Clinicopathologic and molecular characterization of myeloid neoplasms with isolated t(6;9)(p23;q34)

Abstract

Introduction:

The t(6;9)(p23;q34);DEK-NUP214 [t(6;9)] abnormality is found in 0.7–1.8% of patients with acute myeloid leukemia (AML) or myelodysplastic syndromes (MDS). FLT3-ITD mutations are detected in t(6;9) patients. The t(6;9) abnormality is associated with poor outcomes. We studied the clinicopathologic and molecular profiles of patients with AML/MDS carrying t(6;9).

Methods:

We collected clinical data of nine patients with AML/MDS with isolated t(6;9) (median age = 41 years; male/female = 4/5) and genotyped DNAs using whole exome, Sanger, and targeted sequencing.

Results:

Our cohort was characterized by frequent multilineage dysplasia (56%), absence of phospho-STAT3/STAT5 expression, presence of myeloid markers (CD13, CD33, CD34, CD117, HLA-DR) with an aberrant expression of CD7, and poor outcome (median survival of 20 months). Although basophilia has been described in association with t(6;9), we observed lack of marrow basophilia in our cohort. Molecularly, 83% (5/6) of patients with AML/MDS with t(6;9) were characterized by at least one somatic mutation. Among them, four patients showed multiple mutations. FLT3-ITD mutations were detected in 33% of patients (2/6); 80% (4/5) of mutant patients died even after hematopoietic stem cell transplantation.

Conclusion:

Our data demonstrated that AML/MDS patients with t (6;9) have diverse molecular mutations regardless of the presence of FLT3 mutations, which may contribute to their poor survival outcomes.

INTRODUCTION

Acute myeloid leukemia (AML) with t(6;9)(p23;q34); DEK-NUP214 [t(6;9)] is a distinct entity in the 2008 World Health Organization (WHO) classification [1–3]. AML with t(6;9) patients can present as de novo AML with any morphologic subtype of AML, not otherwise specified. The presence of antecedent myelodysplastic syndromes (MDS) features is also commonly seen. The t(6;9) is commonly found in young adult (median age of 35 years) or in childhood AML patients (median age of 13 years). Patients with t(6;9) usually present with anemia, thrombocytopenia, or pancytopenia with circulating blasts. Based on the WHO classification, multilineage dysplasia and peripheral blood/bone marrow (PB/BM) basophilia (defined as >2%) have been associated with patients with t(6;9) abnormality [4]. The prognosis of AML with t(6;9) patients is poor and similar to the one of AML patients with dismal cytogenetic abnormalities [5]. The presence of t(6;9) in AML patients has been associated with poor response to standard chemotherapy and increased postremission relapse [6]. Hematopoietic stem cell transplantation (HSCT) has been shown to improve the prognosis of AML with t (6;9) patients compared to AML with t(6;9) patients receiving chemotherapy only [7, 8]. The t(6;9) is usually present as a sole karyotypic abnormality. Genetically, cells of patients with AML/MDS with t(6;9) are characterized by the presence of a chimeric gene (DEK-NUP214) encoding a 165-kd nucleoporin protein derived by the fusion of DEK (6p23) and NUP214 (9q34). FLT3 mutations have been detected in a large fraction of AML with t(6;9). The presence of FLT3-ITD mutations has been associated with a specific clinical profile (increased leukocytes, BM blasts percentage and poor outcomes). FLT3-TKD mutations have been found in AML with t(6;9) with a much lower frequency [9].

Our aim was to further define the clinicopathologic and molecular profiles of t(6;9) AML/MDS patients possibly explaining their poor prognosis.

MATERIALS AND METHODS

Patients and specimens

We found nine adult cases with AML and/or MDS with t(6;9) as an isolated cytogenetic abnormality by reviewing cytogenetic databases at the Cleveland Clinic and retrospectively reviewed their clinical and pathologic findings. BM or PB mononuclear cells from these nine patients were retrospectively collected for further molecular studies. The study was conducted in accordance with the Declaration of Helsinki. All patients signed a written informed consent approved by the Institutional Review Board of the Cleveland Clinic. Clinical data were collected from medical records as of May 2016. The clinical and pathologic parameters reviewed in this study are listed in Table 1. Length of follow-up was measured from the day of the diagnostic BM biopsy with t(6;9) to the expiration date or the most recent follow-up date in living patients. AML was defined as the occurrence of equal to or greater than 20% blasts in BM or PB. Dysmegakaryopoiesis, dyserythropoiesis, and dysgranulopoiesis were considered to be present when dysplasia was present in greater than 10% of cells in the specific cell lineage. PB and BM basophilia was defined as greater than 2% based on the 2008 WHO classification criteria. Conventional cytogenetics was carried out using standard G-banding techniques in metaphase cells. Findings were reported according to the International System for Human Cytogenetic Nomenclature (ISCN) [10]. Molecular data were reviewed in these patients.

Table 1.

Clinical and pathologic features of patients with myeloid neoplasms with t(6;9)(p23;q34)

	AML	MDS	Total	P
No	6	3	9
Diagnosis
RAEB	0	2
RCMD		1
AML MRC	2
AML with maturation	1
AMMoL	3
Age (years; median)	34.5 (21–70)	59 (41–61)	41 (21–70)	0.20
Sex (M/F)	3/3	1/2	4/5	0.63
WBCs (× 109/L; median)	49.3 (4.6–88.5)	8.4 (2.2–12)	12.0 (2.2–88.5)	0.04
Hb (g/dL; median)	8.8 (6.4–10)	11.7 (8.1–12)	9.1 (6.4–12)	0.26
PLTs (× 109/L; median)	127 (14–450)	44 (27–113)	113 (14–450)	0.28
BM blasts (%; median)	65 (41–88)	7 (2–14)	49 (2–88)	7.45E-05
BM cellularity (%; median)	95	80	92.5	0.38
PB/BM Basophilia (%/%; median)*	0/0	0/0	0/0
Dysplasia (%; median)
Erythroid	50	66	55	0.63
Granulocytes	66	100	77	0.26
Megakaryocytes	33	66	44	0.34
Ring sideroblasts	0	0	0
Multilineage Dysplasia (%; median)	66	100	55	0.26
Treatment
Low-intensity CTX	0	0	0	0.13
High-intensity CTX	2	0	2
Stem cell transplant	4	2	6
Supportive	0	1	1
Clinical outcome
(Expired %; median)	100% (5/5)	0% (0/3)	63%	0.008
Follow-up (mo; median)	17.5 (10–69)	24 (24–81)	20 (10–81)	0.74

No, number; RAEB, refractory anemia with excess blasts; RCMD, refractory cytopenia with multilineage dysplasia; AML, acute myeloid leukemia; MRC, myelodysplasia-related changes; MDS, myelodysplastic syndrome; AMMoL, acute myelomonocytic leukemia; M, male; F, female; WBCs, white blood cells; Hb, hemoglobin; PLTs, platelets; BM, bone marrow; PB, peripheral blood; CTX, chemotherapy; mo, months.

^*BM and PB basophilia by WHO classification: >2%

Whole exome sequencing (WES)

DNAs were extracted from the pretreatment BM specimen of nine patients with AML/MDS with t(6;9) using a Gentra purification kit. DNA (2 μg) was prepped, and 20 millions reads were run on an Illumina HiSeq2000 sequencer. Reads were aligned to the GRCh37/hg19 genome. A bioinformatics algorithm developed in-house was used to score all somatic variants. Tumor nucleotide variants were counted if unique to the tumor by analyzing the variant allele frequency (VAF) in both tumor and germ-line samples (paired CD3 positive). Variants were called as somatic after excluding SNPs (dbSNP135, dbSNP138, dbSNP142 http://www.ncbi.nlm.nih.gov/projects/SNP) and after reviewing germ-line databases (http://exac.broadinstitute.org). A search of public databases of previously observed alterations (http://cancer.sanger.ac.uk/; TCGA) was also conducted.

Multi-amplicon targeted deep sequencing

An Illumina TruSeq custom panel was used to detect coding variants within exons of target genes (Table S1). Target genes amplicon pools were selected following the GRCh37/hg19 assembly and designed by Illumina (http://www.illumina.com/informatics/research/experimental-design/designstudio.html).

Polymerase chain reaction (PCR) and direct sequencing

Genomic DNA was isolated using a Gentra DNA purification kit. DNA (10–20 ng) was used for semiquantitative PCR using specific primers for TP53 (all exons), ASXL1 (exon 12), and N/KRAS (exons 1 and 2). All selected exons of genes underwent direct genomic sequencing by standard techniques on the ABI3730×1 DNA analyzer (Applied Biosystems, Foster City, CA, USA). All mutations were designated as pathogenic on the basis of the observation that they were not detected in DNA of healthy subjects, in germ line sources (CD3 positive cells) derived from paired samples, and in published SNP databases (dbSNP, http://www.ncbi.nlm.nih.gov/projects/SNP), and they were not reported as SNPs in previous publications and were not present in databases of germ-line alterations (http://exac.broadinstitute.org/).

Multiplexed PCR

Multiplexed PCR was used to detect FLT3 mutational status. PCR using fluorescently labeled forward primers was performed on the tyrosine kinase receptor gene containing the juxtamembrane domain and amino acids 835/836 of the second kinase domain. Amplified products were analyzed by capillary gel electrophoresis for in-frame length mutations (ITD mutations) and resistance to EcoRV digestion (TKD mutations).

Immunohistochemistry

Immunohistochemistry was performed using an automated immunostainer [Discovery Ultra, Ventana Medical Systems (VMSI), Tucson, AZ, USA]. Slides were incubated with the pSTAT3 antibody (clone D3A7; Cell Signaling Technology, Danvers, MA, USA) at 1 : 25 dilution for 60 min at room temperature or pSTAT5 antibody (clone Y694/99; Advantex Bio, Houston, TX, USA) at 1 : 50 dilution for 32 min at 37 °C following CC1 standard epitope retrieval (VMSI, 950–124). Visualization was achieved with either OMNIMap anti-rabbit-HRP (VMSI, 760–4311) or OMNIMap anti-rabbit-mouse (VMSI, 760–4310) and ChromoMAP-DAB detection system (VMSI, 760–159). The slides were then counterstained with hematoxylin (VMSI, 760–2021) for 4 min and incubated 4 min with Bluing reagent (VMSI, 760–2037). More than 10% nuclear staining in megakaryocytes, erythrocytes, or granulocytes was defined as a positive result. Stained slides were reviewed by hematopathologists.

Statistical analysis

Overall survival (OS) was calculated from the date of diagnosis to date of death, and surviving patients were censored at date which they were last known to be alive or lost to follow-up. Student’s t-test was used to estimate differences between groups. P ≤ 0.05 was considered significant. Statistical analysis was performed with the statistical software package JMP version 9.0 (SAS Institute, Cary, NC, USA).

RESULTS

Characteristics of patients with myeloid neoplasms and t(6;9)(p23;q34)

We retrospectively collected clinical, pathologic, and genetic data of nine patients with isolated t(6;9) (p23; q34) (AML = 6; MDS = 3). Six AML patients included one AML with maturation, three acute myelomonocytic leukemia (AMMoL), and two AML with myelodysplasia-related changes (MRC). Three MDS patients included two refractory anemia with excess blasts (RAEB) and one refractory cytopenia with multilineage dysplasia (RCMD). There were four males and five females with a median age of 41 years. Median value for WBC count was 12 × 10⁹/L (2.2–88.5), for hemoglobin was 9.1 g/dL (6.4–12), and for BM blast percentage was 65% (41–88) and 7% (2–14) in the AML and MDS group, respectively. PB and BM basophilia was not observed in our cohort. The range of basophils was from 0 to 2% (2% in two patients) (Table 1). Among the parameters analyzed in this study, WBC count was significantly higher in AML than MDS patients with t(6;9) abnormality, while other parameters were not different between the two groups. All patients have hypercellularity for age (median 92.5%; range 75–100%). Multilineage dysplasia in two or more lineage cells was frequently observed in 55% (5/9) of the cases, but there was no significant difference between the two groups. Iron stain was performed in all cases and showed no presence of ring sideroblasts. Flow cytometric immunophenotype was available for eight patients and demonstrated myeloid blasts expressing CD33 (8/8), HLA-DR (8/8), CD13 (7/8), and CD34 (6/8). In addition, there was an aberrant expression of CD7 (3/8; 38% with three patients in the AML group). We also performed immunohistochemistry for pSTAT3 and pSTAT5 without finding any difference in protein expression.

Treatment information was available for all patients. Two patients received high-intensity chemotherapy and six patients underwent HSCT after high-intensity chemotherapy (four patients in the AML group and two patients in the MDS group). High-intensity chemotherapy was administered to two patients. Persistent disease was observed in six patients and complete remission (CR) in three patients. A median follow-up was 24 months in MDS and 17.5 months in AML patients. Median OS in this cohort is 20 months by Kaplan–Meier survival analysis; however, there are too small numbers of patients to compare OS between the two groups (Table 1). In regard to HSCT, 60% of patients who received HSCT had expired and 67% of patients who did not receive HSCT had expired during follow-up (P = 0.85). The patients of our cohort seemed to have shorter OS even after HSCT; however, there was no significant difference in the clinical outcome between the two groups (logrank P = 0.28). The lack of significance may be attributed to the small number of patients in our cohort.

Molecular spectrum of patients with myeloid neoplasms and t(6;9)(p23;q34)

We combined results from WES, targeted gene sequencing, multiplexed PCR, and Sanger sequencing to characterize the molecular profile of myeloid neoplasms with t(6;9) in our cohort and to explore any association with survival outcomes and response to chemotherapy. Five patients were genotyped. One patient was performed for FLT3 mutation, but no further genotypic analysis was available due to lack of material. WES was conducted on DNA derived from five BM specimens, and obtained variants were cross-referenced with public databases to filter for known germ-line polymorphisms and/or SNPs. A multi-amplicon targeted deep sequencing panel was performed to detect known hotspots of MDS and AML-associated genes on five cases. The panel did not include amplicons to detect DEK or NUP214 (Table S1). Molecular screening detected at least one somatic mutation in five patients. Nine genes were found somatically mutated and mapped in independent chromosomes (1, 2, 7, 11, 12, 13), while two genes mapped at the same chromosome (X). In total, we found 11 exonic mutations: duplication (2), frameshift insertion (1), stop gain, (1) and nonsynonymous substitution (7). Targeted gene sequencing detected mutations in WT1, ETV6, IDH1, GATA1, NRAS, CUX1, and KDM6A, while WES detected mutations in NRAS and SMC1A. FLT3-ITD mutations have been associated with the poor prognosis of t(6;9) AML patients. In our cohort, FLT3 mutations were not the only detected genetic events but molecular mutations in other genes were also found. Using multiplexed PCR and targeted sequencing, we found that two patients harbored FLT3-ITD mutations (2/6; 33%). No patient carried missense mutations, deletions, or insertions in FLT3 exon 20. Besides FLT3, NRAS was found mutated in two patients. One patient carrying FLT3 mutation also carried a WT1 (p.A365fs) mutation. Three patients carried multiple somatic mutations: one patient harbored ETV6 (p.R160K), IDH1 (p.R132H), GATA1 (p.S46F), and NRAS (p.G12D), one patient had CUX1 (p.W1514X) and KDM6A (p.P641S) mutations, and one patient carried NRAS (p.G12S) and SMC1A (p.R711L). All the genes we found mutated have previously been found mutated in MDS and/or AML. We conducted a literature search and interrogated publicly available databases of germ-line and somatic mutations of hematologic and solid tumors to verify the somatic status of the mutations (http://cancer.sanger.ac.uk/; http://www.cbioportal.org/; exac.broadinstitute.org; http://www.1000genomes.org; http://evs.gs.washington.edu/EVS). Variant allele frequency was calculated per each mutation and ranged from 3.5% to 90% in the cohort (Table 2). We analyzed the correlation between presence of mutations and survival outcome. Among the five patients harboring mutations, four patients (#1, 2, 3, 4) expired during follow-up (Table 2). Patient #3 was a 39-year-old male with AML-MRC carrying CUX1 and KDM6A mutations who received chemotherapy and expired soon after (Figure 1). Patient #2 who carried negative prognostic genes (ETV6, IDH1, GATA1, and NRAS) had the shorter survival in the cohort (10 months). Patient #4 was a 49-year-old female with AMMoL carrying FLT3 and WT1 mutations that relapsed after chemotherapy and expired 15 months later (Figure 2).

Table 2.

Somatic mutations in patients with myeloid neoplasms carrying t(6;9)(p23;q34)

	1*	2	3	4	5	6
	AML		AML	AML	MDS	MDS
	MRC	AMMoL	MRC	with maturation	RAEB	RAEB
Alive	No	No	No	No	Yes	Yes
FLT3	ITD			ITD
WTl				p.A365 fs 3.5%
ETV6		p.R160K 15%
IDHl		p.R132H 37%
GATAI		p.S46F 26%
NRAS		p.G12D 16%
CUXl			p.W1514X 14%			p.G12S 32%
KDM6A			p.P641S 13%
SMClA						p.R711L 90%

AML, acute myeloid leukemia; MDS, myelodysplastic syndrome; MRC, myelodysplasia-related changes; AMMoL, acute myelomonocytic leukemia; RAEB, refractory anemia with excess blasts; RCMD, refractory cytopenia with multilineage dysplasia.

^*Due to the limited amount of DNA, for patient #1 only FLT3 status could be assessed.

Percentage represents variant allele frequency.

Figure 1.

Morphologic features of a patient with AML MRC and t(6;9) carrying CUX1 and KDM6A. Bone marrow aspirate smear (Wright–Giemsa stain, ×500) from a 39-year-old patient with a diagnosis of AML MRC shows increased blasts with dysgranulopoiesis, dyserythropoiesis, and dysmegakaryopoiesis. Inlet highlights dysplastic megakaryocytes. Red arrow indicates a basophil. Patient corresponds to #3 in Table II. The patient received chemotherapy with no response and expired soon after. Patient carried CUX1 and KDM6A mutations.

Figure 2.

Morphologic features of a patient with AMMoL and t(6;9) carrying FLT3-ITD and WT1. Bone marrow aspirate smear (Wright–Giemsa stain, ×500) derived from a 49-year-old patient with AMMoL shows significantly increased myeloblasts/monoblasts and lack of basophils. Patient corresponds to #4 in Table 2. Patient relapsed after chemotherapy and expired 15 months later. Patient carried FLT3 and WT1 mutations.

DISCUSSION

In this study, we first described the clinicopathologic features of adult AML/MDS with isolated t(6;9) abnormality and then used high throughput technologies such as WES and targeted deep sequencing to comprehensively characterize the molecular profile of patients carrying t(6;9). Our data showed that the AML/MDS with t(6;9) has distinct clinicopathologic features and single or multiple specific somatic mutations which may be associated with the poor prognosis of these patients. Features of MDS have been found in AML with t(6;9) patients. In our cohort, we found frequent multilineage dysplasia (55%) usually associated with unfavorable prognosis. Ring sideroblasts were absent. Although higher degree of basophilia has been described as a pathologic feature in association with t(6;9), some cases in the reports did not demonstrate BM basophilia. It has been suggested that to 9.6% and was present in 6/8 cases [4]. In this current cohort, basophilia may be associated to MDS rather than to the t(6;9) abnormality itself [11]. In one report, basophilia ranges from 1.2% the range of marrow basophilic percentage was scored from 0% to 2%. We observed two AML patients with 2% of basophils. Blasts have usually nonspecific myeloid phenotype with CD13, CD33, CD34, CD38, CD117, and HLA-DR expression. Interestingly, an aberrant expression of CD7 was noted in 3/8 patients with t(6;9). In vitro studies showed that U937 cell line expressing DEK/NUP214 has activated STAT3 and STAT5. Moreover, primary patient samples have increased expression of STAT3 (pS727) by proteomic analysis [12]. Studies from murine models described that STAT3 and STAT5 are phosphorylated in AML independently from FLT3 mutational status [13]. In our studies, no differences in STAT3 and STAT5 immunohistochemistry were found. DEK/NUP214 is considered the genetic event initiating leukemia, but it might not be only responsible for the poor prognosis observed in AML/MDS with t(6;9). While some studies have found an association between FLT3 mutations and inferior outcomes in t(6;9) positive AML patients [14], others have not found any negative impact [15] [16]. In our study, we used high throughput genetic technologies as an ancillary tool to discover alternative genes influencing survival and response to therapy and HSCT. The t(6;9) abnormality predicts poor prognosis in patients following chemotherapy. A study has shown that the clinical outcome in AML with t (6;9) patients receiving allogeneic HSCT is comparable to the one of patients with normal karyotype, although 55% of patients with AML and t(6;9) had unfavorable outcome [17]. An extensive study from the same group reported that CR at the time of HSCT and the M2 FAB classification were prognostic factors in AML with t(6;9) [8]. In our study, patients had poor prognosis even after HSCT and this may be due to the detrimental molecular profile observed in our cohort. In our cohort, molecular screening identified mutations in genes known to have prognostic significance in AML. It has been estimated that the average number of recurrent mutations in an AML genome is 5. In our cohort, we found one patient with a sole mutation, three patients with two mutations, and one patient with four mutations [18]. A recent paper analyzed 15 cases with AML and t(6;9) using targeted sequencing and found that the most commonly mutated genes were FLT3 and KRAS [19]. We found somatic mutations in 83% of the patients. Mutations were more frequent in AML than in MDS patients. Multiple mutations were found in 50% of patients with 80% (4/5) of the mutant patients expiring. Independently of FLT3 mutations, we observed mutations in genes known to predict unfavorable prognosis such as IDH1, NRAS, CUX1, ETV6, GATA1, and WT1. Mutations in WT1 exon 7 and exon 9 have been found in association with FLT3 mutations and are secondary events in AML [20]. Among the two patients with FLT3 mutations, one patient had an OS of 20 months while the other carrying an associated WT1 mutation had an OS of 15 months. The patient with ETV6, IDH1, GATA1, and NRAS had the shorter survival (10 months) in the group. FLT3 is a gene which regulates cellular proliferation and differentiation of hematopoietic stem cells. In our cohort, we also found mutations in genes that act as regulators of hematopoietic proliferation (CUX1, WT1, SMC1A, NRAS) [21–23].

Our data suggest that the presence of single or multiple somatic mutations in patients with t(6;9) may be an important contributing factor to the clinical outcomes. Further analysis on a large number of cases will be able to clarify the contribution of specific somatic mutations on the outcome of cases with t(6;9) by comparing the effect of the sole genetic abnormality in relation to number of mutations.

CONCLUSION

We characterized nine adult cases with AML and/or MDS with isolated t(6;9) and found that AML/MDS patients carrying t(6;9) in our cohort represent multilineage dysplasia, lack of PB/BM basophilia, and poor outcome. Molecular genotyping demonstrated that the poor prognosis of t(6;9) patients may be associated with other unfavorable prognostic gene mutations regardless FLT3 mutational status.

The power of our study resides in the detailed molecular characterization of a cohort of patients with isolated t(6;9). The limited number of cases precludes definitive conclusions related to clinical outcome. However, our study shows that molecular screening with deep next generation genomic technologies is a powerful tool to detect genetic lesions associated with t(6;9) and might be helpful also in cases with less than 20% blasts to monitor disease progression. The survival outcomes of t(6;9) may be influenced by a diverse molecular signature regardless of FLT3 mutational status.

Supplementary Material

si information

Table S1. Multi-amplicon targeted deep sequencing.