CD14 is a pattern recognition receptor involved in the interaction with multiple ligands, including LPS from Gram-negative bacteria and lipoarabinomannan (LAM) from mycobacteria. While the interactions between LPS and soluble CD14 (sCD14) have been analyzed in detail, LAM/CD14 interactions remain uncharacterized due to the lack of suitable functional assays. We describe herein a novel bioassay for the analysis of CD14/ligand interactions. CD14-negative myeloid HL-60 cells up-regulate endogenous CD14 gene expression when stimulated with LPS in the presence of recombinant soluble CD141–348. Using the HL-60 bioassay, we showed that sCD141–348 confers responsiveness not only to LPS, but also to LAM. The response to LAM, but not that to LPS, was highly dependent on LPS binding protein (LBP). The N-terminal half of CD14 was sufficient to mediate HL-60 responses to LAM, since HL-60 cells responded with similar efficiency when stimulated with LAM and LBP in the presence of sCD141–348 or sCD141–152. Thus, the N-terminal 152 amino acids of CD14 contain the site(s) involved in the interaction with LAM and LBP, as well as the residues required for LAM-dependent CD14 signaling.
CD14 is a 55-kDa glycosylphosphatidylinositol (GPI)4-linked glycoprotein that exists in two forms: membrane-bound (mCD14), expressed on monocytes/macrophages and neutrophils (1, 2), and soluble (sCD14), found at high concentrations (2–6 μg/ml) in normal human plasma (3). Membrane-bound CD14 serves as the receptor for LPS, the main component of the cell wall of Gram-negative bacteria (4). Engagement of CD14 by LPS induces a number of biologic responses, including the secretion of inflammatory cytokines (5) that is thought to play a major role in the pathogenesis of septic shock (6). Furthermore, we have shown that LPS/CD14 interactions protect primary macrophages from productive infection by HIV-1 through the induction of suppressive factors, most notably C-C chemokines (7).
Interestingly, monocytes/macrophages and neutrophils are not the only targets of LPS-induced inflammatory responses. Indeed, sCD14 has the ability to confer LPS responsiveness to CD14-negative cells, such as endothelial cells, astrocytes, and epithelial cells (8, 9, 10). Sensitive responses of inflammatory cells to LPS require cooperation between CD14 and LPS binding protein (LBP), a lipid transfer protein that facilitates the binding of LPS to sCD14 or mCD14, but does not seem to participate directly in initiating signal transduction (11, 12). The essential roles of both CD14 and LBP in the response to low concentrations of LPS have been highlighted by in vivo studies with knockout animal models (13, 14).
Recently, it has been shown that CD14 is able to interact not only with LPS, but also with an unexpectedly broad range of bacterial products, including lipoarabinomannan (LAM) from Mycobacterium tuberculosis (15), amphiphilic membrane molecules (16, 17) and peptidoglycan (18) from Staphylococcus aureus, mannuronic acid polymers from Pseudomonas aeruginosa (19, 20), rhamnose-glucose polymers from Streptococcus mutans (21), and chitosans from arthropods (22). Because of this ability to interact with different bacterial products, CD14 is currently viewed as a pattern recognition receptor (15) critically involved in innate immune responses to bacteria (23). Of particular interest is the demonstration that CD14 is the receptor involved in the response of monocyte/macrophages to mycobacterial LAM (15). Anti-CD14 Abs blocked LAM-dependent responses of both CD14-bearing THP-1 cells and CD14-transfected 70Z/3 pre-B cells (15). Furthermore, CD14 has been shown to mediate the uptake of nonopsonized M. tuberculosis by human microglia (24) and the release of TNF-α by monocytes isolated from tuberculosis patients (25). The capacity of mycobacterial constituents to elicit TNF-α secretion by infected macrophages seems to determine the ability of the micro-organism to survive and replicate within the host (26, 27).
Because of the critical role played by CD14 in the immune response to bacteria, the molecular requirements for LPS/CD14 interactions have been extensively dissected using a number of approaches. The U373 bioassay is predicated on the ability of sCD14-containing serum or recombinant sCD14 to confer LPS responsiveness to CD14-negative astrocytes (8, 28). Using mutated and/or truncated sCD14 molecules in the U373 bioassay, it was possible to show that the LPS binding site maps to the N-terminus of CD14 (28) in a region that spans amino acids 57 to 64 and corresponds to the epitope recognized by the neutralizing mAb MEM-18 (29, 30). The epitope recognized by another neutralizing anti-CD14 mAb, 3C10, is in the region between amino acids 7 and 14 and defines yet another functional domain required for cellular signaling, but not LPS binding (31).
By contrast, the U373 bioassay was not suitable for the analysis of LAM/CD14 interactions because U373 cells do not respond to LAM/sCD14 complexes even in the presence of LBP (32) (G. Cosentino, unpublished observation). In an attempt to establish an in vitro assay to characterize sCD14/LAM interactions, we tested several human CD14-negative cells lines for their ability to respond to LPS and/or LAM in a serum-dependent fashion. We show herein that myeloid HL-60 cells up-regulate the expression of the endogenous CD14 gene when stimulated with LPS or LAM in the presence of serum or recombinant sCD14. The N-terminal 152 amino acids of human CD14 are sufficient to impart responsiveness to both ligands in this novel bioassay.
Materials and Methods
Cells and reagents
HL-60 cells were grown in RPMI 1640 supplemented with 10% FCS (Biologic Industries, Beit Haemek, Israel), 2 mM glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. COS-7 cells were cultured in DMEM/high glucose supplemented with 2 mM glutamine, 1% nonessential amino acids, 5 × 10−5 M β-ME, and 10% FCS. LPS from Salmonella minnesota (wild-type) purified by gel filtration, normal human serum (NHS), and cycloheximide were purchased from Sigma (St. Louis, MO). Polymyxin B sulfate was purchased from Calbiochem (La Jolla, CA). Purified mycobacterial LAM (nonmannose-capped LAM (Ara-LAM) isolated from a rapidly growing Mycobacterium species, and mannose-capped LAM isolated from M. tuberculosis Erdman) was provided by Drs. J. Belisle and P. J. Brennan (Colorado State University, Fort Collins, CO). Purified recombinant sCD14 and LBP were prepared as previously described (11).
Anti-CD14 mAbs 3C10 (IgG2b, TIB-228, American Type Culture Collection (ATCC), Manassas, VA) and 63D3 (IgG1, ATCC HB-44) directed against distinct epitopes of human CD14, anti-FcγR2/CD32 mAb IV.3 (IgG2b, ATCC HB-217), anti-myc tag mAb 9E10.2 (IgG1, ATCC CRL-1729), and an IgG1 isotype control (ATCC HB-236, mouse anti-human IgE) were purified from ascites by affinity chromatography on protein G-Sepharose (HiTrap Protein G, Pharmacia, Uppsala, Sweden). Purified anti-CD14 mAb MEM-18 (IgG1) and FITC-conjugated goat anti-mouse Ig were purchased from Caltag (San Francisco, CA) and Becton Dickinson (Mountain View, CA), respectively. Anti-CD54 mAb 14D2D12 (IgG1) was a gift from Dr. Ruggero Pardi (San Raffaele Scientific Institute, Milan, Italy). The Abs were biotinylated using ε-caproylamido-biotin-N-hydroxy-succinimide ester (BioSPA, Milan, Italy). Phycoerythrin (PE)-conjugated streptavidin was purchased from PharMingen (San Diego, CA). The endotoxin content of all cell culture reagents was assessed by the Limulus amoebocyte lysate assay (BioWhittaker, Walkersville, MD) and was always <0.125 endotoxin units/ml.
Construction and expression of sCD141–348, sCD141–152, and sCD141–152 myc
Deletion of the eight C-terminal amino acids of CD14 involved in GPI anchor attachment has been shown to result in the production of a sCD14 molecule, sCD141–348 (11). To obtain sCD141–348, we designed primers (forward, 5′-GGCTGGAACAGGTGCCTAAA; reverse, 5′-TGCTCTAGAGCACTATTACAGCACCAGGGTTCCCGA, nucleotides 1103–1122 and 1327–1344 of GenBank accession no. X06882) that amplify the 3′ end of CD14 DNA, introducing two stop codons and an XbaI site after the codon for amino acid 348 in mature CD14 (underlined). For PCR amplification, a pcDNA1 vector (Invitrogen, San Diego, CA) containing full-length human CD14 cDNA cloned in the HindIII and XbaI sites was used as template. The fragment thus amplified was digested with PvuII and XbaI, and pcDNA1-hCD14 was digested with HindIII and PvuII. The two DNA fragments were isolated on agarose gels and ligated into the HindIII and XbaI sites of pcDNA3 (Invitrogen). Soluble CD141–152 was generated by PCR amplification using pcDNA1-hCD14 as template, and primers 5′-CCCAAGCTTGGGACCATGGAGCGCGCGTCCTGC (that introduces a HindIII site upstream of the codon for amino acid 1 in mature CD14; underlined) and 5′-TGCTCTAGAGCACTATTAGCCTGGCTTGAGCCACTG (that introduces two stop codons and an XbaI site downstream of the codon for amino acid 152 in CD14; underlined). sCD141–152 myc was PCR-amplified using the upstream primer designed for sCD141–152, and a downstream primer 5′-ATGCAGAATTCGCCTGGCTTGAGCCACTG that introduces an EcoRI site immediately 3′ of the codon for amino acid 152 of CD14 (underlined). The two amplification products were digested with HindIII and XbaI or with HindIII and EcoRI, isolated on agarose gels, and cloned into the corresponding sites of pcDNA3 wild-type or pcDNA3-myc. The latter was provided by Dr. Maria Guttinger (San Raffaele Scientific Institute) and was obtained by PCR-mediated insertion of a cassette encoding the 9E10.2 epitope of human c-myc (GAGCAAAAGCTCATTTCTGAAGAGGACTTGAATTGA) (33) in-frame with an EcoRI site. The sequences of all resulting plasmids were verified by the dideoxynucleotide chain termination/extension method using the Sequenase version 2.0 kit (U.S. Biochemical Corp., Cleveland, OH).
For expression, the vectors containing sCD14 cDNA were transiently transfected into COS-7 cells using DEAE-dextran. After transfection, the cells were cultured overnight in DMEM/10% FCS, then washed three times with ice-cold PBS and maintained in serum-free DMEM. Supernatants were collected every 3 days, and the concentration of sCD14 released by transfected COS-7 cells was assessed by ELISA (see below). The presence of sCD14 in the supernatants of transfected COS-7 cells (conditioned medium) was further confirmed by metabolic labeling, followed by immunoprecipitation with anti-CD14 mAb 3C10 or anti-c-myc mAb 9E10.2 (for sCD141–152myc) and SDS-PAGE analysis. Supernatants from COS-7 cells transfected with an empty pcDNA3 vector (mock supernatants) were used as negative controls in all experiments.
ELISA for sCD14 quantitation
The concentrations of sCD141–348, sCD141–152, and sCD141–152 myc released by COS-7 cells transfected with the sCD14 expression vectors were assessed by a sandwich ELISA. Ninety-six-well plates (Maxisorp, Nunc, Roskilde, Denmark) were coated with anti-CD14 mAb 3C10 (20 μg/ml in 0.1 M carbonate buffer, pH 9.0) for 2 h at 37°C, washed with PBS/0.05% Tween 20, and blocked with PBS/5% BSA/0.05% Tween 20 for 2 h at 37°C. After washing, the samples were added and incubated overnight at 4°C in a humidified chamber. After extensive washing, biotinylated anti-CD14 mAb 63D3 or MEM-18 (2 μg/ml) was added to detect sCD141–348 or sCD141–152 and sCD141–152myc, respectively. Following a 2-h incubation at 37°C, the wells were washed extensively and incubated with acetyl-avidin/biotinylated peroxidase complexes (1/2000 in PBS/1% BSA/0.05% Tween 20; BioSPA). A substrate solution (ortho-phenylenediamine dihydrochloride, Sigma) dissolved in 0.05 M phosphate-citrate buffer and mixed with 30% hydrogen peroxide was then added to the wells. The reaction was stopped after a 20-min incubation at room temperature by adding 2 M H2SO4. OD was read at 492 nm. Purified recombinant sCD14 (11) was used as a standard. This assay readily detects ≥2 ng/ml of human sCD14.
Formation of LPS/LAM-sCD14 complexes
Complexes of LPS or LAM with sCD14 were formed before each experiment and were obtained by incubating LPS or LAM (2 μg/ml) with serum-free COS-7 supernatants containing sCD141–348 (2.4 μg/ml) overnight at 37°C (12) in the absence of LBP. The complexes were then added to HL-60 cells at different dilutions, as indicated in the figure legends.
HL-60 cells were resuspended at 6 × 105 cells/ml in RPMI 1640–1% Nutridoma NS (Boehringer Mannheim, Mannheim, Germany; serum-free medium) and stimulated for 24 to 48 h as detailed in Results. In preliminary experiments, we found that polymyxin B at a concentration of 15 μg/ml completely blocked the effects of LPS even when the latter was added at concentrations as high as 1 μg/ml. Polymyxin B (15 μg/ml) was therefore added to all experiments performed with LAM.
Membrane-bound CD14 expression was detected by direct immunofluorescence, as previously described (34). Briefly, HL-60 cells in staining buffer (RPMI 1640/10% dialyzed NHS containing 0.01% sodium azide) were incubated with biotinylated 63D3 mAb or isotype control (5 μg/ml) for 40 min at 4°C in the presence of anti-FcγR2/CD32 mAb IV.3 (10 μg/ml). Cells were then washed and incubated with PE-conjugated streptavidin (1/300) for 30 min. CD54 expression was detected by indirect immunofluorescence using mAb 14D2D12 followed by FITC-conjugated goat anti-mouse Ig. Cells were then extensively washed and fixed in 2% paraformaldehyde. The percentages of positive cells and mean fluorescence intensity were analyzed on a FACScan (Becton Dickinson) by gating on the living cell population, as defined by forward and side light scatter.
HL-60 cells stimulated with preformed LPS/sCD141–348 complexes for 24 h were washed and resuspended in methionine/cysteine-free RPMI 1640 supplemented with 1% Nutridoma NS. LPS/sCD141–348 complexes were then readded, together with 35S cell labeling mix (Amersham, Aylesbury, U.K.; 250 μCi). The cells were cultured for 24 h, collected, and lysed on ice with immunoprecipitation buffer (1% Nonidet P-40, 20 mM HEPES (pH 7.3), 150 mM NaCl, and 2.5 mM EDTA) containing 1 mM PMSF and leupeptin, pepstatin A, antipain, and chymostatin (Sigma; all at 1 μg/ml) for 30 min. Lysates were spun at 14,000 rpm for 15 min at 4°C in a microcentrifuge. The supernatants were then transferred to a new microcentrifuge tube and precleared overnight with normal mouse serum (1 μl) and protein G-agarose beads (Boehringer Mannheim; 40 μl). After a short spin, each supernatant was immunoprecipitated with mAb 63D3 or isotype control (5 μg) and protein G-agarose beads, mixing gently at 4°C for 6 h. The beads were then extensively washed in immunoprecipitation buffer, resuspended in sample buffer, boiled for 5 min, and cooled on ice. The samples were resolved by SDS-PAGE on a 9% reducing gel and autoradiographed using Kodak Biomax MS films (Eastman Kodak, Rochester, NY). The relative intensity of the bands was measured by densitometry using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
Total RNA was isolated as described by Chomczynski (35), and cDNA was synthesized and subjected to semiquantitative RT-PCR as previously described (34). For each PCR reaction, one primer was 5′-labeled using [γ-32P]dATP and T4 kinase polymerase (Life Technologies, Grand Island, NY; 5 U/10 pmol of primer) and added in a 1:10 molar ratio with cold primer. CD14 transcripts were detected as a 470-bp band amplified by forward primer (5′-GGTGCCGCTGTGTAGGAAAGAA) and reverse primer (5′-GTGCCGGTTATCTTTAGGTCCTC; nucleotides 78–99 and 613–635 of GenBank accession no. X06882). As a control, a 438-bp band corresponding to GAPDH transcripts was amplified using appropriate primers (5′-GGGAAGGTGAAGGTCGGAGTC and 5′-CTGATGATCTTGAGGCTGTTG; nucleotides 1456–1476 and 3818–3848 of GenBank accession no. J04038). Amplification was performed on a thermocycler (Omnigene, Hybaid, Teddington, U.K.) for 20 cycles (1 min each at 94, 62, and 74°C) for CD14 transcripts and for 15 cycles (1 min each at 94, 60, and 74°C) for the GAPDH control. The number of amplification cycles was such that the maximum signal intensity for a set of samples was within the linear portion of a product-vs-template amplification curve.
LPS induces serum-dependent expression of CD14 in HL-60 cells
We (7) and others (36, 37) have previously shown that LPS up-regulates the expression of its receptor, mCD14, on human monocytes/macrophages. To determine whether LPS also up-regulates mCD14 on an mCD14-negative cell line, we treated the human myelocytic leukemia cell line HL-60 with LPS (1 μg/ml) in the presence of NHS. Immunofluorescence analysis with anti-CD14 mAb 63D3 showed that a 24-h incubation with LPS resulted in vigorous expression of mCD14 (Fig. 1,A). Induction of CD14 expression by LPS required de novo protein synthesis because it was completely abrogated by the addition of cycloheximide (100 ng/ml; Fig. 1,B). In addition, up-regulation occurred in the presence of NHS but not in serum-free conditions (Fig. 1 C). This result suggests that the signal delivered by LPS to HL-60 cells is mediated by a serum component(s).
The response of HL-60 cells to LPS is mediated by sCD141–348
sCD14 has been shown to retain LPS binding capacity and to confer LPS responsiveness to mCD14-negative cells (8, 9, 10). We therefore asked whether the serum component required for the LPS-dependent induction of endogenous CD14 expression in HL-60 cells was sCD14. For our experiments, we used sCD141–348 that was transiently expressed in COS-7 cells. Figure 2 shows that mCD14 was undetectable on HL-60 cells stimulated for 24 h with LPS or sCD141–348 alone in serum-free conditions. By contrast, the combination of LPS (100 ng/ml) and sCD14 (1.2 μg/ml) induced mCD14 expression that was further enhanced by the addition of recombinant LBP (100 ng/ml). A comparable response was obtained by adding LPS in the presence of NHS (10%).
The ability of sCD141–348 to mediate LPS-dependent induction of CD14 at the protein level was also assessed by immunoprecipitation. HL-60 cells were metabolically labeled with [35S]methionine and cysteine, and CD14 was immunoprecipitated with anti-CD14 mAb 63D3. Figure 3,A shows the presence of a 55-kDa band corresponding to CD14 in lysates from HL-60 cells stimulated with LPS and sCD141–348 for 24 h, but not from cells incubated with either stimulus alone. The CD14 band was specific, since it was absent in samples immunoprecipitated with a control mAb (data not shown). Semiquantitative RT-PCR was used to show that induction of CD14 protein was due to enhanced CD14 mRNA expression. Figure 3 B shows that high levels of CD14 mRNA were expressed in HL-60 cells stimulated with both LPS and sCD141–348, but not in cells treated with individual stimuli. Similar results were obtained by Northern blot analysis (data not shown). Together these data show that expression of mCD14 on LPS-treated HL-60 cells reflects induction of the endogenous CD14 gene.
Interestingly, CD14 was not the only gene triggered by stimulation with LPS, sCD141–348, and LBP. Figure 4 shows that expression of CD54 was undetectable on unstimulated HL-60 cells, but was strongly up-regulated after a 3-day incubation.
HL-60 cells respond to mycobacterial LAM in the presence of sCD141–348 and LBP
The observation that LPS and sCD14 induced mCD14 on HL-60 cells suggested that this assay could also be used to test for interactions between sCD14 and other bacterial ligands. Mycobacterial LAM has been shown to act as a CD14 ligand (15) and to trigger important biologic responses, such as monokine release (25) and uptake of mycobacteria (24). Because the LAM signaling system appears to require a receptor component whose expression is restricted to cells of hemopoietic origin (32), and HL-60 cells represent myelomonocytic precursors (38), we reasoned that sCD141–348 might mediate LAM-dependent responses in our novel HL-60 bioassay. All experiments were performed with Ara-LAM, because mannose-capped LAM had negligible biologic activity (data not shown). A role for contaminating LPS was ruled out by adding polymyxin B (15 μg/ml) in all experiments.
Figure 5 shows that HL-60 cells incubated for 24 h with LAM (100 ng/ml) and sCD141–348 (1.2 μg/ml) expressed low levels of mCD14. However, addition of sCD141–348 and recombinant LBP (100 ng/ml) resulted in a striking up-regulation of mCD14 by LAM. No response was detectable upon treatment with either sCD141–348 or LAM alone (data not shown). Expression of mCD14 was always induced upon stimulation with LAM and NHS (10%), but varied in intensity depending on the batch of NHS.
RT-PCR analysis confirmed that the expression of mCD14 in HL-60 cells stimulated with LAM, sCD141–348, and LBP was paralleled by a marked increase in the level of CD14 mRNA. A weak induction of CD14 RNA was consistently observed in cells incubated with LAM and sCD141–348 without LBP, but not in cells exposed to either stimulus alone (data not shown).
Notably, the addition of LAM, sCD141–348, and LBP up-regulated not only mCD14, but also CD54 expression on HL60 cells (Fig. 4). Together, these results indicate that the combination of sCD141–348 and LBP renders HL-60 cells fully responsive to LAM.
Interactions of LAM and LPS with sCD141–348 and LBP
To compare the abilities of LAM and LPS to interact with sCD141–348 and LBP, we stimulated HL-60 cells in serum-free conditions with increasing concentrations of the bacterial ligands in the presence of constant concentrations of sCD141–348 (1.6 μg/ml) and LBP (100 ng/ml). Figure 6 (top panel) shows that both LAM and LPS dose-dependently up-regulated mCD14 on HL-60 cells. Since the average molecular mass of LAM (39) is approximately 10 times that of LPS (40), on a molar basis the two ligands were comparably efficient in stimulating HL-60 cells. However, a dramatic difference between LAM and LPS became apparent when HL-60 cells were treated with preformed complexes containing LAM or LPS and sCD141–348 in the absence of LBP. Indeed, LAM/sCD141–348 complexes induced a modest response, detectable only at the highest concentration of LAM tested (5 μg/ml). By contrast, LPS/sCD141–348 complexes stimulated HL-60 cells quite efficiently, although LBP appeared to be required for maximal responsiveness. These results indicate that in the HL-60 bioassay, responses induced by LAM, but not by LPS, are critically dependent on the presence of LBP. This hypothesis was confirmed by stimulating HL-60 cells with LAM or LPS (100 ng/ml) in the presence of sCD141–348 (1.6 μg/ml) and increasing concentrations of LBP. Figure 6 (center panel) shows that the LBP dose-response curves for LAM and LPS differ by at least 2 orders of magnitude at the lowest LBP concentrations and become comparable only when LBP is added at the concentrations found in normal human plasma (<0.5 μg/ml) (41). By contrast, LAM and LPS (100 ng/ml) showed nearly equivalent requirements for sCD141–348 when optimal amounts of LBP (100 ng/ml) were present (Fig. 6, bottom panel).
sCD141–152 contains the site(s) required for the response of HL-60 cells to LAM and LBP
The N-terminal half of CD14 (CD141–152) is known to mediate responses to LPS both when expressed as a soluble molecule (28) and when expressed on the membranes of transfected cells (18, 42). Therefore, we asked whether this portion of CD14 is sufficient to mediate HL-60 responses to LAM as well. For these experiments, we produced two forms of sCD14: sCD141–152, truncated at amino acid 152, and sCD141–152myc, in which sCD141–152 is fused with the 9E10.2 epitope of human c-myc (33) at the C-terminus. Figure 7 (left panel) shows that addition of either sCD141–348 or sCD141–152 in the presence of optimal LBP concentrations (100 ng/ml) supported a strong response to LPS (100 ng/ml). The response was still vigorous, although somewhat lower, in the absence of LBP (data not shown). The presence of the c-myc tag did not affect the ability of sCD141–152 to trigger HL-60 responses. Indeed, the dose-response curve was similar for the three proteins, particularly in the microgram range that corresponds to physiologic concentrations of sCD14 (3). These results demonstrate that sCD141–152 retains full activity in the HL-60 bioassay. In addition, HL-60 cells responded efficiently when stimulated with LAM (100 ng/ml) in the presence of sCD141–152 or sCD141–152 myc and LBP (100 ng/ml; Fig. 7, right panel), whereas no significant response was detected in the absence of LBP (data not shown). Again, the dose-response curves were similar for sCD141–348 and sCD141–152, with or without the c-myc tag. Similar curves were also obtained using a lower concentration of LBP (10 ng/ml; data not shown). The response of HL-60 cells to the bacterial ligands was due to sCD141–152 and not to irrelevant proteins in the COS-7 supernatants, since supernatants from mock-transfected cells did not up-regulate mCD14 (data not shown). Furthermore, identical results were obtained adding equivalent concentrations of sCD141–152 myc purified by affinity chromatography on an anti-myc tag mAb-protein G column (data not shown). Thus, the N-terminal 152 amino acids of CD14 contain the site(s) involved in the interaction with LAM and LBP as well as the residues required for LAM-dependent CD14 signaling.
The interactions between LAM and CD14 are likely to play an important role in the pathogenesis of mycobacteria-induced disease and in the ability of the host to successfully contain the infection. However, LAM/CD14 interactions have not been dissected due to the lack of suitable in vitro assays. Indeed, the U373 bioassay that takes advantage of the ability of sCD14 to confer LPS responsiveness to mCD14-negative astrocytoma cells (8, 28) was not applicable in the case of LAM (32) (G.C., unpublished observation). The HL-60 bioassay herein described provides a novel and powerful tool to define the requirements for CD14-mediated responses to LAM in a myeloid cell system.
Our data demonstrate that the N-terminal 152 amino acids of human sCD14 are sufficient to induce HL-60 cells to respond not only, as expected (28, 42), to LPS, but also to LAM. However, there is a striking difference in the requirements for the responses to the two ligands; whereas HL-60 cells were highly sensitive to stimulation with preformed complexes of LPS and sCD141–152 in the absence of LBP, no significant response to LAM was observed unless LBP was added at concentrations within the physiologic range. Although LBP is known to interact with bacterial products other than LPS (e.g., β1–4-linked d-mannuronic acid derived from P. aeruginosa (20)), our findings were unexpected because responses of mCD14-negative cells to sCD14 and LPS do not usually require LBP. The reason for the strict LBP dependency of LAM-induced responses in the HL-60 bioassay is unclear at the moment.
Were the affinity of sCD14 for LAM much lower than that for LPS, LBP-mediated transfer of LAM to sCD14 would become critical for the timely formation of an active LAM-containing complex. However, the experimental conditions of the HL-60 bioassay (overnight preincubation of LAM and sCD14 at 37°C to achieve complex formation, incubation of the cells with the stimulants for 24–48 h) should favor efficient binding between sCD14 and its ligand, circumventing at least in part the need for LBP (12). Thus, we are inclined to believe that the LBP dependency of LAM-induced responses does not simply result from a low affinity in LAM/sCD14 interactions.
One hypothesis to explain our data is that because the HL-60 bioassay measures ligand-induced cellular activation, rather than simple binding of the sCD14/ligand complex, the LBP requirement for LAM-dependent, but not LPS-dependent, responses might reflect a difference in the cell surface receptor(s) involved in triggering the cellular response upon stimulation with the two ligands. In this model, complexes of LPS and sCD14 would be sufficient to engage the still unidentified receptor that mediates cell activation in the absence of mCD14 (43), whereas a different receptor would recognize a ternary complex containing LBP, sCD14, and LAM. We are currently testing the hypothesis that such a complex is formed.
An alternative hypothesis is that the HL-60 cell receptor that interacts with LPS/sCD14 complexes might recognize LAM/sCD14 complexes as well, but only in the conformation that the latter achieve when LBP is part of the complex. In this other model, LBP would not interact directly with the receptor, but would be instrumental in conferring the appropriate conformation to the LAM/sCD14 complex. The finding that both myeloid HL-60 cells and astrocytoma U373 cells respond to LPS and sCD14, but only HL-60 cells respond to LAM, sCD14, and LBP, seems to favor the two-receptor hypothesis. It is possible that only cells of hemopoietic origin possess all the components of the signal transduction machinery required for the LAM-containing complex to bind and induce cellular activation (32).
An additional and not necessarily distinct issue is whether the ability of CD14 to interact differently with different bacterial ligands and LBP reflects the existence of selective binding sites on the CD14 molecule. The interaction of CD14 with LAM, LPS (28, 42), peptidoglycan (18), or products of S. aureus (16) requires the N-terminal 152 amino acids of the protein. However, important differences seem to exist among the modes of binding and/or interaction. Distinct residues in CD14 are known to be selectively involved in the recognition of different LPS ligands, such as LPS from Escherichia coli vs LPS from Porphyromonas gingivalis (44). Similar, but not identical, sequences are critical for the responses of CD14 transfectants to LPS and peptidoglycan (18). On the other hand, the response of CD14-transfected murine pre-B cells to peptidoglycan (18) and those of U373 cells to S. aureus products and sCD14 (16) are both LBP independent, unlike LAM-induced HL-60 cell activation. It will be interesting to determine whether the CD14 residues involved in LAM recognition differ from those already identified for other ligands, and whether different sites on sCD14 are engaged in the presence of LBP.
This work was supported by the National Tuberculosis Project (Istituto Superiore di Sanitá-Ministero della Sanitá), Grant 18 (to D.V.), and by National Institute of Allergy and Infectious Diseases, National Institutes of Health Contract NO1-AI-25147.
Abbreviations used in this paper: GPI, glycosylphosphatidylinositol; mCD14, membrane CD14; sCD14, soluble CD14; LBP, lipopolysaccharide binding protein; LAM, lipoarabinomannan; NHS, normal human serum; Ara-LAM, nonmannose-capped lipoarabinomannan; PE, phycoerythrin; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.