Abstract Glioblastoma (GBM) complexity and heterogeneity requires treatment that addresses those pathobiological features. We have been developing selective cytotoxic agents able to target at the same time several GBM-associated factors. The chosen targets are responsible for the disease progression and/or recurrence as well as for resistance to the existing therapies. As a proof of concept, a Phase I clinical trial of a cytotoxic cocktail targeting IL-13RA2 and EphA2 receptors demonstrated dramatic anti-tumor responses in dogs with spontaneous gliomas that represent the closest translational model to human disease. To this end, we have developed multivalent agents that target four receptors specific to GBM: IL-13RA2, EphA2, EphA3, and EphB2, the combined expression of which covers virtually the whole tumor microenvironment. We have designed a multivalent protein termed QUAD 3.0 that contains an IgG1 scaffold, ephrinA5, which is a ligand for the EphA2, EphA3, and EphB2 receptors, and IL-13.E13K, a mutated version of interleukin 13 (IL-13), which binds preferentially to IL-13RA2. In QUAD 3.0, there is a cysteine at the C-terminal end of the protein to allow site-specific conjugation to cytotoxic cargo. QUAD 3.0 bound effectively to the four receptors in vitro and in vivo. QUAD 3.0 was conjugated to a modified form of Pseudomonas Exotoxin A (PE38QQR) and highly potent DNA binding agents based on modified doxorubicin (WP936, WP1737 and WP1244). All conjugates were highly cytotoxic to established and primary GBM cells with IC50s < 50 nM. We also treated the first dogs with QUAD 3.0-PE38QQR and QUAD 3.0-WP936 at a dose of 1.6 mg/ml using real-time monitored convection-enhanced delivery and observed up to 60% of tumor shrinkage and long-term survival. Thus, multivalent targeted agents demonstrate highly promising anti-tumor activity as single pharmaceutical, off-the-shelf agents. We also expect that our targeted drug candidates produce immune responses against tumors amplifying their cytolytic anti-tumor effect