Evidence for a hyper-reductive redox in a sub-set of heart failure patients

Our study population included heart failure (HF) patients (n = 54) (with systolic or diastolic dysfunction, dilated cardiomyopathy or biventricular dysfunction) and healthy controls (n = 42) who attended the in-patient and out-patient department of cardiology in PSG Hospitals, Tamil Nadu, India. Their characteristics are tabulated in Table 1. Patients with HF who were diagnosed using Framingham criteria were considered for this study [25]. Patients on dialysis, as well as those exhibiting severe liver diseases, malignancies or consuming antioxidant supplements were excluded from the study. While 85% of patients were admitted to the hospital for the first time, the remaining were presented for follow-up visits. The patients were treated as per the existing guidelines at the time. Briefly, the patients were categorized into those displaying heart failure (HF) with reduced ejection fraction (HFrEF) and preserved ejection fraction (HFpEF). The patients with HFrEF received stage-appropriate medications such as ACE inhibitors or Beta Blockers or ARBS alone or in combination, along with diuretics. Mineralocorticoid Receptor Antagonists and Hydralazine were also added if required. Patients with HRpEF had diuretics as their main stay along with the treatment of underlying causes such as coronary artery disease, chronic kidney disease, HF, atrial fibrillation etc. The study was approved by the PSG Institutional Human Ethics Committee (IHEC) and all patients/subjects completed a written informed consent prior to their participation.

Supplies and reagents were purchased from Sisco Research Laboratories (India) unless otherwise specified.

Validation and confirmation of HF in patients were made based on clinical and echocardiography assessment. One echocardiogram was selected for each patient at the time of enrollment. Two-dimensional and M-mode echocardiography, color flow and spectral Doppler as well as annular TDI data were obtained from all patients and healthy controls using an ultrasound system (Philips IE33, Netherlands). Standard views, including the parasternal long axis view, short axis at the papillary muscle level and apical four and two chamber views were recorded. Cardiac chamber dimensions, volumes and left ventricular mass were measured according to current recommendations [26]. Left ventricular ejection fraction (LV EF) was calculated using the teichholz formula [27]. Mitral Doppler signals were recorded in apical 4 chamber view, with Doppler sample volume placed at the tip of the mitral valve leaflets. The following parameters were obtained: early diastolic mitral inflow peak velocity (E), late diastolic mitral inflow peak velocity (A) and their ratio.

Blood samples were obtained by venipuncture using a 21-gauge needle from the patients and healthy control subjects in plain tubes. Serum was separated by centrifugation at 3000 rpm for 5 min at 4 °C and immediately aliquoted and frozen at − 80 °C. A fraction of the serum was treated with 10% meta-phosphoric acid (MPA) to remove proteins and the MPA-supernatant was stored for glutathione redox analysis instantly after the blood collection for all the controls and patients. The stored aliquots of serum were used for analyzing redox status and the activities of antioxidant enzymes.

Serum malondialdehyde levels were measured by the DNPH (2,4-dinitrophenyl hydrazine) derivatization method [28]. Briefly, 125 µl of serum was diluted in 125 µl 1× PBS (pH 7.4), mixed with 50 µl of 6 M NaOH and incubated at 60 °C for 30 min. 125 µl of 35% perchloric acid was then added to the mixture. 250 µl of this mixture was combined with 25 µl of 5 mM DNPH (in 2 M HCl) and incubated in the dark for 30 min. An aliquot of 25 µl of the solution was injected into the HPLC system (Shimadzu, Japan). The mobile phase of 0.2% (v/v) acetic acid and acetonitrile (50:50 v/v) was run at a flow rate of 0.5 ml/min at 25 °C. A C18 column (Agilent, USA) was used and the chromatograms were obtained at 310 nm. Concentration of MDA was measured after comparing with a reference curve using TMP (1,1,3,3-tetramethoxypropane) as a standard.

The spectrophotometric based glutathione reductase/5,5′-dithio-bis (2-nitrobenzoic acid) (DTNB) recycling assay was used to measure reduced glutathione (GSH) levels as per [29‐31] with minor modifications. In short, metaphosphoric acid extracts of serum samples were prepared and treated with triethanolamine (TEAM reagent) to adjust the pH for total GSH quantification according to the manufacturers’ protocols (GSH redox Kit #CS0260, Sigma-Aldrich). An aliquot of TEAM treated samples were mixed with 1.0 mM 2-vinyl pyridine and incubated for 1.0 h at room temperature for GSSG measurements. For enzymatic recycling, the processed samples were treated with a reaction mixture containing 0.25 mM dithiobis-nitrobenzoic acid (DTNB), 0.38 U/ml Glutathione reductase and 60 µl of 0.17 mM NADPH in 0.1 M sodium phosphate buffer (pH 7.4) with 5 mM EDTA. The rate of DTNB formation was measured at A412 nm. Similarly, GSH and GSSG standards were treated and measured to obtain a standard graph to extrapolate the values from serum samples. The concentration of reduced glutathione (GSH) was estimated by subtracting the measured oxidized (GSSG) glutathione levels from the measured total (GSH plus GSSG) glutathione. GSH/GSSG ratio was then determined.

Superoxide dismutase activity in the serum samples were measured using the pyrogallal auto-oxidation method [32]. This assay was carried out in a microtiter plate and the readings were measured using multimode reader (Varioskan Flash, Thermoscientific, USA). Approximately 50 µl of serum was treated with 25 µl ethanol and 15 µl chloroform. The contents were vortexed and centrifuged at 8000 rpm for 5 min and the supernatant was analyzed for SOD activity by mixing with 50 mM Tris–HCl buffer with 1.0 mM EDTA, pH 8.2 and 2.0 mM pyrogallal (in 50 mM HCl). The rate of autoxidation was measured from the increase in absorbance at 420 nm and the values were expressed as U/ml.

The activity of catalase in serum of HF and HC samples was measured by H₂O₂ decomposition kinetics using a spectrophotometer [33]. A mixture containing potassium phosphate buffer (50 mM, pH 7.0), 20 mM H₂O₂ and 2.0 µl serum was used to measure the rate of decrease in optical density at 240 nm. Catalase activity was expressed as U/l.

Glutathione reductase activity was assayed by monitoring NADPH oxidation linked to GSSG reduction [34] using a spectrophotometer. The assay was performed by suspending 10 µl of the serum in sodium phosphate buffer (0.3 M, pH 6.8), 25 mM EDTA, 20 mM GSSG, 2 mM NADPH and measuring the decrease in absorbance at 340 nm over time. The oxidation of 1 µmol of NADPH/min was used as a unit of glutathione reductase activity. Glutathione reductase activity was expressed as U/l.

Glutathione peroxidase (GPx) activity was measured by a spectrophotometric assay using 5,5′-dithio-bis (2-nitrobenzoic acid) (DTNB) according to [35] where the oxidation of GSH in the presence of H₂O₂ takes place, and the unused GSH reacts with DTNB whose absorbance was measured at 412 nm. The enzyme assay was performed by incubating the serum (25 µl) in sodium phosphate buffer containing sodium azide (10 mM), GSH (4 mM) and 2.5 mMH₂O₂ for 5 min. Protein was precipitated using 10% trichloroacetic acid, centrifuged at 5000 rpm for 5 min, and the supernatant was assayed. This assay was carried out in a microtitre plate and the readings were measured using multimode reader (Varioskan Flash, Thermoscientific, USA) after mixing with disodium hydrogen phosphate (0.3 M, pH 7.0) and 100 mM DTNB (in 1% sodium citrate) at 412 nm. GPx activity (U/l) was measured as the amount of GSH consumed/min/mg protein after reaction with GPx and DTNB.

Serum proteins were estimated using Bradford method. A known volume (5 μl) of 1:100 diluted serum was incubated in Bradford reagent (#18004246723, Biorad, USA) and measured the absorbance at 595 nm. The absorbance was extrapolated with the appropriate standard reference curve using bovine serum albumin (0.031–0.5 mg/ml) and the protein concentration was calculated. Specific activity for the antioxidant enzymes were measured after normalizing with the respective protein concentration.

For comparing two groups, Mann–Whitney U test was used for non-normal data and for groups with normal distribution, group mean ± standard deviation (SD) were compared, homogeneity variance test (Bartlett’s test) followed by one way ANOVA was carried out. A p > 0.05 was considered to be statistically significant. Fisher’s exact test was performed to analyse statistical significance of EF fraction and E/A ratio among the different groups of HF patients. For data which was not normally distributed, a nonparametric test, Kruskal–Wallis test was carried out to obtain the p value. All the statistical analyses were carried out using GraphPad Prism software version 5.0 Windows (GraphPad Software, USA).

Circulatory redox state (CRS) was assessed by quantifying the levels of reduced glutathione (GSH), a ubiquitous, small molecular antioxidant redox couple and malondialdehyde (MDA), an end-product of lipid peroxidation (Fig. 1). Compared to the healthy controls (HC; n = 42), circulating GSH levels were unchanged in HF patients (n = 54) (0.09207 ± 0.1215 µM in HF patients vs. 0.06374 ± 0.04887 µM in HC; p = NS; Fig. 1a), suggesting a wide range of redox states among the HF patients. In addition to total glutathione levels, the ratio of reduced to oxidized glutathione (GSH:GSSG) did not differ between HF and HC (4.93 ± 7.25 in HF patients vs. 3.01 ± 1.78 in HC; p = NS; Fig. 1c). In contrast, MDA was significantly higher in HF patients vs. HC (4.342 ± 2.103 vs. 1.566 ± 0.3018 µM, respectively; p < 0.0001) (Fig. 1b), suggesting that increased lipid peroxidation is coupled with the development of HF. Therefore, we next assessed whether the redox index (reduction vs. oxidation) is significantly different between the groups. Comparisons of GSH/GSSG ratio revealed insignificant differences between the groups, indicating that the HF patients seem to exhibit a wide range of redox levels in relation to HC (4.93 ± 7.25 in HF patients vs. 3.01 ± 1.78 in HC; p = NS; Fig. 1c). Therefore, our approach for stratifying HF patients used CRS (i.e. GSH/MDA) to gain more knowledge and understand the pathogenesis. But, comparisons of GSH/MDA ratios between the HF and HC yielded a statistically significant result (0.02 ± 0.0345 in HF patients vs. 0.04 ± 0.028 in HC; p < 0.0001; Fig. 1d). Therefore, we were interested in stratifying the HF patients based on their circulating redox score (CRS) (Fig. 1d). The CRS of healthy subjects was designated as normal redox (NR). A CRS value in any patient lower than the NR displayed by HC was regarded as hyper-oxidative (HO) while a patient exhibiting a GSH/MDA ratio greater than the NR cutoff was classified as hyper-reductive (HR). Surprisingly, our CRS-based stratification revealed three distinct classes within our cohort of HF patients (Fig. 1e). Along these lines, 41% of HF patients were assigned to the normal redox (NR) group whose mean GSH/MDA ratio was similar to HC. Next, 42% of HF patients belonged to the hyper-oxidative (HO) class whose mean GSH/MDA ratio (0.004787 ± 0.002055) was lower than HC group (0.03969 ± 0.02857), and 17% of HF patients exhibited a hyper-reductive (HR) state in which mean GSH/MDA ratio (0.2271 ± 0.03192) was greater than HC group (0.03969 ± 0.02857). Interestingly, the GSH/MDA ratios were statistically significant among all three HF groups (p < 0.0001) by one way ANOVA analysis. These novel observations suggest that redox status plays a variable role in the progression of HF and that both extremes (e.g. oxidative and reductive) may contribute to disease pathogenesis.

To assess the capacity of the endogenous antioxidant defense system in the periphery of HF patients, we further analyzed the activity of circulating antioxidant enzymes such as superoxide dismutase (SOD), glutathione reductase (GR), catalase, and glutathione peroxidase (GPx) among the various groups of HF patients (Fig. 2). SOD (Fig. 2a), Gpx (Fig. 2b), catalase (Fig. 2c), GR (Fig. 2d) did not show any statistically significant differences among NR, HO, and HR groups of HF patients. While the activities of SOD (6.455 ± 2.30 for HO, 6.646 ± 2.556 for NR, 6.400 ± 1.752 for HR vs. 2.571 ± 0.61 U/ml for HC; Fig. 2a), GPx (527.2 ± 343.4 for HO, 413.3 ± 243.7 for NR, 581 ± 310.2 for HR vs. 312 ± 141.4 U/l for HC; p < 0.01; Fig. 2b), and catalase (855.2 ± 776.5 for HO, 787.8 ± 454.9 for NR, 1068 ± 590.3 for HR vs. 689.1 ± 556.8 U/l for HC; p < 0.05; Fig. 2c) were significantly increased, GR activity was significantly lower (11.66 ± 5.5 for HO, 14.29 ± 9.86 for NR, 13.32 ± 6.79 for HR vs. 24.70 ± 28.22 U/l for HC; p < 0.05; Fig. 2d) in all HF patient groups (HO, NR and HR) relative to healthy controls.

To relate our biochemical redox observations in the distinct subsets of HF patient with myocardial structural and functional remodeling, we performed non-invasive echocardiography analyses to determine fractional shortening, ejection fraction and left ventricular (LV) mass. First, we confirmed that HF patients exhibit significantly distinct myocardial mass and systolic function when compared to HC (Fig. 3a–c). Fractional shortening (Fig. 3a), ejection fraction (Fig. 3b) and LV mass (Fig. 3c) did not show any statistically significant differences among the NR, HO, and HR groups of HF patients. However, when compared with HC, all HF patient groups (HO, NR and HR) showed significantly lower fractional shortening (17.03 ± 9.256 for HO, 15.40 ± 5.863 for NR, 16.54 ± 5.185 for HR vs. 36.13 ± 6.020 for HC; p < 0.01) and ejection fraction (33.78 ± 17.62 for HO, 32.47 ± 10.41 for NR, 32.57 ± 9.424 for HR vs. 65.46 ± 7.9 for HC; p < 0.01), but significantly greater LV mass (309.1 ± 107.9 for HO, 281.1 ± 128.2 for NR, 320.1 ± 103.8 for HR vs. 162.6 ± 13.72 for HC; p < 0.01). Representative M-mode echo images for systolic function analysis of NR, HO, HR and healthy control groups are shown in Fig. 3d.

Primary measurements of trans-mitral inflow including the peak early filling (E wave), late diastolic filling (A wave) velocities and the E/A ratio were performed in HF patients as well as HC (Fig. 4a–c). Although there was no statistically significant difference between HF patients and the HC for ME, MA and ME/A ratio, their measures were widely distributed among the groups compared to the controls (Fig. 4d–f). Representative mitral valve Doppler analysis for diastolic function with graphs for ME/A ratios for HO, NR, HR and healthy control groups are shown in Fig. 4g. M E/A ratio differed according to age and gender [36, 37]. Based on this range, we grouped the HF patients into impaired relaxation and restrictive filling.

Among the HF patients, the HO group (55%) had a higher percentage of patients with ejection fraction ≤ 30% than the NR (44.5%) and HR groups (44.5%). Patients with an EF between 31 and 40% were more frequently assigned to the HR group (33.3%) compared to NR (28%) and HO groups (18%). In contrast, HF patients displaying an EF between 41 and 50% were less prone to be classified as HO (9%) when compared to the NR (22%) and HR groups (22%). Intriguingly, HF patients with HFpEF were completely absent in HR groups relative to HO (18%) and NR patients (5.5%) (Fig. 5a).

Using the ME/A ratio as an index, diastolic function was compared between different HF groups and healthy controls (Fig. 5b). To our surprise, none of the HR patients exhibited normal diastolic function (E/A ratio 0.76–1.6); however, 41.3% of HO and 33% of NR patients revealed such E/A ratios. While only 17% of NR patients exhibited impaired relaxation (E/A ratio ≤ 0.75), the incidence of impaired relaxation in HR and HO patients were 37.5 and 23.5% respectively. Finally, the HR group had higher percentage of patients (62.5%) with restrictive filling (E/A ratio ≥ 1.61) as compared with HO (35.2%) and NR patients (50%). Fisher’s exact test was carried out with no statistical significance for EF fraction and E/A ratio among the HO, NR, and HR group of HF patients.