Imaging markers of response to combined BRAF and MEK inhibition in BRAF mutated vemurafenib-sensitive and resistant melanomas

Abstract

A majority of patients with a V600x melanoma respond quickly to BRAF/MEK inhibition (BRAFi/MEKi) and have an obvious clinical benefit. Nearly all the patients after this initial phase will develop resistance. Therefore, non-invasive early markers of response/non-response are needed in order to identify those patients who, due to intrinsic or acquired resistance, do not respond to treatment and would be eligible for alternative treatments. The aim of this study was to investigate the value of magnetic resonance spectroscopy (¹H-MRS) of choline and diffusion-weighted magnetic resonance imaging (DW-MRI) as early markers of response to BRAF inhibition (BRAFi) with vemurafenib alone or in combination with MEK inhibition (MEKi) with trametinib, in BRAFi-sensitive and BRAFi-resistant melanoma xenografts. Tumor response was significantly improved by the combination of BRAFi and MEKi, compared to BRAFi alone, only in sensitive xenografts; thus indicating that vemurafenib-resistant A375R xenografts were cross-resistant to the inhibition of MEK, as confirmed by immunohistochemistry analysis for phosphorylated ERK. In vivo¹H-MRS showed that in sensitive melanoma xenografts, a significant blockage of ERK phosphorylation, but not a decrease in cell proliferation, was required to affect total choline (tCho) levels, thus suggesting that tCho could serve as a pharmacodynamic (PD) marker for agents targeting the MAPK cascade. In addition, early effects of the combination therapy on tumor cellularity could be detected via DW-MRI. In particular, skewness and kurtosis of the apparent diffusion coefficient (ADC) distribution may be useful to detect changes in the diffusional heterogeneity that might not affect the global ADC value.

Keywords: BRAF/MEK inhibitors; choline spectroscopy; diffusion-weighted MRI; melanoma; tumor response.

Figure 1. Growth delays and ERK phosphorylation in sensitive and resistant A375 xenografts

(A) Tumor growth delay was calculated as the mean time for tumors to reach twice their baseline volume. The growth of sensitive A375 xenografts was significantly delayed by both BRAFi as a single agent or in combination with the MEKi (p < 0.001, n = 3–4/group), with an evident benefit of the combination over BRAFi alone (p < 0.001 vs A375-BRAFi). Neither BRAFi alone nor the combination delayed the growth of resistant A375R xenografts (p = 0.16 and p = 0.65 respectively vs A375R-controls, n = 4–5/group). (B) Tumor growth curves in BRAFi-sensitive (upper graph) and BRAFi-resistant (lower graph) melanoma xenografts (n = 3–5/group). (C) Phosphorylation of ERK was significantly blocked in sensitive xenografts collected 4 hours after a single injection of BRAFi alone or in combination with MEKi (^*p < 0.05 vs A375-controls, n = 3/group). As expected, phospho-ERK decrease was not significant in BRAFi-treated A375R xenografts (p = 0.2458 vs A375R-controls, n = 3). However, the BRAFi-resistant xenografts did not retain sensitivity to the inhibition of MEK either (p = 0.1779 vs A375R-controls, n = 3). (D) Phospho-ERK representative staining of melanoma xenografts collected 4 hours after a single injection of the indicated treatments.

Figure 2. Combined BRAF and MEK inhibition decreases the tCho/H2O ratio in BRAFi-sensitive melanoma xenografts

The ratio of total choline peak (tCho) to (unsuppressed) water peak in tumors was measured at baseline, on day 2 and on day 5. Choline levels were affected by the BRAFi/MEKi combination in sensitive (A), but not in vemurafenib-resistant (B) xenografts (^*p = 0.0313 vs A375-baseline, n = 3–9/group). (C) Ki-67 staining was not significantly affected by either treatment (n = 3/group, two-way ANOVA followed by Bonferroni's test for multiple comparisons), whereas phospho-ERK levels (D) in sensitive A375 xenografts were significantly decreased by the combination as soon as 2 days post treatment initiation (^*p = 0.0128, n = 3–5/group). Representative Ki-67 (E) and phospho-ERK (F) staining of sensitive melanoma xenografts at baseline and after 2 or 5 days of treatment.

Figure 3. Effects of the BRAFi/MEKi combination on cell death

(A) Histograms of the ADC values of A375 xenografts obtained from DW-MRI of the same mice as in Figure 4B. The median value for day 0 (full line) and day 5 (dotted line) are shown. Note a clear shift to the left of the median over time in control tumors, a lack of left shift for the BRAFi condition, and a right shift for the combined BRAF and MEK inhibition, as well as a change in the shape of the histograms in treated groups. (B) In sensitive A375 xenografts (left), median apparent diffusion coefficient underwent opposite shifts in BRAFi/MEKi-treated A375 xenografts compared to controls (p = 0.0517 at day 2, p = 0.0333 at day 5 vs A375-controls, n = 4–9/group), whereas it was not affected by BRAFi alone. No significant trends were observed in A375R xenografts after treatment (right). (C) Necrotic area (bars, left y-axis) in A375 xenografts is represented as percent of total tissue section area (^*p = 0.0187 day 2 vs baseline, n = 3–5/group), whereas cleaved caspase 3 levels (symbols, right y-axis) are represented as mean staining index (^**p = 0.0082 day 2 vs baseline, n = 3–5/group). (D) Representative DW images overlaid with corresponding parametric ADC maps acquired at baseline (day 0) and after 5 days of BRAFi/MEKi treatment.

Figure 4. BRAFi/MEKi combination affects diffusional heterogeneity in sensitive melanoma xenografts

Skewness (A–C) and kurtosis (B–D) of the ADC distribution are represented as absolute differences from baseline (mean ± SEM). Both skewness (^*p = 0.0210 vs day 0, n = 4–9/group) and kurtosis (^*p = 0.0103 at day 2, p = 0.0229 at day 5 vs day 0, n = 4–9/group) decreased following combined BRAF/MEK inhibition in BRAFi-sensitive (A–B), but not in BRAFi-resistant (C–D) melanoma xenografts (n = 4/group).