Treating sickle cell anemia: A new era dawns

Abstract

A single mutation in the β-hemoglobin gene allows deoxyHbS to polymerize and initiate the complex pathophysiology of sickle cell disease, a hemolytic anemia with chronic complications punctuated by acute events (Figure 1).1, 2 This commentary encompasses certain assumptions: opinions on how to use new therapeutics are many-data are limited; drugs preventing HbS polymerization are the most efficacious and should be started early; curative cell-based genetic therapy might soon be approved; untoward effects of each new modality will be acceptable; treatment algorithms proposed today will be modified tomorrow as new information accrues; only high-resourced countries can afford the regimens proposed. Hydroxyurea is recommended for nearly all patients. Its restriction to individuals with frequent acute events neglects the importance of chronic hemolysis (Figure 1) Hemolysis is never quiescent; pulmonary and systemic vasculopathy and mortality are closely linked to chronic intravascular hemolysis that begins in childhood but can be asymptomatic for years.3, 4 Hydroxyurea induces fetal hemoglobin (HbF) that retards the polymerization of deoxyHbS. The HbF response to hydroxyurea is heterocellular leaving some erythrocytes “unprotected” from polymer-induced damage; these cells could be clinically meaningful. HbF induction depends in part on a proliferating erythroid bone marrow, which is most vigorous in childhood, as with aging, hematopoietic marrow retreats centripetally. Years of intramedullary sickling and marrow infarction further reduce the erythroid capacity of the marrow.5, 6 In adults with baseline HbF of ~5%, after 2 years of treatment the top quartile of HbF responders had only ~12% HbF.7, 8 The HbF response to hydroxyurea was related to the capacity of the bone marrow to withstand myelosuppression. In contrast, all children seem to respond to hydroxyurea. When hydroxyurea was started between 9 to 18 months of age, HbF changed little during follow-up from its baseline level of 25%.9, 10 Starting hydroxyurea at approxixmately11 months at a dose of ~27 mg/kg, was associated with HbF levels of 33.3% ± 9.1%; 33% of patients had HbF >40% and 29% had hemoglobin concentration >11 g/dL.; acute events were reduced markedly with little toxicity. If this regimen is safe, the results sustained and generalizable, this approach should set a new standard.11 Nonetheless, HbF levels decline with age. When hydroxyurea was started in children <2 years old, HbF was 19.7% after 2 years treatment; following 15 years of continuous therapy HbF was 15.1%. Even when started early in life the HbF response to hydroxyurea might wane.5, 12 Voxelotor binds the N-terminal valine of α-globin decreasing the partial pressure of O2 at which hemoglobin is half saturated (P50). Locking HbS in its oxy conformer prevents HbS polymerization but also retards oxygen delivery. Over several half-lives voxelotor should have a pancellular effect. In a small Phase 3 clinical trial, voxelotor, 1500 mg daily, was associated a 1 g/dL increase in hemoglobin concentration in 59% of patients; biomarkers of hemolysis decreased. The increase in hemoglobin was similar to that of hydroxyurea treated patients.7, 8, 13, 14 Two issues associated with voxelotor treatment require deeper understanding. Firstly, could higher hemoglobin levels increase sickle vasoocclusion? There are precedents for this concern. Senicapoc, an agent affecting cation transport, increased red cell hydration and reduced MC(HbS)C and hemolysis; hemoglobin increased by 0.6 g/dL. Some patients had an increased rate of vasoocclusive events.15 In sickle cell anemia-α thalassemia, HbS polymerization is inhibited because of reduced MC(Hb)SC. These patients have hemoglobin levels 0.5 to 1 g/dL higher than in sickle cell anemia. Although their risk of vasoocclusive-viscosity related complications is increased, propensity to stroke, nephropathy, leg ulcers and priapism is reduced.16, 17 Secondly, blood flow to the brain is maximized in children with sickle cell anemia. The high oxygen affinity of a modified hemoglobin could be harmful.18 HbF also has high oxygen affinity; compound heterozygotes for HbS and gene deletion hereditary persistence of HbF (HPFH) appear well despite 30% HbF in all of their erythrocytes. Hydroxyurea treated patients have lower cerebral oxygen extraction fraction than controls. This might increase the chance of white matter injury.19 Countervailing forces, including hemoglobin levels, blood viscosity, red cell 2,3 BPG, and the high P50 of HbS containing erythrocytes all affect oxygen extraction fraction. According to Wood, “success of any hemoglobin-dissociation curve modifying agents will depend on their ability to both improve functional oxygen carrying capacity and balance the increased difficulty of oxygen unloading.”20 Downstream effects of HbS polymerization include adhesive interactions among endothelial cells, leukocytes, platelets and erythrocytes. P-selectin is involved in these interactions; blocking selectins prevents sickle cell-endothelial adhesion.21-25 Monthly intravenous infusions of crizanlizumab, a P-selectin blocking monoclonal antibody, reduced acute painful episodes by ~45%, a reduction similar to that of hydroxyurea.7, 26 Only 65% of patients completed this study and the overall perception of pain was not affected. Life-long use of an intravenous medication might prove problematic in patients known to have difficult venous access. Myeloablative HLA-identical related donor transplants have event-free survival of 95 to100%. Haploidentical transplants have not achieved this level of success. Unfortunately, only 15% of patients have a matched donor, and no studies have compared the long-term results of transplantation with standard care. Given its curative potential, despite the possibility of graft vs host disease, a case can be made that all patients with a suitable donor be offered transplantation. Additive gene therapy and genome editing that uses autologous stem cells has curative potential. Lentiviral vectors can insert anti-sickling globin genes into CD34+ patient cells. Clinical trials have shown HbS reduced to <50% with lessened hemolysis and acute events.27, 28 Several approaches to reducing the expression of HbF suppressors in CD34+ cells are possible. They include disrupting the erythroid enhancer of BCL11A or the binding sites for BCL11A in the γ-globin gene promoters and down-regulating the expression of BCL11A.29, 30 Clinical trials are targeting BCL11A with CRISPR/Cas, zinc finger nucleases or shRNA (NCT03745287; NCT03653247; NCT03282656). BCL11A has been downregulated in several patients with sickle cell anemia. While follow-up has been short, HbF has risen to nearly 50%, F-cells to >95%, hemoglobin to ~11 g/dL; acute vasoocclusive events have abated.31, 32 Cell-based therapeutics is in its infancy. It has been applied to very few patients and requires myeloablative conditioning. Off-target consequencess of genome editing, effects of the semi-random nature of lentivirus integration and the sustainability of therapeutic effects all await additional long-term study. How might new therapeutics be integrated into treatment? One approach is shown in Figure 2. Because of the primacy of HbS polymerization, the main focus of drug treatment should be to diminish this tendency. Controlled clinical trials have shown that hydroxyurea reduces acute vasoocclusive events and hemolysis with the likely extension of lifespan. Hydroxyurea should be started at maximally tolerated doses in the first year of life. The clinical response should be measured by the rate of acute vasoocclusive events, and the reduction of hemolysis as estimated by commonly available biomarkers. Continued hemolytic anemia that is more than modest provides an indication for adding voxelotor in patients aged ≥12 years to further retard deoxyHbS polymerization. Can the hemoglobin level rise too high provoking vasoocclusive complications? I hypothesize that this will not occur, but clinical trials will required to resolve this. Hydroxyurea both increases hemoglobin and decreases vasoocclusion. A parallel effect on hemoglobin concentration is seen in HbS-HPFH where hemoglobin levels are nearly normal with a HbS concentration of ~70%. The HbS-HPFH patients are asymptomatic because of the pancellular distribution of HbF. The voxelotor-induced increase in hemoglobin concentration might be offset by the pancellular decrease in HbS polymerization. Both voxelotor and HbF have a primary effect on polymerization; neither sickle cell rehydration nor sickle cell anemia-α thalassemia modifies the HbS molecule. If optimal doses of hydroxyurea and voxelotor are not associated with sufficient reduction of acute vasoocclusive complications, crizanlizumab should be added in patients aged ≥16 years. All patients with an HLA-matched sibling donor should be offered allogeneic hematopoietic stem cell transplantation. If gene therapy proves safe and curative, it might be offered to patients lacking a suitable donor for stem cell transplantation or even replace this modality as gene therapy does not require immunosuppression. A combinatorial approach to treatment is unlikely to be subjected to controlled clinical trials. However, there are hints that adding these two drugs to hydroxyurea might confer additional benefit. If the effect of voxelotor is pancellular this would provide an increment in polymerization inhibition beyond hydroxyureaʼs heterocellular effect. Patients taking voxelotor and hydroxyurea had a larger increase in hemoglobin than patients taking only voxelotor.14 In a post-hoc analysis crizanlizumab plus hydroxyurea decreased the likelihood of vasoocclusive events and delayed the time to the first event.37 Predictive laboratory assays tailored to the mechanism of action of each therapeutic would be advantageous. HbF levels and F-cell analysis by FACS only approximates the benefit of HbF inducers. Neither assay measures the distribution of HbF concentrations among F-cells and the intravascular destruction of a small number of cells can impair flow-mediated dilation leading to complications of intravascular hemolysis.38-41 People with HbS-HPFH do not have hemolytic anemia or sickle vasoocclusion because each cell contains ~10 pg of HbF that is sufficient to inhibit HbS polymerization. A high-throughput assay providing an estimate of the distribution of HbF concentrations among F-cells would be prognostically useful. An indirect method of estimating the amount of HbS polymer in single cells in a high-throughput manner, might provide a means of gauging the clinical benefit of any approach to reducing polymerization, including those that do not induce HbF, like voxelotor.42 Reticulocyte count, LDH, serum bilirubin and AST have been used in a principal component analysis to derive a hemolytic component.43 Foretelling the level of hemolysis suppression needed to prevent the common vasculopathic complications of disease would be prognostically valuable. Tricuspid regurgitant velocity (TRV) could be used as an outcome to derive an age-adjusted estimate of the likelihood of increased TRV according to the hemolytic component. Treatment of sickle cell anemia might be titrated to achieve a level of hemolysis similar to HbSC disease where hemolysis-related complications are less.44, 45 Haptoglobin levels are usually unmeasurable in sickle cell anemia; an increase would signify successful reduction of hemolysis. P-selectin blocking reduces sickle vasoocclusion. Microvascular flow can be assessed by laser Doppler velocimetry that is abnormal in steady-state sickle cell anemia.46 Other means of noninvasively measuring microcirculatory flow are available.46, 47 A replicable and physiologically relevant in vitro assay where patient cells can be tested repeatedly would be useful.48-50 Any new treatment will be unaffordable where sickle cell anemia is most prevalent.51 In one estimate (SCI Pharmaceuticals, Altemia Overview, May 2018) ~$900 million was spent annually for healthcare for sickle cell disease. Median hospitalization cost was ~$15 000 per stay; annual cost of care for pediatric patients was ~$68 000; older patients cost ~$200 000. The total cost for 70 years of life expectancy was ~$10 million. Another estimate placed the average cost up to age 45 at ~$1 million but did not account for the higher costs that are likely from age 45 to 70.52 Treatment costs increase logarithmically as therapy moves from hydroxyurea, to voxelotor and crizanlizumab, to hematopoietic stem cell transplantation or gene therapy. Hydroxyurea can be had for ~$1000 per year; voxelotor and crizanlizumab will cost >$100 000 per year; stem cell transplantation and gene therapy will cost >$1 million. Assuming that lifetime cost of caring for sickle cell anemia falls between $1 and $10 million over a median lifespan of ~70 years, a one-time treatment with curative potential, like allogeneic stem cell transplantation or gene therapy appears cost-effective. Lifetime use of hydroxyurea at ~$70 000 is a bargain. A substantial reduction from the estimated charges for voxelotor and crizanlizumab would be needed for these treatments to compete economically with transplantation or gene therapy. New drugs will change the treatment landscape for sickle cell anemia. Their combinatorial use will not result in cure but should alter the course of disease. Initially, drugs will be more widely applied compared with technologically formidable but potentially curative cell-based therapies. Drugs costs will strain healthcare in high-resourced countries; they will be unaffordable where the disease burden is greatest. The paramount challenge will be to make the likely benefits of new treatments available where they are most needed.

Topics & Concepts