Sulfated N-glycans Upregulation in Sera Predicts Early-Stage Breast Cancer in Patients

Analysis of sulfated N-glycans using glycoblotting-based sulphoglycomics workflow

Previously developed methods for the glycoblotting-based sulphoglycomics workflow ³⁰ (Supplementary Figure S1) were employed to analyze sulfated N-glycans from human serum samples. This approach enables the integration of glycoblotting-based on-bead glycan enrichment, weak acid group-selective methylation, and labeling on a single platform, coupled with weak anion exchange (WAX) separation and matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) MS. First, the glycoproteins in the serum samples were digested with PNGase F to release the N-glycans. Subsequently, glycoblotting, a chemically selective enrichment technique, was used to specifically isolate N-glycans. This offers a straightforward and streamlined approach to enhance the detection and analysis of sulfated N-glycans. Notably, the glycoblotting method demonstrated a higher glycan recovery efficiency from biological samples. ³⁰ This means glycoblotting is more effective in separating and enriching sulfated N-glycans from the sample. To eliminate any potential interference during subsequent WAX separation and MALDI-TOF MS, glycoblotting utilizes 3-methyl-1-p-tolyltriazene (MTT) for the methyl esterification of the carboxylic group of sialic acid, whereas the sulfated glycans remain unmethylated. ^{27, 32, 33} Therefore, only sulfated glycans retain their negative charges, enabling selective subsequent enrichment using WAX separation and detection in negative mode. WAX separation was used for selective enrichment of sulfated glycans.

This method employs a 3-aminopropyl silica gel microcolumn, which exhibits characteristics similar to those of weak anion-exchange separation methods. ²⁶ At neutral pH, the primary amine groups in silica gel act as proton acceptors, resulting in a positive charge. This property is crucial for the retention of negatively charged sulfated glycans. Additionally, the sialic acid groups of the glycans were neutralized in advance by methyl esterification with MTT during glycoblotting. This facilitates the separation of sulfated glycans from the more abundant sialylated glycans, which is typically challenging when analyzing low-abundance sulfated glycans using MALDI-TOF MS in negative mode. Before MS analysis, using WAX separation method, neutral and sialylated N-glycans were initially eluted with a solution of 1% AcOH in 50% MeOH. Subsequently, sulfated N-glycans were eluted using a solution of 1% NH₄OH in 5% MeOH (pH 10.5).

MALDI-TOF MS analysis of sulfated N-glycans from healthy human serum samples

As an initial step, MALDI-TOF MS analysis of sulfated N-glycans from healthy human serum samples was performed in negative mode using the glycoblotting-based sulphoglycomics approach. Based on these results, we successfully separated trace sulfated N-glycans from abundant non-sulfated N-glycans (Figure 1). Before enrichment, all non-sulfated N-glycans, including sialylated N-glycans, present in the serum samples were observed, whereas sulfated N-glycans were not identified (Figure 1A). After enrichment, thirteen sulfated N-glycans were detected in the tested samples as [M-H]^- ions (Figure 1B). This result clearly shows that MALDI-MS suppressed sulfated glycans in the presence of abundant non-sulfated N-glycans. Among the sulfated N-glycans, in addition to mono-sulfated N-glycans, several di-sulfated N-glycans have also been identified. These glycans are known to significantly enhance binding interactions with numerous glycan-binding proteins, thereby promoting cancer cell migration and invasion. ³⁴ No phosphorylated N-glycans were detected.

• Download figure
• Open in new tab

Figure 1.

MALDI-TOF MS profiles of BOA-labeled sulfated N-glycans from healthy human serum A) before and B) after enrichment using glycoblotting-based sulphoglycomics workflow. I.S. indicates internal standard.

Compared to non-sulfated N-glycans, the identified sulfated N-glycans are predominantly complex-type structures, with non-reducing terminal sulfo-LacNAc, sulfo-LacdiNAc, sulfo-sialyl-LacNAc, sulfo-lewis-type, and sulfo-sialyl-lewis-type glycan epitopes on N-glycans expressed in human serum samples. Lewis-type epitopes have not been identified for non-sulfated N-glycans (because mono-fucosylated N-glycans are difficult to distinguish from core fucose). In contrast, the major sulfated N-glycans contained multiple fucose residues and presented Lewis-type epitope terminals. Sulfated N-glycans without Lewis-type epitopes also exhibited signal intensity patterns, independent of non-sulfated N-glycans. These findings demonstrate the applicability of the developed method for the analysis of sulfated glycans in serum samples. This method holds promise for identifying new sulfated glycan biomarkers that could be applied in the early detection of BC.

Quantitative analysis of serum sulfated N-glycans between healthy control individuals and BC patients

We then performed a comparative analysis of the sulfated N-glycan profiles found in the serum samples of 20 normal controls (NC) and 76 BC patients to determine whether these glycans could serve as potential biomarkers for the disease. The clinical information of the study participants is summarized in Supplementary Table S1. Figure 2 shows the representative MALDI-TOF MS profiles of BOA-labeled sulfated N-glycans from BC patients and healthy NC serum samples enriched using the glycoblotting method and then fractioned using WAX. Eight mono-sulfated and five di-sulfated N-glycans were detected in all study samples (Supplementary Table S2). These glycans were selected after each peak was evaluated for their quantitative reproducibility using serially diluted standard serum samples (0.2x, 0.4x, 0.6x, 0.8x, 1x, and 1.2x). The peak area of each glycan was normalized to the concentration of the internal standard, and a standard calibration curve was plotted for each glycan (Supplementary Figure S2). Accordingly, glycan peaks with good slope linearity and minimal outlier scores were selected and considered for further statistical analysis.

• Download figure
• Open in new tab

Figure 2.

Representative MALDI-TOF MS spectra of BOA-labeled sulfated N-glycans from A) serum BC patients and B) healthy controls. I.S. indicates the internal standard mono-sulfated biantennary complex-type N-glycan (129 pmol).

An internal standard was used to quantitatively analyze the identified sulfated N-glycans in both patients and controls. The internal standard used was mono-sulfated desialylated biantennary complex-type N-glycan (129 pmol), which was synthesized using chemoenzymatic methods from a sialoglycopeptide (SGP). ³⁵ A detailed list of all the sulfated N-glycans detected is presented in Supplementary Table S2, along with their corresponding glycan compositions and structures. The results showed that the sulfated N-glycans present in the human serum sample were predominantly complex type, and approximately 1% of the N-glycans were hybrid type. Among complex-type N-glycans, bi-antennary N-glycans comprised the highest proportion (96%), tri- and tera-antennary each comprised 2% and 1%, respectively, and most were fucosylated (≥ 75%) and sialylated (≥ 66%). Comparative analysis revealed a remarkable change in the concentration of individual sulfated N-glycans in human serum samples.

The expression levels of the identified sulfated glycans were determined by calculating the peak area of each glycan after normalization to the known concentration of the internal standard used, as shown in Figure 3. Subsequently, these values were compared between NC and BC groups. An independent t-test was used to identify glycans with differential expression patterns between the groups. The results revealed that seven mono-sulfated and two di-sulfated N-glycans were significantly upregulated in the sera of BC patients compared to the control group, as shown in Table 1. All peaks corresponding to sulfated glycans exhibited a notable increase in abundance (p ≤ 0.05) in the sera of BC patients compared with the NC group.

• Download figure
• Open in new tab

Figure 3.

Expression levels and structures of total sulfated N-glycans isolated from NC and BC serum glycoproteins. I.S. indicates an internal standard, mono-sulfated biantennary complex-type N-glycan (129 pmol).

Evaluation of the diagnostic performance of serum sulfated N-glycans

To further examine the diagnostic performance of serum sulfated N-glycans, receiver operating characteristic (ROC) test analysis was employed for glycans that showed statistically significant (p ≤ 0.05) changes in expression through t-test analysis. The area under the curve (AUC) values generated from the ROC test were used to evaluate the discriminative potential to differentiate between the cancer patients and control groups. The discriminating potential of candidate biomarkers that met both criteria (p ˂ 0.05, AUC ≥ 0.8) was considered a candidate biomarker. Consequently, seven serum sulfated N-glycans showed remarkable diagnostic performance and were considered as candidate biomarkers (Table 2). Notably, these glycan structures are mono-sulfated N-glycans. All seven sulfated glycans were strong predictors to distinguish patients with stage I disease from those in the NC group. These candidates were mono-sulfated LacNAc (m/z 1809, 2077, 2272, and 2450), LacdiNAc (m/z 2171), and mono- and bi-sialylated (m/z 2276, 2581)-terminated N-glycans. Three sulfated glycans (m/z 1809, 2077, and 2272) distinguished the stage II patients from the NC group. No significant candidate biomarker was observed for stage III BC patients, whereas four glycans (m/z 2077, 2171, 2276, and 2581) fulfilled the criteria for being considered as biomarkers for stage IV patients. Moreover, three sulfated N-glycans (m/z 2077, 2276, and 2581) demonstrated the same ability to discriminate all BC patients from the NC group.

Multiple comparisons were conducted to visualize the abundance of sulfated N-glycans expressed among the different BC stages and the NC group. Figure 4 shows the dot plot and ROC curves for the identified candidate biomarkers related to stage I and stage II BC patients compared with the NC group. The results revealed that a significant increase in glycan abundance was observed primarily during the early stages of the disease.

• Download figure
• Open in new tab

Figure 4.

The dot plot illustrates the expression levels of serum sulfated N-glycans that were upregulated in BC patients compared to healthy NC, whereas the ROC curve displays the discriminatory power of each glycan in distinguishing between patients and NC. The area under the curve (AUC) value indicates the effectiveness of glycan as a diagnostic marker: A) Discriminate BC-I from NC; B) Discriminate BC-II from NC.

Serum glyco-subclass patterns as predictive indicators for early-stage BC

In addition to assessing and comparing the levels of individual glycans between the BC patients and NC groups, we assessed the patterns of serum glycan subclasses. Specifically, we quantified and compared the levels of serum sulfated N-glycans based on their structural similarities between the healthy and cancer groups. Our findings revealed statistically significant variations in the abundance of glycan subtypes, such as fucosylated, bi-antennary, tri-antennary, tetra-antennary, mono-, bi-, and total-sialylated, and total-sulfated N-glycans, in the serum of patients with stage I and II disease (Figure 5). Based on ROC analysis, fucosylated, bi-antennary, mono- and total-sialylated, and total-sulfated were diagnostic indicators of stage I patients. These glycan subtypes demonstrated high AUC values of 0.979, 0.932, 0.944, 0.892, and 0.941, respectively (p ˂ 0.0001). In stage II patients, fucosylated and tetra-antennary glyco-subclasses had higher AUC values (0.945 and 0.862, respectively). The increased abundance of these specific glycotypes in the early stages of the disease aligns with the expression patterns of individual sulfated N-glycan biomarkers.

• Download figure
• Open in new tab

Figure 5.

The serum glycan expression patterns of glyco-subclasses shared similar structures between the BC stages and NC group. The area under the curve (AUC) value was used to indicate the effectiveness of glycan subclasses as diagnostic markers for early stage BC.

Study population and sample collection

Human serum samples from female BC patients (n=76) at stages I-IV and age-matched healthy individuals (n=20) were collected from Ethiopia during 2015-2016. Stored serum samples were used in this study. Informed consent was obtained from all study participants and the study was conducted in accordance with the ethical standards outlined in the Declaration of Helsinki. The ethical review boards of Addis Ababa University, School of Medicine, Ethiopia, and Hokkaido University, Faculty of Advanced Life Sciences, Japan approved the research protocols. To ensure sample integrity, serum samples were frozen in plastic vials at −80 °C and carefully packed with dried CO₂ in a foam box. Subsequently, the serum samples were transported by airplane to Japan within 72 hours for further analysis.

Materials

Peptide N-glycosidase F (PNGase F) was obtained from New England BioLabs (Ipswich, MA, USA). Ammonium bicarbonate (ABC), 3-methyl-1-p-tolyltriazene (MTT), 2,5-hydroxybenzoic acid (DHB), benzyloxyamine hydrochloride (BOA-HCl), sodium bicarbonate (NaHCO₃), trifluoroacetic acid (TFA), and 3,4-diamino benzophenone (DABP) were purchased from Tokyo Chemical Industry Co. (Tokyo, Japan). BlotGlyco H beads were purchased from Sumitomo Bakelite, Co. Ltd. (Tokyo, Japan), proteinase K was from Roch (Germany), and trypsin was from Sigma-Aldrich Corp. (St. Louis, MO, USA). MultiScreen Slvinert^R filter plates were purchased from Millipore Co., Inc. (Tokyo, Japan).

N-Glycan release from Human Serum Glycoprotein ^{30, 40}

Whole serum from each sample was subjected to enzymatic pretreatment to generate glycans with a reducing terminal. This enzymatic pretreatment facilitated subsequent chemical ligation with hydrazide-functionalized BlotGlyco^® H beads during the glycoblotting process. Hence, 15 µL of human serum and 10 µL of a 12.9 µM mono-sulfated desialylated N-glycan prepared from SGP (Tokyo Chemical Industry Co., Ltd.), used as an internal standard, were mixed with 50 µL of 200 mM NH₄HCO₃. 4 µL of denaturing buffer (5% sodium dodecyl sulfate, 0.4 M dithiothlotol) was added to denature the sample, followed by heating at 100 °C for 10 minutes. Subsequently, 10 µL of 123 mM iodoacetamide (IAA) was added and incubated at room temperature in the dark for 1 hour. The mixture was then subjected to digestion by adding 10 µL of trypsin (40 U/ µL) dissolved in 1 mM hydrochloric acid (HCl) and incubated overnight at 37 °C. Afterward, enzyme activity was halted by heating at 90 °C for 10 minutes and allowed to cool to room temperature. To initiate further processing, 8 µL of reaction buffer (0.5 M Na₃PO₄, pH 7.5) and 10% NP-40 were sequentially added and incubated for 10 minutes at 37 °C. N-Glycans were released using 2 µL of PNGase F (New England BioLabs) with an activity of 5 U/μL, followed by overnight incubation at 37 °C. For further digestion,10 µL proteinase K (0.5 U/ µL) was added and incubated for 3 hours at 37 °C. The enzyme was then heat inactivated at 90 °C for 10 minutes. The resulting sample was dried using SpeedVac and stored at −20 °C until further use.

Enrichment of glycans using glycoblotting^{30, 40–42}

A 250 µL suspension of BlotGlyco^® H beads (Sumitomo Bakelite Co., Ltd.) with a concentration of 10 mg/mL in water was aliquoted into 96 wells of a MultiScreen Solvinert filter plate (Millipore, Billerica, MA, USA), and the water was removed by applying vacuum. The dried sample containing released N-glycans was reconstituted by adding 20 µL of Milli-Q water. The reconstituted sample (20 µL) was added to each well, followed by 180 µL of 2% acetic acid in acetonitrile (AcOH/ACN). The plate was incubated at 80 °C for 45 minutes. Following the incubation, two successive washing steps were carried out using 200 µL each of a 2 M guanidine-HCl in 16.6 mM NH₄HCO₃, water, and a 1% solution of triethylamine in methanol (MeOH). The unreacted hydrazide functional groups were acetyl-capped by adding 100 µL 10% acetic anhydride in MeOH at room temperature for 30 minutes. Any residual acetic anhydride was removed under a vacuum. Each well was washed twice with 10 mM HCl, MeOH, and dioxane. For on-bead methyl esterification, 100 µL of 100 mM MTT in dioxane was added and the mixture was incubated at 60 °C for 90 minutes. Subsequently, two washing steps were performed using 200 µL of dioxane, water, MeOH, and water. To facilitate effective labeling through a trans-iminization reaction, 20 µL of 50 mM BOA-HCl and 180 µL of 2% AcOH/ACN solution were sequentially added. The mixture was then incubated at 80 °C for 45 minutes. N-Glycans labeled with BOA were eluted twice with water (150 µL). The resulting glycan solution was dried using SpeedVac and stored at −20 °C until further use.

Enrichment of sulfated N-glycans using WAX ^{29, 30}

Enrichment of sulfated N-glycans was performed according to previously reported methods. ³⁰ A small cotton plug was inserted into a 200 microliters micropipette tip. Subsequently, 50 µL 3-aminopropyl silica gel (Tokyo Chemical Industry Co. Ltd., 100 mg/mL suspension) was loaded onto the micropipette tip on top of a cotton plug. The gel was allowed to settle under gravity. The packed WAX microcolumn was conditioned and washed twice with 100 µL water, ACN, and 1% AcOH in 95% ACN. After each washing step, the mixture was centrifuged at 500 rpm for 2 minutes. To reconstitute the BOA-labeled N-glycans obtained from glycobolotting, 20 µL of MilliQ water was used. A 5 µL sample aliquot was then dissolved in 150 µL of 1% AcOH in 95% ACN, which was then loaded into the column. The sample was allowed to elute by gravity. The collected eluate was reloaded into the column, and this process was repeated three times to ensure adequate interaction. The column was washed with 150 μL of 1% AcOH in 95% ACN, followed by centrifugation at 500 rpm for 2 minutes to remove unbound and hydrophobic contaminants. Then, 150 µL of 1% AcOH in 50% ACN was added to elute BOA-labeled neutral and methylated sialylated N-glycans. Elution was performed twice, and after each elution, centrifugation was conducted at 500 rpm for 2 minutes. Sulfated N-glycans were eluted using 150 µL 1% NH₄OH in 5% ACN (pH 10.5). As in the previous steps, this elution was repeated twice, and centrifugation at 500 rpm for 2 minutes was performed. Finally, the resulting glycan solution was dried using SpeedVac and stored at −20 °C until use.

MALDI-TOF MS analysis

Mass spectrometric analysis of BOA-labeled sulfated N-glycans was performed using an Ultraflex III (Bruker, Germany) equipped with a reflector in negative ion mode. DABP (10 mg/mL) in 75% ACN with 0.1% TFA was used as a matrix throughout this work^{25, 43}, from which 0.5 µL of each sample and matrix solution were spotted onto an MTP 384 target plate (polished steel TF, Bruker Daltonics) and dried at room temperature. To ensure reproducibility, each sample was spotted in four replicates, and each spectrum was generated by accumulating 1000 laser shots. The average of the four normalized data points for each sample was used for statistical analysis. Data acquisition and processing were performed using the Flexanalysis 3.0 software (Bruker, Germany). The intensity of the monoisotopic peak for each sulfated N-glycan was normalized using 12.9 µM of an internal standard (mono-sulfated biantennary complex-type N-glycans synthesized using chemoenzymatic methods³⁵). The normalized data were used for further statistical analysis and quantitative comparison of sulfated N-glycan profiles. The structural compositions of sulfated N-glycans were assigned using the Expasy GlycoMod Tool and Glyconnect Database provided by the Swiss Institute of Bioinformatics (https://web.expasy.org/glycomod/), and GlycoWorkbench. ^{44, 45} GraphPad Prism 5 software was used for data analysis and graph plotting.

Statistical Analysis

A comparison between the two groups (NC vs. entire BC group) was performed using an independent sample t-test. Multiple comparisons among the NC and clinical stage groups were conducted using Bonferroni one-way analysis of variance (ANOVA). To evaluate the diagnostic potential of individual glycans or glyco-subclasses that showed significant differences, a receiver operating characteristic (ROC) test was performed. The area under the curve (AUC) value generated from the ROC test was used as a measure of the diagnostic accuracy of potential glycan biomarkers. AUC values within the ranges of 0.9-1, 0.8-0.9, 0.7-0.8, and <0.7 are classified as “highly accurate,” “accurate,” “moderately accurate,” and “uninformative tests,” respectively. Differences in mean values were considered significant at a 95% confidence interval (p ≤ 0.05).