Production of Anti-EGFR Gold Nanoparticles for Raman Detection of Glioblastoma Tumor Cells

Maximizing the extent of resection has been shown to be crucial for improved survival in glioblastoma (GBM).¹ Among conventional imaging technologies applied in neurosurgery, 5-aminolevulinic acid (5-ALA) fluorescence-guided resection has helped in augmenting GBM resection, but it still faces notable limitations such as inhomogeneous fluorescence at the periphery of the tumor and false positive signal produced by reactive astrocytes and macrophages.² Moreover, in GBM cells, 5-ALA-induced fluorescence may vary depending on the epidermal growth factor receptor (EGFR) activation and the presence of its constitutively active variant, EGFRvIII.³ These factors complicate the intraoperative setting of tumor resection, leading to large degree of inaccuracy and resulting in tumor relapse due to residual cancer cells left at the edges of the resection cavity.¹ To achieve a sharper delineation between tumor and healthy brain tissue, we propose to use Raman microscopy as a tool for immediate intraoperative GBM-specific cell visualization. For increased sensitivity of Raman scattering, gold nanoparticles (GNPs) are used. When properly engineered, these GNPs provide an accurate mapping of tumor spread.^4,5

Nontoxic, Raman active GNPs were developed by the attachment of a Raman active molecule, and a polyethylene glycol shell. Further, GNPs were conjugated with anti-EGFR antibodies since almost 50% of GBM overexpress EGFR at the cell membrane. The specificity of particles binding was investigated in vitro by testing several incubation times and GNPs concentrations on human GBM cells expressing high to low EGFR levels: LN229wtEGFR, BS153 and U87MG, respectively; while mouse astrocytes, IMA2.1, were used as EGFR-negative cells. Immunofluorescence staining and Raman analysis were performed to visualize GNPs on GBM cells.

Cells incubation with 2*105 anti-EGFR GNPs for 4 hours was found to be the best condition to ensure a clear Raman detection of GBM cells, while minimizing nonspecific binding to the EGFR-negative cell line. With this setting, a comparable high Raman signal intensity was detected among the different GBM cell lines independently on the expression level of EGFR.

These results suggest that GNPs directed against EGFR, visualized with Raman imaging, may be implemented intraoperatively to identify and to remove residual GBM cells leading to better outcomes for patients. Therefore, these GNPs will be first tested in a GBM animal model.

Acknowledgments: This work is supported by Lega Cancro Ticinese, Fondation Gabriel Charitable, Fondo di Ricerca in Neuro-Oncologia