BACKGROUND: Glioblastoma (GBM) remains uniformly lethal despite the progress of conventional therapies, which lack specificity and thus result in high toxicity to normal cells. Immunotargeting therapy such as recombinant immunotoxin (RIT) is an alternative and a Phase I clinical trial of MR1-1-PE38KDEL is undergoing at Duke University Medical Center. To avoid the production of neutralizing antibodies against toxin portion and to allow for repeat treatment cycles, a less immunogenic form of PE, LR-LO10, has been generated by silencing the human B cell epitopes on PE38 and removing the immunodominant T cell epitopes. We have previously developed an affinity-matured glycoprotein NMB (GPNMB)-specific RIT, F6V-PE38, which exhibits significant in vitro and anti-tumor activity against GBM and melanomas. Here we report a newly engineered F6V-LR-LO10 RIT for efficient, targeted therapy of GBM and other GPNMB-expressing cancer patients with normal immunity. METHODS: We have constructed an RIT, F6V-LR-LO10, by fusing a fully human F6V scFv, as VH-(G4S)3-VL, with a truncated form of Pseudomonas exotoxin A, LR-LO10. F6V-LR-LO10 was expressed in E. coli BL21 (DE3), and the inclusion bodies were reduced, refolded, and purified by ion exchange chromatography and gel filtration. The antigen binding of F6V-LR-LO10 was assessed by surface plasmon resonance and flow cytometric analysis. The cytotoxicity was measured by protein inhibition assay on GPNMB-expressing GBM and melanoma cells. The in vivo animal toxicity was tested in nude mice and anti-tumor activity will be assessed in human xenograft mouse models. RESULTS: The F6V-LR-LO10 protein was purified as a monomer of 51 kDa to 95% homogeneity. The F6V-LR-LO10 bound to GPNMB-coated chips with high affinity at KD = 1.1e-8 M. The F6V-LR-LO10 also bound to native GPNMB protein via flow cytometric analysis on GPNMB positive cell lines. We tested the in vitro activity of F6V-LR-LO10 on various glioma and melanoma xenograft cells and demonstrated cell-killing activity. The IC50 of F6V-LR-LO10 on GBM D392 cells was 1.5 ng/mL; the IC50 was 10 ng/mL and 15 ng/mL on melanoma DM443 and DM440 cells, respectively, and that was comparable to the IC50 of F6V-PE38, while negative control RIT did not kill those cells. The animal toxicity (MTD) was greatly reduced by >17 fold for F6V-LR-LO10 than F6V-PE38. CONCLUSIONS: Our experiments showed that a new fully human anti-GPNMB RIT, F6V-LR-LO10, is highly active with higher therapeutic index and is predicted to have low immunogenicity in humans. This next-generation RIT is suitable for further clinical development in therapy of GBM, melanoma, and other GPNMB-expressing malignancies. SECONDARY CATEGORY: Preclinical Experimental Therapeutics.