Pathophysiological and diagnostic importance of fatty acid-binding protein 1 in heart failure with preserved ejection fraction

Elevated intracardiac pressure at rest and/or exercise is a fundamental abnormality in heart failure with preserved ejection fraction (HFpEF). Fatty acid-binding protein 1 (FABP1) is proposed to be a sensitive biomarker for liver injury. We sought to determine whether FABP1 at rest would be elevated in HFpEF and would correlate with echocardiographic markers of intracardiac pressures at rest and during exercise. In this prospective study, subjects with HFpEF (n = 22) and control subjects without HF (n = 23) underwent resting FABP1 measurements and supine bicycle exercise echocardiography. Although levels of conventional hepatic enzymes were similar between groups, FABP1 levels were elevated in HFpEF compared to controls (45 [25–68] vs. 18 [14–24] ng/mL, p = 0.0008). FABP1 levels were correlated with radiographic and blood-based markers of congestion, hemodynamic derangements during peak exercise (E/e’, r = 0.50; right atrial pressure, r = 0.35; pulmonary artery systolic pressure, r = 0.46), reduced exercise cardiac output (r = − 0.49), and poor exercise workload achieved (r = − 0.40, all p < 0.05). FABP1 distinguished HFpEF from controls with an area under the curve of 0.79 (p = 0.003) and had an incremental diagnostic value over the H2FPEF score (p = 0.007). In conclusion, FABP1 could be a novel hepatic biomarker that associates with hemodynamic derangements, reduced cardiac output, and poor exercise capacity in HFpEF.


Results
Baseline characteristics. We enrolled 23 control subjects and 22 HFpEF patients in this study. Compared to control subjects, patients with HFpEF were older and had radiographic and blood-based signs of congestion, evidenced by a higher prevalence of cardiomegaly and greater N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels (Table 1). Sex, body mass index, and prevalence of comorbidities did not differ in HFpEF and control subjects. Pulmonary rales and peripheral edema were rare in both groups. The use of cardiovascular medications was similar between groups. As expected, the H 2 FPEF score was higher in patients with HFpEF than controls. Hemoglobin and creatinine levels were similar between HFpEF patients and controls.
Conventional hepatobiliary markers were on average within the normal range in patients with HFpEF and were similar between groups. However, FABP1 levels were elevated in patients with HFpEF compared to controls (Fig. 1A). The difference remained significant after adjusting for either age (p = 0.04) or any hepatobiliary enzymes (all p < 0.05). FABP1 levels were not correlated with conventional hepatobiliary enzymes (all p > 0. 15). In contrast, FABP1 levels were directly correlated with radiographic and blood biomarkers of congestion (cardiothoracic ratio, r = 0.33, p = 0.03; and NT-proBNP levels, r = 0.50, p = 0.0005, Fig. 1B). Table 1. Baseline Characteristics. Data are mean ± SD, median (interquartile range), or n (%). ACEI, angiotensin-converting enzyme inhibitors; ALP, alkaline phosphatase; ALT, alanine transaminase; ARB, angiotensin-receptor blockers; AST, aspartate transaminase; NT-proBNP, N-terminal pro-B-type natriuretic peptide; FABP, fatty acid-binding protein; HFpEF, heart failure with preserved ejection fraction; T-bilirubin, total bilirubin; and γGT, γ-glutamyl transferase.  (Fig. 2). Serum aspartate transaminase (AST) was also cor-   LV function, hemodynamics, and their relationships with FABP1 during exercise. With low level (20 W) and peak exercise, heart rate, systolic BP, and oxygen saturation were similar between groups (Table 3). Compared to control subjects, mitral E velocity, E/e' ratio, PA pressures, and RA pressure were higher and mitral e' and s' velocities were again lower in HFpEF subjects during 20 W and peak exercise (Fig. 3). Subjects with HFpEF displayed lower CO during peak exercise than controls. Levels of FABP1 at rest were correlated with poor exercise capacity as reflected by lower peak watts achieved and shorter exercise duration (r = − 0. Diagnostic performance of FABP1. As expected, The H 2 FPEF score and NT-proBNP levels demonstrated good discriminatory abilities for identifying HFpEF, with areas under the curve (AUCs) of 0.72 and 0.88 (p = 0.009 and p < 0.0001), respectively. FABP1 distinguished HFpEF from control subjects with an AUC of 0.79 (p = 0.003) whereas other hepatobiliary markers did not (Table 4). FABP1 had an incremental diagnostic value over the H 2 FPEF score (global chi-square 14.1 vs. 6.9, p = 0.007).

Sensitivity analyses.
Of the 45 participants, liver sonographic examinations that were performed within a year from exercise echocardiography were available in six patients. Of the six patients, two patients were found to have mild chronic hepatitis and one had mild fatty liver while no evidence of acute or chronic hepatitis was observed in the others. Sensitivity analyses were then performed excluding the three patients with liver diseases. We found that (1) FABP1 levels remained significantly higher in HFpEF patients than controls (46 [24,70] ng/ mL in HFpEF [n = 21] vs. 17 [13,23]

Discussion
We demonstrated, for the first time to our knowledge, the robust relationships between serum FABP1 and echocardiographic measures characterizing HFpEF. We found that FABP1 but not conventional hepatobiliary markers was significantly elevated at rest in patients with HFpEF compared to controls. Interestingly, FABP1 levels were associated with markers of congestion and alteration of parameters for systolic and diastolic reserve, biventricular filling pressures, pulmonary hypertension, and CO during exercise in HFpEF. Additionally, FABP1 levels were associated with the presence of HFpEF, with an incremental diagnostic value over the H 2 FPEF score. Given that circulating FABP1 is most exclusively derived from the liver, our data suggest that FABP1 could be a novel hepatic biomarker that associates with hemodynamic derangements, lower cardiac output, and reduced exercise capacity in HFpEF.
Potential mechanisms for FABP1 elevation in HFpEF. Biomarkers provide valuable information to understand the specific pathophysiological pathways that relate to the disease 24 . FABPs are relatively small cytoplasmic proteins (14-15 kDa) abundantly expressed in a tissue-restricted manner; therefore, in response to tissue injury, FABPs diffuse more rapidly through the interstitial space and the endothelial clefts to circulation than large proteins such as alanine transaminase (ALT) (96 kDa) or AST (90 kDa) 25 . As such, FABP1 could serve as a biomarker for an earlier phase of hepatic injury where conventional liver markers are not released. In keeping with this notion, our data showed that circulating FABP1 levels were increased in the absence of elevation of AST and ALT. Accordingly, one of the most likely mechanisms for FABP1 elevation in HFpEF patients is a response to early or minimal hepatic injury.   www.nature.com/scientificreports/ However, our simple correlation analyses revealed no significant correlation between FABP1 and AST, ALT, or the other conventional hepatic enzymes, arguing against the hepatic injury as a sole mechanism of an increase in circulating FABP1 in HFpEF. We recently found that FABP1 levels were significantly increased during exercise and were significantly correlated with plasma norepinephrine levels in healthy volunteers (manuscript in preparation). These results suggest that exercise induces circulating FABP1 through the mechanisms involving sympathetic nervous system activation. Thus, it is intriguing to speculate that an elevation of FABP1 in HFpEF is partly due to a hepatic activation of adrenergic signaling. Further studies are needed to determine the mechanisms underlying elevation in FABP1 in patients with HFpEF.

A potential link between FABP1, hepatic injury, and hemodynamic derangements during exercise in HFpEF.
HFpEF is a clinical syndrome that can be characterized by reduced cardiovascular reserve which leads to an elevation in LV filling pressure and secondary pulmonary hypertension during exercise 2,26 . The current study demonstrated correlations of FABP1 levels with radiographic and blood markers of congestion and echocardiographic evidence of hemodynamic derangements (higher E/e' ratio, PASP, and eRAP during peak exercise). The cross-sectional design of our study cannot determine whether hemodynamic derangements caused hepatic injury to promote elevations in circulating FABP1 levels, or whether FABP1 directly worsened LV diastolic function and hemodynamics during exercise. It has been shown that FABP1 is an effective endogenous cytoprotectant, minimizing hepatocyte oxidative damage 27 . The elevation of circulating FABP1 may represent a compensatory mechanism to counteract oxidative stress and inflammation in the liver 21 . Based on these data and ours, we speculate that systemic venous congestion secondary to the elevation in right heart pressures may lead to hepatocyte injury to promote up-regulation of FABP1. Further studies are warranted to determine the mechanisms underlying hepatic injury in HFpEF.
Diagnostic implications. Diagnosis of HFpEF in people presenting with chronic dyspnea is challenging 6,7,10 .
Assessments of clinical characteristics, chest radiography, echocardiography, and blood biomarkers play an important role in the diagnostic evaluation of HFpEF. Natriuretic peptides are the most commonly-used bloodbased biomarker to facilitate diagnosis of HFpEF and a recent guideline statement from the ESC has proposed a scoring system based upon echocardiographic markers of diastolic function as well as natriuretic peptides to determine whether HFpEF is present 7 . However, there are well-known limitations of natriuretic peptides, such as an underestimation in obese patients [28][29][30] . This makes the identification of novel biomarkers that relate to greater elevations of cardiac filling pressures or the presence of HFpEF high priority. It is therefore noteworthy that FABP1 levels were elevated in patients with HFpEF compared to controls, were well correlated with echocardiographic markers of elevated cardiac filling pressures and PA pressures during exercise, and predicted the presence of HFpEF. Although the diagnostic ability of FABP1 was lower than NT-proBNP, FABP1 had an incremental value to identify HFpEF from control subjects over the established H 2 FPEF score. The current data suggest that FABP1 could be a candidate biomarker to help identify HFpEF among patients with chronic dyspnea. Further large-scale studies are required to validate these findings and establish a cut-off for FABP1 levels to allow for incorporation into current diagnostic practice.
Limitations. This is a single-center study from a tertiary referral center. All participants were referred for exercise stress echocardiography for the evaluation of unexplained exertional dyspnea, introducing selection and referral bias. Although the current study and our previous one both focused on FABP1 levels in HF, the two studies are essentially different in three main perspectives: the aim, study design, and population. The primary aim of the previous study was to determine the prognostic value of FABP1 in HF 23 . In other words, the previous study was a longitudinal outcome study in design. On the other hand, the present study was a cross-sectional study to investigate whether FABP1 levels would correlate with echocardiographic markers of intracardiac pressures during supine bicycle exercise. Regarding the population, the previous study included HF patients regardless of EF and control subjects who were referred to coronary angiography. On the other hand, the present study included HFpEF patients and control subjects who were referred to exercise stress echocardiography for the evaluation of unexplained dyspnea. The pathophysiologic role of FABP1 in HF with reduced EF was beyond our scope. Patients with liver disorders were excluded from the analysis based on liver enzymes. Liver sonographic data were available in six patients, in which two had mild chronic hepatitis and one had mild fatty liver. We can- www.nature.com/scientificreports/ not exclude the possibility that some patients who did not undergo liver sonography might have had a liver disease that was not evident from liver enzymes 31 . The presence of hepatitis could have biased the results although key results remained similar even after excluding the three patients. The control group was not normal as they were referred for exercise stress echocardiography in the evaluation of exertional dyspnea and had comorbidities such as hypertension and interstitial pneumonia and relatively higher FABP1 levels than healthy controls, which could also bias the results 22 . However, the fact that the control population was more diseased than a truly normal healthy control population only biases our data toward the null. Given the presence of exertional dyspnea and comorbidity burden, control subjects might be considered as pre-HFpEF, and the inclusion of controls might add greater insight into the continuous relationships between the magnitude of FABP1 elevations and cardiac abnormalities across the spectrum from risk to frank HFpEF. We cannot conclude that these observations were specific to HFpEF, or may be observed in other disorders that cause right heart pressures, such as HF with reduced EF or non-Group II pulmonary artery hypertension. Further studies are required to address this question. The small sample size of this study does not allow to simultaneously adjust multivariable factors to analyze the diagnostic power of FABP1 to distinguish HFpEF from controls.
Conclusions. Serum FABP1 levels are elevated in early HFpEF and the magnitude of elevation is associated with echocardiographic markers of elevated LV filling pressure and PA pressures, systemic congestion, and lower workload. The present study suggests that FABP1 may serve as a potential hepatic biomarker that associates with hemodynamic perturbation, lower cardiac output, and reduced exercise capacity in HFpEF, and that FABP1 may help distinguish HFpEF among subjects with dyspnea. Further studies are required to confirm the current findings.

Material and methods
Study population. This was a cross-sectional study that assessed the association between serum FABP1 levels and Doppler echocardiographic hemodynamics at rest and during supine bicycle exercise. We prospectively enrolled consecutive subjects who were referred to our echocardiographic laboratory for exercise stress echocardiography for the evaluation of unexplained exertional dyspnea between November 2019 and June 2020.  (2) no objective evidence of elevated left heart filling pressures at rest and with exercise (criteria above). Subjects with EF < 50%, significant left-sided valvular heart disease (> moderate regurgitation, > mild stenosis), infiltrative, restrictive, or hypertrophic cardiomyopathy, non-Group II pulmonary artery hypertension or exercise-induced pulmonary hypertension without elevation of E/e' (mPAP with exercise > 30 mmHg with a total pulmonary resistance [i.e., mPAP/CO] of > 3 mmHg min/L) 33,34 , and significant liver disorders (any acute or chronic liver diseases, defined by serum levels of transaminases more than three times the upper limit of normal) were excluded. There was no overlap with our previous study focusing on the prognostic value of FABP1 in HF patients regarding the study subjects 23 . The data underlying this article will be shared on reasonable request to the corresponding author.
Biomarker measurements. Venous blood samples were obtained just before the assessment of exercise stress echocardiography. Serum FABP1 levels were measured using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (abcam, Cambridge, UK). As specified by the manufacturer, the lower limits of detection of serum FABP1 were 9.4 pg/mL. Serum NT-proBNP levels were also determined using another ELISA kit (abcam, Cambridge, UK). Serum hemoglobin, hepatobiliary enzymes, creatinine, glucose, and lipid profiles were measured by routine automated laboratory procedures.
Transthoracic echocardiography. Comprehensive resting echocardiography was performed by experienced sonographers using a commercially available ultrasound system (Vivid E95, GE Healthcare, Horten, Norway). LV volumes and EF were determined using apical 4-chamber views 35 . LV systolic function was assessed based on the EF and the systolic mitral annular tissue velocity at the septal annulus (mitral s'). LV diastolic function was assessed using the E, e' , and the E/e' ratio. Left atrial volume was determined using the biplane method of disks. Stroke volume (SV) was determined from the LV outflow dimension and pulse wave Doppler profile. CO was calculated from the product of heart rate and SV. RV systolic function was assessed using TAPSE and TV s' . RA pressure was estimated from the diameter of the inferior vena cava and its respiratory change. PASP was calculated as 4 × (peak tricuspid regurgitation [TR] velocity) 2 + estimated RA pressure. The mPAP was calculated as 0.61 × PASP + 2 36 . Subjects underwent supine cycle ergometry echocardiography, starting at 20 W for five minutes, increasing 20 W increments in three-minute stages to subject-reported exhaustion. Echocardiographic images were obtained at baseline and during all stages of exercise. All Doppler measures represent the mean of ≥ three beats. All studies were interpreted offline and in a blinded fashion by a single investigator (M.O.).