Dermatan sulfate is involved in the tumorigenic properties of esophagus squamous cell carcinoma.

Active DS-epi1 is overexpressed in ESCC biopsiesA–B, normal or tumor tissue biopsies from patients were stained with anti-DS-epi1 antibody. Dotted line marks the boundary between cancer cells and the surrounding tissue. C–D, mouse wild-type and DS-epi1−/− esophagi (16) were stained to verify the specificity of the anti-DS-ep1 antibody. Scale bar: 20 μm. E, cellular enzymes were detergent-extracted from biopsies, and DS-epimerase activity was measured. Epimerase specific activity in the 8 samples from normal esophagi was 117 ± 69 dpm/h/mg (3H dpm released from the substrate/h/mg of assayed protein; mean ± SD; * p < 0.05 cancer versus normal). F, DS-epi1 was detected by WB in lysates from biopsies. HEK 293 cell lysates transfected with an empty vector or a vector containing DS-epi1 were included as negative and positive controls, respectively. An immunopurified anti-DS-epi1 antibody was used at 4 μg/ml.

Down-regulation of DS-epi1 in ESCC TE-1 cells results in decreased IdoATE-1 cells were infected with lentiviral particles containing DS-epi1 shRNA-a, shRNA-b, or non-target control sequences. A, DS-epi1 protein in control, DS-epi1 shRNA-a, and shRNA–b cell lysates. Immunopurified anti-DS-epi1 antibody was used at 1 μg/ml. B, confocal microscopy immunofluorescence analysis of DS-epi1 (green) and the cis Golgi marker GM130 (red) shows reduced DS-epi1 levels in shRNA-a cells as compared with control. Scale bar: 20 μm. C, size distribution on Superdex Peptide column of metabolically labeled CS/DS chains from TE-1 cells transduced with shRNA control (black line) or DS-epi1 shRNA-a (grey, broken line) and cleaved at IdoA residues by chondroitinase B prior to column separation. CS/DS derived from the large PG versican released into the medium is shown. Two shRNA-a clones were isolated and gave similar results (one is shown for brevity). Similar patterns were obtained from the CS/DS chains of the small PGs decorin and biglycan and from PGs present in the cell layer (see Table I). D, surface staining of non-permeabilized TE-1 cells with the anti-DS antibody GD3A12 that specifically recognizes IdoA residues. The bars indicate mean ± SD of triplicate stainings from flow cytometry analyses. E, confocal microscopy immunofluorescence analysis of IdoA from a similar experiment as in (D). F–G, qRT-PCR of DS-epi1 (F) and DS-epi2 (G) on DS-epi1 shRNA-a and control shRNA cells.

IdoA is involved in cellular HGF binding and HGF-dependent induction of ERK-1/2 signalingA, confocal microscopy analysis of surface-bound recombinant, biotinylated-HGF (green) and IdoA (red) shows partial co-localization (yellow) in shRNA control cells. Scale bar: 10 μm. B, quantification by FACS of binding of exogenously added HGF. Shown are representative data from three independent experiments. Student's t-test was used to test the significance of differences between control and DS-epi1 down-regulated cells (* p < 0.05, ** p < 0.01). C, cells were lysed after 24 h starvation in 0.1% FBS and analyzed for MET protein by WB. D, HGF-mediated induction of pERK-1/2 in control cells. Cells were incubated with or without HGF (2.5 ng/ml) and levels of phosphorylated kinases were determined by antibody array analysis as described in SI Material and Methods. E, lysates from control and DS-epi1 shRNA transduced cells were prepared following incubations in the presence or absence of HGF (2.5 ng/ml) at the indicated time points, and analyzed by WB. Anti-pERK-1/2 (upper panel) and anti-ERK-1/2 (lower panel) antibodies were used at 1:2,000 and 1:1,000 dilutions, respectively.

DS-epi1 is involved in the migration and invasion of TE-1 cellsA–B, wound scratch assay. A, representative phase contrast images taken with a 40× objective. B, quantification of the migrated area as shown in (A). C, transwell migration assay. D, transwell invasion assay using a Matrigel-coated membrane. A–D, each assay was performed in 4–6 replicates, and was repeated at least twice with similar results. Black bars refer to experiments in the presence of HGF (50 ng/ml). White bars refer to experiments without HGF. Student's t-test was used to test the significance of the differences between control and DS-epi1 down-regulated cells. (* p < 0.05, ** p < 0.01, ***p < 0.001) E, cellular lysates were prepared at the indicated time points after cell plating, and analyzed by WB. Anti-FAK and anti-pY397FAK antibodies were used at 1:4,000 and 1:1,000, respectively. Right panel: FAK/tubulin and pFAK/tubulin ratios were calculated by densitometric analysis of WB films using Quantity One. F, cells migrating in the wound scratch assay in the presence of HGF (50 ng/ml) were stained for actin filaments by phalloidin (red) and for focal adhesions by anti-pY397FAK (green). Scale bars, 10 μm. White arrow-heads in upper panels indicate plasma membrane protrusions.

Mass spectrometry analysis of CS/DS from human ESCC biopsies and normal tissueCS/DS was purified from ESCC biopsies and adjacent normal tissue and extensively degraded by a mixture of chondroitinases ABC, AC–I, and B (A–C), or specifically degraded at IdoA residues by chondroitinase B alone (D), as described in SI Material and Methods. The degradation products were separated by size permeation and on-line injected into the mass spectrometer. Monosulfated (m/z = 458) and disulfated (m/z = 538) disaccharides were quantified. Non sulfated disaccharides (m/z = 378) were not considered due to contamination from hyaluronic acid-derived disaccharides after chondroitinase ABC digestion, and as they constituted a minor component (mean 1.8%) of the predominant monosulfated disaccharides after chondroitinase B alone digestion. Measurements were performed in triplicate. P-values were obtained by Wilcoxon signed-rank test (A–D * p < 0.05 cancer versus normal). Results from analyses of normal tissue were set to 1 in A and D.

Increased O-sulfotransferase activities in ESCC tumorsCellular enzymes were detergent-extracted from biopsies, and O-sulfotransferase activities adding a sulfate group to chondroitin (A) or dermatan (B) as substrates, were measured. The labeled products were recovered and quantitatively depolymerized to disaccharides by the action of chondroitinase ABC. The 6-O- or 4-O-position of the added labeled sulfate was determined by HPLC separation of the disaccharides. No labeled disulfated or 2-O-monosulfated disaccharides were observed (data not shown). The data are presented as the ratio of tumor versus normal tissue biopsies (set to 1), both obtained from the same patient. 4-O-sulfotransferase specific activity on chondroitin in normal esophagi tissues was 2,919 ± 1,458 dpm/h/mg (35S incorporated into chondroitin/h/mg of assayed protein; mean ± SD). Assays were run in triplicates. P-values were obtained by subjecting the ratio values to Wilcoxon signed-rank test. Ratio of 4-O-sulfotransferase and 6-O-sulfotransferase activities on chondroitin had p-values of p=0.043 and p=0.08, respectively, when comparing tumor versus normal tissues. Ratio of sulfotransferases activites on dermatan did not reach statistical significance when comparing tumor versus normal.