Emerging drug targets for sickle cell disease: shedding light on new knowledge and advances at the molecular level

ABSTRACT

Introduction

In sickle cell disease (SCD), a single amino acid substitution at β6 of the hemoglobin (Hb) chain replaces glutamate with valine, forming HbS instead of the normal adult HbA. Loss of a negative charge, and the conformational change in deoxygenated HbS molecules, enables formation of HbS polymers. These not only distort red cell morphology but also have other profound effects so that this simple etiology belies a complex pathogenesis with multiple complications. Although SCD represents a common severe inherited disorder with life-long consequences, approved treatments remain inadequate. Hydroxyurea is currently the most effective, with a handful of newer treatments, but there remains a real need for novel, efficacious therapies.

Areas covered

This review summarizes important early events in pathogenesis to highlight key targets for novel treatments.

Expert opinion

A thorough understanding of early events in pathogenesis closely associated with the presence of HbS is the logical starting point for identification of new targets rather than concentrating on more downstream effects. We discuss ways of reducing HbS levels, reducing the impact of HbS polymers, and of membrane events perturbing cell function, and suggest using the unique permeability of sickle cells to target drugs specifically into those more severely compromised.

KEYWORDS:

• Sickle cell disease
• HbS
• expression
• permeability
• cation loss - cell volume
• P_sickle
• KCl cotransport
• Gárdos channel - cytoskeleton
• band 3
• tyrosine kinase

1. The pathogenesis of sickle cell disease

1.1. Etiology

The etiology of sickle cell disease (SCD) is well known [1]. A mutation in one of the genes for hemoglobin (Hb) causes the sixth amino acid of the β chains of the Hb tetramer to become substituted [2]. An uncharged valine moiety replaces the negatively charged glutamic acid residue at this position, thus forming HbS instead of the normal adult HbA [3]. Loss of a negative charge at this point, allied with the shape change in the Hb protein which follows upon deoxygenation, is crucial in enabling neighboring molecules of deoxyHbS to cohere into polymers – βVal6 of one β chain making hydrophobic contacts with β’Phe85 and β’Leu88 of the other. These polymers are highly organized, 14-membered, rigid helical chains [4] with the propensity to distort red cell shape, into the eponymous sickles and other bizarre morphologies. There are also, however, other profound effects on erythrocyte function. The myriad complications of SCD comprise a complex concatenation of events, mirroring the way in which chains of HbS result in polymer formation. However, they all follow from the initial polymerization event [5–8].

The seventh codon in β Hb of HbA which encodes for glutamic acid – GAG – has a single nucleotide substitution in most SCD patients, becoming mutated to GTG and coding instead for valine, thereby producing the abnormal β chain of HbS [3]. As an autosomal gene, two copies are inherited, one from each parent. Generally, having one normal copy of the gene together with one abnormal one, and hence about half the complement of normal adult HbA and half mutated HbS, prevents disease [3] and gives rise to the so-called sickle cell trait condition (or HbAS genotype). Both genes usually need to be mutated for disease penetrance to occur [9]. The commonest genotype for SCD is the homozygous HbSS genotype, with inheritance of two identical copies of HbS, accounting for about two-thirds of patients. A number of other genotypes also occur [3,8–10]. The second most prevalent genotype is the heterozygous HbSC, in which the HbS from one parent is found in combination with HbC from the other, a mutation in which glutamic acid at the same β6 position is substituted with a basic lysine residue – a nucleotide switch from GAG to AAG – affecting about one-third of SCD patients of African origin [11,12].

Rarer genotypes of SCD are also present (see Table 1). In addition, confounding epistatic genes other than those encoding for Hb are notable in that they may affect severity, to increase or decrease morbidity, sometimes markedly. Perhaps, the most obvious is the considerable amelioration in the condition caused by mutations which result in persistence of fetal Hb, HbF – HbF prevents polymer formation; while among the commonest is probably glucose-6-phosphate dehydrogenase (G6PD) deficiency, which compromises red cell antioxidant defense, and usually exacerbates clinical signs. An important practical significance of such epigenetic factors and the existence of several SCD genotypes is that patient subsets will preferentially benefit from different treatments [13]. They may also explain key events in pathogenesis.

Table 1. Genotypes of sickle cell disease.

Genotype	Comments
HbS/HbS	Homozygous sickle cell disease (SCD)
HbS/β thalassemia	Severe double heterozygote for HbS and β° thalassemia and almost indistinguishable from SCD phenotypically; about two-thirds cases
HbSC	Double heterozygote for HbS and HbC with intermediate clinical severity; about one-third cases
HbS/β^+ thalassemia	Mild-to-moderate severity, but variable in different ethnic groups
HbS/hereditary persistence of fetal HbS/HPHP	Very mild phenotype or symptom free
HbS/HbE syndrome	Very rare and generally very mild clinical course
HbS/glucose-6-phosphate dehydrogenase (G6PD) deficiency	Heterozygote for HbS and G6PD, a metabolic enzyme involved in the pentose-phosphate pathway in RBC metabolism. Although G6PD is not an SCD genotype its deficiency markedly affects severity.
HbS/HbS Oman	A supersickling double mutant. This genotype and other examples which exacerbate the tendency to polymerize and sickle are associated with worse prognosis.
Other rare combinations	HbS with HbD Los Angeles, HbO Arab, G-Philadelphia, etc.

1.2. Polymerization

The simple etiology of SCD belies its complex pathogenesis. Key to unraveling the latter – and therefore the logical starting point for the directed design of new molecular targets for therapeutics – is a thorough appreciation of just how the presence of HbS perturbs red cell biology and which extends well ‘beyond polymerisation’ [14].

The critical event is polymerization of HbS following deoxygenation [5], but while some of the resulting sequelae are obvious, some are perhaps less so, and others doubtless remain undiscovered. The propensity of deoxygenated HbS to form polymers resulting in the reversible presence of organized, insoluble conglomerates of Hb molecules, has two immediate consequences: first, an increase in cytoplasmic viscosity [15]; and, second, a change in red cell shape [16]. Both of these severely compromise red cell rheology, lowering the ability of red cells to traverse the microvasulature and increasing red cell fragility, such that red cell lifespan is reduced from the usual 120 or so days to about a 10th that of normal [3]. These less robust red cells undergo intravascular hemolysis with release of free Hb into the plasma, while extravascular hemolysis also occurs, and although markers of red cell destruction may be found in the plasma (e.g. lactate dehydrogenase levels), the exact balance between the two is often unclear. Vascular occlusion may also follow. The best known complications of SCD are thus an ongoing chronic anemia superimposed with acute episodes of ischemia. Currently, however, it remains difficult to predict the extent of either – although certain epistatic genes like those reducing expression of alpha globin or leading to persistence of HbF are known to reduce severity while conversely co-inheritance of glucose-6-phosphate dehydrogenase (G6PD) deficiency exacerbates. Severity in affected individuals can range from subclinical to life-changing, often without known underlying reasons.

1.3. Beyond polymerization

While HbS polymerization is a paradigm of protein function/malfunction often referenced in textbooks of biochemistry, numerous secondary events follow so that red cell behavior, and that of the vasculature as a whole, is often markedly abnormal. The following summarizes some of these.

The first follows from the necessity of having an iron-containing heme cofactor to enable the oxygen-carrying function of Hb. As such, all Hbs are a possible source of oxidant Fe²⁺ but HbS is also notably more unstable than normal HbA [17]. Breakdown products of HbS accumulate within the red cell and include Hb monomers, hemichromes, and free Fe²⁺/Fe³⁺. The pro-oxidant nature of these mean that HbS-containing red cells are under far greater endogenous oxidative threat than is normally the case for circulating red cells [18,19]. In addition, considerable external oxidant challenge arises in SCD patients from ischemic-reperfusion events and the larger numbers of activated endothelial cells, as well as circulating leucocytes; while red cell antioxidant provision may also be compromised [20]. Not surprisingly, many SCD patients show evidence of marked oxidative damage in the vasculature and other tissues. Lipid peroxidation is prevalent, and products such as malondialdehydes and 4-hydroxynonenal accumulate, together with others, like those from ramifications of pathways such as the Maillard reaction.

Second, structural lipids of the membrane bilayer become disorganized with externalization of aminophospholipids, notably phosphatidylserine (PS) [21,22]. In normal red cells, lipid asymmetry confines aminophospholipids to the inner leaflet of the bilayer [23]. Abnormal function of the lipid transporter proteins (the flipase, floppase and translocase), as well as weakening of lipid anchorage, contribute to loss of normal lipid distribution. A significant, but variable, proportion of circulating red cells in SCD patients have exposed PS. Red cell cytoskeletal elements are damaged, both via the physical distortion imposed by the HbS polymers and also from secondary alterations in key protein–protein interactions. Some of these depend on changes in the association between HbS and the anion exchange transport protein (AE1 or ‘band 3’), which has other important roles outlined below. Regardless of the precise mechanisms, ultimately irreversibly sickled red cells (ISCs) accumulate in the circulation which are unable to revert to the normal biconcave disc shape even following reoxygenation [24]. These may play a disproportionate role in pathogenesis.

Third, the solute permeability of the red cell membrane is also markedly abnormal [25]. Cation permeability, in particular, becomes increased. Much of this is passive, in that primary active ATP-driven transport systems are often not directly involved, although ATP depletion can also occur and inhibit, in particular, the plasma membrane Ca²⁺ ATPase (PMCA, or Ca²⁺ pump). Often, however, ensuing solute movements reflect the prevailing electrochemical gradients of the participant ions, with a modest membrane potential of about – 10 mV (inside negative with respect to outside) owing to the Jacob-Stewart cycle [26]. The main alterations in ion and water homeostasis are summarized in the following sections, in particular for the two divalent cations, Ca²⁺ and Mg²⁺, and two of their monovalent counterparts, Na⁺ and K⁺, and effects on red cell volume.

1.4. Permeability

The homeostasis of ions and water in red cells from SCD patients is often grossly abnormal – but, again, variable between individuals. The most obvious abnormality results from the very large inwardly directed Ca²⁺ gradient of about five orders of magnitude. The normal free Ca²⁺ concentration in plasma is around 1.1–1.2 mM, compared with an extremely low intracellular level ([Ca²⁺]_i) of about 30 nM, maintained by the high capacity plasma membrane ATP-driven Ca²⁺ pump (or PMCA), with free [Ca²⁺]_i kept to around 30 nM [27]. For Mg²⁺, the gradients are considerably more modest. Free plasma Mg²⁺ is about 0.5 mM, though it is not controlled nearly as tightly as that of Ca²⁺, and lower levels may abnormally occur in SCD patients. Total intracellular Mg²⁺ is higher, at about 2 mM, although much is bound organic phosphates like ATP and 2,3-DPG, so rather less is free within the red cell cytoplasm. Estimates of the latter have been obtained by a number of different techniques [28,29], and suggest that free Mg²⁺ oscillates between about 0.3 mM in oxygenated red cells, doubling to approximately 0.6 mM upon deoxygenation, with the difference resulting from the reversible binding of certain organic phosphates, especially 2,3-diphosphoglyceric acid (2,3-DPG or 2,3-biphosphoglyceric acid) to deoxygenated soluble Hb thereby reducing cytoplasmic Mg²⁺ buffering. A rise in the membrane permeability to divalent cations which occurs in sickle cells therefore causes a gain in Ca²⁺, while, by contrast, Mg²⁺ levels may become depleted [30], through the elevated passive Mg²⁺ permeability in response to reduced plasma Mg²⁺ or via increased Na⁺/ Mg²⁺ exchange activity [30–34]. These alterations in both divalent cations have important implications for red cell volume regulation. Increased intracellular Ca²⁺ levels activate rapid K⁺ loss via the Ca²⁺-activated K⁺ channel (or Gárdos channel); low intracellular Mg²⁺ is associated with elevated loss of KCl via the KCl cotransporter [35]; both transport pathways mediate solute loss with water following, and hence red cell dehydration [25]. High intracellular Ca²⁺ levels are also implicated in lipid scrambling [23].

The commonest monovalent cations are Na⁺ and K⁺. Predominantly, renal mechanisms regulate plasma values to about 140 mM for Na⁺ and 4–5 mM for K⁺; intracellular values within the red cell are 10 and 100 mM, respectively, through the action of the ATP-driven Na⁺/K⁺ pump and other mechanisms. As membrane permeability to monovalent cations increases, especially in deoxygenated sickle cells, there is a net gain of Na⁺ and loss of K⁺ – an early, seminal finding in red cells from SCD patients [36,37]. At prevailing plasma levels of divalent cations, there is a small imbalance in these passive fluxes such that more K⁺ is lost than Na⁺ enters, contributing to a modest passive net loss of solutes, while, in addition, secondary stimulation of the ATP-driven Na⁺/K⁺ pump, with its 3:2 stoichiometry for Na⁺ efflux and K⁺ influx, worsens this imbalance [38]. Other problems with ion and water homeostasis follow. A more complete treatment of red cell ion transport systems is summarized under the next subheading, as its main significance appears to reside in its importance in the context of red cell volume regulation.

1.5. Dehydration

The important significance of the abnormal permeability of sickle cells for cations is that it mediates irrerversible solute loss and red cell shrinkage [25]. The subsequent elevation in concentration of HbS markedly promotes polymerization as the lag time for this to occur following deoxygenation is inversely proportional to a very high power of [HbS], [HbS]⁻¹⁵⁻³⁰ is often quoted [5]. A small reduction in volume will greatly encourage polymerization as there is more opportunity for the lag time to be exceeded as the red cell traverses more hypoxic regions of the circulation. In this way, the abnormal solute permeability of HbS-containing red cells occupies a critical position in pathogenesis.

A combination of three transport systems is predominantly involved in solute loss and dehydration of sickle cells (Figure 1 [25,38,39];): the deoxygenation-induced cation conductance (sometimes termed P_sickle), the Ca²⁺-activated K⁺ channel (or Gárdos channel) and the KCl cotransporter (or KCC). From a therapeutic approach, understanding the nature of these three systems and how they are controlled may inform new potential therapeutic targets, although aside from P_sickle – and may be not even for this system – none are specific to SCD patients nor to red cells. Several key features are outlined here.

Figure 1. Features of the abnormal permeability of red cells from patients with sickle cell disease. Three cation transport systems are predominantly responsible for solute loss and dehydration: KCl cotransport – probably isoforms KCC1, 3, and 4; the Ca²⁺-activated K⁺ channel or Gárdos channel – or SCNN4; and the deoxygenation-induced cation conductance sometimes called P_sickle. These, together with the Cl⁻ permeability, represent important potential targets for treatment. Phosphatidylserine (PS) is also abnormally externalized, which is partly, but not solely, Ca²⁺-dependent. A key feature of these is how they respond to changes in oxygen tension, with most particularly seen in hypoxic or deoxygenated red cells. P_sickle may be PIEZO1 and is important for increasing the permeability to both Ca²⁺ and Mg²⁺. Detailed explanations are to be found in the text (e.g. see [21,25,38,39,66,67,79,80]).

(a) The dexoygenation-induced cation conductance or P_sickle: P_sickle is a major transport system involved in sickle cell dehydration which has long remained enigmatic. It was originally characterized as a deoxygenation-induced ‘leak’ of monovalent cations [36,37]. These fluxes appear to be electrogenic and apparently activated by the sickling shape change following HbS deoxygenation and polymerization [40–42]. Its conductive nature, coupled with the apparent absence of saturation kinetics, suggest a channel-like entity [38], of which several are found within the red cell membrane [43]. The pathway is sometimes termed ‘P_sickle’ [25]. Subsequently, it was found to be permeable to divalent cations, notably Ca²⁺ and Mg²⁺, as well as monovalent cations [41]. Its main significance is probably increased entry of Ca²⁺, enabling elevation of free [Ca²⁺]_i to sufficiently high levels for activation of the third important transporter, the Ca²⁺-activated K⁺ channel or Gárdos channel. The molecular identity of P_sickle has been the subject of some speculation but, recently, a mechanosensitive cation channel PIEZO1 – which is found in many other tissues as well as red cells – has gained credibility as the possible protein mediating the P_sickle-like permeability [44,45]. If this is substantiated, it will provide a significant advance for understanding the regulation of this abnormal conductance pathway. A number of pharmacological modulators of PIEZO1 exist, both inhibitors like the tarantula spider toxin GsMTx-4 [44], which also inhibits P_sickle [46,47], and activators including Yoda1 [48], which makes it potentially amenable as a new drug target in SCD.

(b) The Ca²⁺-activatedK⁺ channel or Gárdos channel: This powerful channel is named after its discoverer, the Hungarian red cell physiologist George Gárdos. Awareness of its existence dates from the late 1950s when metabolically depleted red cells, which thereby lacked the requisite ATP to power the plasma membrane Ca²⁺ pump, were found to lose K⁺ and rapidly dehydrate [49]. Ca²⁺ accumulation in these poisoned cells will activate the Gárdos pathway which has an EC₅₀ of around 500 nM [27]. Rapid conductive loss of K⁺ ions ensues, coupled electrically by an accompanying efflux of Cl⁻ via a separate population of poorly defined anion channels. The Gárdos channel has been cloned and ascribed hSK4/KCNN4 [50], one of a number of small conductance K⁺ channels, or SK channels (previously known as intermediate conductance or IK channels). Inhibitors include charybdotoxin, clotrimazole, nitrendipine, and TRAM-34 [51–54]. Although participating in a number of other hemolytic anaemias [45], the most obvious manifestation of red cell Gárdos channel activation is found in SCD patients, following Ca²⁺ accumulation upon activation of the P_sickle pathway by deoxygenation and the sickling shape change.

(c) The KCl cotransporter: The proteins responsible for the KCl cotransporters (KCC) are encoded by the SLC12 family of solute transporters, the cation-coupled chloride cotransport family [55,56] (with SLC as the acronym for ‘solute carrier’). There are four known members, isoforms KCC1 – 4, encoded by the genes SLC12A4–7. Of these, KCC2 appears to be confined to the nervous system and is absent from the red cell lineage; all the other isoforms are also found in many other tissues including red cells. All of the others – KCC1, 3 and 4 – are probably present, with some evidence that KCC3, and to a lesser extent KCC4 and KCC1, the most significant in mature circulating red cells [57,58]. Members of the KCC family mediate electrically neutral coupled fluxes of K⁺ and Cl⁻, which given the normal K⁺ and Cl found in vivo will be outwards. KCC also responds to several stimuli which are significant to circulating red cells, notably pH, urea, and oxygen [59] and in some circumstances – but probably not in mature circulating human red cells – is associated with a regulatory volume decrease (RVD) in response to swelling [60].

In HbS-containing red cells, KCC activity is about 10–50-fold higher compared with that in cells from normal individuals [61]. This is partly, but not exclusively, due to increased expression of the protein although it is not known why [57,58]. The transport properties of KCC is also significantly abnormal in SCD patients. Two features stand out. First, KCC activity in normal red cells declines with cell age such that although it is functional in immature and young red cells, its activity declines with red cell age such that it becomes quiescent in mature erythrocytes [62], when it is refractory to the usual physiological stimuli. Notwithstanding, the transporter remains present in the red cell membrane and can be activated pharmacologically – though probably not physiologically – for example by exposure to alkylating agents like N-ethylmaleimide (NEM) [63,64]. This marked loss of activity with age is not seen in red cells from SCD patients, so that the system remains active throughout the red cell lifespan, though activity is still greater in younger (often less dense) cells [61]. KCC can therefore respond to physiological stimuli such as low pH and high urea levels. Second, KCC activity in red cells is sensitive to oxygen tension [65,66]. In normal red cells, activity is high at arterial oxygen tensions, then activity falls along with oxygen level, until its activity is lost in fully deoxygenated red cells, when it is then refractory to physiological stimuli like pH and urea. These stimuli often occur in hypoxia areas of the circulation like active muscle beds and the renal medulla. In sickle cells, the oxygen response is also abnormal – KCC is sensitive to oxygen but in cells from HbSS genotypes, though not from that of the HbSC, after a nadir in activity at around half oxygen saturation of Hb, that at lower oxygen levels increases again [39,66,67]. Overall, activity appears regulated by conjugate protein kinase and phosphatase enzymes [68–70], and possibly by the overall proximity of neighboring proteins, the so-called macromolecular crowding [71], which in red cells will be due exclusively to soluble Hb. Details remain to be completely ascertained, but the insolubility of deoxyHbS may be relevant in this context.

The interaction between these abnormal permeability pathways is well illustrated by the dependence of Gárdos channel activation on Ca²⁺ entry via P_sickle, by acidification and stimulation of KCC which will follow red cell KCl loss, and by the consequences of Mg²⁺-ATP loss. Mg²⁺ depletion will inhibit active transport processes like the Na⁺/K⁺ pump and PMCA, but has also been associated with stimulation of KCC activity [35] and hence dehydration, while Mg²⁺ loading reduces solute loss [72]. Adequate Mg²⁺ is required for hydrolysis of ATP by protein kinases, and lowered Mg²⁺-ATP presumably supports KCC stimulation via its regulatory phosphorylation cascade, although the specific targets and enzymes involved remain uncertain. Again, this is uniquely relevant to red cells from SCD patients.

1.6. Stickiness

Microvascular occlusion is a major feature of SCD affecting disparate tissues and organs. As for many complications, pathogenesis is multifactorial, but there is another obvious association with polymerization of HbS and its perturbation of red cell function, and is thus an early part of pathogenesis. Poor rheology of HbS-containing red cells and increased externalization of PS will reduce their ability to perfuse small blood vessels [15,21,22,73], while their increased fragility and intravascular hemolysis releases Hb into plasma with consequent scavenging of the vasodilator nitric oxide [74,75]. Recent therapies have been designed to target molecules involved in adhesion like the p-selectin inhibitor crizanluzimab (Table 2). The mechanism of PS exposure at least in part is Ca²⁺-induced, as low micromolecular Ca²⁺ concentrations both inhibit red cell flippase activity and stimulate scrambling [76–80] implicating a role for deoxygenation-induced Ca²⁺ entry via the P_sickle pathway. However, other mechanisms may also be important, some of which are Ca²⁺-independent with a number of protein kinases implicated in PS exposure and hence stickiness [80–82]. Nevertheless, the involvement of the abnormal membrane permeability of sickle cells in PS exposure has hitherto not been fully exploited as a therapeutic target (Figure 1). It is also important to realize that many other factors are involved in the increased adhesion of sickle cells to endothelial and other cells.

Table 2. Candidate treatments for sickle cell disease.

Treatment used	Mechanism	References
Blood transfusion	Dilute circulating HbS-containing red cells	[83]
Oxygen therapy	Inhibit polymerization	[84]
Alkali and magnesium	Inhibit polymerization; vaso-dilation	[85]
Acetazolamide	Increase oxygen affinity	[86]
Sodium bicarbonate	Increase oxygen affinity	[87]
Urea	Inhibit HbS polymerization	[88]
Cyanate	Inhibit HbS polymerization	[89]
Carbamyl phosphate	Inhibit HbS polymerization	[90]
Steroids	Inhibit sickling	[91]
Zinc	Inhibit sickling	[92]
Acetylsalicylic acid	Increase oxygen affinity	[93]
Procaine hydrochloride	Inhibit sickling	[94]
Amino acids/peptides	Inhibit HbS polymerization	[95–97]
Valerosol (12C79/BW12C) & Tucaresol (589C80)	Increase HbS O2 affinity	[98,99,186]
Pyridoxal sulfate	Inhibit HbS polymerization	[100]
Hyponatremia	Hypotonic red cell swelling	[212]
Cetiedil	Anti-sickling	[101]
Salicyclic acid	Increase oxygen affinity	[102]
Monensin (sodium ionophore)	Increase RBC solutes & hydration	[213]
l-Phenylalanine benzyl ester	Inhibit HbS polymerization	[103]
Cepharanthine	Inhibit formation of irreversibly sickled cells	[104]
5-Azacytidine	Increase HbF production	[105,177]
Ticlopidine	Anti-platelet	[106]
Bone marrow transplantation	Replace HbS with HbA	[107]
^a/^bHydroxyurea	Increase HbF production	[153–155]
Nitrendipine, nifedipine, & verapimil	Ca^2+ channel inhibitors	[62,108,109]
Erythropoietin	Stimulate F-cell synthesis	[110]
Charybdotoxin (scorpion venom)	Gárdos channel inhibitor	[51]
Bepridil	Gárdos channel inhibitor	[111]
H74	KCl cotransport inhibitor	[223]
Vanillin (p-vanillin)	Increase HbS O2 affinity	[187,197]
p-Hydroxybenzoic acid	Anti-sickling	[112]
Clotrimazole and derivatives	Gárdos channel inhibitors	[52,54]
Butyrate	Increase HbF production	[113,114]
Dipyridamole	Psickle inhibitor	[38]
Nitric oxide	Increase HbS O2 affinity; vaso-dilation	[74,115]
Anionic polysaccharides	Reduce red cell adhesion of sickle	[116]
N-Acetylcysteine	Antioxidant	[117,118]
Magnesium pidolate	Increase red cell Mg^2+ to inhibit KCl cotransport	[33,72]
Monoclonal antibodies to alphaVbeta3	Reduce red cell/endothelium adhesion	[119]
NS1652 & NS3623	Cl^− channel inhibitors	[222]
Poloxamer 188	Increase O2 delivery & reduce vaso-occlusive crisis	[120]
Arginine	Antioxidant	[121]
Heparin	Reduce red cell adhesion to P-selectin	[122]
Niprisan (Nix-0699)	Anti-sickling	[123]
ICA-17043 (‘Senicapoc’)	Gárdos channel inhibitor	[220,221]
Human immunoglobulin	Reduce leukocyte adhesion & cohesion with RBCs	[124]
Vanillin pro-drug (MX-1520)	Increase HbS O2 affinity	[125]
5-Hydroxymethyl-2-furfural (5HMF or Aes-103)	Increase HbS O2 affinity	[199,203]
^al-Glutamine	Antioxidant	[161]
Polynitroxyl albumin	Inhibit inflammation & vaso-occlusion	[126]
Eptifibatide	Anti-platelet	[127,128]
Tinzaparin (heparin/‘Innohep’)	Reduce vaso-occlusive crisis	[129]
Bosentan (endothelin receptor antagonist)	Reduce vaso-occlusive crisis	[130]
INN-298 & INN-312	Increase oxygen affinity	[131]
Propranolol	Reduce adrenaline-induced red cell adhesion	[132]
Prasugrel	Anti-platelet	[133]
TD-1	Increase oxygen affinity	[208]
^a/^bGBT440 (‘Voxelotor’)	Increase oxygen affinity	[162–164]
^a/^bCrizanlizumab (‘Adakveo’; ‘Ryverna’)	Monoclonal antibody to p-selection	[134]
GBT1118	Increase oxygen affinity	[135]
Imatinib	Tyrosine kinase inhibitor	[143]
Ticagrelor	Anti-platelet	[136]
Etavopivat and mitapivat	Pyruvate kinase activator	[137,138]

Notes: ^aindicates compounds which have been approved for SCD treatment by the Federal Drug Administration (FDA), USA, and ^bthose approved for SCD by the European Medicines Agency.

1.7. Band 3

The anion exchanger isoform 1 (or AE1), known in red cells as ‘band 3,’ is found in many tissues notably epithelia, but is especially enriched in red cells where it is present in a copy number of a few million and has an important cytoskeletal function. Among other important roles, Hb interacts with the cytoplasmic tail of band 3 (or cdb3), especially amino acids 12–23. This part of band 3 is highly negatively charged due to its richness in acidic amino acid residues [139,140]. It has also been implicated in numerous red cell functions, notably as one of the two main anchoring points for the cytoskeleton, along with glycophorin A. The oxygen affinity of Hb is important here, with that of HbS often lower compared to HbA due to the increase in red cell 2,3-DPG levels in response to chronic hypoxia, which often accompanies the anemia in SCD patients [141]. In addition, the increased binding of normal Hb to cdb3 on deoxygenation has several important consequences for red cell function [139]. First, displacement of metabolic enzymes means that under deoxygenated conditions, red cell glucose metabolism is diverted toward glycolysis and away from the pentose phosphate shunt, a critical pathway for synthesis of a major red cell antioxidant, reduced glutathione which would provide additional antioxidant defense in the form of NADPH under conditions of higher oxygen tension when greater oxidant damage may be more likely. Second, deoxyHb binding also weakens the cytoskeleton, displacing ankyrin and resulting in a more fragile red cell. Physiologically, this may enable red cells to squeeze through narrower regions of the circulation in hypoxic tissues while recently it has received attention as an important strategy of malaria to aid egress of the mature parasite [142,143], in this case using elevated tyrosine kinase activity – see later – to weaken binding. Third, deoxyHb also can displace the ‘with no lysine’ protein kinase, WNK1 [144]. The WNK1 pathway, perhaps involving a downstream OSPR1 element, is associated with inhibition of KCC (and co-ordinate stimulation of another cation-chloride cotransporter, NKCC) [144,145]. That deoxyHb binds preferentially at this site compared to oxyHb, acting as a switch [146], also has further obvious implications in SCD patients given the increased insolubility of deoxygenated HbS. Furthermore, this region of band 3 is also reversibly phosphorylated, notably at tyrosine residues 8 and 21 [140], with the involvement of protein tyrosine kinases and phosphatases. Tyrosine phosphorylation also increases in sickle cells, especially on deoxygenation [147,148]. Possible consequences are the high levels of KCC activity in deoxygenated sickle cells, weakening of the cytoskeleton and increased red cell fragility, and perturbations in redox potential. The mechanistic details and implications, and their potential as targets for therapeutic intervention, are yet to be fully explored.

1.8. Systemic abnormalities

Although β globin is only expressed in the red cell, as outlined above, Hb polymerization and the resulting damage to the red cell results in a cascade of systemic complications, such that nearly every physiological process is abnormal in SCD. These downstream abnormalities include increased inflammation [149], splenic and immune dysfunction [150], hypercoagulable blood with platelet activation [151], oxidative stress and reperfusion injury [75], vascular-endothelial dysfunction with functional nitric oxide deficiency [74], anemia and a hyperdynamic circulation, hypoxemia and tissue hypoxia [74], and abnormal and ineffective erythropoiesis [152]. All these are potential therapeutic targets in SCD, and also result in progressive organ damage, which in turn creates a range of targets to lessen the impact of SCD: this further includes damage to nearly all organs including the vascular tree, lungs, liver, brain, bones and joints, and kidneys. These downstream possibilities have received considerable attention and a few recent drugs are now available, notwithstanding they have generally proved less fruitful than anticipated – or at least new reliable treatments based upon them are few – perhaps because they are further from the initial critical events in pathogenesis.

2. Current management

In most high-income countries, current management has greatly increased both the lifespan and standard of patients’ lives. Table 2 presents a summary of the development of SCD treatments over the last century. The multiplicity and diverse nature of these various approaches beautifully illustrate how problematical it has proved to design satisfactory, effective treatments for SCD. Routine treatment regimes begin at an early age with prophylaxis against pneumococcal infection and provision of appropriate antibiotic cover, and are coupled with symptomatic supportive treatments reflecting individual complications. Hydroxyurea is now commonly used [153,154]. Previously it was reserved for those with overt clinical problems, although clinical trials now suggest that hydroxyurea is beneficial for everyone with sickle cell anemia (HbSS and HbS/β-thalassemia), and most guidelines recommend starting hydroxyurea in the first few years of life [155–157]. For less well-controlled patients, regular blood transfusions may be used to correct anemia and dilute the proportion of circulating sickle cells, to reduce symptoms or prevent progressive organ damage [158]. Hematopoietic stem cell transplantation offers a curative option, which at present remains mostly for those with HLA-identical siblings and with significant complications, although alternative donors are increasingly used [159,160].

These measures provide a level of protection far superior to that of several decades ago. Notwithstanding, it is widely accepted that more is required and recent additions to drugs licensed for SCD patients include l-glutamine, voxelotor, and crizanlizumab.

l-Glutamine may protect against oxidative damage, through increases in red cell levels of reduced glutathione [161]. One randomized controlled trial suggested a modest reduction in the frequency of painful episodes in adults with SCD, and the drug is licensed for use in the USA and some other countries. The drug has not, however, been approved for use by the European Medicines Agency, because of concern about the larger number of patients dropping out of the treatment arm of the study compared to placebo, and the methodology by which this difference was taken account.

Voxelotor is an Hb adjunct which binds to the α globin chain and increases its oxygen affinity and as oxyHb does not form polymers this reduces Hb polymerization and sickling [162,164]. This results in an increase in Hb of 1−2 g.dl⁻¹, with a 10–20% reduction in reticulocyte count. Although this is potentially beneficial and the drug is increasingly widely used, there are some concerns about whether voxelotor improves or impairs oxygen delivery to some tissues. Notwithstanding, there have been recent reports on clinical improvement in patients with SCD [165–168].

Crizanlizumab is a monoclonal antibody developed against the adhesion molecule p-selectin [134]. It is given as an intravenous infusion every 4 weeks, and has been shown to reduce the frequency of acute pain by about 40%, with relatively few side-effects. P-selectin expression is seen, in particular, in activated endothelial cells and platelets and is involved in many disease processes, not just in SCD. The long-term safety and benefits of crizanlizumab are not yet known.

Gene therapy is also emerging as an effective treatment in clinical trials, with current strategies including gene edition to increase the amount of non-sickling Hb [169], and gene editing focused on boosting HbF production. In the longer term, it may be that gene editing to directly correct the sickle mutation will become established as the treatment of choice, possibly delivered directly in vivo.

A number of clinical trials are currently under way, for example for etapivat and mitapivat. A full list of these can be seen at the clinical trials’ websites (such as ClinicalTrials.gov [229]).

In summary, with hydroxyurea established as the most effective treatment, and despite a few newer treatments emerging over the last few years, none of these options are curative or transformative, and there remains a real need for new, more effective approaches.

3. Emerging molecular targets

Rational identification of new molecular targets requires an appropriate strategy [170]. The etiology and the precipitating cause of all the complications of SCD is the sickle mutation in the β globin gene. Polymerization and instability of HbS probably underlie most of all subsequent problems, with direct effects on the red cell cytoplasm, cytoskeleton, and membrane. What follows thereafter, notwithstanding its importance to disease progression, are secondary downstream consequences of HbS, many of which are not pathognomonic to SCD but are also shared with other vascular or systemic disease processes – which is not to belittle their clinical and therapeutic importance but which perhaps serve to deflect attention from more direct consequences of having a high concentration of an abnormal protein, HbS, within the developing red cell. In the following, therefore, we focus on several areas more specific to SCD: reducing HbS levels; reducing the impact of HbS polymers on red cell function; reducing the impact of membrane events which contribute to pathogenesis; and using the unique permeability of HbS-containing red cells – as an access for drug delivery targeted specifically to sickle cells.

3.1. Drugs which modulate globin expression qualitatively

Regulation of globin genes is complex with different ones sequentially switched on and off during development [171]. Adult Hbs – HbA normally, though HbS in SCD patients – are not transcribed until around parturition, before which various embryonic and fetal globins predominate, notably fetal Hb, HbF, with its γ chain replacing the normal adult β. HbF has particular significance in that it antagonizes the growth of HbS polymers so its presence in sufficient quantities ameliorates or even prevents the complications of SCD. As the transition from fetal HbF to adult Hb, to the normal HbA or the mutated HbS, is usually peripaturient, fetuses and early neonates are free from disease. Hb switching has therefore been well studied, and while its complexities remain unclear, the occurrence of natural persistence of fetal Hb in some individuals raises the possibility of artificially mediated HbF expression as a potential therapy. A number of regulatory genes have been found and include BCL11A, MYB, KLF1, ZBTB7A, and NuRD, many of which represent transcription factors which suppress expression of the fetal variant and favor that of adult Hb [172–174]. Clinical trials are continuing, for example with viral vectors to silence BCL11A [175].

A number of pharmacological agents interfere with DNA processing. Early work was based around the discovery of 5-azacytidine – which is still being tested in combination with tetrahydrouridine [176] – and similar compounds acting against various leukemias, through effects on synthesis [177]. Hydroxyurea represents an extension following on from these and is now one of a handful of licensed drugs for treatment of SCD [153,154,178,179]. Notwithstanding its success and clinical use over more than 30 years, the exact mechanism of action of hydroxyurea still remains unclear. It is a ribonuclease reductase inhibitor, and as such, like 5-azacytidine it is associated with inhibition of DNA synthesis. Furthermore, on the one hand, while therapeutically, hydroxyurea acts at least in part through increased expression of HbF, levels are frustatingly variable between patients, and more efficacious compounds are sought; on the other, supplementary beneficial effects of hydroxyurea also probably occur including better red cell hydration and reduced levels of white cells and the relative importance of these is not clear.

More recent progress has highlighted a number of potential targets which modulate HbF expression including DNA methyltransferase inhibitors, histone deacetylase inhibitors, euchromatic histone-lysine-N-methyltransferases, immunomodulators like thalidomide and its derivatives and reagents acting on the cGMP signaling pathways like the phosphodiesterase-9 inhibitor IMR-687 [180]. A full discussion of these is beyond the scope of this review and readers are directed to several recent publications (e.g. [181,182]).

There is also emerging evidence that erythropoiesis is abnormal or ineffective in SCD. Ineffective erythropoiesis (IE) seems to be caused both by intramedullary HbS polymerization and inflammation triggered by infection, infarction, and hemolysis. While disadvantageous in contributing to bone marrow abnormalities and anemia, IE may usefully lead to increased expression of HbF. A fuller understanding of the mechanisms involved may well suggest future therapeutic targets to increase selection of cells containing more γ globin/HbF [183].

More recent approaches are provided by direct genetic manipulation, specifically inhibiting the molecular apparatus associated with silencing of the HbF gene. One of these relies on provision of siRNA to reduce levels of transcription factors which suppress HbF, like BCL11A. These manipulations are already possible in vitro and could be invaluable clinically. A further technique is correction of the HbS mutation within the genome of erythropoietic stem cells. Caspase 9/CRISPR methodology allows the abnormal nucleotide sequence in the β globin gene encoding for valine of HbS to be amended back to the glutamic acid of HbA. To date, however, this approach has been limited by poor efficiency of homologous recombination at this site. A newer, more effective approach has been proposed recently whereby the corrective edit involves use of other naturally occurring Hb alleles, such as HbG Makassar allele [169].

3.2. Drugs which modulate globin expression quantitatively

The circulating level of HbS can also be lowered by less specific measures to limit synthesis. Iron is a necessary cofactor for heme synthesis and one such approach is to restrict its supply. Because of its potential toxicity as an oxidant, iron homeostasis is under tight physiological control and is particularly dependent on regulated uptake from the small intestine (the ‘iron curtain’). A number of specific proteins determine the magnitude of entry into enterocytes, intracellular chelation to facilitate shuttling from luminal to basolateral faces of the epithelial cells, and export into plasma. The latter involves a unique ferroportin transport protein. Inhibitors such as vamifeport (VIT-2763) have been targeted at this site and appear to produce qualitatively different effects to those produced by simple dietary restriction. Studies in a mouse model suggest that vamifeport has beneficial effects against SCD via this mechanism [184], although translation into the clinic remains at a very early stage.

IE is also a potential target to increase expression of useful Hbs. Although hemolysis predominates in SCD, there is evidence that anemia is partly due to IE, with excessive loss of certain erythroid cells during development in bone marrow. Controlling this might potentially increase Hb levels, although there is a danger that most of the increase will be in HbS levels, which could be detrimental if the hematocrit increases significantly. Drugs such as luspatercept which enhances erythroid maturation via the SMAD pathway are aimed at this target [185].

Finally, regulation of erythropoiesis involves several hormones which include vitamin B₁₂, hepcidin, and erythropoietin. They may have the potential for correcting Hb in SCD, but this has not been fully explored.

3.3. Drugs which target HbS

HbS only polymerizes at low oxygen tension, when the relaxed (R) to tense (T) shape change of Hb occurs – with this transition so named because the deoxygenated heme moiety assumes a domed, taut configuration. The conformational modification exposes a hydrophobic pocket of one β chain which accepts the protruding mutated β6 valine of another, thus enabling neighboring HbS molecules to cohere. Irreversibly sickled cells do occur at arterial oxygen tensions, but these are probably secondary alterations due to persistence of modifications in the cytoskeleton rather than of HbS polymers under oxygenated conditions. Stopping the deoxy shape change of Hb would therefore prevent polymerization and has been a strategy pursued for over 50 years [186–191].

This approach has led to the generation of a number of ‘left-shift’ reagents, because they result in a leftward displacement of the oxygen binding curve of Hb to give an increased oxygen saturation at prevailing oxygen tensions. The increase in oxygen affinity must be enough to reduce polymerization at moderate hypoxia but not so great that it reduces oxygen unloading in tissues where it is required. An additional caveat of these compounds is that their availability and affinity for Hb must achieve sufficient levels to react with the large amount of Hb present in the circulation [192], although more recent estimates of the relative inefficiency of polymer elongation argue that modification of only a small fraction of total HbS is required for their effective [193–196].

Various compounds have been proposed (Table 2), especially following the identification of vanillin [186,187,197–200]. Many successfully increase Hb oxygen affinity but their application has been limited through poor pharmacokinetics or unexpected deleterious side-effects [201,202]. Relatively few have progressed to clinical trials and, to date, of these the most successful ones have been Aes-103/5-hydroxymethylfurfural (5-HMF) [199,202–205] and a compound now licensed as voxelotor (originally ascribed GBT440 from Global Blood Therapeutics) [162,164]. Voxelotor successfully increases Hb levels by a few grams per liter but other clinical benefits are less clear, and an increase in Hb levels alone may not increase oxygen delivery to tissues nor confer clinical benefit [206]. The search continues for newer adjuncts using high throughput screening and computer-design (e.g. TD-1 and successors [207,208]). These drugs are not specific for HbS over HbA as such but modulation of oxygen affinity is unlikely to have deleterious effects on other Hbs.

A modification of this approach is the use of peptides and other small molecules to target the cohesive points between neighboring HbS molecules, rather than to increase oxygen affinity per se [209]. These may involve the best known primary interaction between β6Val with β’Phe85 and β’Leu88, or secondary contacts between other parts of the Hb tetramer. As yet, none have progressed to clinical trials.

Finally, there is also emerging evidence that pyruvate kinase activators have clinical potential in SCD, such as etavopivat and mitapivat. These drugs offer the benefit of increasing pyruvate kinase activity in red cells, which reduces 2,3-DPG and increases ATP levels in erythrocytes. Reduced 2,3-DPG levels lead to increased hemoglobin oxygen affinity, which inhibits hemoglobin polymerization. Two such drugs, mitapivat and etavopivat, are entering early phase clinical trials following encouraging results in cell lines and mice [210,211]. Potential benefits appear to include reduced vaso-occlusion and hemolysis.

3.4. Drugs which target red cell dehydration

The rationale underlying the use of these reagents for ameliorating the progressive pathogenesis of SCD is three-pronged: first, there is a marked dependence of polymerization on the concentration of HbS – the lag time to polymerization following deoxygenation is inversely proportional to a very high power of [HbS] ([HbS]⁻¹⁵⁻³⁰ is often quoted [5]); second, the abnormally high permeability of HbS-containing red cells mediates relatively rapid solute loss and this through specific transport pathways, with water following osmotically to cause cell dehydration; and third, there is no obvious mechanism for shrunken red cells to recover lost volume. Early approaches targeting red cell hydration as a therapeutic weapon against SCD used the simple strategy of water overload [212]. Later, ionophores were tested which partition into any membrane, including that of the HbS-containing red cell, and result in solute and water gain (e.g. monensin [213]). Latterly, the strategy of using specific transport inhibitors has been preeminent. Until the recent interest in PIEZO1 [45]; however, there has existed an apparent anomaly or paradox in that the etiology of SCD, a mutation in Hb, lacks any obvious direct assocation with membrane transport properties. Establishing the connection is important. A further problem arises from the fact that of the main systems participating in solute loss – P_sickle, KCl cotransport and the Gardos channel – only P_sickle was thought unique to sickle cells and; in fact, even this is probably not the case, which raises the significant problem of specificity, to target pathways solely within sickle cells.

P_sickle inhibitors: The P_sickle-like pathway activated in deoxygenated sickle cells was found to be partially inhibited by band 3 inhibitors such as 4,4′-diisothiocyano-2,2′-stilbenedisulphonate (DIDS) and dipyridamole [38,42]. This finding, together with the fact that mutated band 3 molecules may function as ion channels [214], has been partly responsible for the hypothesis that P_sickle might be a manifestation of the anion exchanger. More efficacious compounds were lacking, however. Normal red cells also show a cation conductance, activated physiologically by mechanical distortion, for example by shear stress [215,216], rather than pathologically by the presence of rigid HbS polymers and which is permeable to Ca²⁺. The mechanosensitive channel PIEZO1 is currently the most likely contender for this pathway in red cells, and, by extension, a candidate for P_sickle. However, although some studies have linked common PIEZO1 alleles to SCD patients [217], others found no correlation with P_sickle-like activity [218]. The PIEZO1 entity comprises a very large protein, comprising about 2500 amino acids and a mass of some 300 kDa, with a complex mode of activation and kinetic behavior. Small molecule inhibitors of PIEZO1 exist, like the naturally occurring tarantula venom GsMTx4 and the synthetic partial antagonist DOOKU1 [219], and have similar actions against the P_sickle permeability of deoxygenated sickle cells. No compounds have yet been targeted specifically at the channel in sickle cells. Nonetheless, increased understanding of how this system becomes activated and inactivated has the potential for new effective drugs against SCD.

Gárdos channel and anion conductance inhibitors: The second important pathway is the Gárdos channel which is also found in red cells from normal individuals as well as SCD patients [49]. Usually, the Gárdos channel is quiescent because of low intracellular Ca²⁺ in red cells due to the high capacity PMCA coupled with a very low passive Ca²⁺ leak. Activation of P_sickle in deoxygenated sickle cells alters this balance so that excess Ca²⁺ entry mediates a rise in [Ca²⁺]_i and an increase in Gárdos channel activity. Effective inhibitors of the channel include compounds like the scorpion venom charybdotoxin and the antimicrobial clotimazole [51–54]. In clinical trials, prolonged aministration of clotrimazole led to signs of liver toxicity due to the presence of its imidazole group. Derivatives lacking this side group have been developed and ICA-17043/Senicapoc progressed to clinical trials in SCD patients . Senicapoc effectively increased red cell hydration. The end point chosen in these trials, however, was reduction in pain which was not forthcoming [220,221]. If some other indicator of efficacy had been chosen, such as increased Hb levels (as used for voxelotor) or specific tissue damage, the outcome might have been different. Senicapoc is currently being reassessed as possible treatment for hereditary xerocytosis, and continues to be an interesting option in SCD.

The Gárdos channel mediates conductive K⁺ efflux and is therefore limited by hyperpolarisation so that for higher rates of efflux it must be accompanied by an anion via separate channels. The red cell shows a high background anion conductance although the molecular identity of the anion pathway/pathways involved, perhaps surprisingly, remains unclear. A possible candidate for this pathway is slippage through the anion exchanger, band 3, mediating uncoupled efflux of Cl⁻, rather than conductive anion flux. Compounds to inhibit red cell anion conductance, including DIDS but also higher affinity ones such as NS1652 [222], have been investigated, but the high degree of inhibition required for effect may not be feasible in vivo and none have not yet progressed to trials in SCD patients.

KCl cotransport inhibitors: The third potential target for specific transport pathway inhibitors is KCC. Like PIEZO1 and the Gárdos channel, KCC is found in red cells from normal individuals and also in other tissues, notably transporting epithelia [56]. KCC is also homologous to the Na⁺-linked cotransporter, NKCC, which is present in two isoforms, a housekeeping NKCC1 present in many tissues and a renal-specific NKCC2, the target of the loop diuretics. A major problem, again, therefore is one of specificity. Nevertheless, its significance as one of the three main dehydration pathways of sickle cells warrants attention. A number of bumetanide/furosemide analogues were synthesized by Hoechst in the 1980s [223]. One of these, H74, was shown to be effective in inhibiting KCC in human red cells, with good specificity over NKCC1 but no further developments have been reported. KCC is also notable in its regulation by protein phosphorylation [68–70], and the potential for this as a target is discussed in the next section.

3.5. Drugs which target the red cell cytoskeleton

Red cell fragility increases upon deoxygenation and this is particularly noticeable in HbS-containing red cells. Stability depends on the cellular network of spectrin/actin molecules which is found in juxtaposition to the membrane, to which it is attached through band 3 and glycophorin A [224]. These attachments are all non-covalent and labile, enabling the cytoskeleton to remodel and reform as red cells are squeezed through passages in the microvasculature whose diameter may be considerably less than that of the unstressed red cell (some 8 μm). DeoxyHb binding to band 3 weakens cytoskeletal attachments and may be one of physiologically important effects of red cell oxygen sensitivity. Recently, the effect of tyrosine phosphorylation has received some attention, in particular at residues 8 and 21 [142,143] Phosphorylation of tyrosine residues is increased in the membrane of sickle cells and further by deoxygenation [147,148]. It has also been shown recently that one strategy of the malaria parasite is to increase band 3 tyrosine phosphorylation, to weaken membrane integrity, and which is advantageous for enabling escape of the maturing parasite. Genes for tyrosine kinases are absent from the malaria genome and the parasite makes use of endogenous enzymes, probably following inhibition of their conjugate tyrosine phosphatases by increased oxidative stress. Tyrosine kinase inhibition, to impede parasite release, has been suggested as a potential malaria therapy [142] – for example, with imatinib developed as a leukemia treatment. Imatinib also acts concomitantly to increase red cell deformability in shear stress assays [143]. The latter may provide a rationale for its use in SCD patients, and it has been used in clinical trials. Interestingly, this site in band 3 is also a binding site for WNK1, an inhibitor of KCC, so that targeting this site may also modulate the activity of this important dehydration pathway. Further focus on this cytoplasmic terminus of band 3 is a logical approach.

3.6. Drugs which make use of the unique permeability of HbS-containing red cells

Normal red cells from humans and some other species suspended in low ionic strength (LIS), such as isosmotic sucrose, exhibit new permeability pathways not apparent when they are incubated in salt solutions [225,226]. When similar studies are made in HbS-containing red cells in low ionic strength media, a novel permeability is also revealed but, uniquely, it requires deoxygenation [227,228]. A number of properties are reminiscent of P_sickle – for example oxygen dependence, partial inhibition by DIDS, dipyridamole and left shift reagents, and stochastic activation. It is also permeable to a number of di- and tri-peptides which cannot usually gain access to the red cell interior, and which allow estimation of a putative pore size. This pathway has been proposed as a potential diagnostic tool for SCD, amenable to simple assays which could be carried out in less developed countries [228]. It also provides a potential means of access to the red cell cytoplasm – a sort of ‘Trojan horse’ – for the introduction of inhibitors and other compounds to target the abnormal permeability features of sickle cells. The advantage of this would be that only sickle cells would be accessible. This avenue remains to be investigated fully.

4. Conclusion

In section 1, this article reviews the early events in the pathogenesis of SCD in the belief that key initial events may provide a fruitful focus for new treatments. Section 2 continues as a brief summary of currently licensed drugs. Section 3 represents the main discussion and covers ways of reducing HbS levels, reducing the impact of HbS polymers on red cell function, reducing the impact of membrane events which contribute to pathogenesis, and using the unique permeability of HbS-containing red cells as a potential drug delivery pathway specific to sickle cells, as areas which in our opinion are most likely to yield effective new therapies.

5. Expert opinion

SCD has a considerable impact on health worldwide but also in the USA, the UK, and other areas of Northern Europe. Nevertheless, it is categorized as a ‘rare disease.’ As such, it has received less attention than it might otherwise deserve, although this attitude appears to be changing. The mutated hemoglobin, HbS, is a biochemical paradigm, well known for its ability to polymerize upon deoxygenation, and with a single base mutation causing a single amino acid substitution. This simple etiology, however, belies a complex pathogenesis which extends well beyond polymerization. Unraveling critical early events in pathogenesis is a logical starting point for identifying new molecular targets. As yet, these are still not adequately understood. In this review, we have focused on new and underexplored red cell changes.

Among the red cell changes which follow HbS polymerization are increased stickiness and phosphatidylserine exposure, oxidative damage, fragility, increased cation permeability and shrinkage, and a unique non-electrolyte permeability. HbS also interacts with the cytoplasmic tail of band 3, a region implicated in numerous red cell functions – as a main anchor site for the cytoskeleton, a scaffold for major enzymes involved in glycolysis and antioxidant provision, and also for kinases which modulate membrane permeability. Tyrosine phosphorylation of band 3 increases in sickle cells, especially on deoxygenation. The mechanistic implications and potential for therapeutic intervention of these events, however, are yet to be fully explored. Future work will surely address these key roles of band 3.

Older supportive measures – pneumococcal prophylaxis, antibiotic provision and access to blood transfusions – have markedly reduced mortality and improved the quality of life of SCD patients. More recently, a number of newly licensed drugs, including l-glutamine, crizanlizumab, and voxelotor, as well as hydroxyurea, have become available but none reach the requisite level of efficacy. None of the current treatment options therefore are curative or transformative, and there remains a real need for new, more effective approaches. This point is inadequately understood even by many clnicians.

HbS is an adult form of Hb with earlier ones, including HbF, expressed in the embryo and neonate. HbF does not polymerize and reduces the tendency of HbS to do so, and the value of hydroxyurea is probably mainly through increased HbF expression. More recent approaches are provided by more direct genetic manipulation, such as provision of siRNA to reduce levels of transcription factors like BCL11A which suppress HbF. These maneuvers are already possible in vitro and could become invaluable clinically. A further strategy is correction of the HbS mutation within the genome of erythropoietic stem cells, for example using caspase 9/CRISPR methodology to edit in other naturally occurring Hb alleles, such as HbG Makassar. Circulating level of HbS can also be lowered by less specific measures to limit synthesis like inhibiting iron absorption with transporter inhibitors such as vamifeport targeted against ferroportin proteins. Increasing use of immortalized red cell lines amenable to genetic manipulations will expedite future research.

Other treatments aim to reduce the propensity of HbS to polymerize. Compounds to increase oxygen affinity of HbS and thereby prevent polymerization have been studied for some time and continue to provide a focus of interest. The newly licensed voxelotor works in this way and does raise circulating Hb levels, but doubts remain about its clinical utility, with concern that the reduction in oxygen unloading might offset any potential gains. Others continue to be of interest in clinical trials [229]. A modification of this approach is the use of peptides and other small molecules to target the cohesive points between neighboring HbS molecules, rather than to increase oxygen affinity per se, but, as yet, none have progressed to clinical trials. Lastly, there is also emerging evidence that pyruvate kinase activators may be helpful by increasing hemoglobin levels via the reduction of erythrocyte 2,3-DPG concentration, together with increased ATP levels, leading to reduced complications of anemia and improved red cell health. Mitapivat and etavopivat both act in this way and are entering early-phase clinical trials. Recent reports of beneficial clinical responses to these reagents is encouraging.

The increased cation permeability of red cells to mediate solute loss, red cell shrinkage and hence [HbS], thereby reducing the lag to polymerization following deoxygenation, also remains a possible target. The specific transport system involved – the Gardos channel, and P_sickle, which may or may not be the manifestation of PIEZO1, and the KCl cotransporter – have been all targeted some time ago, using senicapoc, dipyridamole, H74, and other reagents. In particular, the Gardos channel inhibitor, senicapoc, caused increases in hemoglobin and reduced hemolysis but was not developed beyond early-stage trials because it did not reduce the frequency of vaso-occlusive pain; however, in retrospect, its physiological effects seem fairly similar to voxelotor, and it may be worth revisiting its use in SCD. Some of these compounds may remain viable alternative targets, perhaps in combination therapy with drugs based on other mechanisms of action.

Tyrosine kinase inhibitors, for example imatinib, to reduce band 3 phosphorylation and stabilize the cytoskeleton are also currently under clinical trials. Interestingly, however, this site also binds WNK1, a kinase inhibitor of KCl cotransport, and increased phosphorylation to displace the enzyme may therefore co-incidentally alter cotransporter activity. In addition, tyrosine kinase inhibition, in the opposite way to serine-threonine kinase inhibition, can itself lower transport activity. The exact pathways involved, their interactions, and the degree to which their modulation may be beneficial or not requires further elucidation. The many roles of band 3 in red cell physiology and pathology remain incompletely understood and represent a rich ground for future work, as evidenced by recent advances in clinical trials for malaria and SCD using tyrosine kinase inhibitors.

Finally, it may be possible to allow drugs to specifically target sickle red cells by using the unique permeability activated only in sickle cells by incubation in deoxygenated non-electrolyte media, and which appears to involve cytoskeletal elements. This pathway, by allowing drugs to access the cytoplasm, would target them exclusively to the interior of sickle cells, and hitherto has not been explored. Other ways of permeabilizing red cells, such as lysis and resealing or electroporation, are currently under investigation for introducing drugs into the interior, but this novel pathway has not been exploited hitherto.

For all these approaches, advances will be highly reliant on methodological improvements, with high throughput screening techniques – for example, using mechanized patch clamp rigs – advantageous for expediting rapid progress. That there is currently some willingness to expend resources on SCD as an important health problem, allied to some of the more rational approaches discussed here, provides optimism for the future.

Article highlights

• This article reviews the early events in the pathogenesis of sickle cell disease in the belief that key initial events may provide a fruitful focus for new treatments.

• Drugs currently in clinical use are summarized briefly.

• The main discussion covers potential areas where future research may uncover novel treatments.

• In particular, the review concentrates on ways of reducing HbS levels, reducing the impact of HbS polymers on red cell function, reducing the impact of membrane events which contribute to pathogenesis, and using the unique permeability of HbS-containing red cells as a potential drug delivery mechanism – as areas which in our opinion are most likely to yield effective new therapies.

Declaration of interests

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

Reviewer disclosures

A reviewer on this manuscript has disclosed Forma: Research funding from

GBT: Research funding and Honoraria; Novartis Research funding; Amgen Research funding; Agio Honoraria; Sanofi Honoraria.

A second reviewer has disclosed the following: Consultant – DisperSol; Hilton Publishing – scientific advisor; Grant/Research Support – Global Blood Therapeutics, Eli Lilly, Forma Therapeutics; Baxalta – sub-investigator; Other – Aruvant – Data Safety Monitoring Board member.

Peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Acknowledgments

JS Gibson is grateful to Dr Kat Connolly for details concerning the Maillard reaction.

Additional information

Funding

This work was funded by grants from the Medical Research Council (G0901177) and the British Heart Foundation (31966).