Leukemia inhibitory factor protects against graft-versus-host disease while preserving graft-versus-leukemia activity

Abstract

Graft-versus-host disease (GVHD) remains a major complication after allogeneic hematopoietic stem cell transplantation, a widely used therapy for hematologic malignancies and blood disorders. Here, we report an unexpected role of cytokine leukemia inhibitory factor (LIF) in protecting against GVHD development. Administrating recombinant LIF protein (rLIF) protects mice from GVHD-induced tissue damage and lethality without compromising the graft-versus-leukemia activity, which is crucial to prevent tumor relapse. We found that rLIF decreases the infiltration and activation of donor immune cells and protects intestinal stem cells to ameliorate GVHD. Mechanistically, rLIF downregulates IL-12-p40 expression in recipient dendritic cells after irradiation through activating STAT1 signaling, which results in decreased major histocompatibility complex II levels on intestinal epithelial cells and decreased donor T-cell activation and infiltration. This study reveals a previously unidentified protective role of LIF for GVHD-induced tissue pathology and provides a potential effective therapeutic strategy to limit tissue pathology without compromising antileukemic efficacy.

Graphical abstract

Figure 1.

Administering rLIF ameliorates GVHD in mice. (A) Serum LIF levels in B6C3F1 (left) and C57BL/6 (right) mice at different days (D) after TBI (2 × 5.5 Gy), syngeneic BM + T, allogenic BM, and allogenic BM + T. n ≥ 3 mice/group (each dot represents a mouse). Serum LIF levels were measured by ELISAs. (B) The serum LIF levels at 2 days after allo-BMT were positively correlated with the survival length of mice after allo-BMT. B6C3F1 mice were exposed to 2 × 5.5 Gy TBI, followed by transplantation of BM and T cells from C57BL/6 mice. (C) Lethally irradiated B6C3F1 mice received BM alone (BM; n = 3), BM and T cells from C57BL/6 mice (BM + T; n = 13), or BM and T cells along with rLIF treatment (IP, 30 ng/g body weight, twice a day for the period as indicated) (BM + T + rLIF; n = 10 for D-4 to D3; n = 8 for D1 to D3 and D8 to D10). Schematic diagram of experimental procedures is shown on the upper left. Three panels shown are Kaplan-Meier survival curves (upper right), weight loss (lower left), and GVHD score (lower right) of mice after allo-BMT. (D) Lethally irradiated C57BL/6 mice received BM (n = 3), BM + T from BALB/c mice (n = 12), or BM + T + rLIF (n = 9). Mice were treated with rLIF as follows: IP, 30 ng/g body weight, twice a day for 7 days (D-4 to D3). Three panels shown are Kaplan-Meier survival curves (left), weight loss (middle), and GVHD score (right) of mice after allo-BMT. (E,F) Tissues from lethally irradiated B6C3F1 mice receiving C57BL/6 BM (n = 4), BM + T (n = 8) or BM + T + rLIF (n = 8). Tissues from these mice were collected at 14 days after allo-BMT for histopathologic analysis. (E) Representative hematoxylin and eosin–stained images of SI, colon, and liver tissues. (F) The clinical score was analyzed as described in Methods to reflect the histopathologic damage in different tissues. (G) Lethally irradiated C57BL/6 wild-type (n = 11) and LIF knockout (n = 7) mice received BM + T from BALB/c mice. Kaplan-Meier survival curves (upper) and GVHD score (lower) of mice after allo-BMT are presented. Data presented are from at least 3 independent experiments. (A,F) Data are presented as mean ± standard deviation; (C-D,G) data are presented as mean ± standard error of the mean. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, n.s.: not significant; unpaired t test with Welch’s correction (A,F), Spearman’s correlation (B), Kaplan-Meier survival analysis for survival, and analysis of variance for analysis of weight loss and GVHD score.

Figure 1.

Administering rLIF ameliorates GVHD in mice. (A) Serum LIF levels in B6C3F1 (left) and C57BL/6 (right) mice at different days (D) after TBI (2 × 5.5 Gy), syngeneic BM + T, allogenic BM, and allogenic BM + T. n ≥ 3 mice/group (each dot represents a mouse). Serum LIF levels were measured by ELISAs. (B) The serum LIF levels at 2 days after allo-BMT were positively correlated with the survival length of mice after allo-BMT. B6C3F1 mice were exposed to 2 × 5.5 Gy TBI, followed by transplantation of BM and T cells from C57BL/6 mice. (C) Lethally irradiated B6C3F1 mice received BM alone (BM; n = 3), BM and T cells from C57BL/6 mice (BM + T; n = 13), or BM and T cells along with rLIF treatment (IP, 30 ng/g body weight, twice a day for the period as indicated) (BM + T + rLIF; n = 10 for D-4 to D3; n = 8 for D1 to D3 and D8 to D10). Schematic diagram of experimental procedures is shown on the upper left. Three panels shown are Kaplan-Meier survival curves (upper right), weight loss (lower left), and GVHD score (lower right) of mice after allo-BMT. (D) Lethally irradiated C57BL/6 mice received BM (n = 3), BM + T from BALB/c mice (n = 12), or BM + T + rLIF (n = 9). Mice were treated with rLIF as follows: IP, 30 ng/g body weight, twice a day for 7 days (D-4 to D3). Three panels shown are Kaplan-Meier survival curves (left), weight loss (middle), and GVHD score (right) of mice after allo-BMT. (E,F) Tissues from lethally irradiated B6C3F1 mice receiving C57BL/6 BM (n = 4), BM + T (n = 8) or BM + T + rLIF (n = 8). Tissues from these mice were collected at 14 days after allo-BMT for histopathologic analysis. (E) Representative hematoxylin and eosin–stained images of SI, colon, and liver tissues. (F) The clinical score was analyzed as described in Methods to reflect the histopathologic damage in different tissues. (G) Lethally irradiated C57BL/6 wild-type (n = 11) and LIF knockout (n = 7) mice received BM + T from BALB/c mice. Kaplan-Meier survival curves (upper) and GVHD score (lower) of mice after allo-BMT are presented. Data presented are from at least 3 independent experiments. (A,F) Data are presented as mean ± standard deviation; (C-D,G) data are presented as mean ± standard error of the mean. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, n.s.: not significant; unpaired t test with Welch’s correction (A,F), Spearman’s correlation (B), Kaplan-Meier survival analysis for survival, and analysis of variance for analysis of weight loss and GVHD score.

Figure 2.

rLIF administration decreases tissue inflammation and donor immune cell infiltration after allo-BMT. Lethally irradiated B6C3F1 and C57BL/6 mice received BM and BM + T cells from C57BL/6 and BALB/c mice, respectively, along with or without rLIF treatment. (A-B) BM+T increased spleen weight (A) and length of SI (B) in B6C3F1 mice and C57BL/6 mice at 7 days and 10 days after allo-BMT, respectively, which was largely abolished by rLIF administration. (A, left) Representative images of spleen tissues. For B6C3F1 mice: n = 4 for BM; n = 16 for both BM + T and BM + T + rLIF; for C57BL/6 mice: n ≥ 4 for BM; n ≥ 5 for both BM + T and BM + T + rLIF. (C) rLIF administration reduced the expression of majority inflammatory cytokines examined in B6C3F1 mice at 7 days after allo-BMT. Relative mRNA expression levels of TNFα, CXCL1, CCL2, IL-1a, IL-1b, and IL-22 in SI were determined by quantitative real-time PCR assays and normalized with β-actin. n ≥ 5 mice/group. (D) rLIF administration decreased the infiltration of donor CD45⁺ immune cells in epithelium of the intestine (EPI) from B6C3F1 mice at 7 days after allo-BMT. Representative flow cytometry images (left) and quantifications (right) show the percentage of donor CD45⁺ immune cells (H2k^k–CD45⁺) in EPI from B6C3F1 mice at 7 days after allo-BMT. n = 8 mice/group. (E-F) rLIF administration decreased the donor immune cell infiltration in B6C3F1 (left) and C57BL/6 (right) mice at 7 and 10 days after allo-BMT, respectively. (E) The numbers of infiltrating donor cells in spleen, MLN, and LP tissues determined by flow cytometric assays. (F) The numbers of a set of infiltrating donor immune cells, including CD45, CD4, CD8, neutrophil, DC, macrophage and natural killer (NK) cells, in spleen (upper), MLN (middle), and LP (lower) tissues. The number of cells in mice that received BM + T without rLIF treatment was defined as 1. Gating strategies are shown in supplemental Figure 3. n ≥ 7 mice/group. Data are presented as mean ± standard deviation from at least 3 independent experiments. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, unpaired t test with Welch’s correction.

Figure 3.

rLIF administration protects ISCs from GVHD-induced damage. Administration of rLIF significantly increased the number of proliferating crypts (A) and viable ISCs (B) in B6C3F1 mice at 2 days after allo-BMT. (A-B, upper) Representative images of immunohistochemistry staining of Ki67 (A) and Olfm4 (B) in the duodenum and ileum tissues. (A-B, lower) Quantification of viable crypts per field (A) and Olfm4-positive crypts per field (B) in the duodenum and ileum of mice after allo-BMT. n = 30 fields from at least 3 mice/group. (C) IFNγ induced cell death in intestinal organoids derived from B6C3F1 mice, which was largely decreased by rLIF. Representative images of organoid growth (left). Quantification of organoid viability (right). n = 4/group. Data are presented as mean ± standard deviation. ∗∗∗P < .001, n.s.: not significant. Student t test.

Figure 4.

rLIF administration inhibits the elevation of MHC-II expression on IECs and donor T cell activation after allo-BMT. (A-B) Allo-BMT increased MHC-II presentation on IECs, which was reduced by administering rLIF in mice. (A, left) Representative histograms (left) and quantifications of mean florescence intensity (MFI) (right) of MHC-II levels on IECs from B6C3F1 mice at 7 days after allo-BMT. (A, right) Representative histograms (left) and quantifications of MFI (right) of MHC-II levels on IECs from C57BL/6 mice at 10 days after allo-BMT. For B6C3F1 recipients, n = 3 for BM; n = 8 for both BM + T and BM + T + rLIF; for C57BL/6 recipients, n = 3 for BM; n = 10 for both BM + T and BM + T + rLIF. (B, upper) Representative immunofluorescence staining of MHC-II levels on IECs in the SI from C57BL/6 mice at different days after allo-BMT. (B, lower) Quantifications of MHC-II MFI. n ≥ 3 mice/group. (C) The number of donor activated T cells in the MLN (upper) and LP (lower) tissues from B6C3F1 mice at 7 days after allo-BMT with or without rLIF administration. n ≥ 8 mice/group. Data are presented as mean ± standard deviation. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, unpaired t test with Welch’s correction.

Figure 5.

rLIF administration inhibits radiation-induced IL-12 production in DCs through the STAT1 signaling to protect against GVHD. (A) TBI (11 Gy) induced IL-12 production from DCs in MLNs, which was greatly reduced by rLIF administration as examined by using the cytokine panel at 24 hours after TBI in C57BL/6 mice. n = 8 mice/group. (B) TBI (11 Gy) increased percentage (left) and number (right) of IL-12⁺ cells in DCs in MLNs at 24 hours after TBI, which were largely decreased by rLIF administration in IL-12–p40–YFP C57BL/6 reporter mice as examined by flow cytometric assays. n ≥ 4 mice/group. The gating strategy and representative flow images are shown in supplemental Figure 8D. (C) The significantly decreased expression of T-bet in donor CD4⁺ T cells in MLNs from B6C3F1 mice at 7 days after allo-BMT with rLIF administration compared with mice without rLIF administration as determined by flow cytometric assays. n = 8 mice/group. (D) The induction of the expression of Th1 cytokine IFNγ by BM + T in B6C3F1 mice was largely decreased by rLIF administration as determined at 7 days after allo-BMT. Relative mRNA expression levels of IFNγ in the SI were determined by quantitative real-time PCR assays (left). Protein levels of IFNγ in the SI (middle) and serum (right) were determined by ELISAs. n ≥ 4 mice/group. (E-F) Administering rIL-12 largely abolished the protective effect of rLIF on GVHD. Lethally irradiated B6C3F1 mice that received allo-BMT from C57BL/6 mice along with or without rLIF administration were treated with rIL-12 (500 ng/d for 7 days from D-4 to D3) or PBS. (E) The spleen weight (left) and length of SI (right) in B6C3F1 mice measured at 7 days after allo-BMT. (F) MFI of MHC-II on IECs of B6C3F1 mice were determined at 7 days after allo-BMT. n ≥ 3 mice/group. (G-H) Administering rIL-12 largely abolished the inhibitory effect of rLIF on donor immune cell infiltration after allo-BMT. (G) The relative numbers of infiltrating donor cells in the spleen, MLN, and LP tissues from B6C3F1 mice at 7 days after allo-BMT. (H) The relative numbers of infiltrating donor immune cells in the spleen (top), MLN (middle), and LP (bottom) tissues from B6C3F1 mice at 7 days after allo-BMT. n ≥ 6 mice/group. (I) BMDCs were activated by IFNγ (10 ng/ml) and LPS (100 ng/ml) with or without LIF treatment (100 ng/ml) for 6 hours. The mRNA levels of IL12b in BMDCs were determined by quantitative real-time PCR assays and normalized with β-actin. n = 7/group. (J) rLIF treatment increased the phosphorylation levels of STAT1 at Tyr-701 (p-STAT1) in activated BMDC as determined by Western blot assays. (K) rLIF treatment increased the binding of STAT1 to a putative STAT1 binding site in the intron 1 of IL12b gene as determined in BMDCs by chromatin immunoprecipitation assays. (Top) The sequence and location of the putative STAT1 binding site in IL12b gene. A region containing no STAT1 binding site was included as a negative control. n = 5/group. n.d.: non-detectable. (L) Blocking the STAT1 signaling by fludarabine (2 μM) and pravastatin (2 μM), 2 small-molecule STAT1 inhibitors, largely abolished the inhibitory effect of rLIF on IL12b production in activated BMDCs. The mRNA levels of IL12b in BMDCs were determined by quantitative real-time PCR assays and normalized with β-actin. n = 3/group. Data are presented as mean ± standard deviation from 3 independent experiments. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, n.s.: not significant; unpaired t test with Welch’s correction.

Figure 5.

rLIF administration inhibits radiation-induced IL-12 production in DCs through the STAT1 signaling to protect against GVHD. (A) TBI (11 Gy) induced IL-12 production from DCs in MLNs, which was greatly reduced by rLIF administration as examined by using the cytokine panel at 24 hours after TBI in C57BL/6 mice. n = 8 mice/group. (B) TBI (11 Gy) increased percentage (left) and number (right) of IL-12⁺ cells in DCs in MLNs at 24 hours after TBI, which were largely decreased by rLIF administration in IL-12–p40–YFP C57BL/6 reporter mice as examined by flow cytometric assays. n ≥ 4 mice/group. The gating strategy and representative flow images are shown in supplemental Figure 8D. (C) The significantly decreased expression of T-bet in donor CD4⁺ T cells in MLNs from B6C3F1 mice at 7 days after allo-BMT with rLIF administration compared with mice without rLIF administration as determined by flow cytometric assays. n = 8 mice/group. (D) The induction of the expression of Th1 cytokine IFNγ by BM + T in B6C3F1 mice was largely decreased by rLIF administration as determined at 7 days after allo-BMT. Relative mRNA expression levels of IFNγ in the SI were determined by quantitative real-time PCR assays (left). Protein levels of IFNγ in the SI (middle) and serum (right) were determined by ELISAs. n ≥ 4 mice/group. (E-F) Administering rIL-12 largely abolished the protective effect of rLIF on GVHD. Lethally irradiated B6C3F1 mice that received allo-BMT from C57BL/6 mice along with or without rLIF administration were treated with rIL-12 (500 ng/d for 7 days from D-4 to D3) or PBS. (E) The spleen weight (left) and length of SI (right) in B6C3F1 mice measured at 7 days after allo-BMT. (F) MFI of MHC-II on IECs of B6C3F1 mice were determined at 7 days after allo-BMT. n ≥ 3 mice/group. (G-H) Administering rIL-12 largely abolished the inhibitory effect of rLIF on donor immune cell infiltration after allo-BMT. (G) The relative numbers of infiltrating donor cells in the spleen, MLN, and LP tissues from B6C3F1 mice at 7 days after allo-BMT. (H) The relative numbers of infiltrating donor immune cells in the spleen (top), MLN (middle), and LP (bottom) tissues from B6C3F1 mice at 7 days after allo-BMT. n ≥ 6 mice/group. (I) BMDCs were activated by IFNγ (10 ng/ml) and LPS (100 ng/ml) with or without LIF treatment (100 ng/ml) for 6 hours. The mRNA levels of IL12b in BMDCs were determined by quantitative real-time PCR assays and normalized with β-actin. n = 7/group. (J) rLIF treatment increased the phosphorylation levels of STAT1 at Tyr-701 (p-STAT1) in activated BMDC as determined by Western blot assays. (K) rLIF treatment increased the binding of STAT1 to a putative STAT1 binding site in the intron 1 of IL12b gene as determined in BMDCs by chromatin immunoprecipitation assays. (Top) The sequence and location of the putative STAT1 binding site in IL12b gene. A region containing no STAT1 binding site was included as a negative control. n = 5/group. n.d.: non-detectable. (L) Blocking the STAT1 signaling by fludarabine (2 μM) and pravastatin (2 μM), 2 small-molecule STAT1 inhibitors, largely abolished the inhibitory effect of rLIF on IL12b production in activated BMDCs. The mRNA levels of IL12b in BMDCs were determined by quantitative real-time PCR assays and normalized with β-actin. n = 3/group. Data are presented as mean ± standard deviation from 3 independent experiments. ∗P < .05, ∗∗P < .01, ∗∗∗P < .001, n.s.: not significant; unpaired t test with Welch’s correction.

Figure 6.

rLIF administration effectively ameliorates GVHD and preserves the GVL effect. (A-C) Lethally irradiated C57BL/6 mice were transplanted with BM with or without T cells from BALB/c mice along with luciferase-labelled C1498 mouse leukemia cells. The day of BMT was denoted as D0. Recipient mice were treated with vehicle (PBS) or rLIF (IP, 30 ng/g body weight, twice a day, from D-4 to D3). (A) Kaplan-Meier survival curve of mice. (B) Percentage of tumor relapse in mice. (C) Representative bioluminescence images of mice throughout the experiment. Kaplan-Meier survival analysis was used to compare among groups. ∗P < .05, ∗∗∗P < .001, n.s.: not significant. (D) Schematic illustration of the role of LIF in protecting against GVHD. The diagram was prepared by using BioRender software.