Regulation of MUC5B Expression in Idiopathic Pulmonary Fibrosis

Abstract

The gain-of-function mucin 5B (MUC5B) promoter variant, rs35705950, confers the largest risk, genetic or otherwise, for the development of idiopathic pulmonary fibrosis; however, the mechanisms underlying the regulation of MUC5B expression have yet to be elucidated. Here, we identify a critical regulatory domain that contains the MUC5B promoter variant and has a highly conserved forkhead box protein A2 (FOXA2) binding motif. This region is differentially methylated in association with idiopathic pulmonary fibrosis, MUC5B expression, and rs35705950. In addition, we show that this locus binds FOXA2 dynamically, and that binding of FOXA2 is necessary for enhanced expression of MUC5B. In aggregate, our findings identify novel targets to regulate the expression of MUC5B.

Keywords: DNA methylation; gene regulation; idiopathic pulmonary fibrosis; mucin 5B; mucins.

Figure 1.

The mucin 5B (MUC5B) promoter variant, rs35705950, and idiopathic pulmonary fibrosis (IPF) contribute to the expression of MUC5B in the lung. (A) Sample data: samples are divided into control and IPF, and then further divided by the rs35705950 genotype. Sample number is indicated by raw counts and percent of population, sex is represented as percent male, and average age is the mean age (±SD). Smoking data are represented as Ever (those who no longer smoke, but have a history of smoking, and current smokers) and Never (those who report no history of smoking). (B) Lung MUC5B expression levels are plotted with each point representing an IPF (red) or a control (black) sample. Samples are further separated by genotype at the rs35705950 locus, where G is the wild-type allele and T is the common variant. Relative expression of MUC5B on the y axis is calculated by subtracting the change in cycle threshold between MUC5B and GAPDH by 20 (the highest reading). Data are presented as mean (±SEM). This process is used to invert the values for graphical purposes. Statistical analyses (two-tailed t tests) were conducted on noninverted values. **P < 0.0001.

Figure 2.

Methylation analysis of the MUC5B promoter (A). Disease-associated differentially methylated region (DMR). Linear regression modeling of IPF with DNA methylation and covariates (age, sex, smoking status, and genotype at rs35705950) with multiple testing and regional interaction correction. (B) MUC5B expression–associated DMR: linear regression modeling of MUC5B expression and DNA methylation with multiple testing and regional interaction correction. (C) rs35705950-associated DMR: linear regression modeling of the common MUC5B promoter variant with DNA methylation with multiple testing and regional interaction correction. The horizontal line indicates a stringent multiple testing correction threshold for significance.

Figure 3.

Forkhead box protein A2 (FOXA2) binding site is critical for promoter activity. (A) The FOXA2 binding motif in the MUC5B promoter construct was disrupted using site-directed mutagenesis. A “CT” insert, “AAA” → “C” in/del and a “C” → “T” point mutation were made. (B) The MUC5B wild-type (WT) promoter construct clones (blue; pGL4.10 with MUC5B promoter insert) show that the 4.2-kb region directly upstream of the MUC5B transcriptional start site has increased promoter activity relative to the empty vector (gray; pGL4.10 with no insertions). After site-directed mutagenesis of the conserved FOXA2 motif in the WT reporter, the FOXA2-mutant reporter clones (yellow) show a significant decrease in promoter activity (H358 data shown; P = 1.7 × 10⁻⁸). Two-tailed t tests were used for all analyses. Each bar represents an individual clone, with the error bars indicating SEM for technical replicates. EV, empty vector; In/del, dual insertion and deletion.

Figure 4.

Chromatin immunoprecipitation (ChIP)–quantitative PCR (qPCR) occupancy of FOXA2 and RNA polymerase II (A). The two regions targeted by qPCR were at the predicted enhancer site (red) and the MUC5B transcription start site (TSS; blue). The arrows at each site represent the locations of the qPCR primers. In addition, positive control primers were designed at the ENCODE FOXA2 peaks with the highest occupancy (in green in B) and at the transcriptional start site of a transcriptionally active gene for Pol2 (green in C). (B) FOXA2 shows high occupancy at the predicted enhancer with a 201-fold enrichment (P = 1.6 × 10⁻⁹) over the negative controls (gray). (C) RNA polymerase II (Pol2) occupancy is enriched for at the TSS (15-fold, P = 0.015), indicating that MUC5B is being actively transcribed. In addition, Pol2 is enriched for at the enhancer, increasing the likelihood that this is an active enhancer motif (14-fold, P = 0.0009). Two-tailed t tests were used for all analyses with results shown as mean (±SEM).

Figure 5.

Primary tracheal cell FOXA2 ChIP-qPCR binding of FOXA2 to the enhancer was assayed in undifferentiated submerged cultures (B) and differentiated air–liquid interface (ALI) cultures (A). There is increased occupancy of FOXA2 in both the submerged (62-fold, P = 5.1 × 10⁻¹²) and ALI cultures (12-fold, P = 1.3 × 10⁻⁹). The relative occupancy of FOXA2 at the enhancer was increased by 5.1-fold (P = 1.9 × 10⁻⁸) in the MUC5B-expressing ALI cultures compared with the nonexpressing submerged cultures. The red bars represent the predicted enhancer site (see Figure 4A), the green bars represent positive control primers, and the gray bars represent negative control primers. Two-tailed t tests were used for all analyses with results shown as mean (± SEM).

Figure 6.

Effects of short interfering RNA (siRNA) knockdown of transcription factors on MUC5B expression. A total of 24 transcription factors predicted to have binding near rs35705950 were knocked down using siRNA, and the effect on MUC5B expression was measured using qPCR. Fold change in MUC5B expression was calculated using the ∆∆CT method, and significance was calculated using a standard t test with a corrected P value of significance of 0.002. Four transcription factors significantly down-regulated MUC5B expression: FOXA2, HOXA9, STAT3, and ZBTB7A (*P < 0.002). CEBPB, CCAAT/enhancer binding protein β; E2F1, E2F transcription factor 1; ELF-1, E74 like ETS transcription factor 1; EP300, E1A binding protein p300; ESR1, estrogen receptor 1; HDAC2, histone deacetylase 2; HNF4, hepatocyte nuclear factor 4; HOXA9, homeobox A9; JUND, JunD proto-oncogene, AP-1 transcription factor subunit; NFIC, nuclear factor I C; PAX, paired box family of transcription factors; REST, RE1 silencing transcription factor; RXRA, retinoid X receptor α; SP1, Sp1 transcription factor; STAT3, signal transducer and activator of transcription 3; TCF12, transcription factor 12; TEAD4, TEA domain transcription factor 4; ZBTB, transcription repressor. Results are shown as the mean of the replicates (± SEM).