Serum lipoprotein profiles of control and H4LivKO mice. Lipoproteins were separated from 60 μl of pooled mouse plasma samples by FPLC (n = 4 for each group). Profiles from H4Flox (control; open squares) and H4LivKO (solid squares) mice are shown. y axis, concentrations of cholesterol (solid line) and phospholipids (dotted lines). (Inset) Immunoblot analysis of apolipoproteins B, E, A-I, A-II, and Cs in plasma- (left) and in HDL-eluted fractions (right) from control (+/+) and H4LivKO (−/−) mice.

Histological analysis of the livers of H4LivKO mice. (A) Normal liver of a 45-day-old wild-type (+/+) mouse. Hepatocytes around the central vein have few vacuoles present. Shown is an H&E stain. Magnification, ×300. (B) Liver of H4LivKO mouse showing marked vacuolation of most hepatocytes (asterisks) except those around the central vein (CV). Those around the vein are eosinophilic, and one at the left side has a large intracytoplasmic pale eosinophilic inclusion (arrowhead; see panel C). Shown is an H&E stain. Magnification, ×300. (C) Centrilobular hepatocyte with large intracytoplasmic inclusion (arrowhead) in the liver of an H4LivKO mouse. Note the vacuolation of smaller hepatocytes. Shown is an H&E stain. Magnification, ×750. (D) PAS stain of normal wild-type mouse liver showing abundant PAS-positive glycogen (dark red color) in all hepatocytes around the central vein. Livers were fixed in alcoholic formalin. Magnification, ×300. (E) PAS stain of H4LivKO mouse liver showing decrease or loss of PAS-positive glycogen in most hepatocytes but abundant glycogen in centrilobular hepatocytes. Livers were fixed in alcoholic formalin. Magnification, ×300. (F) Frozen section of normal liver from a wild-type mouse, showing many small fat droplets in hepatocytes (red) throughout the liver. Shown is an oil red O stain. Magnification, ×150. (G) Frozen section of H4LivKO liver, showing abundant, large fat droplets in hepatocytes. Shown is an oil red O stain. Magnification, ×150. (H) Ultrastructure of the liver from a H4LivKO mouse showing abundant lipid (1) and mitochondria (m) in the hepatocyte at the top and areas of glycogen (g) in the hepatocyte at the bottom. Staining was with uranyl acetate and lead citrate. Magnification, ×6,000.

Exons 4 and 5 are deleted efficiently and specifically in the livers of H4LivKO mice. (A) Time course of Cre-mediated recombination at the HNF4α locus. Liver RNA from H4LivKO and control animals was subjected to Northern blotting and probed with a fragment of the HNF4α cDNA derived from the targeted exons (exons 4 and 5). (B to D) Tissues were isolated from 45-day-old wild-type (HNF4α +/+; AlbCre −/−), AlbCre (HNF4α +/+; AlbCre +/−), H4LivKO (HNF4α fl/fl; AlbCre +/−), and H4Flox (HNF4α fl/fl; AlbCre −/−) mice. (B) Northern blot analysis of 10 μg of total liver RNA. A truncated RNA species is evident in livers of H4LivKO mice; this species hybridizes to a 3′ fragment of the HNF4α cDNA but fails to hybridize to exons 4 and 5. (C) Western blot analysis of liver nuclei. Twenty micrograms of total nuclear protein was separated on a sodium dodecyl sulfate–10% polacrylamide gel electrophoresis gel and transferred to nitrocellulose. The membrane was probed with an antibody raised to a C-terminal peptide of HNF4α (Santa Cruz Biotechnology). Equal loading of nuclear protein was demonstrated by reprobing the membrane with an antibody directed against mouse histone H1. No full-length or truncated HNF4α protein was detected in livers of H4LivKO mice. (D) Northern blot analysis of total RNA isolated from control tissues of H4LivKO and control animals.

Northern blot analysis of liver RNA from H4LivKO and control animals. Total liver RNA (10 μg) was separated on 0.22 M formaldehyde–1% agarose gels and blotted to GeneScreen Plus membranes. Blots were hybridized in UltraHyb at 42°C for 16 h with the probes indicated. Blots were developed using a Storm 860 phosphorimager and quantified using ImageQuant software. Genes are grouped by function (see Discussion). (A) H4LivKO mice are deficient in expression of genes required for VLDL secretion and HDL synthesis. wt, wild type. (B) H4LivKO mice display increased expression of SR-BI, required for HDL uptake. ABCA1, ATP-binding cassette 1; LDLR, LDL receptor. (C) Expression levels of several nuclear receptors important for lipid homeostasis are unchanged by deletion of hepatic HNF4α. PXR, pregnane-X receptor; FXR, farsenoid-X receptor; LXRα, oxysterol receptor-α; LRH1, liver receptor homologue 1; SHP, small heterodimer partner. (D) Decreased expression of PPARα but increased expression of PPARα target genes in H4LivKO mice. CPT-II, carnitoyl-palmitoyl transferase-II; HMG, 3-hydroxy-3-methylglutaryl. (E) H4LivKO mice display gene expression patterns indicative of defective bile acid homeostasis. OATP1, organic anion transporter protein 1; MDR2, multidrug resistance protein 2. (F) Key genes in the fatty acid synthesis pathway are unaltered in H4LivKO mice. FAS, fatty acid synthase; S14, spot-14; SREBP-1c, sterol receptor element binding protein 1c.

Gene targeting and conditional deletion of exons 4 and 5 of the HNF4α gene. (A) Restriction maps of the wild-type allele, targeting vector, and targeted allele. P1, P2, and P3, probes used to assess recombination events; open boxes, neomycin (Neo) and thymidine kinase (TK) positive- and negative-selection cassettes, respectively; shaded boxes, exons, numbered as described by Taraviras et al. (55); arrowheads, position and orientation of loxP sites. The upside-down Neo shows that the Neo cassette is transcribed in an orientation opposite to that of the HNF4α gene. Restriction sites: E, EcoRI; H, HindIII; B, BamHI; K, KpnI; S, SacI. Sites introduced in the targeting vector are italicized. (B) Southern blot analysis of homologous recombination in ES cells electroporated with the targeting vector. Hybridizing fragments of wild-type (±) and targeted (t) alleles and their respective sizes are indicated. (C) Crosses between mice heterozygous for the targeted allele failed to yield targeted homozygotes (see text). In order to generate a viable, conditional knockout mouse line, heterozygous (t/+) males were crossed with EIIaCre females in order to generate both floxed (fl) and null (−) alleles. Shown is recombination between loxP sites 2 and 3 (although all recombination events, i.e., recombination between loxP sites 1 and 2, 2 and 3, and 1 and 3, were observed). (D) Southern blot analysis of tail DNA from pups derived from the cross described for panel C. Hybridizing fragments of wild-type (±), targeted (t), floxed (fl), null (-), and neomycin (neo) alleles and their respective sizes are indicated. The floxed, null, and neomycin alleles are the result of recombination between loxP sites 2 and 3, 1 and 3, and 1 and 2, respectively. (E) PCR genotyping of mice. (Top) HNF4α genotyping. Tail DNA was amplified using primers flanking loxP site 1 (A) in the floxed allele, yielding products of 241 and 180 bp for the floxed and wild-type alleles, respectively. (Bottom) Cre genotyping. mEH primers served as a positive control for amplification, yielding a fragment of 341 bp in all samples. The presence of the Cre transgene was indicated by the amplification of an additional band of 411 bp with Cre-specific primers.