Sir,

We read with great interest the Letter to the Editor from Shimokawa et al (2011) that reported significant associations between the expressions of the insulin-like growth factor type I receptor (IGF-IR) and those of E-cadherin and γ-catenin in non-small cell lung cancer (NSCLC) biopsies (Shimokawa et al, 2011). These data reproduce previous observations from our group. We have also observed a correlation between the expressions of IGF-IR and E-cadherin, particularly in NSCLC tumours with a high degree of differentiation (Gualberto et al, 2010). Furthermore, using unsupervised Bayesian clustering of epithelial-to-mesenchymal transition (EMT)- and IGF-IR-related markers, we identified three NSCLC subsets that resembled the steps of the EMT and we named epithelial-like, transitional-like and mesenchymal-like (Gualberto et al, 2010). Several markers of the IGF-IR pathway such as nuclear insulin receptor substrate-1 were overexpressed in the transitional subset and a higher objective response rate to the combination of chemotherapy and the anti-IGF-IR antibody figitumumab was observed in patients with transitional tumours (Gualberto et al, 2010). Thus, we agree with Shimokawa et al (2011) that analysis of tumour EMT status may contribute to a better understanding of the sensitivity to anti-IGF-IR therapy.

Shimokawa et al (2011) also pointed out that we have reported an inverse correlation in NSCLC patients between serum IGF-1 and tumour E-cadherin levels (Gualberto et al, 2011). We speculate that the fact that tumour E-cadherin is directly correlated with tumour IGF-IR and inversely correlated with serum IGF-1 may indicate differential functions for the non-activated (low or no ligand) vs the activated (in the presence of IGFs) IGF-1 receptor. Evidence supports the involvement of the IGF-IR in both differentiation and cellular growth and metastasis, with engagement of the receptor by IGFs-inducing neo-expression of mesenchymal markers, E-cadherin downregulation and cell migration (reviewed in Julien-Grille et al, 2005). These observations may have therapeutic implications. Of note, our key finding that high serum (free) IGF-1 levels may identify a subset of NSCLC patients who preferentially benefit from anti-IGF-IR therapy has now been independently corroborated by other groups (Goto et al, 2011; Ramalingam et al, 2011).