Genetic Portrait of Osteoblast-Like Cells Cultured on PerioGlas

PerioGlas (PG; US Biomaterials Corp, Alachua, Fla) is an alloplastic material that has been used for grafting periodontal osseous defects since 1995. In animal models it achieves histologically good repair of surgically created defects. In monkeys^1–3 PG demonstrates biocompatibility and osteoconductive activity. It is mostly resorbed and replaced by bone, and the remaining granules are in close contact with bone. In a rabbit model, PG is able to improve bone healing at the interface between titanium dental implants and bone,4 whereas in ovariectomized rats it enhances newly bone formation into extraction sockets.5 

In clinical trials, bioactive glass is effective as an adjunct to conventional surgery in treating intrabony defects6 and dental extraction sites before dental implant placement, in order to regenerate bone and enhance early fixation of implants.7 However, PG has no regenerative properties with regard to cementum and periodontal ligament.8 

In previous in vitro studies on human cells, the osteoblast cell line MG63 was used as a prototype of human bone cells to test bioglass that found favorable results.9 Human primary osteoblasts were used to investigate the osteogenic potential of a melt-derived bioactive glass (BG). It was shown that BG induces osteoblast proliferation and increases osteoblast activity; thus, it was hypotized that BG could be used as a template for forming bioengineered bone tissue.10 

From a molecular point of view, it has been shown that ionic products of BG dissolution increase proliferation of human osteoblasts and induce insulin-like growth factor II messenger RNA expression and protein synthesis.11 In addition, a gene-expression profiling of human osteoblasts after treatment with the ionic products of BG dissolution was performed by using complementary DNA (cDNA) microarray containing 1176 genes.12 

To our best knowledge, however, no direct analysis on PG has been performed yet. Thus, because the mechanism by which PG stimulates osteoblast activity to promote bone formation is poorly understood, we attempted to address this question by using microarray techniques.

DNA microarray is a molecular technology that enables the analysis of gene expression in parallel on a very large number of genes, spanning a significant fraction of the human genome. Gene expression is performed by a process of (1) RNA extraction, (2) reverse transcription, and (3) labeling of cDNA. Reference (ie, untreated cells) and investigated (ie, cells cultured with PG) cDNA are labeled with different dyes and then hybridized on slides containing cDNA fragments. Then the slides are scanned with a laser system, and 2 false color images are generated for each hybridization with cDNA from the investigated and reference cells. The overall result is the generation of a so-called genetic portrait.13,^–17 It corresponds to upregulated or downregulated genes in the investigated cell system. In the present study we define the genetic effect of PG on osteoblast-like cell line (ie, MG63) by using microarray slides containing 20 000 different oligonucleotides.

Osteoblast-like cells (MG63) were cultured in sterile Falcon wells (Becton Dickinson, Franklin Lakes, NJ) containing Eagle's minimum essential medium supplemented with 10% fetal calf serum (Sigma Chemical Co, St Louis, Mo) and antibiotics (penicillin 100 U/mL and streptomycin 100 μg/mL; Sigma Chemical Co). Cultures were maintained in a 5% CO₂ humidified atmosphere at 37°C.

MG63 cells were collected and seeded at a density of 1 × 10⁵ cells/mL into 9 cm² (3 mL) wells by using 0.1% trypsin, 0.02% ethylene diamine tetraacetic acid in Eagle's buffer free of Ca++ and magnesium for cell release. One set of wells received PG at the concentration of 0.04 g/mL. After 24 hours, when cultures were subconfluent, cells were processed for RNA extraction.

The protocol was the same used in previous experiments.13,^–17 RNA was extracted from cells by using RNAzol (RNAzol Molecular System, San Diego, Calif); 10 μg total RNA was used for each sample. The cDNA was synthesized using Superscript II (Life Technologies, Invitrogen, Milano, Italy) and amino-allyl dUTP (Sigma Chemical Co). Monoreactive Cy3 and Cy5 esters (Amersham Pharmacia, Little Chalfont, UK) were used for indirect cDNA labeling. RNA extracted from untreated cells was labeled with Cy3 and used as a control against the Cy5 labeled treated (PG) cDNA in the first experiment and then switched. For 20 000 human DNA microarray slides (MWG Biotech AG, Ebersberg, Germany) 100 μL of the sample and control cDNAs in DIG Easy hybridization solution (Roche, Basel, Switzerland) was used in a sandwich hybridization of the 2 slides constituting the 20 000 set at 37°C overnight. Washing was performed 3 times for 10 minutes with 1 × saline sodium citrate (SSC), 0.1% sodium dodecyl sulfate at 42°C, and 3 times for 5 minutes with 0.1 × SSC at room temperature. Slides were dried by centrifugation for 2 minutes at 2000 rpm. The experiment was repeated twice and the dyes were switched. A GenePix 4000a DNA microarrays scanner (Axon, Union City, Calif) was used to scan the slides, and data were extracted with GenePix Pro (Axon). Genes with expression levels <1000 after removing local background were not included in the analysis, because ratios are not reliable at that detection level.

After scanning the 2 slides containing the 20 000 human genes in duplicate, local background was calculated for each target location. A normalization factor was estimated from ratios of median. Normalization was performed by adding the log₂ of the normalization factor to the log₂ of the ratio of medians. The log₂ ratios for all the targets on the array were then calibrated using the normalization factor, and log₂ ratios outside the 99.7% confidence interval (the median ±3 times the SD = 0.52) were determined to be significantly changed in the treated cells. Thus, genes are considered significantly modulated in expression when the absolute value of their log₂ expression level is higher than 1.56 or there is a 3-fold difference in expression between treated cells and reference. GenePix Pro software was used to report genes above the threshold and with <10% difference in 3 different statistical evaluations of the intensity ratio, thus effectively enabling an automated quality control check of the hybridized spots. Furthermore, all the positively passed spots were finally visually inspected. Statistical analysis of microarray (SAM) program was then performed, and the SAM score was obtained (t statistic value).13,^–17 The genes differentially expressed in cells treated with PG are shown in Tables 1 and 2, and the SAM plot is shown in Figure 1.

PG is a silicate-based synthetic bone augmentation material that has been used to fill periodontal defects with binding and integration to soft tissue and bone. Previous histologic studies in animal models have shown that PG achieves good repair of surgically created defects.1,^–5 In clinical trials, PG has been shown to be effective as an adjunct to conventional surgery in treating intrabony defects6 and dental extraction sites.7 

With regard to molecular studies, a gene-expression profiling of human osteoblasts after treatment with the ionic products of BG dissolution was performed by using cDNA microarray containing 1176 genes.12 To our best knowledge, however, no direct analysis has been performed on PG. Thus, because the mechanism by which PG stimulates osteoblast activity to promote bone formation is poorly understood, we attempted to address this question by using microarray techniques to identify genes that are differently regulated in osteoblasts exposed to PG. Bone formation can be attributable to ostegenesis (ie, direct stimulation of osteoblast to produce new bone), osteoconduction (which operates like a scaffold), or both processes.

Hybridization of cDNA (derived form MG63 cultured with 0.04 g/mL of PG) to cDNA microarrays allowed us to perform systemic analysis of expression profiles for thousands of genes simultaneously and to provide primary information on transcriptional changes related to PG. We identified several genes whose expression was definitely upregulated and downregulated. Very few genes are upregulated and none has a major regulatory role (Table 1).

Among the downregulated genes (Table 2), some are involved in signal transduction like TGFB3 and GHRHR, a receptor for growth hormone-releasing hormone. Binding of this hormone to the receptor leads to synthesis and release of growth hormone. Mutations in this gene have been associated with isolated growth hormone deficiency, also known as dwarfism of Sindh, a disorder characterized by short stature.18 Another signal transductor is PLCB1, which catalyzes the formation of inositol 1,4,5-trisphosphate and diacylglycerol from phosphatidylinositol 4,5-bisphosphate. This reaction uses calcium as a cofactor and plays an important role in the intracellular transduction of many extracellular signals.19 

Cell-cycle regulation, proliferation, and apoptosis are modulated by PG. PCSK6 belongs to the subtilisin-like proprotein convertase family. Members of this family are proprotein convertases that process latent precursor proteins into their biologically active products. PCSK6 is a calcium-dependent serine endoprotease that can cleave precursor protein at their paired basic amino acid processing sites. Some of its substrates are transforming growth factor beta (TGFB)-related proteins.20 

DAPK3 (death-associated protein kinase 3) induces morphologic changes in apoptosis when overexpressed in mammalian cells; thus, it has been suggested that DAPK3 may play a role in the induction of apoptosis.21 PG acts also on genes related to the immune system. NFAT5 is a member of the nuclear factors of activated T cells family of transcription factors. Proteins belonging to this family play a central role in inducible gene transcription during the immune response.22 CCL26 is 1 of 2 Cys-Cys (CC) cytokine genes clustered on the q arm of chromosome 7. Cytokines represent a family of secreted proteins involved in immunoregulatory and inflammatory processes. The CC cytokines are proteins characterized by 2 adjacent cysteines. The cytokine encoded by CCL26 displays chemotactic activity for normal peripheral blood eosinophils and basophils.23 CD3 is a complex of protein associated to T cell antigen receptor on the T cell surface. CD3G (gamma chain) is 1 of the 4 peptides (gamma, delta, epsilon, and zeta) that form CD3. Defects in CD3G are associated with T cell immunodeficiency.24 

Also, cytoskeleton, cell adhesion, and extracellular matrix components are modulated by PG. FLRT3 is a member of the fibronectin leucine rich transmembrane protein family. It may function in cell adhesion and/or receptor signaling. Its protein structure resembles small leucine-rich proteoglycans found in the extracellular matrix.24 

B4GALT6 is 1 of the 7 beta-1,4-galactosyltransferase (beta4GalT) genes. They encode for type II membrane-bound glycoproteins that seem to have exclusive specificity for the donor substrate UDP-galactose; all transfer galactose in a beta 1,4 linkage to similar acceptor sugars: GlcNAc, Glc, and Xyl. Each beta4GalT has a distinct function in the biosynthesis of different glycoconjugates and saccharide structures. As type II membrane proteins, they have an N-terminal hydrophobic signal sequence that directs the protein to the Golgi apparatus and then remains uncleaved to serve as a transmembrane anchor. By sequence similarity, the beta4GalTs form 4 groups: beta4GalT1 and beta4GalT2, beta4GalT3 and beta4GalT4, beta4GalT5 and beta4GalT6, and beta4GalT7. B4GALT6 is a lactosylceramide synthase important for glycolipid biosynthesis.25 

FMOD is a member of a family of small interstitial proteoglycans. It participates in the assembly of the extracellular matrix as it interacts with type I and type II collagen fibrils and inhibits fibrillogenesis in vitro. It also regulates TGF-beta activities by sequestering TGF-beta into the extracellular matrix.26 Also LAMA4 is downregulated. It belongs to laminins, a family of extracellular matrix glycoproteins that are the major noncollagenous constituents of basement membranes and have a role in adhesion and migration of human bone marrow progenitor cells.27 

Among the receptors are CDH8 (that encodes for a type II classical cadherin from the cadherin superfamily, integral membrane proteins that mediate calcium-dependent cell-cell adhesion²⁸) and CD34 (that is a sialomucin-like adhesion molecule only expressed on a few percent of primitive bone marrow cells; in human bone marrow, virtually all colony-forming unit activity resides in the population expressing human CD34²⁹).

The genes discussed are only a limited number among those differentially expressed and reported in Table 1 and 2. We briefly analyzed some of those with a better-known function.

PG is able to downregulate some functional activities of osteoblast-like cells: it acts on signal transduction especially on TGFBs paracrine network; it inhibits apoptosis and thus increases cell proliferation; it decreases cell adhesion with a consequent enhancement of cell mobility and migration; finally, it acts on bone marrow stem cells (ie, CD34). In conclusion, PG acts on bone formation by stimulating both osteoconduction (as demonstrated by the reduction in cell adhesion) and ostegenesis (as shown by TGFB-related proteins and stem cell markers).

It is worth noting that MG63 is a cell line and not normal osteoblasts. One of the advantages of using a cell line is related to the fact that the reproducibility of the data is higher because there is not the variability of the patient studied. Primary cell cultures provide a source of normal cells but they also contain contaminating cells of different types and cells in variable differentiation states. Moreover, we have chosen to perform the experiment after 24 hours to get information on the early stages of stimulation. It is our knowledge, therefore, that more investigations, with other osteoblast-like cell lines, primary cultures, and different time points, are needed to get a global comprehension of the molecular events related to PG action. Finally, we believe the reported data can be a model to compare different substances with similar effects.

This work was supported by grants from University of Ferrara, Italy (F.C.) and Fondazione CARISBO (F.P.).