Establishment of an Osteocyte-like Cell Line, MLO-Y4


  • Yoichi Kato,

    1. Department of Medicine, University of Texas Health Science Center at San Antonio, San Antonio, Texas, U.S.A.
    Search for more papers by this author
  • Jolene J. Windle,

    1. Cancer Therapy and Research Center, University of Texas Health Science Center at San Antonio, San Antonio. Texas, U.S.A.
    Search for more papers by this author
  • Barbara A. Koop,

    1. Cancer Therapy and Research Center, University of Texas Health Science Center at San Antonio, San Antonio. Texas, U.S.A.
    Search for more papers by this author
  • Gregory R. Mundy,

    1. Department of Medicine, University of Texas Health Science Center at San Antonio, San Antonio, Texas, U.S.A.
    Search for more papers by this author
  • Lynda F. Bonewald

    Corresponding author
    1. Department of Medicine, University of Texas Health Science Center at San Antonio, San Antonio, Texas, U.S.A.
    2. Departments of Biochemistry and Cellular and Structural Biology, University of Texas Health Science Center at San Antonio, San Antonio, Texas, U.S.A.
    • University of Texas Health Science Center Department of Medicine 7703 Floyd Curl Drive San Antonio, TX 78284-7877 USA
    Search for more papers by this author


Although osteocytes are the most abundant cells in bone, their functional role remains unclear. In part, this is due to lack of availability of osteocyte cell lines which can be studied in vitro. Since others have shown that cell lines can be readily developed from transgenic mice in which the SV40 large T-antigen oncogene is expressed under the control of a promoter which targets the cells of interest, we used this approach to develop an osteocyte cell line. We chose as a promoter osteocalcin, whose expression is essentially limited to bone cells and which is expressed more abundantly in osteocytes than in osteoblasts. From these transgenic mice, we isolated cells from the long bones using sequential collagenase digestion and maintained these cells on collagen-coated surfaces which are optimal for osteocyte maintenance and growth. We describe here the properties of a cell line cloned from these cultures, called MLO-Y4 (for murine long bone osteocyte Y4). The properties of MLO-Y4 cells are very similar to primary osteocytes. Like primary osteocytes and unlike primary osteoblasts, the cell line produces large amounts of osteocalcin but low amounts of alkaline-phosphatase. The cells produce extensive, complex dendritic processes and are positive for T-antigen, for osteopontin, for the neural antigen CD44, and for connexin 43, a protein found in gap junctions. This cell line also produces very small amounts of type I collagen mRNA compared with primary osteoblasts. MLO-Y4 cells lack detectable mRNA for osteoblast-specific factor 2, which appears to be a positive marker for osteoblasts but may be a negative marker for osteocytes. This newly established cell line should prove useful for studying the effects of mechanical stress on osteocyte function and for determining the means whereby osteocytes communicate with other bone cells such as osteoblasts and osteoclasts.


Osteocytes are the most abundant cells in bone, but the cells about which we know the least. There are approximately 25,000 osteocytes/mm2 in bone, or ten times as many osteocytes as osteoblasts.(1) Because they are buried in the mineralized matrix, they are relatively inaccessible and have been difficult to study in culture in homogeneous populations. As a consequence, we know little about their function.

Osteocytes are likely derived by terminal differentiation of osteoblasts which become trapped within the forming osteoid tissue during bone formation.(2) Thus, osteoblasts share many markers with osteocytes. However, there are differences in levels of expression of various markers between osteoblasts and osteocytes. For example, alkaline phosphatase (ALP) is expressed at lower levels in osteocytes than in osteoblasts, whereas osteocytes express greater amounts of casein kinase activity than osteoblasts(3) Another important marker that is expressed in greater amounts in osteocytes than in osteoblasts is osteocalcin. Osteocalcin may be important for the osteocyte to maintain an unmineralized area around the cell body and its processes(4) Osteocalcin may function to prevent the mineralization of the adjacent osteoid tissue to allow diffusion of nutrients and waste products to and from the cell. Mikuni-Takagahi and coworkers(3) have described the accumulation of large amounts of osteocalcin around isolated osteocytes. Mice that lack the functional gene for osteocalcin have increased cortical and trabecular bone, which also supports the hypothesis that osteocalcin is a negative regulator of mineralization.(5)

We postulated that since osteocytes are large producers of osteocalcin, bone cells derived from transgenic mice overexpressing T-antigen driven by the osteocalcin promoter would be a potentially abundant source of immortalized cells of the osteocyte phenotype. We chose to use cellular morphology as the initial criterion for cloning a cell line with osteocyte characteristics from cell isolates derived from these mice. Once the murine long bone osteocyte Y4 (MLO-Y4) clonal cells were established, they were characterized and their properties compared with the known properties of primary osteocytes, osteoblasts, and other cells.



Tissue culture media were purchased from GIBCO BRL (Grand island, NY, U.S.A.), fetal bovine serum (FBS) was from BioWhittaker (Walkersville, MD, U.S.A.), and calf serum (CS) was from HyClone Laboratories, Inc. (Logan, UT, U.S.A.). Rat tail collagen type 1, 99% pure, was purchased from Becton Dickinson Laboratories (Bedford, MA, U.S.A.). All other reagents were purchased from Sigma Chemical Co. (St. Louis, MO, U.S.A.), unless otherwise stated.

Establishment of transgenic mice

Construction of the osteocalcin promoter/T-antigen transgene and establishment of transgenic mice have been described previously.(6) In short, a 2.6 kb DNA fragment containing the rat osteocalcin gene promoter region from −2600 to +30 was ligated upstream of the SV40 early region which contains the protein-coding region of large T and small T antigens. DNA was micro injected into the pronuclei of fertilized one-cell mouse embryos. The F2 embryos were derived from matings of CB6F1 (C57B1/6 × BALB/c) males and females obtained from Harlan Sprague-Dawley, Inc. (Indianapolis, IN, U.S.A.). The injected embryos were reimplanted into B6D2F1 (C57BI/6 × DBA2) pseudopregnant females. The presence of the trans-gene in the resulting pups was determined by Southern blot analysis of genomic DNA. Founder transgenic mice were bred to nontransgenic CB6F1 mice to establish the lines of mice.

Cell isolation

Seven 14-day-old transgenic mice were used for isolation of cells from the long bones. Both ends of the tibiae and femurs including the growth plate cartilage were cut off, and the marrow flushed, leaving the bones, which were cut into several pieces. These bone pieces were washed three times with phosphate-buffered saline (PBS), and then the bone pieces were sequentially digested five times using 0.75 mg/ml collagenase in Hank's balanced salt solution for 30 minutes, each at 37°C. Cells were collected with each digestion (fractions 1–5). Fractions 3–5 were combined and used for cell cloning.

Cell culture and cloning

Fractions 3–5 were cultured on collagen-coated plates (rat tail collagen type 1, 0.15 mg/ml) in alpha modified essential medium (α-MEM) supplemented with 5% FBS and 5% CS. After several passages using these culture conditions, osteocyte-like cells with the dendritic phenotype were enriched. The MLO-Y4 cell line was cloned from this osteocyte-like cell enriched population by single colony isolation. Selection was based on expression of the dendritic phenotype.

Measurement of cellular proliferation

Cells, 2 × 103 were plated on collagen-coated or non-coated 48-well plates and cultured with α-MEM + 5% FBS/5% CS or α-MEM + 10% FBS. Media were changed every 3 days. Cell cultures were stopped after 1, 3, 4, 5, 7, 8, or 10 days, and the cells were harvested after trypsin-EDTA treatment. Cell number was measured using a Coulter Counter, model ZF (Coulter Electronics., Inc., Hialeah, FL, U.S.A.).

Determination of T-antigen expression by Western blot analysis

T-antigen protein expression in this cell line was determined by Western blot analysis according to a previously described Western blotting technique.(7) Subconfluent cells were washed twice with PBS and lysed by ice-cold RIPA buffer (50 mM Tris-HCl, pH 7.2, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate) for 5 minutes of incubation at 4°C. The lysates were collected and centrifugea at 14,000 rpm for 10 minutes and the supernatant collected. Aliquots of each sample were treated with 2X electrophoresis sample buffer containing reducing agent and applied to 7% sodium dodecyl sulfide-polyaerylamide gel electrophoresis (SDS-PAGE). Proteins were transferred onto a nitrocellulose membrane by electroblotting. The membranes were blocked with 5% BSA in TBS buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.2) overnight at 4°C. Mouse anti-SV40 T-antigen monoclonal antibody (Oncogene Science, Inc., Cambridge, MA, U.S.A.) diluted 1:50 in TBS buffer + 1% BSA for the primary antibody. Peroxidase-conjugated goat anti-mouse immunoglobulin G (IgG) Fc secondary antibodies were used at a 1:2000 dilution in TBS buffer + 2% skim milk. Bands were visualized using the chemiluminescence detection system as described by the manufacturer (DuPont NEN Research Products, Boston, MA, U.S.A.).

Figure FIG. 1.

Time course showing the morphological appearance of the MLO-Y4 cell line from 3 h to 7 days after seeding on collagen-coated plates. At 3 h the cells are small and stellate. Note the extention and increased branching of dendritic processes over time in culture. At 7 days, cell growth appears contact inhibited. (A) Three hours; 3 h (β) 6 h; (C) 24 h; (D) 58 h; (E) 4 days; (F) 7 days.

RNA preparation and reverse transcription-polymerase chain reaction analysis

Total cellular RNA was isolated from cultures of confluent cells using RNAzol β (Biotecx Laboratories, Inc., Houston, TX, U.S.A.) according to manufacturer's instructions. One 10-cm confluent culture plate was used each for primary osteoblasts, MC3T3-E1, and OCT-1, and 4 confluent 10 cm culture plates were used for MLO-Y4. cDNAs were synthesized from 3 μg of total RNA in a 20 μl reaction mixture containing 1X reverse transcriptase buffer (Pro-mega, Madison, WI, U.S.A.), 0.26 U/μl RNAse inhibitor (Promega). 500 μM dNTP mixture, 10 mM DTT, 50 pmol of oligo d(T)16 primer (Perkin Elmer, Norwalk, CT, U.S.A.), and 20 U of AMV reverse transcriptase (Promega). cDNA, 0.5–2%, was amplified using PCR in a 20 μl reaction mixture containing 1X PCR buffer (Fisher Scientific, Pittsburgh, PA, U.S.A.), 5 pmol of 5′ and 3′ primer, 200 μM dNTP mixture, 2 mM MgCL (Fisher Scientific), and 1 U of Taq DNA polymerase (Fisher Scientific). Amplifications were performed in a DNA Thermal Cycler 480 (Perkin Elmer Cetus, Emeryville, CA, U.S.A.) for 25–35 cycles following the reaction profile: 94°C for 1 minute, 58°C for 1.5 minutes, and 72°C for 2 minutes.

The following primers, annealing temperatures, and cycles were used for each particular cDNA amplification: mouse osteocalcin: 5′-GACAAAGCCTTCATGTCCA AGC, 3′-GTTTGAGACCGTCGAGCCGAAA; 58°C, 25 cycles; mouse osteopontin: 5′-GACCATGAGATTGGCA GTGATTTG, 3′-GTTTCGGTCGGACCTTGTAGT; 58°C. 25 cycles; mouse type I collagen: 5′-AATGGTGAGACGT GGAAACCCGAG, 3′-GGTTTGAGTCTTCTACATCCT CAGC; 58°C, 25 cycles; mouse CD44: 5′-CAAGTTTT GGTGGCACACAGC, 3′-GGTTAAGGAAGCTACCTG GC; 58°C, 30 cycles; mouse OSF-2: 5′-TGGAAGGGAT GAAAGGCTGC, 3′-CGGTGTTTACCA CAGCAGGT; 58°C, 30 cycles.

Figure FIG. 2.

Western blot analysis showing expression of T-antigen by MLO-Y4 cells grown in 10% FBS or 5% FBS/5% CS and OCT-1 cells but not MC 3T3-E1 cells.

Sequence analysis of PCR products

The bands of interest were excised from the gels and TA cloned into pGEM-T vector per suppliers instructions (Promega). Transformation followed by an insertion check was performed before DNA sequencing using a kit for dye terminator cycle sequencing (Perkin-Elmer) which was then read using an Applied Biosystems model 373A DNA sequencer.

Determination of connexin 43 expression by immunocytochemical staining and Western blotting

Subconfluent MLO-Y4 cells grown in 48-well plates were fixed using 3% paraformaldehyde and 2% sucrose in PBS, and permeabilized with 0.05% Triton X-100. The fixed plates were blocked with 5% BSA in TBS buffer for 2 h at room temperature. These cells were then incubated with a 1:125 dilution of anti-connexin 43 monoclonal antibody (mAb; Zymed Laboratories, Inc., San Francisco, CA. U.S.A.). The bound antibody was detected using the Vectastain ABC Kit, followed by staining with VIP substrate according to manufacturer's instructions (Vector Laboratories, Burlingame, CA, U.S.A.). The counter staining was performed using 0.5% methyl green, and the secondary antibody was used alone as a control for background staining. The procedure as listed above as used for T-antigen expression was used for anti-connexin 43 Western blot analysis. The specific antibody for murine connexin 43 was obtained from Zymed Laboratories and used at a 1:1000 dilution. Ten percent SDS-PAGE was used.

Figure FIG. 3.

Growth curve of the MLO-Y4 cell line under various culture conditions (10% FBS or 5% FBS/5% CS, with or without collagen coating of surfaces) compared with the osteoblastic cell line, OCT-1.

Figure FIG. 4.

(A) Photographs of staining for the ALP enzyme (right) and phase contrast (left) of MLO-Y4 cells. Original magnification is X50. Few cells are positive for ALP. Note the extensive cellular processes emphasized by phase contrast. (B) ALP-specific activity of MLO-Y4 as compared with primary osteoblasts (Prim. OBI) and the osteoblastic cell line; OCT-1. The cells were cultured for 15 days using 10% FBS containing culture conditions. The results are presented as the mean ± SD, n = 3. Note the continuous low expression of ALP in the MLO-Y4 cell line. Similar results were observed using 5% FBS/5% CS containing culture conditions (data not shown).

Staining for ALP

Cells, 1 × 106, were plated in a collagen-coated 10 cm culture dish. After 3 or 7 days of culture, the cells were fixed by 10% buffered formalin for 1.0 minutes. After two washes with PBS, freshly prepared 0.033% nitro blue tetrazolium (NBT) and 0.017% bromochloroindoyl phosphate (BCTP) in ALP buffer (100 mM NaCl, 5 mM MgCL, 100 mM Tris-HCl, pH 9.5) was added to the fixed cells and incubated for 15 minutes at 37°C. The reaction was stopped by washing with running water. Stained cells were observed under the microscope (X50, Olympus CK2, Olympus, Lake Success, NY, U.S.A.).

Figure FIG. 5.

Osteocalcin protein and mRNA expression by MLO-Y4 cells. (A) Osteocalcin in conditioned media of the MLO-Y4 clone compared with primary osteoblast cells and two osteoblast cell lines MC3T3-E1 and OCT-1 as determined by RIA. MLO-Y4 cells produce large amounts of osteocalcin whether cultured in 10% FBS or 5% FBS/5% CS compared with primary osteoblast cells, MC3T3-E1, and OCT-1, cultured in 10% FBS (n = 2). (B) Expression of osteocalcin mRNA in the MLO-Y4 cell line when compared with osteoblast cells as determined by RT-PCR. A band is present in the MLO-Y4 lane but not in the OCT-1, MC3T3-E1 or primary osteoblasts (Prim. OBI) lane.

Quantitation of ALP-specific activity

Cells, 4 × 104, were plated on collagen-coated 24-well plates and cultured with α-MEM + 5% FBS + 5% CS or α-MEM + 10% FBS. Media were changed every 3 days. Cell cultures were stopped after 3, 6, 9, 12, or 15 days, and washed twice with PBS. Cells were lysed. twice using freeze-thaw cycles with 200 μl of 0.05% Triton X-100. Ten microliters of lysate was used to determine the protein concentration using the micro Bradford method (Bio-Rad Laboratories), and 10 μl of lysate was used for alkaline phosphatase activity using 1.5 M 2-amino-2-methyl-l-propanol (AMP) buffer (pH 10.3) containing 5 mM P-nitrophenol. phosphate substrate according to a previously described method.(8)

Quantitation of osteocalcin by radioimmunoassay

Cells were plated on collagen coated 48-well plates. After 3 days of culture with α-MEM + 10% FBS or α-MEM + 5% FBS + 5% CS, the conditioned media in each subconfluent well was harvested. Osteocalcin in these conditioned media was measured using a mouse osteocalcin radioimmunoassay (RIA) kit according to manufacturer's instructions (Biomedical Technologies Inc., Stoughton, MA, U.S.A.).

Cell lines used as controls

The osteoblastic cell line, MC3T3-E1, which was established from normal mouse calvaria(9) was a gift of Dr. Hiroaki Kodama (Ohu University, Fukushima, Japan). The osteoblastic cell line, OCT-1, which was established from osteocalcin promoter driven T-antigen transgenic mouse calvaria(6) was a gift of Dr. Di Chen (UTHSC, San Antonio, TX, U.S.A.). Primary osteoblastic cells were isolated from neonatal mouse calvaria by sequential collagenase digestion according to the previously described method of Takahashi and coworkers(10) with a minor modification, using 2-day-old neonatal mice instead of fetal mice. Fractions 3–6 were used.


Establishment of an osteocyte-like cell line from transgenic mice

Before establishing cell lines, it is important to optimize cell culture conditions for that particular cell type, in our case, osteocytes. Using primarily the method of Mikuni-Takagaki and coworkers,(3) we found that isolated osteocyte-like cells from the transgenic mice grew rapidly when cultured in media containing CS compared with no growth with FBS. However, this rapid growth resulted in cells that lost their dendritic morphology which may represent dedifferentiation into an osteoblast-like phenotype. Therefore, 5% FBS was added with 5% CS supplemented culture media for optimal osteocyte growth and maintenance. We also tested collagen-coated surfaces and found them to be an effective way to maintain the osteocyte-like phenotype.

Figure FIG. 6.

(A) Expression of connexin 43 by immunocytochemistry of MLO-Y4 ceils, (a) The entire cell surface, including dendritic processes, is stained by antibody specific for connexin 43. (b) Nonspecific staining using secondary antibody alone. (B) Western blot analysis of cell lysate using antibody specific for connexin 43. Equal amounts of protein (10 μg/lane) were loaded into each lane. Brain tissue was used as a positive control. Note the large amount of connexin 43 (MW = 43 kDa) compared with the positive control and lack of detectable bands in the osteoblast cell lines and primary cells (Prim. OBI).

To establish osteocyte-like cell lines, seven 14-day-old transgenic mice were used for the isolation of cells from the long bones (tibiae, femurs, and humeri) because these bones, in contrast to calvariae, respond to mechanical stress in vivo with an increase in cortical bone. After removing the growth plate cartilage and marrow cells, the bones were cut into several pieces, and then sequential collagenase digestions were performed (fractions 1–5). We observed microscopically that cells present in the osteoid, presumed to be preosteocytes, were also released from trabecular bone along with bone surface cells at the 3rd to 5th collagenase digestions. Therefore, fractions 3–5 were combined and cultured on collagen-coated plates in α-MEM supplemented with 5% FBS/5% CS, the optimized culture condition for osteocytes. After several passages using these culture conditions, osteocyte-like cells with the dendritic phenotype were enriched in fractions 3–5.

The MLO-Y4 cell line was cloned from this population by single colony isolation. Selection was based on expression of the dendritic phenotype. To examine morphological changes with growth and proliferation, MLO-Y4 cells were plated on collagen-coated surfaces at low density and were observed by phase-contrast microscopy over time (Fig. 1). These cells adhere very rapidly to substrate, and after 3–6 h, the cells were small and stellate in shape with many short processes (Figs. 1A–1B). After 1–2 days, these processes became elongated and began to branch (Figs. IC–1E). In the confluent phase, the cellular processes continued to make contact with other cells (Fig. IF). These cells maintained this homogeneous morphology after more than 45 passages. They were also positive for T-antigen expression by Western blot analysis (Fig. 2).

The proliferation rate of the MLO-Y4 cell line was determined and compared with the osteoblast-like cell line, OCT-1, using collagen-coated and plastic surfaces and 10% FBS-containing medium compared with 5% FBS/5% CS. OCT-1 was established from the same type of transgenic mouse, the osteocalcin/promoter T-antigen mice, because MLO-Y4 cells except OCT-1 cells were established from an adult founder mouse.(6) The MLO-Y4 cells grew faster with 5% FBS/5% CS than with 10% FBS. OCT-1 cells grew faster in 10% FBS compared with 5% FBS/5% CS. However, MLO-Y4 growth on the collagen surface was slightly suppressed compared with plastic. This could be due to greater expression of the dendritic morphology on collagen surfaces using either 5% FBS/5% CS or 10% FBS containing media (Fig. 3).

MLO-Y4 cells have low ALP activity but high osteocalcin secretion

Because previous studies have shown that osteocytes express very low ALP activity,(3) we examined ALP activity of MLO-Y4 cells. By immunocytochemistry, MLO-Y4 cells were negative, with only weak staining in the cytoplasm of a small number of cells (Fig. 4A). MLO-Y4 cells expressed very low ALP-specific activity throughout the entire culture period compared with primary osteoblasts and OCT-1 cells cultured in both 10% FBS (Fig. 4B) and 5% FBS/5% CS (data not shown) containing medium.

In contrast, MLO-Y4 cells secreted very high levels of osteocalcin into conditioned medium within only 3 days of plating (11.80 ng/ml with 10% FBS and 11.25 ng/ml with 5% FBS/5% CS) while osteoblast-like cells expressed barely detectable osteocalcin during this same culture period (primary osteoblasts, 0.9 ng/ml; MC3T3-E1, 1.9 ng/ml; and OCT-1, 1.5 ng/ml with 10% FBS) (Fig. 5A). Reverse-transcription polymerase chain reaction (RT-PCR) results supported observations made using the osteocalcin RIA. Osteocalcin mRNA was present in MLO-Y4; however, primary osteoblast cells, MC3T3-E1, and OCT-1 did not present detectable osteocalcin mRNA under identical RT-PCR conditions (Fig. 5B).

ML0-Y4 cells express large amounts of connexin 43

Because it has been shown in situ that osteocytes express connexin 43, a gap junction protein.(11) we examined MLO-Y4 cells for expression of this protein important for cell-cell communication. MLO-Y4 was strongly positive by immunocytochemistry using anti-connexin 43 antibody (Fig. 6A). Staining was found in the cytoplasm and along the long dendritic processes. Strong staining was also observed around the nucleus in some cells. Western blot analysis was also used to compare MLO-Y4 cells to osteoblast-like cells for expression of connexin 43. Connexin 43 was detected in MLO-Y4 cell lysates from both 10% FBS and 5% FBS/5% CS culture conditions. The band in the MLO-Y4 lane was stronger than mouse brain tissue lysate which was used as positive control. In contrast, mouse primary osteoblast cells, MC3T3-E1, and OCT-1 cell lysate were negative although equivalent amounts of protein was loaded (Fig. 6B).

MLO-Y4 cells express similar amounts of osteopontin and CD44 mRNA compared with osteoblasts

Because osteocytes have been reported to express osteopontin, a major matrix protein,(12) and CD44, a neural antigen,(13,14) we examined MLO-Y4 cells for expression of these molecules by RT-PCR. Osteopontin and CD44 mRNA were expressed by both MLO-Y4 and osteoblast-like cells (Figs. 7A and 7B).

MLO-Y4 cells express low amounts of type I collagen mRNA compared with osteoblasts

Type I collagen mRNA expression in MLO-Y4 was not detectable by RT-PCR for 25 cycles, whereas bands were clearly present in osteoblast-like cells (Fig. 7C). When RT-PCR conditions were extended to 30 cycles, the band corresponding to type I collagen was faint but detectable in MLO-Y4 cells (data not shown). This suggests that MLO-Y4 cells express low levels of type I collagen compared with osteoblast-like cells.

MLO-Y4 cells do not express osteoblast-specific factor 2

Osteoblast-specific factor 2 (OSF-2) was recently cloned from an MC3T3-E1 library and is expressed by primary osteoblasts, by MC3T3-E1 cells, and in lung tissue as shown by Takeshita and coworkers.(15) OSF-2 is proposed to be an osteoblast marker; therefore, RT-PCR was performed to determine if MLO-Y4 cells also express this protein. We found that MLO-Y4 cells do not express any detectable OSF-2 mRNA by RT-PCR using 30 cycles. In contrast, a band corresponding to OSF-2 was obvious in the osteoblast-like cells MC3T3-E1, OCT-1 cells, and primary osteoblasts (Fig. 1D). The same results were obtained when the number of PCR cycles was increased to 35 (data not shown).

The two faint upper bands in the MLO-Y4 lane (Fig. 7D) were sequenced, and although they contained the PCR primer sequences, they had no homology to OSF-2. One band had high homology to human transducin-like protein, and the second had no significant homology to any known protein, whereas the OSF-2 band in the osteoblast lanes was 100% homologous to the OSF-2 sequence. The upper two bands may have been transcribed more efficiently in the OCT-1 and MLO-Y4 cell lines as little (OCT-1) or no (MLO-Y4) OSF-2 mRNA was available for transcription. The characteristics of the MLO-Y4 cell line are summarized and compared with osteoblast-like cells in Fig. 8.

Figure FIG. 7.

(A) Expression of mRNA for osteopontin, (B) CD44, (C) type I collagen, and (D) OSF-2 in MLO-Y4 cells as compared with other cell types as determined by RT-PCR, using 25 cycles for osteopontin and type I collagen or 30 cycles for CD44 and OSF-2.


Here we show the establishment and characterization of a cell line with the features ascribed to osteocytes. This cell line was named murine long bone osteocyte, Y4 (MLO-Y4) to emphasize the fact that it was established from long bones, the bones that respond to increased mechanical stress with an increase in bone formation. Because the best marker for mammalian osteocytes at this time is their morphology, (3,16–18) this cell line was selected on the basis of expression of dendritic processes, a characteristic morphologic feature of osteocytes. Although osteocyte-specific antibodies are available for avian cells,(19,20) none are available for mammalian osteocytes at this time.

Figure FIG. 8.

The characteristics of the MLO-Y4 cell line as compared with primary osteoblast cells and to two osteoblast cell lines MC3T3-E1 and OCT-1. ↗, increased activity in the culture period; ⟷, no change during the culture period; +, detectible, —, not detectible; Western, Western blot analysis; EA, enzyme assay; RIA, radioimmunoassay; RT-PCR, reverse transcription-polymerase chain reaction analysis.

MLO-Y4 cells express proteins also expressed by osteoblasts such as osteocalcin, osteopontin, connexin 43, CD44, ALP, and type I collagen but in relative amounts described for osteocytes. The low expression of ALP and high expression of osteocalcin by MLO-Y4 cells supports the hypothesis that the MLO-Y4 cell line is osteocyte-like as reports by others show this pattern of expression in primary osteocytes.(3) However, the low expression of type I collagen in MLO-Y4 cells compared with osteoblasts contrasts with a report by Aarden and coworkers(4) but is in agreement with reports by others(21,22) and by Nijweide and coworkers who have found that type I collagen is produced in relatively low abundance by osteocytes compared with osteoblasts (personal communication). Our results are also in conflict with the reports by Hughes and coworkers(3) and Nakamura and coworkers(14) who used immunohistochemical techniques to show that osteocytes are positive for CD44, whereas osteoblast and lining cells are negative. However, Hassan and coworkers(23) have recently shown that osteoblasts at different stages of maturation express mRNA and protein for CD44 both in vivo and in vitro. In the present study, both osteoblast cell lines tested, MC3T3-E1 and OCT-1, primary osteoblasts, and MLO-Y4 express CD44. Also recently, a human preosteocytic cell line (HOB-01-C1) has been established and characterized(24) Not only was this preosteocytic cell line positive for CD44, but so were the osteoblastic cell lines used for comparison. These data suggest that CD44 is not a specific marker for osteocytes.

Gap junctions are conduits for cell-to-cell communication (for review, see Edelson(25)). Gap junctions penetrate the cell membranes of two communicating cells to allow the flow of low molecular weight signaling molecules such as Ca2+, cAMP, and inositol triphosphate. Gap junctions are composed of structurally related proteins known as connexins. Several connexins have been shown to be expressed by osteoblasts.(26–30) By Northern analysis, MC3T3-E1 cells have been shown to express large amounts of connexin 43 mRNA,(31) and cultured osteoblasts from newborn rat cal-varia also have been shown to express large amounts of this protein.(26) Because the expression of connexin 43 has recently been described for osteocytes in vivo,(11) we examined MLO-Y4 cells for expression of this gap junction protein. We were surprised to find very large amounts of connexin 43 protein produced by the MLO-Y4 cells, especially when compared with equivalent amounts of brain tissue (the positive control). Our data suggest that osteocytes may be the major source of connexin 43 in bone, especially when compared with osteoblasts.

OSF-2, a marker for osteoblasts, was recently cloned from an MC3T3-E1 library and shown to have homology with an insect protein, fasciclin 1, that functions as a hemophilic adhesion molecule.(10) OSF-2 is expressed in primary osteoblasts, MC3T3-E1 cells, and in lung tissue. Brain, heart, kidney, liver, muscle, placenta, spleen, testis, and thymus are negative for this marker. It is not known if OSF-2 is expressed by osteocytes or other bone cells. The cell line MLO-Y4 does not express OSF-2 mRNA when compared with the osteoblast cell lines OCT-1 and MC3T3-E1 and primary osteoblasts analyzed in the same experiments. Therefore, OSF-2, a putative bone adhesion molecule, is a negative marker for this cell line and may be a negative marker for osteocytes in vivo. This remains to be determined.

Two osteoblastic cell lines were previously established from the same strain of transgenic mice used in the present study.(6) These transgenic mice possess a 2.6 kb fragment of the rat osteocalcin promoter driving the expression of SV40 large T-antigen. Several studies have demonstrated the usefulness of using tissue-specific promoters and SV40 large T-antigen for developing immortalized cell lines.(32,33) These cell lines, termed OCT-1 and OCT-2, were classified by their capacity to differentiate into cells with osteoblast characteristics. OCT-1 and OCT-2 were derived from the sequential digestion of the calvaria of one founder transgenic mouse. In contrast, in the present experiments, cells were isolated from the long bones of young 14-day-old mice through a series of digestions designed to select for cells encapsulated within the mineralized bone matrix. Although derived from the same strain of transgenic mice, the osteocyte cell line appears to be distinctly different from and produces large amounts of osteocalcin in contrast to the OCT-1 and OCT-2 cells.

Lanyon and coworkers(35–37) have performed numerous studies examining the effects of mechanical loading on bone in vivo and propose that the osteocyte is the major cell responding to mechanical stress. Sensors and/or tranducers on osteocytes appear to respond to load-induced strain. Glucose 6-phosphate dehydrogenase activity increases transiently in osteocytes soon after loading, and loading also appears to induce PGE2 and PGI2 production by these cells. Isolated osteocytes, but not osteoblasts nor periosteal fibroblasts, react to pulsating fluid flow with a release of PGE2.(37) Lean and coworkers(38) have shown that osteocytes respond to mechanical loading with an increase in mRNA for insulin-like growth factor I within 6 h. Mikuni-Takagaki and coworkers(39) concluded from their studies that isolated osteocytes respond differently from young osteocytes and osteoblasts to both low, physiological strain and to higher magnitudes of strain. It will be of interest to conduct similar studies with the cell line MLO-Y4 to determine if the response of this cell line is similar to isolated primary osteocytes. This will be the focus of future studies.


The authors thank Dr. Darren Ji (UTHSCSA) for the sequences of the oligonucleotides used for the RT-PCR studies and Mr. Kazahiko Katayama (Asahi Chemical Industry Co., Ltd.) for sequencing the PCR products. We also gratefully acknowledge the financial support of Asahi Chemical Industry Co., Ltd. This work was supported in part by National Institutes of Health grant R01-EY09213 (J.J.W.).