Skip to main content
Download PDF

Plasma N-Glycoproteomics in monozygotic twin pairs discordant for body mass index reveals an obesity signature related to inflammation and iron metabolism

Biology Direct volume 20, Article number: 31 (2025) Cite this article

Abstract

Background

N-glycosylation is a complex, post-translational modification which influences protein function and is sensitive to physiological changes. Obesity is associated with alterations in protein function; however, little is known about the glycoproteome in obesity beyond observations of association with types and structures of selected glycopeptides. Most often, due to technical challenges, glycan composition and structure information are missing. Here, we combined label-free data-independent proteomics and targeted quantitative glycoproteomics to study N-glycosylation of plasma proteins in obesity. Using a monozygotic twin study design, we controlled for genetic variation and focused only on the acquired effects of obesity.

Methods

Using plasma samples of 48 monozygotic twin pairs discordant for BMI (intrapair difference > 2.5 kg/m2), we identified using mass spectrometry, differential protein and glycopeptide levels between heavier and leaner co-twins. We used a within-twin paired analysis model and considered p < 0.05 as significant.

Results

We identified 48 protein and 33 N-glycosylation expression differences (p < 0.05) between co-twins. These differences occurred either both in the protein expression and glycoprotein (sometimes in opposing directions) or independently from each other. Haptoglobin protein was upregulated (Fold Change = 1.10, p = 0.001) in heavier co-twins along with seven upregulated glycan compositions at N-glycosylation site Asn241. The complement protein C3 was upregulated (Fold Change = 1.08, p = 0.014) along with one upregulated glycopeptide at Asn85. Additionally, many glycopeptides were upregulated despite non-significant differences in protein-backbone plasma levels.

Conclusion

Differential protein expression related to cholesterol biosynthesis and acute phase signalling as well as N-glycosylation of proteins related to iron metabolism and inflammation can be linked to acquired obesity.

Introduction

The World Obesity Federation, in 2024, estimated that over 1 billion people live with obesity (BMI ≥ 30 kg/m2) [1]. Obesity results from a complex interplay of genetic, epigenetic, and environmental factors [2]. While it has high genetic heritability, it is also influenced by lifestyle, socio-economic factors including education, childhood psychosocial factors, and emotional well-being [3, 4]. Obesity is a major public health concern that is associated with metabolic conditions that include insulin resistance, type 2 diabetes, cardiovascular diseases, and certain types of cancer [5]. Prevention and, failing that, timely treatment of obesity is thus pivotal in addressing this growing epidemic. Improved characterization of molecular alterations in obesity, which may contribute to metabolic derangements, is much needed.

One such molecular alteration is the N-glycosylation of proteins, a complex post-translational modification occurring on approximately half of all plasma proteins [6]. N-glycosylation is one of the most diverse protein modifications since there are as many possible N-glycan structures as structural branching possibilities. During protein production, the multi-span membrane protein complex oligosaccharyltransferase transfers the N-glycan precursors to the N-consensus sequence in the protein as the protein is being translated. After this point, glycosylation patterns start to evolve as glycans are modified by tissue specific enzymes, whose activity change in different pathophysiological states. These N-glycosylation changes can affect protein functionality and activity.

Fucosylation and sialylation are two glycosylation patterns often studied in human diseases [7, 8]. Sialylation of glycans regulates structural stability, cell recognition and communication and is important for immune response [9,10,11]. Fucosylation of glycans helps maintain a steady state in cells and tissues and has an important role in apoptosis [8]. Interestingly, although N-glycosylation is not template-driven, some N-glycans are heritable [12,13,14].

The relationship between proteins and their glycosylation patterns can vary; sometimes the glycoprotein profiles match the protein levels, while at other times, they do not. When observing microheterogenity (more than one variety of glycan at a specific glycosite [15, 16]) of one N-glycosylation site to the protein expression of that particular protein, the basic hypothesis in homeostasis is that they are similar in abundance and variety of the N-glycans produced. When we detect uncoupling of the qualitative or quantitative aspects of the N-glycosylation compared to the protein expression, we know that expression or the activity of the enzymes synthesizing or degrading the glycans have changed. These changes can be specific to a pathophysiological state and affect protein functions and activities [17]. Hence, these N-glycosylation alterations can be used as biomarkers of pathophysiological processes beyond what can be observed in proteins alone.

Aberrant plasma protein glycosylation is associated with a wide range of diseases, including cardiometabolic [18] and immunological disorders [19], and various cancer types [20,21,22]. For instance, proinflammatory immunoglobulin G (IgG) glycome traits associate with higher BMI [23, 24]. In individuals with obesity, higher levels of glycans with antennary fucose, sialylation and low-branching structures were found in the plasma of individuals after diet-induced weight loss [25,26,27]. Meanwhile, weight-loss following bariatric surgery associates with a decrease in core-fucosylated glycans and an increase in sialylated glycans [28].

The traditional method for N-glycan profiling involves cleaving glycans from the peptide chain and then analyzing the free glycan structures to detect global glycan-alterations [29]. Unfortunately, this process leaves the protein of origin for each glycan unknown, with the glycosylation-site information lost as well. As the type of glycan and its specific location within a glycopeptide can drastically change the behavior and characteristics of the protein, glycan/glycosylation site information may provide further valuable insights. Using our own previously developed workflow for quantitative N-glycoproteomics, we can not only quantify N-glycopeptides but also derive their peptide sequence, glycan composition, and proposed glycan structure [19, 21, 22, 30].

Obesity-related changes, in the plasma proteome have been extensively studied, revealing significant alterations in protein expression and function [31,32,33,34,35]. However, very little is known about the glycoproteome in obesity beyond observations of association with types (e.g. fucosylation and sialylation of glycans) and structures (e.g. simple, hybrid, complex) of only selected glycopeptides [23, 24, 36]. In this study, we addressed a critical gap in the literature regarding the effects of acquired obesity on both the global proteome and global glycoproteome levels using methodology that can identify glycan compositions at the glycosylated sites. By studying 48 rare monozygotic (MZ) twin pairs discordant for BMI (within-pair difference in BMI, ≥ 2.5 kg/m2), we identify genetically identical individuals with differing body weight. This is a unique and ideal design to isolate environmental and lifestyle factors from genetic influences. Our findings point at environmental factors that differ between the co-twins. Additionally, we uniquely combine results from both the proteins and glycopeptides to characterize obesity.

Materials and methods

Participants and study design

We selected 48 monozygotic twin pairs discordant for BMI from population-based longitudinal studies (FinnTwin16 (n = 2839 pairs [37]), FinnTwin12 (n = 2578 pairs [38]), and Older Finnish Twin Cohort (n = 2932 pairs [39]). These twin pairs, with within-pair BMI difference (ΔBMI ≥ 2.5 kg/m2), span two age groups (27–42 years old [n = 38 pairs] and 57–69 years old [n = 10 pairs]), with a mean age of 37.8. The participants included 29 female pairs. Smokers were categorised as either current smokers or non-smokers.

The study, approved by the Ethics Committee of the Helsinki University Central Hospital (protocol number 270/13/01/2008), adhered to the Declaration of Helsinki principles. All participants provided written informed consent.

Body composition, energy intake and expenditure

We measured weight, height, whole-body fat (assessed by dual-energy x-ray absorptiometric scans), abdominal subcutaneous and visceral fat (measured using magnetic resonance imaging, MRI), and liver fat (measured using magnetic resonance spectroscopy, MRS) in the study participants as described here [40]. Additionally, we estimated participants’ physical activity with the Baecke questionnaire [41].

Blood laboratory examinations

We collected fasting blood samples for blood biochemistry measures (analyzed from fresh samples using standardized methods at the HUSLAB laboratories) and the glycoproteomics analyses, using stored plasma samples kept at -80 °C.

Proteomics and N-glycoproteomics analysis

Detailed information can be found in Supplementary Methods 1, and Supplementary Fig. 1 summarizes the proteomics and N-glycoproteomics workflows. Briefly, albumin was removed from the plasma samples and remaining proteins digested with trypsin. Peptide fractions and N-glycopeptide were isolated with size exclusion chromatography and both fractions were analyzed separately with LC-MS techniques. Since proteins that are most abundant in the plasma are also glycosylated, we did not deplete these proteins, except for albumin which does not have glycosylation sites. For proteomics, peptides were quantified and identified by LC-UDMSE runs.

For glycoproteomics, N-glycopeptides were first quantified (using size-exclusion chromatography) with LC-MSE runs. 16,000 N-glycopeptide ions were quantified and targeted for fragmentation with LS-MS2 techniques. Where possible, these peptide sequences were elucidated and the N-glycosylation site, the glycan composition, and proposed glycan structures investigated. The glycan compositions are shown, with the following abbreviations of each monosaccharide as follows: H: hexose; N: hexosamine; S; sialic acid; F: fucose. Importantly, the glycan structures are only proposed based on the composition and spectrum; these structures were not experimentally validated.

Statistical analysis

Participant characteristics

Clinical variables are expressed as mean ± standard deviation (SD) for normally distributed variables, and as median (interquartile range) for non-normal variables. Differences between the co-twins (the heavier versus leaner co-twins) were assessed using paired Wilcoxon signed-rank test with p < 0.05 considered statistically significant.

Differential protein and N-glycopeptide analyses

We performed differential protein level analysis to identify differences in protein levels between co-twins. Additionally, we separately identified glycopeptides with different expression levels between co-twins. We used a paired samples analysis model (package limma in R-Bioconductor) to identify the non-genetic and non-shared environmental factors associating with the heavier/leaner status of the co-twins. Because not all twins were concordant for smoking status, we adjusted for it. We considered p < 0.05 statistically significant.

We identified:

  1. 1)

    proteins that were differentially expressed although related glycopeptides were not differentially expressed,

  2. 2)

    glycopeptides that were differentially expressed although related proteins were not differentially expressed,

  3. 3)

    proteins and related glycopeptides that were both differentially expressed,

  4. 4)

    proteins and glycopeptides for which we did not measure or identify their associated counterparts.

Biological pathway analysis

We further enriched the differentially expressed proteins using Ingenuity pathway analysis (Ingenuity Systems, Redwood City, CA, USA).

Clinical measures association analysis

We investigated, using linear mixed models (package R-lme4 [42]) with adjustments for sex, age, smoking and twinship, if any of the differentially (within-twin pairs) expressed proteins and glycopeptides associated with clinical measures.

N-glycosylation structural features analysis

Because sialylation and fucosylation of glycoproteins is altered in many diseases [7, 8], in addition to individually measured glycopeptides, we derived summary measures using the identified glycopeptides to calculate the total amount of sialyation or fucosylation of glycopeptides for each of the identified proteins. We used ratios to indicate, for each person, for each protein, the abundance of sialylated glycopeptides compared to the total abundance of glycopeptides as well as the abundance of fucosylated glycopeptides compared to the total abundance of glycopeptides. We used linear mixed models (package R-lme4 [42]) with adjustments for sex, age, smoking, and twinship to check if the amount of sialylation or fucosylation in a protein associate with clinical measures. Additionally, we also checked if the heavier co-twins had, in total, higher or lower amounts of sialylated or fucosylated proteins across all proteins with identified glycopeptides.

Results

Characteristics of study participants

As designed, the twin pairs displayed substantial discordance in all measures of adiposity. The heavier co-twins of the twin pairs (median BMI 30.5 [27.6–34.8]) had consistently higher measures of adiposity (i.e., body weight, BMI, body fat, fat-free mass, subcutaneous and intra-abdominal fat, adipocyte volume), insulin resistance, triacylglycerol, CRP, liver fat, and lower HDL compared to their leaner co-twins (median BMI 24.6 [22.1–27.7]) (Table 1). No significant differences were detected in total cholesterol, low density lipoprotein (LDL) cholesterol, or physical activity levels between the heavier and leaner co-twins. 14 twin pairs were discordant for smoking status (9 leaner co-twins and 5 unrelated heavier co-twins were smokers) and seven twin pairs were concordant for smoking status.

Table 1 Participant characteristics of 48 monozygotic twin pairs discordant for BMI; mean age 37.8 years (SD ± 14.1); 29 pairs (60%) were female
Full size table

Overview of proteomics and glycoproteomics results

Figure 1 provides an overview of both the proteomics and glycoproteomics results. Overall, out of the 230 identified proteins, 48 were differentially (p < 0.05) expressed between co-twins. Additionally, out of the 108 identified glycopeptides, 33 glycopeptides were differentially (p < 0.05) expressed between co-twins.

Fig. 1
figure 1

Overview of proteomics and glycoproteomics analyses results. Overall, 230 proteins were identified, with 48 proteins differentially (p < 0.05) expressed between heavier and leaner BMI-discordant co-twins. Out of these 48, three proteins were linked to ten glycopeptides which were also differentially (p < 0.05) expressed between co-twins. The remaining 45 proteins were either linked to glycopeptides not differentially expressed between co-twins (2 proteins) or not linked to any of the identified glycopeptides (43 proteins). Additionally, 108 glycopeptides were identified. Out of these 108, 33 glycopeptides were differentially (p < 0.05) expressed between co-twins. Out of these 33, only 29 has proteomics data as well. Ten of these 29 were linked to differentially expressed proteins. The remaining 23 glycopeptides were either linked to proteins not differentially expressed between co-twins (19 glycopeptides) or not linked to any of the identified proteins (4 glycopeptides). Additionally, we found 1986 differentially expressed glycopeptides that were detected but not traced to any of the proteins. (nominal p < 0.05 is considered significant)

Full size image

Most (35 out of 48) of the differentially expressed plasma proteins had higher expression in the heavier co-twins (Table 2, Supplementary Table 1). Several of these 48 proteins have been previously reported to be associated with obesity. For example, adiponectin (ADIPO) was one of the most significantly differentially expressed proteins with lower expression (FC = 0.85 p < 0.001, Table 2) in heavier co-twins.

Table 2 Top ten (by p-value) differentially expressed proteins between the heavier and leaner co-twins in the BMI-discordant pairs (n = 48 pairs/96 individuals). The paired co-twin analysis model was adjusted for smoking status
Full size table

Similar to the protein expression patterns, most (31 out of 33) of the differentially expressed and identified glycopeptides were upregulated in the heavier co-twins; fold changes ranged from 0.84 to 1.40 (Table 3).

Table 3 Differentially expressed glycopeptides (with identified related proteins) between the leaner and heavier co-twins in the BMI-discordant pairs (n = 48 pairs/96 individuals). The paired co-twin analysis model was adjusted for smoking. Glycan composition abbreviations: H: hexoses, N: N-acetylhexosamines, S: sialic acids, F: fucose. Fold change indicates what happens in the heavier co-twins
Full size table

A third of the differentially expressed N-glycopeptides also mapped to differentially expressed proteins

Because some glycopeptide levels may also reflect protein levels, while other glycopeptides may be differentially expressed independently of the protein alterations, we sought to combine the results from the proteomics and glycoproteomics. Out of the 33 differentially expressed glycopeptides (Table 3, Supplementary Table 2), we found ten glycopeptides that mapped to three differentially expressed proteins: haptoglobin (HPT), haptoglobin-related protein (HPTR) and complement 3 (C3). Additionally, we found 19 differentially expressed glycopeptides mapping to eight proteins, mostly immunoglobulin subunits, which (in contrast to their glycopeptides) were not significantly different between the co-twins. The remaining four significant glycopeptides mapped to proteins that were not identified in our proteomics dataset (Table 3).

On a protein level, acute phase signalling and cholesterol biosynthesis were upregulated

The top pathways enriched for these 48 differentially expressed proteins were related to acute phase signaling, DHCR24 signaling (related to cholesterol biosynthesis) and coagulation system (Supplementary Fig. 2), all three showed upregulation in the heavier co-twins.

On a glycoproteome level, differential glycosylation patterns were observed in proteins involved in iron metabolism

HPT, a hemoglobin scavenging protein in the plasma, showed significant protein upregulation (FC 1.10 p = 0.001) in heavier co-twins (Fig. 2; Table 3). Although this fold change is relatively small at the protein level, it may represent larger differences in absolute amount due to its nature of being a high-abundance plasma protein. Out of the four N-glycosylation sites for haptoglobin, we only found differential glycopeptides at site Asn241 indicating glycosylation microheterogeneity at this site. Out of the ten identified glycopeptides for HPT, we found seven different upregulated glycan compositions with fold changes 1.20–1.40. Interestingly, all these glycopeptides were sialylated with at least one sialic acid (Fig. 2; Table 3). One glycopeptide had a hybrid glycan structure, while the others had complex glycan structures. These results show increased sialylation at site Asn241 in the heavier co-twins. As all these glycopeptides have higher fold changes than the protein levels, it can be speculated that these glycosylation patterns are somewhat independent of the protein-level alterations.

Fig. 2
figure 2

Heatmaps of the differential analysis results in relation to the iron metabolism and inflammation functions. The heatmap depicts fold changes between co-twins for the significant differentially expressed proteins/glycopeptides for the two main functions identified i.e. iron metabolism and inflammation. The colors of the heatmap indicate the strength of the fold changes i.e. red indicates higher values in heavier co-twins and blue lower values in heavier co-twins. p < 0.05 was considered significant, * p < 0.05

Full size image

HPTR, a protein with a similar function to haptoglobin, revealed three glycosylation sites and had a lower protein level (FC 0.94, p = 0.035) in the heavier co-twins. We identified eight glycopeptides for haptoglobin-related protein at two N-glycosylation sites (Asn126 and Asn149/153), out of which two glycan structures had significantly higher expression levels in the heavier co-twins (FC 1.15 and FC 1.25, p < 0.05) (Fig. 2; Table 3). In heavier co-twins, at site Asn126, one hybrid glycan structure had higher expression and at site Asn149/Asn153 one complex glycan structure had higher expression. Additionally, we also found eleven N-glycopeptides which were not different between the co-twins, that map to two differentially expressed proteins (HPT and HPTR) (Supplementary Table 2).

The protein levels of the heme-binding protein hemopexin, the iron-binding protein ceruloplasmin and the iron-oxidizing protein transferrin were not different between the co-twins (Fig. 2; Table 3). However, we identified altered glycopeptide abundances for all these proteins. For hemopexin, we found two out of five identified glycopeptides (with complex glycan structures at Asn187 and at Asn453) to be significantly higher in heavier co-twins. Moreover, we identified for ceruloplasmin four out of eight glycopeptides and transferrin three out of seven glycopeptides to be significantly higher in heavier co-twins (Fig. 2). Overall, we found upregulation in the glycopeptides related to iron metabolism; this pattern is not seen on the protein level.

On a glycoproteome level, differential glycosylation patterns were observed in proteins involved in inflammation

Altered complement regulation in plasma is present in early stages of clinically healthy individuals with obesity [32]. In this study, the most abundant complement component, the glycoprotein complement 3, was upregulated (FC = 1.08, p = 0.014) in heavier co-twins. Complement 3 contains two N-glycosylation sites; we found one high-mannose glycopeptide at Asn85 to be higher in heavier co-twins (FC = 1.18 p = 0.039, Fig. 2; Table 3). Both the protein and the glycopeptide levels have similar fold changes, suggesting that the difference in the glycopeptide may reflect the difference in the protein level. We also found one disialylated, complex glycopeptide for complement 4a to be higher in the heavier co-twin (FC = 1.113 p = 0.025, Fig. 2; Table 3) with no significant differences in the related protein levels.

The IgG class of antibodies is composed of four different subtypes of IgG molecules, which can all be glycosylated at one glycosylation site. For IgG2, we identified 13 glycovariants of which five were differentially expressed at glycosylation site Asn176. Of those, two core-fucosylated glycopeptides with one attached fucose were downregulated in heavier co-twins (FC = 0.91 and 0.78 respectively) (Fig. 2; Table 3). The highest upregulated glycopeptide was a large monosialylated complex glycan with the composition S1H11N11 (FC = 1.30 p = 0.003), indicating potential terminal galactosylation and sialylation. Both can have a large impact on the IgG functionality; however a more detailed study of the actual structure of this composition [43] is needed. For IgG1, we identified 16 glycopeptides of which one (with a complex structure) was significantly upregulated in heavier co-twins. Of note, none of the IgG subclass protein levels were different between the co-twins, suggesting that the differences in these glycopeptides are unrelated to IgG-protein levels. Overall, we found upregulation in the glycopeptides related to inflammation; this pattern is not obvious on the protein level.

Two-thirds of differentially expressed proteins associated with both adiposity and insulin resistance measures

Given that we were able to identify molecular level differences in the proteome and glycoproteome in obesity, we next asked how these molecular profiles associated with clinical measures. Across all significant within-twin pair differentially expressed proteins, there were associations (Supplementary Table 3, Fig. 3) with several adiposity measures (BMI, fat percentage, subcutaneous and intra-abdominal adipose tissue amount, liver fat percentage, and fat cell volume), HDL, CRP, and insulin measures (HOMA index and MATSUDA index). Proteins APOC3, APOCH, KIT that were associated with at least one adiposity measure and one insulin measure are responsible for regulation of lipase activity. Proteins APOC2, APOC3, APOH, KIT, LCAT that associated with insulin measures are related to a host of lipid metabolism processes (protein-lipid complex organization, regulation of plasma lipoprotein particle levels, neutral lipid metabolic process, lipid homeostasis, glycerolipid metabolic process, lipid localization).

Fig. 3
figure 3

Heatmaps of the associations of differentially expressed proteins and glycopeptides to clinical outcomes. Standardized coefficients (β) showing associations (in standard deviations) between the protein levels and clinical measurements (figure a on the left) and associations (in standard deviations) between the glycopeptode levels and clinical measurements (figure b on the right). The colors of the heatmap indicate the strength of the association i.e. red indicates a positive and blue a negative association. FDR p < 0.05 was considered significant, * FDR p < 0.05, ** FDR p < 0.01, *** FDR p < 0.001. The glycopeptides are labelled with the protein name followed by the glycan structure with an underscore (i.e. “_”) between them

Full size image

Protein HPT positively associated with most of the adiposity measures as well as HOMA with four of the HPT glycopeptides also showing positive associations with HOMA. Glycopeptides related to IgG and complement proteins also showed associations with adiposity measures and HOMA; CO3 showed a significant association to HOMA on a protein level as well.

Glycopeptides that associate with both measures of adiposity and insulin resistance were sialylated

Across all significant within-twin pair differentially expressed glycopeptides, all glycopeptides that showed associations with at least one adiposity measures (BMI, fat percentage, subcutaneous and intra-abdominal adipose tissue amount, liver fat percentage, and fat cell volume), and at least one insulin measure (HOMA index and MATSUDA index) were also sialylated (Supplementary Table 4, Fig. 3). Most of these glycopeptides are related to iron metabolism proteins.

Total amount of fucosylated or sialylated N-glycopeptides were different between co-twins

Bearing in mind that sialylation and fucosylation of glycans have been associated with several diseases and may also play a role in obesity, we, next, analyzed whether overall compositional features of the glycans, such as glycans being sialylated or fucosylated, were different between the co-twins and whether these features associated with clinical parameters. Across all the glycopeptides that were significantly different between co-twins and having protein level information, we found that heavier co-twins had significantly lower amounts of fucosylated N-glycopeptides (12.9% vs. 14.3%; p = 0.001). Conversely, they had significantly higher percentages of glycopeptides that were sialylated (66.5% vs. 65.3%; p = 0.027) (Supplementary Table 5).

On an individual protein level, heavier co-twins had a higher percentage of sialylated glycans attached to ceruloplasmin (p = 0.025) and a lower percentage of sialylated glycans on both transferrin (p = 0.031) and IGHG1 (p = 0.02073) compared to their leaner co-twins (Supplementary Table 6).

Additionally, we studied if the percentage of sialylation or fucosylation for a protein associated with clinical measures. The composition of glycans attached to transferrin showed that a lower amount of sialic acid was associated with higher adiposity, e.g. amount of subcutaneous fat (Supplementary Table 7, Supplementary Fig. 3) and CRP. HPT also associated negatively with adiposity and CERU associated positively with fat percentage and negatively with MATSUDA.

Discussion

In this study, given the understanding that glycosylation affects protein function, we used a unique study design of MZ twin pairs discordant for BMI to profile the effects of excess body weight on the plasma proteome and glycoproteome. We confirm previous findings that acquired obesity, i.e. independent of genetic predisposition, is associated with increased levels of proteins related to cholesterol biosynthesis and acute phase signaling. Additionally, we reveal a glycopeptide profile related to iron metabolism and inflammatory proteins that associates with acquired obesity, findings that were not evident on a proteome level. We also show that most of the sialylated glycopeptides associate positively with measures of adiposity and insulin resistance. Importantly, we elucidate the glycan structures and sites and combine both proteomics and glycoproteomics results to obtain an enhanced view of acquired obesity. Our study not only reaffirms previous proteomics findings but also unveils new associations with plasma protein N-glycosylation that gives us more insights into acquired obesity not seen on a proteome level.

Glycoprotein profiles are diverse and dependent on protein expression along with the expression of glycan structures on the proteins. On the proteome level, we found upregulation of cholesterol biosynthesis and acute phase signaling, findings already linked to obesity [34]. On the glycopeptide level, we found iron metabolism and inflammation to be upregulated in the heavier co-twins. Not surprisingly, only a third of our differentially expressed glycopeptides also showed associations on proteomics levels. These obesity signatures that are missing from proteomics alone highlight the importance of glycoproteomics in characterizing obesity. For example, the significantly different iron metabolism glycopeptides that were upregulated in obesity only showed upregulation in two out of the five identified significantly differentially expressed proteins. Similarly, the differentially expressed inflammation-related glycopeptides that were upregulated in obesity only showed upregulation in one out of the six differentially expressed inflammation-related proteins.

Obesity has been linked to reduced absorption of iron and low serum iron levels [44, 45]. We found, in the heavier co-twins, higher levels of multiple glycopeptides that are related to iron metabolism [46]. Under normal conditions, iron is transported in the bloodstream by transferrin [46]. Ceruloplasmin oxidizes Fe2+ into Fe3+ to facilitate the binding of iron to transferrin [47, 48]. The scavenger proteins haptoglobin, haptoglobin-related protein and hemopexin protect tissues from oxidative damage by binding free hemoglobin and free heme, respectively [49]. In obesity, a condition of low-grade inflammation, the expression of hepcidin, a key regulator of iron homeostasis is stimulated [50]. Increased levels of hepcidin have been shown to inhibit the release of iron from macrophages thereby reducing intestinal iron absorption and decreasing serum iron levels [51]. Additionally, it is important to note that the proteins and related glycopeptides we discuss here, as part of our findings, are also acute phase proteins. Hence, the differences detected in the heavier versus leaner co-twins may point to both inflammation and related effects in iron metabolism protein functioning due to altered N-glycosylation.

In our study, we observed upregulation of several glycopeptides related to haptoglobin (HPT) and haptoglobin-related protein (HPTR). It is well known that N-glycosylation can affect protein folding [52, 53] and protein-protein interactions [54]. For instance, N-glycosylation can fine-tune the tight haptoglobin and hemoglobin interactions that are widely described as ‘irreversible’ by displaying microheterogeneity in the form of altered fucosylation and glycan branching at a specific N-glycosylation site [55]. Moreover, a molecular dynamic simulation study suggests that different glycans on the transferrin protein affect the layout of the iron binding site residue and the transferrin structure [56]. In our study, we show increased microheterogeneity for haptoglobin and transferrin. We also show that a lower percentage of sialic acids in the glycans attached to transferrin were associated with increased adiposity. It is tempting to speculate that these alterations may impact the protein function. Nevertheless, how the identified glycopeptide characteristics specifically attribute to iron metabolism in obesity remains to be studied.

Our findings also point to upregulation in glycopeptides related to complement 3 (C3) and complement 4 (C4). C3 is a central and the most abundant protein of the complement system, a key system for immune surveillance with a significant role in the pathogenesis of many diseases [57]. We have previously shown that deviant complement regulation in plasma is present in very early stages of clinically healthy individuals with obesity [32]. Here, we confirm our previous findings of complement regulation and observe an unprocessed, high-mannose (i.e. not complex nor hybrid) glycopeptide to be higher in heavier co-twins. Interestingly, in young people with type 1 diabetes, the C3 N-glycome has been reported to have more unprocessed glycan structures [58]. With regards to the C4A glycopeptide, we found significant positive associations with almost all the adiposity measures as well as HOMA making it a sensitive biomarker for both obesity and insulin resistance. Overall, however, it is important to note that the higher glycopeptide level may have been observed due to higher protein levels in the heavier co-twin.

It is also interesting to note that the highest upregulated IgG glycopeptide (composition S1H11N11) in heavier co-twins in our study is sialylated and associated with most of the adiposity measures as well as HOMA. IgG glycosylation has been well studied in cardiometabolic diseases [23, 24] and it has been shown that N-glycosylation determines the structure and immunological function of the IgG, including modulation of pro- and anti-inflammatory signaling and cellular immune response [59]. Inherent differences in IgG subclass specific glycosylation have been shown [60] and the different subclass glycosylation profiles found in our study, particularly in IgG2 compared to the other subclasses, could point towards their different biological role. For instance, the glycan on IgG is important as it interacts directly with the cell surface receptor, fragment crystallizable γ receptor (FcγR) backbone while also altering the orientation of the two CH2 domains, and potentially their flexibility, to affect FcγR binding [61]. It has also been shown that the addition of sialic acid converts IgG from proinflammatory into an anti-inflammatory agent [62]. In mice, reduced IgG sialylation has also been shown to be implicated in obesity-induced insulin resistance [63]. Due to the observational nature of our study, the etiological interpretation of our findings remains speculative.

Fucosylation and sialylation of glycans have been used as predictive biomarkers in many diseases for example inflammatory conditions and cancer [64, 65]. Serum sialic acid (glycoproteins with sialic acid attached to the oligosaccharide side chains) is a marker of acute-phase response [66] and has been shown to be higher in individuals with higher BMI and to associate with triglyceride levels [67]. Our results also show that the heavier co-twins have higher percentages of sialylated glycopeptides. One sialylated glycopeptide of CERU positively associated with triglyceride levels. Although there were no significant differences in HPT fucosylation between co-twins, we found that percentage of fucosylation in protein HPT associated positively with adiposity measures in our study. Higher fucosylation, a hallmark of M1 inflammatory macrophages [68], in our dataset may be hinting at a higher inflammatory profile in the twins with higher BMI in our dataset.

It is important to note that, as in the case of the proteome [69, 70], some N-glycosylation is heritable, with several glycans showing high heritability (> 50%) in plasma [71, 72]. This may explain the small fold changes that we observe for glycopeptides. In general, more abundant glycans tend to have higher heritability [72]. N-Glycans mostly attached to IgGs exhibit high heritability containing core-fucosylated biantennary N-glycans with low levels of sialylation. However, for several IgG glycans, essential contribution of environmental variance has been observed in the TwinsUK cohort [71]. Low heritability of acute-phase proteins N-glycans might reflect physiological changes connected with general metabolic health and/or inflammation [72] and within the twin pairs, this may be regulated at the level of epigenetics [71].

The strength of our study is many-fold. Firstly, out twin study model allows us to only focus on the acquired effects of obesity while controlling for genetics. We also overlay both proteomics and glycopeptide level data to understand the glycoproteome profile. Additionally, we retain the protein of origin for each glycan and the glycosylation-site specific information. Therefore, we can observe the site-specific behavior of individual N-glycosylation sites and consider microheterogeneity next to macroheterogeneity (i.e., unique combinations of glycans at different sites of the same protein) and metaheterogeneity (i.e., variation in glycosylation across multiple sites of a given protein) which is often observed in the traditional method for N-glycan profiling. Nevertheless, if glycan alterations are found at different glycosylation sites within the same protein, we are unable to determine whether these occur together within a protein, or separately. We also cannot determine the types of glycosidic bonds or tell apart monosaccharide epimers with our current analysis methods. Additionally, the functionalities of individual glycan/protein combinations are not well understood and require follow-up studies to understand their effects. Finally, we are not able to pinpoint which of the environmental factors (e.g. eating behaviors [73], socio-economic factors [74], physical activity [75], adverse childhood psychosocial factors [76], emotional wellbeing [77], medication [78]) associate with the differences in the co-twins.

Conclusion

In conclusion, we show that the differential protein expression related to cholesterol biosynthesis and acute phase signalling as well as N-glycosylation of plasma proteins related to iron metabolism and inflammation can be linked to acquired obesity. We also show association on both proteins and glycopeptides to clinical measures. These distinct plasma glycoproteomics profiles in acquired obesity may provide directions for more personalized lifestyle or pharmacological interventions in the prevention of cardiometabolic disease. Notwithstanding, the underlying mechanisms that drive the observed differences require further study and future studies are urgently warranted to obtain detailed insight into these differential metabolic phenotypes.

Data availability

Proteomics and N-glycopeptide data are uploaded to the PRIDE database [79]. The proteomics dataset has an accession PXD041584 and N-glycoproteomics PXD041384.

Abbreviations

DIA:

Data Independent Acquisition

LC-MS:

Liquid Chromatography - Mass Spectrometry

UDMSE:

Ultra Definition Mass Spectrometry

MSE:

Elevated Energy Tandem Mass Spectrometry

CID:

Collision Induced Dissociation

MRM:

Multiple Reaction Monitoring

WE:

Wideband Enhancement

References

  1. World Obesity Federation. London: WOFWOA; https://data.worldobesity.org/publications/?cat=22

  2. Kim MS, Shim I, Fahed AC, Do R, Park W-Y, Natarajan P, et al. Association of genetic risk, lifestyle, and their interaction with obesity and obesity-related morbidities. Cell Metabol. 2024;36(7):1494–503. e3.

    Article  CAS  Google Scholar 

  3. Lin X, Li H. Obesity: epidemiology, pathophysiology, and therapeutics. Front Endocrinol. 2021;12:706978.

    Article  Google Scholar 

  4. Poulsen P, Vaag A, Kyvik K, Beck-Nielsen H. Genetic versus environmental aetiology of the metabolic syndrome among male and female twins. Diabetologia. 2001;44:537–43.

    Article  CAS  PubMed  Google Scholar 

  5. Blüher M. Obesity: global epidemiology and pathogenesis. Nat Rev Endocrinol. 2019;15(5):288–98.

    Article  PubMed  Google Scholar 

  6. Loaeza-Reyes KJ, Zenteno E, Moreno-Rodríguez A, Torres-Rosas R, Argueta-Figueroa L, Salinas-Marín R, et al. An overview of Glycosylation and its impact on Cardiovascular Health and Disease. Front Mol Biosci. 2021;8:751637.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Zhu W, Zhou Y, Guo L, Feng S. Biological function of sialic acid and sialylation in human health and disease. Cell Death Discovery. 2024;10(1):415.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Pekdemir B, Karav S. Exploring the diverse biological significance and roles of fucosylated oligosaccharides. Front Mol Biosci. 2024;11.

  9. Varki A. Glycan-based interactions involving vertebrate sialic-acid-recognizing proteins. Nature. 2007;446(7139):1023–9.

    Article  CAS  PubMed  Google Scholar 

  10. Cohen M, Varki A. The sialome–far more than the sum of its parts. Omics. 2010;14(4):455–64.

    Article  CAS  PubMed  Google Scholar 

  11. Pshezhetsky AV, Hinek A. Where catabolism meets signalling: neuraminidase 1 as a modulator of cell receptors. Glycoconj J. 2011;28(7):441–52.

    Article  CAS  PubMed  Google Scholar 

  12. Huffman JE, Knežević A, Vitart V, Kattla J, Adamczyk B, Novokmet M, et al. Polymorphisms in B3GAT1, SLC9A9 and MGAT5 are associated with variation within the human plasma N-glycome of 3533 European adults. Hum Mol Genet. 2011;20(24):5000–11.

    Article  CAS  PubMed  Google Scholar 

  13. Lauc G, Essafi A, Huffman JE, Hayward C, Knežević A, Kattla JJ, et al. Genomics meets glycomics—the first GWAS study of human N-glycome identifies HNF1α as a master regulator of plasma protein fucosylation. PLoS Genet. 2010;6(12):e1001256.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Lauc G, Huffman JE, Pučić M, Zgaga L, Adamczyk B, Mužinić A, et al. Loci associated with N-glycosylation of human immunoglobulin G show pleiotropy with autoimmune diseases and haematological cancers. PLoS Genet. 2013;9(1):e1003225.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Čaval T, Heck AJR, Reiding KR. Meta-heterogeneity: evaluating and describing the diversity in Glycosylation between sites on the same glycoprotein. Mol Cell Proteom. 2021;20.

  16. West CM, Malzl D, Hykollari A, Wilson IB. Glycomics, glycoproteomics, and glycogenomics: an inter-taxa evolutionary perspective. Mol Cell Proteom. 2021;20.

  17. Reily C, Stewart TJ, Renfrow MB, Novak J. Glycosylation in health and disease. Nat Rev Nephrol. 2019;15(6):346–66.

    Article  PubMed  PubMed Central  Google Scholar 

  18. Cvetko A, Mangino M, Tijardović M, Kifer D, Falchi M, Keser T, et al. Plasma N-glycome shows continuous deterioration as the diagnosis of insulin resistance approaches. Bmj Open Diabetes Res Care. 2021;9(1):e002263.

    Article  PubMed  PubMed Central  Google Scholar 

  19. Joenvaara S, Saraswat M, Kuusela P, Saraswat S, Agarwal R, Kaartinen J, et al. Quantitative N-glycoproteomics reveals altered glycosylation levels of various plasma proteins in bloodstream infected patients. PLoS ONE. 2018;13(3):e0195006.

    Article  PubMed  PubMed Central  Google Scholar 

  20. Kontro H, Joenväärä S, Haglund C, Renkonen R. Comparison of sialylated N-glycopeptide levels in serum of pancreatic cancer patients, acute pancreatitis patients, and healthy controls. Proteomics. 2014;14(15):1713–23.

    Article  CAS  PubMed  Google Scholar 

  21. Saraswat M, Mäkitie A, Tohmola T, Dickinson A, Saraswat S, Joenväärä S, et al. Tongue Cancer patients can be distinguished from healthy controls by specific N-Glycopeptides found in serum. Proteom Clin Appl. 2018;12(6):1800061.

    Article  Google Scholar 

  22. Dickinson A, Joenväärä S, Tohmola T, Renkonen J, Mattila P, Carpén T, et al. Altered microheterogeneity at several N-glycosylation sites in OPSCC in constant protein expression conditions. FASEB BioAdvances. 2024;6(1):26–39.

    Article  CAS  PubMed  Google Scholar 

  23. Perkovic MN, Bakovic MP, Kristic J, Novokmet M, Huffman JE, Vitart V, et al. The association between galactosylation of immunoglobulin G and body mass index. Prog Neuro-psychopharmacology Biol Psychiatry. 2014;48:20–5.

    Article  Google Scholar 

  24. Russell AC, Kepka A, Trbojević-Akmačić I, Ugrina I, Song M, Hui J, et al. Increased central adiposity is associated with pro-inflammatory immunoglobulin G N-glycans. Immunobiology. 2019;224(1):110–5.

    Article  CAS  PubMed  Google Scholar 

  25. Deriš H, Tominac P, Vučković F, Briški N, Astrup A, Blaak EE, et al. Effects of low-calorie and different weight-maintenance diets on IgG glycome composition. Front Immunol. 2022;13(6):995186.

    Article  PubMed  PubMed Central  Google Scholar 

  26. Deriš H, Tominac P, Vučković F, Astrup A, Blaak EE, Lauc G, et al. Susceptibility of human plasma N-glycome to low-calorie and different weight-maintenance diets. Int J Mol Sci. 2022;23(24):15772.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Kim T, Xie Y, Li Q, Artegoitia VM, Lebrilla CB, Keim NL, et al. Diet affects glycosylation of serum proteins in women at risk for cardiometabolic disease. Eur J Nutr. 2021;60(7):3727–41.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  28. Greto VL, Cvetko A, Štambuk T, Dempster NJ, Kifer D, Deriš H, et al. Extensive weight loss reduces glycan age by altering IgG N-glycosylation. Int J Obes. 2021;45(7):1521–31.

    Article  CAS  Google Scholar 

  29. Oliveira T, Thaysen-Andersen M, Packer NH, Kolarich D. The Hitchhiker’s guide to glycoproteomics. Biochem Soc T. 2021;49(4):1643–62.

    Article  CAS  Google Scholar 

  30. Saraswat M, Joenväärä S, Tomar AK, Singh S, Yadav S, Renkonen R. N–Glycoproteomics of human seminal plasma glycoproteins. J Proteome Res. 2016;15(3):991–1001.

    Article  CAS  PubMed  Google Scholar 

  31. Geyer PE, Wewer Albrechtsen NJ, Tyanova S, Grassl N, Iepsen EW, Lundgren J, et al. Proteomics reveals the effects of sustained weight loss on the human plasma proteome. Mol Syst Biol. 2016;12(12):901.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Sahebekhtiari N, Saraswat M, Joenväärä S, Jokinen R, Lovric A, Kaye S, et al. Plasma Proteomics Analysis reveals dysregulation of complement proteins and inflammation in acquired Obesity—A study on rare BMI-Discordant monozygotic twin pairs. Proteom Clin Appl. 2019;13(4):1800173.

    Article  Google Scholar 

  33. Garrison CB, Lastwika KJ, Zhang Y, Li CI, Lampe PD. Proteomic analysis, immune dysregulation, and pathway interconnections with obesity. J Proteome Res. 2017;16(1):274–87.

    Article  CAS  PubMed  Google Scholar 

  34. Doumatey AP, Zhou J, Zhou M, Prieto D, Rotimi CN, Adeyemo A. Proinflammatory and lipid biomarkers mediate metabolically healthy obesity: a proteomics study. Obesity. 2016;24(6):1257–65.

    Article  CAS  PubMed  Google Scholar 

  35. Cominetti O, Núñez Galindo A, Corthésy J, Valsesia A, Irincheeva I, Kussmann M, et al. Obesity shows preserved plasma proteome in large independent clinical cohorts. Sci Rep. 2018;8(1):16981.

    Article  PubMed  PubMed Central  Google Scholar 

  36. Liu D, Li Q, Dong J, Li D, Xu X, Xing W et al. The association between normal BMI with central adiposity and proinflammatory potential immunoglobulin GN-glycosylation. Diabetes Metab Syndr Obes. 2019:2373–85.

  37. Kaidesoja M, Aaltonen S, Bogl LH, Heikkilä K, Kaartinen S, Kujala UM, et al. FinnTwin16: a longitudinal study from Age 16 of a Population-based Finnish twin cohort. Twin Res Hum Genet. 2019;22(6):530–9.

    Article  PubMed  Google Scholar 

  38. Rose SMS-F, Contrepois K, Moneghetti KJ, Zhou W, Mishra T, Mataraso S, et al. A longitudinal big data approach for precision health. Nat Med. 2019;25(5):792–804.

    Article  PubMed Central  Google Scholar 

  39. Kaprio J, Bollepalli S, Buchwald J, Iso-Markku P, Korhonen T, Kovanen V, et al. The older Finnish twin cohort– 45 years of follow-up. Twin Res Hum Genet. 2019;22(4):240–54.

    Article  PubMed  Google Scholar 

  40. Kolk BWvd, Saari S, Lovric A, Arif M, Alvarez M, Ko A, et al. Molecular pathways behind acquired obesity: adipose tissue and skeletal muscle multiomics in monozygotic twin pairs discordant for BMI. Cell Rep Med. 2021;2(4):100226.

    Article  Google Scholar 

  41. Baecke JA, Burema J, Frijters JE. A short questionnaire for the measurement of habitual physical activity in epidemiological studies. Am J Clin Nutr. 1982;36(5):936–42.

    Article  CAS  PubMed  Google Scholar 

  42. Bates D, Mächler M, Bolker B, Walker S. Fitting Linear mixed-effects models using lme4. J Stat Softw. 2015;67(1):1–48.

    Article  Google Scholar 

  43. Štambuk J, Nakić N, Vučković F, Pučić-Baković M, Razdorov G, Trbojević-Akmačić I, et al. Global variability of the human IgG glycome. Aging. 2020;12(15):15222–59.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Aigner E, Feldman A, Datz C. Obesity as an emerging risk factor for Iron Deficiency. Nutrients. 2014;6(9):3587–600.

    Article  PubMed  PubMed Central  Google Scholar 

  45. González-Domínguez Á, Visiedo-García FM, Domínguez-Riscart J, González-Domínguez R, Mateos RM, Lechuga-Sancho AM. Iron metabolism in obesity and metabolic syndrome. Int J Mol Sci. 2020;21(15):5529.

    Article  PubMed  PubMed Central  Google Scholar 

  46. Gkouvatsos K, Papanikolaou G, Pantopoulos K. Regulation of iron transport and the role of transferrin. Biochim Biophys Acta (BBA) - Gen Subj. 2012;1820(3):188–202.

    Article  CAS  Google Scholar 

  47. Patti MCBd, Cutone A, Nemčovič M, Pakanová Z, Baráth P, Musci G. Production of recombinant human ceruloplasmin: improvements and perspectives. Int J Mol Sci. 2021;22(15):8228.

    Article  Google Scholar 

  48. Musci G, Polticelli F, Patti, MCBd. Ceruloplasmin-ferroportin system of iron traffic in vertebrates. World J Biol Chem. 2013;5(2):204–15.

  49. Dev S, Babitt JL. Overview of iron metabolism in health and disease. Hemodial Int. 2017;21:S6–20.

    Article  PubMed  PubMed Central  Google Scholar 

  50. Coimbra S, Catarino C, Santos-Silva A. The role of adipocytes in the modulation of iron metabolism in obesity. Obes Rev. 2013;14(10):771–9.

    Article  CAS  PubMed  Google Scholar 

  51. Ganz T. Hepcidin and iron regulation, 10 years later. Blood. 2011;117(17):4425–33.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  52. Jayaprakash NG, Surolia A. Role of glycosylation in nucleating protein folding and stability. Biochem J. 2017;474(14):2333–47.

    Article  CAS  PubMed  Google Scholar 

  53. Shental-Bechor D, Levy Y. Effect of glycosylation on protein folding: a close look at thermodynamic stabilization. Proc Natl Acad Sci. 2008;105(24):8256–61.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  54. Varki A, Cummings RD, Esko JD, Stanley P, Hart GW, Aebi M et al. Essentials of glycobiology. 4th ed. Cold Spring Harbor (NY): Cold Spring Harbor Laboratory Press; 2022.

  55. Wu D, Struwe WB, Harvey DJ, Ferguson MAJ, Robinson CV. N-glycan microheterogeneity regulates interactions of plasma proteins. Proc Natl Acad Sci. 2018;115(35):8763–8.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  56. Ghanbari Z, Housaindokht MR, Bozorgmehr MR, Izadyar M. The effect of glycosylation on the transferrin structure: a molecular dynamic simulation analysis. J Theor Biol. 2016;404:73–81.

    Article  CAS  PubMed  Google Scholar 

  57. Sahu A, Lambris JD. Structure and biology of complement protein C3, a connecting link between innate and acquired immunity. Immunol Rev. 2001;180(1):35–48.

    Article  CAS  PubMed  Google Scholar 

  58. Šoić D, Štambuk J, Tijardović M, Keser T, Lauc G, Bulum T, et al. Human complement component C3 N-glycome changes in type 1 diabetes complications. Front Endocrinol. 2023;14:1101154.

    Article  Google Scholar 

  59. Cobb BA. The history of IgG glycosylation and where we are now. Glycobiology. 2019;30(4):202–13.

    Article  PubMed Central  Google Scholar 

  60. Baković MPi, Selman MHJ, Hoffmann M, Rudan I, Campbell H, Deelder AM, et al. High-throughput IgG fc N–Glycosylation profiling by Mass Spectrometry of Glycopeptides. J Proteome Res. 2013;12(2):821–31.

    Article  Google Scholar 

  61. Subedi Ganesh P, Barb Adam W. The structural role of antibody N-Glycosylation in receptor interactions. Structure. 2015;23(9):1573–83.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  62. Rudman N, Gornik O, Lauc G. Altered N-glycosylation profiles as potential biomarkers and drug targets in diabetes. FEBS Lett. 2019;593(13):1598–615.

    Article  CAS  PubMed  Google Scholar 

  63. Tanigaki K, Sacharidou A, Peng J, Chambliss KL, Yuhanna IS, Ghosh D, et al. Hyposialylated IgG activates endothelial IgG receptor FcγRIIB to promote obesity-induced insulin resistance. J Clin Investig. 2018;128(1):309–22.

    Article  PubMed  Google Scholar 

  64. Paton B, Suarez M, Herrero P, Canela N. Glycosylation Biomarkers Associated with Age-Related diseases and current methods for Glycan Analysis. Int J Mol Sci. 2021;22(11):5788.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  65. Ma B, Simala-Grant JL, Taylor DE. Fucosylation in prokaryotes and eukaryotes. Glycobiology. 2006;16(12):R158–84.

    Article  Google Scholar 

  66. Crook MA, Pickup JC, Lumb PJ, Georgino F, Webb DJ, Fuller JH, et al. Relationship between plasma sialic acid concentration and microvascular and macrovascular complications in type 1 diabetes: the EURODIAB complications Study. Diabetes Care. 2001;24(2):316–22.

    Article  CAS  PubMed  Google Scholar 

  67. Yerlikaya F, Toker A, Çiçekler H, Arıbaş A. The association of total sialic acid and malondialdehyde levels with metabolic and anthropometric variables in obesity. Biotech Histochem. 2015;90(1):31–7.

    Article  CAS  PubMed  Google Scholar 

  68. Li J, Hsu H-C, Mountz JD, Allen JG. Unmasking fucosylation: from cell adhesion to immune system regulation and diseases. Cell Chem Biology. 2018;25(5):499–512.

    Article  Google Scholar 

  69. Drouard G, Hagenbeek FA, Ollikainen M, Zheng Z, Wang X, Ripatti S et al. Twin study provides heritability estimates for 2,321 plasma proteins and assesses missing SNP heritability. medRxiv. 2024:2024.04.24.24306270.

  70. Sun BB, Chiou J, Traylor M, Benner C, Hsu Y-H, Richardson TG, et al. Plasma proteomic associations with genetics and health in the UK Biobank. Nature. 2023;622(7982):329–38.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  71. Menni C, Keser T, Mangino M, Bell JT, Erte I, Akmačić I, et al. Glycosylation of Immunoglobulin G: role of genetic and epigenetic influences. PLoS ONE. 2013;8(12):e82558.

    Article  PubMed  PubMed Central  Google Scholar 

  72. Zaytseva OO, Freidin MB, Keser T, Štambuk J, Ugrina I, Šimurina M, et al. Heritability of human plasma N–Glycome. J Proteome Res. 2020;19(1):85–91.

    Article  CAS  PubMed  Google Scholar 

  73. Kim SY, Yoo DM, Kwon MJ, Kim JH, Kim JH, Bang WJ et al. Differences in Nutritional Intake, Total Body Fat, and BMI score between twins. Nutrients. 2022;14(17).

  74. Zheng K, Gao W, Cao W, Lv J, Yu C, Wang S, et al. Education, income, and obesity: a nationwide Chinese twin study. Obesity. 2022;30(4):931–42.

    Article  PubMed  Google Scholar 

  75. Berntzen B, Jukarainen S, Kataja M, Hakkarainen A, Lundbom J, Lundbom N, et al. Physical activity, cardiorespiratory fitness, and metabolic outcomes in monozygotic twin pairs discordant for body mass index. Scand J Med Sci Sports. 2018;28(3):1048–55.

    Article  CAS  PubMed  Google Scholar 

  76. Vámosi ME, Heitmann BL, Thinggaard M, Kyvik KO. Parental care in childhood and obesity in adulthood: a study among twins. Obesity. 2011;19(7):1445–50.

    Article  PubMed  Google Scholar 

  77. Kupila SK, Berntzen BJ, Muniandy M, Ahola AJ, Kaprio J, Rissanen A, et al. Mental, physical, and social well-being and quality of life in healthy young adult twin pairs discordant and concordant for body mass index. PLoS ONE. 2023;18(12):e0294162.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  78. Hales CM, Gu Q, Ogden CL, Yanovski SZ. Use of prescription medications associated with weight gain among US adults, 1999-2018: a nationally representative survey. Obesity. 2022;30(1):229–39.

    Article  PubMed  Google Scholar 

  79. Perez-Riverol Y, Bai J, Bandla C, García-Seisdedos D, Hewapathirana S, Kamatchinathan S, et al. The PRIDE database resources in 2022: a hub for mass spectrometry-based proteomics evidences. Nucleic Acids Res. 2021;50(D1):D543–52.

    Article  PubMed Central  Google Scholar 

Download references

Acknowledgements

We thank the twin pairs for their invaluable contribution to this study. The Obesity Research Unit team and the staff at the Finnish Twin Cohort Study are acknowledged for assistance with data collection.Open access funded by Helsinki University Library.

Funding

Open Access funding provided by University of Helsinki (including Helsinki University Central Hospital).

Kirsi H. Pietiläinen was funded by the Research Council of Finland (grant numbers 335443, 314383, 272376 and 266286), Novo Nordisk Foundation (grant numbers NNF20OC0060547, NNF17OC0027232 and NNF10OC1013354), Finnish Diabetes Research Foundation, Sigrid Jusélius Foundation, Finnish Medical Foundation, Gyllenberg Foundation, Paulo Foundation, Finnish Foundation for Cardiovascular Research, Government Research Funds, University of Helsinki and Helsinki University Hospital. Risto Renkonen was also funded from Government Research Funds, University of Helsinki and Helsinki University Hospital. Sini Heinonen was funded by the Research Council of Finland (grant number 338417), Novo Nordisk Foundation (grant number NNF23SA0083953), Finnish Diabetes Research Foundation, Paulo Foundation, Orion Foundation, Paavo Nurmi Foundation and Government Research Funds, University of Helsinki and Helsinki University Hospital. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author information

Author notes

  1. Maheswary Muniandy and Sakari Joenväärä contributed equally to this work and share first authorship.

Authors and Affiliations

  1. Obesity Research Unit, Research Program for Clinical and Molecular Metabolism, Faculty of Medicine, University of Helsinki, Helsinki, Finland

    Maheswary Muniandy, Birgitta W. van der Kolk, Hanna Haltia, Sina Saari, Sini Heinonen & Kirsi H. Pietiläinen

  2. Transplantation Laboratory, Faculty of Medicine, University of Helsinki, Helsinki, Finland

    Sakari Joenväärä, Tiialotta Tohmola & Risto Renkonen

  3. HUS Diagnostic Center, Helsinki University Hospital, Helsinki, Finland

    Sakari Joenväärä, Tiialotta Tohmola & Risto Renkonen

  4. Department of Radiology, HUS Diagnostic Center, Helsinki University Hospital, Helsinki, Finland

    Antti Hakkarainen, Jesper Lundbom & Juho Kuula

  5. Faculty of Medicine Doctoral Program in Clinical Research, University of Helsinki, Helsinki, Finland

    Juho Kuula

  6. Institute for Clinical Diabetology, German Diabetes Center, Leibniz Center for Diabetes Research, Heinrich Heine University, Düsseldorf, Germany

    Jesper Lundbom

  7. Folkhälsan Institute of Genetics, Folkhälsan Research Center, Helsinki, Finland

    Per-Henrik Groop

  8. Research Program for Clinical and Molecular Metabolism, Faculty of Medicine, University of Helsinki, Helsinki, Finland

    Per-Henrik Groop

  9. Abdominal Center, Nephrology, University of Helsinki and Helsinki University Hospital, Helsinki, Finland

    Per-Henrik Groop

  10. Department of Diabetes, Central Clinical School, Monash University, Melbourne, VIC, Australia

    Per-Henrik Groop

  11. Institute for Molecular Medicine Finland, FIMM, University of Helsinki, Helsinki, Finland

    Jaakko Kaprio

  12. Healthy Weight Hub, Abdominal Center, Helsinki University Hospital and University of Helsinki, Helsinki, Finland

    Kirsi H. Pietiläinen

Authors
  1. Maheswary Muniandy
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. Sakari Joenväärä
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Birgitta W. van der Kolk
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Tiialotta Tohmola
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Hanna Haltia
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Sina Saari
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Antti Hakkarainen
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Jesper Lundbom
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Juho Kuula
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Per-Henrik Groop
    View author publications

    You can also search for this author inPubMed Google Scholar

  11. Jaakko Kaprio
    View author publications

    You can also search for this author inPubMed Google Scholar

  12. Sini Heinonen
    View author publications

    You can also search for this author inPubMed Google Scholar

  13. Risto Renkonen
    View author publications

    You can also search for this author inPubMed Google Scholar

  14. Kirsi H. Pietiläinen
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Maheswary Muniandy Formal Analysis (lead); Software (lead); Visualization (equal); writing – original draft (equal); writing – review and editing (equal). Sakari Joenväärä Investigation (lead); Methodology (lead); Visualization (equal); writing – original draft (equal); data curation (lead); writing – review and editing (equal). Birgitta W. van der Kolk Visualization (supporting); writing – original draft (equal); writing – review and editing (equal). Tiialotta Tohmola Investigation (supporting); writing – review and editing (supporting). Hanna Haltia original draft (equal); Visualization (supporting); writing – review and editing (equal). Sina Saari writing – review and editing (supporting) Antti Hakkarainen Investigation (supporting); writing – review and editing (supporting). Jesper Lundbom Investigation(supporting); writing – review and editing (supporting). Juho Kuula Investigation (supporting); writing – review and editing (supporting). Per-Henrik Groop Investigation (supporting); writing – review and editing (supporting). Jaakko Kaprio Resources (supporting); Methodology (supporting); Investigation (supporting); writing – review and editing (supporting). Sini Heinonen Investigation (supporting); writing – review and editing (supporting). Risto Renkonen Resources (equal); Supervision (supporting); Conceptualization (equal); Methodology (supporting); writing – review and editing (supporting). Kirsi H. Pietiläinen Supervision (lead); Resources (equal); Conceptualization (equal); Methodology (lead); writing – review and editing (supporting); Funding Acquisition All authors reviewed the manuscript.

Corresponding author

Correspondence to Maheswary Muniandy.

Ethics declarations

Ethical approval

The study, approved by the Ethics Committee of the Helsinki University Central Hospital (protocol number 270/13/01/2008), adhered to the Declaration of Helsinki principles.

Consent to participate

All participants provided written informed consent.

Consent for publication

The authors affirm that human research participants provided informed consent for publication.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Material 1:

Supplementary Methods

Supplementary Material 2: Supplementary Fig. 1

. summarizes the proteomics and glycoproteomics workflows. For the proteomics experiment, peptides were quantified and identified by LC-UDMSE runs. For the glycoproteomics, the N-glycopeptides were first quantified with LC-MSE runs and statistically significant glycopeptides were later targeted for fragmentation with LS-MS2 techniques and peptide sequence, and glycan compositions identified

Supplementary Material 3: Supplementary Fig. 2

. In total, 48 out of 230 proteins were differentially expressed between co-twins. The top 5 significant pathways from the Ingenuity pathway analysis (IPA) tool (P < 0.001) indicating which pathways are upregulated or downregulated in the heavier co-twins compared to the leaner co-twins. IPA z-scores for the pathways (calculated based on Fisher Exact test): upregulation (z-score > 0, red bar), downregulation (z-score < 0, blue bar), or no directionality (z-score = 0, grey bar). Here, all top 5 pathways were upregulated and so are all shaded in red with darker shades indicating increased upregulation. Results are ranked in order of statistical significance

Supplementary Material 4: Supplementary Fig. 3

. Standardized coefficients (β) showing associations (in standard deviations) between the clinical measurements and percentage of protein sialylation or protein fucosylation. The colors of the heatmap indicate the strength of the association i.e. red indicates a positive and blue a negative association. FDR p < 0.05 was considered significant, * FDR p < 0.05, ** FDR p < 0.01, *** FDR p < 0.001

Supplementary Material 5: Supplementary Table 1

. All within-twin pair differences in plasma proteins between the heavier and leaner cotwins in the BMI-discordant twin pairs. Supplementary Table 2. All within-twin pair differences in the glycopeptides in the BMI-discordant twin pairs (n = 48 pairs/96 individuals). Supplementary Table 3. Associations between protein expression and clinical measurements in the BMI-discordant twin pairs (n = 48 pairs/96 individuals). Supplementary Table 4. Associations between glycopeptide expression and clinical measurements in the BMI-discordant twin pairs (n = 48 pairs/96 individuals). Supplementary Table 5. Difference in percentage of fucosylation/sialylation across all proteins in the BMI-discordant twin pairs (n = 48 pairs/96 individuals). Supplementary Table 6. Difference in percentage of fucosylation/sialylation across each protein in the BMI-discordant twin pairs (n = 48 pairs/96 individuals). Supplementary Table 7. Associations between clinical measures to percentage of fucosylation/sialylation per protein in the BMI-discordant twin pairs (n = 48 pairs/96 individuals)

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Muniandy, M., Joenväärä, S., van der Kolk, B.W. et al. Plasma N-Glycoproteomics in monozygotic twin pairs discordant for body mass index reveals an obesity signature related to inflammation and iron metabolism. Biol Direct 20, 31 (2025). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13062-025-00609-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s13062-025-00609-y

Keywords

Biology Direct

ISSN: 1745-6150

Contact us