Abstract
Gram-negative bacterial LPS is a potent activator of inflammatory responses. The binding of LPS to CD14 initiates signal transduction; however, the molecular processes immediately following this event remain unclear. We engineered an LPS-inducible fibroblast reporter cell line to facilitate the use of molecular genetic techniques to study the LPS signaling pathway. A plasmid containing the human Tac Ag cDNA under transcriptional control of the human E selectin promoter was cotransfected into Chinese hamster ovary (CHO)-K1 cells together with a CD14 expression plasmid. A cell line was obtained, 3E10, which up-regulated expression of Tac following stimulation with LPS. Pools of mutagenized cells were exposed to LPS and then labeled with anti-Tac mAb. Cells that failed to up-regulate Tac expression were enriched by flow cytometry. Thirty clonal mutant cell lines were identified that continued to express CD14 and bind LPS, but failed to express Tac or translocate nuclear factor-κB (NF-κB) following LPS exposure. TNF-α-treated mutant cells continued to express Tac and translocate NF-κB. An analysis of LPS-induced NF-κB activity in heterokaryons derived from polyethylene glycol-fused cell lines indicated that recessive mutations in genes encoding components of the LPS signaling pathway accounted for the signaling defects. To date, two complementation groups have been identified from 11 cell lines analyzed. These data demonstrate that the TNF-α signaling pathway diverges from the LPS pathway early in the signal-transduction cascade despite similarities in LPS- and TNF-α-induced responses. Identification of the genes affected in these mutant reporter cells should identify heretofore-elusive components of the LPS signaling cascade.
Lipopolysaccharide is the major component of the outer leaflet of the Gram-negative bacterial outer membrane. During the course of an infection, LPS can be beneficial to the host by enhancing the immune response (1). However, the release of large amounts of LPS can be harmful to the host by stimulating the excessive production of inflammatory mediators resulting in septic shock (2). An increased understanding of the role of LPS in cellular activation could lead to the design of effective therapies to exploit the beneficial effects of LPS and modulate the harmful effects of excessive exposure to endotoxin.
CD14 is a glycosylphosphatidylinositol (GPI4)-anchored protein (3) expressed by monocytes, macrophages, and activated neutrophils that binds to LPS and initiates cellular activation (4). A central role for CD14 in LPS-mediated responses has been proposed based upon the following observations: 1) specific anti-CD14 mAb inhibit the ability of LPS to stimulate phagocytes (4, 5, 6, 7); 2) a soluble fragment of CD14 (sCD14) is present in blood (8) and facilitates the activation of cells that do not express membrane CD14 (9, 10, 11); 3) transfection of CD14 into Chinese hamster ovary (CHO)-K1 fibroblasts or HT1080 fibrosarcoma cells transformed these cells from an LPS-nonresponsive to an LPS-responsive phenotype. The physiologic relevance of these observations was underscored recently using a CD14-deficient knockout mouse that was 10,000-fold less sensitive to LPS when compared with hemizygous littermates (12).
Macrophages express multiple receptors for LPS in addition to CD14. For example, the macrophage scavenger receptor binds and internalizes the toxic lipid A moiety of LPS. In contrast to CD14, the scavenger receptor does not appear to participate in macrophage activation; rather, it seems to function in the clearance and detoxification of LPS (13). Macrophages also express the β2 integrins, a family of obligate heterodimers that include LFA-1, CR3, and CR4 (CD11a/CD18, CD11b/CD18, and CD11c/CD18, respectively). These leukocyte-restricted adhesion molecules can bind LPS-coated erythrocytes and Gram-negative bacteria (14). Transfection of CHO-K1 cells with any of the β2 integrins increased their sensitivity to LPS, implying a role for these receptors in macrophage-mediated responses to endotoxin (15, 16).
Although both CD14 and CD11/CD18 integrins bind LPS and are associated with signaling, evidence suggests that neither of these proteins functions in transducing a signal following LPS binding. CD14 is a GPI-anchored protein lacking a transmembrane and cytoplasmic domain. It is not directly apparent how this protein can function to transmit a signal across the cell membrane. Consistent with the hypothesis that CD14 does not directly function to induce a transmembrane signaling event is the observation that certain cells (e.g., endothelial cells) that do not express CD14 respond to LPS in the presence of sCD14 normally present in blood. To determine whether the cytoplasmic domains of CD11/CD18 were necessary for signaling, mutant constructs of CD11/CD18 devoid of their cytoplasmic tails were expressed in CHO cells. The mutant receptors were capable of enabling LPS responses when expressed in CHO cells, similar to transfected full-length CR3 constructs (16). Thus, like CD14, CD11/CD18 receptors do not require a cytoplasmic motif to enable an LPS signal. Finally, studies of the LPS antagonists deacylated LPS, lipid IVA (also known as lipid Ia, compound 408), and Rhodobacter sphaeroides lipid A have given additional credence to the idea that CD14 does not directly signal cells after binding to LPS. These LPS antagonists appear to inhibit LPS-induced signal transduction through interactions with a cellular element that is distinct from CD14 (17). One hypothesis that is consistent with these observations is that complexes of LPS and CD14 (or CR3) can interact with a putative transmembrane receptor that functions to transmit a signal across the plasma membrane and initiate an intracellular signaling cascade. The identification of this putative signaling component of the LPS receptor remains an important focus of endotoxin research today.
Progress in defining the LPS signal-transduction pathway has been hampered, in part, by difficulties in working with a ligand that cannot be purified to homogeneity. LPS also cannot be radiolabeled to a high specific activity (discussed in 18 . The observation that macrophages express several receptors that bind to LPS complicates the use of myeloid cells in the study of LPS-induced signal transduction using traditional biochemical methods. In addition to the problems of studying LPS signaling by traditional biochemical approaches, the low efficiency with which myeloid cells can be transfected with plasmid DNA (<10−6) precludes the facile application of molecular genetic approaches using monocyte-derived cell lines. One strategy to circumvent these difficulties is to identify an easily transfectable cell line that does not express other LPS receptors and responds to LPS via CD14. Mutant LPS nonresponder cell lines could then be derived from such cells and used to expression clone genes in the LPS signal-transduction pathway. This approach has been used successfully to define critical elements of important signaling pathways such as the IFN-γ (19, 20) and the TGF-β (21) signal-transducing pathways.
The CHO-K1 fibroblast cell line displays excellent growth characteristics and can be transfected with high efficiency. Established methods for inducing and rescuing mutations in CHO cells have led to the isolation of genes that encode important components of metabolic and signal-transduction pathways (22, 23). Furthermore, stable transfection of CHO cells with a CD14 expression plasmid (CHO/CD14) conferred an LPS-responsive phenotype, as evidenced by the inducible release of arachidonic acid (24), translocation of nuclear factor (NF)-κB from the cytoplasm to the nucleus (25), and the production of IL-6 (H. Heine, R. L. Delude, T. Espevik, and D. T. Golenbock, manuscript in preparation). These characteristics led us to engineer the LPS-responsive 3E10 reporter cell line, which up-regulates surface expression of the human Tac Ag (α-chain of the IL-2R; CD25) following exposure to endotoxin. We report in this work the engineering and use of an LPS-responsive reporter cell line, designated clone 3E10, that we have used to isolate LPS signal-transduction mutants. These mutants fall into two distinct complementation groups that fail to express Tac or translocate NF-κB following exposure to endotoxin. However, these LPS signaling mutants retain responsiveness to murine TNF-α when assessed for NF-κB translocation and Tac reporter expression. These LPS signaling mutants should be useful in identifying genes that encode necessary components of the CD14-dependent LPS signaling pathway by employing established expression-cloning techniques.
Materials and Methods
Reagents
Unless otherwise indicated, reagents were obtained from Sigma (St. Louis, MO). PBS and Ham’s F-12 were obtained from BioWhittaker (Walkersville, MD). Heat-inactivated FBS (LPS <10 pg/ml) was obtained from HyClone Laboratories (Logan, UT). Ciprofloxacin, a broad-spectrum antibiotic with activity against standard lab contaminants as well as mycoplasma, was a gift from Miles Pharmaceuticals (West Haven, CT). Protein-free LPS was obtained from Dr. Nilofor Qureshi (William S. Middleton Veterans Hospital, Madison, WI), and was extracted from Salmonella minnesota R595 (ReLPS). Lipids were prepared at 1 mg/ml in PBS and stored at −20°C. Before use, the suspensions were thawed and sonicated in an 80-W sonicator bath (Lab Supply, Hicksville, NY) for 2 min. The IFN-γ-inducible reporter plasmid, pUMS(GT)8-Tac, was obtained from Dr. M. Aguet (26).
Cell lines and culture
CHO cell lines were cultured in Ham’s F-12 containing 10% heat-inactivated FBS and 10 μg/ml of ciprofloxacin in a 5% saturated CO2 atmosphere at 37°C. Staining of cells for flow cytometry was performed as described (6).
Construction of the Tac reporter plasmid
The PCR was used to amplify a fragment of the human E-selectin (ELAM) promoter from −241 to −54 bp relative to the transcription start site. The 5′ primer (TAATCGATAATCATGCATTTTTGTCATA) and the 3′ primer (CAGAAGCTTATAGGAGGATATTGTCCACAT) were engineered with a ClaI and HindIII restriction site, respectively. A 50-μl amplification reaction containing each primer at 1 μM, 10 μg of human genomic DNA, and 1 U of Taq DNA polymerase (Promega, Madison, WI) was subjected to 30 amplification cycles at 95°C for 1 min, 55°C for 1 min, and 70°C for 1 min. The 205-bp amplification product was digested with ClaI and HindIII, and the fragment was cloned in place of the IFN-γ-inducible element contained within the unique ClaI and HindIII sites in pUMS(GT)8-Tac. The resulting plasmid, pUMS(ELAM)-Tac (Fig. 1⇓A), was subjected to dideoxynucleotide sequence analysis to confirm the integrity of the inserted promoter element.
Construction of an LPS-inducible Tac reporter plasmid. A, Plasmid map of pUMS(ELAM)-Tac, which contains the Tac Ag cDNA region downstream of a fragment of the human ELAM gene promoter driving transcription from the rabbit β-globin gene minimal promoter. B, A schematic diagram showing the location of the NF-κB and CAAT (NF-IL6) binding sites located within the promoter of the human ELAM gene and the PCR generated fragment that was used to generate pUMS(ELAM)-Tac.
Construction of the LPS-responsive 3E10 reporter cell line
We have reported previously that CHO-K1 cells transfected with CD14 responded to LPS by releasing arachidonic acid (24), translocating NF-κB (25), and releasing IL-6 (H. Heine, R. L. Delude, T. Espevik, and D. T. Golenbock, manuscript in preparation). We reasoned that the translocation of NF-κB could be used to drive the expression of a transfected reporter gene under the transcriptional control of an NF-κB-inducible promoter. We obtained a plasmid, pUMS(GT)8-Tac, which contained an upstream viral transcription termination region (UMS), eight repeats of an IFN-γ-binding element (GT), and the minimal promoter of the rabbit β-globin gene upstream of the Tac cDNA (26). We removed the artificial IFN-γ-responsive element from pUMS(GT)8-Tac and replaced it with a segment of the human endothelial leukocyte adhesion molecule (ELAM) promoter including sequences between −241 and −54 bp relative to the transcription start site, as described above (Fig. 1⇑). The resulting vector, pUMS(ELAM)-Tac, contains a cis regulatory element for NF-κB and a CAAT box that binds the CAAT enhancer-binding protein, NF-IL6 (27).
CHO-K1 cells were cotransfected with 5 μg each of pCEP4 (Invitrogen, San Diego, CA) containing the human CD14 cDNA (pCEP4/CD14) and pUMS(ELAM)-Tac, as described (28), except that Ham’s F-12 complete medium was used in place of α-MEM. Stable transfectants were selected in Ham’s F-12 complete medium containing hygromycin B at a concentration of 400 U/ml (LC Laboratories, Woburn, MA). Stable transfectants were transferred to two separate 10-cm dishes and allowed to grow to near confluence in the presence of selective drug. The cells in one dish were left untreated, while the cells in the other dish were stimulated with 100 ng of LPS/ml for 16 h. These cells were harvested and stained with a FITC-conjugated anti-Tac mAb or an isotype-matched mAb specific for the human Tac Ag, to define autofluorescence. LPS-stimulated and anti-CD25-stained cells were subjected to one round of positive selection using a Becton Dickinson FACScanPlus in enrichment mode to select cells with high inducible levels of Tac Ag expression. The sorting gate was set to enrich for LPS-stimulated cells that expressed the highest 5% of fluorescence (measured as mean channel fluorescence). An additional gate was defined, using forward scatter and side scatter as the parameters, to avoid the selection of cells whose enhanced immunofluorescence resulted from their larger size (i.e., cells with a large amount of scatter were excluded). The collected cells were plated in two wells of a six-well dish, allowed to achieve near confluence, and subjected to a second round of positive enrichment after LPS stimulation, as described above. To minimize basal expression of Tac, enriched cells were harvested without LPS stimulation and stained with FITC-conjugated anti-Tac mAb. These unstimulated cells were subjected to one round of negative enrichment by FACS by collecting cells whose immunofluorescence was in the lowest 5% with respect to Tac Ag expression. Finally, the cells were expanded and positively sorted after 18 h of LPS stimulation. After allowing these cells to expand in culture, this enriched population was cloned by limiting dilution in 96-well tissue culture dishes at a cell density of 0.3 to 0.5 cells/well. Individual clones were expanded and tested for LPS responsiveness. One CD14-expressing LPS-responsive cell line was chosen to represent wild-type CHO/CD14 and subcloned. In this way, a clonal cell line that expressed low basal levels of surface Tac, and that up-regulated Tac expression following exposure to low concentrations of LPS, was identified. We designated this clone as 3E10.
Mutagenesis and selection
3E10 reporter cells were seeded at a density of 1 × 106 cells per T25 tissue culture flask and allowed to grow overnight. The following day, the growth medium was replaced with 8 ml of Ham’s F-12 complete medium containing methanesulfonic acid ethyl ester (300 μg/ml), a concentration of mutagen that we (and others (22)) have determined to increase the mutational frequency of ouabain resistance from 1:106 to 1:104 in CHO fibroblasts. Following a 20-h incubation at 37°C, the medium was aspirated and the cells were washed once with 8 ml of Ham’s F-12 complete medium. The cells were allowed to recover for 2 days and then passed 1:10 into a T25 flask. After the cells grew to confluence, they were harvested and stored in liquid nitrogen until later use.
The mutagenized cells were thawed rapidly and plated into a 10-cm-diameter dish containing Ham’s F-12 complete medium with hygromycin B and allowed to grow to approximately 80% confluence. Cells were treated with 100 ng/ml of LPS for 18 h. The cells were labeled with anti-Tac mAb, and hyporesponsive cells were selected by FACS. Three rounds of negative sorting (i.e., only the lowest 5% of cells based on mean FL1 fluorescence) were performed to enrich for cells that failed to up-regulate Tac Ag expression in response to LPS. The resulting population was plated using limiting dilution in a 96-well tissue culture dish to obtain clonal mutant cell lines for analysis. The phenotypes of the mutant cell lines described in this work have remained stable over a period of 9 mo of continuous growth in cell culture.
NF-κB translocation analysis
Nuclear extracts were prepared from 1 × 106 nuclei using slightly hypotonic conditions in a solution containing 0.1% Nonidet P-40, exactly as described (25). A 32P-labeled DNA fragment containing the mouse Ig NF-κB binding site was mixed with 2 μg/lane of nuclear extracts and fractionated on a native 4% polyacrylamide gel. The gels were dried and used to expose x-ray film at −80°C. All gel-shift experiments shown in this report are representative of experiments performed on two or more occasions.
Northern blot analysis
For Northern blot analysis, cells were plated the day before the experiment at 3 × 106 cells per 10-cm-diameter tissue culture-treated Petri dish. The following day, the cells were stimulated for the indicated times, and RNA was extracted from the cells using TriReagent, exactly as directed by the manufacturer (Medical Research Center, Cincinnati, OH). Total RNA (10 μg) was size fractionated and blotted to nylon membranes, exactly as described (29). A random primed DNA probe was generated using the Tac cDNA as template (30). All experiments were performed on two or more occasions.
LPS-binding assay
LPS-binding assays with boron dipyrromethane (BODIPY)-labeled LPS were performed as described, except that the assay was adapted for whole cell binding. Complexes of BODIPY-LPS (100 ng/ml; gift of Rolf Thieringer, Merck, Rahway, NJ) and sCD14 (2 μg/ml; gift of Henry Lichenstein, Amgen, Thousand Oaks, CA) were preformed at 37°C for 6 h. BODIPY-LPS/sCD14 complexes were then added to cells in the presence or absence of human rLPS-binding protein (200 ng/ml; gift of Henry Lichenstein, Amgen) for 20 min at 37°C. Cells were washed with PBS and analyzed for fluorescence by FACS.
Cell fusions
Each mutant cell line was fused to the ouabain-resistant, but hygromycin-sensitive, CHO-K1a cell line, as described (31). The 7.11 mutant cell line was also fused individually to the 3E10, 7.11, 7.19, 8.18, and 8.20 mutant cell lines, and the 8.20 cell line was fused to 3E10, 7.7, 7.9, 7.13, 7.14, 7.15, 7.23, 7.24, and 8.20 cell lines to perform complementation analysis. Briefly, each clone was remutagenized (300 μg of EMS/ml) at a cell density of 106 cells per 25-cm2 flask, and selected in 0.5 mM ouabain in complete medium. In addition, clonal cell lines were transfected as described above with the neomycin resistance plasmid pcDNA3 (Invitrogen), and selected in 1 μg/ml of G418. The fusion partners (2.5 × 105 cells each) were seeded into wells of a 24-well tissue culture dish and grown overnight. The next day, the medium was replaced with 1 ml of 50% polyethylene glycol (w/v; m.w. 3350) in Ham’s F-12 complete medium. The plates were incubated at room temperature for 1 min, and the polyethylene glycol solution was aspirated. Following five washes with PBS, the cells were allowed to recover overnight in Ham’s F-12 complete medium. The following day, the medium was replaced with Ham’s F-12 complete medium containing 0.5 mM ouabain plus hygromycin B (400 U/ml), or ouabain plus G418 (depending on the antibiotic-resistance profile of each fusion partner) to select for heterokaryons. FACS analysis of heterokaryons was performed on two or more separate occasions, and gel-shift analysis was performed on three separate occasions.
Results
FACS analysis of the 3E10 reporter cell line
FACS analysis was performed to study the LPS-induced expression of Tac Ag in the 3E10 reporter cell line. The cells were seeded at a density of 0.5 × 106 cells/well in a six-well tissue culture-treated dish and allowed to grow overnight. The following day, LPS was added to some wells at a final concentration of 100 ng/ml and the cells were incubated an additional 18 h. Cells were harvested and stained with FITC-conjugated anti-human CD14 mAb or an isotype-matched FITC-conjugated anti-human CD20 control mAb. Cell fluorescence was determined using flow cytometry. Resting 3E10 reporter cells expressed low levels of surface Tac Ag (Fig. 2⇓). Stimulation with LPS up-regulated Tac expression, as measured by mean channel fluorescence by approximately 10-fold compared with nonstimulated control cells. This increase in Tac Ag expression was maximal within 18 h of stimulation and decreased to 50% of that level within 48 h of stimulation (data not shown). Thus, we engineered a fibroblast cell line that reports following stimulation with LPS by up-regulating the expression of cell membrane-associated Tac Ag approximately 10-fold when compared with nonstimulated control cells.
Analysis of Tac protein and mRNA expression in the 3E10 reporter cell line. The 3E10 reporter cell line was either left untreated (PBS) or stimulated with 100 ng/ml of LPS (LPS) for 18 h. The cells were harvested and stained with FITC-conjugated anti-Tac mAb, and surface expression levels of the reporter protein were detected using a flow cytometer. LPS-stimulated cells stained with the isotype-matched FITC-conjugated mAb (ISO) were included to show autofluorescence. The inset shows a Northern blot time-course analysis of the 3E10 reporter cell line stimulated with LPS (100 ng/ml) for the times indicated. The lane labeled “-” contains total RNA prepared from CHO/CD14 cells as a negative control. Steady state levels of Tac mRNA were revealed using a 32P-labeled full-length Tac cDNA as probe.
Northern blot analysis of the 3E10 reporter cell line
Northern blot analysis was performed on these cells to confirm that the increase in Tac expression was the result of increased steady state levels of Tac mRNA in cells exposed to LPS. 3E10 reporter cells were seeded at a density of 3 × 106 cells per 10-cm-diameter tissue culture-treated Petri dish and allowed to grow overnight. The following day, the cells were treated with LPS (100 ng/ml) for increasing amounts of time, and total cellular RNA was prepared. Total RNA was size fractionated in an agarose gel containing formaldehyde, then blotted to a nylon membrane and probed using a Tac-specific DNA probe. Resting 3E10 cells expressed very low basal levels of Tac mRNA (inset, Fig. 2⇑, lane 1). Reporter cells treated with LPS for 30 min expressed increased levels of steady state Tac mRNA, and these levels continued to increase until 2 h following LPS exposure (inset, Fig. 2⇑, lane 5). Within 4 h of stimulation, the steady state levels of Tac mRNA had decreased significantly. These results were consistent with the rapid increase and postinduction decrease in NF-κB DNA-binding activity in CHO/CD14 cells (25). Total RNA isolated from resting CD14-transfected CHO-K1 (CHO/CD14) cells were included in the Northern blot analysis as a negative control and were devoid of bands hybridizing to the Tac cDNA probe (inset, Fig. 2⇑, lane 7). A control blot was probed for β-actin mRNA expression, and the steady state levels of β-actin remained constant throughout the experiment (data not shown). Thus, the increase in Tac Ag expression observed in the 3E10 reporter cell line was due, at least in part, to increased steady state levels of the Tac Ag mRNA.
Isolation of 3E10 reporter cell mutants that fail to express Tac in response to LPS
CHO cells appear to be functionally hemizygous (22) for most genetic loci, thus facilitating the identification of mutant cell lines from reasonably small starting numbers of cells after a single round of mutagenesis. Exposure of wild-type CHO-K1 cells to methanesulfonate ethyl ester (300 μg/ml) results in a mutation frequency of approximately 1:10−4 based on the rate at which cells become resistant to ouabain (22). We used these conditions to mutagenize eight dishes of 3E10 reporter cells, as described in Materials and Methods; two of these mutagenized pools (pools 7 and 8) were used to generate LPS response mutants independently, to minimize the enrichment of sibling cells. We exposed mutagenized 3E10 reporter cells to LPS and subjected them to three successive rounds of negative enrichment using a FACS. The sort gate was set to enrich for cells that expressed extremely low to undetectable levels of Tac Ag on their surface (cells with the lowest 5% of mean channel fluorescence). Cells were simultaneously gated for forward and side scatter within the 5th through 95th percentile, to avoid choosing cells that had less surface immunofluorescence because they were smaller than the wild-type cell line. Following the final round of negative enrichment, the cells were cloned by limiting dilution into 96-well dishes. Individual clones were stimulated with LPS and labeled with phycoerythrin-conjugated anti-CD14 and FITC-conjugated anti-Tac mAb. Flow-cytometric analysis revealed that all of the clones expressed CD14 (data not shown). A total of 25 clones of 30 clones screened from pool 7 did not up-regulate Tac Ag in response to LPS. Similarly, 5 of 6 clones from pool 8 had no response to LPS.
We randomly selected a total of 11 clonal cell lines (clones 7.7, 7.9, 7.11, 7.13, 7.14, 7.15, 7.19, 7.23, 7.24, 8.18, and 8.20), representing both independently mutagenized pools of 3E10 reporter cells, for additional analysis. All of these mutant cell lines failed to up-regulate Tac expression following exposure to LPS. Four of the lines are shown in Figure 3⇓. Furthermore, stimulation of the mutants with murine TNF-α up-regulated Tac reporter expression normally compared with the parental 3E10 reporter line (Fig. 3⇓). These results indicated that the integrity of the Tac reporter gene remained intact. Therefore, these mutants appear to result from lesions in genes that encode important components of the LPS signaling pathway.
FACS analysis of the LPS signaling mutants exposed to either LPS or TNF-α. The parental 3E10 reporter line or the LPS signaling mutants were plated in triplicate wells of a six-well tissue culture-treated dish in Ham’s F-12 complete medium at a density of 1 × 106 cells/well and grown overnight. The next day, one well was exposed to PBS (PBS), one well was stimulated with LPS at 100 ng/ml (LPS), and the remaining well was exposed to murine TNF-α at 10 ng/ml (TNF). The plates were incubated for an additional 18 h, then the cells were harvested with 1 mM EDTA in PBS and stained with FITC-conjugated anti-Tac mAb. The y-axis represents cell number, and the x-axis represents the relative cell fluorescence representing Ag expression levels.
3E10 reporter mutant cell lines translocate NF-κB following exposure to TNF-α, but not LPS
Treatment of CHO/CD14 cells with LPS results in the rapid mobilization of NF-κB from the cytoplasm to the nucleus (25). The increase in nuclear NF-κB is evident within 5 min following exposure to LPS and is maximal within 30 min of stimulation. Thus, activation of NF-κB represents an early event in the signaling cascade initiated by LPS. We used the electrophoretic mobility shift assay (EMSA) to determine whether the reporter mutants could translocate NF-κB in response to LPS. We previously reported that CHO-K1 fibroblasts translocated NF-κB to the nuclear compartment following exposure to murine TNF-α (32). Therefore, the mutant cell lines were also stimulated with murine TNF-α. The 3E10 reporter cell line translocated NF-κB to the nucleus following exposure to either LPS (100 ng/ml) or TNF-α (10 ng/ml; Fig. 4⇓). However, each of the 3E10 mutant cell lines failed to mobilize NF-κB in response to LPS. It should be noted that mutants 8.18 and 8.20 show a subtle, almost undetectable increase in nuclear levels of NF-κB following exposure to LPS. The response is minimal, and may be due to the production of a partially functional mutant protein in the LPS signal-transduction pathway, or may be due to variability in background. In contrast to the response to exogenous LPS, all of the mutant lines respond normally to murine TNF-α, as evidenced by translocation of NF-κB to the nucleus. These data infer that the genetic lesions in these mutant cell lines are not the result of generalized effects on NF-κB activation. The observation that the mutant cell lines respond normally to TNF-α strengthens our previous conclusion that the affected genes in the LPS nonresponsive mutant cell lines are specific for the LPS signal-transduction pathway.
Analysis of NF-κB translocation in the LPS signaling mutants exposed to either LPS or TNF-α. The parental 3E10 reporter line or four randomly chosen LPS signaling mutants were plated at 5 × 105 cells/well in tissue culture dishes, as described in Figure 3⇑. The next day, one well was exposed to PBS (P), one well was stimulated with LPS at 100 ng/ml (L), and the remaining well was exposed to murine TNF-α at 10 ng/ml (T). After 45 min of stimulation, nuclear levels of NF-κB were assessed by EMSA. NE, no extract.
LPS signaling mutant 8.20 binds LPS in an LBP-dependent manner
LPS-binding protein (LBP) is a plasma protein that has been shown to accelerate the binding of LPS monomers from LPS aggregates to CD14, thus enhancing the sensitivity of cells to LPS. One possible explanation for the inability of the mutant cell lines to respond to LPS was that they were unable to interact with LPS and LPS/LBP complexes. Therefore, we analyzed one mutant, 8.20, for its ability to bind LPS in a fluorescent assay, using BODIPY-labeled LPS. As BODIPY-labeled LPS monomers exit from their aggregated state and bind to CD14, an increased fluorescent signal can be observed. As shown in Figure 5⇓, both the parental 3E10 line and the mutant 8.20 were capable of binding LPS equally in an LBP-dependent manner.
Analysis of LPS binding to signaling mutant. The parental 3E10 reporter line or the LPS signaling mutant 8.20 was incubated with BODIPY-LPS/sCD14 complexes in the presence or absence of LBP, and assayed for bound LPS by FACS. Mock-transfected neomycin-resistant CHO-K1 (CHO/Neo) cells do not express CD14, and were included as a control.
Fusing mutant 3E10 reporter cells to CHO cells restores the LPS-responsive phenotype
Fusion of mutant and wild-type cells can be used to determine whether the genetic lesions arise through dominant negative mutations or recessive loss-of-function mutations. The generation of such hybrid cells can be used also to determine whether the CD14 present in the mutant cell lines is biologically active, or has been mutated to a nonfunctional receptor, as the only cell line capable of providing CD14 to the resultant heterokaryons is the LPS-response mutant. A hygromycin-sensitive, ouabain-resistant cell line was chosen for hybrid analysis. This line, CHO-K1a, does not express CD14; furthermore, wild-type (nontransfected) CHO cells do not respond to LPS (24, 25). CHO-K1a cells do not contain the pUMS(ELAM)-Tac expression vector. This cell line was fused to each of the individual mutant 3E10 reporter cell lines, and stable fusions were selected by growth in medium containing ouabain and hygromycin B. As a control, we fused a reporter cell line that had undergone mutagenesis and negative selection by FACS, yet retained an LPS-responsive phenotype (clone 7.5). Stimulation of the mutant/CHO-K1a fusion lines resulted in an increase in the surface expression of the Tac Ag following LPS exposure (Fig. 6⇓). This increase in Tac expression was identical when compared with the increase in Tac expression observed with the control 7.5 cell line. Therefore, the mutations that were isolated using the 3E10 reporter act in a recessive fashion.
FACS analysis of the LPS signaling mutants fused to the CHO-K1a cell line. The parental 3E10 reporter line or the LPS signaling mutants fused to the CHO-K1a cell line were plated in duplicate wells of a six-well tissue culture-treated dish at 5 × 105 cells/well in Ham’s F-12 complete medium and grown overnight. The next day, 10 μl of PBS was added to one well (PBS), and the remaining well was exposed to an equal volume of LPS, to achieve a final concentration of 100 ng/ml (LPS). The plates were incubated for an additional 18 h at 37°C, and the cells were harvested and stained with anti-Tac mAb. The dotted line (… ) shows the fluorescence of LPS-stimulated cells stained with an isotype-matched FITC-conjugated mAb. The y-axis represents cell number, and the x-axis represents the relative cell fluorescence representing Ag expression levels. The fused mutant cell lines were given a designation of F#, in which the F represents cell fusion and the # represents the name of the parental mutant line; e.g., F7.11 was produced by fusing the 7.11 mutant to the CHO-K1a cell line. Note that clone 7.5 is a control cell line derived from a population of cells that had undergone mutagenesis and negative selection by FACS and yet retained an LPS-responsive phenotype.
The ability to translocate NF-κB following exposure to LPS was assessed using EMSA. While CHO-K1a cells and the mutant cell lines failed to translocate NF-κB in response to LPS, fusion of the two cell lines produced heterokaryons that translocated NF-κB from the cytoplasm to the nucleus after exposure to endotoxin (Fig. 7⇓). This provides further evidence to support our conclusion that the mutations in the LPS reporter cell were most likely due to loss-of-function (i.e., recessive) mutations. The observation that the genetic lesions can be complemented using the wild-type copy of the mutated gene suggests that an expression-cloning strategy is feasible for the rescue of mutants and concomitant identification of the mutated genes.
Analysis of NF-κB translocation in the LPS signaling mutants fused to CHO-K1a cells. The parental 3E10 reporter line or the LPS signaling mutants fused to the CHO-K1a cell line were plated at 5 × 105 cells/well. The next day, one well was treated with PBS alone (P) and the remaining well was exposed to LPS at 100 ng/ml (L). The plates were incubated for an additional 45 min., and then the cells were harvested. Nuclear extracts were prepared, and nuclear levels of NF-κB were determined using the EMSA. NE, no extract.
Complementation analysis of the Tac reporter mutants reveals that they represent a two-complementation group
To determine whether each of the mutant lines was defective in a unique gene, complementation analysis was performed. Eleven of the thirty LPS signaling mutant cell lines and the parental 3E10 reporter cell line were tested. This analysis revealed that these mutants constitute at least two distinct complementation groups. The larger majority of the mutant cell lines fell into a single complementation group, which we called group A; mutants 7.7 and 7.15 constituted the B complementation group.
The results of experiments using select members of each complementation group are presented in Figures 8⇓ and 9. We selected the entire group of B mutants and selected clones 7.11 and 8.20 as representatives of the A mutants. To perform these experiments, clonal 3E10, 7.11, 7.15, and 8.20 cell lines were transfected with pcDNA3, which conferred resistance to G418. In addition, the 7.7 and 8.20 cell lines were remutagenized with methanesulfonic acid ethyl ester (300 μg/ml), and ouabain-resistant lines were selected in medium containing 0.5 mM of ouabain.
Complementation analysis of the mutant cell lines using FACS analysis. Either ouabain- or neomycin-resistant derivatives (see text) of the parental 3E10 reporter line or each of the LPS signaling mutants were fused to each other and then plated in duplicate wells of a six-well tissue culture-treated dish at 5 × 105 cells/well. The next day, 10 μl of PBS was added to one well (PBS) and the remaining well was exposed to an equal volume of LPS (LPS; final concentration of 100 ng/ml). The plates were incubated for an additional 18 h, harvested, and then stained with anti-Tac mAb (CD25) for flow-cytometric analysis.
Figure 8⇑ shows the results of FACS analysis performed on the various fusion partners, which were selected in G418 and ouabain after polyethylene glycol fusion. As expected, fusion of the parental cell line 3E10 to either mutant 8.20 or 7.7 cell lines produced heterokaryons that responded to LPS and up-regulated Tac Ag expression. Fusion of mutants 8.20 to 7.11 or mutants 7.15 to 7.7 failed to complement the mutations, demonstrating that these cell lines shared genetic lesions. In contrast, the fusion of mutant 8.20 to either mutant 7.7 or 7.15 did restore the LPS-responsive phenotype, demonstrating that 8.20 is lesioned in a gene that is normally expressed in 7.7 and 7.15. The fusion partners were also analyzed for LPS-induced nuclear localization of NF-κB, and the results of this analysis are presented in Figure 9⇓. These results were qualitatively identical to those demonstrated by flow cytometry in the previous figure.
Complementation analysis of the mutant cell lines using EMSA. Either ouabain- or neomycin-resistant derivatives (see text) of the parental 3E10 reporter line or each of the LPS signaling mutants were fused to each other and plated at 5 × 105 cells/well. The next day, one well was treated with PBS alone (P) and the remaining well was exposed to LPS at 100 ng/ml (L). The cells were incubated for an additional 18 h and harvested. Nuclear extracts were prepared and analyzed for the presence of NF-κB by EMSA. NE, no extract.
Discussion
In summary, a useful fibroblast cell line was engineered for use in studying the LPS signaling pathway. Furthermore, this LPS reporter cell line was used to identify several LPS signal-transduction mutant cell lines that constitute at least two distinct complementation groups. These signaling mutations are specific for the LPS signaling pathway, because the mutant cell lines remain responsive to murine TNF-α.
The identification of proteins involved in LPS-induced signal transduction immediately downstream from CD14 is a central question in our understanding of the signaling cascade. Any hypothesis suggesting that LPS and TNF-α share signaling components is not supported by the data presented in this study. Mutant 3E10 reporter cell lines that failed to respond to LPS responded normally to mouse TNF-α (Fig. 3⇑), a finding that suggests that at least some of the most proximal portions of the TNF-α and LPS signaling cascades are unique to each pathway.
While CD14 clearly initiates the process of LPS-induced cellular activation, a mechanistic description of the molecular events that occur subsequent to LPS binding to CD14 remains elusive. CD14 is a GPI-linked protein with no known signaling motifs. Thus, it does not seem likely that CD14 actually effects a transmembrane signaling event. Strong evidence supports a model in which complexes of CD14 and LPS signal cells through interactions with a distinct signal-transducing molecule(s). This derives from studies performed with lipid A analogues, which antagonize the effects of LPS when tested with human cells. Kitchens et al. (17, 33) observed a discrepancy between the concentration of the LPS antagonist lipid IVA needed to inhibit binding of LPS to CD14 (μM-mM range) and the concentration of lipid IVA needed to antagonize LPS-induced signaling (nM-μM range). Thus, LPS inhibitors suppressed signaling at concentrations that were too low to prevent LPS binding to CD14. Previous observations had indicated that several of the LPS analogues that acted as LPS antagonists in human cells had LPS-mimetic activity in cells from mice (34). Human and rodent cell lines that expressed human- or mouse-derived CD14 were engineered. The species-specific effect did not correlate with the species of CD14 expressed in cells, but was dependent on the genetic background of the transfected cell line (32). We hypothesized that CD14-bound LPS causes a signal-transduction event by interacting with a second signaling receptor, which also functions to discriminate between structural differences present in the lipid A moieties of the various LPS antagonists. This putative signal transducer was proposed as the target of LPS antagonists.
The role of sCD14 in the activation of endothelial cells initially suggested that a CD14/LPS complex might form a ternary complex with an as yet unidentified signal-transducing molecule with transmembrane and cytoplasmic signaling domains (10, 11, 35). Sophisticated biochemical approaches to identify such a molecule have not been productive, although the failure of such studies does not eliminate the possibility that such a transmembrane signal transducer exists. Wright and colleagues have recently presented evidence suggesting that CD14 is a lipid transfer protein, and that cell signaling depends upon the movement of CD14 into cells. These authors have observed a lesion in LPS import in cells obtained from C3H/HeJ LPS nonresponder mice. While C3H/HeJ mice express CD14, and bind LPS normally, they fail to internalize LPS presented to cells as sCD14/LPS complexes (36). Thus, Wright has suggested that the first step in CD14-mediated signaling is the movement of LPS into the lipid bilayer of cells. Detmers et al. (37) have suggested that a trypsin-sensitive surface protein aids in the movement of LPS from CD14 into the cell and subsequent LPS-induced signaling. Trypsin treatment of cells had no effect on CD14 expression, but abolished both LPS internalization and LPS-induced activation of adhesion molecules. However, a question that has not been directly addressed using this approach is whether LPS internalization is actually necessary for LPS signaling or if LPS signaling is necessary for rapid internalization of the ligand. The inability of C3H/HeJ cells to respond to LPS may be the cause of reduced LPS uptake. It is likely that the identification and cloning of any surface molecule involved in LPS uptake and signaling will shed great light upon the question of LPS uptake and its relationship to signal transduction. Progress in the understanding of CD14-mediated signaling should be greatly facilitated following the identification of the next signaling molecule downstream of CD14 in the LPS signaling cascade.
The identification and analysis of cell mutants have proven to be an alternative and effective means to binding studies for identifying components of biochemical and signal-transduction pathways. Mutants in a defined pathway have been used to identify genes that can revert the mutant phenotype. Excellent examples of the successful application of this strategy include the characterization of the TGF-β receptor in mutants derived from mink lung epithelium (21), the identification of JAK/STAT pathway components required for IFN-γ-induced signaling in HT1080 fibrosarcoma cells (19, 20), and the expression cloning of peroxisomal assembly factor-2 gene in mutant peroxisome-deficient CHO cells (23).
The use of mutational analysis to study the LPS signaling pathway has been attempted by several investigators, but was apparently thwarted by the low transfection frequencies of the cell lines from which the mutants were derived. Mains and Sibley (38) derived NF-κB signaling mutants in the 70Z/3 pro-B cell line that failed to respond to LPS, but were responsive to IFN-γ. Although the approach was straightforward, later attempts to clone the signaling lesions in these mutant cell lines were not feasible because of a transfection frequency of approximately 1:106. The initial description of a 55-kDa protein necessary for LPS responses, which later proved (via nongenetic approaches) to be CD14, was made by Hara-Kuge and colleagues, who identified mutant J774.1 cells that neither growth arrested in response to LPS nor bound 125I-labeled LPS (39). Like Mains and Sibley, these investigators failed to rescue the lesion in the mutant cell line by expression cloning probably because of the low transfection efficiency of monocyte/macrophage-derived cell lines. Thus, while the mutant analysis of LPS nonresponder B lymphocyte- and macrophage-derived lines is interesting and informative, they are also excellent examples of the problems inherent in performing genetic studies with such cell lines. In contrast, the genetic analysis of CHO-derived mutant cell lines should be significantly easier because the transfection efficiency of CHO cells is approximately 10−2.
It is our hope that the analysis of these mutants, as well as additional mutant lines that we have not yet characterized, will soon constitute a number of complementation groups defining genes that encode necessary components of the LPS signal-transduction cascade. The identification of mutant lines from additional pools of mutagenized stock will also be necessary, because of the tendency to identify siblings using this technique. Ultimately, the goal is to clone genes necessary for LPS responses, including additional signaling components of the LPS receptor. The identification of such genes should facilitate the rationale design of drugs to reduce the morbidity and mortality associated with endotoxin-caused disease.
Footnotes
-
↵1 This was supported by National Institutes of Health Grant R01-GM54060 (to D.T.G.). R.L.D. was supported by a fellowship grant from the Eisai Research Institute (Andover, MA).
-
↵2 Current address: Beth Israel Deaconess Medical Center, Department of Surgery, 330 Brookline Avenue, Boston, MA 02215. E-mail address: rdelude{at}bidmc.harvard.edu
-
↵3 Address correspondence and reprint requests to Dr. Douglas Golenbock, Maxwell Finland Laboratory for Infectious Diseases, 774 Albany Street, Boston, MA 02118. E-mail address: Douglas.Golenbock{at}bmc.org
-
↵4 Abbreviations used in this paper: GPI, glycosylphosphatidylinositol; BODIPY, boron dipyrromethane; CHO, Chinese hamster ovary; EMSA, electrophoretic mobility shift assay; LBP, lipopolysaccharide-binding protein; NF-κB, nuclear factor-κB; sCD14, soluble fragment of CD14.
- Received January 9, 1998.
- Accepted May 8, 1998.
- Copyright © 1998 by The American Association of Immunologists
References
In this issue
