Therapeutic potential of the latest oxygen affinity-modifying agent, GBT021601, for treating sickle cell disease is questionable

Abstract

We think it is important that readers of the British Journal of Haematology should be aware of several important aspects of the latest oxygen-modifying drug, GBT021601, for treating sickle cell disease (SCD) published last year in your journal by Dufu et al.¹ Treating SCD by modifying oxygen affinity is a complex strategy owing to many subtle aspects of sickle haemoglobin (haemoglobin S [HbS]) biochemistry, physiology and kinetics that must be considered.² To make our discussion of this subject more easily understandable, we first provide some important background information.

Haemoglobin has two arrangements of its four subunits, 2α and 2β, that are in rapidly reversible equilibrium.³ One, called T to represent the tension provided by inter-subunit salt-bridges, has a low affinity for oxygen and has the conformation of fully deoxygenated haemoglobin and the other, called R for relaxed (no inter-subunit salt bridges), has a high affinity for oxygen and has the conformation of fully oxygenated haemoglobin.³ As haemoglobin is deoxygenated in the tissues, the shift in the equilibrium population from R to T gives rise to the characteristic sigmoid shape of the oxygen dissociation curve (ODC) that facilitates binding of oxygen in the lungs and unloading it in the tissues. In the case of HbS, deoxygenation results in the formation of fibres that stiffen and distort (‘sickle’) red blood cells (RBC's), the root cause of vaso-occlusion and its consequent pathology in SCD.⁴

Experiments showed that only the T conformation polymerizes and that two different tertiary conformations (t and r) within T also play a role.⁵ These experiments, as well as early studies by Beutler,⁶ led to the concept that binding a drug to shift the conformational equilibrium towards non-polymerizing R would reduce sickling and therefore be a viable strategy for drug treatment of SCD. This approach resulted in many attempts to develop a drug that binds preferentially to R. Voxelotor (GBT440, oxybryta) is the only one so far approved by the FDA that acts by this mechanism. There was no evidence that voxelotor decreases sickle cell crisis frequency, which is of utmost important to the patient. However, crisis frequency was not an end-point in FDA's consideration, which approved the drug because of the increase in haemoglobin levels.

The reason that treatment with voxelotor is controversial is that the reduction in sickling is counteracted by the higher oxygen affinity, which compromises oxygen unloading in the tissues in SCD.^{2, 7, 8} In our 2021 Blood article,⁹ we provided a theoretical answer to the question of how much oxygen is delivered when cells from patients with SCD sickle during the in vivo rapid oxygen pressure decrease that occurs in seconds as RBCs pass through the microcirculation. By evaluating the clinically relevant positive and negative effects on oxygen delivery of decreased sickling and a highly left-shifted oxygen binding curve, respectively, we concluded that, in spite of much reduced sickling and an increase in haemoglobin levels, the net effect would be a decrease in oxygen delivery (see also Worth et al.¹⁰). This prediction was born out by the extended double-blind studies of Howard et al.^{11, 12}

In Henry et al.'s study,⁹ we pointed out that voxelotor is bound so tightly to the HbS molecule, shifting the conformational equilibrium towards R to such an extent, that it effectively locks the HbS conformation in R. The same is true for GBT021601, as evidenced by the biphasic ODC (Figure 1A). The oxygen affinity of R is so high that it unloads little or no oxygen at the oxygen pressures of the tissues. Consequently, although the total Hb increases with drug treatment, the functional Hb actually decreases. As an example, consider an SCD patient with an 8 g/dL Hb level that increases with 30% drug occupancy of HbS to 10 g/dL. 30% of the Hb molecules are non-functional, so the functioning level is now about 7 g/dL. Thus, the anaemia is actually worse!

Haemolysis is less with voxelotor treatment and presumably will also be less with GBT021601 because of decreased RBC fragility from fewer sickling–unsickling cycles, but the long-term effects on organ damage from the predicted decrease in oxygen delivery combined with the functional anaemia just described remain to be determined. It is important to point out that preferential binding of a drug to the R conformation is still a viable strategy, so long as both binding and dissociation are sufficiently rapid (see figure S2 of supplementary information in Henry et al.⁹).

The question, then, is whether there is a property of GBT021601 that will make it more effective than GBT440. The experimental data of Dufu et al. on blood from SCD patients seem accurate and are perfectly consistent with what we have measured.¹ Figure 1A compares our ODC measurements on sickle trait blood for GBT440 (voxelotor) and GBT021601, which are essentially identical. We use trait blood to avoid formation of the very low affinity fibres that would affect the ODC. The biphasic curve for both drugs occurs because of the very tight binding of the drug to HbS, which does not dissociate during the 20- to 30-min experiment. The potentially harmful left shift in the ODC is therefore the same for GBT021601 and voxelotor (GBT440). Figure 1B shows the fraction sickled versus time using an assay previously described,¹³ where the anti-sickling effect of the two drugs is also identical within experimental error (Figure 1B). Dufu et al. do show that there appears to be no decrease in oxygen delivery in their murine model.¹ No oxygen delivery measurements with voxelotor (GBT440) were made in the murine model of Oksenberg et al.,¹⁴ so we cannot make a comparison.

Finally, the ex vivo properties of the two GBT drugs are so similar that it is difficult to imagine that GBT021601 will be a significantly better drug than GBT440, which Pfizer just (September 25, 2024) withdrew from the market, unless it turns out to have fewer side effects.¹⁵ However, for the sake of the patients with this debilitating disease we hope that it is.

Braydon Alaimo measured the oxygen dissociation curves. Belhu Metaferia measured the sickling curves. Troy Cellmer performed the data analysis. William A. Eaton wrote the letter. Swee Lay Thein and H. Franklin Bunn edited the letter.

This work was funded by the intramural programmes of the National Institute of Diabetes and Digestive and Kidney Diseases and the National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD.

The authors declare no conflict of interest.

Written informed consent was obtained for sickle trait donors under the IRB-approved protocol 08DK0004 provided blood samples for the experiments reported in this study.

This study was approved by the institutional review board.

This is an IRB/ethics board-approved study.

There are no materials reproduced from other sources.

Editor´s Note to the letter “Therapeutic potential of the latest oxygen affinity modifying agent, GBT021601, for treating sickle cell disease is questionable” by Drs. Eaton, Alaimo, Metaferia, Cellmer, Thein, and Bunn.

The British Journal of Haematology is committed to allow our readers to reflect on articles published within our columns. All such letter addressing issues in such recently published articles are subject to a peer-review and are sent to the original authors to give them a chance to respond.

Given the critical nature of the basic elements of above mentioned letter, we have made every effort to get into contact with the authors of the article in question. However, every attempt reaching out to them employing both the email addresses given in that article (which bounced), as well one at Pfizer, which can be found on the internet to have bought the company, from which the article originated (and using the address given together with the announcement of the acquisition) and where its 1st author is now employed, have been to no avail.

In this particular instance, and bearing in mind that doubt is cast upon some important issues in the original paper, we have chosen to publish the letter without a rebuttal.

Peter Hokland, Senior Editor.