The study was performed with the approval of the local ethics committee (# RC12 0085, University Hospital of Nantes) and the Commission Nationale de l’Informatique et des Libertés (CNIL # 912335). For technical reasons, only mc IgG could be purified, thus only patients presenting with a mc IgG were included in this study. Thus, we examined 148 patients with mc IgG: 68 MGUS, 6 SM, and 74 MM diagnosed at the French University Hospitals (CHU) of Dijon, Nantes, Paris (Saint-Antoine) and Tours. In this study, all MGUS patients had a mc IgG ≥4 g/L. Sera from 46 HVs and 40 patients diagnosed with MPN were also studied as controls. Written informed consents were obtained from patients in the relevant clinical departments, and in the blood bank for HVs enrolled by the Etablissement Français du Sang (EFS, Nantes, France). A convention has been signed between our laboratory (CRTI—INSERM UMR 1064) and the blood bank (EFS Pays de La Loire).

After clotting, blood samples of patients were centrifuged at 2,200 × g for 15 min at 4°C, serum was collected, and aliquots were stored at −80°C or −20°C, depending on the collecting site. Total IgG concentration in serum was measured with an immuno-nephelemetric assay performed on a Beckman Immage Analyzer (Beckman Coulter, Villepinte, France). The concentration of the monoclonal (component) IgG is estimated by integrating the electrophoretic peak according to the orthogonal mode (the so-called “baseline method”). Purification of pc and mc IgGs and verification of their purity were performed as described (1, 3, 11, 39). Briefly, after separation using electric charge on agarose gel electrophoresis (SAS-MX high resolution, Helena Biosciences, Gateshead, UK), bands corresponding to mc IgG or gamma zone corresponding to pc IgGs were carefully cut and proteins were eluted from gels into PBS. Concentration of the purified pc and mc IgGs was determined using the Nanodrop Spectro-photometer ND-1000 with the IgG extinction coefficient (ε = 1.36 for a solution of 1 mg/mL). The recovered IgG amount after purification varied from 40 to 70% for both mc IgG and pc IgG, depending on experiments and the initial IgG concentration in serum. Purity of each IgG fraction was analyzed by isoelectrophoresis and immunoblotting (homemade isoelectrofocusing gel using a range of pH 3–10, blotting onto PVDF membrane and revelation using an HRP anti-human IgG gamma chain). Only highly purified mc IgG were used for sialylation studies (see Figure S1 in Supplementary Material).

An enzyme-linked lectin assay (ELLA) was developed and used for IgG sialylation detection, and an enzyme-linked immunosorbent assay (ELISA) was developed for total IgG detection, as described (40). Ninety-six well plates (NuncMaxiSorp™) were coated overnight at 4°C with 50 µL of affinipure donkey anti-human IgG, Fcγ-specific fragment Ab (Jackson ImmunoResearch, West Grove, PA, USA) diluted at 1/250 (5.2 µg/mL, ELLA) and 1/1,000 (1.3 µg/mL, ELISA) in 25 mM borate buffer pH9. After 3 washes with 200 µL PBS-Tween 0.05% (Sigma, St. Louis, MO, USA), 100 µL periodic acid (5 mM) per well were added for 10 min at room temperature (RT), protected from light. The plates were then saturated with 100 µL of B-grade bovine gelatin (Sigma, St. Louis, MO, USA) 0.25% in PBS-Tween 0.01%, at 37°C, for 2 h. After three washes, samples were diluted in PBS-Tween 0.1% and deposited in triplicates containing 1.25 ng/well for detection of total IgG, or 100 ng/well for sialylation detection. Total IgG quantity was revealed by incubating the plates with 50 µL of peroxidase affinipure donkey anti-human IgG (H + L) diluted 1/1,000 (0.8 µg/mL, Jackson ImmunoResearch, West Grove, PA, USA) for 1 h. Sialic acid was revealed using 50 µL biotinylated Sambucus nigra agglutinin (SNA) diluted 1/750 (2 µg/mL, Glycodiag, Orleans, France) for 90 min and then 50 µL streptavidin HRP diluted 1/1,000 (1 µg/mL, Vector laboratories, Burlingame, CA, USA) for 1 h, at 37°C. Then 50 µL of TMB, the chromogenic substrate for HRP (Sigma-Aldrich, St. Louis, MO, USA) was added and the reaction was stopped by 50 µL sulfuric acid 0.5 M after 5 min for IgG detection, and after 15 min for sialic acid detection. Optical densities (OD) were measured using Spark 10 M multimode microplate reader (Tecan, Männedorf, Switzerland) at 450 nm. The relative sialylation was expressed as the sialic acid/IgG OD ratio. Control samples were used in all experimental settings, to assess reproducibility.

A 1% agarose gel containing a 10% mixture of 3–10 and 8–10.5 ampholytes was prepared and pre-focalized with acetic acid (0.5 N) and sodium hydroxide (1 N) in order to establish a pH gradient. The pre-focalization was run for 30 min at 250 V and 30 mA for a total of 90 vH. Samples were then dropped at the anode side and the focalization was launched during 90 min at 1,200 V and 50 mA for a total of 900 vH. Sera proteins were then passively transferred on a pre-activated PVDF Immobilon-P membrane (Millipore, Billenca, CA, USA) for a few minutes. Finally, after a saturation step and several washes, the membrane was incubated with the rabbit anti-human IgG (H + L)-peroxydase (Dako, Santa Clara, CA, USA). Revelation was made using HRP revelation system (Sigma, St. Louis, MO, USA).

Purified IgGs were digested by trypsin and glycopeptides were isolated from peptides using two methods, i.e., reversed-phase high-performance liquid chromatography and a protocol involving the commercial ProteoExtract^® Glycopeptide Enrichment Kit (EMD-Millipore, Etobicoke, ON, Canada). Fractions were concentrated for analysis by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-ToF-MS). The instrument used was an UltraFleXtreme™ (Bruker, Bremen, Germany) operated in positive ion, reflective mode.

Frozen aliquots of serum were used to quantify 40 cytokines and 2 soluble cytokine receptors linked to inflammation or/and infection using the Luminex technology (Bio-Plex 200) with Bio-Plex Pro Human Cytokine Panel kits (Bio-Rad, Hercules, CA, USA), following the manufacturer’s instructions.

The Multiplexed Infectious Antigen microArray (MIAA) assay has been described previously (11, 39). The assay was developed to determine the infectious specificity of purified IgG using commercially available Ag or/and lysates from EBV, HCV, cytomegalovirus (CMV), Herpes simplex virus 1 (HSV-1), HSV-2, varicella zoster virus (VZV), Helicobacter pylori (H. pylori), Toxoplasma gondii (T. gondii), and Borrelia burgdorferi (B. burgdorferi). Infectious Ag were purchased from Abcam (Cambridge, United Kingdom), Advanced Biotechnologies Inc. (Columbia, MD, USA) and ImmunoDiag (Hämeenlinna, Finland). Lysates were supplied by Advanced Biotechnologies Inc. (Columbia, MD, USA) and EastCoast Bio (North Berwick, ME, USA). The arrays consist of 8 × 8 matrices that included: (i) 13 Ag: 2 for EBV, 3 for HCV, 1 for T. gondii, 1 for H. pylori, 2 for HSV-1, 2 for HSV-2, and 2 for VZV; (ii) 5 lysates: CMV, T. gondii, H. pylori, HSV-1, and HSV-2; (iii) 2 mixes: one of 5 CMV Ags and one of 2 B. burgdorferi Ags; (iv) 2 negative controls: PBS, and PBS with 0.1% bovine serum albumin (BSA). For hybridization, IgG concentrations were adjusted to 400 µg/mL for serum and from 50 to 200 µg/mL for purified mc IgG. 80 µL of samples were incubated for 2 h at RT. After washing, slides were incubated with a labeled secondary Ab (0.2 µg/mL DylightTM 680 Labeled Goat anti-human IgG (H + L), from SeraCare, Milford, MA, USA; Ref. 5230-0342). Fluorescence signal, detected with the Odyssey infrared imaging system scanner at 21 µm resolution (LI-COR Biosciences, NE, USA) was quantified using the GenePix^® Pro 4 Microarray Acquisition & Analysis Software (Molecular Devices, Sunnyvale, CA, USA).

Data analysis was performed by GraphPad Prism 6.01 software. Patient parameters were expressed as medians and ranges, or/and means ± SEM. The chi-2 test was used for categorical variables. For continuous variables (n ≥ 30) the Student t-test or the one-way ANOVA followed by Tukey’s post hoc test were used. For continuous variables (n < 30), a normality test was systematically performed for each group. When parametric conditions were fulfilled, a Student’s t-test or a one-way ANOVA followed by Tukey’s post hoc test was performed. For non-parametric conditions, a Mann–Whitney U test or a Kruskal–Wallis test followed by Dunn’s post hoc test was performed. The tests used are indicated in Figure and Table legends. A P value below 0.05 was considered statistically significant.

According to the patients’ data, in the present cohort the mc IgG represented 40.8–87.7% (median: 70.8%) of total IgGs in MGUS, and 49.7–95.5% (median: 86.3%) in MM. These percentages represent estimations, since mc IgGs and total IgGs are not measured by the same techniques. However, the data indicated that almost all MGUS and MM patients in the cohort produced some level of pc IgGs, estimated to represent ~30 and ~14% (medians) of total IgGs in MGUS and in MM, respectively. After electrophoresis of the 148 sera on agarose gel, mc IgGs and pc IgGs from each patient were separated. The relevant bands were collected, and proteins were eluted from the gel (Figure 1A). The purity of IgGs was confirmed by immunoblotting after isoelectophoresis. Examples of the efficiency of this method are illustrated in Figure 1B and Figure S1 in Supplementary Material. The typical pattern of mc IgGs resolves into a series of sharp lines that are equidistant approximately 0.05 pH units apart. This difference in pH is due to the deamidation of glutaminyl and asparaginyl residues, yielding aspartic and glutamic acid, respectively (41, 42). The purified pc IgGs appear as a smear after isoelectrophoresis and are completely separated from mc IgG (Figure 1B). Only highly purified mc IgGs (n = 148) and pc IgGs (n = 142) were retained for further analysis (for 6 patients, pc IgGs could not be purified). In addition to deamidation, micro-heterogeneity of mc IgGs is due to carbohydrate differences, especially the sialylation level. Hence, immunofixation of the purified mc IgG, using biotinylated SNA, a lectin specific of sialic acid, shows that the lectin recognized some of the mc IgG bands, indicating that they are not fully sialylated (Figure 1C).

In this retrospective study, we analyzed 148 patients presenting with mc IgG and diagnosed with MGUS (n = 68), SM (n = 6) or MM (n = 74). Clinical data were available for 133/148 patients (59 MGUS, 6 SM, 68 MM). The biological and clinical characteristics of the 133 patients are shown in Table 1. The median age for MGUS, SM, and MM patients at the time of diagnosis was similar, 67.0, 67.1, and 63.6, respectively. The male/female ratio was 55.9% for MGUS, and higher for SM and MM patients: 83.3 and 60.3%, respectively (differences not significant, Chi-2 test). The International Staging System (ISS) and Durie–Salmon Staging (DSS) scores indicated that 32.7% of MM patients presented with ISS stage III at the time of diagnosis (DSS stage III: 44.6%). In MM, mc IgG were studied either at the time of diagnosis (17 patients; 23%), or during or after treatment (40 patients; 54%); for 17 patients (23%), information was not available. In addition, two categories of control sera without mc IgG were used in this study: first, a cohort of 46 HVs (with no chronic inflammation, and no hematological disease); second, a cohort of 40 patients at the time of diagnosis of MPN, a group of chronic hematological malignancies with strong associated inflammation. As expected, MPN patients showed higher leukocyte counts, a higher hemoglobin level, and higher platelet counts compared to MGUS and MM patients.

For MGUS, SM, and MM patients, the degree of sialylation of purified pc and mc IgGs was analyzed separately. For a significant number of patients in this cohort (35/128, or 25.7%: MGUS, n = 19; SM/MM, n = 16), our previous studies had revealed that the purified mc IgG specifically targeted an infectious pathogen (11). Table 2 shows the serological status of patients, as determined by the MIAA assay; the results of these studies reflect the Ab function of pc IgG and mc IgG analyzed together. The MIAA used here allowed testing for panels of commercially available Ag and lysates from nine infectious pathogens: EBV (HHV-4), HCV, CMV (HHV-5), HSV-1 (HHV-1), HSV-2 (HHV-2), VZV, (HHV-3), H. pylori, T. gondii, and B. burgdorferi (39). Overall, the serological status of MGUS patients was similar to that of the general population, which is consistent with the persistence of pc IgGs for most patients. As previously reported, there were lower rates of positive serology for SM/MM patients compared to MGUS patients, likely explained by lower production of pc IgGs in advanced MM disease (11). The characteristics of the 35 patients (19 MGUS and 16 SM/MM) with infectious pathogen-specific mc IgG (“MIAA+ patients”) as determined by the MIAA assay performed with purified mc IgG, are shown in Table 3. The purified mc IgG of 22 patients (12 MGUS and 10 SM/MM) specifically targeted EBV, and the EBNA-1 protein was the main target (20/22 patients) (11). The purified mc IgG from the 13 other patients specifically targeted HSV-1 (n = 6), CMV (n = 2), VZV (n = 2), and H. pylori (n = 3) (11). The sialylation results of mc and pc IgGs from the 35 MIAA+ patients were analyzed separately.

The sialylation state of pc and mc IgGs is shown in Figure 2A. Analysis of the cohort of healthy controls showed that 95.6% of pc IgGs have a relative sialylation level between 0.5 and 1.5. The mean relative sialylation level of IgGs from healthy individuals was 1.023. Of note, 100% of pc IgGs samples from MPN patients had a similar relative sialylation level, between 0.5 and 1.5 (mean: 1.096). By contrast, pc IgGs from 142 MGUS, SM, and MM patients showed a large heterogeneity in relative sialylation level. For 52.1% of MGUS, SM, and MM patients, pc IgGs presented the same sialylation level as pc IgGs from healthy controls and MPN patients, but for 36.6% of MGUS, SM, and MM patients, pc IgGs were more sialylated, and for 11.3% of these patients, pc IgGs were less sialylated. Overall the degree of sialylation of pc IgGs from the MGUS/SM/MM group was significantly higher than for the group of healthy controls (mean = 1.703, p = 0.0007) or the group of MPN patients (p = 0.0062). Regarding mc IgG, for 75% of MGUS and SM/MM patients, the purified mc IgG presented with strongly reduced sialylation; only 4.1% of patients showed a high sialylation of purified mc IgG (Figure 2A). The sialylation level of purified mc IgGs from the MGUS/SM/MM group (mean = 0.443) was significantly lower than pc IgGs from the healthy controls (p = 0.0054), MPN patients (p = 0.0025), and MGUS/SM/MM patients (p < 0.0001) (Figure 2A). When MGUS and SM/MM were compared, the sialylation level of pc IgGs was not different, whereas that of mc IgGs was lower in SM/MM (mean = 0.271) than in MGUS (mean = 0.645, p = 0.0048) (Figure 2B). Regarding SM/MM patients with mc IgGs specific for a pathogen (MIAA+ patients), the sialylation level of pc IgGs (MGUS or SM/MM) was not different from other pc IgGs (Figure 2C) but pathogen-specific mc IgG were significantly less sialylated (mean = 0.117) than all other mc IgGs (mean = 0.486, p = 0.048) (Figure 2D).

In a second series of experiments, we used HPLC-mass spectrometry to confirm the differences in IgG sialylation level observed between purified mc and pc IgGs. Mass spectrometry analysis was performed on 24 samples. Figure S2 in Supplementary Material shows the results obtained for two representative purified mc IgGs that differed in their sialylation level as assessed with ELLA and ELISA. Figure S2A in Supplementary Material shows a purified mc IgG1 that was found to be highly sialylated using the ELLA and ELISA techniques. Conversely, Figure S2B in Supplementary Material shows a purified mc IgG1 found to be poorly sialylated using the ELLA and ELISA techniques. In both cases, the sialylated forms corresponding to G1FS (1 galactose, 1 fucose, and 1 sialic acid) and G2FS (two galactoses, one fucose, and one sialic acid) with a m/z at 3,087 and 3,249, respectively, are indicated with red arrows. The respective percentages of IgG1 in G1FS and G2FS forms were 26.4% (Figure S2A in Supplementary Material) and 5.2% (Figure S2B in Supplementary Material), thus confirming the results obtained by ELLA and ELISA (Figure 2). We then compared the results obtained for six healthy individuals and eight patients (four MGUS and four MM) (Figure 3). The sialylation level of pc IgGs from healthy controls was not different from those of pc IgGs from MGUS/MM patients. Purified mc IgGs from MGUS/MM patients were less sialylated (mean = 7.35%) than pc IgGs from the same patients (mean = 10.21%; paired t-test: p = 0.03) or from healthy controls (mean = 14.05%; p = 0.02), thus confirming the results obtained by ELISA (Figure 2).

For 64 patients (34 MGUS and 30 MM), we quantified in blood serum the level of 40 cytokines and 2 soluble cytokine receptors [IL-1 receptor α (IL-1Rα), IL-2Rα] linked to inflammation and, for certain molecules [interferon (IFN) α₂, IFN-γ, eotaxin, IL-17, IL-22, IL-26, and IL-33] to anti-viral or anti-microbial immune responses (Table 4). Compared to the group of healthy controls, MGUS/SM/MM patients had significant increases in serum levels of 28 molecules: IL-1Rα, IL-2Rα, and 26 cytokines. Among those, 17 cytokines were elevated in the patients group with p < 0.001, compared to healthy controls: IL-6, IL-7, IL-10, IL-13, IL-15, IL-17, IL-33, IFN-α₂, tumor necrosis factor (TNF)-α, granulocyte colony stimulating factor (G-CSF), granulocyte-macrophage CSF (GM-CSF), basic fibroblast growth factor (FGF), vascular endothelial growth factor (VEGF), monokine induced by IFN-γ (MIG, or CXCL9), monocyte chemotactic protein 1 (MCP-1 or CCL2), HGF, and leukemia inhibitory factor (LIF). Most of these cytokines are pro-inflammatory, which indicates a shift toward a pro-inflammatory environment in MGUS and SM/MM patients.

The comparison between the MGUS and SM/MM groups (Figure 4; Table S1 in Supplementary Material) revealed that only four molecules were slightly but significantly higher in SM/MM patients: HGF, IL-11, SDF-1α, and RANTES. Interestingly, HGF is a survival factor which exerts a pro-tumoral and both pro- and anti-inflammatory action. TGF-β₁, TGF-β₂, and TGF-β₃, known for their anti-proliferative and pro-differentiation effect on hematopoietic cells, were more expressed in MGUS than in SM/MM. Since high levels of HGF are considered of poor prognosis in MM, we explored eventual correlations between cytokine levels and β₂-microglobulin, an important biomarker in the prognosis of MM. Figure 5 shows the positive correlations found between β₂-microglobulin concentration and serum levels of IL-9, IL-26, MIP-1β (pro-inflammatory molecules), and PDGF-BB. Of note, IL-26 is involved in anti-microbial immunity.

Hyposialylation of IgGs has been described as a hallmark of pro-inflammatory state in pathologic contexts other than MM. We analyzed cytokine levels in the serum of MGUS, SM, and MM patients according to the sialylation level of their pc IgGs. Figure 6 shows that patients presenting with hyposialylated pc IgGs have significantly higher levels of major pro-inflammatory cytokines (IL-6, TNF-α, TGF-β₁), and also have higher levels of HGF, IL-17, and IL-33, compared to patients with hyper-sialylated pc IgGs (p < 0.05). These results were confirmed when data were analyzed for potential correlations between cytokine levels and the degree of sialylation of pc IgGs or purified mc IgGs. In SM/MM group, the degree of sialylation of pc IgGs was negatively correlated with levels of IL-17 and IL-33, and also with the concentration of leptin (men only) (Figure 7A). The degree of sialylation of purified mc IgGs was also negatively correlated with leptin (men only) (Figure 7B), and positively correlated with the levels of IFN-α₂ and IL-13. Thus, in MM, sialylation of pc IgGs is inversely correlated with the levels of IL-17 and IL-33, two cytokines important for anti-microbial response. The sialylation level of purified mc IgGs increased with the level of 2 anti-inflammatory cytokines: IFN-α₂ and IL-13.

In addition, we found that sialylation of purified mc IgGs was significantly lower for MM patients with advanced disease—Durie–Salmon staging (DSS) III—than for those with DSS I (Figure 8). Similarly, C-reactive protein (CRP) concentration, a common inflammation marker, was inversely correlated with the sialylation level of pc IgGs (MGUS, SM, and MM patients, Figure S3 in Supplementary Material), and the β₂-microglobulin concentration was inversely correlated with the sialylation level of purified mc IgGs (MGUS, SM, and MM patients, Figure S3 in Supplementary Material). Altogether, these findings suggest that mc IgG hyposialylation could be a new marker of disease severity in MM.