Galectin-3 expression in donor T cells reduces GvHD severity and lethality after allogeneic hematopoietic cell transplantation

SUMMARY

Abundant donor cytotoxic T cells that attack normal host organs remain a major problem for patients receiving allogeneic hematopoietic cell transplantation (allo-HCT). Despite an increase in our knowledge of the pathobiology of acute graft versus host disease (aGvHD), the mechanisms regulating the proliferation and function of donor T cells remain unclear. Here, we show that activated donor T cells express galectin-3 (Gal-3) after allo-HCT. In both major and minor histocompatibility-mismatched models of murine aGvHD, expression of Gal-3 is associated with decreased T cell activation and suppression of the secretion of effector cytokines, including IFN-γ and GM-CSF. Mechanistically, Gal-3 results in activation of NFAT signaling, which can induce T cell exhaustion. Gal-3 overexpression in human T cells prevents severe disease by suppressing cytotoxic T cells in xenogeneic aGvHD models. Together, these data identify the Gal-3-dependent regulatory pathway in donor T cells as a critical component of inflammation in aGvHD.

Graphical Abstract

In brief

Mohammadpour et al. provide evidence that galectin-3 in donor T cells plays a role in reducing acute GvHD after allogeneic hematopoietic cell transplantation.

INTRODUCTION

Allogeneic hematopoietic cell transplantation (allo-HCT) is a potentially curative intervention for many patients with hematological disorders and malignancies. After this treatment, donor T cells can eradicate tumor cells through a process called the graft-versus-tumor (GvT) effect.^1,2 However, donor T cells can also attack normal, healthy tissues, typically the skin, gastrointestinal (GI) tract, and liver, leading to acute graft-versus-host disease (aGvHD).^3,4 aGvHD is the primary early complication after allo-HCT due to the highly inflammatory microenvironments in target organs, largely induced by donor T cells.^5,6 After allo-HCT, donor T cells differentiate to effector T cells, producing inflammatory and cytotoxic cytokines such as granulocyte-macrophage colony-stimulating factor (GM-CSF)⁷ and interferon-γ (IFN-γ).^8,9 This can result in tissue destruction and disease.^6,10,11 Therefore, there is a continued need to understand the mechanisms involved in T cell activation after allo-HCT to develop strategies to decrease aGvHD severity and lethality while preserving GvT activity. We previously discovered that β2-adrenergic receptor signaling in donor T cells regulates the expression of Lgals3 (the gene for the glycoprotein galectin-3) after allo-HCT, and it was associated with improved outcomes of aGvHD.¹² Since it is critical to learn how to mitigate aGvHD while preserving the GvT effect following allo-HCT,² we focused on learning more about the role of galectin-3 (Gal-3) in pathways responsible for regulating T cell activity and determining whether high Gal-3 in donor T cells during GvHD is associated with improved outcomes.

Galectins are involved in many physiological functions, such as inflammation, immune responses, cell migration and adhesion, autophagy, and signaling, each of which may contribute to the regulation of GvHD severity.¹³ Gal-3 exists in multiple conformational states: monomer, dimer, and pentamer with ligands including carbohydrate and protein. It is present in soluble, extracellular, membrane-bound, and intracellular states as a manifestation of Gal-3’s pleiotropic effects.^13,14 Intracellular Gal-3 can interact with its ligands, such as B cell lymphoma 2 (Bcl-2) and synexin, to regulate cell apoptosis, while extracellular Gal-3 interacts with CD7 and CD29 on activated T cells to induce apoptosis.^15,16 Within the 15 member galectin family,¹⁷ the protective role of galectin-1 and galectin-9 expressed by recipient tissues in ameliorating GvHD severity and lethality has been shown.^18-20 However, the roles of galectin members expressed by donor T cells during allo-HCT are still unknown. Based on our previous observations¹² and a publicly available dataset,²¹ we focused on the finding that Lgal-3 is the sole galectin family member increased in T cells after allo-HCT. Gal-3 is a unique member of the galectin family as it is composed of a carbohydrate recognition domain and an oligomerization domain, which enables the cross-linking of bound targets. Gal-3 is immune regulatory for distinct immune cell types such as dendritic cells and macrophages,¹⁴ and epithelial cells and myeloid cells express Gal-3, which increases during inflammatory processes such as bacterial infections.²² Furthermore, Gal-3 expression on myeloid cells and tumor cells promotes anti-inflammatory M2 macrophage differentiation and tumor progression, respectively.²³ While resting T cells do not express Gal-3, it is expressed upon T cell activation. Gal-3 localization in the mitochondria correlates with its anti-apoptotic effects; in addition, Gal-3 can be expressed on the cell surface or released from activated T cells, available for binding to its extracellular ligands.^24,25 Extracellular Gal-3 suppresses T cell receptor (TCR) signaling, decreasing T cell cytokine production, including IFN-γ.²⁵ Most importantly, Gal-3 plays a role in increasing the development of memory T cells,²⁴ which differentially maintain anti-tumor GvT efficacy while exhibiting a reduced ability to cause aGvHD.^26,27

In this study, we explored the effect of Gal-3 on donor T cells to develop a potential therapy to reduce aGvHD severity. Using several preclinical murine models of allo-HCT, we demonstrated that increased Gal-3 expression by donor T cells significantly impedes T cell proliferation and function and drives T cell exhaustion. These data reveal an important role of Gal-3 in the regulation of aGvHD severity. From these data, we propose that developing a Gal-3 agonist or Gal-3 overexpression in allogeneic donor T cells may be a useful approach for designing a cellular therapy that reduces the severity and lethality of aGvHD while preserving the GvT effect.

RESULTS

Gal-3 expression in donor T cells after allo-HCT reduces aGvHD severity and lethality

To understand the contribution of Gal-3 signaling during allogeneic, donor T cell-mediated aGvHD, we used major histocompatibility complex (MHC)-mismatched and/or minor histocompatibility antigen (miHA)-mismatched models of murine allo-HCT. For the former, wild-type (WT) BALB/c mice (H-2k^d) were lethally irradiated and transplanted with T cell-depleted bone marrow (TCD-BM) from CD45.2⁺ wildtype (WT) C57BL/6 (B6) (H-2k^b) mice with or without B6 WT or Gal-3^−/− T cells to induce aGvHD. Donor T cells from spleen and liver, obtained on day 7 after allo-HCT, were analyzed for Gal-3 expression. Surface and intracellular Gal-3 expression was upregulated on CD4⁺ and CD8⁺ T cells on day 7 as evidenced by flow cytometry, percentage of positive cells, and mean fluorescence intensity (Figures 1A and S1A). In addition, we utilized a publicly available, preclinical T cell RNA-sequencing dataset to investigate the expression of galectin family member genes in T cells after allo-HCT.²¹ We found that Lgals3 (Gal-3) is the only gene of the eight studied galectin family members with significantly increased mRNA levels in T cells in aGvHD target organs after allo-HCT (Figure S1C). Of note, there were no fetal or developmental abnormalities in Gal-3-deficient mice (data not shown). There were also no significant differences in the T cell phenotypes between naive WT and Gal-3^−/− mice (Figure S1B). To extend our observations of donor T cell Gal-3 expression in aGvHD, lethally irradiated BALB/c (H-2k^d) and C3H.SW (H-2k^b) mice were given TCD-BM with or without B6 (H-2k^b) WT or Gal-3^−/− T cells. The absence of Gal-3 in T cells significantly exacerbates aGvHD severity as evidenced by body weight loss and increased lethality rates in both MHC and miHA-mismatch (B6 → BALB/c; 3.5 × 10⁶ TCD-BM + 0.7 × 10⁶ T cells)(Figure 1B) and miHA-mismatch only (B6 → C3H.SW; 3.5 × 10⁶ TCD-BM + 1.5 × 10⁶ T cells) models (Figure 1C). To dissect the roles of CD8⁺ and/or CD4⁺ T cells in this phenotype, lethally irradiated BALB/c mice were transplanted with TCD-BM with or without B6 CD8⁺, CD4⁺, or CD4⁺CD25⁻ T cells. The absence of Gal-3 in CD8⁺ (B6 → BALB/c; 3.5 × 10⁶ TCD-BM + 2 × 10⁶ CD8⁺ T cells) (Figure 1D), CD4⁺ (B6 → BALB/c; 3.5 × 10⁶ TCD-BM + 0.2 × 10⁶ CD4⁺ T cells) (Figure 1E), or CD4⁺CD25⁻ T cells (B6 → BALB/c; 3.5 × 10⁶ TCD-BM + 0.2 × 10⁵ CD4⁺ T cells) (Figure S1D) significantly increased aGvHD severity and lethality. In all models, the most prominent phenotype was manifested by weight loss and diarrhea, especially in recipients transplanted with Gal-3^−/− T cells. Liver, small intestine, and large intestine were the major target organs, and no signs of skin GvHD were observed. Histologic preparations of the small intestine, large intestine, and liver were examined in a blinded manner to assess aGvHD pathology using a previously established semiquantitative scoring system.^12,28 Gal-3^−/− T cells significantly increased the aGvHD score in small intestine, large intestine, and liver compared with WT T cells (Figure S1E). These data suggest that donor T cells express Gal-3 as an inhibitory molecule of aGvHD after allo-HCT, and Gal-3 expression in donor T cells plays an important role in tempering aGvHD after allo-HCT and acts as a negative regulator of T cell function.

Figure 1. Gal-3 expression on T cells ameliorates the severity and fatality of acute GvHD after allo-HCT.

(A) Lethally irradiated BALB/c mice were transplanted with TCD-BM alone or with 0.7 × 10⁶ B6 WT or Gal-3^−/− T cells. On day 7, splenic cells were harvested and analyzed by flow cytometry. T cells were gated from single live H-2k^b+ H-2k^d− CD45⁺ CD3⁺ cells. The left shows the percentage of surface Gal-3-positive T cells and the right shows the percentage of intracellular Gal-3-positive T cells after allo-HCT (n = 5–8).

(B–E) Body weight and survival of lethally irradiated BALB/c (B, D, and E) or C3H.SW (C) mice after allo-HCT BM with or without (B) 0.7 × 10⁶ B6 WT or Gal-3^−/− pan T cells (n = 24); (C) 1.5 × 10⁶ B6 WT or Gal-3^−/− pan T cells (n = 16); (D) 2 × 10⁶ B6 WT or Gal-3^−/− CD8⁺ T cells (n = 8); and (E) 0.2 × 10⁶ B6 WT or Gal-3^−/− CD4⁺ T cells (n = 8). For comparison of survival curves, a log-rank (Mantel-Cox) test was used in (B), (C), (D), and (E). One-way analysis of variance (ANOVA) with Tukey’s multiple comparisons test was used for Gal-3 expression difference in (A). Two-way ANOVA with Tukey’s multiple comparisons test was used for body weight differences in (B), (C), (D), and (E). In (B-E) graphs, TCD-BM + WT T cells group was compared with TCD-BM + Gal-3^−/− T cells group; **p < 0.01 and ***p < 0.001. Body weights are shown as means ± SEM. The Gal-3 expression data are presented as individual values ± SD.

Gal-3 decreases T cell proliferation while protecting against IFN-γ-mediated apoptosis

We next sought to investigate the mechanisms driving the suppressive function of Gal-3 in T cells. Lethally irradiated BALB/c recipients were transplanted with TCD-BM and T cells from B6 WT or Gal-3^−/− donor mice, and the effects of Gal-3 on T cell proliferation were investigated. At day 7 following allo-HCT, we found that the percentage and number of donor Gal-3^−/− T cells were significantly higher in the spleens of BALB/c recipient mice compared with WT T cells, but interestingly, at day 14, there were no significant differences (Figure 2A). We also found that the number of donor Gal-3^−/− T cells was significantly higher in the livers and lymph nodes of BALB/c recipient mice compared with WT T cells at day 7 after allo-HCT (Figure S2A). Next, we used eF670 dye dilutions to analyze the effect of Gal-3^−/− on T cell proliferation (Figure 2B). Lethally irradiated BALB/c recipients were transplanted with TCD-BM together with eF670-labeled T cells from B6 WT or Gal-3^−/− donor mice, and T cell proliferation was analyzed 3 and 7 days after allo-HCT. These data show that Gal-3 deficiency significantly increases T cell proliferation (Figure 2B). To confirm the proliferation data, we carried out a co-transfer experiment using WT (CD45.1) and Gal-3^−/− (CD45.2) T cells (Figures 2C and 2D). Lethally irradiated BALB/c recipients were transplanted with TCD-BM and a 50:50 mixture of B6 WT and Gal-3^−/− donor T cells, and the effect of Gal-3 on T cell proliferation was investigated. At day 7 after allo-HCT, we found the percentage of Gal-3^−/− (CD45.2) T cells was significantly higher compared with WT T cells (CD45.1), but at day 14, the percentage of WT T cells was higher than that of Gal-3^−/− T cells (Figure 2C). We theorized that this reversal in T cell numbers is associated with activation-induced cell death (AICD).²⁹ It has been shown that Gal-3^−/− T cells release significantly larger amounts of IFN-γ²⁵ and that T cell-derived IFN-γ can directly induce T cell apoptosis.^20,30 We hypothesized that, despite the higher proliferation in Gal-3^−/− T cells compared with WTT cells early after allo-HCT, Gal-3^−/− T cells are more susceptible to apoptosis, resulting in a lower percentage of Gal-3^−/− T cells compared with WT T cells at day 14. To examine this premise, lethally irradiated BALB/c recipients were again transplanted with TCD-BM and a 50:50 mix of T cells from B6 WT or Gal-3^−/− donor mice. The level of apoptosis in T cells was investigated on day 7 after allo-HCT (Figure 2D). The rate of apoptosis was significantly higher in Gal-3^−/− T cells compared with WT T cells, as evidenced by a higher percentage of caspase 8 and annexin V-positive cells in Gal-3^−/− CD8⁺ and CD4⁺ T cells (Figure 2D). To determine whether the apoptosis in Gal-3^−/− T cells is mediated by IFN-γ, lethally irradiated BALB/c recipients were transplanted with TCD-BM and a 50:50 mixture of B6 WT and Gal-3^−/− donor T cells. From day 1 until day 14 post allo-HCT, recipient mice were treated with isotype control or anti-IFN-γ antibodies. The level of apoptosis in T cells was analyzed by flow cytometry on days 7 and 14. As expected, apoptosis was significantly higher in the Gal3^−/− T cells from the isotype control group, and anti-IFN-γ treatment was able to decrease the apoptosis level of Gal3^−/− T cells to the level of WT T cells (Figure 2E). To confirm the role of Gal-3 in protecting allogeneic T cells against IFN-γ-mediated apoptosis, lethally irradiated BALB/c recipients were transplanted with TCD-BM and a 50:50 mixture of B6 WT and Gal-3^−/− donor T cells with isotype or anti-IFN-γ antibody treatment. At day 7 after allo-HCT, we found a higher ratio of Gal-3^−/− T cells to WT T cells, but at day 14 there was no statistically significant differences between groups. In allo-HCT recipients treated with anti-IFN-γ antibody, the percentage of Gal-3^−/− T cells increased compared with WT T cells at day 7 and remained significantly higher at day 14 (Figure 2F). To further elucidate the role of Gal-3 on T cell apoptosis, lethally irradiated BALB/c recipients were transplanted with TCD-BM and T cells from B6 WT or Gal-3^−/− donor mice. At day 7 after allo-HCT, donor T cells (H-2k^b+, H-2k^d−, CD45⁺, TCRβ⁺) were sorted from spleen and cultured with IFN-γ in the presence of CD3 and CD28 for 3 days. The data showed that the level of apoptosis was significantly higher in Gal-3^−/− T cells compared with WT T cells (Figure S2B). We further examined the role of Gal-3 in T cell apoptosis in vitro as previously described.³¹ WT and Gal-3^−/− T cells were cultured with CD3/CD28 beads in the presence of IFN-γ and IL-2 for 72 h. Activated T cells were re-stimulated with CD3 for 20 h in the presence of IFN-γ, and the levels of apoptosis were measured using flow cytometry. The levels of apoptosis in Gal-3^−/− T cells after allo-HCT were significantly higher than in WT T cells after re-stimulation (Figure S2C). These data show that Gal-3 expression in T cells after allo-HCT significantly decreases T cell proliferation while protecting T cells partially from IFN-γ-mediated apoptosis. Thus, Gal-3 participates in balancing T cell proliferation and apoptosis, explaining the significant but not profound differences in GvHD severity and survival induced by WT versus Gal-3^−/− T cells.

Figure 2. Gal-3 expression modulates T cell proliferation and activation-induced apoptosis.

(A) Lethally irradiated BALB/c mice were transplanted with BM plus B6 WT or Gal-3^−/− T cells. At 7 and 14 days after allo-HCT, the percentage and number of splenic T cells were analyzed by flow cytometry by gating single live H-2k^b+ H-2k^d− CD45⁺ CD3⁺ cells.

(B) Lethally irradiated BALB/c mice were transplanted with BM plus eF670-stained B6 WT or Gal-3^−/− T cells. At days 3 and 7 after allo-HCT, T cell proliferation was analyzed by flow cytometry gated on H-2k^b+ H-2k^d− CD45⁺ CD3⁺ cells, demonstrating that Gal-3^−/− T cells had a higher level of proliferation. Proliferating T cells (red) were overlaid on control T cells (blue).

(C and D) Lethally irradiated BALB/c mice were transplanted with BM plus CD45.1 (WT) and CD45.2 (Gal-3^−/−) at 50:50 ratio. (C) Splenic percentage of WT (CD45.1⁺) and Gal-3^−/− (CD45.2⁺) of the BALB/c mice was analyzed by flow cytometry on days 7 and 14. (D) At day 7 after allo-HCT, apoptosis (caspase 8 and annexin V expression) was analyzed in T cells gated from single live H-2k^b+ H-2k^d− showing significantly higher levels of caspase 8 and annexin V in Gal-3^−/− CD8⁺ and CD4⁺ T cells.

(E and F) Recipient mice were treated with isotype or anti-IFN-γ antibody (intraperitoneal, 400 μg per mouse, twice per week). (E) The level of apoptosis was measured by using annexin V on day in CD4⁺ and CD8⁺ cells. (F) T cell proliferation in WT versus Gal-3^−/− T cells was analyzed on days 7 and 14 by flow cytometry. Data were pooled from two individual experiments, each with n = 3–5 per group to obtain a total of n = 6–10 for (A), n = 6 for (B), and n = 9–10 per group in (C–F). For comparison of the means, an unpaired Mann-Whitney test was used in all experiments; *p < 0.05, **p < 0.01, ***p < 0.001. The data are presented as individual values ± SD.

Gal-3 expression in T cells induces T cell exhaustion and suppresses T cell effector function

Our experiments indicated that T cells express Gal-3 after allo-HCT, and this expression ameliorates aGvHD severity and lethality. Further, we showed that Gal-3 decreases T cell proliferation, but inhibits IFN-γ-mediated apoptosis in T cells. We next asked whether Gal-3 changes the activation phenotype/status of T cells. To address this question, lethally irradiated BALB/c recipients were transplanted with TCD-BM and T cells from B6 WT or Gal-3^−/− donor mice. At day 7, splenic WT or Gal-3^−/− donor CD4⁺ and CD8⁺ T cells were isolated, and immune-related gene expression was analyzed by mRNA microarray (nCounter immunology panel, Nanostring, Seattle, WA). This panel measures more than 500 murine immune cell-related genes. Gal-3^−/− CD4⁺ T cells express more cytotoxic-related genes, including Tbx21, Irf7, Bcl6, and Il18, while WT CD4⁺ T cells express more exhaustion-associated genes, including Ptgs2,³² Pdcd2, Pdcd1, and Ctla4 (Figure 3A) and anergy-related genes such as Grail, Ikzf1, and Egr1. In Gal-3^−/− CD8⁺ T cells, the expression of genes related to proliferation and memory, such as Aicda, Cxcr5, Ilf3, and Tcf7, increased, while WT CD8⁺ T cells expressed more suppression- and anergy-related genes, such as Pdgfb, Cd244, Ptgs2, Tgfb, Ikzf1, Ikzf2, and Egr1 (Figure S3A). To assess the role of Gal-3 in T cell exhaustion, lethally irradiated BALB/c recipients were transplanted with TCD-BM and B6 WT or Gal-3^−/− T cells. On day 7, splenic donor T cells were isolated and examined for exhaustion markers and cytokine expression levels. The percentage of exhausted T cells (PD-1⁺ Lag3⁺) was significantly higher in WT T cells compared with Gal-3^−/− T cells (Figure 3B). TIM-3 and SLAMF6 have recently been used to differentiate progenitor and terminally exhausted T cells.³³ With these markers, we showed that progenitor and terminally exhausted T cells were significantly increased in WT T cells versus Gal-3^−/− T cells (Figure 3B). We also found that the expression of T-box expressed in T cells (T-bet; a transcription factor for cytotoxic T cells and Th1 cells),³⁴ GM-CSF (a cytotoxic cytokine involved in aGvHD severity and lethality),^35,36 and IFN-γ (a cytotoxic cytokine involved in GI toxicity)³⁷ was significantly increased in Gal-3^−/− T cells versus WT T cells (Figure 3C). The frequencies of Forkhead box P3 (Foxp-3)-, IL-17a-, and granzyme B-positive T cells were not different in WT and Gal-3^−/− T cells after allo-HCT (Figure S3B). T cell populations in naive WT and Gal3^−/− mice exhibited comparable percentages of T-bet-, Foxp-3-, and IFN-γ-expressing cells (Figure S1B). We measured plasma levels of cytokines after allo-HCT to correlate with immune cell cytotoxicity. Plasma levels of GM-CSF and tumor necrosis factor (TNF)-α on day 7 and of IFN-γ and IL-6 on days 7 and 14 increased significantly, while plasma levels of IL-17 were not significantly different in recipients receiving Gal-3^−/− T cells compared with WT T cells (Figure 3D). These results suggest that Gal-3 significantly suppresses the cytotoxic phenotype of allogeneic T cells by inducing exhaustion.

Figure 3. Expression of Gal-3 in T cells increases T cell exhaustion during severe aGvHD following allo-HCT.

(A) Immune-related gene expression in WT or Gal-3^−/− CD4⁺ T cells sorted from spleens at day 7 after allo-HCT in the B6 → BALB/c model with mice transplanted with TCD-BM combined with 0.7 × 10⁶ C57BL/6 WT or Gal-3^−/− pan T cells using Nanostring nCounter immunology panel mRNA microarray (splenocytes were pooled from 5 or 6 mice before CD4⁺ T cell sorting). The relative gene expression is shown for WT (black) and Gal-3^−/− (red) T cells.

(B) PD-1⁺ Lag-3⁺, Slamf6⁺ Tim-3⁻ (progenitor exhausted T cells), and Slamf6⁺ Tim-3⁺ (terminally exhausted T cells) frequencies in splenic T cells within single live H-2k^b+ H-2k^d− CD45⁺CD3⁺ cells were analyzed 7 days after allo-HCT. Exhausted T cells decreased in recipients of Gal-3^−/− T cells. Representative plots are shown from two independent experiments (n = 10 in total).

(C) T-box expressed in T cells (T-bet)-, granulocyte-macrophage colony-stimulating factor (GM-CSF)-, and interferon (IFN-γ-positive frequencies in splenic T cells within single live H-2k^b+ H-2k^d− CD45⁺CD3⁺ cell populations at day 7 after allo-HCT. Data were pooled from two individual experiments to obtain a total of n = 10 per group. The expression of T-bet, GM-CSF, and IFN-γ in T cells was significantly increased in Gal-3^−/− versus WT T cells.

(D) Plasma levels of inflammatory cytokines of irradiated BALB/c mice transplanted with TCD-BM plus WT or Gal-3^−/− T cells at 7 and 14 days. Plasma GM-CSF, IFN-γ, interleukin (IL)-6, and tumor necrosis factor (TNF)-α on day 7 all increased significantly in Gal-3^−/− T cell recipients. There were no differences in plasma IL-17 levels. Data were pooled from two individual experiments to obtain a total of n = 4–8 per group. For comparison of the means, an unpaired Mann-Whitney test was used in (B), (C), and (D); *p < 0.05, **p < 0.01, ***p < 0.001. Data are presented as individual values ± SD.

Gal-3 expression increased NFAT translocation in T cells

We sought to determine how Gal-3 induces T cell exhaustion. It has been shown that Gal-3 activates TREM2/DAP12 signaling. In turn, TREM2/DAP12 signaling, an important pathway in cell proliferation, activates phospholipase C, leading to a calcium influx that activates calcineurin, which leads to disinhibition of the nuclear import of nuclear factor of activated T cells (NFAT) and its activation.^38,39 On the other hand, NFAT activation (NFAT translocation from cytoplasm to nucleus) alone, without activation of activator protein 1 (AP-1), leads to T cell exhaustion.^40,41 We hypothesized that Gal-3 induces NFAT translocation without activation of AP-1 to promote T cell exhaustion. To test this hypothesis, splenic WT and Gal-3^−/− T cells were isolated and activated with anti-CD3 and anti-CD28 antibodies for 2 h in vitro. Using ImageStream flow cytometry, we found that NFAT translocation was significantly higher in WT T cells versus Gal-3^−/− T cells (Figure 4A). There were no significant differences in c-Jun translocation (Figures 4C and S4C). Further analysis showed that the ratio of NFAT:c-Jun translocation was significantly higher in WT T cells versus Gal-3^−/− T cells (Figure 4D). To confirm these findings in vivo, lethally irradiated BALB/c recipients were transplanted with TCD-BM and T cells from B6 WT or Gal-3^−/− donor mice. On day 7, splenic T cells were stained, and NFAT translocation and phosphorylated c-Jun (as a surrogate for AP-1 activation) were analyzed using ImageStream and flow cytometry. NFAT translocation was significantly greater in WT T cells compared with Gal-3^−/− T cells (Figure 4B), whereas there was no difference in AP-1 activation between WT and Gal-3^−/− T cells (Figure 4C). To further investigate the role of Gal-3 on T cell exhaustion, lethally irradiated BALB/c recipients were transplanted with TCD-BM and T cells from B6 WT or Gal-3^−/− donor mice. At day 7 after allo-HCT, donor T cells were sorted from the spleen and re-stimulated with CD3 and CD28 for 3 days. Then, the levels of IL-2, IFN-γ, and phosphorylated c-Jun (p-c-Jun) were evaluated using flow cytometry. The data showed that the expression of IL-2 and IFN-γ was significantly higher in Gal-3^−/− T cells compared with WT T cells (Figure S4A), but there was no difference in AP-1 activation between WT and Gal-3^−/− T cells (Figure S4B). These findings suggest that Gal-3 induces anergy in T cells and decreases the T cell response to re-stimulation possibly through excessive NFAT translocation. Modulation of TCR signaling is another possible mechanism for Gal-3 to induce T cell anergy. Previous work demonstrated that Gal-3 induces anergy in tumor-associated lymphocytes by inhibiting TCR co-localization with CD8.⁴² Further studies are needed to discover the mechanisms for how Gal-3 regulates NFAT activation and subsequently T cell anergy and exhaustion.

Figure 4. Gal-3 drives T cell exhaustion through NFAT translocation.

(A) T cells were isolated from WT or Gal-3^−/− spleens. Isolated T cells were cultured and activated in the presence of CD3/C28 beads for 2 h, and NFAT translocation was quantitated with ImageStream. Data were pooled from three individual experiments with two replicates in each experiment (n = 6 in total). NFAT translocation was significantly higher in WT T cells versus Gal-3^−/− T cells.

(B) Lethally irradiated BALB/c mice were transplanted with TCD-BM plus WT or Gal-3^−/− T cells. On day 7, NFAT translocation in T cells was quantitated using ImageStream. A significantly higher amount of nuclear translocation was seen in WT compared with Gal-3^−/− CD4⁺ and CD8⁺ T cells. Data were pooled from two individual experiments with four mice per group in each experiment (n = 8 in total).

(C) T cells were isolated from WT or Gal-3^−/− spleens. Isolated T cells were cultured and activated in the presence of CD3/C28 beads for 2 h, and c-Jun translocation, as a surrogate for AP-1 signaling, was quantitated with ImageStream. Data were pooled from three individual experiments with two replicates in each experiment (n = 6 in total). There was no significant difference in c-Jun translocation in WT T cells versus Gal-3^−/− T cells.

(D) The rate of NFAT signaling to AP-1 signaling was calculated by dividing NFAT Rd factor into c-Jun Rd factor. The ratio of NFAT:c-Jun translocation was significantly higher in WT T cells versus Gal-3^−/− T cells. Data were pooled from three individual experiments with two replicates in each experiment (n = 6 in total). For comparison of the means, an unpaired Mann-Whitney test was used in (A–D); *p < 0.05, **p < 0.01, ***p < 0.001. Data are presented as individual values ± SD.

Gal-3 overexpression in human T cells ameliorates aGvHD in a humanized mouse model

We have shown that Gal-3 expression in donor T cells after allo-HCT suppresses the T cell cytotoxic phenotype, decreases T cell proliferation, and ameliorates aGvHD severity and lethality. We studied whether Gal-3 overexpression in T cells can control aGvHD. Overexpression of the Gal-3 protein in human T cells was conducted using retroviral transduction with greater than 85% transduction efficiency (Figure S7). To evaluate the effect of Gal-3 overexpression on aGvHD severity, sub-lethally irradiated NSG-HLA-A*02 (A2) mice were transplanted with control T cells, T^RV-GFP cells, or T^RV-Gal-3 cells from a non-HLA-A2 donor, and aGvHD severity and survival were evaluated. Gal-3 overexpression significantly decreased aGvHD severity and improved survival compared with control groups (Figure 5A). To evaluate the in vivo T cell phenotype after transfer, sub-lethally irradiated NSG-HLA-A2 mice were transplanted with control T cells, T^RV-GFP cells, or T^RV-Gal-3 cells. On day 7 after allo-HCT, T cells were harvested from lung tissue (the main target organ in the NSG model). T cell phenotypes were analyzed using multicolor Cytek Aurora flow cytometry. Gal-3 overexpression significantly increased exhausted and central memory CD4⁺ T cells and decreased the percentage of effector CD4⁺ T cells and regulatory CD4⁺ T cells as seen in t-distributed stochastic neighbor embedding (tSNE) and individual value plots (Figure 5B). There were also significant decreases in effector and memory CD8⁺ T cells transplanted with T^RV-Gal-3 cells as seen in the tSNE plots (Figure 5C). These data suggest that Gal-3 overexpression in T cells ameliorates aGvHD, increases T cell exhaustion, and modulates T cell differentiation.

Figure 5. Gal-3 overexpression in T cells ameliorates the severity and fatality of acute GvHD in a humanized NSG model.

(A) Clinical GvHD score and survival of lethally irradiated NSG mice after allo-HCT with 2 × 10⁶ normal human T cells (control), T cells with retroviral transduction GFP only (retroviral-only control) (T^RV-GFP cells), or T cells with retroviral transduction overexpressing Gal-3 (T^RV-Gal-3 cells). Data were pooled from two individual experiments, each with n = 8–10 per group to obtain a total of n = 16–20 per group. Gal-3 overexpression significantly decreased aGvHD severity compared with controls.

(B) tSNE representation of frequencies of exhausted T cells, central memory cells, effector cells, and regulatory T cells in CD4⁺ T cells within single live CD45⁺CD3⁺ populations from the lungs (target organ in this model) of mice 14 days after allo-HCT obtained by multicolor Cytek Aurora flow cytometry. Frequencies of IFN-γ⁺CD4⁺ T cells were gated from single live CD45⁺CD3⁻ populations. Data were pooled from two individual experiments, each with n = 2–3 per group to obtain a total of n = 5 per group. Gal-3 overexpression significantly increased exhausted and central memory CD4⁺ T cells and decreased the percentage of effector CD4⁺ T cells and regulatory CD4⁺ T cells. There were also significant decreases in naive, effector, and memory CD8⁺ T cells. Two-way ANOVA with Tukey’s multiple comparisons test was used for clinical GvHD score in (A). For comparison of survival curves, a log-rank (Mantel-Cox) test was used in (A). One-way ANOVA with Tukey’s multiple comparisons test was used in (B); *p < 0.05, **p < 0.01, ***p < 0.001. Clinical score is shown as means ± SD. Other data are presented as tSNE plots and individual values ± SD.

Gal-3 expression in T cells is not required for the graft versus tumor effect

To investigate whether T cell Gal-3 affected the GvT effect in the MHC-mismatched HCT model, GFP⁺luciferase⁺ A20 B cell lymphoma cells (A20-GFP-Luc, 1.0 × 10⁵ cells) were injected into lethally irradiated BALB/c recipients 4 h before HCT with WT B6 TCD-BM and low-dose B6 WT or Gal-3^−/− splenic T cells (2 × 10⁵ T cells). Tumor growth was monitored by bioluminescence imaging (BLI). Recipients of A20 cells and TCD-BM alone died before day 24 (Figures 6A and 6B). Tumor growth was controlled in recipients receiving WT or Gal-3^−/− T cells, resulting in significant survival improvement (Figure 6B). Because of the low number of donor T cells, recipient mice did not develop severe GvHD (Figure 6B) but weight loss was significantly higher in mice that received Gal-3^−/− T cells (Figure 6B). To reduce GvHD severity to a minimal level, we further decreased the dose of T cells (from 2 × 10⁵ to 1 × 10⁵). Even at this low dose of T cells, both WT and Gal-3^−/− controlled tumor growth (Figure S5), suggesting that Gal-3 expression at the basal level does not compromise the GvT effect. NKG2D expression on T cells has been associated with an increased GvT effect with minimal effect on GvHD severity.^43,44 We tested whether Gal-3 changes the expression of NKG2D on T cells during the induction of the GvT effect. The expression of NKG2D significantly increased in WT effector cell populations compared with Gal-3^−/− T cells (Figure 6C).

Figure 6. Gal-3 does not compromise the GvT effect.

(A and B) Lethally irradiated BALB/c mice were intravenously injected with 1 × 10⁵ A20 cells followed by allo-HCT with 3.5 × 10⁶ TCD-BM with or without 2 × 10⁵ B6 WT or Gal-3^−/− pan T cells after 4 h. Experiments with A20 tumor cells expressing luciferase (A20 cells) were monitored by in vivo bioluminescence imaging (BLI). Recipients of A20 cells and TCD-BM alone died before day 24. Recipients of A20 cells, TCD-BM, and T cells displayed controlled tumor growth and improved survival, with less GvHD severity when receiving WT T cells compared with Gal-3^−/− T cells (n = 4–8).

(C) BALB/c mice received A20 cells followed by allo-HCT with BM alone or with 5 × 10⁵ T cells. CD44⁺ NKG2D⁺ in T cell frequencies within splenic single live H-2k^b+ H-2k^d− CD45⁺CD3⁺ cell populations on day 14 after transplant were analyzed by flow cytometry. Data were pooled from two individual experiments, each with n = 4 per group to obtain a total of n = 8 per group.

(D) Gal-3 was overexpressed in human T cells (HLA-A2) using retroviral transduction. T cells were co-cultured with luciferase-positive Raji tumor cells (HLA-A2 negative) for 48 h. After 48 h, luciferin was added to the medium and luciferin intensities were measured in different groups using an IVIS Spectrum in vivo imaging system instrument.

(E) Gal-3 was overexpressed in human T cells (HLA-A2) using retroviral transduction. T cells were co-cultured with THP-1 tumor cells (HLA-A2 negative) for 48 h. After 48 h, the levels of apoptosis (annexin V staining) were measured in different groups using flow cytometry. (D and E) Data were pooled from two independent experiments with three repeats per group. Two-way ANOVA with Tukey’s multiple comparisons test was used for body weight and tumor burden difference in (A) and (B). Log-rank (Mantel-Cox) test was used to compare survival curves for (B). For comparison of the means, an unpaired Mann-Whitney test was used in (C); *p < 0.05. For comparison of the means, an unpaired two-tailed t test was used in (D) and (E). Tumor burden, body weight, clinical score, and survival data are shown as means ± SEM. In vitro data are shown as means ± SD. T cell frequencies are presented as individual values ± SD. Representative flow plots (C) and (E) and IVIS images (D) are shown.

To investigate the ability of T^RV-Gal-3 relative to T^RV-GFP cells to kill tumor cells, two different cell lines with HLA disparity (Raji and THP-1) were individually co-cultured with harvested human T cells in vitro. Raji cells were luciferase positive, allowing us to measure the number of tumor cells co-cultured with T cells based on the luminescence signal after luciferin was added to the medium. T cells and Raji cells were cultured for 48 h at different T cell-to-tumor ratios ranging from 1:10 to 1:2. There was no difference in the ability of T^RV-Gal-3 cells to kill Raji tumor cells compared with T^RV-GFP cells (Figure 6D). We also found that Gal-3 overexpression did not compromise the anti-tumor effect of T^RV-Gal-3 cells when co-cultured with the THP-1 tumor cells (Figure 6E). Thus, Gal-3 overexpression in T cells did not impair the T cell anti-tumor effect either in vivo or in vitro, possibly by inducing the expression of NKG2D by effector cells.

The intensity of Gal-3 expression in T cells is associated with lower GI aGvHD pathology scores in patients after allo-HCT

We showed that Gal-3 overexpression decreased aGvHD severity and lethality in a humanized model of GvHD. We next asked if Gal-3 expression by GI T cells correlated with changes in colonic inflammation or evidence of aGvHD pathology in allo-HCT patients undergoing colonoscopies for aGvHD evaluation. The histopathology scoring system of Lerner et al. was utilized.⁴⁵ Colon biopsies from 15 patients with aGvHD pathology scores of 3–4 (severe) and 20 patients with pathology scores of 0–1 (none to mild) were selected and evaluated in a de-identified manner. Colon biopsies with pathology grade 2 were excluded. The biopsies were stained for T cell (CD4 and CD8), Gal-3, and AE1/AE3 (an epithelial marker; pan cytokeratin) and evaluated using the multispectral Vectra Polaris automated quantitative pathology imaging system (Figure S8A). T cells in the colon biopsies were found to express Gal-3. There was a significant association between the percentage of stromal CD8⁺ Gal-3⁺ T cells, but not CD4⁺ Gal-3⁺ T cells, and GI pathology score (Figures S8B and S8C). When analyzing Gal-3 intensity on Gal-3-expressing T cells, we found a significant association between Gal-3 mean fluorescence intensity (MFI) in CD4⁺ T cells and the GI pathology score (Figure S8B). The Gal-3 MFI in CD4⁺ T cells was significantly lower in the epithelium and stroma in biopsies from patients with higher pathology scores (3–4) compared with biopsies from patients with lower scores, 0–1 (Figure S8B). There was a similar trend, but not statistically significant, in the Gal-3 MFI in CD8⁺ T cells in the stroma (Figure S8C). These clinical histopathologic data further suggest the role of Gal-3 expression in T cells, in both CD4⁺ and CD8⁺ T cells, in constraining aGvHD and reducing severity.

DISCUSSION

In this study, we have demonstrated a critical role for donor T cell Gal-3 expression in the development and progression of aGvHD in MHC- and miHA-disparate transplant models. Gal-3 expression was significantly increased on donor T cells and this was associated with decreased aGvHD severity and lethality. We further demonstrated that Gal-3 upregulation in donor T cells also drives exhaustion. Using a humanized model of aGvHD, overexpression of Gal-3 in human T cells dramatically decreased aGvHD lethality and, importantly, without affecting the GvT effect. Furthermore, lower Gal-3 expression in T cells examined in GI colon histopathologic biopsies from allo-HCT patients was significantly associated with more severe aGvHD histopathologic changes. Gal-3 expression in donor T cells, but not in inactivated T cells, suggests that TCR signaling and co-stimulatory factors such as CD28 signaling might play a role in Gal-3 induction.

Gal-3 expression was upregulated in transplanted murine donor T cells, causing significant downregulation of T cell proliferation and cytotoxicity. In the absence of Gal-3, murine donor T cells proliferated more rapidly, shifting toward an IFN-γ-producing type 1 T cell (Th1 and Tc1) phenotype, consistent with other studies examining the aGvHD Th1 response.^46,47 These data are at variance with previous reports that Gal-3 deficiency in T cells favors a Th2 cell response in a preclinical model of Paracoccidioides infection.⁴⁸ These differences in the Gal-3 function in T cells may be attributed to different T cell responses to infections compared with inflammatory diseases such as inflammatory bowel disease (IBD)⁴⁹ or aGvHD. We have shown that activated T cells express Gal-3 and that deletion of Gal-3 in T cells resulted in increased proliferation and apoptosis. Gal-3 deletion also increased T cell IFN-γ production, confirming that Gal-3 suppresses T cell cytokine production.²⁵ Transplanted donor Gal-3^−/− T cells showed higher IFN-γ production resulting in elevated plasma levels. Previous work demonstrated that increased IFN-γ production results in AICD in T cells by promoting caspase 8-mediated apoptosis.^31,50,51 Interestingly, Gal-3 protects T cells against IFN-γ-mediated AICD. The Gal-3 anti-apoptotic effect is consistent with its known anti-apoptotic role in other cell types, including cancer cells.⁵² Gal-3 increases mitochondrial stability, induces BCL-2 expression, and prevents caspase 8 activation in chemotherapy-exposed prostate cancer cells, suggesting that enhancing Gal-3 expression would impair treatment response.⁵³ This demonstrates the complexity of Gal-3 functions: Gal-3 acts as a negative regulator of T cell function and proliferation while protecting T cells from apoptotic death. Within the donor T cells, these two competing pathways partially explain subtle but significant differences in aGvHD severity and lethality in murine models. Gal-3^−/− donor T cells are more cytotoxic, but their numbers decline after allo-HCT compared with WT donor T cells at later time points.

Increased cytokine production, primarily by infiltrating donor T cells in common target organs, such as the liver and GI tract, can drive aGvHD.^54,55 Conversely, inhibiting effector cytokines, including IFN-γ, IL-2, and GM-CSF, can ameliorate aGvHD.⁵⁶ IFN-γ induces cell death in the GI tract stem cells,⁸ while GM-CSF increases epithelial cell damage by inducing reactive oxygen species (ROS) and IL-1β production by donor T cells.²⁸ The increased aGvHD lethality induced by donor Gal-3^−/− T cells may be driven by excess production of inflammatory cytokines. Our results showed that a lower percentage of donor T cells produced inflammatory cytokines (IFN-γ and GM-CSF) when Gal-3 was present. This decrease in inflammatory cytokines may result in decreased epithelial damage in target organs such as GI-tract mucosal cells, as supported by the clinical histopathologic data. Thus, increased aGvHD severity and lethality in murine recipients of Gal-3^−/− T cells are associated with excessive inflammatory cytokine secretion. The percentage of regulatory T cells could also influence aGvHD severity, but we observed that their percentage did not change in murine Gal-3-deficient T cells. Also, overexpression of Gal-3 in human T cells did not increase the frequency of regulatory T cells (Tregs) in the humanized GvHD model, suggesting that Gal-3 in Tregs is not a critical factor in murine GvHD. Further studies are needed to characterize the immunosuppressive function of Gal-3^−/− T cells relative to inflammation to confirm that the protective role of Gal-3 during GvHD is independent of Treg function.

T cells may become exhausted following chronic antigen stimulation, leading to the T cell expression of inhibitory markers, including CTLA-4, Tim-3, Lag-3, TIGIT, and PD-1, and functional impairment, including decreased cytokine production.^20,57-60 T cell exhaustion decreases aGvHD lethality in murine models.^61,62 Increased donor T cell expression of PD-L1 is associated with more severe GvHD.⁶² Increased recipient cell expression of PD-L1 interacts with PD-1 on donor T cells, leading to T cell exhaustion and decreased aGvHD severity.^62,63 Consistent with these data, we determined that Gal-3 drives T cells to either progenitor exhausted or terminally exhausted states, resulting in the reduction of aGvHD lethality. Gal-3^−/− T cells had decreased CTLA-4 expression. This suggests that Gal-3 signaling may decrease GvHD severity through the CD28/CTLA-4 inhibitory pathway. It has been reported that blocking CTLA-4 with CTLA-4-Ig increases severe GvHD and decreases survival after allo-HCT.^60,64 We also found that Gal-3 facilitates NFAT translocation, and the interaction of NFAT with AP-1 is key for T cell activation. However, NFAT can induce T cell exhaustion by binding to promoter regions associated with the expression of inhibitory molecules independent of AP-1.⁴⁰ This results in the impairment of T cell function. Thus, these findings imply that Gal-3 can drive T cell exhaustion through excess activation of NFAT signaling. It has been reported that NFAT deficiency, or blocking of NFAT signaling, ameliorates aGvHD lethality.⁶⁵ It appears that a basal level of NFAT signaling is necessary for T cell activation, whereas excessive NFAT signaling induces T cell exhaustion. Gal-3 can activate Notch signaling⁶⁶ and may induce T cell exhaustion by engaging T cell notch signaling⁶⁷ as another mechanism to drive T cell exhaustion.

A major goal of allo-HCT is to maximize the alloreactive immune response against tumor cells while minimizing alloreactive damage to normal host tissues. Gal-3 expression in donor T cells did not impair the GvT effect in the A20 lymphoma model. Gal-3 was associated with effector cell expression of NKG2D, which plays an important role in balancing GvT activity and aGvHD.^43,44,68,69 In agreement with our data, a recent study has shown that NKG2D expressed by TCR-γδ T cells plays an important role in anti-tumor effects.⁷⁰ Also, adoptive transfer of myeloid-derived suppressor cells (MDSCs) can increase the GvT effect by inducing effector cell upregulation of NKG2D while preventing severe aGvHD.⁴⁴ Future studies will be necessary to clarify the mechanisms behind NKG2D expression induced by Gal-3 and determine the dynamics of its expression in aGvHD after allo-HCT. This is especially pertinent because it has been shown that NKG2D increases aGvHD severity in the days immediately after allo-HCT, yet becomes critical for maintaining GvT effects as time progresses.⁷¹ We evaluated the effect of T cell Gal-3 on the GvT effect using both in vivo and in vitro systems, but future in vivo studies are needed to understand the role of T cell Gal-3 in the GvT effect by depleting or overexpressing Gal-3 in human T cells. The limitations of our work include the need to study the potentially dynamic nature of Gal-3 expression in T cells residing in GI tissue over long time periods in host and donor T cells. Recently, Strobl et al.⁷² showed that host T cells also persist in aGvHD target organs and they express Gal-3. Further studies are also needed to evaluate the role of Gal-3 expression by host epithelial cells in donor T cell differentiation and function. Previous reports have shown that epithelial cell or tumor cell Gal-3 expression regulates T cell activity.^73,74 Gal-3 also changes myeloid differentiation toward M2 macrophages⁷⁵ and is associated with tumor progression²³; thus, induction of Gal-3 by small-molecule modulating agents should be studied in the broader context to prevent possible tumor relapse.

In summary, we have identified donor T cell expression of Gal-3 as a suppressor of aGvHD in murine models and associated decreased Gal-3 expression with histopathologic severity in GI colonic tissue from allo-HCT patients. Donor T cells express Gal-3 during the development of aGvHD, resulting in the modulation of T cell proliferation, function, and apoptosis, ultimately reducing aGvHD lethality. These factors suggest that modulating Gal-3 expression in donor T cells may allow clinicians to manage the balance between aGvHD and the GvT effect.

Limitations of the study

This study has the following limitations. (1) To demonstrate that overexpression of Gal-3 in T cells leads to decreased GvHD while retaining GvT activity, this study utilized two independent experiments. First, allo-HCT using genetic deletion of Gal-3 resulted in increased GvHD and no effect on GvT activity compared with WT T cells. Second, overexpression of Gal-3 led to decreased GvHD. However, due to technical limitations, the GvT activity in Gal-3-overexpressing T cells could not be addressed in this study. It therefore remains a possibility that overexpression of Gal-3 could decrease GvT activity, and the optimal Gal-3 expression dosage and/or timeline remains unknown. (2) In addition, this study focused on the impact of Gal-3 expression on T cell activity and exhaustion to reduce GvHD. However, an alternative mechanism for the observed GvHD reduction is decreased migration of WT donor T cells to GvHD target organs. While we detected changes in chemokine receptors in Gal3^−/− compared with WT T cells, this study did not focus on whether this was involved in the increased GvHD in mice receiving Gal3^−/− T cells. (3) In this study we mainly utilized mice with a genetic deletion of Gal-3, which lack both nuclear and cytoplasmic forms of Gal-3 in T cells. Due to this technical limitation, it is not clear what form of Gal-3 is activating NFAT signaling or protecting T cells from apoptosis. (4) We used ImageStream to study NFAT and AP-1 translocation, but stronger methods such as nuclear fraction western blot could be used as alternative approaches for translocation analysis. (5) Importantly, this study focused on the role of Gal-3 in NFAT translocation to induce T cell exhaustion. However, alternative mechanisms of T cell exhaustion by Gal-3, such as TCR signaling blockade, should be addressed. (6) In addition, future studies should address the signaling cascade involved in inducing Gal-3 expression in T cells post-HCT, the mechanism by which Gal-3 regulates T cell apoptosis after T cell re-activation, and the signaling mechanism by which Gal-3 increases NFAT translocation and activation.

STAR★METHODS

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Hemn Mohammadpour (hemn.mohammadpour@roswellpark.org).

Materials availability

This study did not generate new unique reagents

Data and code availability

• This paper analyzes existing, publicly available data. These accession numbers for the datasets are listed in the key resources table. The Raw T cell mRNA microarray data have been deposited at Figshare and GEO and are publicly available as of the date of publication. DOIs and accession numbers are listed in the key resources table.

• This paper does not report original code.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

KEY RESOURCES TABLE

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
α-mouse CD8	BD Pharmingen™	Clone 53-6.7 Catalog # 553030 RRID: AB_394568
α-mouse CD4	BD Pharmingen™	Clone GK1.5 Catalog # 553730 RRID: AB_396634
α-mouse CD3	BD Pharmingen ™	Clone 17A2 Catalog # 560590 RRID: AB_1727461
α-mouse CD45	BioLegend™	Clone 30-F11 Catalog # 103114 RRID: AB_312979
α-mouse H2b	BioLegend™	Clone AF6-88.5 Catalog # 116516 RRID: AB_1967133
α-mouse H2b	BioLegend™	Clone AF6-88.5 Catalog # 116520 RRID: AB_2721684
α-mouse H2d	BioLegend™	Clone SF1-1.1 Catalog # 116606 RRID: AB_313741
α-mouse RORγT	BD Pharmingen ™	Clone Q31-378 Catalog # 562894 RRID: AB_2687545
α-mouse T-bet	BD Pharmingen ™	Clone 4B10 Catalog # 561264 RRID: AB_10563424
α-mouse Foxp-3	BioLegend™	Clone 150D Catalog #320008 RRID: AB_492980
α-mouse CD45.1	Invitrogen™	Clone A20 Catalog # 12-0453-81 RRID: AB_465675
α-mouse CD45.2	Invitrogen™	Clone 104 Catalog # 17-0454-81 RRID: AB_469399
α-mouse IFN-γ	BD Horizon ™	Clone XMG1.2 Catalog # 563376 RRID: AB_2738165
α-mouse IL-17A	BioLegend™	Clone TC11-18H10.1 Catalog # 506908 RRID: AB_536010
α-mouse IL-10	BioLegend™	Clone JES5-16E3 Catalog # 505008 RRID: AB_315362
α-mouse GM-CSF	Invitrogen™	Clone MP1-22E9 Catalog # 12-7331-82 RRID: AB_466205
α-mouse CD16/32 Block	BD Pharmingen™	Clone 2.4G2 Catalog # 553142 RRID: AB_394656
α-mouse CD44	BioLegend™	Clone IM7 Catalog # 103012 RRID: AB_312963
α-mouse CD62L	BioLegend™	Clone MEL-14 Catalog # 104406 RRID: AB_313093
α-mouse Lag-3	BioLegend™	Clone C9B7W Catalog # 125221 RRID: AB_2572080
α-mouse Lag-3	BioLegend™	Clone C9B7W Catalog # 125210 RRID: AB_10639727
α-mouse Tim-3	BioLegend™	Clone B8.2C12 Catalog # 134004 RRID: AB_1626177
α-mouse Tim-3	BioLegend™	Clone B8.2C12 Catalog # 134010 RRID: AB_2632734
α-mouse Salmf6	BioLegend™	Clone 330-AJ Catalog #134610 RRID: AB_2728155
α-mouse PD-1	BioLegend™	Clone 29F.1A12 Catalog #135231 RRID: AB_2566158
α-mouse PD-1	BioLegend™	Clone 29F.1A12 Catalog #135220 RRID: AB_2562616
α-mouse/human Galectin-3	Biolegend™	Clone eBioM3/38 Catalog # 12-5301-82 RRID: AB_842792
α-human CD8	BD Horizon ™	Clone RPA-T8 Catalog # 751518 RRID: AB_2875513
α-human CD4	Invitrogen™	Clone S3.5 Catalog # MHCD0430 RRID: AB_10371907
α-human CD3	BD OptiBuild ™	Clone SK7 Catalog # 740202 RRID: AB_2739952
α-human CD45RA	BD Horizon™	Clone HI100 Catalog #568712 RRID: NA
α-human CD45RO	BD OptiBuild	Clone UCHL1 Catalog # 748367 RRID: AB_2872786
α-human IFN-γ	BioLegend™	Clone 4S.B3 Catalog # 502539 RRID: AB_11218602
α-human IL-17	BD Pharmingen™	Clone N49-653 Catalog # 560491 RRID: AB_1645418
α-human IL-10	BD Horizon ™	Clone JES-9D7 Catalog # 566276 RRID: AB_2738566
α-mouse Foxp-3	BD Pharmingen™	Clone 150D Catalog # 560401 RRID: AB_1645201
α-mouse Foxp-3	BioLegend™	Clone 150D Catalog # 320008 RRID: AB_492980
α-human PD-1	BioLegend™	Clone EH12.2H7 Catalog # 329920 RRID: AB_10960742
α-human CXCR-3	BioLegend™	Clone G025H7 Catalog # 353708 RRID: AB_10983064
α-human CCR7	BioLegend™	Clone G043H7 Catalog # 353230 RRID: AB_2563630
α-human TIGIT	BioLegend™	Clone A15153G Catalog #372716 RRID: AB_2632931
α-human Tim-3	BioLegend™	Clone F38-2E2 Catalog # 345026 RRID: AB_2565717
α-human CD27	BioLegend™	Clone M-T271 Catalog # 356410 RRID: AB_2561957
α-human CD127	BD Horizon ™	Clone HIL-7R-M21 Catalog # 565185 RRID: AB_2739099
α-human Lag-3	BioLegend™	Clone 11C3C65 Catalog # 369320 RRID: AB_2716125
Anti-mouse IFN-γ (Armenian hamster IgG)	BioXcell	Clone H22 Catalog # BE0312 RRID: AB_2736992
Isotype (Armenian hamster IgG)	BioXcell	Clone BE0091 Catalog # BE0091 RRID: AB_1107773
Ultra-LEAF™ Purified anti-mouse CD3ε Antibody	BioLegend™	Clone 145-2C11 Cat# 100340 RRID: AB_11149115
Ultra-LEAF™ Purified anti-mouse CD28 Antibody	BioLegend™	Clone 37.51 Cat# 102116 RRID: AB_11147170
Ultra-LEAF™ Purified anti-human CD3 Antibody	BioLegend™	Clone OKT3 Cat# 317326 RRID: AB_11150592
Ultra-LEAF™ Purified anti-human CD28 Antibody	BioLegend™	Clone CD28.2 Cat# 302934 RRID: AB_11148949
Biological Samples
allo-HCT patient colon biopsy	Roswell Park Comprehensive Cancer Center	N/A
Peripheral blood mononuclear cells from healthy donors	Roswell Park Database and Biorepository	N/A
Chemicals, peptides, and recombinant proteins
Recombinant Mouse IFN-γ (carrier-free)	BioLegend™	Cat# 575304
Cell Activation Cocktail (with Brefeldin A)	BioLegend™	Cat# 423303
Brefeldin A	BioLegend™	Cat# 420601
eBioscience™ Cell Proliferation Dye eFluor™ 670	Invitrogen™	Cat# 65-0840-85
Percoll PLUS density gradient media	Cytiva	Cat# 17544502
Critical commercial assays
“The Big Easy” EasySep Magnet	StemCell Technologies	Cat# 18001
FOXP3 Fix/Perm Kit	BioLegend™	Cat# 421401
CellTrace™ Violet Cell Proliferation Kit	Invitrogen™	Cat# C34557
LIVE/DEAD™ Fixable Aqua Dead Cell Stain Kit	Invitrogen™	Cat# L34957
Mouse CD90.2 MicroBeads	Miltenyi Biotec	Cat# 130-121-278
Mouse Pan T cell isolation Kit II	Miltenyi Biotec	Cat# 130-095-130
FITC-Annexin V Apoptosis Detection Kit	BioLegend™	Cat# 640922
CaspGLOW™ Fluorescein Active Caspase-8 Staining Kit	Invitrogen™	Cat# 88-7005-42
BD Fixation/Permeabilization Kit	BD Biosciences	Cat # 554714
Deposited data
Preclinical T cell RNA-sequencing dataset	ArrayExpress	Array Express: E-MTAB-5378, E-MTAB-5379, E-MTAB-5380, E-MTAB-5381
Raw and analyzed T cell microarray data	This paper	GEO: GSE225756 Figshare: https://doi.org/10.6084/m9.figshare.21970883, https://doi.org/10.6084/m9.figshare.21970886
Experimental models: Cell lines
Human: THP-1 -Luc cell line	Provided by Dr. Francisco Hernandez-Ilizaliturri (Roswell Park Cancer Institute)	N/A
Human: Raji -Luc cell line	Provided by Dr. Dr. Francisco Hernandez-Ilizaliturri (Roswell Park Cancer Institute)	N/A
Mouse: A20- Luc	Provided by Dr. Xuefang Cao (University of Maryland)	N/A
Experimental models: Organisms/strains
Mouse: C57BL/6J (B6)	The Jackson Laboratory	JAX# 000664
Mouse: Gal-3−/− (C57BL/6J)	Gal-3−/− generated by the lab of Dr. Fu-Tong Liu through Dr. Noorjahan Panjwani (Tufts University).	N/A
Mouse: BALB/cJ	The Jackson Laboratory	JAX# 000651
Mouse: C3H/SW	The Jackson Laboratory	JAX# 000438
Mouse: CD45.1 C57BL/6J (Pep boys)	The Jackson Laboratory	JAX# 002014
Mouse: NSG-HLA-A2	The Jackson Laboratory	JAX# 014570
Software and algorithms
GraphPad Prism_V9	Graphpad	https://www.graphpad.com/scientific-software/prism/
FlowJo™_V10	FlowJo	https://www.flowjo.com/solutions/flowjo

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Animals

B6 (H-2Kb), C3H.SW (H-2Kb), NSG-HLA-A2-HHD, and BALB/cJ (H-2Kd) mice were purchased from the Jackson Laboratory. The Galectin-3 knockout (Gal-3^−/−) B6 mice were provided by Dr. Fu-Tong Liu through Dr. Noorjahan Panjwani, Tufts University, Boston MA. Gal-3^−/− and WT littermates were used as donor T cell sources. All mice were maintained in the Roswell Park Comprehensive Cancer Center (Roswell Park) Lab Animal Shared Resource (LAR) at the standard temperature (~22°C) in specific-pathogen-free housing. Both male and female mice of age 6-8 weeks were used for experiments and were randomly assigned to experimental groups prior to irradiation. All in vivo experiments were performed following the Institutional Animal Care and Use Committee (IACUC)-approved protocol according to Roswell Park animal care guidelines.

Human samples

This study is a retrospective study using deidentified human biopsies from patients aged 40-80 years old that had allo-HCT. Samples were provided by Roswell Park as Formalin-Fixed Paraffin-Embedded (FFPE) tissue slides. Experiments were performed following the institutional approval (BDR123720) at Roswell Park. Due to the nature of the study, informed consent was not required. Peripheral blood mononuclear cells from healthy donors were obtained from the Roswell Park Database and Biorepository. Donors included male and female volunteers with an age range of 35-45 years. Informed consent was obtained prior to sample collection. Due to the limitation in number of patients eligible for the inclusion in the study, the effect of sex on the sample size and outcome was not studied.

Cell lines

Raji and THP-1 cell line were provided by Dr. Francisco Hernandez-Ilizaliturri (Roswell Park Cancer Institute). Both cell lines were tested and authenticated for mycoplasma using the Agilent Mycoplasma Plus PCR Primer Set before experimental use. Cells were maintained in media composed of RPMI-1640, 10% FBS and 1% penicillin & streptomycin. Cells were passaged using 75 cm² flasks at 37°C and 5% CO₂.

METHOD DETAILS

Study design

This study was designed to investigate the role of Gal-3 in donor T cells after allo-HCT and overexpression of Gal-3 as a possible strategy to control aGvHD without impairing GvT effects. We studied the potential role of T cell Gal-3 in aGvHD severity and lethality using multiple experimental mouse models of aGvHD by monitoring body weight, clinical and histopathological aGvHD scores, and recipient survival. We also investigated the role of Gal-3 in T cell proliferation, apoptosis, differentiation, exhaustion, and cytokine production using flow cytometry and multiplex enzyme-linked immunosorbent assay (ELISA). We overexpressed Gal-3 in human T cells to assess the role of donor T cell Gal-3 in a humanized murine model of GvHD. We also used a murine model of GvT and a luciferase-expressing Raji cell line and parental THP-1 cell line to assess the effect of Gal-3 on GvT activity using T cells overexpressing Gal-3. Finally, we studied Gal-3 expression on human T cells in clinical GI tract (colon) biopsies from patients undergoing allo-HCT. Vectra multispectral immunofluorescent (mIF) staining evaluated the possible correlation between T cell Gal-3 expression and histopathology scores in colon GI biopsies scores in de-identified manner. All in vitro and in vivo experiments were repeated independently in triplicate and duplicate, respectively.

Allogeneic hematopoietic cell transplantation (Allo-HCT) for aGvHD

For aGvHD studies, WT BALB/cJ (H-2Kd) or C3H.SW (H-2Kb) mice were used as hosts. BALB/c or C3H.SW mice were irradiated with a single dose of irradiation at 8.5 or 11 Gy (Gy), respectively, from a Cs-137 source on Day −1. One day after irradiation, mice received tail vein injections of 3.5 × 10⁶ T cell-depleted bone marrow (TCD-BM) cells from WT B6 mice with or without 0.7×0⁶ pan T cells isolated from the spleen of WT or Gal-3^−/− B6 (H-2kb) mice (B6→BALB/c model); or 3.5×10⁶ TCD-BM cells only or combined with 1.5×10⁶ pan T cells isolated from WT or Gal-3^−/− B6 mice (B6→C3H.SW model). Bone marrow was depleted of T cells using the Miltenyi CD90.2 isolation kit according to manufacturer’s protocols. Pan T cells were isolated from the spleen using the Miltenyi Pan-T Cell Isolation Kit according to manufacturer’s protocols. In some noted experiments, BALB/c mice received 8.5 Gy of irradiation and were transplanted with 3.5×10⁶ TCD-BM cells only or combined with either 0.2×10⁶ CD4⁺ T cells, 2×10⁶ CD8⁺ T cells or 0.2×10⁵ CD4⁺CD25⁻ T cells. The weight of host mice was recorded every three days. All mice were monitored daily for GvHD scoring (measuring Posture, Fur, Skin, and Mobility). In the aGvHD humanized model, NSG-A2 mice were irradiated with 2.5 Gy on day −1 and transplanted with different types of 2×10⁶ human T cells on day 0 (see Gal3-overexpression in primary human T cells for details).

Donor cell preparation

Donor BM cells were isolated from the femurs of B6 WT mice. TCD-BM was prepared using anti-CD90.2 microbeads (purity >92%) (Miltenyi). The percentage of T cells in TCD-BM after sorting was less than 2% (Figure S6). Donor T cells were purified from the spleens of B6 WT or Gal-3^−/− mice by negative selection using mouse pan T cell isolation kit II (Miltenyi) (purity >93%) (Figure S6). CD4⁺ T cells were purified using a pan T cell isolation kit II plus biotinylated anti-CD8 antibody or CD 8⁺ biotin plus CD25⁺ biotin antibodies, respectively (purity >90%). CD8⁺ T cells were isolated using the CD8⁺ T cell isolation kit.

Flow cytometric analysis

All antibodies and reagents for flow cytometry, recombinant proteins and depleting antibodies are listed in key resources table. The cells were pre-incubated with purified anti-mouse CD16/CD32 monoclonal antibodies for 10 min at 4°C to prevent nonspecific antibody binding. The samples were then incubated with surface staining antibodies for 30 min at 4°C. Fixable viability dye was used to separate live and dead cells. Intracellular transcription factor and cytokine staining was performed using the FoxP3/Transcription Factor Staining Buffer Set and the Fixation and Permeabilization Kit (Thermo Fisher Scientific), respectively. For cytokine staining, cells were in vitro-stimulated with phorbol myristate acetate (PMA; 50 ng/mL) and ionomycin (1 μg/mL; Sigma-Aldrich) in the presence of brefeldin-A (Biolegend) for 4 to 5 h prior to staining.

Luminex^® immunoassay

Blood was collected by retro-orbital bleeding on days 7 and 14 following allo-HCT for plasma isolation. Blood sample vials were placed on ice until all samples had been collected. Sample vials were centrifuged at 4°C for 20 min at 2000RPM. Plasma was collected and frozen at −80°C for subsequent analysis. Mouse cytokine and chemokine 11-plex (Millipore) was performed by the Roswell Park Flow and Image Cytometry Shared Resource, Luminex Division, per the manufacturer’s instructions.

Nanostring gene expression profiling analysis

At day 7 after allo-HCT, CD4⁺ T cells (live, H-2Kb⁺CD45⁺CD3⁺) were sorted from single-cell suspensions of spleens from BALB/c recipient mice for gene expression (Immunology Panel, Nanostring) using BD FACSAria (BD Bioscience) in the Roswell Park Flow and Image Cytometry Shared Resource. After sorting, the total CD4⁺ T cells ranged from 0.5×10⁶ to 1×10⁶ cells. Cells were immediately frozen in liquid nitrogen and then stored at −80°C or on dry ice. Samples were sent to the Roswell Park Genomic Shared Resource for gene expression profiling. Nanostring analysis was performed with the nCounter^® Analysis System from NanoString Technologies utilizing the nCounter Mouse Immunology Kit, which includes 561 immunology-related mouse genes according to manufacturer’s protocols.

Histopathology scoring

At 14 days after allo-HCT, BALB/c host mice transplanted with B6 TCD-BM plus either WT or Gal-3^−/− T cells were sacrificed. Large and small intestines were harvested, fixed in formalin overnight, sectioned, and stained with H&E. The intestinal tissue was examined using an established semi-quantitative scoring system by the study pathologist (JQ) in a de-identified and blinded manner based on established scoring system.^12,28 The scoring system table has been added as Table S1.

Gal-3 overexpression in primary human T cells

The coding region of the human Gal-3 gene was PCR-amplified from cDNA of ovarian tumor tissue. Monomeric green fluorescent protein (mEGFP) gene was amplified from a plasmid mEGFP-N1 (a gift from Michael Davidson; Addgene plasmid # 54767). Gal-3 and mEGFP genes were genetically fused via the P2A-skipping site and cloned in a murine stem cell virus (MSCV)-based retroviral vector⁷⁶ using NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs). As a control, mEGFP alone was cloned in the same retroviral vector. The sequence of the inserts was confirmed by the Sanger sequencing in the Roswell Park Genomics Shared Resource. Retrovirus-producing PG13 cell lines (ATCC^® CRL-10686) were generated as described.⁷⁶ PG13 clones that produced high-titer viruses were established by limiting dilution cloning.

Peripheral blood mononuclear cells (PBMC) from healthy donors, provided by Roswell Park Database and Biorepository, were pre-activated by 10 μg/mL phytohemagglutinin (Remel) for 48 h in RPMI-1640 medium supplemented with 10% FBS, penicillin, streptomycin, and L-glutamine in the presence of rhIL-2 (10 U/mL; Sigma) and rhIL-7 (10 ng/mL; R&D Systems). After 40-48 h, activated T cells were harvested and cultured in a plate that was pre-coated with 10 μg/mL Retronectin (TaKaRa Bio) and 5 μg/mL anti-CD3 monoclonal antibody (OKT3, BioLegend) and loaded with retroviruses in the presence of rhIL-2 and rhIL-7. After 24 h, T cells were harvested and added in a new plate that was pre-coated with Retronectin without anti-CD3 monoclonal antibody and loaded with retroviruses. Transduced T cells were expanded for an additional 3-4 days in the presence of rhIL-2 and rhIL-7 before in vitro characterization or in vivo injection.

ImageStream

WT and Gal-3^−/− T cells were isolated from healthy murine spleens using pan T cell isolation kit (Miltenyi). Isolated T cells were activated with a T cell activation kit (Miltenyi; CD3/CD28 antibodies conjugated to the beads). After two hours of activation, cells were stained for surface markers for 20 min, washed, and fixed in 2% formaldehyde in 1x PBS (40 min), the cells were then permeabilized in 5% normal mouse serum in 1x PBS +0.4% Triton X-100. Next, the cells were stained for intracellular NFAT for 20 min using a rabbit anti-human NFAT1 antibody. After washing once with NGS block (5% normal goat serum in 1X PBS), the cells were incubated with 2 μg/mL Alexa Fluor 647-conjugated donkey anti-rabbit IgG for 15 min. Before running samples on the ImageStream instrument (Amnis^® ImageStream^®XMark II), 10 μL of 0.5 μg/mL DAPI (nuclear stain) was added to each sample and incubated at room temperature for at least 5 min.

Vectra multispectral immunofluorescent (mIF) staining

Clinical colon biopsy specimens that had been evaluated and scored for aGvHD by the Roswell Park Pathology department were selected for low pathologic grade (0-1) and high pathology grade (3-4).⁴⁵ Slides were chosen and analyzed independently in a de-identified and blinded manner. Formalin-fixed Paraffin-embedded (FFPE) 4 mm sections were cut and placed on charged slides. Slides were dried at 65°C for 2 h. After drying, the slides were placed on the BOND RXm Research Stainer (Leica Biosystems) and de-paraffinized with BOND Dewax solution (AR9222, Lecia Biosystems). The mIF staining process involved serial repetitions of the following for each biomarker: epitope retrieval/stripping with ER1 (citrate buffer pH 6, AR996, Leica Biosystems) or ER2 (Tris-EDTA buffer pH 9, AR9640, Leica Biosystems), blocking buffer (AKOYA Biosciences), primary antibody, Opal Polymer HRP secondary antibody (AKOYA Biosciences), and Opal Fluorophore (AKOYA Biosciences). All AKOYA reagents used for mIF staining come as a kit (NEL821001KT). Spectral DAPI (AKOYA Biosciences) was applied once slides were removed from the BOND. They were cover slipped using an aqueous method and Diamond anti-fade mounting medium (Invitrogen ThermoFisher). The mIF panel consisted of the following antibodies (clone, company, and opal fluorophores): TIM-3 (D5D5R, Cell Signaling, Opal Polaris 480), CD4 (4B12, Leica Biosystems, Opal 690), Galectin-3 (9C4, Leica Biosciences, Opal 520), CD8 (C8/144B, Agilent DAKO, Opal 570), LAG-3 (D2G4O, Cell Signaling, Opal 620), Pan Cytokeratin (AE1AE3, Agilent DAKO, Opal Polaris 780).

Tissue imaging and analysis

Slides were imaged on the Vectra^® Polaris Automated Quantitative Pathology Imaging System (AKOYA Biosciences). Further analysis of the slides was performed using inForm^® Software v2.4.11 (AKOYA Biosciences). Whole slide spectral un-mixing was achieved using the synthetic spectral library supplied within in-Form. From the un-mixed images, representative regions of interest (ROIs) were selected to train tissue and cell segmentation. Next, a unique algorithm was created using a machine learning technique in which the operator selects positive and negative cell examples for each marker. These algorithms were then batch applied across the entire tissue section for all samples in the project.

Study approval

All animal studies were reviewed and approved by the Roswell Park institutional animal care and use program and facilities (protocols 1140M and 1143M). All aspects of animal research and husbandry were conducted in accordance with the federal Animal Welfare Act and the NIH’s Guide for the Care and Use of Laboratory Animals (National Academies Press, 2011). Multi-color staining of GI tissue biopsies from patients with aGvHD were approved by Roswell Park Institutional Review Boards (IRB).

QUANTIFICATION AND STATISTICAL ANALYSIS

The body weight and clinical score data are presented as mean ± SEM. Other data are presented as median ±standard deviation. Differences between groups were analyzed using the unpaired Mann–Whitney U test and one-way analysis of variance (ANOVA) for two and more than two groups, respectively. Animal survival (Kaplan-Meier survival curves) was analyzed by log-rank test. The body weight and clinical score difference was analyzed using two-way ANOVA and presented by each time point. Multiple comparisons were assessed using Tukey’s multiple comparison test. To compare the Gal-3 MFI in CD4⁺ T cells between patients with pathology score 0-1 versus 3-4, a weighted two-way t-test was used. A p-value <0.05 was considered statistically significant. Statistical analyses were performed by using GraphPad Prism v7 (La Jolla, CA) and SAS v9.4 (Cary, NC).

Highlights.

• Activated donor T cells express Gal-3 after allo-HCT, leading to decreased activity

• Gal-3 activates NFAT signaling to induce T cell exhaustion

• Gal-3 overexpression results in decreased cytotoxic T cells in aGvHD models

• Gal-3 expression inversely correlates with GvHD severity in patients

INCLUSION AND DIVERSITY

We support inclusive, diverse, and equitable conduct of research.