Sokolowska M. et al 1
[Online repository]
Dysregulation of lipidomic profile and antiviral immunity in response to hyaluronan in severe asthma
Milena Sokolowska, MD, PhD1,#, Li-Yuan Chen, PhD1, Yueqin Liu, PhD1, Asuncion Martinez-Anton, PhD1, CaroleaLogun, MSc1, Sara Alsaaty, MSc1, Rosemarie A. Cuento, MSc2, RongmanCai, PhD1, Junfeng Sun, PhD1, Oswald Quehenberger, PhD3, Aaron M. Armado, MS4, Edward A. Dennis, PhD4, Stewart J. Levine, MD2, James H. Shelhamer, MD1
1Critical Care Medicine Department, Clinical Center, NIH, Bethesda, MD, USA
2Laboratory of Asthma and Lung Inflammation, Cardiovascular and Pulmonary Branch, National Heart, Lung and Blood Institute, NIH, Bethesda, MD
3Department of Medicine, Department of Pharmacology, San Diego, La Jolla, CA, USA
4Department of Chemistry and Biochemistry, San Diego, La Jolla, CA, USA
# Current address: Swiss Institute of Allergy and Asthma Research (SIAF), University of Zurich, Davos, Switzerland; CK-CARE, Davos, Switzerland
Corresponding author:
James H Shelhamer
Critical Care Medicine Department, Clinical Center, NIH
9000 Rockville Pike, Bethesda, Maryland 20892
phone: 301-402-4846; fax: 301-480-3389; email:
Sokolowska M. et al 1
ONLINE METHODS
Participants
The subjects were enrolled under an institutional review board approved protocol (99-H-0076) of the National Heart, Lung, and Blood Institute and all participants provided written, informed consent prior to the study. Severe asthma was defined according to the American Thoracic Society (ATS) Workshop on Refractory Asthma 2000 reportE1and confirmed by the 2013 European Respiratory Society (ERS)/ATS guidelines.E2A patient was included to the severe asthma group if he/she fulfilled one or both major and at least two minor ATS 2000 criteria. Patients who did not fulfil these criteria were enrolled to mild-to-moderate asthma group. Atopy was defined as one or more positive reactions in the skin prick tests, based on the information from the outside physicians or performed at the NIH with cat dander, Dermatophagoidesfarinae, short ragweed and Timothy grass (HollisterStier, Spokane, WA). Sensitivity to aspirin was reported based on the patient’s history. The doses of inhaled corticosteroids were calculated as the budesonide equivalent; the doses of long acting -agonist as a salmeterol equivalent. The sex and age matched control individuals had a negative family history of asthma, a normal chest X-ray and spirometry. All participants had a 1 pack-year or less smoking history. Spirometry was performed after withholding bronchodilatator if symptoms permitted. All asthmatic patients presented bronchodilatator reversibility in at least one spirometry. Subjects were not asked to withhold medications before the blood collection. The levels of corticosteroids or nonsteroidal anti-inflammatory drugs were not measured in the blood of participants at the time of sample collection. Participants’ demographic and detailed phenotypic characteristics are presented in Table E1. Many underlying medical conditions E3-E8 such as nasal polyps, obesity, diabetes, autoimmune disorders, hypertension or cancer might influence the results of our study, thus we examined the differences in their prevalence in our participants (Table E1). Obesity was defined as BMI ≥ 30. We found no differences in comorbidities, except those connected to asthma, such as atopy and allergic rhinitis. Interestingly, there were more subjects with the history of pneumonia or other severe infections in our severe asthma group. Complete blood count with differential (Table E2) and serum IgEwere analyzed for phenotypic characteristics.Blood samples were obtained between June and October. The majority of patients in our cohort were allergic to several allergens, including the perennial allergens such as house dust mites, molds and cat or dog. Some of them were additionally allergic to grass and ragweed.
Peripheral blood mononuclear cells (PBMCs) collection and experimental procedures
Blood specimens from patients with asthma and control subjects were drawn after obtaining informed consent. Complete blood count with differential of all subjects is presented in Table E2. Blood samples were processed immediately using lymphocyte separation medium (LSM) (Lonza, Walkersville, MD), according to the manufacturer’s protocol to obtain PBMCs, a mixture of lymphocytes and monocytes, and a small amount of platelets. After isolation of PBMCs, samples were subjected to cytospins and stained by Diff-Quik (Siemens HealthCare Diagnostics, Newark, DE) to ensure homogenous cellularity between subjects. 4x106 PBMCs were resuspended in 1 ml of RPMI 1640 medium with 2 mM of L-glutamine, supplemented with 10% heat-inactivated FBS (Life Technologies, Thermo Scientific, Waltham, MA) and treated with purified LMW HA (MP Biomedicals; Solon, OH), vehicle, cPLA2 inhibitor (N-{(2S,4R)-4-(Biphenyl-2-ylmethyl-isobutyl-amino)-1-[2-(2,4-difluorobenzoyl)-benzoyl]-pyrrolidin-2-ylmethyl-3-[4-(2,4-dioxothiazolidin-5-ylidenemethyl)-phenyl] acrylamide, HCl; Calbiochem, EMD4Biosciences, Gibbstown, NJ) or cPLA2 inhibitor with LMW HA for 6 h. Each experiment was performed in duplicates. LMW HA used in this study was a purified mixture of fragments between 50 kDa to 600 kDa, as we described previously, together with the detailed quality control.E9 Each experiment was performed in the presence of polymyxin B sulfate (Calbiochem) (10g/ml) to exclude an effect of LPS contamination. The dose and time of the LMW HA stimulation was chosen based on the dose-response and time-course performed in the previous study.E9 Supernatants were collected and kept frozen in -80oC before further analyses. Cell pellets were lysed in RLT buffer, homogenized using QIAshredder columns (Qiagen, Valencia, CA) and stored at -80C before RNA extraction.
Lipidomic profiling by UPLC Mass Spectrometry
All baseline and LMW HA treated samples, as well as 3 random samples with cPLA2 inhibitor from each phenotype were subjected to full lipidomics analysis. For extraction, 200 ul of cell medium was supplemented with a cocktail consisting of 26 deuterated internal standards, and purified by solid phase extraction on Strata-X columns (Phenomenex, Torrance, CA), following the activation procedure provided by the distributor. Samples were eluted with 1 ml of 100% methanol, the eluent was dried under vacuum and dissolved in 50 µl of buffer A consisting of 60/40/0.02 water/acetonitrile/acetic acid = 60/40/0.02 (v/v/v) and immediately used for analysis.
Eicosanoids were analyzed by ultra-high pressure liquid chromatography and mass spectrometry essentially as previously described.E10-E12Briefly, eicosanoids were separated by reverse phase chromatography using a 1.7uM 2.1x100 mm BEH Shield Column (Waters, Milford, MA) and an Acquity UPLC system (Waters, Milford, MA). The column was equilibrated with buffer A and 5 µl of sample was injected via the autosampler. Samples were eluted with a step gradient to 99% buffer B consisting of acetonitrile/isopropanol = 50/50 (v/v). Gradient elution was carried out for 5 min at a flow rate of 0.5 mL/min. Gradient conditions were as follows: 0-4.0 min, 0.1-55% B; 4.0-4.5 min, 55-99% B; 4.5-5.0 min, 99% B;
The liquid chromatography effluent is interfaced with a mass spectrometer and mass spectral analysis was performed on an AB SCIEX 6500 QTrap mass spectrometer equipped with an IonDrive Turbo V source (AB SCIEX, Framingham, MA). Eicosanoids were measured using multiple reaction monitoring (MRM) pairs with the instrument operating in the negative ion mode. The electrospray voltage was -4.5 kV, and the turbo ion spray source temperature was 525 °C. Collisional activation of the eicosanoid precursor ions was achieved with nitrogen as the collision gas, and other mass spectrometer parameters including the declustering potentials and collision energies were optimized for each analyte. The eicosanoids were identified by matching their MRM signal and chromatographic retention time with those of pure identical standards.E13
Eicosanoids were quantitated by the stable isotope dilution method. Briefly, identical amounts of deuterated internal standards were added to each sample and to all the primary standards used to generate standard curves. To calculate the amount of eicosanoids in a sample, ratios of peak areas between endogenous eicosanoids and matching deuterated internal eicosanoids were calculated. Ratios were converted to absolute amounts by linear regression analysis of standard curves generated under identical conditions. Currently, we can quantitate over 150 eicosanoids at sub-fmole levels. Data acquisitions were performed using Analyst 1.6.2 (Applied Biosystems, Foster City, CA) and Multiquant 2.0 (Applied Biosystems) was used for data analysis.
RNA extraction
Total RNA was extracted from cells using QIAshredder columns and RNeasy mini kit and treated with DNase (Qiagen). Total RNA concentration was measured by the Nanodrop ND-1000 spectophotometer (NanoDrop Technologies, Wilmington, DE) and the quality of RNA was confirmed by OD 260/280 ratio. The purified total RNA was stored at -80C until all samples were collected.
Gene arrays analysis
Total RNA (100 ng) was processed for GeneChip analysis using the 3’ IVT Express kit and the U133 plus 2.0 microarray chip (Affymetrix, Santa Clara, CA) following the manufacturer’s directions. The entire data set has been submitted to the National Center for Biotechnology Information Gene Expression Omnibus (NCBI GEO; accession number GSE59019.
Gene expression was preprocessed by the Robust Multichip Analysis (RMA) algorithm using Affymetrix Expression Console. Statistical analyses were performed using the Mathematical and Statistical Computing Laboratory (MSCL) Analyst's Toolbox ( written by P. J. Munson and J. Barb in the JMP scripting language (SAS Institute, Cary, NC). One way ANOVA and post-hoc tests were performed for comparisons between groups, and resulting p-values were converted to False Discovery Rate (FDR) to adjust for multiple testing. Hierarchical cluster analysis was done using JMP version 9. Networks and functional analyses were generated through the use of Ingenuity Pathway Analysis (IPA) (Ingenuity Systems, and verified by Gene Set Enrichment Analysis (GSEA, Broad Institute, resources. E14, E15 IPA-generated canonical and biofunctional pathways were rank ordered according to their significance score. Significance scores of those pathways are indicated by the Benjamini Hochberg P-value, which represents results adjusted for multiple comparisons (FDR).
In detail,we analyzed first the effect of LMW HA on the global gene expression in all subjects enrolled in the study. Principal component analysis and hierarchical clustering revealed that there was a strong effect of HA stimulation in all samples. Separation by HA stimulation was much stronger than separation by the disease phenotype. Therefore, we first compared the HA-treated group to the vehicle-treated group in all subjects. We found that HA stimulation significantly changed expression of 2,273 genes (FDR ≤ 0.001 and FC ≥1.5 or FC≤0.67), among which 1,030 were up-regulated and 1,243 down-regulated (Table E3). Each phenotype analyzed separately for the effect of HA stimulation also yielded a large number of genes significantly regulated by HA (Table E3). The list of significantly changed genes (by FDR ≤ 0.001 and FC ≥1.5 or FC≤0.67) was then analyzed by IPA to find enriched canonical pathways, associated diseases and bio functions. IPA transforms FC < 1 into the negative values and thus all values are presented accordingly, being called FC for simplicity. Analysis of the effects of HA stimulation in combined phenotypes yielded a comparable rank-order list of the most significant pathways and bio functions as the similar analysis in each phenotype separately. Thus, we first report here the results of overall effect of HA in all subjects (Fig. E3, A, B), followed by the significant differences in HA effect between phenotypes. Stimulation of PBMCs with LMW HA led to a significant enrichment of genes in 177 canonical pathways (Table E4). The 12 most significant pathways are presented in Fig. E3, A. In these pathways, HA stimulation led to an overall increase of gene expression (indicated by red color in the stacked bar in Fig E3, A; in detail presented in Tables E5-E9). Profound changes occurred in the gene expression in the Granulocyte Adhesion and Diapedesis pathway (Fig E3, A, Table E5), Triggering Receptor Expression on Myeloid Cells 1 (TREM1) Signaling pathway (Fig E3, A, Table E6) and Role of Pattern Recognition Receptors of Bacteria and Viruses pathway (Fig E3, A, Table E7). There was also a strong enrichment in pathways associated with IL-10 (Fig E3, A, Table E8) and IL-17 signaling (Fig E3, A, Table E9). Overall, the data suggest an association of HA signaling with immune cells trafficking and activation. Indeed, further IPA analysis showed the most significant involvement of HA-induced genes in Immune Cell Trafficking from the Physiological System Development and Function category, Cellular Function and Maintenance from the Molecular and Cellular Functions category as well as Inflammatory Response from the Diseases and Disorders category (Table E10). Not surprisingly, in light of the lipidomics results, Lipid Metabolism was also among the significantly enriched functional categories, with eicosanoid metabolism as one of the main subcategories (Fig. E3, B). We confirmed the LMW HA-induced changes in the expression of selected genes by real-time PCR (Table E11) and by their translation to protein by Multiplex (Fig. E4).
Next, we analyzed the specific effect of LMW HA on gene expression in asthma. We compared the global gene expression changes induced by LMW HA, corrected by the baseline gene expression in each individual (paired difference), in mild-to-moderate asthma and in severe asthma compared to non-asthmatic controls. The clustering relationship in PCA did not show a clear separation among those phenotypic groups. Likewise, by applying one-way ANOVA and post-hoc tests, we did not find any genes that passed the criteria of FDR 0.05 and FC ≥1.5 or FC≤0.67, which is a common observation in gene expression studies in heterogenic phenotypes. Therefore, we applied further IPA analysis on genes which passed the criteria of p<0.05 and FC ≥1.5 or FC≤0.67 to find any changed pathways and bio functions, which were further corrected for multiple comparisons. Applying these conditions, we found 6 canonical pathways significantly changed in severe asthma vs controls (Fig. 2, A), while there were no significantly changed pathways in mild-to-moderate asthma vs controls. Surprisingly, these pathways were created from significantly less upregulated genes as compared to their expression after HA treatment in control subjects; therefore, they are designated by green color in the stacked bar graph (Fig. 2, B). These pathways include genes involved in interferon signaling, antiviral innate immunity, pattern recognition receptors, cell movement and apoptosis. Downstream analysis of bio functions and diseases revealed that the Infectious Disease category was significantly changed and contained several subcategories with the highest prediction score for increase of function (Table E12) in the entire data set. This further corresponded to the significantly changed Antimicrobial Response category, containing a predicted decrease of antiviral response (p=2.94E-7; z-sore=-1.94) and antibacterial response (p=1.66E-6; z-score=-1.94), as well as decrease in cell movement of lymphocytes (6.12E-3; z-score=-2.02) and homing T lymphocytes (5.11E-3; z-score=-1.99). Detailed analysis of these categories and functions led to establishing associations with the canonical pathways and building the model of those associations in severe asthma (Fig. 2, C). This model predicts that less upregulated expression of genes by LMW HA in the Interferon signaling and other canonical pathways in severe asthmatics might lead to impaired activation of antiviral and antimicrobial responses and allow for a subsequent increase in Viral Infection signaling (Fig. 2, C). Surprisingly, the Lipid Metabolism category, including eicosanoid synthesis was not significantly enriched in severe asthma, suggesting that posttranscriptional or posttranslational changes (e.g. phosphorylation) are responsible for the significant increases in the designated lipid species.
Control analyses
We performed several analyses to control for the baseline (unstimulated) phenotype influence on the analysis results. First, as mentioned above, the comparison of LMW HA effect between phenotypes was performed after paired analysis and transformation to fold change as compared to baseline expression in the same subject. Therefore, the potential baseline phenotypic differences should be eliminated. Second, to additionally control whether baseline down regulation or pre-activation of genes could lead to their subsequent less potent activation by LMW HA, we reanalyzed separately baseline, non-stimulated expression of a subset of 84 genes involved in significantly changed canonical pathways and bio functions between controls and severe asthmatics after LMW HA stimulation. We did not find any changes in the basal expression of those genes between different phenotypes (p>0.05 in each gene, data not shown). Third, there were no significant differences of the baseline expression of any analyzed proteins (Fig 2D, Fig. E4). Finally, to exclude the effects of corticosteroids on the analysis results we examined the effect on the Glucocorticosteroid Receptor Signaling pathway. We found that this pathway was unchanged in severe asthma as compared to controls. These results suggest that the significantly less upregulated pathways after LMW HA treatment in severe asthmatics were not likely due to a systemic effect of medications or baseline down or up regulated expression of those genes. Additionally, we performed confirmatory real-time PCR to re-analyze mRNA expression of cPLA2 (PLA2G4A) and various hyaluronan receptors (TLR4, CD44, TLR2, Rhamm (HMMR) and stabilin 2 (STAB2)) with/without LMW HA stimulation. We studied whether the differences in their expression might be responsible for the observed differences in the lipid mediators and/or gene expression profiles between severe asthmatics and controls. Our real-time PCR results confirmed that PLA2G4A, TLR4, CD44, TLR2 and HMMR mRNA expression increased upon LMW HA stimulation, but there was no difference between studied groups (Figure E5). The expression of STAB2 was very low or undetectable in several patients. Therefore, these data suggest that posttranscriptional changes in the signaling pathways might be responsible for the observed differences between severe asthmatics and controls.
Involvement of several eicosanoids is very important in the pathogenesis of AERD (Aspirin Exacerbated Respiratory Disease). Thus, even though we enrolled to our study only 2 patients with aspirin (acetyl salicylic acid; ASA) sensitivity, based on their medical history, as stated in Table E1, we reanalyzed our lipidomic data, including ASA-sensitivity to highlight those patients on the graph (Figure E2). There was no statistical difference between patients with and without aspirin sensitivity (p>0.05, Mann-Whitney test) in severe asthma group. Even though, one of the patients with aspirin sensitivity had the highest level of eicosanoids within the severe asthma group, the other one was grouped differently, depending on which mediator was analyzed. Moreover, often in the mild-to-moderate asthma group, there were patients with higher values of several mediators. We think that this issue, not robust and hugely underpowered in this report, requires further studies in the larger population of subjects.
Real-Time PCR
Reverse transcription was performed using an iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA). Gene expression was assessed using RT-PCR performed on an ABI Prism ViiA7 sequence detection system (Applied Biosystems) using commercially available probe and primers sets (Applied Biosystems) as follows: PTGS2 (COX2), Hs00153133_m1; PTGES, Hs01115610_m1; PLA2G4A (cPLA2a), Hs00233352_m1 ; INHBA, Hs01081598_m1; IL36G, Hs00219742_m1; ALOX5, Hs01095330_m1; FN1, Hs00365052_m1; LEP, Hs00174877_m1; CD44, Hs01075862_m1; TLR4, Hs00152939_m1; TLR2, Hs01014511_m1; HMMR (Rhamm), Hs00234864_m1; STAB2 (stabilin-2), Hs00213948_m1; and iTaq Universal Probes Supermix (Bio-Rad). Gene expression was normalized to GAPDH transcripts and represented as a relative quantification (RQ) compared with vehicle treated control.