Longitudinal heterogeneity in glioblastoma: moving targets in recurrent versus primary tumors

Background: Molecularly targeted therapies using receptor inhibitors, small molecules or monoclonal antibodies are routinely applied in oncology. Verification of target expression should be mandatory prior to initiation of therapy, yet, determining the expression status is most challenging in recurrent glioblastoma (GBM) where most patients are not eligible for second-line surgery. Because very little is known on the consistency of expression along the clinical course we here explored common drug targets in paired primary vs. recurrent GBM tissue samples. Methods: Paired surgical tissue samples were derived from a homogeneously treated cohort of 34 GBM patients. All patients received radiotherapy and temozolomide chemotherapy. Verification of common drug targets included immunohistological analysis of PDGFR-beta, FGFR-2, FGFR-3, and mTOR-pathway component (phospho-mTOR(Ser2448)) as well as molecular, MLPA-based analysis of specific copy number aberrations at the gene loci of ALK, PDGFRA, VEGFR2/KDR, EGFR, MET, and FGFR1. Results: Paired tumor tissue exhibited significant changes of expression in 9 of the 10 investigated druggable targets (90%). Only one target (FGFR1) was found unchanged, since dissimilar expression was observed in only one of the 34 paired tumor tissue samples. All other targets were variably expressed with an 18-56% discordance rate between primary and recurrent tissue. Conclusions: The high incidence of dissimilar target expression status in clinical samples from primary vs. recurrent GBM suggests clinically relevant heterogeneity along the course of disease. Molecular target expression, as determined at primary diagnosis, may not necessarily present rational treatment clues for the clinical care of recurrent GBM. Further studies need to analyze the therapeutic impact of longitudinal heterogeneity in GBM.