The Lewis determinants are structurally related to determinants of the ABO and the H/h blood group systems. They are assembled by sequential addition of specific monosaccharides onto terminal saccharide precursor chains on glycolipids or glycoproteins. On the erythrocyte surface they reside on glycolipids. In contrast to the other blood group antigens, the synthesis of these glycolipids does not occur in erythroid tissues, but they are acquired by the erythrocyte membranes from other tissues through circulating lipoproteins.
The ABH and Lewis glycoproteins possess a common basic structure and their blood group specificity is determined by the sequence and linkage. There are two Lewis antigens, termed Le-a and Le-b. The presence of fucose linked to C-4 of N-acetylglucosamine on a Type 1 chain results in Le-a activity, but a Type 2 oligosaccharide containing fucose linked to C-3 of N-acetylglucosamine on a Type 2 chain results in very weak Le-a activity. The appearance of a second fucose on a type one chain results in the appearance of a new antigenic determinant, Le-b, and the loss of most H and Le-a antigenicity. A Type 2 difucosyl chain has very weak Le-b activity.
The synthesis of the epitopes is dependent on the interaction of two different fucosyltransferases, products of two different loci: FUT2 or the secretor (Se) locus of the H/h blood group system that encodes the alpha (1,2) fucosyltransferase (FUT2), and the FUT3 locus that encodes the alpha (1/3,1/4 fucosyltransferase (FUT3). The final oligosacharide products are the result of the action of these enzymes on the same oligosaccharide precursor substrates of type 1 configuration: Gal beta 1-3 GlcNAc beta-R. An important property of FUT3 should be noted - namely, it can use as substrates both type 1 and type 2 carbohydrate chains resulting in either alpha 1,3 or alpha 1,4 linkages (hence, the FUT3 enzyme is also known as alpha 3/4-Fuc-T).
As stated above, Lewis (Le) a and b epitopes are both products of substrates in type 1 configuration, whereas Le X and Y are the equivalent products when the substrates are in type 2 configuration (see below). Fucosylation by FUT3 gives rise to the Le a epitope, whereas the action of both enzymes results in Le b. Thus, the Le a epitope results from the addition by FUT3 of a fucose in an alpha 1-4 linkage to the GlcNAc on the unsubstituted precursor substrate. This product cannot be further glycosylated, and is the main antigen found in ABH nonsecretor individuals whose Lewis phenotype, therefore, is Le(a+b-).
The Le b epitope occurs when type 1 H epitope generated by the FUT2 locus, or the precursor substrate containing an alpha (1-2)-fucose linked to the Gal, is used as a substrate by FUT3. It is the main epitope found in ABH secretor individuals and is responsible for their Le(a-b+) phenotype.
Clearly, the genotype of an individual at the FUT2 and FUT3 loci determines the Lewis phenotype. In the presence of FUT2 alleles that express type 1 H determinants, the phenotype will be Le (a-b+), but individuals in whom the FUT2 gene is not expressed, it will be Le(a+b-). A Le(a+b+) phenotype may occur in individuals in whom the expression of H is decreased due to a mutation in the FUT2 gene (385A>T). This phenotype may also occur due to incomplete fucosylation of type H precursor sites. Failure to express FUT3 will result in a Lewis null phenotype [Le(a-b-)], irrespective of the Secretor status. However, one should note that the latter phenotype may also be observed upon the absence of Le glycolipids from the membrane, as may occur in patients with certain cancer or other conditions.
Biosynthetic pathways involved in synthesis of Lewis blood group antigens and secretion of ABH antigens into saliva. (Image source: J. Biol. Chem., Vol. 271, Issue 16, 9830-9837, April 19, 1996. )
The 3 genes are on chromosome 19p13.3? in a cluster with two other homologous genes of the family, in the gene order FUT6-FUT3-FUT5 (Reguigne-Arnould et al.) Function of proteins
Primary gene product of FUT3 functional alleles is an alpha(1,3/1,4) fucosyltransferase that transfers a fucose to a N-acetylglucosamine in either an alpha 1,4 linkage if the N-acetylglucosamine is located on a type 1 precursor, or in an alpha 1,3 linkage if it resides on a type 2 precursor. The former configuration is required for the expression of Lewis a or b antigens, wheras the latter is typical of Lewis X and Y. The function of Le a or Le b antigens is not clear but no apparent pathological consequences appear in Lewis null individuals. In sialylated forms they may serve as ligands for E-selectins. Their function in tumor tissues is not clear but they serve as tumor markers. The Le x and sLe x structures play an important role in adhesion phenomena in development and adult tissues.
In many ways the Lewis blood group system is analogous to the ABO and Hh blood group systems: 1) the antigens are indirect gene products and the epitopes are based on specific glycan structures. The structures result from the action of specific glycosyltransferases. 2) Both the ABO and Lewis systems are dependent on the H antigens to serve as precursor substrates that give rise to the A, B or Lewis epitopes. 3) The epitopes of the ABO and Lewis systems both depend on the interaction of products of two genetically independent loci, yet the polymorphisms observed in the population have a different molecular basis: in the ABO system two different alleles of the same gene result in either a N-acetylgalactosamine transferase or a galactose transferase ultimately responsible for the difference between the A and B epitopes, whereas the difference between Le a and Le b epitopes depends primarily on the status of the product of the FUT2 (secretor or Se) locus (19q13.3): if that enzyme is active (secretor status) the fucosylation of the precursor substrate by FUT3 will result in the Le b epitope whereas if the enzyme is inactive (Non-secretor?) the precursor substrate for FUT3 will not be fucosylated and the epitope, with a single fucose introduced by FUT3, will be Le a. In the case of an inactive FUT3 both epitopes will be obliterated and the phenotype will be Le(a-b-), irrespective of the activity state of FUT2. Thus, the introduction by FUT2 of the alpha 1,2 fucose in the absence of an alpha 1,4 adjacent fucose does not result in the Lewis epitope.
As noted, the expression of Lewis epitopes on the erythrocytes is dependent on adsorption from plasma of Lewis-bearing glycolipids. Thus, the Lewis negative phenotype may not occur uniquely as a result of an inactive allele, but may result from several other causes such as a failure of adsorption to the erythocyte or presence of an allele whose product is less active or may be selectively inactive towards glycolipid precursor substrates. Because other members of the fucosyltransferases family may fucosylate type 2 precursor substrates resulting in Le X and Le Y epitopes, the known alleles of FUT6 and FUT7 are included in the list of alleles because they give rise, respectively, to plasma or leukocyte alpha-3 fucosyltransferase. While most humans express FUT6 enzyme in plasma, 9% of individuals on the isle of Java (Indonesia) do not express this enzyme. It is of interest that 95% of plasma FUT6-deficient individuals have Lewis negative phenotype on red cells (in these subjects the status of FUT3 is not clear as yet), suggesting strong linkage disequilibrium between FUT3 and FUT6 genes. Concerning FUT7, so far, only a single variant allele was documented, occurring with an estimated frequency of ~1% in the population.
In the list of alleles, the following sequences (accession numbers) are used as reference: X53578 for Fut3, M98825 for Fut6, U11282 for Fut7.
The antigens of the Lewis system are carbohydrate (sugar) determinants carried either on proteins or lipids. Although they were first detected on red cells, the majority of the biochemical studies have been performed on Lewis substances isolated from plasma or saliva. Generally in both Caucasians and Blacks, the three major phenotypes are Le(a+b-), Le(a-b+) and Le(a-b-). These arise through the interaction of two genes- Lewis and secretor. If a Lewis gene is present the donor will be either Le(a+b-) or Le(a-b+); however, if there is no Lewis gene the red cells type as Le(a-b-). In Blacks the Le(a-b-) type occurs with a frequency of 20-25% as compared to 5% in Caucasians. Furthermore, red blood cells from newborns will type as Le(a-b-) regardless of their genetic makeup as the cells have not had time to adsorb Lewis antigens from the plasma. Another type which is extremely rare in Caucasians and Blacks, ie. Le(a+b+), is found in the Oriental population and appears to be due to a weak secretor gene.
In 1993, Boren et al. reported that the bacteria Helicobacter pylori used the fucose sugar found in the Leb antigen as a receptor to establish infection. H. pylori has been implicated as the causative agent in gastritis, peptic ulcers and gastric carcinoma. Several studies suggest that transplant patients having the Le(a-b-) phenotype have shorter transplant survival times than those who have a Lewis gene. Interestingly, antibodies raised to cancer cells often have specificity within the Lewis blood group.
The common Lewis antigens, Le-a and Le-b, are not intrinsic to RBCs, but are carried on plasma glycosphingolipids that are adsorbed from plasma to the RBC membranes. Their presence or absence in plasma and on RBCs is dependent, in part, on whether a person has inherited one Le or two le genes. The Le gene encodes a fucosyltransferase that adds Fuc in alpha(1-4) linkage to the subterminal GlcNAc of Type I oligosaccharides. (11) The resulting structure has Lea activity. Persons who have inherited the dominant Se(H) gene in addition to Le produce an antigen called Leb. When Leb is produced it is adsorbed preferentially over Lea to RBC membranes. Leb is made when Type I chains are first modified into H by the Se(H) gene transferase. The Le gene transferase then adds fucose to this structure to form Leb. The le gene is an amorph. Persons who are lele produce no Lea and no Leb antigens. RBCs that type as Le(a+b+) are only rarely found when human antisera are used in typing. Such RBCs are seen more frequently when more potent monoclonal anti-Lea and anti-Leb reagents are used.
As noted above the Lewis antigens are not synthesized in erythocyte progenitors. The glycosphingolipids that carry the antigenic Le-a or Le-b epitopes circulate in plasma bound to serum lipoproteins or in form of aqueous dispersions; they become adsorbed to the erythrocyte by a passive adsorption process. That is why their levels may vary; they are usually not detected on erythocytes of newborns but attain adult levels by the age of two. Also their red cell phenotype may become altered in disease or other conditions, such as pregnancy. It appears that the tissue-specific expression pattern of the FUT3 closely corresponds to that of the FUT2; the synthesis of Lewis glycans occurs predominantly in exocrine epithelial cells, mostly of endodermal origin. The digestive track is probably a major, but not exclusive, site of Lewis plasma glycolipid synthesis. They are expressed in colon and pancreas and a number of malignant tumors. Lewis and related antigens may occur as free oligosaccharides in milk and urine or may be protein-bound in a variety of tissues.
The first description of an antibody in the Lewis system was published in 1946 by A.E. Mourant. Lewis system antibodies are some of the most frequently encountered in pre-transfusion or pre-natal screening. Anti-Lea is the most frequent antibody in the Lewis system, is often naturally occurring and is of the IgM class. Anti-Leb exists in two forms: one reacts only with Le(b+) cells of the A2 or O type (anti-LebH) while the other reacts with all Le(b+) cell regardless of ABO type.
Lewis antibodies occur almost exclusively in the sera of Le(a-b-) people, and usually without known RBC stimulus. People whose RBC phenotype is Le(a- b+) do not make anti-Lea because small amounts of unconverted Lea are present in their saliva and plasma. It is unusual to find anti-Leb in the serum of a Le(a+ b-) person. Anti-Lea and anti-Leb may occur together in sera. They are almost always IgM and do not cross the placenta. Because of this, and because Lewis antigens are poorly developed at birth, the antibodies have not been implicated in HDN?. Lewis antibodies may bind complement. Fresh sera that contain anti-Lea (or infrequently anti- Leb) may cause the in vitro hemolysis of incompatible RBCs. In vitro hemolysis is more often seen with enzyme-treated RBCs than with untreated RBCs.
Most Lewis antibodies agglutinate saline-suspended RBCs of the appropriate phenotype. The resulting agglutinates are often fragile and are easily dispersed if RBC buttons are not resuspended gently after centrifugation. Agglutination is sometimes seen after incubation at 37 C, but rarely of the strength seen in tests incubated at room temperature. Some examples of anti-Lea, and less commonly anti-Leb, produce positive indirect antiglobulin reactions, providing complement is present in the reaction mixture and polyspecific antiserum is used.
Sera with anti-Leb activity can be divided into two categories. The most common examples react best with RBCs of group O and A2 These have been designated as anti-Leb(h). Those that react equally well with the Leb antigen on RBCs of all ABO phenotypes are called anti-Leb(l) . Anti-Leb(h) , but not anti-Leb(l) is neutralized by saliva from group O, Le(a- b-) persons who are secretors of H substance.
Two additional antibodies have been given names in the Lewis system although the determinants with which they react are not determined by Lewis genes. Anti-Lec has been reported in one human subject as a cold-reactive agglutinin. This antibody agglutinated the RBCs; of Le(a-b-) people who are sese and are therefore nonsecretors, of H substance. The antibody called anti-Led agglutinates the RBCs of Le(a- b-) secretors. The product defined by anti-Led has been identified as the Type I oligosaccharide to which Fuc has been added at the H-active site. Anti-Led should more correctly be called anti-Type I H. The material that reacts with anti-Lec seems to be the Type 1 chain with no added Fuc molecules. No examples of anti-Le d have been found in humans but both anti-Lec and anti-Led have been successfully produced in goats injected with saliva from Le(a- b-) non-secretors and Le(a- b-) secretors of H, respectively.
The Le-b glycan may be a receptor for Helicobacter pylori (Boren et al.). Lewis antigens are absent in patients suffering from the leukocyte adhesion deficiency syndromes, LADI or LADII. In the latter case the absence of the epitopes is due to a mutation in the GDP-fucose transporter (Luhn et al.; Hirschberg CB)
For a full discussion of the pathology linked with Lewis phenotype, see:
Lewis Antigens in Children
RBCs from newborn infants usually fail to react with both human anti-Lea and anti-Leb and, thus are considered to be Le(a- b-). Some can be shown to carry small amounts of Lea when tested with potent monoclonal or goat anti-Le a reagents. Reliable Lewis grouping of young children may not be possible, as test reactions may not reflect the correct phenotype until 6 years of age.
Among children, the incidence of Le(a+) RBCs is high and that of Le(b+) RBCs low. The phenotype Le(a+ b+) may be observed as a transient phase in children whose phenotypes as adults will be Le(a- b+).
Cord RBCs are agglutinated by certain sera that agglutinate the Le(a+ b- ) and Le(a-b+), but not Le(a-b-), RBCs of adults.33 In serological. tests, these sera appear to contain inseparable forms of anti-Lea and Le b. They define a determinant that has been called Lex, and which is present on the majority of cord RBCs and on the Le(a +) or Le(b +) RBCs of adults. Many serologists, have suggested that anti-Lex may represent a more potent or more avid form of anti-Lea.
Lewis A antigen in herbs
Two Chinese medicinal herbs, Leonurus sibiricus L. and Carthamus tinctorius L., were found to have high levels of Lewis a (Lea)-like determinants. The activity was nondialyzable, resistant to boiling and proteolytic enzymes.
The water-soluble extracts specifically inhibited hemagglutination, hemolysis, and lymphocyte cytotoxicity directed against Lea. “It was suprising that an Lea substance, previously found in human secretions, was also present in flower abstracts.” (Vox Sang 1983;45(4):320-5)
Association of Lewis blood group phenotypes with urinary tract infection in children
J Pediatr. 1994 Jun;124(6):863-8. Jantausch BA, Criss VR, O'Donnell R, Wiedermann BL, Majd M, Rushton HG, Shirey RS, Luban NL.
- Many blood group antigens, genetically controlled carbohydrate molecules, are found on the surface of uroepithelial cells and may affect bacterial adherence and increase the frequency of urinary tract infection (UTI) in adults. Sixty-two children aged 2 weeks to 17 years (mean, 2.3 years) who were hospitalized with fever in association with UTIs caused by Escherichia coli had complete (n = 50) or partial (n = 12) erythrocyte antigen typing to determine the role of erythrocyte antigens and phenotypes in UTI in children; 62 healthy children undergoing nonurologic elective surgery, matched 1 to 1 for age, sex, and race to the patient group, formed the control group. In univariate tests, patients and control subjects did not differ in ABO, Rh, P, Kell, Duffy, MNSs, and Kidd systems by the McNemar test of symmetry (p > 0.05). The frequency of the Lewis (Le) (a-b-) phenotype was higher (16/50 vs 5/50; p = 0.0076) and the frequency of the Le(a + b +) phenotype was lower (8/50 vs 16/50; p = 0.0455) in the patient population than in the control subjects. A stepwise logistic regression model to predict UTI with the explanatory variables A, B, O, M, N, S, s, Pl, Lea, and Leb showed that only the Lea and Leb antigens entered the model with p < 0.1. The Le(a-b-) phenotype was associated with UTI in this pediatric population. The relative risk of UTI in children with the Le(a-b-) phenotype was 3.2 (95% confidence interval, 1.3 to 7.9). Specific blood group phenotypes in pediatric populations may provide a means to identify children at risk of having UTI.
- This article is licensed under the GNU Free Documentation License. Sections excerpted from Blood Group Antigen Gene Mutation Database. See: Blumenfeld OO, Patnaik SK. Allelic genes of blood group antigens: a source of human mutations and cSNPs documented in the Blood Group Antigen Gene Mutation Database. Human Mutation. 2004 Jan; 23(1):8-16. PubMed ID: 14695527