Abstract
Acute decompensated heart failure (ADHF) is associated with a high incidence of acute kidney injury (AKI), an abrupt loss of kidney function associated with a near doubling of mortality at one year. In addition to the direct threat ADHF itself poses to kidney function, several drugs commonly used to treat HF can exacerbate kidney dysfunction and the beneficial effects of these treatments must be weighed against their adverse effects on glomerular perfusion. Serum creatinine (SCr), the gold-standard marker for glomerular filtration, is a delayed and insensitive marker of kidney function that can delay the diagnosis of AKI by 24-72 hours. During this time, irreversible kidney damage can occur. Consequently, there is an urgent need to identify early markers for AKI and long-term renal impairment in ADHF to facilitate timely implementation of supportive measures to minimise kidney damage and improve patient outcomes.
This project performed a search for novel markers of AKI in ADHF across two classes of potential biomarkers, circulating microRNAs (miRNAs) in patients with ADHF and kidney proteins in an ovine model of renal impairment in ADHF. First, a large-scale screen of 375 miRNAs was performed in a well-phenotyped discovery cohort consisting of ADHF patients who incurred AKI (n=19), and age- and gender-matched ADHF patients who did not incur AKI (n=20), to identify a panel of candidate miRNA markers for AKI in ADHF. Subsequently, circulating concentrations of the 10 most promising candidate miRNAs were measured in 200 consecutively-recruited ADHF patients (with a spectrum of kidney dysfunction) to determine whether any of the candidate miRNAs had potential to provide independent prognostic information beyond established markers such as SCr and markers of ADHF severity including amino-terminal pro b-type natriuretic peptide (NT-proBNP).
Second, kidney tissue samples from an ovine model of ADHF were used to characterise the changes in renal protein expression that occur in response to the development of, and recovery from, renal impairment as a result of ADHF, and identify candidate protein markers of AKI in ADHF. Relative protein quantification was performed in kidney tissue from healthy control sheep (n=5), sheep with established ADHF (n=8) and sheep recovered from ADHF (n=7) using Sequential Window Acquisition of all Theoretical Mass Spectra (SWATH-MS). By investigating associations between candidate miRNA concentrations and AKI, and characterising alterations in renal protein expression in response to development of renal impairment following ADHF, this work aimed to identify clinically useful biomarkers which could be used to identify ADHF patients at risk of AKI.
Key findings from this research included identification of 26 candidate miRNAs which were differentially expressed between patients with ADHF who incurred AKI, and patients with ADHF who did not. Of these candidates, only miR-454-3p validated in a larger, independent cohort. Higher concentrations of miR-454-3p prior to discharge from hospital were associated with lower risk of AKI or intermediate kidney dysfunction in ADHF patients and may add value beyond established risk factors including age in predicting AKI (Odds Ratio 0.58, 95% CI 0.34-1.01, adjusted for ethnicity, age, gender and NT-proBNP). The association between miR-454-3p and AKI was consistent across all major ethnic groups in the validation study (European, Māori and Pasifika), but the study was only sufficiently powered to detect this effect in patients of European ancestry.
Through protein expression profiling, this work identified 189 proteins that were altered > 1.2-fold during the development of and recovery from ADHF in an ovine model. Of these, 11 proteins were detectable in serum, plasma or urine and had a fold-change > 1.2 with an adjusted p-value < 0.05. These included seven candidate protein markers of kidney injury (Filamin A, T-complex protein 1 subunit gamma, Talin 1, Chaperonin containing TCP1 subunit 8, T-complex 1, Crystallin zeta, Peroxiredoxin), one potential protein marker of kidney recovery (Apolipoprotein A4) and three potential protein markers of long-term renal impairment (Apolipoprotein E, Tripeptidyl peptidase 1, Chaperonin containing TCP1 subunit 6A). Of these candidates, eight had never previously been associated with AKI. Differentially expressed proteins were enriched in two pro-inflammatory signalling pathways: the Glycoprotein VI signalling pathway was activated during ADHF development (Z-score 2.65, p< 0.01), and the acute phase response signalling pathway was repressed during recovery from ADHF (Z-score -2.25, p< 0.01).
In conclusion, these findings suggest that miR-454-3p and candidate protein biomarkers, potentially together and in combination with existing clinical risk factors, have the potential to improve detection of AKI, kidney recovery and long-term renal impairment in ADHF. New biomarkers for AKI in ADHF may facilitate targeted treatment and monitoring of patients with ADHF.