Canertinib induces ototoxicity in three preclinical models

Highlights

• Canertinib, a new cancer drug candidate blocking ERBB signaling, causes toxicity in hair cells of zebrafish lateral line.

• Canertinib ototoxicity is confirmed in two mouse models.

• One major target of its ototoxicity in mice is outer hair cells.

• Ototoxicity should be monitored for cancer drug candidates with pan-ERBB inhibition properties.

Abstract

Neuregulin-1 (NRG1) ligand and its epidermal growth factor receptor (EGFR)/ERBB family regulate normal cellular proliferation and differentiation in many tissues including the cochlea. Aberrant NRG1 and ERBB signaling cause significant hearing impairment in mice. Dysregulation of the same signaling pathway in humans is involved in certain types of cancers such as breast cancer or non-small cell lung cancer (NSCLC). A new irreversible pan-ERBB inhibitor, canertinib, has been tested in clinical trials for the treatment of refractory NSCLC. Its possible ototoxicity was unknown. In this study, a significant dose-dependent canertinib ototoxicity was observed in a zebrafish model. Canertinib ototoxicity was further confirmed in two mouse models with different genetic backgrounds. The data strongly suggested an evolutionally preserved ERBB molecular mechanism underlying canertinib ototoxicity. Thus, these results imply that clinical monitoring of hearing loss should be considered for clinical testing of canertinib or other pan-ERBB inhibitors.

Introduction

Lung cancer is the number one cause of cancer mortality globally with few screening and treatment options available (Jemal et al., 2011). Non-small cell lung cancer (NSCLC) accounts for 80% of all lung cancer cases (Jemal et al., 2010).

Advanced NSCLCs (Stage IIIb and IV) are treated with cisplatin or combination therapies; however, the five-year survival rate is under 5% (de Castria et al., 2013). Cisplatin, as first-line treatment of many solid tumors, has been extensively studied for its ototoxicity (Lobarinas et al., 2013, Harrison et al., 2014), however, its effectiveness against NSCLC is often subverted (Lipp and Hartmann, 2005, Yorgason et al., 2006, Schacht et al., 2012). Recent research into novel anti-NSCLC drugs has focused on inhibition of the epidermal growth factor receptors (EGFRs/ERBB), which belong to the receptor family of tyrosine kinases (RTKs).

RTK inhibitors, which block ERBB receptors, present novel and exciting therapies for NSCLC (Cowan-Jacob, 2006, Pytel et al., 2009, Awada et al., 2012, Ou, 2012, de Castria et al., 2013, Majem and Pallares, 2013). Most ERBB family members have three domains: an extracellular N-terminus responsible for ligand binding, a transmembrane hydrophobic region, and a conserved intracellular domain with a ligand-activated tyrosine kinase on the C-terminal end (Allen et al., 2003). The ERBB family of receptors includes ERBB1 (EGFR), ERBB2 (Her-2/neu), ERBB3 (Her-3), and ERBB4 (Her-4) (Janne et al., 2007). Aberrant signaling of ERBB family members (Hynes and Stern, 1994, Leung et al., 1997) plays an important role in many cancers including breast (Sainsbury et al., 1985), ovarian (Meden and Kuhn, 1997), prostate (Schwartz et al., 1999), gastric (Hirono et al., 1995), laryngeal (Almadori et al., 1999), and indicate a more aggressive cancer phenotype (Klijn et al., 1994, Zinner et al., 2007). ERBB1 is overexpressed in the majority of NSCLC and is found frequently in lung squamous cell carcinoma (57–100%) whereas ERBB2 is found in lung adenocarcinomas (33%).

Gefitinib and erlotinib are first generation EGFR RTK inhibitors that result in 10–20% lung tumor suppression but do not inhibit other ERBB family members (Wakeling et al., 2002, Barbacci et al., 2003). After treatment with first generation RTK inhibitors, 50% of patients have a secondary missense T790 mutation resulting in acquired resistance due to steric hindrance with a bulkier methionine (Pao et al., 2005, Shepherd et al., 2005, Engelman and Janne, 2008, Sequist et al., 2011, Majem and Pallares, 2013). Although no preclinical studies of their ototoxicity, one clinical case was reported with erlotinib ototoxicity (Koutras et al., 2008). The success of this first generation inhibitors coupled with the known acquired resistance has resulted in a need for new RTK inhibitor therapies with more sustained effects than gefitinib and erlotinib (Han et al., 2008, Butts et al., 2010, Mukherjea et al., 2011, Allen et al., 2012, Yang et al., 2013, Agustoni et al., 2014). Second and third generation ERBB inhibitors are irreversible, have higher binding affinities, inhibit multiple ERBB family members, and have better activity against T790 mutation. Therefore, these new inhibitors, such as canertinib, may provide additional clinical benefits by functioning as a pan-ERBB inhibitor (Fry, 2000, Slichenmyer et al., 2001, Allen et al., 2003, Janne et al., 2007). Although adverse effects such as diarrhea, asthenia, stomatitis, and persistent rash were observed in canertinib clinical trials (Janne et al., 2007), no clinical studies of ototoxicity for these new inhibitors are reported so far.

The NRG1-ERBB signaling pathway plays a significant role in the development and mature function of the inner ear. In the cochlea, NRG1 is expressed in spiral ganglion neurons (SGNs) (Morley et al., 1998, Bao et al., 2003, Bao et al., 2004); whereas hair cells, Schwann cells, and supporting cells express ERBB (Morley, 1998, Hansen et al., 2001, Zhang et al., 2002, Hume et al., 2003). Cochlear innervation is abnormal in NRG1 and ERBB2 null mice and these adult mice have progressive hearing loss (Lee et al., 1995, Meyer and Birchmeier, 1995, Fritzsch et al., 1997, Morley, 1998, Adlkofer and Lai, 2000, Chen et al., 2003, Fritzsch et al., 2004, Hellard et al., 2004, Stankovic et al., 2004, Morris et al., 2006). One chemical compound blocking ERBB signaling could lead to progressive hearing loss in mature female pigmented guinea pigs (Watanabe et al., 2010). Since canertinib can block ERBB receptors, we first assessed possible hair cell toxicity of canertinib in a zebrafish model, and then in two mouse models (C57BL/6J and CBA/CaJ) by auditory brainstem recording (ABR) thresholds, distortion product otoacoustic emissions (DPOAE), and hair cell counts. Canertinib resulted in significant ototoxicity in all three preclinical animal models.