Distribution of Lewis and Secretor polymorphisms and corresponding CA19‐9 antigen expression in a Chinese population

The Lewis (FUT3) and Secretor (FUT2) genes, corresponding to secretion of Lewis ABO (H) histo‐blood group antigen CA19‐9, are highly polymorphic with differences between populations. In this study, the FUT3 and FUT2 genes in 316 Chinese participants were sequenced to detect polymorphisms, and the associated CA19‐9 antigen secretion was also measured. In total, 14 genotypes of FUT3 and 10 genotypes of FUT2 were verified. Le/Le, Le/le 59,508 and Le/le 59 were the main genotypes of FUT3 with frequencies of 53.2%, 10.7% and 3.5%, respectively. Se/Se, Se/se 385 and se 385/se 385 were the main genotypes of FUT2, with frequencies of 21.4%, 18.6% and 16.2%, respectively. The alleles le 1067 and le 508 were found extensively in the Chinese population, and the frequency of allele se 385 was shown to be higher than previously reported. Phenotype analysis revealed that 9.8% of individuals were the Lewis‐negative type and 22.5% were the secretor‐negative type. Combined phenotypes showed that 3.2% of participants were of ‘double‐negative’ phenotype (le, se) and 19.3% were of single dominant non‐secretor phenotype (Le, se). Serum Lewis antigen CA19‐9 levels were significantly different between subgroups and consistent with the defined phenotype. Our study revealed the unique distribution of Lewis and Secretor polymorphisms in a large Chinese population, and decoded the combined genotypes of the two CA19‐9‐related genes.

Lewis-negative individuals (the le/le genotype) have the Lewis (FUT3)-negative phenotype, Le (aÀbÀ), irrespective of the Se genotype. However, Lewis-positive individuals (the Le/Le and Le/le genotypes) are divided into three Lewis-secretor phenotypes according to distinct Secretor genotypes as follows: (a) Le (aÀb+) secretors with the Se/Se or Se/se genotype; (b) Le (a+bÀ) non-secretors with the se/se genotype; and (c) Le (a+b+) partial secretors having homozygosity for the weak Secretor allele [1,2,20,25]. Nucleotide substitutions inactivating the FUT2/3 genes have been found within various populations, and the phenotypes have been determined by the ethnic group-specific genotypes [1,8,16,19,24,25]. Both genotypes of Lewis and Secretor are crucial for an individual's phenotype formation, and the serum CA19-9 value will directly reflect the differences among individuals.
Knowledge of the polymorphisms of the Lewis and Secretor genes in the Chinese population will help in classifying the subgroups and defining an accurate condition for normal phenotypes. In this study, we aimed to examine the prevalence of the five major nucleotide polymorphisms of the FUT3 gene and the four main variation types of the FUT2 gene in a Chinese population. Comparison of serum CA19-9 expression between each phenotype was performed as well. Distribution of the Lewis-negative and secretor-negative phenotypes in the Han ethnic population was also evaluated and compared with previous results in other populations. Since they are based on a reliable method, we expect the results will provide an important reference for disease diagnosis and therapy in the Chinese population.

Participants and genomic DNA isolation
Blood samples were obtained from 316 unrelated and randomly selected healthy individuals of Han ethnicity in the eastern region of China. Either their birthplace or the paternal origin of the participants in this study was in mainland China. Oral informed consent was obtained from all participants in this study. We collected the peripheral blood in tubes containing ethylene diamine tetraacetic acid and isolated the white blood cells. Genomic DNA was extracted from white blood cells using a QIAamp DNA Blood Mini kit (Qiagen, Inc., Hilden, Germany) according to the manufacturer's instructions.

PCR amplification of FUT3 and FUT2 genes
Both FUT3 and FUT2 genes have no intron in the open reading frame and fewer reactions are capable of amplification of the complete mutation region. Three pairs of PCR primers respectively specific for the FUT3 and FUT2 gene segments are shown in Table S1, and sequence design partly referred to those previously reported [16,26]. For each segment amplification, 20 ng of genomic DNA was combined with the primers (7.5 lM for forward and reverse) in a PCR system of final volume 25 lL. Each PCR system contained 5 mM dNTPs, 37.5 mM MgSO 4 , 2.5 lL 10 9 PCR buffer (particular for KOD -Plus-Neo) and 0.5 U KOD -Plus-Neo (Toyobo Co., Osaka, Japan). Thirty cycles were run (2 min at 94°C, 10 s at 98°C, 30 s at T m and 30 s at 68°C, where T m for 385F/385R is 62°C, for 508F/1067R is 65.5°C and for 21F/21R is 60°C), and the products were isolated from agarose gels for sequencing.

Direct DNA sequencing and genotyping
The purified amplification products were sequenced directly for FUT3 and FUT2 genotyping. The anterior half-segment of FUT3 concluding at the 59, 202 and 314 position was directly sequenced with primers 385F and 385R. The bottom half-segment of FUT3 concluding at the 59, 202 and 314 position was directly sequenced with primers 508F and 1067R. The complete segment of FUT2 concluding at the 357, 385, 428 and 739 position was directly sequenced with primers 508F and P1R. The dideoxynucleotide termination sequencing reaction was performed by using the ABI Big-Dye Terminator cycle sequencing system, and the DNA sequence was analyzed by an ABI PRISM 3730 instruments (Applied Biosystems, Carlsbad, CA, USA). Sequencing data were generated by the ABI 3730 Genetic Analyzer platform and were analyzed using CHROMAS software [27]. The genotype present at each SNP site was directly determined by one or two different color peaks on the electropherogram.

CA19-9 antigen measurement and statistical analysis
Serum CA19-9 antigens were measured using an electrochemiluminescence immunoassay on the Roche Cobas e601 (Roche MODU D + P model, D2400-P800) immunoassay analyzer (Roche Diagnostics, Mannheim, Germany). An obtained CA19-9 value of less than 0.06 UÁmL À1 was considered to be undetectable. The frequencies of the gene polymorphisms, as well as the genotypes and phenotypes, were determined by description analysis. Difference of CA19-9 values among each subgroup were tested by one-way analysis of variance and Student's t test was used to compare between each single group. Statistical analyses were performed with SPSS Statistics 19 (IBM Corp., Armonk, NY, USA) and P < 0.05 was defined as statistically significant.

Ethical approval
All procedures performed in studies involving human participants were in accordance with the ethical standards of the Fudan University Shanghai Cancer Center.

Distribution of FUT3 and FUT2 gene polymorphisms in a Chinese population
We amplified the coding regions containing the SNPs in the FUT3 (Lewis) and FUT2 (Secretor) genes and examined the respective polymorphisms. For the Lewis gene, nucleotide 59T>G was the most prevalent mutation, with a frequency of 39.87% (7.59% homozygous and 32.28% heterozygous). Variations 508G>A (25.00%) and 1067T>A (8.23%) manifested in a homozygous form. Variations 202T>C and 314C>T had complete linkage in 18 individuals (5.70%) and no homozygous alleles were encountered (Table 1). For the Secretor gene, four prevalent polymorphisms were also investigated. It is known that the most frequent Secretor gene mutations are detected in Asians, which as a homozygous mutation causes a non-secretor phenotype [20,23,24]. 385A>T was detected in 222 individuals (70.25%) with a T allele frequency of 46.04%. Variation 357A>T was detected with a frequency of 24.68%, and induced a synonymous mutation in Secretor. Mutation at nucleotides 428 and 739 were rarely detected (1.27%) and showed complete linkage (Table 1).

Haplotype analysis of FUT2 and FUT3 genes
Based on combinations of known alleles and their secretor phenotypes, seven haplotypes of the FUT3 gene and five haplotypes of the FUT2 gene were estimated according to each allelic polymorphism. Le and le 59,508 were present as the most common haplotypes in FUT3, with a frequency of 72.94% and 14.72%, respectively. Haplotypes le 59 and le 59,1067 appeared with an equal frequency of 4.75%. Isolated le 508 and le 1067 alleles were rare (0.32% and 0.16%, respectively) and le 59,508,1067 was not detected. le 202,314 combining the 202T>C and 314C>T variations was only found as a co-occurrence, with a frequency of 2.85%. (Table 2).
For the FUT2 gene, Se and se 385 were the most prevalent haplotypes at a frequency of 46.20% and 40.19%, respectively, and thus accounted for 86% of all allele counts. The rarely detected Se 357 and se 357,385 were present at a frequency of 7.12% and 5.85%, respectively. We also noticed that sole variations of se 428 and se 739 were hardly encountered in the Chinese population and only appeared as allele se 357,428,739 at a low frequency of 0.63% ( Table 2).

Distribution of FUT2 and FUT3 genotypes
For FUT3 (Lewis) genotypes, Le/Le, Le/le 59,508 and Le/ le 59 were demonstrated to be the most common positive genotypes (giving a Lewis-functional phenotype), and were detected with a frequency of 55.70%, 17.72% and 7.59%, respectively. Thirty-one individuals (9.82%) exhibited Lewis-negative genotypes (giving the Lewisnull phenotype), in which le 59,508 /le 59,508 (2.17%) was shown to be the most prevalent ( Fig. 1A and Table 2).
For FUT2 (Secretor) genotypes, allele se 385 was the major factor giving rise to 97% of non-secretor phenotypes. In the secretor-positive cohort, Se/se 385 was confirmed to be the most prevalent genotype, with a frequency of 36.08%, and individuals with the Secretor wild-type (Se/Se) were at 17.41%. The non-secretor phenotype was found at a frequency of 22.46% and mostly (21.84%) was identified as se 385 /se 385 . Another non-secretor genotype, se 385 /se 357,428,739 , was much rarer (0.63%). se 428 was previously reported as the major mutation to cause the non-secretor phenotype and to be widely distributed in other races [28], but was not detected in Chinese. In addition, variation at position 428 was also present in conjugation with the 357 and 739 variation, and completely co-present with 739 ( Fig. 1B and Table 2).
A cluster analysis based on each SNP was performed by combining genotypes of the studied objects (Fig. 1). In total, 55 combined genotypes were encountered of which 18 types (84.9%) were present with a frequency greater than 1% ( Table 3). As shown in the genotypes distributions in Fig. 1, subjects homozygous for the functional allele are marked as Le/Le or Se/Se, homozygous mutated loss-of-function alleles are denoted as le/le or se/se, and heterozygotes mutated genotypes are represented by Le/le or Se/se. The particular nucleic acid site variations of the Lewis, Secretor and combined genotypes is showed in Table S2.

Combination of FUT2 and FUT3 phenotype and corresponding CA19-9 value
Corresponding to each genotype, the phenotype of single individuals was investigated. Four combined phenotypes were identified according to Lewis (FUT3) and Secretor (FUT2) genotypes (Fig. 2), namely Le, Se; Le, se; le, Se; and le, se. Among the population, 226 participants (71.52%) were detected as Le, Se (the double-positive phenotype), which manifested as the highest frequency. Le, se (Lewis positive and synchronously secretor negative) and le, Se (Lewis negative and synchronously secretor positive) were present in 62 individuals at a frequency of 19.62% and in 18 individuals at a frequency of 7.28%, respectively. The doublenegative phenotype le, se was shown in 3.17% (10 of 316) of individuals. Table 3 summarized the 18 highfrequency genotypes and corresponding phenotypes.  The serum CA19-9 value of each participant was measured and compared between genotypes. Comparison between the Lewis-functional phenotypes showed that CA19-9 values in participants with the Le/Le genotype were significantly higher than ones with the Le/le genotype (mean 13.07 vs 9.37 UÁmL À1 , P < 0.05). Otherwise, CA19-9 values in the Lewis-null phenotype group were completely undetectable (CA19-9 value < 0.06 UÁmL À1 ) (Fig. 3A). Grouping by secretor phenotype, participants with the se/se phenotype showed a much higher CA19-9 value (mean 17.27 UÁmL À1 ) than the Se/se type (mean 9.43 UÁmL À1 ) and Se/Se type (mean 7.38 UÁmL À1 ) (P < 0.0001) (Fig. 3B). For the combined Lewis and secretor phenotypes, the Le/se group showed the highest CA19-9 value compared with other groups (mean 20 UÁmL À1 , 95% CI: 17.95-22.05 UÁmL À1 ) (P < 0.0001) ( Table 4). Both le/se and le/Se types showed an undetectable CA19-9 value ( Fig. 3C and Table 4).

Population differentiation of the genotype in FUT2 and FUT3 alleles
In the present study, we encountered seven kinds of negative Lewis haplotypes and four kinds of negative Secretor haplotypes. Figure 4 presents the ethnic specificity of putative allelic frequencies among various populations including Korean [9,29], Chinese [10,22,23,30], Japanese [6,25], Thai [26], Caucasian [15,31] and others. Distribution of the FUT3 genotype is approximatively consistent with reported data in Chinese and other Asian populations. Nevertheless, no Asian population study has confirmed the existence of le 508 previously, which is at a frequency of 0.16% in the present study and 0.70% in an Amazonian population [16]. le 59,508 is abundantly distributed in most populations with a frequency from 14% to 31%, but rarely found in Caucasians (1.0% and 1.5%). In contrast, le 202,314 is present in Asians at a much low frequency compared with Caucasians [15,31] (Fig. 4A).
The three most common FUT2 alleles in the present study were Se, se 385 and Se 357 , with a frequency of 46.20%, 40.19% and 7.12%, respectively, and the distribution is highly distinctive in multiple Asian populations including four Chinese groups [22,23,30], except for one group in Taiwan [26] (Fig. 4B). We have noticed that allele se 385 was found at an especially common rate in the current study compared with other populations. On the other hand, the frequency of se 357,428 in the present population is much lower, while the frequency of Se is much higher. No se fus was detected in Asians except for Japanese (4.7-5.7%). In contrast to Xhosa and Caucasians, se 428 in Asians is rare and previous studies revealed a very low frequency. However, our data showed that 428 absolutely conjugated with other mutations in Chinese rather than being an isolated allele. se 357,428,739 was not reported before in other studies, mainly accounted for by a deficiency in nucleotide 739 detection. The variations of Lewis and Secretor allele distributions among different populations are shown Table S3.

Discussion
The distribution of genotype has been reported to have ethnic specificity, and genetic heterogeneity in Lewisnull and non-secretor individuals is obviously present in different populations [2,15,17,18,32]. Even in the Chinese population, the frequency of phenotypes varies with subject selection, study range, detected position and investigation method [10,19,22,26,30].
Integral and systemic comprehension of the FUT3 (Lewis) and FUT2 (Secretor) genotypes in the Chinese population has been insufficient.
In a large cohort of 316 Chinese participants, we performed accurate genotyping and a combined analysis of nine SNPs reported previously in the fucosyltransferase genes [8,10,16,19,23,24]. False-negative reactions are often produced in Lewis blood typing due to weak hemagglutination reactions, which in the main is accounted for by adsorption of glycolipids and by erythrocyte levels [2,16,25]. For further study, FUT2/FUT3 genotyping is the most accurate method for distinguishing the phenotypes [6,7,11,20]. Accordingly, by direct sequencing, five major SNPs of the FUT3 gene and four major SNPs of the FUT2 gene were detected. Additionally, we also discovered 18 primary genotypes and four phenotypes in the Chinese cohort by cluster analysis. It was known that Lewis genotypes in Asians are inconsistent with the polymorphism in Europe and America, and our study has shown the unique distribution of the FUT3/FUT2 gene polymorphism and allied genotype frequencies in a randomly recruited Chinese population. We analyzed the coding DNA sequence range and in FUT3 did not detect any SNPs other than those previously reported, but three individuals with heterozygous 571C>T (rs1800028) in FUT2 were detected.
In the present population, the 'wild-type' (Le/Le) FUT3 gene was detected at a frequency of 55.70%, which is consistent with previous reports in Chinese and Asians. Compared with Caucasians, allele le 59,508 is more abundant and le 202,314 is much rarer in Asians [15,31]. The Amazonian population shows a particular enrichment of the le 59 allele compared with other populations [16]. Allele le 1067 and allele le 508 were reported to be absent in Asians; nevertheless, they were detected at the frequency of 0.32% and 0.16%, respectively, in the present Chinese population. This discrepancy could be attributed to the sensitivity of the detection method. The two alleles were reported with much higher frequency in a study that included three Asian populations, which was mostly caused by omitted the detection of le 59 [19].
The mutated allele le 59 was shown to have the highest prevalence (40.19%) in the present participants, which homozygotes generating an amino acid variation in the transmembrane domain. Variation 59T>G leading to an L20R amino acid substitution was reported to be responsible for a Lewis-negative phenotype probably accounted for by a reduction in Golgi retention [3,7,33]. We classified the isolated 59 mutated allele in the Lewis-negative group (marked as le 59 ), even though multiple studies have counted the allele in the Lewis-positive genotype [10,15,16]. Variations of nucleotides 202 and 314 have usually been reported as present together [32,34], and we found that both T202C and C314T were present at frequencies of 5.70% in the Chinese population and no isolated alleles were found. Similar correlation of two single mutations in FUT2 was also demonstrated for 428 and 739, at a lower frequency of 1.27%.
Earlier reports have revealed that differential histoblood group antigen expression (such as CA19-9) has been insufficiently attributed to FUT3 variation. Enzyme FUT2 competitively binds the substrate from FUT3 and the genotype also influences histo-blood group antigens status. Multiple polymorphic nucleotides in the FUT2 gene reflecting the secretor (Se) and non-secretor (se) type have been reported with ethnic specificity [5,22]. Homozygous 428G>A (non-secretor allele se 428 ) was the first identified missense mutation and was reported in approximately 20% of Caucasians [5,20]. In Asians, variations 357C>T and 385A>T were common, but variation 428G>A was rare [9,23,24,30]. se 385 is reported as the most prevalent allele in Asian populations and showed the highest frequency (40.19%) in the current population. Although the allele se 385 was definitively associated with the nonsecretor phenotype, it was rarely present in European populations (at a frequency of 0.4%) [24]. In  comparison with other populations including Chinese, allele Se (wild-type) and allele se 385 (non-secretor type) appeared at obviously higher frequencies in the current study. Simultaneously, functional allele Se 357 and nonsecretor se 357,385 were much rarer than in other populations (Fig. 4B).
As the synthesized product, the Lewis ABO (H) histo-blood group antigen is routinely utilized as a clinical diagnosis biomarker, especially for malignant gastrointestinal tumor indication [35]. It was previously known that individuals with a Lewis-negative blood group are not able to synthesize Lewis antigens, and that the FUT2/3 gene status determines Lewis antigen synthesis and secretion [25,36]. In the present study, 9.8% of individuals detected as the Lewis-negative genotype were confirmed as Lewis-null and had undetectable serum sLe a (CA19-9) antigen, unrelated to the Secretor genotype. Moreover, we perceived no Le a , Le b antigen synthase, and 22.5% of individuals detected as the Secretor-negative genotype were perceived as non-secretor and had no Le b antigen synthase [25]. Additionally, 19.3% of participants with the 'single Lewis dominant genotype' (Le, se) were detected with significantly more elevated sLe a than those of the 'simultaneously positive type' (Le, Se) (mean 20 vs 9.42 UÁmL À1 , P < 0.0001). A classic study showed that nine groups divided from 400 normal individuals by Le/Se genotype were detected with discrepant serum CA19-9 and DU-PAN-2 values, and DU-PAN-2 measurement was more useable for colorectal cancer diagnosis in Le-negative patients [25]. Recently, Wannhoff et al. [1] have found that differentiation between three FUT2/3 phenotypes improves the clinical practicability of gene-based cut-off values of CA19-9 for cholangiocarcinoma diagnosis. The application of FUT2/3 genotype-based cut-offs improved sensitivity to 82.4% and 100.0% in the intermediate and high biosynthesis groups, respectively [1]. The indicator sensitivity of Lewis antigens would be significantly increased based on the determination of the FUT2/3 genotype, which may have wide application in tumor diagnosis.
In conclusion, this genotyping study of FUT2 and FUT3 indicated the particular distribution of polymorphisms in the Chinese population. Clarifying individuals' FUT2/3 genotype by sequencing might facilitate clinical classification and accurate diagnosis. Frequencies of each allele and the association with phenotype should be investigated in an extended population and at additional polymorphism positions, and the characteristics of separate subgroups warrants further study.

Supporting information
Additional Supporting Information may be found online in the supporting information tab for this article: Table S1. Primers sequence for PCR amplification of the FUT3 (Lewis) and FUT2 (Secretor) genes. Table S2. Particular nucleic acid sites variation of Lewis, Secretor and combined genotypes. Table S3. Comparison of Lewis and Secretor allele frequencies among different populations.