Myelomonocytic skewing in chronic myelomonocytic leukemia: phenotypic, molecular and biologic features and impact on survival

Abstract Background Myelomonocytic skewing is considered as a key pathophysiologic phenomenon in chronic myelomonocytic leukemia (CMML), but its prevalence and potential correlation with phenotypic, genotypic, and clinical features are poorly defined. Methods Skewed differentiation toward the myelomonocytic over erythroid commitment as indicated by an inverse ratio of myelomonocytic/erythroid colonies was investigated in 146 patients with CMML by semisolid in vitro cultures. Results There was a high prevalence of myelomonocytic skewing in patients with CMML (120/146, 82%); whereas, this phenomenon was rare in normal individuals (1/98, 1%). Patients with CMML with myelomonocytic skewing had higher white blood cell and peripheral blast cell counts, and lower platelet values. The number of mutations in genes of the epigenetic and/or splicing category was higher in CMML patients with as compared with patients without skewing. Patients with myelomonocytic skewing had more frequently mutations in RASopathy genes and higher growth factor independent myeloid colony formation. Interestingly, the lack of myelomonocytic skewing discriminated patients with CMML with a particularly favorable prognosis (60 vs 19 months, P = .003) and a minimal risk of transformation. Conclusion Myelomonocytic skewing as determined by semisolid cultures can discriminate subgroups of patients with CMML with a different phenotype, a different genotype, and a different prognosis.


| INTRODUC TI ON
Normal hematopoietic function is maintained by a well controlled balance of myelomonocytic, mega erythroid, and lymphoid progenitor cell populations. This balance may be skewed during pathologic conditions such as hematological malignancies, infections, and autoimmunity but also in aged hematopoiesis. [1][2][3][4][5][6][7] Recently, we have reported that analysis of myelomonocytic skewing in vitro may be useful to investigate skewed differentiation toward the myelomonocytic over erythroid commitment in patients. 8,9 Since the presence of skewing may be associated with a different phenotype, a different mutational landscape and a different prognosis in patients with myeloid malignancies this in vitro test may help to comprehensively study hematopoiesis in patients with complex disturbances of blood formation. Myelomonocytic skewing has been reported in chronic myelomonocytic leukemia (CMML) by analyzing single-cell-derived colonies, but this phenomenon has not been correlated with phenotype and clinical characteristics. 10 The aim of this study was to study the prevalence and a potential correlation of myelomonocytic skewing as determined by semisolid in vitro cultures with phenotypic, molecular, biologic, and clinical features in a large cohort of patients with CMML.

| Patients
In the "Austrian Biodatabase for Chronic Myelomonocytic Leukemia" (ABCMML) clinico-laboratory, real-life data have been captured from 606 patients with CMML from 14 different hospitals over the last 30 years. The ABCMML has been shown to be a representative and useful real-life data source for further biomedical research. 11 In 146 patients with CMML of our ABCMML data from semisolid in vitro cultures were available which were used for this retrospective study. This research has been approved by the ethic committee of the City of Vienna on 10.06.2015 (ethic code: 15-059-VK).

| Colony Assay
In one of our centers (Medical University of Vienna), the assessment of hematopoietic colony formation in vitro has been an integral part of the diagnostic work up in patients with suspected myeloid malignancies for many years. 12 Colony-forming unit-granulocyte/macrophage (CFU-GM) and burst-forming unit-erythroid (BFU-E) growth were assessed in semisolid cultures with and without growth factors as previously described. 13

| Molecular studies
Genomic DNA was isolated from mononuclear cell (MNC) fractions of these blood samples according to standard procedures. The mutational status of CMML-related protein coding genes was determined by targeted amplicon sequencing using the MiSeq platform (Illumina).
Details regarding gene panel, library preparation, and data processing have been reported previously. 11 Only variants with an allelic frequency (VAF) ≥5%, a described population frequency (MAF) <1%, and an annotated pathogenic effect (or probability >90% of being pathogenic) were included, with pathogenicity determined according to databases as shown in Table S1 and published studies.

| Statistical analysis
The log-rank test was used to determine whether individual parameters were associated with OS. OS was defined as the time from sampling to death (uncensored) or last follow-up (censored).
Multivariate Cox regression analysis of overall survival was used to

Novelty statement
• Myelomonocytic skewing as determined by semisolid in vitro cultures has been performed in a large cohort of patients with chronic myelomonocytic leukemia (CMML).
• Myelomonocytic skewing can discriminate subgroups of patients with CMML with a different phenotype, a different genotype and a different prognosis.
• Our findings may be important for the understanding and management of CMML. describe the relation between the event incidence, as expressed by the hazard function and a set of covariates. Dichotomous variables were compared between different groups with the use of the chisquare test. The Mann-Whitney U test was used to compare two unmatched groups when continuous variables were not normally distributed. Results were considered significant at P < .05. Statistical analyses were performed with the SPSS version 19.0.0 (SPSS Inc); the reported P values were 2-sided.

| RE SULTS
There was a high prevalence of myelomonocytic skewing as indicated by an inverse ratio of CFU-GM/BFU-E in patients with CMML (120/146, 82%); whereas, this phenomenon was rare in normal individuals (1/98, 1%). As shown in Table 1, there was no difference in patients with and without myelomonocytic skewing with regard to age and male predominance.

| Impact of myelomonocytic skewing on the phenotype of CMML
The phenotype stratified by the presence or absence of myelomonocytic skewing in patients is shown in Table 1. Patients with CMML with myelomonocytic skewing had higher white blood cell (WBC) and PB blast cell counts, a trend toward lower hemoglobin (Hb) values and significantly lower platelet (PLT) counts as compared with patients without skewing. The incidence of splenomegaly was not significantly different. Figure 1 shows the Kaplan-Meier plots of overall survival in patients with CMML stratified by the presence or absence of myelomonocytic skewing. Interestingly, the lack of myelomonocytic skewing discriminated patients with CMML with a particularly favorable prognosis.

| Impact of myelomonocytic skewing on survival and time to AML transformation of CMML patients
The median survival of patients with CMML with myelomonocytic skewing was 19 months as compared with 60 months in patients without skewing (P = .003). In Figure S1, the Kaplan-Meier plots of other established single prognostic factors such WBC count, Hb value, PLT count, and PB blasts are shown. All these parameters also had a prognostic impact in our study. In order to determine the relation of the prognostic impact of myelomonocytic skewing to other established prognostic factors, several Cox regression analyses were performed adjusting for these factors. As shown in Table S2, myelomonocytic skewing, not unexpectedly, lost its prognostic significance if adjusted for WBC, but retained its significance in the presence of all other parameters. in patients with skewing. There was one patient who had no myelomonocytic skewing at the time of diagnosis and developed AML after 134 months. In Figure S2, the Kaplan-Meier plots for time to transformation is given for WBC count, Hb value, PLT count, and PB blasts. Except for the PB blast cell counts, none of these parameters had a significant impact. As shown in Table S3, myelomonocytic skewing retained its significance in the presence of all these parameters.
In a subgroup of patients, cytogenetic and/or molecular information was available (n = 82). The Kaplan-Meier plots of time to AML transformation in patients with CMML stratified by the presence or absence of genetic variables which were defined as high risk aberrations in the CPSS-Mol Score are shown in Figure S3. 15 There was a trend toward a significant difference (P = .114), but this may be due to the limited number of patients with genetic information.

| Mutational profile of CMML stratified by the presence or absence of myelomonocytic skewing
Mutations in genes of the epigenetic control and the splicing machinery have shown to promote differentiation toward the myelomonocytic cell lineage in preclinical mouse models. 1

| Impact of myelomonocytic skewing on spontaneous myeloid colony formation
In vitro cultures data were available from 146 patients. We recently were able to show that growth factor independent CFU-GM formation is a functional surrogate of RAS-pathway activation. 16,17 The spontaneous formation of CFU-GM in normal individuals (median 4.8/10 5 PBMNC, range 3.5-8.5) has been reported by us previously. 18 The numbers of spontaneously formed CFU-GM in patients with CMML stratified by the presence or absence of myelomonocytic skewing is indicated in Figure 3. The box plots show a large variation in colony numbers between single patients in the two cohorts; however, median CFU-GM numbers per 10 5 MNC were significantly higher in patients with myelomonocytic skewing (md 11, range 0-1127, n = 107) as compared with patients without skewing (md 2, range 0-167, n = 22; P = .0067).

| Temporal relationship of myelomonocytic skewing and RAS-pathway activation in a CMML patient with serial in vitro cultures
In one CMML patient, serial in vitro cultures were performed during the course of disease. As shown in Table 3, the transition to myelomonocytic skewing (from a CFU-GM/BFU-E ratio <1 to >1) was ac-

| D ISCUSS I ON
CMML is a hematopoietic malignancy of the elderly that is characterized by overlapping features of myelodysplastic syndromes (MDS) and myeloproliferative neoplasms (MPN) and an inherent risk of transformation to secondary acute myeloid leukemia.
Several articles extensively reviewed the diagnostic criteria, and the clinical and molecular characteristics of CMML, but the biological features of this disease are not comprehensively reported. [19][20][21][22][23] We have originally shown the in vitro characteristics of CMML in a small number of patients. 24 In this study, we performed cell-  The well known unfavorable impact of leukocytosis and anemia in CMML may be considered as indirect evidence that skewing of hematopoiesis toward the myelomonocytic lineage predict inferior outcome. The prognostic impact of myelomonocytic skewing at the progenitor cell level in CMML, however, has not been shown to best of our knowledge so far. We think that myelomonocytic skewing as demonstrated by in vitro cultures may be a more robust parameter of skewed differentiation toward the myelomonocytic over erythroid commitment because the WBC count and Hb values are more easily confounded by other condition such as infection and bleeding, which usually do not change the progenitor cell compartment. 29,30 We conclude that myelomonocytic skewing can discriminate patients with CMML with a different phenotype and different prognosis. Moreover, myelomonocytic skewing seems to predispose for the emergence of additional molecular aberrations in genes of the RAS-pathway which will finally result in MP-CMML and transformation. Therefore, therapeutic strategies targeting the molecular mechanisms underlying the biologic phenomenon of myelomonocytic skewing may be an attractive approach to impact CMML.

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author, [KG], upon reasonable request.