Col1a1-Driven Transgenic Markers of Osteoblast Lineage Progression


  • S. Dacic,

    1. Department of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, Connecticut, USA
    Search for more papers by this author
  • I. Kalajzic,

    1. Department of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, Connecticut, USA
    Search for more papers by this author
  • D. Visnjic,

    1. Department of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, Connecticut, USA
    Search for more papers by this author
  • A. C. Lichtler,

    1. Department of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, Connecticut, USA
    Search for more papers by this author
  • D. W. Rowe

    Corresponding author
    1. Department of Genetics and Developmental Biology, University of Connecticut Health Center, Farmington, Connecticut, USA
    • Address reprint requests to: Dr. David Rowe, Department of Genetics and, Developmental Biology, MC 1231, (E-2013) University of Connecticut Health Center, 263 Farmington Avenue, Farmington, CT 06030, USA
    Search for more papers by this author


The modular organization of the type I collagen promoter allows creation of promoter-reporter constructs with preferential activity in different type I collagen-producing tissues that might be useful to mark cells at different stages of osteoblastic differentiation. Primary marrow stromal cell (MSC) and mouse calvarial osteoblast (mCOB) cultures were established from transgenic mice harboring different Col1a1 promoter fragments driving chloramphenicol acetyltransferase (CAT). In these models, Col1a1 messenger RNA (mRNA) and alkaline phosphatase (ALP) are the first markers of differentiation appearing soon after the colonies develop. Bone sialoprotein (BSP) is detected 2-3 days later, followed by osteocalcin (OC) expression and nodule mineralization. A 3.6 Col1a1 fragment (ColCAT3.6) initiated activity concomitant with ALP staining and type I collagen mRNA expression. In contrast, a 2.3 Col1a1 fragment (ColCAT2.3) became active coincident with BSP expression. The pattern of transgene expression assessed by immunostaining was distinctly different. ColCAT3.6 was expressed within and at the periphery of developing nodules whereas the ColCAT2.3 expression was restricted to the differentiated nodules. The feasibility of using green fluorescent protein (GFP) as a marker of osteoblast differentiation was evaluated in ROS17/2.8 cells. A 2.3-kilobase (kb) Col1a1 promoter driving GFP (pOB4Col2.3GLP) was stably transfected into the cell line and positive clones were selected. Subcultures lost and then regained GFP expression that was localized in small clusters of cells throughout the culture. This suggests that expression from the 2.3-kb Col1A1 fragment is determined by the state of differentiation of the ROS17/2.8 cells. Col1a1 transgenes should be useful in appreciating the heterogeneity of a primary or immortalized culture undergoing osteoblastic differentiation.


THE OSTEOBLAST (OB) is a highly differentiated extracellular matrix-producing cell that arises from an undifferentiated progenitor through a series of proliferative and maturational steps.(1) The progression of an osteoprogenitor cell to a fully developed OB has been characterized extensively in primary cultures derived from bone marrow and from neonatal mouse calvaria and in a large variety of immortalized cell lines representing various degrees of differentiation. Although it is widely appreciated that the level of differentiation is heterogeneous within the culture, most of the commonly used methods to characterize the differentiation process require analysis of the culture as a homogeneous population of cells.

It is becoming clear that many growth factors and cytokines, as well as certain diseases of bone, exert their effects on bone by perturbing the ordered progression of bone cell maturation in addition to their effects on the fully mature OB.(2, 3) Distinguishing these two aspects of bone biology will require the development of markers specific for cells along the lineage. Following the methods developed for the hematopoietic and immune cell lineage pathways, antibodies to cell surface antigens have been developed for the OB lineage.(4, 5) This approach has been successful particularly on cells directly isolated from bone marrow before the later stages of lineage in which the surrounding matrix plays an important role in differentiation.(6–9) However, at this point isolation of cells using cell surface markers may be less reliable because of the proteolytic steps necessary to free the cells from matrix. Not only does this step potentially destroy surface antigens but the cell may fail to maintain its attained stage of differentiation once removed from its matrix.(10) One approach to overcome these problems is the use of intracellular transgenic markers under the control of a stage-specific promoter.

The Col1a1 promoter appears to have a modular design in which specific domains are used in different type I collagen-producing tissues.(11–13) In regard to OBs, a homeodomain binding TAAT sequence located between −1670 base pair (bp) and −1683 bp is required for the transgene to be active in the OB layer lining newly formed calvarial bone.(14) Calvaria derived from transgenic mice carrying the full-length promoter (ColCAT3.6) express the transgene in organ cultures and primary OB cell cultures. In contrast, when the promoter is deleted to −2.3 kilobases (kb; ColCAT2.3), activity is maintained in calvarial organ cultures but is lost when the cells derived from calvaria are placed in culture.(15) This result suggests that without the sequences upstream of −2.3 kb, the promoter is only functional in fully differentiated OBs, which are present in calvarial organ culture. Differentiation of the OBs is lost when the cells are removed from their matrix. A subsequent study showed that culture of the mouse calvarial OB (mCOB) containing ColCAT2.3 induced chloramphenicol acetyltransferase (CAT) activity under conditions that promote OB differentiation.(14) However, the temporal pattern of ColCAT2.3 activity was not compared with the time course of ColCAT3.6 activity or the induction of endogenous markers of bone differentiation. In this study, we examined the expression of ColCAT3.6 and ColCAT2.3 and endogenous OB markers in mCOB and marrow stromal cell (MSC) cultures undergoing OB differentiation. In addition, expression of green fluorescent protein (GFP; green lantern protein [GLP]) under control of the Col2.3 promoter was studied in ROS 17/2.8 rat osteosarcoma cells. The results indicate that Col1a1 promoter-reporter gene constructs can be used as markers for OB lineage progression.


Construction and maps

The details of these constructs have been described previously.(16) ColCAT3.6, the parental construct, contains 3.6 kb of the 5′ flanking sequence and 115 bp of the rat Col1a1 first exon, terminating just before the translation initation AUG codon (Fig. 1). This fragment is ligated to the CAT gene in the context of SV2 CAT that provides a 3′ splicing sequence terminated by the SV40 polyadenylation signal. ColCAT2.3 is an HindIII deletion of the 5′ flanking sequence to −2.3 kb. ColCAT3.6 and ColCAT2.3 constructs have been used to generate transgenic mice. One line of ColCAT3.6 (line2) and two lines of ColCAT2.3, (line 8, high expression) and (line 5, low expression) were used for these studies.(12)

Figure FIG. 1..

Constructs used in this study. (A) ColCAT3.6 contains 3.6 kb of 5′ flanking sequence and 114 bp of the Col1A1 first exon. This fragment is ligated to the CAT gene in the context of the SV2 CAT and a SV40 polyadenylation sequence. (B) ColCAT2.3 is a HindIII deletion of the 5′ flanking sequence of the parental Col3.6 construct. (C) The pOB4COLCAT2.3 contains Col2.3 promoter fragment, SV40 splice unit, and triple stop sequence for each reading frame before the CAT gene intact. GLP replaced CAT leaving SV40 polyA to terminate the transcript.

The pOB4CAT construct(17) was modified to include GFP. It contains the SV40 early promoter, a single SV40 splice unit, triple stop sequence for each reading frame before the CAT gene, and an SV40 polyA signal. The Xba site at the start of the CAT sequence was converted to BglII and the Xho site upstream of the SV40 splice donor site was converted to Xba. A modified GFP, GLP gene, was removed from pGreen Lantern-1 (Life Technologies, Grand Island, NY, USA) as an NotI fragment and recloned into the NotI site of an SacII to BglII modified pBS SK+. After the GLP was oriented, the Xba site of the polylinker was converted to Xho. GLP was removed as a BglII/Xho fragment and inserted into the modified pOB4CAT at the BglII/Xho sites to make pOB4GLP. The collagen promoter fragment was removed from ColCAT2.3 construct as an HindIII/Xba fragment and placed into pOB4GLP to make pOB4Col2.3GLP.

OB cultures: Preparation of MSC culture

Young adult (2-3 months old) transgenic mice harboring the ColCAT2.3 and ColCAT3.6 transgenes were killed by CO2 asphyxiation. Femurs, tibias, and humeri were dissected free of the attached muscle, the epiphyseal growth plate was removed, and the marrow was collected by flushing with α-modified minimum essential medium (α-MEM) culture medium containing phenol red, 100 U/ml Penicillin, 100 μg/ml streptomycin, and 10% fetal calf serum (FCS; Gibco BRL, Grand Island, NY, USA) through the marrow space with a 25G needle. The cells were dispersed by 4-6 cycles of aspiration and expulsion through an 18G needle followed by filtration through a 70-μm cell strainer (Falcon 2350; Fisher Scientific, Pittsburgh, PA, USA). Approximately 2 × 106 cells/cm2 (4 wells per mouse) were plated in 35-mm culture plates (Falcon 3046 Fisher Scientific). On day 4, half of the medium containing nonadherent cells was exchanged with fresh medium. Medium was changed completely on day 7 to α-MEM supplemented with 50 μg/ml ascorbic acid, 10−8 M dexamethasone, and 8 mM β-glycerophosphate (Sigma Chemical Co., St. Louis, MO, USA). Ascorbic acid was added daily and the medium was exchanged every 2 days for the duration of the experiment. Sister plates were removed at various time points for analysis.

Preparation of mCOB cell culture

Isolation of calvarial cells from 7- to 9-day-old ColCAT2.3 and ColCAT3.6 transgenic mice was performed by a modification of the method of Wong and Cohn.(18) Briefly, after removal of sutures, calvariae were subjected to four sequential 15-minute digestions in an enzyme mixture containing 0.05% trypsin (Gibco BRL) and 0.1% collagenase P (Boehringer Mannheim, Mannheim, Germany) at 37°C on a rocking platform. Cell fractions 2-4 were collected, chilled by the addition of an equal volume of cold Dulbecco's modified Eagle medium (DMEM) containing 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin (Gibco BRL). The fractions were pooled, centrifuged, resuspended in DMEM containing 10% FCS, and filtered through a 70-μm cell strainer. An aliquot of cells was diluted 1:1 with 0.04% trypan blue in phosphate-buffered saline (PBS) and viable cells were counted. Cells were plated at a density of 1-1.5 × 104 cells/cm2 in 35-mm culture plates in DMEM containing 10% FCS. Twenty-four hours later medium was exchanged and 3 days later cultures were fed again. At 1 week of culture, the medium was changed to differentiation medium (α-MEM containing 10% FCS, 25 μg/ml ascorbic acid, and 4 mM β-glycerophosphate). Thereafter, the medium was changed every 2 days and sister plates were removed for analysis.

Transfection of ROS 17/2.8 cells

ROS 17/2.8 cells were cultured in F-12 medium containing 10% FCS, 100 U/m penicillin, and 100 μg/ml streptomycin in 100-mm plates. Stable transfection was done by the calcium phosphate-DNA precipitation method using 10 μg of pOBCol2.3GLP and 1 μg of pSV2Neo as described previously.(16) Cells were placed under selection with 400 μg/ml of G418. Clonal lines were derived from the initial plate by directly scraping and expanding GLP active colonies. Stably transfected cells were examined 3 weeks after transfection under a confocal fluorescent microscope (Zeiss LSM 410; Carl Zeiss, Inc., Thornwood, NY, USA) with an argon-krypton laser at excitation of 488 nm and a 515-nm long-pass emission filter. Before microscopic observation, the media was removed and replaced with PBS to reduce background fluorescence.

In situ analysis of intact cultures

Histochemical staining for alkaline phosphatase (ALP) activity was performed using a commercially available kit (86-R ALP; Sigma Diagnostics, Inc., St. Louis, MO, USA) according to the manufacturer's instructions. Mineralization was assessed using silver nitrate staining according to the von Kossa method. The results of both staining procedures were recorded with a Kodak DCS420 color digital camera (Eastman Kodak Co, Rochester, NY, USA). Photomicrographs were taken with an SPOT digital camera (Diagnostic Instruments, Inc., Sterling Heights, MI, USA), in a bright field mode.

To prepare the cultures for CAT immunostaining, medium was removed and the cultured cells were washed in PBS. The cells were incubated at 37°C with 2 ml of 0.2% collagenase A (Boehringer Mannheim) and 0.2% hyaluronidase (Sigma Chemical Co.) in PBS for 2-10 minutes. After removing the enzymes by washing in PBS, the cells were fixed in 2% paraformaldehyde and 2% sucrose in 0.1 M sodium cacodylate, pH 7.4, for 45 minutes at room temperature (RT). Cells were treated with 0.1% NP-40 for 20 minutes on ice and then with 3% normal goat serum for 15 minutes at RT. Cells were washed three times with PBS after each subsequent step. Rabbit polyclonal antibody to CAT (5Prime-3Prime, Boulder, CO, USA) was diluted to 1:50 in PBA (0.1% bovine serum albumin [BSA] and 0.1% sodium azide in PBS) to a final volume of 700 μl and placed in the cultures for 1 h at RT. Subsequently, 700 μl of fluorescein isothiocyanate (FITC)-conjugated polyclonal goat anti-rabbit immunoglobulin G (IgG; 5Prime-3Prime) at a dilution of 1:300 in PBA was added for 30 minutes at RT in the dark. Controls consisted of cells exposed to nonspecific rabbit IgG alone followed by the secondary antibody and cells exposed to the secondary antibody alone. The cultures were analyzed under a Zeiss Axiovert 100TV fluorescent microscope equipped with an FITC fluorescent filter set and images were recorded with a PXL-EEV37 high-speed digital cooled CCD camera (Photometrics, Tucson, AZ, USA).

A more global impression of the ColCAT expression within the MSC cultures previously labeled with immunofluorescent stain or GLP expression from the ROS17/2.8 cells was obtained with the fluorImager SI system using ImageQuant software (Molecular Dynamics, Sunnyvale, CA, USA). Fluorochromes that are excitable at 488-nm light and that have most of their emission spectra at wavelengths longer than 515 nm are suitable for use. This system has a built-in 515-nm long-pass filter and in our analysis we used an additional bandpass filter 530DF30 to decrease background and improve the sensitivity and dynamic range of the assay. The plates were imaged at a PMT setting of 800 at 16-bit digital resolution.

Analysis of culture extracts

Extracts for analysis of CAT enzymatic activity were obtained from cultured MSC on days 7, 9, 11, 13, 15, 18, and 21 after plating and from calvarial cells on days 7, 14, and 21 after switching to differentiating media. Cultured cells were washed in PBS and scraped in 1 ml of scraping buffer (40 mM Tris, pH 7.8, 150 mM NaCl, and 1 mM EDTA). The suspension was subjected to three freeze-thaw cycles, heated at 65°C for 15 minutes to inactivate endogenous deacetylases, and then centrifuged for 3 minutes to remove precipitated protein. The protein concentration of the soluble supernatant was determined by the bicinchoninic acid (BCA) protein assay using the BCA Protein Assay Kit (Pierce, Rockford, IL, USA) according to the manufacturer's instructions. CAT activity was assayed using a modified fluor diffusion assay.(19) Activity was determined from the linear range of the assay and was expressed as counts per minute (cpm) per hour per microgram of protein.

Total RNA was isolated from the OB cultures using Trizol Reagent (Gibco BRL) according to manufacturer's instructions. In addition, the isopropanol precipitate was redissolved in 300 μl of GTC buffer (4.5 M guanidinium isothiocyanate, 10 mM β-mercaptoethanol, 15 mM sodium N-lauryl sarcosine, and 10 mM Na citrate, pH 7.0) followed by 300 μl isopropanol. The precipitate was washed once with 80% ethanol, drained, and redissolved in 50 μl of diethylpyrocarbonate (DEPC) water. RNA was separated on a 2.2 M formaldehyde 1% agarose gel and transferred onto nylon membrane (Maximum Strength Nytran; Schleicher & Schuell, Keene, NH, USA). The membranes were probed with complementary DNA (cDNA) fragments for rat Col1a1 (pα1R2),(20) mouse osteocalcin (OC) fragment,(21) mouse bone sialoprotein (BSP),(22) and mouse osteopontin (OP).(23) Probes were radiolabeled using the random primer method using (α32P) deoxycytosine triphosphate (dCTP) (3000 Ci/mM; New England Nuclear, Boston, MA, USA) obtaining approximately 1 × 109 cpm/μg. Filters were hybridized 3 × 106 cpm/ml at 42°C in 50% formamide and 5× SSPE (1× SSPE = 0.149 M NaCl, 10 mM NaH2PO4, and 1 mM EDTA, pH 7.4, 1.2× Denhardt's, and 0.5% sodium dodecyl sulfate [SDS]).(24) Filters were washed once in 6× SSPE and 0.5% SDS for 10 minutes at room temperature, once in 0.1× SSPE and 0.1% SDS for 10 minutes at 37°C, and once in 0.1× SSPE and 0.1% SDS for 10 minutes at 65°C. The membranes were sealed in plastic wrap and then exposed at −70°C to Kodak XAR film (Kodak, Rochester, NY, USA) using intensifying screens. Prior hybridization signals were removed by washing in 0.1× SSPE and 0.1% SDS for 20-30 minutes at 80°C.


Temporal expression of collagen transgenes in MSC and mCOB

Sister plates of pooled MSC from transgenic mice harboring the 2.3 and 3.6 ColCAT transgene were examined at eight time points (days 7, 9, 11, 13, 15, 18, 21, and 23) to place the onset of transgene activity in temporal context of visual markers and endogenous messenger RNA (mRNA) markers characteristic of bone cell differentiation. Under our culture conditions, the initial cell monolayer of macrophages and MSCs reach visual confluence between days 7 and 9. At this stage, spindle-shaped cells begin to form ALP-positive groups or colonies (Figs. 2A and 2B). By day 14, colonies progress to three-dimensional structures (nodules) containing cuboidal OB cells in the central zones and more spindle-shaped fibroblast cells in the peripheral zone (Fig. 2C). By day 23, mineralization is present in the central part of the nodules (Fig. 2 days) that contain the cuboidal OB cells. Sister plates were removed during the culture for RNA extraction. Northern analysis (Fig. 2E) showed that Col1a1 expression initiates at day 7 concomitant with ALP staining, while a BSP signal is observed by day 11 and OC is evident by day 15. OP has a characteristic pattern of expression that is strong at day 7, gradually falls during the day-13 to -15 interval, and increases again in parallel with OC expression.

Figure FIG. 2..

Progression of OB differentiation in MSC cultures. (A) ALP expression assessed by histochemical staining (days 7-21). Mineralization of the central nodular portions is detected by von Kossa method (day 23). (B) Photomicrograph shows an ALP- positive fibroblast colony in 9-day-old MSC culture (indicated by arrows). (C) Representative image of differentiating MSC colony on days 15-18. Cuboidal-shaped cells (C) are indicated by small arrows, while surrounding fibroblast-shaped cells (F) are shown by larger arrow. (D) Mineralization (M) of central region of a nodule becomes evident between day 21 and 23 (indicated by arrow). Small arrows show the cuboidal-shaped cell (C) located within a nodule. (E) Northern blot analysis for OB-related gene expression. Each lane contains 15 μg of total RNA extracted at different time points from day 7 to 21. Differentiation of MSCs can be observed on subsequent expression of Col1a1, BSP, and ultimately OC. OP showed two expression peaks, one early in the culture and a later one that is associated with full OB differentiation. The membrane was normalized by hybridization for 18S rRNA expression.

A similar analysis of the pattern of OB differentiation from day 7 to 25 was performed in mCOB cultures (Fig. 3A). On day 7, when visual confluence is attained, a few ALP-positive colonies containing polygonal-shaped cells are seen (P, Fig. 3B, see arrow). The colony number and intensity of ALP staining increases after addition of the differentiating medium (Fig. 3C) and mineralization becomes evident by days 15-18 (Fig. 3A, 3 days). The pattern of the RNA markers of differentiation progresses somewhat more rapidly than the MSC culture with Col1a1 appearing on day 7, followed by BSP between days 7 and 11 and OC by day 18. OP expression remains constant throughout the culture period (Fig. 3E).

Figure FIG. 3..

Progression of OB differentiation in mCOB cultures. (A) ALP expression assessed by histochemical staining (days 7-21) and mineralization by von Kossa method (day 25). (B) Photomicrograph of 7-day-old culture shows ALP-positive cells that have a polygonal shape (P). (C) Representative image of cuboidal (C) and polygonal (P) cells within a 15-day-old culture. (D) Mineral deposition (M) within a nodule of an 18-day-old culture. Clusters of cuboidal cells (C) are still evident within the nodule. (E) Northern blot analysis for OB-related gene expression. Each lane contains 15 μg of total RNA extracted at different time points from day 7 to 25. The membrane was normalized by hybridization for 18S rRNA expression.

Temporal pattern of CAT expression of MSC primary cultures is shown in Fig. 4A. During the proliferative phase of growth (days 7-11), ColCAT3.6 construct becomes active and increases further as the cultures acquire RNA markers of OB differentiation. The ColCAT2.3 construct is undetectable but becomes active when bone nodule formation begins (day 13) at the time that the ColCAT3.6 activity shows its additional increase. This pattern of delayed temporal expression was observed in both the low-expressing and the high-expressing (not shown) lines of ColCAT2.3.

Figure FIG. 4..

ColCAT3.6 and ColCAT2.3 (line 5) transgene expression in primary MSC and mCOB cultures. (A) CAT enzymatic activity in primary MSC culture. Cells were harvesting at different time points and CAT activity of extracts was measured. Data represents the mean and SEM from one of the three representative experiments. The left ordinate shows activity for ColCAT3.6 while the right ordinate is for ColCAT2.3 activity. (B) Northern blot analysis for CAT transgene expression in MSC. RNA was isolated from ColCAT3.6 and ColCAT2.3 MSC cultures at different time points and hybridized with CAT-specific probe. The blot was reprobed with 18S rRNA probe as a control. (C) CAT enzymatic activity in primary mCOB derived from ColCAT2.3 and ColCAT3.6 transgenic mice. Data represent the mean and SEM from one of the three representative experiments.

Sister MSC cultures were harvested for total RNA throughout the culture period. CAT mRNA assessed by Northern analysis paralleled the level of CAT activity extracted from the cells (Fig. 4B). Transgene mRNA in cultures from ColCAT3.6 mice is present at each time point of the culture. In contrast, transgene mRNA from ColCAT2.3 mice is not present at days 7, 9, and 11 but appears at day 13 and increases progressively at later time points.

In neonatal mCOB cultures, ColCAT3.6 activity is very strong on day 7 (Fig. 4C). After the culture is placed in OB inductive medium, a gradual decline in transgene activity is observed unlike the continued rise in activity seen with the MSC cultures. In contrast, the ColCAT2.3 construct shows only minimal activity on day 7 of culture that continued to increase through the culture.

CAT immunostaining

Measurement of CAT enzyme activity suggested that the full-length promoter (ColCAT3.6) initiates activity in undifferentiated type I-collagen producing cells and further augments activity in differentiated OB while the truncated collagen promoter (ColCAT2.3) is active only in differentiated OBs. However, total CAT activity does not adequately show the heterogeneity within the culture system, that is, areas with and without nodule formation or nodules at different stages of development. Therefore, immunofluorescence was performed on MSC using anti-CAT antibodies to visualize transgene expression (Figs. 5A and 5B). During the proliferation and early confluent phase, the ColCAT3.6 transgene showed cytoplasmic staining of cells distributed throughout the culture (Fig. 5B, day 7 and day 9), whereas no staining was present in ColCAT2.3 cultures (Fig. 5A). By day 14, cells within the bone nodules become ColCAT2.3-positive. ColCAT3.6-positive cells showed more intense staining in the nodule and maintained weaker staining in the cells between the nodules. Control cultures labeled with rabbit nonspecific IgG and secondary antibodies gave no fluorescent signal under the same conditions used for specific immunolabeling. Higher power images showed the uniform cuboidal nature of the CAT-positive cells from the ColCAT2.3 construct (Fig. 5C, left panel). In contrast, at least two populations of cells were identified with the ColCAT3.6 construct (Fig. 5C, right panel), one resembling those stained by ColCAT2.3 (Fig. 5C) and the other being a large and elongated spindle-like cell located at the periphery of the nodule.

Figure FIG. 5..

Immunofluorescence detection of CAT transgene. MSC culture were immunostained using rabbit anti-CAT antibodies (primary antibody) and FITC-conjugated goat anti-rabbit secondary antibody. (A) The CAT transgene in the ColCAT2.3 cultures (derived from line 8) is localized only in differentiated bone nodules (N, see arrows, ×5 magnification). (B) ColCAT3.6 transgene is detected at all time points, within the nodule (N) and in surrounding internodular (I) single cell layer (×5 magnification). (C) Higher-power examination of transgene expression. Densely packed ColCAT2.3-positive cuboidal (C)-shaped cells in the center of nodule are identified by the small arrows (left panel). In comparison, positive fibroblastic (F) cells of ColCAT3.6 are detected in the internodular regions (small arrow), along with the presence of both fibroblastic and cuboidal (C)-shaped cells within the nodule (large arrow; right panel). (D) To assess for the transgene distribution throughout the culture dish, the immunostained cultures were examined in the dynamic fluorImager SI system. Cells derived from ColCAT2.3 mice showed high expression of transgene within nodules (dark spots on image), while the ColCAT3.6 cultures showed a broader distribution of transgene throughout the culture.

For the better assessment of transgene distribution throughout the culture dish, we examined the immunostained plates with the fluorImager SI system that allows analysis of total transgene distribution per plate (Fig. 5D). Images obtained by this system showed high expression of transgene within nodules in cultures derived from ColCAT2.3 mice (dark spots on image). However, ColCAT3.6 cultures showed a broad patchlike distribution of transgene expression throughout the culture.

Supravital transgenes

Although the assay for the CAT marker gene has the advantages of simplicity, sensitivity, and low background, it has the disadvantage of not being able to be used to identify intact living cells. Tissues must be extracted for the enzymatic assay or fixed before immunostaining. This precludes the ability to isolate a living subpopulation of cells based on transgene expression.

GFP may be the ideal supravital marker gene for isolation of cells at specific stages of differentiation because of its low background and its autofluorescent properties. This marker gene was inserted into the pOB family of collagen promoter constructs,(25) which is 50- to 100-fold more active than the ColCAT construct design probably because of a more efficient splicing unit than the one in the SV2 design.(26)

ROS17/2.8 cells were stably cotransfected with pOB4Col2.3GLP and SV2Neo plasmids. Fluorescence was detected after 2 weeks of G418 selection. Different levels of fluorescence were observed from cells within the individual clone, with the strongest fluorescence coming from densely packed cells in the central region of the clone (Fig. 6A, upper panel, see arrows). ColCAT2.3-transfected control cells showed no fluorescence. From these positive clones, cell lines were isolated and expanded. Unlike the parental clones, the expanded cells developed as a mixed population of strongly positive, intermediately positive, and negative cells in which the positive cells were organized in clusters rather than uniformly dispersed throughout the culture (Fig. 6B). The distribution of the positive colonies is better appreciated by fluorImager analysis as small spots of varying intensity that are distributed randomly throughout the culture plate (Fig. 6C).

Figure FIG. 6..

Detection of Col2.3GLP autofluorescence in ROS17/2.8 cells. (A) Fluorescent image of ROS17/2.8 clones positive for pOB4Col2.3GLP. Cells were stable cotransfected with pOB4Col2.3GLP or ColCAT2.3 (negative control) and pSV2neo. Images (×10) taken under fluorescent light (upper panel) and transmitted light (lower panel) showed that the intensity of fluorescence varies among the cells within the single clone. The strongest fluorescent signal in a clone comes from the densely packed cells within the center of a colony (see arrows on both panels). (B) From these positive clones, cell lines were isolated and expanded. The resulting cultures developed as a mixed poplation of strongly positive, intermediately positive, and negative cells. (C) Detection by the Molecular Dynamics fluorImager of clusters of pOB4Col2.3GLP-positive cells that developed in the secondary cultures from an original GFP-positive clone.


Primary mCOB and MSC cultures are two widely used nonimortalized in vitro models of OB differentiation. The mCOB culture is initiated from a calvarial digest that contains a mixed population of OB and precursor cells.(27, 28) Primary MSC culture, as described by Friedenstein(29, 30) and further developed by Owen,(31) initiates from a quiescent stem cell that has potential to differentiate into OBs, adipocytes, chondrocytes, or cells supportive of myeloerythropoiesis.(32, 33) Other cell culture models that can be used to study the osteoprogenitor pathway are derived from retinal perivascular cells,(34) aortic myoblasts,(35) periodontal fibroblasts,(36) dental pulp cells,(37) cementoblasts,(38) and primary long bone OBs.(39)

Irrespective of the culture model used, the identification of genes with unique expression in OBs, such as OC and BSP, have been invaluable tools for ensuring that in vitro culture models were reflecting cells that had attained OB differentiation.(28, 40) Other markers such as Col1a1, OP, ON, and ALP, although highly expressed in OBs, are not unique to a specific cell within the OB lineage nor even cells within the MSC lineage. As the field directs its attention to questions of OB lineage, it will be necessary to develop conveniently measured molecular markers that the majority of investigators accept as an agreed-on level of cellular differentiation. For example, BSP is probably the earliest expressed bone-specific marker, while OC is produced at a later stage in the pathway. However, methods to identify individual cells or nodules expressing these endogenous genes within mixed cultures are difficult for routine use. This is where mice carrying the OC or BSP promoter driving a marker transgene will become particularly useful in defining a specific stage of OB differentiation either in cell culture or in intact bone tissue.(41, 42)

The modular organization and inherent strength of the COL1a1 promoter may provide the specificity and activity necessary to drive useful marker genes to identify certain levels of differentiation in type I collagen-producing tissues. The 3.6-kb Col1a1 promoter fragment appears to be active in a wide variety of type I collagen-producing tissues that include pre-OB and OB. When the sequences between −3.6 and −2.3 kb are removed, the strong activity of the promoter continues in the OB cells while it is greatly reduced in other type I-producing tissues.(13) This property may make the −2.3-kb fragment responsive to the transition of a pre-OB (off) to an early OB (on) in cell culture systems. The 2.3-kb-containing transgene appears to activate just after BSP mRNA can be detected but distinctly before OC mRNA is found in both mCOB and MSC culture models. ColCAT3.6 appears earlier in both cultures at the time the Col1a1 mRNA is detected and the colonies begin to stain for ALP activity again in both culture systems. The expression increases further as OB differentiation proceeds in parallel with the onset of ColCAT2.3 activity.

Having a visible marker of OB differentiation may reveal subtleties of OB development that are not appreciated when the culture is examined as a uniform population of cells. Static measurements of colonies in sister culture dishes over time can identify the total colony number and the proportion of colonies that express ALP or the CAT transgene or which mineralize. This type of analysis may be useful in discriminating between growth factors that act at the stage of cellular commitment to OB differentiation (number of OB colonies) or that expand the population of cells within the nodule (size of OB colony formation).

Detection of lineage progression in real time will be more powerful and convenient than repetitive static measurements for detecting an abnormality in lineage progression. In addition, it may provide the ability to isolate a uniform subpopulation of cells within the lineage for more detailed molecular analysis. Modifications of GFP promise that it will be a valuable supravital transgenic marker of cell progression along the developmental pathway of OBs. We wanted to evaluate this possibility using ROS17/2.8 cells, an OB cell line that has features of mature OB differentiation. Using expression of the pOB4Col2.3GLP marker it appears that there is significant heterogeneity of cell differentiation through out the culture. Furthermore, this cell population appears to be capable of reversing their degree of differentiation. Thus, the clonal population of cells after expansion revealed the variability in the transgene expression suggesting the presence of cell populations at different stages of the lineage. It raises the question whether agents that regulate OB genes in ROS 17/2.8 cells act by affecting the number of cells that attain an OB stage of differentiation (as defined by the transgene expression) or act only on cells that have previously attained OB status. As mice expressing the GFP transgene become available,(43) this type of question needs to be asked in primary OB cells to assess if this is a phenomenon of an immortalized OB cell line or a fundamental mechanism for regulating bone formation in vivo.


This work was supported by R01 AR43457. I.K. holds a Michael Geisman Fellowship of the Osteogenesis Imperfecta Foundation and D.V. was a Fogarty International Fellow (TW05309).