Background

GlycA is a NMR spectrometry-derived novel biomarker that reflects protein glycosylation. Although it potentially reflects N-acetylglucosamine residues on any circulating plasma protein, it has been shown that GlycA signal majorly reflects glycosylation of acute-phase proteins, particularly of α1-acid glycoprotein, haptoglobin, α1-antitrypsin, α1-antichymotrypsin, and transferrin [1]. Important changes both in the concentration of acute-phase reaction proteins as well as in their glycosylation pattern are known to occur in response to an inflammatory stimuli, being therefore important determinants of the GlycA signal. For that reason, GlycA is currently being studied as a new marker of systemic inflammation.

Recent studies have shown that GlycA is related to the risk of severe infection [2, 3], as well as to disease activity in lupus [4] and rheumathoid arthritis [5, 6]. In addition, GlycA has also a role in metabolic diseases and cardiovascular events and mortality. In prospective studies, GlycA has been related to overall, cardiovascular and cancer-related mortality [7,8,9,10]. It has also been related to the risk of incident diabetes [11], to insulin resistance, leptin/adiponectin [12] and to sodium excretion [13]. All the above mentioned associations of GlycA have been shown to be independent of C reactive protein (CRP), even though the use of the two biomarkers may provide additive prognostic information, at least for cardiovascular events [9].

Chronic kidney disease (CKD), as defined by reduced estimated glomerular filtration rate (eGFR) and/or the presence of albuminuria, is associated with low-grade inflammation, which was described as an important mechanism underlying CKD progression. However, studies evaluating the role of inflammatory biomarkers on CKD, including hsCRP, have yielded conflicting results [14,15,16,17,18]. Up to the present, GlycA has not been evaluated in the context of CKD. The aim of this study was to investigate the association of GlycA to albuminuria and eGFR in a Brazilian cohort of middle-aged men and women.

Methods

The data used for the present investigation come from the Brazilian Longitudinal Study of Adult Health (ELSA-Brasil), a multicenter prospective cohort study designed primarily to identify risk factors and the natural history of diabetes and cardiovascular disease. The design and preliminary findings of this study have been published elsewhere [19, 20]. Briefly, 15,105 civil servants aged 35 to 74 years from six cities in Brazil were enrolled between August 2008 and December 2010 for baseline examination. Approvals from institutional review boards of all centers were granted, and all individuals signed informed consent. Since GlycA was measured only in the Sao Paulo sample, for the current analysis only data on the 5050 Sao Paulo participants was used.

Interviews and examinations were carried out by trained personnel with strict quality control [21]. Trained nurses measured patients’ weight, height, waist, and hip circumferences. Body mass index (BMI) was calculated by dividing the patients’ weight in kilograms by height in meters squared. Blood pressure was measured with the validated Omron HEM 705CPINT oscillometric device (Omron Co, Kyoto, Japan) after a 5-minute rest with the patient in a sitting position in a quiet, temperature-controlled room (20–24 °C). Three measurements were taken at 1-min intervals and the mean of these measurements was calculated. Baseline laboratorial measurements [22] were done after an overnight fast (urine and blood) and biological fluids were collected and frozen [23]. An oral glucose tolerance test (75 g) was performed in non-diabetic participants [22].

Diabetes was defined as previous medical history of diabetes, use of medication to treat diabetes, fasting plasma glucose ≥ 126 mg/dl, 2-h plasma glucose ≥200 mg/dl, or HbA1C ≥ 6.5% Glomerular filtration rate was calculated using the equations from the Chronic Kidney Disease Epidemiology Collaboration without correction for race [24]. Albuminuria was measured via nephelometry in a 12 h–overnight sample, and categories of albuminuria were defined according to KDIGO: albuminuria A1 as albumin-to-creatinine ratio < 30 μg/mg, albuminuria A2 as ≥ 30 μg/mg creatinine and albuminuria A3 as ≥ 300 mg/g creatinine. GlycA was measured by NMR spectrometry (NMR LipoProfile® test spectra, LabCorp, Raleigh, NC1) in plasma of the 5050 Sao Paulo participants.

In the descriptive analyses, Jonckheere-Terpstra test for ordered alternatives and chi-square test were used to test differences among quartiles of GlycA for continuous and categorical variables, respectively. Uni and multivariate linear regression models were built using albumin-to-creatinine ratio (log-transformed) and eGFR as the dependent variables. For the logistic regression models, the dependent variable was defined as albuminuria A2 or A3, eGFR < 60 ml/min/1.73m2 or both (albuminuria A2 or A3 and/or eGFR < 60 ml/min/1.73m2). Lastly, ROC curves were built on the diagnosis of albuminuria (albuminuria A2 or A3) and of eGFR < 60 ml/min/1.73m2 and the performances of hsCRP and GlycA were compared using c statistics (http://www.vassarstats.net/).

Results

Among the 5050 participants, GlycA had a mean value of 416 ± 67 μmol/L (Fig. 1). In Table 1, descriptive clinical and laboratorial characteristics are shown according to the quartiles of GlycA. GlycA was positively related to several variables, including age, female sex, smoking, BMI, diabetes and insulin resistance biomarkers, blood pressure, albuminuria, hsCRP and lipids. It was also inversely related to alcohol consumption, eGFR and HDL-cholesterol. Figure 1 also shows the scatter plot between albuminuria and GlycA.

Fig. 1
figure 1

Histogram of GlycA and scatterplot between GlycA and albuminuria among 5050 participants

Full size image
Table 1 Baseline clinical and laboratorial characteristics among 5050 participants according to quartiles of GlycA
Full size table

In Table 2, the linear regression models for albuminuria and eGFR as the dependent variables are shown. In the univariable model (model 1), GlycA shows a significantly positive relation to albuminuria and an inverse association to eGFR. In the subsequent models, the significant relation remains even after adjustments for potential confounding variables. In the last model (model 5), it is shown that the association of GlycA to albuminuria is independent of eGFR and vice-versa.

Table 2 Unadjusted and adjusted linear regression models on albuminuria and eGFR among 5050 participants of ELSA-Brasil
Full size table

In Table 3, uni and multivariate models on the diagnosis of albuminuria (albuminuria A2 or A3 versus A1), low eGFR (< 60 ml/min/1.73m2 versus > 60 ml/min/1.73m2) and both (albuminuria A2 or A3 and/or eGFR < 60 ml/min/1.73m2) are shown. In these models, GlycA remains significantly and independently related to albuminuria and/or eGFR.

Table 3 Unadjusted and adjusted logistic regression models on the diagnosis of albuminuria A2 or A3 (versus albuminuria A1), on eGFR < 60 ml/min/1.73m2 (versus eGFR > 60 ml/min/1.73m2) and on both (albuminuria or eGFR < 60 ml/min/1.73m2) among 5050 participants
Full size table

Lastly, Fig. 2 depicts the ROC curves for the diagnosis of albuminuria (albuminuria A2 or A3), for the diagnosis of eGFR < 60 ml/min/1.73m2 and for both. GlycA showed a nearly significant higher area under curve in comparison to hsCRP for albuminuria and for the combined diagnosis of albuminuria eGFR < 60 ml/min/1.73m2 (c statistics p value of 0.06 and 0.08, respectively), but not for eGFR < 60 ml/min/1.73m2 alone (c statistics p value of 0.16).

Fig. 2
figure 2

ROC curves on the diagnosis of albuminuria (a), eGFR <60 ml/min/1.73m2 (b) and both (c) for GlycA and hs-CRP among 5050 participants (p value for C statistics)

Full size image

Discussion

The results of the present study demonstrate that GlycA was significantly and independently related to albuminuria and eGFR in a population of middle-aged men and women. Importantly, this relationship was independent of hsCRP, suggesting that measuring glycosylation could provide additional information to the most widely used inflammatory biomarker. These results are in accordance with previous findings in the literature, which showed that GlycA is related to inflammatory diseases, insulin resistance and DM incidence, cardiovascular disease, and mortality.

The current results have two major implications. First, it hints that GlycA might contribute to CKD assessment and for identification of those at higher risk of CKD progression. Prospective studies are necessary to confirm this hypothesis by evaluating the performance of GlycA to predict renal outcomes. These studies should also evaluate if this new biomarker improves prediction of events in comparison to models using the current established markers of CKD progression.

Secondly, it raises questions on the role of glycosylation on CKD. Protein glycosylation refers to the enzyme-mediated posttranslational process of attachment of glycan chains. It is said to be an N-linkage when the glycan chain is attached to the nitrogen of an asparagine residue or to be O-linkage when the glycan is attached to the oxygen of a serine or threonine residue. While O-linkage glycosylation is more prominent intracellularly and is related to protein signaling and trafficking within the cell, N-linkage is the predominant pattern of glycosylation that occurs in circulating plasma proteins, through the action of glycosyltransferases, glycosidases, and syaliltransferases [25]. Glycosylation patterns are wide, depending on the substrate, enzymes, and monosaccharides available and number of branches being added. Glycosylation is known to alter protein function and is involved in several biological functions such as protein trafficking, protein signaling, ligand-receptor recognition, immunity and distinction between self and non-self and inflammation. For example, glycosylation seems to be important in enhancing adhesion molecules signaling in endothelial cells and marginalization and infiltration of leukocytes through the capillary wall [26]. Glycosylation is a determinant of immunoglobulin function and activation of the complement system [27,28,29,30]. Altering glycosylation might be a pathway cancer cells use to escape immunity and avoid apoptosis [31, 32].

While the determinants of glycosylation are not fully understood, research on the area is expanding, with a growing interest in glycomics. However, particularly in Nephrology, studies on glycosylation are incipient. Specific abnormal glycosylation patterns are being investigated in auto-immune and inflammatory diseases such as IgA nephropathy [33, 34] and multiple myeloma [35]. In the context of general CKD, data on glycosylation is scarce. One recent study showed that 14 traits of IgG glycosylation were related to renal function in CKD patients and in monozygotic twin pairs discordant for renal function [36]. These analyses were centered in patterns of IgG glycosylation.

As stated before, GlycA reflects increased glycosylation in acute-phase reactant proteins, more specifically of α1-acid glycoprotein, haptoglobin, α1-antitrypsin, α1-antichymotrypsin, and transferrin. Acute-phase reaction is a systemic response to several conditions such as infection, trauma, surgery, immunological and inflammatory diseases. It is mediated by several cytokines, with a particular emphasis on interleukin-6 as a major stimulator. It occurs in acute and chronic inflammation and involves several phenomena such as neuroendocrine, hematopoietic, metabolic changes and hepatic changes [37]. The liver increases the production of several proteins, such as CRP, amyloid-A,α1-acid glycoprotein, haptoglobin, complement fractions, mannose-binding lectin, coagulation factors, among others (positive acute-phase proteins), while decreasing the concentration of proteins such as albumin, transthyretin and thyroxine-binding globulin (negative acute-phase proteins). In addition to the increase in the positive acute-phase protein concentration, it has been shown that the liver also modifies post-translational processing of these proteins, with an increase in the glycosylation, mediated by cytokines and glucocorticoid [38, 39]. Studies suggest that both the change in the concentration of acute-phase proteins, as the pattern of glycosylation of these proteins might influence the inflammatory and immune-modulatory functions attributable to these proteins [40,41,42], promoting and intensifying the inflammatory response. Whether the relation between GlycA and eGFR and albuminuria is being mediated by the acute-phase response itself or by the increased glycosylation in the acute-phase response is a question that remains to be answered since in the current analysis these two conditions were inseparable.

A better understanding of determinants and patterns of glycosylation, as a more deep comprehension of how and to what extent glycosylation impacts the biological functions of glycoproteins on chronic inflammatory conditions such as CKD is a promising and needed area of research. Particularly of interest, the association between GlycA and albuminuria, the best “dynamic” marker of CKD risk we have so far, suggests that abnormal glycosilation may be an underlying mechanism mediating inflammation in CKD. This hypothesis leads to questions whether glycosylation could play a role not only in mediating the effect of traditional risk factors for CKD, such as diabetes, insulin resistance, and hypertension, but also be involved in inflammation signaling in primary and secondary glomerulonephritis, diseases where immunological insults are pivotal.

Our study has some limitations. First and most importantly, it is a cross-sectional analysis, and the role of GlycA regarding renal hard outcomes remains to be determined. In addition, the population recruited is essentially a non-CKD population. In order to confirm and better understand the role of GlycA in CKD, the present findings need to be confirmed in other studies, with an emphasis in those addressing CKD incidence and progression. Furthermore, we could not perform other techniques of measuring glycosylation, something that would be interesting to explore and adjust in relation to the GlycA signal.

Conclusion

In conclusion, the results from this cross-sectional study showed that GlycA is significantly and independently associated with albuminuria and eGFR. These findings support further exploration of the role of glycosylation in CKD progression and risk assessment.