We report the therapeutic effects of liver-specific expression of a short-chain acyl-CoA dehydrogenase (SCAD) transgene in the SCAD-deficient mouse model. Transgenic mice were produced with a rat albumin promoter/enhancer driving a mouse SCAD minigene (ALB-SCAD) on both the SCAD normal genetic background and a SCAD-deficient background. In three transgenic lines produced on the SCAD-deficient background, recombinant SCAD activity and antigen in liver mitochondria were found up to 7-fold of normal control values. All three lines showed a markedly reduced organic aciduria and fatty liver, which are sensitive indicators of the metabolic abnormality seen in this disease found in children. We found no detrimental effects of high liver SCAD expression in transgenic mice on either background. These studies provide important basic and practical therapeutic information for the potential gene therapy of nuclear-encoded mitochondrial enzyme deficiencies, as well as insights into the mechanisms of the disease.
Mitochondrial enzyme deficiencies comprise an important group of potentially fatal, incurable human diseases that primarily affect children. These include mitochondrial enzyme deficiencies that are involved in catabolism of amino acids, fatty acids and other intermediary metabolites. One subgroup of these disorders is the fatty acid [beta]-oxidation defects, notably the acyl-CoA dehydrogenase deficiencies (1 ). These diseases often have a recurrent, potentially fatal phenotype resembling Reye syndrome, with severe hypoglycemia, organic acidemia, hyperammonemia and fatty change of the liver, as well as sudden death (1 ).
Nuclear encoded-mitochondrial enzymes are expressed at a wide range of levels in a wide variety of tissues, depending on the metabolic needs of the tissue. The acyl-CoA dehydrogenase genes, involved in the [beta]-oxidation of fatty acids, have the highest level of expression in cardiac muscle, followed by skeletal muscle, kidney and liver (2 ). Independent deficiencies in each of the straight-chain acyl-CoA dehydrogenases [SCAD, medium-chain acyl-CoA dehydrogenase (MCAD), long-chain acyl-CoA dehydrogenase (LCAD) and very long-chain acyl-CoA dehydrogenase (VLCAD)] have been described in human patients (1 ). All are inherited as autosomal recessive traits. Deficiency of MCAD is the most common of the four, with an estimated frequency of 1 in 15 000 (3 ) individuals. Individuals with SCAD deficiency have been reported (4 ,5 ). The metabolic phenotype is primarily that of deficient hepatic adaptation to fasting. A mouse model with a spontaneous deletion mutation in the SCAD gene has been identified and characterized in BALB/cByJ mice (6 -9 ). Molecular analysis has shown the mutation to be a 278 bp deletion at the 3' end of the SCAD gene (9 ). No SCAD antigen is detectable in these mice; likewise, SCAD activity is absent (10 ). The phenotype of this mouse model mimics that seen in human patients, with a severe organic aciduria and development of a fatty liver. High levels of ethylmalonic and methylsuccinic acid and N-butyrylglycine are excreted in the urine (6 ). This mouse model provides a unique opportunity to investigate correction of this and similar [beta]-oxidation defects, as well as other mitochondrial enzyme deficiencies.
In these experiments, we tested the hypotheses that liver-specific expression of SCAD would be sufficient to correct the phenotype of the disorder, and that abnormally high expression of SCAD, which is tightly regulated in vivo, would not be detrimental to hepatocytes. We found previously in preliminary cell culture experiments that putative high level expression of a SCAD cDNA construct appeared to be detrimental to mouse fibroblasts (C.L. Kelly and P.A. Wood, unpublished results). Although SCAD is expressed in multiple tissues, the phenotype of the disorder is primarily a liver disease phenotype. When considering gene therapy for these disorders, it is essential to understand which tissues need to be provided with a wild-type copy of the gene. Since there are great difficulties with in vivo gene transfer and expression, transgenic experiments were used to test the hypotheses posed without encountering the current limitations of in vivo gene transfer.
We tested the therapeutic potential of liver-specific expression of SCAD using a transgene composed of a rat albumin promoter/ enhancer ligated to a mouse SCAD minigene (Fig. 1 ) (ALB-SCAD) on the SCAD-deficient background. To produce transgenic lines on the SCAD-deficient background, female BALB/cByJ mice homozygous for SCAD deficiency were mated with (BALB/cByJ*C57BL/6J) F1 males that are heterozygous for SCAD deficiency. Fertilized eggs from this mating were microinjected with the ALB-SCAD minigene. Seven founder mice were positive for the transgene and were then mated to homozygous SCAD-deficient BALB/cByJ mice to produce transgenic offspring on a homozygous SCAD-deficient background.
Of the seven founder transgenic mice, five passed the transgene to the offspring. Southern analysis showed that four of these lines contained one integration site, with varying copy numbers of the transgene in each line (3-5 copies/genome). The fifth line contained multiple integration sites, and was not studied further. Northern blot analyses of liver RNA demonstrated normal or increased levels of expression of the ALB-SCAD transcript in three lines 590, 597 and 604 (Fig. 2 ). In line 604, there was also a trace amount of expression in the kidney and skeletal muscle. The fourth line with a single integration site, line 603, did not express the transgene. Immunoblot analysis of the liver mitochondrial SCAD antigen from one or more pups from each of the three SCAD mRNA-expressing transgenic lines (590, 597, 604) correlated with the Northern analyses. Enzyme analysis showed increased SCAD activity for these lines. Line 590 had SCAD activity equal to that of normal control mice, while line 597 had 5- to 7-fold the normal mean activity, and line 604 had 2-fold the normal activity (Fig. 3 ). The fatty liver changes characteristic of even the non-fasted SCAD-deficient mice were reduced or totally corrected in the transgenic lines (Fig. 4 ). Likewise, the prominent organic aciduria found in J mutants was markedly reduced to totally eliminated in the transgenic lines (Fig. 5 ). Both of these measures are sensitive indicators of metabolic correction of this disorder.
We have shown here that restoring normal or increased levels of liver SCAD activity in SCAD-deficient mice virtually corrects the organic aciduria and fatty change of liver, key indicators of metabolic well being of patients with many of the mitochondrial enzyme deficiencies. In the normal animal, SCAD is expressed in multiple tissues (2 ); however, for treatment purposes, it will be more practical to genetically correct only one of these tissues. SCAD-deficient heterozygous mice, as in SCAD homozygous normal mice, excrete no abnormal organic acids in urine. Thus, the trace urinary organic acids seen in the ALB-SCAD-corrected transgenic lines are most probably being generated by non-hepatic tissues such as skeletal muscle and heart. From this work, it would appear that the majority of abnormal metabolites are generated by the liver, and restoring enzyme activity to liver would probably provide great benefit to reducing the disease phenotype in the human patient. These results have important practical implications for potential gene therapy correction of this and similar mitochondrial enzyme deficiencies in human patients.
Endogenous expression of SCAD, which is involved in the first step of the [beta]-oxidation of fatty acids, was shown to be highest in cardiac muscle, liver, kidney and skeletal muscle (2 ). Our results demonstrate that increased liver expression of SCAD over background was not detrimental. This is in contrast to situations described by others where increased expression of specific transgenes caused disease in mice (11 -13 ). For example, expression of the phosphenolpyruvate carboxykinase (PEPCK) gene 7-fold above normal in transgenic mice led to the development of non-insulin-dependent diabetes mellitus (11 ).
In summary, we found that increased expression and activity of SCAD, a tightly regulated gene, is not detrimental to liver. Expression of a mouse SCAD minigene driven by a rat albumin promoter/enhancer producing SCAD activity up to seven times the normal activity virtually corrected the mutant phenotype. Therefore, these results would indicate that many of the human mitochondrial enzyme deficiency diseases may potentially be treatable by gene therapy directed primarily to liver.
A 353 bp mouse SCAD cDNA fragment, spanning from position -9 to 344 of the coding sequence (14 ) was digested from the entire coding cDNA with EcoRI, which digested the 5' end from the plasmid, and BspHI, which digested the cDNA internally. A partial mouse SCAD genomic clone (15 ) was digested with BspHI, which cut at the same site in exon 3 as in the cDNA, and XbaI, which cut the remaining genomic sequence from the plasmid. These two fragments were ligated together, giving a mouse SCAD minigene that contained from bp -9 through exon 3 of cDNA sequence, and the SCAD genomic sequence spanning from intron 3 through the polyadenylation signal sequence. This minigene was ligated to two specific linkers. The first, matching up to the 5' end of the minigene, contained an EcoRI overhang, a BamHI overhang and an intact SalI site. The second linker, matching up to the 3' end of the minigene, contained an XbaI overhang, a ClaI overhang and an intact NheI site. This was then ligated to a 1.8 kb portion of the rat albumin (ALB) promoter/ enhancer contained in the plasmid 2219-3 ALB-hGH (16 ) at the BamHI and ClaI sites. The human growth hormone sequence previously had been removed from this construct. The 6.5 kb ALB-SCAD minigene used for microinjection was isolated from this construct by an NheI digest (Fig. 1 ). The microinjection fragment was purified using a low melting point agarose gel isolation, followed by GeneClean (BIO 101) and two ethanol precipitations.
To test the hypothesis that liver-specific expression of the transgene would correct the phenotype seen in these mutant mice, the minigene construct was microinjected into the fertilized ova of BALB/cByJ*(BALB/cByJ*C57BL/6J) F1 mice as described previously (17 ). Originally, the construct was to be injected directly into the fertilized ova of BALB/cByJ mice, but they proved to be inefficient for microinjection, and we were unable to produce any transgenic mice from these injections. The use of the C57BL/6J hybrid strain also introduced a normal SCAD allele into the background of the mice; however, the founders from these injections were mated back to the BALB/cByJ mutant mice, and only transgenic offspring confirmed to be SCAD homozygous deficient were analyzed further. The BALB/cBy and BALB/cByJ mice originated from the Jackson laboratory and were maintained as a colony at the University of Alabama at Birmingham. C57BL/6J*BALB/cByJ F1 mice were obtained from the Jackson Laboratory. To test the hypothesis that increased expression over normal would be detrimental, the transgene construct was microinjected into the fertilized ova of C57BL/6J*SJL/J F2 hybrid mice (17 ). These hybrid mice have normal SCAD activity.
Transgenic mice were identified originally by PCR amplification of the cDNA portion of the minigene. DNA was extracted via established methods from ~2 cm of tail biopsy at 3 weeks of age. The PCR analysis for the presence of the transgene was done with primers specific for the 5' cDNA portion of the transgene (see Fig. 1 ) (TGTTGCGTCAGACATGCCGTGA and TGATAACTCCCGTGGAGGCGCAGG). Amplifications were carried out in a total volume of 100 [mu]l including 1* PCR buffer (50 mM KCl, 10 mM Tris-HCl, 0.1% Triton X-100 and 1.5 mM MgCl2), 1.2 mM dNTPs, 15 pmol of each primer and 4.5 U of Taq polymerase. Southern blots were then done on the transgene-positive DNA samples to determine the endogenous SCAD genotype of these mice. Ten [mu]g of the genomic DNA was digested with EcoRI and NheI. The resulting fragments were separated by electrophoresis on a 0.8% agarose gel, transferred to Zeta probe membrane (BioRad) and hybridized with a partial mouse SCAD cDNA. Probes were radiolabeled with [32P]dCTP by random hexamer priming (18 ). Hybridizations were performed as described previously (9 ).
Total RNA was isolated from five different tissues harvested from F1 progeny of lines 590, 597, 603 and 604 using standard methods (9 ). The tissues analyzed included cardiac muscle, skeletal muscle, liver, kidney and brain. Two separate riboprobes were used for Northern analysis. The first was generated from the full-length coding sequence of the mouse SCAD cDNA subcloned into pGEM-3Zf+ (Promega). The second riboprobe was made from the mouse wild-type SCAD cDNA sequence (exon 9) that was specific to the deletion mutation found in the BALB/cByJ mutant mice, and spanned from nucleotide 1030 to 1164 (9 ). This deletion- specific sequence was subcloned into pCRII (Invitrogen) and used to screen identical blots. This probe was specific to the transgene- derived or normal transcript.
Mitochondria were isolated from liver of transgenic and control mice following standard methods (6 ,19 ).
Immunoblot analysis was performed using pig liver SCAD antibody, as previously described (10 ). Determination of SCAD activity levels in liver mitochondria was performed as described previously (4 ,6 ,10 ).
Urine was collected from control mice, transgenic mice and transgenic negative littermates. All mice were approximately the same age, and were fed the same diet. Urines were extracted and derivatized following our standard protocol then injected and analyzed by gas chromatography-mass spectrometry (6 ).
We thank Richard Palmiter for providing the albumin promoter vector, David Kurtz for assistance with the RNA analysis, and William E. O'Brien and Arthur W. Warman for GC-MS analysis of the urinary organic acids. We also thank the entire staff of the UAB Transgenic Animal/ES Cell Resource for their dedicated assistance with this project. The oligonucleotides used in this work were produced by the UAB Comprehensive Cancer Center oligonucleotide facility, Jeff Engler, Director. This work was supported by National Institutes of Health grant R01-RR02599 (P.A.W.).
*To whom correspondence should be addressed. Tel: +1 205 934 1303; Fax: +1 205 975 4418; Email: cmed004@uabdpo.dpo.uab.edu
Human Molecular Genetics
Pages
Introduction
Results
Discussion
Materials And Methods
SCAD transgene construction
Transgenic mouse production
DNA analysis
RNA analysis
Mitochondrial isolation
Enzymology
Urine organic acid analysis
Acknowledgements
References
REFERENCES
This page is maintained by OUP admin. Last updated Wed Aug 13 15:52:16 BST 1997. Part of the OUP Journals World Wide Web service. Copyright Oxford University Press, 1996