The role of the RAS pathway in iAMP21-ALL

Abstract

Intrachromosomal amplification of chromosome 21 (iAMP21) identifies a high-risk subtype of acute lymphoblastic leukaemia (ALL), requiring intensive treatment to reduce their relapse risk. Improved understanding of the genomic landscape of iAMP21-ALL will ascertain whether these patients may benefit from targeted therapy. We performed whole-exome sequencing of eight iAMP21-ALL samples. The mutation rate was dramatically disparate between cases (average 24.9, range 5-51) and a large number of novel variants were identified, including frequent mutation of the RAS/MEK/ERK pathway. Targeted sequencing of a larger cohort revealed that 60% (25/42) of diagnostic iAMP21-ALL samples harboured 42 distinct RAS pathway mutations. High sequencing coverage demonstrated heterogeneity in the form of multiple RAS pathway mutations within the same sample and diverse variant allele frequencies (VAFs) (2-52%), similar to other subtypes of ALL. Constitutive RAS pathway activation was observed in iAMP21 samples that harboured mutations in the predominant clone (⩾35% VAF). Viable iAMP21 cells from primary xenografts showed reduced viability in response to the MEK1/2 inhibitor, selumetinib, in vitro. As clonal (⩾35% VAF) mutations were detected in 26% (11/42) of iAMP21-ALL, this evidence of response to RAS pathway inhibitors may offer the possibility to introduce targeted therapy to improve therapeutic efficacy in these high-risk patients.

Figure 1

The nature and incidence of RAS pathway mutations in iAMP21-ALL. (a–d) Protein domain and alteration plots for RAS pathway genes that harbour mutations in 26 iAMP21-ALL samples: 44 mutations were identified in NRAS (a), FLT3 (B), KRAS (c) and PTPN11 (d). Each coloured line depicts the type of mutation (mismatch (red), deletion (black), insertion (green) and frameshift (blue)), and the amino-acid change is stated above the position/line. Most of the mutations had been previously reported in cancer (COSMIC), were frequently located in hotspot mutation regions within the gene and often observed as internal tandem duplications (ITD) in FLT3. The coloured circles represent the 26 individual samples; those cases with a single RAS pathway mutation are coloured black and each sample with multiple mutations is defined by a distinct colour. (e) The co-occurring nature of RAS pathway mutations is shown in 26 iAMP21-ALL patient samples. Each sample is labelled on the x axis and the y axis defines the VAF (%) of each mutation. The pattern/colour of each bar represents the mutated gene, as depicted by the key. *PTPN11 p.N58S, present in the matched diagnosis and remission sample. The protein domain and alteration plots were generated using The Protein Painter application in the Pediatric Cancer Genome Project (http://explore.pediatriccancergenomeproject.org).

Figure 2

Nature of RAS pathway mutations in other ALL subgroups. KRAS, NRAS and FLT3 were sequenced in patient samples with high hyperdiploid (n=48) or B-other ALL (n=66). Only exonic regions that were mutated in the iAMP21-ALL cohort were assessed. (a) The incidence of mutations in high hyperdiploid, iAMP21 and B-other ALL ranged from 44 to 85% the proportion of cases with significant mutational burden (⩾35% VAF (black bar)) or subclonal mutation (<35% VAF (patterned bar)) is shown. (b) VAF of KRAS, NRAS and FLT3 mutations in high hyperdiploid, iAMP21 and B-other ALL is demonstrated. Subclonal mutations were identified for each gene, independent of ALL subgroup. (c, d) Protein domain plots for FLT3 mutations in 17 high hyperdiploid (c) (20 mutations) and 5 B-other (d) (10 mutations) patient samples. Each coloured line depicts the type of mutation (mismatch (red), deletion (black), insertion (green) and splice site mutation (gold)), and the amino-acid change is stated above the position/line. The coloured circles represent individual samples; those cases with a single RAS pathway mutation are coloured black and samples with multiple mutations are defined by a white circle.

Figure 3

The association between RAS pathway mutations and CNAs in iAMP21-ALL. iAMP21-ALL patient samples (n=49, 42 diagnostic and 7 relapse) were screened for (a) RAS pathway mutations, (b) RAS pathway CNAs by SNP6.0 array analysis of 30 iAMP21-ALL cases (28 diagnostic and 2 relapse samples) and (c) CNA that affects genes recurrently aberrant in ALL. *Two relapse cases were used to validate the diagnostic-specific nature of RAS gene mutations, but they were not included in the mutation screening cohort. Grey denotes that the genomic assay was not performed. The type and VAF (%) of the mutation and the copy number (CN) status of the gene are represented by the respective colour. CRLF2 abnormalities are shown in purple. Only genes that harboured mutations or CNA are displayed. CNA was frequently observed in genes (BRAF, NF1, CBL) located on chromosomes recurrently abnormal in iAMP21-ALL but they were not observed alongside mutations in cases analysed by targeted or whole-exome sequencing. There was a trend towards co-occurrence of IKZF1 deletions and RAS mutations (Fisher's exact test, P=0.065).

Figure 4

The biological significance of RAS pathway mutations in iAMP21-ALL. (a) Western blot analysis was used to investigate the biological effect of RAS pathway mutations in patient 1 (bold text), relative to positive (KRAS mutated) and negative (wild type (WT)) control samples from our previous studies.^, pERK was observed in the positive control samples and patient 1, which harboured NF1 p.P1667S and NRAS p.Q22K. Absent pERK was observed in the negative control samples. (b, c) Cytotoxicity assays were performed using the MEK1/2 inhibitor, selumetinib, on viable cells from the xenografts of patients 1, 14 and the xenograft material of two iAMP21-ALL patients that did not harbour RAS pathway mutations (patients 45 and 46b) (Supplementary Figure 6). (b) Reduced pERK levels confirmed the inhibitory effect of selumetinib on RAS pathway signalling exclusively in the mutated samples. pERK expression was not detected in the RAS normal xenograft samples (data not shown). (c) Sensitivity to selumetinib was observed in the RAS mutant samples (patients 1 and 14) only. (d) Bar chart of GI50 values after dosing with selumetinib for the xenografts of patients 1, 14, 45 and 46b, compared with RAS mutant and non-mutated ALL samples from our previous study. A similar level of sensitivity to selumetinib was observed in patients 1 and 14, compared with other RAS mutant ALL samples. Similarly, the response observed in patients 45 and 46b was equivalent to non-mutated ALL samples. (e) Transcript expression of FLT3-ITD and normal FLT3 allele in patients 7, 11 and 21, as measured by RT-PCR. FLT3-ITD was expressed in all mutant cases with VAF of 13% (patient 21), 4% (patient 7) and 9% (patient 11). MUTZ5 was used as a control for normal FLT3 expression. A 100-bp DNA ladder (Promega, Madison, WI, USA) was used to determine the PCR product size.