IgG Gal-ratio, a promising serum biomarker for detecting ovarian cancer in as early as stage I disease

Ethical approval was obtained from the Institutional Review Board of Obstetrics & Gynecology Hospital of Fudan University, China (2017-52). This study was conducted in accordance with the Declaration of Helsinki and registered in the Chinese Clinical Trial Registry (Clinical trial number: ChiCTR-DDD-17013356, 13th, Nov, 2017).Patients with pelvic mass or elevated serum CA125 (Roche Diagnostics, Switzerland) levels who later received surgical treatment and were pathologically diagnosed were enrolled into this study from June 2017 to December 2019 at Obstetrics & Gynecology Hospital of Fudan University. Clinical and histopathological characteristics were recorded. Pre-treatment leftover serum samples were collected from the routine clinical laboratory after clinical testing. Since no additional serum sample was collected beyond the standard clinical requirements, written informed consent from participants was not required and was waived by the Institutional Review Board. Serum samples were aliquoted and stored at -80℃ until further analysis. As described in our previous studies, IgG was purified from human serum samples using the IgG Purification Kit Protein A Spin Plate (Thermo Fisher Scientific, Rockford, IL) [ 7 ]. One hundred microliter of IgG-containing elution was then denatured and incubated with 1 µL PNGase F (New England Biolabs, Inc.) for 12 h at 37 °C to release IgG N-glycans. The released oligosaccharides were subsequently purified using porous graphitic carbon (PGC). Oligosaccharides from each sample was spotted in triplicate and analyzed by AXIMA Resonance MALDI-QIT-TOF MS (Shimadzu Corp., JP) equipped with a 337 nm nitrogen laser in reflector positive ionization mode. Tandem mass spectrometry (MS/MS) was utilized to validate the component of the detected glycans. MS spectra was annotated using The GlycoWorkbench software and Progenesis MALDI was used for MS data process. The biantennary, core-fucosylated structures carrying two, one, and no galactose residues are the three most significant glycans identified from IgG, named G2, G1, and G0, respectively. In this study, each sample was spotted in triplicate on the MALDI plate. Thus, a total of three spectra for each sample were averaged, and then subjected to calculation of the degree of galactosylation (IgG Gal-ratio) according to the formula IgG Gal-ratio=G0/(G1 + G2 × 2). Correlation between age, menopausal status and IgG Gal-ratio levels was analyzed using Pearson’s correlation analysis and Spearman’s correlation analysis, respectively. Multiple linear regression model was applied to estimate the impact of age, menopausal status and disease on IgG Gal-ratio levels. IgG Gal-ratio in EOC, BOT, and benign conditions was compared using Student’s t-test. Comparisons of IgG Gal-ratio between different histological types, stages, and grades were performed using one-way analysis of variance (ANOVA). Receiver operating characteristic (ROC) curves were plotted. Area under the curve (AUC) of ROC curves were compared using MedCalc (version 23.3.5). Sensitivity and specificity were calculated to test the diagnostic power of IgG Gal-ratio for EOC. Statistical analyses were performed using PASW Statistics software, version 17.0.

From June 2017 to December 2019, 113 primary EOC patients, 13 BOT patients, and 728 patients with benign gynecological conditions (including ovarian cysts, endometrioma, teratoma, pelvic inflammatory diseases, adenomyoma, uterine myoma, and pelvic tuberculosis et, al.) were enrolled in this study. Median age was 51 (45, 61) years for EOC patients, 39 (26, 44) years for BOT patients, and 41 (34, 46) years for those with benign gynecological conditions. In EOC group, 52% of patients were post-menopausal, while only 9.1% and 8.6% of patients in BOT and benign condition group were post-menopausal, respectively. IgG Gal-ratio in EOC patients was significantly elevated compared to BOT or benign conditions (0.69 ± 0.41 vs. 0.38 ± 0.17 vs. 0.30 ± 0.13, p  < 0.001, Table  1 ; Fig.  1 ). Among the four most common histotypes, IgG Gal-ratio was uniformly elevated, except for endometrioid EOC, which showed only a weak elevation (0.45 ± 0.27). In serous EOC, no difference in IgG Gal-ratio levels was observed between patients with high-grade or low-grade disease (0.72 ± 0.45 vs. 0.65 ± 0.35, p  = 0.605). Table 1 Patient characteristics N Age (y) Mean ± SD Post-menopause N (%) CA125 (U/ml) Mean ± SD Gal-ratio p Ovarian cancer 113 52 ± 11 52 (52.0%) 1037.97 ± 1472.84 0.69 ± 0.41 < 0.001 Serous 83 53 ± 9 43 (55.8%) 1264.19 ± 1647.82 0.71 ± 0.44 Mucinous 6 45 ± 18 2 (33.3%) 245.25 ± 218.29 0.67 ± 0.20 Endometrioid 9 44 ± 15 2 (22.2%) 548.69 ± 534.13 0.45 ± 0.27 Clear Cell 7 52 ± 9 4 (57.1%) 338.15 ± 376.04 0.63 ± 0.24 Unknown 8 NA NA NA NA Ovarian Borderline Tumor 13 37 ± 10 1 (9.1%) 257.42 ± 335.34 0.38 ± 0.17 0.031 Serous 9 37 ± 10 1 (12.5%) 339.58 ± 377.70 0.44 ± 0.17 Mucinous 4 37 ± 13 0 (0%) 72.58 ± 59.98 0.26 ± 0.07 Benign Conditions 728 40 ± 10 62 (8.6%) 95.42 ± 160.17 0.30 ± 0.13 reference Endometriosis 291 37 ± 7 0(0.0%) 128.25 ± 212.76 0.28 ± 0.10 Adenomyoma 129 46 ± 5 10(7.8%) 118.99 ± 104.72 0.28 ± 0.09 PID  † 44 39 ± 8 2(4.5%) 151.71 ± 196.88 0.40 ± 0.23 Teratoma 62 33 ± 9 1(1.6%) 29.79 ± 33.70 0.26 ± 0.08 Pelvic Tuberculosis 3 22 ± 1 0(0.0%) 316.87 ± 133.05 1.11 ± 0.25 Ovarian adenoma 44 44 ± 16 14(32.6%) 35.73 ± 48.86 0.34 ± 0.14 Uterine Myoma 109 44 ± 9 15(14.3%) 33.00 ± 29.91 0.30 ± 0.11 Others 46 47 ± 16 16(39.0%) 46.87 ± 107.46 0.33 ± 0.15 † PID Pelvic inflammatory disease Patient characteristics † PID Pelvic inflammatory disease Fig. 1 IgG Gal-ratio in EOC patients was significantly elevated compared to BOT or benign conditions. EOC-S1, stage I EOC, EOC-S2, stage II EOC, EOC-S3, stage III EOC; EOC-S4, stage IV EOC IgG Gal-ratio in EOC patients was significantly elevated compared to BOT or benign conditions. EOC-S1, stage I EOC, EOC-S2, stage II EOC, EOC-S3, stage III EOC; EOC-S4, stage IV EOC For BOT, which often affects young women and may also wide spread within the abdominal cavity, IgG Gal-ratio was only marginally elevated compared to benign conditions (0.38 ± 0.17 vs. 0.30 ± 0.13, p  = 0.031). Notably, IgG Gal-ratio in serous BOT ( n  = 9) was significantly higher than that in mucinous BOT ( n  = 4) (0.44 ± 0.17 vs. 0.26 ± 0.07, p  = 0.024) or in benign conditions (0.44 ± 0.17 vs. 0.30 ± 0.13, p  = 0.003). In patients with benign gynecological conditions, such as ovarian cysts, endometrioma, adenomyoma, teratoma, pelvic inflammatory diseases, and uterine myoma, IgG Gal-ratio remained low. Surprisingly, patients with pelvic tuberculosis showed a much higher IgG Gal-ratio (1.11 ± 0.25, n  = 3), even higher than that in EOC patients. Since IgG glycosylation status has been reported to be age-related [ 17 ], [ 18 ], we analyzed the correlation of age and menopausal status with IgG Gal-ratio levels. Older age showed weak correlation with increased IgG Gal-ratio levels. Pearson’s correlation coefficients were 0.358 and 0.140 in EOC patients and those with benign conditions, respectively. Additionally, post-menopausal patients tended to have higher IgG Gal-ratio levels. Spearman’s correlation coefficients were 0.375 and 0.316 in EOC patients and those with benign conditions, respectively. Patients in EOC group tend to be older and more likely to be post-menopausal, which correlate with slightly higher IgG Gal-ratio levels. To eliminate these biases, we conducted multiple linear regression analysis, which showed that EOC group was an independent factor associated with higher IgG Gal-ratio levels ( p  < 0.001). To test the diagnostic performance of IgG Gal-ratio for ovarian cancer, an ROC curve was plotted. IgG Gal-ratio showed comparable results to CA125 in differentiating ovarian cancer from non-malignant gynecological conditions (AUC = 0.871 vs. 0.872, Fig.  2 A). When these two markers were combined (Gal-CA125 = IgG Gal-ratio × CA125), the diagnostic performance can be further improved (AUC = 0.913 vs. 0.872, p  < 0.0001, Fig.  2 A). Specificity for detecting ovarian cancer was elevated from 55.9% for CA125 alone to 69.1% for Gal-CA125 at a fixed sensitivity of 90.0% (Table  2 ). Fig. 2 ROC curve of IgG Gal-ratio, CA125, and Gal-CA125 to detect ovarian cancer. A all patients, B detecting stage I ovarian cancer, C patients younger than 45 years old, D patients aged 45 years or older, E pre-menopausal patients, F post-menopausal patients ROC curve of IgG Gal-ratio, CA125, and Gal-CA125 to detect ovarian cancer. A all patients, B detecting stage I ovarian cancer, C patients younger than 45 years old, D patients aged 45 years or older, E pre-menopausal patients, F post-menopausal patients Table 2 Diagnostic performance of Gal-ratio, CA125, and Gal-CA125 for stage I and all stage EOC All Stage I Sensitivity Specificity AUC Sensitivity Specificity AUC Gal-ratio 90.0% 59.5% 0.871 90.0% 46.8% 0.835 CA125 90.0% 55.9% 0.872 90.0% 32.9% 0.761 Gal-CA125 90.0% 69.1% 0.913 90.0% 59.9% 0.851 Diagnostic performance of Gal-ratio, CA125, and Gal-CA125 for stage I and all stage EOC To avoid the influence of age and menopausal status on IgG Gal-ratio levels, we conducted stratified ROC analysis. The AUC of IgG Gal-ratio stayed over 0.80 in patient younger or older than 45 years old, pre- or post-menopause (Fig.  2 C and F). Early diagnosis is critical for improving the outcomes of ovarian cancer patients. In this study, 19 stage I EOC were enrolled, including 8 high grade serous, 3 mucinous, 3 clear cell, and 5 endometrioid. IgG Gal-ratio levels were significantly elevated in these early stage ovarian cancer compared to benign conditions (0.50 ± 0.17 vs. 0.30 ± 0.13, p  < 0.001, Fig.  1 ). AUC for IgG Gal-ratio to detect stage I EOC patients was 0.835 (95% CI: 0.740, 0.929), which is much higher than that for CA125 alone (0.761, 95% CI: 0.647, 0.876, Fig.  2 B). When combined, specificity for Gal-CA125 to detect stage I ovarian cancer was improved from 32.9% for CA125 alone to 59.9% for Gal-CA125 at a fixed sensitivity of 90.0%. (Table  2 ).

In this study, we demonstrated IgG Gal-ratio be a robust serum biomarker for EOC. IgG Gal-ratio was as powerful as CA125 in discriminating EOC patients from patients with BOT or benign gynecologic conditions. For early diagnosis, IgG Gal-ratio was even better than CA125 in detecting stage I EOC patients. Moreover, diagnostic performance can be further improved when these two markers are combined. In benign group, diseases that frequently cause CA125 elevation and interfere with EOC diagnosis were enrolled, including endometrioma, adenomyoma, and pelvic inflammatory diseases. IgG Gal-ratio in these diseases was significantly lower than that in EOC, which proved high specificity of IgG Gal-ratio. However, we found extremely high level of IgG Gal-ratio in patients suffering from pelvic tuberculosis, indicating increased agalactosyl IgG and decreased mono- and/or digalactosyl IgG in these patients. This phenomenon is consistent with a previous report showing elevated agalactosyl IgG in tuberculosis patients and recovery after chemo- and immunotherapy [ 19 ]. Glycosylation is the key post-translational mechanism that regulates function of immunoglobulins, with multiple systemic repercussions for the immune system. Reduction in galactosylated IgG glycan was first found in patients with rheumatoid arthritis [ 20 ]. Later, a reversal of this trend was reported during pregnancy, a relatively immunosuppressive status, where the IgG contained more galactose [ 21 ]. These data form the basis for the widely accepted and repeated notion that decreased IgG galactosylation is a sign of inflammation. Since then, decreased IgG galactosylation has been correlated with a long list of circumstances, including inflammation, infection, cancer or aging [ 22 ], [ 23 ]. Cancers were considered as the product of unresolved inflammation [ 24 ] and compromised immune surveillance [ 25 ]. Thus, changes in IgG galactosylation profile are early events in cancer initiation. Although molecular regulations of the above-described changes are not yet fully understood, in rheumatoid arthritis, a similar trend of reduced IgG galactosylation was shown to result from decreased activity of galactosyltransferase in IgG-producing plasma B cells [ 26 ]. Galactosylation also seems to be, at least in part, driven by the concentration of estrogen. This is evidenced by its raised level during pregnancy [ 27 ] and decreased level after menopause. Additionally, an endocrine manipulation study confirmed estrogens as an in vivo modulator of IgG galactosylation in both women and men, thus suggesting a mechanism for gender-dependent modulation of immune response . Impact of galactosylation on IgG function is still unclear. IgG lacking either sialic acids or galactoses was reported to have weaker binding to FcγR [ 28 ]. However, others reported that removal of galactose residues on the Fc glycan did not alter binding to FcγRIIB [ 29 ]. These data suggest that IgG Fc glycans can differentially influence FcγR binding, and that the outcome may depend on which receptor is present. In terms of clinical practice, the most formidable challenge to date that impacts all glycomics marker application is whether high-throughput and high reproducibility can be achieved [ 30 ]. Fortunately, the rapid and high-throughput human serum N-glycan preparation technology using 96-well plated-based procedures has been developed in recent years [ 31 ], which includes IgG purification, glycan release [ 32 ], and has shed light on the large-scale trials of our IgG agalactosylation marker for practical use. In addition, we independently developed a normalization strategy by calculating the galactosylation ratio (Gal-ratio) using the formula G0 / (G1 + 2 × G2), which allows omission of internal or external standards and has been demonstrated to be highly reproducible and robust for high-throughput analysis in our previous studies [ 16 , 33 – 35 ] .

Despite decades of research and therapeutic advancements, epithelial ovarian cancer (EOC), including fallopian tube cancer and primary peritoneal cancer, remains the most lethal gynecological malignancy in women [ 1 ]. The high mortality rate is largely attributed to the lack of specific symptoms and early diagnostic biomarkers, resulting in late diagnosis and a less than 30% 5-year survival rate among patients in the late stages [ 2 ]. This underscores the urgent need to identify new biomarkers that can facilitate early diagnostic strategies and enable timely and effective interventions. Immunoglobulin G (IgG) is the most abundant glycoprotein in human serum. Each IgG molecule can be functionally divided into two distinctive parts: fragment antigen binding (Fab), responsible for recognizing and binding to specific antigens, and fragment crystallizable (Fc), which exerts effector functions by interacting with complements, Fc-gamma receptors (FcgRs) and neonatal Fc receptors (FcRn) [ 3 ]. Within the Fc region, there is a highly conserved N-glycosylation site at asparagine 297 that is typically occupied by biantennary N-glycans of complex type [ 3 ]. A vast majority of the Fc N-glycans bear 0–2 galactose residues, forming three IgG-Fc glycoforms: G0 (no galactose), G1 (one galactose), and G2 (two galactose) [ 4 ]. IgG glycosylation appears to be highly regulated in both physiological and pathological processes. Changes in IgG glycosylation patterns have been observed in various malignancies, such as gastric cancer [ 5 ], pancreatic cancer [ 6 ], neuroblastoma [ 7 ], and non-malignant diseases[ 8 - 11 ]. In some cases, decreased galactosylation of IgG is associated with disease progression and survival [ 12 ]. In the circulation of EOC patients, higher fraction of IgG carrying agalactosylated N-glycans and, conversely, lower proportions of those carrying monogalactosylated or digalactosylated glycan has been reported by Saldova et al. [ 13 ], Alley et al. [ 14 ], and our group [ 15 , 16 ]. However, these results were based on relatively small cohort studies, leaving many questions unanswered. Is IgG galactosylation pattern a robust EOC marker in real clinical settings? Is the IgG galactosylation level uniformly decreased among EOC of various histotypes and stages? Is the degalactosylation of IgG a specific marker for EOC? What are its levels in ovarian borderline tumors (BOTs) and benign pelvic masses? The present study aimed to address these questions by examining the IgG galactosylation level in a large cohort.