Christenson, Stephanie A

Most genetic variants associated with complex traits are hypothesized to regulate gene expression. To understand the genetics underlying gene expression variability, we characterized 14,324 RNA-sequencing samples from the Trans-Omics for Precision Medicine program and performed expression and splicing quantitative trait locus (e/sQTL) analyses in six tissues and cell types, including whole blood ( n = 6454) and lung ( n = 1291). We detected tens of thousands of secondary cis-e/sQTLs, showing that secondary cis-e/sQTL discovery remains unsaturated. We fine-mapped UK Biobank–derived genome-wide association study (GWAS) signals from 164 traits and identified e/sQTL colocalizations for 10,611 GWAS signals, including 7096 that colocalize with secondary e/sQTLs. Our results suggest that even larger e/sQTL analyses will uncover additional secondary e/sQTLs, further benefiting GWAS interpretation.

Introduction Despite the identification of multiple susceptibility loci by genome-wide association studies (GWAS), considerable chronic obstructive pulmonary disease (COPD) heritability remains unexplained. Aim To identify interaction networks of airway epithelial cell DNA methylation in COPD and further explore potential correlations with airway bacterial composition, as potentially collective regulators of biological pathways influencing COPD severity. Methods Using DNA isolated from bronchial airway brushings of 67 ever-smokers (>20 pack-years) from the SubPopulations and InteRmediate Outcomes Measures in COPD Study (SPIROMICS), we assessed proportion of DNA methylation (β) by epigenome-wide association study (EWAS) and examined associations of differentially methylated CpG probes (DMPs) with risk for moderate-to-severe COPD (N=34) versus absent or mild COPD (N=33). We tested co-methylation modules generated by Weighted Correlation Network Analyses (WGCNA) for associations with moderate-to-severe COPD and with bacterial genus-level relative abundances (16S rRNA sequencing). Results EWAS-identified nominally significant DMPs enriched for lung function GWAS loci. Eigengenes in six WGCNA modules were associated with moderate-to-severe COPD (false discovery rate <0.05). Four of those modules were enriched for forced expiratory volume in 1 s/forced vital capacity GWAS loci, and five overlapped with DMPs from EWAS. Overlapping CpG loci in three COPD-associated modules were adjacent to mucin genes; one had 10 genes highly ranked by connectivity with MUC5B, including important pathway genes: B3GNT6 , DGKI and ITGA8 . CpGs in an independent COPD-associated module showed the most correlations with Rothia , with directionality suggestive of negative associations with moderate-severe COPD. Conclusions Bronchial epithelial DNA methylation modules enriched for lung function GWAS loci associate with COPD severity in SPIROMICS. Potential module relationships to bronchial bacterial composition require further validation.

Abstract Introduction Resilient smokers, who maintain respiratory health despite smoking, remain poorly characterized. Airway epithelial cell gene expression in resilient smokers can provide insight into mechanisms and pathways that may protect smokers from developing smoking-related lung disease. The objective of this study is to identify genes and potential pathways associated with lung-related resilience. Methods Resilient smokers were defined based on a previously published multi-dimensional definition which includes the absence of: 1) spirometric airflow limitation, 2) symptoms (cough/sputum, dyspnea), 3) radiographic (HRCT) abnormalities, 4) exacerbations, and 5) an accelerated FEV1 decline. “Non-resilient” smokers had preserved spirometry but showed abnormalities in at least one domain of resilience. We analyzed airway epithelial brush samples from 21 resilient and 34 non-resilient smokers from SPIROMICS. Differential gene expression was performed using DESeq2 package in R, which estimates differential expression using generalized linear models. We adjusted for sequencing batch, age (continuous), race, gender, and current smoking status. While controlling for these covariates we evaluated differential gene expression between resilient and non-resilient smokers, the interaction between resilience and smoking status, and in each sub-category of resilient (absence of symptoms, radiographic abnormalities, exacerbations, and accelerated decline in FEV1). Results We identified 98 genes that were differentially expressed (FDR ≤0.1) when comparing resilient to non-resilient smokers, 20 of those genes were up regulated in resilient smokers (Figure 1). Notable genes included cadherin-12 (CDH12), aspartate beta-hydroxylase (ASPH), and protein kinase AMP-activated alpha 2 (PRKAA2). CDH12 is involved in cell adhesion and epithelial integrity. ASPH influences cell adhesion and signaling. PRKAA2 plays role in cellular homeostasis and stress response, which may be altered in non-resilient smokers, affecting cellular metabolism and inflammation. BPI folding containing family A member (BPIFA1) was significantly associated with smoking status in the interaction analysis, suggesting its role in modulating the impact of smoking on resilience. BPIFA1 has anti-microbial properties, regulates immune responses, and contributes mucosal surface protection. Sub-category analysis identified 17 genes associated with resilience to symptoms and 2 genes related to physiologic resilience (FEV1 decline), though no significant genes were associated with radiographic or exacerbation resilience. Conclusion Differential gene expression of airway epithelial cells identified distinct expression patterns between resilient and non-resilient smokers. Specifically, genes associated with epithelial cell adhesion, signaling, and cellular homeostasis were up regulated in resilient smokers. These findings suggest that resilient smokers may have transcriptomic profiles and potential mechanisms that mitigate the effects of cigarette smoke.

Abstract Rationale: MUC5B and MUC5AC are neighboring genes on chromosome 11p15 that encode the two predominant mucins secreted in the respiratory airways. Increased airway mucin concentrations and an imbalance between these two mucins drive COPD symptoms and the development of airflow obstruction or COPD, but the epigenomic regulation of this loci has not been characterized in the airways. Methods: We isolated DNA from bronchial airway brushings of 72 participants with &gt;20 pack-years smoking history from the Subpopulations and Intermediate Outcomes Measures in COPD Study (SPIROMICS). 36 participants had GOLD-based moderate-to-severe COPD (FEV1/FVC&lt;0.70, FEV1&lt;80%) and 36 were without or with mild COPD. We used the Illumina Infinium Methylation EPIC BeadChip Array to assess proportion of DNA methylation (β) spanning 689,937 CpG sites. We evaluated proportional DNA methylation (β) at 96 CpG loci within a 103kB region using matrixQTL for cis-expression quantitative trait methylation (eQTM: 10kB flanks relative to MUC5AC/MUC5B) integrating bulk RNA-seq data and mQTL models integrating gene sequencing data to identify DNA variation associated with proportion eQTM methylation. Models included age, sex, BMI, clinical site, current smoking vs past smoking, total pack-years, three ancestral components of ancestry, chipset, and position. Results: We identified a top eQTM, cg00524486, marginally associated with MUC5B expression (coefficient estimate=8.75, punadjusted=0.0014, pFDR=0.065, Figure 1B). This eQTM was associated with CT scan parametric response mapping-based functional small airways disease (PRM-fSAD, beta coefficient[β]=3.64, p=0.011). We tested for DNA variation associated with methylation of this top eQTM which identified two methylation quantitative trait loci (mQTL: rs885454 [β=0.044, minor allele frequency[MAF]=0.09, pFDR=0.01], rs868902 [β=0.020, MAF=0.43, pFDR=0.08, Figure 1A) associated with cg00524486 methylation, of which the more frequent mQTL, rs868902, was marginally associated with PRM-fSAD (p=0.07) in the larger SPIROMICS cohort with MUC5B DNA sequencing data. Conclusions: This preliminary multi-omic study of the lower airway epithelium demonstrates that the expression of MUC5B is epigenetically regulated by neighboring CpG probe methylation (eQTM) represented by DNA variation (mQTL). These studies require further confirmation in a larger dataset of airway samples and support a role for epigenetic variation as a biologic contributor and biomarker of early COPD pathogenesis.

Abstract Rationale Chronic obstructive pulmonary disease (COPD) manifests later in life and is a leading cause of death worldwide. While smoking-related lung pathology is a long-recognized COPD etiotype, emerging evidence suggests that early-life growth conditions also play an important role. A recently-developed index of early-life growth conditions – the difference between attained height and genotype-predicted height (height-GaP) – has been validated against prospectevely ascertained adverse early-life growth conditions. This study used height-GaP to evaluate the association of early-life growth conditions with COPD phenotype and prognosis later in life. Methods Among current and former smokers with spirometry-defined COPD in the SubPopulations and Intermediate Outcomes Measures in COPD Study (SPIROMICS), we computed height-GaP as the difference between attained height and genotype-predicted height. Regression models were fit for respiratory mortality (defined by primary diagnostic code), dyspnea prevalence (Medical Research Council dyspnea rating &gt;1), chronic bronchitis prevalence, airflow obstruction severity as quantified by FEV1Q, and functional capacity (six-minute walk distance), adjusting for age, age2, sex, race-ethnicity, genotype-predicted height, cigarette smoking status and pack-years. Sensitivity analyses accounted for competing risks of death (Fine-Gray subdistribution hazard model) and accounted for later-life height loss (restricting sample to participants under 50 years). Results Among 1,258 participants with COPD (mean±SD age: 65.4±8.0 years, 42.4% female, mean±SD pack-years: 53.6±28.2, mean±SD FEV1%: 63.3±23.9%), there were 202 respiratory deaths over a median 7.5 years of follow-up (Figure). Larger height-GaP deficit was associated with higher respiratory mortality (adjusted hazard ratio [aHR]: 1.30 [95%CI: 1.12 to 1.50] per 1-SD increment in height-GaP deficit). This association remained significant after accounting for competing events (p=0.001) and when restricted to participants under 50 years (p=0.032). Larger height-GaP deficit was associated with higher dyspnea prevalence (adjusted prevalence ratio: 1.13 [95%CI: 1.02 to 1.25] per 1-SD increment in height-GaP deficit), lower FEV1Q (adjusted mean difference: -0.37 [95%CI: -0.45 to -0.29] per 1-SD height-GaP deficit), and shorter 6-minute walk (adjusted mean difference: -16.6 m [95%CI: -23.1 m to -10.0 m] per 1-SD increment in height-GaP deficit), but not chronic bronchitis or CAT score ≥10 (p&gt;0.05 for each). Conclusions Among current and former smokers with COPD, a quantitative index of early-life conditions was associated with higher respiratory mortality, in addition to dyspnea, airflow obstruction severity and reduced functional capacity. These results suggest that early-life growth conditions may influence COPD phenotype and prognosis later in life.

Abstract Rationale: Individuals with chronic obstructive pulmonary disease (COPD) have different clinical trajectories, some with minimal symptom burden, others experiencing recurrent exacerbations, the most severe of which require hospitalization. Existing diagnostics, including spirometry and symptom assessments, lack the precision to identify high-risk patients who may benefit from timely preventive therapies. Nuclear cell-free DNA (ncfDNA) is a well-established marker of systemic tissue injury, implicated in amplifying inflammation by triggering type 1 interferon release by activating DNA-sensing pathways. Because COPD is accompanied by inflammation and tissue injury systemically, we hypothesized that ncfDNA is associated with increased risk of severe exacerbations requiring hospitalization. Methods: Design: We utilized baseline plasma samples randomly selected from 672 individuals with COPD from the Subpopulations and Intermediate Outcomes Measures in COPD Study (SPIROMICS) cohort, which prospectively collected exacerbation numbers and type. The primary endpoint was exacerbation requiring hospitalization. Secondary endpoints included total number of exacerbations (defined as use of oral steroids, antibiotics, or healthcare utilization) and severe exacerbations (defined as requiring emergency room care or hospitalization). Measure: We measured plasma ncfDNA using digital droplet PCR. Analysis: We assessed the association of ncfDNA with exacerbation outcomes using negative binomial regression or Cox regression, adjusting for age, percent post-bronchodilator forced expiratory volume in one second (FEV1), supplemental oxygen use, and smoking status. Results: In total, 57% (n=383) of COPD participants experienced exacerbations, 29% (n=197) at least one severe, and 22% (n=145) requiring hospitalization, the latter at a median of 622 days from enrollment (IQR=287-997). Higher ncfDNA levels at enrollment correlated with measures of COPD severity, including baseline FEV1 (beta = -0.12, 95% CI -0.18 to -0.049, p &lt;0.001). Higher ncfDNA associated with increased risk of COPD exacerbation requiring hospitalization (incidence rate ratio (IRR) 1.40, 95% CI 1.12-1.75, p =0.003) and severe exacerbations (IRR 1.36, 95% CI 1.11-1.67, p=0.002) and a trend towards total exacerbations (IRR 1.11, 95% CI 0.98-1.26, p= 0.09). The primary outcome, risk of time-to-first hospitalization, correlated significantly with higher ncfDNA with hazard ratio (HR) 1.23, 95% CI 1.01-1.49, p=0.037. Conclusion: These results suggest that ncfDNA may serve as a predictor for severe and hospitalization-associated COPD exacerbations. To better understand the relationship between ncfDNA and COPD pathophysiology, follow-up studies should explore longitudinal trends and cfDNA tissue sources, and their link to COPD exacerbations.

Abstract Rationale: We have previously shown reduced total IgG, IgG subclasses, and levels/function of pneumococcal-specific antibodies are associated with increased risk for exacerbations of COPD (ECOPD). The binding of IgG to its receptor, the Fc receptor for IgG (FcγR) on immune cells, has a critical role in determining antibody levels and function. In addition to our research demonstrating that IgG heavy chain gene (IGHG) variants are associated with ECOPD, single nucleotide variants (SNVs) in genes encoding FcγR (FCGR1-3, which encode receptor subtypes FcγRI, FcγRIIa/b/c, and FcγRIIIa/b) may also result in antibody deficiency and increased exacerbation risk. Due to technical challenges in accurately identifying SNVs in FCGR, this region has not been well-characterized. The limited data that exist are from people of European descent, and prior studies have not included African American participants. We therefore sought to investigate the hypothesis that distinct SNVs in genes encoding FcγR are associated with disease severity and history of exacerbations in African American people with COPD. Methods: We developed a novel pipeline to identify SNVs in the coding (CDS) regions of chromosome 1 using short-read whole genome sequence data. The pipeline was benchmarked using paired long-read samples from Genome in a Bottle. SNVs were called within CDS regions and analyzed in a meta-analysis of self-reported African American participants in SPIROMICS and COPDGene (n=3063). Genetic variants were tested for association with COPD severity (GOLD spirometry stage 1-2 vs 3-4) as well as COPD exacerbations, adjusting for age, sex and principal components controlling for population structure. Results: Eight SNVs were significantly associated with frequent ECOPD (≥2 in one year) in the meta-analysis of African American COPDGene and SPIROMICS participants (Table). Seven of these 8 SNVs were found in FCGR3; 4 in the activating FCGR3A subtype, and 3 in FCGR3B. Seven of 11 SNVs associated with COPD severity were in FCGR3 (4 in FCGR3A; 3 in FCGR3B). One SNV (rs371849) was associated with both frequent exacerbations (P=0.002) and COPD severity (P=0.044). Conclusions: We used a cutting-edge variant calling pipeline to identify multiple SNVs in FCGR, primarily involving FcγRIII receptor subtypes, associated with COPD severity and exacerbations in African American participants in SPIROMICS and COPDGene. FcγRIII variants are known to cause impaired antibody levels and function in people of European descent. The structural and functional importance of newly recognized variants in African American cohorts require further study, as they may represent immunogenetic risk factors for antibody deficiency and exacerbations.

Abstract Rationale. Bronchoscopy can directly assess cells in airway epithelium in participants with COPD, but it is invasive, expensive, difficult to perform at scale, and contraindicated in severe disease. Blood is much more easily available but contains only a limited subset of cell types, many of which are less relevant to lung biology. Recent advances in RNA-sequencing technology have allowed for reliable measurement of transcriptomics in a wider variety of more accessible tissues, including hair follicle cells. We collected and performed bulk RNA-seq on hair follicles as an ancillary study within the NIH-funded SubPopulations and Intermediate Outcome Measures in COPD Study (SPIROMICS) and investigated their cell type composition to better understand the opportunities and limitations of this tissue for the study of lung disease. Methods. SPIROMICS recruited individuals with and without COPD aged 40-80 years with ≥ 20 pack-years smoking history. Participant at several sites underwent collection of approximately 10 hair follicles from the head, which were immediately frozen at -80ºC in 2018-21. mRNA in hair follicles was measured by bulk RNA-seq, and cell type deconvolution was performed using GEDIT v1.7 with the provided BlueCodeV1.0.tsv reference matrix. Cell types were collapsed into broader categories by summing the estimated cell type proportions. Results. The cell type deconvolution of the hair follicle bulk RNA-seq from 80 participants with and without COPD estimated that the tissue was largely composed of immune cells (median 36%, including B cells, T cells, neutrophils, macrophages, NK cells, and monocytes) and epithelial cell types (median 31%), with the largest subgroup of epithelial cell types identified as hair follicular keratinocytes (Figure 1). The next largest component was fibroblasts (median 9%), followed by endothelial cells (median 8%). Signals for myocytes (median 5%), melanocytes (median 5%), and chondrocytes (median 5%) were also present. Conclusion. This initial study demonstrated promising signals for a variety of lung-relevant cell types in hair follicles in SPIROMICS, including immune and epithelial cells in addition to fibroblasts and endothelial cells that are not otherwise easily accessible. Hair follicle gene expression profiling shows promise for understanding numerous facets of lung biology and disease that both complements and exceeds that of other noninvasive tissues such as blood.

Abstract Rationale: The club cell secretory protein (CC16), encoded by the SCGB1A1 gene, has an anti-inflammatory role in airways diseases, including asthma (Li, AJRCCM 2023; Voraphani, AJRCCM 2023). We hypothesize that additional pathway genes regulated or co-expressed with CC16 have cumulative effects on asthma severity outcomes by modulating airway inflammation and maintaining lung homeostasis. Methods: As published previously, four genes were downregulated in CC16 knockout experiments and correlated with CC16 levels: BPIFA1, SFTPD, LTF, and LYZ (Iannuzo, Front. Immunol 2023). We developed a four-gene biomarker score in 94 patients with asthma from the NHLBI-sponsored Severe Asthma Research Program (SARP3) who underwent research bronchoscopy for lower airway epithelial brushings, RNA isolation, and bulk RNA sequencing. We log-transformed the raw values of each transcript; obtained standardized residuals from regression models that included sex, age, and batch effects; and summed the standardized residuals of all four proteins to estimate a standardized sum score (Zhai, JACI Allergy 2024). Logistic regression models tested the four-gene score, three-gene score without LTF, and individual transcripts for associations with exacerbations during 12-month follow-up and zero-inflated negative binomial regression for number of exacerbations. Results: 30.85% percent of 94 patients experienced at least one exacerbation at the end of follow-up. The CC16-associated four-gene biomarker score showed protective effects on asthma exacerbation risk with a 54% percent reduction in exacerbation risk for every one-fold higher score (OR=0.46, 95% CI: 0.27-0.78, p =0.0040, Figure 1). Transcript expression of BPIFA1 (OR=0.87, 95% CI:0.66-1.16, p=0.337), SFTPD (OR=0.33, 95% CI:0.04-2.52, p=0.283), LYZ (OR=0.62, 95%CI:0.28-1.39, p=0.248), and LTF (OR=0.36, 95% CI:0.18-0.74, p=0.0053) all indicated protective directionality with respect to exacerbation risk. However, only LTF demonstrated a statistically significant association. When we excluded LTF, the remaining three-gene score showed a trend toward association with risk (OR=0.64, 95% CI:0.40-1.03, p=0.068) and was significantly associated with number of exacerbations (RR=0.49, 95% CI:0.26-0.95, p=0.034). Compared to LTF alone (OR=0.48, PPV=0.50, NPV=0.71, Sensitivity=0.14, AUC=0.65), the combined four-gene score (OR=0.46, PPV=0.60, NPV= 0.73, Sensitivity=0.21, AUC=0.68) showed improved predictive performance. Conclusions: We identified a CC16-associated transcriptomic predictive score for asthma exacerbations, with LTF showing the strongest effect. Although the other three genes showed individually weak and non-significant associations with exacerbations, the three gene score was significant for number of exacerbations.We demonstrate the potential of this transcriptomic score as a predictive biomarker for asthma severity-related outcomes. Further research investigating these genes and their interaction with other CC16-associated immune pathways could provide insight into therapeutic targets.

Abstract Asthma is a heterogeneous disease with variable presentation and characteristics. There is a critical need to identify underlying molecular endotypes of asthma. We performed the largest transcriptomic analysis of 808 bronchial epithelial cell samples across 11 independent cohorts, including 3 cohorts from the Severe Asthma Research Program. Using seven datasets (218 patients with asthma, 148 healthy control subjects) as discovery cohorts, we identified 505 differentially expressed genes, which we validated in the remaining four datasets. Unsupervised clustering using the 505 differentially expressed genes identified four reproducible clusters of patients with asthma across all datasets, corresponding to healthy control subjects, patients with mild/moderate asthma, and patients with severe asthma with significant differences in several clinical markers of severity, including pulmonary function, Type 2 inflammation, fractional exhaled nitric oxide, and maximum bronchodilator reversibility. Importantly, we found the same clusters in pediatric patients using nasal lavage fluid cells, demonstrating the gene signature and clusters are not confounded by age and are conserved in both lower and upper airways. The four asthma clusters may represent a unifying framework for understanding the molecular heterogeneity of asthma. Further study could potentially enable a precision medicine approach of matching therapies with patients with asthma most likely to benefit.