Published ahead of print on December 14, 2006, doi:10.1164/rccm.200601-136OC
© 2007 American Thoracic Society doi: 10.1164/rccm.200601-136OC
Expression of Receptor for Advanced Glycation End Products in Sarcoid Granulomas1 Laboratorio di Biochimica & Genetica, Clinica Malattie Apparto Respiratorio, 2 Istituto di Anatomia ed Istologia Patologica, and 3 Unità di Biometria, IRCCS Policlinico San Matteo, Università di Pavia, Pavia, Italy; 4 Immunologia Clinica, Dipartimento di Medicina Clinica e Sperimentale, Università di Padova, Padova, Italy; 5 Dipartimento Materno-Infantile e di Biologia-Genetica, Università di Verona, Verona, Italy; and 6 Laboratorio di Immunogenetica, Dipartimento di Genetica e Microbiologia, and 7 Dipartimento di Biologia Animale, Università di Pavia, Pavia, Italy Correspondence and requests for reprints should be addressed to Maurizio Luisetti, M.D., Clinica Malattie Apparto Respiratorio, IRCSS Policlinico San Matteo, Via Taramelli 5, 27100 Pavia, Italy. E-mail: m.luisetti{at}smatteo.pv.it
Rationale: The receptor for advanced glycation end products (RAGE) engages a number of ligands implicated in inflammatory processes. The RAGE coding gene maps to the 6p21.32 region, close to the genes DRB1 and BTNL2, which are associated with sarcoidosis. Objectives: We investigated a possible implication of RAGE in sarcoid granulomas.
Methods: RAGE and major ligands (N- Measurements and Results: RAGE, CML, S100A12, and S100B immunoreactivity was observed in all sarcoid granulomas, although at different intensities. The degree of RAGE expression significantly correlated with the degree of S100A12 expression. The –374 TT/AT genotypes, associated with higher RAGE transcriptional activity, were more frequent in the sarcoidosis biopsy group than in control subjects, and the association was confirmed in a second, independent series of 101 patients with sarcoidosis. Conclusions: We showed the association of RAGE and its ligands with sarcoidosis and suggest that an intrinsic genetic factor could be in part involved in its expression. In Italian patients, the –374 T/A polymorphism seems to be significantly associated with this disease.
Key Words: ligands • immunohistochemistry • genetics • real-time polymerase chain reaction
Sarcoidosis is a systemic inflammatory disorder characterized by infiltration of CD4+-activated T cells that give rise to the formation of specific noncaseating granulomas, the hallmark of the disease, in a variety of affected organs (1, 2). Genes mapping within the HLA region are believed to play a significant role in the pathogenesis of sarcoidosis for a number of reasons: (1) products of these genes are involved in the initial event leading to granuloma formation (i.e., presentation of an unknown antigen or antigens by antigen-presenting cells to T lymphocytes) (1–3); (2) various genetic studies have shown that HLA class II alleles are associated with sarcoidosis (4–8); and (3) from an analysis of 55 German families with siblings suffering from sarcoidosis, Schürmann and coworkers demonstrated a strong association maximum multipoint nonparametric linkage score 3.2 with the marker D6S1666 mapping in HLA region at the junction between class II and class III (6p21) (9). Subsequently, a subregion corresponding to the butyrophilin-like 2 (BTNL2) gene was found to be associated with sarcoidosis (10). The gene coding for the receptor for advanced glycation end products (RAGE) maps to the same 6p21.32 position. RAGE recognizes tertiary structures rather than amino acid sequences and therefore acts as a multiligand receptor with the ability to engage classes of molecules (11, 12). Major ligands, termed "advanced glycation end products" (AGEs), are a heterogeneous group of compounds deriving from nonenzymatic glycation and oxidation of proteins and lipids. Generically, AGEs are triggered by oxidant stress and can accumulate in the vessel wall (13). They have been reported to form in several pathophysiologic situations associated with inflammation. Moreover, RAGE binds proinflammatory cytokine–like mediators of the S100/calgranulin family (14), which have been repeatedly associated with granulomatous disorders, including sarcoidosis (15). RAGE is, therefore, heavily implicated in inflammatory processes, particularly contributing to their amplification (11). Considering the previously mentioned points, we hypothesized that RAGE could be involved in the inflammatory process related to granuloma formation in sarcoidosis. In this article, we describe a series of experiments showing that RAGE is expressed in sarcoid granulomas, that its naturally occurring ligands are present in the sarcoidosis inflammatory process, and that the intensity of RAGE expression in sarcoid granulomas varies, and we speculate that a functional polymorphism in the promoter region of the RAGE gene could be implicated, at least partly, in the different degrees of RAGE expression at sites of sarcoidosis activity. Some of the results of this study were previously reported in abstract form (16).
Immunohistochemistry Paraffin-embedded tissue specimens, obtained from patients with sarcoidosis for biopsy confirmation between 1992 and 2003 at the Pathology Department of Pavia, were selected. All subjects were white and from Northern Italy. Anti-RAGE goat antibody and antibodies against RAGE ligands N-  -carboxy-methyl-lysine (CML), S100B, and S-100A12 were immunoreacted on paraffin tissue sections. RAGE immunoreactivity was determined by means of quantitative optical density analysis; RAGE ligand expression was assessed semiquantitatively. Additional information is available in the online supplement.
RAGE and BTNL Gene Single-Nucleotide Polymorphisms
All individuals gave informed consent before entering the study, which was approved by the ethical committees of the institutions involved. The ethical committees approved the use of human tissue for this investigation. RAGE –374 T/A (rs1800624), –429 C/T, 1704 G/T gene polymorphisms, and BTNL2 rs2076530 single-nucleotide polymorphisms (SNPs) were analyzed (details available in the online supplement).
Determination of the Functional Expression of the RAGE –374 T/A Polymorphism
Statistical Analysis
Immunohistochemistry Ninety-nine consecutive sarcoid biopsy samples fulfilling histologic and clinical criteria for sarcoidosis were collected for the immunohistochemical characterization of RAGE and its major ligands in granulomas. Specimens were obtained from 47 male and 52 female patients, with a mean age of 40 ± 3 years (range, 18–71 yr), and included tissues from peripheral lymph nodes (n = 58), lung (n = 20), skin (n = 16), and liver (n = 5). In sarcoid granulomas, RAGE was expressed in the cytoplasm of epithelioid histiocytes and multinucleated giant cells. RAGE immunoreactivity was observed in all sarcoid granulomas; its mean optical density (OD) differed in different cases (range, 0.035612–0.380846), whereas it was fairly homogeneous within each single case (SD range, 0.004–0.05) (Figure 1). Skin and lung granulomas showed the strongest degree of immunoreactivity. RAGE was also expressed in tissue outside the granulomas (details are available in the online supplement).
Immunohistochemical investigations were extended to the major RAGE ligands CML, S100A12, and S100B (Figure 2). CML immunoreactivity was present in epithelioid histiocytes and giant cells, with the intensity of reactivity ranging from low to high. The expression of S100A12 was highly variable in histiocytes and giant cells, ranging from negative to strongly positive. Epithelioid and giant cells of sarcoid granulomas did not express S100B, although this RAGE ligand was present in dendritic cells infiltrating granulomas. At semiquantitative evaluation, we observed an overall higher degree of RAGE ligand expression in granulomas with higher mean RAGE optical density. However, statistical analysis only confirmed a significant correlation between the degree of expression of RAGE and of S100A12 (p < 0.05) (Figure 3) but not of S100B (p = 0.05) or CML (p > 0.05). No correlation was found between the stage of granuloma evolution and RAGE expression.
Functional Variant of the RAGE Gene We examined the mechanisms underlying the varying degree of RAGE expression within sarcoid granulomas. We hypothesized that a possible intrinsic factor within the RAGE gene might play a role. We assessed the frequency of three RAGE SNPs (–374 T/A, –429 T/C, and 1704 G/T) in a control population (Table 1). The first two polymorphisms were selected based on the fact that analysis of the RAGE promoter sequence revealed the presence of potential transcription factor binding sites, which could interfere with the modulation of gene expression (21). The 1704 G/T polymorphism was investigated because of the previously reported involvement in the oxidative stress (13). We then selected the RAGE –374 T/A polymorphism for further investigation because it was the only SNP investigated with a significant minor allele frequency (47%) in healthy individuals (Table 1).
We performed two experiments to assess the transcriptional effect of the –374 T/A RAGE polymorphism. In the first experiment, immunocytochemical analysis of PBMCs isolated from 58 healthy donors showed only weak immunostaining for RAGE in subjects with the AA genotype (Figure 4B), suggesting a low level of RAGE expression on the cell surface, whereas PBMCs from subjects with the AT and TT genotypes (Figures 4C and 4D, respectively) showed stronger immunoreactivity. In the second experiment, we used real-time polymerase chain reaction to make a semiquantitative analysis of PBMCs isolated from 58 healthy blood donors. In our conditions, we identified a trend in which the AA genotype seemed to correlate with a lower transcriptional activity, whereas the TT/AT genotypes seemed to be associated with higher RAGE expression, although this difference did not reach statistical significance (as shown in Figure E1 in the online supplement).
The Frequency of RAGE and BTNL2 Polymorphisms The frequency of the RAGE gene –374 T/A polymorphism in the series of sarcoid biopsies is reported in Table 2. The frequency of the –374 TT/AT genotypes in the group of sarcoidosis biopsies was significantly higher than that in the control subjects (90 vs. 76%;  2 = 8.2781; p = 0.004; OR, 2.7590; 95% confidence interval [CI], 1.3523–5.6291). The –374 AA genotype was significantly less frequent in biopsies from patients with sarcoidosis than in control subjects (OR, 0.3625; 95% CI, 0.1776–0.7395). After stratifying for sex, the –374 AA genotype frequencies remained significantly different between sarcoidosis biopsies from female subjects and control subjects (2 = 4.4817; p = 0.034). This kind of female sex–restricted association is not unusual in our Italian series of patients with sarcoidosis (5). The investigation was extended to a second, confirmatory series of patients with sarcoidosis (n = 101), whose clinical characteristics have been reported (22), and it was compared with a second group of healthy control subjects (n = 65). The findings were confirmed in this second group of subjects, among whom the frequency of the TT/AT genotypes was 94% (2 = 7.7127; p = 0.005 vs. control subjects) (Table 3). Genotypes of the patients with sarcoidosis and of control subjects agreed with the Hardy-Weinberg equilibrium, whereas the genotypes of sarcoidosis biopsies were slightly out of equilibrium (p = 0.0231). When the frequency of the –374 T/A RAGE polymorphism was analyzed in function of the clinical characteristics of the series of patients with sarcoidosis (radiologic stage, relapses, and prognosis), we were unable to detect any significant correlation (data not shown). We also included three series of "positive" controls (i.e., patients with a lung disease whose pathogenesis is believed to be different from that of sarcoidosis). The frequencies in 74 patients with idiopathic pulmonary fibrosis (TT/AT, 80%; AA, 20%) and in 117 patients with systemic sclerosis (TT/AT, 75%; AA, 25%) did not significantly differ from the 327 healthy control subjects, whereas in 101 patients with asthma, the frequencies (TT/AT, 86%; AA, 14%) were slightly but significantly different from healthy control subjects (2 = 3.858; p = 0.0459).
We evaluated the linkage disequilibrium between the two RAGE and BTNL2 SNPs (rs1800624 and rs2076530) and obtained a D' value of 0.2127 (p = 0.028) for control subjects and a D' value of 0.325 (p = 0.028) for patients with sarcoidosis. Table 4 reports the deduced haplotype frequencies in patients with sarcoidosis and in healthy subjects. The frequency of the deduced haplotype BTNL2 G and RAGE –374 T was significantly different between the two groups ( 2 = 4.0015; p = 0.045 vs. healthy control subjects).
Genotype–Phenotype Correlates We looked for correlations between the –374 T/A polymorphism and the degree of RAGE expression in sarcoidosis granulomas. We found a wide dispersion of data (Figure 5). Despite a trend for the group of subjects carrying the T allele to include more samples with OD > 0.2, no statistically significant correlation was found.
We provide evidence that RAGE and its ligands are expressed in sarcoid granulomas. RAGE expression has been extensively documented in normal human tissues (23) and in several inflammatory conditions associated with diabetes (24), amyloidosis (25), and rheumatologic disorders. The concomitant expression of RAGE and its ligands has never been investigated in chronic granulomatous diseases such as sarcoidosis, with the exception of one early report dealing with the presence of S100 immunoreactivity in mononuclear phagocytes and in epithelioid cells of mature sarcoid granulomas (15). We observed that RAGE is consistently expressed, with variable intensity, in epithelioid histiocytes and giant cells in sarcoidosis and that its expression is associated with that of its major proinflammatory ligands. The presence of RAGE in giant cells and epithelioid histiocytes of sarcoid granulomas of mononuclear lineage is in accordance with the documented RAGE expression in mononuclear inflammatory cells and phagocytes. Macrophages have also been reported to express high levels of AGEs, produced by intracellular oxidative degradation of glycated proteins or by endocytosis of receptor-bound proteins in normal and pathologic conditions (26), although AGE expression in epithelioid histiocytes has not been documented. The S100/calgranulins are a family of closely related calcium-binding polypeptides that have been implicated in the regulation of protein phosphorylation, calcium ion homeostasis, enzyme and transcription factor activities, cell growth and differentiation, and in the inflammatory response (27). S100A12 and S100-B have been shown to activate endothelial cells, vascular smooth muscle cells, monocytes, and T cells via RAGE, resulting in the generation of cytokines and proinflammatory adhesion molecules (14, 12). Moreover, the serum level of S100A12 is increased in conditions associated with chronic immune/inflammatory responses, in which immunohistochemical studies have shown its selective expression in polymorphonuclear granulocytes (28–30), and in mononuclear phagocytes in experimental settings (31). We observed S100A12 not only in granulocytes but also in granulomas in over the half of the cases, indicating that the protein can be produced in vivo by stimulated monocytes. Regarding the receptor/ligand interaction, it has been shown that RAGE–ligand binding initiates a sustained period of cellular activation mediated by receptor-dependent signaling (32), leading to increased expression of the receptor itself. The net result is a positive feedback loop in which the ligand-receptor binding increases expression of the receptor. Because the antibody used for the RAGE investigations is specific for the extracellular V-type domain of the protein, which is present on the full-length transmembrane receptor and on the C-terminally truncated secretory forms (sRAGE), we were unable to discriminate the active form of the transmembrane receptor from the secretory forms stored in the cytoplasm. However, it has been shown there is more full-length, membrane-bound receptor than soluble variants in tissue with a low mitotic activity, such as lung (33). Exogenous administration of the soluble splice variant has been demonstrated in experimental settings to have a negative role on RAGE-mediated proinflammatory action because it prevents RAGE ligands from binding to the active, membrane-bound receptor (34). A recent report has shown that sRAGE is the prevalent form expressed intracellularly in normal human tissues (23), suggesting that it may have roles other than binding AGEs; however, the relative distribution of different RAGE isoforms and the antiinflammatory role of endogenous sRAGE in vivo have not been clarified. Analogous to what is observed at sites of diabetic microvascular injury (35), we reported the coexpression of the several RAGE ligands in RAGE-reactive granulomas, suggesting that their interaction may contribute to inflammatory processes related to granuloma formation or progression, at least in some of the sarcoidosis lesions. The significant correlation between the degree of expression of RAGE and of S100A12 (see Figure 3) suggests that the latter has a predominant role in RAGE binding in sarcoid giant and epithelioid cells. However, the activation of the RAGE inflammatory pathway is not specific to sarcoid granuloma because overexpression of RAGE and its ligands was recently described in a series of nonsarcoidoitic granulomatous and of nongranulomatous lung inflammatory disorders (36). These findings are consistent with the hypothesis that RAGE is not a disease-specific causative or initiating factor, but rather a mediator in the amplification of chronic cellular activation and dysfunction (11). The resulting so-called two-hit pathway (11) has been also proposed as a model for sustained inflammation in Crohn's disease and ulcerative colitis (37). This group of inflammatory bowel diseases shares many immunopathophysiologic aspects with sarcoidosis (38). We have shown that the intensity of RAGE immunoreactivity in sarcoid tissue varies widely from strong to absent (see Figures 1 and 2). Although the previously mentioned role of ligand–receptor interaction in the induction of receptor expression might be one plausible explanation, we also explored the possibility of an intrinsic RAGE regulatory factor. According to our RAGE SNP investigation (see Table 1) and according to the concept that the promoter region of a given gene is characterized by a high density of variants potentially affecting the gene's expression (39), we focused on the –374 T/A (rs1800624) polymorphism. Our cytochemistry experiments (see Figure 4) and real-time polymerase chain reaction analysis (see Figure E1) on PBMCs from healthy control subjects seemed to confirm a difference in the degree of RAGE expression according to the –374 T/A genotype, with the ranking being AA < AT/TT. The ex vivo correlates (Figure 5) were less convincing, suggesting that, in the complex inflammatory network within the sarcoid granuloma, the –374 T/A genotype is not the major factor modulating RAGE expression. We expanded our genetic investigations to a possible role of the RAGE –374 T/A polymorphism as a risk factor in sarcoidosis. When we looked at the allele frequencies of the polymorphism, we found that the T allele was significantly more frequent in sarcoid tissue DNA than in a large series of control subjects, whereas the AA genotype was less frequently represented in the sarcoid samples (see Table 2). Similar findings were obtained when a second confirmatory, independent series of patients with sarcoidosis were included in the analysis (see Table 3). The excess of the T allele, in turn related to overexpression of RAGE, would be therefore an intrinsic mechanism for further sustained inflammation in sarcoid granulomas. The RAGE gene is located in a hot spot region of the short arm of chromosome 6, where there is a high density of sarcoidosis-related genes (Figure 6). It lies approximately 400 kb downstream of DRB1, a major genetic factor for sarcoidosis (40–42), and only 211 kb from BTNL2 (10, 43); a gene–gene interaction in this region is, therefore, plausible. Valentonyte and colleagues (10) has suggested that DRB1 is a significant risk factor for sarcoidosis only in the presence of the truncating BTNL2 allele, whereas Rybicki and colleagues (43) suggested that, in white ethnic groups, BTNL2 is a risk factor independently of an association of DRB1 but that a negative interaction exists between BTNL2 and HLA class II alleles in African Americans. Although we did not screen for DRB1 in our study, the combined RAGE and BTNL genotyping revealed a weak linkage disequilibrium between the BTNL2 G and RAGE –374 T variants. This is not surprising because the BTNL2 locus is close to the RAGE gene and because it has been extensively documented that the HLA region is characterized by strong linkage disequilibrium. The proximity of these genes mapping within the chromosome 6 hot spot region makes the interpretation of association studies complicated.
We have no data on the possible role of RAGE, its ligands, and the RAGE polymorphism on the clinical course of the sarcoidosis. This represents a major limitation of this study; however, sarcoidosis in the Italian population always shows a mild clinical course, which also explains our negative results in previous studies aimed at correlating its course with biallelic polymorphisms in other genes, such as ACE (44) or CR1 (22). By contrast, significant correlations were found when we investigated highly polymorphic genes, such as HLA class I and II genes (40). Nevertheless, it would be interesting to explore the distribution of RAGE polymorphism in different ethnic groups in which sarcoidosis is characterized by a less favorable outcome (5). Regarding the therapeutic implications of our findings, in vivo and in vitro studies have suggested that the proinflammatory effects of RAGE activation in an inflamed environment could be markedly attenuated by blocking the RAGE–ligand interaction. It has been demonstrated that blockade of RAGE, by using anti-RAGE F(ab')2 or sRAGE, suppresses inflammation in acute and chronic models (14, 45). In particular, administration of sRAGE to mice decreased infiltration of the tissue by immune/inflammatory cells. Moreover, it was shown that sRAGE treatment in diabetic mice suppressed levels of IL-6, tumor necrosis factor-  , and matrix metalloproteinases-2, -3, and -9 (46). Confirming a role of RAGE or its natural ligands in the progression of sarcoidosis or its resistance to therapy could open new therapeutic perspectives targeted against RAGE. In conclusion, we show here for the first time that RAGE is expressed by sarcoid granulomas. This finding is corroborated by the simultaneous expression of its natural ligands. Although RAGE overexpression might be mediated by a possible feedback loop induced by ligand binding, we have observed indications that an intrinsic genetic factor could be in part involved in regulating RAGE expression. Finally, we show that carriage of the –374 T allele, which is likely to contribute to increased RAGE expression, is significantly associated with sarcoidosis, at least in our series of Italian patients.
The authors thank Marina Gorrini for the development of the manuscript and Dr. Rachel Stenner for editing.
Supported by a Ricerca Corrente grant from IRCCS Policlinico San Matteo (M.L. and I.C.) and by the MIUR-Cofin projects 2004 (E.P.). This article has an online supplement, which is accessible from this issue's table of contents at www.atsjounals.org Originally Published in Press as DOI: 10.1164/rccm.200601-136OC on December 14, 2006 Conflict of Interest Statement: None of the authors has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. Received in original form January 31, 2006; accepted in final form December 11, 2006
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
AT A GLANCE COMMENTARY
