Evaluation of Enhanced Lipid Oxidation and Compensatory Suppression using Natriuretic Peptide in Patients with Cardiovascular Diseases

Malondialdehyde-modiﬁed low-density lipoprotein (MDA-LDL) is recognized as a surrogate marker of lipid oxidation and is associated with arteriosclerosis. However, there are limited reports on the relationship between heart failure and MDA-LDL. Therefore, we aimed to determine whether MDA-LDL is activated in patients with left ventricular (LV) dysfunction and examine our hypothesis that the B-type natriuretic peptide (BNP) masks the enhancement of MDA-LDL in patients with LV dysfunction by its strong antioxidative action. The study population comprised 2,976 patients with various cardiovascular diseases. Patients were divided into four groups depending on the LV ejection fraction (LVEF) or plasma BNP level. A nonparametric analysis with the Kruskal-Wallis test was used to perform an interquartile comparison. In addition, structural equation modeling and Bayesian estimation were used to compare the effects of LVEF and BNP on MDA-LDL. MDA-LDL levels did not significantly change (P > 0.05) with respect to the degree of LVEF among the four groups. In contrast, MDA-LDL levels were significantly decreased (P < 0.001) with respect to the degree of BNP among the four groups. A path model based on structural equation modeling clearly showed a significant effect of LVEF (standardized regression coefficient; β: -0.107, P < 0.001) and BNP (β: -0.114, P < 0.001) on MDA-LDL, with a significant inverse association between LVEF and BNP (correlation coefficient -0.436, P < 0.001). MDA-LDL should be activated in patients with LV dysfunction; however, BNP is thought to exert a strong compensatory suppression on lipid oxidation, masking the relationship between heart failure and lipid oxidation.


Introduction
The involvement of oxidative stress in heart failure has been noted. Basic experiments and clinical studies have shown important evidence that oxidative stress, as represented by reactive oxygen species (ROS), is significantly increased during heart failure, and its action exceeds that of antioxidative stress [1][2][3]. Additionally, nicotinamide adenine dinucleotide phosphate, thioredoxin, and xanthine oxidase have been reported to be related to heart failure [4][5][6][7][8].
Among numerous substances involved in oxidative stress, lipid oxidation is associated with the pathogenesis of various cardiovascular disorders. Malondialdehyde-modified low-density lipoprotein (MDA-LDL) is a low-density lipoprotein cholesterol (LDL-C) that has been modified by malondialdehyde, leading to the production of a large amount of aldehyde when LDL-C degenerates and becomes oxidized [9]. MDA-LDL is recognized as a surrogate marker of lipid oxidation, and previous studies showed that increased serum MDA-LDL levels were associated with coronary artery disease [10][11][12][13]. Moreover, the MDA-LDL-to-LDL-C ratio is a more useful predictor of coronary artery calcification than the MDA-LDL level alone [14]. Although MDA-LDL is important in cardiovascular disease, most reports on it concern atherosclerosis. In contrast, there are very few reports on the relationship between heart failure and MDA-LDL. The reason for this could be that antioxidative factors may be activated to suppress lipid oxidation in heart failure. In fact, the role of lipid oxidation may be stronger than originally thought.
A-type natriuretic peptide (ANP) and B-type natriuretic peptide (BNP) are cardiac hormones with a wide range of potent biological Abbreviations: MDA-LDL, Malondialdehyde-modified low-density lipoprotein; BNP, B-type natriuretic peptide; LVEF, left ventricular ejection fraction; HFpEF, heart failure with preserved ejection fraction; ROS, reactive oxygen species; LDL, low-density lipoprotein; ANP, A-type natriuretic peptide; GMP, guanosine monophosphate; IL, interleukin; 2-D, 2-dimensional. effects, including vasodilation and natriuresis, sympathetic nervous system effects, and inhibition of the renin-angiotensin-aldosterone system [15][16][17][18][19][20][21]. The magnitude of secretion of BNP also varies and depends on the severity of heart failure. Therefore, plasma BNP levels are used as biochemical markers of heart failure in the clinical setting [16,17,[22][23][24][25]. ANP and BNP have been previously shown to be anti-inflammatory hormones [26]. The second messenger of the natriuretic peptide is cyclic guanosine monophosphate (GMP), which reduces oxidative stress [27][28][29]. Additionally, carperitide, an infusion of human ANP, is useful for improving hemodynamics and inhibiting ROS production in patients with heart failure [30]. Furthermore, BNP regulates the production of major inflammatory molecules, such as ROS, reactive nitrogen species, leukotriene B(4), and prostaglandin E(2); modulates the cytokine profile (tumor necrosis factor-alpha, interleukin [IL]-12, and IL-10); and affects cell motility [31]. Although there are many research reports on the relationship between BNP and oxidative stress, only a few papers on MDA-LDL are available in the literature [32]. The action of BNP on MDA-LDL remains an issue to be examined in detail.
It is important to simultaneously examine the effect of the degree of left ventricular (LV) dysfunction on MDA-LDL and the effect of BNP on MDA-LDL; however, in practice, such examination is difficult because the degree of LV dysfunction and BNP have a strong relationship. The objective of this study is, thus, to determine what types of statistical analysis method can be used in combination. Therefore, in this study, we aimed to determine whether MDA-LDL is activated in patients with LV dysfunction and examine our hypothesis that BNP masks the enhancement of MDA-LDL in heart failure with its strong antioxidative action.

Study design and population
The study population comprised 2,976 patients who were consecutively admitted to our institution between March 2012 and June 2016. Patients with a history of percutaneous coronary intervention or coronary artery bypass graft were excluded from this study. Patients were divided into the following groups depending on their left ventricular ejection fraction (LVEF) and plasma BNP level: group A, LVEF < 20%; group B, LVEF 20-39.9%; group C, LVEF 40-59.9%; and group D, LVEF ≥ 60% and group I, BNP level <18.4 pg/ml; group II, BNP level 18.4-99.9 pg/ml; group III, BNP level 100-399.9 pg/ml; and group IV, BNP level ≥400 pg/ml.
Since this study had a retrospective design, instead of obtaining informed consent from each patient, we publicly posted a notice about the study design in detail and contact information at our institution. This study was approved by the ethics committee of the Jikei University School of Medicine (protocol: 24-150[6916]) and complied with the routine ethical regulations of our institution.

Data collection
Cardiac catheterization was performed in all study patients. The LVEF was calculated from left ventriculography.
Blood sampling was conducted for every patient during catheterization. Serum biochemical analyses were performed in our hospital. Blood sampling was performed to examine the serum MDA-LDL, BNP, insulin, fasting blood sugar, glycated hemoglobin A1c, triglyceride, high-density lipoprotein cholesterol, and LDL-C levels. The plasma BNP level was measured in our institution by a central laboratory using the E Test TOSOH II (Tosoh Corporation, Tokyo, Japan; http://www.diagn ostics.jp.tosohbioscience.com/immunoassay/aia-reagents).
The normal values in that system are reported as follows. For men under the age of 45 or women under the age of 55, the MDA-LDL value is 64 ± 18 U/L. For men over the age of 45 or women over the age of 55, it is 83 ± 22 U/L.

Statistical analysis
Continuous variables are expressed as mean ± standard deviation or median. A nonparametric analysis with the Kruskal-Wallis test was used to perform an interquartile comparison. Statistical analyses were performed using IBM SPSS Statistics, version 23.0 (IBM Corp., Armonk, NY, USA), and differences were considered to be statistically significant for P-values <0.05.
In addition, structural equation modeling was performed to compare the effects of LVEF and BNP on MDA-LDL. To focus on the effect of patients with a preserved ejection fraction, additional statistical analysis was performed only in patients with a LVEF ≥ 50%. The implementation procedures of structural equation modeling have been described previously [33]. The obtained structural equation models were tested and confirmed at the significance level for P-values <0.05. Furthermore, we used Bayesian structural equation modeling to clarify the results of the structural equation modeling with a 2-dimensional (2-D) contour image. Path analysis was performed using IBM SPSS AMOS, version 23 (Amos Development Corporation, Meadville, PA, USA).

Study population
The baseline characteristics of the study population (n = 2,976) are shown in Table 1. Table 2A and 2B shows the clinical characteristics of the LVEF group (Table 2A) and BNP group (Table 2B).

Results of comparison between the groups
The MDA-LDL levels did not significantly change (P > 0.05) by the degree of LVEF among the four LVEF groups (

Concept of the proposed path model
The proposed path model is shown in Fig. 3. Logically, BNP and LVEF are expected to potentially confound each other. Paths between variables are drawn from independent to dependent variables, with a directional arrow for every regression model, namely, from LVEF and BNP to MDA-LDL. Table 3 shows the detailed results of structure equation modeling in all study patients. We analyzed the effects of LVEF and BNP on MDA-LDL using a path model (Fig. 3). The path model clearly showed a significant effect of LVEF (standardized regression coefficient; β: -0.107, P < 0.001) and BNP (β: -0.114, P < 0.001) on MDA-LDL, with a significant inverse association between LVEF and BNP (correlation coefficient: -0.436, P < 0.001). Additionally, the statistical analysis in patients with a preserved ejection fraction revealed the same findings (LVEF β: -0.076, P = 0.002 and BNP β: -0.097, P < 0.001) ( Table 4, Fig. 4). This suggests that the MDA-LDL relationship is maintained over a wide range of LVEFs. Furthermore, the path model was clearly demonstrated on the 2-D contour image by Bayesian structural equation modeling (Fig. 5). From this figure, it can be visually understood that the lower the LVEF, the higher the MDA-LDL value, while the higher the BNP value, the lower the MDA-LDL value.

Discussion
MDA-LDL is a marker of oxidative stress, and lipid oxidative stress is deeply involved in the mechanism of arteriosclerotic progression. However, even in heart failure, increased oxidative stress is suggested, and research on MDA-LDL should also be advanced. We have previously studied MDA-LDL [13]. At that time, it was found that MDA-LDL is enhanced in acute coronary syndrome. Although the relationship between heart failure and MDA-LDL was examined at the same time, no significant relationship was found. Even in the current study, a simple comparison of the median values of MDA-LDL and the degree of LVEF did not show a relationship between LVEF and MDA-LDL. Thus, the  results from our previous and the current study are in agreement. Although oxidative stress is increased in LV dysfunction, these two results alone indicate that MDA-LDL is not exceptionally activated in LV dysfunction.
In the examination among the four groups divided by BNP, MDA-LDL suppression appears to have a significant relationship with BNP. Because of this, considering the strong negative relationship between LVEF and BNP, the relationship between LVEF and MDA-LDL should be reconsidered again with another method of analysis. We used structural equation modeling to simultaneously examine the effects of LVEF and BNP on MDA-LDL. This analysis clearly showed that a reduction in LVEF increases the MDA-LDL level. In other words, in LV dysfunction, oxidative stress markers are considered to be enhanced, including MDA-LDL.
In addition, when considering the effect of patients with a preserved ejection fraction, the analysis showed a significant relationship between LV dysfunction and MDA-LDL, which was almost consistent with that found in the entire patient population. The relationship between diastolic disorders and MDA-LDL is interesting. However, since there is no echocardiographic information at this time, we would like to make it a future subject of research. In addition, we also conducted various analyses in our subgroups. For example, patients were divided into four groups based on LVEF, and the same assessment was conducted in each   The MDA-LDL value in group IV (BNP ≥ 400) was found to be lower than that in group I (BNP < 18.4), which was unexpected. A possible explanation could be that this study group was comprised of patients who underwent cardiac catheterization, and so some abnormalities were suspected. In group I (BNP < 18.4), the BNP value was in the normal range, but in reality, there is concern about some kind of heart damage. The MDA-LDL value tended to be the highest in group I (BNP < 18.4), and the MDA-LDL value decreased as the BNP increased. In other words, it is interesting that the MDA-LDL value in group IV (BNP ≥ 400) is the lowest among the four groups, and the value is close to the MDA-LDL value of healthy persons (refer to the normal values of this factor in the Material and Methods section).
On the other hand, we believe that BNP does not show antioxidant activity only at high levels. It is possible that the antioxidant activity of BNP has already been exerted in group I (BNP < 18.4). If there was no BNP, the MDA-LDL value in group I may have been higher. This is because we have reported the following phenomena. Recently, we showed that troponin I is increased in patients with extremely low BNP levels (<4 pg/ml) [34]. Although BNP has various actions in heart failure, including natriuretic action, vasodilating action, inhibitory action on the renin-angiotensin-aldosterone system, and antioxidative stress action, the action of BNP is already exerted even at very low or physiological levels. Subtle control of endogenous BNP may be important. For example, it would be important in obesity to reduce endogenous ANP and BNP levels. Although obesity induces cardiovascular events, it is likely that the low BNP induced by obesity is involved in the process [35]. Since the number of cases with extremely low BNP levels <4 pg/ml is small in this study, we would like to consider this as a topic for further research.
Studies of the mechanism by which BNP lowers the MDA-LDL level will be important. Cyclic GMP is essential, which may directly contribute as antioxidants in cells, while interventions of other factors are also envisaged. Moreover, the involvement of adiponectin may be important. Adiponectin is an adipocyte-specific factor, and its reduction plays a central role in obesity-related diseases, including insulin resistance, atherosclerosis, and cardiovascular disease. A low adiponectin level is inversely associated with an oxidized LDL level [36]. Importantly, natriuretic peptide treatment increases adiponectin gene expression and adiponectin secretion in cultured human adipocytes [37].
The molecules involved in oxidative stress are diverse. Although BNP acts to repress oxidative stress, it is unknown on which factor BNP acts more strongly. A comprehensive analysis of molecules related to oxidative stress is needed in the future. At least in this study, it was clarified that BNP strongly suppresses MDA-LDL even at physiological levels of BNP. In many studies so far, the activity of MDA-LDL in heart failure has been obscured because BNP has a strong inhibitory effect on MDA-LDL.
We reexamined the data using MDA-LDL/LDL-C instead of MDL-LDL (data not shown). Neither BNP nor LVEF had a strong relationship with the MDA-LDL/ LDL-C ratio, although some results tended to be similar. The reason for this is not clear, but the significance of the absolute value of MDA-LDL and the MDA-LDL/LDL-C ratio is probably different; the former was probably effective for this analysis. It will be interesting to determine why in future studies.
Drug analysis may be important. In particular, we have described the relationship between statins and MDA-LDL levels in a previous paper [38] ; that is, statins lower MDL-LDL. Indeed, in this study, the prescription rate of statins was high in the group with a low BNP. Nonetheless, MDA-LDL was higher in the low BNP group; therefore, we believe that statins have no effect on this result.  The association between two factors is shown by two-way arrows.

Study limitations
This study had several limitations. First, among the study population, the proportion of patients with severely impaired cardiac function was low. We need to examine other population groups with highly reduced LVEF. Second, many patients had ischemic heart disease in this study; so it is necessary to perform the same examination in a group of patients with other underlying diseases such as valvular heart diseases and cardiomyopathy. Third, here we showed that, statistically, heart failure increases MDA-LDL and BNP strongly suppresses it. However, these relationships should be supported by biochemical analysis. Basic research on this is required in the future. Forth, the difference between BNP and NT-proBNP has been discussed in recent years [39,40]. Since only BNP was measured in this study, similar studies may be needed for NT-proBNP.

Conclusions
MDA-LDL should be activated in heart failure. However, BNP is thought to exerts a strong compensatory suppression on lipid oxidation, masking the relationship between heart failure and lipid oxidation.

Data statement
Restrictions apply to the availability of data generated or analyzed during this study to preserve patient confidentiality or because they were used under license. The corresponding author will on request detail the restrictions and any conditions under which access to some data may be provided.

Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Declaration of Competing Interest
The authors report no declarations of interest.  The association between two factors is shown by two-way arrows.