Joke S. van de Gevel

The origin of osteoclasts was studied in an in vitro model using organ cultures of periosteum-free embryonic mouse long-bone primordia, which were co-cultured with various cell populations. The bone rudiments were freed of their periosteum-perichondrium by collagenase treatment in a stage before cartilage erosion and osteoclast formation, and co-cultured for 7 d with either embryonic liver or mononuclear phagocytes from various sources. Light and electron microscopic examination of the cultures showed that mineralized matrix-resorbing osteoclasts developed only in bones co-cultured with embryonic liver or with cultured bone marrow mononuclear phagocytes but not when co-cultured with blood monocytes or resident or exudate peritoneal macrophages. Osteoclasts developed from the weakly adherent, but not from the strongly adherent cells of bone marrow cultures, whereas 1,000 rad irradiation destroyed the capacity of such cultures to form osteoclasts. In bone cultures to which no other cells were added, osteoclasts were virtually absent. Bone-resorbing activity of in vitro formed osteoclasts was demonstrated by 45Ca release studies. These studies demonstrate that osteoclasts develop from cells present in cultures of proliferating mononuclear phagocytes and that, at least in our system, monocytes and macrophages are unable to form osteoclasts. The most likely candidates for osteoclast precursor cells seem to be monoblasts and promonocytes.

The consequences of internalization of Staphylococcus aureus by HUVEC with respect to their adhesiveness for human monocytes and granulocytes were investigated. Viable and UV-killed, but not heat-killed, S. aureus were internalized by HUVEC, which required participation of the endothelial cytoskeleton. S. aureus-infected HUVEC displayed increased surface expression of CD106 (VCAM-1), CD54 (ICAM-1), and MHC I molecules. Expression of CD62P (P-selectin), CD62E (E-selectin), CD31 (PECAM-1), and CD102 (ICAM-2) was not affected. Concomitantly, these HUVEC expressed a time- and inoculum size-dependent hyperadhesiveness for monocytes and granulocytes. Monocyte adhesion reached maximal levels (approximately 60% adhesion) 23 h after the initial 1 h period of infection of HUVEC with about 50 bacteria per single HUVEC. To induce maximal (approximately 20%) adhesion of granulocytes, five times higher concentrations of HUVEC-infecting bacteria were required. Using the appropriate mAb, granulocyte adhesion to S. aureus-infected HUVEC was shown to be entirely mediated by the beta2 (CD11/CD18) integrins. Monocyte adhesion to these HUVEC was largely (approximately 70%) dependent on both CD11a/CD18 (LFA-1) and CD49d/CD29 (VLA-4). This demonstrates that infection of HUVEC with S. aureus potentiates CD11/CD18-mediated granulocyte adhesion and shifts the mechanism of monocyte adhesion from being completely CD11/CD18 dependent to one that also utilizes the VLA-4/VCAM-1 dependent pathway. Together, these findings indicate that in response to internalization of S. aureus, vascular endothelial cells may initiate recruitment of monocytes and granulocytes, which may be an important initial event in the pathogenesis of endovascular diseases.

Surface molecules of Staphylococcus aureus are involved in the colonization of vascular endothelium which is a crucial primary event in the pathogenesis of infective endocarditis (IE). The ability of these molecules to also launch endothelial procoagulant and proinflammatory responses, which characterize IE, is not known. In the present study we investigated the individual capacities of three prominent S. aureus surface molecules; fibronectin-binding protein A (FnBPA) and B (FnBPB) and clumping factor A (ClfA), to promote bacterial adherence to cultured human endothelial cells (ECs) and to activate phenotypic and functional changes in these ECs. Non-invasive surrogate bacterium Lactococcus lactis, which, by gene transfer, expressed staphylococcal FnBPA, FnBPB or ClfA molecules were used. Infection of ECs increased 50- to 100-fold with FnBPA- or FnBPB-positive recombinant lactococci. This coincided with EC activation, interleukin-8 secretion and surface expression of ICAM-1 and VCAM-1 and concomitant monocyte adhesion. Infection with ClfA-positive lactococci did not activate EC. FnBPA-positive L. lactis also induced a prominent tissue factor-dependent endothelial coagulation response that was intensified by cell-bound monocytes. Thus S. aureus FnBPs, but not ClfA, confer invasiveness and pathogenicity to non-pathogenic L. lactis organisms indicating that bacterium-EC interactions mediated by these adhesins are sufficient to evoke inflammation as well as procoagulant activity at infected endovascular sites.

Supernatants of human peripheral blood mononuclear cells cultured in the presence of B. gingivalis, showed a strong osteoclast stimulating activity as measured by 45Ca release from fetal mouse long bones in vitro. These supernatants also contained a high concentration of bioactive and immunoreactive interleukin-1 (IL-1), but tumor necrosis factor (TNFa), another osteoclast-activating cytokine, was not detected. Osteoclast activation by the supernatants was inhibited by an antibody against IL-1, whereas ultrapure human IL-1 mimicked the effect of the supernatant. The ability of B. gingivalis to induce IL-1 and OAF production was heat sensitive, as 20 min heating of the bacteria at 120 degrees C caused a 50% loss of activity. In addition, purified B. gingivalis lipopolysaccharide (LPS) had little IL-1 inducing capacity, compared with LPS of Escherichia coli. These data suggest that human peripheral blood cells confronted with B. gingivalis produce large amounts of IL-1 which has strong osteoclast stimulating activity. However, in contrast with E. coli LPS, B. gingivalis LPS does not seem to be the major inducing agent. Thus other bacterial components must be responsible for the observed IL-1 and OAF induction.