Transcriptomic analysis of pulmonary artery smooth muscle cells identifies new potential therapeutic targets for idiopathic pulmonary arterial hypertension

Abstract

Background and Purpose

Pulmonary arterial hypertension (PAH, type 1 pulmonary hypertension) has a 3‐year survival of ~50% and is in need of new, effective therapies. In PAH, remodelling of the pulmonary artery (PA) increases pulmonary vascular resistance and can result in right heart dysfunction and failure. Genetic mutations can cause PAH but it can also be idiopathic (IPAH). Enhanced contractility and proliferation of PA smooth muscle cells (PASMCs) are key contributors to the pathophysiology of PAH, but the underlying mechanisms are not well understood.

Experimental Approach

We utilized RNA‐sequencing (RNA‐seq) of IPAH and control patient‐derived PASMCs as an unbiased approach to define differentially expressed (DE) genes that may identify new biology and potential therapeutic targets.

Key Results

Analysis of DE genes for shared gene pathways revealed increases in genes involved in cell proliferation and mitosis and decreases in a variety of gene sets, including response to cytokine signalling. ADGRG6/GPR126, an adhesion G protein‐coupled receptor (GPCR), was increased in IPAH‐PASMCs compared to control‐PASMCs. Increased expression of this GPCR in control‐PASMCs decreased their proliferation; siRNA knockdown of ADGRG6/GPR126 in IPAH‐PASMCs tended to increase proliferation.

Conclusion and Implications

These data provide insights regarding the expression of current and experimental PAH drug targets, GPCRs and GPCR‐related genes as potentially new therapeutic targets in PAH‐PASMCs. Overall, the findings identify genes and pathways that may contribute to IPAH‐PASMC function and suggest that ADGRG6/GPR126 is a novel therapeutic target for IPAH.

Abbreviations

CPM

counts per million

DE

differential expression

FDR

false discovery rate

GSEA

gene set enrichment analysis

IPAH

idiopathic pulmonary arterial hypertension

PA

pulmonary artery

PAH

pulmonary arterial hypertension

PASMCs

pulmonary artery smooth muscle cells

RNA‐seq

RNA‐sequencing

RV

right ventricle

TPM

transcripts per million

What is already known

• Idiopathic pulmonary arterial hypertension (IPAH) affects multiple cell types within the lung.

• Pulmonary artery smooth muscle cells (PASMCs) in PAH are hyperproliferative and hypercontractile.

What this study adds

• New insights from an unbiased approach (RNA‐sequencing) to define gene expression in patient‐derived IPAH PASMCs.

• The GPCRome (GPCR expression profile) in IPAH PASMCs patients includes an increase in ADGRG6/GPR126.

What is the clinical significance

• Altered expression and potential contribution of numerous genes to IPAH pathobiology.

• Novel therapies for PAH suggested from this pilot study of gene expression in PASMCs.

1. INTRODUCTION

Pulmonary arterial hypertension (PAH) is a progressive and deadly disease affecting the lungs and heart. In PAH, the pulmonary artery (PA) within the lungs undergoes remodelling. The features of this remodelling include neointimal proliferation, medial hypertrophy and formation of plexiform lesions which produce increased pulmonary vascular resistance, increased afterload to the right ventricle (RV) and eventual RV dysfunction and failure. The origin of PAH can be heritable or idiopathic, with many contributing factors such as impaired signalling, metabolic dysfunction, inflammation and oxidative lipids (Bonnet et al., 2007; Giaid et al., 1993; Jasiewicz et al., 2015; Jerkic et al., 2011; Morrell et al., 2001; Richter et al., 2004; Ross et al., 2015; Zamanian et al., 2008). Despite knowledge of these mechanisms, no cure exists for PAH. Current therapy of PAH is limited to symptomatic treatment and does not treat the underlying disease. The 3‐year survival rate remains 50%–60% (Humbert et al., 2010); thus, further understanding of the underlying mechanisms and new therapeutic approaches for PAH are needed.

The pathobiology of PAH involves multiple cell types within the lung. PA endothelial cells (PAECs) demonstrate impaired signalling in PAH, including reduced NO (Giaid & Saleh, 1995) and increased production of growth factors, for example, VEGF (Papaioannou et al., 2009; Voelkel & Gomez‐Arroyo, 2014), PDGF (Perros et al., 2008) and basic FGF (Benisty et al., 2004). Together with PAECs, the PA smooth muscle cells (PASMCs) in PAH respond to local and circulating factors and mediate an imbalance in vascular tone characterized by increased proliferation and contraction (Guilluy et al., 2009). PASMCs cultured ex vivo from PAH patients are aberrant compared to cells from healthy lungs and exhibit increased K⁺ channel activity, elevated intracellular Ca²⁺ and increased Rho kinase activity, leading to increased proliferation and decreased apoptosis (Guilluy et al., 2009; Yuan et al., 1998). These defects indicate cell autonomous dysfunction in PASMCs, but the mechanisms for this dysfunction remain unclear. Thus, further studies of PAH PASMCs, using, in particular, unbiased approaches, may provide new mechanistic insights and reveal novel therapeutic targets.

GPCRs, seven transmembrane domain receptors, are the largest family of membrane receptors. Due to their selective expression in various cell types, their accessibility at the cell surface and ability to initiate signals that alter cell function, GPCRs are excellent drug targets, currently constituting ~35% of the targets for drugs approved in the United States and the European Union (Sriram & Insel, 2018). The canonical signalling from activated GPCRs involves heterotrimeric GTP (G) binding proteins and alteration in cell function via various signalling cascades. In PASMCs, activation of GPCRs that couple to the G_s protein enhances activity of AC to produce cAMP, which activates PKA and thereby can reduce contraction and proliferation (Iyinikkel & Murray, 2018). This is evident with the prostacyclin (IP) receptor, a GPCR for which several drugs have long been approved (Del Pozo, Hernandez Gonzalez, & Escribano‐Subias, 2017) and are in clinical trial (Torres et al., 2019) for PAH treatment. A key goal of this study was to determine if we might identify previously unrecognized GPCRs in PASMCs that could be potentially new therapeutic targets for idiopathic pulmonary arterial hypertension (IPAH).

RNA‐sequencing (RNA‐seq) is a powerful approach for whole‐transcriptome analysis (Stark, Grzelak, & Hadfield, 2019) and provides an unbiased means to define gene expression in healthy and diseased tissues and cells. In this study, we utilized RNA‐seq to characterize PASMCs isolated from the lungs of PAH patients and non‐diseased controls. Using this analysis, we uncovered pathways that characterize PAH PASMCs, defined the GPCR profile of PASMCs from PAH patients and controls, evaluated the expression of current and experimental drug targets for PAH and identified a novel GPCR target—ADGRG6/GPR126—in PAH PASMCs.

2. METHODS

2.1. Cell isolation and culture

All human tissue was handled anonymously and in compliance with the guidelines of the UCSD Institutional Research Board and its approval. PASMCs were isolated from the lungs of patients receiving a lung transplant for IPAH or for control cells, were isolated from rejected donors or were commercially obtained (Table 1). Secondary or greater branches of the PA were removed and cut into small pieces after being washed with sterile PBS with penicillin/streptomycin and amphotericin B. These pieces were placed into digestion media (HBSS, 0.1% Fraction V BSA, 150 U·ml⁻¹ collagenase type II) for 1 h at 37°C in a rocking water bath, then agitated by pipetting and incubated in digestion media with 80 U·ml⁻¹ elastase and 0.05% trypsin/EDTA for 2 h. The remaining tissue was agitated, and the solution was filtered through a 50‐μm strainer and centrifuged at 600 g for 5 min. The pellet was resuspended, and cells were cultured in smooth muscle cell culture media (SCM): DMEM with 1 g·L⁻¹ glucose, 5% FBS, 5 ng·ml⁻¹ FGF, 5 μg·ml⁻¹ insulin, 50 μg·ml⁻¹ ascorbic acid and 5 ng·ml⁻¹ EGF. Cells were maintained in SCM until passage 10–15, at which time they had altered morphology.

TABLE 1.

Information for donors of cells used in this study

Identifier	Disease state	Patient sex	Patient age	Source
PA1217	Control	Male	45	Cell Applications, Inc.
PA1189	Control	Female	17	Cell Applications, Inc.
CHPT1	Control	Male	27	Isolated from donor lungs
CHLONZA	Control	Male	51	Lonza
IPAH4	IPAH	Female	56	Isolated from donor lungs
IPAH1028	IPAH	Female	32	Isolated from donor lungs
IPAH908	IPAH	Male	41	Isolated from donor lungs
IPAHCC014	IPAH	Male	45	Isolated from donor lungs

2.2. RNA isolation and RNA‐seq

RNA was isolated from low passage (2–4) cells using an RNeasy kit with on‐column DNA digestion. RNA‐seq was performed by the UCSD Institute for Genomics Medicine. RNA quality was assessed using the Agilent TapeStation; all samples had an RNA integrity number > 6. RNA‐seq was performed on libraries prepared via the TruSeq stranded mRNA protocol for polyadenylated RNA and sequenced on an Illumina HiSeq 4000 at, on average, 30 million 50 base‐pair single reads.

2.3. RNA‐seq analyses

Data have been deposited at NCBI's Gene Expression Omnibus (GEO) with accession number GSE144274. FASTQ files from the HiSeq instrument were checked for quality using Galaxy (Afgan et al., 2016), prior to analysis using Kallisto with the Ensembl reference transcriptome (Bray, Pimentel, Melsted, & Pachter, 2016). These data were converted to gene‐level expression using Tximport (Soneson, Love, & Robinson, 2015) as transcripts per million (TPM) and in some cases, as counts per million (CPM). The latter was used to assess differential expression (DE) using edgeR (Robinson, McCarthy, & Smyth, 2010). We used a false discovery rate (FDR) <0.05 to define genes whose expression was significantly altered.

Gene set enrichment analysis (GSEA) was performed on DE genes to test for enrichment of specific ontologies (Subramanian et al., 2005). Genes were ranked with the rank calculated as the (log₂ fold change) * (−log₁₀ FDR). The fold change was calculated for IPAH PASMCs compared to control PASMCs, such that positive fold changes indicate higher expression in IPAH cells. The analyses included Gene Ontology (GO) biological processes and Curated Canonical pathways. We also used GSEA results to perform leading edge analysis to identify specific sets of highly enriched or depleted genes/pathways, in order to associate these genes with an up‐regulation/down‐regulation in IPAH. Defining the most strongly and significantly enriched pathways allowed us to identify nodal genes in the set of up‐regulated pathways. We identified additional biological processes and transcription factors using Enrichr (Kuleshov et al., 2016) with genes altered >4‐fold with FDR <0.05.

2.4. Quantitative real‐time PCR

Quantitative real‐time PCR (qPCR) was performed on cDNA prepared from patient RNA using reverse transcriptase as part of the SuperScript II master mix kit. qPCR was ran using iTaq SYBR Green according to the manufacturer's protocol: 30 s for polymerase activation and DNA denaturation at 95°C, followed by 40 amplification cycles of 5 s denaturation at 95°C and 30 s annealing at 60°C, with plate reading at each cycle. The primers used are shown in Table 2.

TABLE 2.

Primer sequences used for qPCR in this study

Primer		Sequence
18s	Forward	GTAACCCGTTGAACCCCATT
	Reverse	CCATCCAATCGGTAGTAGCG
ADGRG6	Forward	AGCGCTTCCTAGAAAGGATGT
	Reverse	AGAGGCGACCTCATCTCTGT
ANLN	Forward	GGCCAATCTTGCTGCAACTA
	Reverse	CCAAAGACGCATCATCAGCA
ARHGAP11A	Forward	CATGCCATTTTCCTGCCTCA
	Reverse	CATGCCTGTAATCCCAGCAC
CCNB1	Forward	TGAGGAAGAGCAAGCAGTCA
	Reverse	GCATCTTCTTGGGCACACAA
CXCL2	Forward	CGCCCAAACCGAAGTCATAG
	Reverse	CTCTGCAGCTGTGTCTCTCT
ICAM1	Forward	AGAGACCCCGTTGCCTAAAA
	Reverse	CAGTACACGGTGAGGAAGGT
STMN1	Forward	TCTCCCCATCCCCTTCCTAA
	Reverse	GTTTCTCCCCTTTAGCCCCT
TRIP13	Forward	AGCAACCTCATCACCTGGAA
	Reverse	ACACCAGGGCGTCTTTATCA
UHRF	Forward	CCAAGGTCAGAGCAAGCATC
	Reverse	GGACTTTTCACGCGGATCTC

2.5. Cell proliferation, transfection and cAMP measurement

Cell proliferation was measured using the CyQuant assay, which quantifies the cell number using fluorescent staining of nuclear material. Cells were plated overnight on a 96‐well plate, and a subset was quantified for baseline (Day 0) fluorescence. Control PASMCs from at least four different cultures were transfected with a plasmid that had a CMV promoter for ADGRG6, or transfection reagent as control (mock). IPAH PASMCs from all four patients were transfected with siRNA for ADGRG6, or control siRNAs used as control. CyQuant was used to obtain fluorescent values every 48 h in at least three technical replicates per patient cell subset; proliferation was calculated at each time point by dividing the values by those obtained at Day 0.

cAMP was measured in transfected cells after 48 h (DiscoveRx). Cells were switched into serum‐free media for 1 h before the assay; some wells were treated with IBMX (50 μM) for 20 min. Cells were then incubated with agonists with or without forskolin (10 μΜ) for 20 min. Following termination of the incubation, cAMP was assayed according to the manufacturer's instructions. Mean luminescence was calculated from wells with technical triplicates; cAMP concentration was calculated using a standard curve.

2.6. Data and statistical analyses

All experiments were completed with at least four patients from each group, as there was a limitation and difficulty in obtaining the patient‐derived cell cultures. The data and statistical analysis comply with the recommendations of the British Journal of Pharmacology on experimental design and analysis in pharmacology (Curtis et al., 2018), wherein the smaller n number used here is deemed exploratory. Statistics were analysed using Prism (GraphPad). Differential gene expression was considered significant when FDR was <0.01. qPCR data used for gene expression validation were analysed with simple linear regression, and when comparative genes are shown, using two‐tailed Student's t test. qPCR data were collected in triplicate of four patients per group.

2.7. Materials

Enzymes (collagenase II and elastase) were obtained from Worthington Biochemical Co. (Lakewood, NJ). Commercially obtained PASMCs were from Cell Applications, Inc. (San Diego, CA; Cat# 352‐05a) and Lonza (Basel, Switzerland; Cat# CC‐2581). RNA isolation kits were from Qiagen (Hilden, Germany). CyQuant assay for proliferation, SuperScript II master mix, penicillin/streptomycin, HBSS with calcium and magnesium and 0.05% trypsin/EDTA were obtained from Thermo Fisher Scientific (Waltham, MA). ADGRG6 plasmid was from AdGene (Cambridge, MA). siRNA and transfection reagent, DharmaFECT, were obtained from GE Dharmacon (Lafayette, CO). iTaq SYBR green was obtained from Bio‐Rad (Hercules, CA). TruSeq stranded mRNA kit was obtained from Illumina (San Diego, CA). Forskolin (FSK), lysophosphatidic acid and collagen IV were obtained from Abcam (Cambridge, UK). Insulin was obtained from Santa Cruz Biotechnology (Dallas, TX). Amphotericin B was obtained from Gemini Bio‐Products (West Sacramento, CA), and Fraction V BSA was obtained from Grainger (Lake Forest, IL). DMEM was obtained from Millipore Sigma (Burlington, MA). Ascorbic acid, bradykinin and IBMX were obtained from Sigma‐Aldrich (St. Louis, MO). FGF was obtained from Miltenyi Biotec (Bergisch Gladbach, Germany), and EGF was obtained from Promega (Madison, WI). Iloprost was obtained from Tocris (Bristol, UK). FBS was obtained from Omega Scientific (Tarzana, CA).

2.8. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Harding et al., 2018), and are permanently archived in the Concise Guide to PHARMACOLOGY 2019/20 (Alexander, Christopoulos et al., 2019; Alexander, Fabbro et al., 2019a, b; Alexander, Mathie et al., 2019).

3. RESULTS

3.1. Comparison of gene expression profiles between control and IPAH PASMCs

To determine the DE of genes in IPAH PASMCs, we conducted RNA‐seq on cultured PASMCs isolated from control and PAH patients. Visualization of the RNA‐seq data via a heat map shows an overall consistency among replicates (Figure 1a). Analysis of clustering of samples via principal component analysis (PCA) for all detected genes in the PASMC samples, normalized in CPM, reveals distinct clusters for control PASMCs and IPAH PASMCs, thus implying large differences in gene expression between the two groups (Figure S1). DE analysis revealed that 858 genes were significantly (FDR <0.05) up‐regulated at least twofold; expression of 499 genes was up‐regulated at least fourfold (Table S1). DE analysis identified 975 genes that were down‐regulated at least twofold in PAH PASMCs; 461 of these genes were down‐regulated at least fourfold (Figure 1b,c). Expression of several genes was validated with qPCR (Figure 1d). Their expression was significantly correlated with values obtained from the RNA‐seq analysis (R ² = 0.9027). The PAH PASMCs had no change of expression of genes with known mutations in heritable PAH (log₂ fold change, FDR: BMPR2, −0.62, −0.43; EIF2AK4, 0.48, 0.36) except an increase in TBX4 (8.00, <0.0001), which has not been previously observed. The samples that we assessed confirmed the altered expression of several other genes implicated in PAH pathobiology: caveolin‐1 (CAV1, increased 4.49‐fold in IPAH, FDR = 0.002); tenascin C (TNC, increased 5.09‐fold in IPAH, FDR = 0.031); pyruvate dehydrogenase kinase 4 (PDK4, increased 7.30‐fold, FDR = .024) and PG‐endoperoxide synthase 2(PTGS2, decreased 0.01‐fold,).

FIGURE 1.

RNA‐seq of IPAH PASMCs reveals differentially expressed genes compared to control PASMCs. RNA‐seq conducted on patient‐derived PASMCs was analysed for differential expression (DE). (a) Heat map of all DE genes in control (n = 4) and IPAH (n = 4) PASMCs shows a consistency in gene expression among replicates. (b) Smear plot of all genes, plotting the log₂ fold change in IPAH PASMCs compared to control PASMCs against the average log₂ counts per million (CPM) amongst all samples; significantly altered genes are indicated in red (FDR <0.05). (c) Number of DE genes increased or decreased in IPAH PASMCs compared to control PASMCs. (d) qPCR confirmation of selected genes; data analysed by simple linear regression

3.2. Pathway analyses reveal increased gene groups related to proliferation

We used GSEA for Biological Processes to assess all significantly (FDR <0.001) up‐regulated and down‐regulated groups of genes from genes ranked according to their fold change and expression level (Figure S2A,B). The five groups of genes with the highest normalized enrichment score are shown in Figure 2a. Nearly all of the significantly up‐regulated gene groups involve cell proliferation, a result confirmed by the GSEA for Cellular Compartments (Figure S3) and Curated Canonical pathways (Figure S4), which indicated the genes were related to the nucleus and cell cycle. To determine the genes responsible for these changes, we utilized the GSEA leading edge analysis on GO Biological Processes (Figures 2b and S5). We found that the most prominent genes in this list are primarily protein kinases associated with proliferation such as the genes encoding cyclin B1 CCNB1, cyclin‐dependent kinase 1 CDK1 and aurora kinases AURKA and AURKB. Analysis of these leading edge genes (using Enrichr) revealed significant association (P‐adjusted <0.05) with several transcription factors (Figure 2c), most notably forkhead box protein M1 (FOXM1) and E2F Transcription Factor 4 (E2F4). Similar transcription factors were identified by the same analysis in Enrichr for up‐regulated genes in IPAH (FDR <0.05, >4‐fold) (Figure S6). GO analysis for molecular function using Enrichr (Figures 2d and S7) revealed up‐regulated gene groups involved in mitosis‐related processes and down‐regulated gene groups that are involved in signalling.

FIGURE 2.

Gene Ontology (GO) analyses reveal novel pathways in IPAH PASMCs. GO analyses used weighted differential expression (DE) from RNA‐seq of IPAH and control PASMCs. (a) Gene set enrichment analysis (GSEA) for GO Biological Processes increased or decreased in IPAH versus control PASMCs, using genes pre‐ranked by their fold change and significance of change, as described in Section 2. (b) Leading edge analysis for the GO Biological Processes associated with up‐regulated genes in IPAH PASMCs, enriched with high significance (FDR <0.001). There were 57 such GO processes, which share a subset of overlapping genes, as shown in this Figure, and in more detail in Figure S4. (c) The 20 highest expressed genes from the leading edge analysis in panel (b) were analysed via Enrichr to identify associated transcription factors, based on associations with transcription factors annotated by ENCODE and ChEA. (d) Enrichment analysis via Enrichr for GO molecular function, for genes either up‐regulated >4‐fold or down‐regulated >4‐fold, in IPAH PASMCs

3.3. Alterations in genes targeted by PAH drugs

Current therapy for PAH involves the activation or inhibition of several proteins, but the expression of genes that encode such proteins in IPAH PASMCs has not been defined. We examined the DE of genes that are current or proposed targets of drugs for the treatment of PAH (Table 3). The main target of prostacyclin analogues, the IP receptor, was significantly down‐regulated (~3.5‐fold) in IPAH PASMCs compared to control PASMCs. The PGE₂ receptor, EP₂, which also binds the drug iloprost (pK_i = 5.7) (Abramovitz et al., 2000), was increased in IPAH PASMCs ~25‐fold. Endothelin receptors and GC activating proteins were not detected in PASMCs. The target of PDE inhibitors, (such as sildenafil) PDE5A, is significantly increased in IPAH PASMCs compared to control PASMCs. The vasoactive intestinal peptide (VIP) receptor, VIPR1, was significantly decreased in IPAH PASMCs; VIPR2 was not detected (Haydar, Sarti, & Grisoni, 2007). An analogue of VIP is currently in clinical trial for PAH (NCT03795428). The PARP1 was increased ~2‐fold in IPAH PASMCs compared to controls; a PARP inhibitor is in clinical trial (NCT03251872). Several other targets being investigated for possible therapy were either not detected or unchanged in IPAH PASMCs, including oestrogen receptors (Lahm et al., 2012; Umar et al., 2011), PDGF receptors α and β (Antoniu, 2012) and mitochondrial 3‐ketoacyl‐CoA thiolase (Sutendra et al., 2010).

TABLE 3.

Differential expression of genes of current drug targets in PAH PASMCs compared to control

Related drug class	Name	Gene ID	Altered in IPAH PASMCs	Log2 FC	FDR
Prostacyclin analogues	PGI2 (prostacyclin) receptor (IP)	PTGIR		−1.8245927	0.06777987
	PG EP4 receptor	PTGER4	No change	0.60060576	0.76825327
	PG EP1 receptor	PTGER1	Not detected	N/A	N/A
	PG EP2 receptor	PTGER2		4.66151166	2.2558E‐05
	PGE2 receptor EP3	PTGER3	Not detected	N/A	N/A
Endothelin antagonists	Endothelin ETA receptor	EDNRA	No change	0.31301671	0.87020708
	Endothelin ETB receptor	EDNRB	No change	0.42309367	0.92492456
GC activators	GC‐activating proteins	GUCA1A, 1B, 1C, 2A, 2B	Not detected	N/A	N/A
PDE inhibitors	PDE5A	PDE5A		2.78016438	1.159E‐05
	PDE6C	PDE6C	Not detected	N/A	N/A
Others being investigated	Oestrogen receptor α	ESR1	No change	0.16208653	0.95001758
	Oestrogen receptor β	ESR2	Not detected	N/A	N/A
	VIP receptor 1	VIPR1		−7.9644464	0.00167047
	VIP receptor 2	VIPR2	Not detected	N/A	N/A
	PARP1	PARP1		1.01711747	0.01589275
	PDGF receptor α	PDGFRA	No change	0.43850118	0.57076391
	PDGF receptor β	PDGFRB	No change	0.47489287	0.62882691
	3‐Ketoacyl‐CoA thiolase, mitochondrial	ACAA2	No change	0.39661902	0.48504612

Abbreviations: FC, fold change; FDR, false discovery rate.

3.4. Expression of GPCRs and related genes in control and PAH PASMCs

As noted in Section 1, GPCRs are important therapeutic targets, including in PAH. Moreover, many of the processes found in the pathway analyses can be modulated by GPCR activation, and several pathways altered in IPAH cells include signalling components. We thus focused on the expression of GPCRs, in particular GPCRs with altered expression that potentially might be exploited therapeutically to alter PASMC function. We found that PASMCs highly express many GPCRs. Figure 3a shows the 10 highest expressed GPCRs in control and IPAH PASMCs with the data expressed in TPM.

FIGURE 3.

PASMCs of IPAH and control patients highly express many GPCRs. (a) Magnitude of expression in transcripts per million (TPM) of the 30 most highly expressed GPCRs in IPAH samples compared to the TPM from control PASMCs. n = 4 per group. (b) All significantly (FDR <0.05) differentially expressed GPCRs in IPAH PASMCs compared to control PASMCs, expressed as log₂ fold change. (c) G protein linkage analyses for GPCRs expressed in control and IPAH PASMCs, based on linkages from the International Union of Basic and Clinical Pharmacology (IUPHAR)/British Pharmacological Society (BPS) Guide to Pharmacology. (d) G protein linkages of GPCRs, weighted on the basis of expression level

Figure 3b shows the log₂ fold change of the 28 GPCRs with significantly altered expression (FDR <0.05). IPAH PASMCs expressed (at >0.2 TPM, the minimum value for detection) a significantly greater number of GPCRs (128 ± 8.4 vs. 141 ± 5.5 in control PASMCs, mean ± SD,). We examined the expressed GPCRs for their linkages to G proteins based on the International Union of Basic and Clinical Pharmacology (IUPHAR)/British Pharmacological Society (BPS) Guide to Pharmacology (www.guidetopharmacology.org) (Figure 3c) and found a similar overall distribution but slightly more receptors linked to G_i in the IPAH PASMCs (N = 48) compared to control PASMCs (N = 44). We weighted these linkages by dividing the sum of the expression level of each gene (in CPM) for each G protein by the total sum of all receptors expressed, and multiplied by 100. This allowed us to determine the overall G protein ‘tone’ of the cells and found similar results for IPAH and control PASMCs (Figure 3d).

We also examined the expression of GPCR‐related genes (Table 4). PDE5A was significantly increased while PDE3A and PDE10A were significantly decreased in IPAH PASMCs. Other PDEs were not changed. We found no changes in gene expression of ACs, G proteins or EPACs but identified a significant decrease in the cyclic nucleotide gated channel (CNGA)‐1; other CNGA genes were not detected in the IPAH PASMCs.

TABLE 4.

Expression of GPCR‐related genes in PAH PASMCs compared to control PASMCs

Type	Gene ID	Log2 FC	FDR
PDEs	PDE1A	1.36768722	0.19279862
	PDE1C	−1.0351015	0.41270203
	PDE2A	−0.8791186	0.72677171
	PDE3A	−3.9218044	0.01773621
	PDE3B	1.12541127	0.53247181
	PDE4A	0.14797201	0.91252035
	PDE4B	0.79063259	0.62235072
	PDE4C	−1.2040347	0.4383037
	PDE4D	−1.218615	0.31239144
	PDE4DIP	−0.0648219	0.95377355
	PDE5A	2.78016438	1.159E‐05
	PDE6D	0.10104132	0.90124954
	PDE7A	1.20759294	0.11462775
	PDE7B	0.93001383	0.34754344
	PDE8A	−0.4491187	0.49376121
	PDE10A	−2.5778578	0.01905484
	PDE12	0.50804432	0.45897269
ACs	ADCY3	0.03857273	0.96470326
	ADCY4	−0.0778801	0.9855517
	ADCY6	−0.9227948	0.05596457
	ADCY7	−0.074738	0.94721759
	ADCY8	3.22638276	0.18972153
	ADCY9	−0.026064	0.98360466
	ADCY10P1	−0.2087408	0.92753358
G proteins	GNA11	0.11956264	0.9037997
	GNA12	−0.1233549	0.86826881
	GNA13	0.40890661	0.46426318
	GNA14	−2.3116221	0.25030075
	GNAI1	−0.3980523	0.57673829
	GNAI2	0.29969138	0.65934871
	GNAI3	0.46488264	0.51862356
	GNAQ	−0.2755371	0.71330998
	GNAS	−0.5746913	0.25372754
	GNB1	0.10542972	0.89762282
	GNB1L	−0.1147831	0.92268944
	GNB2	0.10638378	0.89878217
	GNB2L1	−0.5861035	0.27595366
EPacs	RAPGEF3	0.44243239	0.82304576
	RAPGEF4	0.37854773	0.8657386
Cyclic nucleotide gated channels	CNGA1	−4.4279039	0.04825486
	CNGA2, A3, A4, B1, B2	Not expressed
Popeye domain‐containing	POPDC2	−0.8930059	0.56070418
PKA	PRKACA	0.41858392	0.46615429
	PRKACB	0.15846906	0.84402812
	PRKACG	Not expressed
	PRKAR1A	−0.0271917	0.98104783
	PRKAR1B	0.78116611	0.30496583
	PRKAR2A	0.84870145	0.16803799
	PRKAR2B	1.97595472	0.11540073
GPCR kinases	GRK1, GRK7	Not expressed
	ADRBK1	0.23539162	0.72630081
	ADRBK2	2.06542303	0.09482067
	GRK4	−0.1231783	0.92569456
	GRK5	−0.5410501	0.54255102
	GRK6	0.41108554	0.52215893

Abbreviations: FC, fold change; FDR, false discovery rate.

3.5. ADGRG6/GPR126 is increased in IPAH PASMCs and may be a therapeutic target for PAH

The expression of the adhesion GPCR, ADGRG6 (GPR126), was significantly increased in IPAH PASMCs compared to control PASMCs, a finding that was consistent among replicates (Figure 3b, increased 16.0‐fold, FDR <0.001) and that we confirmed by qPCR (Figure 4a). Based on previous work showing that this receptor acted via G_s to activate AC (Paavola, Sidik, Zuchero, Eckart, & Talbot, 2014; Petersen et al., 2015), we assessed whether ADGRG6 reduced the proliferation in IPAH PASMCs. To test this idea, we overexpressed ADGRG6 in control PASMCs and assayed their proliferation. Transfection of the receptor into control PASMCs reduced their proliferation compared to the mock transfection (Figure 4b). Moreover, siRNA targeting of ADGRG6 in IPAH PASMCs tended to increase proliferation, compared with a scrambled siRNA (Figure 4c). We then tested the cAMP response of PASMCs to several agonists in the absence and presence of FSK (10 μM) and/or IBMX (50 μΜ, a PDE inhibitor): iloprost (100 nM), lysophosphatidic acid (LPA, 1 μΜ), collagen IV (Col IV, 1 μg·ml⁻¹, a proposed agonist of ADGRG6/GPR126 (Paavola et al., 2014), bradykinin (10 nM) or mianserin (10 mM). We found that FSK (alone or together with IBMX) raised cAMP levels in all treatment groups (Figure 4d). Among the GPCR agonists, only iloprost increased cAMP. Consistent with the decreased expression of the IP receptor in IPAH PASMCs (Table 3), cAMP levels were reduced in IPAH PASMCs treated with iloprost, but the cAMP response was restored with ADGRG6‐siRNA transfection

FIGURE 4.

The function of ADGRG6/GPR126 in PASMCs. (a) qPCR for expression of ADGRG6 in control and IPAH PASMCs using 18s as a housekeeping gene (lower ΔCt indicates higher expression). Data shown are means ± SD; n = 4. (b) Proliferation of control PASMCs transfected with plasmid for ADGRG6 or a mock experiment. Data shown are means ± SD; n = 4 separate patients. (c) Proliferation of IPAH PASMCs transfected with siRNA for ADGRG6 or a scrambled siRNA. Data shown are means ± SD; n = 4 separate patients. (d) cAMP in control PASMCs, un‐transfected and after transfection with ADGRG6 plasmid, and in IPAH PASMCs, un‐transfected and after transfection with siRNA for ADGRG6. Cells were treated in the absence or presence of forskolin (FSK, 10 μΜ) and/or IBMX (50 μM) and the indicated agonists. Col IV, Collagen IV; LPA, lysophosphatidic acid

4. DISCUSSION

Currently, the only cure for PAH is lung transplantation. Drug therapy has improved symptoms and signs of PAH, but mortality remains high. The lack of improvement in survival by current therapies is at least in part attributable to limitations in understanding of the mechanisms that underlie the disease. ‘Omics’‐based analyses provide unbiased approaches and may identify previously unknown pathophysiological mechanisms that might be targeted therapeutically for PAH. Until now, few studies have characterized the transcriptome or proteome of PAH patients, and most of these studies utilize tissue or cell types in the lung other than PASMCs (Fraidenburg & Machado, 2018; Gaskill et al., 2016; Hoffmann, Wilhelm, Olschewski, & Kwapiszewska, 2016). Here, we provide what we believe is the first whole‐transcriptome analysis of low passage PASMCs from IPAH patients and controls. This dataset should be useful for comparison/validation in animal studies and aid in further investigations of pathophysiology and potential therapeutics in IPAH.

Previous studies have identified genes with altered expression in PAH. An alteration in bone morphogenetic protein receptor type 2 (BMPR2) has been mechanistically characterized (Lane et al., 2000; Machado et al., 2001; Sztrymf et al., 2008). The IPAH PASMCs that we assessed lacked changes in BMPR2 expression. Expression of several other genes implicated in PAH pathobiology: caveolin‐1, tenascin C, pyruvate dehydrogenase kinase 4 and PG‐endoperoxide synthase were altered, as previously observed (Austin et al., 2012; Gaskill et al., 2016b). Study of CAV1 revealed that IPAH PASMCs are dysfunctional, that is, increased capacitive calcium entry and proliferation, effects ameliorated by disruption of caveolin by cholesterol‐depleting agents and CAV1‐directed siRNA (Patel et al., 2007). It will be of interest to determine if other genes found in the current analysis contribute to cellular functions in IPAH.

GO analyses of significantly increased genes in IPAH PASMCs showed that most of the enriched gene groups are related to cell cycle/cell division, for example, mitotic nuclear division, cell division and chromosome segregation. This result is consistent with evidence that IPAH PASMCs are hyperproliferative, as observed in cells isolated from PAH patients and in pathology from patient lungs in situ (Humbert et al., 2004; Savai et al., 2014). Of note, it has been proposed that PAH may be akin to a cancer in the pulmonary vasculature (Boucherat et al., 2017). Down‐regulated gene groups will also be of interest for study related to PAH, including cytokine‐mediated signalling, and response to topologically incorrect protein inflammation has been implicated as a driver and/or consequence of PAH (Huertas, Tu, Humbert, & Guignabert, 2019; Rabinovitch, Guignabert, Humbert, & Nicolls, 2014), apparently resulting from an increased expression of multiple cytokines (Dorfmüller et al., 2002; Hashimoto‐Kataoka et al., 2015; Olsson et al., 2016). Our data indicate that IPAH PASMCs have a global decrease in cytokine expression, thus implying that PASMCs have a limited (or perhaps a compensatory) role in the inflammation observed in PAH patients.

Examination of gene expression of drug targets that are approved or being investigated revealed changes in certain current drug targets in PAH PASMCs. Based on their increase in expression, EP₂ receptors (PTGER2) and PDE5A (and drugs interacting with them) are likely to regulate PASMC activity in PAH. Interestingly, although the EP₂ receptor is increased, the IP receptor (PTGIR)is decreased in PAH PASMCs, perhaps suggesting a compensatory response in PG/prostacyclin signalling. We found that IPAH PASMCs preferentially express other genes of therapeutic significance. One such up‐regulated gene that may contribute to the increased proliferation in IPAH PASMCs is that for the cyclin‐dependent kinase CDK2 (Figure 2b), which has been recently studied using the CDK inhibitor palbociclib to inhibit the CDK2 pathway. Inhibition of CDK2 reduced proliferation of IPAH PASMCs and PH physiology in two rat models (Weiss et al., 2019). FOXM1 has been implicated in vascular remodelling in PAH (Bourgeois et al., 2018; Dai et al., 2018). Thus, certain genes, their gene products (proteins) and downstream pathways identified here warrant further study as potentially novel PASMC targets for therapy of PAH.

The druggable nature of GPCRs is apparent in their abundant use as therapeutics, including in PAH, and was a key reason why we focused the current study on GPCRs that might contribute to IPAH pathophysiology. Current mainstays of treatment include endothelin receptor antagonists, prostacyclin analogues and IP‐selective agonists (Iyinikkel & Murray, 2018; Lau, Giannoulatou, Celermajer, & Humbert, 2017). Thus, we utilized RNA‐seq to search for other GPCRs currently unknown (‘unknown unknowns’) that might be involved in PAH. We discovered that ADGRG6 (GPR126), an adhesion GPCR, is expressed at very low levels in control PASMCs but relatively high levels in IPAH PASMCs. Activation of this Gs‐coupled GPCR (Paavola et al., 2014; Petersen et al., 2015) is thus predicted to be beneficial for PAH PASMCs. We found that ADGRG6 decreased cell proliferation when transfected in control PASMCs and when knocked‐down, tended to increase IPAH PASMC proliferation. The latter findings suggest that the increase in ADGRG6 in IPAH PASMCs, compared with control PASMCs may be a compensatory mechanism that blunts further proliferation of PAH PASMCs. We found small changes in cAMP in the gain‐ and loss‐of‐function experiments and that treatment with the putative ADGRG6 agonist Collagen IV (Paavola et al., 2014) did not change cellular cAMP levels. However, activation of this GPCR may require mechanical activation, as it is expressed in cell types that undergo mechanical stress (Musa et al., 2019) and activated mechanically by laminin‐211 (Petersen et al., 2015). Further work examining the relationship of ADGRG6 and other factors involved in cAMP signalling and proliferation in PASMCs, and perhaps other smooth muscle cells, is needed.

GPCR expression data can predict functional receptors. We identified many GPCRs expressed in both control and IPAH PASMCs with values greater than 1 TPM, which we have found is a minimum value for expression of functional GPCRs, in particular Gq‐coupled GPCRs, even though GPCR transcripts can be accurately detected at 0.2 TPM (Sriram et al., 2019). Further studies are needed to define the relationship between GPCR expression and the activation of G_s/G_i/G_12/13‐coupled GPCRs. DE of GPCRs can identify receptors that may contribute to pathophysiology and may be potential therapeutic targets, but such receptors must be assessed functionally, perhaps even if their expression is unchanged or decreased in PASMCs from PAH patients (Insel et al., 2019). For example, our analysis reveals that PAR1 (F2R) is very highly expressed in PASMCs. PAR1 may be a therapeutic target whose expression/activity could be decreased to yield beneficial effects in PAH, since activation of PAR1 in porcine PA induces vasoconstriction (Maki, Hirano, Hoka, Kanaide, & Hirano, 2010). Another example is GPRC5B, which has been implicated as a de‐differentiation marker in SMCs (Kaur et al., 2017). Although a decrease in expression occurs in IPAH PASMCs, GPRC5B has a TPM >4 in all samples and thus might be a therapeutic target.

We believe that this study provides a useful dataset for PASMCs isolated from IPAH and control patients. We studied only a small number of patients (n = 4 in each group), but this is somewhat unavoidable due to the rarity of IPAH and difficulty of obtaining samples, as is evident in other studies of IPAH. Even so, using primary cultures of patient‐derived PASMCs, we found several robust, consistent changes in gene expression. We hope that findings in this pilot study will encourage the assessment of additional samples. Our results highlight the utility of RNA‐seq analysis for discovery of genes/pathways and possible targets for therapy, for example, ADGRG6 shown here. Further, our data can help support findings and the potential translatability of results from studies of experimental PAH that identify possible therapeutic targets for PAH in patients.

CONFLICT OF INTEREST

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

M.W.G., P.A.I. and K.S. designed and conceived the study. M.W.G., K.S. and A.M. produced data, analysis, interpretation and figures. All authors contributed to the scientific design and methodology and reviewed the manuscript.

DECLARATION OF TRANSPARENCY AND SCIENTIFIC RIGOUR

This Declaration acknowledges that this paper adheres to the principles for transparent reporting and scientific rigour of preclinical research as stated in the BJP guidelines for Design and Analysis, and as recommended by funding agencies, publishers and other organizations engaged with supporting research.

Supporting information

Figure S1: Principal component analysis (PCA) generated via R (using the ‘prcomp’ command) using gene expression normalized in Log₂ CPM using differential expression from all genes. Sample identifiers are described in Table 1

Figure S2: Normalized enrichment score for GO Biological Processes, the most highly enriched gene sets associated with (A) up‐regulation in IPAH‐PASMCs compared to control‐PASMCs and (B) down‐regulation in IPAH‐PASMCs compared to control‐PASMCs

Figure S3: Normalized enrichment score from the Gene set enrichment analysis (GSEA) for Cellular Compartments the most highly enriched gene sets associated with up‐regulation in IPAHPASMCs compared to control‐ PASMCs

Figure S4: Normalized enrichment score from the gene set enrichment analysis (GSEA) for Curated Pathways the most highly enriched gene sets associated with up‐regulation in IPAHPASMCs compared to control‐ PASMCs

Figure S5: Gene set enrichment analysis (GSEA) leading edge analysis for on GO Biological Processes from genes up‐regulated in IPAH‐PASMCs compared to control‐ PASMCs

Figure S6: Combined Enrichr score for transcription factors from genes up‐regulated in IPAHPASMCs compared to control‐PASMCs

Figure S7: GO analysis for molecular function using Enrichr for (A) up‐regulated and (B) downregulated genes in IPAH‐PASMCs compared to control‐PASMCs

Table S1: Differential gene expression of all genes with false discovery rate (FDR) <0.05 in Control‐ vs. PAH‐PASMCs. FC: fold‐change

Gorr MW, Sriram K, Muthusamy A, Insel PA. Transcriptomic analysis of pulmonary artery smooth muscle cells identifies new potential therapeutic targets for idiopathic pulmonary arterial hypertension. Br J Pharmacol. 2020;177:3505–3518. 10.1111/bph.15074