You are here

Article Text

Article menu

Download PDFPDF

Biomarkers and heart disease
Original article
The effect of glycosylation on plasma N-terminal proBNP-76 levels in patients with heart or renal failure
  1. Toshio Nishikimi1,2,
  2. Masashi Ikeda3,
  3. Yosuke Takeda2,
  4. Toshihiko Ishimitsu2,
  5. Ikuko Shibasaki4,
  6. Hirotsugu Fukuda4,
  7. Hideyuki Kinoshita1,
  8. Yasuaki Nakagawa1,
  9. Koichiro Kuwahara1,
  10. Kazuwa Nakao1
  1. 1Department of Medicine and Clinical Science, Kyoto University Graduate School of Medicine, Kyoto, Japan
  2. 2Department of Hypertension and Cardiorenal Medicine, Dokkyo Medical University, Mibu, Tochigi, Japan
  3. 3Institute of International Education and Research, Dokkyo Medical University, Mibu, Tochigi, Japan
  4. 4Department of Cardiovascular Surgery, Dokkyo Medical University, Mibu, Tochigi, Japan
  1. Correspondence to Professor Toshio Nishikimi, Department of Medicine and Clinical Science, Kyoto University Graduate School of Medicine, 54, Shogoin-Kawara-cho, Sakyo-ku, Kyoto 606-8507, Japan; nishikim{at}kuhp.kyoto-u.ac.jp

Abstract

Objective Pro-brain natriuretic peptide (proBNP)-108 and N-terminal proBNP-76 (NT-BNP) contain seven sites for O-linked oligosaccharide attachment. Currently, levels of glycosylated NT-BNP are probably underestimated because it is not recognised by one antibody in the sandwich assay system. The pathophysiological significance of cardiac and plasma levels of non-glycosylated (nonglyNT-BNP) and glycosylated NT-BNP (glyNT-BNP) in heart failure (HF) and chronic renal failure (CRF) was investigated.

Methods Plasma samples from 186 patients with HF and 76 patients with CRF on haemodialysis were studied, together with 11 atrial tissue samples. To measure nonglyNT-BNP and glyNT-BNP, samples were incubated with or without deglycosylating enzymes and NT-BNP was measured using Roche Elecsys proBNP I. The percentage glyNT-BNP was calculated as glyNT-BNP/(glyNT-BNP + nonglyNT-BNP).

Results In HF, plasma BNP, nonglyNT-BNP and glyNT-BNP levels all increased with increasing disease severity (New York Heart Association class; p<0.0001), though the molar ratio remained constant (molar ratio, BNP:nonglyNT-BNP:glyNT-BNP = 1:2.4:9.6). Before haemodialysis for CRF, plasma BNP and nonglyNT-BNP were somewhat elevated, and glyNT-BNP was markedly increased (molar ratio, BNP:nonglyNT-BNP:glyNT-BNP = 1:8.5:82). After haemodialysis, plasma BNP, nonglyNT-BNP, atrial natriuretic protein and cGMP all declined (p<0.0001), but glyNT-BNP was unchanged. Notably, the percentage of glyNT-BNP was elevated before haemodialysis, and was further increased after haemodialysis (p<0.0001). Atrial tissue levels of BNP, nonglyNT-BNP and glyNT-BNP were similar.

Conclusion The findings suggest that most endogenous plasma NT-BNP is glycosylated and therefore undetectable with the current assay system, and that the relative glycosylation level is increased by haemodialysis.

  • N-Terminal-proBNP
  • glycosylation
  • cardiac hormones
  • heart failure
  • chronic renal failure
  • acute myocardial infarction
  • hypertension—renal
  • adrenomedullin
  • chronic heart failure
  • neurohormones
  • atrial natriuretic peptide
  • natriuretic peptides
  • heart failure treatment
  • heart failure
  • hypertrophy
  • cardiac remodelling
  • growth factors

http://dx.doi.org/10.1136/heartjnl-2011-300102

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Brain natriuretic peptide (BNP; also termed B-type natriuretic peptide) is a cardiac hormone mainly produced and secreted by the ventricles. Ventricular wall stress and/or ischaemia stimulate expression of the BNP precursor proBNP-108,1 2 which is thought to be cleaved to BNP-32 (BNP) and N-terminal proBNP-76 (NT-BNP) prior to its secretion.3 Levels of both BNP and NT-BNP are elevated in heart failure (HF) and chronic renal failure (CRF); they are used for diagnosing and evaluating the severity of HF and CRF, and are predictive of patient prognosis.4

    A recent study has shown that recombinant proBNP derived from mammalian cells has seven sites of O-linked oligosaccharide attachment within the N-terminal portion of the peptide.5 Moreover, it was also shown that proBNP is a major form of immunoreactive BNP in the plasma of patients with severe HF; that most proBNP is glycosylated6; and that NT-BNP is also glycosylated in human plasma.7 Elecsys proBNP I (Roche Diagnostics, Indianapolis, Indiana, USA) is an NT-BNP immunoassay that contains capture monoclonal and signal polyclonal antibodies that recognise NT-BNP[1–21] and NT-BNP[39–50], respectively. Notably, NT-BNP[39-50] contains glycosylation sites at amino acid residues 44 and 48, and a recent study demonstrated that O-linked oligosaccharide attachment almost completely inhibits the binding of the polyclonal signal antibody to the peptide (figure 1).8 9 For that reason, Elecsys proBNP I is thought to measure only non-glycosylated NT-BNP (nonglyNT-BNP). However, neither the amount of glycosylated NT-BNP (glyNT-BNP) present in the plasma of patients with HF or CRF nor the clinical significance of glycosylation is known at present. We therefore measured plasma nonglyNT-BNP and glyNT-BNP levels in patients with HF and patients with CRF on haemodialysis, and considered the clinical significance of both molecular forms of NT-BNP. In addition, to investigate the molecular forms of NT-BNP in the heart before secretion, we also assessed their levels in human atrial tissue.

    Figure 1
    Figure 1

    Schematic representation of nonglycosylated proBNP-108, glycosylated proBNP-108, nonglyNT-BNP and glyNT-BNP. proBNP1-108 can be post-translationally glycosylated at Thr36, Ser37, Ser44, Thr48, Ser53, Thr58 and/or Thr71 in its N-terminal region. ProBNP-108 and glycosylated proBNP-108 are cleaved to form the BNP-32 and nonglyNT-BNP or glyNT-BNP. Arrows indicate the cleavage site. The Roche NT-proBNP assay system uses two antibodies (Abs): a capture antibody targeting NT-proBNP amino acids 1–21 and a signal Ab recognising NT-proBNP amino acids 39–50. Glycosylation at Ser44 and Thr48 inhibits binding of the signal antibody. BNP, brain natriuretic protein; NT, N-terminal.

    Methods

    Informed consent was obtained from each patient; the protocol was approved by the ethics committee of our institute and was carried out in accordance with the recommendations of the ethical committee of Dokkyo Medical University.

    Study 1 (HF)

    We enrolled 186 (108 men and 78 women; age range, 37–91 years; mean age, 67±10 years) Japanese patients with HF. The clinical characteristics of these patients are shown in table 1. The primary causes of HF were valvular heart disease (n=67), ischaemic heart disease (n=52), dilated cardiomyopathy (n=28) and others (n=39). At the time of the study, the patients were being treated with angiotensin-converting enzyme inhibitors/angiotensin receptor blockers (72%), digitalis (38%) and/or diuretics (60%). The New York Heart Association (NYHA) functional classes were as follows: class I (n=82), mean age, 65±11 years; 50 men and 32 women; class II (n=82), 69±11 years; 46 men and 36 women; and class III–IV (n=22), 68±11 years; 12 men and 10 women. All groups had similar creatinine values. A patient with a creatinine value >2.5 mg/dl was excluded from the study.

    View this table:
    Table 1

    Clinical characteristics of patients with heart failure in study 1

    Study 2 (CRF on haemodialysis)

    We enrolled 76 (35 men and 41 women; age range, 32–84 years; mean age, 59±11 years) Japanese patients with end-stage renal failure who had been receiving haemodialysis for >6 months. All patients underwent regular haemodialysis for 3.5 h three times each week. The clinical characteristics of these patients are shown in table 2. The aetiologies of their CRF included chronic glomerulonephritis (n=35), diabetes mellitus (n=25), polycystic kidney disease (n=5), hydronephrosis (n=3), hypertensive nephrosclerosis (n=2) and unknown (n=6). These patients had a history of coronary artery disease (n=8), HF (n=6), vascular disease (n=5) and stroke (n=5). Cardiac function, heart size and blood pressure were relatively well controlled.

    View this table:
    Table 2

    Clinical characteristics of patients with chronic renal filure on haemodialysis in study 2

    Study 3

    In study 3, we assessed levels of BNP, nonglyNT-BNP and glyNT-BNP in atrial tissue from 11 other patients with HF who were in hospital to undergo cardiac surgery. The clinical characteristics of the patients who offered atrial tissue samples are presented in table 3. Resected samples of left atrial tissues from the 11 patients were frozen in liquid nitrogen and stored at −80°C.

    View this table:
    Table 3

    Clinical characteristics of patients with heart failure providing left atrial tissue

    Twelve volunteers with no obvious disease (6 male and 6 female; 50±10 years) served as healthy controls.

    Blood sampling

    Blood (3 ml) was withdrawn via the antecubital vein during study 1 to measure glyNT-BNP, nonglyNT-BNP and BNP, and via the shunt during study 2 before and after haemodialysis to measure glyNT-BNP, nonglyNT-BNP, BNP, atrial natriuretic peptide (ANP) and cGMP. Blood samples were transferred to chilled glass tubes containing disodium EDTA (1 mg/ml) and aprotinin (500 U/ml) and immediately centrifuged at 4°C, after which the plasma was frozen and stored at −80°C until used.

    Assays of plasma ANP, BNP and cGMP levels

    Plasma levels of BNP were measured using a fluorescent immunoenzyme assay (TOSO, Tokyo, Japan) as described previously.10 Levels of ANP were measured using a specific immunoradiometric assay,11 while levels of cGMP were measured using a radioimmunoassay,12 as described previously.

    Measurement of plasma glyNT-BNP and nonglyNT-BNP in HF and CRF

    ProBNP-108 is post-translationally glycosylated to varying degrees at Thr36, Ser37, Ser44, Thr48, Ser53, Thr58 and Thr71 in its N-terminal region,5 and NT-BNP is also glycosylated.7 The Elecsys proBNP I is comprised of a capture monoclonal antibody that recognises NT-BNP[1–21] and a polyclonal signal antibody that recognises NT-BNP[39–50], which has glycosylation sites at amino acid residues 44 and 48 (figure 1).9 Notably, O-linked oligosaccharide attachment almost completely inhibits the binding of the signal antibody to the peptide.8 9 We therefore postulated that NT-BNP measured using Elecsys proBNP I is, in fact, nonglyNT-BNP. To measure total NT-BNP, 250 μl of plasma were diluted with 250 μl of 50 mM phosphate buffer (pH 6.0), after which portions of the diluted plasma (228 μl) were added to 12 μl of phosphate buffer, with or without a cocktail of deglycosylating enzymes, and incubated for 24 h at 37°C, as described.5 7 The enzyme cocktail included O-glycosidase (Roche Diagnostics), neuraminidase (Roche Diagnostics), β-N-acetylglucosaminidase (Sigma, St Louis, Missouri, USA) and β-galactosidase (Sigma) at final concentrations of 6.25, 6.25, 62.5 and 625 mU/ml, respectively. O-Glycosidase and neuraminidase were essential for the deglycosylation, and the enzyme concentrations and incubation period were selected based on the results of preliminary and previously reported studies.5 7 Our preliminary study showed that the enzyme reaction reaches a plateau under the conditions outlined above (figure 2A). The NT-BNP levels were then measured using Elecsys proBNP I, and the value obtained after deglycosylation was postulated to be the total NT-BNP level. The glyNT-BNP was then calculated as total NT-BNP–nonglyNT-BNP, and the percentage glyNT-BNP was calculated as glyNT-BNP/(glyNT-BNP + nonglyNT-BNP) ×100%.

    Figure 2
    Figure 2

    (A) Time course of N-terminal brain natriuretic peptide (NT-BNP) levels, with and without enzymatic deglycosylation. Diluted pooled plasma (n=10) was deglycosylated using 1.25, 6.25 or 12.5 mU/ml O-glycosidase and neuraminidase for 2, 4, 16 or 24 h. (B) Gel filtration high-performance liquid chromatography (HPLC) of NT-BNP from plasma extracts from patients with heart failure (HF), before and after deglycosylation, shows one peak corresponding to nonglyNT-BNP (arrow). (C) Gel filtration HPLC of NT-BNP and BNP from plasma extracts from patients with HF, before and after deglycosylation, shows three peaks corresponding to proBNP-108 (circles, left), nonglyNT-BNP (triangles and squares, centre) and BNP-32 (circles, right). Note that there is no NT-BNP peak corresponding to proBNP-108. A peak corresponding to NT-BNP increased after deglycosylation (squares, centre). Arrows: 1, proBNP-108; 2, nonglyNT-BNP; and 3, BNP-32.

    We also used a previously reported method10 to determine the gel filtration profile of NT-BNP extracted from pooled plasma samples from patients with HF (n=10). In brief, we characterised the molecular forms of NT-BNP in peptide fractions extracted from human plasma using Sep-Pak C18 cartridge condensation. Gel filtration high-performance liquid chromatography (HPLC) was carried out using a TSK gel G2000SWXL column (TOSO). As shown in figure 2B, one peak with NT-BNP immunoreactivity and corresponding to the elution position of the recombinant nonglyNT-BNP (Hytest, Turku, Finland) was increased about fourfold after enzymatic deglycosylation. We also analysed the degree to which this NT-BNP assay system cross-reacted with proBNP using the same method. Using a fluorescence immunoassay system for BNP (TOSO), we measured NT-BNP and BNP immunoreactivity in the same samples and obtained a single peak for NT-BNP (figure 2C, triangles), which was increased about fourfold after enzymatic deglycosylation (figure 2C, squares). In contrast, two peaks were obtained for immunoreactive BNP (figure 2C, circles): the first corresponded to proBNP-108, and the second to BNP-32. There was no obvious immunoreactive NT-BNP peak in the position corresponding to proBNP-108, indicating that the Elecsys assay system had little cross-reactivity with proBNP (<1%) (figure 2C). The deglycosylation procedure did not affect BNP levels.

    Measurement of glyNT-BNP and nonglyNT-BNP in cardiac tissue

    Samples of atrial tissue were boiled in 10 volumes of 1 mol/l acetic acid as described previously.10 The tissue was then homogenised using a Polytron mixer and centrifuged, first at 3000 g and then at 15 000 g, for 15 min each at 4°C. The supernatant was extracted using a Sep-Pak C18 cartridge as described above for plasma. The eluate was lyophilised, dissolved in 30% acetonitrile containing 0.1% trifluoroacetic acid, and subjected to gel filtration HPLC on a TSK gel G2000SWXL column. Some fractions were subjected to the fluorescence immunoassay for BNP-32 and Elecsys II for NT-BNP in the same manner as the plasma samples. Others were dissolved in phosphate buffer; nonglyNT-BNP and glyNT-BNP were measured as described for plasma.

    Echocardiographic measurements

    In studies 1 and 2, an experienced echocardiographer without knowledge of the clinical features of the patients performed the echocardiographic study using a cardiac ultrasound unit (Sonos 5500; Philips Medical Systems, Andover, Massachusetts, USA).13 Aortic diameter, left atrial diameter, interventricular thickness, posterior wall thickness, left ventricular end-diastolic diameter and left ventricular end-systolic diameter were all measured. Fractional shortening (FS), left ventricular mass index (LVMI) and left ventricular ejection fraction (LVEF) were calculated using standard formulae according to the recommendations of the American Society of Echocardiography.14

    Statistical analysis

    All values are expressed as means±SD. The statistical significance of differences between two groups was evaluated using the Fisher exact test or paired Student t test, as appropriate. The distribution of plasma peptide levels was normalised by log transformation, when appropriate. Variables were compared among three groups using one-way analysis of variance followed by the Bonferroni multiple comparison test. Correlation coefficients were calculated using linear regression analysis. Values of p<0.05 were considered significant.

    Results

    Study 1

    Plasma levels of BNP, nonglyNT-BNP and glyNT-BNP in patients with HF are shown in figure 3A. BNP was elevated to 106±72 pg/ml in NYHA class I (normal range <18.4 pg/ml), was increased to a significantly greater degree in NYHA class II, and was even higher in class III–IV. Levels of nonglyNT-BNP also increased with the HF severity, and were 5.6±4.3, 5.6±3.4 and 6.2±3.4 times higher than the BNP levels in NYHA classes I, II and III–IV, respectively. Similarly, levels of glyNT-BNP were increased to 34.7±27.7, 33.5±16.7 and 30.1±27.1 times higher than the BNP levels in the respective NYHA classes (figure 3A).

    Figure 3
    Figure 3

    (A) Plasma levels of brain natriuretic peptide (BNP), non-glycosylated N-teminal BNP (nonglyNT-BNP) and glyNT-BNP in patients with heart failure (HF) in New York Heart Association (NYHA) functional classes I, II and III–IV. *p<0.01 vs NYHA class I; ¶p<0.01 vs NYHA class II; †p<0.01 vs BNP; ‡p<0.01 vs nonglyNT-BNP. (B–D) Relationships between plasma levels of glyNT-BNP and nonglyNT-BNP (B), nonglyNT-BNP and BNP (C), and glyNT-BNP and BNP (D) in patients with HF.

    The percentage glycosylation in NYHA class I HF was 81±5%, suggesting that Elecsys proBNP I measures only ∼20% of the total NT-BNP in this group of patients with HF. The percentage glycosylation did not differ significantly among the HF groups (figure 4), suggesting that glycosylation remains constant, irrespective of the severity of HF. On the other hand, the percentage glycosylation in patients with HF in any NYHA class was lower than in patients with CRF on haemodialysis (figure 4).

    Figure 4
    Figure 4

    Percentage glycosylation in healthy individuals, patients with HF in New York Heart Association (NYHA) classes I, II and III–IV, and patients with chronic renal failure on haemodialysis. *p<0.01 vs healthy individuals; ¶p<0.01 vs NYHA class I; †p<0.01 vs NYHA class II; ‡p<0.01 vs NYHA class III–IV.

    Relationships between glyNT-BNP and ANP, BNP and nonglyNT-BNP

    The relationships between glyNT-BNP and nonglyNT-BNP and BNP are shown in figure 3B,C. Plasma glyNT-BNP levels closely correlated with plasma nonglyNT-BNP and BNP in HF patients (nonglyNT-BNP, r=0.931; BNP, r=0.755; all p<0.0001). In addition, the nonglyNT-BNP and BNP levels were also significantly correlated (r=0.626, all p<0.0001, figure 3D).

    Relationships of glyNT-BNP, nonglyNT-BNP and BNP with LVEF, FS and LVMI

    The relationships of plasma levels of glyNT-BNP, nonglyNT-BNP and BNP with LVEF, FS and LVMI are shown in table 4. BNP, glyNT-BNP and nonglyNT-BNP levels significantly correlated with both LVEF and FS to similar degrees, but not with LVMI (table 4). In addition, the percentage glycosylation did not correlate with either the EF (r=0.06, NS) or creatinine (r=0.07, NS).

    View this table:
    Table 4

    Correlation coefficients and p values for patients with heart failure

    Study 2

    Plasma concentrations of BNP, glyNT-BNP, nonglyNT-BNP, ANP and cGMP in CRF before and after haemodialysis

    Plasma levels of BNP, nonglyNT-BNP and glyNT-BNP in patients with CRF before haemodialysis are shown in figure 5A. BNP was increased to 367±365 pg/ml (normal range <18.4 pg/ml), while nonglyNT-BNP was increased to 7455±8348 pg/ml, or ∼20 times higher than the BNP level. glyNT-BNP was markedly increased to 69 122±69 554 pg/ml, or ∼200 times higher than the BNP level and nine times higher than the nonglyNT-BNP level.

    Figure 5
    Figure 5

    (A) Plasma levels of brain natriureic peptide (BNP), non-glycosylated N-terminal BNP (nonglyNT-BNP) and glyNT-BNP in patients with chronic renal failure (CRF) on haemodialysis. *p<0.01 vs BNP; †p<0.01 vs nonglyNT-BNP. (B–D) Relationships between plasma levels of glyNT-BNP and nonglyNT-BNP (B), glyNT-BNP and BNP (C) and nonglyNT-BNP and BNP (D) in patients with CRF on haemodialysis.

    After haemodialysis, plasma levels of BNP, nonglyNT-BNP, ANP and cGMP were all significantly reduced (p<0.0001) (figure 6A–E), whereas levels of glyNT-BNP were unchanged (p=0.21) (figure 6F).

    Figure 6
    Figure 6

    Plasma levels of brain natriureic peptide (BNP) (A), non-glycosylated N-terminal BNP (nonglyNT-BNP) (B), glyNT-BNP (C), atrial natriuretic peptide (ANP) (D), cGMP (E) and the percentage glycosylation (F) in patients with chronic renal failure before and after haemodialysis.

    The percentage glycosylation before haemodialysis was 91.1±2.2% (figure 4), suggesting that only ∼9% of total NT-BNP was detected using the Elecsys proBNP I assay system. Notably, the percentage glycosylation was significantly increased after haemodialysis (p<0.0001, figure 6F).

    Relationships among glyNT-BNP, BNP and nonglyNT-BNP

    The relationships among the plasma levels of glyNT-BNP, nonglyNT-BNP and BNP in patients with CRF on haemodialysis are shown in figure 5B–D. glyNT-BNP levels closely correlated with nonglyNT-BNP and BNP in patients with CRF (nonglyNT-BNP, r=0.958; BNP, r=0.891; all p<0.0001) (figure 5B,C). In addition, levels of nonglyNT-BNP and BNP were also significantly correlated (r=0.845, all p<0.0001, figure 5D).

    Relationships between cGMP and glyNT-BNP, ANP, BNP and nonglyNT-BNP

    Plasma levels of cGMP, a second messenger of ANP and BNP, significantly correlated with ANP and BNP levels and weakly correlated with levels of nonglyNT-BNP and glyNT-BNP (figure 7A–D).

    Figure 7
    Figure 7

    Relationships between plasma levels of cGMP and brain natriuretic peptide (BNP) (A), atrial natriuretic peptide (ANP) (B), non-glycosylated N-terminal BNP (nonglyNT-BNP) (C) and glyNT-BNP (D) in patients with chronic renal failure on haemodialysis.

    Relationships of glyNT-BNP, nonglyNT-BNP and BNP with LVEF, FS and LVMI

    Plasma BNP is a known marker of left ventricular function and hypertrophy, even in patients with CRF on haemodialysis. Consistent with that, we found that BNP, glyNT-BNP and nonglyNT-BNP all significantly correlated with LVEF, FS and LVMI to similar degrees (table 5).

    View this table:
    Table 5

    Correlation coefficients and p values for patients with chronic renal failure

    Study 3

    In our gel filtration HPLC analysis of atrial extracts from patients with HF, we observed two peaks in NT-BNP immunoreactivity (figure 8A). The larger peak corresponded to the elution point of nonglyNT-BNP, while the smaller peak corresponded to the elution point of proBNP-108. After enzymatic deglycosylation, the amplitude of the larger peak was nearly doubled, suggesting that glyNT-BNP and nonglyNT-BNP were present at similar concentrations in the atrial tissue. In contrast, there was no remarkable change in the other peak. Two peaks of BNP immunoreactivity were also observed. The larger peak corresponded to BNP-32, while the smaller peak corresponded to proBNP-108. The molar ratio of BNP to nonglyNT-BNP before the enzymatic deglycosylation procedure was 1:0.76, while the molar ratio of BNP to total NT-BNP after deglycosylation was 1:1.42. Deglycosylation had no effect on the BNP peaks. In atrial tissue, the cross-reactivity of NT-BNP with proBNP-108 was 11±3% before deglycosylation. Consistent with the gel filtration HPLC analysis, we found that the atrial concentrations of BNP, nonglyNT-BNP and glyNT-BNP were comparable (figure 8B).

    Figure 8
    Figure 8

    (A) Using tissue samples from patients with heart failure (HF), gel filtration high-performance liquid chromatography (HPLC) of N-terminal brain natriuretic peptide (NT-BNP) and BNP in atrial extracts before and after deglycosylation showed three peaks corresponding to proBNP-108 (circles, left), non-glycosylated NT-BNP (nonglyNT-BNP; triangles and squares, centre) and BNP-32 (circles, right). Note that there is a small peak corresponding to proBNP-108. The peak corresponding to NT-BNP increased after deglycosylation (squares, centre). Arrows: 1, proBNP-108; 2, nonglyNT-BNP; 3, BNP-32. (B) Atrial levels of BNP, nonglyNT-BNP and glyNT-BNP in patients with HF. (C–E) Relationships between atrial levels of glyNT-BNP and nonglyNT-BNP (C), glyNT-BNP and BNP (D), and nonglyNT-BNP and BNP (E) in patients with HF.

    The relationships between glyNT-BNP and nonglyNT-BNP and BNP in atrial tissue are shown in figure 8C,D, respectively. Atrial glyNT-BNP levels closely correlated with the levels of nonglyNT-BNP and BNP (nonglyNT-BNP, r=0.611; BNP, r=0.829; all p<0.01) (figure 8C,D). In addition, the nonglyNT-BNP levels also significantly correlated with the BNP levels (r=0.749, all p<0.0001, figure 8E).

    Discussion

    Within the heart, proBNP-108 is normally cleaved to form BNP-32 and NT-BNP, which are then secreted into the circulation,1–4 but recent studies have shown that levels of unprocessed proBNP-108 are increased in HF.15 16 Moreover, the proBNP-108 to BNP-32 ratio is increased to a greater degree in HF with ventricular overload than with atrial overload.10 In addition, proBNP-108 is O-glycosylated at seven sites in the N-terminal portion of the peptide,5–7 which suggests that a significant amount of glyNT-BNP is circulating in the plasma of patients with HF or CRF, though up to now the levels of glyNT-BNP in patients with HF and patients with CRF on haemodialysis were unknown. In the present study, we measured immunoreactive NT-BNP levels in the plasma of these patients using the Elecsys proBNP I, with and without enzymatic deglycosylation. The signal antibody in this assay system recognises the midportion of NT-BNP, but the O-linked oligosaccharide attachments at amino acids 44 and 48 almost completely blocked the binding of this antibody (figure 1).5 7 Our observation that plasma NT-BNP levels appeared to increase for 16–24 h during enzymatic deglycosylation until finally reaching a plateau indicates that NT-BNP was completely deglycosylated at that point.

    Atrial molar levels of BNP and total NT-BNP were similar, suggesting that BNP and total NT-BNP are secreted from the heart in equimolar fashion. In contrast, nonglyNT-BNP levels were 2.5 times higher than the BNP levels in plasma from patients with HF, which is consistent with earlier reports.17 18 Plasma BNP is cleared from the blood via several pathways, including binding to natriuretic peptide receptors-A and -C, excretion in the urine, and metabolism by neutral enzymes and/or unknown proteases in the blood. In contrast, clearance of nonglyNT-BNP largely depends on renal excretion.19 As a result, nonglyNT-BNP has a longer half-life in blood than BNP (∼120 min vs 20 min),3 which probably contributes to the higher plasma nonglyNT-BNP levels.

    In the present study, we initially found that plasma glyNT-BNP was four times higher than nonglyNT-BNP in patients with HF, and that glyNT-BNP levels increased in proportion to disease severity, as was also seen with nonglyNT-BNP and BNP. In addition, the molar ratio of nonglyNT-BNP to glyNT-BNP in plasma was constant and independent of HF severity, suggesting that the half-life of glyNT-BNP in blood is longer than that of nonglyNT-BNP. Post-translational modifications play an important role in protein biosynthesis, stability and biological activity, and glycosylation is one of the most common and diverse post-translational modifications.20 21 Carbohydrate moieties in proteins are known to be important for normal cellular function, enzyme activity and protein–protein interactions.20–22 A recent study by Jiang et al23 showed that proBNP contains a substantial number of O-glycans, which are terminally sialylated. In addition, they showed that the O-glycan inhibitor Ben-gal shortened the half-life of proBNP in the medium of HEK-293 and HL-1 cells.23 We observed that plasma levels of glyNT-BNP were higher than those of nonglyNT-BNP, although the levels were similar in atrial lysates, which is in good agreement with the findings of Jiang et al. Thus, attachment of O-glycans to NT-BNP may increase its resistance to clearance from the blood.

    Previous studies showed that plasma nonglyNT-BNP levels are increased in proportion to the reduction of estimated glomerular filtration.24 25 We also observed that nonglyNT-BNP levels were ∼20 times higher than the BNP levels in patients with CRF on haemodialysis, which is consistent with earlier findings.26 27 Moreover, glyNT-BNP levels were ∼10 times higher than the nonglyNT-BNP levels in those patients and, in contrast to other peptides and substances,28 glyNT-BNP levels did not decline after haemodialysis. Consequently, the percentage glycosylation increased after haemodialysis to >90%, suggesting that >90% of total NT-BNP is invisible to the assay and therefore underestimated. Although the exact mechanism underlying the accumulation of glyNT-BNP in patients with CRF on haemodialysis remains unknown, we would speculate that attachment of O-glycans to NT-BNP probably inhibits its removal during haemodialysis.

    The clinical utility of measuring the extremely elevated glyNT-BNP remains unknown, at present. BNP and nonglyNT-BNP have been used as markers of cardiac function for diagnosing HF, and as prognostic indicators of cardiovascular disease, even in patients with CRF on haemodialysis.26 29 30 Indeed, in the present study, plasma BNP levels correlated significantly with cardiac function (FS and LVEF) and left ventricular mass (LVMI), which is consistent with earlier findings. In addition, we found that cardiac function and structure correlated similarly with glyNT-BNP and nonglyNT-BNP, which suggests that glyNT-BNP may be as useful a marker of cardiac function and/or left ventricular mass as BNP or nonglyNT-BNP.

    We observed modest correlations between cGMP and ANP and BNP in CRF, but a weaker correlation between cGMP and glyNT-BNP (figure 7). Although BNP and glyNT-BNP are encoded by the same gene, their behaviour and metabolism in blood differ somewhat. Plasma levels of glyNT-BNP change more slowly than BNP levels due to its longer half-life, and its molar levels are 80 times higher than those of BNP in patients with CRF on haemodialysis. These pharmacodynamics suggest that accumulation of glyNT-BNP in patients with CRF on haemodialysis reflects the integrated wall stress in the heart, rather than momentary stress. This raises the possibility that glyNT-BNP would be a better prognostic marker than BNP or nonglyNT-BNP in HF and/or CRF, but that remains to be tested.

    Taken together, we found that plasma levels of undetectable glyNT-BNP are increased in patients with HF and patients with CRF on haemodialysis. This remarkable increase in glyNT-BNP is likely to be caused by diminished metabolism. A schematic diagram depicting a model of the production, secretion and metabolism of nonglyNT-BNP, glyNT-BNP and BNP suggested by our findings is shown in figure 9. Whether glyNT-BNP levels are a better marker than BNP or nonglyNT-BNP for diagnosis and more predictive of prognosis in HF and CRF will be determined in future studies.

    Figure 9
    Figure 9

    Schematic representation of a model of the production, secretion and metabolism of non-glycosylated N-terminal brain natriuretic peptide (nonglyNT-BNP), glyNT-BNP and BNP-32. Non-glycosylated proBNP-108 and glycosylated proBNP-108 are produced in myocytes and cleaved to nonglyNT-BNP, glyNT-BNP and BNP-32 by a processing enzyme prior to their secretion. BNP-32 binds to natriuretic receptor-A and exerts such effects as vasodilation, diuresis, natriuresis, inhibition of aldosterone secretion and lipolysis, among others. In contrast, nonglyNT-BNP and glyNT-BNP have no known physiological activity and are excreted mainly in the urine. In patients with chronic renal failure on haemodialysis, nonglyNT-BNP and glyNT-BNP are not excreted in the urine, and are therefore markedly elevated in the blood.

    Acknowledgments

    We thank Ms Masako Minato, Ms Kyoko Tabei, Ms Keiko Fukuda, Mr Kazumi Akimoto and Ms Machiko Sakata for their technical assistance with measuring BNP, nonglyNT-BNP and glyNT-BNP. We thank Mr Yoshibumi Akutsu for his technical assistance with the echocardiography. We also thank Dr Naoto Minamino for helpful advice.

    References

    1. ↵
      1. Nishikimi T,
      2. Kuwahara K,
      3. Nakao K
      . Current biochemistry, molecular biology, and clinical relevance of natriuretic peptides. J Cardiol 2011;57:131–40.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Nishikimi T,
      2. Maeda N,
      3. Matsuoka H
      . The role of natriuretic peptides in cardioprotection. Cardiovasc Res 2006;69:318–28.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Weber M,
      2. Hamm C
      . Role of B-type natriuretic peptide (BNP) and NT-proBNP in clinical routine. Heart 2006;92:843–9.
      OpenUrlFREE Full Text
    4. ↵
      1. Daniels LB,
      2. Maisel AS
      . Natriuretic peptides. J Am Coll Cardiol 2007;50:2357–68.
      OpenUrlCrossRefPubMedWeb of Science
    5. ↵
      1. Schellenberger U,
      2. O'Rear J,
      3. Guzzetta A,
      4. et al
      . The precursor to B-type natriuretic peptide is an O-linked glycoprotein. Arch Biochem Biophys 2006;451:160–6.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Liang F,
      2. O'Rear J,
      3. Schellenberger U,
      4. et al
      . Evidence for functional heterogeneity of circulating B-type natriuretic peptide. J Am Coll Cardiol 2007;49:1071–8.
      OpenUrlCrossRefPubMed
    7. ↵
      1. Seferian KR,
      2. Tamm NN,
      3. Semenov AG,
      4. et al
      . Immunodetection of glycosylated NT-proBNP circulating in human blood. Clin Chem 2008;54:866–73.
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Luckenbill KN,
      2. Christenson RH,
      3. Jaffe AS,
      4. et al
      . Cross-reactivity of BNP, NT-proBNP, and proBNP in commercial BNP and NT-proBNP assays: preliminary observations from the IFCC Committee for Standardization of Markers of Cardiac Damage. Clin Chem 2008;54:619–21.
      OpenUrlFREE Full Text
    9. ↵
      1. Hammerer-Lercher A,
      2. Halfinger B,
      3. Sarg B,
      4. et al
      . Analysis of circulating forms of proBNP and NT-proBNP in patients with severe heart failure. Clin Chem 2008;54:858–65.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Nishikimi T,
      2. Minamino N,
      3. Ikeda M,
      4. et al
      . Diversity of molecular forms of plasma brain natriuretic peptide in heart failure—different proBNP-108 to BNP-32 ratios in atrial and ventricular overload. Heart 2010;96:432–9.
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Nishikimi T,
      2. Saito Y,
      3. Kitamura K,
      4. et al
      . Increased plasma levels of adrenomedullin in patients with heart failure. J Am Coll Cardiol 1995;26:1424–31.
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Nishikimi T,
      2. Karasawa T,
      3. Inaba C,
      4. et al
      . Effects of long-term intravenous administration of adrenomedullin (AM) plus hANP therapy in acute decompensated heart failure: a pilot study. Circ J 2009;73:892–8.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Yoshitomi Y,
      2. Nishikimi T,
      3. Abe H,
      4. et al
      . Comparison of changes in cardiac structure after treatment in secondary hypertension. Hypertension 1996;27:319–23.
      OpenUrlAbstract/FREE Full Text
    14. ↵
      1. Quiones GA,
      2. Douglas PS,
      3. Foster E,
      4. et al
      . ACC/AHA clinical competence statement on echocardiography: a report of the American College of Cardiology/American Heart Association/American College of Physicians-American Society of Internal Medicine Task Force on Clinical Competence. Developed in collaboration with the American Society of Echocardiography, the Society of Cardiovascular Anesthesiologists, and the Society of Pediatric Echocardiography. J Am Coll Cardiol 2003;41:687–708.
      OpenUrlCrossRefPubMed
    15. ↵
      1. Waldo SW,
      2. Beede J,
      3. Isakson S,
      4. et al
      . Pro-B-type natriuretic peptide levels in acute decompensated heart failure. J Am Coll Cardiol 2008;51:1874–82.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Xu-Cai YO,
      2. Wu Q
      . Molecular forms of natriuretic peptides in heart failure and their implications. Heart 2010;96:419–24.
      OpenUrlAbstract/FREE Full Text
    17. ↵
      1. Vickery S,
      2. Price CP,
      3. John RI,
      4. et al
      . B-type natriuretic peptide (BNP) and amino-terminal proBNP in patients with CKD: relationship to renal function and left ventricular hypertrophy. Am J Kidney Dis 2005;46:610–20.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Sakai H,
      2. Tsutamoto T,
      3. Ishikawa C,
      4. et al
      . Direct comparison of brain natriuretic peptide (BNP) and N-terminal pro-BNP secretion and extent of coronary artery stenosis in patients with stable coronary artery disease. Circ J 2007;71:499–505.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Spanaus KS,
      2. Kronenberg F,
      3. Ritz E,
      4. et al.
      ; Mild-to-Moderate Kidney Disease Study Group. B-type natriuretic peptide concentrations predict the progression of nondiabetic chronic kidney disease: the Mild-to-Moderate Kidney Disease Study. Clin Chem 2007;53:1264–72.
      OpenUrlAbstract/FREE Full Text
    20. ↵
      1. Eklund EA,
      2. Freeze HH
      . Essentials of glycosylation. Semin Pediatr Neurol 2005;12:134–43.
      OpenUrlCrossRefPubMed
    21. ↵
      1. Varki A
      . Biological roles of oligosaccharides: all of the theories are correct. Glycobiology 1993;3:97–130.
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Zachara NE,
      2. Hart GW
      . The emerging significance of O-GlcNAc in cellular regulation. Chem Rev 2002;102:431–8.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Jiang J,
      2. Pristera N,
      3. Wang W,
      4. et al
      . Effect of sialylated O-glycans in pro-brain natriuretic peptide stability. Clin Chem 2010;56:959–66.
      OpenUrlAbstract/FREE Full Text
    24. ↵
      1. deFilippi CR,
      2. Seliger SL,
      3. Maynard S,
      4. et al
      . Impact of renal disease on natriuretic peptide testing for diagnosing decompensated heart failure and predicting mortality. Clin Chem 2007;53:1511–19.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Schou M,
      2. Alehagen U,
      3. Goetze JP,
      4. et al
      . Effect of estimated glomerular filtration rate on plasma concentrations of B-type natriuretic peptides measured with multiple immunoassays in elderly individuals. Heart 2009;95:1514–19.
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Bargnoux AS,
      2. Klouche K,
      3. Fareh J,
      4. et al
      . Prohormone brain natriuretic peptide (proBNP), BNP and N-terminal-proBNP circulating levels in chronic hemodialysis patients. Correlation with ventricular function, fluid removal and effect of hemodiafiltration. Clin Chem Lab Med 2008;46:1019–24.
      OpenUrlCrossRefPubMedWeb of Science
    27. ↵
      1. Racek J,
      2. Krlov H,
      3. Trefil L,
      4. et al
      . Brain natriuretic peptide and N-terminal proBNP in chronic haemodialysis patients. Nephron Clin Pract 2006;103:162–72.
      OpenUrlCrossRef
    28. ↵
      1. Wahl HG,
      2. Graf S,
      3. Renz H,
      4. et al
      . Elimination of the cardiac natriuretic peptides B-type natriuretic peptide (BNP) and N-terminal proBNP by hemodialysis. Clin Chem 2004;50:1071–4.
      OpenUrlFREE Full Text
    29. ↵
      1. Nishikimi T,
      2. Futoo Y,
      3. Tamano K,
      4. et al
      . Plasma brain natriuretic peptide levels in chronic hemodialysis patients: influence of coronary artery disease. Am J Kidney Dis 2001;37:1201–8.
      OpenUrlPubMedWeb of Science
    30. ↵
      1. Frankenstein L,
      2. Clark AL,
      3. Goode K,
      4. et al
      . The prognostic value of individual NT-proBNP values in chronic heart failure does not change with advancing age. Heart 2009;95:825–9.
      OpenUrlAbstract/FREE Full Text

    Footnotes

    • See Editorial, p 95

    • Funding This study was supported in part by Scientific Research Grants-in-Aid 18590787 and 20590837 from the Ministry of Education, Culture, Sports, Science and Technology of Japan and by the Science Research Promotion Fund from the Promotion and Mutual Aid Corporation for Private Schools of Japan.

    • Competing interests None.

    • Ethics approval Ethics approval was provided by the ethics committee of the authors' institute and was carried out in accordance with the recommendations of the ethical committee of Dokkyo Medical University.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    Linked Articles