Monoclonal antibodies recognize antigen expressed by osteoblasts

JOURNAL OF BONE AND MINERAL RESEARCH Volume 10, Number 10, 1995 Blackwell Science, Inc.

Monoclonal Antibodies Recognize Antigen Expressed by Osteoblasts D. BENAYAHU,' M. EFRATI,' and S. WIENTROUB'.3

ABSTRACT A marrow stromal osteogenic cell line (MBA-15) was used to create monoclonal antibodies (MoAbs). In this study, we describe a series of MoAbs for mouse marrow stroma (MMS) (MMS-25/17, MMS-85/12, MMS-302/40, and MMS-319/4) that recognized antigens expressed by stromal cells including osteoblastic cells. The MoAbs were screened against various cell and tissue types. MMS-85/12 was positive in detecting an antigen that was highly abundant in osteoblastic cells and primary adherent bone marrow cultures (BMC) but was negative for the marrow adipocytes copartner. The MMS-85/12 MoAb is an IgGl immunoglobulin. The immunohistochemical staining pattern is suggestive of the antigen being associated with the osteoblasts' plasma membrane and with the extracellular matrix constituent secreted by these cells. Western blotting and immunoprecipitation indicated that the antigen that was recognized by MMS-85/12 apparently had a molecular weight of 84 kD. (J Bone Miner Res 1995;lo 1496-1503)

INTRODUCTION TISSIICwith a remarkable composite of heterogenous cell types, extracellular matrix, and minerals. The ccllular constituents that are intimately associated with the bone matrix are marrow elements comprised of stromal and hematopoietic cells. The cells involved in bone physiology are bone-lining cells, bone-forming cells (osteoblasts), and cells surrounded by mineralized matrix (osteocytes). Osteoblasts are primarily responsible for the synthesis, deposition, and organization o f the extracellular matrix that is physiologically mineralized. The fully differentiated osteoblast is located opposite the bone surface and is easily recognized by its cuboidal shape. Osteoblasts differentiate to a definitive stage when forming osteocytes. These stages are preceded by an unknown number of precursors up to the stem cell of the osteogenic lineage. These various cell differentiation stages are difficult to recognize in vivo. Cells isolated in defined cultures are used in the study of the regulation of proliferation and differentiation of the various cell stages in the osteoblastic lineage."-"

B

ONL: I S A SPI:.UAL.IZED CONNI:CI'IVI,

The marrow is a reservoir source for osteoprogenitors, but very little is known about the mechanisms directing their differentiation in the medullary cavity, a process that leads to full osteoblastic expression on the endosteal surface. There is a fine balance between different cellular stages that control the osteoblastic cell renewal and cell loss. It is important to identify the restricted cell populations of the bone marrow which become fully expressed osteoblasts. The establishment of markers for osteoprogenitor cells is crucial both for such identification and for studying their differentiation toward mature osteoblasts. New markers may be of practical importance in understanding basic bone physiology and bone in a disease state. Therefore, we tried t o identify and characterize marker(s) for marrow-derived osteoblasts. For this purpose we used a marrow stroma-derived osteogenic cell line, MBA-15. These cells were characterized for their osteoblastic phenotype in vitro; they expressed alkaline phosphatase activity that was modulated by hormones such as d e ~ a m e t h a s o n e , ( ~ ) retinoic acid,'" and various growth factors.'") The cells responsiveness to PTH and PGE, was measured by changes in intracellular CAMP.'") MBA- 15 cells produced extracel-

'Department of Cell Biology and Histology, Sackler Faculty of Medicinc, Tel-Aviv University, Td-Aviv, Israel. 'Division of Orthopacdics, Sacklcr Faculty of Medicine, Tel-Aviv University, Tel-Aviv, Israel. 'Department of Pediatric Orthopaedics, Dana Children's Hospital, Tel-Aviv Medical Center, Tel-Aviv, Israel.

1496

MONOCLONAL ANTIBODIES AGAINST MARROW STROMAL OSTEOBLASTS lular matrix proteins, mainly of collagen type I,(4)and were also shown to express niRNA for various noncollagenous proteins. such as osteonectin, biglycan, and o~teopontin.(',~' In the presence of ascorbic acid and P-glycerophosphate, the MBA- 15 cultures mineralized and formed hydroxyapatite crystal^.'^*^' In vivo studies of implanted MBA-15 cells exhibited their capability to form bone.'43x)In this study, we describe a series of MoAbs raised against MBA-15 cells. These MoAbs could detect antigen specifically expressed by marrow stromal osteoblast and other stromal cell types, but not by marrow-derived adipocytes.

MATERIALS AND METHODS

Cells Osteogenic marrow stromal MBA-I5 cells derived from SJL/J mice'4) and clonal subpopulations(3) were seeded in tissue culture dishes (Nunc, Denmark) in Dulbecco's modified Eagle's medium (DMEM, Beth Haemek, Israel) containing high glucose supplemented with 10% fetal calf serum (FCS, Bio-Lab, Israel). Stocks of cells were passaged once weekly by removal with a rubber policeman and dispersed in fresh medium. All cultures were incubated at 37°C in a humidified atmosphere of 10% CO, in air. Other stromal cells used were MBA-1.1.1, MBA-2.1, MBA-13.5, and 14F1.1.(''-I I ) The osteoblastic cells of nonstromal origin included in this study were MC3T3-El and ROS 17/28 cell lines and primary cultures of calvaria cells and adherent marrow cells. Other nonosteoblastic cell types, such as myeloblasts (MI), kidney epithelia (Cos), and neuroblastoma (IMS), were grown under the same conditions. Macrophages (14Ml) were grown with the addition of L cellconditioned medium as a source for CSF-I.

1497

Screening assay for selection of MoAb by ELISA The specific antigen determination of secreting hybridomas was identified on cultured cells. Examined cells were plated for 4 days in 96-well plates, after which the plates were washed with phosphate-buffered saline (PBS) and fixed in 0.3% paraformaldehyde in PBS. For the assay itself, the cells fixed in the plates were blocked with skim milk and incubated with the respective primary antibodies overnight. A second antibody peroxidase-conjugated goat anti-mouse immunoglobulin (BioRad Lab. Ltd., Herts, U.K.) (1:3000) was added for 30 minutes and incubated at 4°C. After rinsing, the cultures were incubated at 37°C with substrate solution containing 2 mdml 0-phenylenediamine (OPD) (Sigma Chemical Co., St. Louis, MO) in 50 mM citric buffer pH 5.0 and 30 ~ 1 3 0 % H202.The reaction was stopped after 15 minutes by the addition of 4 N HCI. Optical density (OD) was read on a microplate reader system at 490-405 nm (Molecular Devices, Palo Alto, CA).

Flow cytometric analysis MBA-15 cells were released from culture with trypsin and prepared as single cell suspensions. Mouse bone marrow cells (BMC) and thymocytes were freshly isolated from mice. A total of 1 X 10' cells were fixed in 0.3% paraformaldehyde in PBS and rinsed in 0.2% Triton X-100 (Sigma Chemical Co.) in PBS. Following incubation with the first antibody for 90 minutes on ice with occasional agitation, they were rinsed in PBS and incubated with goat antimouse FITC-conjugated secondary antibody (Zymed Lab., San Francisco, CA) for 30 minutes. Cells were washed in PBS and 1 X lo4 cells were collected for each sample. The accuracy of positive cells fluorescently labeled with various MoAbs was quantified with a four-parameter fluorescenceactivated cell sorter (FACS) (Beckton-Dickinson).

Immunization and hybridization BAL,B/c mice were immunized with 5 X 10" MBA-15 cells per mouse by subcutaneous ( S C ) injection in complete Freund's adjuvant. The initial injection was followed 2 weeks later by an SC injection using incomplete adjuvant followed by another three intraperitoneal (ip) injections every 2 weeks. Ten days later, the sera were tested for the immune response to cultured cells, using colorimetric enzyme-linked immunosorbant assay (ELISA). The animal that showed a high titer received another intravenous (iv) or ip injection with 5 X 10" cells 3 days prior to splenocyte fusion. Splenocytes were fused with NSO myeloma cells using polyethylene glycol 4000 (Merck, Germany) as the fusing agent. The fusion product was plated into 96-well plates in hypoxanthine-aminopterin-thymidine(HAT) medium supplemented with 10% FCS according to the method of Kohler and Milstein,"') at the Hybridoma Unit of the Faculty of Life Sciences, Tel-Aviv University.

IgG subclass determination The isotyping for monoclonal antibodies was determined using the Serotec Kit (Wellcome Ltd., U.K.), employing a test based on red cell agglutination.

Gel electrophoresis and Western blotting analysis The harvested cultured cells were washed twice with PBS, and the pellets of cells were frozen at -20°C. Various organs were removed from the animal, homogenized in TENN buffer (50 mM Tris, 5 mM EDTA, 0.5% NP40, 150 mM NaC1, pH 7.4 with 1 mM phenylmethyl-sulfony1flu~)ride [PMSF] using a Polytron homogenizer (Kinematica, Switzerland) on ice. The extracts were centrifuged for 15 minutes at 7000g and at a temperature of 4°C. Demineralized long bones were extracted by EDTA ( E extracts) and other purified proteins such as osteonectin (ON), bone sialoprotein (BSP) and alkaline phosphatase (ALK-P) were prepared from demineralized bone matrix of various species as described earlier by Fisher et al.'") All samples were prcpared for extraction in sodium dodecyl sulfate (SDS), boiled for 3 minutes and dissolved in a sampling buffer (62.5 mM Tris-HCI, pH 6.8, 2% SDS, P-mercapto-ethanol, and 0.01% bromophenol blue). Two to 5 p g of protein from the samples were used for electrophoresis separation in a 0.75 mm-thick mini-gel of 10% SDS-PAGE polyacrylamide slabs gels with a 2.5% polyacrylamide stacking gel. A horizontal electrophoresis apparatus (BioRad Lab., Ltd.) was used to separate the proteins.'14) The proteins were then

BENAYAHU ET AL.

1498

electroblotted to a nitrocellulose membrane using a semidry transfer blotting apparatus (BioRad Lab., Ltd.), by a modification of Hurley and Frinkelstein's method.('" Nitrocellulose blots were blocked in skim milk and incubated first with MoAb for 30 minutes at 37°C or overnight at 4°C and then with peroxidase-conjugated affinity purified antibody to mouse IgG (BioRad Lab., Ltd.). The blots were exposed to a freshly prepared solution containing either diaminobenzidine tetra-hydrochloride (DAB) (Sigma Chemical Co.) or ECL (Amersham, U.K.). The latter substrate is used for providing higher sensitivity for visualization of the Ag-Ab complex on the blot.

Metabolic labeling and immunoprecipitation of antigens New antigens synthesis was detected by incubation of 3 X lo5 cells in methionine-free medium supplemented with 125 pCi L-[-?3] methionine (Du Pont, NEN, U.K.) at 37°C for 4 h. Following labeling, the cells were lysed according to the method of Jeffries et al."') modified with an addition of 0.5% (viv) NP40, 150 mM EDTA, 5 mM iodoacetamide, 2 mM L-methionine, 0.5% (w1v) BSA, 150 mM NaCI, and 1 mM PMSF in Tris-HCI, pH 6.8. Nuclei were removed by centrifugation at 7000g for 5 minutes. A 10% (v/v) suspension (0.1 ml) of protein A-Sepharose C L 4 B (Pharmacia, Sweden) previously saturated with rabbit anti-mouse IgG was prepared. Hybridoma supernatant was added to aliquots of the mixture of cell lysate and protein A. Immunoprecipitates were released from the beads by boiling them in a sample buffer and then analyzed by SDS-PAGE(I4) followed by fluorography.'

"'

Dot blot of conditioned media Conditioned media (CM) from various stromal cell lines were harvested from confluent cultures. Each medium was removed from cultured cells and replaced by fresh growth medium containing 2% FCS in DMEM free of phenol red. The CM were collected after 48 h, millipore (0.45 pm)filtered, and examined for their antigen expression.

Immunocytochemistly of cultured cells Cells were cultured on glass coverslips fixed for 10 minutes at -20°C in methanol and acetone, washed, and then incubated for 30 minutes in 0.01% H20, to deplete any endogenous peroxidase activity. After washing, the cells were preincubated for 20 minutes in skim milk. The coverslips were incubated overnight at 4°C with the hybridoma culture supernatant washed twice for 5 minutes each time and incubated for 30 minutes with affinity-purified biotinlabeled rabbit anti-mouse (Bio-Makor, Israel) diluted at 1: 100, followed by avidine-peroxidase (Bio-Makor, Israel). The cells were then exposed to 0.1% DAB (Sigma Chemical Co.) used as a substrate and then counterstained with hematoxylin.

HMBA-15

D R O S 17/2.8

Skln Flbro

FIG. 1. Binding of 4 MMS MoAbs to cultured cells was measured by ELISA immunoassay. The immunoassay was performed by the MoAbs binding to fixed cells in %-well plates using peroxidase-conjugated IgG with O D P used as a substrate (described in Materials and Methods).

RESULTS BALB/c mice were immunized with MBA-15 cells and the highly immunized mice were selected for MoAb production.

Screening and selection of hybridomas Two fusion experiments produced 330 hybridomas, and a total of 4 MoAbs (MMS) of interest were selected. Supernatants of hybridomas (SM) were characterized for their Ig subclasses. MMS-85/12 and MMS-25/17 were identified as IgGl, MMS-31914 as IgG3 while MMS-302140 belonged to an IgM subtype. The MoAbs were selected for their binding capacity to cultured cells using ELISA. Their affinity to MBA-15 cells, but not to skin fibroblasts which originated from the same mouse strain, was examined (Fig. 1 ) . The MoAbs were further characterized for their binding efficiency to other osteoblastic cell lines, ROS 17/23 and MC3T3-EI. While all selected MoAbs positively recognized the antigens on tested cells, they differed in the range of their response. MMS-85/12 and MMS-25/17 were bound preferentially to various osteoblastic cells but not to SJL mouse-derived skin fibroblasts. MMS-31914 and MMS302140 showed a higher response to skin fibroblasts and 302/40 had a low affinity toward MC3T3-El (Fig. 1).

Immunohistochemistry of MMS MoAbs Cellular localization of the antigens was performed on cultured cells. Various MoAbs expressed different patterns of staining of the MBA-15 cells (as summarized in Fig. 2 and Table I). MMS-85/12 stained a cell membrane associated with molecules of an ECM protein (Figs. 2A and 2B), MMS-25/17 showed a similar distribution (Figs. 2C and 2D), while MMS-302140 (Fig. 2E) and MMS-31914 (Fig. 2F) stained intracellular components. Similar results were obtained with primary calvaria cells and MC3T3-El cells (data not shown).

MONOCLONAL ANTIBODIES AGAINST MARROW STROMAL OSTEOBLASTS

1499

A

FIG. 2. Immunohistochemistry staining by MMS MoAbs of cultured MBA-15 cells. The MoAbs 85/12 reacted positively with the cell surface and the ECM components deposited between the cells (A and B, X900). MoAb 25/17 reacted to ECM (C and D, X330). MoAbs 302/40 and 31914 reacted with cytoplasmic components ( E and F. X330). The staining was performed using biotin-avidin peroxidase and DAB was used as substrate. TA8I.E

1.

I M M U N O I I I S I O ~ ' H l M I S 1 R Y OF C U L l U R t l l C E L L S

W I T I I MMS

Localizution Extracellular matrix (ECM) Cell surface Cytoplasmatic

MoAbs

85/12 25/17 302140 31914

+

+ ~

+ + -

-

-

+

+

FAG'S analysis Single cell suspensions from MBA-15 cultures were stained and analyzed by flow cytometry. The sorted cells showed a curve of an increased fluorescence in relation to antibody dilution MMS-85/12 (Fig. 3A), MMS-302/40 (Fig. 3B) and MMS-31014 (Fig. 3C).

Western blotting an alysis Cell extracts, various mouse tissues, and bone matrix proteins were first compared by using SDS-PAGE and electroblotting them to a nitrocellulose membrane and then reacting then with the tested MoAbs. Cells suhfype of strornul origin: The MoAb's recognition of antigens expressed by various stromal cells and their appar-

Fluorcscciicc Iiitciisity FIG. 3. Flow cytometry of MBA-15 cells with dose response to MMS MoAbs. MoAbs were diluted at 1: 100, 1:20, and 1.5 and an increased fluorescence is seen. MMS-85/12 (A), MMS-302/40 (B), and MMS-319/4 (C). The arrow indicates fluorescence monitored when stained with the second antibody only. ent molecular weight (MW) are summarized in Table 2. MMS-85/12 and MMS-302/40 detected bands with a MW of 84 kD and 75 kD, respectively. The antigens had a similar distribution on MBA-15, MBA-13.5, MBA-1.1.1, and MBA-2.1 stromal cells. MMS-85/12 did not recognize any antigenic epitope on the 14Fl.l adipose cells lysate. MMS25/17 recognized two bands in all positive stromal cells, as well as an additional third band at a lower MW of 35 kD expressed by MBA-15 cells only. Bone matrix proteins: MMS antibodies were used to screen bone matrix proteins in E-extracts prepared from various animal species. A positive signal was detected with MMS-85/12 in mouse E-extract having an apparent MW similar to that detected in the cell lysates. A weak signal was seen in rat extracts. Neither purified BSP nor ON reacted with the tested MMS MoAbs. Tissues localization: We also used the MoAbs to screen for recognition of an epitope on a variety of tissue extracts

BENAYAHU ET AL.

1500

T A U L2.I ~WESTERN BLOTTING ANALYSIS OF VARIOUS MARROW STROMAL CtLL TYPES AND OSTEOBLASTIC CELLSWITH SPECIFIC MoAkis

No. (,f

Approx.

Strornal cell lines ~-

~~~~

- ~-

~~~

~

~-

~~-

MMS

hunds

MWofAg

MBA 15

MBA 13.5

MBA 1.1.1

MBA 2.1

85/12 302140 25/17

1 1

84 75 30-40

+ + +++

+ + ++

+ + ++

+ +

2or3

++

~

14FI.l -

ND ND

Lysates from cells expressing various antigens were separated by an SDS-PAGE gel and by electroblotting to a nitrocellulose membrane. The Ah-Ag complex was visualized by DAB used as a substrate. The number of bands by various cell examined is indicated by (+), N D = not determined.

Otbcr MoAbs

MMSIS/I?

A

6

C

D

Tliyriiocytes

Fluorescence Intensity

TA~$LI: 3. WES.rEKN BLOTTING ANALYSIS OF MBA-IS CELIS COMPARI