BRAF Kinase Domain Mutations are Common in RAS^wt Chronic Myelomonocytic Leukemia

Abstract

Purpose

The frequency of RAS mutations in chronic myelomonocytic leukemia (CMML) suggests that activation of the MAPK pathway is important in CMML pathogenesis. Accordingly, we hypothesized that mutations in other members of the MAPK pathway might be overrepresented in RAS^wt CMML.

Methods

We performed next generation sequencing analysis on 70 CMML patients with known RAS mutation status using the TruSeq Amplicon Cancer Panel kit (Illumina, San Diego, CA).

Results

The study group included 37 men and 33 women with a median age of 67.8 years (range, 28–86 years). Forty patients were RAS^wt and 30 were RAS^mut; the latter included KRAS=17; NRAS=12; KRAS+NRAS=1. Next-generation sequencing showed 5 patients (7.1% of total group; 12.5% of RAS^wt group) with RAS^wt who had BRAF mutations. All BRAF^mut patients had CMML-1; 2 (40%) with MPN-CMML and 3 with MDS-CMML. The BRAF mutations were of missense type and involved exon 11 in 1 patient and exon 15 in 4 patients. All BRAF^mut patients had CMML-1 with low-risk cytogenetic findings, and none of the BRAF^mut CMML cases were therapy-related. Two (40%) of the 5 patients with BRAF^mut patients transformed to acute myeloid leukemia during follow up. Multivariate Cox proportional hazard regression modeling suggests that BRAF^mut status is associated with overall survival (p=0.04). Additionally, the RAS^mut group tended to have worse OS compared to the RAS^wt group.

Conclusion

In summary, we demonstrate that a subset of patients with RAS^wt CMML harbors BRAF kinase domain mutations that are potentially capable of activating the MAPK signaling pathway.

INTRODUCTION

Chronic myelomonocytic leukemia (CMML) is a heterogeneous clonal myeloid neoplasm characterized by persistent absolute monocytosis in association with myelodysplastic (MDS) and/or myeloproliferative (MP) features.¹ RAS mutations, especially KRAS and NRAS, are among the most common somatic mutations in CMML and are therefore believed to play an important role in its pathogenesis.^2–4 Further supporting this hypothesis, oncogenic NRAS^5–7 and KRAS⁸ mutations have been shown to initiate hematologic malignancies with features of CMML in murine models.

RAS proteins (HRAS, NRAS, and KRAS) are members of the small GTPases superfamily. They play a central role in cell signaling through dual activation of the RAS-RAF-MEK-MAPK-ERK (MAPK) pathway, an important regulator of cell proliferation, differentiation, and survival.⁹ RAF kinases, especially BRAF and CRAF, are potent transducers of RAS-dependent ERK activation. Regulation of BRAF activation is complex and modulated by a variety of conformational and post-translational factors.^10,11 BRAF mutations have been identified in a variety of human malignancies, of which the most common is BRAF^V600E in malignant melanoma.^12,13 As in other kinases, BRAF mutations may lead to increased, low, or impaired kinase activity compared with wild-type BRAF. Notably, however, BRAF mutants with low or impaired kinase activity are capable of constitutive MAPK activation, often through CRAF simulation.¹⁴

Mutational analysis of CMML using next-generation sequencing (NGS) techniques has uncovered recurrent somatic mutations in a variety of gene families, primarily spliceosome genes¹⁵ and genes involved in epigenetic regulation¹⁶. A recent elegant study by Itzykson et al demonstrated that mutations in cell signaling genes seem to represent important secondary events acquired subsequent to an initiating ancestral stem cell event.¹⁷ In view of compelling data suggesting a critical role of RAS mutations in the pathogenesis of CMML^{5,6,8,18–20}, we hypothesized that mutations in other members of the MAPK pathway might be contributing factors in RAS^wt CMML cases and could represent alternate routes to constitutive MAPK/ERK kinase (MEK) activation. Furthermore, identification of MEK activation mechanisms appears critically important for selection of MEK inhibitors as these have variant efficacies depending on whether MEK activation is mediated through RAS or RAF mutations.²¹

In this study, we assessed a group of CMML patients with and without RAS mutations for mutations in other members of the MAPK pathway using targeted NGS-based mutation analysis.

STUDY DESIGN

Study Group

A total of 71 patients diagnosed with CMML according to the World Health Organization classification criteria and with known RAS mutation status were included in this study. RAS mutation status was determined as part of routine clinical workup by PCR-based mutation analysis (codons 12, 13, and 61) as described previously.²² Demographic, laboratory, and clinical data were collected by chart review with emphasis on variables of demonstrated prognostic utility in CMML.²³ Therapies were heterogeneous and included primarily hypomethylating agents. Karyotyping results were used to derive a CMML-specific cytogenetic risk score as described by Such et al.²⁴ This study was approved by the Institutional Review Board of The University of Texas M.D. Anderson Cancer Center.

Next-generation sequencing-based mutation analysis

Next-generation sequencing mutation analysis was performed on 250ng of DNA from unsorted bone marrow or peripheral blood patient samples using the TruSeq Amplicon Cancer Panel kit (TSACP) (Illumina, San Diego, CA). The TSACP interrogates 212 specific genomic areas spanning mutational hotspots in 48 cancer- related genes: AKT1, BRAF, FGFR1, GNAS, IDH1, FGFR2, KRAS, NRAS, PIK3CA, MET, RET, EGFR, JAK2, MPL, PDGFRA, PTEN, TP53, FGFR3, FLT3, KIT, ERBB2, ABL1, HNF1A, HRAS, ATM, RB1, CDH1, SMAD4, STK11, ALK, SRC, SMARCB1, VHL, MLH1, CTNNB1, KDR, FBXW7, APC, CSF1R, NPM1, SMO, ERBB4, CDKN2A, NOTCH1, JAK3, PTPN11, GNAQ and GNA11. Additionally, 7 probe pairs targeting DNMT3A, XPO1, KLHL6, MYD88 and EZH2 were designed by Illumina on request and spiked into the original TSACP pool.

Library preparation was carried out following the manufacturer’s instructions and was analyzed using the Agilent 2200 Tape Station Nucleic Acid System. Successful library preparation, confirmed by the presence of a distinct nucleic acid band at 310bp, was followed by purification using AMPure magnetic beads (Agentcourt, Brea, California) according to the manufacturer’s protocol. One patient sample failed library preparation and was excluded from this study. From each library, equal quantities of the library DNA were isolated and eluted using library normalization beads (TSACP kit) per manufacturer’s instructions. This ensures equal representation of the library from each sample during multiplexed sequencing. Paired end (bidirectional) sequencing was performed using the MiSeq Reagent Kit, V1 (300 cycles) on the MiSeq sequencer (Illumina). Libraries from 10 samples were multiplexed on each sequencing run. Base calling and sequencing run summary were obtained using Real Time Analysis Software (V1.14.23) (Illumina). Sequencing quality was evident by the Q30 scores (one error in 1,000 base pair sequence), and a cutoff of 85% sequence with Q30 score was used as an indication of successful sequencing.

Human genome build 19 (hg19) was used as the reference for sequence alignment. Alignment of the sequence to Human Genome Build 19 reference genome and variant calling was performed by MiSeq Reporter Software 1.3.17. Integrative Genomics Viewer (IGV) was used for visualization of read alignment, presence of variants, confirmation of variant call veracity, and checking for sequencing errors. Distinction between polymorphisms and mutations was conducted through query of annotated public databases (dbSNP and Catalogue of Somatic Mutations in Cancer (COSMIC), historical data on samples analyzed through our pipeline, and literature search. A sequencing coverage of 80× (bidirectional) and a variant frequency >5% in the background of wild type (WT) were used as cutoffs.

Statistical analysis

Descriptive data were compared using Fisher’s exact test for proportions and Wilcoxon rank-sum test for continuous variables. Patients with no event at the time of analysis were censored at date of last follow-up. OS curves were calculated using the Kaplan-Meier method. Differences in survival time between groups of patients were tested using the log-rank test. Multivariate analysis was performed with a Cox proportional hazards regression model. The hazard ratio (HR) and its 95% CI were provided. Given the limited sample size, P=0.10 was set as the threshold for statistical significance. All analyses were performed using R version 2.15.2. Overall survival and time to acute myeloid leukemia (AML) transformation were calculated from the date of first bone marrow diagnosis or date of first presentation to our institution, whichever was earlier.

RESULTS

Study group characteristics

The study group included 70 patients who underwent successful NGS mutation analysis. The patients included 37 men and 33 women with a median age of 67.8 years (range, 28–86 years). At presentation, 36 patients had CMML (33 CMML-1, 3 CMML-2) with low blood counts, and 34 patients had CMML (25 CMML-1, 9 CMML-2) with high blood counts. These patient subsets have been classified by the French-American-British group as MDS-CMML and MPN-CMML, respectively.²⁵ Patients with MDS-CMML had a median white blood count (WBC) of 6 × 10³/μL while patients with MPN-CMML had a median WBC of 25.6 × 10³/μL. Patients had a median follow up of 26 months (range, 2–109 months). The presenting clinical features and outcomes of the study group stratified according to RAS mutation status are summarized in Table 1. The study group included 40 CMML patients with RAS^wt and 30 CMML patients with RAS^mut cases.

Table 1.

Presenting clinical features and outcomes of patients with chronic myelomonocytic leukemia by RAS mutation status.

Feature	RASmut Group	RASwt Group	p-value*
	n=30	n=40
Age (years)
Median	63.7	69.83	0.04
Range	49.4–79.1	46.7–86.2
Gender (%)
Male	63	68	0.8
Female	37	32
WHO classification (n)
CMML-1	22	36	0.11
CMML-2	8	4
Complete blood count
Hemoglobin (g/dL)	12.2	10.1	0.06
White blood count (x103/μL)	13.6	8.1	0.24
Absolute neutrophil count (x103/μL)	6.6	4.2	0.37
Absolute lymphocyte count (x103/μL)	2.0	1.7	0.40
Absolute monocyte count(x103/μL)	2.7	2.0	0.13
Platelets (x103/μL)	93	103	0.14
Blasts (%)	0	0	0.65
Cytogenetic risk group† (%)
Low	3.3	4.0	0.45
Intermediate	76.7	12.5
Poor	20.0	67.5
AML transformation (%)	23.3	20.0	0.26
Overall survival (%)	36.7	12.5	0.09

WHO: World Health Organization; AML: Acute myeloid leukemia.

^*Wilcoxon test for all except Age (t-Test); Sex, WHO classification, Cytogenetic risk group (Log-rank test); Overall survival (Fisher exact test)

^† According to Such et al.²⁴

RAS^mut group

The RAS mutations involved KRAS (n=17), NRAS (n=12), or both KRAS and NRAS (n=1) (Table 2). The TSACP interrogates KRAS exons 2 (codons 1–22), 3 (codons 38–63), and 4 (codons 103–147) and NRAS exons 2 (codons 1–18) and 3 (codons 38–62). KRAS and NRAS mutations in codons 12, 13, and 61 in CMML have been well-characterized.^2,26 There was 99.98% concordance between RAS mutation status by PCR analysis and NGS analysis. One case with KRAS ^K117N by NGS had been missed by PCR analysis since codon 117 was not surveyed in the latter assay (included with RAS^mut group in Table 1).

Table 2.

KRAS, NRAS, and BRAF mutations in chronic myelomonocytic leukemia.

ID	Gene	Mutation	Exon	Variant	Amino acids
15	KRAS	NM_033360.2(KRAS):c.176_178del p.A59del	3	CAGG->G	n/a
7	KRAS	NM_033360.2(KRAS):c.179G>T p.G60V	3	G->T	Glycine>Valine
1	KRAS	NM_033360.2(KRAS):c.183A>T p.Q61H	3	A->T	Glutamine>Histidine
3, 9, 12, 14, 63	KRAS	NM_033360.2(KRAS):c.34G>A p.G12S	2	G->A	Glycine>Serine
30, 33*, 36	KRAS	NM_033360.2(KRAS):c.34G>C p.G12R	2	G->C	Glycine>Arginine
22	KRAS	NM_033360.2(KRAS):c.351A>T p.K117N	4	A->T	Lysine>Asparagine
5, 8, 16, 25	KRAS	NM_033360.2(KRAS):c.35G>A p.G12D	2	G->A	Glycine>Aspartate
19	KRAS	NM_033360.2(KRAS):c.35G>C p.G12A	2	G->C	Glycine>Alanine
6	KRAS	NM_033360.2(KRAS):c.38G>A p.G13D	2	G->A	Glycine>Aspartate
62	NRAS	NM_002524.4(NRAS):c.181C>G p.Q61E	3	C->G	Glutamine>Glutamate
13, 64	NRAS	NM_002524.4(NRAS):c.34G>A p.G12S	2	G->A	Glycine>Serine
62	NRAS	NM_002524.4(NRAS):c.34G>C p.G12R	2	G->C	Glycine>Arginine
23, 27, 28, 32, 33*, 65, 74, 81	NRAS	NM_002524.4(NRAS):c.35G>A p.G12D	2	G->A	Glycine>Aspartate
32	NRAS	NM_002524.4(NRAS):c.35G>T p.G12V	2	G->T	Glycine>Valine
20	NRAS	NM_002524.4(NRAS):c.38G>A p.G13D	2	G->A	Glycine>Aspartate
18	NRAS	NM_002524.4(NRAS):c.38G>T p.G13V	2	G->T	Glycine>Valine
4	BRAF	NM_004333.4(BRAF):c.1782T>A p.D594E	15	T->A	Threonine>Alanine
11	BRAF	NM_004333.4(BRAF):c.1782T>A p.D594E	15	T->A	Threonine>Alanine
24	BRAF	NM_004333.4(BRAF):c.1742A>G p.N581S	15	A->G	Alanine>Glycine
58	BRAF	NM_004333.4(BRAF):c.1790T>A p.L597Q	15	T->A	Threonine>Alanine
82	BRAF	NM_004333.4(BRAF):c.1397G>A p.G466E	11	G->A	Glycine>Alanine

^*Dual KRAS/NRAS mutations.

Among the 17 KRAS^mut patients, 16 (94.1%) presented to our institution with CMML-1 and 1 with CMML-2; 7 (41.2%) had MPN-CMML and 10 had MDS-CMML. Among the 12 NRAS^mut patients, 7(58.3%) presented to our institution with CMML-1 and 5 had CMML-2; 7 (58.3%) had MPN-CMML and 5 had MDS-CMML. Four (23.5%) KRAS^mut and 3 (25%) NRAS^mut patients in the group transformed to AML during follow up. Three NRAS^mut patients and 1 KRAS^mut patients had therapy-related CMML (t-CMML).

All patients with KRAS mutations harbored only one mutation. Two patients with NRAS mutation had two simultaneous NRAS mutations (G12D+G12V; Q61E+G12R) suggestive of the presence of two clones with different NRAS mutations. As described by others², most RAS^mut CMML patients harbored a mutation in either KRAS or NRAS. Kohlman et al identified 3/81 (3.7%) patients with dual KRAS+NRAS mutations; 1 patient (1.4%) had a similar finding in our study group.

BRAF^mut group

We identified 5 CMML patients with BRAF mutations (Table 2). This subset represented 7.1% of the total group and 12.5% of the RAS^wt subset. The TSACP interrogates BRAF hotspots in exons 11 (codons 439–471) and 15 (codons 581–606). All 5 mutations were of missense type with 4 involving exon 15 and 1 located in exon 11. The detection of these BRAF mutations in the aligned reads from sequencing results as visualized by IGV are shown in Figure 2A along with the corresponding depth of sequencing, nucleotide and amino acid changes. The mutation spots in the BRAF kinase domain are illustrated in Figure 2B.

Figure 2. BRAF mutations detected in chronic myelomonocytic leukemia cases.

(A) A total of 5 cases were found to be positive for mutations in the BRAF kinase domain. Representative images of the aligned sequencing reads (red=forward; blue=reverse) with the mutations are shown along with the respective percent variant allele frequency and the sequencing depth (X). (B) A schematic diagram depicting the position and number of mutations occurring in the kinase domain of BRAF. (CR: conserved regions)

All BRAF^mut patients presented to our institution with CMML-1; 2 (40%) patients had myeloproliferative features (MPN-CMML) and 3 patients had low/normal count CMML (MDS-CMML). All 5 patients with BRAF mutations had low-risk cytogenetics and none had therapy-related CMML. Two BRAF^mut patients in the group transformed to AML during the follow up interval.

CMML group with mutations other than RAS/BRAF

The majority of CMML patients in this group could be divided into two broad groups: one patient group associated with genes encoding kinases (n=6) and a second patient group associated with mutations in nonkinase oncogenes (n=5).

In the first group, FLT mutations were detected in 2 patients, one with D835Y and another with internal tandem duplication (ITD). The patient with FLT3^D835Y also harbored an IDH1 mutation during the CMML (myelodysplastic-type) and subsequent AML transformation phase, of which the patient died. Another 2 patients in this group had JAK2^V617F mutation. Two patients in this group had mutations in STK11, a serine threonine kinase that is also known as Liver Kinase B1 (LKB1). A third patient with STK11 mutation also had a KRAS mutation.

In the second group, with mutations involving nonkinase genes, isolated mutations were detected in TP53 (n=3), NPM1 (n=1), and IDH2 (n=1). Notably, 2 patients with TP53 mutation had an identical missense substitution at codon 163 in exon 5 resulting in replacement of the tyrosine residue with aspartate (c.487T>G p.Y163D). This mutation is located within the DNA-binding domain and has been described in other malignancies.^27,28 A third patient with TP53 mutation, notable for a history of colorectal adenocarcinoma treated with chemotherapy and radiation therapy, had therapy-related CMML with intermediate risk cytogenetics and a missense substitution at codon 179 in exon 5 resulting in replacement of the histidine residue with arginine (c.536A>G p.H179R). This TP53 mutation has been demonstrated to correlate with progression to blast crisis in chronic myelogenous leukemia.²⁹ Of note, Solomon and colleagues have shown that TP53^H179R can promote a cancer-related gene signature through augmentation of RAS (HRAS) activity.³⁰

CMML group with no detected mutations

No mutations in the genes/regions assessed by the targeted panel used in this study were detected in 26 patients (37.1%). In this subgroup, 1 patient (3.8%) presented with CMML-1 and 25 patients had CMML-2. These cases included 8 with MPN-CMML and 18 with MDS-CMML. Four (15.4%) patients had therapy-related CMML (t-CMML). Four (15.4%) patients in this group transformed to AML during follow up.

Clinical implications of detected mutational status

When the RAS^mut and BRAF^mut were grouped together on the basis of presumptive pathogenic MAPK perturbations (MAPK^mut, n=36) and compared to the remaining cases (MAPK^wt, n=34), the groups had a significant difference in percentage of bone marrow blasts (p=0.02) and age at diagnosis (p=0.03). We then categorized cases on the basis of NGS findings into five broad categories: (1) RAS^mut, (2) BRAF^mut, (3) cases with mutations in genes encoding kinases, (4) cases with mutations in nonkinase oncogenes, and (5) cases without mutations. Categorized as such, with limitations of sample size in some of the groups notwithstanding, there was a significant difference in time to AML transformation (p=0.016, Log-rank test). Additionally, multivariate Cox proportional hazard regression modeling demonstrated that BRAF^mut status is associated with OS in this study group (p=0.04). (Table 3) Finally, Kaplan-Meier analysis suggests that even though there is no apparent difference in OS among the mutation groups (p=0.429), the BRAF^mut group seems to have a favorable OS. (Figure 3) There was no significant difference in the length of time to AML transformation among the mutation groups.

Table 3.

Multivariate cox proportional hazards regression analysis

Parameter	HR	95% CI	p-value
BRAF mutations	−2.53	(0.007,0.916)	0.04
Mutations in genes encoding kinases	−0.99	(0.031,4.483)	0.44
Mutations in genes encoding nonkinase oncogenes	−1.37	(0.018,3.469)	0.30
No detected mutations	−1.70	(0.025,1.327)	0.09
Gender	0.63	(0.377,9.369)	0.44
Age	0.06	(0.985,1.137)	0.12
Hemoglobin	−0.34	(0.488,1.036)	0.08
White blood count	−0.05	(0.679,1.336)	0.78
Absolute neutrophil count	0.02	(0.678,1.550)	0.91
Absolute lymphocyte count	−0.14	(0.561,1.358)	0.55
Absolute monocyte count	0.16	(0.730,1.891)	0.51
Platelets	0.00	(0.990,1.007)	0.66
Peripheral blood blast %	−3.60	(0.001,1.041)	0.05
Bone marrow blast %	0.04	(0.920,1.175)	0.54

Figure 3. Overall survival analysis.

Patients in the BRAF mutation group tended to have a better overall survival whereas patients in with mutations in genes encoding kinases (“Other-1”=FLT3, JAK2, JAK3, MET, STK11, GNA11, FGFR3) had a worse overall survival. The “Other-2” category includes cases with mutations in nonkinase oncogenes (TP53, NPM1, IDH1, IDH2, EZH2).

DISCUSSION

In this study, we compared mutation analysis data in RAS^mut and RAS^wt CMML patients and show that a subset of the latter group harbor mutations in other genes that play an important role in the activating the MAPK pathway. These data further support murine models and previous reports highlighting the importance of upregulated MAPK activation in the pathogenesis of CMML.^2,3,5–8 Yet, while the most common mechanism of the MAPK activation in CMML is through activating mutations in either KRAS or NRAS, we showed for the first time that BRAF mutations and mutations in genes such as STK11 with demonstrated impact on the MAPK pathway are overrepresented in RAS^wt CMML. These mutations likely function as obligatory secondary mutations that contribute to constitutive activation of the MAPK pathway. Importantly, since RAS^mut and BRAF^mut tumors appear to regulate MEK activation through distinct mechanisms²¹, distinguishing such subsets in CMML will likely have implications for the selection of MEK inhibitors as part of targeted therapy.

BRAF mutations in CMML are heterogeneous and include mutations that lead to increased (BRAF^L597Q), low (BRAF^G466E), or impaired (BRAF^D594) kinase activity.^14,31–34 Among the 5 cases with BRAF mutations (4 in exon 15 and 1 in exon 11) in our study group, 2 cases harbored the BRAF^D594E mutation while the other three mutations were seen in 1 case each. Although mutations involving aspartame-594 residues are common among BRAF mutations¹², to our knowledge the BRAF^D594E mutation has not been described. Mutations at this site (BRAF^D594A and BRAF^D594V) have been shown to result in an inactive kinase protein that is capable of promoting cancer progression through CRAF deregulation.³² Similarly, BRAF^G466E has been shown to result in constitutive MEK activation through CRAF activation despite having a lower kinase activity compared with BRAF^wt.^14,34 The BRAF^L597Q mutation, on the other hand, has been shown to result in a functional oncogene in pediatric acute lymphoblastic leukemia.^31,33 The BRAF^N581S mutation has been reported as a rare event in human malignancies including ovarian and colorectal carcinoma and malignant melanoma.^35–37 Notably, none of the BRAF^mut cases in our group had concurrent RAS mutations suggesting that such mutations might be mutually exclusive in CMML.

In this study, we also show that a small subset of CMML patients had STK11 mutations. The significance of this finding remains to be fully elucidated. Mutations of the tumor suppressor gene STK11 result in the autosomal dominant Peutz-Jeghers syndrome that is characterized by gastrointestinal hamartomas and an increased risk of epithelial and non-epithelial malignancies.^38,39 Inactivating somatic mutations in STK11 have been shown to play an important role in sporadic malignancies including lung cancer⁴⁰ and melanoma⁴¹ by enhancing metastatic potential. Interestingly, the oncogenic effects of BRAF mutations appear to be potentiated through negative regulation of STK11.⁴² One of the CMML patients with STK11 mutation in this study also had a novel KRAS^K117N mutation. Interestingly, KRAS^K117N has been demonstrated to be functional and associated with constitutive activation of the MAPK pathway in a primary melanoma cell line with BRAF^V600E and resistance to vemurafenib.⁴³ Additionally, the oncogenic effects of STK11 function loss are significantly augmented in the presence of activating KRAS mutations.^15,41 While additional studies are needed, the finding of BRAF and STK11 mutations in patients with RAS^wt CMML suggest that neoplastic clones harboring such mutations might be propagated, at least in part, by the proliferative and antiapoptotic effects of MEK activation.

The role that somatic mutations play in the pathogenesis and clonal evolution of CMML is complex and likely underlies the phenotypic heterogeneity of the disease. Itzykson et al recently suggested that selection of TET2^mut progenitor clones in CMML is accomplished through acquisition of secondary mutations affecting cell signaling genes such as KRAS and NRAS during myeloid differentiation.¹⁷ Based on our data, non-RAS mutations such as BRAF and STK11 are likely capable of conferring a growth advantage analogous to RAS^mut. Investigation of the effect of MAPK inhibitors in such cases would be warranted.

In this study group, mutation analysis appears to have clinical relevance. Notably, RAS^mut CMML tended to be associated with worse OS as reported by others⁴⁴, whereas cases with isolated BRAF mutation tended to be associated with a more favorable OS. Interestingly, we identified differences among KRAS^mut and NRAS^mut patients in terms of WHO- and FAB-based disease classifications, findings that warrant additional investigation as they suggest a differential impact of KRAS and NRAS mutations on the progression of CMML.

Some limitations are acknowledged in this retrospective study. Namely, NGS was performed on unsorted samples selected based on availability of RAS mutation status and availability of residual DNA. As such, the frequency of some mutations seen in the present study group might not reflect their broad incidence in CMML. Additionally, since our institution is a tertiary cancer center, a selection bias towards more aggressive disease cannot be excluded. Additional investigation of the functional role of some of the mutations discovered in this study are warranted and being pursued.

In summary, we demonstrate that RAS^wt CMML harbors other mutations that are potentially capable of perturbing the MAPK signaling pathway. Although the scope and mechanisms by which non-RAS mutations that affect the MAPK pathway impact CMML pathobiology remain to be further elucidated, our findings suggest that there are a number of potential therapeutic targets for subversion of MEK activation in patients with CMML.

Figure 1. Summary of mutations detected in chronic myelomonocytic leukemia cases.

The top row indicates case number for each of the samples with mutations identified by next-generation sequencing mutation analysis. The bottom rows represent, respectively, the World Health Organization classification category, cytogenetic risk category, and transformation status to acute myeloid leukemia. Abbreviations: CMML=chronic myelomonocytic leukemia; CG=cytogenetics; AML=acute myeloid leukemia.