BRAFV600E CRC cell lines are sensitive to the selective BRAF inhibitor vemurafenib and insensitive to the EGFR-inhibitor panitumumab
For our investigations, we selected colorectal carcinoma cell lines according to their mutational status (Table 1). In particular, mutations of the EGFR signaling pathway were considered, in the order of the following molecules: EGFR, RAS (including K-RAS, H-RAS, N-RAS), BRAF, TP53, and PI3K.
Two (Colo205, HT-29) harbored the BRAFV600E mutation with wild-type EGFR; the third (SW-48) was BRAF wild-type (wt) with an EGFR mutation.
Both BRAFV600E CRC cell lines are sensitive to the selective BRAF inhibitor vemurafenib and insensitive to the EGFR-inhibitor panitumumab in vitro
To explore their in vitro sensitivity to BRAF blockade, we first treated the three CRC cell lines with the selective BRAFV600E inhibitor, vemurafenib (VEM). As expected, the BRAF wt SW-48 cells were unaffected by the drug (Fig. 1A and 1B). In contrast, the BRAFV600E lines, Colo205 and HT-29, both displayed significantly reduced viability and increased death rates after exposure to VEM, which is consistent with previous reports . To investigate the in vitro sensitivity to EGFR blockade, the cell lines were then exposed to panitumumab (PAN), a fully human recombinant monoclonal IgG2 antibody highly specific against EGFR/ErbB-1/HER1. PAN is indicated for the treatment of metastatic CRC with wild-type KRAS . As shown in Figures 1C and D, however, PAN had no significant growth-limiting effects in any of the cell lines tested, regardless of their EGFR status.
Co-exposure to panitumumab may enhance the efficacy of vemurafenib in some BRAFV600E CRC cells
Since a synergy between EGFR and BRAFV600E inhibition was previously demonstrated [3,11], we investigated the possible combined action of vemurafenib and panitumumab in cell models tested.
As shown in Figure 2A and B, vemurafenib and panitumumab, administered singly or combined, had no effects on cell viability or cell death rates in the BRAF-wt/EGFR-mutant SW-48 line. As for the BRAFV600E CRC cells, the two drugs exerted synergic growth-limiting effects in the Colo205 line (Fig. 2C and D). In contrast, in HT-29 cells, the efficacy of the combination was not significantly different from that of vemurafenib alone (Figures 2E and F).
Our results from data presented in Figure 1 and 2 show that the combined targeting of EFGR and BRAFV600E can be beneficial in BRAFV600E cells.
BRAFV600E CRC cells respond to different doses of the pan-ErbB-family inhibitor afatinib
We then proceeded to explore whether a molecule with a broader action on ErbB family receptors could achieve better results with respect to the highly anti-EGRF specific panitumumab.
We evaluated the responses of the CRC cell lines to afatinib (AFA), a small-molecule receptor tyrosine kinase inhibitor (TKI), which irreversibly blocks signaling activity from all ErbB-family homo- and heterodimeric receptors . Afatinib is currently approved for first-line treatment of patients with EGFR-mutation-positive lung cancer .
Viability and death rates were assessed in all three CRC cell lines after exposure to AFA at concentrations ranging from 0.01µM to 10µM (log10 escalations). As expected SW-48 cells, which harbor an EGFR mutation targeted by afatinib (i.e., p. G719S), were sensitive to low doses of the drug (0.1 and 1 µM) (Figures 3A-B). The responses of the BRAFV600E cell lines—both of which are EGFR-wildtype—differed: in Colo205 cells significant changes in both viability and death rates were already evident after exposure to the lowest concentration tested (0.01 µM) (Figures 3C and D), whereas HT-29 cells were significantly affected only by the highest concentration used (10 µM) (Figures 3E and F).
Vemurafenib and afatinib produce additive growth-limiting effects in CRC BRAFV600E cell lines
We then wanted to investigate if the combination of vemurafenib with afatinib, which is able to target ErbB2 and ErbB4 receptors in addition to EGFR, could achieve better results than the combination of vemurafenib and panitumumab.
We then exposed the cells to vemurafenib and afatinib— the latter at doses of 1µM or 10µM (VEM + AFA1 and VEM + AFA10, respectively).
As shown in Fig. 4A and B, the addition of vemurafenib did not enhance the effect produced in SW-48 cells by exposure to AFA1 or AFA10 alone, which was consistent with these cells’ unresponsiveness to the BRAF inhibitor alone (Fig. 1). In the BRAFV600E lines (Colo205 and HT-29), the two drugs produced additive growth-limiting effects that were afatinib dose-dependent (Fig. 4C-F). In the Colo205 cells, which responded to both VEM and AFA when used alone (in terms of both decreased viability and increased death rates), the response to the combination VEM+AFA was stronger than those achieved with VEM+PAN (Fig. 2C and D), and these effects were even more evident when the higher dose of AFA was used (VEM+AFA10). As for the HT-29 cells, which were not significantly affected by VEM-PAN, the VEM-AFA1 combination was also effective, but the level of efficacy observed in Colo205 cells could be achieved only with the higher dose of AFA (VEM+AFA10).
ErbB2/HER2/Neu is differentially expressed in untreated BRAFV600E CRC cell lines
To explore possible mechanisms underlying the differential responses to the inhibition of the ErbB receptor family members of the BRAFV600E cell lines, we analyzed their expression levels of ErbB2/HER2/Neu, which is targeted by AFA but not by PAN. Unlike EGFR, ErbB2 is activated not by ligand binding, but as a result of its heterodimerization with other ErbB family members. Owing to ErbB2’s constitutive kinase activity, EGFR/ErbB2 heterodimers are more active than EGFR homodimers, but this interaction is limited by the relatively low-level of ErbB2 expression in normal cells . Mutations or amplifications in one of the four ERBB family genes are present in 22 out of 165 (13%) non-hypermutated and 16 out of 30 (53%) hyper-mutated cases . Interestingly, activating mutations and amplifications of ErbB2 account for 7% of cases (Cancer Genome Atlas Network) and patients with HER2-amplified metastatic CRC are less likely to respond to anti-EGFR therapy .
As shown in Figure 5A and B, ErbB2 expression in Colo205 cells was significantly higher than that found in HT-29 (or in the BRAFwt SW-48 cells). This molecular feature could conceivably explain the high sensitivity of Colo205 cells to afatinib shown in Figures 3 and 4. These findings suggest that ErbB2 expression levels in BRAFV600E CRC cells are an important predictor of their responsiveness to ErbB blockade, alone or with BRAF inhibition. Patients whose tumors express high levels of ErbB2 are likely to be sensitive to afatinib monotherapy or combined treatment with afatinib and vemurafenib.
Indeed, in clinical settings, vemurafenib alone proves to be insufficient in CRC tumors owing in many cases to feedback-mediated reactivation of RAS. ErbB2 (HER2-Neu) has been identified as a possible mediator of this feedback. Our experiments showed that the small molecule pan-ErbB family inhibitor afatinib effectively reduces the viability of CRC cells. Moreover, responses to afatinib treatment were heterogeneous. Based on the ErbB2 protein levels we demonstrated in the cell lines we used, it seems reasonable to speculate that high-level expression of ErbB2 might predict an effective response to AFA at lower doses. We therefore propose that combined treatment with vemurafenib and afatinib could be used for BRAFV600E CRCs, after confirming these in vitro results by testing the association in randomized clinical trials. Assessment of ErbB2 (HER2-Neu) expression levels in these tumors should also be used to inform treatment decision-making.