The increasing interest in using MSCs as a source of cellular therapy means that the potential impact of bacterial toxins on the growth and differentiation of MSCs will be a growing concern. Exposure to endotoxin-enriched environment may affect many aspects of MSCs properties such as self-renewal, differentiation potential, production of cytokines & ECM compound. Our study focused on the investigation of the effects of two purified bacteria-derived toxins on the growth and osteogenic differentiation of human MSCs. Cell proliferation was not affected by LPS or LTA, and osteogenic differentiation was promoted by prolonged exposure to LPS but not to LTA. In addition, brief exposure to LPS during osteogenic differentiation reduced the expression of TLR4, whereas prolonged exposure reduced the expression of both TLR4 and TLR2.
Most previous related studies have been conducted on osteoclasts [35–38], than osteoblasts [14–16] or their committed progenitors [17–19]. Among the studies related to the bacterial effects on the osteogenic progenitors, they used either a 6-day chick periosteal osteogenesis model or rat calvaria cells, which contain a subpopulation of osteogenic precursors. The objectives of these studies were to find out the possible adverse effects of bacterial toxin on osteogenesis and they did show bacterial toxin impaired the osteogensis process. We studied the effect of bacterial toxins on osteogenic differentiation using human MSCs from bone marrow aspirate, which is an enriched source of multipotential stem cells. In addition, it is a novel model comparing to the chick or rat models for it allowed us to study how bacterial toxins might affect osteogenic differentiation from MSCs to osteogenic progenitors.
Our study showed that neither LPS nor LTA affected the proliferation of MSCs, even at relative high endotoxin doses. This is in line with the findings of previous study by Hwa Cho et al . We then explored whether these responses pattern may be related to the expression levels of TLR4 and TLR2 genes in MSCs. This hypothesis was partly supported by our observation that very low level of TLR4 nor TLR2 mRNAs expression levels were detectable in MSCs. These low TLR4 and TLR2 expression profiles may serve as a protective mechanism to maintain stem cell survival from the effect of bacterial toxins, even at high concentrations. For cellular therapy purpose, this suggested that MSCs could be safely used in clinical settings even in the microenvironment with bacteria such as oral surgical sites.
TLR4 and TLR2 expressions in MSC-derived osteoprogenitors were minimal during early osteogenic induction but emerged by Day 12. The expression of these two genes seems to be a natural part of osteoblast differentiation in the absence of external bacterial stimuli. This finding agrees with those of previous studies in which osteoblastic cells and osteoblasts constitutively express TLR4, TLR2, and other LPS-signaling molecules (Table 1). This is slightly different from what was observed previously which showed TLR-2, 3, 4, 6 had a higher expression as compared to TLR-1, 5, 9, but such differences may be related to different type of methodology being adopted . Another study on murine BM-derived MSCs also showed TLR1 to TLR8 but not TLR9 mRNA was expressed . Whether such variations in TLRs expression profile has something to do with the state of differentiation require further investigation. Although previous studies examined the basal TLR expression level in MSCs and how they affect MSC differentiation as we did, our study provided information on the real time dynamic changes in the TLR2 and TLR4 mRNA expression profile as they underwent osteogenic differentiation from MSCs under the influence of LPS or LTA. Interestingly, LPS but not LTA could also downregulate the TLR2 and 4 expressions on Day 12 of differentiation, no matter it was under either short or prolonged LPS exposure. Our findings suggested that the TLR2 and 4 expressions could be negatively regulated by endotoxin and such phenomenon has been described in some immune cells such as monocytes.
When testing whether LPS and LTA affect the osteogenic differentiation of human MSCs, we found that short-term challenge by either toxin had no effect, but prolonged LPS challenge upregulated both ALP activity and calcium deposition. This is in line with the findings of Hwa Cho et al. who found an increase in the osteogenic marker genes expression such as the ALP, osteopontin and BMP2 at the 5th and 10th day after induction . We confirmed such findings by measuring the ALP protein level on Day 10.
Documented effects of E. coli LPS on osteoblastic cells are mixed. No inhibitory effect of E. coli LPS were found on osteoblastic cell line MC3T3-E1 and rat calvaria cells [17, 39]. However, Shjoi et al. found an inhibitory effect on ALP activity of SaOS-2, an osteoblast-like osteosarcoma cell line . Factors contributing to discrepancies in results may include heterogeneity of the cell source, the experimental conditions, and the culture system used. For TLR2 ligands, it has been shown that PGN stimulated the osteogenic differentiation of human adipose-derived MSCs in a dose-dependent manner similar to that of LPS . However, no such effect was found concerning LTA used in our study on BM-derived MSCs. Why LPS and LTA affect the MSC differentiation differently in our study remains to be answered, but the specificity and potency of different ligands on the TLRs and also the involvement of different signaling pathways may account for such variations [20, 30–32].
In this study, the effects of LPS exposure during MSC osteogenic differentiation depended on the duration of exposure. This finding might be explained partly by the kinetics of TLR4 gene expression: the lack of an effect from short-term LPS challenge might be due to the delay in expression of TLR4. Therefore, the timing of the activation of TLR genes may play an essential role in altering the osteogenic activity of MSC-derived osteoprogenitors. This finding may explain why studies using osteoblasts and osteogenic precursor cells as effector cells can yield different results after bacterial challenge.
On the other hand, since osteogenic differentiation of MSCs was promoted by prolonged LPS challenge, we further examined whether constitutive TLR gene expression on osteoprogenitors was also affected by LPS. Interestingly, we found that expression of both TLR4 and TLR2 genes was downregulated in osteoprogenitors after continuous LPS exposure. Little has been reported on how bacterial toxins regulate the expression of TLRs on osteoblastic cells. One study  showed an immediate effect of LPS challenge on TLR2 and TLR4 expression of primary murine osteoblasts and MC3T3-E1 cells; unexpectedly, TLR2 but not TLR4 mRNA was upregulated within two hours of E. coli LPS exposure. This raised a question about the specificity of TLRs and some degree of crosstalk between TLR2 and TLR4 may actually exist. In our experiments, prolonged toxin exposure was designed to mimic the situation in which MSCs interact continuously with oral microflora or with bacterial toxins during chronic inflammatory bone disease. Our results suggested that MSC-derived osteoprogenitors could adapt to continuous LPS challenge by reducing TLR4 and TLR2 expression, thereby they can be spared from the toxic effect of these toxins, leading to a paradoxical upregulation of osteogenic activities.
A similar adaptive response phenomenon, known as "endotoxin tolerance," can be found in monocytes and macrophages . Endotoxin tolerance has been well documented as a cell-desensitizing phenomenon that results from sustained exposure of sublethal doses of LPS, which leads to a reduced capacity of the host (in vivo) or of macrophages (in vitro) to further respond to LPS . Furthermore, TLR4 expression on the surface of LPS-tolerant macrophages has been shown to be downregulated , which may account for the molecular mechanism of endotoxin tolerance. Therefore, we suggest that MSC-derived osteoprogenitors acquire an adaptive tolerance under continuous LPS exposure in order to prevent excessive bone and tissue destruction, thereby protecting or preserving the host organ tissues.
Endotoxin tolerance can be found in oral mucosa cells and is associated with chronic periodontitis . In addition, TLR2 and TLR4 mRNAs are significantly downregulated in the gingival tissue of patients with chronic periodontitis as compared to healthy persons , suggesting that the oral mucosa develops endotoxin tolerance in chronic periodontitis. Although experimental endotoxin tolerance involves prior incubation of LPS before subsequent challenge, a study of human gingival fibroblasts-a major constituent of gingival connective tissue that interacts directly with bacteria in periodontitis-has shown that TLR4 expression can be downregulated by LPS without prior LPS incubation . Whether osteoprogenitors or osteoblasts indeed display endotoxin tolerance remains to be confirmed. Interestingly, we also observed downregulation of TLR4 gene expression after osteogenic induction in MSCs that had undergone short-term LPS challenge. These data may suggest that gene might be switched off in the early phase of osteogenic differentiation. Finally, the paradoxical puzzle of inhibition of the TLR receptor expression by LPS was not contradictory to the positive effect on osteogenic differentiation remained to be solved. Possible explanation may include the presence of alternated receptor and signaling pathway for LPS induced osteogenesis.