Neutrophil gelatinase-associated lipocalin as a risk marker in cardiovascular disease

Main characteristics of NGAL

NGAL is a 25 kilodalton (kDa) human protein belonging to the lipocalin superfamily [8]. The human neutrophils are the main source of plasma NGAL, which are constitutively expressed and synthesized at early-myelocyte stages of granulopoiesis in the bone marrow [9]. Consequently, the mature peripheral neutrophils only store NGAL protein in granules.

It has been demonstrated that NGAL functions as an inflammatory modulator of the innate immune system [10]. This is supported by the fact that NGAL is capable of binding siderophores, reducing the amount of siderophore-bound iron available to bacteria inhibiting bacterial growth by depleting their intracellular iron stores [11].

The complex between NGAL and MMP-9 inhibits degradation of MMP-9 thereby extending the proteolytic activity of MMP-9 (Figure 1) [13]. By enhancing the activity of MMP-9, the degradation of the basement membranes and extracellular matrix increases, a process recognised in the progression of unstable atherosclerotic plaques in arteries [5]. In atherosclerotic plaques, the collagen degradation by MMP-9 is high, which causes an unstable fibrous cap and makes the atherosclerotic plaque more susceptible to erosion and rupture [5].

Figure 1:

The role of NGAL and MMP-9 in the atherosclerosis.

At the site of plaque rupture (tunica intima), activated macrophages and mast cells produce proteases (MMP-9), which destabilise the atherosclerotic plaque. The molecules degrade the fibrous cap and collagen and make the plaque more prone to rupture. A complex formation between neutrophil gelatinase associated lipocalin and matrix metalloproteinase 9 (MMP-9) protects and prolongs the proteolytic activity of MMP-9 inducing plaque rupture and risk of thrombus formation [12].

Assessment of NGAL levels in blood and urine

Differences in NGAL cut-off levels for prediction of outcome in diseases and reference values in the healthy population have been observed in numerous reports [14]. Several factors contribute to the variation such as different clinical settings, biological variation, use of different assays, antibody configurations, anticoagulants, and storage conditions [15]. Different methods including commercial and in-house sandwich immunoassays and immunoblotting protocols have been used to determine NGAL in plasma, serum, and urine [16], [17], [18].

Pedersen et al. tested the use of enzyme-linked immunosorbent assay (ELISA) kit, using the manual procedure, (BioPorto Diagnostics) for NGAL in plasma and urine. In healthy adults, the median intra and inter-assay variation was <5% and <10% for plasma and urine, respectively [19]. Similar results have been reported in other studies [18], [20]. In the study by Pedersen et al., the highest inter-assay variation in urine was ≤18.6% and likely explained by interference of sedimentation factors, emphasising the importance of centrifuging urine samples before analysis. Moreover, in plasma it was demonstrated that haemolysis caused NGAL levels to increase, which may be explained by release of NGAL from ruptured leucocytes, a process also observed in serum samples following multiple freeze/thaw cycles. In plasma and urine NGAL samples, levels were not significantly affected by repetitive freeze/thaw cycles [19].

To accommodate standardised assays for inclusion of NGAL in the standard panel to screen kidney diseases, the analytical performance of automated biochemistry analysers have been evaluated (Table 1). Cavalier et al. [32] performed an analytical comparison between a rapid bedside fluorescence-based immunoassay NGAL method (Triage Meter, Biosite Inc.) and the automated Architect NGAL laboratory assay with ethylenediaminetetraacetic acid (EDTA) plasma and urine samples. Both assays correlated well with validated ELISA kits from BioPorto. The Architect assay was shown to provide precise results in the range 22.5–1315 μg/L with maximum error of ±20% and probability of 95%. Using the same conditions, the corresponding range for the Triage method was 619–722 μg/L, indicating high analytical variation of the Triage method [32].

Hansen et al. [20] evaluated the turbidimetric NGAL Test™ from BioPorto with an automated method, Cobas 6000 c501 (Roche Diagnostics, Rotkreuz, Switzerland). The analytical performance was comparable to the automated Hitachi 917 in 40 matched samples of urine, Li-Hep plasma, and EDTA plasma. It was noted that NGAL levels were higher in Li-Hep plasma than in EDTA plasma, thus EDTA was recommended as anticoagulant.

Delanaye et al. [29] studied the biological variation of NGAL by measuring the absolute value of urinary NGAL and NGAL/creatinine ratio using an automated immuno-assay from Abbott Laboratories (Abbott Park, IL, USA) on the Architect platform. As the coefficient of variation (CV) for urinary NGAL was higher than the CV of NGAL/creatinine ratio, the authors recommended the use of NGAL/creatinine ratio to reduce intra-individual variation. Urinary NGAL variation was measured in healthy subjects only, thus the biological variation in patients with AKI or CVD remains unknown.

While several reference ranges for urine NGAL have been published, reference ranges for plasma NGAL have not yet been fully explored. Published reference ranges are summarised in Table 1.

NGAL in atherosclerotic plaques

In atherosclerotic plaques, thrombus formation is triggered by rupture of the plaque cap or endothelial erosion. Inside the plaque, inflammatory activity facilitates plaque instability and disruption [2]. The atheroma possesses inflammatory mediators, which inhibit smooth muscle growth and collagen production, and enhance the activity of MMP’s [12]. This also makes the atherosclerotic plaques unstable and vulnerable, potentially leading to plaque rupture.

Hemdahl et al. [5] analysed the association between NGAL expression in atherosclerotic plaques and myocardial infarction (MI) in mice exposed to hypoxic stress. Mice developing MI had higher concentrations of NGAL and MMP-9 than mice with stable atherosclerotic plaques. In atherosclerotic plaques retrieved from human carotid endarterectomy specimens, expression of NGAL and MMP-9 were detected in macrophages [5]. Yndestad et al. found that NGAL was over-expressed in the myocardium following MI in a rat model, and murine data suggests that mice deficient in the NGAL mouse orthologue were protected from development of aortic aneurysms [6], [33]. In patients undergoing surgery for infrarenal abdominal aortic aneurysms, NGAL/MMP-9 complexes were identified in the thrombus and in the interface between the thrombus and the underlying wall [34]. Furthermore, NGAL has also been associated with intraplaque hemorrhages of human carotid endarterectomy samples [35].

Although causality is not explicitly proven, these findings might suggest a pro-atherogenic action of NGAL via enzymatically active NGAL/MMP-9 complexes. However, further studies are needed to explore whether NGAL is an important and specific marker of atherosclerosis.

NGAL and acute coronary syndrome

The association between NGAL levels and risk scores (clinical and angiographical) has been studied in patients diagnosed with non-ST-segment elevation MI (NSTEMI) and compared with healthy individuals [36]. High levels of NGAL, hs-CRP, and leukocytes were observed in the patient group. Furthermore, NGAL correlated positively with the global registry of acute coronary events (GRACE) risk score (which estimates the 6-month mortality risk in patients with acute coronary syndrome [ACS]). The severity of CAD was assessed by the Gensini score (evaluates collateral circulation of the coronary arteries) and the synergy between percutaneous coronary intervention with taxus and cardiac surgery (SYNTAX) (estimates the complexity of CAD) score showing increased NGAL levels in patients with intermediate and high SYNTAX scores compared to patients with low SYNTAX score [36].

Sahinarslan et al. reported that patients with acute MI had higher levels of plasma NGAL and leukocytes than stable CAD patients [37]. However, there was no correlation between plasma NGAL and the Gensini score. These findings were supported by Choi et al., who reported that NGAL levels were higher in patients with acute MI and unstable angina pectoris than stable CAD patients. However, the findings were not statistically significant, possibly due to a low number of patients diagnosed with ACS (n=11). Also, blood samples were collected at least 3 months after the acute event [38].

Akcay et al. [39] reported a poor prognosis in patients with high NGAL levels and ST-segment elevation MI (STEMI) treated with primary percutaneous coronary intervention (PCI) compared to patients admitted for elective PCI. Similarly, high NGAL levels independently predicted all-cause mortality and major adverse cardiac events (MACE) in STEMI patients treated with PCI [40]. Furthermore, the study showed that high levels of CRP and NGAL were related to poor outcomes, while low levels of NGAL and CRP were associated with decreased risk of MACE. Indeed, the risk still remained high when NGAL was elevated even though CRP was low [40]. This is consistent with findings by Zykov et al., who also reported that high NGAL level, is a strong predictor of all-cause mortality suggesting a high prognostic value of NGAL in patients with STEMI [41].

The majority of these studies (Table 2) suggest that NGAL has the potential to improve risk stratification in patients with ACS.

NGAL in stable coronary artery disease

Only few studies have measured NGAL in stable CAD patients (Table 2), and so far, no causal relation between NGAL and stable CAD has been established.

Paulsson et al. [42] compared levels of neutrophils and MMP-9/NGAL complexes in stable CAD patients with age- and sex matched healthy individuals. Neutrophil levels did not differ between patients and healthy individuals, neither in peripheral blood nor at the site of intense inflammation. However, the concentration of interleukin-8 (IL-8), which stimulates the release of NGAL from neutrophil granules, was higher in patients than healthy individuals. This may contribute to the observed higher concentration of the MMP-9/NGAL complex in patients than healthy individuals and suggest a priming stage of circulating neutrophils in stable CAD patients [42].

Zografos et al. [43] reported that serum NGAL levels in stable CAD patients were significantly increased compared with patients with normal coronary arteries. Another study showed that patients with triple vessel disease had higher serum levels of NGAL than patients with single vessel disease [46]. Moreover, a positive association between NGAL and the SYNTAX score was reported. Patients with high levels of NGAL (>100 ng/mL) and brain natriuretic peptide (BNP) (>25 pg/mL) had a higher SYNTAX score than patients with low levels of NGAL (<100 ng/mL) and BNP (<25 pg/mL) [46].

Recently, Woitas et al. analysed the predictive role of plasma NGAL in CAD patients, including unstable and stable CAD patients and reported that plasma NGAL levels were independently related to all-cause mortality as well as CVD mortality after adjusting for conventional cardiovascular risk factors. However, when adjusting for creatinine levels, plasma NGAL did not predict all-cause mortality and CVD mortality [44]. Thus, the studies indicate that NGAL levels are higher in patients with ACS than in stable CAD patients, and that creatinine is a strong predictor of mortality [37], [45].

NGAL levels and acute kidney injury in cardiac surgery

AKI after cardiac surgery is a relatively frequent event, which causes dialysis, prolonged hospitalisation and increased in-hospital mortality [47]. A short summary of the studies on cardiac surgery and AKI are listed in Table 3.

In one study, 21 children underwent cardiac surgery on extracorporeal circulation and developed postoperative AKI. A significant increase in urinary NGAL concentration was reported in these children [16]. Others found that plasma NGAL concentration tripled and remained significantly elevated in 120 children who developed AKI after cardiac surgery [48].

Similar findings have been reported in studies on adults. In patients who developed AKI after cardiac surgery, urinary NGAL peaked within 2–4 h postoperatively, whereas the concentration of serum NGAL did not increase significantly [47]. Tuladhar et al. [17] also studied the development of AKI after cardiac surgery. The plasma and urinary AUC of NGAL using receiver operating characteristic curves (ROC) were 0.80 and 0.96, respectively. This is in accordance with a study by Koyner et al. showing that the plasma NGAL concentration is inferior to urinary NGAL concentration in predicting AKI [50]. Taken as a whole these studies suggest a relevant role for urinary NGAL as a biomarker of AKI in cardiac surgery patients. However, a recent study by Friedrich et al. demonstrated that urinary NGAL failed to predict AKI in cardiac patients undergoing cardiopulmonary bypass and showed no correlation between urinary NGAL and postoperative AKI classification levels, even though urinary NGAL levels increased after surgery [49].

Although, both urinary and plasma NGAL levels are elevated in clinical conditions with AKI, studies indicate that urinary NGAL may be superior to plasma NGAL in diagnosing and predicting AKI. However, due to heterogeneity of studies and varying diagnostic criteria of AKI, a large variation of diagnostic accuracy has been reported with differing AUC’s (Table 2). Clerico et al. [15] reported AUC’s ranging from 0.54–0.96 for plasma NGAL and 0.61–0.98 for urine NGAL, indicating a large variation in the discrimination of patients with or without AKI. This may mainly be explained by enrolment of low numbers of patients, varying diagnostic criteria for the clinical endpoint AKI, different NGAL assays, non-standardised preanalytical variables (specimen type, storage and anticoagulants), and heterogenic study populations in relation to age and clinical setting.

NGAL in patients with heart failure

Growing evidence suggests that inflammation and matrix degradation in the myocardium play a pathogenic role in HF, which may be related to NGAL [6]. A summary of the studies on HF is provided in Table 4.

Yndestad et al. [6] investigated the role of NGAL in 236 patients with acute post-MI HF and 150 patients with chronic HF. Patients in NYHA class III had significantly higher plasma NGAL levels than patients in NYHA class I or II and controls. Additionally, high baseline plasma NGAL levels were associated with an increased incidence of the composite endpoint (non-fatal MI, stroke, cardiovascular death, and all-cause death) at 27 months (median) follow-up. Finally, a significant association between plasma NGAL and N-terminal prohormone of brain natriuretic peptide (NT-proBNP) was reported. Similar results were found by Bolignano et al. [51] in elderly chronic HF patients.

In a recent study, the predictive value of baseline NGAL was assessed in 61 patients with chronic systolic HF [52]. The primary endpoint was a composite of cardiovascular death and admission for heart failure during 10.6 months (median) follow up. In 15 patients who reached the endpoint, higher baseline NGAL concentrations were observed. The study concluded that plasma NGAL may be a predictor of outcome in patients with HF. Furthermore, the study reported a positive correlation between plasma NGAL and creatinine, yet no association was observed between plasma NGAL and BNP. This finding is in agreement with Lindberg et al. who did not report any association between NGAL and NT-proBNP [56].

In chronic HF patients, urine NGAL levels at admission were associated with increased risk of subsequent worsening of renal function [53]. Another study reported that urinary NGAL was associated with a combined primary endpoint of all-cause mortality and HF hospitalisation in chronic HF patients [57]. Both studies indicated that urinary NGAL may be a direct marker of renal damage and tubulo-interstitial injury in chronic HF patients.

Nymo et al. [54] studied the prognostic value of serum NGAL in elderly with chronic ischaemic HF. When adjusted for demographic and clinical variables, NGAL concentration remained a significant predictor of adverse outcome. However, when adjusted for apolipoprotein A-1 (ApoA-1), estimated glomerular filtration rate (eGFR), hs-CRP, and NT-proBNP, this association was lost. The authors concluded that the clinical use of circulating NGAL in this study was limited.

Maisel et al. [55] investigated the prognostic value of plasma NGAL in patients with acute HF. Patients were followed for 30 days, and the primary combined endpoint was HF readmission or all-cause mortality. Baseline and discharge levels of NGAL were significantly elevated in patients with events compared to event-free patients. Patients with events also had higher BNP levels. The AUC by using ROC analysis was higher for NGAL [0.73] than for BNP [0.65], indicating that plasma NGAL is a good predictor of the combined endpoint after 30 days in acute HF patients. In addition, the combination of high-NGAL and high-BNP or high-NGAL and low-BNP was associated with higher risk of HF readmission or all-cause death than low-NGAL and low-BNP or low-NGAL and high-BNP.

Both plasma and urinary NGAL levels are increased in chronic HF patients. The elevation is dependent on the clinical stage of HF and the presence of renal failure. Plasma NGAL may amplify the predictive value of NT-proBNP and increase the risk of adverse outcome. Thus, the combined elevation of NGAL and NT-proBNP is associated with the worst outcome.

Overall, NGAL demonstrated a questionable discrimination between categories and disease severity of CVD. Discrepancies are also seen between urine and plasma NGAL, where the performance of urine NGAL in most studies is superior to plasma for determining AKI. These differences may be due to different origins of NGAL in urine and plasma in AKI and CVD. During AKI, NGAL is primarily secreted from the kidney, whereas the pool of NGAL in blood may originate from several sources including neutrophils during inflammation, which may contribute to the observed differences in performance of urine and plasma NGAL in AKI and CVD [58], [59].