SEARCH

SEARCH BY CITATION

Keywords:

  • Tamoxifen-dependent Cre;
  • Col2a1 transgenic promoter;
  • chondrogenic deleter line;
  • CreERT in cartilage

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

Cartilage differentiation occurs over a broad time range from early embryonic development, when the mesenchymal condensations that give rise to cartilage models for future bone first appear, and continuing through adult life, when there is ongoing maintenance of articular joint surfaces and re-activation of cartilage formation after fracture. The chondrogenic response also figures in the pathogenesis of degenerative and inflammatory joint diseases. We have generated a transgenic line expressing tamoxifen-dependent Cre recombinase that gives efficient recombination in the chondrogenic lineage, both during embryogenesis and postnatally, and provides a valuable tool for analysis of gene function selectively in chondrogenic cells using conditional genetic approaches. Because the cartilage model of the limb skeleton forms progressively in a proximodistal order during discrete, well-defined time periods, evaluation of the spatial extent of tamoxifen-induced recombination along the limb axis during these time windows has also enabled us to examine the pharmacokinetics of single-dose tamoxifen injections during pregnancy. Developmental Dynamics 235:2603–2612, 2006. © 2006 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

Tamoxifen-regulated CreERT transgenic lines (Feil et al.,1996; Brocard et al.,1997; Vooijs et al.,2001; Hayashi and McMahon,2002; reviewed by Nagy,2000) offer versatile tools to dissect the different functions of genes that play critical roles at multiple stages of development and postnatally and, thus, have the potential to greatly increase our understanding of the regulation of tissues and cell types that form over a broad developmental time range. They have also proven to be very powerful tools for lineage and fate mapping studies in mice (e.g., Kimmel et al.,2000; Harfe et al.,2004; Ahn and Joyner,2004). Another arena in which such tools may be of particular benefit is in the analysis of chondrogenesis, which occurs over a broad time period spanning embryogenesis and postnatal life. Cartilage provides the primary model and foundation upon which most of the future axial and appendicular bony skeleton is built (by means of enchondral bone formation), forms the terminal growth plates for skeletal elongation, and also provides the smooth, frictionless surfaces for mature mobile joints (reviewed by de Crombrugghe et al.,2000; Kronenberg,2003; Archer and Francis-West,2003). A large number of signaling and transcription factors have been implicated in the formation and maintenance of chondrogenic tissue, as well as in the signaling interactions between perichondrium and cartilage that regulate longitudinal growth and, later on, osteogenesis (reviewed by de Crombrugghe et al.,2001; Karsenty and Wagner,2002; Karsenty,2003; Kronenberg,2003). Regulation of the chondrogenic lineage also plays an important role in the pathogenesis and progression of pathological conditions such as inflammatory and degenerative joint diseases during adult life (reviewed by Archer and Francis-West,2003; Buckwalter et al.,2005) and contributes to reparative processes such as fracture healing (reviewed by Ferguson et al.,1998; Uusitalo et al.,2001).

The Col2a1 gene encoding the major cartilage specific (Type II) collagen is expressed fairly early and very robustly within the chondrogenic lineage, as well as in certain other tissues, such as the notochord and the apical ectodermal ridge (AER) of the limb bud, albeit at much lower levels (Cheah et al.,1985,1991; Lui et al.,1995). Furthermore, the regulatory elements driving expression of the Col2a1 gene have been functionally dissected and very well characterized in transgenic mouse lines. A small enhancer region located within the first intron of the Col2a1 gene, in conjunction with ∼1kb of the Col2a1 promoter region (see Fig. 1A), efficiently drives transgenic reporter expression and recapitulates the tissue-specific patterns of Col2a1 gene expression in vivo (Mukhopadhyay et al.,1995; Zhou et al.,1995; Krebsbach et al.,1996; Tsumaki et al.,1999; Ovchinnikov et al.,2000). A conditionally active Cre recombinase under the control of the Col2a1 promoter would enable the use of this robust promoter to drive expression transiently in the chondrogenic lineage without the unwanted side effects of expression in other critical signaling structures in the early embryo (e.g., notochord, AER). Conditional expression would also allow the evaluation of gene roles in postnatal processes without perturbing the developmental chondrogenic program. We have generated, therefore, a CreERT line under the control of the Col2a1 promoter to achieve selective and temporally regulated recombination in tissues expressing Type II collagen transcripts. Screening multiple transgenic founders identified a line that gives both robust and tightly tamoxifen-controlled Cre recombinase activity during embryonic development and early postnatal life, while skeletal growth is still ongoing. The known temporal sequence of formation of different proximodistal components of the limb skeleton has also enabled us to estimate the kinetics of single-dose tamoxifen activity in pregnant animals using this line. Effective tamoxifen levels for embryonic Cre activation were restricted to a fairly narrow temporal window of approximately 12 hr.

thumbnail image

Figure 1. Generation and activity of a Col2CreERT line in early embryos. A: Schematic diagram of transgenic Col2a1CreERT expression construct used to generate transgenic mice, which includes 1 kbp of the Col2a1 promoter, the β-globin intron (IVS), the SV40 polyA additional signal, and the Col2a1 first intron enhancer as a 650 bp fragment downstream of the polyA signal. B: Overview of tamoxifen-dependent recombination in axial and appendicular cartilage during early development (embryonic day [E]10–E12.5). A single dose of 1.5 mg of tamoxifen was injected intraperitoneally (IP) at the times indicated in each panel, and embryos were harvested and stained for LacZ activity at 24 hr postinjection. The lower right panel shows a representative parasagittal section through sclerotomes stained for LacZ reporter activity, of embryo from an E10 tamoxifen injection. No LacZ activity was seen in cartilage when tamoxifen injections were given before E9 (data not shown).

Download figure to PowerPoint

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

CreERT was introduced into an expression vector in between 1 kbp of the native Col2a1 promoter and the Col2a1 intron 1 enhancer cloned in the 3′ noncoding region (Fig. 1A; Tsumaki et al.,1999). Twelve transgenic founders were evaluated in crosses to the Rosa26 LacZ reporter line to generate embryos for analysis of Cre recombinase activity by assessing the LacZ reporter activation. Seven of the founders showed strong Cre activity in these crosses (judged by LacZ reporter activation in mid-stage embryos), but the majority of these (five of seven) also displayed substantial mosaic Cre activity even in the absence of tamoxifen IP (intraperitoneal) injection (data not shown). Our experience with CreERT expression controlled by transgenic promoters has suggested that very strong promoters driving expression are more often associated with some leakiness (loss of tamoxifen dependence), possibly owing to low level proteolysis of CreERT protein to give an active, non–ligand-dependent Cre fragment (E. Nakamura and S. Mackem, unpublished observations). Consequently, there is a trade-off between completely tamoxifen-controlled lines that have very weak expression of CreERT and inefficient induction of recombination, and robustly expressing lines that display rapid and complete recombination, but may have some low-level recombination without tamoxifen. Two founders with minimally leaky (estimated at <0.1%; see Fig. 2A) but efficient tamoxifen-induced recombination were identified and were used to establish lines for more extensive analysis. Both of these lines were extensively evaluated with comparable results; the characterization of Cre recombinase activity for the more strongly expressing of these two lines are presented in this report.

thumbnail image

Figure 2. Time course and recombination efficiency in developing limb components after single-dose tamoxifen injection during mid-embryogenesis. AF: Tamoxifen was injected IP (1.5 mg), and embryos were harvested at embryonic day (E) 13.5 age, after the time intervals indicated in each panel. Top panels show whole embryos; middle panels show isolated forelimbs (on left) and hindlimbs (on right) from each stage, dorsal side with anterior to right and posterior to left; and bottom panels show histological sections through limb cartilage. Sections were taken through long bone limb cartilages, which are already undergoing chondrogenic differentiation over the entire time span from E12–E13.5. Embryos were stained in whole-mount for LacZ activity and multiple sections taken from each time interval were stained for LacZ activity. Representative sections are shown in the lower panels for each time point. E: Recombination, as judged by the percent LacZ+ cells, was essentially complete by 24 hr. E,F: In sections from the 24 and 36 hr times postinjection, note that efficient recombination is also evident in perichondrium.

Download figure to PowerPoint

In preliminary studies, the dose-response of Cre activity to tamoxifen was evaluated. Comparable to the results previously reported by Hayashi and McMahon (2002), 0.5-mg or 1-mg single IP injections to pregnant females at embryonic day (E) 12.5 resulted in clearly mosaic, incomplete recombination, whereas injections of 1.5-mg or higher doses gave similarly efficient recombination levels at 24 hr (see Supplementary Figure S1, which can be viewed at http://www.interscience.wiley.com/jpages/1058-8388/suppmat). To minimize side effects, we subsequently used the lowest dose that reproducibly gave efficient recombination (1.5 mg IP; ∼25-g weight female) to characterize the spatial distribution, extent, and kinetics of recombination in the Col2CreERT line at different stages of embryogenesis. Tamoxifen-related side effects were also reduced by coadministration of progesterone (see the Experimental Procedures section).

To evaluate the overall Cre activity during early embryogenesis, tamoxifen was administered in a single IP dose (of 1.5 mg) beginning from E9, and embryos were harvested at 24 hr postinjection to evaluate the level and pattern of Cre recombinase activity (Fig. 1B). Weak activity in facial condensations and rostral sclerotomes was already evident by E10 (after injection at E9) and progressed caudally along the trunk after single tamoxifen injections given over the ensuing 60-hr period, as well as appearing in the limb buds and the base of the skull region. When IP injections were given at E8.5 or earlier, no significant Cre activity was seen in prechondrogenic tissues using the Rosa26LacZ reporter embryos (data not shown).

By E13.5 most of the cartilage model for skeletal elements has formed in the skull base, face, axial trunk region, and the limbs. To determine the efficiency and time course for complete recombination, tamoxifen was injected at different time intervals ranging from 8 to 36 hr before harvesting embryos at E13.5 for analysis of LacZ activation (Fig. 2). LacZ activation was first detected in scarce cells at 8 hr (Fig. 2B), and considerable mosaic recombination was already evident within 12 hr of injection (∼25%, Fig. 2C). Between 16 and 24 hr, extensive recombination was observed, approaching completion by 24 hr. At 24 hr, some cells were much more weakly stained than others, probably reflecting variable levels of LacZ protein rather than mosaic recombination at this time point (compare Fig. 2A and E). This finding was confirmed by checking reporter expression after 36 hr, at which time all cells clearly expressed LacZ strongly, but the intensity of LacZ activity was still seen to be variable from cell to cell, probably reflecting the range of recombination time. As seen in Figure 2E,F, the forming perichondrium also strongly expressed the CreERT transgene.

Because the cartilage model of the limb skeleton forms progressively in a strict proximodistal (P-D) order during discrete, well-defined time periods, evaluation of the P-D extent of recombination induced by tamoxifen injections given during this time window has also enabled us to evaluate the kinetics of single-dose tamoxifen administered to pregnant female mice. By comparison of LacZ expression driven from the same Col2a1 promoter as used for CreERT with the LacZ reporter activation by a previously described Col2Cre transgenic line (Ovchinnikov et al.,2000), it is evident that Cre-mediated recombination is quite rapid; the Col2Cre activity and Col2LacZ expression appear very similar at all stages evaluated (Fig. 3). However, there is an approximate up to 12-hr delay in comparable expression levels of LacZ activity from either Col2LacZ or reporter activation by Col2Cre transgenes, as compared with the first appearance of endogenous Col2 transcripts (Fig. 3). This delay could partly be related to timing for synthesis of an enzymatically active, large tetrameric LacZ protein, but more likely also reflects some difference in onset of activity of the endogenous promoter vs. the transgenic promoters used. Compared with either the Col2LacZ or Col2Cre transgenes, which behave very similarly, the Col2CreERT activity after tamoxifen injection shows a delay of an additional 12 hr in the appearance of LacZ activity in the limb bud, which is seen consistently during the successive appearance of Col2a1 RNA in each forming P-D skeletal component between E10 to E12.5 (Fig. 3; Table 1). The additional delay indicates that there must be an up to 12-hr lag for absorption, metabolism, and delivery of the more highly active 4-OH-tamoxifen metabolite to reach effective levels in the embryonic circulation. This result is consistent with the time course of recombination shown in Figure 2, where significant levels of recombination became evident at 12 hr after injection.

thumbnail image

Figure 3. Comparison of timing of endogenous Col2 transcript expression, Col2LacZ transgene activity, and of Col2Cre- or Col2CreERT-mediated recombination in forming limb cartilaginous skeleton. In situ hybridization of Col2a1 riboprobe in forelimb (FL) and hindlimb (HL) buds at embryo ages (indicated above) compared with Col2LacZ activity in transgenic embryos at same age, and with activity of Rosa26LacZ reporter in Col2Cre or in Col2CreERT transgenic embryos. For Col2CreERT transgenic embryos, tamoxifen was injected (1.5 mg) 24 hr before harvesting at the age shown. In each panel, limb buds are oriented with the anterior border at the top. Note the endogenous Col2a1 RNA expression in the apical ectodermal ridge (AER), which appears patchy because of partial AER loss owing to the level of protease treatment necessary for good probe penetration into limb mesoderm.

Download figure to PowerPoint

Table 1. Embryonic Timing of Onset of Rosa26LacZ Reporter Activation in Different Limb Components by Col2Cre Compared with Col2CreERT After Single Tamoxifen Dosea
 Time of onset of LacZ activity:Earliest effective Tam. Txb
Rosa26 X Col2CreRosa26 X Col2CreERT (24 hr, short-term)Rosa26 X Col2CreERT (E13.5, long-term)
  • a

    Summarized from Figures 3 and 4 (and data not shown).

  • b

    Tam. Tx, tamoxifen treatment (intraperitoneal [IP] injection time).

  • c

    FL, forelimb.

  • d

    Zeugopod, paired long bone elements: radius, ulna for forelimb and tibia, fibula for hindlimb.

  • e

    Weak, mosaic LacZ activity is evident.

  • f

    HL, hindlimb.

ScapulaE11E11.5E10
HumerusE10.5E11.25eE9.5
FLc zeugopoddE11.5E12E10.5
FL digitsE12.5E13E11.5e
FemurE11.5E12E10.5
HLf zeugopodE12E12.5E11
HL digitsE13>E13E12
thumbnail image

Figure 4. Determination of duration of Col2CreERT activity after single tamoxifen injection from spatial extent of LacZ reporter activation along the limb proximodistal (P-D) axis. A single dose of tamoxifen (1.5 mg) was injected at the times postcoitum as indicated in each panel, and embryos were collected and stained for LacZ activity at embryonic day (E) 13.5. Forelimbs are on the left, and hindlimbs on the right in each panel, with anterior side to left. Note that Cre activity first becomes apparent in the scapula later than in the humerus (see also Table 1); this finding is consistent with the relatively later onset of endogenous Col2a1 RNA expression in the scapular condensation as seen in Figure 3 and previously reported (Pellegrini et al.,2001).

Download figure to PowerPoint

To estimate the duration of effective tamoxifen levels after IP injection, tamoxifen was administered at different times and embryos were allowed to develop until E13.5 before analysis to assess how early a single dose can be given and still achieve efficient Cre activation and recombination within a given, forming P-D limb component (Fig. 4). The P-D order of formation of limb condensations is well established, and the timing of onset of transgenic Col2 promoter activity in the different P-D condensations is clearly seen in Figure 3 (see comparable timing in Col2LacZ and Col2Cre panels), allowing the duration of tamoxifen action to be readily inferred from these data. Tamoxifen administered up to ∼24 hr before the expected onset of LacZ reporter activation by the transgenic Col2Cre within a given P-D limb component was consistently capable of inducing some recombination mediated by Col2CreERT in that particular limb element (Fig. 4; Table 1). Considering the 12-hr lag to first achieve an effective level, this finding indicates that effective tamoxifen levels are maintained (after a single injection) for approximately 12 hr after reaching an effective peak in the embryonic circulation. Given the uniform tissue densities and circulatory distributions in early embryos, these pharmacokinetics are likely to apply for tamoxifen-regulated Cre lines using promoters that are highly expressed in other cell types during early to mid-embryogenesis.

In a previous study that examined these kinetic parameters in embryos for a ubiquitously expressed CAGG-CreERT transgenic line (Hayashi and McMahon,2002), a similar lag time to reach effective tamoxifen levels was observed; however, significant recombination was still reported between 24 and 48 hr after injection. The prior report examined kinetics mainly at E8.5, before establishment of a fetal–placental circulation, and focused largely on dose-dependency of recombination efficiency. In the present report, the kinetics have been evaluated largely after the establishment of placental function, which may alter both drug delivery and clearance rates, and discrete times for the onset of Col2 promoter activity in different P-D limb components were used as endogenous temporal guides for assessing the duration of tamoxifen-induced Cre activity. Probably because of substantially weaker promoter activity, in the previous report (Hayashi and McMahon,2002), much higher doses of tamoxifen were often used to achieve complete recombination (6–9 mg/40 gm) than were found to be adequate for the Col2CreERT line reported here (usually equivalent to 2 mg/40 gm). Such substantially higher doses would also alter the total duration of effective levels for a given half-life.

Notably, with regard to the Col2a1 promoter, the level of LacZ reporter activity in the AER is considerably lower and more temporally confined in both Col2LacZ and Col2CreERT transgenic embryos compared with the Col2Cre transgenic embryos (Fig. 3). LacZ reporter activation due to Col2Cre reflects cumulative recombinase activity over time; consequently considerable LacZ reporter activity is already evident by E10.5 in the AER of Col2Cre embryos, where the Col2a1 promoter is weakly expressed (Lui et al.,1995; and see Fig. 3), and this LacZ activity becomes quite strong in the AER by E12 (Fig. 3). In contrast, the Col2LacZ transgene expression is weaker, being most evident at around E11.5 (for forelimb), and is more akin to endogenous Col2a1 transcript expression levels. Consistent with this low-level expression peaking at approximately E11.5, weak reporter activation by the Col2CreERT transgene is observed in the AER after tamoxifen injection at E11. This weak activity is also 12 hr delayed relative to Col2LacZ transgene expression in AER and is evident at ∼E12 after an E11 injection, but it is not apparent at E11.5 after an earlier injection, again because of the short duration of effective CreERT activity (∼12 hr). Later injections after E11 also did not reveal significant reporter activity in AER. The limited time period during which significant Cre activity occurs in the AER after tamoxifen treatment has the advantage that, over much of the time span for limb development, unwanted effects from AER expression can be minimized using the Col2CreERT line (Fig. 3).

We have also evaluated the efficiency of Col2CreERT-mediated recombination during later stages of embryogenesis (Fig. 5), as well as in the neonatal and early postnatal periods (Figs. 5, 6). Good recombination in growth plates and articular surfaces was evident through at least postnatal (PN) day 16, despite the complete avascularity of joint cartilage. Note that an endogenous galactosidase activity begins to appear in bony tissue at these stages in the wild-type normal skeleton (indicated by brackets in Fig. 5; see also Hogan et al.,1994). However, the cartilaginous joint and growth plate regions are negative for endogenous enzyme activity, as well as LacZ reporter activation, in the absence of tamoxifen treatment, but the Col2CreERT+ animals show efficient activation of the LacZ reporter in growth plates and joints (articular surface perichondrium) after tamoxifen administration (highlighted by asterisks in Fig. 5). We have observed that efficient recombination after tamoxifen is still achieved at PN day 21, but tamoxifen-induced Cre activity declines in adult mice and is essentially undetectable (as tested using the Rosa26LacZ reporter) in 12-week-old animals (data not shown).

thumbnail image

Figure 5. Late fetal and neonatal tamoxifen injection gives efficient Col2CreERT-mediated recombination in cartilage growth plates and articular joint surfaces. A,B: Tamoxifen (6 mg) injected at embryonic day (E) 13.5 and embryos harvested at E16.5. Negative control shows LacZ activity in Col2CreERT embryos at same age without tamoxifen treatment. Bars indicate bony regions of long bones, which begin to show some endogenous galactosidase activity at this age (also observed in wild-type, nontransgenic embryos; see top of D). Note however, there is no LacZ activity seen in cartilaginous tissues in absence of tamoxifen treatment. Asterisks (*) indicate growth plate and articular regions showing LacZ activity induced by tamoxifen treatment. A: Forelimbs with or without tamoxifen treatment. B: Hindlimbs with or without tamoxifen treatment. C: LacZ staining of articular surface (knee joint) after tamoxifen treatment in the immediate neonatal period (0.2 mg × 2 doses) and harvested for staining at postnatal day [PN] day 3). Staining was carried out for short (1- to 2-hr) reaction period, so that endogenous bony galactosidase activity does not obscure detection of LacZ reporter activity in articular cartilage. D: LacZ staining in growth plates of long bones (forelimbs shown) after tamoxifen injection at PN day 1 (1 mg). Brackets indicate bony regions with endogenous galactosidase activity, and asterisks (*) indicate growth plate and joint regions that do not display endogenous activity but show high level LacZ reporter activation due to tamoxifen-activated Cre recombination (as in A,B).

Download figure to PowerPoint

thumbnail image

Figure 6. Postnatal (PN) tamoxifen injection in 2-week-old pups produces efficient Cre-mediated recombination in growth plates and articular cartilage. A: Tamoxifen was injected at PN day 14, and LacZ reporter activation in limb was examined at PN day 16. The middle panel shows the short 1-hr reaction period when no endogenous bony galactosidase activity is detected (compared with left panel control without tamoxifen treatment), and right panel shows prolonged reaction time to demonstrate the high level LacZ activation in growth plates. LacZ reporter activity is easily detected in growth plates (arrows), even after very short staining reaction. B: Representative sections through metatarsal joint after injection at PN day 14. The left panel shows hematoxylin stain of section from Col2CreERT transgenic pup not treated with tamoxifen (no LacZ activity was detectable in sections), and the right panel shows a paraffin section stained for LacZ activity taken from a tamoxifen-treated transgenic pup. Note efficient recombination, as demonstrated by LacZ reporter activation seen in all cells throughout chondrogenic tissues, including the proliferating cartilage zone in growth plates and the perichondrium overlying articular surface cartilage.

Download figure to PowerPoint

As mentioned above, we noted that at ∼E13.5, LacZ reporter activity was also sometimes evident in the early perichondrium layer, presumably reflecting previous Col2CreERT expression in incipient perichondrium forming from mesenchymal condensations concurrent with the onset of chondrogenic differentiation (see Fig. 2). The perichondrium has been shown to be a major source of periosteum and osteoblast progenitors (Colnot et al.,2004; reviewed by Colnot,2005), giving rise to the bone collar including both cortical and trabecular bone in the spongiosa. Hence, dependent on the timing of tamoxifen administration, the Col2CreERT line described here may also have utility in marking and driving recombination in early perichondrial progenitors, and inherited by their differentiated descendants (periosteum and bone). However, tamoxifen injections given after the formation of a distinct perichondrium (E13.5 and later; see, e.g., Figs. 5, 6, and data not shown) generally did not result in any evidence of subsequent LacZ reporter expression in bone, indicating that the likely time-window for Col2CreERT expression in perichondrial progenitors may be transient and restricted to the immediate period during which perichondrium and chondrocytes initially differentiate from the condensed mesenchyme (see Table 1 for “time windows” of chondrogenesis onset; and reviewed by Colnot,2005). To confirm a transient perichondrial expression window, we analyzed the contribution of LacZ+ recombined cells to periosteum and bone osteoblasts in the humerus at E15.5, after single tamoxifen injection at several different times before E13.5. Injection before E12.5 resulted in substantial LacZ activity in bone cortical and trabecular (spongiosa) osteoblasts (Fig. 7D,E,I,J), but Col2CreERT expression in perichondrial progenitors has largely ceased by E12.5 in forelimb long bone, because LacZ reporter activity is no longer detected (Fig. 7C,H). At later times of tamoxifen treatment, only perichondrium lining the articular joint surfaces subsequently displayed LacZ reporter activity (e.g., in Fig. 6), whereas growth plate perichondrium and periosteum were negative. Hence, tamoxifen injection during an early time window when chondrogenic and perichondrial precursors just begin to segregate and differentiate from condensed mesenchyme (between ∼E10 and E13 in limb elements, Table 1) results in recombination in progenitors later contributing to the osteogenic lineage.

thumbnail image

Figure 7. Tamoxifen injections at early condensation stages, when perichondrial and chondrogenic progenitors first segregate, result in Col2CreERT recombinase activity in both cartilage and in the osteoblast lineage descendant from perichondrium. BE: A single dose of tamoxifen (1.5 mg) was given intraperitoneally (IP) at the times indicated in each panel, and embryos were harvested at embryonic day (E) 15.5 and stained in whole-mount for LacZ activity (6–8 hr) to monitor recombination. A: A similar forelimb skeleton stained with Alcian blue and alizarin red to highlight areas of bone formation (red staining, arrowheads). Note that by E12.5 injection time, LacZ activity was subsequently no longer detectable in areas of bone formation (arrowheads in C). GJ: Representative cryosections from the humerus of the stained forelimbs in the adjacent left-side panels (B–E, respectively). F: A representative section (adjacent to G) photographed with Nomarski optics to highlight birefringent areas of bone matrix. Recombination, detected by LacZ activity, was observed in cortical and trabecular osteoblasts after injection at E10.5 and E11.5 (I–J), but osteoblast staining was absent by E12.5 (H). B,G: A Cre-negative control.

Download figure to PowerPoint

These results are consistent with the loss of Col2a1 expression in mature, differentiated perichondrium (e.g., by E13.5–E14 throughout limb; see also Fig. 8 in Colnot et al.,2005). Col2CreERT expression is also shut-off in mature perichondrium, except in the articular surfaces where expression detected by Rosa26LacZ reporter activation persists even postnatally (Fig. 6). Expression of transgenic Col2a1 promoters in early perichondrial progenitors has been observed variably in the past (see Long et al.,2001,2004; Ovchinnikov et al.,2000), possibly correlating with high transgene expression levels. The Col2CreERT line characterized here can also be useful for marking cells and targeting recombination to perichondrium and descendant osteogenic lineage during a narrow time window. Conversely, recombination in the osteogenic lineage may also be avoided by adjusting the timing of tamoxifen administration.

In conclusion, the Col2CreERT line described here gives efficient recombinase activity after a single, moderate dose of IP tamoxifen (∼1.5 mg/25 g animal) that occurs over a fairly narrow temporal window of approximately 12 hr. This line should prove valuable in various molecular genetic manipulations (see Nagy,2000) of genes important in the chondrogenic lineage during embryonic development and during postnatal skeletal growth and also in its response to injury and other pathologic conditions.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

Transgenic Constructs and Generation, Screening, and Breeding of Transgenic Mice

The Col2a1 promoter-driven expression vector (p3013 from Y. Yamada; Tsumaki et al.,1999) contains a single unique Not1 site between the β-globin intron and polyA signal (see Fig. 1A) for insertion of coding sequences to be expressed, such as LacZ (pKN185 from Y. Yamada; Tsumaki et al.,1999). The CreERT coding region (Feil et al.,1996) was isolated as an EcoRI fragment and blunt-end ligated into the NotI cloning site. These Col2CreERT and Col2LacZ expression constructs were injected into FVB/N zygotes to generate transgenic lines, as described by Hogan et al. (1994). Embryos were transferred to foster mothers and recovered after birth for establishing lines. Transgenic founders were identified by DNA extraction from tails and Southern blot analysis using a CreER or LacZ coding probe. Subsequent genotyping of animals in an established line was accomplished using polymerase chain reaction primers from the Cre (5′-GAAAATGCTTCTGTCCGTTTGC-3′ and 5′-ATTGCTGTCACTTGGTCGTGGC-3′; 207 bp amplicon) or LacZ (5′-AATGGCTTTCGCTACCTGGAG-3′ and 5′-TGGTGTTTTGCTTCCGTCAGC-3′; 200-bp amplicon) coding region. The transgenic line selected as giving the strongest recombination with minimal Cre recombinase activity in the absence of tamoxifen, was subsequently bred and maintained to give homozygous animals without any affects on viability or fertility. The LacZ expressing vector was also used to generate several transgenic lines, which were characterized and showed consistent and comparable expression patterns to those previously reported for a Col2a1 promoter containing similar elements (Zhou et al.,1995). The Rosa26LacZ reporter line for detection of Cre activity (Soriano,1999) was obtained from The Jackson Lab repository and genotyped as described. The Col2Cre line used was provided by Richard Behringer and has been described previously (Ovchinnikov et al.,2000). Cre activity was evaluated by crossing Cre transgenic mice with the RosaLacZ reporter line and analyzing LacZ reporter activation in progeny positive for both alleles. Control progeny were harvested either after injection with vehicle alone, or without IP injection for analysis of LacZ activity levels in the absence of tamoxifen.

Tamoxifen Injections and Embryo Analyses

For determination of embryonic ages, noon on the day of the postcoital plug was taken to be E0.5. Tamoxifen (Sigma, T-5648) was administered by IP injection as previously described (Danielian et al.,1998) with a few modifications. To bring the drug more rapidly into solution, tamoxifen was first dissolved at a concentration of 100 mg/ml in 100% ethanol and then diluted to a final concentration of 10 mg/ml into sterile vegetable oil for IP injection. To counteract the mixed estrogen agonist effects of tamoxifen, which can result in late fetal abortions, progesterone (Sigma, P-3972) dissolved in sterile vegetable oil was added for IP injections into pregnant females (usually at half the total tamoxifen dose given, up to a maximum dose of 2 mg of progesterone for single injection). In preliminary experiments, we determined that 1.5 mg of tamoxifen IP was generally adequate to achieve rapid and efficient recombination (complete by 24 hr; for example, see Fig. 2). In our strains, pregnant mice at approximately 10–12 days postcoitus averaged 25 g in body weight. Where not otherwise stated in text and figures, this dose was used. For some later stage tamoxifen injections (E13.5–E16.5) a higher dose of 6 mg was used in experiments shown, but generally doses of 2–3 mg were also effective (data not shown). For some neonatal injections, two IP doses of 0.2 mg were given 24 hr apart. Usually, for early neonatal through PN day 7 pups, a single IP dose of 1 mg was given, and for older postnatal animals, a single IP dose of 2 mg was given to achieve efficient recombination.

For analysis of LacZ activity, embryos were dissected into PBS and fixed with 2% paraformaldehyde and 0.2% glutaraldehyde in PBST (PBS with 0.1% Tween-20) at 4°C for 20 min up to 2 hr, depending on age and size of specimen, then washed twice for 20 min in PBST at 4°C. For analysis of postnatal limbs, the skin was removed, followed by fixation for 2 hr. LacZ staining in whole-mount, or in paraffin sections was performed using X-Gal substrate as previously described (Hogan et al.,1994). Embryos at E13.5 and younger were stained with X-Gal in whole-mount for up to 18 hr. Later stage embryos were stained for 6–8 hr maximum, and postnatal skeletal elements for 2–4 hr maximum to avoid high levels of background staining in bony tissues due to endogenous galactosidase activity. For cryosectioning, stained embryos were postfixed in 4% paraformaldehyde/PBS overnight, then decalcified in 19% ethylenediaminetetraacetic acid (for E15.5 and older; from 0.5–3 days), followed by cryoprotection in 30% sucrose overnight, all at 4°C, before OCT embedding. For whole-mount in situ hybridization, the probe preparation, embryo collection, fixation, and hybridizations were all done according to standard procedures, as described previously (Hogan et al.,1994).

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES

We thank Yoshi Yamada for providing Col2a1 promoter–enhancer and Col2LacZ-expressing plasmids, Daniel Metzger and Pierre Chambon for CreERT coding construct, Bjorn Olsen for mouse Col2a1 riboprobe, Richard Behringer for Col2Cre mouse line, Lionel Feigenbaum (Director, LASP, SAIC/NCI-FCRDC) for pronuclear injections to generate transgenic lines, and Dave Levens for critical reading of the manuscript. This research was supported by the Center for Cancer Research NCI, NIH.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS AND DISCUSSION
  5. EXPERIMENTAL PROCEDURES
  6. Acknowledgements
  7. REFERENCES