Complexity of molecular forms of B-type natriuretic peptide in heart failure
- Toshio Nishikimi1,
- Koichiro Kuwahara1,
- Yasuaki Nakagawa1,
- Kenji Kangawa2,
- Naoto Minamino3,
- Kazuwa Nakao1
- 1Department of Medicine and Clinical Science, Kyoto University Graduate School of Medicine, Kyoto, Japan
- 2Department of Biochemistry, National Cerebral and Cardiovascular Center Research Institute, Osaka, Japan
- 3Department of Molecular Pharmacology, National Cerebral and Cardiovascular Center Research Institute, , Osaka, Japan
- Correspondence to Professor Toshio Nishikimi, Department of Medicine and Clinical Science, Kyoto University Graduate School of Medicine, Shogoin-Kawara-cho 54, Sakyo-ku, Kyoto 606-8507, Japan;
In 1988, a Japanese group isolated B-type (or brain) natriuretic peptide (BNP) from porcine brain extracts by monitoring its relaxant effects on chick rectum.1 Since then studies in humans and rodents demonstrated that BNP is a cardiac hormone mainly expressed in the heart, where its concentration is considerably higher than in brain. BNP possesses a 17-amino acid ring structure containing two cysteine residues, which is essential for its biological activity. Mechanical stress, ischaemia, cytokines and neurohumoral factors, including angiotensin II, stimulate expression of BNP (figure 1),2 and levels of myocardial BNP mRNA and circulating BNP and N-terminal proBNP (NT-proBNP) are markedly increased in patients with congestive heart failure.2 BNP is therefore considered to function as an emergency defence against ventricular overload in disease states.
Molecular complexity of immunoreactive BNP in human plasma
It is thought that human ProBNP is most likely cleaved by furin to BNP and NT-proBNP when it is secreted.2 Once in the plasma, dipeptidyl peptidase IV removes the two N-terminal amino acids (Ser-Pro) of BNP to generate BNP[3–32],3 the levels of which are increased in patients with heart failure.4 Various BNP assay kits (eg, Shionogi, Biosite) similarly detect BNP[3–32] and BNP (figure 2).5 In addition, other aminopeptidases may further digest the N-terminal region of BNP and/or BNP[3–32], and current assay systems likely also cross-react with these forms (figure 2). Consequently, the actual molecular forms of BNP circulating in plasma remain uncertain.
An early study failed to detect native BNP in plasma from New York Heart Association (NYHA) class-IV patients using solid phase extraction combined with liquid chromatography, immunodetection and mass spectrometry, despite BNP levels having been predetermined to exceed 1000 pg/ml. This suggests that native BNP may be altered in such patients.6 Later attempts to quantify active BNP in the plasma of patients with heart failure using various extraction and detection methods, including mass spectrometry, detected BNP[3–32] in all patients along with other forms, including BNP[4–32], BNP[5–32], BNP[5–31], BNP[1–25] and BNP[1–26].7 Incomplete protease inhibition during and after blood collection may partially explain why some investigators detected no BNP in the plasma of heart failure patients. However, when the sum of all the BNP breakdown products was measured in patients using mass spectrometry, it was verified to be only a small fraction of the total immunoreactive BNP determined using the Biosite assay,7 which may be explained by antibody cross-reactivity with proBNP and related molecules (figure 2). A more recent study using mass spectrometry also confirmed the very low levels of biologically active BNP in the plasma of heart failure patients.8 Moreover, clinical measurements of BNP made using an immunoassay correlated poorly with BNP levels measured using mass spectrometry, though they correlated well with BNP degradation fragments such as BNP[3–32], BNP[4–32] and BNP[5–32] (figure 2). These results suggest that clinically measured BNP includes numerous fragments that would be expected to have little compensatory biological activity. Whether molecular forms of BNP are altered in healthy individuals as well as in heart failure patients remains unknown. This is because measuring the subfractions of BNP metabolites is difficult in healthy individuals due to the low levels of BNP present.
These findings highlight the need for more specific clinical immunoassays to address the question of atypical proBNP processing and to accurately measure bioactive BNP concentrations. Mass spectrometry is thought to be a suitable technology for such studies.
Presence of proBNP in bloodstream associated with heart failure
The clinical utility of BNP and NT-proBNP as biochemical markers of heart failure was established with BNP's original discovery as a cardiac hormone.2 However, considerable uncertainty still surrounds the molecular forms of BNP. Earlier studies showing the presence of proBNP in human blood did not garner much attention, but recent studies have shown levels of proBNP to be higher than those of BNP in the plasma of heart failure patients.9 In addition, one recent study further showed that both proBNP and BNP circulate in the plasma of heart failure patients, and that proBNP/BNP ratios vary widely depending on the heart failure status.10
All BNP assays, regardless of the source (eg, Shionogi, Biosite), cross-react with proBNP to some degree because the two antibodies used in the assays recognise epitopes common to BNP and proBNP (figure 2).5 Whether the BNP values obtained with these assays indicate BNP, proBNP or their combination (BNP+proBNP) remains unknown, and the measured increases in plasma BNP levels seen in heart failure may reflect increases in proBNP as well as BNP. Similarly, all NT-BNP assays (regardless of source) cross-react with proBNP, but react little with glycosylated proBNP, as the attached O-saccharide almost completely inhibits antibody binding to the peptide.11
In vitro studies have shown that proBNP is much less able to induce guanosine 3′, 5′-cyclic monophosphate (cGMP) production in vascular smooth muscle and endothelial cells than BNP.12 Plasma cGMP levels are increased in proportion to the severity of mild to moderate heart failure, and correlate with plasma BNP levels. However, the increases in cGMP are attenuated in patients with severe heart failure and a poor prognosis.2 The observed increase in the levels of less hormonally active proBNP in severe heart failure may explain this phenomenon.13 Indeed, one recent study showed that the proBNP/BNP ratio is increased in decompensated heart failure, and that medical therapy reduces plasma BNP and the patients’ symptoms in concert with a reduction in the proBNP/BNP ratio in some cases.10 Elucidation of the mechanism associated with the increased proBNP/BNP ratio should help to clarify the pathogenesis of heart failure and/or pave the way towards novel therapies.
ProBNP and NT-proBNP are O-glycosylated in severe heart failure
Not only do the levels of proBNP increase in heart failure, the degree to which proBNP is O-glycosylated also increases in proportion to heart failure severity.13 Thus understanding the clinical relevance of proBNP glycosylation is a matter of obvious importance. Pressure and volume overload, ischaemia and other conditions stimulate BNP gene transcription.2 BNP mRNA is translated in the endoplasmic reticulum to produce preproBNP. Subsequent removal of the signal peptide yields proBNP, which can be post-translationally glycosylated to varying degrees at several sites in its N-terminal region (Ser36, Thr37, Thr44, Thr48, Thr53, Ser58 and Thr71) while the protein is within the Golgi apparatus.14 The O-glycosylated proBNP is transported to the trans-Golgi network, where it is cleaved to BNP and NT-proBNP by furin.7 Both BNP and NT-proBNP are thought to be secreted via the constitutive secretion pathway without storage in secretory granules (figure 1).
Plasma levels of glycosylated proBNP, but not proBNP, are increased in patients with severe heart failure.12 Why glycosylated proBNP is secreted without processing under conditions of severe heart failure is not fully understood at present. One recent study showed that O-glycosylation at Thr71 in a region close to the cleavage site impairs proBNP processing by furin in HEK293 cells.15 But since the effect of O-glycosylation on furin-catalysed processing has only been evaluated in vitro, the roles of other possible processing enzymes remain unclear. In addition, whether these events occur in cardiac myocytes in the atria and/or ventricles also remains unknown. Further studies using cardiac myocytes will be required to clarify the precise mechanism of proBNP processing.
NT-proBNP is also O-glycosylated in heart failure.11 An assay for NT-proBNP from Roche Diagnostics (Elecsys I) utilises polyclonal antibodies directed against epitopes proBNP[1–21] and proBNP[39–50], and monoclonal antibodies against epitopes proBNP[27–31] and proBNP[42–46] have recently been introduced as Elecsys II. In both of these assays, glycosylation of the middle portion of NT-proBNP (Thr44, Thr48) could affect antibody binding (directed against proBNP[39–50] or proBNP[42–46]), potentially leading to an underestimation of the concentration of circulating NT-proBNP.11 For example, measured levels of plasma NT-proBNP in heart failure increase fourfold to fivefold after enzymatic deglycosylation, as compared with glycosylated NT-proBNP.11 On the other hand, numerous clinical studies have shown that the current NT-proBNP and BNP assays are equally valid for the diagnosis of heart failure. Therefore, an answer to whether new NT-proBNP assays using antibodies directed against non-glycosylated epitopes will result in better clinical applicability awaits further evaluation. Further studies will also be needed to identify the enzymes involved in the processing of non-glycosylated proBNP in the circulation and to determine how much non-glycosylated proBNP is actually present in human plasma.
Here we have briefly summarised the current understanding of the molecular forms of BNP. Several interesting issues remain to be addressed. It is therefore essential to unambiguously identify the molecular forms of BNP present in cardiac tissue and plasma. Also of importance is the relationship between the molecular forms of BNP in plasma and a patient's clinical condition. These studies may contribute not only to more accurate diagnosis of heart failure, but also to the clarification of the pathogenesis of heart failure and the development of new treatments.
We thank Ms Aoi Fujishima (Kyoto University) for her excellent technical assistance and Ms Yukari Kubo (Kyoto University) for her excellent secretarial work.
Funding This study was supported in part by Scientific Research Grants-in-Aid 20590837 and 23591041 from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to TN); a grant (AS 232Z01302F) from the Japan Science and Technology Agency (to TN); a grant from the Suzuken Memorial Foundation (to TN) and the Intramural Research Fund of National Cerebral and Cardiovascular Center (to NM).
Competing interests None.
Contributors TN mainly wrote the manuscript. YN, KK, and NM criticised it and provided the useful discussion.
KN and KK are supervisors and they provided useful comments to the manuscript.
Provenance and peer review Not commissioned; internally peer reviewed.