Therapy-induced remodeling of IgG and plasma N-glycans in breast cancer

Abstract

Background

Protein N-glycosylation plays a key role in cancer biology and may offer insights into tumor behavior and treatment response. This study investigated changes in N-glycosylation of immunoglobulin G (IgG) and total plasma proteins in patients with breast cancer undergoing neoadjuvant chemotherapy.

Methods

A prospective cohort of 34 women with early-stage or locally advanced breast cancer was recruited. Plasma samples were collected before and after neoadjuvant chemotherapy. IgG was isolated by immunoaffinity chromatography, and IgG and total plasma N-glycans were enzymatically released, fluorescently labeled, and analyzed using ultra-high performance liquid chromatography. Glycan traits were expressed as relative abundances and summarized into derived structural features. Longitudinal changes were assessed using linear mixed-effects models adjusted for age and body mass index.

Results

Chemotherapy induced significant decrease in IgG core fucosylation. This effect varied by treatment: anthracycline-based regimen led to decreased core fucosylation and digalactosylation and increased monogalactosylation. The greatest reduction in core fucosylation was observed in patients treated with docetaxel and cyclophosphamide. HER2-targeted therapy was associated with decreased bisecting N-acetylglucosamine. Plasma glycosylation remained largely stable, though oligomannose glycans increased in patients following chemotherapy. Tumor size was significantly associated with several plasma glycan traits, particularly digalactosylation and oligomannosylation.

Conclusions

IgG glycosylation patterns change in a treatment-specific manner during chemotherapy, potentially reflecting immune modulation. Plasma glycan traits are more stable but may reflect tumor burden. These results support the potential of glycan profiling as a biomarker for monitoring the impact of therapy and disease progression in breast cancer.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12885-025-15381-5.

Background

Breast cancer is the most commonly diagnosed malignancy and a leading cause of cancer-related mortality among women worldwide, with over two million new cases annually [1]. The incidence is increasing, particularly in North America, Australasia, and Western and Northern Europe, which are driven by lifestyle factors and expanded screening programs [2]. Despite improvements in early detection and survival rates, particularly in developed countries, clinical management remains limited by tumor heterogeneity and variable treatment responses [1].

Histologically, breast cancer includes invasive carcinoma of no special type (NST), lobular, mucinous, and tubular subtypes, with NST being the most prevalent [3]. Its pathogenesis involves a combination of genetic, hormonal, and lifestyle-related factors. While BRCA1 and BRCA2 mutations are well characterized, they account for only approximately 6% of cases. Other susceptibility genes, such as TP53, PALB2, ATM, CHEK2, and PTEN, contribute to risk to a lesser extent [4]. Hormonal exposures, including menopausal hormone therapy, have also been implicated [5, 6].

Advances in personalized medicine have allowed treatment decisions to be tailored to tumor biology and individual risk. Immunohistochemistry for estrogen receptor (ER), progesterone receptor (PR), HER2, and Ki67 enables molecular subtyping into luminal A, luminal B, HER2-positive, and triple-negative breast cancer (TNBC), each with distinct prognostic and therapeutic implications [7, 8]. Multigene expression panels such as Oncotype DX and Mammaprint are now integral to adjuvant therapy decisions, as supported by large trials including MINDACT and TAILORx [9].

While traditional serum biomarkers such as CA 15 − 3 and CEA are used in monitoring, they are insufficiently specific or sensitive for diagnosis or early relapse detection [10, 11]. Diagnosis and response assessment still rely on imaging and tissue biopsy, both of which are invasive or limited in frequency [12, 13]. Thus, there is a pressing need for novel, minimally invasive biomarkers capable of reflecting disease activity in real time.

Glycosylation, a posttranslational modification involving the enzymatic addition of glycans to proteins or lipids, is increasingly recognized for its role in cancer biology [14]. Altered glycosylation affects cell signaling, immune recognition, and metastatic potential. In breast cancer, shifts in glycan structures contribute to immune evasion, inflammation, and tumor progression [15]. Significant differences in glycosyltransferase activity have been observed between tumor and normal tissues [15–17], and glycosylated protein expression differs markedly between primary tumors and developed metastases [18].

A large proportion of circulating and tissue proteins are glycosylated, making glycoproteins promising candidates for noninvasive biomarker development [15, 16].

Aberrant protein glycosylation, particularly altered N-glycan branching, fucosylation, and sialylation, is a characteristic of breast cancer progression, metastasis and therapy resistance [16, 19, 20]. Specific plasma N-glycan traits, such as hypersialylation, oligomannosylation, increased branching, antennary fucosylation with elevated sialyl Lewis x (sLe^x) epitope correlate with tumor invasiveness, circulating tumor cell burden and metastatic capacity [19, 21–23]. Increased tetra- and hexa-fucosylated haptoglobin glycoforms have also been detected in breast cancer serum, improving differentiation between benign and malignant lesions when combined with classical tumor markers [24]. Immunoglobulin G (IgG), a major circulating glycoprotein, undergoes distinct Fc-glycosylation rearrangement during carcinogenesis and systemic inflammation [25–27]. Altered IgG Fc glycosylation influences antibody-dependent cellular cytotoxicity (ADCC) and inflammatory pathways by modulating Fcγ receptor binding [28]. Both total serum and IgG N-glycan profiling have emerged as promising minimally invasive approaches for cancer biomarker discovery. Elevated levels of branched, core fucosylated, and sialylated total serum N-glycans, along with core fucosylated and agalactosylated IgG structures, enable discrimination of early-stage breast cancer from healthy controls [26]. Dynamic glycosylation changes during chemotherapy and endocrine therapy reflect immune modulation and treatment response [20, 29], while IgG N-glycans have shown predictive value for therapeutic outcomes in gastric cancer [30] and chronic inflammatory diseases [31]. Together, these findings indicate that total plasma N-glycosylation and IgG glycosylation sensitively reflect tumor–host interactions and immune modulation, offering substantial potential for early detection, molecular subtyping and therapy monitoring in breast cancer. Nevertheless, variability among studies reflects tumor heterogeneity and supports the need for further glycomic analyses for reliable biomarker validation [17]. Building on existing evidence, this study investigates IgG and total plasma N-glycosylation patterns in breast cancer patients undergoing neoadjuvant chemotherapy. The analysis focuses on therapy-induced glycan changes and their associations with treatment regimens, clinical response, and tumor size.

Materials and methods

Study participants

This prospective cohort study enrolled 37 women aged 31–80 years with histologically confirmed, early-stage or locally advanced, nonmetastatic invasive breast cancer. Patients were recruited between 2021 and 2023 at the University Hospital for Tumors, Sestre Milosrdnice Clinical Hospital Center in Zagreb, Croatia. All patients were candidates for neoadjuvant chemotherapy on the basis of institutional treatment guidelines.

The inclusion criteria were female sex, age ≥ 18 years, newly diagnosed nonmetastatic invasive breast cancer, and eligibility for systemic neoadjuvant therapy.

The exclusion criteria included a prior history of breast cancer, ductal carcinoma in situ (DCIS), metastatic disease, or the luminal A subtype.

Tumor staging and diagnosis were based on comprehensive clinical and radiological assessments, including bilateral digital mammography, breast and axillary ultrasound, breast MRI, and ultrasound-guided core biopsy. Additional staging modalities (CT, PET/CT, or bone scan) were used as clinically indicated. In premenopausal patients, breast MRI was timed to the luteal phase of the menstrual cycle (days 5–15).

Systemic therapy commenced within 2–4 weeks of diagnosis and staging, in accordance with current evidence-based guidelines and tailored to tumor subtype, nodal status, and patient factors. Patients received standard neoadjuvant chemotherapy regimens including anthracycline-, taxane-, and cyclophosphamide-based protocols, with HER2-positive cases additionally treated with trastuzumab and pertuzumab. Detailed descriptions of all treatment regimens, dosing schedules, and monitoring procedures are provided in the Supplementary Appendix 1.

The response to therapy was monitored with ultrasound (after paclitaxel cycle 4) and MRI (after cycle 12), followed by surgery (either breast-conserving or mastectomy) with sentinel or axillary lymph node biopsy as appropriate. Tumors were staged according to the 8th edition of the AJCC/UICC TNM classification, including both tumor size (T1c–T4b) and nodal involvement (N0–N3c), and were grouped into stages IA to IIIC.

EDTA plasma samples for N-glycan profiling were obtained before starting chemotherapy and immediately prior to surgery, about one month after treatment completion. Although 37 patients were initially enrolled, samples from three were unavailable at either the first or second time point (T1 or T2), resulting in 34 patients included in the final analysis. The interval between the end of chemotherapy and sample collection was consistent across all patients.

Additional laboratory tests included complete blood count, clinical biochemistry, and measurement of tumor markers (CA 15 − 3, CEA). Clinical treatment data were recorded prospectively. All participants provided written informed consent, and the study was approved by the local ethics committee in accordance with the Declaration of Helsinki.

Methods

Isolation of IgG from plasma

IgG was isolated from human plasma via a high-throughput immunoaffinity protocol on 96-well protein G monolithic plates (BIA Separations, Slovenia), as previously described [32]. The plasma samples were diluted 1:7 in phosphate-buffered saline (PBS; Merck, Germany) and applied to the plate. After washing, the IgG was eluted with 0.1 M formic acid (Merck) and immediately neutralized with 1 M ammonium bicarbonate (Acros Organics, USA).

N-glycan release from IgG and total plasma proteins

The isolated IgG was vacuum-dried and denatured with 1.33% sodium dodecyl sulfate (SDS; Invitrogen, USA) at 65 °C for 10 min. Total plasma proteins were treated similarly with 2% SDS. Following denaturation, 4% (v/v) Igepal-CA630 (Sigma Aldrich, USA) was added, and the samples were shaken for 15 min. For each sample, N-glycans were enzymatically released through incubation with 1.2 U PNGase F (Promega, USA) at 37 °C overnight.

Fluorescent labeling and solid-phase extraction

The released N-glycans were fluorescently labeled with 2-aminobenzamide (2-AB) in 30% acetic acid in DMSO, followed by a 2-hour incubation at 65 °C. Excess label and reagents were removed via solid-phase extraction (SPE) with AcroPrep Advance 0.2 μm wwPTFE plates (Pall). The purified glycans were eluted with ultrapure water and stored at − 20 °C.

Hydrophilic interaction ultra-high performance liquid chromatography

The labeled glycans were separated via an Acquity UPLC H-Class system (Waters, USA) equipped with a fluorescence detector (excitation: 250 nm, emission: 428 nm). Separations were performed on a Waters BEH glycan column using 100 mM ammonium formate (pH 4.4) as mobile phase A and acetonitrile as mobile phase B.

For IgG glycans, a linear gradient of 75–62% acetonitrile was applied at 0.4 mL/min over 27 min. For total plasma glycans, a gradient of 70–53% acetonitrile at 0.56 mL/min over 25 min was used.

Chromatograms were processed via Empower 2 software (Waters). Peaks were manually adjusted to ensure consistent integration. The IgG glycan profiles were divided into 24 peaks (IGP1–IGP24), and the plasma profiles were divided into 39 peaks (GP1–GP39). The abundance was calculated as the percentage of the total chromatogram area. Six derived traits were computed for IgG and 16 for plasma proteins, reflecting key glycosylation features such as galactosylation, sialylation, fucosylation, oligomannosylationand bisecting N-acetylglucosamine.

Statistical analysis

The glycan peak areas were normalized via total area normalization and log-transformed via R software (version 4.4.1; R Foundation for Statistical Computing). The glycan peaks were then back-transformed to calculate the derived traits (Supplementary Table 1). To meet model assumptions of normality, all glycan traits were transformed using rank-based inverse normal (RANKL) transformation (mean – 0, SD – 1). Differences in glycan traits between timepoints (before and after treatment) were assessed via paired linear mixed-effects models, with time as a fixed effect and individual samples as random effects. Age and BMI were included as covariates. Pairwise contrasts were tested via the emmeans package [link: https://www.tandfonline.com/doi/abs/10.1080/00031305.1980.10483031], and P values were adjusted via Benjamini–Hochberg false discovery rate correction.

To quantify the power of between-therapy differences in the subgroup analysis, we computed η² from a one-way ANOVA (Δ ~ Therapy), converted to Cohen’s f, and calculated power with pwr.anova.test using the harmonic mean of the unbalanced group sizes (ACdd 22, TC 3, TCbHP 9 → ℎ = 6.12 per group; k = 3; α = 0.05).

Results

This study analyzed 34 paired samples from female patients with breast cancer (age range, 31–80 years) treated at the University Hospital for Tumors, Sestre Milosrdnice Clinical Hospital Center, Zagreb, Croatia, between 2021 and 2023. IgG and total plasma N-glycosylation were assessed before and after neoadjuvant chemotherapy via 24 directly measured and 6 derived IgG glycan traits and 39 directly measured and 16 derived plasma glycan traits (Supplementary Table 1, Supplementary Fig. 1).

Owing to structural similarities among glycans and the limited sample size, statistical analyses were performed on derived glycan traits, which reflect key glycosylation features such as galactosylation, sialylation, fucosylation, and bisecting GlcNAc. Traits were rank-transformed and analyzed via linear mixed-effects models adjusted for age and BMI, with patient ID as a random effect.

Analysis of longitudinal changes in IgG glycosylation revealed that only core fucosylation (CF) significantly decreased following chemotherapy (adjusted P = 0.0047), indicating a potential shift toward more immunostimulatory IgG profiles (Table 1, Fig. 1). No other IgG glycosylation features changed significantly over time. Similarly, stratification by treatment response revealed no significant differences between patients with partial (PR) or complete response (CR) (Supplementary Table 2).

Table 1.

Longitudinal changes in IgG glycan traits adjusted for age and BMI

Glycan Trait (IgG)	Effect (RANK)	Standard Error	P Value	Adjusted P Value
B	−0.1688	0.1205	0.1697	0.5090
CF	−0.3422	0.0933	0.0008	0.0047
G0	0.0734	0.1102	0.5094	0.6499
G1	0.0970	0.1574	0.5416	0.6499
G2	−0.1150	0.1022	0.2682	0.5364
S	0.0284	0.1191	0.8131	0.8131

B Bisected N-glycans, CF Core fucosylated N-glycans, G0 Agalactosylated N-glycans, G1 Monogalactosylated N-glycans, G2 Digalactosylated N-glycans, S Sialylated N-glycans

[P < 0.05] – Bold

Fig. 1.

Longitudinal changes in IgG glycosylation traits before and after treatment. Y-axis shows relative abundance values normalized to baseline. The bold lines represent the mean changes across patients. Each thin line represents an individual patient, color-coded by clinical response status. Dashed colored lines indicate mean changes within each response group, and the solid black line denotes the overall cohort mean change. B, bisected N-glycans; CF, core fucosylated N-glycans; G0, agalactosylated N-glycans; G1, monogalactosylated N-glycans; G2, digalactosylated N-glycans; S, sialylated N-glycans

While the overall longitudinal analysis revealed changes primarily in IgG core fucosylation, further stratification by chemotherapy regimen uncovered additional, treatment-specific glycosylation patterns (Table 2, Fig. 2). In patients receiving the ACdd regimen (n = 22), core fucosylation (CF) and digalactosylation (G2) significantly decreased, whereas monogalactosylation (G1) increased (adjusted P = 0.0476). The TC regimen (n = 3) showed the most pronounced reduction in CF (adjusted P = 0.0267). Additional shifts in sialylation, digalactosylation, and agalactosylation (G0) were observed in this group but did not reach statistical significance, likely due to the limited sample size. In the TCbHP group (n = 9), a significant decrease in bisecting GlcNAc (B) was observed (adjusted P = 0.0476), while other traits remained unchanged.

Table 2.

Therapy-specific changes in IgG glycan traits across two timepoints

Therapy	Glycan Trait (IgG)	Time Effect (RANK)	Time Standard Error	Time P Value	Time-Adjusted P Value
ACdd	B	0.0270	0.1402	0.8483	0.8483
	CF	−0.2964	0.1099	0.0103	0.0476
	G0	0.1508	0.1289	0.2490	0.3735
	G1	0.4489	0.1728	0.0132	0.0476
	G2	−0.3003	0.1143	0.0123	0.0476
	S	−0.1396	0.1388	0.3208	0.4442
TC	B	−0.2666	0.3768	0.4834	0.5801
	CF	−1.0114	0.2958	0.0015	0.0267
	G0	−0.7650	0.3460	0.0328	0.0761
	G1	−0.5887	0.4625	0.2103	0.3441
	G2	0.6755	0.3073	0.0338	0.0761
	S	0.9238	0.3727	0.0175	0.0526
TCbHP	B	−0.6193	0.2202	0.0078	0.0476
	CF	−0.2280	0.1723	0.1937	0.3441
	G0	0.1737	0.2025	0.3965	0.5098
	G1	−0.5313	0.2723	0.0586	0.1171
	G2	0.0736	0.1795	0.6840	0.7243
	S	0.1334	0.2181	0.5444	0.6124

B Bisected N-glycans, CF Core fucosylated N-glycans, G0 Agalactosylated N-glycans, G1 Monogalactosylated N-glycans, G2 Digalactosylated N-glycans, S Sialylated N-glycans, ACdd Doxorubicin, cyclophosphamide dose-dense therapy, TCbHP Docetaxel, carboplatin, trastuzumab and pertuzumab therapy, TC Docetaxel + cyclophosphamide

[P < 0.05] – Bold

Fig. 2.

Therapy-specific longitudinal changes in IgG glycosylation traits. Y-axis shows normalized relative abundance values. The bold lines represent the group’s means. B, bisected N-glycans; CF, core fucosylated N-glycans; G0, agalactosylated N-glycans; G1, monogalactosylated N-glycans; G2, digalactosylated N-glycans; S, sialylated N-glycans

Taken together, these results indicate that chemotherapy induces treatment-specific effects on IgG glycosylation, particularly in traits such as core fucosylation and bisecting GlcNAc, which may have functional implications for antibody-dependent cellular cytotoxicity (ADCC) and immune effector mechanisms.

To assess the robustness of these therapy-specific findings, a post hoc power analysis was conducted across the derived IgG traits, given the small and unbalanced group sizes (22/3/9 paired samples). The observed effects between treatment regimens were moderate to large (Cohen’s f = 0.43–0.61). However, due to unequal group sizes, the achieved power was modest (0.30–0.55; median 0.34). Even for the largest effect (G1, f = 0.61), statistical power remained below the conventional 0.80 threshold.

In contrast to these IgG-specific alterations, the derived glycosylation traits of total plasma proteins remained largely stable across timepoints. The only notable exception was an increase in oligomannose glycans following chemotherapy (adjusted P = 0.0253; Supplementary Table 3, Supplementary Fig. 2). Trend toward increased oligomannosylation was noted in the ACdd group (Supplementary Table 4) and partial responders (Supplementary Table 5) but did not reach statistical significance.

To further explore the biological relevance of plasma glycosylation, we examined associations with clinical tumor size. Exploratory analysis revealed a strong association between tumor size and plasma-derived glycan traits. Due to sample size limitations, direct pairwise comparisons were not feasible; instead, tumor size was evaluated as a predictor in a linear mixed-effects model (Fig. 3). Among the 16 plasma-derived glycan traits, 9 were significantly associated with tumor size. The most pronounced changes were observed in digalactosylation (G2_p) and oligomannosylation (Man), particularly among patients with large or advanced tumors (Fig. 3). The largest shift in glycosylation profiles occurred between intermediate-sized and advanced tumors, suggesting that tumor burden may substantially influence systemic glycan composition.

Fig. 3.

Plasma glycosylation profiles stratified by clinical tumor stage. Error bars indicate 95% confidence intervals. Tumor stages: In situ (< 1 mm), small (1–10 mm), intermediate (10–50 mm), and large (> 50 mm or advanced). The right panel shows Benjamini–Hochberg–corrected P values from ANOVA testing the effect of tumor size. LB, low branching; HB, high branching; S0–S4, increasing sialylation; G0_p–G4, increasing galactosylation; Man, oligomannose N-glycans; B_p, bisected N-glycans; CF_p, core fucosylated N-glycans; AF, antennary fucosylated N-glycans

Discussion

Despite major advances in breast cancer diagnostics and therapy, the lack of sensitive, minimally invasive biomarkers for real-time monitoring remains a critical gap in clinical care. Current tools, such as serum markers (e.g., CA 15 − 3, CEA) and imaging, have limitations in terms of sensitivity, specificity, and timing. Our study provides novel evidence that N-glycosylation of immunoglobulin G (IgG) and total plasma proteins reflect therapeutic effects and indicate that further studies may provide a better understanding of its potential as a useful biomarker for breast cancer.

The most consistent finding across all therapy types was a significant decrease in IgG core fucosylation following neoadjuvant chemotherapy. Given the known role of afucosylated IgG in enhancing antibody-dependent cellular cytotoxicity (ADCC) via increased FcγRIIIa binding [33], this observation may reflect an immunological shift that enhances antitumor immunity. This is particularly relevant in regimens that include taxanes and HER2-targeted therapies such as trastuzumab, both of which have been shown to modulate NK cell-mediated cytotoxicity and amplify ADCC responses [34, 35]. The decline in core fucosylation we observed might therefore reflect a shift toward a more ADCC-permissive IgG repertoire, providing indirect insight into immune activation during treatment.

Therapy-specific glycan shifts were also observed. TC therapy was associated only with decreased core fucosylation, whereas ACdd was additionally associated with decreased digalactosylation and increased monogalactosylation. TCbHP treatment uniquely decreased the level of bisecting GlcNAc. These patterns likely reflect both direct effects of chemotherapy and immunomodulatory interactions. These patterns are consistent with prior reports showing that systemic therapies induce distinct glycomic signatures: for example, endocrine therapies such as tamoxifen and anastrozole differentially modulate IgG glycosylation in luminal breast cancer [29]. Collectively, these results support the concept that systemic therapy imprints a distinct glycomic “fingerprint” on the IgG pool.

By contrast, the total plasma N-glycome was largely stable in our study, with one clear exception: an increase in oligomannose glycans. This observation is consistent with earlier N-glycomic studies on total plasma proteins in breast cancer, which reported therapy-associated changes in global glycosylation patterns [20]. Importantly, a recent meta-analysis confirmed that oligomannose glycans are elevated across multiple human cancers, including breast cancer, and linked this to altered α1,2-mannosidase activity [36]. Similarly, de Leoz et al. showed that oligomannose glycans increase during breast cancer progression in a transgenic mouse model and in patient sera [22]. Ščupáková et al. extended this finding to human metastases, where metastatic lesions exhibited enrichment in oligomannose glycans and downregulation of mannose-trimming enzymes [37]. Consistent with these systemic changes, Choi et al. demonstrated that both the serum levels and glycosylation of alpha-1-acid glycoprotein vary with breast cancer severity, underscoring that individual acute-phase proteins contribute to the altered plasma glycome and may carry stage-relevant glycosylation signatures [38]. Together with our observation that oligomannose glycans correlated with tumor size, these studies support the idea that plasma oligomannosylation may serve as a surrogate for tumor burden and altered glycan-processing pathways.

Importantly, the strength of the glycosylation–tumor size correlation, even in a modest sample, suggests the potential for glycan traits as biomarkers of disease progression. Interestingly, patients with large or advanced tumors presented glycan profiles resembling those of patients with in situ disease, suggesting nonlinear or subtype-specific associations that merit further investigation.

Beyond systemic biomarkers, aberrant N-glycosylation also has mechanistic implications in tumor biology. For example, site-specific N-glycosylation of CD24 is required for its stable cell-surface localization in basal breast cancer, influencing adhesion and immune interactions [39]. Aberrant glycosylation of receptors such as EGFR and integrins similarly alters signaling, migration, and therapy response [40]. Thus, plasma glycome may be understood not only as a composite biomarker but also as a systemic mirror of tumor-intrinsic glycoprotein remodeling.

The strengths of this study include its longitudinal design, the use of high-throughput, robust glycomic methods, and therapy-stratified analysis. However, several limitations must be acknowledged. First, the single-center design and small sample size (n = 34) limit generalizability and statistical power, particularly for subgroup comparisons. Second, the absence of a healthy control group restricts our ability to draw diagnostic conclusions. Third, while plasma glycomics provide a global snapshot of systemic glycosylation, it lacks resolution at the protein carrier and site-specific level; thus, disease-relevant changes in individual glycoproteins may be masked within the composite plasma glycome. Future studies should validate these findings in larger, multi-institutional cohorts with long-term follow-up and survival endpoints. Integrating glycan profiling into multiomic liquid biopsy approaches may also increase its utility in clinical workflows.

In summary, our findings demonstrate that IgG N-glycosylation, particularly core fucosylation, is modulated by chemotherapy and varies with treatment regimen. At the same time, plasma oligomannose glycans correlate with tumor size, in line with previous reports across breast and other cancers. These glycomic signatures may offer a minimally invasive, immunologically relevant tool to monitor treatment efficacy and disease progression, with potential applications in precision oncology. However, future studies are warranted to investigate their biomarker potential in greater depth.

Conclusions

This study demonstrates that IgG N-glycosylation is dynamically remodeled by neoadjuvant chemotherapy in a regimen-specific manner, with consistent decreases in core fucosylation and additional shifts in galactosylation and bisecting GlcNAc. These alterations are not only markers of therapy response but may also reflect enhanced antibody effector functions, including increased ADCC. In contrast, the total plasma N-glycome remained largely stable, with the notable exception of elevated oligomannose glycans, which correlated with tumor size and likely mirror altered glycan processing pathways in cancer. Together, these findings indicate that IgG and plasma glycomic traits capture both immune modulation and tumor burden. Glycan profiling, particularly of IgG, holds promise as a minimally invasive biomarker for monitoring treatment response and may support more personalized therapeutic strategies in breast cancer which should be validated in future studies.

Abbreviations

ACdd

Doxorubicin and cyclophosphamide dose-dense therapy

ADCC

Antibody-dependent cellular cytotoxicity

AJCC/UICC

American Joint Committee on Cancer/Union for International Cancer

ATM

Ataxia Telangiectasia Mutated gene

B

Bisected N-glycans

BMI

Body Mass Index

BRCA1

Breast Cancer Gene 1

BRCA2

Breast Cancer Gene 2

CA 15 − 3

Cancer Antigen 15 − 3

CBC

Complete Blood Count

CHEK2

Checkpoint Kinase 2

CF

Core fucosylation

CR

Complete Response

CT

Computed Tomography

DCIS

Ductal Carcinoma In Situ

DMSO

Dimethyl Sulfoxide

EDTA

Ethylenediaminetetraacetic acid

ER

Estrogen Receptor

G

Galactosylation (general)

G0

Agalactosylated N-glycans

G1

Monogalactosylated N-glycans

G2

Digalactosylated N-glycans

GP

Glycan Peak

HER2

Human Epidermal Growth Factor Receptor 2

IgG

Immunoglobulin G

Ki67

Proliferation marker Ki67

MINDACT

Microarray in Node-Negative Disease May Avoid Chemotherapy trial

MRI

Magnetic Resonance Imaging

NK

Natural Killer cells

NST

No Special Type (histological subtype of breast cancer)

PBS

Phosphate Buffered Saline

PET/CT

Positron Emission Tomography/Computed Tomography

PALB2

Partner and Localizer of BRCA2

PR

Partial Response or Progesterone Receptor (depending on context)

PTEN

Phosphatase and Tensin Homolog

S

Sialylation

S1

Monosialylated N-glycans

S2

Disialylated N-glycans

SDS

Sodium Dodecyl Sulfate

SPE

Solid-Phase Extraction

TAILORx

Trial Assigning Individualized Options for Treatment Rx

TC

Docetaxel and Cyclophosphamide therapy

TCbHP

Docetaxel, Carboplatin, Trastuzumab, and Pertuzumab

TP53

Tumor Protein p53

2-AB

2-Aminobenzamide

Authors’ contributions

IG, DP and GL contributed to the conception and design of the study, supervised the clinical protocol, and critically revised the manuscript. MMV, IK, MG and IM collected clinical data, coordinated patient recruitment, and contributed to interpretation of clinical outcomes. BRT and KR performed laboratory experiments and data acquisition related to glycan profiling. DV conducted statistical analysis and contributed to the interpretation of glycomic data. MMV, BRT and DV drafted the manuscript and created figures and tables with input from all authors. All authors reviewed and approved the final version of the manuscript. They agree to be personally accountable for their individual contributions and to ensure the integrity and accuracy of the work is appropriately investigated and resolved.

Funding

This study was supported by the Horizon Europe program (grant No. 101159018, SynHealth).

Data availability

The UHPLC glycomic datasets generated and analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was approved by the Ethics Committee of the University Hospital for Tumors, Sestre milosrdnice Clinical Hospital Center, Zagreb, Croatia. All participants provided written informed consent prior to inclusion in the study. The study was conducted in accordance with the Declaration of Helsinki.

Consent for publication

Not applicable.

Competing interests

GL declares that he is the founder and owner of Genos Glycoscience Research Laboratory, which offers commercial services of glycomic analysis and has several patents in this field. IG and BRT are employees of Genos Glycoscience Research Laboratory. The other authors declare no conflict of interest.