Endogenous Galectin-1 in T Lymphocytes Regulates Anti-prostate Cancer Immunity

The identification of effective new therapies for prostate cancer (PCa) requires a better understanding of the multiple molecular interactions between tumor cells and their associated microenvironment. In this context, galectin-1 (Gal-1) is a key molecule in the determination of the prostatic carcinoma microenviroment; therefore, it is essential to understand all the molecular processes in which this protein is involved. Most of the previous studies found in the literature have focused on the microenvironment remodeling properties of tumor-secreted Gal-1, through its interactions with the glyco-receptors at the cell membrane and the extracellular matrix. This report shows original aspects of the lectin by focusing on the role of lymphocyte endogenous Gal-1 in controlling anti-prostate tumor immunity. Using a murine preclinical model of prostate cancer, our results demonstrate that endogenous Gal-1 in lymphocytes modulates their proliferative rate and cytotoxic function in conditions of high extracellular Gal-1 concentration, mainly derived from tumor cells. In such conditions, the absence of Gal-1 in T lymphocytes potentiates anti-tumor immune responses. Further studies demonstrated that endogenous Gal-1 in CD4+, but mainly in CD8+T cells, acts as a negative regulator of anti-tumor immunity. In conclusion, prostate tumors require Gal-1 in lymphocytes to evade immune responses. This report lays the foundation for an original immunotherapy strategy for prostate cancer.


INTRODUCTION
Prostate cancer (PCa) is a major challenge for public health worldwide; epidemiological studies place it as the second-screening and the fifth-leading cause of cancer death in adult men (1). Suitable treatments currently exist for this type of cancer if detected early; in contrast, new treatment options are required for advanced stages of the disease (2). Immunotherapy is a validated alternative for the advanced stages of PCa, e.g., Sipuleucel-T (Provenge R ) is the first immune-based therapy approved by the US Food and Drug Administration for this type of cancer (3). In addition, several immunotherapy strategies are currently being evaluated for this and other types of cancers involving active immunization against tumor-associated antigens and blocking immune checkpoint molecules (4,5). In this context, Galectin-1 (Gal-1) is a clue molecule secreted by different types of cancers to promote a tolerogenic microenvironment (6). These tolerogenic effects occur locally at the primary tumor site but also at a distance since Gal-1 is found soluble in plasma and is transported inside microvesicles, such as exosomas (7). Indeed, most of the immunological studies about Gal-1 are focused on its lectin properties resulting from its interaction with β-galactosides and promotion of lattice formation with membrane receptors and extracellular matrix proteins [reviewed in (8)]. This kind of interactions induces lymphocyte apoptosis after activation (9)(10)(11), T cell exhaustion (7) and inhibits T-cell migration across endothelial cells (12); actions that favor the tumor-inefficient immune responses.
However, Gal-1 is also expressed intracellularly in several cell types; it has been reported on transformed tumor cells (13)(14)(15)(16), regulatory T cells (17,18) and antigen-presenting cells (19,20), where the endogenous Gal-1 plays a role as a regulator of the antitumor immune responses. Effector T lymphocytes also express endogenous Gal-1; the absence of Gal-1 in lymphocytes promotes a Th1 response during parasitic diseases (21,22). Few data are available about the role of the endogenous Gal-1 in the control of cytotoxic CD8+ T cell responses. On one hand, Gal-1 negatively regulates CD8+ T cell expansion, survival upon activation, and function fate (23). On the other hand, another study has demonstrated that Gal-1-deficient mice have impaired in vivo anti-viral CD8+ T cell-mediated immune responses (24). There are no data about the role of endogenous Gal-1 expressed by CD8+ T lymphocytes in the control of the anti-tumor properties. However, translating the second concept, in which the absence of Gal-1 impairs immune cytotoxicity, seems controversial, due to the compelling evidence in the literature. In fact, Lgals1-/mice have Foxp3+ Treg cell suppressive dysfunction (17) and are prone to autoimmunity (10,19,(25)(26)(27). More interestingly, the injection of tumorigenic cell lines into Lgals1-/-mice demonstrated reduction of tumorigenesis (28,29), proving the relative importance of the stroma (including immune cells) in determining tumorigenesis potential. Given this scenario, this research aimed to challenge the role of Gal-1 as an endogenous modulator of T cell properties in PCa. Although expressed at low levels in lymphocytes, endogenous Gal-1 plays a major role in the control of the lymphocyte anti-tumor functions in a PCa context. The results allow us to propose the modulation of the endogenous load of Gal-1 in T cells as a novel immunotherapeutic strategy for PCa.

Animals
Animal procedures complied with the Guidelines for the Welfare of Animals in Experimental Neoplasia (UK) and approved by the University of Buenos Aires's Institutional Animal Care and Use Committee (IACUC; FCEN protocol #2014-038). Six to eightweek-old male C57BL/6 mice were housed in the animal facility Abbreviations: PCa, prostate cancer; Gal-1, Galectin-1; TRAMP-C1, prostate tumor cell line used in this preclinical model; WT, wild type mice; Gal-1 KO (Lgals1-/-), Galectin-1 deficient mice; AOD, average stage of division; CFSE, Carboxyfluorescein succinimidyl ester used to trace cell division by flow cytometry; Th, T helper profile of the immune response; DC, dendritic cells; Treg, regulatory T cells; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; BAIM: brachial, axillary, inguinal and mesenteric lymph nodes. of the School of Sciences, University of Buenos Aires (Argentina). Lgals1-/-mice (originally produced by Dr. F. Poirier, France), were provided by Dr. Gabriel A. Rabinovich (Argentina). Ly5.1 C57BL/6 (CD45.1) mice were provided by Dr. Sophie Ezine (CNRS CDTA; Orleans, France). Athymic nude Foxn1 nu mice were acquired from the animal facility of the Veterinary School, National University of La Plata (Argentina).
Stable Gal-1 downregulated TRAMP-C1 cell line was produced by transduction with a shRNA lentivirus previously reported (30). Briefly, after 1 week, transduced (GFP+) cells were purified by cell sorting using a FACSAria II cytometer (BD Bioscience). Purification of the transduced cells was carried out if GFP + cells did not exceed 20%, in order to minimize the number of viral integrations and thus guarantee a minimum perturbation of the genome. As previously reported, this shRNA sequence produced 85% of Gal-1 down-regulation at the protein level.

Murine Samples
Lymph node (brachial, axillary, inguinal and mesenteric; BAIM) samples were harvested and single-cell suspensions were obtained by mechanical disruption.

Tumor Model in Nude Mice Reconstituted With Wt or Lgals1-/-Experienced Lymphocytes
Six to eight-week-old male nude mice were i.v. reconstituted with immune cells from one experienced wild type or Lgals1-/-mice. The day after reconstitution, 1 × 10 6 TRAMP-C1 cells (log phase of growth) were s.c. injected in mice in Matrigel (4-5 mg/ml; Corning). No changes in mice weights were detected during the protocol. Euthanasia was practiced when tumor volume reached 2500 mm 3 . Tumor size was measured every 2 days with caliper and calculated as W 2 × L/2, where W = width and L = length.

Statistical Analysis
At least three biological replicates (i.e., independent experiments) were carried out in each case, unless stated otherwise. Data represent mean ± SD. Prism software (GraphPad) was used. Two groups were compared with the Student t-test for unpaired data (two-sided tests). P < 0.05 was considered statistically significant.

Endogenous Gal-1 in Lymphocytes Negatively Regulates Their Functional Properties in a Tumor Context
A vast literature in different animal models and experimental design are found describing the tumor cells as the main producer of Gal-1 (13,15,(34)(35)(36)(37)(38)(39)(40)(41)(42)(43)(44). Regarding PCa, we previously described that Gal-1 is expressed at high levels in several human PCa cell lines and, more importantly, in patient prostatectomies, being upregulated during disease progression (30,45). Gal-1 is also highly expressed in murine TRAMP-C1 cells (Figures 1A,B) and the downregulation of Gal-1 in these cells by RNAi reduces their immunosuppressive potential. In fact, the in-vitro polyclonal activation of lymphocytes in the presence of Gal-1 downregulated TRAMP-C1 led to higher proliferation of CD8+T cells ( Figure 1C). Consequently, Gal-1 is an important immune escape mechanism activated by the tumors.
Besides tumor cells themselves, other cells in the tumor associated microenvironment play fundamental roles in the determination of the tumor evolution. Understanding how endogenous Gal-1 in each of these cells affects their own properties seems, therefore, essential to have a complete functional picture of the PCa microenvironment. In particular, little is known about the endogenous role of Gal-1 in immune cells. Therefore, we first checked whether Gal-1 is expressed in immune cells and compared its expression levels with those observed in tumor cells. Interestingly, the expression levels of Gal-1 in immune cells [T lymphocytes and antigen presenting cells (APC)] were considerably lower (10-20 times less) than those of tumor cells (Figures 1A,B). Low levels of Gal-1 expression in immune cells do not indicate a priori that lymphocyte endogenous Gal-1 does not play any role as an intrinsic modulator of their properties. Thus, we decided to design an in vitro experimental model where the function of Lgals1-/-lymphocytes would be compared with the wildtype (WT) ones in the presence or absence of wild type tumor (responsible of secreting high levels of "exogenous Gal-1"). As readout, we focused on the function of CD8+ T lymphocytes, as these cells are critical in the anti-tumor cytotoxic response (46). Our results demonstrate that without endogenous Gal-1, CD8+ T cells showed higher proliferative rates (Figure 2). A detailed analysis of these results shows that, in the absence of tumor, there was a trend toward greater proliferation of Gal-1-deficient CD8+ T lymphocytes mainly at low conditions of stimulation (Figures 2A,B, 1µg/mL anti-CD3). However, the results were considerably more significant in the presence of an immunosuppressive microenvironment generated by the tumor cells (Figures 2C,D); the absence of endogenous Gal-1 led to higher CD8+ T cell proliferative capacity even in such conditions. We highlight that, the major source of extracellular Gal-1 in these experimental settings are tumor cells that almost saturate the extracellular medium by the secretion of high quantities of Gal-1 [dosed by ELISA at levels of 20 ng/mL, supported also by the high cellular levels of Gal-1 shown in Figure 1, and evaluated in patients in our previous study (30) and reported at high levels in the serum of cancer patients (47)(48)(49)(50)]. The most potent abrogation of tumor-immunosuppression was achieved by the absence of endogenous Gal-1 in lymphocytes combined with the down-regulation of Gal-1 in tumor cells (Figures 1C,D). While the immune effects of Gal-1 in tumor cells have been extensively evaluated (13)(14)(15)51), the endogenous effect of Gal-1 effect in lymphocytes is original and requires more understanding.
Since a widely described action of exogenous Gal-1 is the induction of cell death on activated T lymphocytes (9), we evaluated if lymphocytes with no endogenous Gal-1 were differentially induced into apoptosis in our tumor experimental model. Figure 2E shows the percentage of apoptotic cells determined by Annexin V staining inside the CD8+ T cell subpopulation, and Figure 2F shows the same staining inside each peak of division of CD8+ T cells. Altogether, these results indicate that the absence of endogenous Gal-1 in CD8+T lymphocytes protects these cells from apoptosis in the presence of tumor cells. As in proliferative studies, the major source of extracellular Gal-1 was the tumor.
Degranulation is a prerequisite to perforin-granzymemediated cytotoxic function by responding to antigen-specific CD8+T cells. The lytic granules are surrounded by a lipid bilayer containing numerous lysosomal-associated membrane proteins (LAMPs, CD107a and b); thus, labeling responding cells with anti-CD107 antibodies and measuring their expression by flow cytometry can directly identify degranulating CD8+ T cells. The results shown in Figure 3 demonstrate that the absence of endogenous Gal-1 in CD8+ T lymphocytes increases their ability to degranulate. This was also observed in conditions without tumor cells and at different percentages of TRAMP-C1 (Figure 3 and data not shown). Interestingly, endogenous Gal-1 in lymphocytes affects not only their percentages of degranulating cells (Figure 3A), but also the CD107a mean fluorescence intensity (MFI). This last result indicates that the intrinsic expression/exposition of CD107a is regulated by this lectin on a per cell basis ( Figure 3B). Altogether, these results imply that endogenous Gal-1 in CD8+T cells an impotant role in the control of their function, particularly in the presence of tumor cells.

Lymphocyte Endogenous Gal-1 Modulates Their Function Independently of Exogenous Factors
To evaluate if the immunosuppressive role of lymphocyte endogenous Gal-1 involves cell-to-cell contacts or the production of soluble factors, we performed transwell experiments. Figure 4 shows that Gal-1 deficient lymphocytes proliferate more extensively than wild-type ones when activated in the presence of tumor (as in Figure 2). We next co-cultured wild-type and Gal-1 deficient lymphocytes (they differed in the expression of CD45 isoform; both were activated and proliferation was evaluated in CD45.2+ CFSE labeled). The presence of wild-type cells reduced the proliferation of Gal-1 deficient CD8+T cells. However, when these two lymphocyte sources were separated through a transwell membrane, such a reduction in proliferation was no longer observed (Figure 4). This experiment clearly demonstrates the fundamental role of endogenous lymphocyte Gal-1 in controlling their function upon activation, a phenomena that is independent of the tumor secretion of Gal-1 and can be controlled through lymphocyte-lymphocyte interactions. This introduces new variants for the current understanding of the role of Gal-1 in defining a tumor-associated tolerogenic context.

Gal-1 Negatively Regulates in vivo Immunity Against PCa
Previously, it has been reported that several tumor cells lose their tumorigenicity when injected into Lgals1-/-mice, compared with their injection into wild-type mice (28,29). We obtained the same behavior in the case of the TRAMP-C1 prostate tumor model. In fact, while 92% of wild-type mice injected s.c. with TRAMP-C1 cells developed tumors (lag time: 47 ± 5; duplication time: 7 ± 2 days, n = 5), no Lgals1-/-mice (0/4) developed tumors after up to 80 days. However, the conclusions obtained from this type of experimental model are complex, and the real impact of immune-mediated effects cannot be established since Gal-1 is absent in all cells of the body, including endothelial cells in which it was previously demonstrated Gal-1 plays a major regulatory role. To evaluate more precisely whether endogenous Gal-1 in lymphocytes affects the capacity of mice to control tumor growth under in vivo settings, we developed the following experimental procedure: anti-tumor lymphocytes were induced by the immunization of Lgals1-/-and wild-type mice with dendritic cells pulsed with TRAMP-C1 lysates. At day 7 postimmunization, immune cells were obtained from the BAIM lymph nodes; no significant phenotype changes were detected at this stage (total absolute cell numbers, %CD4+, %CD8+, %CD4+CD25+Foxp3+Tregs; data not shown). Lymphocytes obtained from the lymph nodes of immunized mice (wild type or Lgals1-/-) were then transferred into an immunodeficient nude mouse (1-to-1 mice ratio). Under such reconstitution settings, all non-immune host compartments (e.g., stroma and particularly endothelial cells) expressed Gal-1 at normal levels, with the exception of T lymphocyte (transferred) cells. This system allowed us to compare the ability of Lgals1-/-and wildtype lymphocytes to control tumor growth in a WT microand macro-environment. An in-depth comparison analysis demonstrated that mice reconstituted with Lgals1-/-lymphocytes showed a marked increase in tumor latency and a slight decrease in tumor duplication time ( Figure 5 and Table 1). In contrast, tumorigenesis, defined as the percentage of mice developing tumors, was 100% in both experimental groups, implying that endogenous Gal-1 on lymphocytes plays an important role in controlling tumor growth in initial phases, but later other molecular pathways can circumvent its action.

Cellular Mechanisms Involved in Lymphocyte Modulation by Endogenous Gal-1
Considering the previous in-vitro and in-vivo results, we analyzed the impact of the endogenous Gal-1 absence in the different subpopulations of cells involved in the immune response. For this purpose, we purified APC, CD4+, and CD8+ T cells to homogeneity and performed in-vitro proliferation assays by mixing the subpopulations of different genotypes in fisiological proportions. The results shown in Figure 6 demonstrate that the absence of Gal-1 in CD4+ and CD8+ T cells results in higher proliferative rates of CD8+ T lymphocytes. Higher effect was observed when Gal-1 was absent in the responding CD8+ T cells compared to CD4+ T cells or all wild-type immune cells (AOD = 1.7 vs. 1.1 or 0.8, respectively); however, maximal CD8+ T cell proliferation occurs when Gal-1 is absent in both lymphoid subpopulations (AOD = 2.8). The absence of Gal-1 in APC did not produce any change in the proliferative potential of CD8+ T lymphocytes (data not shown). Altogether, these results clearly demonstrate that endogenous Gal-1 regulates the functional properties of CD4+ and CD8+ T cells.

DISCUSSION
The pleiotropic functions of Gal-1 are partly due to the complex cellular distribution of this lectin. At the extracellular level, Gal-1 is able to interact with glycosylated receptors (through their Nacetyllactosamine glycans), thereby forming complex membrane lattices [reviewed in (52)]. Through this outside-in process, Gal-1 regulates several cellular properties, acting as a negative regulator of immune responses [reviewed in (53,54)]. Consistently, most of the current Gal-1-based immune strategies proposed for autoimmunity and cancer are focused at the extracellular level.   Additionally, the intracellular regulation of Gal-1 is expected to perform extra regulatory functions. At this subcellular localization, Gal-1 regulates several cellular processes, such as growth, survival, and differentiation; several of these effects are carbohydrate-independent [reviewed in (52)]. It has been demonstrated that the downregulation of the endogenous Gal-1 expressed by tumor cells has a direct effect on the immune response [our results and (13)(14)(15)51)]. As expected, these strategies also modify the extracellular levels of Gal-1 (38), as the transformed cells are the main Gal-1 producers. Interestingly, Gal-1 is also expressed by other cells of the tumor microenvironment; the endogenous function of Gal-1 in those cells may affect tumor growth. Indeed, several reports have demonstrated the decreased tumorigenic potential of various tumor cell lines when injected into Gal-1-deficient mice (28,29). The absence of Gal-1 on endothelial cells of Lgals1-/-hosts is responsible for the loss of tumorigenicity; angiogenesis defaults prevent an adequate supply of nutriments, oxygen, and hormones, factors that are essential for optimal tumor growth (28). While this report does not exclude the importance of this scenario, it proposes a complementary alternative in which tumors require Gal-1 to be endogenously expressed in T lymphocytes to escape from immune attack.
In fact, our results demonstrate that lymphocytes isolated from Gal-1-deficient mice have higher CD8+ T cell proliferative rates and effector functions, even when activated in the presence of high levels of extracellular Gal-1 secreted by tumor cells. This biological observation could be due to a direct endogenous effect of Gal-1 on CD8+ T cells or through indirect effects on other cell types; this report dissected the functional role of endogenous Gal-1 in immune cells. Our results show that the expression of Gal-1 on antigen-presenting cells is indifferent at potentiating CD8+ T cell functions. However, the absence of Gal-1 in CD4+T cells induced higher levels of CD8+ T proliferation. Gal-1 plays a fundamental role in the control of differentiation and function of conventional CD4+CD25+Foxp3+ Tregs (17). While we did not detect changes in the frequency of Tregs in our experimental settings (data not shown), such Tregs are functionally deficient (17). Consequently, CD4+T cell subpopulations from Lgals1-/-and wild-type mice differ at least at this functional level. However, other endogenous modifications are also plausible in the CD4+T cell compartment, as was previously demonstrated in parasitic models of infection (21,22). More interestingly, the absence of endogenous Gal-1 induces higher functional properties of CD8+ T cells. These results agree with a previous report that demonstrated that Gal-1 has a major impact over proliferation burst, lymphocyte survival upon activation, and effector fate (23).
Mechanistically, our transwell experiments demonstrated that the immunoregulation mediated by endogenous Gal-1 in immune cells involves cell-to-cell interactions. In fact, the coculture of Gal-1 deficient together with wild-type lymphocytes abolished the immunopotentiation observed in Lgals1-/-cells alone. In addition, our results demonstrate that soluble factors are not involved in this type of regulation. Altogether, these results led us to hypothesize that Gal-1 deficient lymphocytes lack immune-inhibitory membrane receptors (or, alternatively, acquire immune-activating receptors) that have a direct effect on lymphocyte function. It has been reported that intracellular Gal-1 could affect cell behavior through regulation of the transcriptional machinery (Gal-1 forms part of the spliceosome), miRNA, and signaling pathways (52). Effects coming from intracellular Gal-1 interactors in lymphocytes have not yet been investigated.
Interestingly, the relevant roles of lymphocyte endogenous Gal-1 have recently been proposed in active systemic lupus erythematosus. In fact, patient lymphocytes demonstrate lower levels of endogenous Gal-1 upon activation, a mechanism that may make them prone to autoimmunity (55). In cancer, our results support a scenario where the high levels of exogenous Gal-1 produced by tumor cells are not enough to evade immune attack; tumors require additional licensing by lymphocyte endogenous Gal-1 to be efficiently immunesuppressive. These results define a new and potent target for immunotherapy in PCa. From a clinical point of view, it is interesting to note that the immunopotentiation resulting from the downregulation of Gal-1 in lymphocytes is more significant in the presence of a tumor; consequently, an anti-PCa therapeutic strategy based on these results should have few side effects.

AUTHOR CONTRIBUTIONS
EC and GC performed and analyzed the work. DC data interpretation, critical revision. DL conceived the idea, analyzed data, supervised the project and wrote the paper. All authors discussed the results and contributed to the final manuscript.