Lifelong genetic minipumps

Kathleen M. I. Caron1, Leighton R. James2, Gene Lee3, Hyung-Suk Kim3 and Oliver Smithies3

1 Department of Cell and Molecular Physiology, University of North Carolina-CH, Chapel Hill, North Carolina
2 Department of Medicine, University of Texas at Southwestern, Dallas, Texas
3 Department of Pathology and Laboratory Medicine, University of North Carolina-CH, Chapel Hill, North Carolina


    ABSTRACT
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Most physiologists working with animals are familiar with osmotic minipumps. These surgically implanted devices can, for a limited period, administer a reagent at a constant predetermined rate that is unaffected by concurrent procedures. The investigator can then test the physiological effects of other treatments knowing that the animals’ homeostatic responses will not be able to alter the dose of the pumped reagent. To develop the genetic equivalent of a lifelong minipump, simply inherited as an autosomal dominant, we here combine three of our previously described strategies, genetic clamping, single-copy chosen-site integration, and modification of untranslated regions (UTRs). As a test of the procedure, we have generated a series of intrinsically useful animals having genetic minipumps secreting renin ectopically from the liver at levels controlled by the investigator but not subject to homeostatic changes. To achieve the different dosage levels of these genetic minipumps, we altered the UTRs of a renin transgene driven by an albumin promoter and inserted it into the genome as a single copy at the ApoA1/ApoC3 locus, a locus that is strongly expressed in the liver. The resulting mice express plasma renin over ranges from near physiological to eightfold wild type and develop graded cardiovascular and kidney disease consequent to their different levels of ectopically secreted renin. The procedure and DNA constructs we describe can be used to generate genetic minipumps for controlling plasma levels of a wide variety of secreted protein products.

gene targeting; transgene; hypertension; renin-angiotensin system


    INTRODUCTION
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
FOR NEARLY THIRTY YEARS, OSMOTIC minipumps have been used by investigators to study the physiological effects of a constantly dosed reagent (20). Surgically implanted minipumps have the advantages of ectopically delivering a reagent at a dose chosen by the investigator that is unaffected by the animal’s inherent homeostatic adjustments. Other manipulations can then be performed on the animal without altering the dose of the infused reagent. However, some complications of osmotic minipumps include their transient or short-term nature and the need for surgical implantation and recovery. To overcome some of the limitations of classic osmotic minipumps and to develop improved methods of investigator-controlled gene expression, we have combined three previously described genetic engineering strategies into a procedure that generates a series of mice having the genetic equivalent of lifelong osmotic minipumps of different strengths.

The first strategy requires the construction of a genetically clamped transgene coding for a secreted protein and driven by a promoter that is insensitive to homeostatic feedback (3). The second strategy, single-copy chosen-site integration, controls for the positional effects that influence expression by inserting the transgene into a chromosomal locus neutral with respect to the phenotype of interest (1). The third strategy, modification of the untranslated regions (UTRs) of the transgene, allows the investigator to control the level of expression of the transgene. Many studies (7, 22), including our own (9), have demonstrated that 3'-UTRs and downstream G-rich elements (GREs) play important roles in determining mRNA stability and influence final gene expression levels. Likewise, using both in vitro and in vivo systems, Kozak (12, 13) has demonstrated that lengthening the 5'-UTR of a transgene results in increased levels of protein production. Subsequent investigators have confirmed the importance of the length of the 5'-UTR and its secondary structure on the kinetics and efficiency of protein translation (15, 21). We incorporated modifications of both the 5'- and 3'-UTRs of our transgene to increase or decrease its expression.

We have chosen to illustrate our procedure using renin as the secreted protein, because the resulting animals are useful for studies that require different levels of activity of the renin-angiotensin system (RAS) free from the extremely powerful renin-mediated homeostatic control of the RAS that often hinders investigations of the physiological effects on the cardiovascular system of pharmacological agents or of exploratory gene mutations of interest to the investigator. We demonstrate that the combined result of genetic clamping, single-copy chosen-site integration, and modification of 5'- and 3'-UTRs is a versatile and effective procedure for generating a graded series of lifelong genetic minipumps that result in in vivo plasma renin levels not subject to cardiovascular homeostasis ranging from normal to nearly eightfold wild type, comparable to those seen in human patients with hypertension. This range of expression leads to phenotypes of graded severity that include hypertension, cardiac hypertrophy, cardiac fibrosis, and renal glomerular sclerosis. The resulting mice consequently provide valuable tools for cardiovascular studies requiring different fixed levels of activity of the RAS. The transgenes that produce the three levels of genetic minipump are inherited simply (as autosomal dominants) and so can easily be combined with other genetic manipulations. The genetic minipump concept and the DNA constructs that we have developed should be adaptable to a wide variety of secreted gene products.


    EXPERIMENTAL PROCEDURES
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
The general design of the targeting vectors to generate renin transgenic mice by homologous recombination in embryonic stem cells has been described previously (11). Figure 1 (below) illustrates the three vectors used in the present work. RenTgKC is the vector described as RenTg in reference (3). RenTgARE has as an insert a 75-bp AU-rich element synthetically generated by complimentary oligonucleotides and cloned into a BamHI restriction site immediately 3' to the stop codon of RenTgKC. The insert includes Domain I of the RNA destabilizing element of the c-fos gene (5) (5'-TAAATATCTGAGAATCCATCTTAATAAATAAATTAAAAACACAATAAAAG-3') and terminal restriction sites used for subcloning. RenTgMK has a 22-bp translation enhancing sequence (20-bp MK), designed by Marilyn Kozak (12) (5'-CTCGACCCACTACACA-3'), inserted into an NheI restriction site in the 5'-UTR of RenTgKC.



View larger version (19K):
[in this window]
[in a new window]
 
Fig. 1. Constructs used for generating the RenTg series mice. A: the basal, unmodified renin transgene targeting vector. BamHI restriction sites flank the sequences that code for the secreted Ren2/1d protein. TK, thymidine kinase; epi, a c-myc epitope tag; Neo, neomycin resistance gene. B: the 3'-untranslated region (3'-UTR) destabilizing modification, c-fos ARE, introduced into the basal transgene; its sequence is: 5'-TAAATATCTGAGAATCCATCTTAATAAATAAATTAAAAACACAATAAAAG-3'. C: the 5'-UTR stabilizing modification, 20-bp MK, introduced into the basal transgene; its sequence is 5'-CTAGCTCGACCCACTACACA-3'. D: the endogenous ApoA1/ApoC3 target locus into which homologous recombination introduces the three transgenes. Details of the cloning strategies used to insert the c-fos ARE and 20-bp MK sequences into the basal transgene are given in the EXPERIMENTAL PROCEDURES.

 
For PCR-based genotyping of the three RenTg mouse lines, we use three primers: p1, 5'-TGGGATTCTAACCCTGAGGACC-3'; p2, 5'-CACAGATTGTAACTGCAAATCTGTCG-3'; and p3, 5'-GTTCTTCTGAGGGGATCGGC-3'. The wild-type alleles produce a 430-bp band, and the targeted alleles produce a 330-bp band. In this study we used heterozygous single-copy RenTg male mice on an isogenic SvEv129/6 background.

Renin mRNA levels were measured from total RNA liver and kidney extracts using quantitative RT-PCR with the ABI 7700 sequence detection system as previously described (10). The renin primers were forward, 5'-ACAGTATCCCAACAGGAGAGACAAG-3'; reverse, 5'-GCACCCAGGACCCAGACA-3'; and probe, 5'-FAM-TGGCTCTCCATGCCATGGACATCC-Tamra-3'. The primers and probe for the ß-actin internal standard were forward, 5'-AAGAGCTATGAGCTGCCTGA-3'; reverse, 5'-ACGGATGTCAACGTCACACT-3'; and probe 5'-TET-CACTATTGGCAACGAGCGGTTCCG-Tamra-3' (where FAM, Tamra, and TET are the fluoroprobe marker dyes).

To enrich plasma samples for renin protein, a member of the aspartyl protease group of enzymes which bind to and are inhibited by pepstatin A, we incubated equal amounts of plasma from wild-type and RenTg mice with pepstatin A agarose beads (Sigma P2032) in Triton-PBS Buffer (PBS, pH 7.4, 10 mM EDTA, 1% Triton X-100, 60 mg/ml PMSF) for 1 h with mild shaking at 4°C following a published protocol (17). Pellets were washed twice with Triton-PBS buffer, and renin heavy chain (4) was eluted from the agarose beads in a denaturing loading buffer at 95°C for 5 min. Equal volumes of samples were electrophoresed on 4–15% gradient PAGE gels and immunoblotted using a c-myc antibody (Roche 1814150) that detects the c-myc epitope attached to our modified renin protein (see Fig. 1 below).

Radioimmunoassay for angiotensin I (ANG I), a measure of renin plasma concentration, was performed as previously described (19) using a commercially available kit from NEN Life Science Products.

Blood pressure was measured by a computerized tail-cuff system as previously described (14).

For general histology, 5-µm thick paraformaldehyde-fixed paraffin sections were stained with hematoxylin and eosin (H and E) and Masson’s trichrome reagents.

For renal function studies, mice were placed in metabolic cages for 3 days with free access to food and water. Urine osmolality was determined by freezing point depression. Urine protein was measured at the Animal Clinical Chemistry Core Facility at UNC-CH using a colorimetric assay and the VT250 Chemical Analyzer (Diagnostic Chemical).

Statistical analyses were performed with JMP software (SAS Institute, Cary, NC). Values are presented as means ± SE.

All experiments were approved by the Institutional Animal Care and Use Committee of the University of North Carolina-Chapel Hill.


    RESULTS
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
Generation of RenTg targeting vectors and mouse lines.
Figure 1 illustrates diagrammatically the constructs used to generate three illustrative mouse lines having renin genetic minipumps of different strengths. Each transgene is incorporated into the genome as a single copy at the liver-specific ApoA1/ApoC3 locus (Fig. 1D). The three transgenes are all driven by a strong and well-characterized albumin promoter/enhancer (AlbP/E) that has previously been shown to be active only in the liver (3, 18). The renin coding sequence (Ren2/1d) is a synthetic cDNA consisting of parts of the Ren-2 and Ren-1d genes modified to include glycosylation sites for increased stability, a furin cleavage site to enable prorenin to active renin processing to occur in the liver and allow secretion of active renin into the blood stream, and a c-myc epitope tag for protein immunodetection (16). The renin cDNA in the primary targeting construct (Fig. 1A) is flanked by two BamHI restriction sites, which allow its removal and replacement by cDNA coding for any other secreted protein, so that other genetic minipumps can be easily engineered employing our targeting vector and strategy. The renin transgenes have, at their 3' ends, the ß-globin 3'-UTR (ß-globin), which stabilizes the transcript (9). This construct (Fig. 1A) and the mouse line resulting from it, referred to as RenTg in reference (3), are hereafter designated RenTgKC.

To reduce the expression level of the RenTgKC transgene, we altered the half-life of its mRNA by incorporating immediately 3' to the stop codon an AU-rich element (c-fos ARE) from the c-fos gene (5), which generally destabilizes mRNA and decreases half-life by a factor of about 10 (9). This construct (Fig. 1B) and the resulting mouse line are referred to below as RenTgARE. To increase transgene expression levels over that of the basal RenTgKC line, we inserted a 20-bp oligonucleotide sequence (20-bp MK) devoid of secondary structure (12) into the 5'-UTR of the transgene to increase the translation efficiency of its mRNA. We refer to this construct (Fig. 1C) and the resulting line of mice as RenTgMK, after Marilyn Kozak, who first described the sequence and used this strategy to increase translation (12).

To eliminate positional effects inherent to conventional transgenes, we used homologous recombination in embryonic stem cells to integrate a single copy of each of our transgenes into the genome at the same position in the ApoA1/ApoC3 locus (Fig. 1D). We chose this locus because of its strong expression in the liver (8). An added benefit is that the efficiency of homologous recombination at this locus is high (8); ~80% of neomycin and ganciclovir selected colonies are correctly targeted.

The heterozygous mice used in the present studies have one copy of chromosome 9 with the wild-type ApoA1/ApoC3 locus and the other with the incorporated renin transgene. They are wild type at the natural renin gene locus. In deriving these mice, all matings were with inbred mice of the same strain as the embryonic stem cells (SvEv129/6), and so they are genetically isogenic.

Graded levels of renin transgene expression in the series of RenTg mice.
We have previously demonstrated that expression of our renin transgene is restricted to the liver (3) as a consequence of the use of a liver-specific promoter/enhancer (18). We used quantitative RT-PCR to measure the levels of renin transgene mRNA expression from the livers of the three RenTg mouse lines. Wild-type control animals, as expected, expressed no significant amounts of renin in their livers (5 ± 2 pg/µg total RNA) (Fig. 2A). The RenTgKC and RenTgMK lines had the highest levels of liver mRNA expression (264 ± 32 pg/µg total RNA and 275 ± 60 pg/µg total RNA, respectively) followed by the RenTgARE line (95 ± 20 pg/µg total RNA, p vs. RenTgKC < 0.0001). Thus mRNA expression in the RenTgARE line was decreased to about one-third of that in the RenTgKC line, reflective of the mRNA destabilizing element incorporated into its 3'-UTR. There was no significant difference in the level of transgene mRNA expression between the RenTgMK and RenTgKC lines. These results demonstrate first that these single-copy renin transgenes mediate renin mRNA expression in the liver and second that the incorporation of an AU-rich destabilizing element into the mRNA decreased transgene expression about threefold in vivo.



View larger version (35K):
[in this window]
[in a new window]
 
Fig. 2. Modifications in the 5'- and 3'-UTR cause graded expression of renin transgene mRNA, plasma protein levels, plasma concentration, and endogenous renin gene expression. A: quantitative RT-PCR of renin transgene mRNA from liver extracts of wild-type (WT) and the RenTg series mice. B: Western blot with a c-myc epitope antibody of aspartyl protease-enriched plasma from wild-type and the RenTg series mice. The predicted size of the synthetic renin heavy chain resulting from the transgene is 36 kDa. A prominent band, RenTg(hc) can be visualized at the appropriate size in the two highest expressing lines but is undetectable in the RenTgARE or wild-type plasma. C: plasma renin concentration as nanograms of ANG I produced per milliliter per hour at 37°C. D: quantitative RT-PCR of endogenous renin mRNA from kidney extracts of wild-type and RenTg series mice. *Differs significantly from wild type, P < 0.0001 by ANOVA. #Differs significantly from RenTgKC, P < 0.0001, by ANOVA. Rank order P values are by nonparametric Spearman rho. To avoid the statistically dominant effects of the phenotypes of the RenTgs relative to wild type, the P values for the nonparametric rank order Spearman rho test within the series of RenTg mice and exclude data from the wild-type mice. The numbers of mice in each category are in parentheses in this and subsequent Figs. 36; all mice were male.

 
To test the effectiveness of lengthening the 5'-UTR on in vivo protein translation and renin protein expression, we used an antibody directed against the c-myc epitope tag attached to our modified renin transgene protein to immunoblot aspartyl protease-enriched plasma extracts from wild-type and the renin transgenic mouse lines. As expected and shown in Fig. 2B, epitope-positive protein of the expected size (33 kDa) was not detected in the extracts of plasma from wild-type mice. Nor was Ren2/1d protein detectable in RenTgARE plasma, presumably because its level is below that detectable with the current assay. However, a band of ~33 kDa, which we identify as renin heavy chain (4), is expressed in the extracts of plasma of RenTgKC mice and at approximately twice this level in the RenTgMK mouse plasma. Thus the data indicate that addition of a stabilizing element in the 5'-UTR of the transgene results in increased protein production in vivo relative to that resulting from the same transgene lacking the modification.

To determine whether the ectopically produced renin was being properly secreted as an active form into the circulation, we measured the plasma renin concentration in the series of transgenic mice. As shown in Fig. 2C, plasma renin concentration was ~6 times higher than wild-type levels (24 ± 4 ng ANG I·ml–1·h–1) in the RenTgKC (154 ± 27 ng ANG I·ml–1·h–1; P < 0.0001) and ~8 times higher in the RenTgMK (180 ± 20 ng ANG I·ml–1·h–1; P < 0.0001) mice. In contrast, the RenTgARE mice display the same renin plasma concentration (23 ± 7 ng ANG I·ml–1·h–1) as wild-type mice, demonstrating that decreasing mRNA stability of a transgene by insertion of an ARE element into the 3'-UTR results in decreased protein production.

An important homeostatic consequence of the renin genetic minipump is a substantial shutdown of endogenous renin production in the kidneys to less than 1/3 normal. Shown in Fig. 2D are the amounts of endogenous renin mRNA in the kidneys of wild-type (18 ± 2 pg/µg total RNA), RenTgARE (6 ± 1 pg/µg total RNA), RenTgKC (5 ± 1 pg/µg total RNA), and RenTgMK (3 ± 1 pg/µg total RNA) mice. The differences between the endogenous expression of renin did not reach statistical significance between the three RenTg mice, but there was a trend toward lower (almost undetectable) expression in the higher transgenic expressing lines. Taken together, the data presented in Fig. 2 establish that 5'-UTR and 3'-UTR modifications of a renin transgene result in the graded ectopic production of a secreted physiologically active protein in vivo.

Graded RenTg levels result in physiologically graded cardiovascular disease.
Using a computerized tail-cuff system, we find that the RenTgARE, RenTgKC, and RenTgMK mice have progressively elevated blood pressure (124.9 ± 5, 138 ± 6, 146 ± 5 mmHg) compared with wild-type littermates (116 ± mmHg; P < 0.0001) (Fig. 3A). Thus the renin transgenes with progressively higher levels of expression have progressively elevated blood pressures. Among the series of genetically graded RenTg mice, the graded blood pressures are significant by a Spearman rho rank order test (P < 0.003).



View larger version (22K):
[in this window]
[in a new window]
 
Fig. 3. Graded increases in blood pressure and cardiac hypertrophy in the RenTg series mice. A: mean blood pressure measured with computerized tail cuff. B: left ventricular weight to body weight (LV/BW) ratio. *Differs significantly from wild type, P < 0.001 by ANOVA. Rank Order tests exclude wild type.

 
All of the RenTg mouse lines display marked cardiac hypertrophy relative to wild type, as measured by left ventricular/body weight (LV/BW) ratio and depicted in Fig. 3B (P < 0.0001), and again the extent of the condition progressively increases among the RenTg lines (Spearman rho rank order P < 0.01). The increase in cardiac hypertrophy with increasing levels of RenTg expression is accompanied by an increase in cardiac fibrosis, which occurs at progressively earlier ages, as shown in Fig. 4, which displays representative Masson’s trichrome histological sections of left ventricles from the RenTg series of animals. The mean LV/BW ratios (yellow numbers in Figs. 4 and 6) were significantly increased in all RenTg mice at all ages compared with age-matched wild-type controls (black numbers in bottom line). At 6–8 mo of age, the highest expressing line, RenTgMK, had marked fibrosis that is probably the cause of our previous observation of electrocardiographic disturbances and sudden death in this line (2). The RenTgKC mice, with somewhat lower levels of renin expression, do not develop severe fibrosis until 1 yr of age, and we have not observed any significant premature or sudden deaths in this line. Finally, we note that although the RenTgARE mice have cardiac hypertrophy, they are indistinguishable from age-matched wild-type controls with respect to fibrosis, and like the RenTgKC mice, they do not die prematurely. Thus, in summary, the 5'- and 3'-UTR modifications of the targeted renin transgene produce graded cardiovascular disease that is phenotypically distinguishable in vivo and readily detected microscopically postmortem.



View larger version (71K):
[in this window]
[in a new window]
 
Fig. 4. Cardiac hypertrophy and fibrosis in the series of RenTg mice is progressively affected by RenTg expression and age. Images of histological sections of the left ventricle stained with Masson’s trichrome. Sections were taken from the mid-wall region and include at least one medium-sized artery to reflect the degree of cardiac and vascular fibrosis. Rows are grouped by genotype, and columns are grouped by age. The yellow numbers represent the mean calculated LV/BW ratio for each group of mice. The black numbers in the bottom row represent the mean calculated LV/BW ratio for wild-type mice at different ages. The LV/BW ratios are calculated from an n value of at least 6 animals per group, except the ">1 year" RenTgMK group, for which only two animals survived. *Differs significantly from wild type, P < 0.0001. #Differs significantly from wild type, P < 0.04.

 
Graded kidney disease in the higher expressing RenTgKC and RenTgMK lines.
We compared renal-related metabolic parameters of the RenTg mouse lines to look for any alterations in homeostasis and kidney pathology. We find that the amount of water drunk (Fig. 5A) and urine volumes (Fig. 5B) progressively increase and urine osmolality (Fig. 5C) progressively decreases from wild type to the strongest transgene, indicating a graded inability of the RenTg series mice to concentrate urine. There is also a graded increase in urine protein concentration in the RenTg series mice, which is indicative of kidney glomerular pathology (Fig. 5D). Note, however, that the renal parameters of the RenTgARE mice are indistinguishable from wild type.



View larger version (24K):
[in this window]
[in a new window]
 
Fig. 5. Graded kidney phenotypes as assessed by metabolic parameters in the series of RenTg mice. A: 24-h water intake. B: 24-h urine volume. C: urine osmolality. D: urine protein levels. *Differs significantly from wild type, P < 0.001, by ANOVA. Rank Order tests exclude wild type.

 
Masson’s trichrome staining of kidney sections from 6- to 8-mo-old RenTg mice confirms this graded pathology. Thus the kidneys of the RenTgARE mice are indistinguishable from wild type, but the RenTgKC and RenTgMK kidneys show a progressively greater incidence of glomerulosclerosis, vascular fibrosis, tubulointerstitial fibrosis, immune cell infiltration, and proteinaceous casts (Fig. 6). These results demonstrate that the higher expressing RenTg lines suffer from severe kidney disease, which parallels the deterioration in their kidney function, but that the mice with the genetically clamped RenTgARE do not differ significantly from wild type in any of these renal-related parameters tested. Therefore, the 5'- and 3'-UTR modifications incorporated into the basal RenTgKC transgene result in graded levels of RenTg expression that produce a graded renal phenotype ranging from essentially normal (in the RenTgARE mice) to severely compromised (in the RenTgMK mice).



View larger version (151K):
[in this window]
[in a new window]
 
Fig. 6. Graded kidney phenotypes in the series of RenTg mice assessed by Masson’s trichrome stain of histological sections in 6- to 8-mo-old mice. A: wild type. B: RenTgARE, indistinguishable from wild type. C: RenTgKC. D: RenTgMK. Note arteriolar fibrosis (yellow arrows), glomerular sclerosis (black arrows), and proteinaceous casts (asterisk) in the kidneys of RenTgKC and RenTgMK mice. Immune cell infiltration (i) and tubulointerstitial fibrosis (tf) are shown.

 

    DISCUSSION
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
In this work, we have combined three previously described strategies (genetic clamping, single-copy chosen-site integration, and UTR modifications) to develop a procedure to generate mice having the genetic equivalents of lifelong minipumps that control the plasma concentration of a chosen protein. This approach yielded a series of animals that exhibit increased or decreased expression of a single-copy transgene without changing either the promoter or the chromosomal surroundings of the transgene. The transgenes incorporate two features that remove them from homeostatic feedback of the phenotypes which they affect. First, the promoter/enhancer pair that drives the transgenes and the chromosomal region into which they are inserted as single copies are both active in a tissue that is not influenced by their effects. Second, differences in the levels of expression of the transgenes were obtained by altering the UTRs of their RNAs. In this way, the promoters and chromosomal environment of the transgenes are kept constant. As indicated above, we have chosen to illustrate the effectiveness of the genetic minipump concept by generating a series of mice in which the plasma concentration of active renin is fixed at graded levels not subject to the powerful homeostatic control of the RAS normally mediated by alterations in renin synthesis. The mice have been submitted to the Mutant Mouse Regional Resource Center (MMRRC) at The University of North Carolina at Chapel Hill, Chapel Hill, NC, for access by interested investigators.

The overall effectiveness of our graded renin minipumps was evaluated by cardiovascular phenotyping. Thus, of the five distinct phenotypic measurements that we quantified in the RenTg series [kidney renin mRNA, blood pressure, cardiac hypertrophy, ability to concentrate urine (as judged by combining water intake, urine volume, and urine osmolality), and urine protein], four were significant in a rank order Spearman rho test (P < 0.03). Moreover, when we used the Fisher method (6) to test for the combined significance of these related but independent variables, we found a highly significant (P < 0.0001) overall difference between the three types of RenTg mice. It is therefore clearly evident that by modifying the 5'- and 3'-UTRs of a transgene one can achieve graded control of a secreted protein that, in the case of renin, results in graded cardiovascular phenotypes which are comparable to those seen with conventional osmotic minipumps delivering graded doses of angiotensin II.

Some aspects of the cardiovascular phenotype require comment. First, the mRNA destabilizing element (c-fos ARE) that was incorporated into the 3'-UTR of the RenTgARE transgene achieves the important goal of leading to the ectopic production of near normal plasma renin levels. The RenTgARE mice, consequently, have essentially normal blood pressures, metabolic parameters, and kidneys. Second, we find that the onset and degree of cardiac hypertrophy in the three RenTg lines is earlier and greater than expected from their elevated blood pressures, and we suggest that this is probably due to the additional direct and indirect effects on cardiac tissue of increased embryonic and lifelong RAS activity. We are somewhat puzzled by the cardiac fibrosis and sudden death (2) of 6- to 8-mo-old mice having the RenTgMK transgene that yields only a small (~1/6th) increase in active plasma renin levels relative to the mice having the RenTgKC, which survive normally. However, we note that the RenTgKC mice eventually develop cardiac fibrosis that is nearly indistinguishable from that which develops earlier in the RenTgMK line.

In summary, three strategies (genetic clamping, single-copy chosen-site integration, and UTR modification) can be effectively combined in a gene targeting procedure that results in investigator-controlled genetic minipumps that are simply inherited as autosomal dominants. We have engineered our targeting vector in such a way that it can easily be modified to produce genetic minipumps coding for any secreted protein of interest or for an engineered precursor of a peptide hormone (16). Future development of the genetic minipump concept might include replacing the albumin promoter/enhancer, used in our current examples, with an inducible promoter.


    GRANTS
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 
This work was supported by National Heart, Lung, and Blood Institute Grants HL-71266 (to O. Smithies) and HL-10344 (to K. M. Caron) and by a Biomedical Fellowship, Kidney Foundation of Canada (to L. R. James).


    ACKNOWLEDGMENTS
 
We thank Dr. Nobuyo Maeda for helpful advice and discussions and Gleb Rozanov for technical help.


    FOOTNOTES
 
Article published online before print. See web site for date of publication (http://physiolgenomics.physiology.org).

Address for reprint requests and other correspondence: O. Smithies, D. Phil, Dept. of Pathology and Laboratory Medicine, CB 7525, 701 Brinkhous-Bullitt Bldg., Univ. of North Carolina-CH, Chapel Hill, NC 27599-7525 (E-mail: jenny_langenbach{at}med.unc.edu).

10.1152/physiolgenomics.00221.2004.


    REFERENCES
 TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS
 DISCUSSION
 GRANTS
 REFERENCES
 

  1. Bronson SK, Plaehn EG, Kluckman KD, Hagaman JR, Maeda N, and Smithies O. Single-copy transgenic mice with chosen-site integration. Proc Natl Acad Sci USA 93: 9067–9072, 1996.[Abstract/Free Full Text]
  2. Caron KM, James LR, Kim HS, Knowles J, Uhlir R, Mao L, Hagaman JR, Cascio W, Rockman H, and Smithies O. Cardiac hypertrophy and sudden death in mice with a genetically clamped renin transgene. Proc Natl Acad Sci USA 101: 3106–3111, 2004.[Abstract/Free Full Text]
  3. Caron KM, James LR, Kim HS, Morham SG, Sequeira Lopez ML, Gomez RA, Reudelhuber TL, and Smithies O. A genetically clamped renin transgene for the induction of hypertension. Proc Natl Acad Sci USA 99: 8248–8252, 2002.[Abstract/Free Full Text]
  4. Catanzaro DF, Mullins JJ, and Morris BJ. The biosynthetic pathway of renin in mouse submandibular gland. J Biol Chem 258: 7364–7368, 1983.[Abstract/Free Full Text]
  5. Chen CY, Chen TM, and Shyu AB. Interplay of two functionally and structurally distinct domains of the c-fos AU-rich element specifies its mRNA-destabilizing function. Mol Cell Biol 14: 416–426, 1994.[Abstract]
  6. Fisher RA. Statistical Methods, Experimental Design, and Scientific Inference. New York: Hafner, 1972.
  7. Guhaniyogi J and Brewer G. Regulation of mRNA stability in mammalian cells. Gene 265: 11–23, 2001.[CrossRef][ISI][Medline]
  8. Hatada S, Kuziel W, Smithies O, and Maeda N. The influence of chromosomal location on the expression of two transgenes in mice. J Biol Chem 274: 948–955., 1999.[Abstract/Free Full Text]
  9. Kakoki M, Tsai YS, Kim HS, Hatada S, Ciavatta DJ, Takahashi N, Arnold LW, Maeda N, and Smithies O. Altering the expression in mice of genes by modifying their 3' regions. Dev Cell 6: 597–606, 2004.[CrossRef][ISI][Medline]
  10. Kim HS, Lee G, John SW, Maeda N, and Smithies O. Molecular phenotyping for analyzing subtle genetic effects in mice: application to an angiotensinogen gene titration. Proc Natl Acad Sci USA 99: 4602–4607, 2002.[Abstract/Free Full Text]
  11. Koller BH and Smithies O. Inactivating the beta 2-microglobulin locus in mouse embryonic stem cells by homologous recombination. Proc Natl Acad Sci USA 86: 8932–8935, 1989.[Abstract]
  12. Kozak M. Effects of long 5' leader sequences on initiation by eukaryotic ribosomes in vitro. Gene Expr 1: 117–125, 1991.[ISI][Medline]
  13. Kozak M. Leader length and secondary structure modulate mRNA function under conditions of stress. Mol Cell Biol 8: 2737–2744, 1988.[ISI][Medline]
  14. Krege JH, Hodgin JB, Hagaman JR, and Smithies O. A noninvasive computerized tail-cuff system for measuring blood pressure in mice. Hypertension 25: 1111–1115, 1995.[Abstract/Free Full Text]
  15. Meijer HA and Thomas AA. Control of eukaryotic protein synthesis by upstream open reading frames in the 5' -untranslated region of an mRNA. Biochem J 367: 1–11, 2002.[CrossRef][ISI][Medline]
  16. Methot D, vanKats JP, Lochard N, Tremblay F, Silversides DW, and Reudelhuber TL. Development and application of a biological peptide pump for the study of the in vivo actions of angiotensin peptides. Am J Hypertens 14: 38S-43S, 2001.
  17. Norling LL, Smith TM, and Ingelfinger JR. Stored renin isoforms in the developing rat kidney. Am J Hypertens 11: 213–218, 1998.[CrossRef][ISI][Medline]
  18. Pinkert CA, Ornitz DM, Brinster RL, and Palmiter RD. An albumin enhancer located 10 kb upstream functions along with its promoter to direct efficient, liver-specific expression in transgenic mice. Genes Dev 1: 268–276, 1987.[Abstract]
  19. Sealey JE. Plasma renin activity and plasma prorenin assays. Clin Chem 37: 1811–1819, 1991.[Abstract]
  20. Struyker-Boudier HA and Smits JF. The osmotic minipump: a new tool in the study of steady-state kinetics of drug distribution and metabolism. J Pharm Pharmacol 30: 576–578, 1978.[ISI][Medline]
  21. Wilkie GS, Dickson KS, and Gray NK. Regulation of mRNA translation by 5'- and 3'-UTR-binding factors. Trends Biochem Sci 28: 182–188, 2003.[CrossRef][ISI][Medline]
  22. Wilusz CJ, Wormington M, and Peltz SW. The cap-to-tail guide to mRNA turnover. Nat Rev Mol Cell Biol 2: 237–246, 2001.[CrossRef][ISI][Medline]