The characteristic signalling profiles produced by serelaxin in human vascular cells

Mohsin Sarwar1, Chrishan Samuel2, Ross Bathgate3, Dennis Stewart4, Roger Summers1. 1Monash University, Parkville, Vic 3052, Australia, 2Monash University, Clayton, Vic 3800, Australia, 3University of Melbourne, Parkville, Vic 3052, Australia, 4Corthera Inc, San Mateo CA94402, USA

In a recent phase III clinical trial (RELAX-AHF), infusion of serelaxin, the recombinant form of H2 relaxin, over 48 hr improved short- and long-term clinical outcomes in acute heart failure (1). However, the mechanisms underlying the cardioprotective actions are poorly understood. This study examined the effects of serelaxin administration in human primary vascular cells. qPCR showed expression of RXFP1 mRNA in all cell types but radioligand binding showed cell surface RXFP1 expression in human umbilical vein endothelial (HUVECs: pEC50 8.9±0.5) and smooth muscle cells (HUVSMCs: pEC50 8.8±0.4), human umbilical artery smooth muscle cells (HUASMCs: pEC50 9.3±0.3) and human cardiac fibroblasts (HCFs: pEC50 9.1±0.4), but not human umbilical artery endothelial cells (HUAECs). In venous cells, serelaxin concentration-dependently increased cAMP accumulation in HUVEC (pEC50 9.1±0.4, n=6) and HUVSMC (pEC50 9.6±0.4, n=6) to about 5% of forskolin and also increased cGMP accumulation in HUVEC (pEC50 9.1±0.3, n=7) and HUVSMC (pEC50 9.2±0.4, n=5) to 75% and 30% of DEA-NO, respectively, producing bell-shaped concentration-response curves for both pathways. Serelaxin also increased ERK1/2 phosphorylation (pERK1/2) to 15% of FBS in HUVEC (pEC50 9.2±0.3, n=5) and 20% of FBS in HUVSMC (pEC50 9.1±0.4, n=6). Similar bell-shaped CRCs for cGMP (pEC50 9.1±0.3, n=5) and pERK1/2 (pEC50 9.1±0.40, n=3) were also seen in HCFs where serelaxin increased cGMP and pERK1/2 to about 50% of control responses. In HUASMCs, serelaxin increased cAMP accumulation (pEC50 9.0±0.3, n=4) to 5% of forskolin, cGMP accumulation (pEC50 9.1±0.3, n=6) to 50% of DEA-NO and pERK1/2 (pEC50 9.1±0.4, n=6) to 20% FBS. Pre-treatment with PTX (50ng/mL, 16hr) or wortmannin (100nM, 30min) inhibited almost all serelaxin responses in all cells indicating the involvement of inhibitory G proteins (Gαi) and PI-3-kinase. Unlike HUVEC and HUVSMC, CRCs for cAMP and cGMP accumulation and pERK1/2 in HUASMC were sigmoidal.

The bell-shaped CRCs observed in HUVSMC were altered by Gαi/o inhibition and lipid raft disruption, but not Gαs or βγ inhibition. Pre-treatment of HUASMC (sigmoidal CRC) and HUVSMC (bell-shaped CRC) with NF449 (10μM, 30min, Gαs inhibitor) inhibited the initial phase of the serelaxin CRC, causing a rightward shift in the CRC and a drop in the E-max for cAMP and cGMP accumulation. Pre-treatment of HUASMC with NF023 (10μM, 30min, Gαi/o inhibitor) inhibited the latter part of the serelaxin CRC, causing a drop in the E-max of cAMP without changing potency. Pre-treatment of HUVSMC with NF023 decreased E-max for cAMP and cGMP accumulation and converted a bell-shaped CRC to sigmoid. Disruption of lipid rafts with filipin III pretreatment (1μg/mL, 1hr) had similar effects.

Thus human vascular cells respond to serelaxin with a variety of signalling patterns that differ between arteries and veins. Within the cell types examined here, bell-shaped CRCs that are a hallmark of serelaxin signaling experimentally and clinically were only observed in venous cells and fibroblasts. The CRCs are clearly influenced by Gαi/o associated with membrane lipid rafts and are cell-background dependent. Our studies identify human cell systems that may be useful in determining mechanisms underlying the bell-shaped concentration-response relationships to serelaxin observed clinically.

(1) Teerlink JR et al, Lancet 381: 29, 2013