Prevalence of RUNX1 gene alterations in de novo adult acute myeloid leukemia

Abstract

BACKGROUND

Acute myeloid leukemia (AML) is a complicated disease with uncontrolled hematopoietic precursor proliferation induced by various genetic alterations. Runt-related transcription factor-1 (RUNX1) is commonly disrupted by chromosomal translocations in hematological malignancies.

AIM

To characterize RUNX1 gene rearrangements and copy number variations in newly diagnosed adult AML patients, with an emphasis on the impact of clinical and laboratory features on the outcome.

METHODS

Fluorescence in situ hybridization was used to test RUNX1 gene alterations in 77 newly diagnosed adult AML cases. NPM1, FLT3/ITD, FLT3/TKD, and KIT mutations were tested by PCR. Prognostic clinical and laboratory findings were studied in relation to RUNX1 alterations.

RESULTS

RUNX1 abnormalities were detected by fluorescence in situ hybridization in 41.6% of patients: 20.8% had translocations, 22.1% had amplification, and 5.2% had deletion. Translocations prevailed in AML-M2 (P = 0.019) with a positive expression of myeloperoxidase (P = 0.031), whereas deletions dominated in M4 and M5 subtypes (P = 0.008) with a positive association with CD64 expression (P = 0.05). The modal chromosomal number was higher in cases having amplifications (P = 0.007) and lower in those with deletions (P = 0.008). RUNX1 abnormalities were associated with complex karyotypes (P < 0.001) and were mutually exclusive of NPM1 mutations. After 44 months of follow-up, RUNX1 abnormalities affected neither patients’ response to treatment nor overall survival.

CONCLUSION

RUNX1 abnormalities were mutually exclusive of NPM1 mutations. RUNX1 abnormalities affected neither patients’ response to treatment nor overall survival.

Key Words: Acute myeloid leukemia; Deletion; Disease-free survival; Fluorescence in-situ hybridization; Karyotyping; RUNX1

Core Tip: In the current study, we characterized the runt-related transcription factor-1 (RUNX1) gene rearrangements and copy number variations in patients with newly diagnosed adult acute myeloid leukemia with an emphasis on the impact of clinical and laboratory features on the outcome. RUNX1 abnormalities were mutually exclusive of NPM1 mutations. RUNX1 abnormalities affected neither patients’ response to treatment nor overall survival.

INTRODUCTION

Acute myeloid leukemia (AML) is a heterogeneous hematologic cancer characterized by the clonal growth of myeloid blasts in the blood, bone marrow (BM), and/or other tissues. It is the most prevalent type of acute leukemia in adults[1]. Over the past few decades, the discovery of recurrent structurally balanced and unbalanced chromosomal abnormalities has significantly influenced the clinical management of patients with AML. These chromosomal abnormalities are the most significant prognostic markers where they can define specific clinicopathologic entities of the disease. The current recommendations of the European Leukemia Network for genetic testing in AML are primarily focused on risk stratification in order to identify an effective therapeutic strategy[2].

Runt-related transcription factor 1 (RUNX1) is the founding member of the mammalian core-binding transcription factor family, which also includes RUNX2, RUNX3, and their non-DNA-binding cofactor CBF[3]. The significance of the transcription factor RUNX1 in the t (8; 21) translocation in AML attracted initial interest. Since its discovery, RUNX1 has been found to play significant roles not only in leukemia but also in the development of the normal hematopoietic system[3]. Additionally, RUNX1 has been associated with epithelial tissue development and carcinogenesis[4]. The RUNX1 gene occupies approximately 261 kb on the long arm of chromosome 21. It controls the expression of genes involved in hematopoietic differentiation, ribosome synthesis, cell cycle regulation, p53, and transforming growth factor signaling pathways via interacting with various proteins through its domains[5].

Four forms of acquired RUNX1 genetic abnormalities have been identified in AML: (1) Translocations involving RUNX1 that result in fusion genes; (2) Molecular mutations; (3) RUNX1 amplifications; and (4) Partial or total deletions of the RUNX1 gene. It has been observed that RUNX1 deletions and gains were most common in patients with unfavorable cytogenetics and that the prognosis differed dramatically, being best for individuals with RUNX1 translocations and worst for those with deletions[6].

There have been reports of the RUNX1 gene fusing with over 40 partner genes encoding structurally diverse proteins. Some RUNX1-fusions are frequent in AML, and their partner genes are known to be implicated in recurrent translocations, including RUNX1 RUNX1T1/t (8; 21) (q22; q22), t (3; 21) (q26.2; q22), t (1; 21) (p36; q22), and t (16; 21) (q24; q22). These translocations have been investigated extensively, and their potential prognostic influence is uncertain. Others have been documented in only a small number of trials, and their potential prognostic influence is uncertain[7].

Genetic testing of patients with newly diagnosed AML with RUNX1 fluorescence in situ hybridization (FISH) probe provides the opportunity to identify more instances with minor rearrangements or new partner genes of the RUNX1 locus. This will improve our understanding of the prognosis for these instances and may ultimately aid in the establishment of a very successful therapeutic treatment plan[6] that offers tailored therapy options for a large number of patients and future opportunities to prevent the development of AML[8].

This study aimed to determine the prevalence of RUNX1 gene changes in patients with newly diagnosed AML and their influence on clinical outcomes. Various prognostic markers and other clinical and laboratory results were examined in relation to the expression of RUNX1 genetic variations.

MATERIALS AND METHODS

Patients

Among the 263 adult patients who were diagnosed with AML between January 2018 and July 2019, 77 adult patients with de novo AML were included in this study. All patients were presented to the Outpatient Clinic of the Medical Oncology Department of the National Cancer Institute in Cairo, Egypt. Patients with acute promyelocytic leukemia were omitted from the study since their therapy and prognosis differ significantly from other patients with AML. Additionally, those with a history of hematologic disorders were excluded. The diagnosis of AML was based on morphological assessment of peripheral blood (PB) and BM smears, cytochemistry, immunophenotyping by flow cytometry, conventional cytogenetics, and molecular studies according to French-American-British (FAB) and World Health Organization parameters[9].

All patients underwent standard induction chemotherapy with the 3 + 7 treatment protocol (doxorubicin as a 3-day brief infusion and cytarabine 100 mg/m² as a 7-day continuous infusion). Depending on their risk assessment, patients who achieved complete remission (CR) were offered consolidation with high-dose cytarabine and human leukocyte antigen matching, followed by allogeneic BM transplantation. Refractory cases were given a high-dose cytarabine-based regimen for re-induction.

Clinical endpoints

Response to induction therapy was evaluated between days 14 and 28 post-induction. Response was categorized as CR, partial remission, stable disease, relapsed disease, or refractory disease. CR was defined as BM blasts 5%, absence of blasts with Auer rods, lack of extramedullary illness, neutrophil count > 1.0 × 10⁹/L, platelet count > 100 × 10⁹/L, and independence from red cell transfusions[10]. Patients who attained CR were divided into two groups termed normal recovery or delayed recovery based on whether they achieved CR before or after day 35, respectively[11]. Treatment failures were due to either disease resistance or relapse. The resistant disease was defined as the inability to achieve CR after completion of the initial treatment, with evidence of residual leukemia by PB and/or BM examination. Relapse was defined as BM blasts ≥ 5%, recurrence of blasts in PB, or the development of extramedullary illness.

Disease-free survival (DFS) was only defined for patients who attained a CR. It was calculated from the date of CR until the date of relapse or death, regardless of cause, censoring patients who were still alive at the time of the last follow-up. Overall survival (OS) was estimated from the date of protocol inclusion to the date of death or last follow-up/measured from the date of diagnosis to the date of death or last follow-up.

Cytogenetic analysis

Pretreatment diagnostic conventional karyotyping was applied to BM samples employing G-banded metaphase cells from unstimulated 24-h cultures following the standard techniques. Using the IKAROS imaging system, at least 20 metaphases were analyzed in the majority of cases (Metasystems, Altlussheim, Germany). Through using International System for Human Cytogenetic Nomenclature, the karyotypes were interpreted (ISCN 2016)[12]. FISH was performed according to the manufacturer’s instructions using locus-specific probes XL RUNX1 dual-color Break-Apart probe (MetaSystems) to detect RUNX1 rearrangements, deletions, and amplifications, which together represented total RUNX1 abnormalities.

A minimum of 10 metaphases and 200 interphase nuclei were studied using a fluorescent microscope (AxioImager.Z1 mot; Carl Zeiss Ltd., Hertshire, United Kingdom) with the proper filter settings. The ISIS imaging system was utilized for image capture and processing (Metasystems).

Molecular detection of fusion gene transcripts and mutational analysis

Total RNA was extracted from BM or PB samples using Qiagen RNA Blood Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. RNA was reverse transcribed using a high-capacity complementary DNA reverse transcription Kit (Applied Biosystems, Waltham, MA, United States) for the identification of fusion transcripts t (9; 22) (q34; q11), t (8; 21) (q22; q22), and inv (16) (p13q22) in accordance with the BIOMED-1 guidelines[13]. Mutation analysis of four other significant molecular marker genes, NPM1, FLT3/ITD, FLT3/TKD, and KIT, was carried out using genomic DNA-PCR as directed by the manufacturer.

Immunophenotype analysis

In all cases, blast cells in bone marrow aspiration samples were immunophenotyped using an EPICS XL Coulter Flow Cytometer (Beckman Coulter, Hialeah, FL, United States). A large panel of myeloid markers [myeloperoxidase (MPO), CD13, CD33, CD117, CD14, and CD15], lymphoid markers (CD10, CD19, CD22, CD79a, CD20, Cyto IgM, Kappa and Lambda for B lymphoid series, and CD3, CD2, 4, 8, 7, and 5 for T lymphoid series), and the stem cell marker CD34, as well as CD56 and human leukocyte antigen-DR, were used to confirm the diagnosis of AML.

Statistical analysis

Using version 22 of the statistical software program SPSS, data were analyzed (IBM, Armonk, NY, United States). According to the relevant normality test, quantitative data were described as mean ± standard deviation or median and interquartile ranges or as numbers and percentages for qualitative data. The relationship between RUNX-1 anomaly and patient clinical characteristics was evaluated using the χ² test and/or Fisher exact test, where applicable. Numerical variables from two groups were compared using the Mann-Whitney test. The Kaplan-Meier test was utilized for survival analysis, whilst the log-rank test was utilized to compare survival curves. All tests were run with an alpha level of 0.05 and a confidence interval of 95%.

Compliance with ethical standards

Every patient gave a written informed consent. The study was conducted in accordance with the Helsinki Declaration of 2011 and was approved by the internal review board of the National Cancer Institute and the Faculty of Medicine Research Ethics Committee at Cairo University (Code: MS-38-2020).

RESULTS

Patient characteristics

The median age of the 77 Egyptian patients with de novo AML was 42 (range of 18 to 82) years. Males represented 61% (47/77) of patients. Thirty-eight patients (49.4%) were AML-M2, while 22 patients (28.6%) were AML-M4. Based on genetic findings, the patients with AML were classified according to the European LeukemiaNet genetic risk classification into 18 patients (23.4%) with low risk, 42 patients (54.5%) with intermediate risk, and 17 patients (22.1%) with high-risk stratification. By molecular screening, recurrent translocations were identified in 15/77 cases (19.5%); 7 cases (9.1%) with t (8; 21) (q22; q22), 6 cases (7.8%) with inv (16) (p13q22), and 2 cases (2.6%) with t (9; 22) (q34; q11.2). Forty cases (51.9%) achieved CR, while 36 cases (46.8%) died before day 28. The detailed demographic, clinical, and laboratory characteristics of our patients are illustrated in Table 1.

Table 1 Clinical features of the assessed patients with acute myeloid leukemia.

Parameter	Frequency	Percent	Median (IQR)
Sex
Male	47	61
Female	30	39
Age in years			42 (18-82)
< 50	52	67.5
≥ 50	25	32.5
TLC as × 109/L			20 (1-377)
Hb in g/dL			8.2 ± 2.39
Platelets as × 109/L			32 (1-658)
PB blast as %			53 (0-63)
BM blast as %			69 (20-97)
MCN			45.8 ± 2.62
< 46	10	14.5
46	47	68.1
> 46	12	17.4
BM cellularity
Hypocellular	3	3.9
Normocellular	11	14.3
Hypercellular	63	81.8
Hepatomegaly
Absent	57	74
Present	20	26
Splenomegaly
Absent	60	77.9
Present	17	22.1
Lymphadenopathy
Absent	52	67.5
Present	25	32.5
FAB classification
M0	4	5.2
M1	10	13
M2	38	49.4
M4	22	28.6
M5a	2	2.6
M7	1	1.3
t (8; 21)
Absent	70	90.9
Present	7	9.1
inv16
Absent	71	92.2
Present	6	7.8
t (9; 22)
Absent	75	97.4
Present	2	2.6
Genetic risk
High	17	22.1
Intermediate	42	54.5
Low	18	23.4
FLT3-ITD
Wild	69	89.6
Mutant	8	10.4
FLT3-TKD
Wild	75	97.4
Mutant	2	2.6
C-KIT
Wild	76	98.7
Mutant	1	1.3
NPM
Wild	63	81.8
Mutant	14	18.2
BM blast on day 15			0.03 (0-1)
BM cellularity on day 15
Hypocellular	41	68.3
Normocellular	11	18.3
Hypercellular	8	13.3
BM blast on day 28			0.03 (0-1)
BM cellularity on day 28
Hypocellular	6	14.6
Normocellular	21	51.2
Hypercellular	14	34.1
CR
Negative	37	48.1
Positive	40	51.9
Delayed CR
Negative	70	90.9
Positive	7	9.1
Resistance
Negative	67	87
Positive	10	13
Relapse
Negative	63	81.8
Positive	14	18.2
Death
Negative	20	26
Positive	57	74
Early death
Negative	41	53.2
Positive	36	46.8

BM: Bone marrow; CR: Complete remission; FAB: French-American-British; Hb: Hemoglobin; IQR: Interquartile range; MCN: Modal chromosomal number; PB: Peripheral blood; TLC: Total leucocytic count.

Cytogenetics analysis

Excluding 8 patients who failed to show mitosis, 22 out of 69 patients (31.9%) had normal karyotyping, while 9/69 patients (13.0%) had a complex karyotype including 4 cases with recurrent translocations. The median of the modal chromosome number (MCN) was 45.8 (range: 32-53); 10 patients (14.5%) had hypodiploidy, while 12 cases (17.4%) had a hyperdiploid karyotype (MCN > 46) including 4 cases with concurrent recurrent translocations, 2 cases with inv (16) (p13q22), 1 case with t (8; 21) (q22; q22), and 1 case with t (9; 22) (q34; q11.2).

RUNX1 aberrations in de novo AML

By FISH, RUNX1 gene abnormalities were found in 32/77 patients (41.6%): 16 patients (20.8%) showed RUNX1 rearrangements; 17 patients (22.1%) had RUNX1 amplifications; and 4 patients (5.2%) had RUNX1 deletion encompassing the whole RUNX1 gene (Figure 1A).

Figure 1 Runt-related transcription factor-1 in patients with acute myeloid leukemia. A: Runt-related transcription factor-1 (RUNX1) gene alterations in acute myeloid leukemia (AML) cases; B: G-banded karyotype of a case of t (1; 21). 48, XY, +X, t (1; 21) (p36; q22), +8; C: Interphase fluorescence in situ hybridization using RUNX1 break apart probe showing a split of RUNX1 signal; D: Metaphase fluorescence in situ hybridization using a RUNX1 break apart probe showing a split of the RUNX1 signal. The magnification factor is × 63.

Out of 16 cases with RUNX1 translocations, 7 cases (43.8%) had t (8; 21) (q22; q22), 2 cases (12.5%) had t (1; 21) (Figure 1B-D), 1 case (6.3%) had t (16; 21), and 6 cases (37.5%) had unidentified partner chromosome. One or more copies of chromosome 21 in a hyperdiploid karyotype were gained in 4/17 cases (23.5%), while 10/ 17 cases (58.8%) showed RUNX1 duplications including 4 cases with concurrent RUNX1 translocations and 1 case with isochromosome 21. Three cases (17.6%) failed to show mitosis. Therefore, the differentiation could not be done.

Two out of four cases (50%) had -21; 1 case with a hypodiploid karyotype and another case with concurrent inv16 in a complex karyotype. Deletion of RUNX1 was found in 2/4 cases (50%) in a diploid karyotype including 1 case with concurrent RUNX1 translocations. The clinical characteristics of patients with and without different RUNX1 aberrations at diagnosis are presented in Table 1.

Correlation of RUNX1 aberrations with clinical features and hematological findings

Although no statistically significant correlation between RUNX1 abnormalities and hepatomegaly, lymphadenopathy, or splenomegaly was found (P = 0.434, 0.808, and 0.404, respectively), patients with RUNX1 amplification presented with splenomegaly (P = 0.009) and 52.9% of them had lymphadenopathy (P = 0.076).

Hypercellular BM was more frequent in patients with RUNX1 abnormalities (68.8%) and translocations (62.5%) than normocellular and hypocellular marrow (P = 0.042 and 0.09, respectively).

In RUNX1 translocation positive cases, AML-M2 (43.8%) was the most frequent FAB subtype, and M0 and M7 tended to be more frequent among them than in RUNX1 translocations negative cases (75% vs 25% and 100% vs 0%, respectively; P = 0.019). In addition, RUNX1 translocations were positively associated with MPO expression (P = 0.031). Regarding RUNX1 deletion, all cases were of the myeloid with monocytic phenotype (FAB-M4 and M5) (P = 0.008) and were positively associated with the expression of CD64 (P = 0.050). Otherwise, there was no significant difference between RUNX1 positive and negative cases in different clinical characteristics. The comparison of clinical characteristics of patients with and without RUNX1 gene alterations is shown in (Table 2 and Table 3).

Table 2 Association between all runt-related transcription factor-1 abnormalities and runt-related transcription factor-1 translocations with the clinicopathological features of the patients with acute myeloid leukemia.

Parameter		RUNX1 abnormalities		P value	RUNX1 translocation		P value
		Negative	Positive		Negative	Positive
Age in years, median (IQR)		43.0 (20-69)	35.0 (18-78)	0.472	42.5 (20-78)	30.5 (18-70)	0.156
TLC as × 109/L, median (IQR)		44.7 (1-377)	20.8 (2-191)	0.620	34.8 (1-377)	20.1 (2-107)	0.187
Hb in g/dL, mean ± SD		8.38 ± 2.30	7.93 ± 2.50	0.421	8.15 ± 2.20	8.37 ± 3.10	0.748
Platelets as × 109/L, median (IQR)		24 (12-266)	30 (13-185)	0.549	26 (12-226)	27 (13-185)	0.390
PB blast as %, median (IQR)		60.0 (0-99)	54.0 (5-92)	0.426	53.0 (0-99)	68.5 (33-92)	0.950
BM blast as %, median (IQR)		66 (30-97)	62 (20-90)	0.598	66 (20-97)	70 (36-90)	0.711
MCN, mean ± SD		45.90 ± 0.67	45.80 ± 4.10	0.852	46.00 ± 2.50	45.30 ± 3.10	0.362
Sex, n (%)	Male	28 (62.2)	19 (59.4)	0.817	40 (65.6)	7 (43.8)	0.151
	Female	17 (37.8)	13 (40.6)		21 (34.4)	9 (56.3)
BM cellularity, n (%)	Hypocellular	1 (2.2	2 (6.3)	0.042	3 (4.9)	0 (0.0)	0.009
	Normocellular	3 (6.7)	8 (25.0)		5 (8.2)	6 (37.5)
	Hypercellular	41 (91.1)	22 (68.8)		53 (86.9)	10 (62.5)
Hepatomegaly, n (%)	Absent	35 (77.8)	22 (68.8)	0.434	45 (73.8)	12 (75.0)	0.920
	Present	10 (22.2)	10 (31.3)		16 (26.2)	4 (25.0)
Splenomegaly, n (%)	Absent	37 (82.2)	23 (71.9)	0.404	47 (77.0)	13 (81.3)	0.718
	Present	8 (17.8)	9 (28.1)		14 (23.0)	3 (18.8)
Lymphadenopathy, n (%)	Absent	31 (68.9)	21 (65.6)	0.808	41 (67.2)	11 (68.8)	0.907
	Present	14 (31.1)	11 (34.4)		20 (32.8)	5 (31.3)
MPO, n (%)	Negative	3 (6.7)	5 (15.6)	0.265	4 (6.6)	4 (25.0)	0.031
	Positive	42 (93.3)	27 (84.4)		57 (93.4)	12 (75.0)
CD34, n (%)	Negative	26 (57.8)	16 (50.0)	0.643	35 (57.4)	7 (43.8)	0.330
	Positive	19 (42.2)	16 (50.0)		26 (42.6)	9 (56.3)
CD64, n (%)	Negative	35 (77.8)	22 (68.8)	0.434	45 (73.8)	12 (75.0)	0.920
	Positive	10 (22.2)	10 (31.3)		16 (26.2)	4 (25.0)
CD14, n (%)	Negative	39 (86.7)	29 (90.6)	0.728	53 (86.9)	15 (93.8)	0.447
	Positive	6 (13.3)	3 (9.4)		8 (13.1)	1 (6.3)
FAB, n (%)	M0	1 (2.2)	3 (9.4)	0.241	1 (1.6)	3 (18.8)	0.019
	M1	5 (11.1)	5 (15.6)		9 (14.8)	1 (6.3)
	M2	27 (60.0)	11 (34.4)		31 (50.8)	7 (43.8)
	M4	11 (24.4)	11 (34.4)		19 (31.1)	3 (18.8)
	M5	1 (2.2)	1 (3.1)		1 (1.6)	1 (6.3)
	M7	0 (0.0)	1 (3.1)		0 (0.0)	1 (6.3)
Complex, n (%)	Negative	40 (100.0)	20 (69.0)	< 0.001	50 (94.3)	10 (62.5)	0.001
	Positive	0 (0.0)	9 (31.0)		3 (5.7)	6 (37.5)
t (8; 21), n (%)	Negative	45 (100.0)	25 (78.1)	0.001	61 (100.0)	9 (56.3)	< 0.001
	Positive	0 (0.0)	7 (21.9)		0 (0.0)	7 (43.8)
inv16, n (%)	Negative	41 (91.1)	30 (93.8)	0.670	55 (90.2)	16 (100.0)	0.191
	Positive	4 (8.9)	2 (6.3)		6 (9.8)	0 (0.0)
t (9; 22), n (%)	Negative	44 (97.8)	31 (96.9)	0.806	59 (96.7)	16 (100.0)	0.463
	Positive	1 (2.2)	1 (3.1)		2 (3.3)	0 (0.0)
Cytogenetic risk, n (%)	High	8 (17.8)	9 (28.1)	0.310	12 (19.7)	5 (31.3)	0.542
	Intermediate	24 (53.3)	18 (56.3)		35 (57.4)	7 (43.8)
	Low	13 (28.9)	5 (15.6)		14 (23.0)	4 (25.0)
FLT3-ITD, n (%)	Wildtype	38 (84.4)	31 (96.9)	0.078	53 (86.9)	16 (100.0)	0.126
	Mutant	7 (15.6)	1 (3.1)		8 (13.1)	0 (0.0)
C-KIT, n (%)	Wildtype	45 (100.0)	31 (96.9)	0.416	61 (100.0)	15 (93.8)	0.208
	Mutant	0 (0.0)	1 (3.1)		0 (0.0)	1 (6.3)
NPM, n (%)	Wildtype	33 (73.3)	30 (93.8)	0.034	47 (77.0)	16 (100.0)	0.034
	Mutant	12 (26.7)	2 (6.3)		14 (23.0)	0 (0.0)
BM blast on day 28, n (%)		0.04 (0-1)	0.02 (0-1)	0.082	0.03 (0-1)	0.04 (0-1)	0.789
BM cellularity on day 28, n (%)	Hypocellular	4 (19.0)	2 (10.0)	0.507	5 (16.7)	1 (6.3)	0.831
	Normocellular	9 (42.9)	12 (60.0)		15 (50.0)	6 (54.5)
	Hypercellular	8 (38.1)	6 (30.0)		10 (33.3)	4 (36.4)
CR, n (%)	Negative	23 (51.1)	14 (43.8)	0.644	30 (49.2)	7 (43.8)	0.699
	Positive	22 (48.9)	18 (56.3)		31 (50.8)	9 (56.3)
Delayed CR, n (%)	Negative	43 (95.6)	27 (84.4)	0.093	57 (93.4)	13 (81.3)	0.131
	Positive	2 (4.4)	5 (15.6)		4 (6.6)	3 (18.8)
Relapse, n (%)	Negative	40 (88.9)	23 (71.9)	0.075	52 (85.2)	11 (68.8)	0.128
	Positive	5 (11.1)	9 (28.1)		9 (14.8)	5 (31.3)
Death, n (%)	Negative	9 (20.0)	11 (34.4)	0.192	15 (24.6)	5 (31.3)	0.589
	Positive	36 (80.0)	21 (65.6)		46 (75.4)	11 (68.8)
Early death, n (%)	Negative	21 (46.7)	20 (62.5)	0.247	30 (49.2)	11 (68.8)	0.163
	Positive	24 (53.3)	12 (37.5)		31 (50.8)	5 (31.3)

BM: Bone marrow; CR: Complete remission; FAB: French-American-British; Hb: Hemoglobin; IQR: Interquartile range; PB: Peripheral blood; MCN: Modal chromosomal number; MPO: Myeloperoxidase; RUNX1: Runt-related transcription factor-1; SD: Standard deviation; TLC: Total leucocytic count.

Table 3 Association between runt-related transcription factor-1 amplifications and runt-related transcription factor-1 deletion with clinicopathological features of acute myeloid leukemia patients.

Parameter		RUNX1 amplification		P value	RUNX1 deletion		P value
		Negative	Positive		Negative	Positive
Age in years, median (IQR)		42.0 (18-70)	36.0 (20-78)	0.694	41.5 (18-78)	21.5 (21-22)	0.175
TLC as × 109/L, median (IQR)		44.7 (1-377)	19.2 (2-91)	0.873	29.1 (1-377)	105.3 (19-191)	0.630
Hb in g/dL, mean ± SD		8.33 ± 2.50	7.70 ± 1.70	0.359	8.20 ± 2.40	8.10 ± 1.50	0.950
Platelets as × 109/L, median (IQR)		24 (12-226)	45 (21-185)	0.484	26 (12-226)	24 (18-30)	0.663
PB blast (%), median (IQR)		60 (0-99)	38 (5-72)	0.515	59 (0-99)	47 (40-54)	0.232
BM blast (%), median (IQR)		66 (30-97)	60 (20-90)	0.35	65.5 (20-97)	74.5 (71-78)	0.613
MCN, mean ± SD		45.4 ± 2.5	47.5 ± 2.4	0.007	46 ± 2.1	42.5 ± 7	0.008
Sex, n (%)	Male	34 (56.7)	13 (76.5)	0.168	45 (61.6)	2 (50.0)	0.641
	Female	26 (43.3)	4 (23.5)		28 (38.4)	2 (50.0)
BM cellularity, n (%)	Hypocellular	1 (1.7)	2 (11.8)	0.137	3 (4.1)	0 (0.0)	0.626
	Normocellular	8 (13.3)	3 (17.6)		11 (15.1)	0 (0.0)
	Hypercellular	51 (85.0)	12 (70.6)		59 (80.8)	4 (100.0)
Hepatomegaly, n (%)	Absent	47 (78.3)	10 (58.8)	0.125	53 (72.6)	4 (100.0)	0.568
	Present	13 (21.7)	7 (41.2)		20 (27.4)	0 (0.0)
Splenomegaly, n (%)	Absent	51 (85.0)	9 (52.9)	0.009	56 (76.7)	4 (100.0)	0.57
	Present	9 (15.0)	8 (47.1)		17 (23.3)	0 (0.0)
Lymphadenopathy, n (%)	Absent	44 (73.3)	8 (47.1)	0.076	49 (67.1)	3 (75.0)	0.743
	Present	16 (26.7)	9 (52.9)		24 (32.9)	1 (25.0)
MPO, n (%)	Negative	6 (10.0)	2 (11.8)	0.833	8 (11.0)	0 (0.0)	1
	Positive	54 (90.0)	15 (88.2)		65 (89.0)	4 (100.0)
CD34, n (%)	Negative	34 (56.7)	8 (47.1)	0.584	38 (52.1)	4 (100.0)	0.121
	Positive	26 (43.3)	9 (52.9)		35 (47.9)	0 (0.0)
CD64, n (%)	Negative	45 (75.0)	12 (70.6)	0.758	56 (76.7)	1 (25.0)	0.052
	Positive	15 (25.0)	5 (29.4)		17 (23.3)	3 (75.0)
CD14, n (%)	Negative	51 (85.0)	17 (100.0)	0.194	66 (90.4)	2 (50.0)	0.065
	Positive	9 (15.0)	0 (0)		7 (9.6)	2 (50.0)
FAB, n (%)	M0	2 (3.3)	2 (11.8)	0.368	4 (5.5)	0 (0.0)	0.014
	M1	6 (10.0)	4 (23.5)		10 (13.7)	0 (0.0)
	M2	32 (53.3)	6 (35.3)		38 (52.1)	0 (0.0)
	M4	17 (28.3)	5 (29.4)		19 (26.0)	3 (75.0)
	M5	2 (3.3)	0 (0)		1 (1.4)	1 (25.0)
	M7	1 (1.7)	0 (0)		1 (1.4)	0 (0.0)
Complex, n (%)	Negative	52 (94.5)	8 (47.1)	0.001	57 (87.7)	3 (75.0)	0.436
	Positive	3 (5.5)	6 (35.3)		8 (12.3)	1 (25.0)
t (8; 21), n (%)	Negative	54 (90.0)	16 (94.1)	0.602	66 (90.4)	4 (100.0)	0.516
	Positive	6 (10.0)	1 (5.9)		7 (9.6)	0 (0.0)
inv16, n (%)	Negative	55 (91.7)	16 (94.1)	0.739	68 (93.2)	3 (75.0)	0.282
	Positive	5 (8.3)	1 (5.9)		5 (6.8)	1 (25.0)
t (9; 22), n (%)	Negative	59 (98.3)	16 (94.1)	0.395	71 (97.3)	4 (100.0)	0.737
	Positive	1 (1.7)	1 (5.9)		2 (2.7)	0 (0.0)
Cytogenetic risk, n (%)	High	12 (20.0)	5 (29.4)	0.15	15 (20.5)	2 (50.0)	0.288
	Intermediate	31 (51.7)	11 (64.7)		40 (54.8)	2 (50.0)
	Low	17 (28.3)	1 (5.9)		18 (24.7)	0 (0.0)
FLT3-ITD, n (%)	Wildtype	52 (86.7)	17 (100.0)	0.188	66 (90.4)	3 (75.0)	0.361
	Mutant	8 (13.3)	0 (0)		7 (9.6)	1 (25.0)
C-KIT, n (%)	Wildtype	59 (98.3)	17 (100.0)	0.592	72 (98.6)	4 (100.0)	0.814
	Mutant	1 (1.7)	0 (0)		1 (1.4)	0 (0.0)
NPM, n (%)	Wildtype	47 (78.3)	16 (94.1)	0.173	60 (82.2)	3 (75.0)	0.717
	Mutant	13 (21.7)	1 (5.9)		13 (17.8)	1 (25.0)
BM cellularity on day 28, n (%)	Hypocellular	6 (20.0)	0 (0)	0.152	5 (12.8)	1 (25.0)	0.22
	Normocellular	13 (43.3)	8 (47.1)		21 (53.8)	0 (0.0)
	Hypercellular	11 (36.7)	3 (25.0)		13 (33.3)	1 (25.0)
CR, n (%)	Negative	29 (48.3)	8 (47.1)	0.926	36 (49.3)	1 (25.0)	0.616
	Positive	31 (51.7)	9 (52.9)		37 (50.7)	3 (75.0)
Delayed CR, n (%)	Negative	56 (93.30	14	0.177	66 (90.4)	4 (100.0)	0.516
	Positive	4 (6.7)	3 (25.0)		7 (9.6)	0 (0.0)
Relapse, n (%)	Negative	50 (83.3)	13	0.496	61 (83.6)	2 (50.0)	0.149
	Positive	10 (16.7)	4		12 (16.4)	2 (50.0)
Death, n (%)	Negative	13 (21.7)	7	0.125	19 (26.0)	1 (25.0)	0.964
	Positive	47 (78.3)	10		54 (74.0)	3 (75.0)
Early death, n (%)	Negative	30 (50.0)	11	0.41	39 (53.4)	2 (50.0)	0.894
	Positive	30 (50.0)	6		34 (46.6)	2 (50.0)

BM: Bone marrow; CR: Complete remission; FAB: French-American-British; Hb: Hemoglobin; IQR: Interquartile range; MCN: Modal chromosomal number; PB: Peripheral blood; MPO: Myeloperoxidase; RUNX1: Runt-related transcription factor-1; SD: Standard deviation; TLC: Total leucocytic count.

Association of RUNX1 aberrations with cytogenetic and molecular abnormalities

There was a highly statistically significant relation between RUNX1 copy number variations and MCN where positive cases to RUNX1 amplification tended to have higher MCN, while RUNX1 deletion cases tended to have lower MCN (P = 0.007 and 0.008, respectively). Cases positive to RUNX1 abnormalities, translocations, and amplifications tended to have complex karyotypes compared to RUNX1-negative cases (P = 0.000, 0.001, and 0.001, respectively). There was no statistically significant correlation between all types of RUNX1 abnormalities and other different cytogenetic abnormalities (as -2, -3, -7, -11, -13, -22, +8, +13, +17), inv16, and t (9; 22) (P = 0.623, 0.670, and 0.806, respectively).

To investigate the interaction of gene mutations in the pathogenesis of adult AML, screening of mutational status of four other genes was performed. Among the 32 patients with RUNX1 abnormalities, 4 cases showed additional molecular abnormalities including 2 cases with RUNX1 deletion, of which one had an FLT3-ITD mutation and the other case had concomitant FLT3/TKD and NPM1 mutations, 1 case with t (8; 21) and c-KIT mutation, and 1 case with RUNX1 amplification with NPM1 mutation.

FLT3-ITD mutations tended to be more prevalent in RUNX1-negative cases compared to positive cases. Out of 8 patients with AML with FLT3-ITD mutations, 7 patients (87.5%) were negative for RUNX1 abnormalities, while only 1 patient was positive for RUNX1 abnormalities. However, this relation was statistically insignificant (P = 0.078). Similarly, NPM1 mutations were rarely seen in RUNX1 abnormalities and translocations (P = 0.034 and 0.034, respectively). Otherwise, there were no significant differences between RUNX1 positive and negative cases regarding other cytogenetic and molecular abnormalities, as shown in Table 2 and Table 3 (P > 0.05).

RUNX1 translocations, amplifications, and deletions were more frequent in the intermediate risk group (43.8%, 64.7%, and 50.0%) and the high-risk group (31.3%, 29.4%, and 50.0%) than in the low risk group (25.0%, 5.9%, and 0%), but the relationship was not statistically significant (P = 0.542, 0.150, and 0.288, respectively).

Impact of RUNX1 aberrations on response to treatment and clinical outcome

The response to treatment at day 28 of starting chemotherapy revealed that 36 of 77 (46.8%) cases died before day 28, and 33 cases (42.9%) achieved CR at day 15, 7 cases (9.1%) achieved delayed CR, 10 cases (13%) were resistant to treatment, and eventually 57 cases (74%) died.

Fourteen cases (18.2%) relapsed after achieving CR, 9/14 cases (64.2%) were positive to RUNX1 abnormalities. One patient had t (9; 22) (q34; q11.2), and another patient had inv(16). The detailed karyotype and analysis of the outcome of those patients are summarized in Supplementary Table 1.

Although positive cases to RUNX1 abnormalities tended to have delayed CR and relapse compared to negative cases (71.4% vs 28.6% P = 0.093 and 64.3% vs 35.7% P = 0.075, respectively), the relationship was not statistically significant. The statistical analysis showed the absence of a relationship between the achievement of CR and RUNX1 abnormalities (P = 0.644), RUNX1 translocation (P = 0.699), RUNX1 amplification (P = 0.926), and RUNX1 deletion (P = 0.616), as shown in Table 2 and Table 3.

After a follow-up period of 44.2 months, the present study showed that there was no significant difference between positive and negative RUNX1 aberration cases regarding the OS (Figure 2). Patients with RUNX1 deletion had significantly poorer DFS than those without RUNX1 deletion (mean: 3.03 months vs 27.20 months, respectively; P < 0.001). No other significant differences were observed between any other type of RUNX1 alterations and negative cases regarding DFS (Figure 3).

Figure 2 Kaplan-Meier survival curves. A: Runt-related transcription factor-1 (RUNX1) abnormality; B: RUNX1 amplification; C: RUNX1 translocation; D: RUNX1 deletion on overall survival rates in acute myeloid leukemia patients.

Figure 3 Kaplan-Meier survival curves. A: Runt-related transcription factor-1 (RUNX1) abnormality; B: RUNX1 amplification; C: RUNX1 translocation; D: RUNX1 deletion on disease-free survival of patients with adult acute myeloid leukemia.

DISCUSSION

The role of RUNX1 mutations in cytogenetically normal AML had been identified. However, the prognostic impact of RUNX1 translocations other than t (8; 21) (q22; q22), RUNX1 deletions, and amplifications are still unknown. As a result, addressing such interactions is critical for further risk classification and eventually the development of a successful therapeutic plan.

In this study, FISH was used to screen for RUNX1 gene alterations in 77 newly diagnosed adult patients with de novo AML, and the results were compared to clinical characteristics and prognosis. RUNX1 abnormalities were divided into four categories: (1) RUNX1 translocations; (2) RUNX1 amplifications; (3) RUNX1 deletions; and (4) RUNX1 abnormalities, which encompass all three types.

In agreement with Haferlach et al[6], total RUNX1 abnormalities were detected in 41.6% of cases of which RUNX1 amplification was the most common alteration (22.1%) followed by RUNX1 translocations (20.8%) then RUNX1 deletion (5.2%). Baldus et al[14] investigated 12 patients with AML with complicated karyotypes and chromosome 21 anomalies and showed that amplification of two chromosome 21 areas was frequently seen in AML with complex karyotypes using comparative genomic hybridization, supporting the notion that gain of chromosome 21 material appears to be a nonrandom event implicated in AML. Baldus et al[14] reasoned that this might be related to the function of a specific gene or set of genes.

The clinical and genetic characteristics of patients with and without RUNX1 alterations were compared. In agreement with Yamato et al[15], there were no statistically significant differences in any form of RUNX1 abnormalities with respect to age, sex, total leucocytic count, hemoglobin level, platelet count, PB count, or BM blast cell counts at presentation. Tang et al[16] discovered that male patients had a higher rate of RUNX1 alterations than female patients, while Haferlach et al[17] reported that patients with RUNX1 deletion were considerably older than those with two RUNX1 copies and had a lower WBC count. On the other hand, Said et al[18] used reverse transcription-quantitative PCR to explore the role of RUNX1 gene expression in Egyptian patients with AML and found that male patients had significantly higher RUNX1 expression. This discrepancy can be attributed to the difference in the technique used, a difference in the sample size, and variation in inclusion criteria between the two studies, in spite of conducting both studies on the same race.

The current data showed that splenomegaly is common in patients with RUNX1 amplification, and the majority of them had lymphadenopathy. Hypercellular marrow was also more frequent than normocellular and hypocellular marrow in RUNX1-abnormalities and translocations. No published research had associated RUNX1-abnormalities with hepatomegaly, splenomegaly, lymphadenopathy, or BM cellularity at the time of diagnosis, to our knowledge.

Of interest, 43.8% of patients with RUNX1 translocation were FAB AML-M2 and were associated with MPO expression. Also, M0 and M7 were more prevalent in RUNX1 translocation positive cases than in negative instances. All patients with RUNX1 deletion had a myeloid with monocytic phenotype (FAB-M4 and M5) and were favorably related with CD64 expression. These results matched those of Haferlach et al[6], who found that 45.2% of RUNX1 translocation cases were FAB type M1 and M2. However, in contrast to our findings, they found that in cases of RUNX1 deletion, M0 was the most common AML subtype. This discrepancy can be attributed to the racial and sample size variations. While Said et al[18] found that out of 42% of patients with AML classified as M2, translocation t (8; 21) was found in only 6.6% of the cases. The researchers also discovered no link between FAB subtypes and RUNX1 expression.

In keeping with earlier studies, translocations, amplifications, and deletions of RUNX1 were more common in the intermediate and high-risk groups, although not statistically significant.

In addition, as previously reported[17,19,20], positive cases of RUNX1 amplification have a higher MCN, whereas RUNX1 deletion cases have a lower MCN. This could be explained by the fact that chromosomal gain is nonrandom in patients with AML with hyperdiploid karyotypes, and chromosome 21 is one of the most frequently gained chromosomes in these circumstances.

The present data showed that RUNX1 abnormalities, translocations, and amplifications are associated with more complex karyotypes than RUNX1-negative instances, which supports prior research[6,14,17]. It was concluded that chromosome 21 amplifications are common in people with complicated karyotypes.

In terms of other molecular mutations, 8 cases tested positive for the FLT3-ITD mutation. RUNX1 translocations and amplifications were all negative, but RUNX1 deletion was positive in 1 patient (12.5%). These relationships, however, did not exhibit statistical significance. According to Said et al[18], patients with higher RUNX1 expression were more likely to have FLT3-ITD mutation than patients with lower RUNX1 expression. This suggested that RUNX1 could operate as an oncogene that causes leukemogenesis and as a surrogate marker for other mutations, particularly FLT3-ITD, when expressed at high levels.

Furthermore, NPM1 mutations were mutually exclusive of RUNX1 abnormalities and RUNX1 translocations. This suggests that RUNX1 mutation shares a similar genetic pathway role with NPM1 mutations in leukemia development. This was supported by the study of Zuo et al[21], who stated that NPM1 mutant interacts with PU.1/CEB-PA/RUNX1 transcription factor complexes to block myeloid differentiation. Additionally, NPM1 mutations were found at a lower frequency in RUNX1 copy number variation-positive patients than in negative cases, but the difference was not statistically significant. These RUNX1 mutations have genetic characteristics that are similar to those previously described in patients with AML[6,15,17].

There was no statistically significant association between any form of RUNX1 modification and the achievement of CR or OS in terms of clinical outcome. Although the median DFS differed significantly amongst the three types of RUNX1 changes (9.5, 25.7, and 1.5 months, respectively), RUNX1 amplifications had the best prognosis. Only those with RUNX1 deletion exhibited a considerably worse outcome than cases without the mutation. Consistently, previous research has found that OS varies dramatically between the different forms of RUNX1 alterations, with RUNX1 deletions having the worst outcome[6].

On the contrary, other studies[6,22] found that the outcome differed considerably across RUNX1 alterations and was better in individuals with RUNX1 translocations. Discrepancies in the results could be related to the fact that most research looked at RUNX1 mutations in combination with cytogenetic abnormalities, whereas just a few looked at the impact of RUNX1 translocations, amplifications, and deletions in AML cases from different risk groups. Furthermore, it is thought that the different sorts of techniques used in the studies are a key source of heterogeneity. Said et al[18] found no significant impact of RUNX1 expression on OS or DFS rates, but Chen et al[23] found that RUNX1 mutation was linked to a lower risk-free survival rate.

Nine out of seventeen (53%) RUNX1 amplification positive cases had RUNX1 duplications. Four cases (29.4%) had RUNX1 translocations at the same time, and five cases (29.4%) obtained one or more copies of chromosome 21, four of which were in a hyperdiploid karyotype. All of these are favorable prognostic markers that may help patients with positive RUNX1 amplification live longer. This could explain the disparity in clinical outcomes between our data and that of others.

Furthermore, when inv16 and t (9; 22) cases with favorable prognosis were excluded from statistical analysis, the mean DFS in positive RUNX1 deletion cases was 3.033 months compared to 27.231 months in negative RUNX1 deletion cases. This suggests that RUNX1 deletion has the worst prognosis, even if other strong prognostic markers like inv16 and t (9; 22) are present.

CONCLUSION

Our data presented a pilot study for RUNX1 gene alterations in a cohort of patients with de novo AML. RUNX1 abnormalities were detected in 41.6% of patients. RUNX1 translocations occurred predominantly in FAB M2, M0, and M7 while RUNX1 deletions were of myeloid with monocytic phenotype (FAB-M4 and M5). Cases positive for RUNX1 abnormalities, translocations, and amplifications tended to have complex karyotypes. RUNX1 abnormalities were mutually exclusive of NPM1 mutations. RUNX1 deletion was an independent adverse parameter for DFS. Further trials with larger numbers of RUNX1 abnormal cases are warranted to further highlight the prognostic features and the predictive significance of this abnormality.

ACKNOWLEDGEMENTS

The authors would like to thank all the patients who were included in this study.