Inducible Cre recombinase activity in mouse cerebellar granule cell precursors and inner ear hair cells



A transgenic mouse line expressing the CreERTM fusion protein under the control of the Math1 enhancer was generated. Expression of the transgene in the postnatal mouse was restricted to hair cells of the inner ear and granule neurons in the external granule layer of the cerebellum in a temporally regulated manner. Cre activity was virtually nonexistent in uninduced mice; however, treatment of newborn pups with tamoxifen, leading to nuclear translocation of the fusion protein, resulted in efficient recombination at LoxP sites in the appropriate cell types. Up to two thirds of cerebellar granule neurons and 80–90% of cochlear hair cells underwent Cre-specific recombination. This mouse line provides a powerful tool to dissect gene function at early and late stages in development of the cerebellum and inner ear. Developmental Dynamics 235:2991–2998, 2006. © 2006 Wiley-Liss, Inc.


Genetically altered strains of mice are an important tool to dissect gene function and to model human disease. The ability to express or inactivate genes in a tissue-specific manner (so-called conditional alleles) by using the bacteriophage P1 Cre recombinase-LoxP system or the yeast FLP-FRT variant has dramatically enhanced the power of mouse genetics (Kwan,2002). Nevertheless, depending on the target gene of interest, embryonic lethality may still result from conditional inactivation precluding analysis of gene function at later stages of development and in adult tissues. Furthermore, genes may play multiple developmentally regulated roles in the same tissue or cell type and may also have developmentally regulated expression patterns. Thus, the utility of conditional alleles is limited by the ability to regulate recombinase activity in vivo.

Experimental manipulation of gene expression and protein activity in space and time has been developed by several investigators (Maddison and Clarke,2005). One technique relies on the sequestration of steroid hormone receptors in the cytoplasm by Hsp90 in the absence of ligand and their translocation to the nucleus in a ligand-dependent manner (Feil et al.,1996; Hayashi and McMahon,2002). In addition, specific point mutations have been identified that render these receptors insensitive to endogenous steroid hormones while retaining binding to synthetic analogs, such as tamoxifen. Consequently, a fusion protein between Cre and the altered ligand-binding domain of the estrogen receptor (CreERTM) was demonstrated to mediate tamoxifen-dependent recombination at LoxP sites in vivo (Hayashi and McMahon,2002). This strategy has now been used by numerous investigators with success.

Math1, a mammalian homolog of Drosophila atonal, is a transcription factor involved in the development of the cerebellum and hair cells of the cochlear and vestibular organs (Ben-Arie et al.,1997; Bermingham et al.,1999). Characterization of the regulatory elements of the Math1 gene revealed that a 1.7-kb enhancer located within the 3′ untranslated region is sufficient to direct expression specifically to these locations (Helms et al.,2000). Herein, we describe the generation of a mouse line bearing a transgene expressing CreERTM under the control of the Math1 enhancer and characterize its inducible Cre activity.


The cDNA encoding the CreERTM fusion protein was placed downstream of the 1.7-kb Math1 enhancer (Fig. 1A). This was followed by an internal ribosomal entry site (IRES) and a cDNA encoding human placental alkaline phosphatase (hPLAP). The construct was terminated by an SV40 large T intron and polyadenylation signal. After pronuclear injection of FVB/NJ eggs, one founder was obtained. Transgene integration in this novel transgenic line, hereafter termed Math1-CreERTM, was verified by polymerase chain reaction (PCR) for Cre-specific sequences as well as by Southern blot hybridization, confirming a single integration site (Fig. 1B).

Figure 1.

Generation of transgenic mouse line. A: Diagram of transgenic construct. The cDNAs encoding the Cre-estrogen receptor ligand-binding domain (CreERTM) fusion protein, an internal ribosome entry site (IRES), and the human placental alkaline phosphatase (hPLAP), were engineered downstream of the 1.7-kb Math1 3′ enhancer. B: Southern blot. Genomic DNAs from the transgenic founder mouse and a wild-type littermate were digested with EcoRI (shown in A) to detect a fragment containing the junction of the transgene with the insertion site in genomic DNA. The probe was directed against the Cre sequence as indicated in A.

Transgene expression was characterized by histochemistry for hPLAP activity at different ages and in various tissues (Fig. 2). As expected, no hPLAP activity was detected in wild-type control animals (Fig. 2A,E,I,L,O). Endogenous Math1 is expressed as early as embryonic day (E) 9.5 in cerebellar precursors of the rhombic lip and continues postnatally in the external granule layer (EGL) until this structure disappears by postnatal day (P) 17 (Ben-Arie et al.,2000). Consistently, in postnatal brain from Math1-CreERTM mice, we detected hPLAP activity only in the EGL of the cerebellum (Fig. 2B,C,F,G) and the pontine nucleus (data not shown) and not in other regions of the brain (data not shown) as has been noted previously (Lumpkin et al.,2003). Within the EGL, expression was not uniform throughout all folia, with a greater proportion of expression in lobules VII–X (Fig. 2C). However, the pattern of expression was reproducible between different mice (data not shown). Expression was consistently seen from P0 until P6 at which point a gradual decrease in the intensity of histochemical staining was noted, becoming virtually extinguished by P14 (Fig. 2D,H). Therefore, the postnatal transgene expression pattern in the brain is largely in agreement with what has been reported for the endogenous Math1 gene, with subtle differences in temporal regulation.

Figure 2.

Transgene expression pattern by alkaline phosphatase histochemistry. A–H: Cerebellum. A wild-type control cerebellum at postnatal day (P) 0 is shown (A,E) along with Math1-CreERTM cerebella at P0 (B,F), P6 (C,G), and P14 (D,H). The external granule cell (e), and molecular (m) and internal (i) granule cell layers are indicated. Individual cerebellar lobules are identified by roman numerals. I–Q: Inner ear organs. Stains corresponding to the organ of Corti (I–K), crista ampullaris (L–N), and saccule (O–Q) are presented. Stains derived from wild-type control mice at P0 (I,L,O) are compared with Math1-CreERTM mice at P0 (J,M,P) and P6 (K,N,Q). The inner and outer hair cells of the organ of Corti are indicated by asterisks. Scale bars = 1 mm in A (applies to A–D), 50 μm in E (applies to E–H), 20 μm in I,L,O (applies to I–K, L–N, and O–Q, respectively).

In the inner ear, Math1 is only expressed in inner and outer hair cells in all cochlear turns as early as E13, becoming down-regulated after P4 and disappearing by P7 (Bermingham et al.,1999; Woods et al.,2004). In vestibular end organ, Math1 is also expressed in crista, and utricular and saccular hair cells at P0 (Bermingham et al.,1999). In the inner ear of Math1-CreERTM mice, hPLAP activity was detected in all populations of cochlear and vestibular hair cells at P0 (Fig. 2J,M,P). This finding persisted until P5, when much weaker activity was noted (data not shown), and by P6, activity was undetectable (Fig. 2K,N,Q). This pattern of expression is largely consistent with that reported for Math1 in the inner ear (Bermingham et al.,1999; Woods et al.,2004). No hPLAP activity was detected in a panel of tissues from P0 to P14 Math1-CreERTM mice, including muscle, skin, lung, thymus, heart, liver, spleen, kidney, adrenal gland, pancreas, bladder, uterus, ovary, testis, and epididymus (data not shown). Alkaline phosphatase activity was detected in the gastrointestinal tract; however, nontransgenic control animals also had detectable activity in this organ (data not shown); thus, the origin of the enzyme activity could not be ascertained in the gut.

Reporter mice were used to assess the activity of the CreERTM fusion protein following treatment with tamoxifen. Math1-CreERTM mice were crossed with Rosa26 reporter (R26R) mice in which the LacZ gene encoding β-galactosidase (β-gal) is expressed only after Cre-mediated excision of a LoxP-flanked stop cassette (Soriano,1999). Pups from such crosses were injected intraperitoneally with tamoxifen at a dose of 3–4 mg/40 g body weight or vehicle alone daily for one to three injections. Mice were killed at least 5 days after the last injection, and tissues were analyzed histochemically for β-gal activity.

When mice were treated with vehicle starting in the first 24 hr of life (P0), no β-gal activity was detected in the EGL (Fig. 3A,C,G,K), indicating minimal Cre activity in the absence of tamoxifen. Occasionally, some histochemical staining within cells of the pial membrane was noted in both R26R negative as well as Math1-CreERTM; R26R double-positive mice and, therefore represented endogenous β-gal activity (data not shown). In contrast, when double-positive pups received tamoxifen starting at P0, β-gal activity was noted in the EGL with a distribution paralleling the expression of the transgene (Fig. 3B,D,H,L). Furthermore, scattered activity was also detected in the internal granule layer (IGL), where the transgene is not expressed. This finding is consistent with recombination in precursors in the EGL at P0 and P1 followed by migration to the IGL before analyzing the brains at P6. Mice in which cerebellar development was allowed to proceed to maturation, demonstrated a higher proportion of β-gal–positive granule neurons in the IGL that ranged from 20–66%, depending on the mouse and individual folium analyzed (Fig. 3E,I,M). When tamoxifen was administered starting at P7, a much lower extent of Cre-mediated recombination was detected (Fig. 3F,J,N). Although Cre activity was detected at all mediolateral levels, it was not uniformly spread and was notably absent in the paramedial lobules (Fig. 3B,O–R). Despite transgene expression in the pontine nucleus, only sporadic and inconsistent β-gal activity was detected at this site (data not shown). This result could be explained by differences in the half-lives of CreER and hPLAP in these cells; alternatively, regional differences in tamoxifen access could result in tamoxifen levels below a required critical threshold for CreER activation in the pontine nucleus.

Figure 3.

Cre-mediated recombination in Math1-CreERTM; R26R cerebella by β-gal histochemistry. A,B: Whole-mount staining of postnatal day (P) 6 brains from a mouse injected with vehicle (A) or 3–4 mg/40 g body weight tamoxifen (B) at P0 and P1 are compared. C–N: Representative staining in the cerebella from Math1-CreERTM; R26R mice. C,G,K: Injected with vehicle alone (corn oil) at P0, P1, and P2 and killed at P7. D,H,L: Injected with 3–4 mg/40 g body weight tamoxifen at P0 and P1 and killed at P6. E,I,M: Injected with tamoxifen as above and killed at P17. F,J,N: Injected with 3–4 mg/40 g body weight tamoxifen at P7, P8, and P9 and killed at P16. O–R: Sections from various mediolateral levels of the cerebellum from the brain shown in B are presented. The external granule cell (e), molecular (m) and internal granule cell (i) layers are indicated as well as the paramedian lobule (PM) of the cerebellum. Scale bar = 1 mm in C (applies to C–F), 50 μm in G (applies to G–N), 1 mm in O (applies to O–R).

Tamoxifen-induced Cre activity was also assessed in the inner ear of these animals. In Math1-CreERTM; R26R mice, following three injections of vehicle alone starting at P0, a few rare cells stained for β-gal activity on whole-mount analysis (Fig. 4A, arrow). However, it is unlikely that these cells represented leaky cre activity in hair cells based on their position, and the histochemical staining of over 50 cochlear cryosections that did not identify β-gal–positive cochlear hair cells in vehicle-treated mice (Fig. 4C). No β-gal activity was detected in hair cells of the vestibular end organs in vehicle injected double-positive mice (Fig. 4D,E). Taken together, there appears to be a low degree of Cre activity in the absence of tamoxifen in the inner ear. Sporadic staining was noted in cells localized to the tissues and bones surrounding the inner ear, which may, in part, consist of osteocytes and hematopoietic cells (Tian et al.,2004). Importantly, a similar staining pattern was also seen in R26R-negative control mice, suggesting an endogenous origin for this β-gal activity (data not shown; Tian et al.,2004). When Math1-CreERTM; R26R mice were injected with tamoxifen, specific β-gal activity was detected in both inner and outer hair cells along the entire length of the cochlea (Fig. 4B,F), as well as in the vestibular sensory organs (Fig. 4G,H). Quantitative assessment of the efficiency of Cre induction revealed that a single dose of tamoxifen at P0 resulted in LacZ expression in 40% and 49% of inner and outer hair cells, respectively (Fig. 4I). By adding a second dose 24 hr later, the proportions increased to 80% and 96%, whereas no further improvement was obtained with a third dose. In contrast, Cre-mediated recombination in the vestibular hair cells was somewhat less efficient and continued to trend upward from one to three doses (Fig. 4J). Furthermore, when tamoxifen was administered starting at P6, almost no Cre activity was induced in the hair cells of the cochlea, saccule, or utricule (data not shown). Thus, the pattern of Cre induction coincides with expression of the transgene at the time of tamoxifen administration. Unexpectedly, crista hair cells demonstrated a similar level of β-gal activity when tamoxifen was administered at P6 compared with induction at P0 (data not shown). It is possible that there is relatively less hPLAP expression compared with CreER as sequences downstream of an IRES may be translated less efficiently. Alternatively, the half-life of the Cre protein may be prolonged relative to hPLAP in this population of hair cells, leading to recombinase activity in the absence of detectable hPLAP expression (Fig. 2N). Additionally, the accessibility of tamoxifen may vary with developmental stage resulting in changing sensitivity to induction.

Figure 4.

Cre-mediated recombination in Math1-CreERTM; R26R hair cells by β-gal histochemistry. A,B: Whole-mount staining of postnatal day (P) 6 cochleas from a mouse injected with vehicle (A) or 3–4 mg/40 g body weight tamoxifen (B) at P0 and P1 are compared. Arrows indicate specific staining of inner and outer hair cells. C–H: Mice were injected with vehicle alone (C–E) or 3–4 mg/40 g body weight tamoxifen (F–H) at P0, P1, and P2 and killed at P7. C–H: Representative β-gal histochemistries from the organ of Corti (C,F), crista ampullaris (D,G) and saccule (E,H) are presented. The inner and outer hair cells of the organ of Corti are indicated by asterisks. I,J:Quantitation of CreER activity in inner (IHC) and outer (OHC) hair cells of the organ of Corti (I) and vestibular hair cells (J) is described in The Experimental Procedures section. The percentage (range) of X-gal–positive cells for one, two, and three tamoxifen injections, respectively, in the IHC was 40% (19–52), 80% (74–83), 77% (66–85); for the OHC 49% (41–53), 96% (92–99), 90% (85–95); for the crista ampullaris 10% (5–21), 16% (5–21), 21% (13–31); for the saccule 40% (34–49), 53% (39–74), 62% (50–83); and for the utricle 43% (30–50), 60% (54–69), 64% (52–75). Scale bars in C–E represent 20 μm and apply to F–H, respectively.

The Math1-CreERTM mouse line described in this report represents a very useful reagent for researchers interested in cerebellar granule cell development and function, as well as those in the hearing field. Indeed, the early postnatal period is a time when both structures are undergoing critical and dynamic changes. Most of the development of the cerebellum occurs postnatally. Granule neuron precursors proliferate rapidly in the EGL before exiting the cell cycle, migrating to their final location in the IGL and forming critical synaptic connections with both mossy fibers and Purkinje cells (Goldowitz and Hamre,1998). In fact, the acquisition of critical mutations within the granule neurons at this stage of development can disrupt normal cerebellar development and can also contribute to malignant transformation, resulting in medulloblastoma formation (Marino,2005). This Cre line provides a tool to test such mutations at various points in the maturation of this cell population.

In the inner ear, no successful inducible CreER mouse lines have been reported to date (Tian et al.,2006). Although development of the inner ear begins during embryogenesis, significant reshaping of the innervation pattern takes place within the first postnatal week. Hair cells also undergo dramatic changes with the gain of electromotility right after P7 (Belyantseva et al.,2000) and synaptogenesis of the inner and outer hair cells at around P6 (Sobkowicz et al.,2002). This inducible Cre line will allow for the specific inactivation of genes involved in synaptogenesis and electromotility formation. Thus, we may begin to dissect temporally distinct gene functions between the embryonic stage and the postnatal maturation period before hair cells become fully functional at around P21. Importantly, mice injected with up to three doses of tamoxifen according to the schedule described above retain normal hearing at 1 month of age as determined by ABR threshold measurements (data not shown). In fact, the only manifestation of tamoxifen toxicity was a transient dose-dependent weight loss, which normalized by 2 months of age (data not shown).

Recently, a similar transgenic mouse line was described in which the CreERT2 fusion protein (Indra et al.,1999), was expressed under the control of the identical Math1 enhancer (Machold and Fishell,2005). This report focused on the embryonic expression pattern and induction of the transgene. Cre activity in the inner ear and postnatal induction of Cre activity was not described. The CreERTM version used in our study and CreERT2 contain different mutations in the estrogen receptor ligand-binding domain that may result in different sensitivity to tamoxifen (Feil et al.,1997). However, the expression pattern, expression levels, and therefore, responsiveness to tamoxifen may vary between transgenic lines based on insertion site effects and must be empirically tested. Together, these new Cre transgenic mouse lines provide characterized reagents for genetic manipulations in both early and late stages of the developing cerebellum and inner ear.


Transgenic and Reporter Mice

Standard subcloning techniques were used to construct the transgene, placing the 1.7-kb Math1 3′ enhancer (a gift from J. Johnson) upstream of the cDNA encoding the CreERTM fusion protein (provided by A. McMahon). This sequence was followed by an IRES and the cDNA encoding hPLAP to facilitate detection of transgene expression. The construct was terminated by an SV40 intronic sequence and polyadenylation signal. Following pronuclear injection into fertilized FVB/NJ oocytes and implantation into foster mothers, the transgenic founder was identified by PCR with primers recognizing the Cre sequence (5′ AGCGATCGCTGCCAGGAT 3′ and 5′ ACCAGCGTTTTCGTTCTGCC 3′). Unique integration of the transgene was confirmed by Southern blot using a probe recognizing the Cre sequence. The mouse line was propagated by crossing with FVB/NJ wild-type mice. Cre activity was mapped by crossing into the Rosa26 reporter mouse line (Soriano,1999; Jackson Laboratory, Bar Harbor, ME). Reporter-positive mice were identified by PCR using primers to the LacZ gene (5′ GCTGGGATCCGCCATTGTCAGAC- ATG 3′ and 5′ GCTGGAATTCCGCCGATACTGAC 3′). All experiments involving mice were carried out in compliance with the Animal Care and Use Committee at St. Jude Children's Research Hospital.

Induction of Cre Activity

Tamoxifen (Sigma, St. Louis, MO) was dissolved in corn oil (Sigma) at a concentration of 5 mg/ml at 37°C, and then filter sterilized and stored for up to 7 days at 4°C in the dark. A 30-Ga needle insulin syringe (Becton Dickinson, Franklin Lakes, NJ) was used for intraperitoneal injections in pups. The dose administered to pups varied from 3–4 mg/40 g body weight. An equivalent volume of sterile filtered corn oil alone was used for vehicle injections. Multiple injections in the same mouse were separated by 24 hr.


Animals were perfused with 1× phosphate-buffered saline (PBS) followed by 2% paraformaldehyde (PFA) in PBS. After dissection, tissues were post-fixed overnight in 2% PFA in PBS at 4°C, and then equilibrated to 25% sucrose in PBS for a further 24 hr at 4°C. Inner ears from P10 and older mice were decalcified in 120 mM ethylenediaminetetraacetic acid (EDTA) for 24 to 48 hr at 4°C before equilibration in sucrose. Tissues were subsequently embedded and cryosectioned at a thickness of 12 μm (brain) or 8 μm (inner ear). Tissue slides were washed three times in PBS before staining.

Before the hPLAP assay, endogenous phosphatases were inactivated by incubation in PBS at 65°C for 1 hr. Slides were then equilibrated with AP Detection Buffer (100 mM Tris, pH 9.5, 50 mM MgCl2; 100 mM NaCl) for 10 min then stained in the same buffer containing 0.375 mg/ml nitroblue tetrazolium chloride (NBT) and 0.188 mg/ml 5-bromo-4-chloro-3-indolyl phosphate (BCIP) at room temperature overnight. After extensive washes in PBS, slides were counterstained with Methyl Green (Vector Labs, Burlingame, CA).

For detection of β-gal activity, slides were washed in Rinse A (100 mM NaPO4, pH 7.3; 2 mM MgCl2; 5 mM ethyleneglycoltetraacetic acid [EGTA]) for 30 min followed by Rinse B (100 mM NaPO4 pH 7.3; 2 mM MgCl2; 0.01% sodium deoxycholate; 0.02% NP-40) for 5 min. Slides were then stained at 37°C for 3–4 days in Developing Buffer (1 mg/ml X-gal; 100 mM NaPO4 pH 7.3; 2 mM MgCl2; 0.01% sodium deoxycholate; 0.02% NP-40; 5 mM K3Fe(CN)6; 5 mM K4Fe(CN)6). After washing extensively in PBS, slides were counterstained with Nuclear Fast Red (Vector Labs, Burlingame, CA).

For whole-mount analyses of brains and cochleas, tissues were post-fixed for 5 hr in 2% PFA in PBS at 4°C, washed extensively with PBS and then processed as above for detection of β-gal activity. In the case of cochlear whole-mounts, the otic capsule was punctured and Developing Buffer injected into the round and oval windows. Tissues were then post-fixed again in 4% PFA in PBS at 4°C for 24 hr, washed in PBS, and analyzed.


The percentage of granule neurons which underwent Cre-mediated recombination was determined in five Math1-CreERTM; R26R mice injected with 3–4 mg/40 g body weight tamoxifen at P0 and P1 and killed at P17–51 by counting at least 500 nuclei/lobule (both positively and negatively stained with X-gal) in seven to nine cerebellar lobules per mouse. For quantitation in the inner ear, positively or negatively stained hair cells were identified morphologically and counted from an average of 17 sections each from four Math1-CreERTM; R26R mice that received either vehicle, or one, two, or three doses of tamoxifen at a dose of 3–4 mg/40 g body weight administered daily starting at P0.


We thank J. Johnson and A. McMahon for the gift of reagents. J. Raucci and G. Heath of the St. Jude Children's Research Hospital Transgenic Core provided technical expertise in generating the transgenic mouse. J. Mitchell's role in genotyping mice was indispensable. L.M.L.C. is a recipient of the Jean-François St.-Denis Fellowship in Cancer Research from the Canadian Institutes of Health Research. T.W. is supported by a grant from the Deafness Research Foundation. S.J.B. and J.Z. are supported by grants from the National Institutes of Health and the American Lebanese and Syrian Associated Charities of St. Jude Children's Research Hospital.