Malignant glioma and its potential to serve as a molecular target. We have previously demonstrated selective cleavage of GRP78 in EGFR-positive prostate and breast cancer cells exposed to EGF-SubA; thereby confirming the receptor-binding activity of the EGF moiety and the BTZ043 chemical information proteolytic activity of the SubA moiety [16]. We now extend these studies to explore the potential of both the fusion protein EGF-SubA and the SubA toxin alone to cleave GRP78 in glioblastoma models. As demonstrated in Fig. 2A, EGFSubA demonstrated potent proteolytic activity, cleaving GRP78 atconcentrations ranging from 0.5 to 2.5 pM in established glioblastoma cell lines (U251 and T98G) and the glioblastoma neural stem (GNS) cell line G179. These concentrations were over 20 fold lower when compared to the SubA toxin alone, which required approximately 50 pM to induce GRP78 cleavage, confirming increased potency of the fusion protein EGF-SubA. Time course studies demonstrated maximal cleavage of GRP78 within 16 h of EGF-SubA exposure (Fig. 2B). Conversely, cleavage of GRP78 in normal human astrocytes 15481974 (NHA) required significantly higher concentrations of EGF-SubA when compared to the glioblastoma cell lines, supporting the tumor specificity of this approach. Interestingly, the glioblastoma cell line U87 required considerably higher concentrations of EGF-SubA and SubA toxin to induce GRP78 cleavage. As an initial investigation, based on the mechanism of action of EGF-SubA, we performed western blot analysis to 
determine if the relative expression of EGFR or GRP78 could contribute to the observed differential response of EGF-SubA. As demonstrated in Fig. S2, expression levels of these proteins did not appear to be significantly different between the cancer cell lines tested. As we reported earlier, EGF-SubA induced toxicity is EGFR-dependent, but does not directly correspond to the EGFR expression level, reflecting a more complex cell-specific process of EGFR-mediated internalization and trafficking, as well as the magnitude of ER stress and UPR signaling in a particular cell line16. Therefore, the factors contributing towards relative sensitivity and resistance to EGF-SubA remain an active area of investigation. We went on to determine the influence of EGF-SubA induced cleavage of GRP78 on UPR activation. As describe above, the primary three mediators involved in UPR signaling includeTargeting the UPR in Glioblastoma with EGF-SubAPERK, Ire1, and ATF6. Upon stress, PERK is released from GRP78 to permit homodimerization, autophosphorylation and pathway activation. Similarly Ire1 is activated by dimerization, leading to trans-autophosphorylation; however, pathway activation does not entail a conventional cascade of sequential kinase activation, rather, activation of a cytosolic endoribonuclease activity whose only know substrate is X-box binding protein-1 (Xbp1) mRNA. This alters the Xbp1 translational reading frame leading to activation of a unique UPR specific program. The third mediator, ATF6, is concomitantly released from GRP78, permitting its transport to the Golgi compartment where it is cleaved to generate the cytosolic activated form of ATF6 that translocates to the nucleus [4,6]. In our studies, all three pathways were activated in U251 cells following exposure to EGF-SubA, as determined by PERK phosphorylation (Fig. 2D), nuclear Solvent Yellow 14 localization of cleaved ATF6 (Fig. 2C), and splicing of Xbp1 mRNA (Fig. 2E). However, the EGF-SubA concentrations required to induce X.Malignant glioma and its potential to serve as a molecular target. We have previously demonstrated selective cleavage of GRP78 in EGFR-positive prostate and breast cancer cells exposed to EGF-SubA; thereby confirming the receptor-binding activity of the EGF moiety and the proteolytic activity of the SubA moiety [16]. We now extend these studies to explore the potential of both the fusion protein EGF-SubA and the SubA toxin alone to cleave GRP78 in glioblastoma models. As demonstrated in Fig. 2A, EGFSubA demonstrated potent proteolytic activity, cleaving GRP78 atconcentrations ranging from 0.5 to 2.5 pM in established glioblastoma cell lines (U251 and T98G) and the glioblastoma neural stem (GNS) cell line G179. These concentrations were over 20 fold lower when compared to the SubA toxin alone, which required approximately 50 pM to induce GRP78 cleavage, confirming increased potency of the fusion protein EGF-SubA. Time course studies demonstrated maximal cleavage of GRP78 within 16 h of EGF-SubA exposure (Fig. 2B). Conversely, cleavage of GRP78 in normal human astrocytes 15481974 (NHA) required significantly higher concentrations of EGF-SubA when compared to the glioblastoma cell lines, supporting the tumor specificity of this approach. Interestingly, the glioblastoma cell line U87 required considerably higher concentrations of EGF-SubA and SubA toxin to induce GRP78 cleavage. As an initial investigation, based on the mechanism of action of EGF-SubA, we performed western blot analysis to determine if the relative expression of EGFR or GRP78 could contribute to the observed differential response of EGF-SubA. As demonstrated in Fig. S2, expression levels of these proteins did not appear to be significantly different between the cancer cell lines tested. As we reported earlier, EGF-SubA induced toxicity is EGFR-dependent, but does not directly correspond to the EGFR expression level, reflecting a more complex cell-specific process of EGFR-mediated internalization and trafficking, as well as the magnitude of ER stress and UPR signaling in a particular cell line16. Therefore, the factors contributing towards relative sensitivity and resistance to EGF-SubA remain an active area of investigation. We went on to determine the influence of EGF-SubA induced cleavage of GRP78 on UPR activation. As describe above, the primary three mediators involved in UPR signaling includeTargeting the UPR in Glioblastoma with EGF-SubAPERK, Ire1, and ATF6. Upon stress, PERK is released from GRP78 to permit homodimerization, autophosphorylation and pathway activation. Similarly Ire1 is activated by dimerization, leading to trans-autophosphorylation; however, pathway activation does not entail a 
conventional cascade of sequential kinase activation, rather, activation of a cytosolic endoribonuclease activity whose only know substrate is X-box binding protein-1 (Xbp1) mRNA. This alters the Xbp1 translational reading frame leading to activation of a unique UPR specific program. The third mediator, ATF6, is concomitantly released from GRP78, permitting its transport to the Golgi compartment where it is cleaved to generate the cytosolic activated form of ATF6 that translocates to the nucleus [4,6]. In our studies, all three pathways were activated in U251 cells following exposure to EGF-SubA, as determined by PERK phosphorylation (Fig. 2D), nuclear localization of cleaved ATF6 (Fig. 2C), and splicing of Xbp1 mRNA (Fig. 2E). However, the EGF-SubA concentrations required to induce X.
