FUT2 enhances anti-tumor immunity in pancreatic cancer radiotherapy by driving FBXO2-mediated degradation of NR2F2

Abstract

Pancreatic ductal adenocarcinoma (PDAC) radiotherapy (RT) resistance is frequently mediated by an immunosuppressive tumor microenvironment (TIME). Utilizing an in vivo CRISPR-Cas9 metabolic enzyme screen, we identified fucosyltransferase 2 (FUT2) as a potent non-catalytic enhancer of RT response. Mechanistically, FUT2 scaffolds the E3 ubiquitin ligase FBXO2, facilitating K362 site-specific ubiquitination and proteasomal degradation of the transcription factor NR2F2. This degradation suppresses expression of the immunosuppressive factor Lipocalin-2 (LCN2), which drives CD8⁺ T cell exhaustion and impedes NK cell infiltration, fostering a radioresistant TIME. Interestingly, we observed that RT could reduce FUT2 transcript levels via an METTL14-mediated m⁶A RNA methylation, while NR2F2 was identified to transcriptionally upregulate METTL14, establishing a feedforward inhibitory loop that sustains FUT2 suppression. Clinically, FUT2 expression positively correlates with CD8⁺ T cell infiltration and prolonged survival in RT-treated PDAC patients. Preclinically, combining RT with LCN2-neutralizing antibodies elicited synergistic anti-tumor immunity. These results unveil FUT2 as a regulator of PDAC radiosensitivity via the FUT2-FBXO2-NR2F2-LCN2 axis, offering a promising therapeutic target to overcome RT resistance.

Fig. 1. FUT2 synergistically enhances pancreatic cancer cell death following radiotherapy.

A Schematic illustration of the workflow for in vivo CRISPR knockout screening. B The enrichment of positive and negative candidates in the CRISPR screen. C KPC cells overexpressing EV or FUT2 were subcutaneously injected to C57BL/6 mice that were then treated with or without RT at 14th day (n = 6). Tumor volumes were measured. EV, empty vector. RT, radiotherapy. D Quantification of colony formation in KPC cells overexpressing EV or FUT2 following exposure to varying doses of RT. E KPC cells overexpressing EV or FUT2 were subcutaneously injected to C57BL/6 mice that were then treated with or without RT at 14th day. Quantifications of tumor proliferative and dead index determined by Ki67, TUNEL, and cleaved caspase-3 staining are shown. F–H KPC cells overexpressing EV or FUT2 were subcutaneously injected to C57BL/6 mice that were then treated with RT at 14th day. The tumors were subjected to single-cell sequencing and the subtypes of immune cells are shown. I KPC cells overexpressing EV or FUT2 were subcutaneously injected to C57BL/6 mice that were then treated with or without RT at 14th day. Proportions of the cytotoxic T lymphocytes (CD3⁺CD8⁺) and NK cells (CD3^-NK1.1⁺) in tumors were analyzed by flow cytometry. J Quantification of colony formation in KPC cells overexpressing EV, FUT2 WT or ED following exposure to varying doses of RT. KPC cells overexpressing EV, FUT2 WT or ED were subcutaneously injected to C57BL/6 mice that were then treated with RT at 14th day. Tumor volume (K), quantifications of tumor proliferative and dead index determined by Ki67 and TUNEL staining (L), proportions of the cytotoxic T lymphocytes and NK cells (M) analyzed by flow cytometry are shown. N Forty patients with locally advanced PDAC scheduled for local RT were divided into two groups based on the optimal cutoff of FUT2 IHC score, classified as high or low expression. Survival curves were then generated to compare outcomes between the two groups. Scale bar, 50 μm. O Kaplan–Meier survival analyses of low and high FUT2 expression in pancreatic cancer patients via a Kaplan–Meier plotter database. P CD8⁺ T cell and FUT2 of the pancreatic cancer patients were measured using immunofluorescence co-staining and analyzed by calculating the Pearson correlation coefficient. Scale bar, 20 μm. ^*p < 0.05, ^**p < 0.01; ^***p < 0.001; ^****p < 0.0001; N.S., not significant; two-way ANOVA [(C–E) and (I–M)], log-rank test (N) or Pearson correlation coefficient test (P).

Fig. 2. FUT2 enhances radiotherapy efficacy by suppressing LCN2 expression.

A RNA profiling was performed on KPC cells overexpressing EV or FUT2 12 h post-RT. The differentially expressed genes (DEGs) were analyzed, and the top 10 DEGs are presented. Expression levels of LCN2 mRNA (B) and protein (C) were measured in KPC cells overexpressing EV, FUT2 WT, or ED in the absence or presence of RT, with RNA and protein samples collected at 12 and 24 h post-RT, respectively. D Immunoblotting analyses of indicated proteins in KPC cells expressing shNT, shFUT2, and shFUT2 rescued with rFUT2 WT or ED. E–G Subcutaneous transplantation of EV- or FUT2-overexpressing KPC cells into C57BL/6 mice that were then treated with RT at 14th day. Exogenous LCN2 supplementation was performed as described in the supplementary methods (n = 8). Tumor volume (E), quantifications of tumor proliferative and dead index determined by Ki67 and TUNEL staining (F), proportions of the cytotoxic T lymphocytes and NK cells analyzed by flow cytometry (G) are shown. H, I Flow cytometry and immunoblotting were performed to test the apoptosis of CD8⁺ T cells induced by exogenous LCN2 supplementation. J–L KPC cells were subcutaneously transplanted into C57BL/6 mice that were then treated with RT or LCN2 neutralizing antibodies as described in the supplementary methods (n = 8). Tumor volume (J), quantifications of tumor proliferative and dead index determined by Ki67 and TUNEL staining (K), proportions of the cytotoxic T lymphocytes and NK cells analyzed by flow cytometry (L) are shown. M IHC staining for LCN2 was obtained and analyzed between two groups divided by IHC score of FUT2 in PDAC tissues. Scale bar, 50 μm. N CD8⁺ T cell and LCN2 were measured using immunofluorescence co-staining and analyzed by calculating the Pearson correlation coefficient. Scale bar, 20 μm. ^*p < 0.05; ^**p < 0.01; ^***p < 0.001; ^****p < 0.0001; N.S., not significant; two-way ANOVA [(B) and (E–H) and (J–L)] or Two tailed Student’s t test (M).

Fig. 3. FUT2 represses LCN2 transcription by facilitating NR2F2 destabilization.

A LC–MS analysis was performed on KPC cells overexpressing EV or FUT2 at 12 h post-RT, revealing five major transcription-related factors. B, C Immunoblotting analyses of indicated proteins in KPC cells with the indicated genetic manipulation. Expression levels of LCN2 protein (D) and mRNA (E) were measured in KPC cells overexpressing EV or NR2F2-Flag. Expression levels of LCN2 protein (F) and mRNA (G) were measured in KPC and PANC-1 cells expressing shNT. shNR2F2 and shNR2F2 rescued with rNR2F2. H Immunoblotting analyses of indicated proteins in KPC and PANC-1 cells overexpressing EV, EV with NR2F2-Flag, FUT2, or FUT2 with NR2F2-Flag. I Expression levels of LCN2 mRNA were measured in KPC and PANC-1 cells expressing shNT or shFUT2 and subsequently treated with or without CIA1 (4 μM) for 12 h. J Expression levels of LCN2 mRNA were measured in KPC and PANC-1 cells overexpressing either EV or FUT2 following transfection with Vector, NR2F2 WT, or NR2F2 L364A/L365A. K Comprehensive comparison between RNA-seq and LC-MS results of NR2F2. L Expression levels of NR2F2 mRNA were measured in KPC and PANC-1 cells overexpressing EV or FUT2. M Immunoblotting analyses of indicated proteins in KPC and PANC-1 cells overexpressing EV or FUT2 and subsequently treated with CHX for indicated time. N Immunoblotting analyses of indicated proteins in KPC and PANC-1 cells expressing shNT or shFUT2 and subsequently treated with CHX for indicated time. ^*P < 0.05; ^***P < 0.001; ^****P < 0.0001; N.S., not significant; Two tailed Student’s t test [(E, L)]. or two-way ANOVA [(G), (I, J) and (M, N)].

Fig. 4. FUT2 promotes NR2F2 degradation through ubiquitination at lysine 362 via the proteasomal pathway.

A Immunoblotting analyses of indicated proteins in KPC and PANC-1 cells overexpressing EV or FUT2 and subsequently treated with or without MG132 (10 μM) before harvesting. B KPC and PANC-1 cells overexpressing EV or FUT2 were transfected with Ub-HA and then treated with or without MG132 (10 μM) before harvesting. Immunoprecipitation and immunoblotting analyses were performed with indicated proteins. C NR2F2-Flag was subjected to LC‒MS and identified three potential ubiquitination sites. D KPC and PANC-1 cells overexpressing FUT2 were transfected with Ub-HA and NR2F2-Flag WT or mutants, subsequently treated with MG132 (10 μM) before harvesting. Immunoprecipitation and immunoblotting analyses were performed with indicated proteins. E Screening of the LCN2 promoter region identified one of the potential NR2F2 binding motif sites in both humans and mice. F Dual-luciferase assay was performed to measure LCN2 promoter activity in HEK293T and KPC cells co-transfected with FUT2 and either NR2F2-Flag WT or NR2F2-Flag K362R. G ChIP-qPCR validation of LCN2 in KPC cells overexpressing either EV or FUT2 following transfection with Vector, NR2F2-Flag WT or NR2F2-Flag K362R and immunoprecipitated with anti-NR2F2 antibody. IgG was used as a blank control. H Expression levels of LCN2 mRNA were measured in KPC and PANC-1 cells overexpressing either EV or FUT2 following transfection with Vector, NR2F2-Flag WT or NR2F2-Flag K362R. I IHC staining for NR2F2 was obtained and analyzed between two groups divided by IHC score of FUT2 in PDAC tissues. Scale bar, 50 μm. ^*P < 0.05; ^**P < 0.01; ^****P < 0.0001; N.S., not significant; two-way ANOVA (F–H) or Two tailed Student’s t test (I).

Fig. 5. FUT2 promotes NR2F2 degradation by facilitating its interaction with the E3 ubiquitin ligase FBXO2.

A Flag-FUT2 associated proteins were identified and analyzed using LC‒MS/MS. The Co-IP assays were performed with indicated antibodies to detect the association of FUT2 (B) or NR2F2 (C) with FBXO2 and SKP1 in KPC and PANC-1 cells. D The protein levels of Flag-NR2F2 were detected in KPC and PANC-1 cells with the indicated genetic manipulation and MG132 treatment. E NR2F2 protein levels were assessed in KPC and PANC-1 cells overexpressing FUT2-Flag or EV infected with lentivirus expressing shNT or shFBXO2. F NR2F2 protein levels were assessed in KPC and PANC-1 cells overexpressing FBXO2-HA or EV infected with lentivirus expressing shNT or shFUT2. G The endogenous interaction of FBXO2 and NR2F2 was detected in KPC and PANC-1 cells progressively overexpressing FUT2 by Co-IP assays. MG132 was added to each group. H The endogenous interaction of FBXO2 and NR2F2 was detected in KPC cells expressing shNT or shFUT2 by Co-IP assay. I Schematic diagram of the construction of full-length and truncated FUT2 proteins. J NR2F2 protein levels were assessed in KPC and PANC-1 cells overexpressing FUT2-FL/N/C-Flag or EV. The Co-IP assays in HEK293T cells were performed to test the interaction of NR2F2-HA (K) or FBXO2-HA (L) with FUT2-FL/N/C-Flag. M Schematic of FUT2-enhanced assembly of the NR2F2-FUT2-SCF^FBXO2 holocomplex.

Fig. 6. METTL14-mediated m6A modification represses FUT2 expression following radiotherapy.

A IHC analysis of FUT2 expression was performed on paired tumor samples from 13 PDAC patients before and after radiotherapy. B, C Immunoblotting and qPCR analyses of the FUT2 expression in KPC and PANC-1 cells treated with RT for the indicated time. D The decay rates of Fut2 mRNA were determined at 0, 2, 4, 6 h after treating with actinomycin D (ActD, 5ug/ml) in KPC cells with or without RT treatment. E The potential m6A methylation sites in the Fut2 mRNA 3’UTR from bioinformatics prediction are shown. F Dot blot analysis of m6A levels in total Fut2 RNA from irradiated and control KPC cells at 3 h post-RT. The intensity of the dot blot represents the level of m6A modification (up), and methylene blue staining was used to detect sample loading (down). G, H Immunoblotting and qPCR analyses of the METTL14 expression in KPC and PANC-1 cells treated with RT for the indicated time. I Dot blot analyses of total Fut2 RNA m6A levels in irradiated KPC cells expressing shNT or shMETTL14 at 3 h post-RT. J The decay rates of Fut2 mRNA were determined at 0, 2, 4, 6 h after treating with ActD (5ug/ml) in irradiated KPC cells expressing shNT or shMETTL14. K A luciferase reporter gene containing wild-type or mutant (A-to-G mutation) FUT2 was created for subsequent assays. L Transcript levels of wild-type and mutant FUT2 in KPC cells expressing shNT or shMETTL14 were detected by dual-luciferase assays. M, N EV- or METTL14- overexpressing KPC cells transfected with either Vector or FUT2 were injected subcutaneously into C57BL/6 mice, and the mice were then treated with RT at 14th day. Tumor volume (M), Tumor weight (N) are shown. O Screening of the METTL14 promoter region identified one of the potential NR2F2 binding motif sites in both humans and mice. P, Q Expression levels of Mettl14 mRNA were measured in KPC cells with the indicated genetic manipulation. R Dual-luciferase assay was performed to measure METTL14 promoter activity in HEK293T and KPC cells co-transfected with NR2F2 and either FUT2 WT or ED. S ChIP-qPCR validation of METTL14 in KPC cells overexpressing EV, FUT2 WT or ED following transfection with Vector or NR2F2 and immunoprecipitated with anti-NR2F2 antibody. IgG was used as a blank control. T IHC staining for METTL14 was obtained and analyzed between two groups divided by IHC score of FUT2 in PDAC tissues. Scale bar, 50 μm. ^*p < 0.05; ^**p < 0.01; ^***p < 0.001; ^****p < 0.0001; N.S., not significant; Two tailed Student’s t test [(A), (F), (I) and (T)] or two-way ANOVA [(C, D), (H), (J), (M, N) and (P–S)].

Fig. 7

Schematic model of the mechanism by which FUT2-mediated LCN2 expression determines the efficiency of radiation therapy.