Journal of Dental Research

It is widely accepted that volumetric contraction and solidification during the polymerization process of restorative composites in combination with bonding to the hard tissue result in stress transfer and inward deformation of the cavity walls of the restored tooth. Deformation of the walls decreases the size of the cavity during the filling process. This fact has a profound influence on the assumption-raised and discussed in this paper-that an incremental filling technique reduces the stress effect of composite shrinkage on the tooth. Developing stress fields for different incremental filling techniques are simulated in a numerical analysis. The analysis shows that, in a restoration with a well-established bond to the tooth-as is generally desired-incremental filling techniques increase the deformation of the restored tooth. The increase is caused by the incremental deformation of the preparation, which effectively decreases the total amount of composite needed to fill the cavity. This leads to a higher-stressed tooth-composite structure. The study also shows that the assessment of intercuspal distance measurements as well as simplifications based on generalization of the shrinkage stress state cannot be sufficient to characterize the effect of polymerization shrinkage in a tooth-restoration complex. Incremental filling methods may need to be retained for reasons such as densification, adaptation, thoroughness of cure, and bond formation. However, it is very difficult to prove that incrementalization needs to be retained because of the abatement of shrinkage effects.

The resin-dentin interface formed by two dentin adhesives, Optibond (OPTI, Kerr) and Scotchbond MultiPurpose (SBMP, 3M), was ultramorphologically examined by transmission electron microscopy (TEM). Ultrastructural information from nondemineralized and demineralized sections was correlated. It was hypothesized that the different chemical formulations of the two adhesives would result in a different morphological appearance of the hybrid layer. Ultrastructural TEM examination proved that each of the two dentin adhesive systems was able to establish a micromechanical bond between dentin and resin with the formation of a hybrid layer. However, the interfacial hybridization process that took place to produce this resin-dentin bond appeared to be specifically related to the chemical composition and application modes of both systems. OPTI consistently presented with a hybrid layer with a relatively uniform ultrastructure, electron density, and acid resistance. These three parameters were found to be more variable for the hybrid layer formed by SBMP. Characteristic of SBMP was the identification of an amorphous phase deposited at the outer surface of the hybrid layer. Although both adhesive systems investigated follow a total-etch concept, their specific chemical formulations result in different interfacial ultrastructures that are probably related to different underlying bonding mechanisms. The clinical significance of these morphological findings, however, is still unknown.

Alkaline phosphatase (ALP) activity is involved in the process of calcification in various mineralizing tissues, and it is found at much higher levels in the periodontal ligament (PDL) than in other connective tissues. Since the PDL lies between hard tissues and functions as a cushion mitigating mechanical stress, such as occlusal and orthodontic forces, this stress may modulate ALP activity in PDL cells, which themselves may affect adjacent alveolar bone metabolism. The objective of this study was to determine the level of ALP activity and the gene expression of liver/bone/kidney (L/B/K) ALP in human PDL fibroblasts in response to cyclic tension-forces. Human PDL cells were cultured on flexible-bottomed plates and placed on a Flexercell Strain Unit. Cells were flexed at 6 cycles/min (5 sec strain, 5 sec relaxation) at 6 levels of tension-force (9%, 12%, 15%, 18%, 21%, and 24% increase in surface area) for 5 days. There was no significant difference in cell proliferation between the cells subjected to the tension-force and the controls. There was a 10% and 42% decrease, respectively, in the ALP activity in PDL cells exposed to low (9%) and high (24%) tension-forces, and these decreases were dependent on the magnitude of the tension-force. The finding of inhibited ALP activity in response to tension-force was consistent with the observation that L/B/K ALP mRNA levels were decreased in response to cyclic tension-force. These results suggest that tension-force may affect PDL metabolism, depending on the functional role of ALP.

In periodontal surgery, healing after Guided Tissue Regeneration (GTR) may be explained by differences in functional activities of gingival and periodontal ligament fibroblasts (GF and PDLF). Several studies in vitro have supported this hypothesis, but much remains to be defined. In the present work, gingival and periodontal ligament fibroblasts derived from five healthy subjects were isolated and compared in vitro. The morphology of the cells was observed under scanning electron microscopy (SEM). Several extracellular matrix components (ECM) were studied to compare the effects on fibroblast attachment, proliferation, and protein synthesis. Several biochemical markers were examined in both cellular extract (CE) and conditioned medium (CM). We also examined the muscle differentiation markers alpha-smooth muscle actin, desmin, and smooth-muscle myosin. Finally, we studied the effects of epithelial cells on the proliferation and protein synthesis of the two types of fibroblasts. GF and PDLF appeared identical under the SEM. All ECM components enhanced attachment; however, while collagen types I and IV promoted the attachment of GF, gelatin, laminin, and vitronectin promoted that of PDLF. Most ECM components increased the proliferation rate of GF and the biosynthetic activity of PDLF. The biochemical markers were similarly distributed between the two cell types, except for alkaline phosphatase, which was detected only in the CE of PDLF. Both GF and PDLF strongly expressed alpha-smooth-muscle actin and were negative for desmin; only PDLF were positive for smooth-muscle myosin. Epithelial cells increased the proliferation of both GF and PDLF but had no effect on their biosynthetic activity. These in vitro results may better explain the in vivo functional differences between GF and PDLF.

A variety of pharmaceutical agents has been proposed for use in periodontal therapy to inhibit loss of alveolar bone and to promote regeneration of tissues lost to disease. The purpose of this study was to determine the effects of such agents on periodontal cell-mediated gel contraction, an in vitro process considered representative of wound contraction and remodeling in vivo. Human gingival fibroblasts were cultured in a type I collagen lattice, and contraction was quantified by means of a computer-assisted video imaging system. Cell-gel combinations were prepared with cells both pre-exposed and non-exposed to agents; non-anchored cell-gels were then incubated with agents for various time periods. Agents tested included Demecolcine (an inhibitor of cytoskeletal contraction), growth factors (i.e., TGF-β1, PDGF, and IGF-1), and non-steroidal anti-inflammatory drugs (NSAIDs) (indomethacin, ibuprofen, naproxen, and flurbiprofen). While Demecolcine inhibited gel contraction, TGF-β1 (1-20 ng/mL), PDGF (100 ng/mL), IGF-1 (1000 ng/mL), and [PDGF + IGF], all of which have been reported to enhance wound healing in vivo, promoted lattice contraction in this system. In contrast, NSAIDs inhibited cell-gel contraction. Ethanol, used to solubilize two specific NSAIDs, also inhibited cell proliferation and gel contractile ability, even at very low concentrations. These findings indicate that periodontal cells respond differently to various agents in vitro and may be adversely affected by alcohol. Furthermore, the results of this study suggest that the cell-lattice contraction system holds potential as a method for screening agents for positive or negative effects on cell activity.

Re-epithelialization involves interactions between keratinocytes and the extracellular matrix upon which these cells move. It is hypothesized that keratinocytes are activated when wounded, and the resultant phenotypic change directs re-epithelialization. We have adapted organotypic cultures, in which oral gingival keratinocytes are fully differentiated, to study re-epithelialization following wounding. To elucidate keratinocyte behavior and phenotype during re-epithelialization, we have investigated this process in the presence and absence of the growth factor TGF-β1 and have monitored expression of MMP-1 (Type I collagenase) mRNA by in situ hybridization. In addition, we have followed proliferation and migration of wound keratinocytes by genetically marking these cells with a retroviral vector and by measuring their proliferative index. We found that keratinocytes grown without TGF-β1 were hyperproliferative in response to wounding, and re-epithelialization was complete by 24 h. However, 2.5 ng/mL TGF-β1 induced a transient delay in re-epithelialization, a reduction in proliferation, and fewer clusters of genetically marked cells. Keratinocytes expressed MMP-1 mRNA only when they covered the wounded surface, suggesting that the cells acquire a collagenolytic phenotype during re-epithelializaation and that contact with different ECM components may modulate keratinocyte expression of MMP-1. We conclude that the phenotype of oral keratinocytes is altered during re-epithelialization in vitro and that this process is modulated by TGF-pl. Reepithelialization occurs as keratinocytes are activated to move over the wound bed. Understanding the phenotype of wounded keratinocytes may facilitate treatment of chronic oral wounds and periodontal disease.

Matrix metalloproteinases (MMPs) have an important role in many biological processes, such as tumor metastasis, wound healing, and inflammation. The regulation of MMPs and their inhibitors is still not known in detail, and the aim of this study was to investigate the effects of dexamethasone on cultured oral benign and malignant cell lines. The expression of MMPs in culture was studied: in four gingival (GF) and one periodontal ligament (PLF) fibroblast cell lines; in six gingival keratinocyte (GK) cell lines; and in UNR (UNR-108, Rat Osteogenic sarcoma) and SCC (SCC-25, Human Tonque Squamous Cell Carcinoma) cell lines. In the GFs, PLFs, and UNR cells, only MMP-2 (72 kDa gelatinase) was detected by gelatin zymography, while in most of the GK cell lines only MMP-9 (92 kDa gelatinase) was observed. In confluent SCC cultures, both MMP-2 and MMP-9 were found, while only MMP-2 was seen in rapidly growing SCC cells, demonstrating that cell proliferation influenced gelatinase expression in these cells, but not in the other cell lines studied. Dexamethasone at concentrations of 10-5 mol/L and 10-7 mol/L decreased the production of gelatinases in the GFs and PLFs, but not in the GKs, SCC, or UNR cells. The expression of mRNAs for matrix metalloproteinases (MMP-1 [interstitial collagenase] and MMP-2) and their inhibitors (TIMP-1 and TIMP-2) was also studied in the GFs by Northern hybridization. Dexamethasone markedly decreased the amount of MMP-2 mRNA in the GFs. The mRNA level of MMP-1 decreased even more in the same GFs. The mRNA levels for TIMP-1 and TIMP-2 were also decreased by dexamethasone in the GFs. Cell proliferation influenced the degree to which dexamethasone decreased these mRNA levels. The results indicate that glucocorticoids decrease the levels of MMPs and TIMPs in oral fibroblastic cells, whereas they do not appear to affect the production of gelatinases in either normal or malignant oral epithelial cell lines.

The preparation of a superantigenic fraction F-2 from the culture supernatant of Streptococcus mitis 108, a fresh isolate from human tooth surfaces, was reported previously. Now, to determine the possible pathogenic role of the superantigen in oral mucosal diseases, we examined the cytotoxic effects of human peripheral blood T-cells activated with F-2 on human oral epithelial cells. T-cells activated with F-2 were cytotoxic to the human squamous carcinoma HO-l-N-1 cells derived from the oral mucosa, similar to those activated with Staphylococcus aureus enterotoxin B (SEB). This cytotoxic effect was increased in a dose-dependent manner by the addition of the respective stimulant, F-2 or SEB, to the cytotoxic assay system. F-2 endowed mainly CD8+ T-cells with cytotoxic activity. Pretreatment with human interferon gamma increased the sensitivity of the HO-l-N-1 cells to the cytotoxic effects of F-2-activated T-cells. The F-2-activated T-cells were also cytotoxic to human keratinocytes derived from gingiva. There was no correlation between the degree of cytotoxicity and the levels of tumor necrosis factor alpha in co-cultures of F-2-activated T-cells and HO-l-N-1 cells. A double-chamber plate experiment revealed no cytotoxic effects when the F-2-activated T-cells were separated from the HO-l-N-1 cells. Supernatants of the co-cultures of target and effector cells were not cytotoxic to HO-1-N-1 cells. These findings suggest that the cytotoxic effects of the F-2-activated T-cells on HO-l-N-1 cells were mediated not by soluble factors but by the direct interaction between the activated T-cells and the target cells. The cytotoxicity of the F-2-activated T-cells against HO-l-N-1 cells was markedly inhibited by monoclonal antibodies (MAbs) against CDlla and CD54, but was only slightly inhibited by MAbs against human leukocyte antigen (HLA)-DR and CD2. Thus, the interaction between lymphocyte-function-associated antigen-1 (LFA-1) and intercellular adhesion molecule-1 (ICAM-1) was crucial for the F-2-dependent T-cell-mediated cytotoxicity against oral epithelial cells, while HLA-DR and CD2 molecules are not necessarily involved in the cytotoxicity observed.

Although a functional role for serotonin (5-hydroxytryptamine, 5-HT) has been defined in the salivary glands of some lower species, relatively few data supporting a role for 5-HT in the regulation of mammalian salivary glands have been presented. Our initial results from polymerase chain reaction studies in cells of mammalian submandibular gland origin using consensus sequence primers from G protein-coupled receptors suggested the presence of mRNA for a 5-HT receptor in these cells. Based on this observation, the question of a role for 5-HT in mammalian submandibular gland function was re-addressed, using isolated, perfused rat submandibular glands and dispersed-cell aggregates from this gland. In perfused glands, 5-HT decreased the rate of saliva flow initiated by acetylcholine by about 50% and increased the amount of protein in the saliva two-fold. In dispersed-cell aggregates, 5-HT elicited a concentration-dependent increase in the accumulation of adenosine 3',5' monophosphate (cyclic AMP; EC₅₀ = 660 ± 110 nM). In addition, functional studies, as well as radioligand binding experiments, indicated that the effects of 5-HT are independent of beta-adrenoceptors. Accumulation of cAMP in gland cells was consistent with a direct action of 5-HT on adenylyl cyclase. Similar cyclic AMP responses to 5-HT were observed in cells isolated from mouse and opossum submandibular glands and rat sublingual and parotid glands. Our findings suggest the presence of a 5-HT receptor in mammalian salivary glands coupled to the stimulation of adenylyl cyclase and, at least in rat submandibular gland, involved in modifying the volume and protein content of saliva.

Since lingual nerves may be transected during a variety of oral surgical procedures, including third molar removal, we have investigated two possible methods of repair. Quantitative morphological observations were made on feline chorda tympani and lingual nerves proximal and distal to transection injuries repaired either by epineurial suturing or by insertion of the cut ends into a perforated silicon tube. Proximal to the repair, the most prominent difference was an increase in the number of myelinated axons in the lingual nerve following epineurial suturing but not entubulation. Proximal to the repair site, the number of non-myelinated axons increased in comparison with controls in both chorda tympani and lingual nerves after both procedures, though the difference was statistically significant only in the lingual nerve proximal to entubulation. Distal to the injury, both types of repair showed a reduction in the number, size, and sheath thickness of myelinated axons in comparison with unoperated controls, but the difference in numbers was statistically signficant only distal to repair by entubulation. The number of non-myelinated axons distal to the repair sites was much higher than that in controls, the difference being greater distal to entubulation repair. There were more axons per Remak bundle distal to entubulation repair than to epineurial suturing, suggesting, perhaps, that fewer axons would ultimately become myelinated. Though the morphological differences between the two repair techniques are not as striking as the parallel electrophysiological differences reported previously (Smith and Robinson, 1995a,b), they are consistent with them and support the conclusion that, for transected lingual and chorda tympani nerves, epineurial suturing is the preferred approach. Key words: lingual nerve, injury, oral surgery, taste.

Root-surface caries, like enamel caries, develops as a subsurface type of mineral loss. Very little is known about the composition of the surface zone covering the body of the lesion, and the ultrastructure and composition of carious cementum are not known. The aim of this study was to correlate the ultrastructure and arrangement of the cementum crystals with the distribution of fluoride and calcium in root cementum from human teeth with sound, unexposed, or exposed root surfaces as well as in early stages of root-surface caries. Microradiographically, unexposed specimens showed a relatively homogeneous mineral distribution contrasting with the formation of an apparently highly mineralized surface layer in exposed and, in particular, in carious cementum. The electron-probe findings showed a substantial fluoride peak corresponding to the surface layers in carious tissues in particular, whereas the calcium profile in the surface did not reflect the apparent increase in mineralization. A substantial increase in size of the cementum crystals was found in specimens with formation of the fluoride-rich, well-mineralized surface zone. The crystal lattice intervals when observed along the (001) plane showed a hydroxyapatite spacing. The findings indicated that a significant crystal growth can be achieved in human cementum concomitant with fluoride accumulation.