Fig 2: Troy expression within gastric cancer tissuesA-D. Representative immunostainings of high (A+B), moderate (C) and low (D) Troy expression in tumor cells of intestinal (A+D) or diffuse type (B+C) gastric cancer. E. Analysis of Troy immunostaining of tumor cells was complicated by a noticeable expression within stromal components. F. Most intestinal metaplasia (17 of 19 cases) stained positive for Troy (all original magnifications 250x).

Fig 3: The Wnt target gene Troy can be detected in gastric cancer cell lines and overexpression of variant 2 suppresses the tumor phenotype of MKN45 cellsA. Expression profile of Troy and its variant 2 in different cell lines compared with non-neoplastic human stomach tissue [n=3 for non-neoplastic tissue (NT) and tumor (TU), n=2 for cell lines]. Expression was normalized to SDHA in all cell culture experiments. B. Expression of Troy in MKN45 cells is not significantly changed after 5azaC treatment for 72 h (n=3). C. qRT- PCR analysis of Troy and LGR5 expression after stimulation of MKN45 with 10 nM Wnt3a (n=2) alone or in combination with 10 nM RSPO1 and/or RSPO2 (n=3-4). D. - E. MTT based proliferation analysis of stable transfected MKN45 (D) and MKN74 (E) overexpressing Troy variant 2 compared to control transfected cells (n=3). F. Colony formation assay evaluates the ability for clonal expansion of Troy variant 2 transfected MKN45 (n=8). All data are represented at mean +/- SD with *p=0.051; **p=0.108; ***p=0.031).

Fig 4: Detection of Troy on protein level is relevant for gastric cancer prognosisA. Distribution of the Histoscore results for non-neoplastic mucosal epithelium (NT) and corresponding cancer cells (TU; n=52). Compared with normal tissue, neoplastic cells showed a less pronounced Troy expression. For further analysis, the group was divided at the median (positive, when Histoscore = 50, compared to low expressing or negative cases). B. Survival analysis for Troy-positive and Troy-negative patients using Kaplan-Meier plots. Troy expression proved especially relevant for prognosis of intestinal type GC, whereas stroma cell expression had no obvious predictive power. C. qRT-PCR analysis on tumor samples and matching non-neoplastic mucosa with primers detecting all isoforms of Troy (left) or specifically transcript variant 2 (right). Expression was normalized to succinate dehydrogenase subunit A (SDHA), calpain 2 (CAPN2) and cyclophilin C (CYCC) as housekeeping genes and logarithmized (n=50).

Fig 5: Troy variant 2 suppresses Wnt target gene expressionAfter 20 h of stimulation with 10 mM LiCl, TCF/Lef1 dependent reporter activity measured by luciferase signal is decreased in Troy overexpressing MKN45-cells compared with the pcDNA control line (n=2 independent experiments; mean ± SD; p=0.001).