|Home | About | Journals | Submit | Contact Us | Français|
Recent literature has indicated the feasibility of microarray analysis in the characterization of chronic sinusitis. We hypothesized that previously unexplored inflammatory mechanisms would be involved in the pathophysiology of noneosinophilic chronic rhinosinusitis with nasal polyps (NE-CRSwNP) and that this technology could be used to identify the gene expression of these novel and previously known mediators.
Patients with CRSwNP failing medical therapy were prospectively enrolled and NP tissue was removed at time of surgery. NE-CRSwNP was diagnosed based on clinical parameters including absence of allergic disease and confirmed with histopathology showing lack of eosinophilic infiltration. Messenger RNA (mRNA) transcripts extracted from study and control patients were then subjected to microarray analysis using Affymatrix based chips. Validation of findings was then confirmed via quantitative reverse transcription polymerase chain reaction (qRT-PCR).
Microarray analysis revealed activation of pathways involved in antigen presentation, cellular movement, hematopoiesis, carcinogenesis, apoptosis, and cell signaling. Previously unexplored genes of interest were identified and their differential regulation was validated via qRT-PCR. Our data showed up-regulation of innate inflammation genes (IL-6, IL-8, and monocyte chemoattractant protein 1), hypoxia-induced inflammation 1α, and fibrosis (tenascin) and lack of up-regulation of genes associated with allergic, eosinophilic inflammation (IL-4 and IL-13). Additionally, the genes for CXCL1 and autocrine motility factor receptor were novelly identified to be up-regulated
This study explores the utility of gene microarray technology in identifying unexplored targets of immune dysregulation in NE-CRSwNP. Furthermore, the data characterize the immunologic profile of NE-CRSwNP as it differs from other forms of CRSwNP, in particular, those known to be associated with eosinophilic inflammation.
Chronic rhinosinusitis (CRS) is a chronic inflammatory condition affecting the paranasal sinuses and nasal cavities, which has a significant adverse impact on the quality of life and daily functioning of its sufferers.1–3 This single term, however, most likely represents multiple different diseases with or without various stages of severity along their spectra. A summary statement comprising expert opinion within this field divides CRS into two easily distinguishable subsets: CRS without nasal polyps (CRSsNP), characterized by a chronic inflammatory infiltrate and the absence of NP; and CRS with NPs (CRSwNP).2,3 Research into these two subtypes has indicated a significant association between the presence of eosinophilia and nasal polyposis4,5 but other work has also indicated that the level of this eosinophilia has been variable6,7 and that certain forms of CRSwNP have a higher degree of eosinophilia such as allergic fungal sinusitis8 and aspirin exacerbated respiratory disease.9 However, a certain subset of NP specimens does not show significant eosinophilia and microscopic evaluation of them has indicated other significant differences between their histopathological appearance and that of polyps with eosinophilia.10 These noneosinophilic (NE) polyps tend to show characteristics not seen in their eosinophilic counterparts including profound glandular hypertrophy, mononuclear cell infiltrate, fibrosis, and the presence of mast cells.10 This noneosinophilic CRSwNP (NE-CRSwNP) would appear to represent a different subset of CRSwNP for which the elucidation of its specific cytochemical and inflammatory mediators is required to better understand and treat the disease.
Microarray analysis of complementary DNA (cDNA) has become a powerful tool in the analysis of gene expression profiles. By assessing the relative amount of messenger RNA (mRNA) in a sample compared with a control, this technique can quickly indicate which genes have been activated in the disease state and that may help explain its pathophysiology. This relatively new and revolutionary technique has recently been used in the characterization of human NP tissue.11–17 However, none of these studies have specifically investigated NE polyps in the nonallergic patient. The purpose of this study, therefore, was to use cDNA microarray technology to identify the gene expression profile in polyp samples from patients with NE-CRSwNP when compared with control tissue.
Study approval was obtained from the Institutional Review Board of the University of Virginia.
Patients presenting to the University of Virginia Health System who met the diagnostic criteria for CRSwNP18 and who failed medical therapy were eligible for the study. Patients without a history of inhalant allergies who chose to undergo endoscopic sinus surgery were then enrolled in the study. Additional exclusion criteria included the presence of cystic fibrosis, sinonasal tumor, or immunodeficiency. Preoperative oral steroids were not used.
Control tissue was harvested from the ethmoid or sphenoid sinuses of patients undergoing nonsinusitis-related surgery (e.g., orbital decompression or cerebrospinal fluid leak repair). At the time of surgery, sinus tissue was harvested and placed in sterile normal saline. Initial gene array studies were performed on two control and two NECRSwNP subjects. Subsequent studies were performed on separate cohorts of six additional subjects for each the control and polyp samples.
A portion of each polyp was placed in 4% paraformaldehyde (Sigma, St. Louis, MO) overnight at 4°C. The next day the polyps were washed in phosphate-buffered saline and stored in 70% ethanol until paraffin embedding. Paraffin embedding, tissue sectioning, and H&E staining were performed by the University of Virginia Histology Core. Samples were examined by microscopy. Inclusion of tissue sample required four or less eosinophils in each of 10 high-powered (400×) fields.
For gene chip analysis, an adjacent portion of two of the polyp samples were frozen in liquid nitrogen. RNA extraction and gene chip analysis were performed by the UVA Biomolecular Research Facility using the Affymetrix (Santa Clara, CA) HgU133 plus 2.0 gene chip (which analyzes 38,500 genes) as previously described.19 Gene ontology data mining was then performed using the Expression Analysis Systematic Explorer downloaded from the Database for Annotation, Visualization, and Integrated Discovery 2.0 program (www.david.niaid.nih.gov).20 Gene expression data sets containing the information of gene identifier, fold change, and p values were submitted to the Expression Analysis Systematic Explorer for analysis. In these studies, a fold change cutoff of 1.5 and value of p < 0.05 was set for inclusion in the data set. Several genes were then selected for verification and additional study via reverse-transcription and quantitative polymerase chain reaction (qRT-PCR).
Total RNA was extracted from tissue using RNA-Bee isolation solvent (Tel-Test, Inc., Friendswood, TX). Conversion of the mRNA to cDNA was performed using a Taqman Reverse Transcription kit (Roche, Branchburg, NJ) according to manufacturer’s instructions. Specific mRNA levels were quantified using realtime PCR. Primer pairs for the genes analyzed were sequenced and obtained from Integrated DNA Technologies (Coralville, IA) or ordered from SuperArray Biosciences Corp. (Frederick, MD). Information regarding these is listed in Table 1. Relative gene expression was calculated using the comparative crossing threshold (CT) method as described elsewhere21 with β-actin as a reference housekeeping gene. This method was similarly used to compare gene expression between the control and polyp tissue such that fold difference was calculated via the formula 2−ΔΔCt.19
IL-6, IL-8, and CXCL1 levels were measured in the supernatants of protein extracted from NP or control samples. Tissue samples were minced and added to 500 µL of cell lysis buffer (Bio-plex cell lysis kit; Bio-Rad, Hercules, CA) and dounce homogenized. Samples were sonicated and centrifuged at 3800 × g for 4 minutes. Supernatants were collected and any pelleted debris was discarded. IL-8 and IL-6 levels were determined using a Bio-Plex bead-suspension assay (Bio-Rad) and CXCL1 was measured using a commercial ELISA kit (R & D Systems, Minneapolis, MN). The sensitivity of the assay for CXCL1 was 10 pg/mL, for IL-6 was 2 pg/mL, and for IL-8 was 1.4 pg/mL. Cytokine concentrations were normalized to total protein to allow comparisons between different samples.
Samples were deparaffinized and hydrated to distilled water. Endogenous peroxidase activity was eliminated by incubating sections for 30 minutes at room temperature in 200 mL of methanol with 3 mL of 30% hydrogen peroxide. Epitope unmasking was performed by heating sections for 20 minutes in unmasking solution (Vector Laboratories, Burlingame, CA). Specific staining for CXCL1 was performed using a 1:250 dilution of rabbit polyclonal antibody ab9804 (AbCam, Cambridge, MA). Isotype control rabbit IgG (Southern Biotech, Birmingham, AL) was used at the same concentration as the CXCL1 antibody. Blocking and signal amplification were performed using the RTU Vectastain Universal Elite ABC Kit (Vector Laboratories) according to manufactures directions. Samples were developed for 3 minutes with diaminobenzidine (Dako Cytomation, Carpinteria, CA) and counterstained 1 minute with Mayer’s hematoxylin (Vector Laboratories). Samples were washed in distilled water and dehydrated through alcohol and xylenes and mounted.
Two polyps each from both the study and the control populations were submitted for microarray analysis. This small sample was chosen because this was an exploratory study and initial analysis between the two samples showed a high correlation (data not presented). A total of 120 genes were identified, which showed both an absolute fold change of ≥1.5 and a value of p < 0.05. Of these, 58 genes were up-regulated, and 62 were down-regulated. A breakdown of the various functional groups represented by these genes is presented in Fig. 1. An exhaustive list of the significantly differentially regulated genes is provided in Table 2. Many of the traditionally identified inflammatory modulators and genes associated with CRS were differentially expressed, although not significantly so. These included those associated with Th2 or allergic inflammation (IL-4, fold change 1.28; IL-13, −2.79) and those involved in innate immunity (IL-6, 5.13; IL-8, 2.3).
We next analyzed several genes by qRT-PCR on separate larger cohorts of samples. Despite a lack of significant fold change of the previously known inflammatory mediators on microarray analysis, 12 of these genes were selected for RT-PCR investigation to determine their involvement in this subset of disease. These included characteristically Th2-mediated genes (IL-4, IL-13, thymus and activation-regulated chemokine, RANTES, and thymic stromal lymphoprotein), Th1 mediated genes (interferon [IFN] γ), and proinflammatory or fibrosis-associated genes (stem cell factor [SCF], IL-6, IL-8, IL-11, tenascin-C, and hypoxia-inducible inflammation [HIF] 1α). The microarray data were also further evaluated for novel genes of interest including representative genes thought to be important in cell trafficking, inflammatory cell recruitment, neoplasia, and inflammation. Five previously unexplored genes were selected for RT-PCR analysis. The selection was based on either a significant fold change or a substantially high absolute fold change even if this change was not statistically significant. Information regarding the selected genes is presented in Table 3 and the fold change results are represented graphically in Fig. 2. Of the previously known modulators/mediators of inflammation, the analysis shows a likely clinically significant fourfold or greater up-regulation for SCF, IL-8, tenascin-C, and HIF1-α (p < 0.01). Of the novel genes, a fourfold up-regulation is seen with CXCL1 (p < 0.05), ATP-binding cassette B11 (ABCB11), and autocrine motility factor receptor (AMFR). IL-4, IL-13, and IFN-γ were substantially down-regulated compared with control tissue, whereas IL-11 and CLCA2 showed a general equivalency of expression between control and NE-CRSwNP.
Using either Bioplex or ELISAs for IL-6, IL-8, and CXCL1 (Gro-α) on protein extracted from control and NE-CRSwNP tissue, we verified that the increased gene expression observed by qRT-PCR corresponded to increased protein production (Fig. 3).
Staining of the tissue indicated that the up-regulation of CXCL1 was noted to occur within the epithelial layers of the polyp and its glandular structures (Fig. 4).
CRS is a general term for what appears to be numerous distinct pathways to inflammation of the paranasal sinuses for which the distinction between the presence or lack of polyps can be of significant therapeutic benefit.22 However, even within CRSwNP, multiple subsets of disease with respectively different treatments exist including allergic fungal sinusitis and aspirin-exacerbated respiratory disease. Nevertheless, polyps characterized by eosinophilia are only seen ~80% of the time in western cultures23 and perhaps substantially less in certain Asian populations.24,25 The remaining portion of patients has NPs that are significantly lacking in an eosinophilic infiltration, but this entity has not been subject to the same level of scrutiny as the eosinophilic variety as we sought to provide with the current study and microarray analysis.
Several studies using microarray technology in the evaluation of NPs have been published in recent years. Benson and colleagues have published several articles regarding their initial study on the effects of corticosteroids on the gene profile of NPs.14 Later, Fritz and colleagues15 and Figuerido et al.12 compared polyp tissue with adjacent inflamed mucosa and showed significant differences in the gene expression profiles. Liu et al.26 and Wang et al.16 also found differences comparing polyps with sphenoid sinus mucosa and inferior turbinate mucosa, respectively. However, these several studies all evaluated CRSwNP in general, without specific attention to any subset therein. Orlandi et al.,11 on the other hand, did evaluate subsets of CRSwNP, but unlike our study, theirs focused on two variants of eosinophilic CRSwNP (E-CRSwNP). As such, no study has yet specifically explored NE-CRSwNP through either microarray technology or extensive RT-PCR of genetic transcripts.
In this study, we found 120 genes that were either significantly differentially expressed in polyp tissue when compared with normal sinus tissue controls, but the various inflammatory cytokines typically seen in CRS were not found to have a significant change in regulation. This may reflect the inherent variability of gene expression and the heterogeneity of this disorder in addition to the low power analysis of the small sample size used for this screening technique. These problems seem to be common to microarray analysis and other authors have indicated that this can lead to both spurious data as well as underrecognizing important differences.27 Pooling of specimens may have been able to compensate for this problem, but this method has also been discouraged elsewhere.28
Some trends were, however, observed with these cytokines and several of these were selected for further study with RT-PCR. Although not every prediction was confirmed, the data were of importance. Most compelling was the lack of evidence for expression of genes associated with allergic inflammation including Th2-associated cytokines (e.g., IL-4 and IL-13), consistent with our belief that this is a unique disorder and the down-regulation of genes associated with eosinophilic inflammation (eotaxin and eosinophil cationic protein) on the gene array. In contrast, we did see a fourfold elevation of SCF (KITLG). As a stimulator of the differentiation and proliferation of mast cells, its up-regulation in the polyp tissue is consistent with our previous histological findings of the presence of mast cells in this tissue. The fibrosis in these polyps may well be mediated through this type of mast cell process, which has been implicated in other fibrotic diseases such as rheumatoid arthritis, scleroderma, and idiopathic pulmonary fibrosis.28 Both transcripts and protein levels for IL-8 (CXCL8) and CXCL1 were also noted to be significantly elevated in the polyp tissue. These chemokines are both attractants for neutrophils indicating a role for this immune cell in the disease. Although neutrophils were not seen in our histological samples, it may be that polyp formation in this disease develops in the setting of chronic infection, biofilm formation, and purulent exudates that are remote from the polyp.
As discussed, our previous studies indicated an increased fibrotic stroma in these types of polyps and the nearly 10-fold increase in tenascin-C in the qRT-PCR is consistent with this observation. Tenascin-C is often transiently expressed in acute tissue injury and inflammation where it regulates fibroblast migration.29 Its presence in a chronic disease such as NECRSwNP implies a “persistent acute” injury to the tissue as a part of the overall disease process resulting in a more fibrotic appearing polyp.
The role that hypoxia might play in upper airway inflammatory disease is a potentially important one.30 HIF-1α is an inducible transcription factor expressed in hypoxic conditions that is involved in activation of glycolytic and inflammatory pathways.31 NP-derived fibroblasts have been specifically shown to increase their production of HIF-1α in response to hypoxic conditions32 and a corresponding in vivo up-regulation was significantly shown in this study, consistent with the reported hypoxic milieu observed in CRS.33 Although the increased expression of HIF-1α may simply be a stress response, the implication can not be ignored that hypoxia may play a significant role in the pathogenesis of NE-CRSwNP. Improved aeration with a secondary reduction of HIF-1α could contribute to the therapeutic benefit observed after surgery.
Thymus and activation-regulated chemokine and thymic stromal lymphoprotein were not found to be significantly up-regulated and IL-4, IL-13, and IFN-γ, the hallmark cytokines for Th2- and Th1-mediated disease, respectively, were found to be down-regulated. This correlates well with a recent study by Kim et al.23 where on immunohistochemical study of NE polyps, 0/20 expressed only CCR5+ (Th1) cells, 6/20 expressed only CCR3+ (Th2) cells, 4/20 expressed both, and 10/20 (50%) expressed neither Th1 or Th2 cells.
Our investigation also found the expression of two other genes to be substantially increased on both microarray and RT-PCR data, ABCB11 and AMFR. The former is a bile salt exporter linked to drug resistance and for which studies investigating its presence in hepatic inflammation and stimulation with IL-6 have actually shown a down-regulation.34 It is not immediately clear why this gene was found to be up-regulated in sinus tissues, especially under a condition of IL-6 up-regulation, and as such, this deserves additional evaluation in future studies.
AMFR, on the other hand, is a gene that codes for the receptor for autocrine motility factor, a cytokine seen to be produced by tumor cells,35 and its up-regulation in our study correlates well with other studies on NPs indicating up-regulation of carcinogenic genes.14,15 Although a large differential expression of SPINK7 and CLCA2 was seen during microarray analysis, this was not confirmed by RT-PCR.
When viewed in the context of the RT-PCR results, the predicted direction of change for the studied genes was confirmed in 11/16 (68.8%). It is uncertain as to why a greater correlation was not seen. Perhaps this relates to the inability for population estimates to be made based on a very limited sample size, especially in what may be a heterogeneous entity. Nevertheless, the microarray was able to provide some very useful confirmable data regarding some of the novel genes identified.
NE-CRSwNP is a separate histologically identifiable subset of CRSwNP. This article represents the first dedicated study investigating its gene expression profile through microarray analysis. Although not all findings were validated, the expression of SCF, IL-6, IL-8, tenascin-C, and CXCL1 and lack of up-regulation of IL-4, IL-13, and IFN-γ indicate that the pathophysiology of NE-CRSwNP may be more related to an innate immune response as opposed to a directly Th1- or Th2-mediated process. The identification of novel genes in this process, ABCB11 and AMFR, deserve additional study to determine their exact role in this disease.
Supported by a research grant from the Richmond Eye and Ear Foundation and National Institutes of Health Grants R01-AI47737 and P01-AI50989
Presented at the American Rhinologic Society meeting, Orlando, Florida, May 1, 2008