Effect of voxelotor on murine bone marrow and peripheral blood with hematopoietic progenitor cell mobilization for gene therapy of sickle cell disease

Abstract

In preparation for hematopoietic stem cell mobilization and collection, current ex vivo gene therapy protocols for sickle cell disease require patients to undergo several months of chronic red cell transfusion. For health care equity, alternatives to red cell transfusion should be available. We examined whether treatment with GBT1118, the murine analog of voxelotor, could be a safe and feasible alternative to red cell transfusion. We found that 3 weeks of treatment with GBT1118 increased the percentage of bone marrow hematopoietic stem cells and upon plerixafor mobilization, the percentage of peripheral blood hematopoietic stem cells. Our data suggest that voxelotor should be further explored for its potential safety and utility as preparation for hematopoietic stem cell mobilization and collection.

Introduction

Gene therapy is a potentially curative treatment for sickle cell disease (SCD), but feasibility and cost are major roadblocks to its implementation in low and middle income countries. Current ex vivo gene therapy protocols require the patient to undergo several months of chronic red cell transfusion before hematopoietic progenitor cell (HPC) mobilization and collection^1–4. Such preparation is thought to accomplish two goals: 1) decrease the risk of vaso-occlusive pain during HPC mobilization and collection, and 2) improve oxidative, hypoxic, and inflammatory stress on the bone marrow (BM) microenvironment that may contribute to low HPC numbers, quality and mobilization^5–7.

Nevertheless, 20% percent of transfused patients still have severe vaso-occlusive pain following HPC mobilization and collection, even with HbS percentages as low as 12% or processing as little as 15 liters⁸. Red cell transfusion also poses transfusion-specific risks such as delayed hemolytic transfusion reactions and hyperhemolysis⁹. In sub-Saharan Africa, the safety and feasibility of transfusion is limited by significant risks for contracting HIV, hepatitis B, and hepatitis C; low blood supply; and high cost¹⁰. A growing number of patients may be difficult to transfuse due to medical issues such as red cell alloimmunization or poor venous access, religious beliefs, or personal concerns¹¹.

Outside of pre-clinical models, the degree to which transfusion improves the BM microenvironment is unclear. Despite transfusion, 73% patients with SCD undergoing HPC mobilization and collection still require 2 to 5 days of apheresis, and 47% of patients still need at least 2 mobilization cycles⁸. In a cross-sectional study, markers of hemolysis, anemia, or inflammation and kidney, liver, or lung function were not improved in chronically transfused compared to untreated patients with HsBB or HbSβ0¹². Neither did baseline peripheral blood (PB) concentrations of hematopoietic stem cells (HSC), as measured by CD34+ bright cells or Lin⁻CD34⁺CD90⁺CD38⁻CD45RA⁻ cells, or stress erythropoiesis, as measured by glycophorin-A positive HPCs^6,13,14, improve with chronic transfusion compared to treatment-naïve patients (on no SCD-targeted therapy). Baseline PB CD34+ concentrations are relevant because they correlate with the PB CD34+ concentration or yield upon HPC mobilization with plerixafor, both in SCD patients^15,16 as well as healthy donors.¹⁷

Voxelotor may be a safe and feasible alternative to red cell transfusion for preparation prior to HPC mobilization and collection for gene therapy. Like hydroxyurea (HU) but without its myelosuppressive effects, it decreases the sickling capacity of endogenous red cells, and like transfusion and HU, it can increase hemoglobin (Hb)¹⁸.

We thus studied in female 4-month-old SS Townes mice (stock number 013071, The Jackson Laboratory) whether prior to plerixafor mobilization, conditioning with the mouse analog of voxelotor, GBT1118, might be a safe and effective alternative to red cell transfusion for HPC mobilization and collection for gene therapy. Plerixafor mobilization in the Townes mouse model has previously been shown to foreshadow results in humans¹⁹.

Materials and Methods

All animal studies were approved by the Institutional Animal Care and Use Committee of the New York Blood Center and carried out in strict accordance with their guidelines. All mice were sacrificed after a 3-week Teklad 2020 diet (Envigo) containing 4 g/kg GBT1118 or vehicle control, as stable Hb modification with GBT1118 is attained by day 14 of chronic dosing.²⁰ A subset of mice also received 10 mg/kg SC plerixafor (Sanofi-Aventis) or equivalent volume saline control prior to sacrifice 1–2 hour afterwards. PB, bone marrow (BM), and platelet-poor plasma were collected for analyses by hematology analyzer (Siemens Advia), flow cytometry (BD Biosciences LSR Fortessa), and ELISA (R&D Systems). Lin−Sca1+ckit+Flt3− staining was used to identify PB HSC²¹ and Lin−Sca1+ckit+CD48−CD150+ staining was used to identify BM HSC²². Because of the elevation of PB erythroid, myeloid, and lymphoid progenitors in SCD^23,24, we analyzed PB HSC by first gating on CD45+ cells to gate out CD45− negative mature erythroid progenitors and then gating on lineage-negative progenitors²⁵. Groups were tested for normality and then compared using the appropriate t-test (Mann–Whitney U for non-parametric data). Significance was assessed at the p ≤0.05 level. Figures show mean ± standard error of the mean.

Results

Verification of potency of GBT1118 and plerixafor.

The expected increase in Hb/Hct and decrease in retic % upon GBT1118 treatment versus vehicle control was verified¹⁸ (Supplemental Figure 1A–C). There was a decrease in WBC count with GBT1118 treatment (Supplemental Figure 1D), as observed in humans.²⁶ Also verified in both groups was the expected increase in WBC count after plerixafor treatment^16,19. (Supplemental Figure 1E).

Increased bone marrow and PB HSCs after GBT1118 conditioning alone (prior to plerixafor mobilization).

The % of BM HPC (LSK) was not significantly increased with GBT1118 treatment, but when more specific HSC markers were used²², the % of BM HSC was significantly increased with GBT1118 treatment (Figure 1A). The GBT1118 group also exhibited a trend towards an increased total # BM HSC per femur (Figure 1B). These increases in BM HSC were associated with increases in PB HPC (LSK) and HSC (LSKF) concentrations (Figure 2A–B).

Figure 1. BM HPC in mice treated with GBT1118 or vehicle control alone or followed by plerixafor (+P).

N=3 in both groups alone, N=5 in both groups with plerixafor. LSK = Lin−Sca1+ckit+; HSC = Lin−Sca1+ckit+CD48−CD150+.

A. BM LSK and HSC as % of Lin− cells. B. # BM LSK and HSC per femur.

Figure 2. PB HPCs in mice treated with GBT1118 or vehicle control alone or followed by plerixafor (+P).

N=3 each for control and GBT1118, 5 for control+P and 6 for GBT1118+P. LSK=Lin−Sca1+ckit+; LSKF = Lin⁻Sca1⁺ckit⁺Flt3⁻.

A. LSK cells/mL B. LSKF cells/mL C. LSKF as % of Lin− cells. D. LSKF as % of CD45+ cells. E. LSKF as % of CD45+ Lin− cells.

The % CD45+ cells ranged from 60–80% per sample.

Improved plerixafor-induced PB HSC mobilization after GBT1118 conditioning.

Plerixafor-induced mobilization of LSK cells/mL PB or LSKF cells/mL PB was not significantly increased with GBT1118 treatment (Figure 2A and B), possibly due to the interindividual variability in mobilization even within a single mouse strain²⁷. The % of mobilized LSKF cells was not increased with GBT1118 treatment when using Lin− gating only (Figure 2C), but when CD45+ gating was added to gate out CD45− mature erythroid progenitors which are increased in SCD^23,24, the % of mobilized LSKF cells was significantly higher with GBT1118 treatment (Figures 2D–E).

No significant increase in post-plerixafor inflammatory markers with GBT1118 conditioning.

Plasma VCAM-1, P-selectin, and E-selectin levels post-plerixafor were not significantly different between GBT1118 versus control mice (Supplemental Figure 2A). Interestingly, there were trends toward increased % and concentration of aged neutrophils in GBT1118 mice (Supplemental Figure 2B).

Discussion

Our results show that with 3 weeks of GBT1118 treatment versus vehicle control in SCD mice, the % of BM HSC increased (Figure 1A). This 1.6 fold increase is similar to the 1.7 fold increase seen with red cell transfusion or SA compared to SS mice⁷. There was also a trend towards increased # BM HSC per femur (Figure 1B). GBT1118 treatment also increased PB HPC and HSC concentrations (Figure 2A–B) and upon plerixafor mobilization, the % of PB HSC increased (Figure 2D–E). We hypothesize that BM HSCs may increase with voxelotor because anemia may damage the BM microenvironment^5,7 and voxelotor not only improves anemia but may increase O₂ delivery within a hypoxic BM environment²⁸.

That GBT1118 increased PB LSKF concentrations pre-plerixafor differs from our findings after red cell transfusion, where PB LSKF concentrations decreased⁷. We hypothesize that the baseline increase in PB CD34+ cells observed in SCD^8,16 may be due to persistence of some BM dysfunction despite GBT1118 treatment, thus causing some of the increased BM HSC induced with GBT1118 treatment to “spill” into the PB^5,7.

In the setting of mobilization for gene therapy, such a “spillover” may not be unfavorable, because the baseline PB CD34+ count correlates with the mobilized PB CD34+ count^15,16. Furthermore, the increased HSC in the BM after GBT1118 treatment still appear to mobilize into the PB after plerixafor, as coincident with the PB HSC increase (Figures 2D–E), the BM HSCs in the GBT1118 group were no longer increased (Figure 1A–B).

Regarding safety, analysis of inflammatory biomarkers after plerixafor showed no significant differences in GBT1118 treated versus vehicle control mice. The trend towards increased % and concentration of aged neutrophils is notable, however, and should be examined after GBT1118 treatment alone (without plerixafor), as it might contribute to suboptimal improvement in vaso-occlusive crises with voxelotor¹⁸.

Our results suggest that conditioning with voxelotor prior to plerixafor-based HPC collection may be a safe and efficacious alternative to conditioning with red cell transfusion. Voxelotor conditioning may have the potential to increase equity to gene therapy treatment and should be further explored.