Erlotinib in gliomas: should selection be based on EGFR and Akt analyses?

Gliomas are the most common primary malignant neoplasm of the central nervous system in adults. Despite tangible progress, results remain disappointing, and median survival is 10 – 12 months ( 1 ) . Several new drugs have been developed that interfere with specifi c cellular targets in cancer cells. Among them, small-molecule inhibitors of the tyrosine kinase domain of the epidermal growth factor receptor (EGFR) were shown to be effective in the treatment of solid malignancies. Gefi tinib (Iressa; AstraZeneca, Wilmington, DE) and erlotinib (Tarceva; Genentech, South San Francisco, CA) are both orally active, selective EGFR tyrosine kinase inhibitors (EGFR-TKIs) that have produced impressive responses in a small subgroup of lung cancer patients ( 2 – 5 ), with a statistically signifi cant improvement in survival compared with best supportive care in the case of erlotinib ( 6 ) . EGFR represents a particularly attractive target in malignant gliomas because the receptor is expressed, amplifi ed, or mutated in most cases ( 7 – 9 ) . Moreover, because intracranial delivery of many agents is limited, small-molecule TKIs offer a pharmacologic advantage for brain malignancies ( 10 , 11 ) . A phase II trial, evaluating effi cacy and tolerability of gefi tinib in 53 recurrent glioblastoma patients not selected for any biologic characteristic showed that the drug may have modest activity ( 12 ) . Thirteen percent of patients had no disease progression at 6 months, suggesting a potential benefi t in a small, and so far undefi ned, subgroup of patients. In non – small cell lung cancer (NSCLC), mechanisms underlying TKI sensitivity have been described recently, allowing clinicians to select patients with the highest probability of a clinical response. The presence of activating EGFR gene mutations and/or EGFR gene amplifi cation has been shown to be associated with TKI sensitivity ( 13 – 15 ), particularly in patients with EGFR-dependent activation of the antiapoptotic phosphatidylinositol 3-kinase (PI3K) – Akt pathway ( 15 – 17 ) . In other diseases, the best procedure to select patients for EGFRTKI therapy remains unknown. No EGFR gene mutations have been described in glioma cell lines ( 13 ) or in specimens from glioma patients ( 13 , 18 ), indicating that these mutations have probably no role in the sensitivity of gliomas to EGFR-TKIs. In this issue of the Journal, Haas-Kogan et al. ( 19 ) reported their experience in 41 glioma patients treated with erlotinib within a phase I study. Erlotinib was administered alone or in combination with the alkylating agent temozolomide. In this study, no EGFR gene mutation was found, confi rming the limited role of EGFR gene sequencing as a predictive test for response to EGFR-TKIs in gliomas. Response to erlotinib was more frequently observed in patients with EGFR expression ( P = .07) or EGFR amplifi cation ( P = .08); in patients with glioblastoma multiforme, these associations were statistically signifi cant ( P = .03 and .02, respectively). Nevertheless, among the 10 patients with EGFR gene amplifi cation, six did not respond to the therapy, suggesting that other mechanisms are involved. In lung cancer studies ( 15 – 17 ), activation of the PI3K – Akt pathway has been shown to play an important role in EGFR-TKI sensitivity. Therefore, the statistically signifi cant inverse association between Akt phosphorylation and response to erlotinib reported in this study ( 19 ) is surprising. None of the 22 tumors expressing phosphorylated Akt responded to the drug, whereas all tumors responding to the drug lacked phosphorylated Akt. Based on these fi ndings, the authors concluded that, in patients with glioblastoma multiforme, high levels of EGFR expression and low levels of phosphorylated Akt are associated with a better response to erlotinib therapy. How then should we select glioma patients for EGFR-TKI therapy? The association of EGFR gene amplifi cation and expression with response to erlotinib in glioblastoma multiforme is probably the most interesting fi nding of this study. These fi ndings are corroborated by a recent study ( 16 ) of NSCLC patients reporting that EGFR gene amplifi cation, detected by fl uorescence in situ hybridization, and EGFR expression, detected by immunohistochemistry, were statistically signifi cantly associated with improved response and survival. Unfortunately, the small number of patients and the potentially confounding effect of temozolomide therapy do not allow fi rm conclusions to be drawn. Statistical analysis was conducted in the whole population regardless of whether patients did or did not receive temozolomide therapy, and the effect of chemotherapy on response was not considered. Although pharmacokinetic analysis showed that temozolomide did not modify the level of erlotinib in plasma ( 19 ), we cannot exclude a positive effect of temozolomide in responding patients treated with both drugs. The inverse association with Akt activation and the evidence that six of 10 patients with EGFR gene amplifi cation did not respond to the therapy suggest that EGFR is critical for glioma cell survival in only a small fraction of patients. Because specimens were collected at the time of primary diagnosis or of resection of recurrent disease, it is also not possible to assess the impact of previous or concurrent therapies on the EGFR pathway activations ( 20 ) . Therefore, we cannot exclude the possibilities that Akt was only apparently not phosphorylated in the four responding tumors with EGFR gene amplifi cation and that the PI3K – Akt pathway remained inactive or Akt activation was not EGFR dependent in the six tumors with EGFR amplifi cation that were refractory to the therapy. Preclinical data showed that Akt could be activated by a non – EGFRdependent mechanism, such as loss of PTEN or insulin-like growth factor receptor 1 (IGFR-1) gene expression. Data from cell lines suggest ( 21 ) that IGFR-1 mediates resistance to antiEGFR therapy through continued activation of the PI3K – Akt pathway. Although, in this study ( 19 ), the authors determined