Te proteoglycan (A) or were fixed and immunostained with an antibody to 
chondroitin sulfate proteoglycan and nuclei were stained with DAPI (B). (C) U87 cells were treated or not with TG for 48 h, total RNA was isolated and SR-3029 chemical information levels of 1485-00-3 manufacturer Col1a1 mRNA were measured by real time PCR. N = 3 experiments. (D) U87 cells were treated with control or OASIS siRNAs as in Figure 4A, then with or without TG for 16 h and levels of Col1a1 mRNA measured by real time PCR. N = 3 independent experiments (Bars are SEM; *p,0.05; TG vs without TG, t-test). doi:10.1371/journal.pone.0054060.ginduction of the UPR in human glioma cell lines. Our results also suggest that GRP94 may be an OASIS 12926553 target gene in glioma cells. In osteoblasts and pancreatic b-cell lines, however, OASIS does not induce GRP78 expression [16,18]. Rather in these cell 
types OASIS induces expression of genes involved in extracellular matrix production and protein transport, among others. Rat C6 glioma cells exposed to TG had increased levels of chondroitin sulfate [33]. However, TG treatment of U373 and U87 cell lines resulted in reduced cellular chondroitin sulfateproteoglycan expression. The reason for this is reduction is unclear, although it may relate to the concentration of TG used and length of treatment. Importantly, however, we observed that even in the absence of chemically-induced ER stress, the levels of cellular chondroitin sulfate proteoglycans were reduced in U373 and U87 cells treated with OASIS siRNA. Thus, OASIS may be responsible for maintaining chondroitin sulfate proteoglycan expression even under control conditions in glioma cell lines that express this protein. This is possible given that glioma cells exhibitOASIS in Human Glioma CellsFigure 6. OASIS knock-down perturbs U373 cell migration. (A) U373 cells transfected with 100 nM control (GFP) or OASIS siRNAs for 72 h before scratching the 90 confluent cell monolayers. The cells were then incubated for the indicated times and DIC images were obtained. The approximate cell migration is indicated by the white 23727046 lines. The images were analyzed by ImageJ and the wound area was quantified (B) (*p,0.05; OASIS siRNA vs. control siRNA, n = 3). Note the almost complete healing of the wound in control cells and poor migration of cells in which OASIS was knocked-down. doi:10.1371/journal.pone.0054060.gmild activation of the UPR even under basal conditions (Figure 4D). Another extracellular matrix component shown to be regulated by OASIS in osteoblast cells is the collagen gene Col1a1 [16]. Interestingly, the induction of this gene by ER stress was not affected by OASIS knock-down, suggesting that OASIS is not required for Col1a1 induction in glioma cells or that perhaps another ATF6 family isoform is able to compensate for OASIS loss allowing for Col1a1 induction in glioma cells. The in vitro migration assay identified that OASIS silenced cells have poor migration efficiency. The exact mechanism of this effect is unknown, although presumably OASIS is required for maintaining ECM components that are required for efficient cell migration. Gliomas are characterized by aggressive growth and invasiveness, which is closely related to cell-ECM interactions [23,25,26,27,37,38,39]. Given that OASIS can be induced by ER stress and allows the cell to modulate its matrix we hypothesize that OASIS is likely to be beneficial under chronic, but low intensity ER stress, such as may occur under hypoxic conditions in vivo [22]. This will allow the cel.Te proteoglycan (A) or were fixed and immunostained with an antibody to chondroitin sulfate proteoglycan and nuclei were stained with DAPI (B). (C) U87 cells were treated or not with TG for 48 h, total RNA was isolated and levels of Col1a1 mRNA were measured by real time PCR. N = 3 experiments. (D) U87 cells were treated with control or OASIS siRNAs as in Figure 4A, then with or without TG for 16 h and levels of Col1a1 mRNA measured by real time PCR. N = 3 independent experiments (Bars are SEM; *p,0.05; TG vs without TG, t-test). doi:10.1371/journal.pone.0054060.ginduction of the UPR in human glioma cell lines. Our results also suggest that GRP94 may be an OASIS 12926553 target gene in glioma cells. In osteoblasts and pancreatic b-cell lines, however, OASIS does not induce GRP78 expression [16,18]. Rather in these cell types OASIS induces expression of genes involved in extracellular matrix production and protein transport, among others. Rat C6 glioma cells exposed to TG had increased levels of chondroitin sulfate [33]. However, TG treatment of U373 and U87 cell lines resulted in reduced cellular chondroitin sulfateproteoglycan expression. The reason for this is reduction is unclear, although it may relate to the concentration of TG used and length of treatment. Importantly, however, we observed that even in the absence of chemically-induced ER stress, the levels of cellular chondroitin sulfate proteoglycans were reduced in U373 and U87 cells treated with OASIS siRNA. Thus, OASIS may be responsible for maintaining chondroitin sulfate proteoglycan expression even under control conditions in glioma cell lines that express this protein. This is possible given that glioma cells exhibitOASIS in Human Glioma CellsFigure 6. OASIS knock-down perturbs U373 cell migration. (A) U373 cells transfected with 100 nM control (GFP) or OASIS siRNAs for 72 h before scratching the 90 confluent cell monolayers. The cells were then incubated for the indicated times and DIC images were obtained. The approximate cell migration is indicated by the white 23727046 lines. The images were analyzed by ImageJ and the wound area was quantified (B) (*p,0.05; OASIS siRNA vs. control siRNA, n = 3). Note the almost complete healing of the wound in control cells and poor migration of cells in which OASIS was knocked-down. doi:10.1371/journal.pone.0054060.gmild activation of the UPR even under basal conditions (Figure 4D). Another extracellular matrix component shown to be regulated by OASIS in osteoblast cells is the collagen gene Col1a1 [16]. Interestingly, the induction of this gene by ER stress was not affected by OASIS knock-down, suggesting that OASIS is not required for Col1a1 induction in glioma cells or that perhaps another ATF6 family isoform is able to compensate for OASIS loss allowing for Col1a1 induction in glioma cells. The in vitro migration assay identified that OASIS silenced cells have poor migration efficiency. The exact mechanism of this effect is unknown, although presumably OASIS is required for maintaining ECM components that are required for efficient cell migration. Gliomas are characterized by aggressive growth and invasiveness, which is closely related to cell-ECM interactions [23,25,26,27,37,38,39]. Given that OASIS can be induced by ER stress and allows the cell to modulate its matrix we hypothesize that OASIS is likely to be beneficial under chronic, but low intensity ER stress, such as may occur under hypoxic conditions in vivo [22]. This will allow the cel.