Downregulation of the DNA-Binding Activity of Nuclear Factor-κB p65 Subunit in Porphyromonas gingivalis Fimbria-Induced Tolerance

Abstract

Porphyromonas gingivalis fimbriae induce high levels of nuclear factor-κB (NF-κB)-dependent cytokine release upon primary but not secondary stimulation of monocytic cells (FimA tolerance). In this study, fimbriae induced Toll-like receptor-mediated activation of both p50 and p65 subunits of NF-κB upon primary cellular activation. However, activation of the transactivating p65 subunit (but not of the transcriptionally inactive p50 subunit) was significantly inhibited in fimbria-restimulated cells. Moreover, expression of a NF-κB-dependent reporter gene was inhibited upon secondary stimulation with fimbriae. NF-κB p65 downregulation may thus contribute to induction of FimA tolerance.

Lack of homeostatic balance between pro- and anti-inflammatory activities during infection may result in a state of disease. The innate immune system has developed the ability to downregulate excessive inflammatory reactions that could potentially contribute to tissue destruction, as in gram-negative septic shock (24). In this respect, prior exposure of monocytic cells to enterobacterial lipopolysaccharide (LPS) results in a transient state of diminished proinflammatory cytokine expression to a subsequent challenge with LPS (4, 6, 14). This hyporesponsive state, termed “endotoxin tolerance,” is mediated by transient reprogramming of Toll-like receptor (TLR) signal transduction pathways. Such alterations include reduced expression (12) or activation (16) of interleukin-1 (IL-1) receptor-associated kinase 1 (IRAK-1), suppressed degradation of IκB proteins (4), and changes in the subunit composition of nuclear factor-κB (NF-κB) (10). The NF-κB complex most commonly consists of p50/p65 heterodimers or p50/p50 homodimers. Upregulation of the p50 subunit, which favors the formation of transcriptionally inactive p50/p50 homodimers, has been implicated as a contributory mechanism in endotoxin tolerance (10). Reduced expression of TLRs has also been correlated with tolerance (18, 22). Innate immune cells also become tolerant to non-LPS proinflammatory molecules through generally different mechanisms, although the existence of common signaling pathways renders induction of cross-tolerance between different molecules possible (7, 17, 22). The understanding of tolerance mechanisms is important, because it may be exploited for prophylaxis of the septic shock syndrome or other severe inflammatory conditions.

The fimbriae of Porphyromonas gingivalis, a documented periodontal pathogen that has more recently been implicated as well in atherosclerosis (reviewed in reference 8), mediate host tissue degradation—possibly though induction of proinflammatory cytokines (reviewed in reference 11). Induction of IL-1β and tumor necrosis factor-α (TNF-α) by native fimbriae or their recombinant fimbrillin subunit (rFimA) in human monocytic THP-1 cells depends on fimbrial interactions with the TLR-CD14-CD11/CD18 pattern recognition receptor complex (7). The ability of fimbriae to stimulate cytokine release by interacting with these pattern recognition receptors has been confirmed in primary human monocytes by an independent group (19). We previously found that THP-1 cells pretreated with native fimbriae or rFimA become tolerant and respond with significantly reduced cytokine release to subsequent stimulation with either molecule (FimA tolerance) (7). The mechanisms underlying FimA tolerance could not be adequately explained by downregulation of the pattern recognition receptors involved. Specifically, surface expression of TLR2, TLR4, CD11a, or CD11b is essentially not downregulated in FimA tolerance, whereas expression of CD14 or CD18 is only modestly reduced (7). We thus hypothesized that induction of FimA tolerance may involve postreceptor alterations. Because NF-κB plays an important regulatory role in TLR-mediated innate immune and inflammatory responses, we investigated possible involvement of this transcription factor in FimA tolerance.

We first determined whether fimbriae indeed activate NF-κB in a TLR-dependent mode in differentiated THP-1 cells. THP-1 cells were differentiated with phorbol myristate acetate for 3 days (7) and subsequently allowed to rest for an additional 20 h in RPMI 1640 medium containing 10% fetal bovine serum (medium-pretreated cells). The following day, the cells were either incubated with medium only or were stimulated for 1 h with 1 μg of LPS-free native fimbriae or rFimA per ml, purified as described earlier (7). To neutralize any potential contamination of the culture medium with traces of exogenous LPS, polymyxin B sulfate (10 μg/ml) was added to the medium in all experiments with protein stimuli in this study. Stimulation with fimbriae was performed in the absence or presence of 10 μg of monoclonal antibodies (MAbs) per ml to TLR2 (TL2.1; e-Bioscience, San Diego, Calif.), TLR4 (HTA125; eBioscience), both TLR2 and TLR4, or isotype-matched control immunoglobulin. Cellular extracts were analyzed for NF-κB activation as previously described (6), using a quantitative enzyme-linked immunosorbent assay (ELISA)-based transcription factor assay (Active Motif, Carlsbad, Calif.) specific for the activated forms of p50 and p65 subunits of NF-κB (i.e., the detecting antibodies recognize epitopes on either p50 or p65 that are accessible only when NF-κB is activated and bound to its target DNA). All NF-κB activation data (Fig. 1 and Tables 1 and 2) were evaluated by analysis of variance and the Dunnett multiple comparison test by using the InStat program (GraphPad Software, San Diego, Calif.).

FIG. 1.

Activation of the p50 and p65 subunits of NF-κB by fimbriae (A and B) and LPS (C and D). Differentiated THP-1 cells were rested at 37°C for 20 h with medium only (M) and subsequently were stimulated for 1 h with native fimbriae (NF; 1 μg/ml) (A), rFimA (RF; 1 μg/ml) (B), E. coli LPS (EcLPS; 0.1 μg/ml) (C), and P. gingivalis LPS (PgLPS; 1 μg/ml) (D), in the presence or absence of anti-TLR MAbs or an isotype-matched control. To assess NF-κB activation in FimA tolerance, THP-1 cell groups were pretreated for 20 h with NF (A) or RF (B) prior to restimulation for 1 h with the same stimuli. NF-κB activation in cellular extracts was assessed with an NF-κB p50/p65 ELISA-based kit. The data are shown as means ± standard deviations of triplicate determinations. Asterisks indicate statistically significant (P < 0.05) inhibition of NF-κB subunit activation by anti-TLR MAb treatment or by secondary stimulation with native fimbriae or rFimA. The data shown above are from one of two independent experiments that yielded similar results.

TABLE 1.

Downregulation of the DNA-binding activity of the p65 subunit of NF-κB in FimA tolerance in primary monocytes^a

Cell stimulation (primary/secondary)	NF-κB subunit activation (OD450)
	p50	p65
Medium/medium (baseline control)	0.121 ± 0.042	0.068 ± 0.031
Medium/native fimbriae	1.715 ± 0.301	1.510 ± 0.208
Native fimbriae/medium	1.270 ± 0.195	0.141 ± 0.044*
Native fimbriae/native fimbriae	1.497 ± 0.264	0.402 ± 0.095*
Medium/rFimA	1.314 ± 0.138	1.280 ± 0.203
rFimA/medium	1.045 ± 0.249	0.152 ± 0.065*
rFimA/rFimA	1.178 ± 0.184	0.302 ± 0.087*

^aHuman monocytes were pretreated for 20 h with medium only, native fimbriae, or rFimA (both at 1 μg/ml). Subsequently, the cells were washed with warm medium, rested for 2 h, and restimulated for 1 h as indicated (stimuli used at the same concentration as in pretreatment). NF-κB activation in cellular extracts was analyzed with an NF-κB p50/p65 ELISA-based kit. OD₄₅₀, optical density at 450 nm. The results are shown as means ± standard deviations of triplicate determinations. Asterisks indicate statistically significant (P < 0.05) inhibition of the DNA-binding activity of NF-κB subunits in tolerant cells with or without restimulation.

TABLE 2.

Downregulation of the DNA-binding activity of the p65 subunit of NF-κB in FimA tolerance in THP-1 cells^a

Cell stimulation (primary/secondary)	NF-κB subunit activation (OD450)
	p50	p65
Medium/medium (baseline control)	0.187 ± 0.068	0.052 ± 0.019
Medium/native fimbriae	2.124 ± 0.548	1.378 ± 0.279
Native fimbriae/medium	1.779 ± 0.354	0.176 ± 0.035*
Native fimbriae/native fimbriae	2.007 ± 0.432	0.389 ± 0.072*
Medium/rFimA	1.693 ± 0.267	1.401 ± 0.321
rFimA/Medium	1.382 ± 0.298	0.209 ± 0.039*
rFimA/rFimA	1.538 ± 0.403	0.432 ± 0.090*

^aTHP-1 cells were pretreated for 20 h with medium only or with 1 μg of native fimbriae or rFimA per ml. Subsequently, the cells were washed, rested for 2 h, and restimulated for 1 h as indicated with the stimuli at the same concentration as in pretreatment. NF-κB activation in cellular extracts was analyzed with an NF-κB p50/p65 ELISA-based kit. OD₄₅₀, optical density at 450 nm. The results are shown are means ± standard deviations (n = 3). Asterisks indicate statistically significant (P < 0.05) inhibition of the DNA-binding activity of NF-κB subunits in tolerant cells with or without restimulation.

We found that native fimbriae (Fig. 1A) and rFimA (Fig. 1B) activated both the p50 and p65 subunits of NF-κB in medium-pretreated THP-1 cells. However, the ability of native fimbriae or rFimA to activate p50 and p65 was significantly (P < 0.05) suppressed by anti-TLR2 or anti-TLR4 MAbs and was almost completely inhibited in the presence of both MAbs (Fig. 1A and B). These data show that fimbriae activate NF-κB through TLR2 and TLR4. To verify the function-blocking specificity of the anti-TLR MAbs used, we tested their ability to inhibit NF-κB activation by known TLR2 or TLR4 agonists. As expected, anti-TLR4 (but not anti-TLR2) inhibited Escherichia coli LPS (tested at 0.1 μg/ml in Fig. 1C) (6), whereas anti-TLR2 (but not anti-TLR4) inhibited P. gingivalis LPS (tested at 1 μg/ml in Fig. 1D) (6), in accordance with earlier reports (2, 9, 14). The ability of fimbriae to activate NF-κB via both TLR2 and TLR4 is not unprecedented, because this property is shared by at least two other proteins: heat shock protein 60 (HSP60) and HSP70 (1, 21). However, the activation of the p65 subunit, but interestingly not of p50, was compromised in tolerant THP-1 cells: i.e., cells pretreated for 20 h with 1 μg of either native fimbriae (Fig. 1A, right corner) or rFimA (Fig. 1B, right corner) per ml prior to restimulation with the same stimuli. Indeed, the DNA-binding activity of p65 was significantly (P < 0.05) lower in fimbria-restimulated cells than in medium-pretreated cells exposed to the fimbrial preparations for the first time (Fig. 1A and B).

We confirmed and extended these findings by using primary monocytes purified from human peripheral blood (Table 1). The monocytes were pretreated at 37°C for 20 h with medium only or with 1 μg of native fimbriae or rFimA per ml. Subsequently, the cells were washed, allowed to rest for 2 h, and finally were either restimulated for 1 h with the same stimuli or incubated with medium only. The DNA-binding activity of the NF-κB p65 subunit (but not of the p50 subunit) was significantly inhibited (P < 0.05) in tolerant monocytes (i.e., native fimbria- or rFimA-pretreated cells) with or without secondary stimulation (Table 1). Since the p65 subunit is required for initiating transcription (20), its downregulation suggests a possible mechanism for the observed suppression of fimbria-induced cytokine expression in tolerant cells (7; data not shown). Interestingly, upon prolonged primary stimulation with native fimbriae or rFimA, monocytes maintained high levels of NF-κB p50 activity (even without further stimulation in secondary treatment) in contrast to the diminishing p65 activity (Table 1). These findings regarding the opposite trends in p50 versus p65 activity were confirmed in THP-1 cells (Table 2) and may be the net result of two different processes. (i) NF-κB (p50/p65) activation in monocytic cells results in upregulation of the expression of p105 (the precursor of p50) and its processing into p50, whereas p65 expression is unaffected (3, 5, 25). (ii) The activation of heterodimeric NF-κB (p50/p65) depends on the degradation of IκBα; suppressed IκBα degradation in tolerant cells inhibits NF-κB activation (15; our unpublished observations with FimA-tolerant cells). Although this would result in suppressed DNA-binding activities for both the p50 and p65 subunits, the net p50 activity after prolonged cellular activation would be compensated for by new synthesis of p50 from its p105 precursor. Indeed, p50/p50 homodimers do not interact with IκBα (or other inhibitory IκB proteins) (20), and thus their DNA-binding activity would not be inhibited, which is consistent with our findings that p50 activity remains high.

We then determined whether secondary stimulation with fimbriae results in suppressed NF-κB-dependent transcription of a reporter gene. Specifically, we used the stably transfected CD14⁺ Chinese hamster ovary (CHO) reporter cell lines 3E10/huTLR2 and 3E10/huTLR4, which express inducible cell surface CD25 under the control of an NF-κB-dependent promoter in response to TLR2 and TLR4 agonists, respectively (13). Cell surface expression of CD25 was evaluated by flow cytometric analysis after cell staining with fluorescein isothiocyanate-conjugated anti-CD25 MAb, as previously described (6). Both native fimbriae and rFimA activated CD25 expression, although the former appeared to be more active in the TLR2 clone (Fig. 2A), whereas the latter appeared more active in the TLR4 clone (Fig. 2B). These differences between native fimbriae and rFimA subunits may be contributed by the presence in the former of quantitatively minor but biologically active molecules, which are natural components of the native fimbrial superstructure (7, 23; H. Sojar et al., abstract 263 from the 72nd International Association of Dental Research Session, J. Dent. Res. 73:431, 1994). The streptococcal fibrillar M6 protein (provided by Susan Hollingshead, University of Alabama at Birmingham) was used as a control and found to activate neither clone (Fig. 2A and B). However, the ability of native fimbriae to induce NF-κB-dependent expression of CD25 in 3E10/TLR2 cells was inhibited upon restimulation (Fig. 2C). Similar induction of tolerance was observed in the 3E10/TLR4 clone after secondary stimulation with rFimA (Fig. 2D).

FIG. 2.

Activation of NF-κB-dependent expression of a reporter gene (CD25) by native fimbriae (NF) or rFimA (RF). 3E10/huTLR2 (A) and 3E10/huTLR4 (B) cells were incubated for 20 h with medium only (M) or with 1-μg/ml native fimbriae, rFimA, or streptococcal fibrillar M6 protein (SM6) as a negative control. To assess NF-κB-dependent expression of CD25 in FimA tolerance, 3E10/huTLR2 (C) and 3E10/huTLR4 (D) cells were respectively pretreated for 5 h with medium only and native fimbriae (C) or medium only and rFimA (D), followed by 20 h of restimulation with the same stimuli. Surface expression of CD25 was evaluated by flow cytometry. The histograms shown above are from one of three independent experiments that yielded similar results.

In summary, secondary stimulation of monocytic cells with native fimbriae or rFimA resulted in significant downregulation of the DNA-binding activity of the p65 subunit of NF-κB but not of the p50 subunit. A relative predominance of p50 subunits would favor the formation of p50/p50 homodimers, which function as negative transcription regulators (20). This is because the p50 molecules possess DNA-binding activity but lack a transactivation domain (20). Upon binding to the promoter region, p50/p50 homodimers not only are unable to activate transcription, but they also block the binding of transactivating p50/p65 complexes resulting in repressed transcription of target genes (10, 20). The notion that the observed fimbria-induced downregulation of the NF-κB p65 DNA-binding activity may result in transcriptional repression is supported by our finding that the expression of an exclusively NF-κB-dependent reporter gene was inhibited upon secondary stimulation with either native fimbriae or rFimA. These results suggest that NF-κB p65 downregulation may, at least partly, contribute to induction of FimA tolerance.