Evolving Acquired Vemurafenib Resistance in a BRAF V600E Mutant Melanoma PDTX Model to Reveal New Potential Targets

Abstract

Malignant melanoma is challenging to treat, and metastatic cases need chemotherapy strategies. Targeted inhibition of commonly mutant BRAF V600E by inhibitors is efficient but eventually leads to resistance and progression in the vast majority of cases. Numerous studies investigated the mechanisms of resistance in melanoma cell lines, and an increasing number of in vivo or clinical data are accumulating. In most cases, bypassing BRAF and resulting reactivation of the MAPK signaling, as well as alternative PI3K-AKT signaling activation are reported. However, several unique changes were also shown. We developed and used a patient-derived tumor xenograft (PDTX) model to screen resistance evolution in mice in vivo, maintaining tumor heterogeneity. Our results showed no substantial activation of the canonical pathways; however, RNAseq and qPCR data revealed several altered genes, such as GPR39, CD27, SLC15A3, IFI27, PDGFA, and ABCB1. Surprisingly, p53 activity, leading to apoptotic cell death, was unchanged. The found biomarkers can confer resistance in a subset of melanoma patients via immune modulation, microenvironment changes, or drug elimination. Our resistance model can be further used in testing specific inhibitors that could be used in future drug development, and combination therapy testing that can overcome inhibitor resistance in melanoma.

Keywords: BRAF V600E; PDTX; acquired resistance; melanoma; patient-derived tumor xenograft model; preclinical resistance model; resistance evolution.

Figure 1

Tumor growth of melanoma PDTX model over 6 generations of mice.

Figure 2

Gene expression patterns in untreated and treated BRAF V600E melanoma PDTX model. (A) The variance among samples was analyzed in primary component analysis, showing distinct groups of the untreated (black), generation 3 (green), and generation 5 (red) samples. (B) Expression pattern of 3991 differentially expressed RNAs versus the untreated samples. The x-axis represents expression in generation 5, and the y-axis the expression in generation 3 tumors.

Figure 2

Gene expression patterns in untreated and treated BRAF V600E melanoma PDTX model. (A) The variance among samples was analyzed in primary component analysis, showing distinct groups of the untreated (black), generation 3 (green), and generation 5 (red) samples. (B) Expression pattern of 3991 differentially expressed RNAs versus the untreated samples. The x-axis represents expression in generation 5, and the y-axis the expression in generation 3 tumors.

Figure 3

Relative mRNA expression changes of melanoma PDTX tissue upon vemurafenib treatment. All panels show relative expressions, normalized to control RPLP0. All measurements were analyzed in a common study to ensure the common threshold values for cT determination. Gene expressions (GSTA1, FKBP1A-SDCBP2, GZMB, AGAP9, CP, GNRHR) that share an expression peak in G3 samples but are not consistently changed. *: p < 0.05; **: p < 0.01.

Figure 4

Relative mRNA expression changes of melanoma PDTX tissue upon vemurafenib treatment. All panels show relative expressions, normalized to control RPLP0. All measurements were analyzed in a common study to ensure the common threshold values for cT determination. Genes systematically upregulated (GPR39, CD27, SLC15A3, PDGFA, ABCB1) or downregulated (IFI27) after vemurafenib treatment. *: p < 0.05; **: p < 0.01.

Figure 5

Protein expression patterns upon vemurafenib treatment in melanoma. (A) Quantitative analysis of positive and negative cells from whole-section scanned images. N = 3 tumors in each groups. (B) Representative immunohistochemical staining pictures of P-gp, GPR39, F4/80, and GNRHR proteins. Black arrows indicate F4/80-positive macrophages. Scale bars: 50 µm. (C) Western blot images showing protein level changes of CD27 and IFI27. Three control and three treated animal samples were measured.

Figure 6

All samples carried full length BRAF mRNA. Both with primers on exons 8 and 9 (300 bp product) and primers 3–9 (1000 bp product), we identified the full length BRAF mRNA but not the exon 4–8 deleted variant, which could have been visualized in the exon 3–9 probe, at 200 bp size.

Figure 7

Protein levels and activation of different signaling pathway proteins, BRAF V600E, and activated CRAF, and p53 expression. The figure shows three parallel animals’ tumor samples from the treatment-naïve group and from generation 5. GAPDH was used as an endogenous control to ensure equal protein levels of every sample. (A) Neither AKT, mTOR, MAP2K1/2 (MEK 1/2), or P44/42 (ERK 1/2) was found to be systematically activated, though the ERK 1/2 expression was found elevated in one of the three resistant tumors. Moreover, PDGFRB, which showed a 4-times elevation in RNAseq data, was shown not to be overexpressed on the protein level. (B) BRAF V600E and p-c-RAF expressions remained unchanged on the vemurafenib treatment. (C) p53 accumulation is similar in both the naïve and resistant (and treated) samples.

Figure 8

Quantitative analysis of apoptosis by counting p53 positive cells. (A) We used the scanned tissue slides to analyze three view areas per slide, examining in total above 7000 cells in each tumor slide. (B) We trained QuPath 0.4.3. software to identify negative and positive cells (red and blue marks, respectively) under visual control in order to use the best image analysis parameters. Scale bars: 50 µm. (C) p53-positive apoptotic cell ratio in control (N = 4) and treated (N = 3) tumors. (D) Average p53-positive apoptotic cell ratio in control and treated tumors.

Figure 8

Quantitative analysis of apoptosis by counting p53 positive cells. (A) We used the scanned tissue slides to analyze three view areas per slide, examining in total above 7000 cells in each tumor slide. (B) We trained QuPath 0.4.3. software to identify negative and positive cells (red and blue marks, respectively) under visual control in order to use the best image analysis parameters. Scale bars: 50 µm. (C) p53-positive apoptotic cell ratio in control (N = 4) and treated (N = 3) tumors. (D) Average p53-positive apoptotic cell ratio in control and treated tumors.