Selectins are calcium-dependent (C-type) lectins that recognize specific glycan residues as their ligands and mediate the adhesion of immune cells to the endothelium. To date, three members of the selectin family have been identified, P-selectin, E-selectin and L-selectin (18). L-selectin is highly expressed on hematopoietic stem cells and mature leukocytes, and is rapidly shed by proteolytic cleavage upon leukocyte activation (18–20). P- and E-selectin are known as “vascular selectins” because they can be expressed on endothelial cells. P-selectin is constitutively expressed by endothelial cells and platelets where they are stored in Weibel-Palade bodies and α-granules, respectively. Therefore, they can be translocated to the cell surface within minutes after a proinflammatory stimuli such as thrombin and histamine, making them the most important adhesive molecules in acute injury. On the other hand, E-selectin is not constitutively expressed by endothelial cells, but their expression is strongly upregulated by inflammatory cytokines such as interleukin 1β (IL-1β) and tumor necrosis factor α (TNFα) through binding of NF-κB to regulatory domains in the E-selectin promoter. The latter is not possible in the case of P-selectin, as the P-selectin promoter in humans lacks binding sites for NF-κB (21, 22). Because of this property, E-selectin is considered the most important adhesive molecule involved in leukocyte capture in chronic inflammation. The involvement of selectins in the development of many acute and chronic inflammatory conditions is dependent on the selectin-ligand axis, with N-glycosylation playing an important role. The interaction required for leukocyte capture on the endothelial surface is dependent on the affinity of selectins for sialofucosylated glycan epitopes expressed on both endothelial and immune cells. These include sialyl Lewis x (sLex), sialyl Lewis a (sLea), and 6-sulfo sialyl Lewis x (6-sulfo sLex) epitopes, which are responsible for mediating leukocyte capture/rolling during inflammation and are relevant to the successful homing of lymphocytes to lymph nodes (15, 23, 24). E-selectin binds to sialofucosylated N-glycans on E-selectin ligand-1 (ESL-1) and CD44 glycovariant, hematopoietic cell E-/L-selectin ligand (HCELL), to support leukocyte extravasation to the site of inflammation (25–27). Interestingly, Pachón-Peña et al. demonstrated the potential use of glycoengineered HCELL on human adipose-derived mesenchymal stem cells (A-hMSCs) to direct their migration to sites of tissue injury/inflammation, thus enabling relevant immunomodulatory and regenerative functions (28). In addition, the sLex epitope found on APPs may modulate the binding of leukocytes to E-selectin (29, 30). In the case of L-selectin, Mitoma et al. demonstrated a critical role for 6-sulfo sLex-decorated N-glycans found on CD34, a major component of the L-selectin ligand, in the leukocyte trafficking in the high-endothelial venules (HEV) of the peripheral lymph node (17). Interestingly, Huopaniemi et al. showed that co-regulated expression of CMP-sialic acid and GDP-fucose transporters, essential for the synthesis of 6-sulfo sLex, occurs in inflammation, which is not common in physiological conditions. Therefore, it has been suggested that there must be an inflammation-induced transcriptional regulation for Golgi membrane transporters that support trafficking of substrates necessary for synthesis of 6-sulfo sLex N-glycans (31). Furthermore, sulfonation of sLex epitopes on N-glycans is thought to be the result of N-acetylglucosamine-6-O-Sulfotransferase-1 (GlcNAc6ST-1) activity, but further studies are needed to uncover how this synthesis is regulated under physiological and pathological conditions (32). Moreover, Beta-Galactoside-Alpha-2,3-Sialyltransferase 4 (ST3GAL4) is the primary sialyltransferase regulating the synthesis of sLex epitopes in human myeloid leukocytes (33), the expression of which, together with the expression of 6-sulfo sLex, has been shown to be increased by TNFα in chronic lung disease. Thus, disruption of ST3GAL4 function in human myeloid cells may represent a potential target for anti-cell adhesion and anti-inflammatory therapy (16). In addition, fucosyltransferases such as FUT7 and FUT9 are involved in the synthesis of the Lex epitope, and FUT7^-9^- dual knockdown has been demonstrated to significantly decrease the selectin-dependent interaction between leukocytes and endothelial cells (34). Interestingly, an in vivo study has shown that the cytokines IL-6 and/or IL-8 can induce a significant increase in the expression of alpha-1,3/4-fucosyltransferases in mucosal cells, contributing to an increased level of sLex epitopes and thus to dysregulated transendothelial migration. The latter might potentiate the persistent lung inflammation and tissue damage in cystic fibrosis (CF) (35).

After leukocyte rolling and capture by selectins, firm endothelial adhesion of leukocytes is mediated by ICAM-1 and VCAM-1. They are members of the immunoglobulin supergene family that are expressed during chronic inflammation on vascular endothelium, lymphocytes, and macrophages (36). The expression of ICAM-1 and VCAM-1 is under the direct influence of proinflammatory cytokines, such as TNFα, which increase their levels on the endothelial surface (37). In response to inflammation, increased ICAM-1 and VCAM-1 cell surface levels enhance adhesive interactions with their ligands on leukocytes, Lymphocyte function-associated antigen 1/Macrophage-1 antigen (LFA-1/MAC-1) and Very late antigen-1 (VLA-1), respectively (38, 39). N-glycosylation is a crucial factor that can significantly affect the ligand binding of ICAM-1 and VCAM-1 (38, 40–43). In general, the transition from the homeostatic to the inflammatory state of the endothelium is characterized by a decrease in N-glycan complexity and increased expression of high mannose and hybrid structures (44–47). This has been demonstrated to be a consequence of proinflammatory stimulation, possibly by inhibition of early mannose-trimming enzymes (α-mannosidase) (42). Not surprisingly, increased presence of high-mannose ICAM-1 (HM-ICAM-1) results in high-affinity leukocyte binding (38). In particular, this phenomenon is seen in CD16⁺ proinflammatory monocytes, which have a higher affinity for HM-ICAM-1 molecules in atherosclerotic lesions compared with complex α-2,6-sialylated ICAM-1 (48, 49). However, this is not the only mechanism by which the pathological state is maintained in chronic inflammation, as far as N-glycosylation is concerned. In systemic lupus erythematosus (SLE), diabetes, and rheumatoid arthritis (RA), this is regulated by inhibition of galectin-mediated immunosuppressive prevention of ICAM-1/LFA-1 interaction (50), aberrant expression of ICAM-1 N-glycans due to high glucose (51), and activity of glycosyltransferases (52). Another possible candidate for inflammatory modulation is VCAM-1, where removal of α2,6-sialic acid increases leukocyte trafficking (41). The mechanism of action does not involve stronger binding with VLA-1 on leukocytes but with galectin-3 (Gal-3), which supports leukocyte adhesion (53).

The final step in the process of leukocyte extravasation is transendothelial migration (TEM). Adhesion molecules mainly associated with this phenomenon are PECAMs (54). PECAM-1 is a member of the immunoglobulin (Ig) superfamily selectively expressed on the surface of immune cells and is highly enriched at the intercellular interface of endothelial cells (54). The main mechanism responsible for interaction of PECAMs with leukocytes involves homophilic binding (55). As PECAM-1 is highly N-glycosylated, Lertkiatmongkol and her group showed that homophilic binding interactions of human PECAM-1 are supported by α2,3-linked, but inhibited by α2,6-linked sialic acid residues (56). This was previously demonstrated by Doring et al. who presented solid evidence for the importance of α2,3-linked sialic acid in leukocyte activation, adhesion, and recruitment to inflamed vessels (57). In agreement with this, it was later shown that a variety of proinflammatory cytokines secreted in chronic inflammation can downregulate the levels of the (extracellular) enzyme Beta-Galactoside Alpha-2,6-Sialyltransferase 1 (ST6GAL1), responsible for the synthesis of α2,6-linked sialic acid, and by that way maintain the inflammatory state (58, 59).

Neutrophils are polymorphonuclear leukocytes that have long been known to be key players in pathogen recognition and elimination in acute inflammation, but their role in chronic inflammatory and autoimmune diseases, such as psoriasis, RA and SLE, has also been described (62, 63). Understanding the underlying mechanisms of neutrophil activation, migration, survival, and executive function may open new avenues for the treatment of chronic inflammation. N-glycosylation has been shown to contribute to important effector functions of neutrophils, such as extravasation, phagocytosis, degranulation, and formation of neutrophil extracellular traps (NETs) (64) ( Figure 3 ). On their surface, neutrophils express the N-glycosylated MAC-1 integrin, which consists of two subunits - CD11b and CD18. This complex is involved in the regulation of neutrophil trafficking and interaction with other immune cells (65). Kelm et al. performed an analysis of glycan epitopes expressed on the neutrophil MAC-1 surface and observed decreased sialylation together with an increase in the Lex motif and high mannose content in chronic inflammation. These changes were significant in inflammatory bowel disease (IBD) because blocking the terminal Lex motif reduces dysregulated transepithelial migration of neutrophils, presumably by inhibiting the binding of MAC-1 to ICAM-1 molecules expressed on the surface of the inflamed epithelium (66, 67). Moreover, the Lex motif expressed on MAC-1 mediates binding to DC-SIGN expressed on DCs, thus providing an indirect link between innate and adaptive immunity. Monocytes and macrophages also express MAC-1, but since they lack the Lex motif, this trait is exclusively dependent on neutrophils (68). Therefore, the Lex motif may represent a novel target for modulating inflammation in chronic diseases in which tissue damage is mediated by dysregulated neutrophil trafficking. After neutrophils accumulate at the site of inflammation, their immunomodulation depends on efficient degranulation, phagocytosis, and NET formation. Interestingly, granule glycoproteins show differential, stage-dependent glycosylation that decorates them with hyper-truncated chitobiose core, paucimannose and complex monoantennary N-glycans (64). Interestingly, it was recently demonstrated that N-acetyl-Beta-D-Hexosaminidase (Hex) enzyme is elevated in many inflammatory diseases (69–71) and catalyzes formation of paucimannosidic glycans found on neutrophilic granules (72). Those paucimannosidic N-glycoforms of human neutrophil elastase (HNE) show an enhanced ability to suppress the growth of P. aeruginosa (PA), presumably by bypassing interactions with its suppressive counter-binding partner, A1-antitripsyne (A1AT) (73). However, PA was demonstrated to uptake host’s sialylated N-glycans, making these bacteria suitable binding partners for the inhibitory siglec-9 receptor expressed on neutrophils. As a result, neutrophils show reduced levels of reactive oxygen species (ROS) and released elastases which leads to reduced formation of NETs and increases survival of PA (74), eventually leading to chronic lung inflammation and tissue destruction as seen in CF patients susceptible to respiratory infections caused by PA. Therefore, further studies are needed to draw conclusions about neutrophil glycosylation role in CF pathogenesis and completely illuminate these processes. Another modulatory potential of neutrophil activity lies in alpha-1-acid glycoprotein (AGP-1). AGP-1 can stimulate neutrophil activation by inducing an increase in cytosolic calcium concentration through interactions with the neutrophil receptors siglec-5 and/or siglec-14, which preferentially bind α2-3 or α2-6 sialic acid. The latter is presumably true for acute inflammation as hyperfucosylation of AGP-1 in chronic inflammation leads to increased expression of sLex on AGP-1, a motif that is not a ligand for siglec-5 nor siglec-14 (75). Furthermore, in addition to the liver, AGP-1 is expressed by activated neutrophils. In contrast to the N-glycans expressed on hepatic AGP-1 (hAGP-1), neutrophil-expressed AGP-1 (nAGP-1) carries mainly high-mannose, nonsialylated, and mono-sialylated N-glycans (76). Taken together, these data may indicate that chronic inflammation in some cases attenuates neutrophil recruitment and activation in favor of other, more potent leukocytes.

Monocytes and tissue macrophages are part of the mononuclear phagocyte system, which plays a central role in inflammation through antigen presentation, phagocytosis, and cytokine-mediated immune modulation (77). These mononuclear leukocytes are considered hallmarks of the transition from acute to chronic inflammation, as their accumulation is the result of cytokine-induced neutrophil apoptosis and increased production of monocyte chemoattractant protein (MCP-1) (78). Not surprisingly, several studies have uncovered different mechanisms in which monocytes and macrophages are involved in the modulation and maintenance of chronic inflammation. In particular, these have been demonstrated in cardiovascular disease (CVD) (79), RA (80, 81), chronic obstructive pulmonary disease (COPD) (82), diabetes (83, 84), and IBD (85). In the last decade, N-glycosylation has gained much attention as a tool by which inflammation orchestrates the immune response of monocytes and macrophages. There are three main steps involved in the accumulation of macrophages in the inflamed environment: recruitment of monocytes from the circulation, differentiation into macrophages, and activation of macrophages at the site of inflammation (77). All three steps are under the direct influence of altered N-glycosylation ( Figure 3 ). First, after monocytes are recruited by MCP-1 (83), their passage through the endothelial layer requires a complex system of interactions with heavily N-glycosylated adhesion molecules. Previous studies have shown that proinflammatory cytokines such as TNFα can upregulate the expression of adhesion molecules (ICAM-1, VCAM-1, E-selectin) (12, 13) and regulate their N-glycosylation (86). This was confirmed by Chacko et al., who further identified the mannosidases MAN1A2 and MAN1C1 as subjects of decreased expression by TNFα. These mannosidases catalyze the early removal of mannose, which is required for the conversion of high mannose to hybrid and subsequently complex N-glycans. Using THP-1 monocytes and PPARy ligands, they also demonstrated that the dual function of TNFα, stimulation of adhesion molecules and regulation of their N-glycosylation, is controlled by independent pathways, underscoring the importance of high-mannose N-glycans for monocyte trafficking (87). Recently, Regal-McDonald and his team specified that MAC-1 receptors expressed on the intermediate, proinflammatory, subclass of monocytes (CD14⁺CD16⁺) have a higher affinity for HM-ICAM-1 compared with the classical subclass (CD14⁺CD16^-). They also showed that the monocytes bind with higher affinity to HM-ICAM-1 than to α2,6-sialylated ICAM-1 (48, 49). After migration through the endothelium, monocytes differentiate into M0 macrophages, which can further polarize into different subclasses of macrophages stimulated by different cytokines - classically activated M1 macrophages, alternatively activated M2 macrophages, CD169 macrophages, or TCR macrophages (88). During differentiation from monocytes to macrophages, ST6GAL1 is downregulated, resulting in the loss of α2,6-sialic acid. Importantly, ST6GAL1 downregulation results from cleavage by Beta-Site APP-Cleaving Enzyme 1 (BACE1), which is dramatically upregulated during macrophage differentiation (89). This may also occur during differentiation into M1 macrophages, as it has already been shown that M2 macrophages, associated with anti-inflammatory effects, exhibit higher ST6GAL1 production compared with M1 (90). Moreover, α2,6-sialylation is included in the regulation of macrophage survival. Liu et al. showed that ST6GAL1 mediated α2,6-sialylation of TNFα death receptor 1 (TNFR1) expressed on primary macrophages inhibits apoptosis (91). This discovery further highlights the protective properties of α2,6-sialylation due to its promotion of proinflammatory M1 apoptosis, and survival of anti-inflammatory M2. In addition to sialylation, fucosylation also affects M1/M2 polarization in chronic inflammation. In RA, M1 macrophages were shown to express 5-10 times more fucosyltransferases (FUTs), catalyzing terminal and subterminal fucosylation (FUT1, FUT3, FUT7, and FUT9), than their monocyte progenitors, whereas this was not observed for FUT8, responsible for core fucosylation (92).Interestingly, terminal fucosylation is important for the synthesis of the sLex epitope, which is considered proinflammatory (93), whereas core fucosylation has more anti-inflammatory properties (94). Inhibition of terminal FUTs such as FUT1/2 leads to a shift in M1 differentiation toward M2 macrophages (92). This study was the first to highlight the potential of terminal fucosylation as a novel hallmark of inflammatory M1 macrophages. Another subclass of macrophages worth mentioning is CD169⁺. These macrophages express high levels of CD169, also known as sialoadhesin or Siglec-1, and are strategically positioned at the entry site of lymphoid tissue, where they bind and remove pathogens from the lymphatic fluid and blood (95). In addition to “gatekeeper” CD169⁺ macrophages that constitutively express Siglec-1, its expression can be upregulated in other tissue macrophages upon exposure to a type I interferon (IFN-I) (96). Siglec-1 belongs to the sialic acid-binding immunoglobulin-like lectins (Siglecs) (97), preferentially binding to α2,3-linked sialic acids (98), and differing from other siglecs as it has a long extracellular region (17 Ig domains) that lacks intracellular signaling motifs (99). In their review, O’Neill et al. introduced Siglec-1 as a macrophage-specific marker of chronic inflammation and emphasized its contribution to cell-cell and cell-matrix interactions of macrophages in inflammation (100). Also, soluble Siglec-1 (sSIGLEC-1) has been presented as a marker of monocyte and macrophage activation as well as a marker of interferonopathy in SLE and other inflammatory disease (101). Recently, Wang et al. showed that there is a TNFα mediated reduction in α2,6-, but not α2,3-terminal sialylation in RA (59). The latter would support interactions of proinflammatory synovial fibroblasts with pathogenic macrophages via Siglec-1, whose expression is upregulated in macrophages in RA (59, 102). On the other hand, Tanno et al. showed decreased expression of Siglec-1 on alveolar macrophages in COPD, significantly reducing their phagocytic capabilities against microbial pathogens and thus maintaining the inflammatory state (103). This clearly demonstrates how inflammation alters glycosylation and glycan binding molecules to maintain inflammatory environment, rather than the other way around. Mesenchymal stem cells (MSCs) have recently gained much attention as potential macrophage immunomodulators in chronic inflammation. As mentioned previously, they can be α(1,3)-exofucosylated to express the sLex-decorated CD44 ligand (known as HCELL), allowing them to migrate to the site of inflammation (28, 104). Further evidence suggests that targets of MSC mediated immunosuppression include macrophages, as there is evidence of the ability of adipose derived MSCs to shift macrophages from the M1 to the M2 phenotype (105, 106). Whether binding of HCELL ligand to Siglec-1 expressed on macrophages may be a possible mechanism of immunosuppression, thereby stopping macrophage interaction with other immune cells, is a question that remains to be answered. Another family of glycan-binding proteins involved in macrophage-mediated immunomodulation are the galectins. Among the 15 galectins identified to date expressed by immune cells, galectin-1 (Gal-1) and galectin-3 (Gal-3) show significant expression in macrophages and monocytes (107, 108). Both Gal-1 and Gal-3 possess a conserved carbohydrate recognition domain (CRD) that recognizes glycans containing a terminal N-acetyllactosamine (Galβ1,4GlcNAc or LacNAc). However, the terminal α2,6-sialylation present on LacNAc prevents the binding of Gal-1, whereas this was not observed for Gal-3 (109). Since the contribution of Gal-1 and Gal-3 to immunomodulation has been discussed in great detail elsewhere (108, 110, 111), we were focused here only on recent discoveries regarding the interplay between these galectins and macrophages. According to several different studies, soluble Gal-3 is mainly associated with proinflammatory functions (112–114). However, Di Gregoli et al. in their recent work proposed a new mechanism in which high levels of circulating Gal-3 in inflammation could be a consequence of matrix metalloproteinase 12 (MMP-12) dependent cleavage of Gal-3 from the surface of macrophages. Moreover, Gal-3 negative macrophages tend to exhibit proinflammatory properties, whereas Gal-3 positive macrophages exhibit pro-resolving and profibrotic properties (115). The latter is consistent with previously published data that revealed a novel mechanism of alternative M2 macrophage activation through binding of Gal-3 to its heavily N-glycosylated CD98 membrane receptor while emphasizing the association between M2 macrophages and increased fibrosis (116). Thus, although Gal-3 does not show direct proinflammatory effects, it is still part of the axis in maintaining the inflammatory state. On the other hand, Gal-1 seems to be a “jack of all trades” in resolving inflammation (111). Regarding the macrophages, Gal-1 is known to promote the differentiation of macrophages into the M2 profile (111, 117, 118). Yaseen et al. recently uncovered a positive feedback loop involving interferon β (IFN-β) mediated expression of Gal-1 in proinflammatory macrophages, thereby promoting their reprogramming into a pro-resolving phenotype with high expression of IFN-β (119). However, the specific mechanism is still unknown. One possibility is autocrine stimulation of proinflammatory macrophages by binding self-expressed Gal-1 to N-glycan ligands on their surface that normally lack the inhibitory α2,6-linked sialic acid. The Gal-1/IFN-β feedback loop is thought to occur at the time of termination of acute inflammation, so any misstep (e.g., insufficient Gal-1 expression) could lead to the development of a chronic inflammatory state, as decreased Gal-1 has been observed in several chronic inflammatory states (120–122).

DCs are antigen-presenting cells with the ability to take up antigens in the periphery and expose them to lymphocytes, thus bridging the gap between innate and adaptive immune responses (123). A specific subset of DCs derived from monocytes (Mo-DC) plays a key role in inflammation (124). The surface of Mo-DCs is covered with glycoproteins decorated predominantly with sialylated glycans (125). Sialylation of DCs is regulated during both differentiation and maturation, and has been found to significantly affect DC functions such as antigen uptake, phagocytosis, and T cell priming (126) ( Figure 3 ). Immature Mo-DCs are often tolerogenic because they have high levels of α2,6- and α2,3-sialylated N-glycans that are recognized by inhibitory siglecs (127) and galectins (128), respectively. By binding to α2,3-sialic acid-decorated CD43/CD45 clusters expressed on DCs, Gal-1 has been shown to support differentiation of tolerogenic DCs, thereby promoting interleukin 10 (IL-10) mediated T cell tolerance and suppression of autoimmunity (129). However, DC maturation in the presence of proinflammatory stimuli results in significant downregulation of expression and activity of ST6GAL1 and ST3GAL4 (130, 131), which may cause phenotype switch to inflammatory DCs. In contrast to the strictly tolerogenic activity of Gal-1, there are conflicting data regarding the control of inflammatory and tolerogenic DC phenotypes mediated by Gal-3 (128, 132, 133), reflecting the fact that this is a context- and tissue-dependent phenomenon. Additionally, by regulating T cell differentiation, DCs may also indirectly contribute to altered glycosylation of IgG molecules. Gringhuis et al. identified the molecular mechanism by which fucose specific triggering of DC-SIGN leads to increased Interleukin 27 (IL-27) expression by DCs, which promotes differentiation of T follicular helper (Tfh) cells known to affect IgG production by B cells (134, 135). Interestingly, DC-SIGN preferentially binds N-glycans with fucose incorporated into the Lex epitope (136), which is abundantly expressed by various immune cells during inflammation, suggesting another potential mechanism for maintaining the inflammatory state. While the presence of sialic acids has a tolerogenic effect on DCs, fully desialylated DCs exhibit a much more potent phenotype – high expression of major histocompatibility complex (MHC) molecules, secretion of inflammatory cytokines, phagocytosis, and activation of inflammatory T cells (137). Although the exact mechanism is still unclear, sialidases such as neuraminidase 1 and 3 (NEU1 and NEU3) are thought to contribute to the desialylation of DCs (138–140). This hypothesis is also supported by the fact that sialidases are abundant and involved in the pathology of many inflammatory diseases (141). However, Lübbers et al. have recently demonstrated an alternative pathway for the induction of tolerance by DCs independent of their sialylation status, driven by the immunoregulatory sialic acid-siglec axis. Specifically, binding of α2-3-sialic acid to Siglec-9 expressed on the surface of DCs alters metabolic pathways and cytokine signaling and reprograms DCs to enhance regulatory T cell/T helper type 1 (Treg : Th1) ratio balance (142). Collectively, these data highlight the importance of glycan recognition by DCs in controlling both inflammation and its resolution.

NK cells are known for their role in cell-mediated cytotoxicity and secretion of proinflammatory cytokines (143), which are critical for both the promotion of inflammation and immune regulation (144). The effector functions of NK cells are regulated by a series of activating and inhibitory receptors expressed on their surface, with glycosylation playing a key role in receptor-ligand recognition ( Figure 3 ). FcγRIIIa (CD16a) is the most abundantly expressed activating receptor on circulating NK cells (145), and its role in antibody dependent cell mediated cytotoxicity (ADCC) is well established (146). While it is established that modulation of IgG N-glycome significantly affects its binding to FcγRIIIa (94), Several studies made observations that underscore the importance of N-glycosylation of FcγRIIIa for IgG binding affinity. Tremendous increase in binding affinity of proinflammatory afucosylated IgG was observed when oligomannose N-glycans were present on FcγRIIIa (147, 148), which correlated with decreased expression of α-mannosidase in NK cells (149). Furthermore, higher levels of sialylated complex N-glycans on FcγRIIIa were shown to correlate with lower affinity for antibody binding (150), which was also observed for the activating NK cell receptor 2B4 (CD244) (151). In their recent review, Rosenstock and Kaufmann describe an important contribution of sialic acids to the functions of NK cells, both through the expression of sialic acid-binding receptors and by having sialic acids on their surface (152). Two of these receptors are Siglec-7 and Siglec-9, which have an inhibitory function on NK cells. While Siglec-7 mainly recognizes tumor-expressing gangliosides (153), Siglec-9 has a high affinity for α2,6- and α2,3-linked sialic acids, including the sLex epitope (154). Cytokines such as interleukin 2 (IL-2) and interferon α (IFN-α) have been shown to increase the level of sialylation on the surface of NK cells (155, 156). Although increased sialylation is usually considered to be anti-inflammatory, the functional role of these sialic acids may be to mask Siglec-9 through cis interactions, and thus preventing the inhibition of NK cells that would occur through trans binding of sialic acids. The importance of Siglec-9 in NK cell immunoregulation has been demonstrated in liver inflammation, where decreased Siglec-9 expression has been associated with disease progression (157). Although glycosylation in NK cells is functionally important, there is little information on the underlying mechanisms that alter N-glycosylation of NK cells during inflammation because of their relatively low abundance. However, the development of methods to enrich human NK cells from a single donor (149) may be a first step toward a more detailed analysis of inflammation induced N-glycosylation changes in NK cells.

T cells (T lymphocytes) have a central role in the adaptive immune system. Briefly, after differentiation from thymocytes to naïve T cells, T cells leave the thymus and enter the periphery. There, exposure to antigens by antigen presenting cells (APCs) such as macrophages and/or DCs along with concomitant cytokine stimulation triggers maturation of naïve T cells. In general, mature T cells carry a unique T cell receptor (TCR) and can express either CD4 or CD8 molecules, allowing the identification of CD4⁺ T helper cells (Th) and CD8⁺ cytotoxic T lymphocytes (CTLs). While CTLs can exert direct cellular cytotoxicity, Th cells are required for the initiation of humoral and cell-mediated immune responses. Thus, they can be divided into several subtypes based on functions and the production of specific cytokines - Th1, Th2, Th17, Tfh and Treg cells (159). The involvement of T cells through various mechanisms in the development and progression of chronic inflammation is undisputed (160–163). T cell function in inflammation is highly pleotropic and dependent on intra- and intercellular communication, which is often mediated by N-glycans and their corresponding binding partners ( Figure 4 ). In this regard, alterations in the N-glycome of T cells can significantly affect their activation, differentiation, survival, and cytokine production, often leading to autoimmunity, chronic inflammation, or cancer (164). Under homeostatic conditions, galectins are the major immune regulators of T cells, with Gal-1, Gal-3, and Gal-9 consistently showing immunosuppressive effects. The role of galectins in immunomodulation of T cells has been discussed in detail by several authors (111, 128, 129, 165). Therefore, we will specifically discuss galectin functions mediated by N-glycosylation in chronic inflammation, along with the latest findings on the underlying mechanisms affecting N-glycosylation in and by inflammation itself. Gal-1 and Gal-3 preferentially bind to branched N-glycans containing the LacNAc motif found on their T cell counter-receptors such as CD7, CD45, CD43 and TCR. This leads to inhibited transendothelial migration and induced apoptosis of T cells (108). The aforementioned binding is under the direct influence of the activity of glycan-modifying enzymes and the availability of corresponding substrates. One such enzyme is Golgi Beta-1,6-N-acetylglucosaminyltransferase V (MGAT5), which catalyzes the biosynthesis of tetra-antennary N-linked glycans, the preferred intermediates for elongation with (poly) LacNAc and ligands for galectins. MGAT5 expression in T cells is altered in chronic inflammatory diseases at both the genetic and protein levels. At the genetic level, several MGAT5 single-nucleotide polymorphisms (SNPs), associated with reduced expression of the MGAT5 enzyme, have been found to correlate with pathological changes in T cell glycosylation in chronic diseases such as IBD, COPD, and multiple sclerosis (MS) (47, 166, 167). Deficiency in the N-glycosylation branching pathway increases susceptibility to development of severe forms of disease due to the lack of galectins’ binding substrate and consequently their inability to inhibit the exuberant Th1/Th17 immune response (168, 169). In addition, N-glycosylation alterations may occur under the influence of various cytokines; an interesting study showed that in chronic viral infection, IL-10 induced expression of MGAT5 in CD8⁺ T cells promotes the formation of the Gal-3 lattice and increases the antigen activation threshold. Normally, this would be considered an anti-inflammatory mechanism, but this restriction in viral infection allows rapid replication of the pathogen, and thus leading to the establishment of persistent chronic inflammation (170). In terms of cytokine-mediated T cell N-glycome regulation, IL-2 is one of the most involved. Based on their research in MS, Grigorian and colleagues elegantly explained the paradoxical impacts of IL-2 on N-glycan branching and MGATs in T cells. Interestingly, IL-2 reduces N-glycan branching in resting T cells, whereas it has the opposite effect in activated T cells. This is thought to be a consequence of IL-2 induced upregulation of MGAT1, an enzyme that catalyzes the biosynthesis of mono-branched N-glycans, in resting T cells. MGAT1 has a ∼250-fold higher affinity for UDP-GlcNAc than MGAT5, thus increased MGAT1 expression inhibits further N-glycan branching by limiting UDP-GlcNAc availability to MGAT5. In contrast, in active T cells, TCR signaling appears to increase levels of MGAT5 and UDP-GlcNAc, thereby exploiting IL-2 induced upregulation of MGAT1 to increase N-glycan branching by providing more substrates for downstream enzymes (47, 171). IL-2 is also involved in T cell differentiation. It suppresses the formation of Th17 and Tfh while promoting the development and activation of Treg cells (172–174). The latter is critical for maintaining immune homeostasis, as Treg cell dysfunction is associated with several inflammatory diseases. Therefore it is no surprise that low-dose IL-2 therapy has shown improvement in various autoimmune and inflammatory conditions (175–177). Also, mature Treg cells on their surface carry IL-2R receptor consisting of three subunits, IL-2Rα (CD25), IL-2Rβ (CD122), and IL-2Rγc (CD132), of which CD25 is heavily N- and O-glycosylated (178). Reduced branching decreases surface expression and retention of CD25, inhibits proper IL-2 binding, and eventually prevents Treg cell activation which consequently promotes inflammation (179). In addition to glycosyltransferase activity, substrate availability is another critical factor for successful N-glycan branching. The hexosamine biosynthetic pathway (HBP) is the main source of UDP-GlcNAc, which is required for N-glycan branching. De novo synthesis of UDP-GlcNAc is characterized by the conversion of fructose-6-phosphate to glucosamine-6-phosphate by the rate-limiting enzyme glutamine-fructose-6-phosphate transaminase (GFPT). To complete the conversion, GFPT also requires glutamine. Thus, the synthesis of UDP-GlcNAc by HBP may directly compete with glycolysis and glutaminolysis for fructose-6-phosphate and glutamine, respectively (179, 180). Inflammatory Th1 and Th17 undergo a metabolic switch from oxidative phosphorylation to glycolysis and glutaminolysis during inflammation (181). Therefore, by switching to glycolysis alone during inflammation, Th1/Th17 indirectly starve the hexosamine pathway of fructose-6-phosphate and consequently UDP-GlcNAc. In addition, Th17 cytokines were shown to induce down-regulation of GFPT, UDP-GlcNAc and branching in abundantly present proinflammatory T cells. These data suggest that glycolysis drives Th17 over Treg differentiation, with Th17 cytokines further maintaining reduced N-glycan branching (179). Therefore, a potential treatment for autoimmune diseases could be with metabolites of the hexosamine pathway (180). As can be seen, alteration of N-glycan branching seems to have dual function in promoting inflammation; it abrogates immunosuppression by galectins and shifts fate toward inflammatory T cells. Nevertheless, N-glycan branching is not the only feature that influences immune modulation and polarization of T cells. It has long been known that Gal-1 preferentially kills proinflammatory Th1 cells over anti-inflammatory Th2 and Treg cells. The latter is explained by the fact that Th2 and Treg cells have higher expression of ST6GAL1, which is responsible for the synthesis of terminal α2,6-sialic acids, compared with Th1 cells, and are thus protected from galectin-mediated apoptosis (109, 130, 182). Not surprisingly, the expression of ST6GAL1 is altered in chronic inflammation. In SLE, the expression of ST6GAL1 is increased in autoimmune-activated T cells, which inhibits the binding of Gal-1 and thus contributes to the pathophysiology of SLE (183). Moreover, besides the lymphocyte-specific ST6GAL1, there is a soluble form of ST6GAL1 released from the liver which is also involved in the immunomodulation of T cells. Interestingly, in mice with hepatocyte-specific ablation of ST6GAL1, there was an increase in local inflammation and a decrease in systemic Ag tolerance projected via increased T cell activation, and thus greater susceptibility to T cell dependent inflammatory diseases. Paradoxical as this may seem with respect to galectin inhibition, this clearly demonstrates that galectins are not sufficient to carry T cell immunosuppression alone. This is consistent with the recent discovery that liver macrophages expressing the α2,6-sialic acid-specific Siglec, CD22, can inhibit α2,6-sialic acid decorated T cells, which provides an alternative liver-driven mechanism for maintaining systemic immune homeostasis (184). APCs also have a key role in T cell polarization and activation. Sialylation of antigens has been shown to cause a shift in the differentiation of effector T cells toward tolerogenic Treg through the sialic acid-siglec axis on DCs. This could open a new way to treat patients suffering from autoimmune diseases or allergies (142, 185). Finally, another important glycosylation trait on T cells that is altered in chronic inflammation is fucosylation. The TCR receptor requires core fucosylated N-glycans for its proper activation and function. This is mediated by the Alpha-1,6-Fucosyltransferase, FUT8. In SLE and IBD, the expression of FUT8 is strongly upregulated, resulting in a hyperfucosylated TCR and thus hyperactivated T cells that contribute to the pathophysiology of the aforementioned diseases (186, 187). On the other hand, core fucosylation is required for the expression of programmed cell death receptor 1 (PD-1), which is responsible for attenuating TCR signaling, resulting in depleted and unresponsive T cells (188). The hyper-core fucosylation induced upregulation of PD-1 expression could then provide an explanation for the impairment of T cells in chronic viral infections (189, 190). Unfortunately, the underlying mechanism of upregulated core fucosylation in chronic inflammation is still unclear and is a topic for further study. In summary, inflammation has apparently found every loophole in the N-glycosylation life cycle of T cells to turn the tide in its favor. Therefore, it is necessary to consider N-glycosylation during the development of anti-inflammatory therapy, and particularly in case of a therapy specifically targeting critical steps in the transition from homeostasis to inflammation.

B cells, also called B lymphocytes, are the major central effector immune cells in the humoral branch of adaptive immunity. During inflammation, naïve or memory B cells are exposed to antigens by APCs under co-stimulation of Th cells in the germinal center (GC) in secondary lymphoid organs. This induces activation and rapid proliferation of B cells and selection of high-affinity B cell receptors (BCRs) (191, 192). B cells expressing a high-affinity receptor enter the periphery, where they differentiate into plasma cells that secrete large amounts of antibodies (193). Once antibodies encounter their antigen, pathogen, or infected cells, their functions include neutralization, ADCC, phagocytosis, and complement-dependent cytotoxicity (CDC) (194). In addition to their function as precursors of antibody-secreting plasma cells, B cells are involved in suppression of T cells and secretion of relevant cytokines that control adaptive immunity (195, 196). N-glycosylation has a tremendous impact on B cell proliferation, differentiation, and effector functions ( Figure 5 ), but research on this topic lags far behind that of T cells. Nevertheless, there are implications that altered N-glycosylation in B cells may contribute to the development of various chronic inflammatory (autoimmune) diseases.

In B cells, among the best understood roles for lectin-glycan interactions are those of sialoglycans and Siglecs in BCR signaling. Sialic acids are often referred to as inhibitory “self-signals” because of their high local concentration on the surface of B cells. Thus, it is not surprising that Siglecs are considered major B cell immunomodulators (197). B cells express siglec-2 (CD22) and siglec-10 (ortholog of mouse siglec-G), both of which preferentially bind α2,6-sialic acid and act as inhibitory co-receptors of the BCR to maintain peripheral tolerance and prevent autoimmunity (198, 199). Thus, varying degrees of autoimmunity have been observed in mice lacking CD22, Siglec G, or both (200–202). Interestingly, ST6GAL1 deficient B cells show suppressed BCR signaling, yet mice deficient in both CD22 and ST6GAL1 showed restored BCR signaling, emphasizing the importance of α2,6-sialic/siglec axis in immunomodulation of BCR signaling (203). In addition to ST6GAL1, sialic acid acetyl esterase (SIAE) is another enzyme involved in regulation of BCR signaling. SIAE deacetylate sialic acid ligands, thereby allowing CD22-mediated inhibition of BCR signaling (204). Several studies showed that congenital mutations in SIAE are associated with an increased risk of autoimmune disease in humans (205–207). Nevertheless, while the contribution of sialic acid/siglec interactions to B cell function has been extensively defined (97, 199, 208, 209), sialoglycans represent only a fraction of the glycans involved in interactions regulating B cell immunity. Recently, Giovannone and colleagues discovered that B cells express significant amounts of tri- and tetra- complex N-glycans decorated with (poly)LacNAcs. Interestingly, the (poly)LacNAc structures were linear on naïve B cells but modified by Glucosaminyl (N-acetyl) Transferase 2 (GCNT2) with I-branches in GC B cells. In addition, I-branches were found to selectively impair B cell binding to Gal-9. Specifically, Gal-9 was found to be predominantly expressed by naïve B cells and to bind mainly the glycoprotein receptor CD45 carrying (poly)LacNAc decorated N-glycans. Functionally, binding of Gal-9 to CD45 induces inhibitory signaling through interaction with CD22, abrogated BCR calcium flux, and attenuated B cell activation. On the contrary, in GC B cells, Gal-9 mediated inhibition is down-modulated by the combined downregulation of Gal-9 and upregulation of GCNT2 (210). Although this is a novel BCR regulatory axis involving Gal-9 and GCNT2, further studies are needed to unravel the underlying mechanisms controlling the expression of these proteins. However, increased expression of Gal-9 has been reported in various autoimmune inflammatory diseases (211), therefore the exact involvement of Gal-9 in B-cell (dys)function in chronic diseases remains to be elucidated. In addition to Gal-9, other galectins such as Gal-1 and Gal-3 are also known to regulate BCR signaling, plasma cell differentiation, and survival (212–214). Interestingly, besides Gal-9, I-branches have also been shown to selectively impair B cell binding to Gal-3 but not Gal-1 (210, 215). A proposed explanation includes the fact that Gal-3 and Gal-9 preferentially bind to internal LacNAc residues, while Gal-1 favors binding to (poly)LacNAc termini (128). This suggests that I-branches may inhibit Gal-3 and Gal-9 binding, whereas terminal modifications such as α2,6-sialylation by ST6GAL1 may more selectively inhibit Gal-1. In addition, branched N-glycans are not exclusively a feature of B cell regulation but also of B cell mediated T cell regulation in autoimmunity. Branched N-glycans have been reported to suppress B cell triggered proinflammatory Th1/Th17 differentiation by promoting Toll-like receptor-2 (TLR2) and Toll-like receptor-4 (TLR4) endocytosis and downstream APC activity in B cells, thereby reducing inflammatory demyelination in a murine model of MS. At the same time, it was observed that minimal branching promotes surface retention of BCR and its co-receptor CD19, thereby stimulating adaptive B cell function. Although branching may represent another way to prevent Gal-9 binding and enhance BCR signaling, MGAT1 deficiency has been observed to prevent both branching and poly-LacNAc synthesis, leading to a decrease in BCR/CD19 surface expression and BCR signaling in this MS model (216). Because altered expression of glycosyltransferases may be a mechanism of differential regulation of galectin activity and receptor expression in B cells, future studies are needed to determine potential roles of these enzymes and their corresponding effector molecules in the development of autoimmune diseases. Lastly, it was demonstrated that core fucosylation of IgG-BCR mediates antigen recognition, along with cell signal transduction via BCR and antibody production (217).

The final, but not less important, role of B cells is the secretion of immunoglobulins (Igs) - the major executive glycoproteins of the humoral adaptive immune response. In humans, five classes of immunoglobulins exist: IgG, IgA, IgE, IgM, and IgD. All human Ig classes are N-glycosylated, with N-glycans attached to the conserved glycosylation regions on the fragment crystallizable (Fc) and/or on the variable fragment antigen binding (Fab), where new glycosylation sites can be acquired during somatic hypermutation (218). N-glycans can affect the structural stability and conformation of immunoglobulins as well as their effector functions (219). While alterations in N-glycosylation of IgG have been observed in several chronic (inflammatory) diseases and discussed in detail elsewhere (3), not much is known about alterations in the N-glycosylation profile of other immunoglobulins under pathological conditions. In the following paragraphs, the current knowledge about this topic is summarized with the focus on the possible underlying mechanisms mediated by inflammation that could contribute to the alterations in N-glycosylation of Igs.