Discovery of Novel Pyrazolopyrimidines as Potent, Selective, and Orally Bioavailable Inhibitors of ALK2

Abstract

Activin receptor-like kinase 2 (ALK2) is a transmembrane kinase receptor that mediates the signaling of the members of the TGF-β superfamily. The aberrant activation of ALK2 has been linked to the rare genetic disorder fibrodysplasia ossificans progressiva (FOP) and diffuse intrinsic pontine glioma (DIPG) that are associated with severely reduced life expectancy in pediatric patients. ALK2 has also been shown to play an essential role in iron metabolism by regulating hepcidin levels and affecting anemia of chronic disease. Thus, selective inhibition of ALK2 has emerged as a promising strategy for the treatment of multiple disorders. Herein, we report the discovery of a novel pyrazolopyrimidines series as highly potent, selective, and orally bioavailable inhibitors of ALK2. Structure-based drug design and systematic structure–activity relationship studies were employed to identify potent inhibitors displaying high selectivity against other ALK subtypes with good pharmacokinetic profiles.

The activin receptor-like kinase (ALK) family comprises seven receptors that mediate signal transduction of the transforming growth factor β (TGF-β) superfamily.¹ ALK family members share common structural elements including an extracellular ligand-binding domain, a transmembrane region, and a cytoplasmic tail with a serine/threonine kinase domain. In particular, ALK2, also known as activin A receptor, type I (ACVR1), participates in the signaling of bone morphogenetic protein (BMP) and phosphorylates BMP receptor-responsive SMAD proteins 1/5/8. The activated SMAD1/5/8 next forms heteromeric complexes with co-SMAD4 and translocates to the nucleus to regulate gene transcription. ALK2 is ubiquitously expressed in the body and plays important roles in bone development and the regulation of muscles, brain, and reproductive systems.^2,3

Gain-of-function mutations in the ACVR1 gene have been identified as the cause of fibrodysplasia ossificans progressiva (FOP), a rare disease characterized by the ossification of soft tissues.⁴ Hyperactivation of ALK2 signaling is also associated with diffuse intrinsic pontine glioma (DIPG), a fatal pediatric brainstem cancer.⁵ Furthermore, ALK2 has been proposed to attenuate anemia of chronic disease (ACD) by regulating hepatic hepcidin expression and serum ion levels.⁶⁻⁸ ACD frequently develops in patients with a wide range of infectious or autoimmune disorders, as well as cancer and kidney diseases, and contributes to the morbidity associated with these conditions. Recent studies have identified that targeting BMP-SMAD signaling with ALK2/3 inhibitors could offer an effective therapeutic approach to improve the phenotype of iron-refractory iron-deficiency anemia.⁹ Additional disorders, such as diffuse idiopathic skeletal hyperostosis and primary focal hyperhidrosis have also been found to be associated with ALK2 overactivation.^10,11 Although several inhibitors of ALK2 have been reported as potential therapeutic agents, their development was hindered because of low selectivity against other ALK subtypes.¹²⁻¹⁶ Herein, we report discovery of a novel pyrazolo[4,3-d]pyrimidine series as potent, selective, and orally bioavailable inhibitors of ALK2.¹⁷

To identify ALK2 inhibitors, we first conducted a high-throughput screening of our in-house compound library employing a homogeneous time-resolved fluorescence (HTRF) based activity assay. Compound potency was determined by measuring the loss of activity of ALK2 to phosphorylate peptide substrate. Following hit confirmation and initial hit expansion, pyrazolo[4,3-b]pyridine 1 (Table 1) was identified as an early lead showing moderate ALK2 kinase inhibition in the biochemical enzyme assay (IC₅₀ = 95 nM). Preliminary structure–activity relationship (SAR) study of the piperidine ring and the core bicycle are highlighted in Table 1. While a “nitrogen walk” from pyrazolo[4,3-b]pyridine 1 to pyrazolo[3,4-c]pyridine 2 yielded no significant increase in enzyme inhibition, introduction of an additional nitrogen atom provided compound 3 with >10-fold improved potency (IC₅₀ = 8.1 nM). Satisfied by the initial SAR exploration on the bicyclic core, pyrazolo[4,3-d]pyrimidine was selected for further investigation. Evaluation of substitutions on the piperidine ring revealed that 2,6-cis-dimethylpiperidine 5 (IC₅₀ = 2.5 nM) significantly improved ALK2 potency in comparison to the unsubstituted piperidine and 2-methylpiperidine analogs (compounds 3 and 4). Next, a HTRF assay was developed to measure the effect of our compounds on phosphorylation of SMAD1 in cells. HeLa cells were stimulated with BMP7 and the compounds ability to inhibit the BMP7-induced SMAD1 was measured. Although the initial lead 1 displayed no significant cellular activity (pSMAD1 IC₅₀ > 3 μM), compound 5 was identified as a moderate inhibitor of ALK2 kinase activity in cells (pSMAD1 IC₅₀ = 384 nM).

Table 1. Preliminary SAR Exploration.

Encouraged by the potency of compound 5 in both enzyme and cellular assays, we evaluated the binding mode of this promising lead with an ALK2 kinase structure through in silico docking studies (PDB code: 3MTF) (Figure 1). We propose that 5 binds to the ATP-binding pocket of ALK2, with the pyrazolopyrimidine fragment making critical hydrogen bonds in the hinge region with the backbone CO and NH of His284 and His286, respectively. In addition, a hydrogen bond is observed between the nitrogen of the pyridine ring and Tyr285 side chain. The protonated methylpiperazine group is largely solvent exposed in front of the ATP-binding pocket forming a hydrogen bond with Asp293. The cis-2,6-dimethylpiperidine occupies the hydrophobic pocket under the P-loop, making a 30° dihedral angle to the hinge binding core. One of the methyl groups is oriented toward a small hydrophobic groove formed by Leu343 and Ala353. In addition to the shape complementarity, displacement of a loosely bound (unfavorable) water molecule in the hydrophobic pocket by the dimethylpiperidine moiety could also account for the potency of compound 5. Overlaying published ALK2 structures with the binding mode of 5 identified several ALK2 structures having a crystallographic water in the I5 conserved water cluster (KLIFS notation^18,19) that resides in the same location as one of the methyl groups on the piperidine ring (Figure S1). Displacement of the I5 water might contribute to the improved potency observed by 5 in comparison to the unsubstituted piperidine and 2-methylpiperidine analogs.

Figure 1.

Proposed binding mode of compound 5 in the ATP pocket of ALK2. (A) Interaction of 5 with ALK2 in the hinge region. (B) Side view showing the dimethylpiperidine moiety in proximity to the catalytic Lys235 and Asp354.

Analysis of the binding mode of compound 5 suggested an opportunity to enhance potency by expanding to the ATP phosphate pocket. We rationalized that substitution at the 4-position of the piperidine ring would offer the proper trajectory to engage with the catalytic lysine Lys235 and/or DFG-motif residue Asp354 either by a direct hydrogen bond²⁰ or through a water-mediated hydrogen bonding network.²¹ To that end, polar functional groups were incorporated where select examples are highlighted in Table 2. Replacing the 2,6-dimethylpiperidine group with the corresponding morpholine analog 6 led to a 10-fold drop in ALK2 enzymatic potency (IC₅₀ = 23 nM). However, compound 6 remained potent in the cellular assays (pSMAD1 IC₅₀ = 245 nM), indicating favorable physicochemical properties for cellular uptake.²² Conversion of the piperidine ring to piperazine (compound 7), a ring system capable of serving as hydrogen bond donor, significantly diminished ALK2 activity. The piperazine group was next employed as a synthetic handle to further incorporate hydrogen bond acceptors. Examination of a variety of functional groups at the N-4 position identified amide, urea, and carbamate moieties provide inhibitors with good ALK2 activity (e.g., 8-13). From this series of compounds, urea 11 (pSMAD1 IC₅₀ = 39 nM) and carbamate 12 (pSMAD1 IC₅₀ = 27 nM) displaying 10-fold improvement in cellular potency over 5 were selected for further characterization.

Table 2. SAR Exploration of Piperidine Analogs.

^aNT = not tested.

One of the challenges in the development of ALK2 inhibitors has been obtaining selectivity over other ALK subtypes because of the high degree of structural homology within the family. In particular, ALK1 shares 79% sequence identity with ALK2 in the kinase domain, and none of the ALK2 inhibitors reported in the literature displayed a favorable selectivity over ALK1 at the time we initiated our discovery campaign. Given our interest in a truly selective ALK2 inhibitor, we profiled the lead compounds 11 and 12 against other members of the ALK family that share high sequence homology with ALK2, including the BMP type I receptors ALK1 and ALK3 as well as the TGF-β type I receptor ALK5 (Table 3). Both compounds displayed favorable selectivity against ALK3 (300–400×) where profiles against ALK1 (∼10×) and ALK5 (30–40×) were modest. To access the biological effect in iron metabolism, we measured the inhibitory activity of our compounds against hepcidin levels in Huh7 liver cell line. Cells were stimulated with BMP7 and the compounds ability to inhibit the secreted hepcidin was measured. Importantly, compounds 11 and 12 were capable of reversing the BMP-7-induced production of hepcidin in cells, with IC₅₀ values of 28 and 7.3 nM, respectively. Further analysis of physicochemical properties and ADME profiles revealed that compounds 11 and 12 exhibit low Caco2 permeability and metabolic stability in human liver microsomes, which translated to high systemic clearance and low oral absorption in rats following oral administration (Table 4). In the rat PK experiment, the ratio of compound concentrations in the liver (the target organ of anemia) versus plasma (L/P ratio) at 2 h post p.o. dosing was also measured. Compound 12 demonstrated a high tissue distribution to the liver (L/P > 350), suggesting a favorable targeted drug profile. Overall, 12 displays an improved ALK2 potency in enzyme and cellular assays with favorable ADME and rat PK profile values compared to compound 11. Therefore, 12 was subjected to further SAR investigation.

Table 3. Lead Development.

^aFold selectivity over ALK2 is indicated in parentheses. NT = not tested.

Table 4. In Vitro ADME Data and Rat Pharmacokinetic Profile^a of Selected Compounds.

	in vitro ADME			rat PK
compd	h-PB (%free)	h-IntCL (L/(h kg))	Caco2 Pm (× 10–6 cm/s)	Cmax (p.o., nM)	AUCinf (p.o., nM h)	T1/2 (h)	Vss (L/kg)	CL (L/(h kg))	F (%)	L/Pb
11	4.3	1.1	0.4	151	669	3.8	9.2	4.1	46	48
12	9.9	0.9	0.6	75	441	3.1	21.9	7.2	49	355
16	4.3	1.0	4.5	434	4050	5.7	7.6	1.1	85	30

^a1 mg/kg i.v. and 3 mg/kg p.o.

^bThe liver/plasma ratio of compound concentration was collected at 2 h post p.o. dosing.

First, the pyridinylpiperazine fragment of 12 is explored to enhance the ALK2 selectivity as well as the pharmacokinetic profile via fine-tuning the physicochemical properties. We reasoned that the low Caco2 permeability could be addressed by capping the free basic nitrogen of the piperazine moiety. Indeed, conversion to the N-methylpiperazine resulted in compound 14 with significant improvement in cellular permeability (Caco2 = 2.5 × 10^–6 cm/s) versus compound 12 (Caco2 = 0.6 × 10^–6 cm/s). This structural modification later proved important in achieving good oral exposure and bioavailability. Compound 14 showed a slight improvement in selectivity against ALK5 (75-fold), whereas selectivity against ALK1 remained modest (8-fold). Following an extensive SAR campaign, we discovered the selectivity could be significantly improved by changing the pyridine ring to fluorophenyl (compound 15, Table 3). However, this change led to a drop in ALK2 cellular potency (pSMAD1 IC₅₀ = 119 nM). Removal of the fluorine yielded compound 16 with potent ALK2 cellular activity (pSMAD1 IC₅₀ = 61 nM, hepcidin IC₅₀ = 34 nM) and excellent selectivity against the other ALK isoforms. In kinase inhibition assays employing K_m ATP concentrations, 16 displayed 19, 919 and 102-fold ALK2 selectivities against ALK1, ALK3, and ALK5, respectively. At 1 mM ATP concentration mimicking the cellular environment, the selectivity for ALK2 versus ALK1 improved to 31-fold. In addition, compound 16 exhibited good selectivity over other kinases in our in-house panel of 56 kinases, where only FGFR3 (IC₅₀ = 176 nM) was identified as having an IC₅₀ value less than 1000 nM (Table S3). Pleasingly, compound 16 demonstrated an excellent rat PK profile (Table 4). Following oral dosing at 3 mg/kg, 16 exhibited high exposure, low systemic clearance, and good oral bioavailability (85%) with favorable liver tissue distribution (L/P = 30). In addition, 16 was not a potent CYP inhibitor with IC₅₀ values greater than 25 μM for all major CYP isoforms tested (1A2, 2B6, 2C8, 2C9, 2C19, 2D6, and 3A4) and showed a low hERG patch clamp activity (<1% inhibition at 1 μM).

With such favorable profile, compound 16 was selected as the lead candidate and further profiled in mouse PK/PD investigation. Following oral administration into female C57BL/6 mice at 10 mg/kg, 16 demonstrated high absorption, with liver exposures (AUC_0–24 = 376 μM h) 14-fold higher than plasma exposures (AUC_0–24 = 26.2 μM h). In this study we measured the effect of ALK2 inhibition on hepcidin expression, which is regulated through the BMP-SMAD signaling pathway. Treatment of mice with compound 16 led to reduction of hepcidin concentration in plasma for at least 24 h, with maximal reduction (ca. 80% versus control) observed at 4 h post oral dosing (Table S4). The excellent PK profile and the observed in vivo suppression of hepcidin, a regulator of iron metabolism, support further development of 16 and related analogs as potential therapeutic agents for the treatment of anemia.

In summary, we have discovered a series of pyrazolopyrimidines as potent and selective ALK2 inhibitors exploiting in-house high-throughput screening and structure-guided drug design. Compounds with good ALK2 potency in enzyme and cellular assays were obtained by leveraging the interaction with the ATP phosphate pocket. Iterative multiparameter improvement of activity and physicochemical properties led to the discovery of compound 16 with high selectivity against other ALK subtypes and a favorable pharmacokinetic profile. Importantly, initial PK/PD studies of 16 in mice demonstrated significant hepcidin suppression in vivo, potentially blocking the negative effect of hepcidin on iron metabolism. The pyrazolopyrimidines reported herein represent a new chemotype that is structurally distinct from previously reported ALK2 inhibitors and is suitable for further development to treat anemia of chronic disease as well as other ALK2-associated disorders including FOP and DIPG.

Glossary

Abbreviations

ALK

activin receptor-like kinase;

TGF

transforming growth factor

FOP

fibrodysplasia ossificans progressiva

DIPG

diffuse intrinsic pontine glioma

ACVR1

activin A receptor, type I

BMP

bone morphogenetic protein

ACD

anemia of chronic disease

HTRF

homogeneous time-resolved fluorescence

SAR

structure–activity relationship

KLIFS

the kinase–ligand interaction fingerprints and structure database

ADME

absorption distribution, metabolism, and excretion

PK

pharmacokinetics

PD

pharmacodynamics

L/P ratio

ratio of compound concentrations in liver versus plasma

p.o.

oral administration

I.V.

intravenous

h-PB

in vitro protein binding of compound in human plasma

h-IntCL

intrinsic clearance in human liver microsomes

P_m

permeability

C_max

maximum concentration

AUC

area under the curve

T_1/2

half-life

V_ss

volume of distribution at steady state

CL

systemic clearance

F

oral bioavailability

CYP

cytochrome P450

hERG

human ether-go-go related gene

FGFR

fibroblast growth factor receptor

Clog P

calculated log P

TPSA

topological polar surface area.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsmedchemlett.2c00206.

• Experimental details for in vitro and in vivo studies; calculated physicochemical properties, synthetic procedures, and characterization of all compounds; supplementary data for docking studies; kinase profiling of compound 16 (PDF)

Author Contributions

The manuscript was prepared by M.H.N. and edited with O.A. with contributions from all authors.

The authors declare no competing financial interest.