Saracatinib is an efficacious clinical candidate for fibrodysplasia ossificans progressiva

Abstract

Currently, no effective therapies exist for fibrodysplasia ossificans progressiva (FOP), a rare congenital syndrome in which heterotopic bone is formed in soft tissues owing to dysregulated activity of the bone morphogenetic protein (BMP) receptor kinase ALK2 (also known as ACVR1). From a screen of known biologically active compounds, we identified saracatinib as a potent ALK2 kinase inhibitor. In enzymatic and cell-based assays, saracatinib preferentially inhibited ALK2, compared with other receptors of the BMP/TGF-β signaling pathway, and induced dorsalization in zebrafish embryos consistent with BMP antagonism. We further tested the efficacy of saracatinib using an inducible ACVR1Q207D-transgenic mouse line, which provides a model of heterotopic ossification (HO), as well as an inducible ACVR1R206H-knockin mouse, which serves as a genetically and physiologically faithful FOP model. In both models, saracatinib was well tolerated and potently inhibited the development of HO, even when administered transiently following soft tissue injury. Together, these data suggest that saracatinib is an efficacious clinical candidate for repositioning in FOP treatment, offering an accelerated path to clinical proof-of-efficacy studies and potentially significant benefits to individuals with this devastating condition.

Keywords: Bone Biology; Drug screens; Growth factors; Protein kinases; Therapeutics.

Figure 1. DSF screening identifies saracatinib as a potent inhibitor of ALK2.

(A) Chemical structure of saracatinib. (B) Plot comparing the thermal shift (ΔT_m) of ALK2 and ALK5 in response to different clinical compounds. Selected compounds of interest are labeled, including saracatinib. A red dashed line is drawn as an approximate guide to mark equipotency for ALK2 and ALK5. (C) Overview of the ALK2 co-crystal structure with saracatinib. (D) Interactions of saracatinib with the ATP-binding pocket of ALK2.

Figure 2. Saracatinib is a selective inhibitor of BMP versus TGF-β type I receptor activity in cells.

(A) IC₅₀ inhibition curves for saracatinib against purified recombinant ALK1-6 kinase domains were determined using a LANCE Ultra TR-FRET kinase assay (PerkinElmer). Reactions contained 10 nM kinase, 50 μM peptide substrate, and 10 μM ATP. Data shown are plotted as mean ± SD. (n = 3 independent replicates). (B) Representative inhibition curves for saracatinib against constitutively active BMP (caALK1, 2 and 3) and activin/TGF-β (caALK4 and 5) type I receptors, based on the activity of BMP responsive promoter element LUC (BRE-Luc) and TGF-β responsive LUC (CAGA-Luc) reporters in C2C12 and 293T cells, respectively. Data shown are representative of more than 3 independent experiments, with data plotted as mean ±SEM. (n = 3 replicates). (C) IC₅₀ inhibition curves for saracatinib against the signaling induced by indicated ligands based on the activity of BRE-Luc (BMP ligands) and CAGA-Luc (activin/TGF-β ligands) reporters stably expressed in MDA-MB-231 cells. Data shown are plotted as mean ± SD. (n ≥ 3 independent replicates). (D) Western blot analyses showing the inhibitory activity of saracatinib against BMP7-induced phosphorylation of SMAD1/5, as well as TGF-β–dependent phosphorylation of SMAD2 in C2C12 cells.

Figure 3. Saracatinib inhibits the neofunction of ALK2^R206H.

Western blot analysis of phospho-SMAD1/5 levels following treatment with saracatinib and either BMP6 or activin A in (A) FOP patient-derived fibroblasts cells (GM00513) or (B) WT fibroblasts cells (ND34770). Cell lines were validated by DNA sequencing (top panels). Data are representative of multiple experiments using fibroblasts from 2 independent FOP patients (GM00513 female 16 years of age and GM00783 male age unknown, Coriell Institute).

Figure 4. In vivo efficacy of saracatinib in the ACVR1^Q207D-transgenic mouse model of FOP.

Neonatal CAG-Z-eGFP-caALK2-transgenic mice were injected with Ad.Cre (1 × 10⁸ PFU i.m. P7) and treated with 25 mg/kg/d saracatinib or vehicle orally for 28 days. (A) All mice expressed the eGFP reporter in the injected left hindlimb, and 100% of vehicle-treated mice (6 of 6) developed radiographic HO and severe loss of passive range of motion (original magnification, ×0.8). (B) Treatment with saracatinib essentially abrogated radiographic HO in 4/4 mice, preserving range of motion (2-way ANOVA with Sidak’s test for multiple comparisons; P = NS, days 1–7; **P < 0.01, days 8–10; ***P < 0.001, days 11–28; all vs. vehicle treatment), data depicted as median ± IQR, n as indicated. (C) Treatment with saracatinib had no significant impact on normal growth based on weight gain as compared with controls (2-way ANOVA, P = NS), data depicted as mean ± SEM, n as indicated. (D) Representative hindlimb scoring system. All range of motion scoring was performed by 2 separate operators who were blinded to the treatment condition of the animal.

Figure 5. Dose-dependent in vivo efficacy of saracatinib in the FOP-knockin mouse model.

Neonatal Acvr1^{[R206H]FlEx/+]} mice were injected with a low dose of Ad.Cre (1 × 10⁸ PFU i.m. P7) and treated with 25 mg/kg/d saracatinib or vehicle orally for 28 days (grey shaded region) and observed for a total of 90 days. Representative mice are shown with x-ray radiography, Micro-CT, and Alizarin red hindlimb prep. (A) Sixty percent of vehicle-treated mice (3/5) developed HO and severe loss of passive range of motion at 90 days (original magnification, ×0.7). Treatment with saracatinib prevented radiographic HO in 5/5 treated mice, and (B) preserved range of motion in 4/5 mice at 90 days (2-way ANOVA with Sidak’s test for multiple comparisons, **P < 0.01, day 61; ****P < 0.0001, day 90 vs. control), data depicted as median ± IQR, n as indicated. ACVR1^{[R206H]FlEx/+}-knockin mice were injected with a high dose of Ad.Cre (5 × 10⁹ PFU i.m. P7) and treated with varying doses of saracatinib or vehicle orally for 40 days. Radiographic HO and associated impaired range of motion were observed to progress over 40 days of treatment. Treatment with saracatinib at 2.5, 5, and 10 mg/kg/d protected mice from radiographic HO, and preserved passive range of motion (C–E). Two-way with Dunnett’s test for multiple comparisons. (C) *P < 0.05, days 14–15, 24–25, and 27–28; ***P < 0.001, days 19–23; ****P < 0.0001, days 29–39 vs. vehicle treatment. (D) ***P < 0.001, days 19–24; **P < 0.01, days 29–34; ****P < 0.0001, days 35–39 vs. vehicle treatment. (E) *P < 0.05, days 19–24; **P < 0.01, days 29–33; ****P < 0.0001, days 34–39 vs. vehicle treatment. Data depicted as median ± IQR, n as indicated.

Figure 6. Saracatinib treatment preserves neonatal and juvenile orthotopic skeletal growth.

(A) Treatment of neonatal mice with saracatinib at 25 mg/kg/d exerted a mild impact on weight gain during the first 28 days of treatment (P < 0.0001 compared with vehicle treatment, 2-way ANOVA), data depicted as mean ± SEM, n as indicated. (B) Treatment of neonatal mice with saracatinib at 2.5—10 mg/kg/d generally did not impact weight gain in comparison to vehicle, except for the 10 mg/kg/d dose at days 33 and 39 (P < 0.05, 2-way ANOVA with Dunnett’s test for multiple comparisons), nor did it impact neonatal and juvenile orthotopic skeletal growth, based on preserved femur length (C), and preserved femur bone mineral density (D) at 2.5, 5, and 10 mg/kg/d oral doses of saracatinib versus vehicle treatment (data shown as mean ± SD, n = 8, 4, and 4 femurs analyzed, respectively; P = NS, 1-way ANOVA with Dunnett’s test for multiple comparisons).