Biomarkers have become central to the management of cardiovascular disease, and their potential clinical applications continue to grow. This article reviews galectin-3 (Gal-3), a novel biomarker of fibrosis and remodeling, and its clinical application in heart failure (HF).
Case Study One
A 63-year-old African-American female presents to the emergency department with acute dyspnea. She has no prior history of HF, myocardial infarction, or coronary disease. Her risk factors are positive for diabetes and hypertension. Her renal function is normal, with a creatinine of 0.9mg/dl. She is morbidly obese, with a body mass index (BMI) of 41. Her admission brain natriuretic peptide (BNP) is in the low intermediate range at 172pg/ml, and her Gal-3 is markedly elevated at 35.2ng/ml. Assessment of her left ventricular (LV) function is performed by echocardiography and reveals an LV ejection fraction (LVEF) of 21%. Her symptoms improve and 30 days later a repeat BNP is only 89pg/ml. She is not re-hospitalized, but four months after her initial presentation is found to be deceased on follow-up contact.
Case Study Two
A 51-year-old Caucasian male with known HF, coronary artery disease, and multiple cardiac risk factors, including hypertension, diabetes, and smoking, presents with acute dyspnea. On evaluation, his BNP is markedly elevated at 1,227pg/ml; however, his Gal-3 level is low at 7.97ng/ml and his renal function is normal with a creatinine of 0.9mg/dl. A transthoracic echocardiogram is performed and reveals an LVEF of 25%. His symptoms improve and 30 days following initial evaluation his BNP decreases modestly to 963pg/ml. He remains alive at one year post-presentation.
HF is a common diagnosis, accounting for over 1.1 million discharges per year with a prevalence of 5.8 million cases in the US.1 Despite widespread adoption of guideline-based therapy, including angiotensin-converting enzyme (ACE) inhibitors and beta-blockers, there remains an upward trend for HF admissions.2 For those admitted with acute decompensated HF (ADHF), mortality in the next 30 days is six times that of patients with chronic HF with a survival of only 86%.3,4
Furthermore, re-hospitalization of HF patients is a significant factor in the utilization of healthcare resources. A recent review of Medicare claims data for re-hospitalization identified patients with discharge diagnosis of HF as the group with the highest 30-day re-hospitalization rate (27%).5 The estimated direct and indirect cost of HF in the US for 2010 is over $39 billion.1 It is clear that the management of patients with HF is a huge burden for the afflicted patients and their families, as well as for the healthcare system. Therefore, it is imperative to appropriately risk-stratify patients to identify those who need the closest follow-up. Traditional risk factors such as age, diabetes, and smoking, as well as symptom severity, can all indicate those at risk. However, these alone are often inadequate to stratify risk. As a result, novel blood-based biomarkers have received substantial attention for their ability to both objectively prognosticate patients with HF and potentially provide insight into the underlying pathophysiology of the HF. Natriuretic peptide tests (BNP and N-terminal prohormone BNP [NT-proBNP]) are now established biomarkers for the diagnosis and prognosis of those with HF, with substantial data to support these indications, and have been given a class IIa indication for prognostication by the National Academy of Clinical Biochemistry.6
Common biomarkers used in cardiovascular disease such as the natriuretic peptides, cardiac troponins, and C-reactive protein are generally elevated as a result of the disease process and do not substantively contribute to disease progression. There remains a need to identify future markers that can not only prognosticate but also directly play a role in HF progression, making them potential targets for acute or chronic intervention. Gal-3 is a prototype of such a marker (see Figure 1). Gal-3 is a galactoside-binding lectin 30kD in size that is highly expressed, is secreted by macrophages, and is a potent mitogen for fibroblasts in vitro.7 Gal-3 represents an intriguing link between inflammation and, ultimately, fibrosis—both ubiquitous findings in patients with HF.
A role for Gal-3 in the progression of HF first came to attention in a sequence of basic science and animal studies performed by Sharma et al.8 First identified by a DNA micro-array, the Gal-3 gene was the most prominently over-expressed gene in HF rat myocardium, with a greater than five-fold rise compared with compensated hypertrophied myocardium. Second, the protein was also found in the highest quantities in rat hearts with the highest degree of cardiac fibrosis. These rats also went on to develop HF.
Gal-3 was found to be co-localized with infiltrating macrophages within the myocardium. Normal rats had neither Gal-3 nor infiltrating macrophages. These findings were confirmed in human heart myocardial biopsies, with higher levels of Gal-3 in those with poorer LVEFs. Levels of Gal-3 were lowest in the myocardial biopsies of patients with aortic stenosis with normal EFs and without hypertrophy.8 To determine whether Gal-3 was directly related to fibrosis, the investigators found that intrapericardial infusion of Gal-3 into normal rat hearts resulted in decreased EF, increased heart weight, and an abundance of collagen I compared with rat hearts that had received a placebo infusion. In summary, Gal-3 has a direct effect on myocardial fibrosis and development of HF. Furthermore, its expression is seen prior to the onset of HF in hearts exhibiting hypertrophy, potentially identifying it as a marker of those at risk for progressing from American College of Cardiology (ACC)/American Heart Association (AHA) stage B (evidence of structural heart disease such as hypertrophy, but without symptoms) to stage C (symptomatic HF).9
Recently, the Gal-3 assay was refined to a high-precision enzyme-linked immunosorbent assay (ELISA) with a limit of quantitation of 0.97ng/ml (≤20% coefficient of variation [CV]) and a measurement range of 0.96–130ng/ml. Gal-3 measurement is quite robust with this assay. Both serum and ethylenediaminetetraacetic acid (EDTA) plasma can be collected, stored at room temperature for 15 days, frozen at -20oC or -70oC for six months, and undergo at least six freeze–thaws with minimal change in Gal-3 results. A CV <10% is seen at a value of ~6.0ng/ml and is ~7% at ~21ng/ml. A normal range was defined in 1,092 HF-free subjects, with the 95th percentile for Gal-3 found at 20.3ng/ml.10
The initial clinical experience with Gal-3 was in evaluating patients presenting clinically with acute dyspnea. Gal-3 and NT-proBNP levels were measured in 599 of these subjects presenting to the emergency department (ADHF was present in 35%).11 Diagnostically, Gal-3 was inferior to NT-proBNP (area under the curve 0.72 versus 0.94) for identifying ADHF from other etiologies. For prognosis, by contrast, Gal-3 was a superior predictor of death or death/recurrent HF at 60 days in a multivariate adjusted analysis of the 209 subjects with HF. Synergism was also seen between Gal-3 and NT-proBNP, with the poorest outcomes for both end-points seen in subjects with elevated levels of both markers.
Recently, the patient cohort was expanded to include multiple centers with additional patients, many of whom are African-American, recruited from urban-based centers, including the University of Maryland and affiliates in Baltimore, as well as the University of California, Davis. Gal-3 and NT-proBNP were measured during the index dyspnea presentation to the emergency department. In addition, the follow-up time for all-cause mortality was extended to four years. The results of this analysis were presented at the 2009 Heart Failure Society of America (HFSA) meeting.12 The baseline characteristics of the population are shown in Table 1 and include 310 patients. The mean age is 68.2±14.4 years, 39% are female, 69% are Caucasian, and 79% have either New York Heart Association (NYHA) class III or IV HF.
Compared with NT-proBNP, Gal-3 level appears to be less influenced by severity of symptoms (see Figure 2). However, despite only a modest relationship with presenting symptoms, Gal-3 shows a graded association with all-cause mortality. At approximately one year into follow-up, subjects with the Gal-3 levels in the lowest quartile (<12.0ng/ml) had a mortality rate of only 10.3% versus 32.5% for those with Gal-3 values in the highest quartile (>22.3ng/ml; see Figure 3). Importantly, Gal-3 levels continued to work synergistically with NT-proBNP levels for prediction of all-cause mortality in this heterogeneous cohort with extended follow-up. Kaplan-Meier curves based on receiver operator curve optimal cut-off levels for both markers show that the best prognosis was for those subjects with low levels of both markers and the poorest prognosis was in those subjects with elevated levels of both tests (see Figure 4). Importantly, an elevated Gal-3 identified subjects at increased risk despite a low NT-proBNP level. To confirm that Gal-3 levels remained independent of demographic factors, symptom status, generalized inflammation (as represented by C-reactive protein level), and NT-proBNP, a Cox multivariate analysis was performed on this expanded cohort over the duration of follow-up (see Table 2). Gal-3 was one of the most powerful predictors of mortality, with a hazard ratio of 2.18 per log unit for all-cause death with an associated χ2 of 10.5 (p=0.001) compared with NT-proBNP, and a hazard ratio of only 1.21 per log unit with an associated χ2 of 4.9 (p=0.026).
Gal-3 has also been studied in chronic stable HF populations. The Deventer–Alkmaar heart failure study (DEAL-HF) included a total of 240 patients (232 with Gal-3 measured) with stable NYHA functional class III or IV HF who were randomized between 2000 and 2003 to a disease management program or standard of care. The primary end-points were all-cause mortality and hospitalization for HF over 12 months.
Lok et al. recently investigated the prognostic value of Gal-3 in this cohort, extending an all-cause mortality follow-up for up to 6.5 years with a mean follow-up of 4.0±1.9 years. The baseline characteristics of the population by Gal-3 quartiles are shown in Table 3. Progressively higher levels of Gal-3 were associated with increased age, worse renal function, lower BMI, and higher NT-proBNP levels. Overall, Gal-3 remained an independent predictor of mortality in these stable HF patients (Gal-3 hazard ratio 1.24, 95% confidence interval 1.03–1.5 after adjustment for age, gender, renal function, and NT-proBNP).13
A second recent evaluation of chronic HF patients, the Coordinating Study Evaluating Outcomes of Advising and Counseling in Heart failure (COACH), examined another chronic HF population, and the results were presented at the 2009 HFSA. van Veldhuisen et al. measured Gal-3 in 592 patients with stable NYHA class II–IV HF with an LVEF <35% who were followed for all-cause mortality and HF admissions for approximately two years. Again, Gal-3 was found to be an independent and powerful prognostic marker in this patient population, even when adjusting for other baseline risk factors, including age, gender, natriuretic peptides, and renal function.14
Conclusions and Future Directions for Galectin-3 in Heart Failure
Gal-3 is a protein that has been directly implicated in the inflammatory process and subsequent fibrosis in multiple organ systems, including the heart. Unlike markers such as the cardiac troponins or the natriuretic peptides, there is ample evidence that Gal-3 is directly involved in the pathophysiology of cardiac injury and progression to HF, making it a potential target for future therapy. In this review specific case examples are shown and data are presented showing that Gal-3 could have an important complementary role to the established natriuretic peptide tests. Given what is known about its role in accelerating cardiac fibrosis and based on the statistical analysis from the studies outlined above, Gal-3 may ultimately be a more powerful prognosticator in the HF setting than the natriuretic peptides. Furthermore, although we are relatively early in the evaluation of Gal-3 as a cardiac biomarker, there now exists a robust ELISA that will allow testing in populations with more subtle manifestations of disease, with anticipated lower Gal-3 levels than those found in HF patients. Subsequent evaluation of Gal-3 in populations with ACC/AHA stage B (structural heart disease without symptoms) or even stage A (risk factors without structural heart disease or symptoms) may ultimately find a marker that identifies those who are at greatest risk for progressing to symptomatic HF. Lastly, and perhaps most intriguing, is the possibility that inhibition of the pro-fibrotic actions of Gal-3 may be a target for the prevention or treatment of HF. Ultimately, both of these scenarios are likely to be the subject of much future research with this new cardiac biomarker.