bulgaricus (1% viability) resulted in degradation of proteins Src inhibitor and peptides and such degraded proteins, if exposed on the bacterial cell wall, may be the cause of the increased cytokine production. Taking into account the bacterial viability

after lyophilization, this works out to a 6 : 1 ratio of live bacteria to splenocytes. Baba et al. (2008) found that a low bacterial to dendritic cell (DC) ratio results in a reduction of cytokine (IL-10, IL-12p70 and TNFα) production. Thus, the increase in cytokine production reported here is probably due to the dead bacteria. The ability of L. casei to induce IL-12p40 increased by more than threefold, IL-10 by 10-fold and TNFα by 2.4-fold (Table 1) (P<0.001). Lactobacillus bulgaricus and L. rhamnosus induced significantly more IL-10 (1.5- and 3.8-fold, respectively) (P<0.001) after being lyophilized, but there was no change in TNFα or IL-12p40 production (Table 1). Lyophilization changed the order of cytokine induction by these bacteria such that Doxorubicin in vivo for TNFα: L. bulgaricus>L. casei>L. rhamnosus; for IL-12p40: L. bulgaricus=L. casei>L. rhamnosus; and for IL-10: L. bulgaricus>L. rhamnosus>L. casei.

To determine whether the cytoplasmic components or the cell wall architecture disruption are the cause of the increased cytokine secretion, we carried out contact inhibition experiments. In the presence of the membrane inserts, the production of TNFα and IL-10 (by both live and lyophilized lactobacilli) was abrogated and drastically reduced, respectively (Fig. 2a and b) (P<0.001), indicating that direct contact between lactobacilli and spleen cells was important for cytokine induction. This reflects Carbohydrate the necessity for the engagement of membrane receptors and/or phagocytosis. The low level of IL-10 production was probably due to soluble bacterial products as in the presence of the membranes, there was still significantly more IL-10 than in the

media alone (P<0.05). The roles of TLRs in lactobacilli stimulation of splenocytes were evaluated using TLR-blocking antibodies or oligonucleotides. As both TLR1 and TLR6 require an association with TLR2 for activation, blocking TLR2 will effectively block interactions with either of these receptors as well; thus, we used anti-TLR2 antibodies. The anti-TLR2 antibody had a negligible effect on L. casei-induced TNFα production, while there was a 20% and 60% reduction in TNFα production by L. bulgaricus and L. rhamnosus, respectively (Fig. 3a). Lactobacillus rhamnosus-stimulated IL-10 secretion was abrogated after TLR2 blocking (by 80%), while IL-10 induction by L. bulgaricus was reduced by 30% (Fig. 3b). IL-12 production was independent of TLR2 (Fig. 3c). When spleen cells were treated with anti-TLR4 antibody or anti-TLR9 oligonucleotides, the production of all three cytokines remained unchanged, indicating that TLR4 and TLR9 had little influence on the induction of these cytokines by lactobacilli (data not shown).