|Home | About | Journals | Submit | Contact Us | Français|
The purpose of this review is to give a comprehensive overview of transgenic mouse lines suitable for studying gene function and cellular lineage relationships in lung development, homeostasis, injury and repair. Many of the mouse strains reviewed in this article have been widely shared within the lung research community and new strains are continuously being developed. There are many useful transgenic lines that work to target subsets of lung cells, but it remains a challenge for investigators to select the correct transgenic modules for their experiment. This review covers both the tetracycline and tamoxifen inducible systems and will primarily focus on conditional lines that target the epithelial cells. We point out the limitations of each strain so investigators can choose the system that will work best for their scientific question. Current mesenchymal and endothelial lines are limited by the fact that they are not lung specific. These lines will be summarized in a brief overview. In addition, useful transgenic reporter mice for studying lineage relationships, promoter activity and signaling pathways will complete our lung specific conditional transgenic mouse-shopping list.
In recent years tremendous progress has been made in understanding the cellular processes underlying lung development, homeostasis and repair. This has been facilitated by the use of lung cell type-specific transgenic mouse lines. Initially lung specific promoters [1-3] were used to constitutively drive transgenes in a cell type specific manner. This mostly resulted in embryonic lethal phenotypes and consequently limited adult studies. The development of conditional transgenic mice allowed the temporal and spatial control of gene expression, overcoming many lethal phenotypes, and allowing the analysis of lung-specific gene knock-outs, identification of progenitor cells, lineage tracing, and studies of progenitor proliferation and differentiation capacity. The most widely used systems are the doxycycline system (tTA and rtTA) and the Cre-LoxP system, but others exist and their use is becoming more widespread. As transgenic mouse strains have been used their strengths, limitations, and the strategies required for optimal experimental design, have become apparent. In this review we will discuss the mouse strains that have been shown to be the most useful for manipulating gene expression in the lung and also highlight areas where new mice would be extremely beneficial for the community.
Cre-Lox technology was introduced in the 1980s [4, 5] and patented by DuPont Pharmaceuticals. It was successfully applied to mice in 1998 . The technology is based on the ability of the P1 bacteriophage recombinase (Cre) to direct site-specific DNA recombination between pairs of LoxP sites. Such recombination in a “Cre-lox” mouse can permanently either inactivate, or activate, a gene of interest.
Typical Cre-Lox experiments require two transgenic animals: a Cre strain and a LoxP strain. A Cre mouse contains a Cre recombinase transgene under the control of a tissue-specific promoter (Fig. 1A), whereas a LoxP mouse contains two LoxP sites that flank a genomic segment of interest, the “floxed” locus. Depending on the location and orientation of the LoxP sites in a Cre-Lox mouse, Cre recombinase can initiate deletions, inversions, and translocations of the floxed locus . The floxed loci can be designed to allow either permanent inactivation, or activation, of the gene of interest (Fig. 1B). Mutated LoxP sites, which allow recombination between various independent LoxP sites, have been successfully used to rapidly target genes and generate multicolor reporter mice, ‘brainbow’ and ‘confetti’, which are useful reporters for clonal analysis of progenitor cells [8-10].
The site of Cre activity (cell type specificity) is dependent on the availability of tissue-specific or cell-specific promoters. Moreover, tissue-specific Cre expression can be combined with temporal-specific activity. Time-specific Cre activation can be achieved by combination with the doxycycline system (Fig. 1D), or by the use of Cre fusion proteins. Cell type specific Cre strains have been widely used in the lung for lineage tracing and permanent gene activation or deletion (for example [11, 12]).
A very useful technique is to track the descendents of stem/progenitor cells (lineage-tracing) by crossing a Cre mouse with a reporter mouse strain that permanently expresses a reporter gene following Cre activity. The Z/AP and Z/EG reporter mice were initially used for lineage-tracing experiments [13, 14]. However, studies have shown that these reporters are not expressed in all cell types and can be silenced in adult tissue. The Rosa26 reporter (Rosa26R) variants are now the most widely-used reporter strains available. The ROSAβgeo26 (GtROSA26) line was initially derived from pools of ES cells infected with the retroviral gene trap vector Gen-ROSAβgeo . After cloning the ROSA26 locus , it was used to produce a variety of Cre reporter lines starting with lacZ and expanding it to a vast repertoire of cytosolic, membrane bound and nuclear florescent lineage-tags . The ROSA26 locus is particularly useful for generating these strains as it is expressed robustly in most cell types and is gene-targeted at high efficiency, thus numerous other cassettes have been targeted to the ROSA26 locus.
To allow temporal control of Cre activity, fusion proteins have been constructed between Cre and the ligand-binding domain of steroid hormone receptors (Fig. 1A). The most commonly used variant is a fusion between Cre and a mutated ligand-binding domain of the estrogen receptor (CreERT2) [18-20]. ERT2 binds weakly to endogenous estrogens and strongly to 4-hydroxy tamoxifen (4OH-T), the active metabolite of the synthetic steroid tamoxifen (tmx). Administration of tmx or 4OH-T by itself can be toxic, resulting in various embryonic phenotypes if administered up to about E11.5, or abortion if administered at later stages . Moreover, tmx dosing can cause a transient increase in blood pressure in adult mice . For these reasons it is important to titrate the tmx dose to the minimum required for each experiment. Most investigators dose their animals with tmx, which is converted to 4OH-T in the liver. 4OH-T can also be administered directly, but kinetics of Cre activation and drug metabolism will be different.
The CreERT2 fusion protein is cytoplasmic. Upon binding to tmx, CreERT2 translocates to the nucleus where it accesses the LoxP sites. Earlier CreER and CreERT2 versions can be somewhat leaky when expressed from a strong promoter. However, such strains can still be highly informative if the correct controls are performed . Recombination rates are very sensitive to the levels of CreERT2 expression  and to the length of time the protein spends in the nucleus. This is dependent on tmx dose and frequency of dosing. Elegant studies using a cartilage-specific CreERT fusion protein demonstrated that subsequent to a single intraperitoneal injection of tmx into a pregnant female, reporter activity could be detected within 8 hours and that recombination was complete within 24 hours . In addition, it has been reported that administration of tmx to pregnant females via oral gavage, rather than intraperitoneal injection, results in more efficient labeling and less embryonic toxicity . The CreER2T system has been widely used in the lung, although the extent of recombination, and most effective tmx dosing strategy, must be determined empirically for each CreERT2 mouse strain [27, 28]. In addition, the recombination rate between each pair of LoxP sites also varies and must be determined experimentally . Another variant of the CreER system has the ERT fused to both the amino and carboxy terminals of Cre (known as mER-Cre-mER) . The mER-Cre-mER has not been used widely in pulmonary research.
Multiple transgenic strains with widespread expression of tamoxifen-inducible Cre have been generated and are potentially useful in the adult lung for deletion of genes with cell type specific expression. Representatives of these lines are beta-actin-Cre  and actin-CreERT , which have been used successfully in the developing lung mesenchyme , both direct Cre expression in multiple lung cell types. Similarly, a CMV-CreERT mouse line directs expression of CreER in most cell types , and has been successfully used to study the role of Sox2 in the adult tracheal epithelium . In addition, four independent RosaCreERT2 strains exist that allow ubiquitous expression of CreERT2 from the ROSA26 locus [29, 35-38].
The doxycycline and Cre/CreER systems are both dependent on the availability of a cell type specific promoter to restrict gene expression to the cells of interest. However the variety of cell type specific promoters is limited. The split-Cre system was developed to overcome these limitations . In this system the Cre protein is split into two halves that are expressed from different promoters. Individually these two parts of the protein are inactive. When both promoters are activated in the same cell, inter-molecular complementation occurs and Cre is functional. This system has been successfully applied to the mouse brain to target populations of rare stem cells, which had previously been defined by flow cytometry only . An inducible split CreERT2 system has also been developed and shown to function in vitro . In vivo application of the split Cre in the lung will be beneficial to advance the field of lineage relationships and progenitor cells.
Off target effects of the Cre recombinase have been reported in several systems , including the lung [43, 44]. These off target effects are probably due to endogenous cryptic LoxP sites within the mammalian genome that cause cytotoxic chromosomal rearrangements when activated [45, 46]. Not every Cre strain shows off-target effects and this variation is likely to result from differences in levels of Cre protein expression. Cre toxicity has led to off-target phenotypes and made interpretation of some experiments difficult. In particular, some Cre lines demonstrate Cre activity in the germline, which results in recombination of the floxed allele independent of the regulatory element driving Cre ( and B. R. Stripp personal communication). Such sex-specific effects can usually be avoided by transmitting the Cre via the female or male germline respectively. The use of dox or tmx-dependent Cre strains limits the amount of time Cre spends in the nucleus and decreases Cre toxicity, which also highlights the importance of using the minimum dose of dox or tmx for each experiment. In addition, it is crucial to perform the correct controls: 1) dox and tmx treatment of single transgenic mice, 2) dox and tmx treatment of mice with all transgenes in a non floxed background 3) untreated mice containing all transgenes in the floxed background. Cre mouse strains, often in combination with doxycycline (dox) or tmx, are probably the most widely used animals for manipulating gene expression in the lung.
The FLP-FRT system is similar to the Cre-Lox system and is becoming more frequently used in mouse-based research. It involves using Flippase (FLP) recombinase, derived from the yeast Saccharomyces cerevisiae . FLP recognizes a pair of FLP recombinase target (FRT) sequences that flank a genomic region of interest. RosaFLPe is a mouse line with a ubiquitous FLP expression . A useful reporter mouse for this recombination is the Flp indicator mouse expressing alkaline phosphatase from the ROSA26 locus . However, despite many attempts it has been difficult to generate lung specific Flippase mice (A.K. Perl, J. Whitsett unpublished data). Nevertheless, a Flp-inducible allele of K-ras was recently activated in the lung using a lentivirus-encoded FLP protein .
The tetracycline (tet) inducible system was developed independently to the Cre-LoxP system and has different advantages and limitations. The tet system in vivo consists of two transgenic mouse lines, an activator line and an operator line. The activator line expresses either the tetracycline activator (TA, tet-off) or the reverse tetracycline responsive transactivator (rtTA, tet-on) in a tissue specific manner. The operator line carries a transgene of interest under control of the (tetO)7CMV operator (tetO). In double transgenic mice, doxycycline either causes the tTA to bind the tetO sequence, suppressing transcription (Tet-off), or causes the rtTA to bind to the tetO sequence, activating transcription of the gene of interest (Tet-on) (Fig. 1C) [52, 53]. The major advantage of this system is to reversibly turn genes on and off and study the effects of genes at specific times. In the lung this system was successfully used in 2000 to conditionally activate FGF7 during lung development . The technical aspects of using doxycycline to activate genes were described using a luciferase reporter mouse [55, 56] and the limitations of the system were reported in 2006 .
The use of the rtTA system was then further expanded by the combination of the rtTA system with activation of a tetO-Cre transgene [58, 59]. The cell specific, dox dependent Cre expression enables permanent endogenous gene inactivation or transgene activation at any given time during development, or in adult animals (Fig. 1D) . This dox-dependent Cre activation has been shown to be very useful for in vivo cell lineage labeling and conditional deletion of genes that otherwise would result in lethal phenotypes. It has also recently been used to conditionally deplete specific cell populations and study epithelial regeneration .
Combining tissue specific Cre lines with Rosa26rtTA transgenic mice results in Cre-mediated rtTA expression (Fig. 1D). These mice have Cre-inducible expression of rtTA and can be used to achieve spatially and temporally controlled transgene expression in a wide variety of settings simply by crossing to any existing mice carrying cell type-specific Cre recombinase and tet-O-regulatable responder genes [61, 62].
Off-target toxicity of rtTA and doxycycline has been reviewed previously [43, 57, 63]. Briefly, these toxicities were unrelated to the effects of the transgene and varied both with specific mouse strain and genetic background. Off-target effects can influence both lung morphogenesis and perinatal survival and most often result in airspace enlargement. In addition, doxycycline is a matrix metalloproteinase inhibitor and has been shown to promote pulmonary hypertension after hypoxia, and to attenuate mucin production [64-66]. More recently it has been shown that tracheal Clara cells are sensitive to doxycycline treatment . Doxycycline is stored in tissues and will be released over time. The timing and duration of treatment needed to target subsets of lung epithelial cells has been described and we highly recommend limiting time and dose of doxycycline exposure [58, 59]. With the new generation rtTA constructs, that are more sensitive to low doses of doxycycline (for example, rtTA2S-M2), off-target effects of doxycycline will become less significant, but dosing regimens will be more important [67, 68]. In all dox experiments, it is vital to perform the correct controls. To control for phenotypes not related to the activation/inactivation of the gene of interest, single transgenic mice, and mice containing all transgenes, should be tested in the absence of doxycycline. Littermate controls should be used where possible to minimize strain or age related variability. Experiments should also be controlled for weight and gender differences.
There are multiple methods for producing Cre or tTA/rtTA mouse strains. These include generating transgenic animals by pronuclear injection, BAC transgenics (Bacterial Artificial Chromosome) and gene targeting by homologous recombination in ES cells (“knock-ins”). The method used affects the properties of the resulting mouse strain and it is important to be aware of the benefits and limitations of each method.
Many of the mice we mention in this review are available form the Jackson Laboratories (http://jaxmice.jax.org). Most importantly the Jackson Laboratories has also developed useful databases. These include, MGI (Mouse Genome Informatics), which provides access to integrated data on mouse genes and genome features, from sequences and genomic maps to gene expression and disease models (http://www.informatics.jax.org/). The IMSR (International Mouse Strain Resource), which is a searchable online database of mouse strains and stocks available worldwide, including inbred, mutant, and genetically engineered mice (http://www.findmice.org/). And the MPD (Mouse Phenome Database), which is a collaborative collection of baseline phenotypic data on inbred mouse strains. The MPD includes data sets, protocols, projects and publications, and SNPs (http://phenome.jax.org/).
The lung buds from the foregut endoderm around embryonic day (E) 9.5 in mice and continues to develop by branching morphogenesis. By contrast, the trachea and esophagus separate from the foregut just ventral to the lung buds between E10.0 and E11.5 and then increase in length and diameter during development. A number of mouse strains suitable for conditional gene activation in the developing lung have been developed using genes expressed in the foregut endoderm. However, with the exception of Sftpc lines, these are not unique to the developing lung. The choice of line depends partly on the timing of activity required: 1) in the undivided foregut (that is, throughout the entire lung, trachea and esophagus), or 2) more lung-specific. As differentiated cell types appear during lung development they can be targeted with cell type specific mouse lines.
Sonic hedgehog (Shh) is expressed in a highly dynamic pattern during embryogenesis. The Shh-Cre line is a knock-in of a GFP-Cre fusion protein  and has been shown to activate reporter recombination in the ventral foregut endoderm by E9.5, prior to lung budding and tracheal/esophageal separation . It has successfully been used to study both tracheal and lung development [11, 75, 76].
Islet1 is expressed in many tissues during embryogenesis including the limb and heart, but is enriched in the lung epithelial precursors at E9.5 . The Islet1Cre line is a knock-in strain  and drives Cre expression throughout the pharyngeal endoderm, including the pulmonary epithelial precursors, and in a subset of mesenchymal cells . Although it is not epithelial specific, it has been successfully used to study Fgf8 function during lung development .
Transgenic mice driving Cre from a 4.8 kb fragment of the Gata5 promoter (Gata5Cre) mediate recombination in the developing epicardium and also throughout the developing lung endoderm [81, 82]. However, onset of recombination and potential activity in the developing trachea has yet to be determined.
The transcription factor Id2 is dynamically expressed in various cell types during embryogenesis and in the adult. The Id2-CreERT2 line is a knock-in of the CreERT2 fusion . In the developing lung this strain displays tamoxifen-dependent Cre-mediated recombination in the distal epithelial tips and, at a lower level, in a sub-set of mesenchymal cells.
Nkx2-5 is one of the earliest cardiac-specific markers in vertebrate embryos. The Nkx2-5 Cre line is a knock-in  and drives recombination in the developing proepicardium and subsequently throughout the myocardium and the first pharyngeal arch . Recombination occurs in both the foregut endoderm and surrounding mesoderm prior to E9.5 before lung budding or tracheal-esophogeal separation. Nkx2.5-Cre has successfully been used to inactivate epithelial expression of Sox2 in the developing respiratory tract .
Nkx2.1 (also known as Ttf1 and Titf1) is expressed in the domain of the ventral foregut, which will give rise to both the lung and trachea. Nkx2.1 is also expressed in thyroid progenitors and regions of the developing brain . An Nkx2.1-Cre BAC transgenic mouse  functions throughout the lung and tracheal epithelium and has been successfully used in a number of studies [88, 89].
Subsequent to budding, the lung epithelium transcribes Sftpc (SpC, or Surfatant associated protein C). Since the human SFTPC promoter is active once endoderm has been committed to becoming lung it is very useful for studying early lung development. The SFTPC lines are discussed below. As lung development proceeds more restricted progenitor cells, such as bronchiolar progenitors, are hypothesized to exist . Tools for manipulating gene expression specifically in such progenitor cells would be highly desirable.
The 3.7 Kb fragment of the human SFTPC promoter is one of the most widely used promoters to generate constitutive and conditional transgenic mouse strains that target the respiratory epithelium . The advantage of the SFTPC promoter is that it is very lung-specific and off-target effects on other organs are extremely rare. In the adult lung SFTPC promoter activity is restricted to alveolar type II cells and subsets of cuboidal bronchiolar cells. The most widely distributed strains are the SFTPC-rtTA [54, 56, 90] and SFTPC-Cre . The SFTPC-rtTA lines are particularly useful since dox application from E6.5 to E10.5 only targets the progenitor pool of the distal lung epithelium, the parathyroid and the thymus. However, targeting of neuroendocrine cells with these lines was not observed . During organogenesis expression levels of morphogenic genes dynamically change spatially and temporally. With the rtTA line it was possible to activate and inactivate signaling pathways that regulate morphogenic genes in specific compartments at defined times. These studies led to a better understanding of temporal windows for FGF signaling and allowed the process of lung organogenesis to be dissected in greater detail [59, 92, 93]. Most prenatal studies were done using SFTPC-rtTA line 1, which expresses rtTA at high levels, but also shows dox independent gene activation especially around birth and postnatally. Dox independent expression can result in embryonic lethal phenotypes, as it was the case for overexpression of FGF7 and VEGF [54, 90]. To overcome these limitations, a second founder line (SFTPC-rtTA line2) was characterized , which expresses lower levels of rtTA and demonstrates less off-dox effects. While line 2 has been demonstrated to work after E14.5 and in adult mice, activation in the developing embryonic endoderm has not been tested. Both lines have been widely distributed in the scientific community.
The SFTPC-Cre transgenic strain contains the human SFTPC promoter fragment, which drives a rabbit β globin intron, followed by Cre recombinase. SFTPC-Cre directs recombination throughout the lung epithelium starting at E10.5 . This strain has also been widely used and has contributed significantly to our understanding of lung development [94, 95]. There are reports of toxic effects on some genetic backgrounds [95, 96]. More recently, it has been demonstrated that the SFTPC-Cre line directs recombination in the male germline, which may have confounded previous studies resulting in the apparent toxicity (B.R. Stripp, personal communication). As long as this transgene is transmitted through the female germline it is very useful to study embryonic lung development.
The Sftpc-CreERT2-rtTA is a knock-in of both CreERT2 and rtTA cassettes just after the stop codon of the endogenous Sftpc gene . This very flexible strain drives both CreERT2 and rtTA expression in mature adult type II cells. It has already proved to be useful for lineage-tracing experiments and will undoubtedly also be widely used for manipulating gene expression.
Other SFTPC-lines are summarized in Table 1. There are alternative promoters (e.g. ABCA3, C/EBPα), which could be potentially used to target alveolar type II cells. However, these promoters will not be lung specific.
Aquaporin 5 (Aqp5), is expressed predominantly in salivary and lacrimal glands, cornea, trachea, and distal lung [98, 99]. In rat and human lungs, Aqp5 is specifically expressed in alveolar type I (AT1) cells and not in alveolar type II (AT2) cells. In mice Aqp5 expression has also been found in AT2 cells . A Cre-IRES-DsRed cassette has been inserted into exon 1 of the endogenous Aqp5 locus generating the Aqp5-Cre-IRES-DsRed, or ACID mouse . Analysis with the ROSA-mT/mG reporter , demonstrated that recombination had occurred in a very high fraction of AT1 cells in the distal lung and not in AT2 cells. However, AT2 recombination in other genetic backgrounds cannot yet be ruled out. This is the first transgenic mouse engineered to express Cre in AT1 cells and it should be very useful for studies of AT1 turnover and function.
Podopladin, or T1 alpha, a gene with unclear function, is expressed in mouse AT1 cells and lymphatics . By contrast, the rat podopladin gene is specific for AT1 cells. A modified rat BAC containing internal ribosome entry site (IRES)-green fluorescent protein (GFP) in the podoplanin 3′UTR has been generated (RTIbac) . RTIbac-transgenic mice expressed rat podoplanin in AT1 cells and in the brain, and expression in AT2 cells, airways, and vascular endothelium was not detected. Modifications of this BAC to express the rtTA or Cre recombinase could make this construct useful for targeting ATI cells.
Secretoglobin1a1 (Scgb1a1, also known as CCSP, CC10 and CCA) is expressed in all bronchiolar Clara cells, and at lower levels in most tracheal Clara-like cells. In the rat Scgb1a1 is also expressed in AT2 cells. A rat Scgb1a1 promoter fragment has been used for making various transgenic lines.
A 2.3Kb fragment of the rat Scgb1a1 promoter is sufficient to direct expression in mouse Clara cells . This promoter was subsequently used to generate two independent Scgb1a1-rtTA mouse lines. The first line  has been widely distributed within the research community and used successfully in multiple studies. Lineage-tracing showed that this strain has efficient activity in many bronchiolar Clara cells and also a sub-set of AT2 cells . The Scgb1a1-rtTA line 2 targets most Clara cells, but retains little AT2 cell activity [43, 60]. Scgb1a1 is also active in the uterus. However, using luciferase reporter mice no luciferase activity was detected in whole uterus homogenates .
The rat Scgb1a1 promoter has also been used to generate a Scgb1a1-rtTA2S-M2 which uses the newer more-sensitive version of rtTA . This strain is reported to have no basal activity and increased doxycycline sensitivity. However, it also has some activity in AT2 cells.
An Scgb1a1-Cre transgenic strain was generated using the rat promoter fragment inserted upstream of the coding sequence for Cre [105, 106]. Lineage tracing shows that this strain directs recombination in the bronchiolar cells, but not in any alveolar cells , demonstrating that the insertion site of the transgene has a strong effect on expression pattern.
An Scgb1a1-CreER™ “knock-in” mouse strain was generated by inserting an IRES-CreER™ cassette into the 3′ UTR of the endogenous mouse Scgb1a1 locus . This line was used for detailed lineage-tracing studies, which revealed that it provides specific, tamoxifen dose-dependent Cre activity in up to 90% of bronchiolar Clara cells, and up to 7% of AT2 cells. However, it also displays some tamoxifen-independent activity.
A knock-in Tgfb3-Cre line  has recently been used to manipulate Notch signaling in the postnatal lung airways . Reporter analysis suggested that this strain targets the majority (~90%) of Clara cells.
There is evidence to suggest that not all Clara cells are functionally equivalent [110, 111]. Transgenic strains, which target specific sub-sets of Clara cells would be highly desirable for the lung research community.
Foxj1 is a transcription factor which is expressed in all multiciliated cells including those of the lung, oviducts, ependyma and testes , and various cells with motile cilia [113-115]. A 1Kb fragment of the human FOXJ1 promoter was shown to be sufficient to direct reporter gene expression specifically in all of these cell types in adult mice . This promoter was subsequently used to generate FOXJ1-Cre , and FOXJ1-CreERT2 transgenic mice . Both lines drive efficient recombination in ciliated cells of the respiratory tract and have been useful for gene knock-out and lineage-tracing studies. In particular, the FOXJ1-CreERT2 mice were used to determine the average half-life of ciliated cells in the mouse airways . The FOXJ1-CreERT2 strain also unexpectedly directs recombination very efficiently in pericytes (J.R. Rock, B.L.M. Hogan, personal communication). This may reflect a low level of endogenous pericyte Foxj1 expression, or be due to the insertion site of the transgene. Similarly, a recent paper has shown that the same human FOXJ1 promoter can drive expression in human ciliated cells, but also some basal cells, growing at an air-liquid interface . Basal cell expression was not observed in the transgenic animals. However, any future transgenic strains generated with this promoter should be screened for basal cell and pericyte activity.
Published gene expression data have so far not identified a gene that is expressed exclusively in airway basal cells . A split-Cre, or viral, approach may be necessary for airway-specific basal cell genetic manipulation.
Keratin 5 (Krt5) and Keratin 14 (Krt14) promoters have been used to target basal cells in the airway epithelium. All mouse and human airway epithelial basal cells express Krt5 [120, 121]. A human 6kb KRT5 promoter fragment was cloned by the Fuchs lab and successfully used to target epidermal basal cells [122, 123]. Using the same promoter fragment KRT5-CreER2T transgenic mice were generated and used for cell lineage tracing in the airways. These studies demonstrated that airway basal cells are stem cells . This strain has subsequently been used for studying the control of basal cell function  and should prove to be generally useful for manipulating gene expression in tracheal basal cells. However, the KRT5-CreER2T transgenic strain is limited as it directs recombination in only about 15% of basal cells in the adult trachea. Moreover, the high levels of transgene activity in the skin and oral epithelium makes this strain extremely difficult to use for studies of oncogenes, as the mice will develop skin and oral tumors before the trachea is affected.
Krt14 is expressed in roughly 30% of mouse tracheal basal cells [120, 121, 125]. Transgenic mice containing the human KRT14 promoter linked to CreERT were generated for use in the skin . In the trachea these KRT14-CreERT transgenic mice allow tmx-induced recombination in an extremely small population of basal cells at steady-state . Following naphthalene injury, most surviving basal cells upregulate Krt14 and also express the transgene . This mouse has not yet been used to manipulate gene expression in airway basal cells. However, a K14-rtTA mouse  has been used, in combination with the tetO-Cre strain , to direct gene expression in the trachea . These data show the importance of careful control experiments, as in the K14-rtTA strain tracheal Clara cells were sensitive to dox exposure.
Pulmonary neuroendocrine (NE) cell differentiation depends on genes which are conserved in the nervous system of many organisms, for example Ascl1, NeuroD, Rb, and Gfi1 [129-132]. A rat NE cell specific promoter has been identified  and recently used in an adenovirus to direct gene expression specifically to NE cells . While no transgenic mouse strains, that specifically target NE cells have been generated, such strains could be very useful for studying NE cell function, or their putative role as an airway epithelial stem cell niche .
There is still much disagreement over the numbers of different lung mesenchymal cell types and their best markers . The challenge of the available mesenchymal mouse strains is that they are not lung-specific and that many of their expression patterns are highly variable depending on the integration of the transgene. Pod1 (also known as Tcf21) is highly expressed in the mesenchyme of the developing lung, kidney, heart and intestine and may be a useful promoter for more restricted mesenchymal gene manipulation . Tbx genes may also provide useful mesenchymal promoters. Tbx2-5 are expressed in the developing lung mesenchyme  and are often used as reporters of embryonic mesenchymal fate [139, 140]. However, the expression patterns of Pod1 and the Tbx genes in the adult lung mesenchyme have yet to be determined.
Dermo1-Cre (also known as Twist2-Cre) is a knock-in of Cre  that displays robust recombination in mesenchymal and mesothelial lineages in the lung . It has been widely used for manipulating gene expression in the developing lung mesenchyme [143, 144].
Mesp1-Cre is a knock-in of Cre, replacing the Mesp1 coding region, and drives expression throughout the anterior mesoderm from early gastrulation . It has been successfully used to manipulate gene expression in the developing lung mesoderm .
Fibroblast specific protein 1 (FSP1), also known as S100A4, has been reported as a fibroblast specific gene, but is also induced in epithelial cells during injury and tumor progression [146, 147]. The FSP1 promoter has been used to generate various FSP1-Cre mice with variable success . The use of these mice to study mesenchymal cells in the lung remains controversial.
Adipocyte lipid-binding protein 2 (aP2, also known as fatty acid binding protein 4, Fabp4) is expressed in alveolar type II cells and interstitial lipofibroblasts. The mouse aP2 promoter was used to generate aP2-Cre and aP2-CreERT2 mice . While this aP2-Cre line does not target AT2 cells it does target a subset of alveolar fibroblasts. Induction of recombination with the aP2-CreERT2 in adult mice was not observed (A.K. Perl, unpublished data).
The SM22-alpha (SM22α or transgelin) promoter was used to generate SM22-rtTA mice. This line provides a “Tet-On” tool that allows the inducible expression of genes in smooth muscle cells . Expression is mainly in the vascular smooth muscle and the SM22 promoter was used to generate several transgenic and knock-in CreERT2-expressing lines with varying expression patterns: with the highest levels detected in the aorta, intestine and uterus. However, none of these lines is particularly efficient, even in the vascular smooth muscle .
SMA (Smooth Muscle alpha Actin, or Acta2) is expressed in all smooth muscle cells in the adult, and also transiently in myocardiocytes and skeletal muscle during embryonic development [152, 153]. In lung parenchyma SMA is expressed during alveolarization, realveolarization, and during the development of lung fibrosis after bleomycin or hyperoxic injury [154-157]. Mice with a murine αSMA-Cre transgene express Cre in the airway smooth muscle and lung vasculature . This line is not inducible which limits postnatal studies. More recently an αSMA-CreERT2 BAC transgenic line has been generated and shown to exhibit tamoxifen-dependent Cre activity in all adult smooth muscle, including the lung airways and vasculature . These mice have off target Cre activity only in a small number of cardiomyocytes and should be very useful for future lung studies.
SMMHC (Smooth Muscle Myosin Heavy Chain, or Myh11) is expressed in all smooth muscle cells. Two independent mouse strains expressing Cre from a fragment of the mouse SMMHC promoter have been generated. The expression of the transgene is somewhat variable [160, 161]. This promoter leads to spurious CRE activity in some tissues due to expression in male and female germline . By contrast, a SMMHC-CreERT2 BAC transgenic strain shows inducible Cre activity in all smooth muscle, including the airway smooth muscle and lung vasculature . This strain should be useful for manipulation of gene expression in perivascular and peribronchiolar smooth muscle in the postnatal lung.
To direct Cre expression to the mesothelium of internal organs including the liver, gut and lung a Wt1-Cre YAC (yeast artificial chromosome) transgenic strain was generated , This has been shown to be active throughout the lung mesothelium from early developmental stages [163, 164]. However, it may be expressed at low levels in mesenchymal lineage of the embryonic lung and needs to be used with caution (B.L.M. Hogan personal communication). An inducible version would be useful for adult studies.
Multiple independent transgenic strains express Cre recombinase under the control of the mouse Tek promoter (Endothelial-Specific Receptor Tyrosine Kinase, also known as Tie2), Some of these have been very widely used [165, 166] [167, 168]. In these strains, reporter gene activity was detected in most endothelial cells and blood islands of the extra embryonic mesoderm by E7.5, in the dorsal aorta by E8.5 and in all blood vessels and some blood cells examined at E11.5, indicating that Cre was active in early vascular progenitors, endothelial cells and some hematopoietic cells. This promoter leads to spurious Cre activity in some tissues due to expression in male and female germline .
Wnt signaling pathways play divergent roles during development, homeostasis and repair and play a major role in stem cell proliferation and differentiation. Three transgenic reporter lines for Wnt pathway activity have been generated 1), TOPGAL, which reports epithelial Wnt signaling , 2) BATGAL with sporadic epithelial and mesenchymal activity  and 3) Axin2-lacZ, which is useful to study proximal lung and mesenchymal Wnt signaling . A recent study compared these lines during development and after naphthalene injury . A new reporter line, TCF/Lef:H2B-GFP, has not yet been tested in the lung .
Signals through the Notch receptors are used throughout embryonic development and in the adult to control cellular fate choices. CP-EGFP (also known as TNR) transgenic mice have a transgenic Notch reporter with an enhanced green fluorescent protein (EGFP) placed under the control of 4 tandem copies of a CBF1 (also known as Rbpj) responsive element (4 CBF1 binding site consensus sequences and the basal SV40 promoter) [174, 175]. This strain has been shown to faithfully report Notch activity in the adult trachea . N1IP::CRELOW and N1IP::CREHI are knock-ins of Cre, replacing the Notch1 coding region (Notch1 Intramembrane Proteolysis) and allow lineage studies of descendents of cells after Notch 1 activation [164, 177]. However, this Cre line identifies each cell lineage, which has previously experienced Notch activity and does not report on current signaling events. Comparison of various Notch reporter lines will shed better light on cell fate decisions and lineage relationships and lead to a better understanding of stem cell biology and interactions of epithelium, mesenchyme, mesothelium and endothelium during development and repair.
The lung is exposed to the external environment and multiple groups have taken advantage of this by administering viruses to manipulate gene expression in the adult mouse lung epithelium. The most widespread system is intranasal administration of an adenovirus expressing Cre from a ubiquitous CMV promoter (AdenoCre) to activate the expression of oncogenes and model lung cancer . A similar adenovirus-based approach has been taken to transiently overexpress specific genes throughout the lung epithelium (for example, ). More recently, adenoviruses using Scgb1a1, rat SftpC, or rat CGRP promoter fragments to direct Cre expression to specific epithelial cell types have been developed . Lentiviral vectors containing specific promoters for manipulating gene expression in restricted adult lung epithelial cell types have also been reported [180, 181]. In addition, Adeno-Associated Virus (AAV) transduction of mouse lung epithelial progenitors has also been demonstrated . The use of viral systems is likely to become more widespread over the next few years, particularly for epithelial studies.
Transgenic mice have been instrumental in developing our current understanding of lung embryonic development, adult homeostasis and repair. However, it is important to remember that all transgenic approaches have limitations, which can only be overcome by integrating findings from different lines and performing all the appropriate controls. New developments in mouse conditional genetics have the potential to further enhance our understanding of lung development and disease. Moreover, optimizing mouse strains of the existing doxycycline and Cre systems will increase flexibility and improve experimental design. For example, using the newer more dox-sensitive rtTA gene activation (rtTA2S-M2), or extremely low doses of tamoxifen in CreER based transgenic mice, will allow recombination in single cells and enable clonal cell type-specific gene manipulation. Due to the lack of lung-specific mesenchymal and endothelial gene expression, more lines need to be characterized for their usefulness in targeting specific subsets of mesenchymal and endothelial cells. On the other hand, complex targeting systems, such as the split-Cre will be helpful to target subsets of epithelial progenitor cells or specific mesenchymal cell lineages. Recently the applicability of the Flipase system to the lung has been demonstrated by combining Cre and FLP to independently control recombination of p53 and kRas in lung tumor progression . Development of tools based on flippase and viruses will further expand the combinatorial use of the existing mouse lines and help to develop newer lines, possibly overcoming the problems of off-target activation, lack of cell type specificity and lack of adult regulation. In addition, the generation of new publicly-available floxed alleles by the International Mouse Knock out Consortium (http://www.knockoutmouse.org/) and the use of transgenic mice expressing conditional RNA interference constructs, should facilitate mouse conditional genetic analysis.
Jeffrey Whitsett, Jason Rock, Barry Stripp and Brigid Hogan very kindly shared unpublished data. In addition, we thank Jeffrey Whitsett, Brigid Hogan, Jim Bridges, and members of our laboratories for critical comments on the manuscript.