Writers and readers of sialylation in immunoregulation in cancer

Abstract

Sialic acids are the terminal monosaccharides of the glycocalyx that critically shape cell–cell interactions and are strongly implicated in regulating immune recognition and tissue homeostasis. In cancer, aberrant sialylation rewires the tumor microenvironment by enhancing ligands of the inhibitory Siglecs, suppressing immune effector functions, and facilitating metastatic dissemination. This review provides a comprehensive synthesis of the dual role of sialyltransferases (the “writers”) and Siglecs/Selectins (the “readers”) in cancer progression. We examine the structural and functional diversity of these molecules, their dysregulation in malignancy, and their impact on tumor–immune dynamics. Finally, we highlight emerging therapeutic strategies, including sialyltransferase inhibitors, sialidase conjugates, and Siglec-targeted immunotherapies, which collectively position the sialome as a tractable frontier in cancer treatment.

Sialic acids (Sias) are a diverse family of nine-carbon acidic monosaccharides typically located at the terminal positions of glycan chains in glycoproteins and glycolipids, forming a crucial component of the cellular glycocalyx. Their unique chemical properties, most notably a carboxyl group with a pKa of approximately 2.6 (1), confer a persistent negative charge at physiological pH, profoundly influencing electrostatic interactions and cellular communication. The “sialome,” defined as the full repertoire of sialylated glycans displayed at the cell surface, is highly heterogeneous across tissues and species (1). This molecular heterogeneity is further amplified by the structural diversity of underlying glycans (N-glycans, O-glycans, and glycolipids) and a wide range of chemical modifications, including O-acetylation, sulfation, which constitutes a complex biochemical signature essential for immune self-recognition and homeostasis (2).

The sialome is dynamically regulated by the coordinated action of enzymes named sialyltransferases (STs) and neuraminidases, which act as the “writers” or “erasers,” respectively, of the sialome, while two families of lectins named Selectins and sialic acid–binding immunoglobulin-like lectins (Siglecs) are considered as “readers” of these sialylated patterns. Lectins recognitions of these sialylated ligands mediate fundamental physiological processes such as immune regulation, cellular adhesion, and signaling (2).

Aberrant sialylation is a well-recognized hallmark of malignant transformation, first identified in the late 1960s and sialoglycoconjugates play a key role in many disease processes (3, 4). In most cancers, sialylation levels increase, and during tumor progression, cancer cells direct their metabolism towards increased synthesis of sialic acids and control sialylation of glycoproteins involved in metastatic processes (5, 6). Tumor cells frequently exhibit hypersialylation, an overabundance of sialylated glycans, resulting in profound remodeling of the glycocalyx. For example, this thickened, highly charged barrier promotes immune evasion by shielding tumor cells from natural killer (NK) cell–mediated cytotoxicity (7), complement activation, via enhanced recruitment of factor H (2), and other immune effector mechanisms (8).

Dysregulated sialylation also affects the metastatic cascade, influencing cell detachment, migration, and adhesion (5, 9, 10) through altered charge distribution and receptor engagement (11). Sialylated tumor-associated carbohydrate antigens (TACAs), such as sialyl-Tn (sTn), sialyl-Lewis X (sLe^x), sialyl-Lewis A (sLe^a), polysialic acid (polySia), and gangliosides (GD2, GD3) (Fig. 1), are recurrently overexpressed across multiple cancer types (12), often correlating with poor prognosis and increased metastatic potential, immune evasion, and therapy resistance. These changes are largely driven by the dysregulation of STs (3, 13). For example, ST6GalI-mediated α2-6-sialylation of N-acetyllactosamine is markedly upregulated in several cancers, while ST6GalNAcI-dependent sTn synthesis—often facilitated by mutations of the molecular chaperone COSMC, required for proper T-synthase activity—is commonly found in gastric, pancreatic, and breast tumors (14). Such alterations are closely linked to oncogenic pathways, including Ras and c-Myc, establishing a direct connection between glycosylation changes and cancer progression. Similarly, overexpression of Selectins and their sialylated ligands, sLe^x and sLe^a, promotes tumor–endothelium interactions, facilitating metastasis (15).

Figure 1.

Sialylated TACAs reported in cancers. Major sialylated carbohydrate structures aberrantly expressed on tumor cells include sLe^x, sLe^a, sT, sTn, polySia. These antigens decorate glycoproteins and glycolipids such as mucins (MUC1, MUC16), influencing adhesion, migration, and immune recognition. Their overexpression promotes immune evasion and metastatic dissemination through interactions with Siglecs and Selectins.

Beyond their structural roles, tumor-associated sialoglycans act as immune checkpoints. The Siglec–sialoside axis is increasingly recognized as a novel immunoregulatory pathway, comparable to canonical checkpoint pathways such as Programmed cell Death protein 1 and its ligand (PD-1/PD-L1) and cytolytic T-lymphocyte–associated protein 4 (CTLA-4), and represents a promising target for immunotherapy. Given the growing recognition of sialylation as a determinant of tumor progression and therapy resistance, this review provides an up-to-date synthesis of the roles of STs, Siglecs, and Selectins in cancer, highlighting their molecular interplay, contributions to tumor immune evasion and therapeutic resistance, and their potential as targets for innovative immunotherapies.

Expression of sialylated signatures in cancers

Since the early 2000s, glycan-based biomarkers have gained increasing importance due to the essential role of glycoconjugates in cellular functions and interactions (16). Altered glycosylation patterns and TACAs are frequently observed in cancer cells (17), where they act as oncofetal antigens resembling those found during early development. Aberrant sialylation is widely recognized as a hallmark of cancer (3, 4), associated with tumor progression, metastasis, and disease status where sialylated TACAs modulate immune responses and are currently used as valuable clinical biomarkers as represented in Figure 1.

The sLe^a epitope is a sialylated antigenic determinant mostly reported on O-glycans in gastrointestinal tumors (18) and often elevated in pancreatic, gastric, and colorectal cancer (CRC) patient serum (19, 20). sLe^a is notably used as a sialylated marker named CA19-9 for the diagnostic of pancreaticobiliary tumors and the monitoring of pancreatic cancer (21, 22). In colon cancer, sLe^a expression is associated with E-selectin binding, whereas the disialylated form of Lewis A (di-sLe^a) interacts with Siglec-7 (23). Synthesized by ST6GalNAcVI (24), di-sLe^a expression decreases upon malignant transformation, while sLe^a levels increase, which may contribute to enhanced cancer aggressiveness and metastasis (23).

The sLe^x epitope has been reported both on various sialylated glycoconjugates with different biomarkers-associated names like carcinoembryonic antigen on N-glycans (19) for E-Selectin ligand in nonsmall cell lung cancer (NSCLC), in breast, gastric cancers (20, 25), CRC differentiation (26), and CA125 antigen on mucin-type O-glycans MUC16 in epithelial ovarian cancer (27). Their expression levels often correlate with an increase of metastasis through extravasation (28) and help in early diagnosis, prognosis, and monitoring in various cancers (21, 22). Sulfated forms of sLe^x have been reported on Gal and GlcNAc residues—6′- and 6-sulfo sLe^x, respectively. Antibodies are currently developed against 6′-sulfo sLe^x, which supports L-selectin–mediated adhesion (29) and is a high-affinity ligand for Siglec-8 (30) and against 6-sulfo sLe^x, which is a preferential ligand for Siglec-9 (31).

The Sd^a/Cad antigen has been reported either on N-glycans, core-1 and core-3 O-glycans, as well as on the GM2 ganglioside (32). Sd^a expression is notably downregulated in CRC, in balance with sLe^x (33). The α-fetal protein (AFP)-L3 is a specific antigen with aberrant core-fucosylation on AFP N-glycans. This biomarker has been used since 2001 (34) in hepatocellular carcinoma (HCC) detection and is still used for early diagnosis (35).

The sTn antigen, also known as CD175s, is an oncofetal antigen mainly expressed on MUC1, a glycoprotein bearing mucin-type O-glycans. sTn has been reported in breast cancer (36), as well as in other cancers, including gastric (37) or pancreatic ductal adenocarcinoma (PDAC) (38). Along with Tn antigen (Tn), sTn is used as a prognostic marker and to monitor cancer progression (14). Altered expression of sialyl Lewis and sTn structures notably promotes tumor growth and metastatic spread (17). The sialyl-T Thomsen-Friedenreich antigen (sT) is mostly reported on the O-glycans of MUC1 in prostate, ovarian, lung, pancreas, and breast cancers (39, 40, 41), while its disialylated version, the disialyl-T, has been found on human pancreatic cells (42) and is associated with Siglec-7 recognition (43, 44). CD65s, also named VIM-2, is the sialylated form of the O-glycoprotein CD65 used as a biomarker in leukemia. Notably in acute myeloid leukemia (AML), CD65s is a key ligand for E-Selectin and play a role in the extravascular infiltration and the growth of the cancer cells (45).

Different gangliosides have been associated with cancers, and their overexpression also serves as biomarkers and for monitoring disease progression. Fucosyl-GM1 is a sialylated structure reported in small cell lung cancer (SCLC) (46), high levels of GM2 have been noted in melanoma (47), and the gangliosides GD2, GD3, and their O-acetylated forms are highly expressed in melanoma, neuroblastoma, acute lymphoblastic leukemia, breast cancer, glioma, SCLC, and osteosarcoma (48, 49, 50).

N-glycolylneuraminic acid (Neu5Gc) is one of the aberrant sialylations found at higher levels on glycoconjugates in cancer tissues, cells, and serum samples compared with healthy conditions (51). Neu5Gc is thought to be acquired from dietary sources and may be preferentially incorporated into cancer cells due to their elevated metabolic rate. Under hypoxic conditions, the expression of the putative sialic acid transporter Sialin is increased, promoting enhanced Neu5Gc uptake and incorporation into glycoconjugates (11). Neu5Gc has been detected as a serological biomarker for breast and ovarian cancers and in urine samples from patients with bladder and prostate cancer (51, 52). It has also been found to be associated with the levels of the ganglioside GM3 in CRC, glioma, lung cancer, lymphoma, leukemia, melanoma, and neuroblastoma. Tumor-associated Neu5Gc-containing gangliosides, considered “xeno-autosialylation” products (11), represent potential immunotherapy targets, particularly under the hypoxic conditions that are characteristic of solid tumors. Although human enzymes are not able to synthetize Neu5Gc due to the loss of CMAH (1), it is not excluded that Neu5Gc biosynthesis could be de novo in human cancer cells (51). Likewise, no direct biological evidence has been demonstrated yet that human STs could transfer Neu5Gc, compared to other vertebrates STs (53). Interestingly, several human Siglecs, such as Siglec-10, show a stronger binding preference for ligands containing Neu5Gc over those with N-acetylneuraminic acid (Neu5Ac) (54).

PolySia chains are primarily found on the N-glycans of NCAM, notably in neuroblastomas, where they can reach a degree of polymerization of up to 55 residues (55). PolySia is often associated with higher-grade cancers, including CRC or NSCLC metastasis and aggressiveness (56), gliomas (57), and breast cancer (58). Its expression correlates with increased metastasis (59), reduced immune responses (60), and poor prognosis (61, 62). All these biomarkers reflect not only disease presence but also dynamic changes in the tumor sialome, which may influence therapy response and immune escape.

Sialyltransferases: The writers

Structure, function, and classification

STs are the last sequentially acting enzymes involved in the biosynthesis and regulation of sialylated glycoconjugates (63, 64). They are a family of type II transmembrane glycosyltransferases located primarily within the Golgi apparatus, where they catalyze the transfer of sialic acid from the activated donor CMP-Neu5Ac to the terminal positions of N- and O-linked glycans on glycoproteins, as well as glycolipids, with the inversion of stereochemistry at the anomeric center. This enzymatic “capping” of glycoconjugates generates a wide range of sialylated epitopes, which critically shape the physicochemical properties and recognition potential of the cellular glycocalyx (65). Twenty STs have been identified in the human genome and are classified into four major subfamilies according to their activity and modular nature in the family of glycosyltransferases GT29 (66). Despite their relatively low sequence identities, all 20 ST sequences are characterized by four conserved peptide motifs known as sialylmotifs (66, 67). The four STs families are named according to the type of glycosidic linkage formed and their acceptor substrate preferences.

The ST3GAL family is composed by six members, which catalyze the formation of α2-3-linkages on galactose (Gal) residues, the ST6GAL family with two members involved in the catalysis of α2-6-sialylation of Gal on N-glycans, whereas the six members of the ST6GALNAC family could mediate α2-6-sialylation of GalNAc on O-glycans and on glycolipids. The ST8SIA family is also composed by six members, which catalyze the transfer of sialic acid through an α2-8-linkage onto a sialylated structure that can be either N-glycans, O-glycans, or gangliosides, depending on the enzyme (64). ST genes are polyexonic and widely dispersed across multiple chromosomes, with their transcriptional and post-transcriptional levels tightly regulated by numerous tissue-specific and pathophysiology-dependent transcription factors and promoters (64, 66). In cancer, several transcription factors positively regulate ST expression. For example, Sp1 and Oct-1 promote ST6GAL1 transcription in colon adenocarcinoma (68) and leukemia cell lines (69), and at least four main promoters (P1–P4) are known to control ST6GAL1 expression (70). Likewise, Sp1, USF, and NF-κB enhance ST3GAL1 and ST3GAL5 expression in colon adenocarcinoma (71), ST3GAL6 in prostate cancer (72), and HNF-1, DBP, Sp1, and NF-κB regulate ST8SIA1 expression (73). In addition to these well-characterized positive regulators, numerous other transcription factors and regulatory elements identified in the GeneHancer Regulatory Element database may modulate ST expression in a context-dependent manner. Epigenetic mechanisms also contribute to ST regulation and are frequently altered in pathological conditions. Notably, ST6GAL1 (70) is reported hypermethylated in glioblastoma (GBM) (74) and bladder cancer (75). Noncoding nucleic acid transcripts, also called miRNAs, regulate gene and mRNA expression of STs. In cancers, downregulatory and upregulatory miRNAs have been identified for ST6GAL1 and ST6GAL2 genes (76). STs can also be influenced by different other factors in the Golgi environment such as pH, ions, redox homeostasis, posttranslational modifications, and cooperation with other glycosyltransferases as enzymes complexes (77) that could have a significant impact on their enzymatic activity (64).

The enzymatic activity of STs is sometimes oversimplified under a “one enzyme–one linkage” paradigm, which is often interpreted to mean that each ST catalyzes only one type of glycosidic linkage. However, this view does not capture the true diversity of ST function (65). While many STs display strong specificity for both the linkage formed and the nature of their acceptor substrates (Fig. 2), others exhibit overlapping substrate preferences, functional redundancy, or a degree of enzymatic promiscuity (78), enabling them to act on multiple acceptors or generate the same glycan structure via alternative pathways. This interplay of specificity, partial redundancy, and promiscuity reflects the evolutionary history of the ST family (66).

Figure 2.

Specific sialylated glycan structures made by individual sialyltransferases. Each colored frame corresponds to one ST family: ST6GAL (green), ST3GAL (blue), ST6GALNAC (red), and ST8SIA (yellow). These enzymes catalyze the transfer of sialic acid in different linkages—α2-6-, α2-3-, or α2-8—to Gal, GalNAc, or Sia residues, giving rise to the major sialoglycan motifs found on glycoproteins (GP) and glycolipids (GL). Their combined activity underlies the structural diversity and biological specificity of the human sialome.

Some STs exhibit very strict specificity toward their acceptor substrates. For example, ST3GalV exclusively synthesizes the GM3 ganglioside, hence the alternative name GM3 synthase, and ST3GalI and ST3GalII display redundant enzymatic activities, catalyzing the transfer of Sia in an α2-3-linkage to the same terminal Gal residue of Galβ1-3GalNAc (type 3 disaccharide), contributing to O-glycan biosynthesis. Despite their overlapping enzymatic activities, their tissue-specific expression differs: ST3GAL1 is predominantly expressed in the spleen and weakly in the brain, whereas ST3GAL2 shows the opposite pattern in mice (79).

ST6GalI and ST6GalII are active on LacNAc substrates (Galβ1-4GlcNAc) and also on the GalNAc residue of LacDiNAc substrates (GalNAcβ1-4GlcNAc), with an increased acceptor specificity for ST6GalII (80) explained by their molecular evolution (81) and the structural studies (82), which highlight that specific amino acid changes, as cancer mutations, could modulate STs enzymatic specificity.

The ST6GALNAC family also displays variability in substrate specificity. ST6GalNAcI has enzymatic specificity towards the Tn antigen to produce the sTn antigen both in vitro and in vivo (83, 84), whereas ST6GalNAcII can synthesize the sTn antigen in vitro (84) but primarily generates the sT antigen in vivo (79, 85). ST6GalNAcIII and ST6GalNAcIV show even more restricted specificity, acting on a limited set of O-glycoproteins and glycolipids to generate GD1α (63). Conversely, ST6GalNAcV and ST6GalNAcVI catalyze the synthesis of di-sLe^a on GlcNAc residues of gangliosides (24).

Among the ST8SIA family, ST8SiaI and ST8SiaV are mono-α2-8-STs responsible for gangliosides GD3 and GT3 biosynthesis, respectively (66, 86) although ST8SiaV can be active on other glycolipids, such as GM1b, GD1a, GT1b, and GQ1c. ST8SiaVI is involved on diSia formation onto O-glycoproteins like CD24 (87). ST8SiaIII is an oligo-α2-8-ST, which catalyzes the transfer of one or many Sias onto glycoproteins and glycolipids to form oligoSia chains with degree of polymerization = 3 to 7 (88) and could be responsible for polySia initiation. ST8SiaII and ST8SiaIV are poly-α2-8-STs responsible for polySia formation on glycoproteins with notably longer polySia generated for ST8SiaIV than ST8SiaII (88). Together, these transcriptional, epigenetic, and biochemical regulatory mechanisms, along with substrate- and tissue-specific enzyme activities, ensure dynamic and context-dependent control of the cellular sialome, notably during cancer progression and aggressiveness.

Dysregulation of sialyltransferases expression in cancer

Sialylation occurs in all human tissues, but ST expression is not uniform across them. According to data from The Human Protein Atlas and normalized transcriptomic pipelines, ST expression levels have been quantified across different tissues (89). ST3GAL1 and ST6GAL1 are ubiquitously expressed with ST3GAL1 showing high expression in the heart, kidney, thyroid, and liver, and ST6GAL1 being strongly expressed in the liver. Other broadly distributed enzymes, such as ST3GAL2 and ST8SIA1, are expressed at lower levels. Certain tissues exhibit higher and more specific ST expression: for instance, the brain shows elevated levels of ST3GAL5, ST6GALNAC6, ST8SIA2, ST8SIA3, ST8SIA4, ST8SIA5, and ST8SIA6; the reproductive system expresses ST3GAL4, ST3GAL5, and ST6GALNAC2, while digestive tissues express ST3GAL3 and ST6GALNAC1 at higher levels.

ST expression is frequently dysregulated in pathological states, particularly in cancer, where hypersialylation is a hallmark feature (3, 4, 13). Aberrant ST expression also constitutes a hallmark of oncogenic glycosylation, leading to hypersialylation, altered cell–cell and cell–matrix interactions, and the generation of TACAs. Quantitative glycomics and large-scale transcriptomic analyses have shown that ST overexpression can increase tumor sialylation by up to 60% (3), correlating with enhanced proliferation, migration, angiogenesis, metastasis, and immune evasion and therapy resistance. Several STs are thus directly implicated in cancer progression and represent potential therapeutic targets.

The oncogenic sialome is not shaped by individual enzymes acting alone but by the coordinated activity of multiple STs that together generate complex, high-affinity ligands for immune receptors. For instance, the formation of Siglec-7 ligands on mucins requires the combined action of ST3GalI and ST3GalII, which introduce α2-3–linked sialic acids, and of ST6GalNAcII, ST6GalNAcIII, and ST6GalNAcIV, which extend these structures into disialylated motifs (90). Likewise, the synthesis of 6-sulfated sLe^x, a potent L-Selectin ligand, depends on ST3GalI/II-mediated α2-3-sialylation in synergy with glycan sulfotransferases. These examples illustrate how STs act in a coordinated network to construct complex glycan signatures, fine-tuning tumor–immune interactions. Such structural complexity not only enhances the binding affinity for Siglecs and Selectins but also creates therapeutically actionable vulnerabilities, making these pathways attractive targets for disrupting immune evasion and metastatic dissemination. This interplay ultimately links ST activity to the immune regulatory functions of Siglecs, setting the stage for their pivotal role in modulating tumor progression.

ST3GAL family

The ST3GAL family has emerged as a key regulator of oncogenic glycosylation, modulating immune checkpoint engagement, epithelial-mesenchymal transition (EMT), angiogenesis, and metastatic potential. The tissue-specific expression and context-dependent roles of its six members underscore their dual potential as prognostic biomarkers and therapeutic targets across diverse cancer types. Their differential expression plays a central role in shaping the cancer glycome, influencing tumor growth, immune evasion, metastasis, and therapy resistance (13, 91). Global cancer genome analyses (TCGA) reveal a widespread overexpression of all ST3GAL enzymes in skin cutaneous melanoma, with more restricted upregulation in other cancers: ST3GAL1/2/3/4 in PDAC, ST3GAL3 in PDAC metastasis, ST3GAL1/5/6 in HCC, and ST3GAL1/3 in ovarian cancer, where their protein expression contribute to chemoresistance (92, 93, 94). Dysregulated expression patterns are also reported in osteosarcoma for ST3GAL1/2/3/4, in renal cancer where ST3GAL1 is overexpressed in contrast to ST3GAL5, in CRC where ST3GAL2 is upregulated and ST3GAL6 downregulated, and lung cancer where ST3GAL5 and ST3GAL6 are downregulated (91). In ovarian cancer, α2-3-sialylated glycoproteins are elevated, driven by ST overexpression, which promotes growth through EGFR and FGFR1 activation and facilitates metastasis via Selectin-mediated interactions and EMT (9).

ST3GalI is particularly well-studied for its role in modifying O-glycans. It catalyzes the sialylation of core-1 O-glycans T antigen, generating sialyl-T antigen and preventing further elongation of glycan chains. Overexpression of ST3GalI in breast and pancreatic cancers enhances the levels of sialyl-T antigens on MUC1, which can engage Siglec-7 and Siglec-9, leading to the suppression of NK cell activity and facilitating immune evasion (95). ST3GalI substrates include GFRA1, VASN, DAF/CD55, and neuropilin-1 (91). Notably, the α2-3-sialylation of neuropilin-1 by ST3GalI promotes tumor migration through EGF/EGFR signaling in metastatic breast cancer (96). In prostate cancer, ST3GalI-dependent Siglec ligands act as glyco-immune checkpoints (95). High ST3GAL1 mRNA expression also defines a GBM subpopulation with invasive and self-renewal properties and correlates with poor prognosis (97). In ovarian cancer, ST3GAL1 mRNA and ST3GalI protein levels are elevated, driving cell growth, invasion, and paclitaxel resistance, partly via induction of TGF-β1–mediated EMT both in vitro and in vivo (92). Similar associations have been reported in endometrial cancer, where its expression correlates with advanced stage, deep myometrial invasion, and VEGF-A upregulation (98), and in intrahepatic cholangiocarcinoma, promoting proliferation, invasion, and migration while inhibiting apoptosis (99).

ST3GalII, despite sharing functional similarities with ST3GalI, can also sialylate glycolipids such as GM1 and GD1b and is associated with the biosynthesis of stage-specific embryonic antigen-4 (100). High expression of ST3GalII in leukemia cancer cells can substitute ST3GalI activity in creating sialylated core 1 structures (43) Its upregulation in CRC correlates with tumor proliferation, migration, and invasiveness, while knockdown reduces tumor growth in xenografts (101). ST3GAL2 expression is also upregulated in androgen receptor–negative castration-resistant prostate cancer, where it modifies MUC1 (CA15-3) and SSEA-4 (102). In melanoma, ST3GAL1 and ST3GAL2 are transcriptionally and translationally upregulated compared to benign melanocytic nevi, where the α2-3-sialylation of CD98hc is critical for melanoma growth (103). This regulation is partly mediated by miRNAs that coregulate ST3GalI and ST3GalII protein levels and modulate CD98hc sialylation (104).

ST3GalIII and ST3GalIV primarily act on type I and type II chains (Galβ1-3/4GlcNAc), mediating the synthesis of sLeˣ, a key E-Selectin ligand involved in leukocyte trafficking and tumor metastasis (105, 106). ST3GAL3 expression is upregulated in pancreatic, prostate, and ovarian cancers, is associated with metastasis in PDAC (94), and modulates sLe^x expression in breast cancer cells (107). ST3GalIV is highly expressed in gastric tumors and contributes to P-Selectin binding in colon cancer and promotes metastasis in PDAC via activation of the ER stress PERK–eIF2α–ATF4 signaling pathway and mitochondrial dysfunction (108). It also drives sLeˣ biosynthesis in gastrointestinal cancers, modulating tumor cell motility (109). ST3GAL4 levels are overexpressed in breast carcinoma, cholangiocarcinoma, and AML compared to normal tissues (110). It is also upregulated in NSCLC, where it modulates α2-3-sialylation of receptor protein tyrosine kinases such as MET, contributing to Osimertinib resistance (111). Moreover, silencing of ST3GAL4 reduces proliferation, invasion, and migration of NSCLC cells by downregulating AKT1, ERK1/2, and NF-κB pathways (112). In breast cancer, ST3GalIV overexpression drives tumorigenesis by enhancing aerobic glycolysis and correlates with lactate dehydrogenase A expression (113). Elevated ST3GalIV levels are also associated with macrophage infiltration in the tumor microenvironment and poor prognosis in gliomas (110).

ST3GAL5 expression is context-dependent: its downregulation in bladder cancer correlates with poor prognosis, reduced infiltration of M0 macrophages, and increased tumor proliferation, whereas its overexpression inhibits tumor growth both in vitro and in vivo (114). Other STs (ST3GAL3, ST6GALNAC3, ST6GALNAC5, ST6GALNAC6, ST8SIA1, and ST8SIA6) have also downregulated mRNA expression in bladder cancer (114). Conversely, high ST3GAL5 expression in kidney renal clear cell carcinoma (KIRC) correlates with poor prognosis and promotes tumor progression (115). ST3GalV is also linked to EMT and VEGF-induced angiogenesis, with varied expression across cancers: downregulated in ovarian, pancreatic, prostate, and lung cancers but upregulated in leukemia, melanoma, lymphoma, kidney, esophageal, and head and neck cancers (116).

ST3GalVI, like ST3GalIII and ST3GalIV, synthesizes Galβ1-4GlcNAc and sLe^a antigen. It enhances P-selectin glycoprotein ligand-1 (PSGL-1) biosynthesis, supporting the homing and survival of multiple myeloma (MM) cells within the bone marrow niche (117). In prostate cancer, ST3GalVI upregulation increases sLe^a expression, correlating with disease aggressiveness (118). Additionally, ST3GalVI can sialylate EGFR, modulating tumor behavior, although it may exert a protective role in lung cancer (91).

ST6GAL family

The α2-6-sialylation on N-glycoproteins profoundly influences receptor stability, signaling, and cell–cell interactions, making these enzymes critical regulators of cancer-associated glycosylation. ST6GalI is by far the most extensively studied of this subfamily and consistently dysregulated ST in cancer. It is frequently overexpressed across a wide range of malignancies, including CRC and metastasis, ovarian, prostate, leukemia, gastric, and breast cancers and strongly associated with metastasis and poor prognosis (70, 119, 120, 121, 122, 123). Elevated ST6GalI levels have been also documented in PDAC and HCC, where this enzyme promotes β1-integrin sialylation, enhancing tumor invasion, chemoresistance, and protection from Fas-mediated apoptosis. In ovarian cancer cell models (PA-I, SK-OV-3), ST6GalI upregulation confers an invasive phenotype by altering integrin function and confers cisplatin resistance (121). The role of ST6GalI extends to prostate cancer, where its overexpression is linked to tumor growth and bone metastasis (124). Accordingly, α2-6-sialylated tri- and tetra-antennary N-glycans are abundant in prostate tumors but limited in normal prostate tissue (125). In rectal cancer, ST6GalI expression increases following chemoradiation, correlating with treatment resistance (126). In PDAC, it is markedly upregulated and cooperates with oncogenic KRAS^G12D to promote tumor progression and mortality in mice models (127). ST6GalI also exhibits protumorigenic functions in GBM, where high α2-6-sialylation of specific surface receptors such as PDGFRB supports aggressive tumor behavior (128). ST6GalI also contributes to modulate cancer metabolism and microenvironmental adaptation. In ovarian cancer cells, its activity correlates with oxidative metabolism rates and promotes cell invasion under hypoxic conditions (129). It is noteworthy that ST6GalI can be cleaved and released into the extracellular milieu or in exosome-like particles, where it compensates for intrinsic enzymatic activity and boosts proliferation and invasiveness in breast cancer cells (130). Functionally, α2-6-sialylation mediated by ST6GalI on death receptors such as TNFR1 and Fas confers protection against apoptosis, as shown in HEK293 cell models (131). The upregulation of specific microRNAs that enhance ST6GAL1 transcription and translation has also been implicated in driving abnormal α2-6-sialylation in multiple cancer types (76).

In contrast, ST6GalII exhibits lower and more tissue-specific expression and displays preferential activity toward LacdiNAc and LacNAc residues (80). Despite being less studied, ST6GalII shows differential regulation in several cancers: it is notably overexpressed in NSCLC, SCLC, neuroblastoma, and ovarian cancers according to the TCGA database, while being downregulated in HCC (132). In HCC, these alterations are associated with liver inflammation and may exert immunosuppressive effects by modulating immune cell infiltration during tumor progression (132). Altogether, the ST6GAL family, particularly ST6GalI, emerges as a central player in tumor progression, chemoresistance, metabolic reprogramming, and immune evasion. Their context-dependent expression underscores their potential as prognostic biomarkers and therapeutic targets across diverse cancer types.

ST6GALNAC family

The ST6GALNAC family plays a pivotal role in reshaping the cancer O-glycoproteome, producing sialylated TACAs that influence tumor growth, invasion, immune evasion, angiogenesis, and resistance to therapy (89). They emerge as key regulators of tumor-associated O-glycosylation, driving the formation of truncated and hypersialylated glycans that facilitate immune evasion, metastatic dissemination, and therapy resistance. Their tumor- and subtype-specific expression, for example, ST6GalNAcI in aggressive lung adenocarcinoma (LUAD), ST6GalNAcV in metastatic prostate cancer, or ST6GalNAcVI in NSCLC, highlights their dual potential as prognostic biomarkers and therapeutic targets in precision oncology (13, 24, 133).

ST6GalNAcI is one of the most extensively characterized ST6GALNAC family members, catalyzing the premature sialylation of GalNAc residues to produce the sTn antigen, a truncated O-glycan that prevents elongation of mucin-type glycans (83). ST6GALNAC1 transcript levels, ST6GalNAcI protein levels, and sTn are overexpressed in CRC, bladder, breast, ovarian, gastric, and pancreatic cancers and is associated with poor prognosis, invasiveness, and immune evasion (13). The activity of ST6GalNAcI is often potentiated by COSMC mutations, which disrupt core 1 β3-Gal-T–specific molecular chaperone function and lead to the accumulation of sTn (89). In NSCLC, ST6GalNAcI is upregulated in an aggressive Kras^Trp53 Ad-Cre LUAD subtype, where it sialylates NECTIN2 and MUC5AC, enhancing T cell immunosuppression, promoting angiogenesis, and facilitating liver metastasis (134). In breast cancer, ST6GalNAcI contributes to cell migration and invasion by altering O-glycosylation and promoting EMT, though it does not seem to impact proliferation (135). In CRC, lower ST6GALNAC1 expression correlates with poor survival, posttranslationally regulated by miRNA on protein levels in malignant tissues (136).

ST6GalNAcII shares overlapping functions with ST6GalNAcI (79), particularly in synthesizing sTn antigen, but has distinct associations with tumor aggressiveness. It correlates with poor survival in CRC and enhances invasive capacities in breast cancer (136). Like ST6GalNAcI, its expression correlate with Siglec-15 expression in CRC (136), underscoring its role in glyco-immune checkpoint modulation (137). Additionally, ST6GALNAC1 and ST6GALNAC2 expression have been proposed as biomarkers for NSCLC, in relation to altered glycosylation of mucins such as MUC4, MUC5B, MUC13, MUC16, MUC20, and MUC21 (137).

Other members of the ST6GALNAC family exhibit tumor- and context-specific functions. St6galnac3 and St6galnac4 expression has been correlated with disialyl-T expression in mice (138), while disialyl-T levels have been detected in human pancreatic cells (42) and associated with ST6GALNAC4 expression and Siglec-7 ligand biosynthesis in acute leukemia cell lines (139). ST6GalNAcIV is overexpressed in HCC, where it sialylates TGFBR2, enhancing TGF-β signaling and driving cell proliferation, migration, and invasion in vitro and in vivo (140). ST6GALNAC4 knockdown not only inhibits these malignant properties but also increases the infiltration of CD8⁺ T cells producing IFN-γ, TNF-α, granzyme B, and perforin, suggesting that ST6GalNAcIV fosters immune evasion (140).

ST6GalNAcV has a specific role in metastatic prostate cancer, being the only family member consistently and significantly upregulated in this context. It enhances ERK phosphorylation and promotes tumor cell invasion, under the transcriptional regulation of GATA2, with their co-expression serving as a poor prognosis signature for prostate cancer patients (133).

ST6GalNAcVI contributes to the biosynthesis of di-sLe^a antigen, a potential Siglec-7 ligand (23). Its expression varies depending on tumor subtype; elevated ST6GALNAC4 and ST6GALNAC6 expression levels are associated with poor survival in lung squamous cell carcinoma, whereas in LUAD, high ST6GALNAC6 expression correlates with better survival (137). In renal cancers, ST6GALNAC6 mRNA levels are significantly reduced compared to normal tissues, affecting the biosynthesis of disialylated globosides (141). Beyond transcriptional regulation, mutational events also impact the ST6GALNAC family. In cervical cancer, 14 STs are reported as mutated—particularly ST6GalNAcIII, ST6GalI, and ST8SiaV—that serve as predictive factors for prognosis and therapeutic response (142).

ST8SIA family

The ST8SIA family contributes to the formation of high-density α2-8-linked sialylated structures that reshape the tumor glycocalyx, modulating cell–cell interactions, immune recognition, and metastatic behavior. By generating Siglec ligands and engaging immune checkpoint pathways, these enzymes facilitate immune escape while influencing tumor plasticity and therapeutic responses, underscoring their potential as biomarkers and targets for cancer immunotherapy (4, 13, 17).

ST8SiaI, also known as GD3 synthase, catalyzes the production of b-series gangliosides which are highly expressed in melanoma and brain tumors and support cell growth, migration, and invasive behavior. ST8SIA1 mRNA overexpression in GBM drives the synthesis of GD3, promoting cell proliferation, clonogenicity, migration, and invasion, whereas its silencing significantly suppresses these oncogenic features (143). In breast cancer, ST8SIA1 expression is upregulated, particularly in tumors harboring specific p53 mutations, where it confers a survival advantage by modulating mitochondrial function and inducing resistance to p53-mediated apoptosis (144). Clinically, high ST8SIA1 expression correlates with increased tumor aggressiveness and poor prognosis across multiple cancers such as breast cancer (145), GBM, gliomas, melanoma, and neuroblastoma (146). ST8SiaI also contributes to cancer stem cell self-renewal properties, drives EMT, hyperactivates inflammatory pathways, and facilitates metastatic dissemination (146).

The two polysialyltransferases, ST8SiaII and ST8SiaIV, are key contributors to tumor biology where their overexpression in neuroblastoma, SCLC, and other tumors enhance metastatic potential and modulate immune interactions (147). ST8SiaIV expression levels shows particularly strong associations with chronic inflammation, as its high expression correlates with pro-inflammatory interleukin genes (IL-16, IL-21, IL-2), suggesting a link to an inflammatory tumor microenvironment (148). Furthermore, high ST8SIA4 expression in KIRC correlates with reduced responsiveness to immune checkpoint blockade therapy and associated with increased SIGLEC10 expression, pointing toward a Siglec–sialoglycan axis as an immune-modulatory pathway in this tumor context (149). A broader analysis of diffuse gastric cancers revealed elevated levels of ST8SIA1, ST8SIA4, ST3GAL3, ST6GALNAC5, and ST6GALNAC6 compared to intestinal gastric cancer subtypes, strongly associated with EMT features and correlates with multiple Siglecs (-1, -2, -3, -7, -8, and -9), suggesting their role in tumor cell migration, metastasis, and immune evasion (148).

ST8SiaIII has few links to cancer but has a role in the biosynthesis of gangliosides promoting proliferation, migration, and clonogenicity in GBM cells (150). Likewise, ST8SiaV show weak association with cancer progression, although TCGA dataset suggests that decreased expression may correlate with poor survival in CRC patients (13).

ST8SiaVI has a more specialized role, modifying the O-glycosylation of CD24 and generating di-sialylated structures that act as Siglec ligands (44). In melanoma and CRC, its overexpression promotes tumor growth and worsens survival outcomes, while inducing tumor-associated macrophages (TAMs) to adopt an M2-like immunosuppressive phenotype in murine models, independently of PD-L1 signaling (151). Functionally, ST8SiaVI-mediated CD24 sialylation enhances CD24 membrane localization, reinforcing an immune-suppressive, protumorigenic phenotype. CD24 itself is upregulated in breast, KIRC, and HCC with its expression strongly correlating with ST8SIA6, particularly in breast cancer (87). These observations align with Siglec-7, Siglec-9, and Siglec-10 expression and significantly correlate with macrophage infiltration in over 90% of the cancers (87). Interestingly, in colon cancer, high ST8SIA6 mRNA expression is instead associated with improved immune infiltration and enhanced responsiveness to immunotherapy, suggesting a context-dependent role and involvement in Siglec ligand biosynthesis for this enzyme (152).

Siglecs and Selectins: The readers of the sialome

Siglecs and Selectins are two major families of carbohydrate-binding proteins that act as “readers” of the sialylated glycocalyx, translating complex glycan structures into cellular responses. Together, they serve as critical mediators of immune regulation, cell adhesion, and intercellular communication in both physiological and pathological contexts, including cancer.

Siglecs are type I transmembrane glycoproteins predominantly expressed as cell surface receptors on immune cells, where they fine-tune immune responses by recognizing sialylated glycoconjugates (8, 153). In humans, 15 Siglecs have been identified, falling into two main subgroups: conserved Siglecs (-1, -2, -4, and -15), which maintain relatively stable evolutionary roles, and the more rapidly evolving CD33-related Siglecs (-3, -5 to -11, -14, and -16), which display diversified ligand preferences (54, 154). Structurally, Siglecs possess an N-terminal V-set Ig-like domain for sialic acid recognition and binding, anchored by a conserved arginine that engages the carboxyl group of Sia, followed by one or more membrane-proximal C2-set Ig-like domains that provide structural spacing and stability. The length of the extracellular domain varies considerably between family members, with Siglec-1 having the longest ectodomain (16 Ig-like domains) whereas CD33 and Siglec-15 contain only a single C2-set domain, a structural diversity that can influence receptor accessibility and binding in dense glycocalyx environments (8, 153).

The majority of Siglecs are inhibitory, with cytoplasmic tails containing immunoreceptor tyrosine-based inhibitory/switch motifs (ITIMs and ITSMs) that upon phosphorylation by Src family kinases, recruit the SH2 domains of SHP-1 and SHP-2 phosphatases, leading to dephosphorylation of key signaling intermediates and suppression of activating pathways including MAPK, NF-κB, and calcium mobilization pathways (153, 154). In contrast, a subset of Siglecs, such as Siglec-14, Siglec-15, and Siglec-16, lack intrinsic inhibitory motifs but possess a positively charged residue within their transmembrane domain, enabling constitutive association with the adaptor proteins DAP12, which can initiate activating signaling cascades (155). Individual Siglec–glycan interactions are typically of low affinity in the monovalent state (8). However, biological engagement and outcome are significantly influenced by factors such as density and multivalency (153). Additionally, fine-tuning through modifications, such as sulfation, on Gal or GlcNAc residues (90, 156, 157) and acetylation on sialic acid (158) further modulate these interactions. Siglecs are notably helpful to define the profile of myeloid-derived suppressor cells (MDSCs) involved in the immunosuppressive activities in the TME (159, 160).

Functionally, Siglec engagement can occur in cis, when sialoglycan ligands are located on the same cell surface as the receptor, often constraining receptor mobility and masking ligand-binding sites, or in trans, when ligands are presented on opposing cells, soluble glycoproteins, or microbial surfaces (8). The balance between cis masking and trans engagement is dynamic and can be altered during inflammation, cellular activation, or malignant transformation, thereby modifying signaling thresholds and immune cell responsiveness. Through these structural features, signaling properties, and context-dependent modes of ligand recognition, Siglecs act as key immunomodulatory checkpoints, dampening excessive activation, preserving self-tolerance, and integrating signals that shape both innate and adaptive immune responses (154). By engaging inhibitory Siglecs expressed on NK cells, MDSCs, macrophages, and dendritic cells (DCs) (Fig. 3), cancer cells dampen cytotoxicity and inflammatory responses, fostering immune tolerance within the tumor microenvironment (161, 162). TAMs support tumor cells immune evasion by secreting factors inhibiting T cells and NK cells (163).

Figure 3.

Landscape of Siglecs and Selectins interactions known with their sialylated ligands in the tumor cell microenvironment of cancer. Schematic representation of known interactions between tumor cell–associated sialoglycans and their cognate receptors on immune and endothelial cells. Siglec–sialoside recognition mediates immune suppression, whereas Selectin–ligand binding (sLe^x, sLe^a) facilitates tumor cell adhesion, rolling, and metastasis within the vasculature.

Ligand specificity varies across family members leading to the concept of Siglec–sialylated ligands axis where each cancer type display redundant or specific Siglec–Siglec ligands interactions recognized now as a new class of glyco-immune checkpoint (161, 164). Similar to PD-1/PD-L1 and CTLA-4 pathways, they deliver negative regulatory signals that blunt antitumor immunity. Notably, tumors with high sialylation profiles often display reduced responses to conventional immunotherapies, underscoring the need to integrate sialylation-targeted strategies into current treatment paradigms. Siglecs engagement suppresses ROS production (Siglec-3, Siglec-5, Siglec-7, Siglec-9, and Siglec-14), inhibits T-cell proliferation and cytotoxicity (Siglec-3, Siglec-5, Siglec-7, Siglec-9, and Siglec-10), promotes M2-like macrophage polarization (Siglec-1, Siglec-3, Siglec-7, Siglec-9, Siglec-10, Siglec-15, -and Siglec-16), and reduces antigen presentation (165).

Siglecs and their ligands in cancers

Siglec-1, also named CD169 or sialoadhesin, is expressed on activated monocytes, DCs, and specialized macrophage subsets in lymph nodes, spleen, and some tumor sites. Its ligands are preferentially α2-3-sialylated glycoconjugates. Unlike most other Siglecs, Siglec-1 lacks canonical intracellular signaling motifs, such as ITIMs or the positively charged residue required for DAP12 coupling. Its short cytoplasmic domain suggests that Siglec-1 functions primarily as an adhesion receptor rather than a signaling molecule. Siglec-1 expression in TAMs and tumor-draining lymph nodes are often correlated with better clinical outcomes in pancreatic, HCC, gastric cancers, and GBM (166), likely through enhanced CD8⁺ T and NK cell recruitment. In contrast, elevated Siglec-1 levels are associated with tumor aggressiveness and progression in endometrial, bladder, and breast cancers (167).

Siglec-2, also named CD22, is expressed predominantly on B-cells and is a well-characterized inhibitory co-receptor of BCR signaling (168), where it preferentially binds α2-6-sialyllactosamine ligands (169). CD22 primarily recognizes sialylated ligands in B-cell malignancies, including leukemia and lymphomas, modulates B-cell behavior within the TME, and is currently being targeted in combination immunotherapies (170). Prostate cancer progression also involves Siglec-2 engagement fostering M2-like macrophage polarization (124).

Siglec-3, or CD33, is widely expressed on myeloid progenitors, macrophages, monocytes, microglia (153). Its dysregulated activity is associated with Alzheimer’s disease progression where two proteins isoform of CD33 modulate microglial cell responses (171). High expression of Siglec-3 is also associated with lower survival in AML and has long been used as a target for immunotherapy (172).

Siglec-5 and Siglec-14 are homologous paired receptors, expressed both on neutrophils and monocytes. Siglec-5 is an inhibitory receptor whereas Siglec-14 is a DAP12-associated activatory receptor that drives pro-inflammatory responses. Interactions with their ligands regulate neutrophil–tumor cell interactions and could be targeted to improve antibody-dependent cellular cytotoxicity in solid tumors (173). Siglec-5 is also reported as an inhibitory immune checkpoint for human T cells (174).

Siglec-6 has restricted expression on mast cells, certain trophoblasts, and memory B cells. It displays endocytic behavior and recognizes glycolipids in a lipid-context dependent way and can mediate vesicle uptake (175), a property that could be co-opted by tumors as an inhibitory receptor. Siglec-6 expression is increased on circulating T cells and associated with lower survival in bladder cancer patients (176), upregulated in CRC (177), and is involved in migration and adhesion of B cells to bone marrow and spleen in blood cancers like CLL (178).

Siglec-7 is an inhibitory receptor found mainly on NK cells and certain myeloid populations such as monocytes, macrophages, and DCs and has emerged as a critical player in dampening antitumor innate immunity (44, 179, 180). Tumor cells frequently present sialylated and truncated mucin-type O-glycans, gangliosides, and disialylated motifs that engage Siglec-7 and inhibit NK-cell cytotoxicity. Siglec-7 prefers α2-8–linked disialylated ligands but can also recognize α2-3 and α2-6 sialic acids (44, 90) and bind gangliosides like GD3 (181). STs such as ST6GalNAcIV, ST6GalNAcVI, and ST3GAL families synthesize high-affinity ligands for Siglec-7 (23, 95). Oncogenic drivers such as MYC can enhance disialyl-T ligand biosynthesis via upregulation of ST6GalNAcIV, linking oncogenesis directly with immune escape (139). CD43 is described as a carrier of Siglec-7 ligands, which exhibits mainly disialyl core 1 O-glycans (43, 163). Siglec-7 ligands are expressed on T cells and regulate T-cell activation and polarization (182). In leukemia, CD162/PSGL-1, CD43, and CD45 are disialyl-T carriers which serve as Siglec-7 ligands and their overexpression protects leukemia B cell from NK cell cytotoxicity (43, 180). Tri-sT antigen has been reported on O-glycans of PODXL and MUC13 as Siglec-7 ligands in CRC cell line (Fig. 3). This glycan is synthetized by a combined action of ST3GalI, ST6GalNAcI, ST6GalNAcIII, and ST8SiaVI (44). Elevated Siglec-7 and Siglec-9 expression levels are associated together in cancer progression like immune evasion, TAMs differentiation and poor patient outcomes in prostate (95, 183) in triple-negative breast tumors (184), and in pancreatic cancers (185, 186).

Siglec-8 is an inhibitory receptor expressed on eosinophils and mast cells, whose canonical high-affinity ligand is 6′-O-sulfated sLe^x (30). Its direct involvement in cancer and tumor immune evasion appears less prominent than that of Siglec-7, Siglec-9, and Siglec-10. However, emerging evidence indicates that Siglec-8 is expressed in breast cancer, where its levels correlate with tumor-associated MUC1 expression (187), suggesting potential interactions mediated by MUC1-associated sialylated glycans.

Siglec-9 shares relatively high sequence homology with Siglec-7 but preferentially binds α2-3- and α2-6-sialylated ligands such as MUC16/CA-125 (188) while Siglec-7 prefers α2-8-linkages. Siglec-9 is broadly expressed across neutrophils, monocytes, subsets of DCs, NK cells, and activated T cells. It functions as a key inhibitory checkpoint in both myeloid and lymphoid compartments. Siglec-9 exhibits broad specificity for sialylated ligands, preferentially for sLe^x and 6-O-sulfated-GlcNAc sialo-epitopes (31) and protein carriers like MUC1 and MUC16 (163) (Fig. 3). AML cells express high levels of Siglec-9 ligands, which are synthesized by ST3GalIV and contribute to both phagocytosis sensitivity and immune evasion (189). In NSCLC, tumor-infiltrating T cells upregulate Siglec-9 expression, and interaction with Siglec-9 ligands inhibits T-cell activation, promotes immune escape, and enhances lung tumor growth in vivo (190). Similarly, breast cancer cells express Siglec-9 ligands that interact with neutrophils expressing Siglec-9, thereby inhibiting neutrophil cytotoxic activity both in vitro and in vivo (191). In ovarian tumor cells, the interaction of Siglec-9 with MUC16 attenuate T cell and NK cell function (188). Siglec-9 upregulation in GBM also correlates with phagocytosis inhibition and reduced survival in patients (192). Knockout of the murine ortholog Siglec-E led to the production of pro-inflammatory cytokines, more efficient T-cell priming, immune response, and resulted in prolonged survival in the murine models of GBM (192, 193).

Siglec-10 is an inhibitory receptor expressed on macrophages, some B cells, leukocytes, and DCs. CD24–Siglec-10 engagement protects tumor cells from macrophage-mediated phagocytosis, analogous and complementary to the CD47–SIRPα axis (194). In ovarian and breast cancers, Siglec-10 has been associated with CD24 and their expression levels correlates with higher rates of metastasis. Blocking this interaction increases phagocytosis and reduces tumor growth in vivo xenografted mice (194). In the TME of cervical cancer, Siglec-10 is expressed on infiltrated DCs and its interaction with ligands is a glyco-immune checkpoint and correlates with poor patient prognosis (195). Strategies for enhancing TAMs phagocytosis in HCC and solid tumors have been developed, like bispecific fusion proteins of SIRPα/Siglec-10 has been developed to block both CD47/SIRPα and CD24/Siglec-10 checkpoints (196). However, recent studies suggest that Siglec-10 ligands are more complex than CD24 (197), pointing out a broad affinity for multiple classes of N-linked sialoglycans (198).

Siglec-11 and Siglec-16 are another paired receptor, with an inhibitory effect for Siglec-11 and DAP12 activating Siglec-16 (199). Expressed on microglia and macrophages, their effects are counter-balanced to influence pro- or anti-inflammatory activation in brain mainly and supposedly in spleen, intestine, and liver (199). Siglec-11 on microglial cells recognizes polySia on human neuroblastoma cells (200) and this interaction confers a protective role inhibiting complement activation in inflammatory process (201). Siglec-11 ligands are reported to be sensitive to RNAse treatment, suggesting that they composed in part of the emerging area of glycoRNAs (202). In gastric cancer, Siglec-11 promotes M2 macrophages polarization, facilitate cancer progression through immune evasion (203) while in GBM, Siglec-16–polySia interactions are association with proinflammatory macrophage activation and higher survival of patients (204).

Siglec-15 is an evolutionarily conserved Siglec expressed on osteoclasts and macrophages, which is an activatory receptor when its intracellular region is coupled to DAP12 and DAP10. Structural insights into Siglec-15 and its binding epitope shown preferential binding mode to α2-3- and α2-6-sialylated structures (90, 205) and identified protein carriers like CD44 and CD11b (163) (Fig. 3). However, enhancing sTn expression on lung and breast adenocarcinoma cells did not lead to increased Siglec-15 binding (206). Glycan microarray analyses showed that sialylation on internal GlcNAc residue increase Siglec-15 binding (206, 207) and that ST6GalNAcVI is responsible for this internal sialylation in vitro (207), although this remains to be validated in cells. Upregulated across many human malignancies and on macrophages (208), Siglec-15 is recognized as a PD-1/PD-L1–independent critical immune checkpoint that suppresses T-cell responses in vitro and in vivo (208). This function has been described on macrophages in lung adenocarcinoma (209). Siglec-15 is notably upregulated in TAMs and trigger the production of immunosuppressive TGF-β in breast cancer (41). In colon adenocarcinoma, Siglec-15 and PD-L1 are associated with poor prognosis for patients, including a stroma component involved in immune evasion (210). Siglec-15 is a potential target for normalization cancer immunotherapy (208) as shown for metastatic breast cancer where targeting Siglec-15–ligands axis aim to treat bone metastasis and its spread to other organs (211).

Selectins and their ligands in cancers

Selectins are type I transmembrane glycoproteins comprising three members with distinct expression patterns and complementary roles (212). L-Selectin is constitutively expressed on most leukocytes, mediating their recruitment to sites of inflammation; P-Selectin is stored in the α-granules of platelets and in the Weibel–Palade bodies of endothelial cells and is rapidly mobilized to the cell surface upon activation; E-Selectin is induced on activated endothelial cells. Under physiological conditions, Selectins initiate leukocyte tethering and rolling on the vascular endothelium during inflammation and lymphocyte homing (15, 212). This process relies on the recognition of the sialylated and fucosylated sLe^x and sLe^a antigens. These TACAs are displayed on various glycoproteins, including PSGL-1, ESL-1, CD44 (Fig. 3), and mucins such as MUC1 or MUC5AC (20), as well as on other carriers, where ligand density, multivalency, and sulfation further enhance Selectin binding affinity (212).

In cancer, the physiological trafficking mechanism used by the Selectins is hijacked to promote immune evasion, tumor progression, and foster metastatic dissemination (10). By overexpressing sLe^x and sLe^a on Selectin ligands, tumor cells acquire the ability to bind with high affinity to E- and P-Selectins on activated endothelium and platelets, as well as to L-Selectin on leukocytes. The extent of E-Selectin expression on the vascular wall, together with the presence of its sialylated ligands on tumor cells, critically determines the efficiency of adhesion and extravasation at specific metastatic sites (15, 106, 213). Mechanistic studies have shown, for instance, that E-Selectin–mediated binding directs colon carcinoma cells to the liver microvasculature, where E-Selectin is abundantly expressed, thereby facilitating organ-specific metastasis (213).

In parallel, the formation of platelet–tumor cell aggregates via P-Selectin provides circulating tumor cells with a protective shield against immune clearance and shear stress, thereby enhancing their metastatic potential (214). P-Selectin expression is elevated in gastric and breast cancers but reduced in melanoma and CRC, suggesting tumor type–specific regulation associated with disease progression (214).

Clinically, high levels of sLe^x and sLe^a are consistently associated with poor prognosis and metastatic potential in CRC, breast (215), NSCLC (19), pancreatic, leukemia (45), and gastric carcinomas. Their carriers, such as MUC1, MUC5AC, α1-acid glycoprotein, and CEACAM5 (216), have been proposed as biomarkers of aggressive disease (10, 20).

STs like ST3GalVI tightly regulates sLe^a and sLe^x expression on E- and P-Selectin ligand, modulating interactions with Selectins and contributing notably to immune evasion or drug resistance in MM (117, 217). More recently, deep glycoproteomic profiling has uncovered additional physiologically relevant E-Selectin ligands enriched in sLe^x and sLe^a, which may serve as poor-prognosis biomarkers, notably in CRC (218). Collectively, these findings underscore the central role of Selectin–ligand interactions in cancer dissemination and identify them as promising targets for therapeutic intervention.

Strategies to study and target sialylation in cancer

Experimental approaches to decipher the sialome

Dissecting the complexity of tumor-associated sialylation and identifying Siglec ligands requires multimodal strategies as detailed in this review (219). Siglec–Fc fusion proteins have been optimized on different cell lines and tissues (220), using flow cytometry–based approaches (221, 222). Siglec ligands have been profiled through Siglec conjugations with antibodies for mapping on different cell lines (156, 220), with phages and the multivalent liquid lectin array platform on cells in vitro and in vivo (223) and on liposomes using SpyTag-SpyCatcher affinity, for in vivo and ex vivo tissue profiling, showing efficient immunomodulation (224).

Chemo-enzymatic synthesis has also been applied to study Siglec ligands affinity, like sulfated and sialylated structures on microarray (225, 226), for Siglec and Selectin ligands studies like Siglec-2 ligands (227, 228) or Siglec-7 ligands (229). Chemo-enzymatic editing on cell surface glycans using recombinant STs have been investigated for Siglec-2 and Selectins ligands in B cell lymphoma (228).

Cell-based platforms have been used to screen for high affinity Siglec ligands, notably by employing unnatural sialylated epitopes for detection (230) or by generating expanded sialome cell libraries (90). CRISPR/Cas9-mediated cell line engineering and overexpression systems have also been developed to examine how sulfation on sialylated ligands enhance Siglec interactions (156) and to understand the glycosylation pathway involved in sulfation process of Siglec ligands (231) and to identify which STs contribute to Siglec ligands biosynthesis (189). Alternatively, liposome-based formulations enable to probe Siglec–glycolipid interactions in a multivalent membrane-like environments (175) or testing binding of liposomes-conjugated Siglec ligands on different immune cells and explore Siglec modulation of FcγR signaling (179).

In in vivo models, more than 20 genetically engineered Siglec-KO mice models have been used to study Siglec biology (232) although murine Siglecs could display different binding than their human homolog, as shown with a range of gangliosides (233).

Development of sialylation-targeted inhibitors

STs are attractive drug targets due to their central role generating immunosuppressive sialoglycans. Strategies to inhibit these enzymes include both synthetic and natural products, with notable success in small-molecule inhibitors (234) (Fig. 4). Among these, the peracetylated 3F_ax-Neu5Ac acts as a global metabolic inhibitor of sialylation, while derivatives and natural products have shown potent activity (4). P-3F_ax-Neu5Ac use have shown to reduce tumor hypersialylation and delay tumor growth in vivo (235, 236).

Figure 4.

Immunotherapy strategies based on Siglecs/Selectins interactions with their sialylated ligands in the tumor cell microenvironment of cancer. Schematic overview of current and emerging therapeutic approaches designed to disrupt or exploit sialoglycan–lectin interactions in cancer. Strategies include inhibition of STs, blockade of Siglecs or Selectins with monoclonal antibodies (Drug-mAb), use of soluble recombinant Siglecs (sSiglecs) as decoy receptors, sialic acid mimetics formulated as nanoparticles (SAM-NPs), and antibody-based Bispecific T-cell Engager (BiTE) or CAR-T cell therapies targeting sialoglycan epitopes.

Sialic acid mimetics have been synthesized since 2002 with high affinity for Siglec-2 ligands by modifying the C-9, C-5, and C-2 positions on the sialic acid backbone (237, 238). Modifications incorporating alkyne or azide groups allow immobilization and testing on microarrays via CuAAC click chemistry (237). C-5 carbamates–fluorinated sialic acids inhibit sialylation with increased potency compared to C-5 amide analogs and show long-lasting effects (239). Likewise, C-4 modifications such as 4AzNeu5Ac are metabolically incorporated and tolerated for Siglec-7 interactions (240). In vitro assays have identified aryl-modified and nucleoside-modified derivatives as inhibitors of specific ST isoenzymes. Notably, nucleoside-modified analogs demonstrated promising inhibition of ST8SiaII activity compared to ST3GalI and ST6GalI (241). A comparative kinetic analysis of 13 human STs using a microplate-based sensitive assay, named MPSA (242), has provided key insights for targeted inhibitor development. Furthermore, a kinetic trapping approach has enabled the chemoenzymatic synthesis of a photo-crosslinkable CMP-3F_ax-Neu5Ac analog to probe ST selectivity and guide selective inhibitor design (243).

In vivo, applications of 3F_ax-Neu5Ac span multiple cancer contexts: in breast cancer cell lines, it enhances the phagocytic activity of TAMs (244); in MM, it reduces E-Selectin interactions, potentially restricting tumor cells to the bone marrow for improved therapeutic targeting (245). Another Sia-derivative inhibitor, P-SiaFNEtoc, has been shown to block bone metastases from prostate tumors in syngeneic mouse models (124). Similarly, an oral sialylation inhibitor has shown promising inhibition of STs activity in vitro and in PDAC cells, leading to tumor growth inhibition and TME remodeling in PDAC model (246). Other glycan-targeted molecules, such as sLe^x mimetics, have shown potent immunomodulatory effects in vitro and in vivo by binding both E- and P-Selectins (247).

As a comparison of inhibitor effects, CRISPR/Cas9-mediated desialylation through Cmas knockout has been shown to increase T-cell infiltration and improve survival in pancreatic cancer models (248), as well as to enhance immune responses and survival in CRC xenografts compared with anti-PD-L1 treatment (249).

Cancer immunotherapy targeting key players in sialylation

Immunotherapy has been revolutionized by immune checkpoint inhibitors targeting PD-1/PD-L1 and CTLA-4 (161). However, therapeutic resistance has driven the exploration of alternative checkpoints, including Siglecs and their sialylated ligands. Approaches include sialic acid mimetics, glycan-based approaches, sialidase-conjugated strategies, Siglec-blocking antibodies, CAR-T cells, and combination therapy (163, 227, 232) (Fig. 4). Various immune cells like T cells, NK cells, neutrophils, TAMs, and other infiltrating leukocytes express Siglecs that can be therapeutically targeted to restore antitumor activity (154).

Using a ST inhibitor, desialylation of stromal cells enhances T-cell activation in a Siglec-dependent manner in CRC (250) and helps unmask CD38 expression in MM, thereby potentiating NK cell cytotoxicity and the efficacy of anti-CD38 immunotherapies (251). Clinical developments include E-Selectin antagonists in MM and AML, Siglec-2 and Siglec-3 antibody–drug conjugates like inotuzumab and gemtuzumab ozogamicin in AML and patients (227, 252), enzalutamide to modulate Siglec-7 and Siglec-9 in prostate cancer (95), Siglec-15 antagonists like NC318 in metastatic solid tumors, and glycan vaccines for various cancers such as breast, CRC, melanoma, and lung as extensively reported by (155). Upregulated in TAMs, Siglec-15 suppresses T-cell function independently of PD-1 and is under clinical evaluation using NC318 antagonist (161). Therapeutics such as AL009 disrupt Siglec-mediated immunosuppression in mice tumor models (252). Anti-Selectin antibodies, including crizanlizumab (P-Selectin blockade) and uproleselan (E-Selectin inhibition), show promise in AML treatment (4). Monoclonal antibodies targeting CD24/Siglec-10 have shown preclinical and early clinical potential, though higher specificity is needed (253). Anti-Neu5Gc antibodies have also been explored (52).

To prevent metastasis, strategies involve the use of Selectin inhibitors such as GMI-1271 (uproleselan), which blocks tumor-endothelial adhesion and thereby disrupt a critical step in the metastatic cascade. Complementing this pharmacologic approach, cell-based immunotherapies are also being explored (Fig. 4). These include Siglec-directed CAR-T cells and bispecific T-cell engagers, which target specific Siglec markers expressed in various hematologic malignancies—for instance, Siglec-3 in AML, Siglec-6 in B-cell malignancies, and Siglec-2 in certain leukemias (252). Together, these strategies represent a multipronged effort to interfere with both tumor-intrinsic processes, such as proliferation and survival, and immune-mediated mechanisms of tumor progression.

Glycan editing strategies include tumor-targeted antibody-sialidase conjugates and engineered bi-sialidases (Fig. 4) such as E-602, which simultaneously remove multiple Siglec ligands to restore NK, macrophage, and T-cell function (124, 254, 255). However, sialidases have tumor-associated activity and specificity that need to be explored for potential therapeutic applications (256). For example, NEU1 facilitates EMT in pancreatic cancer and its expression negatively correlates with CRC metastasis and NEU3 is particularly associated with gangliosides structures in various cancers (3, 52). Desialylation via trastuzumab-sialidase conjugates in breast cancer models increases immune infiltration and survival in a Siglec-E–dependent manner (257).

Combination therapies may offer maximal benefit. For instance, inhibition of ST3GalI combined with VEGF-A blockade enhance antiangiogenic therapy (98). Similarly, Siglec-10 blockade can boost the efficacy of anti–PD-1/PD-L1 treatment (195), and Siglec-9 targeting has shown synergistic effects in GBM models (192, 193).

Conclusions

Sialylation is recognized as a central orchestrator of tumor–immune system interactions, influencing every stage of cancer progression, from immune evasion and metastasis to therapy resistance. The “writers” of sialylation, namely sialyltransferases, dynamically remodel the tumor glycocalyx, generating structurally diverse and functionally potent sialoglycans. These structures are decoded by “readers,” particularly inhibitory Siglecs and Selectins, which reprogram immune responses, dampen cytotoxicity, and facilitate metastatic dissemination. Together, this bidirectional network constitutes a glyco-immune checkpoint that tumors exploit to secure their survival and expansion. Clinically, tumors that upregulate STs or that present high density of sialylated glycoconjugates exploit multiple Siglecs for immunosuppression. Other glycan-binding molecules such as galectins also play critical immunomodulatory roles (16, 162). Galectin-1, for example, promotes MDSCs activity and angiogenesis (258). Loss of α2-6-sialylation on myeloid cells in melanoma and CRC correlates with poor prognosis and resistance to immune checkpoint inhibitors (258). A better understanding of these interactions is crucial for the development of synergetic approaches that will combine the delivery of selective STs inhibitors, the use of tumor-directed sialidases to degrade sialoglycans, and/or the blocking of Siglec-mediated and/or Selectin-mediated signaling will represent a promising avenue to overcome immune resistance and potentiate current immunotherapies.

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.

Reviewed by members of the JBC Editorial Board. Edited by Robert Haltiwanger