TLR2-mediated cell stimulation in bacterial vaginosis

Abstract

Bacterial vaginosis (BV) is associated with preterm labor, pelvic inflammatory disease (PID) and increased HIV acquisition, although the pathways that mediate these pathological effects have not been elucidated. To determine the presence of Toll-like receptor (TLR)-ligands and their specificity in BV, genital tract fluids were collected from women with and without BV by cervicovaginal lavage (CVL). The CVL samples were evaluated for their ability to stimulate secretion of proinflammatory cytokines and to activate NFkappaB and the HIV long terminal repeat (LTR), indicators of TLR activation, in human monocytic cells. Stimulation with BV CVLs induced higher levels of IL-8 and TNFalpha secretion, as well as higher levels of HIV LTR and NFkappaB activation, than CVLs from women with normal healthy bacterial flora. To identify which TLRs were important in BV, 293 cells expressing specific TLRs were exposed to CVL samples. BV CVLs induced higher IL-8 secretion by cells expressing TLR2 than CVLs from women without BV. Surprisingly, BV CVLs did not stimulate cells expressing TLR4/MD2, although these cells responded to purified lipopolysaccharide (LPS), a TLR4 ligand. BV CVLs, in cells expressing TLR2, also activated the HIV LTR. Thus, these studies show that soluble factor(s) present in the lower genital tract of women with BV activate cells via TLR2, identifying a pathway through which BV may mediate adverse effects.

Figure 1. BV CVLs induce higher levels of NF κB activation, HIV LTR activation and cytokine secretion by THP-1 monocytic cells than normal CVLs

A) THP-1 cells were co-transfected with an NF κB-luciferase reporter and a Renilla-TK (RL) internal control plasmid. At 18 hours post-transfection, cells were stimulated with CVL (10% of culture volume) and the indicated controls in triplicate for 6 hours. Data are expressed as relative units corresponding to the ratio of NF κB activity to RL. Two-tailed Mann-Whitney U analysis: *p<.0001; BV=23, Normal=17 B) THP-1 cells were co-transfected with an HIV-LTR luciferase reporter and the RL internal control plasmid as in Figure 1A. Following a 24-hour stimulation with 10% CVL or the indicated controls, the cells were lysed and a dual luciferase assay was performed. Each CVL was tested in triplicate in two separate assays. The average of both experiments is shown. Two-tailed Mann-Whitney U analysis: **p=.0015; BV=23, Normal=17 C) Culture supernatants from the HIV LTR experiment shown in Figure 1B were assayed for TNFα by ELISA. Two-tailed Mann-Whitney U analysis: ***p=.0002; BV=23, Normal=17 D) The culture supernatants from the HIV LTR experiment shown in Figure 1B were analyzed for IL-8 by ELISA. The horizontal line within the experimental scatter plot represents the median. Two-tailed Mann-Whitney U analyses: ****p=.0033; BV=14, Normal=16. IL-8 present within the CVL itself was measured by a flow cytometric cytokine multiplex assay and was subtracted from the total to determine the amount of induced IL-8. In all panels, the error bars represent the SEM. The horizontal lines within the experimental scatter plot denote the median. The dark horizontal lines above the scatter plots and below the asterisk or asterisks define the comparison groups that were used for the indicated statistical analysis.

Figure 2. NF κB activation by both American and Brazilian BV CVLs

Equal volumes of CVL from either the Brazilian or American cohorts were pooled and tested in an NF κB activation assay. THP-1 cells were transfected with the NF κB reporter and RL internal control plasmids and stimulated as in Figure 1A. Samples were tested in quadruplicate. One experiment representative of two is shown. Error bars represent the SD. Two-tailed Unpaired Student t test, *p<.0001, **p= .0046.

Figure 3. BV CVLs stimulate IL-8 secretion via TLR2

293 cells stably expressing CD14 (open symbols) or TLR2/CD14 (closed symbols) were stimulated with 5% CVL or the indicated controls in triplicate for 24 hours. Supernatants were analyzed for IL-8 secretion by ELISA as in Figure 1D. Error bars represent the SEM. The horizontal line within the scatter dot plot represents the median. Two-tailed Mann-Whitney U analyses: *p= .0043; N.S.= not significant, p>.05; BV=21; Normal=15

Figure 4. BV CVLs do not stimulate IL-8 secretion via TLR4/MD2

293 cells stably expressing TLR4 (open symbols) or TLR4/MD2 (closed symbols) were stimulated with 5% CVL or the indicated controls in triplicate for 24 hours. Supernatants were analyzed for IL-8 secretion as in Figure 1D. Each CVL was tested in two separate experiments. The average of the two experiments is shown. Error bars represent the SEM. The horizontal line within the scatter dot plot represents the median. Two-tailed Mann-Whitney U analyses: N.S.= not significant (p>.05); BV=23; Normal=17

Figure 5. BV CVLs induce HIV LTR activation via TLR2

293 cells stably expressing CD14 (open symbols) or TLR2/CD14 (closed symbols) were co-transfected with an HIV LTR luciferase reporter and the RL plasmid. The cells were stimulated with 5% CVL or the indicated controls for 6.5 hours. HIV LTR activity was normalized to the RL internal control. Each CVL was tested in triplicate. Error bars represent the SEM. The horizontal line represents the mean. Two-tailed one sample t test analyses: *p=.005 (compared to saline control), **p<.0001 (compared to saline control).