Galectin‐3 Predicts Left Ventricular Remodeling of Hypertension

Abstract

Previous studies have suggested that galectin‐3 is an important mediator of cardiac fibrosis. The aim of this study was to investigate the utility of galectin‐3 in identifying early left ventricular remodeling (LVRM) in patients with hypertension. A total of 107 patients with hypertension and 108 controls were enrolled in this study. The levels of galectin‐3 were significantly greater in hypertension patients with LVRM compared with those without LVRM. Multivariate regression analysis demonstrated that body mass index and galectin‐3 were independent predictors of LVRM in the hypertension group. Only left ventricular mass was independently correlated with serum galectin‐3 levels in patients with hypertension. The receiver operating characteristic analysis showed an area under the curve for galectin‐3 of 0.698 (P<.001), with an optimal cutoff of 9.43 ng/mL. Therefore, galectin‐3 is independently correlated with LVRM and can be regarded as a valuable biomarker of early cardiac remodeling of hypertension.

The rising prevalence of hypertension (HTN) is a global health problem, attributed largely to a complex interaction between genes and environmental factors.1 In 2000, the global prevalence of HTN was 26.4%, affecting an estimated 972 million people worldwide.2 Initially, HTN usually does not cause symptoms, but sustained HTN over time is a major risk factor for left ventricular (LV) remodeling (LVRM), which is characterized by hypertrophy of cardiomyocytes, proliferation of fibroblasts, and deposition of collagens.3 Fibroblasts proliferate and collagen I and III fibers accumulate within the interstitium and perivascular regions of myocardium. These events induce the development of fibrosis, which predisposes to systolic and diastolic dysfunction of the left ventricle, diminishes coronary flow reserve, and induces ventricular arrhythmia.4 Thus, recognition of HTN patients with LVRM, prior to the onset of irreversible outcomes, seems extremely urgent. Cardiac imaging, such as echocardiography and cardiac magnetic resonance imaging, is not recommended because it time‐consuming and expensive. On the contrary, biomarkers theoretically offer a convenient, objective, safe, and biological strategy to screen and identify individuals with LVRM.5

As a member of the lectin family, galectin‐3 (Gal‐3) is secreted by activated macrophages and involved in inflammation and fibrosis, which are pivotal mechanisms in the development of cardiac remodeling.6, 7 In rodent models, exogenous Gal‐3 administration has been shown to promote fibrosis and HF, and genetic or pharmacologic inhibition of Gal‐3 attenuates fibrosis and cardiac dysfunction in response to profibrotic stimuli.8, 9, 10 Several clinic studies have also demonstrated a prognostic role for Gal‐3 in acute and chronic heart failure (HF) patients for LVRM and mortality.7, 11, 12, 13 Therefore, Gal‐3 has been proposed both as a risk marker and a risk mediator, thus qualifying it as a culprit biomarker of cardiac fibrosis.14

While initial data have been reported on the association between Gal‐3 and LVRM in HF patients,7, 11, 12, 13, 15 there is currently lack of evidence on the relationship between Gal‐3 and LVRM in HTN patients. Therefore, the aim of our study was to investigate the utility of Gal‐3 in identifying early LVRM in patients with HTN. This may initiate therapy to favorably alter the course of progression to cardiac symptoms and reduce the onset of future HF.

Methods

Participants

According to the 2013 European Society of Hypertension/European Society of Cardiology guidelines,16 HTN is defined as a systolic blood pressure ≥140 mm Hg (SBP) and/or a diastolic blood pressure (DBP) ≥90 mm Hg. A total of 107 patients with HTN and 108 age‐ and sex‐matched controls were enrolled in this study. Patients with histories of diabetes mellitus, coronary artery disease, cardiac dysfunction, atrial fibrillation, pulmonary HTN, hepatic and renal failure, malignant cancer, pulmonary fibrosis, anemia, or thyroid disease were excluded from the study. The control group consisted of participants without histories of any disease and who underwent routine physical examination in outpatient clinics. The study protocol was approved by a hospital ethics committee and the investigation conformed to the principles outlined in the Declaration of Helsinki (1997). Each participant gave written informed consent.

Baseline demographic and clinical characteristics, including age, sex, pulse rate, blood pressure (BP), body mass index (BMI), and body surface area (BSA) were recorded for all participants.

Measurement of Biomarkers

After an overnight fast, blood samples were collected into tubes containing ethylenediaminetetraacetic acid. The blood was immediately processed and frozen at −80°C for later measurement of B‐type natriuretic peptide (BNP) and Gal‐3. The BNP analysis was performed with a commercially available immunoassay (Alere Inc, Waltham, MA) on an Alere Triage MeterPro analyzer according to the manufacturer's instructions. Gal‐3 was analyzed using enzyme‐linked immunosorbent assay kits (eBioscience Company, San Diego, CA) according to the manufacturer's instructions.

Echocardiography

A total of 215 participants underwent routine M‐mode and two‐dimensional echocardiography. Antero‐posterior left atrial diameter (LAD), LV posterior wall thickness (LVPWTD), interventricular septal diastolic thickness (IVSTD), LV diastolic dimension (LVDD), LV systolic dimension (LVDS), and LV ejection fraction (LVEF) were measured according to American Society of Echocardiography guidelines (2005).17 LV mass (LVM) was calculated as follows: LVM (g)=0.8×[1.04×(LVDD+IVSD+ LVPWD)³–LVDD³]+0.6.18 LVM index (LVMI) was calculated as follows: LVMI (g/m²)=LVM(g)/BSA(m²). Finally, relative wall thickness (RWT) was calculated with the formula (2×LVPWTD/LDDD). According to 2015 American Society of Echocardiography guidelines,19 LVRM has three types: concentric hypertrophy with LVMI >115 g/m² (male)/95 g/m² (female) and RWT >0.42; eccentric hypertrophy with LVMI >115 g/m² (male)/95 g/m² (female) and RWT ≤0.42; and concentric remodeling with LVMI ≤115 g/m² (male)/95 g/m² (female) and RWT >0.42.

Statistical Analysis

Categorical variables were described as frequency, and descriptive statistics were expressed as mean±standard deviation ($\overline{\mathit{\chi }}$±s). The Kolmogorov‐Smirnov test was used to determine the normality of distribution of the variables. The continuous variables that displayed normal distribution were compared with the independent Student t test, while nonparametric test was used to compare abnormally distributed continuous variables. Concentration of BNP and Gal‐3 were log‐transformed to achieve normality. Univariate and multivariate regression analyses were performed to determine the independent predictor of LVRM. Spearman's correlation analysis and linear regression analysis were performed to investigate risk factors related to levels of Gal‐3. Receiver operating characteristic (ROC) curves were used to evaluate the utility of Gal‐3 and BNP for the early diagnosis of LV remodeling in HTN patients. The optimal cutoff point for identifying LVRM was identified. Statistical analyses were performed using SPSS version 19.0 (SPSS Inc; IBM, Armonk, NY). The two‐tailed significance level was set to P<.05.

Results

Baseline Demographics and Gal‐3 Levels in the Study Population

A total of 107 patients with HTN and 108 age‐ and sex‐matched controls were enrolled in this study. Baseline demographic, clinical, laboratory, and echocardiographic parameters of the study population are listed in Table 1. BMI, BSA, IVSTD, LVPWTD, LVDD, RWT, LVM, and LVMI were significantly elevated in the HTN group compared with the control group. Adjusting for age, sex, and BMI, serum Gal‐3 levels were higher in patients with HTN vs those without (9.07±3.76 ng/mL vs 5.65±1.69 ng/mL, P<.001). Similarly, mean concentrations of BNP were significantly higher in the patients in the HTN group than in those in the control group (29.28±1.94 pg/mL vs 11.57±4.33 pg/mL, P<.001).

Table 1.

Baseline Demographic, Clinical, Laboratory, and Echocardiographic Parameters of the Study Population

Characteristic	Control Group (n=108)	HTN Group (n=107)	P Value
Age, y	58.37±5.97	57.60±11.21	.597
Male, No. (%)	57 (52.7)	58 (54.2)	.867
BMI, kg/m2	24.42±2.88	25.86±3.20	.016
BSA, m2	1.81±0.16	1.88±0.16	.022
Pulse rate, beats per min	76.5±13.5	78.4±12.7	.247
Glomerular filtration rate, mL/min/1.73 m2	88.03±24.24	79.20±21.30	.706
Cr, mg/dL	0.83±0.22	0.87±0.30	.722
LAD, mm	30.76±2.90	32.40±4.11	.09
IVSTD, mm	9.03±1.50	11.42±2.23	<.001
LVPWTD, mm	7.87±0.81	9.18±1.35	<.001
LVDS, mm	27.34±2.18	28.72±3.22	.138
LVDD, mm	43.29±2.99	44.96±3.74	.032
LVEF, %	66.08±2.05	65.75±2.69	.998
FS, %	37.53±2.35	36.50±2.44	.095
RWT	0.37±0.04	0.41±0.06	<.001
LVM, g	117.81±23.87	165.44±48.63	<.001
LVMI, g/m2	65.28±12.19	88.28±24.94	<.001
BNP, pg/mL	11.57±4.33	29.28±1.94	<.001
Gal‐3, ng/mL	5.65±1.69	9.07±3.76	<.001

Abbreviations: BMI, body mass index; BNP, B‐type natriuretic peptide; BSA, body surface area; Cr, creatinine; FS, fractional shortening; Gal‐3, galectin‐3; HTN, hypertension; IVSTD, interventricular septal diastolic thickness; LAD, left atrial dimension; LVDD, left ventricular diastolic dimension; LVDS, left ventricular systolic dimension; LVEF, left ventricular ejection fraction; LVM, left ventricular mass; LVMI, left ventricular mass index; LVPWTD, left ventricular posterior wall thickness; RWT, relative wall thickness. P<.05 was considered significant.

Gal‐3 in HTN Patients With Nonremodeling vs Remodeling of the Left Ventricle

According to the American echocardiography standard (2015) #bib107 participants with HTN were divided into an LVRM group (n=55) and an LV nonremodeling (LVNRM) group (n=52). Baseline demographic, clinical, laboratory, and echocardiographic parameters of patients in the LVRM and LVNRM groups are listed (Table 2). BMI, SBP, IVSTD, LVPWTD, LVDD, LVDS, RWT, LVM, and LVMI were significantly elevated in the LVRM group compared with the LVNRM group. Compared with the LVNRM group, Gal‐3 levels in the LVRM group were significantly elevated, after adjusting for age, sex, and BMI (12.47±2.29 ng/mL vs 7.27±2.17 ng/mL, respectively; P<.001). Nevertheless, there was no difference in serum BNP levels in patients in the LVNRM group vs those in the LVRM group (29.93±13.29 pg/mL vs 30.19±14.48 pg/mL, P=.701).

Table 2.

Comparison of Characteristics Between Patients of LVRM and LVNRM

Characteristic	LVRM Group (n=55)	LVNRM Group (n=52)	P Value
Age, y	58.04±7.66	57.09±10.32	.487
Male, No. (%)	30 (54.5)	28 (53.8)	.304
Pulse rate, beats per min	76.3±11.2	80.1±13.7	.617
BMI, kg/m2	26.73±2.46	25.27±2.07	.011
BSA, m2	1.89±0.14	1.87±0.12	.342
Systolic blood pressure, mm Hg	159.4±15.6	150.3±14.2	.034
Diastolic blood pressure, mm Hg	95.7±10.6	92.2±10.3	.082
Glomerular filtration rate, mL/min/1.73 m2	78.23±18.77	81.10±19.72	.543
Cr, mg/dL	0.85±0.20	0.88±0.34	.649
Medications at presentation, %
Angiotensin‐converting enzyme inhibitor	38	34	.643
Angiotensin receptor block	44	39	.843
β‐Blocker	16	18	.323
Calcium channel blocker	57	50	.119
Diuretic	22	27	.217
LAD, mm	33.74±4.32	31.44±3.43	.12
IVSTD, mm	13.04±1.37	10.28±1.96	<.001
LVPWTD, mm	9.85±1.32	8.45±1.65	.042
LVDS, mm	30.63±3.34	27.74±2.96	.034
LVDD, mm	46.76±3.86	43.14±4.76	.032
LVEF, %	65.08±3.78	66.45±3.93	.498
FS, %	36.04±3.12	36.83±2.17	.114
RWT	0.43±0.03	0.39±0.06	.01
LVM, g	184.36±27.36	134.64±35.38	<.001
LVMI, g/m2	95.46±15.33	76.68±14.57	<.001
BNP, pg/mL	30.19±14.48	29.93±13.29	.701
Gal‐3, ng/mL	12.47±2.29	7.27±2.17	<.001

Abbreviations: BMI, body mass index; BNP, B‐type natriuretic peptide; BSA, body surface area; Cr, creatinine; FS, fractional shortening; Gal‐3, galectin‐3; IVSTD, interventricular septal diastolic thickness; LAD, left atrial dimension; LVDD, left ventricular diastolic dimension; LVDS, left ventricular systolic dimension; LVEF, left ventricular ejection fraction; LVM, left ventricular mass; LVMI, left ventricular mass index; LVNRM, left ventricular nonremodeling; LVPWTD, left ventricular posterior wall thickness; LVRM, left ventricular remodeling; RWT, relative wall thickness. P<.05 was considered significant.

Correlation Analysis Between LVRM and Baseline Characteristics

The univariate regression analysis model showed that BMI (odds ratio [OR] #bib2.16; 95% confidence interval [CI], 1.15–3.82; P=.006), SBP (OR, 1.54; 95% CI, 1.03–2.42; P=.038), and Gal‐3 (OR, 11.53; 95% CI, 3.75–23.66; P<.001) were associated with LVRM. Multivariate regression analysis demonstrated that only BMI (OR, 1.35; 95% CI, 1.09–1.78; P=.022) and Gal‐3 (OR, 14.76; 95% CI, 5.39–27.76; P<.001) were independent predictors of LVRM in the HTN group (Table 3).

Table 3.

Univariate and Multivariate Regression Analysis Demonstrating the Relationship Between Baseline Characteristics and Presence of Left Ventricular Remodeling

	Univariate Analysis			Multivariate Analysis
	OR	95% CI	P Value	OR	95% CI	P Value
BMI	2.16	1.15–3.82	.006	1.35	1.09–1.78	.022
SBP	1.54	1.03–2.42	.038	1.29	0.72–1.99	.29
Gal‐3	11.53	3.75–23.66	<.001	14.76	5.39–27.76	<.001

Abbreviations: BMI, body mass index; CI, confidence interval; Gal‐3, galectin‐3; OR, odds ratio; SBP, systolic blood pressure. P<.05 was considered significant.

Correlation Analysis Between Gal‐3 Levels and Baseline Characteristics

In Spearman correlation analysis, age (r=0.301, P=.002), BMI (r=0.101, P=.005), SBP (r=0.225, P=.020), RWT (r=0.131, P<.001), and LVM (r=0.367, P<.001) were found to be significantly correlated with serum Gal‐3 levels in patients with HTN. Among age, BMI, SBP, RWT, and LVM, only LVM (β=0.498; 95% CI, 0.027–0.050; P<.001) was independently correlated with serum Gal‐3 levels in patients with HTN in multivariate linear regression analysis (Table 4).

Table 4.

Correlation and Linear Regression Analysis Between Baseline Characteristics and Galectin‐3

Characteristics	Spearman Correlation Analysis		Multivariate Linear Regression Analysis
	r	P value	β	95% CI	P value
Age	0.301	.002	0.267	0.030–0.149	.093
BMI	0.101	.005	0.353	0.046–0.170	.106
SBP	0.225	.020	0.342	0.073–0.167	.367
RWT	0.131	<.001	0.873	0.014–0.179	.057
LVM	0.367	<.001	0.498	0.027–0.050	<.001

Abbreviations: BMI, body mass index; CI, confidence interval; LVM, left ventricular mass; RWT, relative wall thickness; SBP, systolic blood pressure. P<.05 was considered significant.

Gal‐3 Predicts LVRM of HTN

To assess whether serum Gal‐3 and BNP levels could predict the remodeling of the left ventricle in patients with HTN, ROC curves for Gal‐3 and BNP were calculated. The ROC analysis showed an area under the curve (AUC) for Gal‐3 of 0.698 (P<.001), with an optimal cutoff of 11.43 ng/mL yielding a sensitivity of 0.677 and a specificity of 0.711. In contrast, the AUC for BNP for diagnosis of LVRM was 0.53 (P=.347). A comparison of the two ROC curves is shown in the Figure. Gal‐3 had significantly greater AUC than BNP (P<.001).

Figure 1.

Combined receiver operating characteristic (ROC) curves for B‐type natriuretic peptide (BNP) and Galectin‐3 (Gal‐3) for the prediction of left ventricular remodeling with hypertension. The ROC analysis for Gal‐3 showed an area under the curve (AUC) of 0.698 (P<.001). The ROC analysis for BNP showed an AUC of 0.53 (P=.347).

Discussion

In this study, we demonstrated that serum Gal‐3 level is independently correlated with LVRM in HTN patients. Meanwhile, ROC analysis showed that Gal‐3 could be regarded as a valuable biomarker of early cardiac remodeling with an optimal cutoff of 11.43 ng/mL. This finding may be of clinical importance because serum Gal‐3 ≥11.43 ng/mL likely indicates early LVRM and adverse prognosis in HTN patients. More attention should be paid to these high‐risk patients and appropriate treatment strategies should be used to reverse cardiac remodeling.

HTN, the most common disease in the community population, is viewed as a high‐risk factor of cardiac disease. In hypertensive patients, LVRM is frequently seen and has been considered an adaptive response to hemodynamic overload imposed by systemic HTN.20 Heterogeneity has been reported in the prevalence of LV geometric patterns among hypertensive patients. In one review, a total of 30 studies and 37 #bib700 patients were considered.21 LV hypertrophy was defined by 23 criteria, and the prevalence ranged from 36% to 41% in the pooled population. On the whole #bib27.1% and 23.7% patients were found to fulfill the more and less conservative criteria for eccentric LV hypertrophy, respectively. The corresponding numbers for concentric LV hypertrophy were 17.3% and 19.3%, respectively. Furthermore, hypertrophic myocardium shows fibrosis, which may result in HF.22 While most HF therapies are initiated at the irreversible stage with poor prognosis, identification of fibrosis prior to impairment of LV function may offer an opportunity to initiate preventive treatment at the early stage of HTN.

Gal‐3 is a 26‐kDa protein and a member of the galectin family, which are a group of carbohydrate‐binding proteins with a specific amino acid sequence that are able to recognize β‐galactosides.23 It is reported that Gal‐3 is a vital factor in fibrosis formation in different organs.24, 25, 26 When secreted by macrophages, Gal‐3 acts on fibroblasts and initiates a profibrotic program. In a study by Sharma and colleagues,7 myocardial biopsies obtained at an early stage of hypertrophy showed that expression of Gal‐3 was specifically increased in the rats that later rapidly developed HF. They also found that Gal‐3 binding was seen in cardiac fibroblasts and extracellular matrix. Recombinant Gal‐3 induced cardiac fibroblast proliferation and collagen production. In a transverse aortic constricted mice model,27 genetic disruption and pharmacologic inhibition of Gal‐3 attenuated cardiac fibrosis, LV dysfunction, and subsequent HF development. Recently, Gal‐3 has been regarded as a biomarker of fibrosis for the prognosis of heart disease. In a study with 599 patients presenting with dyspnea,28 the investigators found that Gal‐3 levels were higher in patients with acute HF and were correlated with 60‐day mortality. Hence, they demonstrated that Gal‐3 might be a novel marker for the diagnosis and prognosis of acute HF. In another study,15 240 patients with stable chronic HF were followed for 8.7 years. The results showed that Gal‐3 was positively correlated with change in LV end‐diastolic volume. In addition, Gal‐3 was a significant predictor of mortality after long‐term follow‐up. Among patients with ST‐elevation myocardial infarction infarction undergoing percutaneous coronary intervention, Gal‐3 was a strong independent predictor of 30‐day major adverse cardiac outcome.29 These collective experimental evidences suggest that, as a mediator of cardiac fibrosis, Gal‐3 may be correlated with cardiac remodeling at the early stage of HTN.

In our study, a total of 107 HTN patients without HF and reduced LVEF were enrolled. Compared with the control group, IVSTD, LVPWTD, LVDD, RWT, LVM, and LVMI were significantly elevated in the HTN group. This result indicated that these HTN patients were probably at an early stage of HTN, with cardiac hypertrophy and fibrosis but not irreversible cardiac dysfunction. Still, we found that concentrations of BNP and Gal‐3 were elevated in these early‐stage HTN patients. Next, according to the echocardiography standard #bib107 HTN participants were divided into an LVRM group and an LVNRM group. Gal‐3 levels in the LVRM group were significantly elevated compared with those in the LVNRM group, and were independently positively correlated with LVRM. In multivariate linear regression analysis, only LVM was independently correlated with serum Gal‐3 levels in patients with HTN. The ROC analysis for Gal‐3 showed that serum Gal‐3 could predict early cardiac remodeling in patients with HTN. BNP is an established serum marker for diagnosis and prognosis in acute and chronic HF.30, 31 In our study, BNP levels were elevated in HTN patients; however, there was no difference in the LVNRM and LVRM groups and they were not correlated with early cardiac remodeling in HTN patients. The reason is likely that BNP is released by the myocardium as a result of myocardial stretching in the chronic HF.32 In the early stage of HTN, active fibrosis may precede clinical manifestations of HF by many years. As a biomarker of fibrosis, it is Gal‐3 rather than BNP that is correlated with early cardiac remodeling of HTN.

We also found that BMI was an independent predictor of LVRM. This is consistent with prior findings from the Framingham Heart Study, where BMI was strongly related with LVM, LV wall thickness, and LV internal dimensions after adjusting for age and BP.33 In response to the metabolic demand of increased fat mass, obese patients exhibit increased systemic blood volume and cardiac output. The left ventricle then dilates and, for this reason, obesity has been predominantly associated with eccentric hypertrophy.34, 35 When HTN accompanies obesity, pressure overload is added and exerts an exponential effect on the prevalence of LVRM. An analysis of a cohort of 4176 hypertensive patients showed that normal‐weight, overweight, and obese patients exhibited prevalence rates of LV hypertrophy of 12% #bib25%, and 48%, respectively.35 However, we did not find BMI to be independently correlated with serum Gal‐3 level in patients with HTN in multivariate linear regression analysis.

Study Limitations

There are some limitations in our study. As mentioned above, various fibrotic conditions, such as liver cirrhosis, idiopathic lung fibrosis, and chronic pancreatitis, are associated with upregulation of Gal‐3.24, 25, 26 In order to exclude potential impact of these diseases on the final conclusion, we should have performed more intensive examinations in our study patients, not simply syndrome and histories. In addition, because of the relatively limited number of enrolled patients and single center, our results are to be considered as preliminary findings. Further studies involving a larger number of patients and multicenters are needed to confirm these results. Finally, antihypertensive drugs, especially angiotensin‐converting enzyme inhibitors and angiotensin receptor blockers, not only control BP but also reverse LVRM and delay the development of HF. It is possible that the different medications used by the patients had an effect on Gal‐3. Because we performed a cross‐sectional study, we could not determine causal relationships among Gal‐3, BP, and antihypertensive drugs. These results must be confirmed in prospective studies.

Conclusions

Our findings suggest that Gal‐3 is independently correlated with LVRM and can be regarded as a valuable biomarker of early cardiac remodeling of HTN.

Disclosure

The study was supported by the National Natural Science Foundation of China (No. 81370333).

Author Contributions

Yongwei Yao carried out the main experiment and drafted the manuscript. Dongli Shen collected and measured the biomarkers. Rong Chen finished statistical analysis. Chunyang Ying and Chenghua Wang participated in echocardiography. Junfang Guo performed physical examination of the participants. Guohui Zhang conceived the study, designed the experiments, and helped revise the manuscript. All authors read and approved the final manuscript.

J Clin Hypertens (Greenwich). 2016;18 506–511. 10.1111/jch.12757. © 2015 Wiley Periodicals, Inc.