Therapeutic targeting and rapid mobilization of endosteal HSC using a small molecule integrin antagonist

Abstract

The inherent disadvantages of using granulocyte colony-stimulating factor (G-CSF) for hematopoietic stem cell (HSC) mobilization have driven efforts to identify alternate strategies based on single doses of small molecules. Here, we show targeting α9β1/α4β1 integrins with a single dose of a small molecule antagonist (BOP (N-(benzenesulfonyl)-L-prolyl-L-O-(1-pyrrolidinylcarbonyl)tyrosine)) rapidly mobilizes long-term multi-lineage reconstituting HSC. Synergistic engraftment augmentation is observed when BOP is co-administered with AMD3100. Impressively, HSC in equal volumes of peripheral blood (PB) mobilized with this combination effectively out-competes PB mobilized with G-CSF. The enhanced mobilization observed using BOP and AMD3100 is recapitulated in a humanized NODSCIDIL2Rγ(-/-) model, demonstrated by a significant increase in PB CD34(+) cells. Using a related fluorescent analogue of BOP (R-BC154), we show that this class of antagonists preferentially bind human and mouse HSC and progenitors via endogenously primed/activated α9β1/α4β1 within the endosteal niche. These results support using dual α9β1/α4β1 inhibitors as effective, rapid and transient mobilization agents with promising clinical applications.

Figure 1. R-BC154 and BOP preferentially bind human and murine HSC and progenitors in a divalent cation and integrin-dependent manner.

(a) Chemical structure of R-BC154 (1) and BOP (2). (b) Use of R-BC154, BOP and BIO5192 to demonstrate specific detection of binding to α₉ (grey) in human and (c) murine overexpressing cell lines (LN18 and CHO respectively) compared with α₄ (black) in the presence of 1 mM Ca²⁺/Mg²⁺ n=3 and is representative of at least two independent experiments. (d) Dose response of R-BC154 binding to human CB MNC. n=3 and is representative of two independent experiments. Binding in the presence of cations is significantly greater at all concentrations except 0 nM (P<0.005). (e) Representative populations of CB HSC (CD34⁺CD38⁻), progenitors (CD34⁺CD38⁺) and committed cells (CD34⁻CD38⁺). (f) R-BC154 binding to CB MNC committed cells, progenitors and HSC in the presence of 1 mM Ca²⁺/Mg²⁺ alone, or in combination with BOP or BIO5192. Data are representative of three individual samples. US, unstained. α₉—specific binding via α₉β₁, α₄—specific binding via α₄β₁. (g) Specific binding of R-BC154 binding to α₉β₁ on CB-committed cells (CC), progenitors (prog) and HSC. n=3 (h) R-BC154 binding to human BM (huBM). Data are representative of three individual samples. (i) Specific binding of R-BC154 to huBM α₉β₁. n=3 (j) α₉β₁ expression on huBM progenitors and HSC (grey solid line, dashed line IgG1 isotype control). Data are representative of three individual samples. (k) R-BC154 binding to humanized NSG (huNSG) BM. Data are representative of three individual samples. (l) Specific binding of R-BC154 to huNSG BM α₉β₁. n=3. (m) α₉β₁ expression on huNSG BM progenitors and HSC (grey solid line, dashed line IgG1 isotype control). Data are representative of three individual samples. (n) R-BC154 binding to mu BM-committed cells, progenitors (Lin⁻Sca-1⁺c-kit⁺; LSK) and HSC (LSKCD150⁺CD48⁻; LSKSLAM). Data are representative of three individual samples. (o) Specific binding of R-BC154 to mu BM α₉β₁. n=3. CB CD34⁺ cells were transplanted into 6.5 week old male (n=1) and female (n=2) NSG mice (k,l). Around 7–8-week-old female C57 mice were used for n and o. All data are mean±s.e.m. Analysis using one-way ANOVA (g,i,l,o) or two-way ANOVA (d), *P<0.05, **P<0.01, ***P<0.005, ****P<0.001.

Figure 2. R-BC154 targets HSC and progenitors via endogenously activated/primed α₄/α₉ integrins in endosteal BM.

(a) Schematic of central BM (cBM; blue) and endosteal BM (eBM; orange). (b) in vivo R-BC154 binding to huNSG central (blue) and endosteal (orange) BM huCD34⁺ cells compared with uninjected control (grey). Data are representative of three animals. (c) In vitro R-BC154 binding to huNSG central and endosteal BM progenitors (CD34⁺CD38⁺ cells) and HSC (CD34⁺CD38⁻ cells) in the absence of exogenous cations compared with uninjected control. Data are the representative of three samples. (d) Quantified R-BC154 binding from c, n=3. (e) In vitro R-BC154 binding to murine central and endosteal BM progenitors (LSK cells) and HSC (LSKSLAM cells) in the absence of exogenous cations compared to uninjected control. Data are representative of >10 samples. (f) Quantified R-BC154 binding from e, n=5. (g) Representative XFM images depicting the elastic scattering, Compton scattering and calcium distribution profile of a diaphyseal cross-section of non-decalcified murine femur. White arrow depicts calcium gradient radiating away from the endosteal surface (red dotted line). The hashed white line at ∼100 μm from the bone surface represents the interface between eBM and cBM. (h) Calcium content across representative line a to b in g, which begins ∼20 μm and extends to ∼220 μm from the endosteal surface. Data are expressed as arbitrary units (a.u.) n=10 different lines a to b. Representative of three animals. (i) Quantitative analysis of calcium, magnesium and manganese in murine BM central and endosteal extracellular fluid, (j) membrane bound (k) and intracellular. n=3. The endosteal extracellular component has been corrected for mechanically caused release (Supplementary Fig. 2a). CB CD34⁺ cells were transplanted into 9-week-old female NSG mice (b–d). Around 7–8-week-old female mice were used for e–h. All data are mean±s.e.m. Analysis using t-test, *P<0.05, ***P<0.005, ****P<0.001.

Figure 3. The α₉β₁/α₄β₁ integrin antagonist BOP rapidly mobilizes murine long-term repopulating HSC.

(a) Dose-dependent mobilization of murine progenitors (LSK, white square) and HSC (LSKSLAM, grey square) with BOP. Data are pooled from two biological repeats; n>3 per group. (b) Time-course of progenitor and HSC mobilization with 10 mg kg⁻¹ BOP. Time 0 is the vehicle control group. Data are pooled from two biological repeats. n=5 Individual animals per time point, not repetitive analysis. (c) Representative dot plots of PB progenitors and (d) HSC 1 h after a single dose of vehicle, BOP, AMD3100 or BOP+AMD3100 in combination, n>3. (e) Analysis of progenitor and HSC content in PB after 1 h treatment with vehicle, BOP, AMD3100 or BOP+AMD3100 in combination. (f) Total PB progenitor and HSC content 1 h after a single dose of vehicle (n=5), BIO5192 alone (n=5) or in combination with AMD3100 (n=5), or BOP in combination with AMD3100 (n=3). (g) Schematic of limiting dilution transplant analysis of mobilized RFP PB (five recipients per group). Representative dot plots of recipient BM multi-lineage engraftment. (h) Survival of recipients transplanted with 30 μl of RFP-mobilized PB. (i) Long-term HSC frequency in PB mobilized following BOP (green line), AMD3100 (red line) or BOP+AMD3100 (blue line). Around 7–8-week-old female (a–e) and male (f) C57 mice were used. 7–8 week old male (n=15) and female (n=10) RFP donors and male C57 recipients were used for g–i. All data are mean±s.e.m. Analysis using one-way ANOVA, *P<0.05, **P<0.01, ***P<0.005, ****P<0.001.

Figure 4. BOP and AMD3100 combination is a rapid and effective alternative to G-CSF.

(a) Analysis of PB progenitor (LSK) and HSC (LSKSLAM) content following a single dose of BOP (n=4), 4 days of G-CSF (n=3) or 4 days of G-CSF+1 dose of BOP (n=3). Data are representative of three independent experiments. (b) Analysis of PB progenitor and HSC content following a single dose of BOP+AMD3100 (n=8), 4 days of G-CSF (n=3), 4 days of G-CSF+1 dose of AMD3100 (n=9) or 4 days of G-CSF+1 dose of BOP+AMD3100 (n=8). Data are a pool of two independent experiments. (c) Schematic of serial competitive transplant assay. (d) Proportion of PB RFP⁺ white blood cells (WBC) in 1° recipients. Each data point is an individual recipient from one of two transplants. 1st=crosses, 2nd=circles. Dashed line is mathematically expected engraftment level. (e) Analysis of PB RFP⁺ and GFP⁺ lymphoid (B220⁺ and CD3⁺) and myeloid (Gr1⁺/Mac1⁺) engraftment in 1° recipients. n=8; pooled from 1st and 2nd experiments. (f) 2° recipient PB analysis. Each data point is an individual recipient. Dashed line is mathematically expected engraftment level. (g) Analysis of PB RFP⁺ and GFP⁺ lymphoid (B220⁺ and CD3⁺) and myeloid (Gr1⁺/Mac1⁺) engraftment in 2° recipients 20 weeks post transplant. Data are grouped based on the 5 individual 2° donors. n=4 2°recipients per donor. UD=undetectable. Around 7–8-week-old female C57 mice were used for a and b. Around 6–7-week-old male RFP and GFP and 9-week-old female C57 recipients were used for d and e. Secondary C57 recipients were 7 weeks old (f,g). All data are mean±s.e.m. Analysis using one-way ANOVA (a,b), or t-test (d,f) *P<0.05, **P<0.01, ***P<0.005, ****P<0.001.

Figure 5. The combination of BOP and AMD3100 effectively mobilizes human CD34⁺ stem and progenitors in humanized NSG (huNSG) mice.

(a) Representative flow cytometric plot of huNSG PB. (b) Analysis of huCD45⁺CD34⁺ cell content in huNSG PB following mobilization. Data are expressed as fold increase in CD34⁺ cells per ml PB relative to saline control and each data point represents an individual animal. Red bar, mean. CB CD34⁺ cells were transplanted into 6–10-week-old female and male NSG mice. Analysis using one-way ANOVA, **P<0.01, ****P<0.001.