PVRL1 as a Candidate Gene for Nonsyndromic Cleft Lip With or Without Cleft Palate: No Evidence for the Involvement of Common or Rare Variants in Southern Han Chinese Patients

Introduction

C left lip with or without a cleft palate (CL/P) is among the most common birth defects worldwide, and has been reported to affect 0.4 to 2.0 per 1000 infants born alive (Schutte and Murray, 1999). Approximately 70% of all CL/P cases occur as isolated, sporadic birth defects, known as nonsyndromic CL/P (NSCL/P), while the remaining 30% occur as a part of more than 300 different syndromes with the Mendelian inheritance pattern, in which CL/P is only one manifestation (Spritz, 2001). Many of the genes responsible for the Mendelian clefting syndromes have been identified, and research has implicated several of them in NSCL/P as well. These general CL/P predisposing genes include IRF6, p63, MSX1, and PVRL1 (Lidral et al., 1998; Sozen et al., 2001; Zucchero et al., 2004; Leoyklang et al., 2006).

The poliovirus receptor related-1 (PVRL1) gene encodes nectin-1, a cell–cell adhesion molecule (OMIM #600644). Studies in mice have revealed that the mRNA of PVRL1 is highly expressed in the medial edge epithelium of the developing palate (Suzuki et al., 2000), suggesting that normal PVRL1 function is involved in fusion of the palatal shelves during palatogenesis (Cobourne, 2004). However, the precise role of this gene in normal human development and the onset of CL/P remains to be defined.

Suzuki et al. (2000) published the first evidence that PVRL1 mutations, especially the W185× homozygous loss-of-function coding mutation, result in a rare autosomal recessive syndrome CL/P-ectodermal dysplasia-1 (CLPED1, OMIM #225000). A follow-up study from the same research team identified heterozygosity of the nonsense W185× mutation as a genetic risk factor of NSCL/P in a northern Venezuelan population (Sozen et al., 2001). These findings established PVRL1 as a promising candidate gene of NSCL/P, and several subsequent investigations have been conducted on the coding regions of PVRL1 (Item et al., 2003; Scapoli et al., 2004, 2006; Turhani et al., 2005; Avila et al., 2006; Ichikawa et al., 2006; Tseng et al., 2006; Sozen et al., 2009; Zhao et al., 2009; Shu et al., 2011). Nevertheless, results from different ethnic groups have been conflicting (the findings of these studies are summarized in Table 1).

We recently analyzed a set of 100 Han Chinese NSCL/P patients to detect the presence of PVRL1 missense mutations previously identified in Filipino NSCL/P cases (G361V in exon 6 of the β isoform [rs7940667], and S112T and T131A in exon 2 of the α isoform) (Avila et al., 2006) and in Thai patients (V395M in exon 6 of the α isoform [rs141253617]) (Tongkobpetch et al., 2008), and to further resequence the affected exons (2 and 5 of the α isoform). Surprisingly, neither of these mutations nor novel variants were detected in our Chinese study population (Shu et al., 2011). These data demonstrated the complex etiology of NSCL/P, which may involve population-specific genetic risk factors.

The majority of NSCL/P-associated PVRL1 investigations to date have been designed in an attempt to identify rare mutations that map directly to the coding regions. Such a strategy is based on the “rare variants-common diseases hypothesis,” which proposes that a significant proportion of the inherited susceptibility to common human diseases may be due to the summation of the effects of a series of low-frequency dominantly and independently acting variants of a variety of different genes, each conferring a moderate but readily detectable increase in relative risk. The problem is that such rare variants will mostly be population specific because of founder effects resulting from genetic drift (Bodmer and Bonilla, 2008). As a result, investigations of different ethnic groups will yield distinctive, and possibly conflicting, data. For example, although the PVRL1 W185× nonsense mutation was identified as a genetic risk factor of NSCL/P in the northern Venezuelan population (Sozen et al., 2001), a follow-up study conducted in Italians (Scapoli et al., 2004), Iowans, Filipinos (Avila et al., 2006), and Taiwanese (Tseng et al., 2006) failed to replicate this association. Thus, an alternative hypothesis, named “common variants-common diseases,” which considers common variants as tags of true disease loci, may be more suitable for assessing the common etiology of NSCL/P through populations. Therefore, we applied two strategies that perform a comprehensive gene association analysis of PVRL1 in Han Chinese NSCL/P patients. First, we used a case–control study design that tests for a possible association with common variants using a set of 45 tagging single-nucleotide polymorphisms (SNPs), which overlapped ∼50 kb at the 5′-end to 50 kb at the 3′-end of PVRL1. Second, we resequenced trios of affected child-normal parents with a positive NSCL/P family history to search for rare high-penetrance mutations. Using these approaches, both the rare- and common- hypotheses were tested in order to provide a more comprehensive view of the association between PVRL1 and NSCL/P. To our knowledge, this is the first study of PVRL1 to have used relatively more tagging SNPs covering all the coding regions along with promoters, introns, regulatory sequences, and splice sites, to search for common NSCL/P-associated mutations.

Materials and Methods

Results

All SNPs were in HWE, except rs11217365 and rs7933848 (p-values of 0.03659 and 0.001725, respectively), which were excluded from subsequent statistical analysis (Table 3).

Two of the 39 tested SNPs, rs12797352 and rs4409845, had p-values that were near the significance threshold (p-values of 0.03339 and 0.0424, respectively) in the allelic test. Another two SNPs, rs7129848 and rs7128327, had slightly lower p-values (0.0176 and 0.009567, respectively) in the genotypic test. However, neither of these results withstood correction of multiple testing (Table 3). Likewise, the sliding window haplotype analyses of two, three, or four SNPs detected no association between the marker sets and NSCL/P (data not shown). The power to detect the association with a marker having a MAF of 0.2 was determined to be nearly 80% (by assuming a range of odds ratios [ORs] of 1.3–6.0).

The entire coding sequence of the PVRL1 α isoform was analyzed by PCR and direct sequencing for the entire set of 45 trios. We failed to detect any novel rare sequence variants in this set of study participants. In all trios, however, the known SNPs that were detected were homozygous for the common alleles (data not shown).

Discussion

In the present study, 470 patients with NSCL/P and 693 healthy controls were evaluated to verify whether common mutations of the PVRL1 gene constitute a genetic risk for NSCL/P within the southern Han Chinese population. Although no strong evidence of an associated marker was identified in our current dataset, one SNP (rs7128327, located 30 kb upstream of PVRL1) showed a trend toward statistical significance (p-value of 0.009567 by the χ ² test at the genotypic level, but the adjusted p-value was not significant at 0.009567×39≈0.37). The interpretation of this type of nominal significance should be carried out very carefully. First, it is possible that the marker may be etiologic or tagging another true etiologic locus; in this way, the current sample size may be too small to generate a strong enough p-value that can withstand multiple-testing correction. To evaluate this presumption, we determined that the present study has 80% power to detect a risk marker with an effect size comparable to those of previous genome-wide association studies (GWASs) (OR of the previously identified NSCL/P-associated SNPs ranged from 1.3 to 6.0) (Birnbaum et al., 2009; Beaty et al., 2010; Mangold et al., 2010). This suggests that some other markers may exist with moderate effects that are below the OR range of 1.3–6.0; such markers would be undetectable or not significant enough based on our sample size. A second possibility is that false positives may have been generated from the population stratification. To accept this presumption, it is first necessary to determine whether a pre-existing population stratification was present in the area from where the samples were obtained. Although enrollment in the present study was based on a seemingly homozygous genetic background of the southern Han ethnicity, evidence of sub-population structures were recently reported. Chen et al. (2009) demonstrated that distinct sub-populations existed within the Guangdong Province, the region from which subjects had been obtained in the present study. Therefore, in the absence of accurate methods that control the potential confounding of such intricate sub-population structures (such as principal component analysis or the transmission disequilibrium test used in the trio study), the population stratification may have biased our study's results. The third possibility to explain the nominal significance observed in this study relates to possible errors in the genotyping assay or enrollments; these types of confounding factors are not unexpected, given the large number of markers tested (de Assis et al., 2011). Considering our data collectively and conservatively, we, thus, speculate that the SNP rs7128327 may not have a significant influence on NSCL/P susceptibility in patients from the Guangdong Province of China, but further studies with methods that control potential confounding effects are warranted.

Thus, the hypothesis that common variants in PVRL1 play a significant role in the development of NSCL/P in the Han Chinese population was disproven. Interestingly, a recent GWAS aimed at searching for common NSCL/P susceptibility genes scanned a mixture of genomic samples of Asian origin (including Chinese samples), but failed to identify any genome-wide associated signal for the chromosomal 11q23 region harboring the PVRL1 gene (Beaty et al., 2010). This finding agrees with the negative results of our targeted investigation of the Han Chinese population.

It is possible that rare, private variants with a relatively high penetrance that map within PVRL1 may be the causative mutation of NSCL/P in the Han Chinese populations (i.e., the rare variants-common diseases hypothesis). However, in our evaluation of exons 1–6 of the PVRL1 α isoform in 45 trios of southern Han Chinese descent, no novel or previously associated mutations were identified. This finding was consistent with results from previous studies in Japanese (Ichikawa et al., 2006), Taiwanese (Tseng et al., 2006), and Chinese populations (Zhao et al., 2009; Shu et al., 2011). However, other studies of NSCL/P patients from Filipino (Avila et al., 2006) and Thai (Tongkobpetch et al., 2008) Asian origins identified several rare causative mutations. Given the complicated heterogeneous nature of NSCL/P and the number of potentially confounding factors (Carinci et al., 2003), this level of discordant results among various ethnicities of a larger race classification is not unexpected (Scapoli et al., 2005). For example, our research team previously investigated an NSCL/P SNP, rs987525 of 8q24, which had been identified as being strongly associated in Caucasians, but we did not detect any such association in our Han Chinese dataset (Xu et al., 2011). Thus, our study's findings further emphasize the importance of recognizing pathogenetic differences between ethnic groups to better understand the NSCL/P disease process and develop more effective targeted molecular therapies.

In conclusion, the present study found no support for the hypothesis that common or rare variants in PVRL1 play a significant role in the development of NSCL/P in the southern Han Chinese population. However, some limitations exist in the present study that should be considered when interpreting our findings. First, this was a hospital-based association study, and selection bias might exist, owing to the potential population stratification just discussed. Future studies of genetic association with NSCL/P will likely benefit from including components such as case–parents trio study to improve the sensitivity of their analysis and accuracy of their results. Second, a small number of affected child-parent trios (n=45) were evaluated in the resequencing phase of our study, and only the α spliced isoform of the PVRL1 gene mRNA was assayed. The α isoform was selected as the exclusive focus, as it encodes the cell-surface transmembrane receptor nectin-1, which mediates the cell–cell adhesion function. Thus, the presence of rare high penetrance mutations in other familial cases or other spliced isoforms (β and γ) of PVRL1 cannot be absolutely excluded on the basis of our results.

Conclusion

The present study provided no support for the hypothesis that common or rare variants in PVRL1 play a significant role in NSCL/P development in the southern Han Chinese population.