Mutant NRAS and KRAS lead to the activation of the RAS/RAF/MEK/ERK pathway in approximately 50% of multiple myeloma (MM). Blocking this pathway with MEK1/2 inhibitors (MEKi) such as trametinib (Tram) is a therapeutic option but the response rate in MM varies between 30-50% (Heuck et al, Leukemia 2015). In MM it is unknown whether RAS mutation status correlates with sensitivity to Tram. The purpose of this study was to characterize factors which predict response to Tram and to identify mechanism mediating resistance.
We established the IC50 of Tram using MTT assays in 32 MM cell lines (HMCL) including 16 RAS mutant positive (RASm+), and 15 wildtype RAS (RASm-), and 1 BRAF mutant (BRAFm+) line which acted as a positive control. HMCLs were classified according to the IC50 value as sensitive (<0.05μM); intermediate sensitive (IMS) (0.05 and 10μM); and resistant (>10μM). All lines underwent immunoblotting for pERK at baseline and following treatment with serial concentrations of Tram to identify correlation of activation with sensitivity. BrdU incorporation analyzed by FACS was performed to determine the molecular action of Tram. A lentiviral mediated expression system was used to engineer a MAF overexpressing cell line in a RASm+ HMCL lacking MAF (MMRASm+MAF) and silencing MAF in two lines with co-occurring MAF and RASm+ (MMRASm+shMAF). The clinical characteristics of 84 relapsed RASm+ patients who received Tram either as a single agent or in combination with other anti-MM therapies were also examined.
6/16 (37.5%) of RASm+ HMCLs were sensitive to Tram, 5/16 (31.1%) were IMS and 5/16 (31.1%) resistant. There was no difference in sensitivity to Tram between KRASm+ (IC50 = 9.5μM, n = 11) and NRASm+ (IC50 =12.5μM, n=4, p=0.65). In contrast, 13/15 (87%) RASm- HMCLs were resistant to Tram. Mechanistically, Tram blocked cell cycle progression in Tram-sensitive RASm+ cells with an increase in G0/G1 phase (22.25%) and a decrease in S phase (16.76%) compared with untreated controls (p<0.01). These in-vitro findings correlated with our clinical experience where there was no difference in progression free or overall survival (OS) between KRASm+ (n=46) or NRASm+ (n=38) patients treated with Tram.
Given that RASm+ lines exhibit a range of sensitivity (IC50 0.005-30μM), we further defined additional molecular modifiers that influence sensitivity by comparing KRASm+ alone with KRASm+ co-occurring with t(14;16) (n=4) or t(4;14) (n=3). The mean IC50 value of Tram (0.025μM) in 5 RASm+ was significantly lower than those with a co-occurring t(14;16) (IC50 =16μM; P<0.001) or a t(4;14) (IC50 =11.6μM, P<0.05). Among 15 RASm- HMCLs, the IC50 (37.5μM, n=9) of Tram was much higher in t(14;16) HMCL than in t(4;14) (IC50=9.16μM, n=3, P<0.01). A similar pattern was seen in patients, with a shorter OS when RASm+ co-occurred with t(14;16) or t(4;14) (n=17 OS=0.99yrs) than those with RASm+ alone (n= 67, OS=1.85yrs, P=0.01). These findings provide evidence that co-occurring MAF expression renders RAS+ MM less sensitive to MAKP inhibition.
To determine if activation of MAPK pathway caused by RAS mutation was associated with sensitivity to MEKi, we assess baseline and post Tram treatment levels of p-ERK1/2, a downstream substrate of MEK1/2 activation. MM HMCLs displayed variable basal levels of pERK. There was no correlation between basal levels of pERK and RASm+, with the highest baseline levels of pERK being seen in HMCLs with a t(14;16) or t(14;20) and wildtype RAS, intermediate levels in RASm+ with a t(14;16) and lowest in RASm+ alone. The concentrations of Tram for complete inhibition of pERK were higher in HMCLs with co-occurring t(14;16) ( >or = 500 nM) than for RASm+ alone.
To further confirm the ability of MAF to confer resistance of RASm+ cells to MEKi, we examined the sensitivity of MMRASm+/MAF to Tram.MTT assay showed that the IC50 of MM RASm+/MAF cells was 50% higher than that of MMEV cells (P <0.001). Similar results were observed in two other HMCLs with transient ectopic expression of MAF indicating that increase in MAF protein in RAS mutant cells reduces sensitivity to MEKi. In contrast, silencing of MAF in two RASm+ HMCLs significantly increased sensitivity (75% decrease in IC50) to Tram, compared with MMRASm+shCont (P<0.001).
Our results demonstrate that RASm+ MM is sensitive to Tram but the co-occurrence of t(14;16) leading to MAF overexpression confers resistance both in-vitro and in patients.
Morgan:Univ of AR for Medical Sciences: Employment; Celgene: Consultancy, Honoraria, Research Funding; Takeda: Consultancy, Honoraria; Bristol Meyers: Consultancy, Honoraria; Janssen: Research Funding. Davies:Celgene: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Janssen: Consultancy, Honoraria.
Asterisk with author names denotes non-ASH members.