| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Review |
Departments of1 Pathology and Laboratory Medicine, 2 Medicine/Cardiology, 3 Physiology and Biophysics, and 4 Biochemistry and Molecular Biology, University of Louisville School of Medicine, Louisville, KY 40292.
aAddress correspondence to this author at: Department of Pathology and Laboratory Medicine, University of Louisville School of Medicine, Louisville, KY 40292. Fax 502-852-7674; e-mail rvaldes{at}louisville.edu.
 |
Abstract |
|---|
 Top Abstract Introduction and DemographicsStrategies for Developing...ConclusionReferences |
|---|
Methods: We searched Medline via Ovid for studies published during the period 1966–2003 regarding various biomarkers suggested for HF. Our review focused on developing strategies for discovering and using new biomarkers, particularly those potentially linked to pathophysiologic mechanisms. We also point out strategic advantages, limitations, and methods available for measuring each of the currently proposed markers.
Results: Biomarkers reviewed include those released from the heart during normal homeostasis (natriuretic peptides), those produced elsewhere that act on the heart (endogenous cardiotonic steroids and other hormones), and those released in response to tissue damage (inflammatory cytokines). The concept of using a combination of multiple markers based on diagnosis, prognosis, and acute vs chronic disease is also discussed. In view of recent advances in our understanding of molecular biochemical derangements observed during cardiac failure, we consider the concept of myocardial remodeling and the heart as part of an endocrine system as strategies.
Conclusion: Strategically, biomarkers linked to mechanisms involved in the etiology of HF, such as dysregulation of ion transport, seem best suited for serving as early biological markers to predict and diagnose disease, select therapy, or assess progression.
|
Introduction and Demographics |
|---|
|
TopAbstractIntroduction and DemographicsStrategies for Developing...ConclusionReferences |
|---|
pathophysiology of hf (remodeling and molecular basis)
Clinical aspects.
HF occurs when the heart cannot maintain adequate output to the peripheral tissues or can do so only at increased filling pressure (6). HF is a complex clinical syndrome manifested by signs and symptoms of low cardiac output and pulmonary and/or systemic congestion. Symptoms of HF include fatigue, cough, nocturia, exertional dyspnea, orthopnea, paroxysmal nocturnal dyspnea, and wheezing (7). Abnormalities in left ventricular function and neurohormonal regulation are major characteristics of this condition (6). The New York Heart Association (NYHA) has developed a four-tiered classification system based on the extent of physical activity required to precipitate symptoms. Class I patients have fatigue or dyspnea only with strenuous activity; class II patients have these symptoms with moderate activity. Class III patients develop symptoms with activities of daily living, whereas class IV individuals are symptomatic at rest (8). Although this classification has been used extensively to assess diagnostic and therapeutic measures for HF, it is, at its core, a subjective approach (9).
To address this issue, recently published guidelines from the American College of Cardiology and the American Heart Association have introduced a new classification scheme based on staging the progression of HF and treatment strategies (10)(11). Patients with stage A disease are at high risk of developing HF but do not exhibit any apparent structural abnormality of the heart. Stage B patients have structural abnormalities without having had symptoms. Stage C patients have structural abnormalities of the heart with current or previous symptoms of HF. Individuals with end-stage symptoms of HF who are refractory to standard treatment are considered stage D patients (11).
HF can occur secondary to either systolic dysfunction, which is characterized by reduced contractility and decreased pump function, or diastolic dysfunction, which is characterized by impaired diastolic filling and increased chamber stiffness. Patients with systolic failure have ejection fractions <45%, whereas those with diastolic failure have a preserved ejection fraction (3). Recognition of HF can pose a diagnostic challenge for primary care physicians, who often make the initial diagnosis. Patient history and physical examination may fail to provide a definitive diagnosis, and additional testing such as chest radiographic echocardiography, complete blood count, chemistries, urinalysis, electrocardiogram, and thyroid-stimulating hormone measurements are used to aid in diagnosis (3).
Molecular aspects and tissue remodeling.
One pathophysiologic hallmark of HF is pathologic remodeling of the cardiac tissue. This phenotypic transition is believed to be initiated by alterations in cardiac load after an index myocardial injury (e.g., myocardial infarction or myocarditis) and modulated by the subsequent long-term activation of various neurohormonal and autocrine/paracrine systems (12). The adaptational and remodeling responses of the heart during functional overload are depicted in Figs. 1
and 2
(13). After an initial myocardial insult (such as a large myocardial infarction), there is a reduction in cardiac output and increased myocardial load. To maintain cardiac output and systemic perfusion, an array of compensatory systems are activated, including the adrenergic nervous system, the renin–angiotensin–aldosterone system, arginine vasopressin, endothelin, inflammatory cytokines, and growth factors. Working in concert, these pathophysiologic components impart growth stimuli that engender changes in myocardial gene expression that lead to deleterious functional and structural changes at the myocyte, myocardial, and chamber levels, a process collectively referred to as left ventricular remodeling.
. Depending on the nature of the load imposed (and the degree of neurohormonal activation), the resulting hypertrophy can be attributable to sarcomeric replication in parallel (concentric hypertrophy) or in series (eccentric hypertrophy).
|
|
In addition to functional alterations, remodeling is also manifested by biochemical derangements within the myocytes. From a biochemical perspective, dysregulation of calcium-flux homeostasis is associated with the loss of contractility and inotropic reserve. The biochemical and structural elements that seemingly fail and lead to dysregulation of ion fluxes and contraction are not yet fully understood. However, recent observations, and the concept that the heart may itself be part of an endocrine axis (16)(17), suggest new strategies for developing potential biomarkers for this disease.
|
Strategies for Developing Biomarkers of HF |
|---|
|
TopAbstractIntroduction and DemographicsStrategies for Developing...ConclusionReferences |
|---|
); we also discuss the present and projected clinical usefulness in diagnosis and management. At present, the three general types of biomolecules that hold promise as biomarkers of HF are (a) B-type natriuretic peptide (BNP), which is the most recently established and is secreted from the heart in response to myocardial stretch; (b) mammalian cardiotonic steroids (ouabain- and digitalis-like factors and bufadienolides), as hormones potentially involved in the mechanism of the disease process; and (c) inflammatory cytokines, as biologically active molecules that may be elaborated by both the failing heart and by peripheral tissues in response to tissue injury and altered perfusion.
|
natriuretic peptides
The natriuretic peptides are molecules secreted by the heart in the process of maintaining normal cardiovascular homeostasis. Approximately two decades ago, A-type natriuretic peptide (ANP) was isolated from the rat atrium and was demonstrated to have vasodilating properties (22). This factor was also shown to increase renal NaCl excretion by inhibiting ion transport in the medullary collecting duct (23). A few years later, Sudoh et al. (24) isolated another peptide from porcine brain that had properties similar to those of ANP; that peptide is now referred to as BNP. Subsequent studies indicated that BNP is produced in the cardiac ventricle; its typical concentrations in blood are approximately one-sixth of those of circulating ANP (25). The C-type natriuretic peptide (CNP), which is released by endothelial cells, and urodilatin are additional members of the natriuretic peptide family; however, neither has natriuretic properties (26)(27). Fig. 3
depicts the amino acid sequences for the three major natriuretic peptides. BNP is the largest, with 32 amino acids. ANP has 28 amino acids, and CNP is the smallest, with 22 amino acids (26). These peptides have 11 amino acids identically positioned between two cysteine molecules, forming a loop through their disulfide bridge (Fig. 3
). The genes encoding for ANP and BNP in humans are localized on chromosome 1, and the chromosome that encodes for CNP is on chromosome 4. Across species, ANP and CNP are highly conserved, whereas BNP exhibits species specificity (28). Three natriuretic peptide receptors, termed NPR-A, NPR-B, and NPR-C, have been cloned (26)(27)(28). NPR-A, the most abundant receptor type, is located in adipose tissue (29), adrenal gland, lung, and aorta (30). NPR-B is produced in high concentrations in the uterus, ovary, lung, and brain, whereas NPR-C is produced in the atria and adrenal tissues (30). Renal tissue produces more NPR-C than the adrenals (29). Both ANP and BNP bind to NPR-A and elicit their natriuretic, vasodilatory, and renin inhibitory effects via the second messenger, cGMP (31). All three peptides are cleared by binding to NPR-C and subsequent endocytosis and degeneration by lysosomes, as well as by the actions of neutral endopeptidase 24.11 (28)(29)(30)(31).
|
Natriuretic peptides in normal physiology and in HF.
Before the discovery of ANP in the mammalian atrium by de Bold et al. (22), atrial cells had been suspected of having secretory functions because they have granule-containing substances (26). Injection of atrial extracts into anesthetized animals and observation of the increase in renal excretion of sodium and chloride, as well as in urinary volume, set the stage for the discovery of ANP (22). As depicted in Fig. 4A
, ANP is derived from the cleavage of a proANP1–126 peptide into N-terminal proANP1–98 and ANP99–126 in the atrial granules by a membrane protease responding to atrial stretch (26). The protease responsible for this cleavage is thought to be corin (32). Additional proteolysis of the pro-ANP1–98 fragment leads to the formation of pro-ANP79–98, pro-ANP1–30, and pro-ANP31–67. Various immunoassays for the measurement of these fragments in plasma have been reported (26)(27). The mean (SD) physiologic concentration of pro-ANP31–67 is the greatest at 1422 (790) pmol/L, followed by pro-ANP1–98 [731 (628) pmol/L], pro-ANP1–30 [708 (251) pmol/L], and ANP [5.6 (3.6) pmol/L]. BNP is also generated by proteolytic cleavage of a 132-amino acid peptide (pre-pro-BNP) in myocytes into a nonactive N-terminal fragment with 108 amino acids (pro-BNP), which is further cleaved into the N-terminal pro-BNP (NT-pro-BNP) and an active hormone with 32 amino acids (BNP; Fig. 4B
) (26)(27)(28).
|
Natriuretic peptides as biomarkers of HF.
Several studies have assessed the potential clinical utility of ANP, BNP, and their related counterparts as biomarkers in diagnosis of HF. In an early study assessing the plasma concentrations of ANP in HF, Cody et al. (33) obtained mean (SD) concentrations of plasma immunoreactive "atrial natriuretic factor" of 11 (0.9) pmol/L in 70 healthy individuals and 71 (9.9) in 31 patients with HF. Another study, involving 211 patients before cardiac catheterization, showed that the plasma C-terminal (cANP99–126) and N-terminal fragments of pro-ANP [(NT-ANP)26–55 and (NT-ANP)80–96] were independent predictors of low ejection fraction (34). In a prospective study involving 180 patients who had been evaluated for left ventricular function, the mean NT-ANP concentration (using an antibody directed to amino acids 1–25) in NYHA class I individuals (n = 70) was 243 pmol/L with a range of 27–922 pmol/L (35). The mean for this marker in 25 control individuals was 28 pmol/L. Authors concluded that plasma NT-ANP of 54 pmol/L or greater offered a sensitivity of 90% and specificity of 92% in diagnosing patients with left ventricular dysfunction who were still asymptomatic. In patients with NYHA classifications of class II and above, the clinical sensitivities and specificities for detection of HF were all >92%. The N-terminal fragment of pro-ANP has also been shown to be a useful predictor of increased rate of hospitalizations and diuretic dosage requirement in patients with HF (36). The diagnostic performance of N-terminal pro-ANP was compared with the performance of BNP and NT-pro-BNP in a clinical study involving 57 patients with left ventricular dysfunction (37). According to ROC analysis, the best diagnostic performance was by BNP, with a mean (SD) area under the curve of 0.75 (0.06) vs NT-pro-BNP [area under the curve of 0.67 (0.07)] and N-terminal pro-ANP [area under the curve of 0.69 (0.07)]. The authors of this study concluded that BNP was the best marker for diagnosing patients with impaired left ventricular ejection fraction with a diagnostic sensitivity and specificity of 73% and 77%, respectively. NT-pro-BNP followed closely with a sensitivity of 70% and specificity of 73%. N-Terminal pro-ANP exhibited lower sensitivity and specificity: 59% and 61%, respectively.
Bay et al. (38) recently evaluated whether NT-pro-BNP could differentiate between normal and reduced left ventricular ejection fraction in an unselected consecutive group of hospital patients. The negative predictive value of this biomarker in identifying patients with reduced left ventricular ejection fraction was 98%. BNP has also been shown to be a marker of diastolic abnormalities as determined by echocardiography (39). A BNP value of 62 ng/L in blood had a sensitivity of 85%, a specificity of 83%, and an accuracy [defined as the ability to detect left ventricular (LV) dysfunction, also defined as diagnosing LV diastolic dysfunction] of 84% for detecting diastolic dysfunction. The use of antibodies with different cross-reactivities to natriuretic peptides may partly account for the wide variations reported in their performance as biomarkers in HF. In patients with ventricular assist devices, a reduction in plasma BNP concentrations was indicative of functional recovery during mechanical circulatory support (40). In that study, HF patients who had been weaned off support and did not require transplantation exhibited faster decreases in their plasma BNP concentrations during the first week on the device. In fact, increased concentrations of BNP early after myocardial infarction have been associated with adverse left ventricular remodeling. The mean (SD) concentrations of BNP in serum samples collected at the time of presentation and 2 months later for 133 patients who had survived their first myocardial infarction were 629 (76) and 334 (21) ng/L, respectively. These abnormal BNP values were also significantly associated with mortality within 1 year in these patients (41). In a large trial involving 2525 individuals, patients with plasma BNP concentrations >80 ng/L a few days after the onset of ischemic symptoms were most likely to have recurrent myocardial infarction, progression of HF, or death (42). The 10-month mortality rate in these patients was at least three times higher than the rate for those with BNP values <80 ng/L. In another large prospective trial (The Breathing Not Properly Multinational Study), the utility of BNP in the emergency diagnosis of HF with preserved ejection fraction was evaluated (43). In this study, >1500 patients with acute dyspnea had been enrolled, and those with ejection fractions >45% were considered to have nonsystolic CHF. BNP measured on arrival at the hospital had a clinical sensitivity of 86%, negative predictive value of 96%, and an accuracy of 75% (at a cutoff of 100 ng/L) in diagnosing abnormal diastolic dysfunction. Therefore, both ANP and BNP are potentially useful in diagnosis and prognosis of HF. In addition, BNP also appears to be useful in assessing risk and predicting the outcome in patients with myocardial infarction and failure.
Analytical issues and limitations.
Several immunoassay designs, including RIA, enzyme immunoassay, ELISA, and IRMA, have been developed for detection and quantification of natriuretic peptides and their fragments, including ANP, pro-ANP (various fragments), BNP, and NT-pro-BNP. Currently, the Food and Drug Administration (FDA)-cleared assays for diagnosis of CHF include the point-of-care immunoassay for measuring BNP in whole blood by BioSite Corporation and the NT-pro-BNP immunoassay by Roche Diagnostics, which is to be used as an aid in diagnosis of CHF. Many other manufacturers are also making available either of these two assays on their automated laboratory analyzers. To date there has been no reported organized attempt to standardize the various assays for measuring natriuretic peptides. With BNP and NT-pro-BNP having been introduced on the testing menus of clinical laboratories, selection of a common calibrator to assure agreement among assays measuring these peptides seems appropriate. Other important issues, such as identifying which fragments to measure (e.g., biologically active vs inactive peptides), differences in clearance rates of various fragments, intracellular pools of ANP and BNP and their differential release in response to physiologic stimuli, degradation of natriuretic peptides by proteases in plasma after blood collection, and the proper sample storage conditions need to be appropriately considered (26).
The specimen appropriate for measuring BNP in the BIOSITE fluorescent immunoassay is whole blood or EDTA plasma (44). Testing should be performed within 4 h of collection; otherwise, the separated plasma should be stored at -20 °C. Clerico et al. (45) have shown that ANP and BNP concentrations in healthy adults increase with age, particularly in females. Using the recommended cutoff of 100 ng/L, the manufacturer has reported clinical sensitivities of 69–92% and specificities of 76–100% for diagnosis of HF for males and females in various age groups tested (44). One example of discrepancies related to these issues is given in the recent study by Vogeser and Jacob (46), who report that the BIOSITE Triage BNP assay gave higher results than the Shinonoria assay (Shionogi & Co) with a correlation slope of 1.52 and a y-intercept of -7.0 ng/L (n = 70; r = 0.94). The Shinonoria RIA is a sandwich-based assay format with two monoclonal antibodies, one of which is directed toward the COOH terminus and the other toward the ring structure (47). The known fragmentation of these polypeptides in blood may have advantages with regard to measuring the total BNP secreted. This has recently been proposed as a method for quantifying total BNP in plasma by use of induced enzymatic degradation before analysis and subsequent measurement of a common fragment (48). The utility of this approach remains to be determined, but preliminary studies showed some promise. Overall, the issue of standardization and whether one marker would have an advantage over the other remains open. Furthermore, an important issue to be considered involves the different units in which results are reported for natriuretic peptides. As evident above, we have kept the units reported in various articles as originally published. The literature on both BNP and NT-pro-BNP contains both pmol/L or ng/L in reporting results. It is imperative that appropriate reference intervals be used when a result is reported. It is also essential that clinical laboratory professionals, through either recommendations or standardization, make reporting units less diverse.
In summary, natriuretic peptides are the only FDA-cleared biomarkers for diagnosis of HF to date. BNP is also used as a medication for treatment of HF. Concentrations of natriuretic peptides are also increased in the plasma of patients with ischemic heart disease. Differing clinical sensitivities and specificities have been reported for the different forms of the natriuretic peptides. It appears that two FDA-cleared forms (BNP and NT-pro-BNP) have adequate clinical efficiencies in diagnosis of HF based on the above-mentioned studies; however, neither has been rigorously evaluated as a predictor of HF in symptom-free individuals.
mammalian cardiotonic steroids
The mammalian cardiotonic steroids are examples of ligands secreted by a distant gland that affect ion transport in cardiovascular tissues and potentially contribute to molecular events involved in the progression of HF. Mammalian cardiotonic steroids (cardenolides) are a family of endogenous compounds analogous in structure and function to the plant-derived digitalis or ouabain compounds or to those found in amphibians, such as the bufadienolides. In this review, we focus on digoxin-like immunoreactive factors (DLIFs) and ouabain-like factors (OLFs) as examples of endogenous ligands and potential biomarkers that are implicated in cardiovascular homeostasis. It should also be noted that the bufadienolides also show promise as endogenous modulators of natriuresis and blood pressure (49). Evidence now indicates that these factors exist as a family of compounds, the secretion of which is believed to be metabolically controlled (49)(50). Preliminary evidence suggests that both DLIFs and OLFs have chemically reduced counterparts known as dihydro-DLIF and dihydro-OLF, respectively (49)(50)(51). A recent review of these compounds as a new class of steroid hormones has recently been published (52), and we will not repeat those details here. Cardenolides are, by and large, synthesized and secreted into the circulation by the adrenal glands (DLIFs and OLFs) and the hypothalamus (OLFs). Recent studies using animal and human culture systems have shown that radioactive carbon can be incorporated in the structures of both DLIFs and OLFs and their related derivatives (49). These findings demonstrate that the compounds are not from exogenous contaminants and are indeed produced de novo in mammalian cells.
Mammalian cardenolides in cardiovascular homeostasis.
In general, the similarity of DLIFs to digitalis makes them particularly attractive as endogenous ligands regulating cardiac activity. Increased concentrations of endogenous cardenolides, both DLIF and OLF, have been reported in various pathologic conditions such as renal failure, liver failure, hypertension, and HF, as well as in physiologic conditions such as pregnancy (53). The common feature of these conditions is the state of hypervolemia. As discussed below, endogenous cardenolides exist in picomolar quantities in the plasma of healthy individuals, and their concentrations increase up to 10-fold in HF (54)(55). To be used as a biomarker for HF, the fact that these compounds are increased in hypervolemia should be considered.
The importance of these compounds in cardiac homeostasis is underscored by the well-established effect of digoxin on cardiac function. The receptor for digoxin and other cardiac glycosides is the sarcolemmal sodium pump (Na+,K+-ATPase). The failing heart is more sensitive to the actions of digitalis than the healthy heart. This phenomenon is thought to be attributable to a decrease of 
50% in the abundance and activity of sodium pump molecules found in the failing heart (56). Analogous to digoxin and ouabain, DLIF and OLF inhibit the catalytic activity of this multisubunit protein. The sodium pump maintains ion-transport homeostasis in the myocyte by transporting three Na+ ions out of and two K+ ions into the cell at the expense of ATP (57). The link between DLIF and OLF and cardiac function is believed to be through intracellular Na+ concentration-dependent effects on the bidirectional Na+-Ca2+-exchanger (51) as depicted in Fig. 5
(58). Increased intracellular Na+ concentrations, effected by glycoside-mediated (or cardenolide-mediated) inhibition of the sodium pump, lead to an increase in the intracellular Ca2+ concentration that loads the sarcoplasmic reticulum (SR). This subsequently enhances Ca2+ release in the SR and increases the force of contraction in the failing heart (51)(58). Thus, by its indirect effects on the Na+-Ca2+ exchanger and reduction in Ca2+ efflux, inhibition of the sodium pump ultimately causes an accumulation of Ca2+ in the cytoplasm and SR, which leads to higher Ca2+ concentrations in subsequent release cycles, in turn leading to increased force of contraction (58). In addition to the inhibition of the sodium pump and effects on the intracellular Na+ and Ca2+ concentrations, several alternative mechanisms for the role of cardenolides in HF have been proposed. For example, the demonstrated linkage of the sodium pump and the Ras/mitogen-activated protein kinase cascade (59) suggests a signal transduction function for this enzyme. Ouabain has also been shown to induce the genes for 
-skeletal actin and ANP, inhibit the genes expressing the 3-subunit of the sodium pump, generate reactive oxygen species, and induce Ras-dependent protein synthesis (60). Furthermore, Leenen et al. (61) have demonstrated that ouabain-like activity increased two- to threefold in the brains of hamsters and in rat models of HF. They also showed that brain OLF is involved in increasing the resting sympathetic tone characteristic of HF.
|
Mammalian cardenolides as biomarkers of HF.
In HF, a condition of volume overload exists. In this as well as in other volume-overload conditions, such as pregnancy and renal or hepatic impairment, plasma concentrations of cardenolides have been shown to be increased (53). Endogenous mammalian cardenolides are thought to be effector ligands that regulate the sodium pump as a receptor. Therefore, studies investigating mammalian cardenolides in HF have focused on measuring the concentrations of effectors (e.g., DLIF, OLF, and marinobufogenin) in the plasma of patients diagnosed with HF together with measurement of receptor capacity (i.e., the sodium pump) and variations in its concentrations in disease (62). We will review each of these in turn.
Many different antibodies have been used to measure mammalian cardenolides. Hence, the reported concentrations of these factors in plasma of healthy individuals or patients have varied widely but are directionally consistent in demonstrating increases in HF. For example, Lin et al. (54) used three separate enzyme immunoassays (one monoclonal and one polyclonal antibody for measuring OLF and another antibody for DLIF) to determine the concentrations of mammalian cardenolides in blood collected from healthy individuals. They reported that the mean (SE) physiologic concentration of OLF was 53 (5) ng/L as measured by a polyclonal antibody and 48 (7) ng/L as measured by a monoclonal antibody. The mean DLIF concentration measured by a polyclonal antibody was 312 (94) ng/L. Liu et al. (55) measured the concentration of DLIF in 50 patients diagnosed with HF and in 39 control individuals. Using RIA, they demonstrated that the mean (SE) DLIF concentration in plasma of healthy individuals was 23.3 (2.2) ng/L, 10-fold lower than the 295 (67.4) ng/L measured in patients with NYHA class IV HF. Interestingly, the mean for DLIF reported by Lin et al. (54) in healthy individuals was 312 (94) ng/L, which was 14-fold greater than the concentration reported by Liu et al. (55). This mean was approximately equal to the DLIF mean concentration reported by Liu et al. (55) in their class IV HF patients.
The circulating concentrations of OLF were measured by RIA in a study of 47 patients with asymptomatic left ventricular dysfunction (63). The mean (SD) plasma concentrations of OLF in healthy individuals was 17 (12) ng/L [29.4 (20.6) pmol/L], and the mean for patients with asymptomatic left ventricular dysfunction was 31 (15) ng/L. On the basis of these results, the authors reported that plasma concentrations of OLF increased in patients with asymptomatic left ventricular dysfunction with no symptoms (P <0.05). These data suggest that OLF could serve as an early biomarker for predicting the onset of disease (63). The mean concentration of ouabain in the plasma of patients with ejection fractions <21% was one-fourth of the mean concentration in individuals with ejection fractions >21% (64). An endogenous bufadienolide inhibitor of sodium pump, marinobufagenin, has also been shown to be increased in HF, and its plasma concentrations seem to correlate with severity of disease (65). Despite the evidence presented, it is not yet established whether the increase in plasma concentrations of endogenous inhibitors of the sodium pump in patients with HF has a causative link to the disease or is merely an epiphenomenon.
Studies on the alterations in sodium pump activity and abundance in HF have been controversial. Two families of ouabain-binding sites in rat myocardium have been described: one with a Kd of 1.8–3.2 x 10-8 mol/L and the other with a Kd of 1.0–8.0 x 10-6 mol/L. Both of these high- and low-affinity binding sites bind and release ouabain at a four- to fivefold slower rate in hypertrophied hearts than in healthy hearts (66). However, the number of high-affinity binding sites reportedly is increased in hypertrophy, whereas the low-affinity binding sites remain unchanged (67). On the other hand, Velotta et al. (68) have reported that in humans no differences in affinities of the three subunits of the sodium pump exist. They also noted no difference in the Kd values for these isoforms in failing and in nonfailing human hearts. In a recently published study, Despa et al. (69) reported that neither the sodium pump Vmax nor Km decreased in failing rabbit hearts. In another study, using a vanidate-facilitated [3H]ouabain-binding method, Bundgaard and Kjeldsen (56) showed that the physiologic sodium pump concentration in human myocardium is 700 pmol/g wet weight. They also reported that the number of sodium pumps decreased up to 89% in patients with ejection fractions <20%. However, sodium pump mRNA concentrations did not differ between healthy and failing human myocardium (62). The mean (SE) sodium pump activity in humans measured in red blood cells from NYHA class IV patients was 3.04 (0.33) mmol Pi/mg of protein, approximately one-half the activity measured in class I patients (55). Therefore, the data on humans seem to support the notion of altered sodium pump function and protein concentrations in human HF. It is possible that discordant results from the above-mentioned animal studies might be attributable to species differences.
Analytical issues and limitations.
As evidenced by the above-referenced studies, many issues need to be resolved before endogenous cardenolides can be used as biomarkers for diagnosis or prognosis of HF. Some of these issues include variations in the cross-reactivities of endogenous factors with the various antibodies used in clinical studies; the extent of protein binding of endogenous factors in plasma and whether assays differ in detecting the bound fractions; and the materials used for calibration of the immunoassays for detection and measurement of the endogenous cardenolides. All reported studies to date have used plant-derived cardenolides as counterparts to the endogenous factors in development of the antibodies and calibrators in immunoassays and as reference materials in chromatographic procedures. It is possible that differences in ratios of the various endogenous cardenolides can serve as biomarkers for early detection and diagnosis of HF. Development of assays with sufficient sensitivities (e.g., more sensitive immunoassays or mass spectrometry) may be necessary to circumvent the current challenges in measuring endogenous cardenolides.
In summary, cardiotonic steroids hold promise as biomarkers for predicting HF progression provided that assays with adequate sensitivity (picomolar range) are developed. The rationale for this is that these compounds regulate sodium pump activity, which in turn affects intracellular calcium homeostasis—an important component of remodeling in the failing heart. It has not yet been established which mammalian cardenolide would be the best candidate as a biomarker for HF or which combination in blood would best be suited as early biomarkers.
cytokines
The cytokine family of proteins includes interleukins, interferons, tumor necrosis factors, growth factors, and colony-stimulating factors (70). Cytokines are soluble proteins released by various cell types, such as leukocytes, endothelial cells, and fibroblasts, intended to modulate cell adhesion, migration, inflammation, angiogenesis, and tissue destruction or repair (70). Most cytokines are released by more than one cell type and can enhance or diminish their own release (71). Cytokines can also be produced by cardiovascular cell types, including cardiomyocytes, and impart a variety of myocardial effects that resemble the HF phenotype, including contractile depression (72), myocyte growth and induction of a fetal gene program (73), myocyte apoptosis (74), activation of matrix metalloproteinases (75), and oxidative stress (76). Inflammatory cytokines such as tumor necrosis factor- (TNF-), interleukin (IL)-1ß, IL-2, and IL-6 have been implicated in the pathophysiology of HF (77). Various clinical observations, such as cardiac decompensation, peripheral vascular responsiveness, weight loss, and anorexia in patients with chronic HF, have been attributed to the release of cytokines (78). Indeed, increased circulating concentrations of TNF-, IL-1ß, and IL-6, as well as soluble receptors for TNF- (TNFR1 and TNFR2) and IL-6 (IL-6R) have been reported in patients with HF, with plasma concentrations correlating with disease severity and prognosis (78)(79).
Mechanism of release in HF.
The sources of increased cytokine production in HF are likely multiple and include the immune system, peripheral tissues, and the failing heart itself (80). Systemic production can occur as a result of tissue hypoxia, endotoxin release by the bacteria in the gut as a result of the congestive process, or other noncardiac sources (77). Additionally, multiple studies have demonstrated that the failing heart is capable of elaborating cytokines such as TNF-, IL-1ß, and IL-6 (79)(81) and that hemodynamic load and adrenergic activation may play a role (82)(83). End-diastolic wall stress is thought to contribute to myocardial cytokine production (77), because mechanical unloading of the failing heart decreases production of TNF- (84). Furthermore, there appears to be substantial cross-talk between the ß-adrenergic system (which is chronically activated in HF) and inflammatory cytokine production. In rats chronically treated with low-dose L-isoproterenol (2.4 mg · kg-1 · day-1 for 7 days), mRNA and protein production of TNF-, IL-1ß, and IL-6 have been shown to be increased in the myocardium (83). In contrast, untreated control animals had no detectable myocardial cytokines. These cytokines were produced locally in the heart; there was no spillover into the systemic circulation. It has also been demonstrated that 12 weeks after large myocardial infarction in rats, there is substantial up-regulation of TNF-, IL-1ß, and IL-6 mRNA and protein in the remaining remodeled failing myocardium (82).
Beta blockade is now a potent tool in treating HF even in end-stage disease. Concomitant treatment with the ß-adrenergic receptor antagonist metoprolol improved myocardial function while selectively reducing TNF- and IL-1ß production, suggesting that one mechanism through which adrenergic blockade improves remodeling is by modulation of inflammatory cytokine production. Indeed, clinical studies have shown that chronic beta blockade in HF reduces mortality, reverses left ventricle remodeling, and attenuates pathologic gene activation (85). Furthermore, it has been suggested that one beneficial effect of beta-blocker treatment in HF patients can be attributable to normalization of the immune response and control of cytokine release, thereby limiting their detrimental effects on the heart (86). Thus, catecholamine-mediated cytokine release can contribute to the pathogenesis of HF.
Cytokines as biomarkers for CHF.
Regardless of the origin of their release, serum concentrations of several cytokines, such as TNF- and IL-1, IL-2, IL-6, IL-8, neopterin, and hepatocyte growth factor, have been measured in the plasma of patients with HF (86)(87)(88)(89). In this review, we focus on the clinical studies that have examined TNF- and IL-6. More than two decades ago, TNF (also known as cachetin) was first shown to cause necrosis of tumors by Carswell et al. (90). It is released as a 233-amino acid prohormone and is then cleaved to a 157-amino acid mature protein known as TNF- (86). The plasma half-life of TNF- is only 12–17 min, and it is cleared by binding to soluble TNF receptors (78). TNF- modulates the expression of many genes, including adhesion molecules, IL-1, IL-6, and inducible nitric oxide synthase (77). The effects of TNF- are concentration dependent and are mediated by autocrine, paracrine, and endocrine pathways (89). Administration of IL-18 to a mouse model of viral myocarditis has been reported to prevent the expected myocardial necrosis by decreasing the expression of TNF- responsive genes (91).
Levine et al. (92) provided the first report that serum concentrations of TNF- in 33 patients with HF were 10-fold greater than values in age-matched controls. They demonstrated that these increases were independent of renal clearance but were associated with the activation of the renin–angiotensin system. The mean (SE) plasma TNF- concentrations in patients with various heart diseases other than HF was reported to be 90 (30) ng/L compared with patients with HF [510 (260) ng/L] and those with cardiac cachexia who have reduced peripheral blood flow both at rest and with intervention [6190 (2760) ng/L] (93). In HF patients, increased TNF- concentrations have also been associated with poor functional status and exercise intolerance (94). Despite such intolerance, it has been shown that aerobic exercise training modestly decreases serum TNF- concentrations in patients with HF by a mean of 3.6 (7.2) ng/L over 12 weeks (95). Birks et al.(96) reported that in malfunctioning donor hearts, concentrations of both TNF- and IL-6 were greater than those in transplanted donor hearts that were functioning well. Mounting evidence has implicated TNF- in the pathogenesis of HF, especially in the later stages of the disease. Serum concentrations of TNF- decrease significantly (P <0.001) with improvement in cardiac function as a result of therapeutic interventions (97). In fact, many undesirable effects, such as left ventricular remodeling and dysfunction, pulmonary edema, alterations in mitochondrial metabolism, and cachexia, have been attributed to TNF- (98). It has been shown that plasma concentrations of TNF- can serve as a biomarker for both recent-onset HF (NYHA class I–III) and in end-stage patients requiring left ventricular assist device support (99).
Similar to TNF-, serum IL-6 concentrations have also been shown to be increased in patients with HF. Recent work by Raymond et al. (88) has suggested that serum IL-6 is increased with asymptomatic left ventricular dysfunction before overt clinical HF. In their study, the mean plasma concentration of IL-6 was 8.1 (2.1) ng/L in patients with asymptomatic (n = 14) left ventricular dysfunction compared with 2.8 (1.1) ng/L in controls (n = 32). This cytokine is perhaps involved in the progression of left ventricular dysfunction and could potentially serve as a marker of risk or a target for therapy (88).
In another study involving HF patients with cardiomyopathy (n = 40), the mean (SD) serum IL-6 concentration was 18 (19) ng/L (87). Although this value represents a rather large variation among the patients tested, the CV for the assay used was <6%. The IL-6 concentration corresponding to the 98th percentile of the control group was <5 ng/L. In a different study involving 18 consecutive NYHA class IV patients, a mean value of 14.7 (16.2) ng/L (range, 2.9–50.9 ng/L) was reported (100). The corresponding value for the healthy volunteers (n = 14) was 1.8 (2.0) ng/L (range, 0.3–8.3 ng/L). In the donated heart study by Birks et al. (96), a trend was noted for higher IL-6 concentrations in the hearts not used in transplantation [16.5 (2.9) ng/L] compared with 13.9 (1.6) ng/L in the hearts used. The IL-6 mRNA concentration was also more than twofold greater in the unused hearts than in the hearts actually suitable for transplantation (96). In another cohort of >140 heart transplantations, plasma concentrations of TNF- and IL-6 were shown to increase before the appearance of histologic signs of rejection (101). In particular, increased serum concentrations of IL-6 or its expression in the heart can serve as a biomarker of myocardial function. It has also been suggested that pharmacologic modulation of cytokine production in donor hearts can improve the probability that the donated hearts will be used in transplantation (96).
Analytical issues and limitations.
Many cytokines, including TNF- and IL-6, are frequently measured in immunologic research. As stated above, they are being considered as biomarkers for HF and other cardiac function-related assessments, such as donor heart use or rejection status in transplantation. Before considering cytokines as biomarkers for cardiac ailments, it is important to consider that these proteins are also released as part of the body’s defense in dealing with infections and other threatening situations. On the other hand, the severity of HF has been suggested to increase as a result of the spillover of the cytokines from a failing heart into the circulation (102). Many cytokines, including TNF- and IL-6, have been measured in the plasma of patients by various immunoassay formats, such as ELISA, enzyme immunoassay, and colorimetric techniques (87)(100)(103). In a study comparing three different commercial assays for measuring human TNF-, Kreuzer et al. (103) reported significant differences in values obtained by various assays for patient samples. They also showed that even in samples to which TNF- had been added, ELISAs by various manufacturers gave significantly different values. The presence of cytokine-binding proteins in plasma as well as autoantibodies in some individuals and the differential responses of various ELISAs to these proteins and antibodies have been suggested to be part of the underlying basis for the observed variations (103). As mentioned above, the variations in IL-6 and TNF- concentrations measured in HF patients are broad as indicated by means (SD) of 18 (19) and 24 (16) ng/L, respectively (87). Masson et al. (104) addressed the issue of intrapatient variability in plasma cytokine concentrations. They noted that between-visit agreement for many cytokines, including IL-6, was <35% during a 3-week study period. In choosing cytokines as biomarkers for HF, issues such as soluble receptor binding, autoantibodies binding, within-patient variations, and interassay differences should be addressed.
In summary, increases in the plasma concentrations of various cytokines in HF patients have been associated with the extent of the disease. The cytokines have also been shown to be produced by the hypertrophied and failing heart. Whether the increased concentrations of cytokines in HF are the result of or a direct cause of worsening HF has not been determined. However, it known that cytokines are generally increased in the late stages of HF. It is imperative to consider their role in the pathogenesis of HF, but whether their measurement could be useful to reproducibly predict or diagnose the function of a failing heart has not been shown.
other potential biomarkers
In addition to natriuretic peptides, mammalian cardenolides, and cytokines, alterations in plasma concentrations of various other analytes have been reported in HF patients.
Other markers.
Some of these include troponin I (105) and T (106)(107), adrenomedullin (108)(109), leptin (110), bradykinin (111), carnitine (112), P-selectin (113), hepatocyte growth factor (114), serum complements (115), parathyroid hormone-related protein (116), and ketone bodies (acetoacetate and ß-hydroxybutyrate) (117). The urinary dopamine concentration has also been reported to be increased in HF (118). We will briefly discuss the troponins and adrenomedullin.
Troponins.
Currently, many in vitro diagnostics manufacturers offer FDA-cleared immunoassays on their automated platforms for analysis of troponins. Troponin I and T have been established as markers of cell injury and are used for diagnosis of acute coronary syndromes. These proteins are released from myocytes as a result of cell death attributable to apoptosis or necrosis. Cardiac troponin T concentrations have been shown to be significantly increased in patients with HF [0.140 (0.439) µg/L; n = 33] compared with healthy controls [0.0002 (0.001) µg/L] (107). Cardiac troponin I has also been reported as a potential biomarker for HF (105). In HF, apoptosis is caused by many factors, including release of catecholamines, inflammatory cytokines, hypoxia, nitric oxide, and mechanical stretch (106). However, the exact mechanism underlying the release of troponins remains unclear.
Adrenomedullin.
This vasodilating protein has been shown to be increased in both experimental and clinical HF (108)(109). In a canine pacing-tachycardia HF model, the plasma concentration of adrenomedullin was increased approximately threefold over controls (109). Immunohistochemical studies of various tissues from the HF dogs showed that myocytes as well as various renal cells exhibited more staining than controls. In patients with HF, circulating concentrations of adrenomedullin increase as a result of increased secretion from failing myocardium (108). In individuals diagnosed with NYHA class III failure, the mean plasma concentration of adrenomedullin was 31.5 (3.0) ng/L, which was greater than in controls [14.4 (2.7) ng/L]. In the class IV patients, the mean concentration of this protein was 66.1 (9.4) ng/L (108). Similar to natriuretic peptides, adrenomedullin has also been evaluated as a therapeutic agent in HF. Intravenous adrenomedullin infusion in patients with HF caused natriuretic, diuretic, and vasodilatory effects (119).
|
Conclusion |
|---|
|
TopAbstractIntroduction and DemographicsStrategies for Developing...ConclusionReferences |
|---|
|
Footnotes |
|---|
, tumor necrosis factor-; and IL, interleukin. 
|
References |
|---|
|
TopAbstractIntroduction and DemographicsStrategies for Developing...ConclusionReferences |
|---|
-1 sodium pump, is a marker of congestive heart failure severity. J Hypertens 2002;20:1189-1194.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
-subunit isoforms have a similar affinity for cardiac glycosides. Am J Physiol Cell Physiol 2001;281:1336-1343.
provokes a hypertrophic growth response in adult cardiac myocytes. Circulation 1997;95:1247-1252.-induced apoptosis in cardiac myocytes. Involvement of the sphingolipid signaling cascade in cardiac cell death. J Clin Invest 1996;98:2854-2865.[Web of Science][Medline]
[Order article via Infotrieve]
and myocardial matrix metalloproteinases in heart failure: relationship to LV remodeling. Am J Physiol Heart Circ Physiol 2002;282:H1288-H1295. and atrial natriuretic peptide in marine model of viral myocarditis. Life Sci 2002;70:1225-1234.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
levels are associated with the exercise intolerance and neurohormonal activation in chronic heart failure patients. Cytokine 2001;15:80-86.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
in patients with heart failure. Am J Cardiol 2001;88:805-808.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
and interleukin-6 in myocardium and serum of malfunctioning donor hearts. Circulation 2000;102:352-358.
and interleukin-6 as rejection markers in a cohort of 142 heart-grafted patients followed by endomyocardial biopsy. Eur Heart J 1997;18:885-886.. J Immunol Methods 1996;195:49-54.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
system. Int J Cardiol 2000;76:117-122.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
The following articles in journals at HighWire Press have cited this article:
|
|
 |
|
  K. H. Miettinen, J. Lassus, V.-P. Harjola, K. Siirila-Waris, J. Melin, K. R. Punnonen, M. S. Nieminen, M. Laakso, and K. J. Peuhkurinen Prognostic role of pro- and anti-inflammatory cytokines and their polymorphisms in acute decompensated heart failure Eur J Heart Fail, April 1, 2008; 10(4): 396 - 403. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. von Haehling, E. A. Jankowska, N. G. Morgenthaler, C. Vassanelli, L. Zanolla, P. Rozentryt, G. S. Filippatos, W. Doehner, F. Koehler, J. Papassotiriou, et al. Comparison of Midregional Pro-Atrial Natriuretic Peptide With N-Terminal Pro-B-Type Natriuretic Peptide in Predicting Survival in Patients With Chronic Heart Failure J. Am. Coll. Cardiol., November 13, 2007; 50(20): 1973 - 1980. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Vassalle, M. G. Andreassi, C. Prontera, M. Fontana, L. Zyw, C. Passino, and M. Emdin Influence of ScaI and Natriuretic Peptide (NP) Clearance Receptor Polymorphisms of the NP System on NP Concentration in Chronic Heart Failure Clin. Chem., November 1, 2007; 53(11): 1886 - 1890. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. P. Goetze, R. Mogelvang, L. Maage, H. Scharling, P. Schnohr, P. Sogaard, J. F. Rehfeld, and J. S. Jensen Plasma pro-B-type natriuretic peptide in the general population: screening for left ventricular hypertrophy and systolic dysfunction Eur. Heart J., December 2, 2006; 27(24): 3004 - 3010. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Clerico, G. C. Zucchelli, A. Pilo, and M. Emdin Clinical Relevance of Biological Variation of B-Type Natriuretic Peptide Clin. Chem., May 1, 2005; 51(5): 925 - 926. [Full Text] [PDF]
|
|
|
|
|
|
F. S. Apple, M. Panteghini, J. Ravkilde, J. Mair, A. H.B. Wu, J. Tate, F. Pagani, R. H. Christenson, A. S. Jaffe, and on Behalf of the Committee on Standardization of M Quality Specifications for B-Type Natriuretic Peptide Assays Clin. Chem., March 1, 2005; 51(3): 486 - 493. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
