SEARCH

SEARCH BY CITATION

Keywords:

  • Osteoblast;
  • Bone;
  • α1(I)-collagen;
  • osteocalcin;
  • Cre recombinase

Abstract

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

Cell- and time-specific gene inactivation should enhance our knowledge of bone biology. Implementation of this technique requires construction of transgenic mouse lines expressing Cre recombinase in osteoblasts, the bone forming cell. We tested several promoter fragments for their ability to drive efficient Cre expression in osteoblasts. In the first mouse transgenic line, the Cre gene was placed under the control of the 2.3-kb proximal fragment of the α1(I)-collagen promoter, which is expressed at high levels in osteoblasts throughout their differentiation. Transgenic mice expressing this transgene in bone were bred with the ROSA26 reporter (R26R) strain in which the ROSA26 locus is targeted with a conditional LacZ reporter cassette. In R26R mice, Cre expression and subsequent Cre-mediated recombination lead to expression of the LacZ reporter gene, an event that can be monitored by LacZ staining. LacZ staining was detected in virtually all osteoblasts of α1(I)-Cre;R26R mice indicating that homologous recombination occurred in these cells. No other cell type stained blue. In the second line studied, the 1.3-kb fragment of osteocalcin gene 2 (OG2) promoter, which is active in differentiated osteoblasts, was used to drive Cre expression. OG2-Cre mice expressed Cre specifically in bone. However, cross of OG2-Cre mice with R26R mice did not lead to any detectable LacZ staining in osteoblasts. Lastly, we tested a more active artificial promoter derived from the OG2 promoter. The artificial OG2-Cre transgene was expressed by reverse transcriptase-polymerase chain reaction in cartilage and bone samples. After cross of the artificial OG2-Cre mice with R26R mice, we detected a LacZ staining in articular chondrocytes but not in osteoblasts. Our data suggest that the only promoter able to drive Cre expression at a level sufficient to induce recombination in osteoblasts is the α1(I)-collagen promoter. © 2002 Wiley-Liss, Inc.


INTRODUCTION

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

Gene targeting technology is the method of choice to study gene function in vivo (Capecchi, 1989). This development has allowed the field of skeletal biology to make tremendous progress in the past 12 years (Karsenty, 1999). However, many questions related to the nature and the complexity of genetic pathways affecting cell differentiation or cell function remain to be elucidated. In some instances progress in deciphering these pathways has been hampered by the fact that the disruption of some genes leads to embryonic lethality before skeletal development is initiated. In other instances, deletion of a gene of interest leads to perinatal lethality, therefore, precluding the study of the function of this gene in adult animals.

These two difficulties could be alleviated by improvements in gene targeting technology allowing cell- and time-specific gene deletion (Rossant and McMahon, 1999). The most frequently used technique to perform this conditional gene deletion experiment is based on the use of the Cre-Lox system (Rajewsky et al., 1996; Le and Sauer, 2000). The bacteriophage P1 Cre recombinase efficiently excises DNA flanked by two directly repeated LoxP recognition sites when expressed in cultured mammalian cells and in mice (Sternberg and Hamilton, 1981). Thus a prerequisite for cell-specific gene deletion experiments is to generate mouse strains expressing Cre recombinase at high levels only in the cell type of interest (Rossant and McMahon, 1999). Here, we compare the ability of several osteoblast-specific promoter fragments to induce osteoblast-specific gene deletion. This analysis indicates that, at the present time, osteoblast-specific deletion can be only achieved during embryonic development by using the 2.3-kb fragment of the mouse α1(I)-collagen promoter. Promoter fragments acting later during osteoblast differentiation were inefficient in driving Cre expression in osteoblasts.

RESULTS AND DISCUSSION

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

Comparison of Promoter Activities in Cell Culture

Three different promoters were tested. The 2.3-kb α1(I)-collagen promoter fragment (Rossert et al., 1995), the 1.3-kb fragment of the osteocalcin gene 2 (OG2) promoter, and an artificial promoter containing six copies of two osteoblast-specific cis-acting elements, OSE2 and OSE1, present in the OG2 promoter (Ducy and Karsenty, 1995) (Fig. 1A). To compare their activities and to test their cell-specificity, each of these promoter fragments were first fused to the luciferase reporter gene and assayed in DNA transfection experiments. As shown in Figure 1B, the α1(I)-collagen promoter fragment was the most active promoter in ROS 17/2.8 osteoblastic cells followed by the artificial OG2-Luc that, as expected, was more active than the 1.3-kb OG2-Luc construct. The activity of these promoters was osteoblast-specific as they were not active in F9 cells or in C2C12 cells, two cell lines that have no feature of osteoblasts (Fig. 1B and data not shown).

thumbnail image

Figure 1. A: Schematic representation of the constructs used in DNA transfection experiments. The α1(I)-promoter cassette contains a 2.3-kb proximal fragment of the α1(I)-collagen promoter. The OG2 promoter cassette contains a 1.3-kb fragment of the OG2 promoter. The artificial OG2-Cre promoter cassette contains 6 repeats of an OSE2 element, 6 OSE1 binding elements, and the α1(I)-collagen TATA box. All promoter cassettes were fused to a luciferase reporter gene and a SV40 polyadenylation signal. B: ROS17/2.8 osteoblastic cells or F9 cells were transfected with either α1(I)-collagen promoter-Luc, artificial OG2-Luc, or 1.3- kb of proximal OG2 promoter-Luc construct. Data are expressed as means of luciferase activity/β Gal ratios. Errors bars indicate standard errors of the mean.

Download figure to PowerPoint

Generation and Analysis of Transgenic Mice

Each of these promoter fragments was then fused to the Cre recombinase cDNA, and these constructs were used to generate transgenic mouse lines (Fig. 2A). Transgene integration in the mouse genome was assessed by genomic PCR (Fig. 2B). We analyzed four α1(I)-Cre mouse lines, two OG2-Cre mouse lines, and one artificial OG2-Cre mouse line. For α1(I)-Cre and OG2-Cre mice, similar results were obtained in all transgenic mouse lines at any given developmental stage studied. Reverse transcriptase-polymerase chain reaction (RT-PCR) was used to determine Cre expression for every mouse transgenic line. As shown in Figure 2C, the 2.3-kb α1(I)-collagen promoter and the 1.3-kb OG2-Cre promoter directed Cre expression specifically in bone. These two transgenes were not expressed in any other tissue tested. On the other hand, Cre was expressed in bone and cartilage in artificial OG2-Cre transgenic mice (Fig. 2C).

thumbnail image

Figure 2. A: Structure of the α1(I)-Cre, OG2-Cre, and artificial OG2-Cre transgenes. The Cre recombinase gene was placed under the control of the 2.3-kb α1(I)-collagen promoter, the 1.3-kb OG2 promoter, or the artificial OG2 promoter. All the constructs included an MT-1 polyadenylation signal. B: Transgenic animals were genotyped by polymerase chain reaction (PCR) using Cre-specific primers. C: Tissue-specific expression of α1(I)-Cre, OG2-Cre, and artificial OG2-Cre transgenes was demonstrated by reverse transcriptase-PCR followed by southern hybridization by using Cre-specific primers and a Cre-specific probe.

Download figure to PowerPoint

In Vivo Recombination

To define the ability of Cre to induce recombination specifically in osteoblasts, offspring of each Cre expressing transgenic line were bred with a previously described targeted gene-trap strain, the ROSA26 reporter strain (R26R) (Soriano, 1999). In these mice, a DNA fragment containing four polyadenylation signals flanked by two LoxP sites was introduced by gene targeting downstream of the LacZ reporter inserted in the ROSA26 locus. Upon Cre expression and recombination, the removal by Cre of the DNA fragment flanked by LoxP sites and containing the four polyadenylation signals allows LacZ expression. This event can be monitored by LacZ staining.

LacZ staining of α1(I)-Cre;R26R mice was performed in 14.5 days post coitum (dpc) and 16.5-dpc embryos, and in 5-day-old transgenic mice. In 14.5-dpc embryos, LacZ staining was observed in the skull and in all long bone ossification centers (Fig. 3A). Except for a weak staining located in the skin of the digits and the face, no other tissue stained in blue (Fig. 3A). In 16.5-dpc embryos and 5-day-old mice, LacZ staining could be detected in all bones of the α1(I)-Cre;R26R mice whether they develop through endochondral or intramembranous ossification (Fig. 3B–E). No staining was detected in any other tissue. To determine whether Cre was expressed in osteoblasts at a level allowing recombination, we performed histologic analysis of the bone and counterstained them with eosin and hematoxylin. Sections of long bones revealed LacZ staining in osteoblasts but not in chondrocytes (Fig. 3F,G). On these sections, all identifiable osteoblasts stained blue. No staining was detectable in R26R control mice (Fig. 3H). Taken together, these data indicate that Cre was expressed in osteoblasts in the α1(I)-Cre transgenic mice at a level allowing recombination at the ROSA26 locus.

thumbnail image

Figure 3. Osteoblast-specific recombination determined by LacZ staining of α1(I)-Cre;R26R mice at 14.5 days post coitum (dpc) (A), at 16.5 dpc (B), and in 5-day-old pups (C). LacZ staining of bones of the skull (D), rib, and vertebrae (E). Longitudinal sections of long bone showing that the LacZ staining in 5-day-old mice is restricted to cuboid cells attached to bone trabeculae, i.e., osteoblasts (F,G). No LacZ staining is visible in osteoblasts of R26R control as seen here on longitudinal section of long bone (H).

Download figure to PowerPoint

We performed similar crosses by using OG2-Cre mice or artificial OG2-Cre mice to determine whether these promoter fragments could be used to study gene function in differentiated osteoblasts. In the OG2-Cre;R26R mice, no LacZ staining was visible in any tissue (Fig. 4A). To confirm this observation at a cellular level, we analyzed long bone sections. Consistent with this observation, no osteoblasts were stained for β-galactosidase activity on these sections (Fig. 4C). Therefore, we concluded that, despite the bone-specific expression of the transgene, the expression level of Cre in osteoblasts was probably below the threshold necessary to cause recombination. Surprisingly, in the artificial OG2-Cre;R26R mice, LacZ staining was detectable in cartilage (Fig. 4B). Chondrocytes of articular cartilage were lightly stained on histologic sections of 5-day-old mice long bones (Fig. 4D). In those mice, LacZ staining was also observed in bone blood vessels (data not shown). However, no staining was visible in osteoblasts (Fig. 4E). This result indicates that the expression observed by RT-PCR in bone could be due to the presence of blood vessels and of chondrocytes in the epiphysis of the bone samples we used.

thumbnail image

Figure 4. Absence of LacZ staining in OG2-Cre;R26R mice (A) and staining of cartilages in artificial OG2-Cre;R26R mice (B). On histologic sections of long bones, no staining is visible in osteoblasts (arrow) of OG2-Cre;R26R mice (C). Histologic analysis of bone of artificial OG2-Cre;R26R reveals a staining in articular cartilage (arrow) of the long bone (D) but not in osteoblasts (arrow in E).

Download figure to PowerPoint

Taken together, our results indicate that the α1(I)-collagen promoter is the only promoter among the ones tested that is able to drive expression of Cre specifically in osteoblasts at a level sufficient for homologous recombination to occur in vivo. A note of caution should be made as ROSA26 is an easy locus to target (Friedrich and Soriano, 1991). Therefore, although the fact that α1(I)-Cre allows homologous recombination at this locus is encouraging, it does not mean that it will be the case for all loci of interest. The failure of the artificial OG2 promoter or the native OG2 promoter to drive Cre expression at a level sufficient for efficient recombination limits for now our ability to perform efficient osteoblast-specific gene deletion postnatally. This approach will have to await the generation of additional artificial promoters to generate cell-specific gene deletion.

EXPERIMENTAL PROCEDURES

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

DNA Constructions

Construction of the 2.3-kb α1(I)-Luc and OG2-Luc (p1316-Luc) plasmids has been described previously (Ducy and Karsenty, 1995; Kern et al., 2001). α1(I)-Cre was constructed by inserting the Cre recombinase gene, including the polyA region of pCBM9 (Sauer and Henderson, 1990) into pJ2300LacZ (Rossert et al., 1995), a plasmid containing the 2.3-kb α1(I)-collagen promoter. To create the artificial OG2 promoter, six copies of an oligonucleotide containing the OSE2 element of the OG2 promoter (Ducy and Karsenty, 1995) were cloned in front of an oligonucleotide covering the region from −132 to −74 of the OG2 promoter. Six copies of an oligonucleotide containing the OSE1 sequence (Ducy and Karsenty, 1995) were cloned downstream of this cassette. The HindIII fragment of pK1-Luc containing the α1(I)-collagen TATA box (Karsenty and de Crombrugghe, 1990) was cloned downstream of these oligonucleotides. Artificial OG2-Luc and artificial OG2-Cre were generated by inserting the blunted end ClaI/SacI artificial OG2 promoter fragment from artificial OG2 into SmaI blunted of p4Luc, a promoterless luciferase expression vector and in pCBM9 Cre containing vector, respectively. The 1.3-kb KpnI/SalI fragment of the OG2-promoter from pIIBS 1.3 OG2 (Ducy and Karsenty, 1995) was inserted into pCBM9 to generate OG2-Cre.

Transfection Experiments

ROS17/2.8 osteoblasts were cultured in Dulbecco's modified Eagle's medium, F-12 and F9 teratocarcinoma cells in Eagle's minimal essential medium. Both media were supplemented with 10% fetal bovine serum. All reagents were purchased from Gibco. DNA transfection experiments were performed by using the calcium phosphate coprecipitation method with 5 μg of reporter plasmid constructs and 2 μg of pSV-βGal. β-Galactosidase assay results were used to normalize the luciferase assay results for transfection efficiency. All transfections were repeated at least three times in triplicate with different DNA preparations.

Generation and Analysis of Transgenic Mice

The constructs α1(I)-Cre, OG2-Cre, and artificial OG2-Cre were linearized, inserts were purified by two rounds of agarose gel electrophoresis and injected into the pronuclei of fertilized FVB mouse oocytes (Jackson Laboratories, Inc.), which were then implanted in the oviducts of pseudopregnant CD1 foster mothers (Charles River Laboratories) for development to term. Animals expressing the transgenes were identified by PCR on tail genomic DNA in the presence of 10% dimethyl sulfoxide. The Cre-specific primers used were Cre5′, 5′-CCTGGAAAATGCTTCTGTCCGTTTGCC-3′; Cre3′, 5′-GAGTTGATAGCTGGCTGGTGGCAGATG-3′. ROSA26 reporter mice were a generous gift of Philippe Soriano.

Tissue Distribution and Expression Level of the Cre Transgenes

Total RNA from different tissues were extracted with TRIAZOL reagent (Invitrogen, Inc.), according to the manufacturer's instructions. For RT-PCR analysis, cDNAs were synthesized from 1 μg of DnaseI-treated total RNA with 0.2 μg random primers by using the SuperscriptII RnaseH reverse transcriptase kit (Invitrogen, Inc.). Southern blot hybridizations were performed according to standard procedure by using a Cre-specific probe.

Xgal Staining and Histologic Analysis

After a 2-hr fixation in 2% paraformaldehyde, 0.2% glutaraldehyde, 5 mM EGTA, 0.1 mM MgCl2, and 0.1 M NaPO4 pH 7.3, Cre-transgenic;R26R embryos and pups were washed in a solution containing NP-40 detergent and stained for 3 hr with X-Gal (5-bromo-4-chloro-3-indoyl-D-galactosidase) 1 mg/ml as described (Frendo et al., 1998). Skin and internal organs were removed and stained in parallel for 16.5-dpc embryos and 5-day-old mice. Specimens were embedded in paraffin, sectioned at 10 μm, and counterstained with eosin-hematoxylin or eosin alone.

Acknowledgements

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

We thank Dr. Philippe Soriano for his generous gift of ROSA26 reporter mice, reagents, and advice, and Dr. Patricia Ducy for critical reading of the manuscript.

REFERENCES

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