Selective Nanobody-Derived Minibodies Targeting Galectin‑1 and -7 Reveal Non-Redundant Glyco-Immune Functions and Therapeutic Potential in Triple-Negative Breast Cancer

Abstract

Galectins (GALs) act as glyco-immune checkpoints that modulate tumor immunity, but their overlapping expression complicates functional dissection and targeted inhibition. Here, we explore the distinct molecular and immunological roles of GAL-1 and GAL-7, two GAL sover expressed in triple-negative breast cancer (TNBC), and describe the development of highly selective nanobody-derived minibodies (G1M1 and G7M8) that specifically target each protein. In TNBC cells, GAL-1 and GAL-7 induced different cytokine profiles: GAL-1 increased pro-tumor mediators such as G-CSF and VEGF-A, while GAL-7 promoted immunomodulatory cytokines, highlighting their nonredundant functions. In vivo, both G1M1 and G7M8 reduced pro-tumor cytokines and boosted antitumor immunity, either alone or combined with anti-PD-1 therapy. Notably, G1M1 alone achieved results comparable to anti-PD-1 therapy in limiting lung metastases, increasing CD4⁺ T-cell infiltration, and decreasing PD-1⁺ Tregs. These findings demonstrate that selective GAL blockade with engineered minibodies is a promising strategy to expand immunotherapy options in TNBC.

Introduction

Immunotherapy has marked a paradigm shift in the field of oncology by offering new treatment options that can lead to durable responses and improved survival rates across various types of cancer. It has particularly opened new avenues for treating aggressive cancers with limited therapeutic options, such as triple-negative breast cancer (TNBC). As more specific molecular drivers are identified, novel therapeutic strategies have emerged, particularly in the field of immunotherapy. The success of immunotherapy currently relies on inhibiting immune checkpoints, which target the immunosuppressive mechanisms driven by cancer cells, thereby allowing the patient’s immune system to mount an effective antitumor response. Today, immunotherapy using immune checkpoint inhibitors, such as anti-PD-1, is part of the first-line treatment for patients with aggressive cancers, including TNBC. − Unfortunately, the success rate of immunotherapy varies significantly among patients. Not all patients respond to immunotherapy, and predicting which patients will respond remains a major challenge. , Moreover, many patients who initially respond to immunotherapy may later develop resistance to it.

Galectins (GALs) are a family of evolutionarily conserved animal lectins that are broadly distributed across species, ranging from invertebrates to vertebrates. They were first identified in the electric eel (Electrophorus electricus) as low-molecular-weight, β-galactoside-binding soluble proteins. Since their initial discovery, 15 GALs have been sequentially numbered based on their structural characteristics and the number of carbohydrate recognition domains (CRDs). While GALs have various homeostatic roles within cells, they are predominantly recognized for their functions outside the cell. In pathological conditions, GALs are secreted into the extracellular environment through nonclassical pathways. Once outside the cell, they bind multivalently to repeating units of high-density N- and O-glycans on membrane receptors via their glycan-binding sites, located within their CRDs. This interaction is believed promote the formation of organized “lattice” structures that modulate receptor diffusion, compartmentalization, and endocytosis. Although it is not clear how this model translates in vivo within the complexity of living tissues, it provides a conceptual framework for understanding how GALs organize glycosylated receptors lacking classical signaling domains.

GALs play critical roles in cancer progression, including tumor growth, angiogenesis, immune evasion, metastasis, and therapeutic resistance. , GALs are often expressed at abnormally high levels in both cancer cells and stromal cells within the tumor microenvironment (TME). One of their best-characterized roles is their ability to modulate the immune response. , GALs contribute to the establishment of an immunosuppressive TME through various mechanisms, including the killing of activated T cells, the promotion of Tregs proliferation, and the alteration of cytokine profiles by promoting the secretion of anti-inflammatory and immunosuppressive cytokines by immune and cancer cells. This immunomodulatory function has attracted considerable attention due to its potential to improve the effectiveness of cancer immunotherapy. − This is particularly relevant for GAL-1, which has been implicated in mechanisms of resistance to immune checkpoint blockade and broader immune suppression in the tumor microenvironment. − Accordingly, an increasing number of studies, including our study, have demonstrated the effectiveness of galectin-specific antibodies and minibodies that block extracellular functions of galectins. For example, anti-GAL-1 and anti-GAL-9 antibodies have shown antitumor efficacy in preclinical models, most notably by preventing galectin-mediated immune evasion. ,, A recent report using camelid antibodies against GAL-2 has also shown the potential of extracellular galectins to modulate macrophage polarity. These findings have led to the emerging concept of a “glyco-immune checkpoint”, which is under active investigation in the field of onco-immunology as it opens new avenues for research and therapy that could potentially transform the management of cancer and other diseases. , In aggressive breast cancer subtypes, both GAL-1 and GAL-7 have been associated with chemoresistance, poorer survival, and metastasis. − However, the mechanisms by which GAL-1 and GAL-7 contribute to breast cancer progression differ in how well they are understood. The role of GAL-1 has been studied more extensively: extracellular GAL-1 can interact with β1-integrin on breast cancer cells to activate FAK/c-Src/ERK/STAT3/survivin signaling, thereby promoting proliferation, migration, invasion, and chemoresistance. In parallel, extracellular GAL-1 can foster immune evasion by expanding and activating regulatory T cells, consistent with elevated serum GAL-1 levels and stromal GAL-1 expression correlating with poor prognosis. Another widely accepted model is that extracellular galectins may exert their protumoral activity by suppressing antitumor immunity. This model is consistent with elevated circulating and stromal GAL-1 levels correlating with poor patient outcomes. By contrast, the mechanistic basis for GAL-7′s association with aggressive disease remains less clear; it may involve both intracellular functions and context-dependent extracellular effects that are still under investigation. Thus, although galectins also exert important intracellular functions, continued investigation and selective inhibition of their extracellular activities remain particularly critical, as these interactions at the cell surface play central roles in shaping the tumor microenvironment and driving galectin-dependent cancer progression.

The use of glyco-immune checkpoint inhibitors targeting galectins, however, faces two important obstacles. First, our understanding of the functional differences between galectins in a tumor context remains limited. Most cancers simultaneously express two or more galectins, whose roles are difficult to unravel. Although galectins share relatively low sequence identity, their conserved structural scaffold and glycan-binding motifs enable them to recognize overlapping classes of β-galactoside-containing receptors. Consequently, these galectins are likely to exhibit both shared and distinct ligand-specificities and functions, consistent with their partially convergent yet nonredundant roles in tumor and immune regulation. Second, the development of galectin-specific inhibitors remains a major technical challenge, complicating the advancement of targeted and effective therapeutic strategies. Given their critical role in cancer, considerable efforts have been devoted to developing GAL inhibitors. However, despite nearly two decades of research, the development of effective GAL antagonists has met with limited success. The greatest challenge in using such drugs is to achieve high selectivity, a formidable task considering the striking structural similarity between the glycan-binding sites of GALs. To overcome these challenges, our team has recently developed a new series of GAL inhibitors based on camelid single-chain antibodies, known as nanobodies (Nbs). , Nbs are attractive for cancer treatment because of their small size, high stability, and strong binding affinity, which enable them to penetrate solid tumors efficiently. They also have access to epitopes that are often hidden from conventional antibodies and can be easily engineered for therapeutic versatility. Moreover, recent safety data indicate that camelid-derived Nbs have low immunogenicity in humans, particularly when humanized. Most clinical trials, including those with caplacizumab, the first camelid-derived single-domain antibody approved for clinical use in the United States and Europe for the treatment of acquired thrombotic thrombocytopenic purpura, show minimal antidrug antibodies. We have demonstrated that Nbs bind GALs with high affinity and specificity. Most importantly, they effectively block the biological activities of GALs, notably GAL-induced T cell apoptosis. ,

In the present work, we explore strategies to overcome two major challenges that have limited the development of glyco-immune checkpoint inhibitors: (1) the lack of knowledge about how different galectins contribute to tumor progression, particularly since most cancers express multiple galectins with overlapping yet distinct functions, and (2) the difficulty of developing inhibitors with high specificity for individual galectins. To address these issues, we used human and mouse preclinical TNBC models, focusing on GAL-1 and GAL-7. Our in vitro comparative study using MDA-MB-231 cells revealed that stimulation with GAL-1 and GAL-7 elicits, at least in part, unique cytokine production programs, including several cytokines known to modulate the immune landscape. Consistent with these findings, using an in vivo syngeneic immunocompetent mouse model of TNBC, we demonstrate that two nanobody-derived minibodies, G1M1 and G7M8, which selectively target GAL-1 and GAL-7, respectively, exhibit differences in the breadth and targets of cytokine suppression, suggesting that they may act through distinct yet complementary pathways. Both minibodies reduced a wide array of pro-tumorigenic cytokines when combined with anti-PD1, surpassing the effects of either agent alone. Importantly, G1M1 treatment was as effective as anti-PD1 in reducing lung metastases, and both minibodies further enhanced the immune remodeling of primary tumors, often augmenting responses to anti-PD1. Together, these findings reduce key barriers limiting the therapeutic use of glyco-immune checkpoint inhibitors by demonstrating how two distinct galectins can contribute to cancer progression through complementary mechanisms, and by showing that specific inhibitors of GAL-1 and GAL-7 not only reshape the tumor immune microenvironment but also hold strong therapeutic promise when used alone or in combination with existing immune-checkpoint inhibitors.

Results

GAL-1 and GAL-7 Trigger Differential Cytokine Secretion Programs in TNBC Cells

Because the cellular sources and modes of galectin signaling in breast tumors remain uncertain, we first established an in vitro stimulation model using TNBC cells to characterize the cytokine programs elicited by recombinant GAL-1 and GAL-7 and to provide a functional baseline for assessing the activity of our galectin-neutralizing minibodies. For this, we used MDA-MB-231 cells, which constitutively express GAL-1 but not GAL-7. Cytokine profiling of MDA-MB-231 cells showed that stimulation with recombinant GAL-1, GAL-7, or both generated distinct secretion patterns compared to untreated controls (PERMANOVA, p = 0.001; Figure A). Comparative analysis revealed that GAL-7 mainly increased cytokines such as CX3CL1, IL9, IL18, and CCL11, while GAL-1 increased cytokines such as IL1α, CCL7/MCP3, G-CSF, CXCL1, and VEGFA (Figure B). Quantitative profiling confirmed that each GALs promotes a unique cytokine signature, with combined stimulation producing additive or intermediate effects across multiple factors (Figure C). These findings indicate that GAL-1 and GAL-7 activate distinct but complementary cytokine production programs in TNBC cells, emphasizing their cooperative role in shaping the tumor cytokine environment. These experiments therefore represent a controlled in vitro model to dissect galectin-dependent cytokine programs and to evaluate the activity of galectin-targeting minibodies.

1.

Model system to compare cytokine responses of TNBC cells to exogenous GAL-1 and GAL-7. (A) Principal coordinate analysis (PCoA) of cytokine profiles from MDA-MB-231 cells treated for 16h with exogenous with GAL-1, GAL-7, or GAL-1 + GAL-7 compared to controls. Ellipses represent 95% confidence intervals (PERMANOVA, p = 0.001). (B) Scatter plot comparing cytokine induction by GAL-7 versus GAL-1. Each dot represents a cytokine; the solid line indicates equal induction. Cytokines above the line are preferentially induced by GAL-7 (e.g., CX3CL1, IL-9, IL-18, CCL11), whereas those below are more strongly induced by GAL-1 (e.g., IL-1α, CCL7, G-CSF, VEGF-A). (C) Relative cytokine secretion from MDA-MB-231 cells stimulated with GAL-1, GAL-7, or GAL-1 + GAL-7 compared to untreated controls (dashed lines). Bars show mean ± SEM *p < 0.05; **p < 0.01; ***p < 0.001 versus control (global test with multiple-comparison correction).

Development and Characterization of GAL-Specific Minibodies

For our in vivo experiments, we used minibodies generated by fusing Nbs to an IgG1 Fc domain. This design prolongs serum half-life while maintaining a smaller size than traditional antibodies and ensures compatibility with clinically used IgG1-based therapeutics. This was accomplished by fusing the amino acid sequences of previously developed and characterized GAL-1 and GAL-7 Nbs to a human IgG1 Fc fragment (∼40 kDa) and purifying the resulting fusion proteins from the supernatants of genetically engineered HEK-293 cells (Figure A,B). , To confirm their specificity, we tested the binding of both minibodies to human and mouse GALs using ELISA. G1M1 showed strong binding to human and mouse GAL-1 and very low binding to GAL-3 or mouse and human GAL-7 (Figure C). Similarly, G7M8 showed strong affinity for human and mouse GAL-7, minimal binding to GAL-3, and no detectable binding to mouse or human GAL-1 (Figure D). Both minibodies effectively prevented GAL-induced T-cell apoptosis (Figure E,F).

2.

Characterization of GAL-7 and GAL-1 minibodies. (A) Schematic representation of minibodies generated from previously characterized nanobodies. (B) SDS-PAGE analysis of G1M1 and G7M8. (C) Binding specificity of G1M1 and (D) G7M8 to mouse and human GAL-1 or GAL-7, respectively, compared to other GALs. (E) Inhibition of T cell apoptosis by G1M1 and (F) G7M8. (E, F) Controls included cells treated with GAL-1 or GAL-7 alone, and with lactose (Lac) as a positive control for apoptosis inhibition. p-values were calculated by one-way ANOVA followed by Dunnett’s multiple comparisons test. (C–F) Results are representative of three independent experiments and data are presented as mean ± SD.

Preclinical Assessment of G1M1 and G7M8 in TNBC

To study the therapeutic potential of G1M1 and G7M8, E0771 cells were orthotopically injected into the mammary fat pad of C57BL/6 mice. This model is commonly used to determine the therapeutic potential of new drugs, including the anti-PD-1 regimens. , Once the tumors were palpable, both minibodies were tested alone or in combination with anti-PD-1 (n = 13–15 mice/group), which is known to slow tumor progression of E0771 cells (Figure A). Our results show that anti-PD-1 treatment significantly delayed primary tumor growth in terms of volume and weight (Figure B–D). Treatment with G1M1 or G7M8 alone did not significantly reduce primary tumor growth. Interestingly, a significant increase in mouse weight was observed in groups treated with a combination of anti-PD-1 and G1M1 or G7M8 compared to anti-PD-1 alone (Figures E and S5). We then determined whether minibodies, alone or in combination with anti-PD-1, could reduce the burden of lung metastasis. Longitudinal sections of lung tissue showed multiple metastatic lesions of varying sizes (Figure F). As expected, treatment with anti-PD-1 alone significantly reduced the metastatic burden (Figure G,H), as indicated by fewer lung nodules and a lower incidence of lung metastasis in mice. G1M1 alone reduced the number of lung metastases and significantly decreased the percentage of mice with lung metastases compared to the control group, an effect comparable to anti-PD-1 in limiting metastatic spread. G7M8 alone reduced the incidence of metastasis from 81.82% in the control group to 42.86%, suggesting a potential biological effect, although the p-value (p = 0.10) did not reach the conventional level of statistical significance.

3.

Therapeutic effects of G1M1 and G7M8, alone or combined with anti-PD-1, in a TNBC mouse model. (A) Schematic of the experimental design. (B) Tumor growth kinetics. (C) End point tumor volume, (D) tumor weights, and (E) mice body weight. (B–E) Data are presented as mean ± SD. Group sizes: IgG2a + IgG1, n = 14; anti-PD-1 + IgG1, n = 13; G1M1 + IgG2a, n = 13; G1M1 + anti-PD-1, n = 13; G7M8 + IgG2a, n = 15; G7M8 + anti-PD-1, n = 13. p-values were calculated using one-way ANOVA followed by Dunnett’s multiple comparisons test. (F) Representative lung histology section (H&E staining). Metastatic nodules are outlined in black. Scale bar corresponds to 2.5 mm. (G) Number of lung metastatic nodules per histological section. p-values were calculated using the Kruskal–Wallis test followed by Dunn’s post hoc test. (H) Percentage of mice with lung nodules per group. Pairwise comparisons of metastasis frequency were performed using Fisher’s exact test (two-sided) on 2 × 2 contingency tables. Bars represent the percentage of mice with lung metastases. (G, H) Group sizes: IgG2a + IgG1, n = 11; anti-PD-1 + IgG1, n = 12; G1M1 + IgG2a, n = 13; G1M1 + anti-PD-1, n = 12; G7M8 + IgG2a, n = 14; G7M8 + anti-PD-1, n = 13. (C–E, G, H) Bars represent the mean ± SD.

Modulation of Tumor-Infiltrating Lymphoid T Cells in Response to Treatment with G1M1 and G7M8

To assess whether G1M1 and G7M8 could alter the immune landscape of the primary tumor, we analyzed lymphoid infiltration in the primary tumors of treated mice. We first examined CD45⁺ leukocyte infiltration and found a significant increase in the number of CD45⁺ cells in the tumors of mice treated with anti-PD-1 alone compared to the control group (p = 0.01) (Figures A and S1). An even greater increase was observed when anti-PD-1 was combined with G7M8 (p = 0.001). The effect of G7M8 may be attributed to its ability to enhance CD4⁺ T cell infiltration, a trend that approached statistical significance in comparison to anti-PD-1 alone (p = 0.06) (Figures B and S2). In the case of G1M1, it led to a significant increase in CD4⁺ T cell infiltration compared to control treatment and anti-PD-1 alone. We did not observe any statistically significant changes in the percentage of CD8⁺ T cells or CD4⁺ annexin V⁺ T cells in any of the mouse groups (Figure C,D). As for the Treg compartment, a key factor in GAL-1-mediated immunosuppression, we observed a statistically significant reduction in CD4⁺FoxP3⁺PD-1^± following treatment with G1M1 or G7M8 (Figures E,F and S3). G1M1 alone was particularly effective in reducing the percentage of CD4⁺FoxP3⁺PD-1⁺ Tregs, a population associated with worse overall survival in nivolumab-treated patients with metastatic clear cell renal cell carcinoma (Figure F). G1M1 and G7M8 had a marked effect on CD8⁺ Annexin V⁺ cells, with statistically significant differences observed between control treatment groups (p = 0.01–0.02) (Figures G and S4). G1M1 alone was particularly effective in increasing the proportion of CD8⁺ annexinV⁺PD-1^– cells (Figure H). Moreover, G7M8 in combination with anti-PD-1 and G1M1 alone significantly increased the percentage of CD8⁺AnnexinV⁺PD-1⁺ cells compared to the control group (Figure I). Collectively, these results demonstrate that G1M1 and G7M8 distinctively modulate tumor-infiltrating lymphocyte populations, both as monotherapy and in combination with anti-PD-1. Their immunomodulatory effects contribute to reshaping the tumor immune microenvironment, particularly through the regulation of distinct CD4⁺ and CD8⁺ T-cell subpopulations.

4.

Profiling of tumor-infiltrating lymphocytes. C57BL/6 mice were orthotopically injected with E0771 cells into the fourth mammary fat pad to induce primary tumors, which were collected at end point for immune profiling (n = 5). (A) CD45⁺ cells, (B) CD4⁺ T cells, (C) CD8⁺ T cells, (D) CD4⁺AnnexinV⁺ T cells, (E) CD4⁺FoxP3⁺PD-1^– T cells, (F) CD4⁺FoxP3⁺PD-1⁺ T cells, (G) CD8⁺AnnexinV⁺ T cells, (H) CD8⁺AnnexinV⁺PD-1^– T cells, and (I) CD8⁺AnnexinV⁺PD-1⁺ T cells. p-values were calculated using one-way ANOVA followed by Dunnett’s, Tukey’s or Šídák’s post hoc multiple comparisons test (A, B, D, F–I), or the Kruskal–Wallis test followed by Dunn’s post hoc test (C, E), depending on data normality as assessed by the Shapiro–Wilk test. (A–I) Bars represent the mean ± SD.

Systemic Cytokine Modulation by Galectin-Targeting Minibodies

We next investigated how G1M1 and G7M8 influence systemic cytokine profiles when administered alone or together with anti-PD1. Treatment with anti-PD1 alone modestly reduced the levels of several cytokines, including G-CSF, CXCL10, and MCP-1 (CCL2) (Figure ). In contrast, both minibodies exerted broader modulatory effects, particularly in combination with anti-PD1. G1M1 monotherapy reduced chemokines such as 6Ckine/Exodus 2 (CCL21), whereas its combination with anti-PD1 led to consistent suppression of G-CSF, IP-10 (CXCL10), and MCP-1 (CCL2), cytokines linked to the recruitment and expansion of immunosuppressive myeloid cells. G7M8 monotherapy decreased TNF-α, but its effect was markedly amplified in combination with anti-PD-1, resulting in lower levels of G-CSF, IP-10 (CXCL10), MCP-1 (CCL2), KC (CXCL1), and fractalkine (CX3CL1), many of which are involved in sustaining inflammatory circuits and metastatic signaling in TNBC. Multivariate analysis further confirmed that combination treatments were distinct from both controls and monotherapies. Both G1M1 + anti-PD-1 and G7M8 + anti-PD-1 clustered separately from the single-agent groups, reflecting the emergence of unique cytokine signatures. Notably, both combinations increased IFN-β1, a cytokine associated with enhanced antigen presentation and cytotoxic T-cell recruitment in TNBC models (Figure A). − In contrast, the treatment of mice with G1M1 and G7M8, alone or in combination with anti-PD1, resulted in the systemic reduction of a large number of protumorigenic cytokines and growth factors (Figure B). These observations suggest a mechanistic synergy in which galectin blockade alleviates myeloid-driven immunosuppression and dampens protumorigenic cytokines, thereby broadening the immunostimulatory window of PD-1 blockade. Thus, galectin inhibition complements checkpoint therapy by modulating nonredundant immunoregulatory pathways that collectively reshape the cytokine milieu toward a more antitumor state (Figure C).

5.

Cytokine, chemokine, and growth factor levels in mouse serum. Concentrations of cytokine, chemokine, and growth factors were measured in serum samples collected from mice (n = 5–10) at day 27/28. Statistical analyses were performed using one-way ANOVA followed by Dunnett’s post hoc test, or Kruskal–Wallis test followed by Dunn’s post hoc test, based on data normality as assessed by the Shapiro–Wilk test. Lines indicate group mean.

6.

Systemic cytokine modulation by galectin-specific minibodies in TNBC-bearing mice. PCoA of systemic cytokine profiles showing clustering of treatment groups that were statistically increased (A) or decreased (B) in mice treated with G1M1 or G8M7 alone or combined with anti-PD1. (C) Heatmap of scaled cytokine abundance across groups.

Discussion

In the present work, we have addressed the emerging concept of galectins as glyco-immune checkpoints by focusing on GAL-1 and GAL-7, two lectins linked to immune evasion, cytokine modulation, and poor outcomes in TNBC. Our study reveals a dual novelty: first, in vitro profiling showed that GAL-1 and GAL-7 elicit distinct cytokine secretion programs, underscoring that they act through distinct and complementary pathways. Second, using newly developed nanobody-derived minibodies (G1M1 and G7M8) targeting GAL-1 and GAL-7, respectively, we report that selective targeting of each galectin not only blocks its activity with high specificity but also reshapes tumor biology in vivo. More specifically, we found that (1) G1M1 alone was as effective as anti-PD-1 in reducing the number of metastatic lesions in the lung and the percentage of mice with metastasis; (2) Analysis of tumor-infiltrating lymphocytes showed that G1M1 significantly increased CD4⁺ T cell infiltration while reducing the number of Tregs; (3) G1M1 was also effective alone or in combination with anti-PD-1 to reduce the production of tumor-promoting cytokines. In the case of G7M8, to our knowledge, it is the first antibody to inhibit the biological activity of GAL-7. Although it did not significantly reduce the number of metastases under the conditions tested in our study, it is worth highlighting its ability to positively modulate the antitumor response, both locally and systemically. These observations suggest significant potential for the future development of these antibodies as therapeutic tools and for optimizing their efficacy across various clinical settings. Collectively, these results establish GAL-1 and GAL-7 as nonoverlapping immunoregulatory drivers in TNBC, highlighting their blockade as a promising strategy to enhance checkpoint immunotherapy.

Our in vitro data demonstrate that GAL-1 and GAL-7 drive divergent cytokine programs, highlighting their nonredundant roles in TNBC. While GAL-1 primarily promoted pro-tumorigenic factors such as G-CSF, CXCL1, and VEGF-A, GAL-7 induced mostly immunomodulatory cytokines. This divergence suggests that GAL-1 and GAL-7 act through complementary rather than overlapping pathways to shape the tumor cytokine milieu. Previous studies have extensively documented the role of GAL-1 in immune evasion and resistance to checkpoint blockade; however, our findings identify GAL-7 as an equally important, yet mechanistically distinct, immune modulator. Here, we used MDA-MB-231 cells. As previously reported, MDA-MB-231 cells constitutively express GAL-1, but not GAL-7; thus, the observed effects reflect the cellular response to exogenous GAL-7 rather than autocrine stimulation. In the case of GAL-1, it mimics the tumor microenvironment, where high local concentrations of GAL-1 are produced by tumor and stromal cells, amplifying immunomodulatory and pro-tumor signaling. This is a real scenario because extracellular GAL-1 is produced by a wide range of cell types, including immune cells (such as T cells, B cells, dendritic cells, and plasma cells), endothelial cells, and epithelial cells. Thus, while endogenous expression establishes a baseline regulatory tone, exogenous GAL-1 captures the enhanced extracellular effects that are physiologically relevant in cancer progression. Together, these observations provide evidence that targeting both galectins may be necessary to fully disrupt the immunosuppressive networks that drive TNBC progression.

The hypothesis that GAL-1 and GAL-7 play divergent roles in cancer was further supported by our flow cytometric analysis of infiltrating lymphocyte populations, which revealed that G1M1 and G7M8 remodel the immune microenvironment through distinct yet complementary mechanisms. G1M1 was particularly effective at enhancing CD4⁺ T-cell infiltration while reducing regulatory T cells, including PD-1⁺ Tregs previously associated with poor survival under immunotherapy. In contrast, G7M8 increased overall leukocyte entry into tumors and, when combined with anti-PD1, further expanded the CD4⁺ compartment, suggesting a cooperative effect with checkpoint blockade. These nonredundant patterns of immune remodeling indicate that the selective inhibition of GAL-1 and GAL-7 differentially shapes the tumor microenvironment, providing complementary avenues to enhance antitumor immunity in TNBC. Despite these mechanistic differences, both minibodies converge functionally on the activation and engagement of cytotoxic CD8⁺ T cells, a process accompanied by increased apoptosis consistent with activation-induced cell death (AICD) or intense tumor-killing activity. The enrichment of CD8⁺Annexin V⁺ cells thus likely reflects a robust effector response rather than impaired T-cell viability. This coordinated activation indicates that both minibodies foster productive antitumor immunity by stimulating and maintaining dynamic turnover of the cytotoxic pool. In contrast, PD-1 blockade may prevent activation-induced apoptosis, thereby preserving effector T cell numbers but potentially altering the physiological equilibrium between the activation, cytotoxicity, and contraction phases within the tumor microenvironment. Further preclinical studies are warranted to dissect these distinct pathways and confirm the complementary mechanisms by which GAL-1 and GAL-7 inhibition enhance antitumor immunity in TNBC.

One of the critical questions addressed in this study was whether our minibodies exhibit biological activity in vivo and, if so, whether they hold potential as therapeutic agents for future clinical applications. This is a fundamental question with implications beyond the GALs studied in this work. Our in vitro results using the apoptosis assay were encouraging. Of note, the pro-apoptotic activity of GAL-1 toward T cells is well established, whereas a similar function of GAL-7 is less documented. Previous in vitro studies have shown that GAL-7 can induce apoptosis GAL-7 of T cells and PARP-1 cleavage, − consistent with the results of Wu and colleagues who reported, using an in vivo mouse model and samples from patients with esophageal cancer, that GAL-7 expression is negatively correlated with CD4⁺ T-cell infiltration (consistent with the increased infiltration of CD4⁺ cells in mice treated with G7M8). It is important to note, however, that in our study, the apoptosis assay was designed as a proof-of-principle functional test to evaluate the galectin-neutralizing activity of our minibodies under controlled in vitro conditions. Our in vivo results further confirmed that G1M1 and G7M8 are promising candidates for cancer treatment, particularly in immunotherapy. For example, we found that G1M1 alone demonstrated comparable efficacy to anti-PD-1 in reducing the number of lung metastases and decreasing the proportion of mice with metastases by more than 50%. This is a very promising result, especially considering that most patients are refractory to anti-PD-1 and that GAL-1 is overexpressed in approximately 40% of TNBC patients. Moreover, both minibodies positively modulated the immune landscape of primary tumors and peripheral immune profiles, and in many cases, enhanced the efficacy of anti-PD-1 therapy. A notable example is the combination of G1M1 with anti-PD-1, which increased IFN-β1 production by more than 35-fold, a cytokine known to suppress cancer stem cell properties in TNBC. It also significantly reduced the secretion of MIP-3α (CCL20) induced by tumor-derived GAL-1, which has recently been proposed as one of the pathways that limits the efficacy of anti-PD-1. Another interesting finding was our results with fractalkine (CX3CL1). Our in vitro results showed that GAL-7 induced CX3CL1 secretion by MDA-MB-231 cells. Conversely, in vivo, G7M8 suppressed its systemic release. CX3CL1 and its receptor (CX3CR1) are implicated in TNBC by promoting cancer cell metastasis, invasion, proliferation, and growth. High CX3CL1 levels are associated with increased risk of brain metastases, and targeting the CX3CL1/CX3CR1 axis can impair tumor seeding and growth, suggesting a potential therapeutic target for TNBC.

It is worth noting that we observed G1M1 inhibiting the formation of lung metastases without significantly affecting the growth of the primary tumor. While this may seem surprising at first glance, it is not uncommon in cancer biology, including breast cancer. Several preclinical studies have reported that targeted agents can selectively suppress metastatic dissemination without affecting primary tumor size, likely reflecting distinct molecular mechanisms driving metastasis. − The observation that a treatment reduces metastases without significantly affecting primary tumor growth can be attributed to fundamental biological differences between primary and metastatic tumor cells, as well as to distinct immunological and microenvironmental contexts. This reflects a dissociation between the mechanisms governing local tumor expansion and those driving metastatic dissemination and colonization, such that certain therapies may selectively interfere with the metastatic cascade without directly affecting the primary tumor (references). It is conceivable that galectins act during the late stages of metastasis, consistent with the “seed and soil” hypothesis, most notably by their role on the formation of the metastatic niches. This has been shown for several galectins, including GAL-1. After extravasation, disseminated tumor cells encounter a hostile microenvironment characterized by immune surveillance and metabolic stress. Galectins may thus facilitate their survival and adaptation by suppressing immune cytotoxicity, modulating adhesion and matrix interactions, and promoting stress resistance. These functions would enable metastatic “seeds” to persist and colonize the distant “soil,” even when effects on primary tumor growth are limited.

The development of specific GAL minibodies also provides valuable tools for better understanding the roles of GALs in cancer and other diseases. Both antibodies are camelid antibodies, which are well-suited for high specificity due to their unique mode of antigen recognition, enabling them to target epitopes that are often inaccessible to conventional antibodies. , A case in point is the development of a highly specific Nbs targeting matrix metalloproteinases (MMPs), a family of structurally homologous proteases for which achieving selective inhibition has historically been very challenging. Nbs can be efficiently generated using synthetic libraries, allowing them to be humanized during the initial screening and grafted onto IgG frameworks, as demonstrated in this study. , Because of their high specificity, such antibodies are powerful research tools for understanding the roles of individual GAL family members, not only in cancer but also in other diseases. To date, the lack of such specific tools has limited our ability to validate, in vivo, the roles of GALs hypothesized from in vitro studies. Another interesting aspect of Nbs is their effectiveness as imaging tools to identify potential responders to immunotherapy. We have recently shown that GAL-1 and GAL-7-specific Nbs, when conjugated with the NOTA chelator and labeled with copper-64 ([⁶⁴Cu]­Cu), can be used as a radiotracer for PET imaging in a TNBC mouse model. ,

Galectins exist as both intracellular and extracellular proteins, a dual localization that underlies their classification as alarmins, molecules capable of acting inside the cell while also being released to exert extracellular functions. − Among these, intracellular galectins have garnered significant attention, as their functions can be readily explored through genetic engineering of cells. In contrast, studying extracellular galectins has proven more challenging, largely due to the lack of galectin-specific antibodies. To date, only the extracellular role of GAL-1 is well established, whereas that of other galectins, including GAL-7, remains less well-defined. Because GAL-1 is expressed by many host cell types, our minibodies likely neutralize galectins derived from both tumor and nontumor sources, as they are designed to target extracellular galectins regardless of origin. Like GAL-1 and GAL-3, GAL-7 is abundantly present in the cytosol and nucleus, where it fulfills distinct cellular functions. Mounting evidence indicates that GAL-7 can also be released into the extracellular environment under specific conditions. Extracellular GAL-7 has been detected in advanced breast tumors, in the secretomes of breast and lung cancers, and in multiple myeloma, where elevated circulating levels correlate with adverse clinical outcomes. Its release has also been observed in nonmalignant tissues, such as placental trophoblasts during pre-eclampsia. Importantly, GAL-7 differs from GAL-1 in that its expression is far more restricted: while GAL-1 is widely expressed across the tumor microenvironment and in many normal tissues, GAL-7 expression is largely confined to specific epithelial compartments. This limited tissue distribution may represent a therapeutic advantage, as it likely reduces the risk of off-target effects. Indeed, galectin-7 deficiency primarily affects epithelial stress responses and wound healing, whereas galectin-1 deficiency leads to broad immune, metabolic, and skeletal abnormalities. These differences highlight why GAL-7 represents an appealing preclinical target. Further studies will be required, however, to determine whether potential side effects of galectin inhibition arise predominantly from the loss of intracellular versus extracellular functionsan issue that our newly developed minibodies will be particularly well suited to address.

While our study provides strong preclinical evidence for the distinct and complementary roles of GAL-1 and GAL-7 in TNBC, several limitations should be acknowledged. First, the findings are based on a limited number of murine models, which, although widely used in immuno-oncology research, may not fully capture the complexity and heterogeneity of human TNBC. Here, we intentionally used human TNBC (MDA-MB-231) cells for in vitro cytokine profiling and murine TNBC (E0771) cells for our in vivo experiments to match the immunocompetent mouse model. This dual-species approach was chosen to capture the conserved galectin-driven mechanisms across human and murine systems while maintaining physiological relevance within an intact immune context. Importantly, the cytokine responses elicited by GAL-1 and GAL-7 in MDA-MB-231 cells were consistent with their distinct immunologic roles observed in vivo, supporting the translational value of our model system. Second, cytokine responses are inherently variable across experimental systems, and further studies are needed to assess their reproducibility and clinical significance. Mechanistically, it will be essential to dissect whether the effects of GAL blockade arise primarily from direct modulations of tumor cells, immune cells, or both. Importantly, our results suggest that combining G1M1 and G7M8 could enhance the therapeutic efficacy of glyco-immune checkpoint inhibition, raising the possibility of even more versatile approaches, such as bispecific minibodies or diabodies derived from the nanobody platform. Of note, we chose the human IgG1-Fc format because it is the most widely used subclass for therapeutic antibody development, offering well-characterized effector functions and favorable pharmacokinetics. Human IgG1 has sufficient affinity for mouse Fcγ receptors to enable informative preclinical assessment of biological activity, including immune effector mechanisms. Accordingly, future studies will be needed to compare different formats to confirm whether the observed effects are indeed driven by Fcγ-mediated mechanisms or simply reflect the blocking of galectin GBS. Additionally, it will be crucial to determine whether these inhibitors also enhance responses to standard treatment modalities, such as chemotherapy. Together, these future directions will help pave the way toward clinical translation, not only in TNBC but also in other GAL-1 and GAL-7-driven malignancies, including ovarian and esophageal cancers.

Conclusions

Our findings have important therapeutic implications, as they address two major barriers in the development of glyco-immune checkpoint inhibitors. First, by demonstrating that GAL-1 and GAL-7 orchestrate distinct cytokine and immune remodeling programs, our work provides new mechanistic insights into how individual galectins uniquely contribute to TNBC progression. Second, the development of G1M1 and G7M8 establishes that highly selective galectin inhibitors are both technically achievable and biologically effective in preclinical models. While the current in vivo evidence is derived from a single syngeneic model, the concordance between our in vitro and in vivo data reinforces the robustness of this conclusion. Future validation in additional models, particularly warm- and cold-TNBC models, will further substantiate the translational significance of selective galectin blockade as a strategy to complement existing immune checkpoint therapies.

Experimental Section

Reagents and Cell Lines

Jurkat, E0771, and MDA-MB-231 cells were obtained from the American Type Culture Collection (ATCC) and maintained in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). All cell lines were tested negative for Mycoplasma upon arrival at our laboratory. Human GAL-1, GAL-3, GAL-7, and mouse GAL-7 were produced and purified as previously described. Recombinant mouse GAL-1 was purchased from Sino Biological (cat. #50100-MNAE). GAL-1- and GAL-7-specific minibodies were generated by Hybrigenics by fusing nanobodies to a human IgG1 Fc fragment. Purified recombinant G1M1 and G7M8 minibodies were produced by Sino Biological using the Expi293 expression system and purified using protein A affinity chromatography. All synthetic compounds were confirmed to be >95% pure by HPLC. In the case of antibodies, they were at least 91–94% pure, as defined by densitometric analysis of Western blots. Endotoxin levels were assessed (<1 EU/mg). The antimouse PD-1 antibody (cat. #BE0146), rat IgG2a isotype control (cat. #BE0089), and human IgG1 isotype control (cat. #BE0297) were obtained from Bio X Cell.

Cell Culture Supernatant and Serum Profiling

MDA-MB-231 cells were plated in 24-well plates and cultured to 80% confluence. Cells were serum-starved for 24 h. After starvation, the medium was changed to DMEM supplemented with 1% FBS, and the cells were treated for 24 h with PBS (control) and 5 μM of GAL-1 and/or GAL-7. Supernatants were collected, centrifuged twice at 3,000g for 10 min, and stored at −80 °C. For serum profiling, whole blood samples were collected by cardiac puncture from isoflurane-anesthetized mice. Serum was isolated according to standard procedures. Briefly, blood samples were incubated at room temperature (RT) for 30 min to allow clot formation, then centrifuged at 2,000g for 10 min at 4 °C. The supernatants were collected and stored at −80 °C. Multiplex analyses of cytokine and chemokine markers were performed by Eve Technologies using the Luminex xMAP technology.

ELISA

Human recombinant GAL-1, GAL-3, and GAL-7, and mouse recombinant GAL-1 and GAL-7, were diluted to 10 μg/mL in PBS and coated onto a flat-bottom 96-well polystyrene plate (Costar) for 1 h at RT. The plates were washed with PBS and blocked with PBS containing 1% bovine serum albumin (PBA) for 1 h at RT. After washing, G1M1 and G7M8 were diluted to the indicated concentrations in PBA, added to the wells, and incubated for 1 h at RT. The binding of minibodies to GALs was detected using an antihuman IgG antibody (Sigma, cat. #I8635) diluted 1:60,000 in PBA and incubated for 1 h at RT, followed by incubation with an antirabbit HRP-conjugated secondary antibody (Promega, cat. #W401B) diluted 1:10,000 in PBA and incubated for 1 h at RT. After the final washes, 3,3 ′,5,5′-tetramethylbenzidine (TMB) Liquid Substrate System (Sigma, cat. #T8665) was added. The reaction was stopped after 15 min with 0.16 M sulfuric acid. The optical density was measured at 450 nm using a Tecan plate reader.

Apoptosis Assay

Jurkat T cells were incubated for 4 h at 37 °C with 4 μM GAL-1 or 20 μM GAL-7 and incubated overnight at 4 °C with or without a 5-fold molar excess of G1M1 or G7M8, respectively. As a positive control for apoptosis inhibition, 100 mM lactose was preincubated with GALs under the same conditions. Following incubation, cells were washed with binding buffer (0.01 M HEPES, 0.14 M NaCl, 2.5 mM CaCl₂, pH 7.4) before staining with Annexin V (0.63 μg/mL, Biolegend, cat. #640906) for 15 min at RT, followed by the addition of propidium iodide (0.25 μg/mL, Sigma, cat. #P4170). Apoptosis was measured using a FACSCalibur flow cytometer (BD Biosciences). 5,000 events were recorded for each sample.

Preclinical Mouse Models

C57BL/6 female mice were obtained from Charles River Laboratories. Mice were orthotopically injected into the fourth mammary fat pad with 1 × 10⁶ E0771 cells. Once the tumors became palpable, animals were randomly allocated to groups, and the mice were treated intraperitoneally with antibodies (100 μg/mouse) for 10 days. Anti-PD-1 and its isotypic control were administered every 2 days. Minibodies were administered daily. Mouse body weight and tumor volume were measured every 2–3 days. Tumor volume (V) was calculated using caliper measurements of tumor length (L) and width (W) according to the following formula: V = (L × W ²)/2. On days 27 and 28 of the study, the mice were euthanized. Blood, tumors, and lungs were collected for post-mortem analysis. The tumor weights were recorded using a scale. Lung tissues were fixed with 4% paraformaldehyde. Tissue embedding, sectioning, and staining were performed at the Histology Core Facility at the Institute for Research in Immunology and Cancer (IRIC) at the Université de Montréal. Investigators performing metastasis quantification were blinded to group identity.

Flow Cytometry Analysis of Tumor-Infiltrating Lymphocytes

Tumor-infiltrating lymphocytes were isolated from tumor tissues (n = 5) using the Mouse Tumor Dissociation Kit (Miltenyi Biotec) and gentleMACS Dissociator (Miltenyi Biotec), according to the manufacturer’s instructions. For cell surface staining, the following fluorochrome-conjugated monoclonal antibodies were used: PE-CF594 Rat Anti-Mouse CD45 (BD Horizon), FITC Anti-Mouse CD3 (BioLegend), APC-Cy7 Rat Anti-Mouse CD4 (BD Pharmingen), Alexa Fluor 700 Rat Anti-Mouse CD8a (BD Pharmingen), Brilliant Violet 605 Anti-Mouse CD279 (PD-1) (BioLegend), and Annexin V Conjugates (Invitrogen) (Table S1). For intracellular transcription factor staining, cells were fixed and permeabilized using the Anti-Human FoxP3 Staining Set APC (eBioscience), according to the manufacturer’s protocol. Flow cytometry data were acquired using an LSR Fortessa flow cytometer (BD Biosciences) and analyzed with FlowJo software v10.8.1 (BD Biosciences). Detailed flow cytometry analyses are shown in Figures S1–S4. Investigators performing gating and flow cytometric analyses were blinded to group identity.

Bioinformatics and Statistical Analysis

For in vitro and in vivo cytokine analysis, aside from the results shown in Figure , calculations and statistical analysis were performed with R version 4.5.0 (R Core Team, 2025). Key packages used included dplyr (version 1.1.4), ggplot2 (version 3.5.2), FSA (version 0.10.0) for nonparametric posthoc Dunn’s test, broom (version 1.0.10) for tidying Tukey’s HSD results, and rstatix (version 0.7.2) for the Games-Howell test and Welch posthoc analysis. − The different concentrations of cytokines obtained during the in vitro and in vivo experiments were first subjected to variance tests (Levene) and normality tests (Shapiro). Ordinary one-way ANOVA followed by Tukey HSD was applied for cytokines for which both Levene and Shapiro tests indicated a normal distribution. In cases where Levene’s test indicated homogeneity of variance but Shapiro’s test indicated a violation of normality, the Kruskal–Wallis test followed by Dunn’s test with BH correction was used. Finally, when Levene’s test indicated a violation of homogeneity but Shapiro’s test indicated normality, Welch’s ANOVA and Games-Howell tests were employed. To visualize differences in cytokine profiles, we conducted a Principal Coordinate Analysis (PCoA) based on pairwise comparisons of cytokine levels across treatment groups. First, cytokine data were transformed into pairwise comparison matrices, with adjusted p-values derived from statistical tests (Dunn’s test, Tukey’s HSD, and Games-Howell) for various cytokines. Missing pairs were treated as no differences (p_adj = 1). The similarity matrix was computed as 1 – padj1 – p_{\text­{adj}}­1 – padj, followed by the calculation of Euclidean distances between cytokine profiles. PCoA was performed using classical multidimensional scaling (cmdscale), yielding two dimensions (PCoA1 and PCoA2) representing the relationships among the cytokines based on their treatment effects. The PCoA coordinates were then plotted to illustrate biological trends among the cytokines. To visualize the absolute distance of cytokine levels from the control group (IgG2a + IgG1) across different treatment groups, a heatmap was created. The data were extracted from a distance matrix, focusing on comparisons that included the control group. Each cytokine’s distance from the control was represented as a color gradient, with darker shades indicating greater differences. For all other results of this study, statistical analyses were conducted using GraphPad Prism (v10.1.1). Data are presented as mean ± standard deviation (SD) and were analyzed using one-way ANOVA or the Kruskal–Wallis test, depending on the assumptions of normality and homogeneity of variances, as assessed using the Shapiro-Wilk and Levene’s tests, respectively. Metastatic incidence statistics were evaluated using a two-sided Fisher’s exact test on 2 × 2 contingency tables. While p < 0.05 was used as the threshold for statistical significance, we also reported results with 0.05 ≤ p < 0.10 to highlight potential biological effects that warrant further validation.

Glossary

Abbreviations Used

AICD

activation-induced cell death

CRD

carbohydrate recognition domain

GAL

galectin

MMPs

matrix metalloproteinases

Nb

nanobody

PCoA

principal coordinate analysis

TME

tumor microenvironment

TNBC

triple-negative breast cancer

The data generated in this study are available within the article and its Supporting Data files.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jmedchem.6c00142.

• Additional experimental details on flow cytometry analysis, mouse body weight kinetics, and flow cytometry markers and dilutions (PDF)

The study was conceived by R.N., P.G.J.d.B., D.C., and Y.S.P. All authors contributed to the data interpretation and critical appraisal of the study. They also conducted the experiments and/or contributed to the experimental design and analysis of the results. R.N. and Y.S.P. drafted the manuscript, with input from all authors at all stages. All authors reviewed and approved the final manuscript.

The authors acknowledge support from the Arbour Foundation (R.N.), Glyconet (Y.S.P.), the Quebec Consortium for Drug Discovery (CQDM) (Y.S.P.), Fonds de Recherche du Québec-Nature and Technology (FRQNT) (S.F. and F.F.), and the Canadian Institutes of Health Research (CIHR) (R.N., N.D., and Y.S.P.). Part of this work was also supported by the Molecular and Cellular Glycobiology Facility and the Molecular Interactions and Materials Characterization Facility (MIMC) of INRS, which are funded by a Canada Foundation for Innovation (CFI) Major Science Initiatives (MSI) fund awarded to GlycoNet Integrated Services. A.L. holds the Jeanne and J. Lévesque Research Chair in Immunovirology from the J. Louis Lévesque Foundation. We thank Audrey Dubé for her assistance with the in vivo studies.

All animal handling was performed in accordance with protocol #2110–03, which was approved by the Institutional Committee for the Protection of Animals of the INRS-Centre Armand-Frappier Santé Biotechnologie. These protocols follow the guidelines for animal care as set out by the Canadian Council on Animal Care, as described in the Guide to the Care and Use of Experimental Animals.

The authors declare the following competing financial interest(s): D.C., R.N., and Y.S.P. are co-inventors on patent applications filed by the Institut National de la Recherche Scientifique (INRS) related to this work, all of which concern the use of nanobodies and minibodies to inhibit biological, physiological, and/or pathological processes involving GAL-1 and GAL-7. All other authors have no competing interests to declare.