Cardiac-Specific Inducible and Conditional Gene Targeting in Mice

Background of Mouse Gene Targeting and Transgenic Technology

Pluripotent cells that supported transplantable teratomas were reported in spontaneous teratomas from specific mouse strains.^12,13 Teratocarcinoma cell lines were derived from those teratomas,¹⁴ and their pluripotency was demonstrated by blastocyst injection, but with little germline efficiency.^15,16 ES cells, on the other hand, were isolated directly from the inner cell mass of mouse embryos,^17,18 were subsequently shown to colonize the germline at practical frequencies,¹⁹ and were shown to be capable of transmitting transgenes through the germline²⁰ in a generationally stable fashion.²¹ Gene targeting was achieved in ES cells through homologous recombination,^22,23 and those ES cells could be used to stably transmit the targeted gene into a new strain of mouse.^24,25 Finally, it was demonstrated that genes not expressed in ES cells could be targeted at the same frequency as expressed genes.²⁶ Reviews on developments in gene targeting are available.^9,27,28 These studies have provided the technical foundation for the development of mouse models that have led to a wave of second-generation genetically engineered mice with more complex genetic combinations and more highly controlled gene regulation.²⁹ Genetically engineered mouse models are now being used in a preclinical setting to identify and validate novel diagnostic and therapeutic approaches for cardiovascular diseases.^{3,5–7,30,31}

Conditional and Inducible Gene Targeting

Conventional gene targeting, in which gene function is ablated in all cells of the body, is estimated (T.D.) to have produced most of the first 1000 gene KO mice, but it has limitations.⁹ In about 30% of the cases, KO mice exhibit embryonic or early postnatal lethality, which underscores the absolute gene requirement for embryogenesis and therefore obscures the analysis of the potential gene function at postnatal stages of life. Development of approaches for conditional (spatial control) and inducible (temporal control) mutagenesis have been applied to overcome these limitations.^32,33 The first successful in vivo description of cre transgenic mice was reported in 1992,³⁴ followed by another study demonstrating that Cre recombination also worked in mouse ES cells.³⁵ These initial reports revolutionized the conditional KO mouse field. In mice harboring conditional alleles, the target gene is modified in the germline so that it has the potential for future mutation but maintains wildtype gene function. This is done by flanking a critical gene region with loxP sites, which themselves are not in critical regions. The loxP sequences are recognition sites for Cre recombination. Through genetic combination with a tissue-specific cre transgene, those critical sequences between the loxP sites are recombined out, resulting in loss of gene function in that tissue. In addition to the Cre-loxP recombination system, FLP-FRT and Dre-rox systems are also available.^36,37

Temporal control over ablation of gene function can be achieved by using tamoxifen (Tam)-inducible and tetracy-cline (Tet), or its analog doxycycline (Dox)-inducible recom-binases.^{38 – 40} The combination of both conditional and inducible gene targeting approaches has led to tissue-specific genetic alterations under temporal control. Temporal control over the conditional gene targeting event has enabled investigation of gene function during embryonic development and over time in adult mice as a model of aging. In addition, adenoviral cre expression vectors can be delivered by intravenous or stereotaxic injection into mice, which can be effective tools for localized temporal regulation of cre expression in the target organ.^41,42

There are also problems with the Cre-loxP system that must be kept in mind. Several studies have indicated significant Cre toxicity in mammalian cells and transgenic mice.^{43– 45} This is consistent with the presence of cryptic loxP sites (1.2 primary cryptic loxP sites per Mb DNA) in the mouse genome, which could cause illegitimate DNA recombination. The Cre toxicity can be minimized by using floxed cre expression cassette in Cre mice, which terminates its own expression via self-deletion of the floxed cre.⁴⁶ Another strategy to reduce Cre toxicity is to use regulated cre expression by inducible Cre mice (see below). Another technical concern regarding the use of Cre mice is that these mice either can exhibit phenotypic abnormalities and altered expression of genes at and around the site of random transgene insertion caused by the inherent problems associated with the transgenic mouse technology (ie, cre transgenic mice) or Cre mice may have subtle undesirable phenotypes because of their heterozygous KO status (cre knockin mice). Some of these concerns can also be minimized by using appropriate controls such as mice carrying the cre transgene but not its loxP-flanked target gene and Cre mice with a floxed heterozygous allele. There is no data to suggest any pitfalls in the use of FLP and Dre mice. Overall, numerous successful experiments using Cre mice indicates that despite many pitfalls the conditional gene targeting remains by and large the most useful technique in manipulating mouse genes in a tissue-specific fashion.

Conditional gene targeting typically requires crossing mice carrying a gene with loxP sites flanking an exon (hereafter termed “floxed” mice) with mice carrying a cre transgene (hereafter termed Cre mice). These floxed alleles are introduced in mice using traditional gene targeting approaches, and the exon deletion is activated when the cre transgene is expressed. The Cre (P1 phage-derived) and Flip (FLP, yeast-derived) recombinases recognize 34 bp loxP and FRT (flip-pase recognition target) sites, respectively (Figure 2). FRT and loxP sequences have an asymmetry such that recombination between similarly oriented recognition sites in the same DNA strand results in a deletion of the intervening sequence. The recombination leaves a single loxP or FRT site. These recognition sites are not present in the mouse genome.

Because a positive selectable marker gene (eg, neomycin) that is often introduced to modify the locus could disrupt or alter the gene function or result in a hypomorphic locus, strategies have been developed for removal of the marker gene after the targeted allele has been identified or established in the germline. The commonly accepted strategy is to flank the selectable marker cassette with FRT sites and remove the intervening region with FLP recombinase.^29,36 Another established procedure is to keep the marker gene beside one of the loxP sites and place a third loxP site on the other side of the marker after the critical exon that will be conditionally deleted. The marker gene can be efficiently removed by a partial Cre excision process, which leaves behind the floxed critical exon in the targeted locus.^29,47 These strategies are shown in Figure 3 which provides an example of a European Conditional Mouse Mutagenesis Program/Knockout Mouse Project (see below) gene targeting strategy that contains features for expression marking, removal of selectable marker gene, and conditional/inducible gene inactivation capabilities.⁴⁸

Resources for Targeted ES Cells and Mice

Before embarking on a costly and time-consuming gene-targeting project it is best to determine whether the targeted gene is already present either in an ES cell line or a mutant mouse strain. The International Knockout Mouse Consortium (http://www.knockoutmouse.org/) is generating ES cells and mice with KO and conditional/inducible-capable alleles for all protein coding genes.²⁹ The International Knockout Mouse Consortium reports targeted ES cell lines for over 15,000 genes (including 9000 conditional targeted alleles⁴⁹), and mice for about 10% of these.⁵⁰ The consortium consists of the Knockout Mouse Project, European Conditional Mouse Mutagenesis program, North American Conditional Mouse Mutagenesis Project, and Texas A&M Institute for Genomic Medicine. Targeted ES cells can be purchased by academic researchers at nonprofit organizations at affordable prices. The requirements for generating these targeted alleles were that they not only KO the gene, but that they have the potential for conditional/inducible inactivation, provide a marker gene for expression analysis, and have the capacity to remove promoter interference by the selection gene. Mice generated by the Knockout Mouse Project are reposited by the National Center for Research Resources (http://www.ncrr.nih.gov/comparativemedicine/resourcedirectory/rodents.asp). In addition, the Jackson Laboratories has a large mutant mouse resource (http://mousemutant.jax.org/), and the National Cancer Institute’s Frederick Cancer Research facility has a Mouse Models of Human Cancer repository (http://mouse.ncifcrf.gov/). These repositories are also a primary source for finding Cre and Flp mouse lines for Cre/LoxP and Flp/Frt gene targeting.

Tam-Inducible Conditional Gene Targeting

Cre activity can be tightly controlled in space and time. Inducible and global or cell-type-specific conditional gene deletion has been most frequently carried out by fusion proteins of Cre with the mutant ligand-binding domain of human estrogen receptor (CreER).^51,52 Both CreER and its more refined version, the CreER^T2, do not bind to endogenous estradiol, but they bind to synthetic estrogen receptor ligand 4-hydroxytamoxifen or Tam with high affinity (Figure 4). Generally, in the absence of hormone, estrogen receptors are largely located in the cytoplasm. Estrogen binding to the receptor translocates the receptor from the cytoplasm to the nucleus where it binds to DNA and regulates transcription of target genes.⁵³ However, the CreER recombinases are inactive and remain in the cytoplasm but can be activated and translocated to the nucleus by Tam, thereby allowing for external temporal control of Cre activity.^51,54 A CreER^T2 mouse can be a transgenic or a knockin of the CreER^T2 coding region into an endogenous gene.

ROSA26-CreER^T2 mice were made by targeting CreER^T2 into the ROSA26 locus.⁵⁵ The ROSA26 locus located on mouse chromosome 6 is a constitutively expressed gene in all cell types. ROSA26 mutant mice are normal and viable. The high rate of homologous recombination at the ROSA26 locus has made it a preferred site for knocking in a single-copy transgene to achieve a stable and ubiquitous or conditional transgenic overexpression during development or in the adult.^56,57 Thus, ROSA26-CreER^T2 mice have ubiquitous Tam-induced Cre recombination in adult mice.⁵⁸ In addition, cell-type–specific CreER^T2 mice such as cardiac-specific αMHC-CreER^T2 (also called αMHC-MerCreMer) mice have been effectively used in cell-type–specific Tam-inducible gene targeting.³³ In addition, Tam-inducible Cre recombination has been effectively used to modify gene function in mouse embryos in utero.^55,58,59 Tamoxifen can be administered via food (custom-made chow containing 0.4–1 g/kg Tam, Harlan), water intake (0.5–1 mg/mL), 5 consecutive days of either intraperitoneal injection (100 mL of 10 μg/mL Tam) or oral gavage (200 mg/kg).

A disadvantage of the Tam approach to induce constitutive or cell-type-specific gene manipulation in utero is the risk of embryo abortion and death (can occur after 4 days of treatment), so it is useful only during those 4 days.^55,58,59 Because RNAs and proteins have different half-lives and can function at different concentration ranges, it is possible that residual RNA or proteins present before the induced inactivation could compensate for gene inactivation. Assessment of both RNA and protein can test for this potential problem. Intestinal problems from Tam administration have also been reported in homozygous ROSA26-CreER^T2 mice.⁵⁸ There are also some concerns that Tam can cause behavioral alterations.⁶⁰ The effects of Tam on CreER^T2 mice alone, CreER^T2 mice with the floxed allele, and mice harboring the conditional allele alone should be compared to the experimental mice (CreER^T2 mice with the floxed-out conditional allele). Despite these potential problems, numerous researchers have successfully used this approach to investigate cardiac-specific gene function in the physiology and pathophysiology of the heart.^7,11

Tetracycline-Inducible Conditional Gene Targeting

Tet-mediated recombination is another widely used method for inducible gene targeting.⁶¹ This approach is reversible, it can control gene expression in either cell- or lineage-specific manners, and it can be temporally regulated. This system is derived from E. coli and has been adapted for use in eukaryotes. The 2 major components of this system are the Tet repressor (TetR) and the tet operator (tetO), which were originally derived from the Tn10 Tet resistance operon.⁶² Both Tet and its analog doxycycline (Dox) interact with TetR and are well tolerated and widely used in mammalian systems. Both Tet-OFF and Tet-ON approaches have been used to regulate gene expression in mice.⁶³ In the tTA (Tet-controlled transactivator) or Tet-OFF system Tet or Dox prevents binding of tTA to the Tet-responsive promoter (contains minimal cytomegalovirus [CMV] promoter and 7 tandem tetO [tetO₇] sequences), and thus abolishes gene transcription. In contrast, the reverse Tet controlled transactivator (rtTA) or Tet-ON system requires Tet or Dox for binding to and inducing a Tet-responsive promoter (Figure 5).

The first demonstration of the Tet system in transgenic mice used the tTA approach in which E. coli’s Tn10 transposon fused to the VP16 transactivation domain from herpes simplex virus was used as a TetR for the Tet effector molecule.⁶⁴ Gene induction became detectable in a few hours and reached its maximum level within a day.⁶³ Several studies have indicated that the rtTA system is better suited for temporal control of rapid induction of gene expression.⁶³

Recently, many investigators have combined the Tet-inducible system with the Tam-inducible Cre-loxP approach (Figure 5).⁶³ This allows more flexibility producing more clinically relevant mouse models of human diseases. Several laboratories have integrated the Tet-inducible and Cre-loxP approaches and developed mice based on rtTA-dependent Dox-mediated gene induction following a Cre-mediated deletion to obtain a temporal and spatial or cell-, tissue-, and organ-specific gene targeting.^65,66 Thus, gene targeting can be manipulated in specific cell types and lineages with a flexibility that is difficult to achieve with other methods. To make this system more sophisticated, a Tet-inducible CreER^T2 has been targeted to the ROSA26 locus.⁶⁷ By introducing Tet induction, this Tam-inducible Cre driver line permits highly selective dual Tet and Tam regulation of loxP recombination which could open new avenues for spatiotemporally controlled gene targeting in mice.

There are pitfalls inherent in Tet-inducible technology that is worth mentioning. The Tet system can be leaky.⁶⁸ This can occur either through weak binding of rtTA to tetO even in the absence of Tet or Dox effectors, or through unwanted basal activity of the Tet-responsive promoter even in the absence of rtTA. This leakiness can result in unexpected phenotypes when Tet is absent.⁶⁹ Although this leakiness can be tolerated in many experimental systems in which a phenotypic outcome are desired, it can be limiting when gene function is being investigated. Use of several Cre transgenic lines could obviate some of these difficulties. Refinements have also been made to rtTA (ie, rtTA2S-M2) to increase the sensitivity of the system and to reduce background expression.⁷⁰ Additional control over background expression can also be achieved by using an inducible transcriptional repressor that can bind to the Tet-responsive promoter and thereby block background expression. For example, the TetR protein can be used to minimize leakiness in the system.⁷¹ In theory it is possible that the application of a tTA repressor with rtTA in the absence of Tet could prevent leaky expression through its binding to the TRE. This means that in the presence of Tet, Tet would bind to the tTA repressor and dissociate from the TRE, while at the same time the Tet-rtTA complex would bind to the TRE and induce transcription. Although initial efforts to test these possibilities have produced mixed results,⁷² refinements could be promising.

Another potential problem with the Tet system is that of tissue toxicity caused by overexpression of TetR proteins.⁶⁸ This can also be avoided by producing multiple transgenic lines expressing varying levels of Tet repressor. Taken together, it is very important that proper control animals are used in Tet-inducible gene targeting experiments to avoid misleading results due to these pitfalls. Overall, the ability to regulate gene activity in spatiotemporal and reversible fashions has made the Tet-inducible approach a favorite technology of numerous mouse geneticists.

Targeting Single Nucleotide Polymorphisms Into the Mouse Genome

There are numerous examples of single nucleotide polymorphisms (SNPs) in both exons and introns, and also important functionally relevant regulatory noncoding sequences and enhancer sequences within intronic regions.⁷³ Several epidemiological studies have suggested that SNPs may be involved in the susceptibility and pathogenesis of cardiac disease in humans,⁷⁴ and it is known that functionally relevant SNPs can be in noncoding regulatory and intronic regions. For example, about one-third of mammalian coding transcripts are silenced by microRNAs (miRNAs), powerful regulators of gene expression that are involved in cardiac development and pathogenesis of cardiac disease.^75,76 Although conventional or conditional gene targeting has been used in introducing SNPs, which were created by site-directed mutagenesis in the targeting vector, in most instances a residual loxP site is left behind, sometimes accompanied by a selectable marker gene such as neomycin.⁷⁷ The concern is that insertion of a selectable marker and/or loxP or FRT site(s) in noncoding intronic regions could cause lethality, or a hypomorphic or other unwanted phenotype.^{78 – 80} This makes it difficult to separate the effect of neomycin or loxP from the effect of the SNP. At the vector design stage incorporating bioinformatic approaches such as TraFac,⁸¹ which provides information on the conserved cis sequences within a gene, or ORF and miR analysis to detect other embedded genes, can usually identify regions in the gene that should be avoided when placing loxPs, FRTs, and selectable cassettes. However, these approaches are not completely accurate, and there are undoubtedly other functionally relevant sequence motifs not yet identified.

Consequently, it may be worthwhile not to leave behind residual sequences that in and of themselves could complicate phenotype analysis. We have previously reported on the use of a Tag and Exchange gene targeting approach to introduce point mutations in the basic fibroblast growth factor gene to ablate the function of protein isoforms resulting from alternative, in frame translational start sites.^{82– 85} Alternative technologies have been used by others to introduce subtle mutations without the use of recombinase systems that necessarily leave residual sequences in intronic regions. The “Hit and Run” system was used to introduce a small mutation in the Hox2.6 gene without leaving residual sequences,⁸⁶ and an “In and Out” approach used a sequence insertion and removal procedure to put a 4-bp insertion into the Hprt gene.⁸⁷ Improvement of these approaches would be useful for future functional analysis on SNPs.

Nuclease-Mediated Gene Targeting

Zinc finger nucleases (ZFNs) are custom-engineered DNA-binding modules fused to the catalytic domain of Fok1 endonuclease that have been adopted for targeted editing of the mouse and rat genome (Figure 6).^88,89 ZFNs are eukaryotic transcription factors that bind 3 bp of DNA per finger; binding site specificity depends on key residues within the ≈ 30aa ZFNs. The FokI cleavage domain can render a double-strand break at any user-specific sequence in the genome but must dimerize to be functional. Therefore, it is possible to cut within a unique genomic sequence of 18 to 24 nucleotides by designing a dimerizing pair of ZFNs, each containing a set of 3 to 4 strategically chosen ZFNs joined to a FokI catalytic domain (up to 6 fingers per monomer have been used). A donor DNA can be exogenously supplied to affect repair of the break by a homology-driven homologous recombination process, resulting in targeted gene replacement with a desired sequence. Alternatively, a double strand break stimulates the cellular process of nonhomologous end joining and results in the mis-repair of the DNA sequence. This error-prone process results in a KO genotype/phenotype.⁹⁰

ZFNs can be introduced directly into fertilized single cell embryos in plasmid or mRNA form. The embryos are implanted into the foster mother and the newborn founder animals are identified by standard screening methods. KO mice and rats can be generated in about 4 months at remarkably high efficiencies (monoallelic 10%–15%, biallelic 1%) compared to the conventional ES cell based gene targeting approach in mice that usually takes about 9 months to a year. The ZFN break-induced events are somewhere between 1,000- and 1,000,000-fold more efficient than the conventional ES cell targeting and there is no need to use selection to find the targeted products. A major difficulty in reproducibly applying the ZFN technology is the fact that it is difficult to design the zinc finger set with the highest probability of successfully cutting the user-specified location in the genome. There is also a potential risk of off-target cleavage by ZFNs. This is also due to a lack of known fingers for some nucleotide triplets and context effects on the specificities of individual fingers in an array.

Another class of engineered nucleases, transcription activator-like effector (nucleases (TALENs) appears to overcome some of the challenges facing the ZFN technology.⁹¹ Similar to ZFNs, TALENs are made by fusing the catalytic domain of FokI to transcription activator-like effector repeat modules that are custom combined to target a unique stretch of genomic nucleotide sequence. Transcription activator-like effectors harbor ≈ 34 amino acid repeat domains that each binds 1 target base in the host genome with base specificity determined by 2 variable amino acids (Figure 6).⁹² TALENs have been successfully used in creating heritable changes in both zebrafish and rats.^93,94 However, there are no gene-targeted mouse models generated using TALENs technology. Advantages of TALENs are that they are more designable than ZFNs because each module of TALENs binds 1 bp of DNA instead of 3 in ZFNs, and the TALEN specificity seems to be less influenced by neighboring DNA sequences. The TALENs also seem to be better than ZFNs at being restricted to their designed target. However, a limitation with TALENs is that the constructs for the TALENs are bigger than for ZFNs, largely because each module of 34 amino acids recognizes only 1 bp in the TALENs; whereas, each ZFN module of 30 amino acids recognizes 3. Although it is possible in principle to conditionally modify an allele for a tissue-specific disruption or overexpression using nuclease-mediated gene targeting techniques, a proof of principle experiment is still missing. Overall, nuclease-mediated gene targeting technology is at the cutting edge and needs to be perfected before it becomes ready to be applied for making conditional and inducible gene KOs and transgenic mice.

Mouse Models for miRNA Expression

miRNAs are 21 to 23 nucleotide long single-stranded RNAs. miRNAs are involved in numerous diverse cellular and developmental processes, and in the pathogenesis of several diseases, including cardiovascular disease.⁹⁵ Sequence-specific base pairing of miRNA with target mRNAs reduces protein stability and translation. At least 30% of protein-coding genes appear to be conserved targets for miRNAs.⁹⁵ The influence and importance of miRNAs as regulators of gene expression has led to their intensive study in the cardiovascular system.⁹⁶ Only a limited number of individual miRNA KO mice have been described in the literature. Overall, several mouse models of miRNA expression have been successfully made by individual investigators using conditional mouse gene targeting approaches.^97,98

The application of inducible technology in the field is growing rapidly.⁹⁹ The analysis of miRNA KO alleles is complicated by the presence of several miRNA genes with related predicted specificities at multiple loci.¹⁰⁰ Consequently, analysis of miRNA function will require the creation of tissue- or cell type-specific conditional and inducible as well as compound mutant mice of miRNA family members. The miRNA KO resource is a research toolbox of mutant alleles that can be used as reagents for studying miRNA function in a range of complementary approaches.¹⁰¹ The targeted alleles have been designed to be adaptable research tools that can be efficiently altered by recombinase-mediated cassette exchange to create reporter, conditional, and other allelic variants. Other examples of possible applications might include the variation of miRNA copy number and mutating individual miRNAs within clusters. This miRNA KO resource can be searched electronically and is available from ES cell repositories (http://www.knockoutmouse.org/martsearch).⁵⁰ Finally, it is yet to be seen whether emerging technologies such as nuclease-mediated gene targeting can successfully manipulate miRNA in conditional or inducible fashion.

Manipulating Alternatively Splicing of Genes Through Gene Targeting

Many mammalian genes undergo alternative splicing.¹⁰² Alternative splicing results in the differential inclusion or exclusion of a specific exon in different cell types, retention of introns, exclusion of a portion of an exon, and mutually exclusive inclusion of exons.¹⁰³ Importantly, any of these differential patterns have the potential to alter the open reading frame of the resultant mRNA or modify the cis-regulatory elements that control mRNA stability or translation. Thus, the proper regulation of alternative splicing is important for maintaining cellular protein repertoire, and changes in splicing can significantly alter cellular function. Gene targeting in mice has been used to create in vivo models to study the regulation and consequences of alternative splicing.^{104 –106} The evidence accumulated so far argues for an important role of alternative splicing in mammalian development and tissue function, and highlights the importance of various protein isoforms in vivo.

Various approaches have been used to generate mice with modified alternatively spliced forms (Figure 7). Briefly, conditional mouse genetics approaches can be used to manipulate alternative splicing of a gene in a tissue-specific fashion. Furthermore, a stop codon can be placed to eliminate expression of all alternatively spliced isoforms carrying an upstream exon, but allowing expression of downstream exons containing alternatively spliced variants. Another similar approach is to destroy the splice sites of the target exon. Furthermore, 1 exon can be replaced by another exon or an unrelated coding region to obtain the expression of an isoform in a tissue, where it is usually not expressed. This technique requires careful consideration of the targeting construct, as exonic regulatory sequences of the deleted exon might be important for tissue-specific splicing of the gene. Random retroviral insertion of marker genes and subsequent sequencing of the integration site has also generated many commercially available targeted ES cells. In this method, stop codons, loxPs, and other gene modification can be made along with the marker gene into the gene. A conventional gene KO is used to eliminate the whole or parts of the gene to disrupt specific alternatively spliced isoforms. In the case of a partial removal of a gene, the expression of the resultant truncated form of alternatively spliced variant mRNA and protein should be taken into consideration. Lastly, the utility of ZFNs and TALENs in conditional or inducible gene targeting to manipulate alternative splicing in mice remains to be tested.

Cardiac Cell Type-Specific Inducible Gene Targeting in Mice

Several important genes that are involved in human cardiovascular disease are expressed in multiple cardiac cell types (Figure 1). In addition, the foundation of many adult congenital cardiovascular diseases can be traced to abnormal cardiovascular development.^3,6,107 Thus, careful mutational analysis of genes in multiple cardiac cell lineages during development and into the adult heart is required to delineate the complex function of many genes in the heart. Tam- and Tet-inducible strategies have been applied to achieve this end^6,7,108 and have been successful in demonstrating gene function in cardiac structure and function, cardiac physiology and pathophysiology, cardiomechanics, cardiac calcium handling, and cardiac stress response.¹¹ In clinical terms, the development of cardiac-inducible/reversible gene targeting approaches provide translational potential.⁹ Because of the success and feasibility of cardiac-specific inducible gene targeting, the number of inducible cardiac cell-type-specific Cre mice is increasing (see the tetmouse database: www.creline.org). Here, we briefly describe the importance and practical aspects of some of the major cardiac lineage-specific CreER^T2 and rtTA driver lines (Table).

Perspectives

Site-specific recombinases have proven to be remarkable tools for many aspects of mouse experimentation. However, these tools have limitations. Knowledge of their strengths and weaknesses is vital for optimizing their application. There are several good Cre driver lines to induce gene manipulations in myocardium. There is a need for more Cre driver lines for inducible gene targeting in epicardium. There is no satisfactory Cre driver line for inducible conditional gene targeting in cardiac valve endocardium and valve mesenchyme for both mouse embryos and adult mice. Serious efforts are needed to identify, develop, and validate these resources for the cardiac valve community.

Because conditional KO or transgenic mouse technology limits gene mutation or overexpression to a specific cell type/tissue/organ or to a specific life stage, generation of improved mouse models of adult cardiac diseases heavily depends on the progress and innovation in inducible conditional gene targeting technology. The difference between conventional conditional and inducible conditional gene targeting approaches has recently been exemplified by Braf gene inactivation in forebrain neurons where both common and distinct phenotypes were found.¹⁵¹ This also applies to cardiovascular disease studies. For example, although considerable progress has been made in the past 20 years in delineating mechanisms underlying cardiovascular disease by using conventional gene targeting strategies, future progress will come largely from inducible conditional gene targeting approaches. These can better model the human disease conditions and identify pathways involved in unintended side-effects of potential therapies.