AML cells with wild-type p53 exhibited enhanced apoptosis following Gal-3 inhibition (48). A similar connection between wt p53 and Gal-1 has also been described in gliomas (49). Another work shows that, in p53^-/- pancreatic ductal adenocarcinoma, Gal-1 is upregulated and confers enhanced tumour aggressiveness and shorter survival of the mice (50). Campion and colleagues proved that transfection of breast cancer cell lines with vectors encoding mutant p53 resulted in the upregulation of another galectin with tumour-promoting properties, Gal-7. Doxorubicin treatment of breast cancer cells harboring mutant p53 also produced the same result, via NF-κB activation, but this did not occur when the mutant p53 was knocked down (51). Finally, Gal-7 was found by the same group to accelerate degradation of wt p53 in a carbohydrate-binding-independent manner (52). Cecchinelli and colleagues demonstrated that the homeodomain-interacting protein kinase 2 (HIPK2), a serine-threonine kinase responsible for the activation of p53 via phosphorylation, negatively regulates Gal-3. HIPK2-mediated repression of galectin-3 is an important step following the induction of the p53-dependent apoptosis pathway, while forced expression of Gal-3 prevents it (53). Lavra and colleagues further showed that impaired HIPK2 in thyroid carcinomas results in a strong presence of Gal-3, despite a wild type p53 (54). Taken together, these findings suggest that mutated or dysfunctional p53, or defects in factors upstream in the p53-mediated apoptosis pathway may trigger galectin upregulation with a cell-protecting role, upon stress signals ( Figure 2 ).

Numerous works indicate the upregulation of galectins in response to hypoxia and nutrient deprivation, in a mechanism regulated by HIF-1α and HIF-2 (56–61). Hypoxia is known to be a fundamental force behind tumour aggressiveness and resistance to various therapeutic modalities, including checkpoint inhibition immunotherapy. It has been clearly demonstrated that the degree of tumour hypoxia correlates with immunosuppressive M2 macrophage infiltration, while it is inversely proportional with intratumoral CD4⁺ and CD8⁺ T cells, the main mediators of anti-tumour immunity (62, 63). Wang and colleagues recently drew attention to macrophage-derived Gal-3 in breast cancer models, showing that M2-polarized macrophages are activated and secrete Gal-3 specifically in hypoxic regions, promoting angiogenesis and vascular mimicry in endothelial monolayers (HUVECs). Induction of Gal-3 was found to be caused by ROS production and NF-κB activation and nucleation. As macrophage depletion was already known to reverse the consequences of hypoxia and abrogate angiogenesis, they were able to reproduce the same result by selectively targeting Gal-3 expression (64). In NSCLC, hypoxia-induced Gal-3 increased localization of RhoA to the plasma membrane, thus activating cell motility (65). Le and colleagues were among the first to report that Gal-1 upregulation is a direct result of hypoxia in HNSCC, being inversely proportional to T-cell infiltration (66). Later, it was further demonstrated that Gal-1 upregulation in hypoxic conditions is HIF-1-dependent and that Gal-1 knockdown significantly reduces hypoxia-induced invasion and migration of colorectal carcinoma cell lines (60).

In a very recent publication, Funkhouser and colleagues studied in parallel the expression of 50 known cancer-critical genes in tumor fragments from patients with breast and lung cancer and the profiles of galectins in the serum of these patients. They discovered that gain-of-function mutations in the KIT proto-oncogene are consistently associated with elevated serum levels of galectins -1, -3, -8, and -9 in breast cancer and galectin-1 in non-small cell lung cancer patients. KIT codes for the receptor tyrosine kinase c-KIT, often upregulated in various cancers, leading to enhanced cell survival and proliferation. The authors suggest that c-KIT may have a role in the increase of galectin abundance in circulation. This could be either through its altered glycosylation pattern in the presence of a mutation, causing increased galectin secretion, or as a result of intracellular mechanisms following its activation by the stem-cell factor (SCF), leading to increased transcription of galectins (67).

Galectin-9 expression in tumour cells was found to be induced by extracellular cytokine signaling. More precisely, in growing tumours, DCs and cancer cells secrete IFNβ, while T cells produce IFNγ. Both cytokines were found to upregulate not only Gal-9 expression by cancer cells, but also its secretion. Interestingly, expression but not secretion of Gal-9 required a functional EGFR (68). Jabbari and colleagues, also showed that Gal-9 expression is induced by various types of chemotherapy, which had enhanced cytotoxicity when combined with blockade of Tim-3, the ligand of Gal-9 on the T cell surface (69).

Galectins are present in the nucleus, influencing gene expression at the transcriptional and post-transcriptional level. At the former, Gal-3 was found to upregulate oncogenes Cyclin D1 (70) and MUC2 (71), by enhancing and stabilizing the binding of their respective transcription factors, CRE and AP-1, to their promoter sites. Direct interaction with the transcription factor TTF-1 also endows cancer cells with superior proliferative capacity (72). Moreover, in the nucleus of breast carcinoma cells, Gal-1 is reported to interact with FoxP3, a transcription factor suppressing potential oncogenes MYC, ERBB2 and SKP2, dampening its tumour-suppressive features (73). Another transcription factor, FoxD1, is reported to form a positive regulatory loop with Gal-3, promoting cancer aggressiveness (74). In the context of its perturbed distribution between the nucleus and the cytoplasm, Gal-3 was found to translocate to the nucleus via the importin-α/β complex, mediated by the protein’s C-terminal region (75). Of note, phosphorylation of Gal-3 at Ser⁶ has been shown to be crucial for the exertion of its malignant gene regulatory activity and could be a target of therapeutic experimentation (72). Post-transcriptionally, Gal-3 and Gal-1 have been shown to participate in the pre-mRNA splicing machinery, without binding RNA directly, in a carbohydrate-independent manner. Various works demonstrate their interaction with core polypeptides of the spliceosome, inducing significant alterations in the splicing of several cancer-related genes (76–79). By separately inhibiting Gal-1 and Gal-3, it was proven that they bind a common splicing partner, thus being functionally redundant (78) and double inhibition significantly altered the splicing of several genes (77). Interestingly, the amino-terminal domain of Gal-3 was found to inhibit the incorporation of Gal-1 into the spliceosome (78).

A recent publication focuses on Gal-3 activity in response to stress. Changes in the environment can generate conditions of cell stress, resulting in a multitude of metabolic changes. Forcing stress by thapsigargin administration, this group revealed that Gal-3 localizes in the endoplasmic reticulum (ER)-mitochondria interphase. By downregulating the mRNA half-life of USP14 and SEL1L, Gal-3 prevented the activation of the ER-associated misfolded protein degradation pathway (ERAD) and blocked pre-apoptotic mitochondrial fission (80). These data emphasize on the importance of Gal-3 in preserving ER-mitochondrial integrity and cell survival upon stress. Other works demonstrated a link between Gal-3 and the Bcl2 protein family. Bcl2 is an anti-apoptotic protein, controlled by the master regulator of apoptosis, p53. Harazono and colleagues showed that Gal-3 bears a conserved motif also identified in Bcl2 (NWGR), which allows binding and sequestering the pro-apoptotic protein BAX. This leads to BAX oligomerization, in a similar way to Bcl2 and ultimately to the prevention of apoptosis (55, 81) and resistance to chemotherapy (82) ( Figure 2 ).

Another notable finding is the contribution of Gal-1 and Gal-3 to Ras-mediated oncogenic signaling. In many cancers, mutations in the Ras genes endow cells with enhanced growth and resistance to cell death mechanisms. Both galectins were found to interact with Ras proteins, stabilizing them in the cell membrane, but inducing different oncogenic pathways. Gal-1 increased the ERK downstream pathway and Gal-3 the Raf-1 and PI3-K activity (55, 81, 83). Further works demonstrated that Gal-3 is an integral component of the K-Ras-GTP nanocluster in the plasma membrane, determining the magnitude of its activity (84). Disruption of Gal-3 binding to the cell surface receptor integrin αvβ3 was suggested to be a potential therapeutic target in tumours with uncontrollable K-Ras activity (85). Besides, it was demonstrated in neutrophils that intrinsic Gal-3 expression reduces the generation of ROS and that this can be abrogated by inhibition of Gal-3 with modified citrus pectin (86).

In hypoxic tumour microenvironments, increased glucose uptake is essential for cell survival, as cells switch from oxidative phosphorylation to glycolysis (87) [also elegantly explained in (88)]. Studies in diabetes shed significant light on the implication of intracellular Gal-3 in glucose metabolism. In type 2 diabetes, Gal-3 is upregulated in various tissues and organs. Application of a high-fat diet in Gal-3 KO mice resulted in reduced tissue glucose uptake and greater hyperglycemia, compared to wt mice. This is consistent with a reduction in the expression of the glucose transporter Glut-4 in the absence of Gal-3, with a concomitant reduction in insulin levels (89). In humans, in hypoxic breast and lung carcinomas, Gal-3 has been also shown to be simultaneously upregulated with another glucose transporter, Glut-1 (61). These data suggest that Gal-3 upregulation could potentially sustain a high degree of glucose uptake, even in the absence of insulin signals. Reducing glucose availability by switching from a high-fat to a normal diet downregulates the glucose transporter Glut-4 [discussed in (90)]. However, extracellular Gal-3 has opposite effects on cellular glucose uptake. For example, obesity is characterized by a strong detection of Gal-3 in circulation. Using an obesity-simulating model, Li and colleagues proved that circulating Gal-3 induced systemic insulin resistance. By direct interaction with the insulin receptor (IR), extracellular Gal-3 mediated Glut-4 downregulation and reduced glucose uptake (91). In the context of cancer, this would mean that malignant cells releasing large amounts of Gal-3 could gain access to an increased glucose reservoir by blocking its uptake by adjacent tissues in a paracrine way ( Figure 3 ).

In a remarkable discovery, Gal-3 ablation was shown to downregulate the expression of human telomerase reverse transcriptase (hTERT) in gastric cancer cells. Constitutive expression of the normally silent hTERT is a key factor for the immortalization of cancer cells (92), being proportionally correlated with poor patient outcome. La and colleagues observed a link between the mRNA expression levels of hTERT and Gal-3 in gastric cancer cell lines. Knocking out one of them, they discovered that the other one is downregulated. Ablation of hTERT triggers decreased cell proliferation and increased senescence. Gal-3 knockdown or knockout reduced expression of hTERT and led to subsequent senescence which was rescued by forced hTERT overexpression. Immunoprecipitation experiments proved that hTERT activity is regulated by direct interaction with the N-terminal part of Gal-3 (93).

The effects of intracellular galectins on cell motility and metastatic processes are complex, not unequivocal and sometimes controversial. In general, the cellular abundance of various galectins strongly affects the cell surface expression of biomolecules, especially glycoproteins and glycolipids with repercussions on cell adhesion properties towards the neighboring cells and the extracellular matrix. However, the consequences in terms of long-range cell motility and metastatic processes are variable depending on the context either in vitro, in animal models or in the patients. This complexity is well illustrated by 2 reports dealing with Gal-3 and published the same year, in 2019. Working with the B16 murine melanoma model, Hayashi et al. have either overexpressed Gal-3 in the malignant cells or knocked-out endogenous Gal-3. Both experiments showed that intracellular Gal-3 tends to reduce the metastatic potential of B16 cells, at least in part by negative regulation of the integrin-β3 (94). In contrast, working with the 4T1 murine breast carcinoma model, Pereira et al. showed that Gal-3 tends to increase the metastatic potential of these malignant cells with a concomitant reduction of the surface expression of syndecan-1 and a greater expression of chondroitin sulfate proteoglycans such as versican (95). Apparently with regard to Gal-1, published data are more consistent suggesting that it frequently enhances the metastatic potential of malignant cells. For example, in the 4T1 breast carcinoma model, Gal-1 ablation reduces tumour growth and metastatic spreading. However, the contribution of Gal-1 to the metastatic process in this model seems to be mainly mediated by its immune suppressive effects (96). Using the human gastric carcinoma cell line MGC-803 in athymic mice You et al. showed that Gal-1 promotes the metastatic potential of the malignant cells with a mechanism involving the Sphingosine-1-phosphate receptor 1 (S1PR1). In the same study, the authors reported that the abundance of Gal-1 and S1PR1 detected by IHC on human gastric tumour tissue sections was correlated with a greater frequency of lymph node metastases and a poorer outcome (97). Quite different observations were reported for intracellular Gal-9 in human breast carcinomas. Its high expression in tumour tissue sections is correlated with less metastatic lesions and a better clinical outcome. Simultaneously the authors reported greater cell aggregation and reduced adhesion to the extracellular matrix for MCF7 overexpressing Gal-9 which supposedly was consistent with an anti-metastatic function of intracellular Gal-9 (98).

Extracellular galectins are also suspected to contribute to the metastatic processes especially by their role in the formation of metastatic niches. For example, Gal-3 secreted by malignant cells enhances osteoclast activity in combination with the cytokine RANKL in skeletal metastatic lesions of breast and prostate carcinomas (99). In these pathological contexts, Gal-3 is secreted in a soluble form, either intact (breast carcinoma) or cleaved (prostate carcinoma). Galectins can also be secreted in association with exosomes. For example, this has been reported by our group for Gal-9 and by other groups for Gal-3 (21, 100). To our knowledge the role of tumour exosomes carrying galectins has been reported for immune suppression associated with tumour growth but so far not for the creation of metastatic niches (21, 101).

The impact of galectins on tumour vasculature has been an object of study for more than ten years. Angiogenesis is a well-known pre-requisite for cancer growth (102). Thijssen and colleagues had previously developed an angiostatic peptide, anginex, with demonstrated inhibitory activity on tumour growth, but for a long time were unable to identify its direct target. In 2006, they demonstrated that Gal-1 was the receptor for anginex on the cell surface in various tumour models. Gal-1 exhibited selectively increased patterns of expression in endothelial cells lining tumour vasculature, correlating with staining for proliferation marker Ki67. Binding of anginex to Gal-1 resulted in significant reduction of microvessel density (103). In in vivo multiple myeloma models, shRNA-mediated Gal-1 inhibition downregulated pro-angiogenic genes, including MMP9 and CCL2, and upregulated the anti-angiogenic SEMA3A and CXCL10. This led to decreased microvascular density and lower tumour burden (57). Increased Gal-9 expression in endothelial cells and tumour vasculature compared to normal tissues has also been reported (104). Evidence for Gal-3 has also emerged, as dos Santos and colleagues showed that cancer-cell-secreted Gal-3 is bound by endothelial cells, altering the metabolic balance between DLL4 and JAG1, two factors with opposing roles in angiogenesis. More precisely, Gal-3 was found to increase the half-life of the second over the first, leading to a more pronounced VEGFR2 expression and faster tumour growth in mice bearing LLC tumours. On the contrary, this effect was prevented in LGALS3 ^−/− mice (105).

Induction of immune tolerance is one major aspect of galectin contributions to the development of human malignancies. This is accomplished through 3 main types of biological processes: 1) alterations of the malignant or stromal cell phenotype distorting their interactions with the immune system; 2) Modifications of the extracellular chemical contexture in the tumour microenvironment; 3) Direct phenotype modifications of the cells of the innate and adaptative immune system. The first type is mainly promoted by intracellular galectins contained in malignant cells while the two others are mainly supported by extracellular galectins.

Downregulation of the surface expression of MHC-I molecules is one way to impair recognition of target cells by CD8⁺ T-cells. Mathew and Donaldson have shown that endogenous Gal-3 is involved in the internalization of surface MHC-I molecules. Briefly, HeLa cells were treated for 48h with N-acetylglycosamine resulting in enhanced branching of surface glycans and internalization of MHC-I. This last effect was prevented by siRNA knockdown of endogenous Gal-3. Reciprocally, MHC-I internalization was stimulated by extracellular addition of recombinant human Gal-3 (106, 107). In nasopharyngeal carcinoma, a human malignancy involving latent EBV infection, Gal-9 associated with malignant cells distorts their interactions with the immune system in another way. It interacts with STING inside the malignant cells, accelerating its degradation by enhanced ubiquitination. STING degradation then promotes IL-1β and IL-6 secretion by malignant cells which stimulate differentiation of bystander immature myeloid cells into MDSCs which have a strong immunosuppressive potential (108). Galectins can also affect stromal cell interactions with the immune system. Nambiar and colleagues demonstrated that secreted Gal-1 upregulates PD-L1 and Gal-9 expression on endothelial cells. Confirming that this is the result of chronic STAT activation, they proved that Gal-1-mediated impact on PD-L1 and Gal-9 in the tumour endothelium promotes T-cell exclusion, as Gal-1 blockade significantly boosted T-cell infiltration and conferred enhanced response to immunotherapy (109). However, in the context of patients undergoing surgery for hepatocellular carcinomas (HCC), Sideras et al. reported quite distinct observations. Concomitant pre-operative high levels of Gal-9 and PD-L1 in both tumour tissue and circulation, associated with high CD8⁺ TIL counts, predicted very accurately an improved overall and recurrence-free survival in hepatocellular carcinoma (HCC) (110). This could be explained if we take into consideration that both Gal-9 and PD-L1 are upregulated following IFNγ signaling (111, 112), possibly reflecting strong ongoing immunoreactivity, backed by the strong TIL infiltration.

Outside the field of cancer, it was found that Gal-1 induction in kidneys exposed to high glucose concentration in type 1 and type 2 diabetes accounts for the development of tissue fibrosis by increasing expression of fibronectin (113). This may have repercussions for cancer biology. Indeed, as explained by Cox and Erler, fibrosis is often involved in cancer development. In tumours, disrupted extracellular matrix (ECM) homeostasis leads to enhanced tumour stroma and desmoplasia (114). Massive collagen deposition in the tumour stroma by activated fibroblasts promotes ECM stiffness, favoring tumour immune escape and progression. Extracellular galectins can also reduce the rise of the local immune response by scavenging inflammatory products, for example end products of advanced glycation (AGEs) or advanced lipidation (ALEs) which result from enhanced glucose metabolism. Binding of such products to their receptors generates an inflammatory response. Extracellular Gal-3 has the opposite effect because it binds AGEs and ALEs, preventing interaction with their receptors, thus contributing to their clearance in an immunosuppressive manner (115).

Some galectins also have the power to directly antagonize cytokine accession to their target cells. Described by Gordon-Alonso and colleagues, secreted Gal-3 accumulates in the tumour extracellular matrix and blocks diffusion of IFNγ by binding its glycans (116). This prevents the creation of a CXCL9/CXCL10/CXCL11 chemokine gradient downstream of an IFNγ signal, which is essential for the recruitment of primed T cells (117). Secreted galectins have been shown to accumulate in the extracellular matrix and the tumour stroma, thus potentially shielding it by blocking the liberation of immune-stimulating signals outside the tumour. Another pivotal work by Blouin and colleagues elegantly supplements the effect of Gal-3 but also of Gal-1 on IFNγ signaling (118). More specifically, unhindered binding of exogenous IFNγ to the IFNγ receptor complex (IFNγR), consisting of the subunits 1 and 2, triggers activation of the JAK/STAT pathway for downstream signaling. Excessive glycosylation of IFNγR2 caused conformational changes in this subunit, eventually impairing JAK activation. Mass spectrometry analysis revealed a 3-fold superior binding of Gal-1 and Gal-3 to the hyper-glycosylated subunit, whereas depletion of the two galectins restored normal signaling. Respectively, the addition of recombinant Gal-3 led to a concentration-dependent decrease of STAT1 tyrosine phosphorylation in IFNγR2 WT fibroblasts. These data imply that the amount of extracellular Gal-1 and Gal-3 bound to IFNγR2 is a key regulator of the subsequent activation of the JAK/STAT signaling pathway by IFNγ. Besides, it is well established that N-glycosylation of human IFNγ is of great importance for its successful secretion and dimerization, protecting it from the action of proteases at the extracellular milieu (119, 120). Generation of stable Gal-9 knockouts of the bladder cancer cell line MB49 led to gradual decline of tumour growth over iterative xenograft transplantations. Replacing cell injections with iterative transplantations of small tumour fragments into new receiver mice, Baloche and colleagues were able to study tumour evolution through time, as the tumour and the immune infiltrate were allowed to “age” together. After 4 passages, Gal-9 WT tumour maintained the same growth rate, while KOs stopped growing. Interestingly, early passage KO tumours had an immune-excluded phenotype, while late passage small tumours displayed robust T cell infiltration and readily produced IFNγ, CXCL9, CXCL10 and CCL5 (121). It is not clear from this work if Gal-9 has a direct interaction with IFNγ, as previously described for Gal-3, but its impact on the chemokine gradient hints towards this assumption. Taken together, these data suggest that Gal-1, Gal-3 and may be Gal-9 thwart cytokine-mediated communications, hampering immune persistence within the tumour and potentially weakening systemic immunity.

When acting on cells of the immune system, extracellular galectins tend to use diverse and convergent mechanisms to facilitate tumour immune escape. 1) Towards myeloid and non-myeloid cells of the innate immune system, they promote a suppressive phenotype; 2) Towards CD4⁺ and CD8⁺ lymphocytes they tend to strengthen the activity of suppressive cells and to neutralize effector T-cells either by induction of apoptosis, exhaustion or conversion towards a suppressive phenotype. However, like for most regulators of the immune system, the effects of galectins are rarely unequivocal and there are reports of immunostimulatory effects. For example, Gal-9 has been shown to induce cell death in a specific population of T-regs which express Tim-3 and are often present inside tumours, with a strong suppressive activity (68).

Among myeloid cells, DCs (dendritic cells) have a major role in the initiation of the immune response. There are several reports about immunosuppressive effects resulting from functional alterations of DCs by extracellular galectins. For example, extracellular Gal-1 binds CD69 on the surface of DCs, blocking Th17 differentiation and generating immunotolerant DCs (122). In breast carcinomas, binding of extracellular Gal-9 to Tim-3 at the surface of DCs suppresses their production of CXCL9, resulting in a significantly decreased anti-tumour response, especially in the context of a treatment by paclitaxel (123). In older works, extracellular Gal-1 was also found to bind polymorphonuclear (PMN) cells, inhibiting their chemotaxis and endothelial transmigration with possible suppression of these inhibitory activities by addition of lactose (124).

Extracellular galectins also have strong influence on macrophages. Gal-3 promotes the selective polarization of tumour macrophages towards the pro-tumourigenic M2 type and increases the production of the immunosuppressive cytokines IL-4, IL-10 and IL-13 [reviewed in (40)]. In turn, Jia and colleagues demonstrated that M2-derived Gal-3 leads to enhanced recruitment of M2 macrophages, in a positive feedback loop mode (125). Gal-9 has also been identified to promote M2 polarization, potentially through ligation with the Dectin-1 surface receptor and CD206 (34, 126). The plasticity of tumour-associated macrophages between M1 and M2 polarization is a crucial component in the connection between cancer and inflammation [presented in (127)].

Galectins can impair NK cell functions by various mechanisms in diverse types of contexts. In the murine glioma model GL26, Gal-1 released by malignant cells or present on their surface plays a key role in blocking tumour rejection, mainly by inhibitory effects on NK cells (128). Extracellular Gal-3 has been identified as a ligand of NKp30, a NK cell surface receptor normally triggering NK cell activation and IFNγ release (129, 130). Gal-3 binding to NKp30 results in the inhibition of NK-mediated tumor cell lysis in vitro and in xenografted malignant cells (129). However, somehow inconsistent data have been published, based on experiments made using the B16 murine melanoma model. Syngenic mice KO-ed for Gal-3 (Gal-3^-/-) seem to have a better anti-tumour NK response (131). This might result from indirect changes in NK cell physiology, while the inhibitory effect of NKp30 binding is direct. Finally, extracellular Gal-9 has been reported to impair NK cell functions in human and mice, inhibiting both cytotoxicity and cytokine production (132). This is consistent with several publications dealing with the protective effect of Gal-9 released by trophoblastic cells at the foeto-maternal interface. This effect is due to a large extent to NK cell inhibition (133, 134).

MDSCs (myeloid-derived suppressor cells) promote cancer development by various mechanisms, mainly by a suppressive activity on effector T cells. For example, they produce peroxynitrites inactivating CD8⁺ T lymphocytes by nitrosylation of their TCR. They also release TGF-β and IL-10 which favor Treg development. In addition, they have pro-angiogenic activity and can enhance the survival of tumor stem cells (135). Extracellular Gal-1 was reported to enhance the activity of MDSCs. In myeloma, Gal-1 released by malignant cells binds CD304 at the surface of monocyte-derived MDSCs enhancing their expansion. In myeloma patients, there is a correlation between the abundance of circulating M-MDSCs and the plasma concentration of Gal-1 (136). In the murine Lewis Lung Carcinoma model Gal-3 has been reported to enhance the migration of MDSCs in the tumor microenvironment, triggered by cisplatin treatment. This effect seems to be related to the binding of Gal-3 to CD98 at the surface of MDSCs (137). Well before that, Gal-9 was shown to enhance the expansion of granulocytic MDSCs in mice (138). More recently, a study on primary and secondary resistance to anti-PD1 antibodies in lung carcinoma patients has suggested a role for MDSC expansion linked to the activation of the Gal-9/Tim-3 axis. There is also evidence from in vitro experiments that Gal-9 can stimulate MDSC differentiation and expansion by mechanisms independent of Tim-3 involving Gal-9 entry in target cells and STING degradation (108).

Gal-1 expression in the tumour inversely correlates with CD4⁺ and CD8⁺ T cell infiltration in patient samples (9). It was revealed though, that elevated tumor Gal-1 expression in colorectal carcinoma correlates, in both mouse tumours and patients, with a signature of CD8⁺CD122⁺PD-1⁺ Tregs, contributing to an immunosuppressive phenotype and favoring poor prognosis (139). On the T cell membrane, Gal-1 binds mainly CD45, but also CD7, CD43 and CD4, while Gal-3 binds LAG-3 and, secondarily, CD45, while Gal-9 binds Tim-3, CD44 and PD1, promoting apoptosis or exhaustion (68, 140) (summarized in Figure 4 ). The interaction of cancer cell-derived Gal-9 with Tim-3 on Th1 and Th17 cell surface has been identified to favor pro-tumour Th2 responses, as CD4⁺ T cells of this subtype do not express Tim-3. In B cell leukemia, strong expression of PD-L1 and Gal-9 on leukemic B cells correlated with impaired T cells, co-expressing PD1 and Tim-3. PD-1 blockade only yielded modest benefit. In contrast, forced restoration of Akt/mTORC1 signaling in T cells upregulated the glucose transporter Glut1 and glucose uptake and downregulated PD1 and Tim-3 (141). As a matter of fact, unlike naïve subsets, activated T cells switch to glycolysis and heavily rely on glucose uptake. In contrast, activated Tregs continue to rely on oxidative phosphorylation (142). Interestingly, although Gal-1 and Gal-3 bind their respective T-cell ligands with similar affinities, studies suggest that only the latter induces direct apoptosis (143). Gal-1 is suggested to bind T cells and mainly modify their profiles, as it attenuates IFNγ production and triggers IL-10, without altering significantly T-cell viability (144).

In the highly metastatic breast cancer murine model 4T1, silencing of Gal-1 markedly reduced formation of metastases in the lung, associated with reduced Tregs in tumours, lymph nodes and the spleen. Injection of WT 4T1 cells on the opposite flank of the mouse restored the typical metastatic potential, suggesting a systemic modulation of the immune response by Gal-1 (96). Kared et al. have reported in the context of Hepatitis C a balance between Gal-9 released by Tregs and IL-21 released by Th17 cells, with a capacity of IL-21 to protect effector CD4⁺ and CD8⁺ T cells from cell death and/or exhaustion induced by Gal-9. Predominance of the Gal-9 stimulus would favor the transition to a chronic disease (145). Finally, Yang and colleagues recently discovered that Gal-9 binds both Tim-3 and PD-1, but binding of the second prevents the induction of Gal-9/Tim-3-mediated apoptosis (68). It is noteworthy that anti-PD1 currently used in immunotherapy blocks PD-L1 but not Gal-9 binding to PD1 (68). Similarly, anti-Tim-3 currently used in immunotherapy blocks phosphatidylserine and CEACAM1 but not Gal-9 binding to Tim-3 (146).

Overall, the data presented in this chapter clearly show that Gal-1, Gal-3 and Gal-9 are all major players in cancer-related immune suppression. Many of their effects converge towards the transformation of the tumour microenvironment into a non-accessible, immune-excluded compartment. The so-called “cold” tumours with an immune-excluded phenotype consistently present an abnormal expression of galectins (147). This is in line with the fact that separate depletion of each of these galectins in tumour models overexpressing them let to increased T-cell infiltration of the tumour and enhanced antitumour immunity (109, 121, 148–150).

Capalbo and colleagues showed that a high expression of Gal-3 in biopsies of PD-L1⁺ NSCLC tumours prior to treatment with anti-PD1 successfully distinguished the non-responders from the responders (151). This work is an example of galectins’ potential to provide precious information regarding the immune landscape of a diagnosed tumour and selectively assess whether it qualifies for direct immunotherapy or not.

As lactose has been identified as a pan-galectin inhibitor, lactose-containing human milk glycans (HMGs) have been shown to bind human galectins, except for Gal-2. However, one issue that needs to be tackled is that only 1% of HMGs reach circulation, as shown in infants (152). Set aside natural inhibitors, the development of the first synthetic galectin inhibitors has been constantly gaining interest in the last years. The most advanced forms of inhibitors with high affinity include modified large complex carbohydrates (like plant-derived pectins) and small-molecule inhibitors, developed according to the 3D structure of specific galectins (153) (summarized in Table 2 ). In radiotherapy-resistant prostate cancer PCa cells, the treatment with the Gal-3 inhibitor modified citrus pectin (MCP) led to radiosensitization of the cells, increasing ROS production and apoptosis (155). Recently, modified apple pectin (belapectin or GR-MD-02) exhibited significant synergy with anti-OX40 immunotherapy in mice bearing various tumour models. Combined administration of the two treatments prolonged survival, enhancing CD8⁺ T-cell infiltration and reducing monocytic MDSCs and MHC-II^hi macrophages (156). Given the size of such molecules however, the ability of modified pectins to target intracellular galectins remains questionable. On the other hand, emerging small-molecule inhibitors have exhibited variable ability to inhibit intracellular galectins, gaining increasing interest (157). Oral administration of GB1107, a new Gal-3 antagonist, along with PD-L1 blockade boosted the CD8⁺ T cell infiltration in lung adenocarcinoma, reversed macrophage polarization towards M1 and increased the levels of produced IFNγ (158). Another oral Gal-3 inhibitor, GB1211 was recently shown to substantially increase the binding of pembrolizumab and atezolizumab, in cancer cell lines overexpressing PD-1 and PD-L1 respectively (159), thus complementing the findings of Capalbo and colleagues (151). In Gal-1-overexpressing hepatocellular carcinoma (HCC), the novel Gal-1 inhibitor OTX008 compensated the downregulated negative regulator of Gal-1 miR-22, blocking cell growth and reducing tumour size, in combination with the angiogenesis inhibitor sorafenib (33). In the same work, it was found that Gal-1 overexpression correlated with bad prognosis and enhanced growth and metastasis by upregulating the Golgi transmembrane protein endoplasmic reticulum 1 (RER1). OTX008 was further found to downregulate proliferation, invasion and angiogenesis in a large variety of cancer cell lines in vitro, but also in immunocompromised mice bearing ovarian carcinoma xenografts (160). Mainly, OTX008 was proven to counteract Gal-1-mediated activation of the ERK1/2 and AKT-dependent survival pathways. An inhibitor of Gal-3, GCS-100, proved to have significant synergy with BH3 mimetics (Bcl2 antagonists). Treatment of AML cells with GCS-100 increased the sensitivity to BH3 mimetics, leading to enhanced apoptosis, largely depending on the presence of a WT p53 (48).

Other ways to target galectins under investigation include the use of antibodies, aptamers and nanoparticles. The use of a Gal-1-targeting DNA aptamer (AP-74 M-545) prevented the binding of Gal-1 to CD45 and increased CD4⁺ and CD8⁺ T cell influx, suppressing tumour growth in immunocompetent but not in immunocompromised mice (161). In two distinct works, targeted disruption of Gal-1–N-glycan interactions eliminated hypoxia-driven angiogenesis and suppressed tumourigenesis in vivo. In addition, these works confirmed the impact of hypoxia on abnormal vascularization, as anti-Gal-1 antibodies gave promising results in increasing the efficacy of anti-VEGF treatment (56, 162). In mice bearing Gal-1-overexpressing glioblastomas, intranasal administration of nanoparticles loaded with Gal-1-targeting siRNA remarkably transformed the TME. Gal-1 blockade through nose-to-brain transport led to reversed macrophage polarization and increased CD4⁺ and CD8⁺ T-cell infiltration, while exhibiting synergy with dendritic cell vaccination and PD-1 blocking (163). Interestingly, Tiraboschi and colleagues found that low dose docetaxel treatment inhibits Gal-3 expression in prostate cancer cells, leading to a reconditioned TME and a superior response to tumour vaccination in a post-surgery mouse model (164). Targeting of Gal-9 is also under investigation, as newly developed Gal-9-neutralizing antibodies have exhibited an ability to prevent Gal-9-mediated T–cell death in mouse tumour models (165). One such antibody, Lyt-200, is currently under clinical investigation in phase I/II trials for its safety and efficacy after encouraging data from treated primates.