Expression of end-binding protein 1 (EB1), a potential response-predictive biomarker for lisavanbulin, in glioblastoma and various other solid tumor types.

Abstract

3118 Background: EB1, a protein located on the plus-ends of microtubules is involved in microtubule function and has been associated with glioblastoma (GBM) stem-cell-ness and more aggressive disease. Lisavanbulin (BAL101553) is a prodrug of the lipophilic small molecule BAL27862, that promotes tumor cell death by modulating the spindle assembly checkpoint and has been shown in rodents to efficiently penetrate the brain. Data from GBM mouse models and recent phase 1 clinical data (Lopez et al. ESMO 2020) suggest that EB1 is a response-predictive marker for lisavanbulin in GBM. A phase 2 study is ongoing to confirm this hypothesis (NCT02490800). A proof-of-concept in GBM would support an expansion of EB1-directed lisavanbulin clinical development in non-GBM tumors, which requires prevalence estimates of EB1-positivity in non-GBM tumor types. Methods: Tissue samples from GBM and other tumor types were stained for EB1 using a CE-marked immunohistochemistry Clinical Trial Assay (Targos Molecular Pathology GmbH, Kassel Germany). EB1-positivity was assessed by a board-certified pathologist based on the percentage of tumor cells showing moderate or strong staining for EB1, using thresholds of ≥50%, ≥60% and ≥70% of tumor cells with EB1 positivity. Whole transcriptome sequencing (WTS) using RNAseq was performed in a subset of tissue samples to develop a potential RNA-based predictive response signature for lisavanbulin. Results: 73 GBM tissue samples and 333 tissue samples from 13 other cancer types were stained for EB1. The strongest overall signal for EB1-positivity was obtained for medulloblastoma, neuroblastoma and GBM. In addition, moderate or strong EB1-staining in ≥50% of tumor cells was observed in samples from colorectal cancer (CRC), non small-cell lung cancer (NSCLC), metastatic melanoma, small-cell lung cancer (SCLC) and triple-negative breast cancer (TNBC). An expanded staining campaign is ongoing in these cancer types. Initial results from the ongoing WTS analyses show marked differences in gene expression profiles between EB1-positive and -negative cases. Conclusions: Strong EB1-positivity is infrequent but occurs in a variety of tumor types, with the strongest signals in medulloblastoma, neuroblastoma and GBM. A phase 2 study is ongoing to assess prospectively whether EB1 is a response-predictive biomarker for lisavanbulin in GBM.[Table: see text]