A mouse model for α-methylacyl-CoA racemase deficiency: adjustment of bile acid synthesis and intolerance to dietary methyl-branched lipids

Abstract

α-Methylacyl-CoA racemase (Amacr) deficiency in humans leads to sensory motor neuronal and liver abnormalities. The disorder is recessively inherited and caused by mutations in the AMACR gene, which encodes Amacr, an enzyme presumed to be essential for bile acid synthesis and to participate in the degradation of methyl-branched fatty acids. To generate a model to study the pathophysiology in Amacr deficiency we inactivated the mouse Amacr gene. As per human Amacr deficiency, the Amacr^−/− mice showed accumulation (44-fold) of C27 bile acid precursors and decreased (over 50%) primary (C24) bile acids in bile, serum and liver, however the Amacr^−/− mice were clinically symptomless. Real-time quantitative PCR analysis showed that, among other responses, the level of mRNA for peroxisomal multifunctional enzyme type 1 (pMFE-1) was increased 3-fold in Amacr^−/− mice. This enzyme can be placed, together with CYP3A11 and CYP46A1, to make an Amacr-independent pathway for the generation of C24 bile acids. Exposure of Amacr^−/− mice to a diet supplemented with phytol, a source for branched-chain fatty acids, triggered the development of a disease state with liver manifestations, redefining the physiological significance of Amacr. Amacr is indispensable for the detoxification of dietary methyl-branched lipids and, although it contributes normally to bile acid synthesis from cholesterol, the putative pMFE-1-mediated cholesterol degradation can provide for generation of bile acids, allowing survival without Amacr. Based upon our mouse model, we propose elimination of phytol from the diet of patients suffering from Amacr deficiency.

INTRODUCTION

α-Methylacyl-CoA racemase (Amacr), which is a member of the CAIB-BAIF CoA transferase family (1), has been recently connected to both inborn errors of lipid metabolism and cancer in humans. Amacr-deficiency causes adult-onset sensory motor neuropathy (2) as well as the more severe infantile-onset type of liver dysfunction (3). Another link to human disorders was the finding that Amacr is strongly overexpressed in prostate (4,5) and colon (6) cancer tissues, as revealed by mRNA level assessment. Recent immunohistochemical monitoring of Amacr led to a proposal that this protein can be used as an adjuvant diagnostic indicator for prostatic malignomas in clinical settings (7).

Amacr catalyzes the racemization of acyl-CoA's of α-methyl-branched acids such as di-/trihydroxycoprostanoic acids (D/THCA), which are C27 intermediates of bile acid synthesis, and pristanic acid, which is a phytol metabolite (8–10). Because of the stereo-specificity of acyl-CoA oxidases acting on α-methylacyl-CoAs in liver peroxisomes, where these substrates are β-oxidized, Amacr is regarded as indispensable for their chain shortening (11–13). This is supported further by observations that patients with Amacr deficiency show accumulation of the intermediates of bile acid synthesis and elevated levels of α-branched fatty acids (2,3). Amacr plays also a role in the pharmacological action of 2-arylpropionic acids, e.g. Ibuprofen^® (14,15), which are used as non-steroidal anti-inflammatory drugs. AMACR in human and Amacr in mouse are present as a single gene (locus 5p13.2–q11.1 in human and 15B1 in mouse) and the polypeptide encoded by the open reading frame carries both mitochondrial N-terminal and peroxisomal C-terminal targeting sequences, allowing alternative targeting into these organelles (16–18).

We have generated an Amacr-deficient mouse (Amacr^−/−), which, in agreement with Amacr-deficient patients, showed accumulation of both unconjugated and conjugated bile acid precursors (C27) and a decrease of bile acids (C24). In spite of the disturbed bile acid pattern, the phenotype of the Amacr^−/− mice was clinically symptomless. Although being reduced, the remaining bile acids demonstrated that the Amacr-dependent bile acid synthesis can be partially by-passed (metabolic redundance), allowing essential lipid absorption. However, the Amacr^−/− mice showed intolerance to phytol feeding, suggesting that elimination of methyl-branched fatty acids and their derivatives depends critically on the activity of Amacr.

RESULTS

Disruption of the Amacr-gene in mouse

To disrupt the mouse Amacr gene, exons 1–3 and introns 1–2 were replaced with the neomycin-resistance gene by homologous recombination in R1 ES-cells as described in Materials and Methods and shown schematically in Figure 1. The disruption was confirmed with Southern blot analysis of genomic DNA (Fig. 1B). In line with this, real-time quantitative PCR showed that Amacr was expressed in the liver and kidney of wild-type mice, whereas the disruption resulted in undetectable Amacr expression in mutant mice (Table 1). Immunoblotting with an antibody against mouse Amacr showed the absence of the Amacr-polypeptide in the liver of the homozygous mutant, but its presence in the livers of wild-type and heterozygous mice (Fig. 1C). Immunoblotting from kidney gave the same result (data not shown). These, together with the disappearance of Amacr activity in liver and kidney of Amacr^−/− mice (data not shown), suggested that Amacr is the only enzyme responsible for this activity in these tissues. The Amacr activity was also determined to be absent, within the limit of the assay system used, in the brains and skeletal muscles of all animal groups.

The clinical phenotype of Amacr^−/− mice

Breeding between the heterozygous Amacr^+/− mice yielded litters of normal size representing normal Mendelian distribution of 22% (Amacr^+/+), 54% (Amacr^+/−) and 24% (Amacr^−/−) in over 100 offspring analyzed. This suggests that the intrauterine and postnatal survival rate of Amacr^−/− offspring is not reduced. Amacr^−/− offsprings were normal in appearance, generally healthy and their growth and bodyweights were comparable to age- and sex-matched littermate controls. After reaching adulthood, both males and females were fertile at least up to 6 months of age.

Bile acid precursors accumulate and bile acids are reduced in Amacr^−/− mice

Negative ion mass spectrometry was used, as described in the Materials and Methods, to analyze the content of taurine and glycine conjugated and unconjugated forms of 35 bile acid precursors and bile acids in the bile-fluid of Amacr^+/+, Amacr^+/− and Amacr^−/− mice (see Supplementary Material, Table 1). The wild-type and Amacr^+/− mice had almost exclusively C24 bile acids (the sum of the concentrations being 198±103 mM for wild-type, 179±84 mM for heterozygote) in the bile and only negligible amounts of C27 precursors (the sum being ∼1 mM), while the Amacr^−/− animals had significantly reduced C24 species (98±34 mM) but a strongly increased concentration of C27 precursors (44±11 mM). In all animal groups, the concentrations of taurine conjugated bile acids and bile acid precursors were 100–200 times higher than the concentrations of glycine conjugates and 15–75 times higher than the concentrations of unconjugated metabolites. This is in line with the notion that mice are known to secrete mainly taurine conjugated bile acids (19). Figure 2 shows a representative sampling of taurine conjugated metabolites in mouse bile.

The bile acid patterns in bile were also reflected in sera. Typically, the serum concentrations of both conjugated and unconjugated C27 intermediates were increased by 100–450% in Amacr^−/− mice as compared with wild-type and the concentrations of C24 bile acids were decreased by 50–85% (see Supplementary Material, Table 2). Similarly disturbed patterns for conjugated and unconjugated bile acid metabolites were measured also in liver tissue (see Supplementary Material, Table 3).

No indication of disturbed lipid or lipid-soluble vitamin absorption in Amacr^−/− mice

The decline in the bile acid pool prompted us to study intestinal lipid absorption and possible steatorrhea in Amacr^−/− mice. On regular chow, the total fatty acid contents of the stools of Amacr^+/+ and Amacr^−/− mice were similar (0.10±0.01 mmol/day; n=6, in both groups). Although the groups were subsequently exposed to a triolein-enriched diet (5%, w/w), the fatty acid contents of the stools remained similar (0.14±0.04 mmol/day in both groups).

The analysis of the fatty acid content of stools of mice even on a high-triolein diet showed that the absorption of lipids was not altered upon disruption of Amacr. However, a vitamin K deficiency has been suggested in Amacr-deficient patients due to an elevation of the values in the prothrombin complex assay, which were normalized upon administration of vitamin K (20). Therefore, this assay was taken as an indicator for absorption of lipid-soluble vitamins and function of vitamin K-dependent coagulation cascade. The observed plasma prothrombin complex assay values of wild-type mice (92±20%; n=10) did not differ from the values of Amacr^−/− animals (89±11%; n=10).

Changes in serum lipid values on a phytol-enriched diet

The apparent difference between clinical symptoms in Amacr-deficient patients (2,3) and their absence in Amacr^−/− mice was unexpected. Aiming to uncover the factors behind this difference, 3-month-old mice were fed up to the age of 4.5 months with a chow supplemented with phytol. Phytol is metabolized to phytanic acid (C20) and, further, by peroxisomal α-oxidation, to pristanic acid (C19), which, after activation to a CoA ester, is a substrate for Amacr (8). The dose used (0.5%, w/w) has been described as being tolerated by wild-type mice (21).

After the commencement of the phytol feeding, one out of seven Amacr^−/− mice died on day 9. At the end of the 6-week observation period the phytol feeding had resulted in accumulation of both phytanic and pristanic acids in the serum of surviving Amacr^−/− mice (n=6) when compared with wild-type mice (n=7). On regular chow, the concentrations of phytanic and pristanic acids were 0.06±0.01 and 0.020±0.004 mM in Amacr^−/− mice and 0.07±0.01 and 0.025±0.004 mM in Amacr^+/+ mice. However, on the phytol diet their respective concentrations were elevated (0.52±0.28 and 0.23±0.07 mM; P<0.05 in both cases) in Amacr^−/− mice, but not (0.10±0.01 and 0.022±0.002 mM) in Amacr^+/+ mice (Fig. 3A and B).

When other fatty acids in the serum were analyzed, the concentrations of palmitic (C16 : 0), oleic (C18 : 1) and stearic (C18 : 0) acids were the same in both Amacr^+/+ and Amacr^−/− mice on regular chow (Fig. 3A). Phytol feeding resulted in significantly lower concentrations of these fatty acids in Amacr^−/− mice compared with wild-type animals (Fig. 3B). When comparison was done between mice on regular and phytol chows the stearic acid was decreased by 72% and 84% in both Amacr^+/+ and Amacr^−/− mice, respectively, whereas oleic acid increased by 44% in Amacr^+/+ mice (Fig. 3A and B).

The total serum concentrations of cholesterol and triacylglycerols were also measured from animals on different diets. On both regular and phytol chows, the cholesterol concentrations in Amacr^−/− mice were reduced to about half of that found in the wild-type (n=7; P<0.05 in both cases). On the regular diet, the concentrations of triacylglycerols were the same in both animal groups, whereas, upon phytol feeding, the concentration of triacylglycerols in Amacr^−/− mice reduced by 52% (P<0.05) from the value found in Amacr^+/+ mice (Fig. 4A).

Alterations in physical parameters of Amacr^−/− mice

The bodyweights of Amacr^−/− and Amacr^+/+ mice were the same at the time of introduction of the phytol diet (initial weight), but after 6 weeks on the diet, Amacr^−/− mice weighed 10% less than Amacr^+/+ mice (P<0.05; Fig. 4C). When the weights of epididymal fatpads were monitored and expressed as percentages from the body weight (Fig. 4D), it was observed that fatpads of the Amacr^−/− mice on the phytol diet were 62% less than those of Amacr^−/− mice on regular chow (P<0.05). As compared with the same animal groups on the regular chow, the phytol feeding caused hepatomegaly. However, the increase in the weight-percent of the livers was higher in Amacr^−/− mice (38%; P<0.05) than in Amacr^+/+ mice (17%; P<0.05). No change in food intake was observed between the animals on different diets (data not shown).

Since Amacr-deficient patients suffer from sensory motor neuropathy with upper motor neuron signs in legs (2), mice at the age of 4.5 months were tested for motor coordination and muscular performance by submitting them to the hanging and grid tests, described earlier (22). In the hanging test both Amacr^+/+ (n=6) and Amacr^−/− mice (n=6) reached the end of the 50 cm thread in the 1 min time limit. Amacr^−/− mice also performed similarly to the wild-type mice in the grid test by holding on the grid for 1 min. After phytol-feeding, Amacr^+/+ and Amacr^−/− mice were also subjected to same tests but no differences in performing the tests were observed compared to animals on normal diet.

Phytol causes liver injury in Amacr^−/− mice

When mouse tissues (liver, kidney, skeletal and heart muscles, lung, spleen, eye and brain) were examined using light microscopy, no morphological differences were found in different Amacr genotypes, either on the regular or on the phytol chow, except in the livers. Lobular accumulation of lymphocytes and Kupffer's cells, multivacuolar degeneration and coagulation necrosis were visualized in the livers of Amacr^−/− but not in Amacr^+/+ mice fed with phytol, indicating occurrence of liver damage (Fig. 5A). Oil red O staining of frozen liver sections did not reveal lipid droplets, suggesting that the observed vacuolization in Amacr^−/− mice fed with phytol was not due to accumulation of lipids. Electron microscopy of specimens from the same animals revealed proliferation of smooth endoplasmic reticulum and peroxisomes, which were enlarged and unevenly shaped (Fig. 5B). To examine the liver damage further, the serum concentrations of alanine amino transferase (ALAT) and alkaline phosphatase (APHOS) were measured in mice on different diets. The values of the wild-type animals were taken as references, and they did not change significantly upon phytol feeding. While there was no elevation of APHOS in Amacr^−/− mice on the normal diet, the phytol diet increased APHOS by 175% (P<0.05; Fig. 4B). ALAT in Amacr^−/− mice on the normal diet as well as on the phytol diet was unchanged (Fig. 4B). The ALAT and APHOS values of the Amacr^−/− mouse that died on the phytol diet were excluded from the calculations (850% increase in ALAT and 900% increase in APHOS).

The effects of clofibrate and phytol are different

To further elucidate the role of phytol and concomitant peroxisome proliferation in development of the disease state in the Amacr^−/− mice, the animals were fed with a diet containing clofibrate (0.5%, w/w), a well-documented peroxisome proliferator (23). There were no changes in the liver morphology at light microscopical level, but electron microscopy revealed peroxisome proliferation upon clofibrate in both animal groups. However, the peroxisomes appeared to be normal in size and shape, in contrast to those proliferated in Amacr^−/− mice on the phytol diet (Fig. 5B). Also, a slight elevation of APHOS in serum (Fig. 4B) and increase of the liver weight-percent (Fig 4D) were observed on clofibrate diet. The levels of ALAT, cholesterol, bile acid metabolites, palmitic, oleic, stearic, pristanic and phytanic acids were similar compared with wild-type animals (Figs 3 and 4).

Both inactivation of Amacr and phytol diet evoke changes in gene expression

Treatment of Amacr-deficient mice with phytol caused liver injuries and changes in lipid metabolites. To dissect the ensuing metacrine signaling, the cDNAs from mouse livers were subjected to real-time quantitative PCR to analyze the expressions of a selected set of 25 representative genes (Table 1), encoding peroxisomal proteins of lipid metabolism (Abcd2, Acox1, Acox2, Amacr, pMfe1, pMfe2), proteins involved in cholesterol and bile acid transports and lipid hydroxylations (Abca1, Abcb4, Abcb11, Cyp26a1, Cyp4b1, Cyp7a1) and transcription factors regulating lipid metabolisms (Fxr, Lxrα, Lxrβ, Pparα, Pparβ/δ, Pparγ, Rarα, Rarβ, Rarγ, Rxrα, Rxrβ, Srebp1, Srebp2). On regular chow, the expression ratios (Amacr^−/−/Amacr^+/+ mice) were elevated (1.7–4.5) for Abcb11, Acox1, Acox2, Cyp7a1 and pMfe1 and decreased (0.4) for Cyp26a1. On phytol diet, with the exception of Abcb11, the expression ratios for the same genes were 1.6–18.2 and 0.5 (Table 1). Additionally, on the phytol diet, the expression ratios for Abca1, Abcb4, Abcd2, Lxrβ, Pparγ, Rarβ, Rarγ, Rxrβ and Srebp2 were elevated (1.6–6) and decreased (0.4) for Srebp1 0.4 (Table 1). When tested in separate experiments the expression ratios (Amacr^−/−/Amacr^+/+ mice) for selected genes in oxysterol metabolism, namely for Cyp3a11, Chol25OH and Cyp7b1 were elevated (1.6, 1.4 and 1.5), whereas the ratios for Cyp46a1 and Cyp39a1 were decreased (0.5 and 0.6) in mice fed with regular chow (Table 1).

DISCUSSION

We describe here the generation of an Amacr-deficient mouse strain by targeted disruption of mouse Amacr gene via homologous recombination in embryonic stem cells. The gene inactivation was complete, as demonstrated by the lack of Amacr mRNA, polypeptide recognized by antibody to mouse Amacr and undetectable Amacr activity in the tissues tested. The latter observation also showed that Amacr is the only protein with α-methylacyl-CoA racemase activity in these tissues.

In view of the presumed essential role of Amacr in the peroxisomal chain-shortening of C27 bile acid precursors and ultimately in the synthesis of C24 bile acids (11–13) as well as the disease state of the patients suffering from Amacr-deficiency (2,3), the observation of symptomless phenotype of Amacr^−/− mice was unexpected. In healthy humans, the pool of bile acids and bile acid precursors in serum consists of C24 bile acids and C27 metabolites have been determined to be undetectable (3,24). In contrast, in the Amacr-deficient patients the share of C24 bile acids drops to 14–18% and C27 precursors account for 82–86% of the pool (3,24). The shares of C24 and C27 metabolites are 97 and 3% (7 and 93%) in Amacr^+/+ (Amacr^−/−) mice, respectively. These values demonstrate that the disturbance in the bile acid synthesis due to Amacr deficiency is stronger in mouse than in man.

The lack of steatorrhea in Amacr^−/− mice, even when kept on a high-triolein diet, and the normal function of the vitamin K-dependent coagulation factors demonstrated that the mice had no generalized defect in lipid absorption. Interestingly, sterol 27-hydroxylase (CYP27A1)-deficient mice, similar to Amacr^−/− mice, have decreased levels of bile acids and still the absorption of lipids is normal in these mice (25). The observed undisturbed lipid absorption may in part be explained by the reduced, but still considerable, amount of C24 bile acids found in Amacr^−/− mice and in part by the substitution of bile acid function with C27 precursors (which are the only form of bile acids in many amphibia and reptiles) (26).

The phytol-enriched diet evoked the disease state in the Amacr^−/− but not in Amacr^+/+ mice. Sick mice developed hepatomegaly, lobular accumulation of lymphocytes and degeneration of hepatocytes together with elevated serum levels of APHOS. As histological liver damage was not observed in the mice fed with clofibrate, which showed enlargement of livers and proliferation of peroxisomes (which were normal in size and shape in contrast to large deformed organelles of Amacr^−/− mice on the phytol diet), the effects of phytol cannot be ascribed alone to metabolic changes associated to peroxisome proliferation (e.g. production of free oxygen radicals). Therefore we assume that the acute toxic effects are mediated by phytol metabolites.

The observed peroxisome proliferation in livers of Amacr^−/− mice is probably transmitted via peroxisome proliforator activated receptor α (PPARα), since accumulating phytol metabolites, phytanic and pristanic acids (Fig. 3), are known ligands of PPARα (21,27). Furthermore, the expressions of Abcb4, Abcd2, pMfe1 and Acox1, which encode peroxisomal proteins and are controlled by PPARα (28–30), were induced, while the expression of Cyp4b1, which is controlled by PPARγ (31,32), was not induced. One further observation was the decrease of stearic acid in serum of both wild-type and Amacr^−/− mice upon phytol but not clofibrate feeding (Fig. 3). This is an additional finding showing different metabolic responses to phytol and clofibrate. A further experiment to analyze the effect of phytol feeding on oleic acid level showed that the expression of stearoyl-CoA desaturase 1 (Scd1) was increased in animals on phytol diet (unpublished data).

Patients with peroxisomal multifunctional enzyme type 2 (pMFE-2) deficiency as well as pMFE-2 knock-out mice accumulate C27 bile acid precursors in their serum, suggesting that the peroxisomal chain-shortening of the cholesterol tail is dependent on this enzyme (33–35). The metabolic flux linking C27 bile acid precursors to pMFE-2 goes through activation of D/THCAs to CoA esters, racemization by Amacr to their (25S)-diastereomers and subsequent oxidation to (24E)-Δ²⁴-D/THCA-CoAs. The hydratase component of pMFE-2 converts (24E)-Δ²⁴-D/THCA-CoAs into (24R,25R)-hydroxy-D/THCA-CoAs, which are oxidized by the dehydrogenase part of the same enzyme to 3-keto-D/THCA-CoAs. The hydratase part of peroxisomal multifunctional enzyme type 1 (pMFE-1) can also catalyze the hydration of (24E)-Δ²⁴-D/THCA-CoA (36), but the product is (24S,25S)-hydroxy-D/THCA-CoA, which is a substrate for neither pMFE-1 nor pMFE-2. (24S,25S)-Hydroxy-D/THCA-CoA is a substrate for Amacr, producing (24S,25R)-hydroxy-D/THCA-CoA, a substrate for pMFE-1 (36,37). The combined action of the hydratase in pMFE-1, Amacr and the dehydrogenase in pMFE-1 has been used to explain the C24 bile acids found in pMFE-2 deficiencies (37).

However, (24S,25R)-hydroxy-D/THCA-CoA can also be predicted to arise in an Amacr-independent manner through oxysterols via at least two different sequences of hydroxylations in the body (Fig. 6). When characterizing the source of the residual bile acids in CYP27A1-deficient mice, which do not develop cerebrotendinous xanthomatosis (CTX) as deficient humans do (25), Honda et al. (38) demonstrated that CYP3A4, which corresponds to CYP3A11 in mice, among other hydroxylations, catalyzes (24S)-hydroxylation of oxysterols. This hydroxylation results in the generation of (24S,25R)-metabolites, potential precursors of substrates for pMFE-1. In line with the idea that CYP3A11 contributes to the bile acid synthesis, the gene expression of Cyp3a11 was moderately increased in livers of Amacr-deficient mice.

Alternatively, (24S,25R)-metabolites can also be generated from cholesterol by cholesterol 24-hydroxylase (CYP46A1), which is expressed in brain, liver and testis in mice (39,40). However, the contribution of this enzyme was challenged by the observation that the expression level of Cyp46A1 was decreased in livers of Amacr^−/− mice.

Additional credence for the proposed participation of pMFE-1 in bile acid synthesis in Amacr-deficient mice came from the observation that, upon disruption of Amacr, the expression of pMfe1 was increased, but that of pMfe2 was not. The proposition that pMFE-1 contributes to the metabolism of oxysterols and ultimately to the synthesis of bile acids in an Amacr-independent manner can be tested in the future by generating a double knock-out, e.g. crossing Amacr^−/− and pMfe1^−/− mouse strains. It is noteworthy that the pMFE-1-deficient mice are symptomless (41,42).

The existence of at least two parallel, although partially overlapping, peroxisomal β-oxidation pathways with different substrate specificities results in highly variable clinical presentations in different deficiencies in these pathways. The pMFE-2 deficiency in mouse causes severe growth retardation, premature death of some of the siblings, accumulation of methyl-branched fatty acids and bile acid intermediates, and impeded peroxisomal oxidation of very long fatty acids (33). Although, the biochemistry of both Amacr^−/− and pMfe2^−/− mice shows many common features, the severe phenotype of the latter mice can be at least partially explained in terms of accumulation of very long chain fatty acids. This phenotype was not detected in Amacr^−/− animals with the serum concentrations of C24 : 0 and C26 : 0 being 18.3±3.9 (16.4±3.5) and 29.5±6.6 (30.7±8.8) µM, respectively, in Amacr^+/+ (Amacr^−/−) mice.

Comparison between phenotypes of Sterol carrier protein-X/Sterol carrier protein-2 (SCP_X/SCP2; a SCP_X 3-ketothiolase) and Amacr-deficient mice reveal striking similarities. The Scp_X/Scp2^−/− mice developed normally up to 6 months of age and morphology of their major organs appeared to be normal. However, Scp_X/Scp2^−/− mice are sensitive to phytol, developing liver disease with ultimate death within 3 weeks on a phytol diet (43). Peroxisomal β-oxidations, whether proceeding via either pMFE-2 or pMFE-1, merge at 3-ketoacyl intermediates. In the case of bile acid synthesis both MFE's generate 3-ketoD/THCA-CoA, a substrate for SCP_X 3-ketothiolase. Although Scp_X/Scp2^−/− mice show accumulation of abnormal bile acid intermediates, normal bile acids are still synthesized (44), suggesting that the specificity of peroxisomal thiolase(s) is (are) not as strict as is generally thought.

Bile acids are known to regulate the feedback from the bile acid pool to bile acid synthesis by acting as ligand(s) for transcription factors, such as farnesoid X receptor (FXR) (45–48). Amacr^−/− mice had low concentrations of chenodeoxycholic acid, the physiological ligand of FXR (44), which explains why the expression of Cyp7a1, which is downregulated by FXR, was enhanced in Amacr^−/− mice (Table 1). Cyp7a1 encodes cholesterol 7α-hydroxylase (CYP7A1) (49). CYP7A1-deficient mice appeared normal at birth, but died within the first 18 days of life, suffering from symptoms such as fat malabsorption (50). However, the postnatal lethality was reversed by bile acid and vitamin supplementation.

When compared with wild-type controls, the Amacr^−/− mice had about 50% reduction in serum cholesterol in all feeding groups. Interestingly, a moderate reduction in serum cholesterol has been described also in Amacr-deficient patients (24). The decrease in serum cholesterol in Amacr^−/− mice can be explained at least partially in terms of accelerated channeling of cholesterol towards bile acids. Although the Amacr^−/− mice did not show a defect in intestinal triacylglycerol absorption, the possible isolated defect in cholesterol absorption and its contribution to the low serum cholesterol is currently under investigation in our laboratory. In contrast to cholesterol, a significant decrease in serum triacylglycerols as well as in the size of the epididymal fatpad was developed only upon phytol feeding in Amacr^−/− mice.

We conclude that the key physiological role of Amacr is the elimination and detoxification of methyl-branched fatty acids. The lack of symptoms in Amacr^−/− mice on standard diet can be explained in that the regular laboratory-chow, low with phytol, is actually a therapeutic diet. Similar elimination of phytol from the diet might be a treatment for patients suffering from Amacr-deficiency. Amacr is still an important enzyme for bile acid synthesis, but the proposed oxysterol-pMFE-1 route may circumvent the Amacr-dependent pathway and provide a fall-back system that suffices for survival of Amacr-deficient individuals.

MATERIALS AND METHODS

Generation of Amacr^−/− mice

Fragments of 2.5 and 4.8 kb of the BACM-13L1-clone (derived from 129/SvJ mouse strain by Genome systems Inc., St Louis, MO, USA), which was previously shown to contain Amacr (16), were cleaved with AseI, SmaI, ClaI and EcoRI restriction endonucleases and ligated to the targeting vector as 5′ and 3′ flanking regions of the PGK-neo cassette, which was enveloped with two Lox-P-sites (Fig. 1A). The neomycin-resistance (neo) gene was for positive and thymidine kinase (TK) gene for negative selection. The targeting vector was linearized with SalI restriction endonuclease before electroporation into R1 embryonic stem (ES) cells (derived from mouse strains 129/Sv-+^c-Tyr, + ^p x 129/SvJ). The ES cells, 10⁷, were electroporated in a 1 ml cuvette at 240 V and 500 µF using a Bio-Rad gene pulser. After 24 h, the cells were placed under selection with 200 µg/ml G418 (GIBCO/BRL, Gaithersburgh, MD, USA) and 2 µM ganciclovir (Syntex, Palo Alto, CA, USA) for 7–9 days.

ES cell colonies that survived the double selection were isolated and screened by Southern blot analysis for correct disruption of exons 1–3 with the PGK-neo-cassette. Southern analysis was performed on genomic DNA extracted from ES-cells with Blood and Cell Culture DNA Midi Kit (Qiagen, Hilden, Germany) and digested with EcoRI restriction endonuclease. The blot was hybridized at 65°C using ³²P-labeled (Random Primed Labelling Kit, Amersham Pharmacia Biotech, Buckinghamshire, UK) 1 kb fragment upstream of the 5′ flanking region as an external probe (Fig. 1A). The filter was autoradiographed in Phosphocassette and the cassette analyzed with Phosphorimager and Image Quant software (Molecular Dynamics, Buckinghamshire, UK). The normal (wild-type) allele gave a band corresponding 13 kb and the targeted (knock-out) allele a band corresponding 7 kb (Fig. 1B).

Amacr^+/− ES-cell aggregates from positive clones were microinjected into C57BL/6 blastocysts to generate chimeric mice, which transmitted disrupted Amacr through the germline. Genotypes of mice were identified with PCR analyses of tail DNA samples. Three primers were used in PCR analyses (Fig. 1A): the forward primer for the normal allele was a sequence in intron 2 (5′-ACCTTCCTGACTGTCCAAAGTAT-3′) and the forward primer for the targeted allele was a sequence in the neo-cassette (5′-TCGCATTGTCTGAGTAGGTGTCA-3′). A reverse primer for both alleles was a sequence in intron 3 (5′-AAGCAAAGGACCTCTTGCCAGA-3′). The sizes of PCR products for the normal allele and the targeted allele were 800 and 450 bp, respectively.

Heterozygous offspring were interbred to produce Amacr^−/−, Amacr^+/− and Amacr^+/+ mice, which were used in all experiments. All animal experimentation was done on adult mice (4.5 month), unless stated otherwise. The plan for this research was approved by the University of Oulu committee of animal experimentation.

Immunoblotting and activity measurements

For SDS–PAGE and immunoblotting, tissue samples from mouse liver, kidney, brain and muscle were homogenized respectively. Small pieces of frozen (−20°C) tissues were weighed and homogenized in PBS to give 1 : 100-homogenates. SDS–PAGE and following immunoblotting was then performed on these 1 : 100-homogenates using a polyclonal anti-rabbit antibody for mouse Amacr. For protein determination the 1 : 100-homogenates were diluted 1 : 10 with PBS and protein concentrations were determined at 595 nm with a photometer using Protein Assay chemistry (BIO-Rad Laboratories, Hercules, CA, USA). Measurements of Amacr-activity were performed as previously described (8).

Real-time quantitative PCR

Expression levels of Amacr and 24 additional genes were measured by real-time quantitative PCR using cDNA, synthesized from total RNA with First Strand cDNA Synthesis Kit (MBI Fermentas, Heidelberg, Germany). The RNA was extracted from mouse livers with QuickPrep^TM Total RNA Excraction Kit (Amersham Pharmacia Biotech). Real-time quantitative PCR was performed with the ABI PRISM 7700 Sequence Detection System (Perkin Elmer, Boston, MD, USA) and TaqMan chemistry as previously described (51). Sense and antisense primers for the genes were used and the products were detected using fluorogenic probes. The results were normalized to 18S-rRNA quantified from the same samples. 18S-fluoro was used as a probe for 18S amplicon detection.

Blood samples

Mice were weighed and anesthetized by intraperitoneal administration (10 µl/g body weight) of a mixture of 0.7% Ketalar^® (Pfizer, Espoo, Finland) and 0.004% Domitor^® (Orion Pharma, Espoo, Finland). The blood was collected by orbital bleeding with subsequent euthanasia by cervical dislocation and harvesting of the bile and tissues. Analysis of serum cholesterol, triacylglycerols, ALAT and APHOS were done by TG-diagnostics, Oulu, Finland. The prothrombin complex assay was done by the clinical laboratory of the University Hospital of Oulu, Finland.

Mass spectral analysis of bile acid derivatives and fatty acids

Mass spectral data were obtained using an APEX II FT–ICR mass spectrometer controlled by Xmass 6.0.0 software and equipped with a 7 T active shielded magnet, an APOLLO electrospray ion source and a syringe pump (Bruker Daltonics, Billerica, MA, USA). The sample flow rate was 2 µl/min. Nitrogen was used as the drying and nebulizing gas. The drying gas temperature was maintained at 150°C. The detection range was set to 150–600 m/z. The accumulation time in the hexapole was set to 0.5 s. For each measurement, 256 scans were accumulated. When measuring in the positive ion mode a voltage of −4.5 kV was applied to the sprayer needle and the capillary exit voltage was set to 70 V. When measuring in the negative ion mode a voltage of 3.5 kV was applied to the sprayer needle and the capillary exit voltage was set to −115 V. The routine mass accuracy was <1 ppm, the mass resolution >25 000. All raw data were processed using the Xmass software. Calibration for curves were linear for deoxycholic acid, cholic acid, glycocholic acid, taurodeoxycholic acid, taurocholic acid and the dimethylaminoethyl esters of pentadecanoic acid and heptadecanoic acid from 5 nM to 50 µM.

Sample preparation.

Bile samples of 20 µl were diluted with 980 µl H₂O for determination of bile acid derivates. A 10 µl aliquot of standard solution [10 µM lithocholic acid, 10 µM glycolithocholic acid and 200 µM tauro lithocholic acid in CH₃CN–10 mM NH₄OAc (50/50; v/v)] and 188 µl CH₃CN–10 mM NH₄OAc (50/50; v/v) was added into 2 µl of this bile-dilution and 8 µl of the resulting solution was used for the recording of negative ion mass spectra.

For determination of bile acid derivates from serum, 50 µl serum, 1 ml H₂O, 10 µl of standard solution [1 mM pentadecanoic acid and heptadecanoic acid, 0.05 mM lithocholic acid, 0.025 mM glycolithocholic acid and 0.005 mM tauro-lithocholic acid in CH₃CN–10 mM NH₄OAc (50/50; v/v)] were added to 25 µl RP-18 and mixed. The supernatant was removed and the residue was washed with 1 ml H₂O. Bile acids were extracted three times with 400 µl CH₃CN–H₂O (50/50; v/v). Extracts were evaporated and the residue was resuspended in 30 µl CH₃CN–10 mM NH₄OAc (50/50; v/v). Recording of mass spectral data was done as described above.

For determination of bile acid derivates in liver, 20 mg of liver was homogenized in 180 µl H₂O using a Potter homogenizer. A 50 µl aliquot of standard solution [2.5 mM pentadecanoic acid, 0.2 mM lithocholic acid, 0.2 mM glycolithocholic acid and 20 µM taurolithocholic acid in CH₃CN–10 mM NH₄OAc (50/50; v/v)] and 250 µl PBS were added to the homogenate and the resulting suspension was mixed. A 50 µl aliquot of RP-18 was added to the supernatant, the suspension mixed, the supernatant removed, the residue washed twice with 1 ml H₂O and lipids were extracted three times with 400 µl CH₃CN–H₂O (50/50; v/v). The combined extracts were evaporated under a stream of nitrogen at 95°C and the residue was resuspended in 50 µl CH₃CN–10 mM NH₄OAc (50/50; v/v). A 4 µl aliquot of this solution was used for the recording of negative ion mass spectra.

For determination of fatty acids in serum, 10 µl serum and 10 nmol each of pentadecanoic acid and heptadecanoic acid were suspended in 0.2 ml 0.5 M HCl in CH₃CN–H₂O (9/1) and heated for 1 h at 95°C. A 190 µl aliquot of 1 M KOH was added and the solution was again heated for 1 h at 95°C. A 100 µl aliquot of 5 M HCl was added and the fatty acids were extracted twice with 500 µl hexane. The combined extracts were evaporated. A 50 µl aliquot of a solution of 5% oxalylchloride in CH₃CN (v/v) was added and evaporated after 5 min heating at 50°C. A 50 µl aliquot of a solution of 5% dimethylaminoethanol in CH₃CN was added and evaporated after 5 min. The residue was resuspended in 1.5 ml methanol–H₂O–AcOH [49.5/49.5/1; (v/v/v)]. Recording of mass spectral data was done in positive mode as described above.

Histological analysis

Tissue samples for light microscopy were fixed in 4% paraformaldehyde, embedded in paraffin, sectioned with a Leica RM 2165 microtome and stained with hematoxylin–eosin. Frozen samples were sectioned with a Reichert-Jung, 2800 Frigocut cryomicrotome and the Oil red O staining was done with standard methods. The sections were examined with a Olympus BX50 light microscope equipped with Analysis^® software (Soft Imaging Systems, Münster, Germany).

For transmission electron microscopy, tissue samples were fixed in 2.5% glutaraldehyde, postfixed in 1% osmiumtetroxide, dehydrated in acetone and embedded in Epon Embed 812 (Electron Microscopy Sciences, Fort Washington, PA, USA). Thin sections were cut with a Reichert Ultracut ultramicrotome and examined in a Philips CM100 transmission electron microscope. Images were captured by CCD camera equipped with TCL-EM-Menu version 3 (Tietz Video and Image Processing Systems, Gaunting, Germany).

Special diets

Chow was supplemented with 5% (w/w) triolein (C18 : 1, [cis]-9, Sigma Aldrich, Steinheim, Germany) to study the malabsorption of lipids. Titrimetrical determination of total fatty acids from collected stools was done by Yhtyneet Laboratoriot, Helsinki, Finland. To study the effects of accumulating phytol derivatives and peroxisome proliferation, either 0.5% (w/w) phytol (Sigma Aldrich) or 0.5% (w/w) clofibrate (ICN Biomedicals, Aurora, OH, USA) were added into the chow.

Statistics

All values are expressed as mean±standard deviation (SD) unless stated otherwise. All data were analyzed by the Student's t-test for significant differences between mean values of each group.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online.

ACKNOWLEDGEMENTS

We wish to thank graduate student Eija Selkälä for her enthusiastic work in the laboratory and Dr A.-P. Kvist and laboratory assistant Marika Kamps for excellent technical assistance. This study was supported by Sigrid Juselius Foundation, Academy of Finland and Research and Science Foundation of Farmos.

Figure 1.Amacr-disruption strategy and verification of the gene inactivation. (A) Schematic drawing of the knock-out strategy. The 1 kb genomic DNA fragment used as an external probe in Southern analysis of EcoRI-digested DNA is marked as Probe. Primers used in determination of mouse genotypes from tail DNA are denoted as small arrows. All exons are denoted as boxed numbers 1–5. NEO=neomycine-resistance gene and TK=thymidine kinase gene. Restriction sites for endonucleases are marked as A=AseI, S=SmaI, C=ClaI and E=EcoRI. (B) Southern blot of liver DNA shows bands corresponding to the wild-type allele (13 kb) in Amacr^+/+ mouse, the disrupted allele (7 kb) in Amacr^−/− mouse and both alleles in Amacr^+/− mouse. (C) Western blot of liver homogenates shows the presence and the absence of the band corresponding to Amacr (41.7 kDa) in Amacr^+/+, Amacr^+/− and Amacr^−/− mice. The asterisk indicates the predicted position of the absent truncated Amacr (∼25 kDa); 360 ng of protein was applied to each lane. The molecular mass standards are given on the left.

Figure 2. Taurine conjugated bile acids and bile acid derivatives in mouse bile. The taurine conjugated metabolites shown are: di-OH-C24 : 1=dihydroxycholenoic acid; di-OH-C24 : 0=chenodeoxycholic acid (dihydroxycholanoic acid); tri-OH-C24 : 1=cholenoic acid (trihydroxycholenoic acid); tri-OH-C24 : 0=cholic acid (trihydroxycholanoic acid); tetra-OH-C24 : 1=muricholenoic acid (tetrahydroxycholenoic acid); tetra-OH-C24 : 0=muricholic acid (tetrahydroxycholanoic acid); di-OH-C27 : 1=dihydroxycoprostenoic acid; di-OH-C27 : 0=dihydroxycoprostanoic acid (DHCA); tri-OH-C27 : 1=trihydroxycoprostenoic acid; tri-OH-C27 : 0=trihydroxycoprostanoic acid (THCA); tetra-OH-C27 : 1=tetrahydroxycoprostenoic acid; tetra-OH-C27 : 0=varanic acid (tetrahydroxycoprostanoic acid). Heights of the bars indicate the means±SD of five to seven independent observations. Note the logarithmic scale.

Figure 3. Fatty acids in mouse serum. Serum contents of palmitic acid (C16 : 0), oleic acid (C18 : 1), stearic acid (C18 : 0), pristanic acid (C19 : 0/pr) and phytanic acid (C20 : 0/ph) in mice on (A) normal chow, (B) phytol diet and (C) on clofibrate diet. Heights of the bars indicate the means±SD of five to seven independent observations.

Figure 4. Serum lipids, liver marker enzymes and bodyweights of Amacr^+/+ and Amacr^−/− mice. (A) Serum concentrations (mM) of cholesterol (CHOL) and triacylglycerol (TRI). (B) Serum levels (U/l) of alanine amino transferase (ALAT) and alkaline phosphatase (APHOS) in serum. (C) Initial weights (g) and the weights (g) at the end of the 6-week feeding experiment. (D) Percentages of bodyweights (%) of livers and epididymal fatpads. Genotypes and diets concerning all panels are indicated at the bottom of the figure. Heights of the bars indicate the means±SD of five to seven independent observations.

Figure 5. Microscopical examination of livers from Amacr^+/+ and Amacr^−/− mice. (A) Light microscopy of haematoxylin–eosin-stained sections of livers of untreated (1,2) and phytol-treated (3,4) mice in 84× magnification demonstrating that phytol-treatment caused degeneration in Amacr^−/− mice. (B) Electron microscopy of livers of untreated (1,2), phytol-treated (3,4) and clofibrate-treated (5,6) mice. The most striking enlargement and proliferation of peroxisomes is seen in phytol-treated Amacr^−/− mice. m=mitochondria, p=peroxisome and the bar is 50 nm.

Figure 6. The proposed Amacr-independent route for bile acid synthesis. The schematic picture shows the Amacr-pMFE-2-dependent route for branched-chain fatty acid catabolism (on the left) and bile acid synthesis (in the middle). The proposed route for bile acid synthesis, which utilizes (24S)-hydroxylated metabolites of cholesterol and the (3S)-hydroxyacyl-CoA dehydrogenase activity of pMFE-1, and bypasses Amacr and pMFE-2, is shown on the right. (24S)-hydroxylated intermediates can be generated in reactions catalyzed by CYP3A11 or CYP46A1 (see text for details).

Table 1.

The expression of a selected set of genes in livers of Amacr^+/+ and Amacr^−/− mice upon phytol feeding. The expression ratios for genes were calculated as expression in Amacr^−/−/Amacr^+/+ mouse, for animals fed with regular and phytol chows, respectively. Values are means of five to seven independent observations

Gene	Accession no.	Expression ratio (Amacr−/−/Amacr+/+)		Function of the gene product
		Regular chow	Phytol chow
Abca1	NM_013454	1	1.7	Phospholipid/cholesterol transporter (rate limiting enzyme in reverse cholesterol transport)
Abcb4	NM_008830	1.3	3.5	Essential for secretion of phospholipids into the bile (multidrug resistance 2)
Abcb11	NM_021022	1.8	1	Bile salt export pump
Abcd2	NM_011994	1	6.0	Promote fatty acid transport into peroxisomes
Acox1	NM_015729	1.7	2.4	Acyl-coenzyme-A oxidase 1 (palmitoyl)
Acox2	NM_053115	1.7	1.8	Acyl-coenzyme-A oxidase 2 (branched-chain)
Amacr	NM_008537	0 in Amacr−/−	0 in Amacr−/−	Racemize branched-chain lipids
Cyp26a1	NM_007811	0.44	0.49	Deactivates all-trans retinoic acid
Cyp4b1	NM_007823	1	1	Monooxygenase (putative substrates: steroids and eicosanoids)
Cyp7a1	NM_007824	4.5	2.2	Cholesterol 7α-hydroxylase (rate-limiting enzyme in bile acid biosynthesis)
Fxr	NM_009108	1	1	Farnesoid X receptor (bile acid receptor)
Lxrα	NM_013839	1	1	Liver X receptor (oxysterol receptor)
Lxrβ	NM_009473	1	1.6	Liver X receptor (oxysterol receptor)
pMfe1	AK004867	2.9	18.2	Peroxisomal multifunctional enzyme type 1
pMfe2	NM_008292	1	1	Peroxisomal multifunctional enzyme type 2
Pparα	NM_011144	1	1	Peroxisomal proliferator-activated receptor α
Pparβ/δ	U10375	1	1	Peroxisomal proliferator-activated receptor β/δ
Ppparγ	NM_011146	1	2.5	Peroxisomal proliferator-activated receptor γ
Rarα	NM_009024	1	1	Retinoic acid receptor α
Rarβ	S56660	1	3.1	Retinoic acid receptor β
Rarγ	NM_011244	1	5.6	Retinoic acid receptor γ
Rxrα	NM_011305	1	1	Retinoid X receptor α
Rxrβ	NM_011306	1	1.8	Retinoid X receptor β
Srebp1	AB017337	1	0.36	Sterol regulatory element binding protein 1
Srebp2	XM_127995	1	2.3	Sterol regulatory element binding protein 2
Cyp3a11	NM_007818	1.6	—	Testosterone 6β-hydroxylase
Cyp46a1	NM_010010	0.46	—	Cholesterol 24-hydroxylase
Cyp39a1	NM_018887	0.59	—	Oxysterol 7α-hydroxylase
Chol25OH	NM_009890	1.4	—	Cholesterol 25-hydroxylase
Cyp7b1	NM_007825	1.5	—	Oxysterol 7α-hydroxylase