Association of miR-34b/c rs4938723 and TP53 Arg72Pro Polymorphisms with Neuroblastoma Susceptibility: Evidence from Seven Centers

Abstract

Neuroblastoma is a pediatric malignancy arising from the developing peripheral nervous system. p53 and downstream effector miR-34b/c have critical tumor suppressing functions. TP53 Arg72Pro (rs1042522 C > G) and miR-34b/c rs4938723 (T > C) polymorphisms have been known to modify cancer susceptibility. This study was performed to validate the association of these two polymorphisms and neuroblastoma risk with 819 cases and 1780 controls. Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) were used to assess the strength of the associations. False positive report possibility analysis was adopted to dissect out real significant associations from chance findings. We found that both TP53 Arg72Pro (CG/GG vs. CC: adjusted OR = 0.82, 95% CI = 0.69-0.98) and miR-34b/c rs4938723 (TC/CC vs. TT: adjusted OR = 0.64, 95% CI = 0.54-0.75) were associated with decreased neuroblastoma susceptibility. Stratify analyses further confirmed the protective effect among some subgroups. Moreover, subjects with variant alleles of both polymorphisms were associated with more significantly decreased neuroblastoma risk (CG/TC vs. CC/TT: adjusted OR = 0.38, 95% CI = 0.28-0.50; GG/TC vs. CC/TT: adjusted OR = 0.43, 95% CI = 0.30-0.63) than those carrying variant allele of either one polymorphism (CC/TC vs. CC/TT: adjusted OR = 0.51, 95% CI = 0.37-0.69; CG/TT vs. CC/TT: adjusted OR = 0.71, 95% CI = 0.55-0.92), suggesting cumulative effects of the polymorphisms. False positive report possibility analysis further verified that our findings are noteworthy. Overall, we confirmed that miR-34b/c rs4938723 and TP53 Arg72Pro conferred decreased neuroblastoma risk and two polymorphisms exerted stronger protective effects against neuroblastoma than either one alone.

Introduction

Neuroblastoma is an extracranial neuroendocrine tumor, affecting approximately 25 to 50 individuals per million [1]. Tumor may occur in the adrenal glands and/or sympathetic ganglia. The majority of tumors (90%) are diagnosed in children younger than 10 years old, and the median age at diagnosis is about 18 months old. Neuroblastoma is a group of heterogeneous diseases. Its clinical presentation and prognosis vary greatly dependent on the tumor biology, including molecular genetics. Neuroblastoma is also a complex genetic disease [2], [3], [4]. Apart from driver gene mutations [2], genome-wide association studies (GWASs) have identified a number of neuroblastoma susceptibility loci in the CASC15, BARD1, DUSP12, DDX4, IL31RA, HSD17B12, LMO1, HACE1, LIN28B, MLF1, and CPZ genes [5], [6], [7], [8], [9]. As a complementary to agnostic approach, traditional candidate gene method is also frequently used to investigate genetic variation in protein coding sequences. Recently, studies by candidate gene approaches have found some genetic variations associated with neuroblastoma risk in NEFL, CDKN1B and BARD1 genes [10], [11], [12], [13].

Tumor suppressing protein p53 is a transcription factor. It suppresses tumorigenesis by orchestrating the transcriptional activation of multiple target genes to fight against DNA damage, cellular stress, and excessive mitogenic stimulation [14]. Numerous genes are involved in p53 tumor suppressor network, including p21, cyclin G, MDM-2, GADD-45, PTEN, and TSC-2. Moreover, p53 network was further complicated by the finding that p53 can execute tumor suppressing function by transcriptionally regulating microRNAs (miRNAs), especially miR-34 family [15], [16]. miR-34 family consists of three mature miRNAs, miR-34a, miR-34b, and miR-34c. miR-34 family encoding genes are direct targets of p53 in the transcriptional level, with evolutionarily conserved p53 binding sites located upstream the miRNA encoding sequence [15], [16].

Genetic variations may change the expression levels and structures of tumor repressor genes and alter tumor repressing function. A genetic polymorphism Arg72Pro at codon 72 (rs1042522 C > G) of p53 protein can affect protein function biochemically and biologically [17], [18], [19]. Moreover, a functional common single nucleotide polymorphism (SNP) rs4938723 T > C was identified in the promoter region of the pri-miR-34b/c encoding gene [20]. This SNP is located in a typical CpG island, 423-bp upstream from the transcription start site. These two SNPs have been broadly investigated for their association with cancer susceptibility.

Due to the close relationship between p53 and miR-34b/c, a number of studies were launched to investigate the risk effects of TP53 rs1042522 C > G and miR-34b/c rs4938723 T > C polymorphisms jointly on the different types of cancer, including primary hepatocellular carcinoma, intracranial aneurysm, nasopharyngeal carcinoma, colorectal cancer, cervical cancer, and papillary thyroid carcinoma [21], [22], [23], [24], [25], [26]. However, whether they jointly confer susceptibility to neuroblastoma needed to be explored in a large well-designed case control study. Therefore, we investigated the association of these two SNPs with neuroblastoma risk in Chinese children by performing this case control study with 819 cases and 1780 controls.

Materials and Methods

Study Population

Only patients with newly diagnosed and histopathologically confirmed neuroblastoma were qualified to be recruited for this study. Healthy controls were frequency-matched to cases on the basis of age and gender. Totally, 819 controls and 1780 cases were separately recruited from seven hospitals in China, including Hunan Children’s Hospital (162 cases and 270 controls), Guangzhou Women and Children’s Medical Center (275 cases and 531 controls) [27], [28], The Second Affiliated Hospital of Xi’an Jiaotong University (76 cases and 186 controls) [29], The Second Affiliated Hospital and Yuying Children’s Hospital of Wenzhou Medical University (36 cases and 72controls) [30], The First Affiliated Hospital of Zhengzhou University (118 cases and 281 controls) [31], [32], Children’s Hospital of Shanxi (33 cases and 176 controls), and Anhui Provincial Children’s Hospital (119 cases and 264 controls) [33]. Written informed consent was obtained from all participants or their guardians. The institutional review boards of Guangzhou Women and Children’s Medical Center, the First Affiliated Hospital of Zhengzhou University, Anhui Provincial Children’s Hospital, Hunan Children’s Hospital, the Second Affiliated Hospital of Xi’an Jiaotong University, Children Hospital and Women Health Center of Shanxi, The Second Affiliated Hospital and Yuying Children’s Hospital of Wenzhou Medical University authorized this study.

SNP Selection and Genotyping

Two potentially functional SNPs were chosen for this study based on previous publications [23], [28], [34]. The TP53 Arg72Pro (rs1042522 C > G) is a common nonsynonymous SNP, generating two biochemically and biologically polymorphic variants, p53 Arg and p53 Pro [19]. The rs4938723 T > C polymorphism is located in the promoter region of pri-miR-34b/c, which was initially reported to be associated with an elevated risk of developing primary hepatocellular carcinoma [23]. Genomic DNA was isolated from venous blood samples donated by participants, use the TIANamp Genomic DNA blood kit (Tiangen Biotech, Beijing, China). Allelic discrimination TaqMan assay was employed to genotype SNPs in 384-wellplates with strict quality control [35], [36], [37], [38], [39]. Assay was run in the ABI 7900 HT Sequence Detection System (Applied Biosystems, Foster City, CA). Individuals involved in genotyping remain blind to status of blood donor.

Statistical Analysis

Frequency distributions of demographic variables and genotype were compared between cases and controls using χ² test. Hardy–Weinberg equilibrium (HWE) was checked for the frequency distribution of target SNPs among control subjects, using the goodness-of-chi-squared test. Unconditional logistic regression was used to generate odds ratios (ORs) and 95% confidence intervals (CIs) in order to estimate the association of studied polymorphisms with neuroblastoma risk. OR and 95% CI were estimated under different genetic models, I) homozygous (VV vs. WW), II) heterozygous (WV vs. WW), III) dominant (WV/VV vs. WW), IV) recessive (VV vs. WW/WV), and allele contrast (V vs. W), with W and V indicating wild-type and variant allele of a SNP, respectively. Multivariate analysis was conducted using unconditional logistic regression, with adjustment for age and gender. To dissect out real significant associations from chance findings, we performed the false positive report possibility (FPRP) analysis for the significant findings. As indicated by previous publication [40], we used a prior probability of 0.1 to interrogate OR of 1.50/0.67 (risk/protective association) with the significance level of FPRP predetermined as 0.2. The association with a FPRP value of <0.2 was considered noteworthy. All statistical analyses were two-sided and carried out using SAS software (Version 9.1; SAS Institute, Cary, NC, USA). A significance level of P < .05 was applied without extra specification.

Results

Association of miR-34b/c rs4938723 and TP53 Arg72Pro Polymorphisms with Neuroblastoma Susceptibility

No significant difference was detected between cases and controls for age (P = .395) and gender (P = .832) for combined subjects (Supplemental Table 1). Both of the two studied SNPs were shown to exert protective effects against neuroblastoma (Table 1). TP53 rs1042522 C > G polymorphism was associated with decreased neuroblastoma susceptibility [CG vs. CC adjusted OR (AOR) = 0.80, 95% CI = 0.67-0.97; CG/GG vs.CC: AOR = 0.82, 95% CI = 0.69-0.98]. miR-34b/c rs4938723 T > C polymorphism also conferred reduced neuroblastoma susceptibility (TC vs. TT, AOR = 0.54, 95% CI = 0.45-0.65; additive model: AOR = 0.84, 95% CI = 0.74-0.95; TC/CC vs. TT: AOR = 0.64, 95% CI = 0.54-0.75; CC vs. TC/TT: AOR = 1.40, 95% CI = 1.08-1.81; C vs. T: AOR = 0.83, 95% CI = 0.73-0.95).

Table 1.

Associations Between TP53 and miR-34b/c Polymorphisms and Neuroblastoma Susceptibility

Genotype	Cases (N = 819)	Controls (N = 1780)	P†	Crude OR (95% CI)	P	Adjusted OR (95% CI)‡	P‡
TP53 rs1042522 C > G (HWE = 0.541)§
CC	285 (34.80)	544 (30.58)		1.00		1.00
CG	375 (45.79)	891 (50.08)		0.80 (0.67–0.97)	.022	0.80 (0.67–0.97)	.022
GG	159 (19.41)	344 (19.34)		0.88 (0.70–1.12)	.299	0.88 (0.69–1.11)	.285
Additive			.072	0.92 (0.82–1.04)	.164	0.92 (0.82–1.03)	.156
Dominant	534 (65.20)	1235 (69.42)	.032	0.83 (0.69–0.98)	.032	0.82 (0.69–0.98)	.031
Recessive	660 (80.59)	1435 (80.66)	.963	1.01 (0.82–1.24)	.963	1.00 (0.81–1.24)	.988
C	945 (57.69)	1979 (55.62)		1.00		1.00
G	693 (42.31)	1579 (44.38)	.162	0.92 (0.82–1.03)	.162	0.92 (0.82–1.03)	.153
miR-34b/c rs4938723 T > C (HWE = 0.276)§
TT	455 (56.66)	808 (45.44)		1.00		1.00
TC	242 (30.14)	796 (44.77)		0.54 (0.45–0.65)	<.0001	0.54 (0.45–0.65)	<.0001
CC	106 (13.20)	174 (9.79)		1.08 (0.83–1.41)	.564	1.08 (0.83–1.41)	.578
Additive			<.0001	0.84 (0.74–0.95)	.006	0.84 (0.74–0.95)	.006
Dominant	348 (43.34)	970 (54.56)	<.0001	0.64 (0.54–0.75)	<.0001	0.64 (0.54–0.75)	<.0001
Recessive	697 (86.80)	1604 (90.21)	.010	1.40 (1.08–1.81)	.010	1.40 (1.08–1.81)	.011
T	1152 (71.73)	2412 (67.83)		1.00		1.00
C	454 (28.27)	1144 (32.17)	.005	0.83 (0.73–0.95)	.005	0.83 (0.73–0.95)	.005

OR, odds ratio; CI, confidence interval.

^†χ² test for genotype distributions between neuroblastoma cases and cancer-free controls.

^‡Adjusted for age and gender.

^§There were missing values for genotyping that failed.

Stratification Analysis

We further performed stratification analysis to dissect the effects of confounding factors on the strength of the association, including age, gender, sites of origin and clinical stages (Table 2). Regarding the protective effect of TP53 rs1042522 CG/GG genotypes, significant association resided in children old than 18 months (AOR = 0.78, 95% CI = 0.62-0.97), males (AOR = 0.79, 95% CI = 0.62-0.99), and those with tumors in mediastinum (AOR = 0.68, 95% CI = 0.51-0.92). There was no modification of this result by clinical stages. In contrast, the protective effect of the miR-34b/c rs4938723 TC/CC genotypes remained significant among all subgroups, except for strata with tumor in mediastinum and “others”.

Table 2.

Stratification Analysis of TP53 and miR-34b/c Polymorphisms with Neuroblastoma Susceptibility

Variables	rs1042522 (cases/controls)		AOR (95% CI)†	P†	rs4938723 (cases/controls)		AOR (95% CI)†	P†
	CC	CG/GG			TT	TC/CC
Age, month
≤18	110/232	216/508	0.89 (0.67–1.17)	.393	181/334	139/406	0.63 (0.48–0.82)	.0006
>18	175/312	318/727	0.78 (0.62–0.97)	.029	274/474	209/564	0.64 (0.52–0.80)	<.0001
Gender
Females	122/241	235/526	0.87 (0.67–1.14)	.304	210/342	142/425	0.54 (0.42–0.70)	<.0001
Males	163/303	299/709	0.79 (0.62–0.99)	.043	245/466	206/545	0.72 (0.58–0.90)	.004
Sites of origin
Adrenal gland	90/544	168/1235	0.82 (0.62–1.08)	.161	160/808	97/970	0.51 (0.39–0.66)	<.0001
Retroperitoneal	93/544	188/1235	0.88 (0.68–1.16)	.369	154/808	117/970	0.63 (0.49–0.82)	.0005
Mediastinum	80/544	123/1235	0.68 (0.51–0.92)	.012	98/808	102/970	0.87 (0.65–1.16)	.341
Others	20/544	49/1235	1.09 (0.64–1.85)	.759	39/808	29/970	0.62 (0.38–1.01)	.056
Clinical stages
I + II+ 4 s	154/544	288/1235	0.83 (0.66–1.03)	.089	238/808	200/970	0.70 (0.57–0.86)	.0009
III + IV	121/544	231/1235	0.84 (0.66–1.07)	.147	204/808	139/970	0.57 (0.45–0.72)	<.0001

AOR, adjusted odds ratio; CI, confidence interval.

^†Adjusted for age and gender, omitting the corresponding stratify factor.

Combined Effect Analysis

To explore the combined effect of SNPs in miR-34b/c and TP53 gene, we tested the association between inferred genotype combinations and neuroblastoma susceptibility (Table 3). The following genotype combinations were shown to decrease susceptibility to neuroblastoma when compared to combination of wide type genotype (CC/TC vs. CC/TT: AOR = 0.51, 95% CI = 0.37-0.69; CG/TT vs. CC/TT: AOR = 0.71, 95% CI = 0.55-0.92; CG/TC vs. CC/TT: AOR = 0.38, 95% CI = 0.28-0.50; GG/TC vs. CC/TT: AOR = 0.43, 95% CI = 0.30-0.63). However, the carriers of combination of variant genotype (GG/CC) did not show significantly decreased risk probably because of small sample size (AOR = 0.91, 95% CI = 0.52-1.60). Furthermore, we found that subjects with variant alleles of both polymorphisms have smaller ORs (0.38 for CG/TC; 0.43 for GG/TC) than those carrying variant allele of either one polymorphism (0.51 for CC/TC; 0.71 for CG/TT). It suggests that combined protective effects conferred by two SNPs are stronger than either one alone and the former are less likely to develop neuroblastoma than the latter.

Table 3.

Inferred Genotypes of miR-34b/c and TP53 Gene and Their Association with the Neuroblastoma Susceptibility

Genotypes		Cases	Controls	OR (95% CI)	P	AOR (95% CI)‡	P‡
rs1042522	rs4938723	(n = 803)†	(n = 1777)†
CC	TT	163 (20.30)	246 (13.84)	1.00		1.00
CC	TC	85 (10.59)	236 (13.28)	0.50 (0.37–0.69)	<.0001	0.51 (0.37–0.69)	<.0001
CC	CC	33 (4.11)	61 (3.43)	0.76 (0.48–1.20)	.237	0.76 (0.48–1.21)	.246
CG	TT	206 (25.65)	407 (22.90)	0.71 (0.55–0.91)	.008	0.71 (0.55–0.92)	.008
CG	TC	109 (13.57)	405 (22.79)	0.38 (0.28–0.50)	<.0001	0.38 (0.28–0.50)	<.0001
CG	CC	50 (6.23)	78 (4.39)	0.90 (0.60–1.34)	.591	0.89 (0.60–1.34)	.588
GG	TT	86 (10.71)	154 (8.67)	0.78 (0.56–1.08)	.135	0.78 (0.56–1.08)	.139
GG	TC	48 (5.98)	155 (8.72)	0.43 (0.30–0.63)	<.0001	0.43 (0.30–0.63)	<.0001
GG	CC	23 (2.86)	35 (1.97)	0.92 (0.52–1.61)	.764	0.91 (0.52–1.60)	.750

OR, odds ratio; AOR, adjusted odds ratio; CI, confidence interval.

^†There were missing value for genotyping failed.

^‡Obtained in logistic regression models with adjustment for age and gender.

False Positive Report Possibility Analysis

The results of association studies are often questioned by false positivity. To address this issue, the FPRP analysis was performed to test the credibility of our significant findings (Table 4). FPRP analysis determines whether a statistically significant finding is noteworthy by collectively considering statistical power of the study, the calculated P value, and the prior probability of reality of the association, which is more objective than statistical significance based on a P < .05 alone [40]. With a prior probability of 0.25, all our significant findings are deserving of attention. While probability was lowered to 0.1, all tested results remained noteworthy, except for the association under the dominant model and the association for older children and males for rs1042522 C > G polymorphism. When the standard of prior probability become more strict (0.001), results for rs1042522 C > G polymorphism became not deserving of attention, but most of results for rs4938723 T > C maintained to be noteworthy, even with much smaller prior possibilities. It suggests that latter is higher penetrant SNP than the former. Overall, FPRP analysis confirmed the credibility of our results.

Table 4.

False-Positive Report Probability Analysis for Significant Findings

Genotype	OR (95% CI)	P†	Statistical power‡	Prior probability
				0.25	0.1	0.01	0.001	0.0001
rs1042522 C > G
CG vs. CC	0.80 (0.67–0.97)	.022	0.992	0.062	0.166	0.687	0.957	0.996
CG/GG vs. CC	0.83 (0.69–0.98)	.032	0.991	0.089	0.226	0.763	0.970	0.997
>18	0.78 (0.62–0.98)	.032	0.910	0.095	0.240	0.777	0.972	0.997
Males	0.78 (0.62–0.99)	.041	0.911	0.119	0.288	0.817	0.978	0.998
Mediastinum	0.68 (0.50–0.91)	.011	0.532	0.056	0.152	0.663	0.952	0.995
rs4938723 T > C
TC vs. TT	0.54 (0.45–0.65)	<.0001	0.021	0.000	0.000	0.000	0.000	0.000
TC/CC vs. TT	0.64 (0.54–0.75)	<.0001	0.291	0.000	0.000	0.000	0.001	0.005
CC vs. TT/CT	1.40 (1.08–1.81)	.010	0.708	0.041	0.114	0.585	0.934	0.993
C vs. T	0.83 (0.73–0.95)	.005	0.999	0.015	0.043	0.331	0.833	0.980
TC/CC vs. TT
≤18	0.63 (0.49–0.82)	.0007	0.342	0.006	0.018	0.168	0.671	0.953
>18	0.64 (0.52–0.80)	<.0001	0.352	0.001	0.002	0.017	0.151	0.641
Females	0.54 (0.42–0.70)	<.0001	0.060	0.000	0.000	0.005	0.052	0.353
Males	0.72 (0.58–0.90)	.004	0.734	0.015	0.043	0.333	0.834	0.981
Adrenal gland	0.51 (0.39–0.66)	<.0001	0.024	0.000	0.000	0.003	0.027	0.217
Retroperitoneal	0.63 (0.49–0.82)	.0005	0.339	0.004	0.013	0.127	0.596	0.937
I + II+ 4 s	0.70 (0.57–0.86)	.0009	0.666	0.004	0.012	0.118	0.575	0.931
III + IV	0.57 (0.45–0.72)	<.0001	0.089	0.000	0.000	0.002	0.024	0.201
Genotypes§
CC/TC vs. CC/TT	0.50 (0.37–0.69)	<.0001	0.065	0.001	0.002	0.025	0.208	0.725
CG/TT vs. CC/TT	0.71 (0.55–0.91)	.008	0.847	0.026	0.075	0.471	0.900	0.989
CG/TC vs. CC/TT	0.38 (0.28–0.50)	<.0001	0.001	0.000	0.000	0.000	0.000	0.000
GG/TC vs. CC/TT	0.43 (0.30–0.63)	<.0001	0.016	0.002	0.007	0.071	0.434	0.885

OR, odds ratio; CI, confidence interval.

^aχ² Test was used to calculate the genotype frequency distributions.

^‡Statistical power was calculated using the number of observations in each subgroup and the corresponding ORs and P values in this table.

^§The genotypes were constructed in the order of rs1042522 and rs4938723.

Discussion

The importance of p53 in tumor suppression can be partially reflected by the fact genetic alterations (e.g., mutations) in the p53 signaling pathway are implicated in nearly all types of human cancers. In response to DNA damage, cellular stress, and excessive mitogenic stimulation, p53 is activated to trigger apoptosis, cellular senescence or cell cycle arrest to maintain homeostasis [16], [41]. As a component of p53 tumor suppression network, human miR-34a and miR-34b/c genes are mapped to Chr.1p36 and Chr.11q23 [42]. Mechanistic study revealed that miR-34b/miR-34c can inhibit proliferation of ovarian cancer cell [43]. miR-34 also acts as a tumor suppressor in neuroblastoma by targeting MYCN [44] and CD44 [45], suggesting the implication of miR-34 family in nueroblastoma. mRNA expression profiling analysis revealed that miR-34 suppressed cell cycle genes of neuroblastoma IMR32 cells [44]. miR-34 also induced apoptosis of neuroblastoma cells and inhibited DNA synthesis [44]. miR-34b/c is processed from a common primary transcript (pri-miR-34b/c). In response to stimuli (e.g., DNA damage), p53 promotes the expression of miR-34 by transcriptionally activating the miRNA-encoding gene; miR-34 in turn induced cell cycle arrest by facilitating the degradation of transcripts of target genes including CCNE2, CDK4 and the MET [16]. Therefore, miR-34 is important downstream effectors of p53 signaling cascades [16], [41]. Epigenetic inactivation of miR-34 gene by CpG methylation has been observed in several types of cancer [42].

We have previously explore the association between TP53 gene rs1042522 C > G polymorphism and neuroblastoma susceptibility in Chinese children, with 256 patients and 531 controls [34]. Because the sample size was relatively small, the association only reached borderline significant (CG vs. CC: OR = 0.72, 95% CI = 0.51-1.02, P = .065) [34]. Diskin et al. reported that the association of TP53 gene rs35850753 and rs78378222 polymorphisms and susceptibility to neuroblastoma [45]. However, Cattelani et al. found lack of association between minor allele of TP53 rs1042522 and neuroblastoma risk in an Italy population with 288 healthy subjects and 286 neuroblastoma patients. Alternatively, the same study revealed significant association between minor allele of rs1042522 and poor neuroblastoma prognosis [46], validating the role of this SNP in the neuroblastoma. It is not uncommon to generate conflicting results for observational association case-control studies. Association results could be also affected by sample size, sampling strategy, genotyping method, geographic region, and ethnicity.

miR-34b/c rs4938723 has been reported to be associated with the risk of a wide spectrum of cancer, including esophageal squamous cell carcinoma, childhood acute lymphoblastic leukemia, hepatocellular carcinoma, gastric cancer, and prostate cancer [26], [47], [48], [49], [50], [51], [52], [53]. With a study population of 393 cases and 812 controls, we for the first time reported a protective association between the miR-34b/c rs4938723 and neuroblastoma risk [28].

In this study, we aimed to validate our findings above and evaluate combined effects of these two SNPs on neuroblastoma risk in a larger study. In the current study, the triple sample size of 819 cases and 1780 controls allowed us to detect significant association between TP53 rs1042522 C > G polymorphism and neuroblastoma susceptibility under the heterogeneous and dominant model. Moreover, the association of the miR-34b/c rs4938723 with neuroblastoma risk was validated in this study. These two SNPs may exert protective effects cumulatively. We found that subjects with variant alleles of both polymorphisms are less likely to develop neuroblastoma than those carrying variant allele of either one polymorphism. FPRP analysis indicated that most of our significant findings are noteworthy with a prior probability of 0.1.

There is evidence indicating that the TP53 Arg72Pro polymorphism may affect the function of p53 [17], [18], [19], [54]. For instance, this nonsynonymous common SNP not only changed the primary structure of the protein, but also led to differential migration rate during sodium dodecyl sulfate polyacrylamide gel electrophoresis [17]. A study showed that the exogenous p53 Arg was significantly more vulnerable than p53 Pro to the ubiquitin-mediated degradation in p53-null Saos-2 cells when exposed to human papillomavirus (HPV) E6 protein [18]. Moreover, p53 Arg and p53 Pro differed in term of their abilities to transcriptionally activate target genes, to induce apoptosis, and to suppress the transformation of primary murine fibroblasts [19]. However, the underlying molecular mechanisms for its association with reduced neuroblastoma susceptibility need to be clarified.

Recently, two meta-analyses revealed that the roles of the miR-34b/c rs4938723 in cancer susceptibility are tissue dependent [20], [55]. The rs4938723 polymorphism was shown to significantly increase the risk of hepatocellular carcinoma but decreased the risk of developing esophageal squamous cell carcinoma, colorectal cancer, and acute lymphoblastic leukemia. Several possibilities may help to explain such conflicting situation. This T to C transition polymorphism is positioned in the promoter region of pri-miR-34b/c, within a typical CpG island specifically. According to bioinformatics analysis, this SNP may affect predicted GATA-X transcription factors’ binding to the promoter of pri-miR-34b/c gene so as to alter its expression levels. Given transcription factors regulate gene expression in a tissue-specific way, this SNP may affect different transcription factors’ binding to the promoter, thereby either upregulating or downregulating transcription in different tissues. Moreover, the same microRNA may target different genes in the different tissues, and thereby modify cancer susceptibility in the tissue-specific manner.

This study also has limitations to be addressed. First, only two functional SNPs in the p53 tumor suppression network were investigated. Second, selection bias might be inevitable in this hospital-based case and control study. Third, although this was the largest association study for neuroblastoma susceptibility in Chinese children, the sample size was still moderate, especially for stratification analysis and inferred genotype analysis. Finally, our findings should be interpreted with caution since only Chinese Han population was recruited.

In conclusion, we validated the association of TP53 Arg72Pro and miR-34b/c rs4938723 polymorphisms with neuroblastoma susceptibility in Chinese children with a multi-center case-control study. These two SNPs may confer decreased neuroblastoma susceptibility cumulatively.

The following are the supplementary data related to this article.

Supplemental Table 1

Frequency distribution of selected characteristics in neuroblastoma cases and cancer-free controls

Conflict of Interest

None.