Pathobiology of Sickle Cell Disease

Molecular Context

The sickle mutation in the HBB gene is a GAG→GTG conversion that creates a β ^6Glu→Val substitution and thereby forms β ^S globin chains. Genes for other β-globin variants are allelic to the β ^S gene and have a codominant impact. Examples include genes for the normal β chain (β ^A ), other β mutants (e.g., β ^C , β°, or β ⁺ thalassemia), and deletional hereditary persistence of fetal Hb (HPFH). Compound heterozygosity for β ^S and each one of these results in well-defined clinical syndromes, such as HbAS (sickle trait), HbSC disease, HbS–β-thalassemia, and HbS-HPFH. Eight percent of African Americans have a β ^S gene, 3% have β ^C , 1.5% have β-thalassemia, and 0.1% have HPFH. Among African Americans, about 1 in 600 births results in the homozygous state, HbSS, and about 1 in 400 results in some form of sickle cell disease, which additionally includes the compound heterozygous variants other than HbAS. Worldwide, about 78% of HbSS births now occur in sub-Saharan Africa, 14% in India, 4% in the Americans, 3% in the Eastern Mediterranean, and 1% in Europe.

The HBB gene resides in a β-like gene cluster within which are various non-exonic polymorphic sites. Different combinations of these define discrete β-locus backgrounds, referred to as the Senegal, Benin, Bantu, Cameroon, and Arab-India haplotypes. Each refers to an ethnographic region in which the sickle mutation arose and achieved high gene frequency, typically peaking at 0.15 to 0.18. In most cases, the sickle gene resides on one of these five major haplotypes. Recent DNA sequence analysis of these haplotypes has suggested that, in fact, the sickle allele arose just once about 7300 years ago in central Africa and spread to these 5 haplotypes as the result of migration and continued selective pressure from malaria.

Origin, Selection, and Dispersion of the Sickle Gene

The residence of both β ^A and β ^S alleles on the distinct regional β cluster haplotypes suggests that the sickle mutation was selected for independently in the five regions. The β ^C mutation arose only once. Historical and biological data argue that the frequency of the β ^S gene greatly expanded in South Asia about 4000 years ago and in Africa about 3000 years ago, following the introduction of iron tools in each case. That led to the adoption of an agricultural system that promoted both increased human habitation density and favorable breeding conditions for the Anopheles mosquito, which in turn enabled the development of endemic Plasmodium falciparum malaria. In this context, high fixed β ^S gene frequencies were reached because of a balanced polymorphism, such that heterozygotes (HbAS) have an adaptive advantage over either homozygote. Thus, the Old World geographic distributions of the sickle gene and historical endemic malaria are notably concordant, suggesting that the sickle gene represents “a biologic solution to a cultural problem.”

In hyperendemic areas, falciparum malaria uniformly infects the young and is the primary cause of death for children with HbSS. However, those with HbAS are less likely to develop severe malaria, a protective effect appearing early in childhood and possibly increasingly until age 10. At the level of the RBC, this protection reflects steps after the initial parasite invasion. One proposed mechanism links protection to the instability of HbS, immune status, and splenic function. That is, infection of HbAS RBCs with P. falciparum leads sequentially to augmented Hb denaturation, clustering of membrane protein band 3, attraction of band 3 autoantibody, complement binding, and enhanced erythrophagocytosis even of the early ring forms. Thereby, an accelerated clearance of parasitized RBC by the functional spleen could protect those with HbAS, while HbS homozygotes would lose this protection as they acquire functional asplenia. In synergy with this scenario, the presence of HbS impairs endothelial cytoadherence of infected RBC, thereby diminishing cerebral symptoms and impeding the RBC sequestration that would otherwise protect them from splenic exposure. An alternative (or complementary) immune mechanism may be developing antibodies to falciparum antigens on the RBC surface.

Additionally, the blunted malarial susceptibility in HbAS reflects complex interrelationships between the sickle gene, other globin genes, host biology, and environmental factors. For example, the protective benefit of HbAS is lost if there is concurrent α thalassemia (which lowers the proportion of HbS). Malarial severity is affected by polymorphisms in non-globin genes such as CR1 (complement receptor 1), CD36 , TGFB1 (transforming growth factor β), HMOX1 (heme oxygenase 1), and TLR4 (toll-like receptor 4). Both carbon monoxide (CO) and nitric oxide (NO) blunt the severity of experimental malaria. And certain micro RNAs, enriched in HbS-containing RBC, can inhibit P. falciparum growth.

Eventually, the sickle gene spread geographically by means of commerce, migration, and the slave trade. This dispersion has been tracked by analyses of the regional β haplotypes, a biological marker that largely corroborates predictions of gene flow derived from historical records. As a generalization, it spread on the Benin haplotype to North Africa and then across the Mediterranean. All three major African haplotypes are present in the western Arabian Peninsula; but on the eastern side, the sickle gene tends to be on the Arab-India haplotype. This is also true in India, although sub-Saharan haplotypes are represented as well. In the Americas, the β ^S gene is mostly found on the Benin, Senegal, and Bantu haplotypes.

Hemoglobin S Charge and Tetramer Assembly

The formation of Hb tetramers requires the proximate assembly of stable dimers from unlike monomers (e.g., α + β → αβ), an event governed by electrostatic attraction. The normal α and β chains are positively and negatively charged, respectively. In heterozygous states for β-globin mutants, β-chain competition for dimer assembly is a determinant of the relative proportions of the Hb variants. Mutant β chains with less negative charge form αβ dimers more slowly; the relative rates for dimer association are αβ ^A > αβ ^S > αβ ^C , with αβ ^A dimers formed about twice as rapidly as αβ ^S dimers. This explains why those with HbAS typically have only 40% HbS and why the proportion of HbS exceeds this in HbSC disease. It also explains the effect of concurrent α-thalassemia on the proportion of HbS in sickle trait—availability of α chains becomes limiting, the percentage of HbS typically drops from 40% to 35% (one α deletion), 30% (two α deletions), or less than 25% (three α deletions).

Hemoglobin S Stability and Oxidant Formation

HbS is modestly unstable, observed in vitro as instability to various applied stresses. Two stresses that are most clearly physiologic involve Hb oxidation. HbS has an abnormal redox potential compared with HbA that may underlie its modestly (~40%) increased auto-oxidation rate. Yet, HbS exhibits markedly (~340%) augmented instability and oxidation upon interaction with aminophospholipids of the membrane’s inner leaflet. This instability leads to the accumulation of various Hb and iron forms at the cytosol–membrane interface. The resulting localization of abnormal oxidative biochemistry promotes a number of prominent defects of the sickle RBC membrane. Other contributors to oxidant stress within sickle RBC are addressed later.

Hemoglobin S and Polymerization

OxyHbS, oxyHbA, and deoxyHbA have very high solubilities, but deoxyHbS aggregates into densely packed polymers, a process that is fully reversible upon reoxygenation. This abnormal property, the fundamental pathogenesis of the sickling disorders, causes the eponymous RBC shape to change because of polymer-mediated distortion (see Fig. 42.1 ).

Polymer Structure

Deoxygenation transforms soluble HbS into a highly viscous and semisolid gel that behaves thermodynamically similar to a crystal in equilibrium with a solution of individual tetrameric Hb molecules. Even complete deoxygenation does not convert all deoxyHbS to polymers. The insoluble phase is a collection of domains of aligned polymers, the basic unit of which is a double-strand in which two strings of deoxyHb tetramers make multiple contacts with each other ( Fig. 42.2 ).

Each HbS tetramer has two β ^S chains, the β ₁ and β ₂ . DeoxyHbS undergoes a slight structural shift so that the A helix β ^6Val “donor” site of the β ₂ chain in one tetramer can contact an EF helix “acceptor” site (formed mainly by β ^85Phe , β ^88Leu , and β ^70Ala ) in the β ₁ chain of a tetramer in the neighboring single string. This critical, lateral association can be made only when HbS is in its deoxy conformation. In HbA this EF helix hydrophobic pocket is not a favorable acceptor site for the charged β ^6Glu of the β ^A . In HbS, the β ^6Val in the β ₁ subunit is located so it cannot participate in such contacts. However, the β ₂ chain of the second single string can form chemically similar β ^6Val -dependent contacts with the β ₁ chain of the first single string. There are multiple additional axial and lateral contacts, but these are largely the same for deoxyHbA and deoxyHbS and are not themselves sufficient to stabilize a polymeric structure.

In the physiologic form of the polymer, the component strings of HbS molecules in a double-strand are half-staggered and have a slight twist, creating a fiber that is approximately 21 nM in diameter and is composed of one central and six peripheral double strands. The crystal formed in vitro lacks the twist, but its molecular structure is known in great detail.

HbS Solubility

The RBC’s hydration state dominates the physical-chemical behavior of HbS. The solubility of deoxyHbS (approximately 16 g/dL, measured under laboratory conditions) is much lower than the RBC mean cell Hb concentration (MCHC, ~34 g/dL). So, even partial cellular deoxygenation can raise deoxyHbS concentration above its solubility limit, allowing polymerization to occur. The biophysical effect of macromolecular crowding (boosting a protein’s activity far above that predicted from concentration alone) confers nonideal behavior upon cytoplasmic constituents, augmenting the likelihood for polymerization at any given degree of deoxygenation.

Equilibrium and Polymerization

In vitro studies carried out under (nonphysiologic) equilibrium conditions of stable oxygen tension and long-time scale corroborate crystallographic identification of critical amino acids involved in atomic contacts by revealing the influence of other Hbs on HbS solubility ( Fig. 42.3 ). When different Hbs are mixed together, the tetramers dissociate into dimers that intermix and randomly reassemble in a binomial distribution to reform tetramers. This clarifies the impact of naturally occurring, intracellular Hb mixtures. In a mixture of HbS and HbA, overall solubility is improved because the hybrid αβ ^S /αβ ^A tetramer integrates into polymer only one half as well as the αβ ^S /αβ ^S tetramer (see Fig. 42.3A ). The addition of HbF to HbS has a greater sparing effect because neither the αγ/αγ nor the hybrid αβ ^S /αγ tetramer can be incorporated into polymers. In this regard, HbC has the same effect as HbA, and HbA ₂ has the same effect as HbF. This sparing effect of HbA is such that much lower Hb oxygen saturation is required for a polymer to form in HbAS than in HbSS RBCs (see Fig. 42.3B ).

Kinetics and Polymerization

Laboratory measurements of polymerization kinetics, enabled by inducing (nonphysiologic) near-instantaneous and complete conversion of HbS from R (oxy) to T (deoxy) state, reveal a delay until polymer forms explosively. This inherent delay time is inversely related to an extraordinarily high power of the initial Hb concentration. For example, the delay is 10 ms at HbS 40 g/dL, but it is 100,000 seconds at HbS 20 g/dL ( Fig. 42.4A ). HbS solutions and sickle RBCs behave similarly in this regard. Delay times must vary enormously from cell to cell because they are dominated by the marked heterogeneity in MCHC (i.e., shorter delay for more dehydrated cells) and are influenced by the presence of any non-S Hb (i.e., longer delay for the presence of HbA, C, or F) (see Fig. 42.4E ). Admixture of 20% to 30% HbA with HbS (simulating HbS-β ⁺ -thalassemia) increases the delay time 10- to 100-fold, and admixture of 20% to 30% HbF with HbS increases it by 10 ³ - to 10 ⁴ -fold.

The mechanism of such polymer formation is understood to proceed by a two-step, double-nucleation process (see Fig. 42.4F ). Accordingly, the initial homogeneous nucleation takes place in bulk solution, during which small numbers of tetramers associate, with accumulation not favored until a critical nucleus size, develops (estimated to be 30 to 50 tetramers). Only then can new tetramers be added lengthwise to form a large polymer. After this occurs, heterogeneous nucleation causes explosive, autocatalytic polymer formation as new fibers form and extend on the surface of the preexisting polymer. It is the time until this explosive formation occurs that laboratory experiments detect as the inherent “delay time.”

The striking irreproducibility of long delay times reflects the underlying stochastic formation of a single (or very few) homogeneous nucleation event(s) in cells that form polymer slowly—and thus assume a sickled shape reflecting membrane deformation by highly elongated polymer (see Fig. 42.4B ). In contrast, short delay times are highly reproducible and reflect the simultaneous formation of multiple nucleation sites in cells that polymerize rapidly—and thus assume a less dramatic, irregular shape reflecting membrane deformation by multiple short polymer domains (see Fig. 42.4C and D ).