Dynamic glycolytic reprogramming effects on dendritic cells in pancreatic ductal adenocarcinoma

Abstract

Background: Pancreatic ductal adenocarcinoma tumors exhibit resistance to chemotherapy, targeted therapies, and even immunotherapy. Dendritic cells use glucose to support their effector functions and play a key role in anti-tumor immunity by promoting cytotoxic CD8⁺ T cell activity. However, the effects of glucose and lactate levels on dendritic cells in pancreatic ductal adenocarcinoma are unclear. In this study, we aimed to clarify how glucose and lactate can impact the dendritic cell antigen-presenting function and elucidate the relevant mechanisms.

Methods: Glycolytic activity and immune cell infiltration in pancreatic ductal adenocarcinoma were evaluated using patient-derived organoids and resected specimens. Cell lines with increased or decreased glycolysis were established from KPC mice. Flow cytometry and single-cell RNA sequencing were used to evaluate the impacts on the tumor microenvironment. The effects of glucose and lactate on the bone marrow-derived dendritic cell antigen-presenting function were detected by flow cytometry.

Results: The pancreatic ductal adenocarcinoma tumor microenvironment exhibited low glucose and high lactate concentrations from varying levels of glycolytic activity in cancer cells. In mouse transplantation models, tumors with increased glycolysis showed enhanced myeloid-derived suppressor cell infiltration and reduced dendritic cell and CD8⁺ T cell infiltration, whereas tumors with decreased glycolysis displayed the opposite trends. In three-dimensional co-culture, increased glycolysis in cancer cells suppressed the antigen-presenting function of bone marrow-derived dendritic cells. In addition, low-glucose and high-lactate media inhibited the antigen-presenting and mitochondrial functions of bone marrow-derived dendritic cells.

Conclusions: Our study demonstrates the impact of dynamic glycolytic reprogramming on the composition of immune cells in the tumor microenvironment of pancreatic ductal adenocarcinoma, especially on the antigen-presenting function of dendritic cells.

Keywords: Antigen-presenting function; Dendritic cell; Glycolysis; Pancreatic ductal adenocarcinoma; Tumor microenvironment.

Fig. 1

PDAC tumor glycolysis levels affect the numbers of DCs and CD8⁺ T cells. A Relative luminescence (RLU) levels of glucose and lactate in human PDAC tumors and paired adjacent normal pancreatic tissues (n = 20 patients). B We analyzed public scRNA-seq data and found that the relative LDHA expression levels were higher in cancer cells from primary pancreatic tumors of PDAC patients (Tumor) than those in normal pancreatic epithelial cells derived from individuals without PDAC (Normal). C Fold changes of glucose consumption and lactate production by PDOs. The results relative to Organoid 497 are shown (n = 3, 48 h). D qRT-PCR analysis of the mRNA expression levels of GLUT1, LDHA, and MCT4 in PDOs relative to their expression levels in Organoid 497. The results were normalized to β-ACTIN mRNA expression levels (n = 3). E, F Representative IHC staining of (E) CD8⁺ cells and (F) CD11c⁺ cells in high LDHA PDACs (n = 29) and low LDHA PDACs (n = 19). Scale bar, 100 μm. G, H The analysis of the public PDAC scRNA-seq data revealed the proportions of (G) CD8⁺ T cells and (H) DCs among the CD45⁺ cells from normal pancreatic tissues of individuals without PDAC (Normal) and primary pancreatic tumors of PDAC patients (Tumor). I, J Overall survival analysis (Kaplan–Meier curve analysis) of PDAC patients, combining data from (I) Kyushu University Hospital (n = 29 high LDHA, n = 19 low LDHA) and (J) TCGA database (n = 89 high LDHA, n = 89 low LDHA)

Fig. 2

Increased glycolysis cells lead to the development of a low-glucose and high-lactate TME. A The FCM representative plot (left) and quantification (right) of 2-NBDG staining in control and increased glycolysis (IG) cells (n = 3). ‘Relative MFI’ denotes 2-NBDG MFI of stained samples relative to the matched unstained cells. B ECAR results of control and IG cells (n = 5). C Glucose consumption and lactate production of the control and IG cells (n = 3, 24 h). D Proliferation rates of the control and IG cells (n = 5). E, F Orthotopic transplantation of control and IG cells (n = 5) into (E) BALB/c-nu mice and (F) C57BL/6 mice for 18 days, followed by quantification of tumor weights (one of the BALB/c-nu mice transplanted with IG cells had died). G–J FCM analysis of (G) the percentages of tumor-infiltrating CD45⁺ cells in the control and IG tumors (n = 5) of C57BL/6 mice, and the percentages of (H) tumor-infiltrating CD8⁺ T cells, (I) DCs, and (J) Mo-MDSCs among the CD45⁺ cells. K–M Representative IHC staining of (K) CD8⁺ cells, (L) CD11c⁺ cells, and (M) CD11b⁺ cells in tumors of C57BL/6 mice. Scale bar, 100 μm. N Relative fold changes of glucose and lactate levels in the control and IG tumors (n = 5)

Fig. 3

Decreased glycolysis cells lead to the development of a high-glucose and low-lactate TME. A Representative western blot analysis results of Ldha and β-actin protein levels in whole cell lysates of shNC, shLDHA-1, shLDHA-2, and shLDHA-3 cells. B ECAR results of shNC and shLDHA cells (n = 5). C Glucose consumption and lactate production by shNC and shLDHA cells (n = 3, 24 h). D Proliferation rates of shNC and shLDHA cells (n = 5). E, F Orthotopic transplantation of shNC and shLDHA cells (n = 5) into (E) BALB/c-nu mice and (F) C57BL/6 mice for 20 days, followed by quantification of tumor weight (one of the BALB/c-nu mice transplanted with shNC cells had died). G–J FCM analysis of (G) the percentages of tumor-infiltrating CD45⁺ cells in shNC and shLDHA tumors (n = 5) of C57BL/6 mice, and the percentages of (H) tumor-infiltrating CD8⁺ T cells, (I) DCs, and (J) MDSCs among the CD45⁺ cells. K–M Representative IHC staining of (K) CD8⁺ cells, (L) CD11c⁺ cells, and (M) CD11b⁺ cells in tumors of C57BL/6 mice. Scale bar, 100 μm. N Relative fold changes of glucose and lactate levels in shNC and shLDHA tumors (n = 5)

Fig. 4

High-glucose and low-lactate conditions promote the DC antigen-presenting function and enhance CD8⁺ T cell activity. A The UMAP plots of 5,393 CD45⁺ cells (shNC tumors, n = 5) and 1,974 CD45⁺ cells (shLDHA tumors, n = 5). CD45⁺ cells were classified into 11 subsets using known marker genes. B The UMAP plots of 1,125 DCs (shNC tumors) and 712 DCs (shLDHA tumors) were classified into four subsets using known marker genes. C The dot plots of representative genes related to DC subsets and Z-scores normalized log2. The center represents the average expression level across all single cells with a color scale from 1 to -1. D Gene Ontology (GO) terms associated with the differentially expressed genes between DCs from shLDHA tumors and shNC tumors (analyzed using Metascape). E–G Violin plots showing the relative expression levels of (E) B2m in cDC1, (F) H2-Aa and CD74 in cDC2, and (G) Sl2ac1 in DCs. H The GSEA results showing the enrichment of RESPONSE_TO_TYPE_II_INTERFERON in DCs. I, J Violin plots showing (I) the relative expression levels of IFN-γ and PDCD1 in CD8⁺ T cells and (J) the cytotoxicity signature scores of CD8⁺ T cells

Fig. 5

Cancer cells with increased glycolysis can inhibit the antigen-presenting function of BM-DCs. A Representative images of BM-DCs co-cultured with control and IG cells (n = 3), respectively, using ALI-3D co-culture. B FCM analysis of the MHC I, MHC II, CD80, and CD86 expression levels in BM-DCs. C Representative images of BM-DCs co-cultured with shNC and shLDHA cells (n = 3), respectively, using ALI-3D co-culture. D FCM analysis of the MHC I, MHC II, CD80, and CD86 expression levels in BM-DCs. After 2D culture of cancer cell supernatant and BM-DCs for 48 h, the MHC I, MHC II, CD80, and CD86 expression levels of BM-DCs were analyzed by FCM. E Control and IG supernatants (n = 3). F shNC and shLDHA supernatants (n = 3)

Fig. 6

The BM-DC antigen-presenting function is inhibited by mitochondrial OXPHOS dampening in low-glucose and high-lactate conditions. A, B BM-DCs were cultured in control RPMI and CM RPMI for 48 h, and the expression levels of MHC I, MHC II, CD80 and CD86 were analyzed by FCM, (A) without LPS (A) and (B) with 100 ng/mL LPS. Control RPMI (11 mM glucose and 0 mM lactate) and conditioned media (CM) RPMI (1 mM glucose and 15 mM lactate). C Real-time analysis of the OCR of pre-cultured BM-DCs in control and CM RPMI for 48 h (n = 5). D Representative mitochondria (TIM23) images of pre-treatment BM-DCs in control and CM RPMI for 48 h (left) and the quantification of the number of mitochondria (right) (n = 5). E Relative ATP levels of BM-DCs in control and CM RPMI (n = 3, 48 h). F Representative images (left) and the percentages of BM-DC (CD11c⁺ MHC II⁺) differentiation rates (right) in control and CM RPMI (n = 3, 9 days)