Interleukin-11 is important for vascular smooth muscle phenotypic switching and aortic inflammation, fibrosis and remodeling in mouse models

Transforming growth factor beta-1 (TGFβ1) is a major driver of vascular smooth muscle cell (VSMC) phenotypic switching, an important pathobiology in arterial disease. We performed RNA-sequencing of TGFβ1-stimulated human aortic or arterial VSMCs which revealed large and consistent upregulation of Interleukin 11 (IL11). IL11 has an unknown function in VSMCs, which highly express the IL11 receptor alpha, suggestive of an autocrine loop. In vitro, IL11 activated ERK signaling, but inhibited STAT3 activity, and caused VSMC phenotypic switching to a similar extent as TGFβ1 or angiotensin II (ANGII) stimulation. Genetic or therapeutic inhibition of IL11 signaling reduced TGFβ1- or ANGII-induced VSMC phenotypic switching, placing IL11 activity downstream of these factors. Aortas of mice with Myh11-driven IL11 expression were remodeled and had reduced contractile but increased matrix and inflammatory genes expression. In two models of arterial pressure loading, IL11 was upregulated in the aorta and neutralizing IL11 antibodies reduced remodeling along with matrix and pro-inflammatory gene expression. These data show that IL11 plays an important role in VSMC phenotype switching, vascular inflammation and aortic pathobiology.


Human VSMCs Cell Culture
Patients aged ≥21 and ≤81 years of age undergoing coronary bypass grafting at the National Heart Centre Singapore were recruited to the study. Patients with valvular heart disease or previous atrial intervention were excluded. Aortic biopsies and/or left internal mammary artery (LIMA) trimmings were obtained and used to outgrow primary aortic and arterial VSMCs respectively. Vascular biopsies were prepared by the removal of the tunica adventitia and endothelium under a dissecting microscope.
Experiments were carried out at low cell passages (≤ passage 4) and cells were growth restricted with 0.2% fetal bovine serum (FBS) in basal M231 for 24h before treatment with recombinant human IL11 (5 ng/ml), TGFβ1 (5 ng/ml), or ANGII (100nM) in serum-free M231 for 24h. Unstimulated VSMCs for the same duration under basal M231 medium was included as controls. For MEK/ERK inhibition studies, VSMCs were stimulated with IL11 in the presence of U0126 (10 µM) reconstituted in dimethyl sulfoxide (DMSO) compared to vehicle controls.

RNA-sequencing
Total RNA was isolated using Trizol Plus RNA mini kit (12183555, Life Technologies) and quantified using Qubit RNA high sensitivity assay kit (Life Technologies). RNA integrity number (RIN) was determined with the LabChip GX RNA Assay Reagent Kit (PerkinElmer). TruSeq Stranded mRNA Library Prep kit (Illumina) was used to assess transcript abundance following standard manufacturer's instructions. Briefly, poly(A)+ RNA was purified from 0.8 -1 µg of total RNA with RIN > 7, fragmented, and used for cDNA synthesis, followed by 3' adenylation, adaptor ligation, and PCR amplification. The final libraries were quantified using KAPA library quantification kits (KAPA Biosystems) on StepOnePlus Real-Time PCR system (Applied Biosystems).
The quality and average fragment size of the final libraries were determined using LabChip GX DNA High Sensitivity Reagent Kit (PerkinElmer). Libraries were pooled and sequenced on a NextSeq 500 benchtop sequencer using 75-bp paired-end sequencing chemistry. Raw sequencing data (.bcl files) were demultiplexed into individual FastQ read files with Illumina's bcl2fastq v2.16.0.10 based on unique index pairs. The adaptor sequences and low-quality reads/bases were trimmed using Trimmomatic v0.36 59 and the read quality was assessed with FastQC v0.11.5. Highquality reads were mapped to Ensembl human GRCh38 v92 reference genome using Spliced Transcripts Alignment to a Reference (STAR) v2.5.2b 60 . STAR alignment options were --outFilterType BySJout --outFilterMultimapNmax 20 --alignSJoverhangMin 8 --alignSJDBoverhangMin 1 --outFilterMismatchNmax 999 --alignIntronMin 20 --alignIntronMax 1000000 --alignMatesGapMax 1000000. Strandspecific raw counts of uniquely mapped reads (paired-end) were summarized with featureCounts to get gene-level quantification of genomic features: featureCounts -t exon -g gene_id -s 2 -p. Differential expression analyses were performed in R 3.5.2 using the Bioconductor package DESeq2 1.22.2, using the Wald test for comparisons and including the variance shrinkage step setting a significance threshold of 0.05. Genes with less than 1 count in all samples were removed to speed up the analysis. The sample ID was included in the model to account for interpatient variability and the unstimulated condition was always used as reference.
For GO over-representation analysis, differentially expressed genes (FDR < 0.05) were tested for statistically significant over-representation against the background of all expressed genes using the R package gProfileR 61 with "strong" hierarchical filtering. Gene Set Enrichment Analyses (GSEA) were run using the fgsea library 62 , pre-ranking the gene list by the "stat" column of the DESeq2 results output and using 10 5 permutations.

FANTOM5 data processing
Gene expression data in primary cell types with replicates were downloaded from the FANTOM5 web resource 22 . Since the FANTOM5 data is at the level of transcription start site expression derived from Cap Analysis of Gene Expression (CAGE) sequencing, we calculated the gene level expression by summing all counts that were assigned to a given gene. These were normalized by library size to calculate the tags per million (TPM) for each gene. To compare the expression profiles of IL11RA and IL6R, we extracted the TPM for these two genes across 511 different primary cell types encompassing cell types from all lineages. In each case, for which the expression of either IL11RA or IL6R is above the level of noise, we highlight these cell types and categorize them as described in the FANTOM5 cell type ontology.

Operetta high content immunofluorescence analysis
High-content immunofluorescence imaging and quantification of VSMC activation were performed as previously described 14 . VSMCs were seeded in 96-well CellCarrier plates (6005550, PerkinElmer) at a density of 5x10 3 cells per well and following experimental conditions, fixed and permeabilized as before. To measure cell proliferation, 5-Ethynyl-2´-deoxyuridine (EdU; 10 µM) was added 1h poststimulation and detected after 24h of stimulation using the Click-iT EdU Labelling kit (C10350, Life Technologies) according to manufacturer's protocol. Cells were blocked using 0.5% BSA and 0.1% Tween-20 in PBS before incubation with primary anti-Collagen I (1:500), anti-SM22α (1:200), and anti-myocardin (1:200) and visualized using Alexa Fluor-488-conjugated secondary antibody (1:1000). Cells were counterstained with DAPI (D1306, Invitrogen) in blocking solution. Plates were scanned and images were collected with the Operetta high-content imaging system 1483 (PerkinElmer). The measurement of collagen I fluorescence intensity per area was done using Columbus software version 2.7.1 (PerkinElmer). Fluorescence intensity (collagen I and myocardin expression) was normalized to the number of cells detected in the field and recorded for seven fields per well. EdU +ve and SM22α +ve cells were quantified using Harmony software version 3.5.2 (PerkinElmer) and the percentage of VSMCs was determined for each field.

Scratch wound migration assay
Scratch wound assays were performed on fully confluent monolayers of human and murine VSMCs (passage 3-4) cultured on 35mm culture dishes. After synchronizing in low serum media (0.2% FBS-M231) for 24h, a linear scratch was created with a sterile pipette tip followed by treatment with: either IL11 (5 ng/ml), TGFβ1 (5 ng/ml) or ANGII (100 nM) with or without either IL11 neutralizing antibody (2 μg/ml, X203; Aldevron) or IgG isotype control (2 μg/ml, IgG; Aldevron) for 24h (human VSMCs) to 48h (murine VSMCs). The wound area was imaged at 0h and 24h or 48h and quantified using ImageJ software with the MRI wound healing tool plugin (http://dev.mri.cnrs.fr/projects/imagej-macros/wiki/Wound_Healing_Tool). Migration of VSMCs was calculated as a percentage of wound area recovered at endpoint compared to 0h. A total of 6 to 10 random regions were analyzed per treatment and averages reported.

Invasion assay
The invasive capacity of VSMCs were assessed using Boyden chamber invasion assays (Cell Biolabs). VSMCs were cultured in serum-free M231 medium for 24h prior to cell invasion assays. Equal numbers of VSMCs were seeded in duplicates onto apical chambers containing Matrigel for invasion assays. VSMCs invaded towards the bottom chamber containing 2% FBS-M231 medium with the respective chemoattractant. After 24h of incubation, media was removed, and non-invasive cells were removed using cotton swabs. Invaded cells were stained with cell staining solution (Cell Biolabs) and colorimetrically quantified at 540 nm. For antibody inhibition studies, VSMCs were pre-treated with X203 or IgG control antibodies (both 2 µg/ml) for 15 min prior to addition of chemoattractant.
Passage 3 to 4 mouse aortic VSMCs were used for scratch wound migration assay and cultured for collagen secretion assay. Murine VSMCs were either treated with recombinant mouse TGFβ1 (5 ng/ml), IL11 (5 ng/ml), or ANGII (100nM) in M231 basal medium. Control littermates were injected with an equal amount of corn oil (C8267, Sigma-Aldrich) vehicle (n = 12). Mice were euthanized at 14 days from the first TAM dose.

Transverse aortic constriction (TAC)
TAC surgeries were performed on C57BL/6J male mice (n = 24) purchased from In Vivos Pte Ltd, Singapore as described 65 . Age-matched sham controls underwent the same operative procedure without ligation (n = 10). Trans-thoracic two-dimensional Doppler echocardiography confirmed increased pressure gradients (>40 mmHg) indicative of successful TAC. Mice were randomized to receive post-operative antibody treatment conducted by IP injections of either X203 or IgG control antibodies (n = 12 per group) at a dose of 20 mg/kg twice per week for two consecutive weeks starting 24h following TAC. Mice were euthanized at 2 weeks post-TAC prior to maladaptive heart failure response 65 and the proximal ascending aorta was excised.

Angiotensin II (ANGII) infusion
ANGII pump infusions were performed as previously described 14 . C57BL/6J male mice were implanted SQ with an osmotic minipump (Alzet model 1004, Durect) containing either angiotensin II (ANGII, 2 mg/kg per day; n = 28) or an identical volume of saline for controls. ANGII-treated mice were randomized to receive postoperatively injections of either X203 or IgG control antibodies at a dose of 20 mg/kg twice per week for four consecutive weeks. Mice underwent aortic echocardiography and were euthanized at 4 weeks post-ANGII and the thoracic aorta was excised. No mortality associated with ANGII infusion was observed.

Aortic Echocardiography
In vivo trans-thoracic echocardiography was conducted using Vevo 2100 with a MS400 linear array transducer (VisualSonics), 18-38 MHz by a trained echocardiographer (NGZT) blinded to genotype and treatment groups. Mice were anaesthetized with 3% isoflurane and maintained at 1% isoflurane with the body temperature maintained at 37°C on a heated platform. Chest and neck hair were removed using depilatory cream and a layer of acoustic coupling gel was applied to the thorax. Aortic root and ascending aortic diameters were assessed from B and mmode of parasternal long-axis view, using inner edge-to-inner edge according to USA and European guidelines 66 . Peak aortic flow velocity was obtained by applying pulsed-wave Doppler across the aortic valve from the aortic arch at suprasternal view. All measurements were averaged over three cardiac cycles. Aortic dimensions were referenced to body weight per animal to account for differences in body mass.

Enzyme-linked immunosorbent assay
VSMCs were seeded at 10,000 live cells per cm 2 in 35mm culture dishes and maintained on complete M231 medium as previously mentioned. Following 0.2% FBS-M231 growth restriction, VSMCs were stimulated with 1 ml of TGFꞵ (5 ng/ml) or ANGII (100nM) as previously described. The levels of IL11 in equal volumes of VSMC culture medium were quantified with the human IL11 Quantikine ELISA kit (D1100, R&D Systems) and normalized against unstimulated conditions. The levels of TIMP-1 and MMP2 in cell supernatant were quantified using the human TIMP-1 and MMP2 Quantikine ELISA kit (DTM100 and MMP200, R&D Systems) respectively. All assays were conducted as per manufacturer's instructions.

Colorimetric Assays
To detect for secreted collagen by VSMCs, cell culture supernatant was first concentrated using polyethylene glycol solution (90626, Chondrex) prior to quantification using the Sirius red collagen detection kit (9062, Chondrex) according to manufacturer's instructions.

Quantitative polymerase chain reaction (qPCR)
Total RNA was extracted from snap-frozen tissues using RNAzol RT (Sigma-Aldrich) followed by Purelink RNA mini kit (Invitrogen) purification. The cDNA was synthesized using iScript cDNA synthesis kit (Bio-Rad) as per manufacturer's instructions. Gene expression analysis was performed in duplicates using fast SYBR green (Qiagen) technology on the ViiA 7 real-time PCR system (Applied Biosystems). Expression data were normalized to Gapdh mRNA expression levels and fold-change was calculated using the 2 -ΔΔCt method. The primer sequences are available in Supplementary Table S3 online.
Total fibrosis was measured by ImageJ (v1.52a, NIH) with Color deconvolution-Masson Trichrome vector (programmed by Landini) in photomicrographs taken at 100X magnification of each section (n = 5-6 per group) and expressed as a percentage of collagen stained area over total tissue area. Media thickness was measured as the intima-media distance using the incremental distance tool at a calibrated step size of 10μm on Image-Pro Premier 9.2 (Media Cybernetics) and reported as an average of 62-142 measurements across 4 random photomicrographs (400X magnification) per section (n = 5-6 per group).
Adventitial area and smooth muscle cell (SMC) nuclei were quantified in H&E stained sections as described 67 across 4 random photomicrographs (400X magnification) per section (n = 5-6 per group). Color deconvolution-H&E vector was performed. The adventitial area was expressed as a percentage of total tissue area.
SMC nuclei were counted with the Cell Counter plugin (programmed by De Vos) and expressed as nuclei counts over media area. Elastin area and elastin breaks were quantified in VVG stained sections with color deconvolution-ROI vector selecting for elastin across 4 random photomicrographs (400X magnification) per section (n = 5-6 per group). Elastin area fraction was expressed as the elastin positive area over total tissue area. Elastin breaks were counted and expressed normalized to the media area.

Statistical Analyses
Data are presented as mean ± standard deviation (SD) or median ± IQR unless otherwise stated. Statistical analyses were conducted using GraphPad Prism