Kinetic assay shows that increasing red cell volume could be a treatment for sickle cell disease

Abstract

Although it has been known for more than 60 years that the cause of sickle cell disease is polymerization of a hemoglobin mutant, hydroxyurea is the only drug approved for treatment by the US Food and Drug Administration. This drug, however, is only partially successful, and the discovery of additional drugs that inhibit fiber formation has been hampered by the lack of a sensitive and quantitative cellular assay. Here, we describe such a method in a 96-well plate format that is based on laser-induced polymerization in sickle trait cells and robust, automated image analysis to detect the precise time at which fibers distort ("sickle") the cells. With this kinetic method, we show that small increases in cell volume to reduce the hemoglobin concentration can result in therapeutic increases in the delay time prior to fiber formation. We also show that, of the two drugs (AES103 and GBT440) in clinical trials that inhibit polymerization by increasing oxygen affinity, one of them (GBT440) also inhibits sickling in the absence of oxygen by two additional mechanisms.

Trial registration: ClinicalTrials.gov NCT02380079.

Keywords: drugs; hemoglobin S; screening assay; sickle cell; treatment.

Conflict of interest statement

The authors declare no conflict of interest.

Figures

Fig. 1.

Kinetics of HbS polymerization and connection to disease pathogenesis. (A) Schematic kinetic progress curve showing the delay period (lag phase) before the detection of any fibers. (B) Cartoon of microcirculation showing the flow of a cell containing no fibers—and therefore easily deformable—through a muscled arteriole, a capillary, and a venule. As shown by Mozzarelli et al. (10), the delay time at in vivo oxygen pressures is sufficiently long that the vast majority of cells escape the microcirculation before sickling has occurred. (C) Cartoon showing that, if the delay time is sufficiently short or disappears because the intracellular HbS concentration is so high that fibers or seeds remain even after oxygenation in the lungs (11), then sickling can occur in the microcirculation, and the cell becomes much less flexible, introducing the possibility of vaso-occlusion by producing a “log jam” of cells. Cells that escape (B) and sickle in the larger vessels rapidly unsickle upon oxygenation in the lungs. Increasing the delay time or shortening the transit time through the microcirculation—for example, by decreasing the adherence of cells to the vessel walls (12)—should therefore decrease clinical severity. If the system were at equilibrium, all cells would be sickled at tissue oxygen pressures (10), and the disease would not be survivable after fetal hemoglobin is replaced by HbS (13).

Fig. 2.

Representative images of red cells before photolysis (−5 s) and at 0.2, 10, and 60 s after initiating cw photolysis (Movie S1). The contrast is less at −5 s because the absorption maximum of HbCO is 419 nm, whereas that of deoxy Hb is 430 nm, which corresponds to the wavelength of maximum transmission of the bandpass filter. At +0.2 s, no cells have sickled; 50% are sickled at +10 s, and 71% are sickled at +60 s. Data from the analysis of the cells circled in white is shown in Fig. 3.

Fig. 3.

Kinetic progress curves and sickling time distributions. (A) Images of single red cells from a trait donor at 0.2 s and 60 s after initiating photolysis, with curves corresponding to the three metrics used to detect cell distortion—area (blue), density ratio (red), and eccentricity (green). The curves for each metric were scaled so that the highest and lowest values were taken as one and zero. The noise reflects the absolute amplitudes. Analysis of photolysis experiments on CO-saturated red cells from an AA donor showed no change in any of the three metrics at all times up to 60 s. (B) Distribution of sickling times for ∼21,000 cells from 96 wells. (C) Cumulant of distribution (red curve). The dark blue dashed curve was calculated from the universal curve of delay time vs. supersaturation (the initial activity divided by the equilibrium activity) (33), by using the concentration distribution for normal cells of Lew et al. (34) multiplied by a factor of 1.03, activity coefficients calculated from SI Appendix, Eq. S2 (SI Appendix, Fig. S3), and solubilities at each concentration of the concentration distribution from SI Appendix, Eq. S1 by using a hemoglobin composition of 38% HbS and 62% HbA (the composition for this patient is 38% HbS, 58% HbA, 3.7% HbA2, and 0.3% HbF). Importantly, it is assumed that every cell has the same hemoglobin composition.

Fig. 4.

Calculations of fraction sickled vs. time in sickle trait cells corresponding to a predicted therapeutic effect. The plot shows the calculated fraction of trait cells sickled vs. time after initiating cw laser photodissociation of CO. Control (blue curve) was calculated as for Fig. 3C (described in SI Appendix) using a composition of 38% HbS and 62% HbA. Orange curve shows the level of inhibition predicted to be potentially therapeutic and corresponds to the sickling times in trait cells comparable to the sickling times in cells from patients with SC disease compared with homozygous SS disease cells. Red curve shows the level of inhibition predicted to be potentially curative and corresponds to sickling times in trait cells comparable to the sickling times in cells from patients with sickle cell disease and pancellular hereditary persistence of fetal hemoglobin (S/HPFH) compared with homozygous SS disease cells. Orange and red curves are calculated in SI Appendix, Fig. S5. The shaded areas reflect the uncertainties in the calculation.

Fig. 5.

Mean corpuscular volume and fraction sickled at 60 s for compounds that decrease intracellular hemoglobin concentration by swelling cells. (Upper) Mean corpuscular volume (MCV) measured with Coulter Counter. (Lower) Fraction sickled at 60 s after photolysis. Blue inverted triangles are for 4-h incubations, and red upright triangles are for 24-h incubations. The green horizontal line is our estimate of the minimum level of inhibition required for a therapeutic effect. The error bars represent one SD from the mean of 12 determinations [i.e., the fraction sickled for each of 12 wells at a single candidate-drug concentration relative to the average fraction sickle for the 12 wells of the control (no drug) in each of 12 wells, where both the mean and the SD were weighted by the number of cells in the image collected from each well]. There are no errors shown for the MCV, because the ratio of the SD to the mean of five determinations is <1%. The fraction sickled vs. time for each experiment is given in SI Appendix, Fig. S13.

Fig. 6.

Mean corpuscular volume and fraction sickled at 60 s for compounds in clinical trials that increase oxygen affinity. (Upper) Mean corpuscular volume measured with Coulter Counter. (Lower) Fraction sickled at 60 s after photolysis. Blue inverted triangles are for 4-h incubations, and red upright triangles are for 24-h incubations. The yellow vertical bar corresponds to the 9 ± 1 μM maximum concentration of the drug in the therapeutic range (41). There is little or no contribution to the inhibition of sickling by GBT440 that can potentially result from incomplete CO photodissociation because of the higher CO rebinding rate, as shown by experiments in which the fraction sickled at all GBT440 concentrations is independent of laser power density (SI Appendix, SI Text and Fig. S11). Errors were calculated as in the legend to Fig. 5. The fraction sickled vs. time for each experiment is given in SI Appendix, Fig. S13.