Association of ARID5B gene variants with acute lymphoblastic leukemia in Yemeni children

Abstract

Studies have shown an association between ARID5B gene polymorphisms and childhood acute lymphoblastic leukemia. However, the association between ARID5B variants and acute lymphoblastic leukemia among the Arab population still needs to be studied. The aim of this study was to investigate the association between ARID5B variants with acute lymphoblastic leukemia in Yemeni children. A total of 14 ARID5B gene single nucleotide polymorphisms (SNPs) were genotyped in 289 Yemeni children, of whom 136 had acute lymphoblastic leukemia and 153 were controls, using the nanofluidic Dynamic Array (Fluidigm 192.24 Dynamic Array). Using logistic regression adjusted for age and gender, the risks of acute lymphoblastic leukemia were presented as odds ratios and 95% confidence intervals. We found that nine SNPs were associated with acute lymphoblastic leukemia under additive genetic models: rs7073837, rs10740055, rs7089424, rs10821936, rs4506592, rs10994982, rs7896246, rs10821938, and rs7923074. Furthermore, the recessive models revealed that six SNPs were risk factors for acute lymphoblastic leukemia: rs10740055, rs7089424, rs10994982, rs7896246, rs10821938, and rs7923074. The gender-specific impact of these SNPs under the recessive genetic model revealed that SNPs rs10740055, rs10994982, and rs6479779 in females, and rs10821938 and rs7923074 in males were significantly associated with acute lymphoblastic leukemia risk. Under the dominant model, SNPs rs7073837, rs10821936, rs7896246, and rs6479778 in males only showed striking association with acute lymphoblastic leukemia. The additive model revealed that SNPs with significant association with acute lymphoblastic leukemia were rs10821936 (both males and females); rs7073837, rs10740055, rs10994982, and rs4948487 (females only); and rs7089424, rs7896246, rs10821938, and rs7923074 (males only). In addition, the ARID5B haplotype block (CGAACACAA) showed a higher risk for acute lymphoblastic leukemia. The haplotype (CCCGACTGC) was associated with protection against acute lymphoblastic leukemia. In conclusion, our study has shown that ARID5B variants are associated with acute lymphoblastic leukemia in Yemeni children with several gender biases of ARID5B single nucleotide polymorphisms reported.

Keywords

Childhood leukemia ARID5B single nucleotide polymorphism haplotype genotyping

Introduction

Acute lymphoblastic leukemia (ALL) is the most common pediatric cancer among patients younger than 15 years, representing 26% of all tumors and 78% of leukemias in these patients.¹ At least 54,000 new ALL cases in Asia are estimated to occur each year.² In Yemen, the estimated incidence rate of leukemia is 10% of all cancers, with a mortality rate of 12.3%.³ The available data on the epidemiology of childhood leukemia in Yemen are limited, but it has been reported that the highest frequency of leukemia occurs in children under the age of 4 years.⁴

Leukemic clones have been detected early in pre-natal stages in both individuals who develop and do not develop the disease.^5,6 Additional genetic lesions are required to activate the leukemogenesis process.^7,8 Genome wide association (GWA) studies on populations of European ancestry have revealed that ARID5B single nucleotide polymorphisms (SNPs) are associated with childhood ALL.^9–11 Several follow-up studies have confirmed the association of ALL risk with ARID5B genetic variations.^12–20

ARID5B belongs to the family of AT-rich interaction domain transcription factors²¹ that plays a critical role in embryogenesis and cell growth regulation.²² Childhood ALL shows substantial variations in disease incidence across geographic areas, with higher prevalence in European than that reported in Asian populations.²³ Different populations are well distinguished by race/ethnicity-related genetic polymorphisms, which may lead to these incidence disparities, in addition to environmental risk factors.^24,25 It is important to study the effects of these genetic variations in different ethnicities to understand the differences in susceptibility to ALL and infer the causes and mechanism of ALL progression.

To the best of our knowledge, the impact of ARID5B variants on childhood ALL in Yemen has not been studied. Therefore, the goal of this study was to determine the possible association of ARID5B gene SNPs and haplotypes with ALL in Yemeni children of Arab Asian ancestry.

Materials and methods

Ethical approval

Approval for this case–control study was obtained from the Medical Ethics Committee of the University Malaya Medical Centre (RF No.: 989.39). Permission was obtained from the hospitals authorities in Sana’a, Yemen before commencing the study. Written informed consent was obtained from all parent or guardian respondents.

Subjects and collection of data

Leukemic children aged between 2 and 15 years who had been diagnosed with ALL and were under treatment at government hospitals in Sana’a, Yemen, were recruited to this study as the case group. The diagnosis of ALL was based on bone marrow and peripheral blood morphology and histochemistry. For the control group, healthy children with normal complete blood counts were approached to participate in this study. Venous blood samples (4 mL) were collected under sterile conditions from each child into two labeled Vacutainer® containing K2EDTA: one for hematologic analysis and the second for DNA extraction.

Hematologic analysis

Complete blood counts were performed using a Mythic™ 22 AL fully automated 22-parameter hematology analyzer (Orphee SA, Switzerland).

Genetic analysis

ARID5B intronic SNPs—rs7073837, rs10740055, rs7089424, rs10821936, rs4506592, rs10994982, rs7896246, rs10821938, rs7923074, rs6479778, rs4948487, rs6479779, rs2893881, and rs10994990—were chosen based on the findings of previous studies.^{9–11,15,17,19,26,27} Extraction of genomic DNA from peripheral blood leukocytes was performed using the FlexiGene DNA Kit (Qiagen, Valencia, CA, USA). ARID5B SNPs genotyping was conducted by Fluidigm 192.24 Dynamic Array and read by a BioMark HD Reader (Fluidigm Corporation, San Francisco, CA, USA) according to the manufacturer’s protocol. The call rates for ARID5B SNPs rs10740055, rs10821936, rs4506592, rs10994982, rs10821938, and rs4948487 were > 96% in both control and ALL groups. The call rates for rs7073837, rs7089424, rs7896246, rs6479778, and rs6479779 were ≥86% in the control group and ≥90% in the ALL group, while rs7923074 showed a call rate ≥86% in both groups. The concordance rate was 100% based on 40% blind duplicate samples. Controls without template were analyzed in each run of the Fluidigm 192.24 Dynamic Array to ensure no DNA contamination was present.

Statistical analysis

Hardy–Weinberg Equilibrium (HWE) analysis was achieved with DeFinetti software (http://ihg.gsf.de/cgi-bin/hw/hwa1.pl; the Institute of Human Genetics). SNP & Variation Suite v8.x software (Golden Helix, Bozeman, MT, USA) was used to analyze linkage disequilibrium (LD) and for building up haplotypes and diplotypes of related SNPs. Statistical analyses were done using Social Package of Statistical Science (SPSS) 20.0 (LEAD Technologies, Inc., Charlotte, NC, USA). Hematologic parameters were log-transformed and then transformed back and presented as geometric means. Comparisons between groups were assessed by independent sample t-test. The evaluation of the association between ALL and ARID5B SNP genetic models (recessive, dominant, and additive), haplotypes, and diplotypes was performed by logistic regression analysis controlled for age and gender as covariates. A significant difference was considered when p < 0.05.

Results

A total of 136 children with ALL and 153 non-leukemic (control) children were included in this study. The demographic information and hematologic test results of the children are provided in Table 1.

Table 1.

Demography and hematologic parameters.

Parameters	Control (n = 153)	ALL (n = 136)	p value
Age (years)	9.59 (9.04–10.18)	6.61 (6.04–7.22)	<0.001
Gender
Female	67	54
Male	86	82
Hemoglobin (g/dL)	13.95 (13.78–14.13)	11.1 (10.74–11.52)	<0.001
Red blood cells count × 10¹²/L	5.23 (5.16–5.31)	4.05 (3.92–4.18)	<0.001
Hematocrit %	42.06 (41.60–42.53)	32.82 (31.70–33.99)	<0.001
Mean corpuscular volume (fL)	80.34 (79.41–81.28)	81.18 (79.82–82.57)	0.32
Mean corpuscular hemoglobin (pg)	26.69 (26.28–27.10)	27.50 (26.98–28.02)	<0.05
Mean corpuscular hemoglobin concentration (g/dL)	33.19 (32.96–33.42)	33.87 (33.63–34.13)	<0.001
White blood cells count × 10⁹/L	6.01 (5.72–6.31)	6.69 (5.57–8.04)	0.26
Neutrophil count × 10⁹/L	2.29 (2.10–2.50)	2.06 (1.75–2.40)	0.23
Lymphocyte count × 10⁹/L	2.83 (2.70–2.97)	2.41 (2.03–2.83)	0.06
Monocyte count × 10⁹/L	0.55 (0.51–0.59)	0.50 (0.43–0.57)	0.22
Eosinophil count × 10⁹/L	0.17 (0.14–0.21)	0.14 (0.09–0.18)	0.23
Basophil count × 10⁹/L	0.03 (0.03–0.04)	0.05 (0.04–0.06)	<0.05
Platelets count × 10⁹/L	327.9 (317.0–339.2)	139.68 (111.9–174.2)	<0.001

ALL: acute lymphoblastic leukemia.

Results presented represent geometric means (95% confidence interval of means), ALL versus normal control group evaluated by independent samples t-test.

ARID5B SNPs rs7073837, rs10740055, rs7089424, rs10821936, rs4506592, rs10994982, rs7896246, rs10821938, rs7923074, rs6479778, rs4948487, and rs6479779 were consistent with HWE (p > 0.05) in the control group. Two SNPs rs2893881 and rs10994990 deviated from HWE (p = 0.02, 0.0001, respectively) so they were no longer analyzed.

The logistic regression models (controlled for age and gender) revealed that the additive genetic models of ARID5B SNPs rs7073837 rs10740055, rs7089424, rs10821936, rs4506592, rs10994982, rs7896246, rs10821938, and rs7923074 were significantly associated with ALL (odds ratio (OR) = 1.66, 1.60, 1.46, 1.69, 1.44, 1.58, 1.75, 1.48, 1.7; p = 0.008, 0.009, 0.047, 0.004, 0.04, 0.01, 0.004, 0.02, 0.005, respectively) (Table 2).

Table 2.

Association of ARID5B polymorphisms with acute lymphoblastic leukemia among Yemeni children.

ARID5B SNPs	Alleles	Control		Leukemic		Recessive model		Dominant model		Additive model
ARID5B SNPs	Alleles	Freq.	11/12/22	Freq.	11/12/22	OR (95% CI)	p value	OR (95% CI)	p value	OR (95%CI)	p value
rs7073837	C/A	0.38	55/54/24	0.53	27/62/34	1.75 (0.92–3.33)	0.09	2.44 (1.33–4.47)	0.004	1.66 (1.14–2.40)	0.008
rs10740055	A/C	0.45	49/71/33	0.56	29/63/44	1.96 (1.11–3.48)	0.02	1.83 (1.03–3.27)	0.04	1.60 (1.13–2.27)	0.009
rs7089424	T/G	0.36	55/59/19	0.46	38/58/29	2.19 (1.08–4.45)	0.03	1.50 (0.86–2.63)	0.16	1.46 (1.01–2.13)	0.047
rs10821936	T/C	0.33	74/57/22	0.45	44/62/30	1.89 (0.98–3.65)	0.058	2.00 (1.19–3.38)	0.009	1.69 (1.19–2.41)	0.004
rs4506592	G/A	0.38	61/68/24	0.46	42/62/32	1.74 (0.92–3.29)	0.09	1.58 (0.93–2.69)	0.09	1.44 (1.02–2.05)	0.04
rs10994982	G/A	0.44	51/69/33	0.55	31/61/44	2.02 (1.13–3.58)	0.017	1.8 (0.99–3.09)	0.05	1.58 (1.12–2.23)	0.01
rs7896246	G/A	0.32	66/51/17	0.44	40/59/26	2.14 (1.02–4.48)	0.04	2.08 (1.20–3.61)	0.01	1.75 (1.19–2.57)	0.004
rs10821938	C/A	0.41	57/68/28	0.50	39/59/38	1.93 (1.07–3.49)	0.03	1.57 (0.91–2.70)	0.11	1.48 (1.05–2.09)	0.02
rs7923074	C/A	0.41	51/55/26	0.53	29/53/36	2.12 (1.12–4.02)	0.02	2.02 (1.11–3.66)	0.02	1.70 (1.17–2.47)	0.005
rs6479778	C/T	0.23	83/42/10	0.28	67/50/10	1.08 (0.38–3.06)	0.88	1.38 (0.79–2.41)	0.26	1.20 (0.78–1.85)	0.40
rs4948487	A/C	0.48	42/75/36	0.52	33/64/39	1.46 (0.83–2.58)	0.19	1.31 (0.73–2.33)	0.36	1.27 (0.90–1.80)	0.18
rs6479779	G/C	0.42	49/72/26	0.48	38/61/32	1.65 (0.87–3.13)	0.13	1.38 (0.78–2.43)	0.27	1.34 (0.93–1.94)	0.12

Freq.: risk allele frequency; SNP: single nucleotide polymorphism; OR: odds ratio; CI: confidence interval.

Controlled for age and gender. In the additive model, genotype of homozygote for the non-risk allele 11, heterozygote 12, and homozygote for the risk allele 22. The recessive model was defined as 22 versus 12 + 11 and dominant model as 22 + 12 versus 11. The outliers (studentized residual is greater than 2.0 or less than −2.0) were excluded. Bold values are significant.

The dominant genetic models revealed a significant association of rs7073837 (OR = 2.44, p = 0.004), rs10740055 (OR = 1.83, p = 0.04), rs10821936 (OR = 2.00, p = 0.009), rs7896246 (OR = 2.08, p = 0.01), and rs7923074 (OR = 2.02, p = 0.02) with ALL. Similarly, rs10740055 (OR = 1.96, p = 0.02), rs7089424 (OR = 2.19, p = 0.03), rs10994982 (OR = 2.02, p = 0.017), rs7896246 (OR = 2.14, p = 0.04), rs10821938 (OR = 1.93, p = 0.03), and rs7923074 (OR = 2.12, p = 0.02) showed a significant association with ALL under recessive genetic models. ARID5B SNPs rs6479778, rs4948487, and rs6479779 showed no association (Table 2).

In relation to gender, a significant association with ALL under recessive genetic models at SNPs rs10740055 (OR = 4.89, p = 0.003), rs10994982 (OR = 4.92, p = 0.002) and rs6479779 (OR = 5.18, p = 0.009) was found in females, while rs10821938 (OR = 2.61, p = 0.02) and rs7923074 (OR = 2.51, p = 0.035) were significantly associated with ALL in males (Table 3). The additive genetic models of rs7073837 (OR = 1.90, p = 0.04), rs10740055 (OR = 2.13, p = 0.01), rs10994982 (OR = 2.08, p = 0.02), and rs4948487 (OR = 1.87, p = 0.047) in females showed significant association with ALL. Conversely, rs7089424 (OR = 1.72, p = 0.03), rs7896246 (OR = 1.88, p = 0.01), rs10821938 (OR = 1.75, p = 0.02), and rs7923074 (OR = 1.69, p = 0.03) revealed a significant association under additive genetic models in males. The additive genetic model of rs10821936 was significantly associated with ALL in both males and females (OR = 1.70, p = 0.02; OR = 2.00, p = 0.03, respectively) (Table 3).

Table 3.

Associations by gender for ARID5B SNPs with acute lymphoblastic leukemia among Yemeni children.

ARID5B SNPs	Alleles	Gender	Control		Leukemic		Recessive		Dominant		Additive
ARID5B SNPs	Alleles	Gender	Freq.	11/12/22	Freq.	11/12/22	OR (95% CI)	p value	OR (95% CI)	p value	OR (95% CI)	p value
rs7073837	C/A	Male	0.36	35/25/14	0.51	18/38/20	1.54 (0.66–3.58)	0.32	2.40 (1.12–5.16)	0.03	1.61 (1.00–2.59)	0.05
rs7073837	C/A	Female	0.42	20/29/10	0.55	9/24/14	2.47 (0.88–6.94)	0.09	2.48 (0.92–6.75)	0.07	1.90 (1.02–3.55)	0.04
rs10740055	A/C	Male	0.45	29/37/20	0.54	18/40/24	1.49 (0.71–3.11)	0.29	2.02 (0.96–4.27)	0.07	1.49 (0.95–2.34)	0.08
rs10740055	A/C	Female	0.45	20/34/13	0.58	11/23/20	4.89 (1.74–13.7)	0.003	1.55 (0.62–3.88)	0.35	2.13 (1.17–3.88)	0.01
rs7089424	T/G	Male	0.34	34/30/10	0.49	22/35/20	2.44 (0.98–6.10)	0.056	1.98 (0.94–4.19)	0.07	1.72 (1.06–2.81)	0.03
rs7089424	T/G	Female	0.40	21/29/9	0.43	16/23/9	1.38 (0.45–4.17)	0.57	1.08 (0.46–2.56)	0.86	1.14 (0.63–2.06)	0.67
rs10821936	T/C	Male	0.32	43/31/12	0.46	27/35/20	2.10 (0.90–4.87)	0.09	2.09 (1.06–4.11)	0.03	1.70 (1.08–2.66)	0.02
rs10821936	T/C	Female	0.34	31/26/10	0.44	17/27/10	2.13 (0.70–6.43)	0.18	2.34 (0.99–5.54)	0.05	2.00(1.09–3.68)	0.03
rs4506592	G/A	Male	0.38	34/39/13	0.47	26/35/21	2.03 (0.89–4.62)	0.09	1.61 (0.81–3.22)	0.18	1.52 (0.97–2.39)	0.07
rs4506592	G/A	Female	0.38	27/29/11	0.45	16/27/11	1.37 (0.49–3.84)	0.55	1.51 (0.66–3.47)	0.33	1.32 (0.76–2.31)	0.33
rs10994982	G/A	Male	0.44	30/36/20	0.53	19/39/24	1.49 (0.71–3.11)	0.29	1.99 (0.95–4.17)	0.07	1.48 (0.95–2.31)	0.08
rs10994982	G/A	Female	0.44	21/33/13	0.57	12/22/20	4.92 (1.76–13.7)	0.002	1.42 (0.58–3.46)	0.45	2.08 (1.15–3.75)	0.02
rs7896246	G/A	Male	0.71	10/24/41	0.46	24/35/18	2.03 (0.81–5.10)	0.13	2.88 (1.37–6.06)	0.005	1.88 (1.15–3.05)	0.01
rs7896246	G/A	Female	0.35	25/27/7	0.42	16/24/8	2.27 (0.65–7.96)	0.20	1.47 (0.63–3.44)	0.37	1.39 (0.75–2.55)	0.29
rs10821938	C/A	Male	0.41	31/40/15	0.51	24/33/25	2.61 (1.20–5.67)	0.02	1.85 (0.90–3.81)	0.09	1.75 (1.12–2.75)	0.02
rs10821938	C/A	Female	0.40	26/28/13	0.48	15/26/13	1.25 (0.49–3.20)	0.64	1.20 (0.51–2.81)	0.67	1.16 (0.68–1.98)	0.59
rs7923074	C/A	Male	0.41	27/32/14	0.54	18/30/23	2.51 (1.07–5.90)	0.035	2.09 (0.94–4.66)	0.07	1.69 (1.04–2.75)	0.03
rs7923074	C/A	Female	0.40	24/23/12	0.52	11/23/13	1.64 (0.62–4.31)	0.32	2.11 (0.86–5.22)	0.11	1.58 (0.90–2.77)	0.11
rs6479778	C/T	Male	0.20	49/22/4	0.32	36/34/8	1.31 (0.34–5.03)	0.70	2.18 (1.06–4.49)	0.03	1.72 (0.97–3.06)	0.07
rs6479778	C/T	Female	0.27	34/20/6	0.20	31/16/2	0.33 (0.03–3.20)	0.34	0.74 (0.31–1.78)	0.50	0.71 (0.34–1.47)	0.35
rs4948487	A/C	Male	0.50	24/38/24	0.52	20/39/23	1.15 (0.56–2.37)	0.70	1.39 (0.66–2.93)	0.39	1.18 (0.76–1.84)	0.46
rs4948487	A/C	Female	0.46	18/37/12	0.53	13/25/16	2.26 (0.88–5.78)	0.09	1.17 (0.47–2.92)	0.73	1.87 (1.01–3.47)	0.047
rs6479779	G/C	Male	0.43	29/35/17	0.51	20/39/21	1.22 (0.55–2.68)	0.63	1.56 (0.74–3.31)	0.25	1.25 (0.79–1.99)	0.34
rs6479779	G/C	Female	0.42	20/37/9	0.43	18/22/11	5.18 (1.50–17.9)	0.009	1.19 (0.49–2.89)	0.70	1.75 (0.91–3.35)	0.09

Freq.: risk allele frequency; SNP: single nucleotide polymorphism; OR: odds ratio; CI: confidence interval.

Controlled for age. In the additive model, genotype of homozygote for the non-risk allele 11, heterozygote 12, and homozygote for the risk allele 22. The recessive model was defined as 22 versus 12 + 11 and dominant model as 22 + 12 versus 11. The outliers (studentized residual is greater than 2.0 or less than −2.0) were excluded. Bold values are significant.

Nine SNPs are located in one haplotype block with significant LD (Figure 1 and Table 4). The frequencies of haplotypes and diplotypes that were <0.03 in the total sample were excluded from the additional analysis. The haplotype (CGAACACAA) was significantly associated with ALL risk (OR = 4.29, p = 0.015). In contrast, the frequencies of the haplotypes (CCCGACTGC) and (CGCACATGA) were higher in the control group than in leukemic children (OR = 0.24, p = 0.048; OR = 0.13, p = 0.076, respectively). The diplotypes revealed no significant association with acute lymphoblastic leukemia (Table 5).

Figure 1.

Pairwise linkage disequilibrium among ARID5B SNPs in Yemeni children.

Table 4.

The estimated values of linkage disequilibrium analysis between the nine ARID5B SNPs in Yemeni children.

D′	rs6479778	rs6479779	rs7073837	rs10994982	rs10740055	rs7923074	rs10821936	rs7896246	rs10821938
rs6479778	–	0.999	0.652	0.721	0.713	0.553	0.612	0.579	0.644
rs6479779	0.396	–	0.381	0.387	0.391	0.295	0.333	0.379	0.263
rs7073837	0.168	0.143	–	0.885	0.892	0.763	0.979	0.989	0.741
rs10994982	0.176	0.126	0.685	–	0.993	0.857	0.980	0.875	0.968
rs10740055	0.167	0.125	0.679	0.959	–	0.873	0.999	0.896	0.984
rs7923074	0.113	0.064	0.564	0.662	0.671	–	0.947	0.946	0.939
rs10821936	0.197	0.086	0.713	0.625	0.633	0.640	–	0.916	0.983
rs7896246	0.179	0.104	0.715	0.488	0.500	0.628	0.825	–	0.899
rs10821938	0.166	0.069	0.528	0.789	0.791	0.814	0.747	0.612	r²

SNP: single nucleotide polymorphism.

The correlation coefficients D′ and r² between nine SNPs were shown in the above and below diagonal, respectively.

Table 5.

Association of common haplotypes and diplotypes with acute lymphoblastic leukemia among Yemeni children.

rs6479778, rs6479779, rs7073837, rs10994982, rs10740055, rs7923074, rs10821936, rs7896246, and rs10821938	Frequency		OR (95% CI)	p value
	Control (n = 120)	ALL (n = 111)	OR (95% CI)	p value
Haplotypes
CCAACACAA	13 (0.11)	15 (0.14)	1.38 (0.58–3.28)	0.47
CCCGACTGC	11(0.09)	5 (0.05)	0.24 (0.06–0.99)	0.048
CGAACACAA	7 (0.06)	16 (0.14)	4.29 (1.33–13.83)	0.015
CGCGACTGC	26 (0.22)	18 (0.16)	0.62 (0.29–1.34)	0.22
TCAACACAA	35 (0.29)	41(0.37)	1.71 (0.89–3.28)	0.11
CGCACATGA	9 (0.08)	2 (0.02)	0.13 (0.01–1.24)	0.076
Diplotypes
CCAACACAA–CGCGACTGC	6 (0.05)	8 (0.07)	1.42 (0.42–4.83)	0.58
CGCGACTGC–CGCGACTGC	20 (0.17)	13 (0.12)	0.66 (0.28–1.55)	0.34
TCAACACAA–CGCGACTGC	17(0.14)	17 (0.15)	0.96 (0.41–2.25)	0.93
CCCGACTGC–CGCGACTGC	8 (0.07)	2 (0.02)	0.13 (0.01–1.12)	0.063
CGAACACAA–CGCGACTGC	4 (0.03)	5 (0.05)	0.83 (0.18–3.85)	0.82
TCAACACAA–CCCGACTGC	4 (0.03)	6 (0.05)	1.54 (0.35–6.76)	0.57
TCAACACAA–CGAACACAA	6 (0.05)	5 (0.05)	1.09 (0.29–4.09)	0.90
TCAACACAA–TCAACACAA	3 (0.03)	6 (0.05)	2.33 (0.45–12.17)	0.32

OR: odds ratio; CI: confidence interval.

Controlled for age and gender. The outliers (studentized residual is greater than 2.0 or less than −2.0) were excluded. Bold values are significant.

Discussion

We evaluated the association between ARID5B SNPs and ALL in Yemeni children who are of Arab Asian descent. Several genome wide and candidate gene association studies have reported strong associations between ARID5B SNPs and ALL risk in different populations.^{9–12,15,17–20,27–30} Previously, GWA studies have been conducted in populations of European, Hispanic, and African American ancestry. Our findings in Yemeni population are in agreement with these studies, confirming ARID5B SNPs as general susceptibility markers for ALL. The main findings in our study are that rs7073837, rs10740055, rs7089424, rs10821936, rs4506592, rs10994982, rs7896246, rs10821938, and rs7923074 polymorphisms are associated with an increased risk of childhood ALL. The strongest associations were with rs10821936 and rs7896246 and these genetic effects seem to be additive. The effects of rs7073837, rs10740055, rs10994982, and rs4948487 were more pronounced in females than in males, while rs7089424, rs7896246, rs10821938, and rs7923074 were significantly associated with ALL in males. Noteworthy, rs10821936 was the only SNP found to be associated with ALL in both males and females.

B-cell progenitors and bone marrow cellularity are decreased in homozygous Arid5b knockout mice, confirming the role of ARID5B in B-cell lineage development.^31,32 Abnormal expression of the ARID5B gene during fetal life has been found to inhibit B-lymphocyte development and contribute to leukemogenesis.¹⁵ It has been suggested that ARID5B variations may affect the function of this gene during the maturation of B-progenitor cells, leading to an increased risk of B-ALL.^11,31 ARID5B SNPs rs7089424, rs7073837, rs10821936, and rs10821938 were found to have putative roles in the regulation of gene transcription, thereby affecting either ARID5B expression or splicing, likely producing different isoforms or preventing the binding of different transcription factors.^33,34

In comparison to other Asian populations, studies in Thailand and Taiwan showed no association between rs7089424 and ALL;^32,35 however, we found that this SNP was associated with ALL risk in Yemeni children. Different populations are likely to have different LD patterns and may not share the same susceptibility loci. The different levels of ALL risk are related to the genetic variations in different ethnicities.²⁰ Furthermore, these genetic variations may interact with different ancestral backgrounds, genetic make-up, and environmental factors in multiple ways to either increase or decrease the susceptibility to ALL in different areas. Conversely, we are in line with a Korean study that revealed an association between rs7089424 and ALL in Korean children.¹⁴

This study shows that rs10821936 and rs7896246 were the most significant signals detected for ARID5B SNPs and were in strong LD with each other (r² = 0.83), correlating with reports of rs10821936 being highly related to ALL susceptibility in European, French-Canadian, and Chinese populations.^11,15,30 In addition, Xu et al.¹⁹ reported that the strongest association between this SNP and ALL risk is in Whites and Hispanics. A recent meta-analysis study of 13 case–control studies involving 7147 patients and 31,969 controls reported that ARID5B rs10821936 SNP is associated with a high risk for ALL.³⁶ Functional studies are needed to explore how this SNP contributes to the increased risk of ALL.

Due to conflicting studies relating to possible gender-specific associations with ALL, a stratified analysis of ALL between males and females was also performed. Although some previous studies have reported that the association between ARID5B SNPs and ALL is more marked in females,^12,16 Healy et al.¹⁵ contradicted those studies by reporting a male bias in their gender-specific association analysis. However, there have been other reports showing no differences between males and females in the influence of ARID5B variations on the increased risk of ALL.^19,27,29 As a consequence of these diverging results, the effect of gender on ALL risk is as yet unknown.

We performed haplotype analysis and demonstrated that the haplotype (CGAACACAA), consisting of the risk alleles of eight SNPs, was associated with a fourfold increase in ALL risk. The associations of these individual ARID5B SNPs are not independent. However, the haplotype (CCCGACTGC) includes non-risk alleles which have revealed a protective effect.

Conclusion

We have conducted the first study on the association of ARID5B gene variants with ALL risk in Yemeni children. Some ARID5B SNPs are associated with increased susceptibility to ALL in females than in males and vice versa. In addition, the ARID5B haplotype containing SNP risk alleles supports the contribution of ARID5B SNPs to an increased risk of childhood ALL. Further studies with a larger sample size are warranted to confirm our results. Also, functional studies are needed to determine the causative variants of ARID5B gene and their effects on increased childhood ALL risk. Understanding how these SNP variations affect the overall structure and function of the ARID5B protein would be of important value in the diagnosis of childhood leukemia, improving risk-directed treatments and disease outcomes.

Footnotes

Acknowledgements

The authors are grateful to all children and their parents for their valuable participation. They also appreciate the kind cooperation of all medical staff and nursing in this study at Government hospitals in Sana’a city, Yemen. B.A., S.M., R.S.A., and S.M.N. contributed to study design; B.A. contributed to sample collection; B.A., R.H.A., and S.D.S. performed the lab work and statistical analyses; B.A., S.M., S.M.N., and M.F.M.R. contributed to reagents/materials/analysis tools; and B.A., S.M., S.M.N., R.S.A., and M.F.M.R. wrote the paper.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

This study was supported by grants from the University of Malaya (HIR/420001-E000046) and IPPP (PG039/2013A).

References

Howlader

Noone

Krapcho

. SEER cancer statistics review, 1975–2010. Bethesda, MD: National Cancer Institute, 2013.

Yeoh

Tan

C-K

. Management of adult and paediatric acute lymphoblastic leukaemia in Asia: resource-stratified guidelines from the Asian Oncology Summit 2013. Lancet Oncol 2013; 14: e508–e523.

World Health Organization. GLOBOCAN 2012: estimated cancer incidence, mortality and prevalence worldwide in 2012. Geneva: World Health Organization, 2012.

Ba-Saddik

. Childhood cancer in Aden, Yemen. Cancer Epidemiol 2013; 37: 803–806.

Greaves

. In utero origins of childhood leukaemia. Early Hum Dev 2005; 81: 123–129.

Greaves

Maley

. Clonal evolution in cancer. Nature 2012; 481: 306–313.

Mullighan

Goorha

Radtke

. Genome-wide analysis of genetic alterations in acute lymphoblastic leukaemia. Nature 2007; 446: 758–764.

Mullighan

. The molecular genetic makeup of acute lymphoblastic leukemia. Hematology Am Soc Hematol Educ Program 2012; 2012: 389–396.

Papaemmanuil

Hosking

Vijayakrishnan

. Loci on 7p12.2, 10q21.2 and 14q11.2 are associated with risk of childhood acute lymphoblastic leukemia. Nat Genet 2009; 41: 1006–1010.

10.

Prasad

Hosking

Vijayakrishnan

. Verification of the susceptibility loci on 7p12.2, 10q21.2, and 14q11.2 in precursor B-cell acute lymphoblastic leukemia of childhood. Blood 2010; 115: 1765–1767.

11.

Treviño

Yang

French

. Germline genomic variants associated with childhood acute lymphoblastic leukemia. Nat Genet 2009; 41: 1001–1005.

12.

Gutierrez-Camino

Lopez-Lopez

Martin-Guerrero

. Intron 3 of the ARID5B gene: a hot spot for acute lymphoblastic leukemia susceptibility. J Cancer Res Clin Oncol 2013; 139: 1879–1886.

13.

Ellinghaus

Stanulla

Richter

. Identification of germline susceptibility loci in ETV6-RUNX1-rearranged childhood acute lymphoblastic leukemia. Leukemia 2012; 26: 902–909.

14.

Han

Lee

Park

. Genome-wide association study of childhood acute lymphoblastic leukemia in Korea. Leuk Res 2010; 34: 1271–1274.

15.

Healy

Richer

Bourgey

. Replication analysis confirms the association of ARID5B with childhood B-cell acute lymphoblastic leukemia. Haematologica 2010; 95: 1608–1611.

16.

Orsi

Rudant

Bonaventure

. Genetic polymorphisms and childhood acute lymphoblastic leukemia: GWAS of the ESCALE study (SFCE). Leukemia 2012; 26: 2561–2564.

17.

Pastorczak

Gorniak

Sherborne

. Role of 657del5 NBN mutation and 7p12.2 (IKZF1), 9p21 (CDKN2A), 10q21.2 (ARID5B) and 14q11.2 (CEBPE) variation and risk of childhood ALL in the Polish population. Leuk Res 2011; 35: 1534–1536.

18.

Vijayakrishnan

Houlston

. Candidate gene association studies and risk of childhood acute lymphoblastic leukemia: a systematic review and meta-analysis. Haematologica 2010; 95: 1405–1414.

19.

Cheng

Devidas

. ARID5B genetic polymorphisms contribute to racial disparities in the incidence and treatment outcome of childhood acute lymphoblastic leukemia. J Clin Oncol 2012; 30: 751–757.

20.

Yang

Trevino

Yang

. ARID5B SNP rs10821936 is associated with risk of childhood acute lymphoblastic leukemia in blacks and contributes to racial differences in leukemia incidence. Leukemia 2010; 24: 894–896.

21.

Lin

Song

. Recent advances in the ARID family: focusing on roles in human cancer. Onco Targets Ther 2014; 7: 315–324.

22.

Patsialou

Wilsker

Moran

. DNA-binding properties of ARID family proteins. Nucleic Acids Res 2005; 33: 66–80.

23.

Linet

Brown

Mbulaiteye

. International long-term trends and recent patterns in the incidence of leukemias and lymphomas among children and adolescents ages 0–19 years. Int J Cancer 2016; 138: 1862–1874.

24.

Absher

Tang

. Worldwide human relationships inferred from genome-wide patterns of variation. Science 2008; 319: 1100–1104.

25.

Bolufer

Barragan

Collado

. Influence of genetic polymorphisms on the risk of developing leukemia and on disease progression. Leuk Res 2006; 30: 1471–1491.

26.

Chokkalingam

Hsu

L-I

Metayer

. Genetic variants in ARID5B and CEBPE are childhood ALL susceptibility loci in Hispanics. Cancer Causes Control 2013; 24: 1789–1795.

27.

Lautner-Csorba

Gezsi

Semsei

. Candidate gene association study in pediatric acute lymphoblastic leukemia evaluated by Bayesian network based Bayesian multilevel analysis of relevance. BMC Med Genomics 2012; 5: 42.

28.

Hsu

Chokkalingam

Briggs

. Association of genetic variation in IKZF1, ARID5B, and CEBPE and surrogates for early-life infections with the risk of acute lymphoblastic leukemia in Hispanic children. Cancer Causes Control 2015; 26: 609–619.

29.

Linabery

Blommer

Spector

. ARID5B and IKZF1 variants, selected demographic factors, and childhood acute lymphoblastic leukemia: a report from the Children’s Oncology Group. Leuk Res 2013; 37: 936–942.

30.

Wang

Chen

. Association of three polymorphisms in ARID5B, IKZF1and CEBPE with the risk of childhood acute lymphoblastic leukemia in a Chinese population. Gene 2013; 524: 203–207.

31.

Paulsson

Forestier

Lilljebjörn

. Genetic landscape of high hyperdiploid childhood acute lymphoblastic leukemia. Proc Natl Acad Sci USA 2010; 107: 21719–21724.

32.

Lin

Chang

. High-resolution melting analyses for genetic variants in ARID5B and IKZF1 with childhood acute lymphoblastic leukemia susceptibility loci in Taiwan. Blood Cells Mol Dis 2014; 52: 140–145.

33.

Lee

Shatkay

. F-SNP: computationally predicted functional SNPs for disease association studies. Nucleic Acids Res 2008; 36: D820–D824.

34.

Samuelsson

Alonso

Ruiz-Larroya

. Frequent somatic demethylation of RAPGEF1/C3G intronic sequences in gastrointestinal and gynecological cancer. Int J Oncol 2011; 38: 1575–1577.

35.

Vijayakrishnan

Sherborne

Sawangpanich

. Variation at 7p12.2 and 10q21.2 influences childhood acute lymphoblastic leukemia risk in the Thai population and may contribute to racial differences in leukemia incidence. Leuk Lymphoma 2010; 51: 1870–1874.

36.

Guo

L-M

J-S

. ARID5B gene rs10821936 polymorphism is associated with childhood acute lymphoblastic leukemia: a meta-analysis based on 39,116 subjects. Tumour Biol 2014; 35: 709–713.