Treatment Strategies for Glucose-6-Phosphate Dehydrogenase Deficiency: Past and Future Perspectives

Abstract

Glucose-6-phosphate dehydrogenase (G6PD) maintains redox balance in a variety of cell types and is essential for erythrocyte resistance to oxidative stress. G6PD deficiency, caused by mutations in the G6PD gene, is present in ~400 million people worldwide, and can cause acute hemolytic anemia. Currently, there are no treatments for G6PD deficiency. We discuss the role of G6PD in hemolytic and non-hemolytic disorders, treatment strategies attempted over the years, and potential reasons for their failure. We also discuss potential pharmacological pathways, including glutathione metabolism, compensatory NADPH production routes, transcriptional upregulation of the G6PD gene, highlighting potential drug targets. The needs and opportunities described here may motivate the development of a therapeutic for hematological and other chronic diseases associated with G6PD deficiency.

G6PD Deficiency as a Hematologic Disorder

Glucose-6-phosphate dehydrogenase (G6PD; see Glossary) regulates glycolytic flux through the pentose phosphate pathway (PPP) and plays a central role in redox homeostasis; it produces nicotinamide adenine dinucleotide phosphate (NADPH), an essential cofactor for glutathione (GSH/GSSG) regeneration. G6PD is a major source of NADPH and loss of G6PD function is detrimental in erythrocytes; it causes G6PD deficiency (G6PD^def), a potentially hemolytic disorder [1].

G6PD^def is an X-linked genetic disorder present in 300-400 million people worldwide [2]. To date, over 200 non-synonymous mutations in the G6PD gene have been reported [3] and in attempt to understand the genotype/phenotype correlation, the majority of mutations have been classified according to their residual enzymatic activity and clinical outcome as outlined by the WHO: Class I, <10% residual activity with chronic hemolytic anemia (CHA), Class II, <10% residual activity with intermittent acute hemolytic anemia (AHA), and Class III, 10-60% residual activity with AHA in response to an oxidative trigger [4]. Although Class I variants are the most severe, they are rare. Their low allele frequency is likely the result of G6PDs essential role in development; knock-down of G6PD is embryonic lethal in mice [5]. In contrast, Class II and III variants are more common, however they are largely asymptomatic throughout life when following preventative measures – dietary/treatment restrictions [1].

Although Class I mutations are rare and Class II/III largely asymptomatic, self-resolution from AHA is not guaranteed and blood transfusions not always viable. Additionally, treatment with 8-aminoquinolines, the only anti-malarial drugs capable of eliminating the dormant liver forms of Plasmodium Vivax, induces AHA in G6PD^def, with G6PD^def found in geographic areas with a high incidence of malaria [6,7]. This, combined with lack of reliable testing for G6PD^def, presents a major challenge in the eradication of malaria [8,9]. Thus, prophylactic or combination therapies may be attractive strategies for treatment of G6PD^def and the eradication of malaria.

Although hemolysis is the main human pathology associated with G6PD^def, G6PD^def has been implicated in other chronic diseases [10]. Additionally, G6PD^def has been thought to contribute to Coronavirus Disease 2019 (COVID-19) complications. Despite these health risks, G6PD^def is not routinely screened for and there are no active efforts to develop treatments to address it. The belief that common variants have a mild pathology that can be overcome by avoiding trigger foods and drugs or treated with blood transfusions during a hemolytic crisis hinders the development of a therapeutic. Here, we review past treatment strategies and discuss molecular pathways and potential drug targets that may motivate the development of therapies for G6PD^def-related disorders.

The effect of G6PD^def in Non-Hemolytic Disorders

The contribution of G6PD^def to the pathophysiology of chronic diseases will only be briefly discussed, as dedicated reviews exist for G6PD^def in viral infections [11], hyperbilirubinemia/kernicterus [12], diabetes [13], cardiovascular disease [14] and neurodegeneration [15].

Viral Infections, including COVID-19

G6PD^def may increase cell susceptibility to viral infections such as hepatitis B [16], dengue fever, enterovirus 71, and coronavirus 229E [11]. Additionally, viral infections may trigger hemolytic complications in G6PD^def patients [17-19]. Recently, two case-reports found that COVID-19 triggered hemolysis and methemoglobinemia in G6PD^def patients without an identifiable eliciting drug [20,21]. Methemoglobinemia has been reported in other COVID-19 cases [22] and anti-malarial drugs known to induce hemolysis in G6PD^def, hydroxychloroquine and chloroquine, have been used to treat COVID-19. To date, seven case studies reported hydroxychloroquine/chloroquine triggered hemolysis in G6PD^def COVID-19 patients [23-27].

Neonatal Jaundice, Kernicterus, Hyperbilirubinemia

G6PD^def is a major and well-known risk factor for jaundice and hyperbilirubinemia in newborns as hemolysis increases bilirubin levels in the serum [28]. In newborns, bilirubin can cross the blood brain barrier and cause neurological issues such as kernicterus or acute bilirubin encephalopathy [29]. Interestingly, newborns with kernicterus have a higher chance of developing sensorineural hearing loss [30], suggesting a role of G6PD^def in auditory disorders. Currently, the pathophysiology for kernicterus/hyperbilirubinemia-induced sensorineural hearing loss is unknown and only a few studies have explored the role of G6PD^def in auditory disorders [31,32]. Additionally, the role G6PD plays in kernicterus-induced brain damage remains elusive.

Diabetes

High glucose levels inhibit G6PD activity and increases oxidative stress [33]. Glucose-mediated G6PD inhibition reduced insulin secretion from ß-islets and induced ß-cells apoptosis, with overexpression of G6PD rescuing high glucose mediated dysfunction [34]. Additionally, glucose-mediated G6PD inhibition resulted in elevated ROS in podocytes resulting in podocyte injury [33], a complication of diabetes that can lead to diabetic kidney disease [35]. These findings suggest that glucose-mediated inhibition of G6PD and G6PD^def contribute to loss of ß-cell function and may increase susceptibility to diabetic complications. G6PD^def has been reported as a risk factor for diabetes [14].

Cardiovascular Diseases

Aldosterone has been found to inhibit G6PD activity, with G6PD inhibition reducing bioavailable nitric oxide, vascular reactivity, and the contractility of vascular smooth muscle cells, coronary arteries and cardiomyocytes [14]. Additionally, G6PD activity modulates genes related to thrombosis, atherosclerosis, contractility, and blood vessel calcification [36] and mice with reduced G6PD activity develop spontaneous pulmonary hypertension with pulmonary artery and right heart remodeling [37]. However, it remains unclear if dysfunctional reactivity and contractility caused by G6PD^def lead to a pathogenic phenotype in humans as G6PD^def has been reported to be both cardio-protective and cardio-damaging [14,38].

Neurodegenerative Disorders

Neurons are highly dependent on G6PD and the PPP to combat ROS [15]. Inhibiting G6PD in neurons increases ROS, induces apoptosis, and induces neurodegeneration in both aging and neurodegenerative disease models [39-41]. A recent genome-wide CRISPR screen identified loss of G6PD function as synthetic lethal when combined with mitochondria dysfunction [42], a known contributor to the pathogenesis and pathophysiology of neurodegenerative disorders [43].

Interventions to Address G6PD Deficiencies

Antioxidants

Ascorbate

Ascorbate, more commonly known as vitamin C, is a potent antioxidant; however, it has prooxidant effects in human erythrocytes in vivo [44] and high-dose ascorbate (e, g., ≥ 6mg/day [45]) induces AHA in G6PD^def individuals (Table 1). Despite this, lower doses have been used to treat methemoglobinemia in G6PD^def patients (Table 1) [23,46-48]; four case studies reported use of ascorbate to treat methemoglobinemia induced by rasburicase or other AHA triggers (Table 1). In all cases, ascorbate resolved methemoglobinemia. However, blood transfusions were used as co-treatments, masking the safety and efficacy profile of low-dose ascorbate.

Table 1.

Summary of clinical cases reporting the use of ascorbate in G6PD^def.

Hemolytic Trigger	Dose	Treatment	Dose	Response to Trigger						Type of Study	References
Ascorbate	80g/day, IV	Blood transfusion		X	X		X	X	X	case study	[118]
Ascorbate	40g/3x/week, IV + 20-40g/daily/oral			X	X		X			case study	[119]
	Increased to 80g/3x/week, IV
Ascorbate	30g/day, IV	Blood transfusion	8 days	X					X	case study	[120]
	60g/day, IV
Fava Beans Ascorbate	~10 beans							X		case study	[121]
	250-500mg/day, oral
Ascorbate	4-6g/6hr, oral	Blood transfusion			X		X			case study	[122]
Ascorbate	30g/day, IV	Blood transfusion		X	X	X	X			case study	[123]
Ascorbate	75g/day, IV	Blood transfusion		X	X	X	X			case study	[124]
Ascorbate	2g/day, IV	-			X		X			case study	[125]
Rasburicase	6mg	Ascorbate	5g/6hrs, IV	X		X	X			case study	[46]
		Blood transfusion
Rasburicase	6mg	Ascorbate	1g/day, oral	X	X	X				case study	[47]
		Blood transfusion
Chloroquine + Covid-19	600mg/day	Ascorbate	1g/4x/day	X		X				case study	[23]
		Blood transfusion
Fava Beans Ciprofloxacin Acetaminophen	500mg/2x/daily 1000mg/3x/daily	Ascorbate	1g/4x/day	X	X	X	X			case study	[48]
		Blood transfusion
				Hemolysis/Anemia	Dark Urine	Methemoglobinemia	Respiratory Distress /↓O2	Death	Renal Failure

A proposed mechanism for high-dose, ascorbate-induced, AHA in G6PD^def is described (Figure 1). Upon oxidant exposure, ascorbate is oxidized to dehydroascorbic acid (DHA) in the plasma and DHA is rapidly uptaken by erythrocyte glucose transporter 1 (GLUT1). Once inside, DHA is reduced to ascorbate by GSH, reducing intracellular GSH levels. Additionally, ascorbate participates in iron redox reactions; upon oxidation of the hemoglobin (Hb) iron, ascorbate can interact with the Fe³⁺ of methemoglobin, resulting in its oxidation to an ascorbate radical and concomitant reduction of MetHb (Fe⁺³) to OxyHb (Fe⁺²). Consequently, methemoglobinemia is inhibited; however, hyperoxy radicals, generated from the iron redox reaction, remain uncaptured [49]. Thus, both DHA reduction and hyperoxy radicals generated via iron redox reactions contribute to a hemolytic phenotype in G6PD^def patients. Additionally, ascorbate, DHA, and glucose are all internalized by GLUT1; therefore, ascorbate may alter glucose uptake and consequent PPP-mediated GSH recycling, especially in G6PD^def.

Figure 1.

A potential mechanism for ascorbate induced AHA in G6PD^def. a) Ascorbate concentration increases in the plasma as a result of IV or oral administration. b) In the presence of oxidative stress, ascorbate is oxidized to DHA in the plasma. c) Subsequently, DHA and ascorbate are imported into erythrocytes, with DHA having a higher preference for import. d) Once inside, DHA is reduced by GSH, depleting the already small GSH pool in G6PD^def. e) The ascorbate, imported from the plasma or generated via DHA reduction, can interact with oxidized iron to generate radicals. f) Both GSH depletion and iron-generated radicals likely contribute to cell damage and a hemolytic phenotype in G6PD^def.

Ascorbyl 6-palmitate (A6P), a lipid-soluble ester analog of ascorbate, acts as an antioxidant in erythrocytes in vitro [50,51] and increases ascorbate concentration in neuronal tissues, suggesting A6P crosses the blood brain barrier [52]. Other amphiphilic forms of ascorbate may have pharmacological value for the treatment of neurodegenerative and auditory disorders associated with G6PD^def, since they may cross tightly controlled barriers [53]. It remains unclear whether they will provide greater benefits than other lipid-soluble antioxidants such as vitamin E.

α-Tocopherol

α-Tocopherol, or vitamin E, is a lipid-soluble antioxidant which protects the cell membrane against oxidative stress. In G6PD^def patients, α-tocopherol improved hematological parameters (Table 2); it increased erythrocyte half-life (from 23 to 25 days), Hb levels, and reduced reticulocytosis in the most severe Class I and II G6PD^def [54-56]. However, challenging patient erythrocytes in vitro with H₂O₂ revealed no change in the degree of hemolysis, despite improvements in hematological parameters [55]. Additionally, three smaller clinical studies reported α-tocopherol did not improve hematological parameters or patient outcome [54,57,58] (Table 2). Although α-tocopherol efficacy remains inconclusive and studies have yet to be revisited, this is the first long-term preventive therapy attempted for G6PD^def. Additionally, treatment with α-tocopherol appeared safe at the indicated doses, suggesting that AHA is not a common feature of antioxidants.

Table 2.

Summary of clinical reports using antioxidants or NAC in G6PD^def.

Treatment	Dose	Duration	Outcome	Type of Study	Reference
α-tocopherol	adults: 800IU/day, oral children: 400 IU/day, oral	60 days	↑ Hb levels ↑ erythrocyte number ↑ packed cell volume ↓ reticulocytosis ↓ serum bilirubin	not registered	[126]
α-tocophero-acetate	800 IU/day, oral	12 weeks, 1 year	↑ Hb levels ↑ erythrocyte half-life ↓ reticulocytosis	not registered	[55]
α-tocophero-acetate	1000 IU/day, oral	24 weeks, 1 year	no improvements		[127]
α-tocopherol	800 IU/day, oral	8 weeks	↑ Hb levels ↑ erythrocyte half-life ↓ reticulocytosis	not registered	[56]
α-tocopherol-acetate	800 IU/day, oral	16 weeks	↑ erythrocyte survival ↓ reticulocytosis Failed to abolish hemolytic process	case study	[54]
α-tocopherol	400 IU/day, oral	4 weeks	no improvements	case study	[58]
α-tocophero-acetate	2000 IU/day, oral	4 weeks	no improvements	case study	[57]
α-lipoic Acid	600mg/day, oral	28 days	↑ GSH, TAC, and catalase activity ↓ lipid peroxidation in both groups	not registered	[62]
α-lipoic Acid	600mg/day, oral	28 days	↑ TAC at resting did not increase response to acute exercise stress	NCT02937363	[63]
N-acetylcysteine	initial dose-10g, IV 5g/4hr, IV	56 hrs	↑ hemolysis, unclear if due to NAC	case study	[128]
N-acetylcysteine	1200mg/x2, oral	1 year	↑ brief psychiatric rating score	case study	[129]
N-acetylcysteine	10g/x3/day, IV 600mg/x2/day, IV 600mg/x2/day, IV	1 day 7 days 5 days	↓ hemolysis indices after each treatment	case study	[25]

Astaxanthin and ß-Carotenoids

Astaxanthin, a lipid soluble carotenoid pigment, increased the activity of purified G6PD^WT in a dose-dependent manner, with a maximum 30% activation achieved at 0.64 mM astaxanthin, suggesting a direct mechanism [59]. Other carotenoids have been shown to exert antioxidant effects in erythrocytes in vivo, though they did not protect against GSH depletion [60,61].

α-Lipoic acid

Clinical studies find α-lipoic acid (ALA) improves erythrocyte total antioxidant capacity (TAC) in G6PD^WT and G6PD^def subjects (NCT02937363ⁱ) [62,63]. However, although ALA improves TAC in the context of exercise induced oxidative stress, it does not change erythrocyte redox response to acute exercise (Table 2) [1]. Taken together, ALA improves TAC in G6PD^WT and G6PD^def at baseline; however, like α-tocopherol, it may not improve response in the face of an acute stressor.

Lessons from Antioxidant Treatments

To date, the majority of treatments have focused on antioxidant therapies; however, these interventions have either triggered hemolysis (ascorbate) or have not prevented AHA in the face of a stressor (α-tocopherol, ALA), despite improvements in hematological parameters. It remains unclear if basal hematological indexes translate into meaningful clinical endpoints. A more appropriate measurement may be a surrogate endpoint; a carefully designed in vitro stress assay on blood withdrawn from human subjects starting and following a treatment regimen. Though, this may present challenging since drugs that cause AHA in vivo may not translate in vitro. A surrogate endpoint may reduce the duration length and cost of clinical trials and be a better indicator of drug efficacy with hematological indices better served as secondary endpoints. Additionally, animal models of G6PD^def may serve as good preclinical models for testing safety and efficacy of new therapeutics; zebrafish, mouse, and C. elegans [1].

GSH Metabolism and NAC Supplementation

Whereas NADPH-dependent GSH regeneration is one way by which the GSH pool can be restored, other metabolic pathways are capable of generating GSH. Studies reveal G6PD^def erythrocytes are metabolically different. At baseline, they have elevated levels of glycolytic metabolites related to NADH and ATP production, with a net increase in ATP levels [64]. However, when challenging G6PD^def erythrocytes with the oxidant diamide, GSH is depleted, and erythrocyte metabolism shifts towards GSH biosynthesis. Consequently, glutamate levels are elevated and metabolites related to ATP consumption increase, triggering glucose uptake by GLUT1 [65]. Yet, in G6PD^def, metabolic reprogramming is insufficient to restore GSH levels towards those seen in G6PD^WT individuals [64].

Modeling GSH metabolism in erythrocytes predicts that a 50% increase in plasma L-cysteine (LC) will improve GSH levels when the GSH oxidation rate is increased by 20% [66]. However, a 50% increase in plasma LC levels may be difficult to achieve in vivo; high doses of LC are required to compensate for poor bioavailability but are limited by toxicity [67]. N-acetyl-cysteine (NAC), an LC prodrug, has been developed to improve LC bioavailability and can be delivered orally at high doses, with minimal side effects [68]. In the plasma, NAC undergoes redox exchange reactions with L-cystine, to generate plasma LC at a concentration capable of achieving the maximal rate of GSH synthesis in erythrocytes [69] and LC/NAC increase the rate of GSH biosynthesis in erythrocytes in vitro [70]. However, despite its reputable role as a therapeutic antioxidant, few cases report the use of NAC to treat G6PD^def (Table 2).

Recently, a case-study reported a G6PD^def patient with COVID-19 was treated with hydroxychloroquine, which triggered AHA, and NAC resolved hematological indexes and hemolysis on three different occasions [25]. The patient received 90 g of intravenous (IV) NAC, which immediately improved hematological indexes. One week after NAC treatment was discontinued, hematological indexes began to decline, and the patient received 1.2 g daily IV NAC for one week. Again, NAC quickly resolved hematological indexes. Upon discontinuation of the second round of NAC, hematological indexes declined with NAC treatment improving hematological indexes for a third time, suggesting NAC may be beneficial for the treatment of G6PD^def in AHA.

The use of LC or NAC as a therapeutic for G6PD^def may expand to other G6PD^def-related disorders as LC increases expression of G6PD, glutamate-cysteine ligase modifier subunit (GCLM), glutamate-cysteine ligase catalytic subunit (GCLC), in G6PD^def monocytes and endothelial cells [71,72]. A combination therapy including oral NAC and a GCL activator, the rate limiting enzyme of GSH biosynthesis, may improve GSH levels in erythrocytes, protecting them against oxidative stress. Mathematical modeling of erythrocyte metabolism estimates a 50% increase in GCL V_max may reduce GSH pool depletion induced by an oxidative stressor [66]. A figure summarizing G6PD^def erythrocyte metabolism and potential modulation via NAC is presented (Figure 2).

Figure 2.

Metabolic rewiring in G6PD^def erythrocytes and NAC as a potential therapeutic route. G6PD^def erythrocytes have reduced (red) and elevated (teal) metabolites, favoring GSH production, however LC is limiting. a) NAC concentration increases in the plasma as a result of IV or oral administration. b) NAC undergoes redox exchange reactions in the plasma to generate LC. c) LC and NAC are imported into erythrocytes, where (d) NAC is deacylated to produce LC. (e) LC is used for GSH biosynthesis, (f) which may restore the GSH pool. NAC treatment combined with a GCL activator may maximize GSH biosynthesis and overcome GSH imbalance in G6PD^def erythrocytes. Abbreviations: 1,3 BPG, 1,3-bisphosphoglycerate; 2PG, 2-phosphoglycerate; 3PG, 3-phosphoglycerate; 6PG, 6-phosphogluconate; 6PGD, 6-phosphogluconate dehydrogenase; 6PGL, 6-phosphogluconolactone; E4P, erythrose 4-phosphate; F6P, fructose 6-phosphate; GGC, gamma-glutamylcysteine; G6P, glucose 6-phosphate; G3P, glyceraldehyde 3-phosphate; PEP, phosphoenolpyruvate; R5P, ribose 5-phosphate; Ru5P, ribulose 5-phosphate; S7P, sedoheptulose 7-phosphate; X5P, xylulose 5-phosphate.

Generation of NADPH via Complementary Pathways

NADPH is regenerated in the cytosol via three metabolic routes: G6PD utilizes glucose to produce NADPH via the PPP, whereas malic enzyme 1 (ME1) and isocitrate dehydrogenase 1 (IDH1) use tricarboxylic acid (TCA) cycle intermediates; ME1 and IDH1 are reliant on mitochondrial redox balance and subsequent substrate trafficking to the cytosol [73]. The contribution and responsiveness of each pathway to NAPDH production varies according to tissue and cell type as well as genetic and environmental pressures. For example, erythrocytes and neurons strongly rely on G6PD whereas adipocytes and liver cells are more reliant on IDH1 and ME1 [14,74,75]. Furthermore, adipocytes rely more on G6PD under hypoxic conditions [74]. The lack of complimentary pathways to recycle NADPH increases cell susceptibility to oxidative stress.

In erythrocytes, the absence of mitochondria limits the contribution of IDH1 and ME1. Although erythrocytes lack mitochondria, they still express TCA cycle enzymes as well as IDH1 and ME1 [76] and are capable of metabolizing TCA intermediates to generate NADPH. Precursors of complementary routes have been found to accumulate in G6PD^def erythrocytes, suggesting ME1 and IDH1 pathways are activated to maintain NADPH levels [64]. However, this is insufficient to counteract erythrocytes vulnerability to oxidative stress in G6PD^def.

It remains to be determined whether boosting complimentary routes with exogenous substrates, such as citrate, or genetic/pharmacological interventions to increase ME1 or IDH1 levels and/or activity, is sufficient to counteract the excessive oxidative stress and hemolysis seen in G6PD^def erythrocytes. In this case, the development of small molecule activators of complimentary routes, focusing on regenerating NADPH, might become a feasible strategy to improve the antioxidant capacity of G6PD^def in erythrocytes as well as other cell types associated with chronic disorders.

G6PD Upregulation via Transcriptional Regulators

During erythropoiesis, erythroblasts undergo successive changes, including nuclear extrusion [77]. Although this process is essential for erythrocyte function, this presents a therapeutic challenge as erythrocytes are incapable of synthesizing new protein. However, targeting transcription in early-stage erythroblasts, before nuclei extrusion, or in other cell types with nuclei, is an attractive strategy for increasing G6PD protein levels and treating conditions associated with G6PD^def.

Histone deacetylase inhibitors (HDACi)s have gained attention for their ability to modulate protein transcription, and in the context of G6PD, have been shown to increase PPP flux by upregulating and downregulating G6PD expression in cancer cells [78,79]. Interestingly, treatment of patient derived erythroblasts and non-hemopoietic cells with HDACi sodium butyrate (NaBu) increased transcriptional upregulation of the G6PD gene, while having no effect on other genes in the glycolytic or PPP pathway. The specificity of G6PD induced upregulation was attributed to increased histone hyperacetylation by histone acetyltransferases (HAT)s and increased HDAC, HAT, DNA polymerase II, and SP1 (transcription factor) binding to the G6PD core promoter (Figure 3). SP1 binding induced G6PD transcription, mRNA production, and protein translation, which lead to a subsequent increase in Class I-III G6PD^def protein levels and activity, restoring it to that of WT or greater. Suberoylanilide hydroxamic acid (SAHA), another HDACi, produced similar results to NaBu, demonstrating transcriptional activation is not HDACi specific [80-82].

Figure 3.

The process and consequence of HDACi-mediated transcriptional activation of the G6PD gene in erythroid cells. a) HDACi exposes the G6PD core promoter; b) SP1, HAT, HDAC, and DNA polymerase localize to the core promoter to induce G6PD transcription in erythroid precursor cells and c,d) subsequently increase protein levels in new erythrocytes. However, HDAC5 and HDAC2 have been found to disrupt erythropoiesis and therefore, inhibition of these two HDACs may have adverse side effects. Other transcriptional modulators, such as a Nrf2/Keap1 protein-protein-interaction inhibitor, may also increase G6PD transcriptional output. e) G6PD variant biochemical properties will likely influence the efficacy of a transcriptional activator, with protein stability being a major determinant of residual protein levels in aging erythrocytes.

Although HDACis are capable of restoring Class I-III G6PD variant activity to that of WT or greater in erythroid precursor cells, histone deacetylation is important for DNA compression, nuclear extrusion, and erythroid maturation [83-85]. Therefore, it is not surprising that side-effects reported for multiple FDA approved HDACis include thrombocytopenia and anemia [86]. To date, two HDACs have been implicated in erythropoiesis: HDAC5 [84] and HDAC2 [87]. An HDACi that does not target HDAC5 or HDAC2 may limit side effects – if inhibition is not required for transcriptional upregulation of the G6PD gene. Improving HDAC isoform selectivity is a common strategy to reduce off-target effects and improve the safety profile of HDACis [88].

Due to their cytotoxic effects, the majority of HDACis are approved for use in hematological and neoplastic malignancies [89]. However, the use of HDACis in non-neoplastic indications is expanding; HDACis are in early stage clinical trials for Niemann-Pick type C1 (NCT02124083ⁱ), Alzheimer’s (NCT03056495ⁱ), epilepsy (NCT03894826ⁱ), and Crohn’s disease (NCT03167437ⁱ) [90], with Depakene® (valproic acid) being used as a first-line therapy for migraine prophylaxis and found to be well-tolerated [91,92]. The use of HDACis at their current stage may not be ideal for G6PD^def patients, however, the development of more selective HDACis may circumvent toxicity. Additionally, the use of HDACis provide a proof-of-concept; transcriptional upregulation increases G6PD activity in erythrocyte precursor cells. Therefore, other transcriptional targets, with improved target specificity, may reduce off target effects while providing benefit to G6PD^def patients.

A second molecular target of interest is transcription factor nuclear erythroid receptor 2 (Nrf2). Under basal conditions, Nrf2 levels are suppressed by Keap1-dependent-ubiquitination and subsequent proteasomal degradation. However, in the presence of electrophiles and oxidants, cysteine modifications deactivate Keap1 and in turn increase Nrf2 levels. Nrf2 exerts antioxidant effects by binding to antioxidant response elements and upregulating transcription of antioxidant defense enzymes, including G6PD [93].

Currently, Tecfidera® (dimethyl fumarate), indicated for use in multiple sclerosis, is the only FDA approved activator of Nrf2. Natural products such as curcumin, sulforaphane, and flavonoids have also been shown to activate Nrf2 [94,95]. Unfortunately, these molecules are electrophiles and lack target specificity, making them toxic [96,97]. A small molecule or peptide inhibitor of the Keap1/Nrf2 protein interaction may increase target selectivity and overcome toxicity and remains an attractive target [98]. As with all drugs, too much of anything can be a bad and chronic activation of Nrf2 may cause reductive stress [93].

AHA caused by G6PD^def is largely described by two biochemical enzyme parameters: catalytic activity and enzyme stability [99,100]. Protein stability is particularly important in erythrocytes since they are anucleate with a 120-day lifespan devoid of protein synthesis. We foresee transcriptional modifiers to be most beneficial for G6PD variants which maintain enzymatic stability, such as Class II and III variants; protein stability will ensure increased protein levels achieved by transcriptional upregulation will be maintained for significant period of time. Although Class I variants generally have reduced stability, it has yet to be determined whether a significant increase in G6PD activity for new erythrocytes, regardless of stability, will be sufficient to improve a CHA phenotype. Additionally, transcriptional regulators may be beneficial in cell types that have nuclei and are associated with G6PD^def related chronic disorders.

Small Molecule Activators of Clinical G6PD Variants

Recently, we sought a structural chaperone that could correct the most common variants and identified one lead compound, AG1. AG1 improved the activity and stability of several Class II and Class III variants, including the common Mediterranean and A⁻ and Canton, by stabilizing dimeric G6PD in vitro, in cellulo, and in vivo models [101]. As a result, AG1 increased NADPH and GSH levels, reduced ROS, and protected cells from oxidative stress.

These results demonstrate that a single small molecule can partially correct the catalytic activity and stability of multiple common G6PD variants. Although AG1’s effect on catalytic activity is important, stabilization of dimeric G6PD is likely an important feature; enzyme stability is a major determinant of residual activity in erythrocytes [99,100]. Therefore, a small molecule that increases the stability of G6PD may be an effective therapeutic, though its efficacy may not be evident until tested in an in vivo model.

The discovery of AG1 has inspired the pursuit of other G6PD structural chaperones. In recent computational studies, AG1 was used as a scaffold for a structure-based pharmacophore screen where five compounds were identified as potential activators of G6PD [102]. In addition, several AG1 analogs were developed; although none activated G6PD better than AG1, several displayed AG1-like activation and are better suited for in vivo studies – they lack a disulfide bond [103].

Recently, the structure of the most severe Class I variants revealed a common structural defect [104]. This common defect in Class I variants presents an opportunity for a single molecular chaperone that corrects this common defect. Future work focuses on search for activators of most severe Class I variants and identifying a small molecule that stabilizes the tetrameric state; it has been shown tetrameric G6PD has increased catalytic efficiency over dimeric G6PD [105]. Such compounds may also be beneficial in chronic diseases that are associated with the downregulation of G6PD [106], reduced GSH levels [107], or increased ROS [108], such as neurodegenerative disorders [15], but do not have G6PD^def, as demonstrated by overexpression studies [109,31].

Gene Therapy

Gene therapy is gaining recognition for its use in inherited genetic disorders and gene therapies for cystic fibrosisⁱⁱ, spinal muscular atrophyⁱⁱⁱ, and retinal dystrophy^iv have already received FDA approval, with therapies for sickle cell anemia and ß-thalassemia in clinical trials [110]. Recently, Zynteglo® (betibeglogene autotemcel), a gene therapy for ß-thalassemia, received conditional market approval in Europe [111,112]. Approval was granted after Zynteglo® reversed blood transfusion dependency in 12 of 13 patients in early phase clinical trials [113]: a remarkable feat for genetic hematological therapies.

A gene therapy for G6PD^def will likely overcome the therapeutic challenges that arise from the genetic diversity of the disorder. However, the development of a gene therapy for G6PD^def presents a major challenge; the cost per Zynteglo® treatment course is set at $1.8 million with high costs common in gene therapy [114]. Although rare, G6PD^def Class I variants have a severe phenotype, impacting the patient’s quality of life [115,116]. Early studies find retroviral transduction of the human G6PD gene into mouse and human hematopoietic stem cells resulted in stable lifelong expression of human G6PD in primary and secondary bone marrow transplant recipient mice, increasing G6PD activity two-fold [117]. suggesting gene therapy for the most severe variants of G6PD^def is warranted.

Concluding Remarks and Future Perspectives

Although G6PD-associated enzymopathy has been discovered more than 60 years ago, there are still no therapies for the treatment of G6PD^def. The fact that the most severe form of G6PD^def is rare and although Class II and III variants are common, they are often asymptomatic, unless exposed to an oxidative stressor, reduces interest by pharmaceutical companies. Additionally, the large number of mutations and divergent consequences on protein structure, make a “one-size fits all” therapy unlikely. Even with a single point mutation, correcting enzyme dysfunction is inherently challenging.

Past failures with antioxidant therapies make this strategy less attractive with the majority of explored therapies not having reached the stage of clinical trials. Therapeutic routes warranting further investigation include: 1) NAC with the potential for a combination therapy; a GCL activator or transcriptional regulator; 2) modulation of NADPH pathways; 3) transcriptional regulators of G6PD such as HDACi or Nrf2; 4) small molecule chaperones that activate G6PD directly and 5) gene therapy. Due to high treatment costs, gene therapy is likely a good candidate for the most severe Class I variants, with non-genetic approaches best for less severe variants.

Therapeutic routes and molecular targets are summarized (Figure 4, Key Figure) and questions regarding these routes as drug targets remain largely unanswered (see Outstanding Questions). We present potential treatment strategies for G6PD^def in hopes of motivating the development of new therapies. The interest of the pharmaceutical industry will increase only after more epidemiological studies confirm that G6PD^def is an important unmet need for a variety of common acute and chronic diseases.

Figure 4.

Summary of some of the chronic diseases associated with G6PD^def and strategies to treat G6PD related disorders. Treatment strategies include increasing GSH biosynthesis by activating GCL and co-supplementing with LC/NAC; activating NADPH compensatory enzymes ME1, IDH1; upregulating G6PD protein levels via HDAC inhibitors (HDACi) or Nrf2; small molecule chaperones increasing catalytic activity and stability of G6PD; antioxidants; and gene therapy (this strategy is likely for most severe Class I variants).

Outstanding Questions.

• Although antioxidants do not alleviate a G6PD deficient hematologic phenotype, are they effective for treatment of chronic diseases associated with G6PD deficiency?

• Are hematological indexes suitable clinical endpoints for clinical trials in G6PD deficient patients? If so, how should improvements in hematological indexes be translated to a clinical phenotype?

• Should a surrogate endpoint, subjecting patient blood to oxidative stressors, be used for patient selection in a trial and/or a primary endpoint to establish efficacy for clinical trials in G6PD deficient patients?

• Can improving GSH production via targeting GSH biosynthesis prevent hemolysis induced by oxidative stress?

• Can other NADPH producing pathways, such as IDH1 and ME1, be targeted to compensate for G6PD deficiency in chronic diseases or in a hemolytic context?

• Is improving G6PD protein levels in erythroblast sufficient to restore enzymatic activity? If so, is variant stability a major predictor of efficacy?

• What HDACs are important for erythropoiesis and G6PD transcription? Are HDAC2 and HDAC5 required for G6PD transcription?

• Can NRF2 upregulate G6PD transcription in erythroblasts? Are benefits comparable to an HDAC inhibitor?

• Is stabilization of G6PD protein sufficient to improve clinical phenotype?

• Does cost/benefit ratio for gene therapy deem it as practical for treatment of G6PD deficiency?

Highlights.

• G6PD deficiency is better known as a hematologic disorder. However, G6PD plays a role in other chronic disorders such as diabetes, cardiovascular disease, viral infectivity and complications, and neurodegeneration.

• Antioxidant therapy is the most commonly explored treatment strategy for hemolysis caused by G6PD deficiency. However, it is largely ineffective, with ascorbate triggering AHA in erythrocytes.

• Therapeutic opportunities for G6PD deficiency include improved glutathione biosynthesis via NAC administration, exploiting NADPH compensatory pathways, small molecules which correct enzyme dysfunction by directly binding to G6PD, small molecules that increase the transcriptional output of G6PD, and gene therapy.

Glossary

Bilirubin

a byproduct of heme catabolism, often noted for its color (yellow pigment), found to have antioxidant and oxidant properties and to have pathological consequences.

Glucose-6-phosphate dehydrogenase (G6PD)

A housekeeping enzyme responsible for catalyzing the conversion of glucose-6-phosphate and NADP⁺ into 6-phosphoglucono-δ-lactone and NADPH.

G6PD deficiency (G6PD^def)

A genetic disorder, caused by mutations in the G6PD gene, which leads to the hemolysis of red blood cells.

Glutathione (GSH)

An endogenous antioxidant that protects the cell from oxidative stress by neutralizing ROS; upon reaction with ROS, GSH is converted to GSSH, where it is recycled back into GSH by glutathione reductase, using NADPH as a cofactor.

Histone Deacetylase Inhibitor (HDACi)

A class of anti-cancer small molecules that target histone deacetylases.

Methemoglobinemia

a disorder characterized by excessive levels of methemoglobin. Due to loss of hemoglobin O₂ binding, symptoms are largely related to O₂ deficiency.

Nicotinamide adenine dinucleotide phosphate (NADPH)

A cofactor that is used for biosynthetic reactions and plays a major role in protecting the cell against oxidative stress.

Pentose phosphate pathway (PPP)

A branch of glucose metabolism, parallel to glycolysis, where glucose-6-phosphate is diverted to produce cellular NADPH and ribose-5-phosphate; metabolites important in redox homeostasis, nucleotide and fatty acid synthesis.