Role of endothelial cells in pulmonary fibrosis via SREBP2 activation

Idiopathic pulmonary fibrosis (IPF) is a progressive lung disease with limited treatment options. Despite endothelial cells (ECs) comprising 30% of the lung cellular composition, the role of EC dysfunction in pulmonary fibrosis (PF) remains unclear. We hypothesize that sterol regulatory element-binding protein 2 (SREBP2) plays a critical role in the pathogenesis of PF via EC phenotypic modifications. Transcriptome data demonstrate that SREBP2 overexpression in ECs led to the induction of the TGF, Wnt, and cytoskeleton remodeling gene ontology pathways and the increased expression of mesenchymal genes, such as snail family transcriptional repressor 1 (snai1), α-smooth muscle actin, vimentin, and neural cadherin. Furthermore, SREBP2 directly bound to the promoter regions and transactivated these mesenchymal genes. This transcriptomic change was associated with an epigenetic and phenotypic switch in ECs, leading to increased proliferation, stress fiber formation, and ECM deposition. Mice with endothelial-specific transgenic overexpression of SREBP2 (EC-SREBP2[N]-Tg mice) that were administered bleomycin to induce PF demonstrated exacerbated vascular remodeling and increased mesenchymal transition in the lung. SREBP2 was also found to be markedly increased in lung specimens from patients with IPF. These results suggest that SREBP2, induced by lung injury, can exacerbate PF in rodent models and in human patients with IPF.


RNA-sequencing library construction and data processing
Total RNA was isolated from human umbilical vein endothelial cells (HUVECs) infected with Ad-SREBP2 or Ad-null (empty vector). mRNA was isolated using mirVana mRNA isolation kit (Thermo Fisher Scientific). Standard Illumina protocols were used to construct the RNA libraries and sequencing. Analysis was performed using base calling and quality scoring by using Real-Time Analysis version 2 (RTA v2) on the NextSeq 500 system. Data were demultiplexed and converted to FASTQ files using Bcl2fastq conversion software v1. 8.4. Sequence reads were trimmed of their adaptor sequences and masked for low complexity or low-quality sequence.

ATAC-sequencing library construction and data processing
Nuclei were isolated from ~5000 cells in lysis buffer (10 mM Tris-HCl, 10 mM NaCl, 3 mM MgCl2, and 0.1% NP-40). Nuclear pellets were resuspended in transposition buffer (Nextera DNA Library Preparation kit, Illumina) and incubated at 37°C for 30 min. DNA were then purified using Qiagen MinElute PCR purification kit. Libraries were sequenced using the Illumina HiSeq 2000. ATAC-seq data analysis followed the official pipeline of ATAqC (4) specification of the Encyclopedia of DNA Elements (ENCODE) consortium. Adapter sequences were trimmed from the raw reads by using cutadapt (5). Reads were aligned to the Human GRCh38 genome using bowtie2 (6). After read alignment, SAMtools (7), Picard's MarkDuplicates (8), and bedtools (9) were used to remove multi-mapped reads (MAPQ < 30), remove PCR duplicates reads, and convert alignment BAM format to tagAlign format, respectively. The retained alignment results were used to call peaks by MACS2 (10) with the false discovery rate (FDR) threshold of 0.01. Results of peak signal were imported into WashU Epigenome Browser for further visualization and analysis (Supplemental Figure 4). Dataset is available at GSE121781.

Western blot and qPCR analyses
Total protein was isolated using NP-40 and resolved by SDS-PAGE. Gels were transferred to polyvinylidene fluoride (PVDF) membranes and immunoblotted with antibodies against SREBP2 (Abcam, ab30682), snai1 (Cell Signaling, clone L70G2), aSMA (Abcam, ab124964), vimentin (Cell Signaling, clone D21H3), and b-actin (Santa Cruz Biotechnology, clone C4). Immunoblotted bands were visualized using enhanced chemiluminescence (Milipore). For qPCR, total RNA was isolated from cells or tissue using Trizol (Life Technologies). The extracted RNA was reverse transcribed using 5X PrimeScript reverse transcriptase (Takara), and normalized to b-actin as the internal control. Relative mRNA expression was determined using iQ TM SYBR Green PCR supermix (Bio-Rad) in the Bio-Rad Real-time detection system. Data was normalized using 2 -ΔCT [-(CT of target gene-CT of reference gene)] , followed by the 2 -ΔCT of each experimental group divided by that of the averaged control groups. Primers are listed in Table 1 below.

Tissue collection, histological and immunofluorescence analyses
The left lung was fixed in 4% paraformaldehyde (PFA), while the right lung was snap frozen in

MSP-qPCR
Genomic DNA was isolated using QIAamp DNA Mini Kit (Qiagen) from indicated cells, and then bisulfite converted using EpiTect Bisulfite kit (Qiagen). Following bisulfite conversion, DNA methylation status was quantified by qPCR with primers (listed below in Table 2) that specifically recognize the methylated cytosine. UBB was used as an internal control.

Promoter binding site prediction
Sterol regulatory elements (SREs) that SREBP2 putatively binds was predicted on snai1, aSMA, N-Cad, and vimentin promoter regions (-1000 to +500 from the transcriptional start sites [TSS]) using the weight matrix-based program MATCH (11). CpG islands were predicted using WashU EPI Genome Browser.
Briefly, ECs were cross-linked with 0.75% formaldehyde at room temp for 10 min and quenched with 125 mM glycine. Cells were lysed in RIPA buffer and sonicated to fragment the DNA to a size of 500-1000 bp. Cell lysates were then incubated with Dynabeads (Invitrogen) conjugated to the indicated antibodies. Immunoprecipitated DNA fragments were reverse cross-linked and DNA binding was quantified as percentage of input using qPCR (see list of primers below in Table 3).

DNMT1 Activity assay
DNMT1 activity was measured using a DNMT1 Assay kit (Abcam) following the manufacturer's instructions. Briefly, nuclear extracts of BLM treated HUVECs were isolated using a Nuclear Extraction kit (Abcam). Nuclear extracts were then incubated with capture reagents and antibodies specific to active DNMT1 using ELISA. Colorimetric intensity was used to measure DNMT1 activity.

Lung EC isolation
Harvested lungs were immediately minced and incubated with Type 1 Collagenase (Thermo Fisher Scientific) for 45 min. The solution was then triturated using 12 cm cannulas, filtered at 70 µm,