IL-23 is a proinflammatory cytokine consisting of a p19 subunit and a p40 subunit that is shared with IL-12. IL-23 is overexpressed in and around tumor tissues, where it induces local inflammation and promotes tumor development. Many tumor cells produce large amounts of lactic acid by altering their glucose metabolism. In this study, we show that lactic acid secreted by tumor cells enhances the transcription of IL-23p19 and IL-23 production in monocytes/macrophages and in tumor-infiltrating immune cells that are stimulated with TLR2 and 4 ligands. DNA elements responsible for this enhancing activity of lactic acid were detected in a 2.7-kb 5′-flanking region of the human IL-23p19 gene. The effect of lactic acid was strictly regulated by extracellular pH. Furthermore, by inducing IL-23 overproduction, lactic acid facilitated the Ag-dependent secretion of proinflammatory cytokine IL-17 but not IFN-γ by TLR ligand-stimulated mouse splenocytes. Interestingly, this effect was observed even in the absence of TLR ligand stimulation. These results suggest that rather than just being a terminal metabolite, lactic acid is a proinflammatory mediator that is secreted by tumor cells to activate the IL-23/IL-17 proinflammatory pathway but not the Th1 pathway. Targeting the lactic acid-induced proinflammatory response may be a useful approach for treating cancer.

Interleukin-17A (IL-17A) is a cytokine produced by T(h)17 cells that plays an important role in inflammatory and autoimmune diseases and cancer. Stimulation with IL-6, transforming growth factor-β , IL-21, IL-1β and IL-23 is required for differentiation of T(h)17 cells and the production of IL-17A. Recently, we reported that tumor-derived lactic acid enhances the toll-like receptor (TLR) ligand-mediated expression of IL-23, leading to increased IL-17A production. Tumor cells secrete large amounts of lactic acid due to the up-regulation of glycolysis, which is known as the Warburg effect. Even without TLR ligand stimulation, lactic acid enhanced antigen-dependent IL-17A production from splenocytes in an IL-23-dependent manner. Here, we show that macrophages and effector/memory CD4(+) T cells are the primary cell types involved in the ability of lactic acid to boost IL-17A production. Although lactic acid suppressed the proliferation of T(h)1 and T(h)17 cells, T(h)17 cells still secreted large amounts of IL-17A. CD40 ligand-CD40 interactions were involved in the up-regulation of IL-17A by lactic acid through IL-12/23p40 production. A new cytokine containing the IL-12/23p40 subunit, but not IL-23, IL-12 or the IL-12p40 homodimer, is a candidate for involvement in the up-regulation of IL-17A. IL-1β also increased IL-17A expression; however, IL-1β, CARD9 and MyD88 signaling pathways activated by known intrinsic inflammatory mediators were hardly required for the enhanced activity induced by lactic acid. Our results show that lactic acid functions as an intrinsic inflammatory mediator that activates IL-23-dependent and -independent pathways, resulting in the promotion of chronic inflammation in tumor microenvironments.

O-glycans are suitable targets as novel and useful tumor markers. The structures of O-glycans in human sera from four healthy controls were precisely analyzed to obtain the reference O-glycan database. O-glycans were prepared from sera by hydrazine treatment followed by fluorescent labeling with aminopyridine and identified using two-dimensional mapping, enzymatic digestion and mass spectrometry (MS) together with methanolysis and the use of newly synthesized sulfated oligosaccharides as standards. O-glycans, present at more than 0.01% of the total O-glycans, were analyzed, and 18 kinds of acidic and 2 kinds of neutral glycans were identified. NeuAcα2-3Galβ1-3N-acetylgalactosamine (GalNAc) (61-64%), NeuAcα2-3Galβ1-3(NeuAcα2-6)GalNAc (15-26%) and Galβ1-3GalNAc (6-14%) were major components while other sialylated glycans, Galβ1-3(NeuAcα2-6)GalNAc, Galβ1-4GlcNAcβ1-6(NeuAcα2-3Galβ1-3)GalNAc and NeuAcα2-3Galβ1-4GlcNAcβ1-6(NeuAcα2-3Galβ1-3)GalNAc were relatively minor components, accounting for ∼1-2%. Very minor glycans accounting for ∼0.01-0.1% of the total include (i) the neutral glycan, Galβ1-4GlcNAcβ1-6(Galβ1-3)GalNAc, (ii) sialylated glycans, having sialyl Tn antigen, agalacto and trisialylated structures, (iii) fucosylated glycans forming blood type H antigen, blood type A antigen, blood type B antigen, Lewis X antigen and sialyl Lewis X antigen and (iv) sulfated glycans, having 6-sulfo and 3'-sulfo structures. Two kinds of clinically applied tumor markers namely sialyl Tn antigen and sialyl Lewis X antigen in healthy controls sera were revealed to be present at ∼0.1-0.2% of the total. However, other markers such as CA19-9 and DU-PAN-2 were not found, suggesting the relative amounts of these glycans to be <0.01%. These detailed O-glycan profiles will help to find novel carbohydrate tumor markers.

We have analyzed the structures of glycosphingolipids and intracellular free glycans in human cancers. In our previous study, trace amounts of free N-acetylneuraminic acid (Neu5Ac)-containing complex-type N-glycans with a single GlcNAc at each reducing terminus (Gn1 type) was found to accumulate intracellularly in colorectal cancers, but were undetectable in most normal colorectal epithelial cells. Here, we used cancer glycomic analyses to reveal that substantial amounts of free Neu5Ac-containing complex-type N-glycans, almost all of which were α2,6-Neu5Ac-linked, accumulated in the pancreatic cancer cells from three out of five patients, but were undetectable in normal pancreatic cells from all five cases. These molecular species were mostly composed of five kinds of glycans having a sequence Neu5Ac-Gal-GlcNAc-Man-Man-GlcNAc and one with the following sequence Neu5Ac-Gal-GlcNAc-Man-(Man-)Man-GlcNAc. The most abundant glycan was Neu5Acα2-6Galβ1-4GlcNAcβ1-2Manα1-3Manβ1-4GlcNAc, followed by Neu5Acα2-6Galβ1-4GlcNAcβ1-2Manα1-6Manβ1-4GlcNAc. This is the first study to show unequivocal evidence for the occurrence of free Neu5Ac-linked N-glycans in human cancer tissues. Our findings suggest that free Neu5Ac-linked glycans may serve as a useful tumor marker.

We previously reported on the accumulation of a substantial amount of free N-acetylneuraminic acid (Neu5Ac)-containing complex-type N-glycans in human pancreatic cancer cells (Yabu M, Korekane H, Takahashi H, Ohigashi H, Ishikawa O, Miyamoto Y. 2013. Accumulation of free Neu5Ac-containing complex-type N-glycans in human pancreatic cancers. Glycoconj J, 30(3):247-256). In the present paper, we further extend our cancer glycomic study of human prostate cancer. Specifically, we demonstrate that, in addition to the free Neu5Ac-containing N-glycans, significant amounts of free deaminoneuraminic acid (KDN, 2-keto-3-deoxy-D-glycero-D-galacto-nononic acid)-containing N-glycans had accumulated in the prostate cancer tissues from four of five patients. Indeed, in one of the four cases, the free KDN glycans accumulated as major components in prostate cancer tissue. The structures of the KDN-containing free oligosaccharides were analyzed by a variety of methods. Specifically, we used fluorescent labeling with aminopyridine combined with two-dimensional mapping, KDNase digestion and mass spectrometry to facilitate identification. The analysis also utilized newly synthesized KDN-linked oligosaccharides as standards. The prostate-specific glycans were composed of five species having the following sequence, KDN-Gal-GlcNAc-Man-Man-GlcNAc (α2,6-KDN-linked glycans being the dominant form). The most abundant free KDN-containing N-glycan was KDNα2-6Galβ1-4GlcNAcβ1-2Manα1-3Manβ1-4GlcNAc followed by KDNα2-6Galβ1-4GlcNAcβ1-2Manα1-6Manβ1-4GlcNAc. This is the first study to show unequivocal chemical evidence for the occurrence of KDN glycoconjugates in human tissues together with their detailed structures. These oligosaccharides might be developed as tumor markers, especially for prostate cancer.

We have attempted to identify a novel glycan tumor marker. Pyridylaminated (PA) O-glycans were prepared from sera, and the corresponding O-glycan profiles were constructed by HPLC separation. By comparing the serum O-glycan profiles from healthy controls with those of cancer patients, we identified a marker candidate, core 1 sialyl Lewis A (NeuAcα2-3Galβ1-3(Fucα1-4)GlcNAcβ1-3Gal) (abbreviated C1SLA), whose concentration appeared to be weakly correlated with CA19-9 values. To quantify this glycan, we developed a selected reaction monitoring (SRM) assay that used a stable isotope, tetradeuterium-labeled pyridylamino (d4-PA) glycan, as an internal standard. The analyte (d0-PA-C1SLA) and the internal standard (d4-PA-C1SLA) were subjected to SRM analyses after two types of HPLC separation. Serum levels of C1SLA, determined as the relative ratio to total O-glycans, were then measured. These analyses revealed that (i) C1SLA is a CA19-9-related glycan, (ii) the mean value of C1SLA in normal controls is 3.41 ppm, (iii) the level of C1SLA was significantly higher in samples of stages II-IV stomach cancers (P = 0.0036) as well as pancreatic cancers (P < 0.0001) compared to that of normal controls, (iv) the relationship between C1SLA and CA19-9 varies from poor to weak depending on the cancer, and (v) C1SLA could be valuable as a diagnostic adjunct for cancer.