Cutting Edge: Lipopolysaccharide Induces Physical Proximity Between CD14 and Toll-Like Receptor 4 (TLR4) Prior to Nuclear Translocation of NF-κB

Materials and Methods

Results and Discussion

Previous studies have shown that CD14 is up-regulated by LPS (20, 21). To determine whether TLR4 expression is affected by LPS, adherent monocytes were stimulated with 200 ng/ml LPS or unstimulated for 0 to 30 min, fixed, then labeled with rhodamine-conjugated TLR4. The results showed that TLR4 was expressed on the membrane of unstimulated cells, and that TLR4 expression was increased by LPS, reaching a peak value at 15–30 min. Moreover, the inclusion of serum was found to enhance TLR4 expression. Cells were incubated with 200 ng/ml LPS in the presence or absence of 10% heat-inactivated serum. In the absence of serum, controls and LPS-treated cells demonstrated fluorescence intensities of 10,430 ± 7,810 and 28,670 ± 9,160 counts/min (n = 3; p < 0.05), respectively. In the presence of serum, control and LPS-treated samples exhibited intensities of 11,000 ± 8,520 and 96,860 ± 7,270 counts/min (n = 3; p < 0.001). Therefore, TLR4 expression is up-regulated by LPS.

To study the physical proximity of CD14 and TLR4, RET microscopy was used. RET microscopy, in both imaging and spectroscopic formats, can be used to detect lateral proximity of membrane proteins. This is accomplished by detecting the migration of excitation energy from donor (FITC) to acceptor (TRITC) chromophores attached to different proteins (22). Monocytes were allowed to adhere to endotoxin-free glass coverslips. Cells were incubated with either 200 ng/ml LPS or medium alone for 30 min at 37°C. Cells were fixed and stained with FITC-conjugated anti-CD14, rhodamine-conjugated anti-TLR4, or both at 4°C in the dark. The fluorescence images showed that anti-CD14 and anti-TLR4 were expressed on monocytes, and that the fluorescence of both receptors was enhanced after LPS stimulation (Fig. 1⇓). Moreover, RET emission images were only observed on cells treated with LPS, suggesting a role for LPS in promoting the physical proximity between these two molecules.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Representative fluorescence micrographs of labeled human monocytes. Cells were incubated with 200 ng/ml LPS in 10% heat-inactivated FCS (LPS+) or buffer alone (LPS−) at 37°C for 30 min. A–F, Cells were fixed and labeled with both anti-CD14 and anti-TLR4 at 4°C for 30 min. Using appropriate optical filters, CD14 (FITC) fluorescence (A and B), TLR4 (rhodamine) fluorescence (C and D), and RET (E and F) are shown. Enhanced labeling and RET are observed in the presence of LPS. G and H, After fixation, other samples were blocked with normal goat serum and stained with FITC-conjugated anti-NF-κB mAbs. Substantial translocation into the nucleus is seen after 60 min at 37°C in LPS-treated cells (H), but not in control cells, which display cytoplasmic staining (G) (n = 3–8, where n = the number of independent experiments) (×1000).

LPS-induced RET between CD14 and TLR4 was quantitatively studied using spectrophotometry. Cells were labeled with FITC-conjugated anti-CD14, rhodamine-conjugated anti-TLR4, or both reagents. First, we measured the emission spectra of single cells labeled separately with either FITC-conjugated anti-CD14 or rhodamine-conjugated anti-TLR4 (Fig. 2⇓). Compared with that of control cells incubated with only culture medium, the intensities of labeled CD14 or TLR4 were substantially enhanced on cells stimulated by LPS, as noted above. Spectra were then obtained from cells that were labeled with both FITC-conjugated anti-CD14 and rhodamine-conjugated anti-TLR4. Spectra from control cells showed a small single peak at ∼560 nm. In contrast, when cells were incubated with 200 ng/ml LPS, the spectrum showed two peaks at ∼560 and ∼590 nm, indicating the presence of RET. Moreover, RET was not observed between anti-CD14 and anti-Mo5, anti-Mo1 (CD11b/CD18), or anti-CD87 in the presence and absence of LPS (data not shown). Thus, RET is specific for anti-CD14 and anti-TLR4 in the presence of LPS.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Representative emission spectra of labeled monocytes are shown. A–C, Cells were incubated with either 200 ng/ml LPS or culture medium only at 37°C for 30 min. Cells were stained with FITC-conjugated anti-CD14 (donor) and/or rhodamine-conjugated anti-TLR4 (acceptor). LPS enhances TRL4 and CD14 expression as illustrated in A and B, respectively. Although RET cannot be discerned in the absence of LPS, substantial amounts of RET can be observed in the presence of LPS (C). D–F, Cells were treated with 200 ng/ml LPS at 37°C for 30 min then labeled with both reagents. After fixation, cells were stained with anti-CD14/anti-TLR4 at ratios of 1:5, 1:1, or 5:1. Spectral features are clearly dependent upon the donor/acceptor-labeling ratio of the cells (n = 4).

To provide further evidence in support of energy transfer, we tested cells labeled with different ratios of donor and acceptor chromophores. RET emission intensity depends upon the ratio of donor and acceptor chromophores (22). Spectra from cells labeled with different ratios of FITC-conjugated anti-CD14 and rhodamine-conjugated anti-TLR4 reagents are shown in Fig. 2⇑, D–F. As anticipated, emission spectra vary with the donor-to-acceptor ratio. The relative heights of the two peaks vary in a dose-dependent fashion, with the 1:5 donor-acceptor ratio having the greatest relative emission at 590 nm.

Although these data support an LPS-dependent physical association of CD14 and TLR4, they do not discriminate between the possibilities that LPS binding triggers 1) a lateral association of CD14 and TLR4 in the plasma membrane, and 2) the delivery of preassociated CD14 and TLR4 from an intracellular pool to the plasma membrane. To address this issue, we sought to up-regulate CD14 and TLR4 expression in the absence of LPS. Therefore, we treated cells with the calcium ionophore A23187 (2 μM, 20 min, 37°C) in HBSS containing 5 mM Ca²⁺. This procedure up-regulated CD14 and TLR4 expression to the same extent as LPS (Fig. 3⇓, A and B). However, CD14-to-TLR4 RET was not observed in these cells (Fig. 3⇓C, A23187 control). Thus, CD14 and TLR4 are not delivered to the plasma membrane in a preassociated form. When LPS (200 ng/ml) is incubated with A23187-treated cells, RET is again observed (Fig. 3⇓C). Hence, the physical proximity of CD14 and TLR4 is induced at the plasma membrane.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

Effect of A23187 on expression of CD14 and TLR4, and their association. Monocytes were incubated at 37°C for 15 min in the presence or absence of 2 μM A23187 then labeled with anti-CD14 and/or anti-TLR4 reagents as described above. Cells were fixed with 3.7% glutaraldyhyde. A23187 dramatically enhances the expression of both CD14 (A) and TLR4 (B) on cells. When cells were labeled with both reagents, RET was not observed (C). However, RET was detected when A23187-treated cells were incubated with LPS (C) (n = 4).

Previous studies have shown that the lipid A moiety of LPS is responsible for its key biological features (23). Therefore, we tested the ability of lipid A to promote the molecular proximity of CD14 and TLR4 under conditions identical with those described above for LPS. Again, RET was observed (data not shown). However, a lipid A antagonist does not support CD14-TLR4 RET (Q. Jiang, K. Miyake, S. Qureshi, and H. Petty, unpublished observations). Thus, CD14-TLR4 complex formation parallels certain biological features of the system.

To correlate the physical association of CD14 and TLR4 with downstream elements of signal transduction, we examined NF-κB translocation to the nucleus. In these experiments, monocytes were incubated with 200 ng/ml LPS or culture medium alone for 60 min at 37°C. Cells were fixed as described in Materials and Methods. After blocking with 10% normal goat serum, cells were stained with FITC-conjugated anti-NF-κB mAb. As expected, LPS induced NF-κB staining inside the nucleus (Fig. 1⇑H). In contrast, there was no nuclear staining for cells exposed to HBSS alone (Fig. 1⇑G). RET was observed at the earliest time points tested, whereas roughly 1 h was required to obtain nuclear staining for NF-κB. Thus, physical proximity of CD14 and TLR4 precedes NF-κB nuclear translocation.

In this study we have shown that LPS triggers close physical proximity between CD14 and TLR4. TLR4 is one member of the Toll-like receptor family. TLR4 requires CD14 to participate in the process of LPS-induced signaling, including NF-κB activation (21, 24). The positive RET signal between TLR4 and CD14 indicates that the donor and acceptor chromophores are ≤7 nm apart; on a molecular level, this would roughly correspond to CD14 and TLR4 being nearest protein neighbors. A CD14/LPS/TLR4 complex is consistent with other reports suggesting that LPS is in close physical proximity with TLR4 during signaling, although the binding affinity of LPS for TLR4 might be low (25). We suggest a model wherein LPS induces physical proximity between CD14 and TLR4. LPS is delivered from the serum protein LPS-binding protein to membrane-bound CD14. LPS-charged CD14 has been previously shown to interact with the leukocyte integrin CD11b/CD18 before cell adherence (9); this may contribute to certain immediate responses of leukocytes to LPS. The close physical proximity of CD14 and CD11b/CD18 is lost after adherence. It now appears that LPS-charged CD14 may then promote changes in gene expression by physically interacting with TLR4. We suggest that the LPS/CD14 complex can associate with various membrane partners to elicit different aspects of the inflammatory response. The ability of CD14 to shuttle between CR3 and TLR4 is analogous to urokinase receptor (another GPI-linked protein) shuttling between CR3 and CR4 during neutrophil polarization (26, 27).

We have also shown that the up-regulation of CD14 and TLR4 are very early events in LPS stimulation of human monocytes. Although the up-regulation of CD14 has been previously reported (20, 21), the up-regulation of human TLR4 has not. However, TLR4 expression declines after LPS stimulation in murine peritoneal macrophages (28, 29). It is possible that this difference could be accounted for by species or cell type; differences in human and murine TLR4 responses have been reported (e.g., 25). However, this difference could also be accounted for by the fact that we are studying changes that take place over several minutes, whereas the down-regulation of TLR4 expression occurs over a period of hours and is temporally associated with LPS internalization and changes in TLR4 gene expression.

Although these studies have established a molecular proximity between CD14 and TLR4, they have not established the supramolecular composition of these signaling complexes. That is, there may be other components of the signaling complex. One possibility is CD11b/CD18. However, attempts to measure RET between TLR4 and CD11b/CD18 have not been successful (our unpublished data). It may be that these proteins are part of the same membrane domain or present in an array on a membrane, but too far apart to be detected with RET. However, other microscopic methods, such as fluctuation correlation spectroscopy, might contribute to further dissecting these signaling complexes within the membranes of intact living cells.