TP53 and p21 (CDKN1A) polymorphisms and the risk of systemic lupus erythematosus

The p53 and p21 proteins are important regulators of cell cycle and apoptosis and may contribute to autoimmune diseases, such as systemic lupus erythematosus (SLE). As genetic polymorphisms may cause changes in protein levels and functions, we investigated associations of TP53 and p21 (CDKN1A) polymorphisms (p53 72 G > C—rs1042522; p53 PIN3—rs17878362; p21 31 C > A—rs1801270; p21 70 C > T—rs1059234) with the development of systemic lupus erythematosus (SLE) in a Southeastern Brazilian population. Genotyping of 353 female volunteers (cases, n = 145; controls, n = 208) was performed by polymerase chain reaction, restriction fragment length polymorphism and/or DNA sequencing. Associations between TP53 and p21 polymorphisms and SLE susceptibility and clinical manifestations of SLE patients were assessed by logistic regression analysis. Protective effect was observed for the genotype combinations p53 PIN3 A1/A1-p21 31 C/A, in the total study population (OR 0.45), and p53 PIN3 A1/A2-p21 31 C/C, in non-white women (OR 0.28). In Whites, p53 72 C-containing (OR 3.06) and p53 PIN3 A2-containing (OR 6.93) genotypes were associated with SLE risk, and higher OR value was observed for the combined genotype p53 72 G/C-p53 PIN3 A1/A2 (OR 9.00). Further, p53 PIN3 A1/A2 genotype was associated with serositis (OR 2.82), while p53 PIN3 A2/A2 and p53 72 C/C genotypes were associated with neurological disorders (OR 4.69 and OR 3.34, respectively). Our findings showed that the TP53 and p21 polymorphisms included in this study may have potential to emerge as SLE susceptibility markers for specific groups of patients. Significant interactions of the TP53 polymorphisms with serositis and neurological disorders were also observed in SLE patients. The polymorphisms TP53 rs1042522 (G > C) and TP53 rs17878362 (16 bp Del/Ins) were associated with SLE risk in whites. In whites, the combined genotype TP53 rs1042522 GC- TP53 rs17878362 A1A2 and the haplotype TP53 rs1042522 C-rs17878362 A2 represented higher SLE risk. Combination of TP53 rs17878362 (16 bp Del/Ins) and p21 rs1801270 (C > A) protected against SLE in non-white women. TP53 and p21 (CDKN1A) polymorphisms may be SLE susceptibility markers for specific groups. The polymorphisms TP53 rs1042522 (G > C) and TP53 rs17878362 (16 bp Del/Ins) were associated with SLE risk in whites. In whites, the combined genotype TP53 rs1042522 GC- TP53 rs17878362 A1A2 and the haplotype TP53 rs1042522 C-rs17878362 A2 represented higher SLE risk. Combination of TP53 rs17878362 (16 bp Del/Ins) and p21 rs1801270 (C > A) protected against SLE in non-white women. TP53 and p21 (CDKN1A) polymorphisms may be SLE susceptibility markers for specific groups.


Introduction
Systemic lupus erythematosus (SLE) is an autoimmune disease characterized by a breakdown in self-tolerance, dysregulation of lymphocytes number and subsets in association with an increased autoantibody production [1,2].Although the interaction of genetic, immunological and environmental factors is widely accepted, this complex pathophysiological process is not fully understood [3].There is a central role for an excess of autoantigens derived mainly from altered programmed cell-death mechanisms like NETosis, pyroptosis and apoptosis [4] in association with a deficient clearance of extracellular vesicles associated autoantigens [5,6].
It has already been suggested that apoptosis plays an important role in the elimination of autoreactive lymphocytes in SLE patients.Cell cycle control, in its turn, is responsible for maintaining lymphocyte homeostasis as it participates in lymphocyte differentiation, effector function, memory development, tolerance induction, and apoptosis [5].Therefore, defects in the regulation of apoptotic process and cell cycle may cause loss of tolerance, generating autoimmune responses [1].
One of the major regulators of cell cycle and apoptosis is the p53 tumor suppressor protein.The protein p53 acts as a transcription factor and regulates the expression of genes involved in several cellular processes, including cell cycle arrest, DNA repair and activation of apoptosis [7].In SLE patients, it has been observed high levels of p53 expression and anti-p53 antibodies, with a significant correlation between p53 levels and SLE activity index [8].
The TP53 gene, which encodes the p53 protein, is located on chromosome 17 (17p13.1),contains 11 exons, and its expression is controlled by two promoter regions giving rise to 12 different p53 protein isoforms [9].This gene has numerous polymorphisms, and some of them are known to alter protein function [10].Two TP53 polymorphisms have been more widely studied.The single nucleotide polymorphism (SNP) rs1042522 is characterized by substitution of guanine to cytosine at the second position of the codon 72, which is located in exon 4 of the TP53 gene [NM_000546.5(TP53):c.215C> G (p.Pro72Arg)].Experimental evidence suggests that the p53Arg (CGC codon) variant is more effective in activating the apoptosis pathway in altered cells than the p53Pro (CCC codon) variant, which, in turn, showed to be more efficient in the activation of cell cycle arrest [11].The second polymorphism, rs17878362, is a 16 base pairs duplication in intron 3 of the TP53 gene (NM_000546.5(TP53):c.96+ 25_96 + 40ACC TGG AGG GCT GGGG(1_2)intron variant) [12].The 16 bp duplication allele is associated with lower levels of p53 transcript, which might be due to modification of mRNA processing [13].
One of the mechanisms by which p53 acts on cell cycle arrest is through the transcriptional activation of the gene encoding the p21 protein, a kinase dependent cyclin inhibitor (CDKI).In DNA damaged cells, the p21 protein acts on different cyclin/CDK complexes, promoting cell cycle arrest at the G1/S and G2/M checkpoints [12].p21 inhibits DNA synthesis by binding to proliferating cell nuclear antigen, and also affects key molecules of the apoptotic process, such as p53, and contributes to cell senescence [14].In addition, the p21 protein seems to be associated with susceptibility to autoimmune diseases, especially to SLE [15].
The p21 protein is encoded by the p21 (WAF1, CIP, CDKN1A) gene, which is located in the chromosomal region 6p21.2[16].Changes in the p21 (CDKN1A) gene sequence may result in p21 inactivation, which in turn, may lead to an increased activation and proliferation of self-reactive T cells, and apoptosis [17].More than 40 polymorphisms have already been identified in the p21 gene.One of the most studied polymorphism is rs1801270, characterized by a C to A substitution at the third base of codon 31 of the p21 (CDKN1A) gene (NM_000389.5(CDKN1A):c.93C> A (p.Ser31Arg)).This SNP results in the substitution of serine (AGC codon) by arginine (AGA codon) in a conserved region of the p21 protein [18].The second most studied polymorphism, rs1059234, is a C to T substitution at 20 nucleotides after the stop codon at exon 3 in the 3' untranslated region of the gene (NM_000389.5(CDKN1A):c.*20C> T).By altering mRNA stability and inducing its faster degradation, this SNP leads to a change in the p21 protein level [18].
In recent years, associations of multiple genes and susceptibility to SLE have been analyzed, and among these candidate genes are p21 (CDKN1A) and TP53 [15,19].The number of studies published so far is still small, and there are no reports of studies evaluating TP53 and p21 (CDKN1A) polymorphisms in Brazilian SLE patients.Taking all this into consideration, the objective of this work was to evaluate possible associations of TP53 rs1042522 (G > C), TP53 rs17878362 (16 bp Del/Ins), p21 rs1801270 (C > A) and p21 rs1059234 (C > T) polymorphisms with risk of systemic lupus erythematosus, some SLE clinical manifestations and/or age at onset of symptoms.

Study population
The population included in the group of cases was composed of women with SLE (n = 145) according to the American College of Rheumatology (ACR) criteria [20] who were selected sequentially during routine outpatient consultations at the Rheumatology Department of the State University of Rio de Janeiro.The control group included healthy women without any complain (n = 208) who attended for a routine outpatient clinic at the same institution.Sociodemographic data and clinical manifestations were obtained with a standardized questionnaire and medical records review.The presence of any autoimmune disease was excluded after anamnesis and physical examination performed by an experienced rheumatologist.Skin color/ethnicity was established by self-classification as phenotypic proxy for ancestry previously proved to have a high concordance in Southeastern Brazil [21].All participants signed an informed consent previously approved by the University Hospital Ethics Committee (#321 and #909).

DNA extraction and genotyping of polymorphisms in the TP53 and p21 (CDKN1A) genes
A total of 5 mL of peripheral blood from all participants was collected in Vacutainer tubes containing ethylenediamine tetraacetic acid (EDTA) and transported on ice to the Molecular Biology Laboratory, Department of Biochemistry/IBRAG/UERJ, where molecular analyses were performed.
DNA was extracted from mononuclear cells following a salting out method described by Vargas-Torres et al. [22].
The polymorphisms rs1042522 and rs17878362 in the TP53 gene were analyzed by the polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) techniques.A primer pair (Forward: 5′GAG ACC TGT GGG GAA GCG AA-3′ and Reverse: 5′GGA AGC CAG CCC CTCAG-3′) was specifically designed for amplification of the genomic region containing both TP53 gene polymorphisms and generated a 476 or 492 bp fragment depending, respectively, on the absence or presence of 16 bp duplication in the intron 3 region.
The amplification reaction was performed in a thermocycler (model MJ96 + Biocycler) using the following program: an initial denaturation step at 94 °C for 5 min; 30 cycles of denaturation at 94 °C for 45 s, primer pairing at 61 °C for 45 s and extension at 72 °C for 45 s; and a final step extension at 72 °C for 10 min.The PCR products were analyzed by electrophoresis on 6% polyacrylamide gels followed by ethidium bromide staining, which allowed the identification of the three TP53 rs17878362 (16 bp Del/Ins) genotypes composed of alleles without (A1) and with (A2) 16 bp duplication.
The TP53 rs1042522 (G > C) polymorphism was analyzed by digestion of PCR products (7 μL) with 2 U of the BstUI enzyme (BioLabs) in a final volume of 12 μL adjusted with sterile deionized water.Samples were incubated at 60 °C for approximately 16 h and then electrophoresed on 2% agarose gels stained with ethidium bromide.
Genotyping identification was performed by two independent researchers.In order to validate the genotyping results, 10% of the total samples were reanalyzed using PCR or PCR-RFLP techniques.In addition, representative samples of each genotype were confirmed by direct sequencing of PCR products using an ABI 3130 sequencing platform (Applied Biosystems), following the manufacturer's instructions.

Statistical analysis
Deviation from the Hardy-Weinberg equilibrium was estimated within the study groups (cases and controls) by using the chi-square (χ 2 ) test.
Associations between TP53 and p21 (CDKN1A) polymorphisms and SLE susceptibility in the whole population and in skin color/ethnicity-stratified subgroups were assessed by logistic regression analysis using four genetic models, codominant, dominant, recessive, and overdominant.Combinatorial interactions of the four studied polymorphisms, linkage disequilibrium, haplotype frequency estimation (Expectation maximization algorithm), and haplotype association analyses were also performed.The magnitude of association between TP53 and p21 (CDKN1A) polymorphisms and systemic lupus erythematosus was estimated by calculating odds ratio (OR) and 95% confidence interval (CI).Interaction analysis between the TP53 and p21 polymorphisms and some clinical manifestations of SLE and age at onset of symptoms were performed.Reference categories were the most frequent allele for each polymorphism, the homozygous genotypes for these alleles [24], the combination of these homozygous genotypes, and haplotypes formed by the most frequent variants.Adjusted OR was calculated controlling for age and/or skin color/ethnicity.
Statistical analyses were performed using the software GraphPad Prism version 6.05 (GraphPad Software, Inc., San Diego, CA) and the SNPStats program (Institut Català d'Oncologia, Barcelona, Spain) [25], a webtool for genetic association analysis.A p value below 0.05 was considered statistically significant.

Association of TP53 and p21 (CDKN1A) polymorphisms with SLE risk
All samples were genotyped with respect to at least one of the four polymorphisms, TP53 rs1042522 (G > C), TP53 rs17878362 (16 bp Del/Ins), p21 rs1801270 (C > A), and p21 rs1059234 (C > T).PCR or PCR-RFLP patterns of genotypes corresponding to the four analyzed polymorphisms and their confirmation by DNA sequencing can be seen in Fig. 1.Genotype distribution of the four polymorphisms was in Hardy-Weinberg equilibrium in both study groups, cases and controls (p > 0.05) (Table 1).
No significant results were observed in comparative analyzes of genotype and allele distributions in the groups of cases and controls regardless the genetic model used (Table 1).
Concerning skin color/ethnicity-classification there was no significant difference for the p21 polymorphisms (data not shown).However, both polymorphisms in the TP53 gene showed significant results in the subgroup of white women (Table 2).Statistically relevant differences were observed in the distribution of TP53 rs1042522 (G > C) polymorphism between cases and controls, in both codominant (GC vs. GG: adjusted OR 2.73; 95% CI

Association of combined genotypes and haplotypes regarding the TP53 and p21 (CDKN1A) polymorphisms with SLE risk
Results of genotype combination analysis are shown in Table 3. Considering the entire study population, the TP53 rs17878362 A1A1-p21 rs1801270 CA combined genotype was less frequent in the case group than in controls (OR 0.45; 95% CI 0.23-0.87;p = 0.018).In non-white individuals, the frequency of the genotype combination rs17878362 A1A2-rs1801270 CC was lower among SLE patients (OR 0.28; 95% CI 0.10-0.80;p = 0.015).

Table 4
Haplotype frequency estimation and haplotype association analysis, with respect to the TP53 polymorphisms, rs1042522 (G > C) and rs17878362 (16 bp Del/Ins), and the SNPs in the p21 (CDKN1A) gene, rs1801270 (C > A) and rs1059234 (C > T), in the entire study sample and in the groups stratified according skin color/ethnicity

Discussion
In the present work, the TP53 rs1042522 (G > C) and TP53 rs17878362 (16 bp Del/Ins) polymorphisms were associated with risk of SLE in the subgroup of white women (Table 2).The rs1042522 C allele carriers and rs17878362 A2 allele carriers presented two-, and seven-fold higher risk of developing SLE (p = 0.037 and p < 0.001, respectively).
To date, we could not find any association studies regarding the TP53 rs17878362 (16 bp Del/Ins) polymorphism and SLE or the TP53 rs1042522 (G > C), p21 rs1801270 (C > A) and p21 rs1059234 (C > T) polymorphisms and the development of SLE in Brazilian populations.Moreover, the few published studies related to the importance of TP53 and p21 polymorphisms for the development of SLE present discordant results [15,19,[26][27][28][29].A meta-analysis on association between TP53 polymorphisms and SLE showed significant results only after stratification by ethnicity, confirming the rs1042522 C allele as risk factor for SLE development in Asians (Koreans) [30].We are aware that some studies describing Brazilian genetic structure have reported divergences between self-identified skin color and genetic ancestry [31,32].However, most clinical and genetic studies all around the world employ self-declaration of skin color [33], and this criterion previously proved to have a high concordance in the Southeastern region of Brazil [21].
The divergence of susceptibility genes found among various ethnicities may be related to the influence of local environmental factors, which may interfere with epigenetic modifications [3,28,34,35].Besides, allele distributions of the above four polymorphisms vary widely among different populations and may have different impact on susceptibility for SLE [36,37].Particularly, Brazilian population presents a mixture of genetic ancestry, mostly European and African, which may partially explain the present results [32].Even though a single polymorphism may represent a valuable biomarker of a complex disease, such as SLE, it is believed that multiple small-effect genetic variants influence disease susceptibility [38].Therefore, genotype combination and haplotype association analyses were also performed to assess the combined effect of the four polymorphisms in the TP53 and p21 (CDKN1A) genes on the risk of SLE.Some genotype combinations were absent in some groups or subgroups stratified according to skin color/ ethnicity (Table 3).These findings are not unexpected, since some individual genotypes were found at a very low frequency in our study sample (Tables 1 and 2).Noteworthy, a combined effect of both TP53 polymorphisms seems to exist among Whites, since the rs1042522 GC-rs17878362 A1A2 combined genotype were ninefold more frequent in cases than in controls (p = 0.008) (Table 3).Although a significant difference does exist, due to the limited sample size when the total study population was stratified according to skin color/ethnicity, the statistical power is not strong enough to establish the true association with SLE development.
It has been previously shown that the p53 Arg (rs1042522 G allele) variant has a greater ability to activate apoptosis pathways, while the p53Pro (rs1042522 C allele) variant demonstrates greater efficacy in promoting cell cycle arrest [11], a mechanism that may be associated with SLE development [1,2].Besides, removal of self-reactive B and T cells in SLE patients and mice with lupus-like syndrome could be impaired by defective apoptotic processes [39,40].This might support the hypothesis that the p53 Pro variant may reduce the clearance of auto-reactive lymphocytes in patients with SLE by reducing the efficiency of apoptosis [41].On the other hand, the rs17878362 A2 allele is associated with lower levels of p53 transcription [42].Thus, the suggested combined effect of TP53 polymorphisms could be explained by a lower expression of the p53 protein and the presence of the p53 Pro variant, which is less effective in apoptosis process.
On the other hand, TP53 rs17878362 A1A1-p21 rs1801270 CA (OR 0.45) and TP53 rs17878362 A1A2-p21 rs1801270 CC (OR 0.28) combined genotypes showed protective effects on SLE development in the entire study sample and in the subgroup of non-white women, respectively.Of note, the allele TP53 rs17878362 A2 has been associated with lower constitutive levels of p53 mRNA in lymphoblast cell lines, suggesting that this polymorphism may interfere in mRNA splicing, thus influencing transcript stability [13].A consequence of that includes changes in the p53 protein activity, which may negatively alter its response to apoptosis activation, cell cycle arrest, and DNA repair, all recognized as important mechanisms for the development of SLE.Regarding the p21 rs1801270 (C > A) polymorphism, previous studies reported that both p21 variants exhibit similar activity in relation to CDK inhibition and tumor suppressor activity, although A allele (Arg) seems to be associated with a decrease in mRNA expression [43].In addition, p21 deficiency proved to be associated with loss of immune tolerance in mice [44].Taken together, the protection against SLE attributed to these genotype combinations cannot be explained by the combined influence of the individual genotypes on transcript levels, and/or levels and functions of proteins.However, it is important to emphasize that both p53 and p21 proteins may play different roles, with p21 not always acting as a downstream signaling component of p53 [45] and are involved in several pathways [46] that could be responsible for the observed effect.
In this study, a statistically significant difference was also observed for distribution of the TP53 haplotypes between white SLE patients and controls, with the haplotype rs17878362 A2-rs1042522 C being associated with risk of SLE development (OR 9.67) (Table 4).This estimated risk effect is higher than the effects observed for the individual alleles rs1042522 C (OR 2.23) and rs17878362 (OR 7.00) and corroborate the risk of SLE associated with the rs1042522 GC/rs17878362 A1A2 combined genotype.To determine if the polymorphism are in cis or in trans other methods, such as NGS (Next Generation Sequencing) or genotyping of mother-father-child trios followed by robust bioinformatics tools have to be used [47,48].We believe that our findings may pave the way for further research in this area.
In this work, the relationship between the studied polymorphisms and clinical characteristics of SLE patients was also observed.The SNP TP53 rs1042522 (G > C) was associated with risk for both serositis and neuropsychiatric manifestations, while the TP53 rs17878362 (16 bp Del/ Ins) polymorphism was associated only with neuropsychiatric disorders.Serositis is a common manifestation related to lupus, and its pathogenesis is still not well described.It has been proposed that the appearance of serositis may be associated with an inflammatory response resulting from an abnormal influx of Ca 2+ [49].Although increasing number of genetic associations with pathways involved in innate and adaptive immunity systems has been observed [50], more genetic studies concerning neuropsychiatric disorders are needed.
In conclusion, we found associations of the TP53 (rs1042522 and rs17878362) and p21 (rs1801270) polymorphisms, individually or in combination, with SLE development, particularly in skin color/ethnicitystratified subgroups.In addition, interaction analysis revealed that TP53 polymorphisms might be associated with serositis and neuropsychiatric disorders in SLE patients.Our results suggest that ethnic variations of these polymorphisms should be considered in association studies for complex diseases, such as SLE, as widely shown by different authors [30,31,35,51].To our knowledge, this is the first study to investigate associations of TP53 and p21 gene polymorphisms with the development of SLE in a Brazilian population.Our findings may provide the basis for further studies on association of TP53 and p21 polymorphisms with the development of SLE in different and larger populations, since these polymorphisms may have potential to emerge as SLE susceptibility markers for specific groups of patients.
Adjusted for age and skin color/ethnicity.p values > 0.05 d

Table 2
Analysis of interaction between the TP53 rs1042522 (G > C) and TP53 rs17878362 (16 bp Del/Ins) polymorphisms, skin color/ ethnicity and development of SLE

Table 3
Genotype interactions of the TP53 rs1042522 (G > C), TP53 rs17878362 (16 bp Del/Ins), p21 rs1801270 (C > A) and p21 rs1059234 (C > T) polymorphisms between the group and subgroups of cases and controls stratified by skin color/ethnicity n number of volunteers a Genotype combinations not represented in at least one group or subgroup are not shown Statistically significant results are in bold.b

Table 5
Interaction analysis between the polymorphisms TP53 rs17878362 (16 bp Del/Ins) and TP53 rs1042522 (G > C) and clinical manifestations (serositis and neuropsychiatric disorders) of SLE patients n number of volunteers, NA not applicable Statistically significant results are in bold: a p = 0.021; b p = 0.054; c p = 0.048; d p = 0.015; e p = 0.031; f p = 0.015