Main

TP53 encodes the multi-functional tumor suppressor transcription factor p53 which has a crucial role in maintaining genomic stability in somatic cells exposed to oncogenic or genotoxic stress, thus preventing tumor formation.1 In response to a wide range of stress signals, p53 accumulates in the nucleus and regulates the expression of a large panel of genes involved in the control of cell cycle arrest, apoptosis, cell senescence, DNA repair and energy metabolism. One of the transcriptional targets of p53 is leukemia inhibitory factor (LIF), the gene encoding LIF. LIF is a secreted cytokine with broad roles in the control of lymphocyte proliferation and differentiation. It has also been identified as a critical factor for blastocyst implantation.2 Control of p53 over LIF expression is operated through a p53-response element located in intron 1 and conserved in both mouse and human LIF genes.3

Recent studies have demonstrated that p53 regulates female reproduction and blastocyst implantation through LIF. Implantation is a critical step in mammalian embryonic development during which the blastocyst establishes close interactions with the uterus, leading to the formation of the placenta supporting fetal development.4 Hu et al.4 have demonstrated that p53 regulates LIF expression in the uterus of female mice. p53-deficient mice express lower levels of LIF than their p53-competent counterparts and show impaired blastocyst implantation and consequently, impaired fertility.

There is strong evidence that genes at critical regulatory nodes in the p53 pathway are under evolutionary selection5, 6 and that SNPs in the p53 pathway influence human fertility.7 Of these, one of the most studied is TP53 polymorphism in exon 4 (PEX4 of the TP53 gene), widely known as p.P72R (C/G, rs1042522). This single-nucleotide polymorphism (SNP) located at the second position of the codon 72 consist in either an ancestral C allele whose frequency in African populations is around 0.70 or a derived G allele whose frequency in European and Asian populations varies from around 0.50 to 0.80. Presence of the C allele results in a proline in codon 72, and presence of the G allele, in an arginine. These polymorphic protein variants significantly differ in their biological properties and there is evidence that R72p53 has higher transcriptional activity toward a particular subset of p53 target genes, including LIF, than P72p53.8 Previous studies have identified an association between TP53 PEX4 and infertility7 or endometriosis (END).9, 10, 11 It has been suggested that the effect of PEX4 on LIF expression and fertility may account for population differences in the distribution of PEX4 alleles in different parts of the world. These differences may reflect subtle adaptation to environmental constraints affecting fertility. However, the magnitude of the PEX4 effect on infertility associated with different pathological causes remains controversial.12, 13

PEX4 is in strong linkage disequilibrium with another common polymorphism located in its close vicinity, PIN2 (polymorphism in intron 2; rs1642785; G/C). The PIN2 G allele has been associated with human papillomavirus persistence14 and individuals with two copies of the PIN2 G allele have been reported as having an increased risk of osteosarcoma.15 Recently, it has been shown that another polymorphism in intron 3 of the TP53 gene, PIN3 (Polymorphism in Intron 3, rs17878362, 16 bp duplication, N=non-duplicated, D=duplicated) overlaps with a G-quadruplex motif, which regulates p53 mRNA splicing generating an alternatively spliced form, which supports the synthesis of an isoform of p53 lacking the N-terminal transactivation domain (Delta40p53).16 PIN3 D allele is associated with increased risk of colorectal,17 lung18 and breast cancer,19 whereas the N allele has been reported in association with an average acceleration of 19 years in the mean age at first cancer diagnosis in a Brazilian cohort of TP53 germline mutation carriers.20 The effects of this polymorphism in END or infertility have not been investigated so far.

Although the association between END and infertility is well known (END affects up to 50% of women with infertility),21 the cause of infertility in the disease is not fully understood but is thought to involve hormonal,22 immunological,23 genetic,24 proliferative (endometrial) and uterine alterations.25 We hypothesized that TP53 polymorphisms that alter p53 function may be associated with in vitro fertilization (IVF) failure and with END-associated infertility.

Results

Patients and healthy study subjects did not differ significantly regarding self-attributed skin color (Supplementary Table S1). Overall, a self-denomination of ‘white’ color predominated in all study subgroups (END, FIV, Unselected and Fertile). In terms of reproductive history, the mean number of pregnancies in women of the fertile and unselected for fertility groups was 3.62±1.9 and 3.22±2.1, respectively. In the later, nulliparity was observed in 2.6%.

Women in the fertile and unselected for fertility groups presented higher mean age at recruitment (42.68±12.8 years and 43.2±12.7 years, respectively) as compared with END (32.87±4.7 years) and IVF (31.65±3.2 years) groups. Hardy–Weinberg equilibrium was achieved in all study groups for PIN2, PIN3 and PEX4 (all P>0.05, Supplemental Materials, Table S2. Genotypic and allelic frequencies of the TP53 polymorphisms are shown in Table 1. In all four study subgroups, PIN3 and PEX4 allele frequencies did not differ significantly from those previously described in European populations (Supplementary Table S3).

Table 1 Genotypic and allelic frequencies of selected TP53 polymorphisms between Fertility Unselected, Fertile, END and IVF groups
Full size table

Single marker analysis (Table 1) revealed a significant association between PIN2 (rs1642785) genotypes and IVF (P=0.016), and a borderline association with the END group (P=0.052) when compared with the Fertile group. There was an increased frequency of the PIN2 C allele in both the END and IVF groups. When analyzing TP53 PIN3 (rs17878362) polymorphism, a clear difference between IVF and END groups was observed when compared with the Fertile group. Allele D (the duplicated allele) was enriched in patients in both groups as compared with Fertile (P=0.042 and P<0.0004 for the END and IVF groups, respectively). For TP53 PEX4 (rs1042522), a statistically significant difference between both the END and IVF groups and the Fertile group was also demonstrated, with enrichment of the PEX4 C allele in both groups (P=0.007 and P=0.009, respectively).

When the Fertile and Unselected groups were compared, we observed that the allelic frequencies of PIN2 G and PEX4 G were significantly higher in the Fertile group, whereas PIN2 and PEX4 genotype distribution did not differ between groups. Similarly, PIN3 genotypic or allelic frequencies did not differ between groups (Table 1). For both the END and IVF groups, the allelic frequencies of PIN2, PIN3 and PEX4 differed significantly from those observed in the Fertile group. The allelic frequencies of PIN3 in infertile women (either END and IVF groups) differed significantly from both the Fertile and Unselected groups.

Haplotype analysis showed strong linkage disequilibrium between TP53 PIN2 and PEX4 (D′=1; r2=0.94 in all studies groups, Supplementary Table S4) as previously described.20 Therefore, we have only considered TP53 PIN3 and TP53 PEX4 in further analyses and in our discussion. We carried out a binary logistic regression analysis to evaluate the effect of TP53 haplotypes with regard to END and IVF. Figure 1 shows the distribution of the most frequent haplotypes encountered (see Figure 1). Table 2 shows the odds ratios for the END and IVF groups of the most frequent haplotypes when compared with the reference N-G haplotype. Haplotypes D-C and N-C were related to higher risk for END (P=0.002, P=0.001, respectively) and failure of IVF (P=0.018 and P=0.002, respectively) when compared with the Fertile group. However, when the Unselected group (unselected for fertility) was used as the comparison group in the logistic regression model, the risk association with haplotypes D-C and N-C was not observed (data not shown).

Figure 1
figure 1

Distribution of the most frequent haplotypes among Unselected, Fertile, END and IVF groups. Haplotypes frequencies are shown as ‘%’. Haplotypes were constructed as PIN2 (G/C) – PIN3 (N/D) – PEX4 (C/G)

Full size image
Table 2 Binary logistic regression model for TP53 haplotypes regarding PIN3 (N/D) and PEX4 (C/G) using the Fertile group as reference
Full size table

Discussion

In this study, we have analyzed the distribution of three common polymorphisms in the TP53 gene (PIN2, PIN3 and PEX4) in infertile women with failure of IVF treatment or with END-associated infertility. Our results demonstrate an association between these two forms of infertility and TP53 alleles PIN3 D and PEX4 C, suggesting that variations in p53 activity specified by these polymorphisms may be involved in the pathogenesis of both conditions. These results support previously reported observations on associations between PEX4 and infertility, in particular IVF failure. Furthermore, these results provide clear evidence in favor of an association between TP53 polymorphism and infertility-related END. Regarding TP53 PIN3, several studies have evaluated the association between this polymorphism and lung18 or breast cancer,19 but to our knowledge, no previous study has analyzed its association with infertility or END.

Previous studies have shown associations between PEX4 C allele and END,10, 26, 27 whereas others fail to demonstrate this association.11, 12 These controversies may be due to the environmental and genetic background of the studied populations but also because of differences in illness classifications (END is sometimes asymptomatic and often can only be diagnosed by laparoscopy). In our study, the END group was carefully diagnosed according to the American Society for Reproductive Medicine (ASRM) and women with END were excluded from both IVF and Fertile groups after laparoscopic examination. In addition, all four study groups described here were quite homogeneous in terms of self-reported skin color (a feature used as proxy for ‘race’ or ancestry background in Brazil) corroborating with previous population-based studies that demonstrate predominance of European genomes in this specific region.28, 29, 30, 31, 32 This observation was further confirmed by comparative analysis of PIN3 and PEX4 allele frequencies encountered here and those previously described in European/European-derived and African/African-derived populations, showing that in all four study groups, allelic distribution was not statistically different from the observed in Europeans/Europeans-derived.

Kay et al.33 were the first to associate TP53 PEX4 C allele with women experiencing recurrent implantation failure. Other studies also associated the PEX4 C allele with the occurrence of idiopathic recurrent miscarriages34 and implantation failure,35 and Kang et al.7 demonstrated that PEX4 C was significantly enriched among IVF patients, serving as risk factor for implantation failure. Our results are in agreement with these previous findings regarding the TP53 PEX4 C allele and confirm this allele as a risk factor for both END-associated infertility and IVF failure in a different sample set.

These results and the findings of our study suggest that PIN3 and PEX4 polymorphisms present specific functional differences in p53 protein variants, having an impact on events that are critical for embryo implantation and/or early development. In the case of PEX4, there is experimental evidence from cell and animal studies that the p53 protein encoded by the PEX4 C allele (P72p53) is more efficient in initiating senescence that the product of the PEX4 G allele (R72p53), which in turn appears to have a stronger effect on p53-mediated apoptosis and suppression of cell transformation.36 In the case of PIN3, presence of the TP53 PIN3 D allele has been associated with reduced levels of TP53 mRNA in lymphoblastoid cell lines.17 Whether this effect also occurs in vivo remains to be determined. Marcel et al.16 demonstrated that TP53 PIN3 is located within a GC-rich region of intron 3 that form G-Quaduplex structures, which modulate splicing of intron 2. In silico models predict that PIN3 may alter the topology of these G-quadruplex structures, thus modifying the patterns of p53 mRNA isoform expression. The p53 isoform encoded by alternatively spliced p53 retaining intron 2 lacks the N-terminal domain containing the main transactivation activity of p53, thus resulting in an N-terminally truncated protein, which binds DNA but does not activate transcription through p53-response elements. It is important to emphasize that in our study TP53 PIN3 presented an allelic distribution that was significantly different in infertile women (either END or IVF groups) when compared with women selected for fertility but also when compared with women from a community sample and unselected for fertility, suggesting that genetic variations in PIN3 may have a critical effect on infertility. Further experimental studies are needed to evaluate the possible impact of p53 isoforms in regulating these biological events, especially their impact on transactivation of key genes involved in the early stages of gestation, such as LIF. Haplotypes D-C and N-C were related to higher risk for END and IVF only when a group of women selected in favor of normal fertility (the Fertile group) was used as comparison group; this was not observed when the comparison group included women unselected for reproductive history. This observation suggests that specific haplotypes of TP53 may be associated with high fertility features. Given the associations between specific SNPs and infertility, it is reasonable to assume that particular combinations of SNPs might provide a genetic marker for women with high fertility features.

Our results support that TP53 polymorphisms have a role in both END-associated infertility and IVF failure; although current evidence points to a strong effect of the PEX4 polymorphism in embryo implantation and fertility, other SNPs in TP53, especially PIN3, may have a key role in the modulation of this process and in other biological processes related to early embryonic development. PIN3 was the only SNP that showed differential frequencies in infertile women (either END and IVF groups) when compared with either fertility-selected or unselected groups, whereas PIN2 and PEX4 only showed a differential distribution in END and IVF patients when compared with a group of patients at the other extreme of the phenotype (fertile group).

In conclusion, the data presented here add to the current evidence that variations in expression and activity of p53 may have an effect on the expression of key genes related to the control of cellular growth and invasion, which have been associated with END (BAX, FAS, PIG11, PTEN), as well as on genes associated with embryo implantation (LIF). Infertility associated to END could be related, at least in part, to embryo implantation failure in a mechanism similar to that seen in other infertile women without END. It may also involve other mechanisms affecting early embryonic development as well as cell–cell communications during the pre-implantation and implantation phases. In agreement with this hypothesis, previous studies have demonstrated lower implantation and pregnancy rates in endometriotic patients.37 TP53 polymorphisms, especially PIN3 and PEX4 may have an interest as biomarkers and could add to the development of a clinically relevant genetic profile that would be of great help for clinicians to identify patients at higher risk for IVF failure. The results of this study should be confirmed in larger cohorts with well defined phenotypes of END and infertility and long-term follow-up data. They also emphasize the importance of a clear definition of clinical phenotypes and of study design when analyzing the effects of specific polymorphisms on fertility.

Materials and Methods

Patients and subjects

All patients and subjects were informed about the procedures of the study when invited to participate and signed a consent form at inclusion. The research project was approved by the Institutional Ethics Committee (Hospital de Clinicas de Porto Alegre – GPPG 05-182; GPPG 09-430).

At inclusion, patients and subjects were also asked to provide a description of their perceived skin color. In Brazil, skin color is normally used to define an equivalent to ‘race’ or ancestry background.32, 38 We used the words ‘White’ and ‘non-White’ to identify women who defined themselves with some term that suggests only European ancestry and with other terms that suggest some level of African ancestry (such as mulato or pardo), respectively. No term that reports some level of Amerindian ancestry was used by volunteers.

Patients and subjects were divided into four study groups. The IVF Group consisted of 115 women (<35 years) with at least one IVF failure, defined as a failure after IVF cycle treatment with transfer of two or more top quality embryos (8 cell embryos with <20% fragmentation). Briefly, inclusion criteria of this group were: age <35 years, exclusion of END by laparoscopy and the main factor was of mild masculine (oligospermia) or tubal origin. All patients in this group were submitted to conventional IVF. Patients with previous thyroid disease, positive anti-lupus or anti-cardiolipin antibodies and trombophilias were also excluded from our sample. Controlled ovarian hyperstimulation was performed with the use of recombinant human FSH and pituitary suppression with GnRh antagonist (fixed day-6 protocol). Ovulation was induced by 6500 IU recombinant hCG when at least three follicles had reached a diameter of >17 mm, and transvaginal follicle aspiration was performed 36 h later under ultrasound guidance. Embryos were classified according to the cumulative embryo classification, taking into account cleavage speed, blastomere symmetry, extent of fragmentation and the presence or absence of multinucleated blastomeres.

The END group comprised 98 infertile women with minimal or mild END as diagnosed by laparoscopy recruited at the Gynecology Service of Hospital de Clinicas de Porto Alegre (HCPA), in Southern Brazil. Infertility was defined as the inability of a couple to achieve pregnancy after 1 year of regular unprotected sexual intercourse.39 Other causes of infertility were excluded by hysterosalpingography, sperm evaluation and hormonal measurements whenever necessary. END diagnosed during laparoscopy was categorized according to the classification proposed by the ASRM.39

The Fertile group consisted of 134 women with no history of infertility, who already had children without any difficulties or assisted reproduction and underwent laparoscopy for tubal ligation at HCPA. END was excluded in women from IVF and Fertile groups. In addition, we studied a group of 300 asymptomatic women, who volunteered for a community-based breast cancer screening program in Southern Brazil (from the same geographic recruitment area of the patients included in the IVF and Fertile groups). This group (‘Unselected’) was unselected with respect to fertility or infertility-related symptoms, as described elsewhere.40

Genotyping

Genomic DNA was extracted from peripheral blood using the Ilustra blood genomic Prep Mini spin Kit (GE Healthcare, Piscataway, NJ, USA) as described by the manufacturer. Genotypes and haplotypes defined by the three TP53 gene polymorphisms (PIN2 rs1642785 G/C, PIN3 rs17878362 16 pb duplication and PEX4 rs1042522 C/G) were determined by Amplification Refractory Mutation System as previously described.20

Statistical analysis

The clinical characteristics of the women in all study groups were compared by one-way analysis of variance. Differences in genotype/allele distribution between IVF, END, Fertile and Unselected groups were evaluated using χ2-analysis, also used to test for Hardy–Weinberg equilibrium. Linkage disequilibrium was assessed calculating D′ value (the relative magnitude of D as compared with its theoretical maximum, calculated as D/Dmax) as described by Lewontin.41

Binary logistic regression analysis was carried out to estimate the odds ratios with 95% confidence intervals in order to assess the influence of TP53 haplotypes for END and IVF using the Fertile group as reference. Haplotype frequencies were calculated by direct count. Statistical analysis was performed using the SPSS 18.0 statistical package. All reported P-values are two-tailed and considered statistically significant when 0.05⩾.