Glucosyltransferases of viridans streptococci are modulins of interleukin-6 induction in infective endocarditis

Abstract

The glucosyltransferases (GTFs) of viridans streptococci, common pathogens of infective endocarditis, are extracellular proteins that convert sucrose into exopolysaccharides and glucans. GTFs B, C, and D of Streptococcus mutans are modulins that induce, in vitro and in vivo, the production of cytokines, in particular interleukin-6 (IL-6), from monocytes. The roles of S. mutans GTFs in infectivity and inflammation in situ were tested in a rat experimental model of endocarditis. No significant differences in infectivity, in terms of 95% infective dose and densities of bacteria inside vegetations, were observed between laboratory strain GS-5 and two clinical isolates or isogenic mutant NHS1DD, defective in the expression of GTFs. In aortic valves and surrounding tissues, IL-6 was detected by Western blots and immunostaining 24 h after GS-5 infection, was maintained over 72 h, and was followed by production of tumor necrosis factor alpha but not IL-1beta. Animals infected with NHS1DD showed markedly lower levels of IL-6 (less than 5% of that of parental GS-5-infected rats), while tumor necrosis factor alpha was unaffected. In contrast, animals infected with NHR1DD, another isogenic mutant expressing only GtfB, showed a much smaller reduction (down to 56%). These results suggest that GTFs are specific modulins that act during acute inflammation, inducing IL-6 from endothelial cells surrounding the infected valves without affecting bacterial colonization in vegetations, and that IL-6 might persist in chronic inflammation in endocarditis.

FIG. 1.

Southern and Western blot analysis of the parental GS-5 and isogenic mutants NHR1DD and NHS1DD. (A) Hybridization patterns of chromosomal DNA digested with EcoRI (lane 1) or PstI (lane 2) with [α-³²P]dCTP (Amersham Pharmacia)-labeled DNA probes recognizing both gtfB and gtfC. The deletion of gtfB/C resulted in the absence of two EcoRI-digested bands (7.3 and 4.6 kb) and two PstI-digested bands (7.3 and 6.4) in NHS1DD. The deletion of gtfC resulted in the absence of a 7.3-kb EcoRI fragment and a 7.3 PstI fragment in NHR1DD. (B) Hybridization patterns of chromosomal DNA digested with HindIII (lane 1) or HpaI (lane 2) with [α-³²P]dCTP (Amersham Pharmacia)-labeled DNA probes specifically recognized gtfD. Insertional mutation in the gtfD with a tetracycline resistance cassette resulted in an increase (2.0 kb) in size of HindIII- or HpaI-digested fragments in both NHS1DD and NHR1DD. (C) Protein profiles of from GS-5 (lane 1), NHR1DD (lane 2), or NHS1DD (lane 3). Total proteins were resolved on a 7.5% polyacrylamide gel stained with silver. (D) Western blot analysis of GTFs. Total proteins from GS-5 (lanes 1 and 2), NHR1DD (lane 3), or NHS1DD (lane 4) were probed with purified anti-GtfB/C/D rabbit IgG (PJS-2). No GTFs were found in NHS1DD, and only GtfB was expressed on NHR1DD. Molecular size in kb or kDa is depicted in the left border of panels A, B, and C.

FIG. 2.

Vegetation formation in rat experimental endocarditis. The correct placement of the inserted transvalvular catheter through aortic valves was confirmed by X radiograph (upper panel). A typical vegetation, mostly small with warty nodules, could be visualized 24 h after intravenous challenge with S. mutans GS-5 at an inoculum higher than 10⁷ CFU (lower panel). Normal aortic valves are shown in the middle panel.

FIG. 3.

Infectivity and virulence of GS-5 and NHS1DD in rat experimental endocarditis. (A and B) Weight of vegetation and infectivity titration. Each point represents one rat and a dash represents the mean weight of vegetation formed. (C and D) Bacterial densities in vegetations. Each point represents one rat and a dash represents the mean bacterial density. The ID₉₅ for GS-5 and NHS1DD was similar, 10⁷ CFU. (E and F) Bacterial densities in the spleens. Each point represents one rat, and the bars represent the mean bacterial densities. No statistically significant difference was found in vegetation weights or bacterial densities recovered from either the vegetations or spleens between GS-5- and NHS1DD-infected rats at all inocula except at 10⁶ in the spleen. Diamonds in A, C, and E, GS-5; circles in B, D, and F, NHS1DD; bar, mean for GS-5 or NHS1DD.

FIG. 4.

Detection of IL-6 in situ in frozen sections of noninfected valves or valves infected with different strains. (A) Normal aortic valve (hematoxylin and eosin, 100×). (B) Immunostaining for IL-6 secretion in catheter-injured valve, uninfected. (C) NHS1DD-infected vegetations and perivalvular regions. (D) GS-5-infected aortic valve. (ABC method; B to D, 100×). Positive IL-6 secretion in endothelial linings.

FIG. 5.

Detection of IL-6 in situ in frozen section of spleens infected with GS-5 (A) or NHS1DD (B). Fewer IL-6-secreting neutrophils or macrophages were detected in disseminated NHS1DD infection after prolonged staining, as indicated by the higher background (B). (ABC method, 100×).

FIG. 6.

Western blot analysis of infected valves and surrounding tissues. Aortic valves and surrounding tissues from three different animals were dissected under a stereoscope and prepared as described in Materials and Methods. β-Actin was used as an internal control to normalize the total amount of proteins loaded in each lane. IV, catheter-injured valve; NV, normal valve.