Generation and characterization of MEK and ERK inhibitors- resistant non-small-cells-lung-cancer (NSCLC) cells

The RAS/RAF/MEK/ERK pathway is one of the most downregulated pathway in cancer [1‐4].

Among the proteins belonging to this pathway, KRAS is the only one for which no drugs directly targeting its activity are available. B-RAF, MEK and ERK inhibitors are available and some already marketed for specific indications. Vemurafenib (B-RAF inhibitor) has been approved for the treatment of late stage and BRAF mutated melanoma as well as for the treatment of Erdheim-Chester Disease (a rare histiocytosis marked by recurrent BRAFV600E mutations in more than half of patients). Another B-RAF inhibitor, dabrafenib received approval for the treatment of BRAF mutated melanoma while trametinib, a MEK inhibitor has been approved for the same indication, as well as for the treatment of B-Raf mutated NSCLC, both as single agent and in combination with dabrafenib. For ERK, an inhibitor, Ulixertinib, has recently entered clinical investigation as first in class drug [5].

Both MEK and, particularly, B-RAF inhibitors have shown activity in NSCLC tumors with altered expression of mutational status of their target, although in several instances the good response was limited by the appearance of acquired resistance [2, 6‐9].

Several mechanisms associated to resistance development have been reported [10‐17]. They go to changes in target splicing, upregulation, changes in phosphorylation of downstream proteins, different metabolism modulation.

To try to better elucidate the mechanisms associated to resistance to RAS/RAF/MEK/ERK pathway inhibitor and to try to overcome the resistance, it is mandatory to have preclinical models of resistance.

Being MEK/ERK inhibitors a reasonable option for the treatment of NSCLC [7], we developed, starting from a human NSCLC cell lines sensitive to MEK and ERK inhibitors, two distinct sublines resistant to MEK-162 (MEK inhibitor) and SCH772984 (ERK inhibitor), respectively. Here we report the initial molecular and pharmacological characterization of these cells, which will be helpful for the development of alternative treatments or combinations.

Almost invariably, drugs specifically targeting kinases (TKI) in clinical use developed resistance, which strongly affects their potential use. As examples, drugs targeting EGFR kinases (gefitinib, erlotinib or afatinib), both as reversible or irreversible ATP competitors, while producing very high response in patients with EGFR mutated NSCLC, with a significant increase in Progression Free Survival (PFS), the majority of patients relapses with a tumor no more responsive to these drugs [19‐21]. The same is true for ALK inhibitors [22]. In both cases several mechanisms of resistance have been proposed, the main being the development of secondary mutations altering the binding of the drugs to the ATP-binding pockets of the kinase domain of the targets [23‐25]. Fortunately, for these drugs, third and fourth generations compounds have now been synthetized and have shown activity in first generation TKIs-resistant tumors thus significantly increasing survival of patients with NSCLC [26, 27]. Similarly, for other tumors, B-RAF inhibitors (such as vemurafenib) showed impressive complete responses in melanoma patients which are challenged by the appearance of relapsing tumors resistant to these drugs [10, 16, 28] .

From these and other examples, it is therefore important to have models to study drug resistance for new drugs approaching the clinic. Here we studied in particular MEK and ERK inhibitors the first, in clinical use for the treatment of NSCLC [29, 30] and the second just entering initial clinical evaluation [5]. We were able to select, starting from a human NSCLC cell line sensitive to both ERK and MEK inhibitors, two cell lines with a stable resistance to the drugs. The cell lines showed cross-resistance one to each other but not to drugs targeting the same pathway upstream (such as sorafenib). It is not surprising that MEK and ERK inhibitors are cross-resistant each other, being their modulation of downstream targets (S6 and RSK) similar, however, we do not have, at present, an explanation for the lack of cross resistance with the B-RAF inhibitor sorafenib. In fact, it would have been expected cross resistance also for this drug. We can only speculate that inhibiting B-RAF could lead to (in addition to the canonical downstream proteins) alteration in other pathways not altered in H727/MEK and H727/SCH cells, thus justifying its activity. We can not exclude that sorafenib could act in these cells with a mechanism not related to B-RAF inhibition.

The two cell lines generated here did not show cross-resistance with drugs acting on parallel pathways such as the PI3K/akt/mTOR being the three inhibitors tested for the three main players of the pathway (i.e. PI3K, akt, mTOR) similarly active against sensitive and resistant cells. The resistant cells seem also not to have defects in DNA repair, at least as judged from the results obtained with the DNA damaging agents or DNA repair inhibitors tested here. A special note should be done for doxorubicin. This drug, in fact, showed a certain degree of cross-resistance in the H727/SCH cell line but not in the H727/MEK cell lines. We attribute this to the evidence that H727/SCH cells have a higher expression of MDR-1 gene compared to both parental H727 and H727/MEK cells. Being doxorubicin a drug substrate for MDR-1 the results are not surprising. We do not have an explanation for the higher MDR-1 expression observed in the H727/SCH cells. What we can reasonably exclude is that SCH772984 is a substrate of MDR-1. In fact, the parental H727 cells already express high levels of MDR-1, significantly higher than other NSCLC cell lines. However, H727 cells are very responsive in vitro to SCH772984 while the other NSCLC cell lines are not (both for A549 and H460 cells the IC50 of SCH772984 is > 10.000 nM, compared to the value of 135 nM in H727 cells). In addition, the similar downregulation of ERK phosphorylation shown in the present manuscript at equimolar concentration of SCH772984 in sensitive and resistant cells, indicates that the drug efficiently reaches the target inside the cells.

At molecular levels, we showed evidence that in resistant cells (both to MEK and ERK inhibitors) drug treatment did not induce a decrease in S6 phosphorylation, an event mediated by the RSK kinase, while this was clearly observable in parental cells. The involvement of S6 in resistance was recently shown in melanoma cells resistant to MEK inhibitors [12]. In these cells, RSK phosphorylation was unaffected while here, treatment of both H727/SCH and H727/MEK did not result in decreased phosphorylation of RSK (which was evident in parental H727 cells) suggesting that the cellular context can induce different molecular alterations.

In conclusion, we have generated two NSCLC sublines resistant to MEK and ERK inhibitors, which show no cross-resistance to BRAF inhibitors and no cross-resistance to the agents inhibiting the parallel pathway PI3K/akt/mTOR. A lack of downregulation of RSK mediated S6 phosphorylation was observed in both resistant cells after treatment suggesting that this, as already shown in melanoma cells could be an important determinant of sensitivity to MEK and ERK inhibitors.

These cells, which show class-specific resistance, will be important for the development of new agents and combination able to circumvent resistance to MEK and ERK inhibitors. Although we did not test yet if the in resistance obtained in vitro can be confirmed in vivo too, we have several examples showing that MDR-1 unrelated drug resistance (as in this case) is maintained when cells are implanted in vivo [18, 31, 32]. Being ERK inhibitors at their initial clinical testing [5], the availability of models to study resistance in vitro and in vivo is of particular relevance.