A novel long-acting relaxin-2 fusion, AZD3427, improves cardiac performance in non-human primates with cardiac dysfunction

Abstract

Aims: Relaxin-2, a well-known human hormone primarily associated with pregnancy, has shown promising cardiovascular benefits in both pre-clinical models and clinical trials. However, its therapeutic potential has been limited due to the short half-life and the short duration of treatment. To address this, we developed AZD3427, a novel long-acting relaxin-2 analogue, and assessed its efficacy during prolonged treatment in a large animal model with cardiac dysfunction.

Methods and results: Extensive protein engineering resulted in AZD3427, a novel fusion protein, which closely mimics the natural hormone's structure and consists of a single relaxin-2 and the Fc fragment of human IgG1 to extend its half-life. AZD3427 exhibits an improved pharmacokinetic profile, allowing for weekly or less frequent, subcutaneous dosing, and maintains the pharmacology profile of relaxin-2 with signalling via relaxin family peptide receptor 1 (RXFP1) in cell systems. The effects of chronic RXFP1 agonism with AZD3427 were investigated in a non-human primate (NHP) model with systolic dysfunction and metabolic syndrome. Administration of AZD3427 over a 21-week period led to significant improvements in cardiac function, as evidenced by increased ejection fraction, cardiac output, and stroke volume, as well as reduced systemic vascular resistance. Importantly, no adverse events related to treatments were observed and there were no concomitant changes in heart rate or blood pressure. During the 18-week washout period, the observed effects gradually disappeared.

Conclusion: Prolonged administration of AZD3427, a long-acting relaxin receptor RXFP1 agonist, resulted in remarkable improvement in cardiac function in a NHP model. Findings of this study are an important translational step to developing future therapies and support further clinical development of AZD3427 as a novel treatment for patients with heart failure.

Keywords: AZD3427; Heart failure; Non-human primate; RXFP1; Relaxin-2.

Graphical Abstract

Figure 1

Study design of AZD3427 treatment in NHP with reduced EF.

Figure 2

Schematic representation of relaxin-2 and Fc–relaxin AZD3427. (A) NMR structure of relaxin-2 (PDB/MMDB ID 2MV1), and schematic representation of relaxin-2 composed of chain A and chain B. Structure–critical disulphide bonds are included in the NMR structure model. (B) Heterodimeric Fc–relaxin is depicted as separate Fc-A and Fc-B monomers with complementary knob-in-hole mutations in the Fc chains (left) and assembled heterodimer AZD3427 (right) which folds during production in CHO cells. (C) Structure of AZD3427 generated in silico using homology modelling with MD equilibration with the heterodimeric Fc chains depicted on the top of model (red and dark grey) and relaxin-2 heterodimer chains depicted at the bottom of the model (blue and green) with linkers and disulphide bonds included.

Figure 3

Stimulation of cAMP production by the RXFP1 agonists in recombinant cells lines. Recombinant cell lines expressing either (A) human, (B) cynomolgus, or (C) mouse RXFP1 were treated with dilutions (Log M) of AZD3427 or recombinant human relaxin-2. The cAMP production was measured in cell lysates using an homogeneous time-resolved fluorescence (HTRF) assay and is shown as % Effect normalized to the maximum effect stimulated by a saturating concentration of relaxin-2. Typical dose–response curves are shown on the graphs. Table shows pEC₅₀ values ± standard deviations from independent experiments (n = 6–8) and EC₅₀ geometrical mean values from all experiments.

Figure 4

Pharmacokinetic profiles following once-weekly subcutaneous dosing of AZD3427 in NHPs with reduced EF. (A) NHP with maintained exposure throughout the study period (mean ± SD). n = 5 for 0.015 mg/kg and n = 6 for 0.15 and 3 mg/kg dose group each. (B) Individual NHP with ADA affecting the AZD3427 exposure, shown as individual values (two NHPs in all three dose groups). The shaded area represents samples below limit of quantification.

Figure 5

Functional cardiac and haemodynamic parameters in NHP model with reduced EF after 21 weeks of AZD3427 treatment and 18 weeks of washout. (A) EF improvement was dose dependent, (B) stroke volume, (C) CO, (D) HR, (E) MAP, and (F) SVR. The graphs show mean ± SEM: n = 14 for the vehicle, n = 7 for the 0.015 and 0.15 mg/kg dose groups, and n = 8 for 3 mg/kg.

Figure 6

Plasma renin concentrations trends over 21 weeks of AZD3427 treatment and 18 weeks of washout. NHPs with maintained AZD3427 drug exposure are shown (seven NHPs excluded due to ADA that impacted PK). The graphs show mean ± SEM: n = 14 for the vehicle, n = 5 for the 0.015 and 0.15 mg/kg dose group, and n = 6 for 3 mg/kg.