control). B, RNA expression of CYP7A1, CYP8B1 and CYP27A1 in hepatocytes, normalized to cyclophillin analyzed by quantitative real time PCR. Expression of human genes were analyzed in hepatocytes isolated from humanized FRG (Tx-Mice) and compared to isolated human hepatocyte controls (Human). Statistics were performed by a non-parametric Mann-Whitney U test.

Figure 3. Expression of human RNA.

A, Expression of human CYP7A1 in humanized FRG mice (TxFRG) treated with FGF19 (TxFRG+FGF19) compared to human control. Statistics were performed by a non-parametric Kruskal-Wallis ANOVA. The overall significance of the experiment was p<0.05. Expression of human CYP8B1 (B), CYP27A1(C), FXR (D) and SHP (E) in livers of humanized mice (TxFRG) treated with FGF19 (TxFRG+FGF19). Human liver RNA was used for reference (n = 9).

Figure 4. Expression of mouse RNA.

A, Cyp7a1, B,Cyp27a1, C, Cyp8b1, and E, SHP in livers of both FRG and humanized FRGN mice (TxFRG), with or without FGF19 (TxFRG+FGF19, FRG+FGF19). Statistics were performed by a 1-way ANOVA on log-transformed data followed by LSD test.