Chapter 21 Blood cell antigens and antibodies

erythrocytes, platelets and granulocytes

Erythrocytes

Red Cell Antigens

Since Landsteiner’s discovery in 1901, that human blood groups existed, a vast body of serological, genetic and biochemical data on red cell (blood group) antigens has been accumulated. More recently, the biological functions of some of these antigens have been appreciated.

A total of 30 blood group systems have been described (Table 21.1). Each system is a series of red cell antigens, determined either by a single genetic locus or very closely linked loci. In addition to the blood group systems, there are six ‘collections’ of antigens (e.g. Cost), which bring together other genetically, biochemically or serologically related sets of antigens and a separate series of low-frequency (e.g. Rd) and high-frequency (e.g. Vel) antigens, which do not fit into any system or collection. A numeric catalogue of red cell antigens is being maintained by an International Society of Blood Transfusion (ISBT) Working Party.¹

Table 21.1 Blood group systems recognized by the ISBT Working Party

Apart from those of the ABO system, most of these antigens were detected by antibodies stimulated by transfusion or pregnancy.

Alternative forms of a gene coding for red cell antigens at a particular locus are called alleles and individuals may inherit identical or non-identical alleles. Most blood group genes have been assigned to specific chromosomes (e.g. ABO system on chromosome 9, Rh system on chromosome 1). The term genotype is used for the sum of the inherited alleles of a particular gene (e.g. AA, AO) and most red cell genes are expressed as codominant antigens (i.e. both genes are expressed in the heterozygote). The phenotype refers to the recognizable product of the alleles and there are many racial differences in the frequencies of red cell phenotypes, as shown in Table 21.2.

Table 21.2 Frequencies of red cell phenotypes in US Black and White populations

Red cell antigens are determined either by carbohydrate structures or protein structures. Carbohydrate-defined antigens are indirect gene products (e.g. ABO, Lewis, P). The genes code for an intermediate product, usually an enzyme that creates the antigenic specificity by transferring sugar molecules onto the protein or lipid. Protein-defined antigens are direct gene products and the specificity is determined by the inherited amino acid sequence and/or the conformation of the protein. Proteins carrying red cell antigens are inserted into the membrane in one of three ways: single pass, multipass or linked to phosphatidylinositol (GPI-linked). Only a few red cell antigens are erythroid-specific (Rh, LW, Kell and MNSs), the remainder being expressed in many other tissues. The structure and functions of the membrane proteins and glycoproteins carrying blood group antigens have been reviewed by Daniels.² An illustration of the putative functions of molecules containing blood group antigens is provided in Table 21.3.

Table 21.3 Putative functions of molecules containing blood group antigens

However, the main clinical importance of a blood group system depends on the capacity of alloantibodies (directed against the antigens not possessed by the individual) to cause destruction of transfused red cells or to cross the placenta and give rise to haemolytic disease in the fetus or newborn. This in turn depends on the frequency of the antigens and the alloantibodies and the characteristics of the latter: thermal range, immunoglobulin class and ability to fix complement. On these criteria, the ABO and Rh systems are of major clinical importance. Anti-A and anti-B are naturally occurring and are capable of causing severe intravascular haemolysis after an incompatible transfusion. The RhD antigen is the most immunogenic red cell antigen after A and B, being capable of stimulating anti-D production after transfusion or pregnancy in the majority of RhD-negative individuals.

ABO System

Discovery of the ABO system by Landsteiner marked the beginning of safe blood transfusion. The ABO antigens, although most important in relation to transfusion, are also expressed on most endothelial and epithelial membranes and are important histocompatibility antigens.³ Transplantation of ABO-incompatible solid organs increases the potential for hyperacute graft rejection, although ABO-incompatible renal transplantation can be successfully carried out with plasmapheresis in addition to immunosuppression of the recipient.⁴ Major ABO-incompatible stem cell transplants (e.g. group A stem cells into a group O recipient) will provoke haemolysis, unless the donation is depleted of red cells.

ABO Antigens and Encoding Genes

There are four main blood groups: A, B, AB and O (Table 21.4). In the British Caucasian population, the frequency of group A is 42%, B 9%, AB 3% and O 46%, but there is racial variation in these frequencies.⁵ The epitopes of ABO antigens are determined by carbohydrates (sugars), which are linked either to polypeptides (forming glycoproteins) or to lipids (glycolipids).

Table 21.4 ABO blood group system

The expression of ABO antigens is controlled by three separate genetic loci: ABO located on chromosome 9 and FUT1 (H) and FUT2 (Se), both of which are located on chromosome 19. The genes from each locus are inherited in pairs as Mendelian dominants. Each gene codes for a different enzyme (glycosyltransferase), which attaches specific monosaccharides onto precursor disaccharide chains (Table 21.5). There are four types of disaccharide chains known to occur on red cells, on other tissues and in secretions. The Type 1 disaccharide chain is found in plasma and secretions and is the substrate for the FUT2 (Se) gene, whereas Types 2, 3 and 4 chains are only found on red cells and are the substrate for the FUT1 (H) gene. It is likely that the O and B genes arose by mutation of the A gene. The O gene does not encode for the production of a functional enzyme; group O individuals commonly have a deletion at nucleotide 261 (the O1 allele), which results in a frame-shift and premature termination of the translated polypeptide and the production of an enzyme with no catalytic activity. The B gene differs from A by consistent nucleotide substitutions.⁶ The expression of A and B antigens is dependent on the H and Se genes, which both give rise to glycosyltransferases that add L-fucose, producing the H antigen. The presence of an A or B gene (or both) results in the production of further glycosyltransferases, which convert H substance into A and B antigens by the terminal addition of N-acetyl-D-galactosamine and D-galactose, respectively (Fig. 21.1). Because the O gene produces an inactive transferase, H substance persists unchanged as group O. In the extremely rare Oh Bombay phenotype, the individual is homozygous for the h allele of FUT1 and hence cannot form the H precursor of the A and B antigen. Their red cells type as group O, but their plasma contains anti-H, in addition to anti-A, anti-B and anti-A,B, which are all active at 37°C. As a consequence, individuals with an Oh Bombay phenotype can only be safely transfused with other Oh red cells.

Table 21.5 Glycosyltransferases produced by genes encoding antigens within the ABO, H and Lewis blood group systems

Gene	Allele	Transferase
FUT1	H	α-2-L-fucosyltransferase
	h	None
A	A	α-3-N-acetyl-D-galactosaminyltransferase
B	B	α-3-D-galactosyltransferase
O	O	None
FUT2	Se	α-2-L-fucosyltransferase
	se	None
FUT3	Le	α-3/4-L-fucosyltransferase
	le	None

Figure 21.1 Pathways from HAB blood group genes to antigens. *Glycosyltransferase H transfers L-fucose; A transfers N-acetyl-D-galactosamine; B transfers D-galactose; O is inactive.

Serologists have defined two common subgroups of the A antigen. Approximately 20% of group A and group AB individuals belong to group A₂ and group A₂B, respectively, the remainder belonging to group A₁ and group A₁B. These subgroups arise as a result of inheritance of either the A¹ or A² alleles. The A₂ transferase is less efficient in transferring N-acetyl-D-galactosamine to available H antigen sites and cannot utilize Types 3 and 4 disaccharide chains. As a consequence, A₂ red cells have fewer A antigen sites than A₁ cells and the plasma of group A₂ and group A₂B individuals may also contain anti-A₁. The distinction between these subgroups can be made using the lectin Dolichos biflorus, which only reacts with A₁ cells. The H antigen content of red cells depends on the ABO group and, when assessed by agglutination reactions with anti-H, the strength of reaction tends to be graded O > A₂ > A₂B > B > A₁ > A₁B. Other subgroups of A are occasionally found (e.g. A₃, A_x) that result from mutant forms of the glycosyltransferases produced by the A gene and are less efficient at transferring N-acetyl-D-galactosamine onto H substance.⁶

The A, B and H antigens are detectable early in fetal life but are not fully developed on the red cells at birth. The number of antigen sites reaches ‘adult’ level at around 1 year of age and remains constant until old age, when a slight reduction may occur.

Secretors and Non-Secretors

The ability to secrete A, B and H substances in water-soluble form is controlled by FUT2 (dominant allele Se). In a Caucasian population, about 80% are secretors (genotype SeSe or Sese) and 20% are non-secretors (genotype sese) (Table 21.6). Secretors have H substance in the saliva and other body fluids together with A substances, B substances or both, depending on their blood group. Only traces of these substances are present in the secretions of non-secretors, although the antigens are expressed normally on their red cells and other tissues.

Table 21.6 Secretor status in the Caucasian population

An individual’s secretor status can be determined by testing for ABH substance in saliva (see p. 504).

ABO Antigens and Disease

Group A individuals rarely may acquire a B antigen from a bacterial infection that results in the release of a deacetylase enzyme. This converts N-acetyl-D-galactosamine into α-galactosamine, which is similar to galactose, the immunodominant sugar of group B, thereby sometimes causing the red cells to appear to be group AB. In the original reported cases, five out of seven of the patients had carcinoma of the gastrointestinal tract. Case reports attest to the danger of individuals with an acquired B antigen being transfused with AB red cells, resulting in a fatal haemolytic transfusion reaction following the production of hyperimmune anti-B.⁷

The inheritance of ABH antigens is also known to be weakly associated with predisposition to certain diseases. Group A individuals have 1.2 times the risk of developing carcinoma of the stomach than group O or B; group O individuals have 1.4 times more risk of developing peptic ulcer than non-group O individuals; and non-secretors of ABH have 1.5 times the risk of developing peptic ulcer than secretors.⁸ The ABO group also affects plasma von Willebrand factor (VWF) and factor VIII levels; group O healthy individuals have levels around 25% lower than those of other ABO groups.⁹ ABO blood group appears to mediate its effect by accelerating clearance of VWF but the mechanism is not yet clear¹⁰ ABH antigens are also frequently more weakly expressed on the red cells of persons with leukaemia.

ABO Antibodies

Anti-A and anti-B

ABO antibodies, in the absence of the corresponding antigens, appear during the first few months after birth, probably as a result of exposure to ABH antigen-like substances in the diet or the environment (i.e. they are ‘naturally occurring’) (Table 21.4). This allows for reverse (serum/plasma) grouping as a means of confirming the red cell phenotype. The antibodies are a potential cause of dangerous haemolytic transfusion reactions if transfusions are given without regard to ABO compatibility. Anti-A and anti-B are always, to some extent, immunoglobulin M (IgM). Although they react best at low temperatures, they are nevertheless potentially lytic at 37°C. Hyperimmune anti-A and anti-B occur less frequently, usually in response to transfusion or pregnancy, but they may also be formed following the injection of some toxoids and vaccines. They are predominantly of IgG class and are usually produced by group O and sometimes by group A₂ individuals. Hyperimmune IgG anti-A and/or anti-B from group O or group A₂ mothers may cross the placenta and cause haemolytic disease of the newborn (HDN). These antibodies react over a wide thermal range and are more effective haemolysins than the naturally occurring antibodies. Group O donors should always be screened for high-titre anti-A and anti-B antibodies, which may cause haemolysis when group O platelets or plasma are transfused to recipients with A and B phenotypes.

Plasma-containing blood components from such high-titre universal donors should be reserved for group O recipients.

Anti-A₁ and anti-H

Anti-A₁ reacts only with A₁ and A₁B cells and is occasionally found in the serum of group A₂ individuals (1–8%) and not uncommonly in the serum of group A₂B subjects (25–50%). However, anti-A₁ normally acts as a cold agglutinin and is very rarely reactive at 37°C, when it is only capable of limited red cell destruction. There have been a few reports of red cell haemolysis ascribed to anti-A₁, which some authors have questioned because, although the antibodies reacted only with A₁ red cells, no attempts were made to absorb them with A₂ cells, which would have revealed their anti-A specificity.

Anti-H reacts most strongly with group O and A₂ red cells and also normally acts as a cold agglutinin. A notable, but rare, exception is the anti-H that occurs in the Oh Bombay phenotype, which is an IgM antibody and causes lysis at 37°C (Table 21.4) so that Oh Bombay phenotype blood would be required for transfusion.

Lewis System

Lewis Antigens and Encoding Genes

The Lewis antigens (Le^a and Le^b) are located on soluble glycosphingolipids found in saliva and plasma and are secondarily absorbed into the red cell membranes from the plasma.

The Le gene at the FUT3 (LE) locus is located on chromosome 19 and codes for a fucosyltransferase, which acts on an adjacent sugar molecule to that acted on by the Se gene. Where Se and Le are present, the Le^b antigen is produced; where Le but not Se is present, Le^a is produced; and where Le is not present, neither Le^a nor Le^b is produced. After transfusion of red cells, donor red cells convert to the Lewis type of the recipient owing to the continuous exchange of glycosphingolipids between the plasma and red cell membrane.

Neonates have the phenotype Le(a−b−) because low levels of the fucosyltransferase are produced in the first 2 months of life.

Lewis Antibodies

Lewis antibodies are naturally occurring and are usually IgM and complement binding. In vitro, their reactivity is enhanced with the use of enzyme-treated red cells, when lysis may occur. However, only rare examples of anti-Le^a that are strictly reactive at 37°C have given rise to haemolytic transfusion reactions and there is no good evidence that anti-Le^b has ever caused a haemolytic episode. Explanations for the relative lack of clinical significance include their thermal range, neutralization by Lewis antigens in the plasma of transfused blood and the gradual elution of Lewis antigens from the donor red cells. Consequently, it is acceptable to provide red cells for transfusion that have not been typed as negative for the relevant Lewis antigen but are compatible with the recipient plasma when the compatibility test is performed strictly at 37°C.

Lewis antibodies have not been implicated in haemolytic disease of the fetus or newborn. The role of Lewis in influencing the outcome of renal transplants is unclear.

The P System and Globoside Collection

The Rh haplotypes are named either by the component antigens (e.g. CDe, cde) or by a single shorthand symbol (e.g. R¹ = CDe, r = cde). Thus, a person may inherit CDe (R¹) from one parent and cde (r) from the other and have the genotype CDe/cde (R¹r). The haplotypes in order of frequency and the corresponding shorthand notation are given in Table 21.7. Although two other nomenclatures are also used to describe the Rh system, namely, Wiener’s Rh-Hr terminology and Rosenfield’s numeric notation, the CDE nomenclature, derived from Fisher’s original theory, is recommended by a World Health Organization Expert Committee¹³ in the interest of simplicity and uniformity. The Rh antigens are defined by corresponding antisera, with the exception of ‘anti-d’, which does not exist. Consequently, the distinction between homozygous DD and the heterozygous Dd cannot be made by direct serological testing but may be resolved by informative family studies. It is still routine practice to predict the genotype from the phenotype on the basis of probability tables for the various Rh genotypes in the population (Table 21.7). However, in women with immune anti-D and a history of an infant affected by HDN, RH DNA typing is used in prenatal testing for the fetal D status to decide on the clinical management of the pregnancy, e.g. the need for monitoring for fetal anaemia using middle cerebral artery Doppler ultrasound. Suitable sources include amniotic fluid (amniocytes) and trophoblastic cells (chorionic villi) or after 15 weeks’ gestation, maternal blood can be used because it contains fetal DNA.^14,¹⁵ In practice, multiplex polymerase chain reaction (PCR) is used, with more than two primer sets, to detect the different molecular bases for D-negative phenotypes in non-Caucasians. RH DNA typing also has applications in paternity testing and forensic medicine. There are racial differences in the distribution of Rh antigens, e.g. D negativity is more common in Caucasians (approximately 15%), whereas R^o (cDe) is found in approximately 48% of Black Americans but is uncommon (approximately 2%) in Caucasians. The Rh antigens are present only on red cells and are a structural part of the cell membrane. Complete absence of Rh antigens (Rh-null phenotype) may be associated with a congenital haemolytic anaemia with spherocytes and stomatocytes in the blood film, increased osmotic fragility and increased cation transport. This phenotype arises either as a result of homozygosity for silent alleles at the RH locus (the amorph type) or more commonly by homozygosity for an autosomal suppressor gene (X), genetically independent of the RH locus (the regulator type). Rh antigens are well-developed before birth and can be demonstrated on the red cells of very early fetuses.

Table 21.7 The Rh haplotypes in order of frequency (Fisher nomenclature) in Caucasians and the corresponding short notations

Fisher	Short Notations	Approximate Frequency (%)
CDe	R¹	41
cde	r	39
cDE	R²	14
cDe	R⁰	3
C^wDe	R^1w	1
cdE	r″	1
Cde	r′	1
CDE	R^Z	Rare
CdE	r^y	Rare

Antibodies

Fisher’s nomenclature is convenient when applied to Rh antibodies, and antibodies directed against all Rh antigens, except d, have been described: anti-D, anti-C, anti-c, anti-E and anti-e. Rh antigens are restricted to red cells and Rh antibodies result from previous alloimmunization by previous pregnancy or transfusion, except for some naturally occurring forms of anti-E and anti-C^W. Immune Rh antibodies are predominantly IgG (IgG1 and/or IgG3), but may have an IgM component. They react optimally at 37°C, they do not bind complement and their detection is often enhanced by the use of enzyme-treated red cells. Haemolysis, when it occurs, is therefore extravascular and predominantly in the spleen.

Anti-D is clinically the most important antibody; it may cause haemolytic transfusion reactions and was a common cause of fetal death resulting from haemolytic disease of the newborn before the introduction of anti-D prophylaxis. Anti-D is accompanied by anti-C in 30% of cases and anti-E in 2% cases. Primary immunization following a transfusion of D-positive cells becomes apparent within 2–5 months, but it may not be detectable following exposure to a small dose of D-positive cells in pregnancy. However, a second exposure to D-positive cells in a subsequent pregnancy will provoke a prompt anamnestic or secondary immune response.

Of the non-D Rh antibodies, anti-c is most commonly found and can also give rise to severe haemolytic disease of the fetus and newborn. Anti-E is less common, whereas anti-C is rare in the absence of anti-D.

Kell and Kx Systems

Antigens and Encoding Genes

A total of 34 antigens have been identified (K1–K34), but three very closely linked sets of alleles are clinically important: K (KEL1) and k (KEL2); Kp^a (KEL3), Kp^b (KEL4) and Kp^c (KEL21); and Js^a (KEL6) and Js^b (KEL7). These antigens are encoded by alleles at the KEL locus on chromosome 7, but their production also depends on genes at the KX locus on the X chromosome. The K antigen is present in 9% of the English population. The Kp^b antigen has a high frequency in Caucasians; the Js^b antigen is universal in Caucasians and almost universal in black races.

The Kell protein is a single-pass glycoprotein and is believed to be complexed by a disulphide bridge to the Kx protein, which is multipass with 10 putative transmembrane domains. It has considerable sequence homology to other neutral endopeptidases.

In the McLeod phenotype, red cells lack Kx and there is a marked decrease in all Kell antigens, an acanthocytic morphology and compensated haemolysis. The McLeod syndrome is X-linked with slow progression to cardiomyopathy, skeletal muscle wasting and neurological defects.

Kell Antibodies

Immune anti-K is the most common antibody found outside the ABO and Rh systems. It is commonly IgG1 and occasionally complement binding. Other immune antibodies directed against Kell antigens are less common. The presence of some of these antibodies, such as anti-k, anti-Kp^b and anti-Js^b, may cause extensive difficulties in the selection of antigen-negative units for transfusion.

Duffy System

Duffy antigens and encoding genes

The Duffy (Fy) locus is on chromosome 1 and encodes a multipass protein with seven or nine putative transmembrane domains.

The locus has the following alleles: Fy^a, Fy^b, which code for the co-dominant Fy^a and Fy^b antigens, respectively; Fy^x, which is responsible for a weak Fy^b antigen; and Fy, which is responsible when homozygous for the Fy(a−b−) phenotype in black races. This Fy gene is identical to the Fy^b gene in its structural region but has a mutation in the promoter region, resulting in the lack of production of red cell Duffy glycoprotein.

The Fy glycoprotein (also known as Duffy antigen receptor for chemokines, DARC) is a receptor for the CC and CXC classes of proinflammatory chemokines and is expressed on vascular endothelial cells and Purkinje cells in the cerebellum, but its precise role as a potential scavenger of excess chemokines is unknown. The Fy glycoprotein is also a receptor for Plasmodium vivax.

Duffy antibodies

Anti-Fy^a is much more common than anti-Fy^b and all other Duffy antibodies are rare apart from anti-Fy³ (to both Fy^a and Fy^b), which occurs in some African/Afro-Caribbean patients, in whom Fy(a−b−) antigen status is common. They are predominantly IgG1 and are sometimes complement binding.

Kidd (JK) System

Kidd antigens and encoding genes

Genes at the HUT 11(JK) locus on chromosome 18 encode for a multipass protein, which carries both the Kidd antigens and the human erythroid urea transporter. The codominant alleles, Jk^a and Jk^b, represent a polymorphism on HUT 11, which differs by a single amino acid substitution at position 280 (Asp/Asn).

The Jk(a−b−) phenotype is very rare and is caused by homozygous inheritance of the silent allele, Jk, at the JK locus or by inheritance of the dominant inhibitor gene In (Jk) unlinked to the JK locus. These Jk(a−b−) cells are resistant to lysis by solutions of urea and have a selective defect in urea transport.

Kidd antibodies

Anti-Jk^a is more common than anti-Jk^b; both are usually IgG. Kidd antibodies are usually complement binding, which is thought to be because most of them contain an IgG3 fraction. Anti-Jk3 is produced by individuals of the rare Jk(a−b−) phenotype.

Kidd antibodies can be difficult to detect because they often show dosage (may only react with cells showing homozygous expressions of Jk^a or Jk^b), they fall to undetectable levels in plasma and they are often present in mixtures of alloantibodies. A previous history of antibodies is therefore important, to avoid a post-transfusion haemolytic reaction, due to an anamnestic response by an antibody that was below the level of detection before transfusion.

MNSs System

MNSs antigens and encoding genes

GYPA and GYPB are closely linked genes on chromosome 4 and encode glycophorin A (GPA) and glycophorin B (GPB), respectively. Both GPA and GPB are single-pass membrane sialoglycoproteins. M and N are alleles of GYPA (encoding the M and N antigens on GPA) and S and s are alleles of GPYB (encoding the S and s antigens on GPB). Many rare variants have been described owing to gene deletions, mutations and segmental exchanges.

The U antigen is found on the red cells of Caucasians and 99% of black races. U-negative individuals are, with rare exceptions, S−s− and lack GPB or have an altered form of GPB.

MNSs antibodies

Anti-M is a relatively common antibody that may be IgM or IgG. Rare examples are reactive at 37°C when they can give rise to haemolytic transfusion reactions. Anti-M very rarely gives rise to HDN.

Anti-N is uncommon and of no clinical significance.

Anti-S and anti-s are usually IgG; both rarely have been implicated in haemolytic transfusion reactions and HDN.

Anti-U is a rare immune antibody, usually containing an IgG1 component. It has been known to cause fatal haemolytic transfusion reactions and occasional severe HDN.

Other Blood Group Systems

Lutheran system

The antigens in the Lutheran system are not well-developed at birth and as a consequence there are no documented cases of clinically significant haemolytic disease owing to Lutheran antibodies.

Anti-Lu^a is uncommon and rarely of clinical significance. Anti-Lu^b has caused extravascular haemolysis.

Yt (Cartwright) system

The antigens Yt^a and Yt^b are found on GPI-linked acetylcholinesterase. Some examples of anti-Yt^a have caused accelerated red cell destruction.

Colton system

The antigens in the Colton system, Co^a and Co^b, are carried on the water-transport protein, channel-forming integral protein (CHIP-1). Anti-Co^a and the rarer anti-Co^b are both sometimes clinically significant.

Dombrock system

The antigens in the Dombrock system include Do^a and Do^b and also include the high-incidence antigens Gy^a, Hy and Jo^a. Antibodies of this system are usually weak, but all should be considered as potentially significant.

Clinical Significance of Red Cell Alloantibodies

The significance of the alloantibodies described, with respect to the nature of the haemolytic transfusion reaction they produce, is provided in Table 21.8. The majority of haemolytic transfusion reactions, however, are the result of ABO incompatibility.¹⁶

Table 21.8 Antibody specificities related to the mechanism of immune haemolytic destruction

Blood Group System	Intravascular Haemolysis	Extravascular Haemolysis
ABO,H	A, B, H
Rh		All
Kell	K	K, k, Kp^a, Kp^b, Js^a, Js^b
Kidd	Jk^a	Jk^a, Jk^b, Jk³
Duffy		Fy^a, Fy^b
MNS		M, S, s, U
Lutheran		Lu^b
Lewis	Le^a
Cartwright		Yt^a
Colton		Co^a, Co^b
Dombrock		Do^a, Do^b

Mollison et al.¹⁷ analysed the significance of blood group antigens other than those of the ABO system and D by looking at the prevalence of transfusion-induced red cell alloantibodies, excluding anti-D, -CD and -DE (Table 21.9). Rh antibodies, mainly anti-c or anti-E, accounted for 53% of the total and anti-K and anti-Fy^a accounted for a further 38%, leaving only about 9% for all other specificities. A similar distribution of the different red cell antibodies was found in a smaller group of patients who experienced immediate haemolytic transfusion reactions (HTR). However, the figures for delayed HTR showed a striking increase in the relative frequency of Jk antibodies, which reflects the outlined characteristics of Jk antibodies.

Table 21.9 Relative frequency of immune red cell alloantibodies*

Haemolytic disease of the fetus and newborn has not been associated with antibodies directed against Lewis antigens and only very mild disease is produced by anti-Lu^a and anti-Lu^b. With these exceptions, all other IgG antibodies directed against antigens in the systems mentioned should be considered capable of causing haemolysis in this setting.

The significance of the many other blood group antigens not referred to in the text is summarized in Table 21.10. However, it should be noted that the antibodies listed are usually wholly or predominantly IgG and would be detectable in routine pretransfusion testing using the indirect anti-globulin test (IAT).

Table 21.10 ‘Minor’ blood group antigens

It is difficult to find suitable blood for transfusion to a patient whose plasma contains an antibody, such as anti-Vel, which has a specificity for a high-frequency antigen and which can cause severe haemolytic transfusion reactions. In addition to using blood from a frozen blood bank and calling up rare phenotype donors, autologous blood could be considered for planned elective procedures and if necessary, the compatibility of red cells from close relatives (particularly siblings) can be investigated. Antibodies such as anti-Kn^a are commonly found and not clinically important, but their presence may cause delay in the provision of blood until their specificity has been determined.

Mechanisms of Immune Destruction of Red Cells

Immune-mediated haemolysis of red cells depends on the following:¹⁸

1. The immunoglobulin class of the antibody (for all practical purposes, antibodies directed against red cell antigens are either IgM or IgG or both)

2. The ability of the antibody to bind complement.

3. Interaction with the reticuloendothelial system (mononuclear phagocytic system). The most important phagocyte participating in immune haemolysis is the macrophage, predominantly in the spleen.

The mechanism of immune haemolysis also determines the site of haemolysis:

a. Intravascular haemolysis owing to sequential binding of complement components (C1 to C9) cascade and the formation of the membrane attack complex (MAC; C5b678(9)_n). This is characteristic of IgM antibodies, but some IgG antibodies can also act as haemolysins. Red cells are usually destroyed by intravascular complement lysis in ABO incompatible transfusion reactions (see p. 543). Most other alloimmune red cell destruction is extravascular and mediated by the mononuclear-phagocytic system.

Red cell autoantibodies may also cause intravascular lysis, especially the IgG autoantibody of PCH (see p. 277) and some autoantibodies of the cold haemagglutinin disease (CHAD) (see p. 277). Complement-mediated intravascular lysis may also occur in drug-induced immune haemolysis (see p. 289).

b. Extravascular haemolysis by the mononuclear phagocytic system is characteristic of IgG antibodies and occurs predominantly in the spleen. This is caused by non-complement-binding IgG antibodies or those that bind sublytic amounts of complement. Macrophages have Fcγ receptors for cell-bound IgG and sensitized red cells may be wholly phagocytosed or lose part of the membrane and return to the circulation as microspherocytes Spherocytes are less deformable and more readily trapped in the spleen than normal red cells; this shortens their lifespan. In addition to Fc receptor-mediated phagocytosis, antibody-dependent cell-mediated cytotoxicity (ADCC) may also contribute to cell damage during the close contact with splenic macrophages. Red cells are destroyed external to the monocyte membrane by lysosomal enzymes secreted by the monocyte.¹⁹

Complement components may enhance red cell destruction. Complement activation by some IgM and most IgG antibodies is not always complete and the red cell escapes intravascular lysis. The activation of complement stops at the C3 stage and, in these circumstances, complement can be detected on the red cell by the antiglobulin test using appropriate anticomplement reagents. The first activation product of C3 is membrane-bound C3b, which is constantly being broken down to C3bi. Red cells with these components on their surface adhere to phagocytes (monocytes, macrophages and neutrophils), which have complement receptors, CR1 (CD35) and CR3 (CD 11b/CD 18). These sensitized cells are rapidly sequestered in the liver because of its bulk of phagocytic cells (Küpffer cells) and large blood flow, but no engulfment occurs. When C3bi is cleaved, leaving only C3dg on the cell surface, the cells tagged with ‘inactive’ C3dg return to the circulation, as in chronic cold haemagglutinin disease. However, when IgG is also present on the cell surface, C3b enhances phagocytosis and under these circumstances both liver and spleen are important sites of extravascular haemolysis. Hence, C3b and C3bi augment macrophage-mediated clearance of IgG-coated cells and antibodies binding sublytic amounts of complement (e.g. Duffy and Kidd antibodies) often cause more rapid destruction and more marked symptoms than non-complement binding antibodies (e.g. Rh antibodies).

Macrophage activity is an important component of cell destruction and further study of cellular interactions at this stage of immune haemolysis may provide an explanation for the differing severity of haemolysis in patients with apparently similar antibodies. In vitro macrophage (monocyte) assays have been used sometimes to supplement conventional serological techniques to assess this aspect of immune haemolysis.²⁰

Factors that may affect the interaction between sensitized cells and macrophages include the following:

1. IgG subclass. IgG1 and IgG3 antibodies have a higher binding affinity to mononuclear Fcγ receptors than IgG2 and IgG4 antibodies.

2. Antigen density. This affects the number of antibody molecules bound to the cell surface.

3. Fluid-phase IgG. Serum IgG concentration is a determinant of Fc-dependent mononuclear-phagocytic function. Normal levels of IgG block the adherence of sensitized red cells to monocyte Fc receptors (particularly FcγR1) in vitro. Haemoconcentration within the splenic sinusoids is probably a major factor in minimizing this effect in vivo, which may explain why the spleen is about 100 times more efficient at removing IgG-sensitized cells than the liver despite the greater macrophage mass and higher blood flow of the latter organ.

The initial effect of high-dose intravenous IgG is to cause blockade of macrophage FcγR. This reduces the immune clearance of antibody-coated cells and has particular application in the management of autoimmune thrombocytopenia and post-transfusion purpura.

4. Regulation of macrophage activity. Cytokines are known to be important in the upregulation of macrophage receptors. Interferon gamma enhances macrophage phagocytic activity by increasing the expression of FcγRI in vitro and in vivo and also activates FcγII without increasing the number of these receptors.²¹

Interleukin-6 also enhances FcγRII activation and increased activity of the CR1 receptor occurs through the action of T-cell cytokines and through chemotactic agents released in the inflammatory response.²² The increased levels of proinflammatory cytokines and other biological mediators and their effects on the activity of the monocyte phagocytic system have been monitored in patients with systemic inflammatory response syndrome.²³ It is therefore possible that release of cytokines during viral and bacterial infections could, at least in part, trigger some episodes of autoimmune cell destruction.

The rate of immune destruction is therefore determined by antigen and antibody characteristics and the level of activation of the monocyte phagocytic system.