Jr^a, a high-prevalence blood group antigen, was first reported in 1970 by Stroup and Macllroy [1]. Anti-Jr^a usually does not cause severe transfusion reactions or hemolytic disease of the fetus and neonate, however, the antibody may rarely develop fatal hemolysis and intrauterine fetal death [2],[3]. Revealed to be a truncated product of ATP-binding cassette, membrane G₂ (ABCG₂) gene [4],[5] in 2012, the Jr^a antigen was assigned to the 32nd blood group system, JR (ISBT 032), by the International Society of Blood Transfusion in 2012 [6]. Among Japanese blood donors, the frequency of Jr(a–) of around 0.06–0.07% is more frequent compared to other ethnic populations [7],[8]. The Jr(a–) phenotype is caused by any null alleles of the ABCG₂ gene, 20 or more of which have been reported [9]. As for the Japanese population, the Jr(a–) phenotype is brought by mainly (>70%) homozygotes of ABCG₂*01N.01 (c.376C>T, p.Gln126X), while the remaining <30% is brought by heterozygotes ABCG₂*01N.01 and other ABCG₂ null allele [10],[11],[12].

Moreover, the presence of ABCG₂*01W.01 (c.421C>A, p.Gln141Lys) diminishes Jr^a antigen expression on red cells of individuals with the Jr(a+) phenotype [13]. Thus, Jr^a expression on RBCs should be suppressed in individuals with heterozygotes of ABCG₂*01N.01 and ABCG₂*01W.01. On the other hand, we have suspected that another genetic variance c.-262C>T increases the amount of Jr^a of RBCs. This implies that expression of Jr^a is also controlled by unknown mechanisms and may influence pathogenesis associated with anti-Jr^a.

The association between the above alleles and Jr^a expression on RBCs throughout the normal population has rarely been reported. Having a higher frequency of Jr(a–) individuals in the Japanese population than in other ethnic groups, this is important when looking for compatible RBCs for transfusion purposes, since only 0.06% are Jr(a–) among blood donors in Japan [8], but transfusion cells must be Jr(a–) and not Jr(a+^w). Hence, we studied the associations between each of three Jr^a associated alleles, c.-262C/T, c.376C/T, and c.421C/A, and the expression levels of Jr^a on RBCs.

Samples

The blood samples used were from 474 random blood donors who gave written informed consent for research use at the time of blood donation. The study design has been approved by the institutional ethics committee of the Japanese Red Cross Society (#2016-030).

DNA sequence analysis

Genomic DNA was isolated from whole blood with a kit (DNA blood mini kit, QIAamp, Qiagen, Tokyo, Japan). Polymerase-chain reaction (PCR)-sequence specific primers (SSP) were performed to determine the three SNPs, c.-262C/T, c.376C/T, and c.421C/A. Detailed information of primers and conditions of PCRSSP are shown in Figure 1 and Table 1.

RNA secondary structure prediction

RNA secondary structure was estimated by the RNAstructure Web server (http://rna.urmc.rochester. edu/RNAstructureWeb/Servers/Predict1/Predict1.html) using the MaxExpect algorithm, which calculates partition functions for base-pair and single-strand probabilities to find the most probable structure [14].

Jr^a antigen intensity analysis by flow cytometry

The antigen expression intensity was evaluated by flow cytometry with FACSCalibur^TM (Becton Dickinson, Tokyo, Japan) using in-house monoclonal anti-Jr^a (HIRO-133) and fluorescence-conjugated (ab′)2 to human IgG (H+L) (Abliance, Germany) as a second antibody. The amount of Jr^a antigen was expressed as mean fluorescence intensity (MFI).

Statistics

The estimated amount of Jr^a expressed on RBCs by flow cytometry was compared among healthy donors with different genetic variations at three loci of ABCG₂ genes with Microsoft Excel software (Microsoft Excel 2016, Microsoft, USA) using Student’s t-test. Values of p < 0.01 were considered statistically significant.

Of the 474 blood donors, none were found to be c.376T homozygote corresponding to Jr(a–). Instead, the wild type group (Group I) with no variance at three loci (c.-262C, c.376C, and c.421C), a variant group (Group II) at least one locus at either or both c.421C>A and/or c.376C>T, and another variant group (Group III) with mutation at c.-262C>T, regardless of loci c.421 and c.376, were found in 185 (39.0%), 263 (55.5%), and 26 (5.5%) subjects, respectively, as shown in Table 2.

Comparing the Jr^a expression on RBCs of the two variants groups with Group I, the Group II showed significantly (p < 0.01) lower expression by 20–75%. Especially, when c.376C was altered to T, which generates premature termination of the gene, Jr^a was more seriously diminished than one locus SNP at c.421C>A.

Expression of Jr^a of Group III with c.-262C>T had significantly upregulated compared to that Group I or Group II nearly 1.5-fold as shown in Table 2 and Figure 2.

To understand why c.-262C>T increases expression of Jr^a, we estimated the most probable RNA structure of the entire 5' untranslated region of 546 bases (UTR) and evaluated the effect of the c.-262C>T mutation. As shown in Figure 3, a 238-bp segment from c.-455 to c.-218 likely formed a major branch (red box) in the 5' UTR. A single base substitution at c.-262 was predicted to cause a large structure transition by losing a -262C: -244G pair at a stem (indicated by an arrow in Panel A), which may increase the number of unpaired nucleotides (indicated by an arrow in Panel B).

In this study, we found that nearly 40% of blood donors that we have studied had a 20–25% downregulated expression of Jr^a due to a SNP at locus c.421C>A, that may not be a serious obstacle in correctly typing Jr^a using human-derived anti-Jr^a reagents. On the other hand, a variance of c.376C>T alone (4% of donors), or a combination of c.376C>T with c.421C>A (2% of donors) diminished Jr^a antigen expression seriously; the latter expressed Jr^a only 25% of that expressed by the wild type. In this case, it is likely that Jr(a+^w) may be mistyped as Jr(a-) due to its weak signal when weak anti-Jr^a reagents are used. Similarly, patients’ weak anti-Jr^a cannot be detected correctly but falsely judged negative when weaker Jr(a+) cells are used for antibody screening.

Interestingly, we found that 5.5% of Japanese blood donors had a variance of c.-262C>T that upregulated the Jr^a expression significantly by more than 50% of that of wild type. This enhancement was also observed in individuals having c.421C>A. The variance c.-262C>T may compensate the reducing effect of c.421C>A if the ABCG₂ variant encoded by the gene with c.421C>A is functionally equivalent to that with c.421C.

RNA secondary structure was predicted that c.-262 largely influenced the 5' UTR structure. It is tempting to speculate that the open structure formed by c.-262 may recruit the initiation complex for translation or increase the rate of ribosomal scanning, thereby enhancing the frequency of translation.

In the setting of pregnancy of women with anti-Jr^a, the fetuses/newborns born range from being asymptomatic to having life-threatening hydrops; however, the mechanism of pathogenesis and progression to a serious condition have yet to be elucidated [15]. The present study may explain the pathogenesis and offer a solution. Mother with anti-Jr^a must be theoretically ABCG₂ gene null. If her fetus is heterozygous of ABCG₂ null allele and other ABCG₂ gene then the severity of the fetus could possibly depend on the expression level of ABCG₂. Hence, we predict that the fetus with c.-262T may be attacked severely. When the hypothesis is experimentally and clinically established, ABCG₂ genotyping should be helpful in managing fetuses and newborns that are born to mothers with anti-Jr^a.

A SNP c.-262C>T, and two others, c376C>T and/or c.421C>A in ABCG₂ gene, defined up- and downregulated expression of Jr^a, respectively, which may contribute to elucidating the mechanism of the development of severe forms of anti-Jr^a-induced hemolytic disease of the fetuses and newborns.