Endotoxin Up-regulates Interleukin-18: Potential Role for Gram-Negative Colonization in Sarcoidosis

Sarcoidosis is a multisystem disease of unknown etiology characterized by a granulomatous inflammatory process (1, 2). It is generally assumed that the pathogenesis of the disease results from genetically susceptible hosts being exposed to particular environmental stimuli. Postulated environmental stimuli include infectious agents and noninfectious environmental factors. The fact that there have been numerous reports of familial clustering of this disease indicates that genetic predisposition plays an important role (3). The presence of noncaseating granulomas is a feature of sarcoidosis and 90% of patients have pulmonary symptoms (4, 5). The primary manifestation of this disease is an accumulation of mononuclear inflammatory cells, mostly activated CD4^sup +^ type 1 helper T lymphocytes (Th1 cells), which produce interleukin (IL)-2 and interferon (IFN)-? (6-9).

IL-18 plays an important role in the Th1 response by upregulating IFN-? in synergy with IL-12 (10, 11). IL-18 is primarily a monocyte/macrophage-derived cytokine (12) but is also expressed in dendritic cells, Kupffer cells, keratinocytes, synovial fibroblasts, epithelial cells, and osteoblasts (13). IL-18 and its receptor have been shown to be present at elevated levels in the lungs of patients with sarcoidosis (14-17). It is also an important factor in the regulation of IL-2 expression by pulmonary CD4^sup +^ T lymphocytes through AP-1- and nuclear factor-?B-mediated signaling pathways (17).

Mutations in the regulatory regions of cytokine genes have been associated with susceptibility to an increasing number of complex diseases. One such mutation is found in the IL-18 5'-untranslated region (UTR) promoter (18). The single-nucleotide polymorphism (SNP) -607^sup C/A^ is predicted to be located at a binding site for cAMP-responsive element-binding protein (CREB) (19). At this position a change from cytosine (C) to adenine (A) will disrupt a putative CREB-binding site. Previous studies have shown that this SNP is associated with diabetes, necrotizing enterocolitis, rheumatoid arthritis, the development of postinjury sepsis, and also with sarcoidosis (20-24).

To further understand the role of IL-18 in sarcoidosis we examined the effect of various stimuli and genetic factors on the expression of IL-18. We found that incubation of monocytes with active sarcoid epithelial lining fluid (ELF) increased IL-18 levels significantly, whereas incubation with ELF from patients who had recovered from sarcoidosis (recovered patient ELF) did not. We examined IL-18, IL-12. and IFN-? levels in active sarcoid ELF, recovered patient ELF, and control subjects and found elevated cytokine concentrations in patients with active sarcoidosis. The most striking data were observed during endotoxin analysis of ELF; it was found that sarcoid ELF contained significantly higher levels than did normal subject or recovered patient ELF. The addition of polymyxin B to sarcoid ELF abrogated IL-18 activity. To further elucidate this we demonstrated the presence of gram-negative bacterial DNA in sarcoid bronchoalveolar lavage (BAL) fluid and further identified Haemophilus influenzae and Moraxella catarrhalis as the gram-negative bacteria present. These observations suggest a potential role for a gram-negative bacteria-driven IL-18 response in sarcoidosis and further characterize the Th1 response in this disorder. Some of the results of these studies have been previously reported in the form of an abstract (25, 26).

METHODS

Study Population

Informed consent for BAL fluid collection, using a protocol approved by the Ethics Committee of Beaumont Hospital (Dublin, Ireland), was obtained from 11 patients with biopsy-proven stage I (54.5%) or stage II (45.5%) sarcoidosis, 6 patients recovered from sarcoidosis, and 11 non-smoking control subjects. Patient and control subject data are presented in Table 1. Patients recovered from sarcoidosis had clinical and radiologic data consistent with quiescence of active disease, having recovered either spontaneously or after treatment with prednisolone. and had a mean duration of disease of 17.6

A homogeneous Irish population was investigated for the -607^sup A/C^ mutation: 94 patients with sarcoidosis (stage 0, I, II, or III) and 97 healthy control subjects (male:female ratio, 61:33 and 42:55, respectively: mean age ± SEM, 39 ± 3.6 and 75 ± 4.0 yr, respectively).

BAL fluid from nine individuals with chronic obstructive pulmonary disease (four males and five females; FEV^sub 1^, 50 ± 6% predicted) was also evaluated for IL-18 levels.

Cell Culture

Human THP-1 monocytic cells were cultured in Roswell Park Memorial Institute 1640 containing 10% fetal calf serum, 2 mM L-glutamine, penicillin, and streptomycin (Life Technologies, Paisley, UK).

Promoter Analysis

THP-1 cells (1 × 10^sup 6^) were transiently transfected with a human IL-18 5'-UTR-pGL3 (hIL-18 5'-UTR-pGL3) luciferase plasmid (18); cells were left untreated or stimulated with tuberculin-purified protein derivative (PPD, 5 TU; Statens Serum Institut, Copenhagen, Denmark), beryllium sulfate (BeSO^sub 4^, 100 µM), aluminum sulfate (Al^sub 2^[SO^sub 4^]^sub 3^, 25-100 µM), zirconium sulfate (Zr[SO^sub 4^]^sub 3^, 25-100 µM), or LPS (as indicated, Pseudomonas aeruginosa; Sigma-Aldrich, Dublin, Ireland), and/or 10 µl of sarcoid or recovered patient ELF. Twenty-four-hour poststimulation reporter gene expression was determined (29). See the online supplement for additional details.

IL-18 mRNA Analysis

IL-18 transcripts were quantified by reverse transcriptase-PCR (RT-PCT) (29), using primers 5'-TGGCTGCTGAACCAGTAGAA-3' and 5'-ATAGAGGCCGATTTCCTTGG-3' (IL-18) and primers 5'-AAC TCT GGT AAA GTG GAT-3' and 5'-TAC TCA GCG CCA CCA GCA TGC-3' (glyceraldehyde-3-phosphate dehydrogenase). See the online supplement for additional details.

IL-18 Protein Production

IL-18 was measured in supernatants of THP-1 cells (1 × 10^sup 6^/well) by ELISA (R&D Systems, Abingdon, UK). See the online supplement for additional details.

Endotoxin Analysis

LPS activity was measured in BAL fluid samples, using a quantitative chromogenic Limulus amebocyte assay (QCL-1000; Cambrex, Wokingham, UK). See the online supplement for additional details.

Genotyping the -607^sup C/A^ Polymorphism

DNA extracted from peripheral blood leukocytes, using the Wizard Genomic DNA purification kit (Promega, Southampton, UK), was used as a template for allele-specific PCR to genotype the -607^sup A/C^ polymorphism as described previously (18). See the online supplement for additional details.

Bacterial Species-specific PCR

DNA was extracted from BAL fluid by phenol-chloroform extraction and ethanol precipitation. Analysis of DNA for 16S rRNA was described previously (30). Species-specific PCR for H. influenzae (31), M. catarrhalis (32), Pseudomonas spp. (33), Legionella pneumophilia (34), and Chlamydia spp. (35) were described previously (Table 2). See the online supplement for additional details.

Statistical Analysis

Data are expressed as means ± SEM. p values less than 0.05 were considered to be significant (Prism 3.03; GraphPad Software, San Diego, CA). ?^sup 2^ and Fisher exact tests were used to calculate odds ratios and 95% confidence intervals for mutation frequencies in patients with sarcoidosis versus control subjects. Both sarcoidosis and control populations were tested for conformity to exact Hardy-Weinberg equilibrium test for genotype counts at biallelic loci available at the NHLBI Innate Immunity Programs for Genomic Applications (see http://innateimmunity.net/IIPGA2/Bioinformatics/exacthweform).

RESULTS

Cytokine Levels in Sarcoidosis and Recovered Patient ELF

We have previously reported elevated levels of IFN-?, IL-12, and IL-18 in sarcoid versus recovered patient ELF. To determine whether this is a feature of the granulomatous process or an idiosyncratic feature of the individuals with sarcoidosis, we measured ELF cytokine levels in active and recovered sarcoid patients. IFN-?, IL-12, and IL-18 levels were all significantly (p < 0.05) lower in recovered individuals compared with levels in the same individuals with active disease (Figures 1A-1C). We also evaluated IL-18 levels in ELF from nine individuals with chronic obstructive pulmonary disease, a condition associated with gram-negative infection. Levels of IL-18 in chronic obstructive pulmonary disease ELF were similar to active sarcoid levels (1,001 ± 267 and 1,262 ± 363 pM, respectively).

Effect of Sarcoid and Recovered Patient ELF on IL-18 Levels in THP-1 Cells

To determine whether activity of the hIL-18 5'-UTR promoter region was directly activated by sarcoid ELF, THP-1 cells were transiently transfected with a pGL3-5'-UTR-luciferase construct. Compared with unstimulated cells, luciferase activity was threefold higher after stimulation with sarcoid ELF (Figure 2A). When recovered patient ELF was used as the stimulant, no induction of hIL-18 5'-UTR activity was evident. Semiquantitative RT-PCR and ELISA were employed to quantify the effects of sarcoid ELF on IL-18 gene and protein expression in THP-1 cells. Sarcoid ELF significantly increased IL-18 gene expression (Figure 2B). A similar effect was seen on IL-18 protein expression (Figure 2C; 27.4 ± 0.29, 302.7 ± 35.07, and 133.7 ± 27.3 pg of IL-18 per milliliter for control, sarcoid, and recovered ELF, respectively).

Effect of Purified Protein Derivative of Mycobacterium tuberculosis, Beryllium, Aluminum, Zirconium, and LPS on IL-18 Activity

Given that the formation of granulomas also occurs in tuberculosis and chronic beryllium disease, we examined the effect of PPD of M. tuberculosis and BeSO^sub 4^ on IL-18 expression. THP-1 cells were transiently transfected with an hIL-18 5'-UTR-luciferase construct and stimulated with PPD (5 TU), BeSO^sub 4^ (100 µM), or LPS (10 µg) for 24 h. Both BeSO^sub 4^ and LPS significantly upregulated promoter activity (Figure 3A) and also induced IL-18 gene and protein expression in THP-1 cells (Figures 3B and 3C). The BeSO^sub 4^ effect was not inhibited by polymyxin B (data not shown). PPD stimulation did not increase hIL-18 5'-UTR promoter activity (Figure 3A). However, PPD did induce IL-18 mRNA and protein expression in THP-1 cells (Figures 3B and 3C). Dose-response experiments using Al^sub 2^(SO^sub 4^)^sub 3^ (25-100 µM) or Zr(SO^sub 4^)^sub 3^ (25-100 µM) as agonist also increased IL-18 protein production from THP-1 cells (Figures 3D and 3E). with Al^sub 2^(SO^sub 4^)^sub 3^ having a considerably more potent effect than all other agonists.

Determination of Endotoxin Levels in BAL Fluid

Given the potent effect of LPS on IL-18 promoter activity and gene and protein expression, we quantified endotoxin levels in sarcoid, recovered patient, and normal subject BAL fluid. Sarcoid BAL fluid was found to contain a notably higher level of endotoxin (26.0 ± 3.7 EU) compared with recovered patient BAL fluid (7.7 ± 4.3 EU) and healthy control samples (13.5 ± 3.0 EU; Figure 4A). The physiologic level of endotoxin in the sarcoid samples equated to 52 ± 6 ng of LPS per milliliter. The addition of polymyxin B, which binds to and neutralizes LPS, to sarcoid BAL fluid significantly (p < 0.001) reduced its effect on hIL-18 5'-UTR promoter activity in vitro (Figure 4B). As a control we determined the effect of polymyxin B on non-LPS-induced IL-18 expression. As shown in Figure 4C, polymyxin B had no inhibitory effect on BeSO^sub 4^-induced IL-18 protein production (53 ± 2.3 and 51 ± 0.6 pg of IL-18 per milliliter for BeSO^sub 4^ and BeSO^sub 4^ plus polymyxin B, respectively).

Genetic Analysis of the hIL-18 5'-UTR -607^sup C/A^ Mutation

The hIL-18 5'-UTR -607^sup C/A^ polymorphism disrupts a putative CREB-binding site. A Swedish study revealed that patients with multiple sclerosis and homozygous for cytosine at position -607 had higher levels of IL-18 mRNA compared with patients homozygous for adenine (18). Genotype and allelic frequencies of the -607^sup C/A^ SNP were therefore examined in a homogeneous Irish sample population of 94 patients with sarcoidosis and 97 healthy control subjects. Genotype frequencies were in agreement with Hardy-Weinberg equilibrium. Figure 5A shows the genotypes of three patients: Patient 1 is homozygous for the C allele and Patients 2 and 3 are -607^sup A/C^ heterozygotes. Of the 94 patients with sarcoidosis, 51 were homozygous for the C allele (54.3%), 5 patients were homozygous for the A allele (5.3%), and 38 patients were heterozygotes (40.4%; Figure 5B). The C allele was significantly increased in the sarcoidosis population (p = 0.029; odds ratio, 1.96; 95% confidence interval, 1.258-3.052) compared with the control population. A statistically significant increase was observed for -607^sup C/C^ in the sarcoidosis group (p = 0.0265; ?^sup 2^, df = 7.259, 2). The -607^sup A/A^ genotype was noted as the rare genotype in this sample population as only 5 patients (5.3%) with sarcoidosis and 13 control individuals (13.4%) were homozygous for the A allele.

Effect of the hIL-18 5'-UTR -607^sup C/A^ SNP on Promoter Activity

Our previous experiments examining hIL-18 5'-UTR promoter activity used a plasmid that carried an adenine at position -607. Given that the -607^sup C^ allele was the most predominant in the sarcoidosis group, it was important to determine whether the -607^sup C/A^ SNP leads to a functional effect on IL-18 activity. THP-1 cells were transiently transfected with a pGL3 5'-UTR-luciferase construct that carried either the A or C allele at position -607. After stimulation with physiologic levels of LPS (50 and 100 ng) promoter activity was significantly (p < 0.001) increased (Figure 5C) to levels similar to those induced by 10 µg of LPS (Figure 3A). There was, however, no notable difference between alleles after stimulation with 50 ng of LPS (-607^sup A^ allele, 1.8 ± 0.06 LUVg protein; -607^sup C^ allele, 2.25 ± 0.02 LU/µg protein) or 100 ng of LPS (-607^sup A^ allele, 2.8 ± 0.21 LU/µg protein; -607^sup C^ allele, 2.75 ±0.17 LU/µg protein).

16S Ribosomal RNA and Bacterial Species-specific PCR

Because of the significantly higher levels of endotoxin present in sarcoid BAL fluid, we sought to determine from what gramnegative bacterial species the endotoxin was derived. To do this we performed 16S rRNA-specific PCR, using primers that amplify a region of the conserved 16S ribosomal RNA gene common to all bacteria. Amplification of a product indicates the presence of bacterial DNA in a sample. We detected the presence of bacterial DNA in 19 of a total of 26 active sarcoidosis BAL fluid samples (73%). To identify the species from which this bacterial DNA was derived, species-specific PCR was employed. DNA from H. influenzae and M. catarrhalis was detected in 69 and 54% of the sarcoid BAL fluid samples, respectively (Figure 6). DNA from Pseudomonas, Chlamydia, or Legionella was not detected. PCR analysis of recovered patient BAL fluid did not detect any bacterial DNA. Analysis of normal BAL fluid detected bacterial DNA in 46.6% of samples; however, this was not derived from any of the species tested above.

DISCUSSION

The current study illustrates that bacterial endotoxin may play an important role in the regulation of IL-18 in sarcoidosis. Sarcoid ELF, but not recovered ELF, induced IL-18 5'-UTR promoter activity and in turn induced IL-18 mRNA and protein production in THP-1 cells. Depletion of LPS by the addition of polymyxin B impaired this effect. Determination of endotoxin levels in sarcoid, recovered patient, and normal BAL fluid samples showed a higher level of LPS present in sarcoid BAL fluid. A subsequent PCR-based method identified the presence of H. influenzae and M. catarrhalis DNA in sarcoidosis BAL fluid. These findings establish a potential role for a gram-negative bacteria-driven IL-18 response in sarcoidosis.

Since sarcoidosis was first described in 1877, much controversy surrounds the elusive causative agent of this disease (36). Exposures to inorganic and organic environmental moieties, such as clay soil, pine tree pollen, and mineral salts, have all been put forward as possible antigens involved in the etiology of sarcoidosis (1). Because the hallmark of this disease is the formation of granulomas, agents that may elicit this response have also been implicated, such as mycobacteria, aluminum, and zirconium (37-39). Interestingly, our studies show that aluminum and zirconium can positively regulate IL-18 protein production from THP-1 cells. The role of infectious agents in sarcoidosis has always been plausible given that some studies have shown an increased frequency of serum antibodies to pathogens such as Propionibacterium acnes and Chlamydia pneumonias (40,41). Propionibacteria have also been isolated from sarcoid lesions and. using quantitative PCR, propionibacterial rRNA was detected in all 15 patients with sarcoidosis in a Japanese study (42, 43). The data presented in this study now implicate H. influenzas and M. catarrhalis in the pathogenesis of sarcoidosis. All the patients with active sarcoidosis had bacterial DNA from either or both of these two species; however, there was no specific pattern. These gram-negative bacteria have been implicated in chronic obstructive pulmonary disease, pneumonia, sinusitis, and chronic bronchitis (44-47). It is possible that patients with sarcoidosis who do not recover, either spontaneously or with treatment, may have impaired bacterial clearance mechanisms, and this could be related to the ability of these individuals to mount an IL-18 response. Although LPS was also detected in normal BAL fluid the levels were significantly lower and derived from different bacterial species than the LPS in the sarcoid samples. Furthermore, the LPS levels in control and recovered sarcoid samples were lower than the lowest effective dose of LPS capable of inducing a physiologic response in the THP-1 cells (50 ng/ml).

IL-18 has the ability to induce IFN-? production from Th1 cells in synergy with IL-12 (48). We and others have shown that IL-18 is increased in the lungs of patients with sarcoidosis and it has been shown to upregulate IL-2 expression and to activate CD4^sup +^ T lymphocytes (16, 17). IL-18 was shown to be produced by alveolar macrophages and has been found in airway columnar epithelial cells in biopsies from patients with sarcoidosis (16, 49). The IL-18 receptor is a member of the IL-1 receptor/Toll-like receptor (TLR) family and the regulation of the Th1 response has been shown to involve TLRs (50). TLR4 signaling is critical for IL-18 production from macrophages (51) and one study has shown that TLR4-deficient mice showed significantly lower IL-18 levels compared with wild-type mice (52). Studies have also implicated IL-18 in the host response to bacterial infection, with IL-18-deficient mice displaying impaired innate immune responses during pneumococcal pneumonia (53) and Escherichia coli peritonitis (54). In the present study we have detected the presence of gram-negative bacterial DNA in active sarcoid BAL fluid; however, this bacterial DNA is not the factor responsible for inducing IL-18 expression as DNase treatment of sarcoid BAL fluid has no effect on IL-18 inducibility (data not shown). IL-18 levels during the active phase of pulmonary sarcoidosis were also shown to be elevated compared with those of recovered patients.

Sarcoidosis is pathologically similar to chronic beryllium disease-some people have even coined the term "sarcoidosis of known etiology" to describe chronic beryllium disease-yet the disorders develop in response to different antigens, and culminate in the development of noncaseating granulomas. Chronic beryllium disease is characterized by a hypersensitivity to beryllium salts, and previous studies have shown that BeSO^sub 4^ stimulates bronchoalveolar cell release of tumor necrosis factor a, IL-6, and IL-18 (55, 56). In this study, BeSO^sub 4^ was shown to upregulate IL-18 expression from THP-1 cells in vitro. It has been demonstrated elsewhere that IL-18 is induced in response to PPD antigen from M. tuberculosis (57, 58). We have shown that although PPD upregulates IL-18 mRNA and protein expression, it did not act on the 5'-UTR promoter, perhaps suggesting that PPD may induce IL-18 expression by targeting the IL-18 intron-1 promoter. This study indicates that both PPD and BeSO^sub 4^ induce IL-18 mRNA and protein production. Interestingly, M. tuberculosis catalase-peroxidase, but not PPD, has been identified as a tissue antigen associated with sarcoidosis and may be a target of the adaptive immune response driving granulomatous inflammation in sarcoidosis (59). However, BAL fluid from the patients in this study had no previous exposure to beryllium and showed negative cultures and biopsies for tuberculosis infection.

In addition to environmental factors, there is evidence that genetic factors contribute to the pathogenesis of sarcoidosis (1). Much evidence suggests that genetic variability in human leukocyte antigen genes plays a role in sarcoidosis (60). In ACCESS (A Case Control Etiologic Study of Sarcoidosis) the human leukocyte antigen DRB1*1101 allele was associated with sarcoidosis in a population composed of black subjects and white subjects (61). Mutations in genes for angiotensin-converting enzyme, tumor necrosis factor a, C-C chemokine receptor 2, and the vitamin D receptor are also thought to be involved in this disorder (62). The -607^sup C/A^ SNP present in the IL-18 5'-UTR promoter is located at a putative CREB-binding site (18). Patients with multiple sclerosis and homozygous for the -607^sup C/C^ genotype have higher levels of IL-18 compared with patients with other genotypes (18). In this Irish population, the -607^sup C/C^ genotype and the -607^sup C^ allele were found to be the most frequent: however, functional analysis did not reveal any differential activity between the -607^sup A^ and -607^sup C^ allelic variants in response to LPS. In the Japanese sarcoidosis population, the -607^sup A/C^ genotype and the -607^sup A/C^ allele were predominant (24). Taken together, these data suggest that the -607^sup C^ allele may be a genetic risk factor for sarcoidosis and may be in linkage disequilibrium with a functional mutation elsewhere.

In summary, we have demonstrated that gram-negative bacteria may have a role in initiating the inflammatory response in sarcoidosis. This opens up new possibilities for elucidation of the pathogenesis of this condition and therapeutic interventions.

Conflict of Interest Statement: D.M.K. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. C.M.G. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. G.M. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. M.O. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. P.M.G. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. C.C.T. does not have a financial relationship with a commercial entity that has an interest in the subject of this manuscript. S.J.O. spoke at conferences sponsored by AstraZeneca, Merck Sharp & Dohme, and Boehringer Ingelheim. He also received industry supported grants from AstraZeneca. N.G.M. acted as consultant to Aventis Behring and Baxter Biosciences and served on the Advisory Board for Genaera.

Acknowledgment: The authors thank Dr. Stephen Smith (Department of Clinical Microbiology, Trinity College Dublin) and Dr. Fidelma Filzpatrick (Department of Microbiology, Royal College of Surgeons in Ireland) for their assistance. pGL35'-UTR vectors and Chlamydia pneumoniae DNA were generously provided by Dr. Vilmantas Giedraitis (Karolinska Institute, Sweden) and Prof. Catherine O'Connell (University of Arkansas for Medical Sciences), respectively.

© 2005 American Thoracic Society Provided by ProQuest LLC. All Rights Reserved.