NF-κB is activated during acute inflammatory states as well as in other injury response disease states. Several pathologic states in squamous tissue injury response are characterized by increased squamous proliferation. This study was performed to investigate the hypothesis that Pseudomonas aeruginosa LPS is able to activate a proliferative phenotype in squamous cells via NF-κB induction and that this NF-κB-mediated response may be abrogated with the classic anti-inflammatory agent indomethacin. EMSA, luciferase reporter gene experiments, Western blots, and cellular proliferation assays were performed in normal and transformed human keratinocytes after stimulation with P. aeruginosa LPS. EMSA and luciferase reporter gene assays showed a 3- to 5-fold induction of active NF-κB in human keratinocyte cell lines after stimulation with P. aeruginosa LPS. The stimulation correlated with significantly increased cellular proliferation. As one potential mechanism for this increase in proliferation, an NF-κB-specific activation of cyclin D1 was observed. Both the NF-κB induction and proliferation response were inhibited with indomethacin and in dominant negative stable transfection clones. P. aeruginosa LPS activates proliferation of human keratinocytes, potentially through the induction of NF-κB and cyclin D1. These findings suggest that bacterial components can contribute to proliferative disease states in squamous epithelium through NF-κB activation.
It is becoming clear that bacterial infection and host inflammatory responses are intimately associated with the development of a full range of abnormal proliferative states in host tissues up to and including the development of cancer (1). Prototypically, Helicobacter pylori has been linked to mutagenic free radical formation leading to chronic gastric inflammation and the stimulation of cellular proliferation (2, 3, 4). Mechanisms for H. pylori-induced cellular proliferation and carcinogenesis are under active investigation, but other model systems for complex interactions between bacterial pathogens and host responses, resulting in the development of cellular proliferation and the cancer phenotype are not well described. A variety of pathologic states in squamous tissue injury, including cholesteatoma, chronic burns, and aerodigestive carcinogenesis, are characterized by increased proliferation in response to stimuli that occur during chronic inflammation. One possible reason for this response is NF-κB activation (5).
Many extracellular events can trigger the activation of NF-κB in inflammatory cells. Bacterial LPS was one of the first known activators of B cells and is closely linked to the discovery of NF-κB (6). NF-κB has long been known to be a mediator of the effects of LPS (7). The production of multiple cytokines, such as TNF-α (8), IL-1 (9), IL-6 (10), and GM-CSF (11), is up-regulated in macrophages by NF-κB after LPS stimulation. Reports of bacterial product NF-κB activation in keratinocytes are far less frequent; LPS has been shown to induce the expression of another NF-κB-dependent gene, M-CSF (12). In normal, unstimulated keratinocytes, NF-κB has been shown to inhibit programmed cell death (13). This phenomenon effectively confers protection from premature apoptosis to proliferating benign epidermal cells. To this point, the overexpression of NF-κB’s active subunit, p65, is associated with the production of cytokeratin markers associated with proliferation, such as cytokeratin K6, but mechanisms for proliferative responses of the cells were not investigated (14). Conversely, it has recently been demonstrated that NF-κB inhibition may actually trigger human epidermal invasive cancer, an effect opposite of what might be expected. This result has been observed in other studies (15, 16). Therefore, generalities regarding the effect of NF-κB activation in even a single cell type may be dependent upon the whether the activation is constitutive or induced and may depend on the type of stimulation.
Studies have attempted to elucidate cellular mechanisms and circumstances by which NF-κB is able to mediate changes in cellular proliferation. For example, in a nonepithelial, nonkeratinocyte model system, NF-κB has been shown to induce cyclin D1 gene expression and cellular proliferation by binding to specific sequences in the cyclin D1 promoter (17, 18). Cyclin D1 is an integral protein in cell cycle progression (19). It is synthesized in response to external activators and forms a complex with CDK4 and CDK6 that is then able to phosphorylate the retinoblastoma tumor suppressor protein. Retinoblastoma tumor suppressor protein regulates the transcription of multiple genes that are indispensable for S phase entry. In squamous tissues, cyclin D1 is overexpressed in squamous cell carcinomas of the head and neck (20) and cholesteatoma (21), two epithelial hyperproliferative conditions. Cyclin D1 has been established in vivo as a pure mediator of proliferation in epithelial cells, because its overexpression in the epithelial tissues of transgenic mice results in epidermal hyperproliferation, but has no effect on terminal differentiation (22). The effect of any activators of NFκB on cyclin D1 or cyclin D kinase expression in keratinocytes is unknown.
Cholesteatoma is a squamous disease state characterized by 1) hyperproliferative keratinocyte epithelial in-growth, angiogenesis, and local bony invasion into the middle ear and mastoid cavity; and 2) chronic middle ear infection associated with P. aeruginosa (23). Therefore, it represents one injury response process associated with chronic exposure to bacterial toxins instead. It is also a process that has features in common with aerodigestive carcinogenesis. A large number of NF-κB dependent proinflammatory cytokines, PGs, and growth factors have been identified during squamous tissue inflammatory states, including chronic middle ear infection (24), cholesteatomagenesis (25), and even head and neck cancer (26, 27). Therefore, NF-κB may be an important regulator of events in the squamous inflammatory milieu in a variety of processes. Because P. aeruginosa bacterial products have been linked to NF-κB activation in other epithelial cell experimental models of both chronic and acute inflammatory dysregulation (28, 29), its effects were further studied in this series of experiments.
The purpose of this study was to demonstrate that 1) other models of disorganized epithelial cell proliferation exist during bacterially induced injury exist in addition to H. pylori infection and gastric adenocarcinogenesis; and 2) activation of injury response gene programs by transcription factors such as NF-κB may be responsible for establishing a permissive environment for abnormal cell proliferation during chronic inflammation. Specifically, this study was performed to determine whether P. aeruginosa LPS can induce NF-κB and cell proliferation in a keratinocyte model system as an in vitro model of early response gene activation during squamous inflammation and carcinogenesis and to determine whether this activation results in an NF-κB-dependent, cyclin D1-mediated, induction of proliferation. As a potential therapy for abnormal proliferation, indomethacin was used to examine whether it had any abrogating effect on keratinocyte NF-κB activation and the proliferative response.
Materials and Methods
Rhek-1A (Rhek) human epidermal keratinocytes were used for most studies (provided by Dr. J. S. Rhim, National Cancer Institute, Frederick Cancer Research and Development Center). The Rhek-1A cell line was immortalized by SV40 and is nontumorigenic in nude mice (30). The Rhek cell line has been used in other studies of keratinocyte proliferation, carcinogenesis, and cytotoxicity (31, 32, 33). Rhek cells were used throughout the project as a prototypic squamous epithelial model system. Human epidermal keratinocytes were used to confirm all findings established in the Rhek cells to demonstrate the suitability of Rhek cells for such experiments. The Rhek cells were maintained in EMEM supplemented with 10% FBS and penicillin/streptomycin (50 μg/ml) at 37°C in 5% CO2 as adherent monolayer cultures. Normal human epidermal keratinocytes were purchased commercially from Cambrex. They were cultured according to the instructions of the manufacturer in keratinocyte serum-free growth medium to maintain a proliferative phenotype. This medium is supplemented with 0.1 μg/ml human recombinant epidermal growth factor, 5.0 mg/ml insulin, 0.5 mg/ml hydrocortisone, 50 mg/ml gentamicin, 50 μg/ml amphotericin B, and 7.5 mg/ml bovine pituitary extract (BioWhittaker) at 37°C in 5% CO2
Stable transfection and development of control and dominant negative clones
The cDNA expression plasmid pCMX-IκB-α/M contains a mutation at S36 of the NH2 terminus, and a COOH-terminal PEST sequence mutation and its empty vector pCMX were a kind gift of Dr. Inder Verma (Salk Institute, La Jolla, CA) (34, 35). This stably expressed mutated protein is resistant to extracellular signal-induced phosphorylation and subsequent degradation thus not allowing for activation of NF-κB (36). RHEK cell cultures at 70% confluence were transfected with pCMX-IκB-α/M (2 μg/well) and the pCMV-neo resistant (2 μg/well) or empty vector pCMX (2 μg/well) and pCMV-neo resistant in OptiMEM medium containing 10 μg/ml Lipofectamine (Invitrogen Life Technologies). After 4 h, the medium was removed, and the cells were placed in complete medium for 72 h. Then, cells were placed in selection medium containing G-418 (750 μg/ml). After 14 days, neo-resistant colonies were selected from each well using cloning discs embedded in trypsin and establish as a subcultures.
Preparation of nuclear extracts for EMSAs
Extracts from the cell lines in log-growth phase were prepared according to the methods of Dignam et al. (37) and Lee et al. (38) as follows. Cells were grown in 75-cm2 flasks to 60–80% confluence. Cells were stimulated with 0, 10, 50, or 100 μg/ml P. aeruginosa LPS for 60 min. To determine the potential blockade of NF-κB activation, cells were pretreated with 5 or 20 μg/ml indomethacin in DMSO; the groups not pretreated with indomethacin received DMSO only for 90 min. Cell suspensions were prepared by trypsin-EDTA treatment, and the cells were washed twice in cold PBS and pelleted at 500 × g. The cell pellets were then incubated on ice for 5 min with lysis buffer (10 mM Tris-HCl, 60 mM KCl, 1 mM EDTA, 0.5% Nonidet P-40, fresh 0.1 M DTT, and protease inhibitor). The lysates were microfuged for 5 min at 500 × g. The supernatant was collected and saved as the cytoplasmic extract. Next, an equal volume of nuclear extract buffer (20 mM Tris-HCl, 420 mM NaCl, 0.2 mM EDTA, 25% glycerol, fresh 0.1 M DTT, and PMSF) was added to the remaining nuclear pellet. This was incubated on ice for 10 min, then centrifuged at 15,000 × g for 10 min. The remaining nuclear extracts were aliquoted and stored at −80°C. The protein concentrations of the extracts were determined in triplicate using the bicinchoninic acid modification of the biuret reaction (Protein Assay Kit; Pierce) scaled for microtiter plate analysis. BSA was used as a standard, and the plates were read at 580 nm on a Biotek 311 microtiter plate reader. Standard curves were generated using computer software. The correlation coefficients for the functions were >0.95 in all experiments.
dsDNA oligonucleotide probes for NF-κB and Oct-1 were synthesized commercially (Promega). The consensus sequences used were 5′-AGTTGAGGGGACTTTCCCAGGC-3′ for NF-κB and 5′-TGTCGAATGCAAATCACTAGAA-3′ for Oct-1. Consensus oligonucleotide probes were labeled with T4 polynucleotide kinase (Promega) and [γ-32P]ATP (6000 Ci/mmol; Amersham Biosciences).
Binding reactions were performed using 5 μg of nuclear extract protein incubated for 25 min at 20°C in buffer containing 20 mM HEPES (pH 7.9), 4.6 mM MgCl2, 63 mM KCl, 11% glycerol, 1 mM DTT, 1 μg of poly(dI-dC) (Amersham Biosciences), and 60,000 cpm 32P-labeled nucleotide probe. For unlabeled competitions, extracts were preincubated with a 100-fold excess of each probe before addition of labeled probe. The binding complexes were resolved from reactions on 5% polyacrylamide gels in 0.25× Tris-borate-EDTA buffer at 20°C and run for 90 min at 200 V. The gels were dried and exposed by autoradiography to intensifying screens (Eastman Kodak) for 12–24 h and imaged with an OptiQuant instant imager (Packard Instruments). All reactions were performed with at least three separate nuclear extract preparations. The procedures for nuclear extract preparation were identical for all nuclear extract preparations. Positive controls for NF-κB p65/p50 heterodimers were included (provided by Dr. K. Brown, National Institutes of Allergy and Infectious Diseases, National Institutes of Health). These proteins were provided as whole cell extracts from p50 and p65 transfected COS cells after PMA stimulation.
Ab supershift analyses of the composition of NF-κB were performed. Anti-p65 and anti-p50 NF-κB subunit Abs were purchased from Rockland Laboratories. Supershift reactions were performed under identical binding conditions for both Abs. Supershift Abs (1 μg) were added 30 min before binding to the reaction mixtures with the labeled probes. All EMSA experiments were repeated three times with similar results.
Transient transfection and luciferase assays
The pIgκB Luc reporter construct containing two Ig G-κ chain NF-κB binding sites and the luciferase gene have been previously described (39) and were provided by Dr. K. Brown (National Institutes of Allergy and Infectious Diseases, National Institutes of Health). The CD1 reporter plasmids were gifts from Dr. R. Pestell (Albert Einstein College of Medicine). They contain the cyclin D1 promoter upstream form the luciferase gene as previously described (40). The −1745CD1Luc reporter plasmid contains the full sequence cyclin D1 promoter upstream from the luciferase gene. The −66NF-κBmut plasmid has the cyclin D1 promoter with a site-specific mutation at the NF-κB binding site. Cell line cultures at 70% confluence were cotransfected with luciferase reporter plasmid (2 μg/ml) and a pCMV-βGal reporter construct (0.4 μg/ml) in OptiMEM medium containing 5 μg/ml Lipofectamine (Life Technologies). After 4 h, the medium was removed, and the cells were washed and placed in complete medium overnight. The next day the cells were pretreated with indomethacin at 5 or 20 μg/ml in DMSO or with DMSO alone for 90 min. The cells were then stimulated for 12 h with 10, 50, or 100 μg/ml P. aeruginosa LPS. After stimulation, the relative luciferase activity was determined with the Dual Light reporter gene assay (Tropix) and a Monolight 2010 plate luminometer (Analytical Luminescence Laboratories) according to the manufacturers’ instructions. All reporter gene experiments were repeated three times with similar results.
Cell proliferation assays
Cell count experiments for Rhek cells were performed by culturing the cells in T-25 flasks at an initial concentration of 35,000 cells/flask. Cells were then pretreated for 90 min with 5 or 20 μg/ml indomethacin in DMSO or in DMSO (0.1%) alone. Cells were stimulated with 10, 50, and 100 μg/ml P. aeruginosa LPS and allowed to grow for 72 h at 37°C in 5% CO2. The cells were then detached from the flask with trypsin-EDTA solution. Cell suspensions were counted using a standard hemocytometer. Each experimental group was repeated in duplicate.
MTT assays (Roche) were performed in 96-well microtiter plates. Rhek cells and normal human epidermal keratinocytes were plated in appropriate medium at a concentration of 5000 cells/well and allowed to grow overnight. The next day, the cells were pretreated with 5 or 20 μg/ml indomethacin in DMSO or in DMSO (0.1%) alone for 90 min before stimulation with 10, 50, or 100 μg/ml P. aeruginosa LPS. Each experimental condition was performed with six replicates in serum-free medium. Cells were then incubated for 3 days at 37°C in 5% CO2. MTT reagent was added to each well, and the plates were reincubated at 37°C in 5% CO2 for 4 h. The resultant cellular precipitant formazan dye was then solubilized with MTT detergent. After overnight incubation at room temperature in the dark, the absorbance was measured using a Biotek microplate reader at 570 nm. All MTT experiments were repeated three times with similar results.
Thymidine incorporation assays
Ten thousand cells per well were seeded in a 96-well plate in appropriate serum-free medium for 24 h. The next day, cells were treated with four serial dilutions of 100 μg/ml P. aeruginosa LPS. Each experimental condition was performed with six replicates in serum-free medium. The cells were then incubated for 24 h at 37°C in 5% CO2. One microcurie of [3H]thymidine in 500 μl was added to each well, and the cells were incubated for 8 h at 37°C in 5% CO2. Medium was the removed, and the cells were washed with 500 μl of ice-cold 1× PBS. Ice-cold 5% TCA (500 μl) was then added to each well, and the plates were left at 4°C for 30 min. The TCA solution was then aspirated, and each well was again washed with 500 μl of ice-cold 1× PBS. At room temperature, 250 μl of 0.5 N NaOH/0.5% SDS was added to each well, solution was pipetted up and down, and each well was counted with a 96-well scintillation counter.
Western blots for cyclin D1 were performed on the cytoplasmic extracts (see above for method of extraction) of the cells exposed to 0, 10, 50, or 100 μg/ml P. aeruginosa LPS for 24 h. After a bicinchoninic acid protein assay, as described above, 20 μg from each sample was incubated with an equal volume of sample buffer for 5 min at 95°C and then loaded onto a 10% acrylamide gel and run by electrophoresis at 20 mAmp for 60–90 min. The protein was transferred onto a nitrocellulose membrane using a semidry gel transfer unit. The membrane was blocked for 1 h, and then primary cyclin D1 Ab (Santa Cruz Biotechnologies) was added overnight. The next morning, the membrane was washed and incubated with secondary Ab for 1–5 min. The horseradish fluoroilluminescence detection protocol was followed. The blots were exposed to Bio-Max film for 10 min and developed.
Similarly, Western blots were performed for IκB-α. Cell lines were grown to 70–80% confluence in T-25 flasks in appropriate medium. The cells were then pretreated with 20 μg/ml indomethacin in DMSO or in DMSO alone for 90 min. Cells were then stimulated for 3 h with 100 μg/ml P. aeruginosa LPS. Cytoplasmic extracts were then prepared as described above. Western blots were performed as described above using primary IκB-α Abs (Santa Cruz Biotechnologies).
The statistical difference between experimental and control groups for all experiments was determined by two-sided Student’s t test. The significance level was set at p < 0.05.
Activation of NF-κB in keratinocyte cell lines with P. aeruginosa LPS
To explore whether NF-κB can be induced by LPS in keratinocytes, the function and composition of NF-κB binding proteins were determined in Rhek cells and normal human epidermal keratinocytes by luciferase reporter assay and EMSA after stimulation with increasing doses of P. aeruginosa LPS. Fig. 1,A demonstrates a statistically significant 2.25- and 3.1-fold functional induction of NF-κB in Rhek cells after 60-min stimulation with 10 and 100 μg/ml LPS, respectively. Additional analysis in this cell line by EMSA demonstrated differences in nuclear NF-κB binding activity consistent with the functional activation shown with the luciferase assay (Fig. 1,B). Competition experiments with unlabeled, wild-type oligonucleotide probe in a 100-fold excess confirm that the binding activity detected in Fig. 1,B was specific for NF-κB. Integrity and equal protein loading amounts of the nuclear extracts were demonstrated by performing Oct-1 DNA binding assays, showing similar Oct-1 binding across experimental groups. To elucidate the specific nature of the EMSA bands, competition experiments were conducted with 1:1, 1:10, 1:25, and 1:50 molar ratios of labeled to unlabeled specific NF-κB oligonucleotides and nonspecific PPAR-γ oligonucleotides (Fig. 1 C). The NF-κB band disappeared with increasing specific cold competition, but did not with increasing nonspecific competition.
The composition of the protein binding to NF-κB was examined by supershift analysis (Fig. 1,D). Nuclear proteins from Rhek cells hybridized with NF-κB consensus oligonucleotides were incubated for 60 min with anti-p65 and anti-p50 Abs and compared by EMSA. Fig. 1 C demonstrates a supershift with both anti-p65 and anti-p50 Abs. Control p50/p65 heterodimer proteins were run side-by-side with the experimental extract, confirming the nature of the EMSA bands.
To determine whether the functional activation of NF-κB seen in the transformed keratinocytes was replicable in normal human keratinocytes, the above experiments were confirmed in normal keratinocytes. Much like in Rhek cells, luciferase reporter assays showed a statistically significant 1.75- and 3.98-fold functional activation of NF-κB in normal keratinocytes with 10 and 100 μg/ml LPS, respectively (Fig. 2,A). Again, the EMSA binding experiments showed similar increases in nuclear NF-κB roughly corresponding to the functional activation seen with the luciferase assays (Fig. 2,B). The integrity of nuclear extracts was also demonstrated in the normal keratinocytes with EMSA for Oct-1 binding (Fig. 2 B). Previously there have been little data linking pathologically important bacterial LPS with NF-κB-associated pathologic states in squamous tissues. We have demonstrated that P. aeruginosa LPS is able to activate NF-κB in two separate keratinocyte systems.
Inhibition of P. aeruginosa-induced NF-κB activation with indomethacin in keratinocyte cell lines
We hypothesized that anti-inflammatory agents may play a therapeutic role in pathologic states of squamous tissues associated with increased proliferation. Therefore, we tested indomethacin, a prototypic, clinically useful, anti-inflammatory drug, as an inhibitor of the observed LPS induction of NF-κB in keratinocytes. Fig. 3 A depicts luciferase reporter assays in Rhek cells, again demonstrating similar 1.68- and 3.06-fold functional activation of NF-κB with 50 and 100 μg/ml LPS. Experimental groups treated for 90 min with 5 and 20 μg/ml indomethacin before stimulation with 100 μg/ml LPS showed significant down-regulation of NF-κB activation. Treatment with 5 μg/ml indomethacin before maximal LPS stimulation reduced NF-κB activation to levels similar to the induction produced by the smaller LPS dose. Pretreatment with 20 μg/ml indomethacin before maximal LPS stimulation decreased the activity of NF-κB to levels below constitutive activity.
EMSA binding assays confirmed the luciferase findings (Fig. 3 B). Again, increases in nuclear NF-κB DNA binding were seen with LPS stimulation, roughly corresponding to the functional activation demonstrated with the luciferase assays. Pretreatment with indomethacin reduced this nuclear NF-κB DNA binding to below constitutive levels using the maximum indomethacin dose. Oct-1 DNA binding experiments confirmed the integrity and equal loading of the nuclear extracts.
Similar experiments were performed in normal human epidermal keratinocytes. As in the Rhek cells, indomethacin was able to reduce the significant induction of NF-κB after LPS stimulation. Again, pretreatment with 20 μg/ml indomethacin was able to reduce NF-κB activity to below constitutive levels (Fig. 3,C). EMSA binding assays confirmed the luciferase assay findings in the normal keratinocytes by demonstrating an obvious decrease in nuclear NF-κB binding with indomethacin pretreatment (Fig. 3 D). These results indicate that indomethacin has the capacity to block both induced and constitutive NF-κB activities in human keratinocyte cell lines.
Western blot for IκB-α in Rhek cells after LPS stimulation and indomethacin pretreatment
Under normal conditions NF-κB is present in the cytoplasm bound to IκB-α, rendering it inactive. Upon extracellular stimulation, IκB-α is phosphorylated, leading to its dissociation from NF-κB subunits (41, 42). IκB-α is then degraded in a proteosome-dependent mechanism (43), and NF-κB becomes activated, translocating into the nucleus. A Western blot was performed to determine whether the observed indomethacin inhibition of NF-κB occurs through prevention of IκB-α degradation. The presence of IκB-α was determined after maximal stimulation (100 μg/ml) with LPS and after pretreatment with 20 μg/ml indomethacin.
IκB-α disappeared in Rhek cells after 3-h stimulation with LPS. Ninety-minute treatment with 20 μg/ml indomethacin before LPS stimulation was able to prevent the disappearance of IκB-α (Fig. 4). These results suggest that indomethacin prevents the release of p50 and p65 from IκB-α, thus blocking the stimulatory effect of LPS.
Cell proliferation assays
Experiments were performed to determine whether the functional activation of NF-κB seen in the keratinocyte cell lines correlated with an increased proliferative response. First, MTT assays were performed on Rhek cells. Fig. 5,A demonstrates a significant increase in cell proliferation rates 3 days after stimulation with 50 and 100 μg/ml P. aeruginosa LPS. A similar significant increase in proliferation after 3 days of stimulation with LPS was seen in normal keratinocytes (Fig. 5,B). To determine whether indomethacin is able to block this increase in cell proliferation, the experiments were repeated by pretreating Rhek cells and normal keratinocytes with 5 and 20 μg/ml indomethacin before maximal LPS stimulation (100 μg/ml). Much as indomethacin is able to inhibit the activation of inducible and constitutive NF-κB, it causes a significant decrease in keratinocyte proliferation (Fig. 5, C and D). Other experiments, using p65-overexpressing Rhek cell clones, showed increased proliferation rates similar to those seen with LPS stimulation, as shown in the above experiments (data not shown). This finding coupled with the finding that an NF-κB inhibitor, such as indomethacin, is able to block the LPS-induced increase in keratinocyte proliferation leads us to suspect that the mechanism by which LPS causes the increased keratinocyte cellular proliferative response is by NF-κB activation.
Finally, the MTT assay results were confirmed by cell count experiments (Fig. 5 E). Rhek cells were plated in T-25 flasks at an initial concentration of 35,000 cells/flask, then stimulated with increasing doses of LPS with or without indomethacin pretreatment in serum-free medium. The cells were allowed to grow for 3 days in culture. Again, there was a marked increase in the flasks in which the cells had been treated with LPS. Indomethacin was able to block this increase.
Inhibition of NF-κB activation and cellular proliferation in pCMX-IκB-α/M dominant negative clones
Because the actual mechanism for inhibition of NF-κB by indomethacin is not known, more direct inhibition of NF-κB with a dominant negative IκB construct was performed. Rhek cells (Fig. 6,A), Rhek cell clones stably transfected with pCMX empty vector (Rhek-pCMX; Fig. 6B), and Rhek cell clones stably transfected with pCMX-IκB-α/M plasmids (Rhek-pCMX-IκB-α/M) were stimulated with 0, 5, 10, and 25 μg/ml LPS for 60 min before nuclear extraction. EMSA experiments were then performed as described above. These gel results are depicted in the top panels of Fig. 6. Note how Rhek (Fig. 6,A, top panel) and Rhek-CMX (Fig. 6,B, top panel) cells demonstrated activation of nuclear NF-κB with increasing doses of LPS as expected. Rhek-pCMX-IκB-α/M (Fig. 6 C, top panel) cells, in contrast, showed little constitutive or stimulated NF-κB activity.
MTT assays for cellular proliferation after stimulation with 0, 5, 10, and 25 μg/ml P. aeruginosa LPS for 72 h were also performed in these cells as described above. These proliferation results are depicted in the bottom panels of Fig. 6. Rhek (Fig. 6,A, bottom panel) and Rhek-CMX (Fig. 6,B, bottom panel) showed statistically increased proliferation as measured by MTT assay compared with the dominant negative Rhek-pCMX-IκB-α/M (Fig. 6 C, bottom panel) clones, which did not.
Thymidine incorporation assays in Rhek and pCMX-IκB-α/M clones
To confirm the results found with the MTT assays, thymidine incorporation assays for cellular proliferation were performed on Rhek cells and pCMX-IκB-α/M dominant negative clones. Fig. 7,A demonstrates a statistically significant increase in irradiated thymidine incorporation after simulation with P. aeruginosa LPS in Rhek cells. In the dominant negative cell clones, however (Fig. 7 B), no such increase in irradiated thymidine incorporation was noted. These results demonstrate that with direct, specific NF-κB inhibition, the increases in cellular proliferation due to LPS stimulation do not occur.
Cyclin D1 luciferase reporter gene assays
We hypothesized that the LPS-induced increase in cell proliferation was mediated through an NF-κB-dependent pathway. NF-κB is known to have binding sequences in the cyclin D1 promoter and is a mediator of cyclin D1 effects in mesodermally derived transformed cell lines (17, 44). To investigate this hypothesis, we transiently transfected Rhek cells with a cyclin D1 promoter luciferase plasmid and stimulated the transfected cells with 100 μg/ml P. aeruginosa LPS for 60 min. Fig. 8,A shows a significant increase in cyclin D1 promoter activity in Rhek cells after LPS stimulation. LPS is unable to significantly cause this activation of the cyclin D1 promoter in the presence of indomethacin. Because we have shown that indomethacin blocks the LPS-induced activation of NF-κB, it is likely that this increase in cyclin D1 promoter activity is a reflection of NF-κB activation and binding. To further confirm this suspicion, we repeated the experiment, this time transiently transfecting Rhek cells with a cyclin D1 promoter luciferase plasmid with a mutation at the NF-κB binding site (−66CD1Luc). Fig. 8 B shows that in these cells, LPS was unable to induce activation of the cyclin D1 promoter. Thus, when NF-κB is prevented from binding the cyclin D1 promoter, P. aeruginosa LPS is unable to activate the cyclin D1 promoter. This confirms our suspicion that LPS is activating cyclin D1 in Rhek cells via its up-regulation of NF-κB functional activity. This finding probably explains the increase in cellular proliferation that we found in keratinocytes after P. aeruginosa stimulation.
Finally, we performed a Western blot, which confirmed the up-regulation of cyclin D1 in Rhek cells 24 h after treatment with LPS (Fig. 8 C). In the presence of indomethacin (and NF-κB inhibition), cyclin D1 was not up-regulated.
It has been established that bacterial products have the capacity to stimulate NF-κB in inflammatory cells and that these activations are important in host humoral immunity. There are few data linking pathologically important bacterial products with injury response and abnormal proliferation in squamous tissues. In this study we demonstrate that NF-κB is activated in both normal and transformed keratinocytes after exposure to a pathologically important bacterial product, specifically P. aeruginosa LPS. This activation correlates with a significant increase in NF-κB-dependent cell proliferation, because inhibition of NF-κB reverses the proliferative response. LPS activates the cyclin D1 promoter in keratinocytes via the induction of NF-κB binding to the cyclin D1 promoter, a mechanism probably responsible for the observed proliferative response. Finally, a prototypic anti-inflammatory drug, indomethacin, has the capacity to inhibit both the proliferative process as well as the molecular events (NF-κB activation of cyclin D1 promoter) underlying it. NF-κB-mediated activation of cyclin D1 has not been described in epithelially derived cells.
Bacterial toxins have been implicated as mediators of nonepithelial cell proliferation in other systems (45, 46). It is not well established that bacterial products can induce a proliferative phenotype in squamous epithelia. In fact, outside of the field of H. pylori-induced proliferation of gastric mucosa, findings of bacterial products directly contributing to pathologic proliferative states are rare. Our findings have implications for the pathogenesis of a multitude of inflammatory associated squamous disease processes. The activation of NF-κB with bacterial products (such as P. aeruginosa LPS) may trigger a proinflammatory and hyperproliferative cellular phenotype that, when deranged, can lead to neoplasia (such as cholesteatoma) or even contribute downstream gene products to the chronic inflammatory milieus that leads to malignancy (47). H. pylori infection induces cell proliferation (48) and overexpression of IL-8 and cyclooxygenase-2 in gastric epithelial cells (49), both events that may play a causative role in gastric carcinoma. Similarly, NF-κB-mediated proinflammatory dysregulation has been associated with both the development of pancreatic cancer (50) and the progression to metastases (51).
The production of cytokines, angiogenic factors, proliferation markers, cell cycle activators, and apoptosis resistance are all NF-κB-mediated cellular events that can contribute to abnormal cellular proliferation and even the malignant phenotype (52, 53). Constitutive NF-κB activation has been shown to be associated with squamous proliferation and tumor growth (27) and resistance to apoptosis (53). Drugs that inhibit NF-κB may serve as potential therapeutic agents in these abnormal proliferative processes (52, 53, 54, 55). Antioxidants (56), phytochemicals (57), and anti-inflammatory mediators (58) may exert some of their antiproliferative functions through suppression of NF-κB. Our finding of NF-κB inhibition with indomethacin suggests that it may also have a potential therapeutic role in abnormal proliferative states of squamous tissues, including cancer progression. In this matter, indomethacin has been touted as a potential novel chemotherapeutic agent for oral cancer by other investigators (58).
The mechanisms by which bacterial LPS is able to mediate intracellular proinflammatory responses, such as NF-κB nuclear translocation, have been partially delineated. Genetic studies and gene transfer experiments indicate that LPS intracellular effects are mediated through TLR signaling (59, 60). TLRs are a family of transmembrane proteins with a leucine-rich extracellular domain and an IL-1R-like cytoplasmic homologue domain (61). They appear to be the main intracellular signaling molecules of CD14, an extracellular glycosylphophatidylinositol membrane-anchored protein known to bind to LPS (60, 61). Together, TLR and CD14 form the LPS-receptor complex. The most important TLR family member involved in Gram-negative LPS-receptor complexes appears to be TLR4 (62). Another TLR family member, TLR2, has been reported to signal Gram-positive bacterial lipoprotein-mediated intracellular events (63). Although how TLRs mediate innate immune responses in leukocytes has been studied extensively (64), investigations dealing with their presence and function in keratinocytes are few. Recently, it has been shown that human cornea keratinocytes express functionally active TLR4 (65). Additional studies have shown that cultured skin keratinocytes also express active TLR4 that, when bound to LPS, induces NF-κB nuclear translocation and proinflammatory cytokine secretion (62). Immunohistochemical profiling studies have shown cell membrane TLR4 expression in the midepidermal layers of normal human skin (66). To this end, we performed initial studies in which Rhek cells were grown to 80% confluence directly on slides and subsequently fixed with 4% paraformaldehyde. Immunohistochemical analyses failed to demonstrate the presence of TLR4 in these cells (data not shown). It is possible that immunohistochemistry may not be sensitive enough to detect TLR4 in cultured cells. Future studies using flow cytometry techniques should more sensitively determine whether TLR4 are involved in NFκB signaling processes in cultured keratinocyte systems.
Cholesteatomagenesis is a prototypic state of disorganized squamous proliferation and invasion, although cholesteatomas do not metastasize nor are they genetically unstable. Cholesteatomas almost universally arise in the setting of inflammation and infection, and their formation is related to internal molecular dysregulation and external stimuli from Pseudomonas and other bacterial toxins, proinflammatory cytokines, and growth factors. Inflammatory granulation tissue always appears with invading epithelium in active human cholesteatomas and experimental animal cholesteatomas (25). Many of the proinflammatory mediators and cytokines seen in otitis media and cholesteatoma (24) are under transcriptional control of NF-κB. TNF-α, IL-1, and IL-6 are among these proinflammatory molecules. The most common pathogen in cholesteatoma is P. aeruginosa (23). Our studies demonstrate that P. aeruginosa LPS induces NF-κB activation in human keratinocytes, causing a significant increase in proliferation, a phenomenon likely to contribute to cholesteatoma pathogenesis and behavior.
In conclusion, we have shown that P. aeruginosa LPS up-regulates NF-κB and cellular proliferation in human keratinocytes and RHEK cells. The increase in proliferation is probably mediated by cyclin D1 in an NF-κB-dependent fashion. Indomethacin, a prototypic, clinically useful, anti-inflammatory drug abrogates these events. Clearly, future studies will focus on the degree to which the squamous phenotype is altered by these bacterial toxin-mediated events on the continuum of normal to malignant keratinocyte growth. It is unlikely that LPS alone would result in the conversion of squamous cells to fully malignant cells. However, other products generated in the milieu, in addition to underlying cellular genetic abnormalities, might easily constitute a permissive environment for very disordered growth of cells in a variety of pathologic states, including the development of malignancy. The development of anti-inflammatory strategies in such processes is of obvious benefit.
The authors have no financial conflict of interest.
We acknowledge Drs. Simon Wright and Richard Pestell for technical advice and support.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Lion’s 5M Research Foundation.