Molecular genetics and genomics of the ABO blood group system
Review Article

Molecular genetics and genomics of the ABO blood group system

Fumiichiro Yamamoto

Immunohematology and Glycobiology, Josep Carreras Leukaemia Research Institute (IJC), Badalona, Spain

Correspondence to: Fumiichiro Yamamoto. Immunohematology and Glycobiology, Josep Carreras Leukaemia Research Institute (IJC), Badalona, Spain. Email: fyamamoto@carrerasresearch.org.

Abstract: The A and B oligosaccharide antigens of the ABO blood group system are produced from the common precursor, H substance, by enzymatic reactions catalyzed by A and B glycosyltransferases (AT and BT) encoded by functional A and B alleles at the ABO genetic locus, respectively. In 1990, my research team cloned human A, B, and O allelic cDNAs. We then demonstrated this central dogma of ABO and opened a new era of molecular genetics. We identified four amino acid substitutions between AT and BT and inactivating mutations in the O alleles, clarifying the allelic basis of ABO. We became the first to achieve successful ABO genotyping, discriminating between AA and AO genotypes and between BB and BO, which was impossible using immunohematological/serological methods. We also identified mutations in several subgroup alleles and also in the cis-AB and B(A) alleles that specify the expression of the A and B antigens by single alleles. Later, other scientists interested in the ABO system characterized many additional ABO alleles. However, the situation has changed drastically in the last decade, due to rapid advances in next-generation sequencing (NGS) technology, which has allowed the sequencing of several thousand genes and even the entire genome in individual experiments. Genome sequencing has revealed not only the exome but also transcription/translation regulatory elements. RNA sequencing determines which genes and spliced transcripts are expressed. Because more than 500,000 human genomes have been sequenced and deposited in sequence databases, bioinformaticians can retrieve and analyze this data without generating it. Now, in this era of genomics, we can harness the vast sequence information to unravel the molecular mechanisms responsible for important biological phenomena associated with the ABO polymorphism. Two examples are presented in this review: the delineation of the ABO gene evolution in a variety of species and the association of single nucleotide variant (SNV) sites in the ABO gene with diseases and biological parameters through genome-wide association studies (GWAS).

Keywords: Blood group ABO system; ABO genes; A and B glycosyltransferases; A and B oligosaccharide antigens


Received: 10 November 2020; Accepted: 22 February 2021; Published: 30 September 2021.

doi: 10.21037/aob-20-71


The dawn of the molecular genetics of the ABO blood group system

Immunohematology and serology studying erythrocyte antigens and their antibodies, respectively, were the initial fields of ABO research. However, it subsequently expanded into a wide variety of scientific disciplines. Antigens A and B can also be expressed on epithelial cells and endothelial cells, in addition to red blood cells (RBCs), depending on the blood types of the individuals. Consequently, ABO compatibility is also essential in cell/tissue/organ transplantation, as well as blood transfusion. The antigens A and B are oligosaccharides, GalNAcα1-3(Fucα1-2)Gal and Galα1-3(Fucα1-2)Gal, respectively. Therefore, they are subjects of glycobiology. In the late 1950s, a hypothesis was put forward on the biosynthetic pathways of these antigens: A and B alleles at the ABO genetic locus encode A and B transferases (AT and BT), which catalyze the biosynthesis of A and B antigens by transferring an N-acetyl-D-galactosamine (GalNAc) and a galactose (Gal), respectively, to the same substrates, H substances (Fucα1-2Gal), which are present on the RBCs of individuals of group O. Therefore, ABO has become subjects of genetics, biochemistry and enzymology. A and/or B antigens can also be found in secretion, such as saliva and seminal fluid, as well as hair and skin. Biological samples containing A/B antigens can provide crucial evidence in crime scene investigation. In fact, the ABO polymorphism was the most widely used evidence to exclude innocent people from possible suspects before DNA typing became popular in the last decade. Furthermore, the ABO polymorphism is not limited to humans. Consequently, ABO is the subject of research on genetic evolution.

In 1990, 90 years after the discovery of ABO blood groups, Clausen, Hakomori, and I cloned the A, B, and O allelic cDNAs from stomach and colon cancer cell line cells expressing differential ABO phenotypes, and determined their nucleotide sequences and deduced amino acid sequences (1,2). By correlating the sequence differences with the expression of A/B antigens, we were able to demonstrate, what is now called, the Central Dogma of ABO. AT and BT are the same in size with 354 amino acids (353 in alternatively spliced transcripts), but 4 amino acid substitutions differentiate these two enzymes with different sugar specificities. They are arginine (R), glycine (G), leucine (L), and glycine at codons 176, 235, 266, and 268 in AT encoded by A allele (A101), while they are glycine, serine (S), methionine (M), and alanine (A) in BT encoded by B allele (B101). We also identified a single nucleotide inactivating deletion of guanine at nucleotide 261, 261delG, in most O alleles (O01), while some other O alleles (O02) have additional nucleotide substitutions, in addition to 261delG. Using allele-specific differences in nucleotide sequences and restriction fragment length polymorphism (RFLP), we successfully achieved the first ABO genotyping. AA and AO genotypes, as well as BB and BO genotypes, were also discriminated. Later, we also found another type of O allele (O03) lacking 261delG, but containing an inactivating glycine to arginine substitution at codon 268, G268R, and a functionally insignificant arginine to glycine substitution at codon 176, R176G (3). We also identified a single nucleotide deletion (1060delC) at the C-terminus and a substitution of proline to leucine at codon 156 (P156L) in A2 alleles (A201) (4), an aspartic acid to asparagine at codon 291 (D291N) in an A3 allele (A301) (5), a phenylalanine to isoleucine substitution at codon 216, F216I, in an Ax allele (Ax01) (6), and an arginine to tryptophan substitution at codon 352, R352W, in a B3 allele (B301) (5). We also demonstrated that a cis-AB allele (cis-AB01) and a B(A) allele (B(A)01) that specify the expression of the A and B antigens by single alleles encode AT-BT chimeras (6,7). The alleles that we characterized are summarized in Figure 1.

Figure 1 Twelve ABO alleles characterized molecularly in our group. Nucleotide and deduced amino acid sequences were compared, and only differences from the A101 allele are shown. This table was modified from previous publications (8,9).

Gold hunting for additional ABO alleles

When analyzing the sequences of the cDNA clones, we found that several clones contained unspliced intron sequences. Taking advantage of the sequence information, we showed that PCR and DNA sequencing of the amplified fragments could be used for the characterization of the majority of the coding sequence of the ABO gene. In fact, we have shown that the coding sequences of the last two coding exons (exons 6 and 7), which correspond to approximately 80% of the complete coding sequence of the soluble form of AT, could be amplified by PCR. The DNA fragment spanning exon 6 was amplified, using a pair of oligo primers (a sense strand primer in intron 5 and an antisense strand primer in intron 6), while the exon 7 sequence was amplified in two overlapping DNA fragments, using a sense primer in intron 6 and an antisense primer in exon 7, and also using a sense primer in exon 7 and an antisense primer in the 3'-untranslated region. We used these primers to characterize mutations in the alleles of A and B subgroups and mutations in the cis-AB and B(A) alleles described above (3-7). Our characterization of twelve ABO alleles was the beginning of the search for new ABO alleles. Other researchers followed suit, employing the same strategy. Some used the same primers for PCR and characterized additional ABO alleles, while others used primers of their own design, all based on our published sequences. We also found at least one primer that had the same nucleotide sequence as ours, but with a different primer name. Anyway, thanks to the availability of the combined strategy of PCR and sequencing, mutations in dozens of ABO alleles were characterized even before ABO genomic cloning clarified the entire genetic organization of the ABO gene in 1995 (10).

Many researchers contributed to the characterization of these additional ABO alleles. Among them are Olsson and Chester (Sweden), Ogasawara, Hosoi, Suzuki and Fukumori (Japan), Yu and Lin (Taiwan), Yip (China), and Seltsam and Blasczyk (Germany) [see the list of references in review articles (11-14)]. In addition to the ABO system, other blood group systems were also molecularly characterized. Originally, we used an allele nomenclature combining the phenotype with the number showing the order of discovery in parentheses, A1(1) for example to indicate the first A1 allele identified (8). As the number of alleles identified increased, it became apparent that discoveries were often made without the knowledge of others. Along with the use of different nomenclatures by different research teams, this caused a confusion. To alleviate chaos, in 1999 Blumenfeld established the Blood Group Antigen Gene Mutation Database (BGMUT) through an initiative of the Human Genome Variation Society (HGVS) (15), and I was invited to be a curator for the ABO section of the database. Assuming that 99 alleles for individual ABO phenotypes would be sufficient for some time, we employed a nomenclature of alleles per phenotype followed by a 2-digit number showing the order of discovery. The nucleotide and deduced amino acid sequences of the first characterized A1 allele, A101, were used as standards (9,11). In 2006, BGMUT became a part of NCBI’s dbRBC (database Red Blood Cells) resource at NIH. More than 250 ABO alleles were deposited before the unfortunate closure of the database in 2016 due to the passing of Dr. Blumenfeld. Meanwhile, the International Society of Blood Transfusion (ISBT) established a Working Party on Red Cell Immunogenetics and Blood Group Terminology chaired by Storry, and prepared catalogs of blood group antigens and alleles. Olsson used ABO allele terminology, which was similar to ours based on phenotype followed by a number (http://www.isbtweb.org/working-parties/red-cell-immunogenetics-and-blood-group-terminology/).


Functional characterization of ABO mutations

Thanks to the efforts of many, a variety of ABO alleles were molecularly characterized. These alleles include A1, A2, A3, Ael, Aint, Am, Aw, Ax, cis-AB, B(A), B, B3, Bel, Bw, Bx, and O. Different types of mutations were found, including synonymous mutations, missense mutations, nonsense mutations, frameshift mutations, and splicing mutations. Later studies identified mutations in the regulatory elements of transcription. Furthermore, several alleles were found to contain different types of mutations. See review articles (11-14) for individual references to the original articles.

Identifying potential mutations is one thing, but proving their functional importance is another. Therefore, it was necessary to examine the effects of the identified mutations on the appearance of A and/or B antigens. For this evaluation, we developed a functional assay system that employs DNA transfection of AT/BT expression constructs, as well as those containing mutations, and subsequent immunological detection of A/B antigens. We used HeLa cells derived from uterine cancer. We first performed ABO genotyping of these cells and found that Ms. Henrietta Lacks had a group O phenotype based on the presence of 261delG in both alleles of the ABO gene. We also showed that HeLa cells express H substance on the cell surface (16). It was assumed that if a eukaryotic expression construct encodes a protein with AT and/or BT activities, a GalNAc and/or a Gal can be transferred to H substances, and A and/or B antigens can be produced, respectively. And these antigens can be detected immunologically, using anti-A and/or anti-B antibodies.

To examine the feasibility of the approach, we prepared the cDNA expression constructs of A101 (AT) and B101 (BT) in a eukaryotic expression plasmid vector pSG5 (16). We then transfected the DNA of these constructs into HeLa cells and examined the appearance of the A and B antigens. As we anticipated, we observed the expression of A and B antigens, respectively (Figure 2). Once the AT and BT activities of the original AT and BT constructs were confirmed, AT constructs containing 261delG specific for the O01/O02 alleles or R176G and G268R substitutions specific for the O03 allele were prepared, and their AT activity was examined. We did not detect any appearance of A antigens, which showed that these are inactivating mutations (3). Although it was not determined whether both substitutions of R176G and G268R were necessary for inactivation in that experiment, a single G268R substitution was found to be sufficient to abolish AT activity in a later experiment (17). Similarly, we introduced the A201 allele-specific P156L substitution and 1060delC deletion into the A101 construct and observed a decrease in AT activity. Because P156L substitution was also found in some A1 alleles and the introduction of P156L alone did not decrease AT activity, we concluded that 1060delC was responsible for the A2 phenotype (4). We also introduced D291N substitution specific for an A301 allele in AT and observed a decrease in AT activity (8).

Figure 2 Functional analysis of mutations in O, A2, and A3 alleles. The appearance of A and/or B antigens was examined in cells transfected with DNA from the AT, BT and their derivative constructs containing mutations specific to O, A2, and A3 alleles. The word “ND” means “not determined”. This table was modified from our previous publication (8).

In addition to subgroup-specific mutations, we also used this system to analyze differences in sugar specificity. First, we determined which of the four amino acid substitutions are responsible for the differential sugar specificity between AT and BT. Are all the four amino acid substitutions required? Or just a few are enough? We answered these questions. We constructed 14 AT-BT chimeras that are different at those four positions having the amino acid of AT or BT, and examined the sugar specificities of GalNAc/Gal (16). The following results were obtained as shown in Figure 3. When positions 3 and 4 have amino acids of AT (AA), only AT activity was observed. When they were from BT (BB), only BT activity was observed. When they were AB in this order, the amino acid in the 2nd position determined the specificity. Only AT activity was detected when it was from AT (AAB), while weak BT activity was also present, in addition to AT activity, when it was from BT (BAB). And finally, when positions 3 and 4 were BA in this order, strong AT and BT activities were observed. We concluded that amino acids at codons 176 and 235 have no and slight effects, respectively, while the amino acids at codons 266 and 268 are decisive in determining the sugar specificity. Despite our success in evaluating the functional significance of mutations in ABO alleles that we characterized, this approach was not widely used, possibly due to the inability to conduct molecular and cellular biological investigations in immunohematology/serology laboratories. Consequently, the functionality of many of the mutations found in the ABO gene has yet to be determined.

Figure 3 AT/BT specificity/activity of the AT-BT chimeras. The appearance of A and/or B antigens was examined in cells transfected with DNA from the AT, BT, and 14 AT-BT chimeric expression constructs. This table was modified from our previous publication (16).

ABO polymorphism in the era of genomics

In the current ISBT Blood Group Allele Table (ISBT 001 v1.1 171023), 207 ABO blood group alleles are listed, including 84 A, 49 B, 12 cis-AB/B(A), and 62 O. However, recent advances in next-generation sequencing (NGS) have generated sequence data for more than half a million human genomes. If the sequences of two alleles in a genome were counted separately, data could have been generated for more than a million ABO genes. Figure 4 shows the nucleotide and deduced amino acid sequences of the coding region of the reported human ABO gene cDNA (as of August 20, 2020). To prepare this figure, the original (transcript: ABO-203 ENST00000611156.4 in the Ensembl database) was added with three nucleotides and one amino acid surrounding the 261delG. This manipulation was necessary because the reference sequence in the database was from an O allele with the 261delG deletion, and because two nucleotides were mistakenly removed as an intron to fit in the reading frame after the deletion. Different types of single nucleotide variants (SNVs) are shown in different colors.

Figure 4 SNVs in the ABO gene cDNA sequence. SNVs are shown in the O allele transcript (ABO-203 ENST00000611156.4) from the Ensembl database, which was modified to encode AT. Different types of SNVs are highlighted in different colors. SNV, single nucleotide variant.

Many SNVs have been identified within the coding region of the ABO gene. Non-synonymous SNVs that produced a change in the amino acid sequence of the encoded protein were observed at 549 nucleotides out of 1,065 nucleotides in the coding sequence including the termination codon. The number of SNVs is higher if synonymous differences are added. And the number will increase even more if the SNVs found in introns and regulatory regions are added, because the human ABO gene spans over 24,818 base pairs (chromosome 9: 133,250,401–133,275,219) where the majority is non-coding sequence. Assuming that alleles are defined as genes that occupy the same genetic locus but have different sequences, the number of alleles can be in the millions, because they can have different combinations of SNVs. It was clear that the allele nomenclature in use, which is based on numbering, would soon disappear due to a rapid increase in new alleles identified from NGS. In predicting this situation, in 2012 we proposed the use of allele nomenclature that lists SNVs that are different from the nucleotide and amino acid sequences of the standard A101 allele (14).


Predominance of ABO phenotyping over ABO genotyping

The ABO polymorphism is specified by a single genetic locus ABO. This is correct in a general sense. However, it can also be wrong. For example, individuals of the Bombay and para-Bombay phenotypes may be wrongly genotyped to exhibit an apparent A/B/AB phenotype if functional A and/or B alleles are present. This is due to the fact that these individuals lack functional alleles in the FUT1 (H/h) and/or FUT2 (Se/se) genes that encode α1,2-fucosyltransferases to synthesize H substances, the precursor substrates for A and B transferases. Without H substances, even functional AT and BT are unable to synthesize A and B antigens, respectively. Although the incidence of Bombay or para-Bombay individuals is extremely low, this illustrates a potential difficulty in ABO genotyping. In other words, the ABO phenotype can be deduced from the ABO gene sequence determined by ABO genotyping, but it may not be accurate. Even if there are no mutations in the coding region of the ABO gene, mutations may be present in regulatory elements. Even in the presence of functional FUT1/FUT2 and A/B genes, donor nucleotide-sugars (UDP-GalNAc/UDP-Gal) are necessary for the synthesis of A and/or B antigens to occur in the trans-Golgi apparatus.

In addition, the ABO phenotypes of most of the samples, whose genomic sequences were determined, have not been characterized. Furthermore, the functional significance of most of the SNVs that were identified in the ABO gene has not been determined. In this circumstance, the ABO phenotypes deduced from the ABO gene sequences may not be completely correct, although the A, B, or O phenotype can be determined with greater than 99% precision, using SNV data characterized by function. Nonetheless, useful information can be obtained on the possible molecular mechanisms that explain differential specificity/activity. It is also true that DNA typing has been used successfully to treat hemolytic disease of the fetus and newborn (HDFN) and to identify optimally matched donations for patients with pre-existing antibodies or predisposed to alloimmunization. However, a cross-match test is required prior to transfusion to determine whether the donor blood is compatible with the blood of the intended recipient because incorrect determination of ABO phenotypes by genotyping can result in very detrimental clinical outcomes. In conclusion, ABO phenotyping appears to continue to prevail over ABO genotyping.


ABO genes in species other than Homo sapiens

The ABO polymorphism was initially identified in humans. It was later found in other animals. Moor-Jankowski performed an extensive analysis of the ABO polymorphism of primates (18,19). It was found that chimpanzees analyzed were either A or O phenotype, while gorillas were B phenotype. These anthropoid ape species express A/B antigens on RBCs like humans do, but lower primates do not. Still, they can express these antigens in respiratory/digestive epithelium and/or in secretions such as saliva. We determined partial nucleotide sequences and deduced amino acid sequences of the ABO genes from several primate species (20). We observed that the amino acids corresponding to codons 266 and 268 of human AT/BT were conserved in primates, depending on the A/B status. They were leucine (L) and glycine (G) in the A alleles of chimpanzee, orangutan, macaque, and baboon, while they were methionine (M) and alanine (A) in the B alleles of gorilla and baboon. The results confirmed our previous finding from functional assays that demonstrated the importance of these amino acids for differential sugar specificities of AT/BT. Antigens A and/or B are also expressed in animals other than primates (21). Due to its usefulness for animal experimentation, we characterized the mouse ABO gene (22). We showed that the murine gene encodes a cis-AB transferase capable of transferring both GalNAc and Gal in vitro, although A/B expression was very weak in vivo. We also studied the swine AO system (23). Pigs exhibit A or O phenotype. We cloned porcine AT cDNA and demonstrated AT activity. We also showed that the porcine O allele lacks most of the structural gene. It was in contrast to human O alleles that have minimal alterations, such as a single nucleotide deletion (261delG) (2) or an inactivating single amino acid substitution (G268R) (3).

In the era of genomics, genomes of many species of organisms have been sequenced. And this sequence data is available in public databases, such as Ensembl and Genbank, along with annotations and other useful information. We retrieved the nucleotide and deduced amino acid sequences of ABO genes from a variety of species and generated phylogenetic trees to study the evolution of ABO genes (24). In addition, we used our expertise in biochemistry, immunology and glycobiology. We previously demonstrated that codons 266 and 268 of human AT and BT are crucial in determining the sugar specificity of GalNAc and Gal (16,17). We also showed that the corresponding codons are also crucial in determining the sugar specificity of the enzymes encoded by the ABO genes in primates, mice, and pigs (20,22,23,25). Therefore, we constructed several dozen human AT expression constructs that possess the amino acid substitutions at codons 266 to 268, and also 263 to 268 in some, corresponding to codons in the ABO genes of various species. We then examined the specificity and activity. The results are shown in a code table that associates the amino acid sequences with the potential AT/BT specificity/activity (Figure 5) (24). Since amino acids at codon 235 were known to affect specificity/activity, there was no guarantee that the specificity/activity assignment using the table was 100% accurate. However, combined with other information, we hoped to gain new insight. In fact, we made several unexpected discoveries. In humans and primates, excluding primordial primates, A, B, and/or O genes are located at the single genetic locus ABO as alleles. Therefore, the ABO polymorphism of primates is the result of allelism. However, it was shown that there are species that contain multiple ABO genes and/or gene fragments linked in tandem. There are also some species with ABO genes in different chromosomal regions and/or different chromosomes (Figure 6A). There are even some that have non-allelic A and B genes. For example, rats can exhibit AO polymorphism, having a functional A allele or a non-functional O allele at the same genetic locus. However, rats also possess additional B gene sequences at different chromosomal locations.

Figure 5 The amino acid motif—A/B specificity code table. Sugar specificity was determined experimentally for human AT with a variety of amino acid substitutions around codons 266–268 or 263–268 to associate amino acid motifs with A/B specificity. Amino acids are shown in one-letter symbols. This table was taken from our previous publication (24).
Figure 6 Phylogenetic trees of mammalian and bacterial ABO genes. (A) Phylogenetic tree of mammalian ABO genes possessing multiple gene copies. Round and triangle symbols indicate seemingly complete and partial gene sequences. The colors red, green, yellow, and blue indicate the deduced A, B, AB, and O specificities, respectively, while the color black indicates “undetermined”. Tripeptide sequences corresponding to codons 266–268 of human AT/BT are shown in one-letter symbols in parentheses after the species names. (B) Phylogenetic tree of bacterial ABO genes. The A/B specificity is color coded as shown above in (A). These figures were duplicated from our previous publication (24).

We also studied the evolution of bacterial ABO genes. Some strains of bacteria were known to express A, B, AB, or O phenotype based on studies conducted by Springer in the 1960s (26). He is famous for his elucidation of the sensitization by A/B antigens expressed in bacteria in the intestinal flora as a possible acquisition of what are called “natural antibodies” (27). In that experiment he demonstrated that chickens grown under sterile conditions did not develop anti-B antibodies, while chickens fed diets contaminated with the O86 strain of Escherichia coli bacteria that express the B antigen developed anti-B antibodies. Bacterial B and A genes were cloned from Escherichia coli O86 strain and Helicobacter mustelae, respectively, in 2005 and 2008 (28,29). We used the deduced amino acid sequences of the bacterial ABO genes from a few dozen bacterial genome sequences deposited in databases, and generated the phylogenetic trees. We then assigned the specificity of AT/BT based on the motif of the amino acid sequence, using our code table. Bacterial ABO genes formed two separate clusters, one with A genes and the other with B/AB genes, showing vertical transmission (24). However, Bacteroides genes were found in both clusters, implying that horizontal transmission may have occurred in the evolution of the bacterial ABO gene. The results are shown in Figure 6B.


Generation of the species-dependent ABO polymorphism

The repertoire of the ABO polymorphism varies, depending on the species, as mentioned above. Primates that exclude primordials have single gene ABO polymorphism, while several other species, such as rats, have multiple gene ABO polymorphism. We looked for the molecular mechanisms that generated changes in the number of genes and that converted the multigenic polymorphism into unigenic polymorphism. In humans, the ABO gene is mapped together with an evolutionarily related GBGT1 gene on chromosomal region 9q34 (human GRCh38/hg38: 133,250,401–133,275,219 for ABO and 133,152,948–133,163,945 for GBGT1). In some species, functional GBGT1 genes encode Forssman glycolipid synthases (FSs) that catalyze the biosynthesis of Forssman glycolipid antigen (FORS1) of the FORS blood group system (30). However, the human GBGT1 gene is not functional (31), having two inactivating amino acid substitutions G235S and Q296R (32), excluding rare individuals showing the Apae phenotype, who have the R296Q reversion (33,34).

We analyzed the genes in the chromosomal regions neighboring the ABO and GBGT1 genes in other species (35). Two chromosomal fragments were found flanking these genes in different combinations of orientations, indicating that chromosomal translocations/inversions occurred during the evolution of the species (Figure 7). Importantly, the ABO and GBGT1 genes were found at the ends of the chromosomal fragments. Because genetic alterations, such as duplications and deletions, are frequent at the junctions of chromosomal rearrangement, we hypothesized that this may be the molecular force that has driven the species-dependent divergence of the ABO gene. Furthermore, we observed that the orientations of these two chromosome fragments are the same in primates, excluding Otolemur garnettii. Closer examination showed that lipocalin genes (LCN1/3/4) are located on both sides of the ABO gene in most primate species (Figure 8). We therefore reasoned that the recombination events flanking these genes may have played an important role in the establishment of ABO alleles in primates from non-allelic ABO genes, and that the ABO polymorphism has been inherited in a trans-species manner since then.

Figure 7 Orientations of two chromosomal fragments. The blue and orange arrows indicate the directions of two chromosomal fragments. Various combinations were observed, possibly due to translocations/inversions that occurred during the evolution of the species. The ABO and GBGT1 genes are found at the ends of the fragments. The figure was duplicated from our previous publication (35).
Figure 8 Lipocalin genes (LCN1/3/4) neighboring the ABO genes in many primates. The chromosomal regions surrounding the ABO and GBGT1 genes are shown. The LCN1/3/4 gene sequences are indicated with purple asterisks. The figure was duplicated from our previous publication (35).

ABO association with diseases observed by genome-wide association studies (GWAS)

ABO-incompatible blood transfusion can lead to hemagglutination and lysis of RBCs, kidney failure, and occasional death of the recipients. ABO-incompatible cell/tissue/organ transplantation can lead to acute rejection. However, these are reactions against artificial medical practices. Although ABO-incompatible pregnancy can cause HDFN, symptoms are mild and do not generally require any treatment. Since the discovery of ABO, dozens of diseases have been associated with the ABO polymorphism (36). In most of these studies, statistics were used to assess the significance of the association, comparing the ABO distributions between the diseased population and the corresponding healthy population. However, the selection of healthy population may have the potential to introduce bias. After the Human Genome Project determined the nucleotide sequence of the first human genome in 2003, many human genomes have been sequenced. This resulted in the identification of hundreds to thousands of SNVs in the ABO gene as mentioned above, as well as millions of SNVs throughout the entire genome. Recent advances in genomics have made it possible to conduct GWAS that examine the association of numerous SNVs distributed in more than 25,000 genes throughout the human genome with specific diseases and/or biological parameters. In contrast to the ABO gene-driven approach whereby the association between ABO polymorphism and a given disease is investigated, GWAS identifies anonymous SNVs with high association and is therefore more objective with less chance of introducing bias.

Figure 9 shows some representative diseases and biological parameters that were found associated with SNVs in the ABO gene. Diseases include pancreatic cancer, venous thromboembolism (VTE), coronary artery disease (CAD), severe cerebral malaria, Graves’ disease, stomach/duodenal ulcers, type 2 diabetes, and coronavirus disease 2019 (COVID-19). Serum levels of more than a dozen soluble glycoproteins have been associated with SNVs in the ABO gene, including von Willebrand factor (vWF) and coagulation factor VIII (FVIII) (37), intercellular adhesion molecule 1 (ICAM-1) (38), tumor necrosis factor alpha (TNF-alpha) (although assay specific) (39), alkaline phosphatase (40,41), E-selectin (42,43), P-selectin (44) and angiotensin-converting enzyme (45).

Figure 9 Diseases and biological parameters associated with SNVs in the ABO gene by GWAS. Representative diseases and biological parameters found associated by GWAS with SNVs in the ABO gene are shown, along with the odds ratio (OR), β-coefficient and the P values. CEA, carcinoembryonic antigen; vWF, von Willebrand factor; FVIII, factor VIII; EA, European ancestry; RBCC, red blood cell count; TSH, thyroid-stimulating hormone; Ht, hematocrit; ICAM-1, intercellular adhesion molecule 1; ALP, alkaline phosphatase; TNFA, tumor necrosis factor alpha; IL-6, interleukin 6; Hb, hemoglobin; SNVs, single nucleotide variants; GWAS, genome-wide association studies.

The association of ABO polymorphism to disease can be direct or indirect. The latter may be mediated by differential concentrations of factors. In this review, only cardiovascular diseases including VTE, malaria, ulcer and COVID-19 are detailed below, restricted to diseases whose molecular mechanisms of the ABO association have been determined in some way by functional analysis.


A differential plasma level of coagulation FVIII causing differential susceptibility to VTE

The ABO locus was found associated with VTE and CAD by GWAS (46-48). In the VTE study, the authors analyzed 317,000 SNVs in 453 VTE cases and 1,327 controls, and identified three SNVs with significant association (P value below 1.7×10−7). One SNV was located at the coagulation factor V (FV) locus (P=8.1×10−10) and the other two (rs505922 and rs657152) were located at the ABO locus (P=1.5×10−14 and 2.2×10−13, respectively) (46). ABO genotyping of additional 1,700 VTE cases and 1,400 controls was performed, analyzing three additional SNVs that specify A2 (rs8176750), B (rs8176746), and O (rs8176719) alleles, which showed that individuals in groups O and A2 have a lower risk of VTE. The ABO gene has also been associated with myocardial infarction in the presence of coronary atherosclerosis (47). Eleven SNVs with the highest associations were mapped at the ABO locus. The odds ratio (OR) for A/B/AB vs. O was 1.44. A meta-analysis of 14 GWAS studies of CAD comprising 22,233 cases and 64,762 controls of European descent, followed by genotyping of top association signals in 56,682 additional individuals, identified SNV rs579459 at the ABO locus as having the 5th highest association with an OR of 1.1 (48).

The different rates of synthesis, secretion and clearance of vWF lead to differential plasma levels of vWF. Because FVIII is stable in blood only when it binds to vWF, the plasma concentration of vWF dictates the plasma concentration of FVIII. And the major determinant of the plasma levels of these glycoproteins is the ABO polymorphism, and individuals in group O have levels approximately 25% lower than those in group A (49). Consequently, the molecular cause of the ABO association with cardiovascular disease can be explained, at least partially, by ABO-dependent differential plasma concentrations of vWF and FVIII. A high level of plasma FVIII may result in an increased risk of ischemic heart disease and VTE in the non-O groups, while a low level may cause excessive bleeding in the O group (50,51). In addition to vWF and FVIII, some other factors showing ABO dependence may also be involved in susceptibility to cardiovascular disease.


Stronger adhesion to capillaries of group A erythrocytes infected with Plasmodium falciparum parasites than group O erythrocytes infected with parasites

In 1967, Athreya and Coriell reported that group B confers a selective advantage over malaria infection (52). Since then, many articles have been published linking ABO phenotypes with susceptibility to malaria. It was even proposed that malaria had played an important role in shaping the current distribution of ABO polymorphism in the world (53). However, the association was not evident in some studies. Critically analyzing the literature that reported the association/non-association between ABO and Plasmodium falciparum malaria, Cserti and Dzik showed that individuals in group O tend to exhibit a favorable outcome compared to individuals in group A (54). In 2008, using SNVs rs8176719 (A/B vs. O) and rs8176746 (A/O vs. B), the association between severe malaria and SNVs at the ABO locus was examined (55). The A/B alleles that produce functional transferases were associated with an increased risk of severe malaria compared to the O allele with 261delG (OR =1.18, P=2×10−7). The following year, a GWAS study on severe malaria was published (56). In that study, 19 genetic loci exhibiting a significant association with the threshold of P<10−4 were identified, including β-hemoglobin gene (HBB). The signal peak in HBB coincided with the position of the causative variant of S-hemoglobin (HbS). In that study, none of the SNVs in the ABO gene were associated with malaria. However, later studies associated severe malaria with the ABO gene (57-60).

RBCs infected with P. falciparum roll on and adhere to the microvascular endothelium and then disappear from the circulation (sequestration) to form aggregates (rosettes) with uninfected RBCs and/or platelets. By obstructing blood microcirculation and reducing the supply of oxygen and substrates, rosette formation results in cerebral malaria (61). There is experimental evidence showing that these processes depend on ABO polymorphism. A higher binding affinity of non-O group erythrocytes than O group is reflected in larger rosettes (62). The terminal mono- and tri-saccharides of A and B antigens, H disaccharide, and fucose specifically inhibited rosettes formation of group A/B RBCs infected with the parasites. Selective enzymatic digestion of the A antigen from the surface of uninfected RBCs totally eliminated the parasite’s preference to form rosettes with these RBCs (63,64). The P. falciparum erythrocyte membrane protein 1 (PfEMP1) encoded by the parasite gene mediates this adhesion to interact with a repertoire of human host proteins. The semi-conserved head structure of the parasite protein was shown to bind A antigen (65).


Greater susceptibility of group O individuals to ulcers by more abundant Lewis b expressed on the duodenal epithelium

GWAS identified two susceptibility loci: one in the PSCA gene encoding the prostate stem cell antigen in chromosomal region 8q24 and the other in the ABO gene at 9q34, analyzing a total of 7,035 individuals with duodenal ulcer and 25,323 controls in Japan (66). SNV rs505922 in ABO was associated with duodenal ulcer in a recessive model (OR =1.32; P=1.15×10−10). In fact, the ABO and ulcer association was one of the first associations identified using a targeted approach, in which individuals in group O were shown to have a higher susceptibility (67). The relative incidences reported were 0.73 (A vs. O) and 0.80 (B vs. O) for duodenal ulcers and 0.87 for gastric ulcers (both A vs. O and B vs. O). In 1984, Marshall demonstrated that infection with the bacterium Helicobacter pylori causes gastritis and ulceration of the stomach/duodenum and that patients can be cured of peptic ulcer by eradicating the bacteria with antibiotics and acid secretion inhibitors (68). Subsequently, the relative risk of non-secretors/secretors was calculated to be 1.9 (69). In 1993, Boren demonstrated that fucosylated H type 1 and Lewis b (Leb) antigens mediate the attachment of H. pylori to the human gastric mucosa and that the soluble glycoproteins that present Leb or antibodies against the Leb antigen inhibit bacterial binding (70). Furthermore, the conversion of Leb to ALeb by the addition of a terminal GalNAc by AT decreased bacterial binding. Based on the observation, the lower availability of Leb receptors in non-O groups individuals was assumed to be responsible for the reduced infectivity. However, subsequent studies showed that different H. pylori strains exhibit different sugar specificity, complicating understanding of the interactions between host blood group antigens and bacterial blood group antigen-binding adhesion (BabA). In fact, it was later discovered that more than 95% of the strains bound to fucosylated blood group antigens without showing a preference for group O. Instead, these generalists bind to A and B antigens, as well as H substance. Only 5% of the strains specifically bind to H substances. However, these specialists occupy a higher percentage (60%) among the Amerindian strains of South America (71). The specialization of H. pylori appears to have coincided with the unique predominance of group O in Amerindians. Selection cycles to increase or decrease bacterial adherence were proposed to contribute to BabA diversity and to replace generalists by specialists gradually. In other words, positive selection for pathogens due to the abundance of a specific host population group was suggested, as opposed to negative selection for a specific host population group caused by infectious agents.


ABO blood groups and SARS-CoV-2 infectivity/COVID-19 progression

Unfortunately, we are in the midst of a COVID-19 pandemic caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). Although most infected people have a few or no symptoms, some progress to severe pneumonia, multiple organ failure, and even death (72). Older people are at high risk, and eighty percent of deaths have occurred in people with at least one underlying comorbidity, particularly cardiovascular disease/hypertension, overweight/obesity, and diabetes (73).

A GWAS study of COVID-19 has recently been published (74). More than 8 million SNVs of 835 patients with respiratory failure due to severe COVID-19 and 1,255 control participants from Italy and 775 patients and 950 control participants from Spain were analyzed. Two chromosomal regions with significant associations were identified at 3p21.31 and 9q34.2. Furthermore, the frequency of risk alleles for lead variants was higher in the more severely ill and mechanically ventilated patients compared to the less severely ill patients who received only supplemental oxygen. The association in 9q34.2 was mapped to the ABO gene. A higher and lower risk was calculated for individuals in group A (OR =1.45, P=1.48×10−4) and group O (OR =0.65, P=1.06×10−5) by group-specific analysis. A more recent GWAS study of critical illness in COVID-19 has identified several additional genetic loci with a significant association (75).

Prior to the GWAS work, the ABO association to COVID-19 was reported in a couple of manuscripts posted on medRxiv, the preprint server for health sciences (76-78). In the article recently published in the British Journal of Hematology, Li compared ABO blood group distribution among 265 SARS-CoV-2 infected patients and 3,694 healthy controls. A significantly higher proportion of group A individuals was found in patients than in healthy controls (39.3% vs. 32.3%, P=0.017), while the proportion of group O individuals was significantly lower in patients (25.7% vs. 33.8%, P<0.01) (79). This trend was also observed in other studies (75-78,80,81). Furthermore, genetic testing company 23andMe posted online preliminary unpublished data from its ongoing COVID-19 study (82). It analyzed the genetic and survey data of more than 750,000 participants, and calculated the percentages of individuals with different ABO groups who reported COVID-19: 1.3%, 1.4%, 1.5%, and 1.5% for groups O, A, B, and AB, respectively, among all participants, and 3.2%, 3.9%, 4.0% and 4.1% among health professionals. A statistically significant protective effect of group O was observed (OR =0.86, P<0.0001 against acquisition, OR =0.81, P=0.05 hospitalization) in the entire population and also among health professionals (OR =0.81, P<0.0001 against acquisition). This new approach based on mass genome analysis and survey questions has proven very powerful in identifying significant genetic associations.


Inhibition of SARS-CoV-2 infection by natural antibodies

SARS-CoV-2 exhibits a broad organ tropism, infecting and proliferating in the epithelial cells of the respiratory and digestive tracts where the A and/or B antigens are expressed depending on the ABO phenotype of the individual. SARS-CoV-2 is membrane encapsulated and its infection in human cells is mediated by the binding of viral Spike (S) glycoproteins embedded in the membrane with angiotensin-converting enzyme 2 (ACE2) receptors present on the cell surface. S proteins can carry A and/or B glycan antigens, reflecting the ABO phenotype of the cells where viruses are produced. In an experimental cellular model of SARS (and not SARS-CoV-2), mouse monoclonal or human polyclonal anti-A antibodies were shown to partially inhibit the physical interaction between viral S proteins carrying A antigens and cellular ACE2 proteins (83). Taking into account that the ABO blood group polymorphism was shown to influence susceptibility to SARS (highest and lowest risk for individuals in groups A and O, respectively) (84), it was suggested that this molecular mechanism is responsible for the ABO-dependent differential susceptibility to SARS. In fact, similar experimental observations have been reported of HIV and measles viruses expressing A or B antigens (85,86).

The infectivity of SARS-CoV-2 among individuals with various ABO phenotypes is shown schematically in Figure 10. On the one hand, viruses produced in individuals of groups A, B, AB, and O can express A, B, A and B antigens, and none, respectively. On the other hand, individuals in groups A, B, AB, and O have anti-B, anti-A, none, and anti-A/anti-B/anti-A,B antibodies, respectively, following Landsteiner’s law. Thus, like “matched” and “mismatched” blood transfusion, these antibodies are expected to react with corresponding antigens on viral particles and inhibit, at least partially, interpersonal infection in certain combinations (88). For example, SARS-CoV-2 viruses produced in cells of group A individuals (group A viruses) can express A antigens and infect group A or AB individuals without such antigen-antibody reactions, as shown by red arrows. However, infection of group B or O individuals who possess anti-A antibodies may be somewhat inhibited (as shown by black dotted arrows). The important thing here is that the inhibition is directional and that it may or may not be 100% efficient. Therefore, infection can occur even if the corresponding antibodies are present. More importantly, once infection is established, newly produced SARS-CoV-2 viruses exhibit the same ABO phenotype as the infected individual, and these antibodies no longer inactivate them. In other words, natural antibodies are only relevant for the prevention of initial attacks, but irrelevant for the subsequent productive infection. Ironically, more protected group O individuals can produce group O SARS-CoV-2 viruses that are capable of effectively infecting individuals with any ABO phenotypes, including group O, without the protection of natural antibodies. In other words, having the group O phenotype does not protect against group O viruses. This protection works better in the ABO-heterogeneous population than in ABO-homogeneous populations. As a result, countries with the highest frequency of O phenotype, such as Ecuador (75%) and Peru (70%), also suffer from the COVID-19 pandemic. Natural antibodies of the IgA class may be primarily responsible for mucosal immunity, although natural antibodies of other classes, especially anti-A,B IgG, may also function. The latter class of antibodies may explain a greater protective effect reported in group O than in group B (P<0.001) (89).

Figure 10 ABO-dependent differential inhibition of SARS-CoV-2 infection. With and without inhibition of the SARS-CoV-2 infection by natural antibodies is shown schematically with dotted black and solid red arrows, respectively. This figure was modified from our previous publication (87).

In addition to natural antibodies that affect the infectivity of SARS-CoV-2, the ABO polymorphism can indirectly affect the progression of COVID-19. As mentioned above, individuals in group O have 25% lower plasma levels of vWF and FVIII, and have a lower risk of thrombosis, pulmonary embolism, and VTE (46). Most of these glycoproteins are produced in vascular endothelial cells where A and/or B antigens are expressed, depending on the ABO phenotype of the individual. Because severe COVID-19 can dysregulate vascular tone and permeability and induce cytokine storms and redox stress, the ABO polymorphism may influence COVID-19 differently. However, the progression of the disease also depends on other factors, several of which are more relevant than the ABO polymorphism, with much higher ORs. For example, the ORs for people over the age of 85 are 13 and 630 for hospitalization and death, respectively, compared to the young adults aged between 18 and 29 years (73). In this circumstance, the effects of ABO on disease severity and mortality can be easily masked. Recently, a couple of articles have been published that reported a question about the association between ABO and COVID-19 (90,91). For example, Boudin compared ABO blood group distributions in 1,279 crew members confirmed/suspected of being infected with SARS-CoV-2 and 409 crew members who were exposed but not infected at the same time and in the same location on a French aircraft carrier (92,93). Univariate analysis did not show a statistically significant association between ABO polymorphism and SARS-CoV-2 infection, although a slightly higher distribution was observed for group A (40.7% vs. 37.4%) and lower for group O (43.2% vs. 46.2%) in the infected population. It is one thing that the ABO blood group polymorphism inhibits SARS-CoV-2 infection and affects the progression of COVID-19. Another is whether or not these effects are actually observed in a population.


Conclusions

Recent advances in molecular genetic/genomic analysis of the ABO system have answered many important biological and medical questions. These studies have also contributed to the development of ABO genotyping and the unbiased identification of diseases associated with SNVs in the ABO gene by GWAS, which are beneficial for clinical work. There are still a variety of problems to solve, as well as technical and economical limitations to overcome. However, future studies are expected to open a novel venue in our better understanding of the ABO blood group system and useful clinical translations.


Acknowledgments

The author acknowledges the efforts of many who have contributed to the advanced understanding of the molecular genetic/genomic basis of the ABO blood group system. He apologizes for not citing many relevant studies due to space limitations.

Funding: None.


Footnote

Provenance and Peer Review: This article was commissioned by the Guest Editor (Yann Fichou) for the series “Molecular Genetics and Genomics of Blood Group Systems” published in Annals of Blood. The article has undergone external peer review.

Conflicts of Interest: The author has completed the ICMJE uniform disclosure form (available at http://dx.doi.org/10.21037/aob-20-71). The series “Molecular Genetics and Genomics of Blood Group Systems” was commissioned by the editorial office without any funding or sponsorship. The author has no other conflicts of interest to declare.

Ethical Statement: The author is accountable for all aspects of the manuscript and ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.


References

  1. Yamamoto F, Marken J, Tsuji T, et al. Cloning and characterization of DNA complementary to human UDP-GalNAc: Fuc alpha 1->2Gal alpha 1->3GalNAc transferase (histo-blood group A transferase) mRNA. J Biol Chem 1990;265:1146-51. [Crossref] [PubMed]
  2. Yamamoto F, Clausen H, White T, et al. Molecular genetic basis of the histo-blood group ABO system. Nature 1990;345:229-33. [Crossref] [PubMed]
  3. Yamamoto F, McNeill PD, Yamamoto M, et al. Molecular genetic analysis of the ABO blood group system: 4. Another type of O allele. Vox Sang 1993;64:175-8. [Crossref] [PubMed]
  4. Yamamoto F, McNeill PD, Hakomori S. Human histo-blood group A2 allele, one of the A subtypes, is characterized by a single base deletion in the coding sequence, which results in an additional domain at the carboxyl terminal. Biochem Biophys Res Commun 1992;187:366-74. [Crossref] [PubMed]
  5. Yamamoto F, McNeill PD, Yamamoto M, et al. Molecular genetic analysis of the ABO blood group system: 1. Weak subgroups: A3 and B3 alleles. Vox Sang 1993;64:116-9. [Crossref] [PubMed]
  6. Yamamoto F, McNeill PD, Yamamoto M, et al. Molecular genetic analysis of the ABO blood group system: 3. A(X) and B(A) alleles. Vox Sang 1993;64:171-4. [Crossref] [PubMed]
  7. Yamamoto F, McNeill PD, Kominato Y, et al. Molecular genetic analysis of the ABO blood group system: 2. cis-AB alleles. Vox Sang 1993;64:120-3. [Crossref] [PubMed]
  8. Yamamoto F. Molecular genetics of the ABO histo-blood group system. Vox Sang 1995;69:1-7. [Crossref] [PubMed]
  9. Yamamoto F. Molecular genetics of ABO. Vox Sang 2000;78:91-103. [PubMed]
  10. Yamamoto F, McNeill PD, Hakomori S. Genomic organization of human histo-blood group ABO genes. Glycobiology 1995;5:51-8. [Crossref] [PubMed]
  11. Yamamoto F. Cloning and regulation of the ABO genes. Transfus Med 2001;11:281-94. [Crossref] [PubMed]
  12. Yip SP. Sequence variation at the human ABO locus. Ann Hum Genet 2002;66:1-27. [Crossref] [PubMed]
  13. Yamamoto F. Review: ABO blood group system--ABH oligosaccharide antigens, anti-A and anti-B, A and B glycosyltransferases, and ABO genes. Immunohematology 2004;20:3-22. [Crossref] [PubMed]
  14. Yamamoto F, Cid E, Yamamoto M, et al. ABO research in the modern era of genomics. Transfus Med Rev 2012;26:103-18. [Crossref] [PubMed]
  15. Blumenfeld OO. Mutation databases and other online sites as a resource for transfusion medicine: history and attributes. Transfus Med Rev 2002;16:103-14. [Crossref] [PubMed]
  16. Yamamoto F, Hakomori S. Sugar-nucleotide donor specificity of histo-blood group A and B transferases is based on amino acid substitutions. J Biol Chem 1990;265:19257-62. [Crossref] [PubMed]
  17. Yamamoto F, McNeill PD. Amino acid residue at codon 268 determines both activity and nucleotide-sugar donor substrate specificity of human histo-blood group A and B transferases: In vitro mutagenesis study. J Biol Chem 1996;271:10515-20. [Crossref] [PubMed]
  18. Moor-Jankowski J, Wiener AS, Rogers CM. Human blood group factors in non-human primates. Nature 1964;202:663-5. [Crossref] [PubMed]
  19. Moor-Jankowski J, Wiener AS. Blood group antigens in primate animals and their relation to human blood groups. Primates Med 1969;3:64-77. [PubMed]
  20. Kominato Y, McNeill PD, Yamamoto M, et al. Animal histo-blood group ABO genes. Biochem Biophys Res Commun 1992;189:154-64. [Crossref] [PubMed]
  21. Oriol R, Candelier JJ, Taniguchi S, et al. Major carbohydrate epitopes in tissues of domestic and African wild animals of potential interest for xenotransplantation research. Xenotransplantation 1999;6:79-89. [Crossref] [PubMed]
  22. Yamamoto M, Lin XH, Kominato Y, et al. Murine equivalent of the human histo-blood group ABO gene is a cis-AB gene and encodes a glycosyltransferase with both A and B transferase activity. J Biol Chem 2001;276:13701-8. [Crossref] [PubMed]
  23. Yamamoto F, Yamamoto M. Molecular genetic basis of porcine histo-blood group AO system. Blood 2001;97:3308-10. [Crossref] [PubMed]
  24. Yamamoto F, Cid E, Yamamoto M, et al. An integrative evolution theory of histo-blood group ABO and related genes. Sci Rep 2014;4:6601. [Crossref] [PubMed]
  25. Saitou N, Yamamoto F. Evolution of primate ABO blood group genes and their homologous genes. Mol Biol Evol 1997;14:399-411. [Crossref] [PubMed]
  26. Springer GF, Williamson P, Brandes WC. Blood group activity of Gram-negative bacteria. J Exp Med 1961;113:1077-93. [Crossref] [PubMed]
  27. Springer GF, Horton RE. Blood group isoantibody stimulation in man by feeding blood group-active bacteria. J Clin Invest 1969;48:1280-91. [Crossref] [PubMed]
  28. Yi W, Shao J, Zhu L, et al. Escherichia coli O86 O-antigen biosynthetic gene cluster and stepwise enzymatic synthesis of human blood group B antigen tetrasaccharide. J Am Chem Soc 2005;127:2040-1. [Crossref] [PubMed]
  29. Yi W, Shen J, Zhou G, et al. Bacterial homologue of human blood group A transferase. J Am Chem Soc 2008;130:14420-1. [Crossref] [PubMed]
  30. Haslam DB, Baenziger JU. Expression cloning of Forssman glycolipid synthetase: a novel member of the histo-blood group ABO gene family. Proc Natl Acad Sci U S A 1996;93:10697-702. [Crossref] [PubMed]
  31. Xu H, Storch T, Yu M, et al. Characterization of the human Forssman synthetase gene. An evolving association between glycolipid synthesis and host-microbial interactions. J Biol Chem 1999;274:29390-8. [Crossref] [PubMed]
  32. Yamamoto M, Cid E, Yamamoto F. Molecular genetic basis of the human Forssman glycolipid antigen negativity. Sci Rep 2012;2:975. [Crossref] [PubMed]
  33. Svensson L, Hult AK, Stamps R, et al. Forssman expression on human erythrocytes: biochemical and genetic evidence of a new histo-blood group system. Blood 2013;121:1459-68. [Crossref] [PubMed]
  34. Hult AK, McSherry E, Moller M, et al. GBGT1 is allelically diverse but dispensable in humans and naturally occurring anti-FORS1 shows an ABO-restricted pattern. Transfusion 2018;58:2036-45. [Crossref] [PubMed]
  35. Yamamoto F. Evolutionary divergence of the ABO and GBGT1 genes specifying the ABO and FORS blood group systems through chromosomal rearrangements. Sci Rep 2017;7:9375. [Crossref] [PubMed]
  36. Mourant AE, Kope*c AC, Domaniewska-Sobczak K. Blood groups and diseases: a study of associations of diseases with blood groups and other polymorphisms. Oxford monographs on medical genetics. Oxford; New York: Oxford University Press, 1978.
  37. Smith NL, Chen MH, Dehghan A, et al. Novel Associations of Multiple Genetic Loci With Plasma Levels of Factor VII, Factor VIII, and von Willebrand Factor. Circulation 2010;121:1382-92. [Crossref] [PubMed]
  38. Paré G, Chasman DI, Kellogg M, et al. Novel association of ABO histo-blood group antigen with soluble ICAM-1: results of a genome-wide association study of 6,578 women. PLoS Genet 2008;4:e1000118 [Crossref] [PubMed]
  39. Melzer D, Perry JR, Hernandez D, et al. A genome-wide association study identifies protein quantitative trait loci (pQTLs). PLoS Genet 2008;4:e1000072 [Crossref] [PubMed]
  40. Yuan X, Waterworth D, Perry JR, et al. Population-based genome-wide association studies reveal six loci influencing plasma levels of liver enzymes. Am J Hum Genet 2008;83:520-8. [Crossref] [PubMed]
  41. Kamatani Y, Matsuda K, Okada Y, et al. Genome-wide association study of hematological and biochemical traits in a Japanese population. Nat Genet 2010;42:210-5. [Crossref] [PubMed]
  42. Paterson AD, Lopes-Virella MF, Waggott D, et al. Genome-wide association identifies the ABO blood group as a major locus associated with serum levels of soluble E-selectin. Arterioscler Thromb Vasc Biol 2009;29:1958-67. [Crossref] [PubMed]
  43. Qi L, Cornelis MC, Kraft P, et al. Genetic variants in ABO blood group region, plasma soluble E-selectin levels and risk of type 2 diabetes. Hum Mol Genet 2010;19:1856-62. [Crossref] [PubMed]
  44. Barbalic M, Dupuis J, Dehghan A, et al. Large-scale genomic studies reveal central role of ABO in sP-selectin and sICAM-1 levels. Hum Mol Genet 2010;19:1863-72. [Crossref] [PubMed]
  45. Chung CM, Wang RY, Chen JW, et al. A genome-wide association study identifies new loci for ACE activity: potential implications for response to ACE inhibitor. Pharmacogenomics J 2010;10:537-44. [Crossref] [PubMed]
  46. Trégouët DA, Heath S, Saut N, et al. Common susceptibility alleles are unlikely to contribute as strongly as the FV and ABO loci to VTE risk: results from a GWAS approach. Blood 2009;113:5298-303. [Crossref] [PubMed]
  47. Reilly MP, Li M, He J, et al. Identification of ADAMTS7 as a novel locus for coronary atherosclerosis and association of ABO with myocardial infarction in the presence of coronary atherosclerosis: two genome-wide association studies. Lancet 2011;377:383-92. [Crossref] [PubMed]
  48. Schunkert H, Konig IR, Kathiresan S, et al. Large-scale association analysis identifies 13 new susceptibility loci for coronary artery disease. Nat Genet 2011;43:333-8. [Crossref] [PubMed]
  49. Gill JC, Endres-Brooks J, Bauer PJ, et al. The effect of ABO blood group on the diagnosis of von Willebrand disease. Blood 1987;69:1691-5. [Crossref] [PubMed]
  50. O'Donnell J, Laffan MA. The relationship between ABO histo-blood group, factor VIII and von Willebrand factor. Transfus Med 2001;11:343-51. [Crossref] [PubMed]
  51. Jenkins PV, O'Donnell JS. ABO blood group determines plasma von Willebrand factor levels: a biologic function after all? Transfusion 2006;46:1836-44. [Crossref] [PubMed]
  52. Athreya BH, Coriell LL. Relation of blood groups to infection. I. A survey and review of data suggesting possible relationship between malaria and blood groups. Am J Epidemiol 1967;86:292-304. [Crossref] [PubMed]
  53. Weatherall DJ. Genetic variation and susceptibility to infection: the red cell and malaria. Br J Haematol 2008;141:276-86. [Crossref] [PubMed]
  54. Cserti CM, Dzik WH. The ABO blood group system and Plasmodium falciparum malaria. Blood 2007;110:2250-8. [Crossref] [PubMed]
  55. Fry AE, Griffiths MJ, Auburn S, et al. Common variation in the ABO glycosyltransferase is associated with susceptibility to severe Plasmodium falciparum malaria. Hum Mol Genet 2008;17:567-76. [Crossref] [PubMed]
  56. Jallow M, Teo YY, Small KS, et al. Genome-wide and fine-resolution association analysis of malaria in West Africa. Nat Genet 2009;41:657-65. [Crossref] [PubMed]
  57. Timmann C, Thye T, Vens M, et al. Genome-wide association study indicates two novel resistance loci for severe malaria. Nature 2012;489:443-6. [Crossref] [PubMed]
  58. Band G, Le QS, Jostins L, et al. Imputation-based meta-analysis of severe malaria in three African populations. PLoS Genet 2013;9:e1003509 Erratum in: PLoS Genet 2013 Jun;9(6). doi:10.1371/annotation/adc2beaf-4bee-4e22-925b-6788d62fe029. Malaria Genomic Epidemiological Network [corrected to Malaria Genomic Epidemiology Network]. [Crossref] [PubMed]
  59. Malaria Genomic Epidemiology Network. A novel locus of resistance to severe malaria in a region of ancient balancing selection. Nature 2015;526:253-7. [Crossref] [PubMed]
  60. Malaria Genomic Epidemiology Network. Insights into malaria susceptibility using genome-wide data on 17,000 individuals from Africa, Asia and Oceania. Nat Commun 2019;10:5732. [Crossref] [PubMed]
  61. White NJ, Ho M. The pathophysiology of malaria. Adv Parasitol 1992;31:83-173. [Crossref] [PubMed]
  62. Carlson J, Wahlgren M. Plasmodium falciparum erythrocyte rosetting is mediated by promiscuous lectin-like interactions. J Exp Med 1992;176:1311-7. [Crossref] [PubMed]
  63. Barragan A, Kremsner PG, Wahlgren M, et al. Blood group A antigen is a coreceptor in Plasmodium falciparum rosetting. Infect Immun 2000;68:2971-5. [Crossref] [PubMed]
  64. Rowe JA, Handel IG, Thera MA, et al. Blood group O protects against severe Plasmodium falciparum malaria through the mechanism of reduced rosetting. Proc Natl Acad Sci U S A 2007;104:17471-6. [Crossref] [PubMed]
  65. Chen Q, Heddini A, Barragan A, et al. The semiconserved head structure of Plasmodium falciparum erythrocyte membrane protein 1 mediates binding to multiple independent host receptors. J Exp Med 2000;192:1-10. [Crossref] [PubMed]
  66. Tanikawa C, Urabe Y, Matsuo K, et al. A genome-wide association study identifies two susceptibility loci for duodenal ulcer in the Japanese population. Nat Genet 2012;44:430-4, S1-2.
  67. Aird I, Bentall HH, Mehigan JA, et al. The blood groups in relation to peptic ulceration and carcinoma of colon, rectum, breast, and bronchus; an association between the ABO groups and peptic ulceration. Br Med J 1954;2:315-21. [Crossref] [PubMed]
  68. Marshall BJ, Warren JR. Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration. Lancet 1984;1:1311-5. [Crossref] [PubMed]
  69. Dickey W, Collins JS, Watson RG, et al. Secretor status and Helicobacter pylori infection are independent risk factors for gastroduodenal disease. Gut 1993;34:351-3. [Crossref] [PubMed]
  70. Borén T, Falk P, Roth KA, et al. Attachment of Helicobacter pylori to human gastric epithelium mediated by blood group antigens. Science 1993;262:1892-5. [Crossref] [PubMed]
  71. Aspholm-Hurtig M, Dailide G, Lahmann M, et al. Functional adaptation of BabA, the H. pylori ABO blood group antigen binding adhesin. Science 2004;305:519-22. [Crossref] [PubMed]
  72. Huang C, Wang Y, Li X, et al. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet 2020;395:497-506. [Crossref] [PubMed]
  73. Prevention CfDCa. Coronavirus Disease 2019 (COVID-19), People at Increased Risk for Severe Illness. Available online: https://wwwcdcgov/coronavirus/2019-ncov/need-extra-precautions/indexhtml (accessed on 20 August 2020) 2020.
  74. Severe Covid-19 GWAS Group, Ellinghaus D, Degenhardt F, et al. Genomewide Association Study of Severe Covid-19 with Respiratory Failure. N Engl J Med 2020;383:1522-34.
  75. Pairo-Castineira E, Clohisey S, Klaric L, et al. Genetic mechanisms of critical illness in COVID-19. Nature 2021;591:92-8. [Crossref] [PubMed]
  76. Zhao J, Yang Y, Huang H, et al. Relationship between the ABO blood group and the COVID-19 susceptibility. Clin Infect Dis 2021;73:328-31. [Crossref] [PubMed]
  77. ZietzMTatonettiNP. Testing the association between blood type and COVID-19 infection, intubation, and death.medRxiv Preprint 2020. doi: .10.1101/2020.04.08.20058073
  78. Zeng X, Fan H, Lu D, et al. Association between ABO blood groups and clinical outcome of coronavirus disease 2019: Evidence from two cohorts. medRxiv Preprint 2020. doi: 10.1101/2020.04.15.20063107.10.1101/2020.04.15.20063107
  79. Li J, Wang X, Chen J, et al. Association between ABO blood groups and risk of SARS-CoV-2 pneumonia. Br J Haematol 2020;190:24-7. [Crossref] [PubMed]
  80. Hoiland RL, Fergusson NA, Mitra AR, et al. The association of ABO blood group with indices of disease severity and multiorgan dysfunction in COVID-19. Blood Adv 2020;4:4981-9. [Crossref] [PubMed]
  81. Barnkob MB, Pottegard A, Støvring H, et al. Reduced prevalence of SARS-CoV-2 infection in ABO blood group O. Blood Adv 2020;4:4990-3. [Crossref] [PubMed]
  82. 23andMe. 23andMe finds evidence that blood type plays a role in COVID-19. 2020. Available online: https://blog.23andme.com/andme-research/andme-finds-evidence-that-blood-type-plays-a-role-in-covid-19/ (accessed on 20 August 2020).
  83. Guillon P, Ruvoen-Clouet N, Le Moullac-Vaidye B, et al. Association between expression of the H histo-blood group antigen, alpha1,2fucosyltransferases polymorphism of wild rabbits, and sensitivity to rabbit hemorrhagic disease virus. Glycobiology 2009;19:21-8. [Crossref] [PubMed]
  84. Cheng Y, Cheng G, Chui CH, et al. ABO blood group and susceptibility to severe acute respiratory syndrome. JAMA 2005;293:1450-1. Erratum in: JAMA 2005 Aug 17;294(7):794. Cheng, Yufeng [corrected to Cheng, Yunfeng]. [PubMed]
  85. Arendrup M, Hansen JE, Clausen H, et al. Antibody to histo-blood group A antigen neutralizes HIV produced by lymphocytes from blood group A donors but not from blood group B or O donors. AIDS 1991;5:441-4. [Crossref] [PubMed]
  86. Preece AF, Strahan KM, Devitt J, et al. Expression of ABO or related antigenic carbohydrates on viral envelopes leads to neutralization in the presence of serum containing specific natural antibodies and complement. Blood 2002;99:2477-82. [Crossref] [PubMed]
  87. Yamamoto F, Yamamoto M, Muñiz‐Diaz E. Blood group ABO polymorphism inhibits SARS‐CoV‐2 infection and affects COVID‐19 progression. Vox Sang 2021;116:15-7. [Crossref] [PubMed]
  88. Yamamoto F. ABO blood groups and SARS-CoV-2 infection. ResearchGate 2020. doi: 10.13140/RG.2.2.24833.20329. Available online: https://www.researchgate.net/publication/340209045_ABO_Blood_Groups_and_SARS-CoV-2_Infection 10.13140/RG.2.2.24833.20329
  89. Gérard C, Maggipinto G, Minon JM. COVID‐19 and ABO blood group: another viewpoint. Br J Haematol 2020;190:e93-4. [Crossref] [PubMed]
  90. Latz CA, DeCarlo C, Boitano L, et al. Blood type and outcomes in patients with COVID-19. Ann Hematol 2020;99:2113-8. [Crossref] [PubMed]
  91. Dzik S, Eliason K, Morris EB, et al. COVID-19 and ABO blood groups. Transfusion 2020;60:1883-4. [Crossref] [PubMed]
  92. Boudin L, Janvier F, Bylicki O, et al. ABO blood groups are not associated with risk of acquiring the SARS-CoV-2 infection in young adults. Haematologica 2020;105:2841-3. [Crossref] [PubMed]
  93. Flegel WA. COVID-19: risk of infection is high, independently of ABO blood group. Haematologica 2020;105:2706-8. [Crossref] [PubMed]
doi: 10.21037/aob-20-71
Cite this article as: Yamamoto F. Molecular genetics and genomics of the ABO blood group system. Ann Blood 2021;6:25.