Gradient
Search

Development of selective bispecific Wnt mimetics for bone loss and repair

The Wnt signaling pathway is intricately connected with bone mass regulation in humans and rodent models. They designed an antibody-based platform that generates potent and selective Wnt mimetics. Using this platform, they engineer bi-specific Wnt mimetics that target Frizzled and low-density lipoprotein receptor-related proteins and evaluate their effects on bone accrual in murine models. These synthetic Wnt agonists induce rapid and robust bone-building effects, and correct bone mass deficiency and bone defects in various disease models, including osteoporosis, aging, and long bone fracture. Furthermore, when these Wnt agonists are combined with antiresorptive bisphosphonates or anti-sclerostin antibody therapies, additional bone accrual/maintenance effects are observed compared to monotherapy, which could benefit individuals with severe and/or acute bone-building deficiencies. Their data support the continued development of Wnt mimetics for the treatment of diseases of low bone mineral density, including osteoporosis.


Wnt signaling components expression analysis from various cells and species.

Fzd1–10, along with Lrp5, and Lrp6 mRNA expression in a mesenchymal stem cells (MSC), b C3H10T1/2 cells, c MC3T3 cells, d mouse femur, and e human cancellous femoral head quantified by qPCR and normalized by the housekeeping gene Actin. Bars represent means ± SD. dn = 3 mice and en = 3 human femoral heads.


Diseases of low bone mineral density (BMD), including osteoporosis, present serious global health concerns for both men and women. Osteoporosis-related fragility fractures are associated with significant morbidity and mortality. It has been estimated that the lifetime risk for osteoporotic fracture at age 50 is ~20% for men and 50% for women. Using Fracture Risk Assessment Tool (FRAX) models, 158 million people (21 million men and 137 million women) over age 50 years were estimated to be at high risk for an osteoporotic fracture worldwide in 2010, with an expected doubling of this number by 204. Antiresorptive agents and bone anabolic drugs are the mainstays of therapy for osteoporosis, but different therapies are still needed to address the clinical unmet needs and provide alternative therapeutic options.


Radiographic and serum analyses of bone accrual induced by Wnt mimetics.

a Representative x-ray (top row) and micro-computed tomography (micro-CT) (bottom row) images at day 28. For x-ray, individual sample regions of interest (ROI) quantified the longitudinal (b) average whole-body bone mineral density (BMD). a Red arrows point to increased bone mineral (radiographic contrast) in the distal femur and proximal tibia. Red dotted regions demarcate the ROI analyzed for c average lumbar vertebra (L4–L6). d Serum samples obtained at days 0, 7, 14, 21, and 28 were assessed for procollagen type-1 N-terminal propeptide (P1NP). For each time point, the serum P1NP levels for all test groups are normalized to the average of vehicle-treated P1NP serum concentration. e Longitudinal average tibial BV/TV at days 0, 14, and 28. f Day 28 ex vivo endpoint micro-CT quantification of the distal femoral epiphyseal BV/TV and g mid-shaft cortical thickness collected at termination. Statistical significance was determined by two-way ANOVA for (b, d, e) and one-way ANOVA for all others. Graphs represent mean values ± SD, wherein *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Statistical analyses where the following letters define groups (a = Vehicle, b = Anti-Sclerostin, c = 18R5-Dkk1c). b Day 14, a vs. b, p = 0.0049; a vs. c, p = 0.0031; Day 21, a vs. b, p = 0.0005; a vs. c, p = 0.0002; Day 28: a vs. b, p = 0.0006; a vs. c, p ≤ 0.0001. c a vs. b, p ≤ 0.0001; a vs. c, p ≤ 0.0001. d Day 7, a vs. b, p ≤ 0.0001; a vs. c, p = 0.017; Day 14, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001. e Day 14, a vs. b, p ≤ 0.0001; a vs. c, p ≤ 0.0001; Day 28, a vs. b, p = 0.0033; a vs. c, p ≤ 0.0001. f a vs. b, p = 0.0388; a vs. c, p ≤ 0.0001. g a vs. b, p ≤ 0.0001; a vs. c, p = 0.0009. b–gn = 10 mice/group.


In addition, other conditions of low bone mineral density, including osteogenesis imperfecta, renal osteodystrophy, and disuse osteopenia have proven difficult to treat and therapies are needed to overcome this treatment gap. Wnt (“Wingless-related integration site” or “Wingless and Int-1” or “Wingless-Int”) ligands and their signaling activators play key roles in controlling the development, homeostasis, and regeneration of many essential organs and tissues, including bone. Specifically, osteoblasts, osteocytes, and osteoclasts (the three major cell types of bone), all have been shown to be regulated by Wnt/β-catenin-dependent signaling. This has been observed along the entire osteoblast lineage, including in osteocytes where modulation of sclerostin is required in response to mechanical load-induced bone accrual. Additionally, Wnt signaling has been shown to reduce the ratio of receptor activator of nuclear factor-κB ligand (RANKL)/osteoprotegerin (OPG) expression, which indirectly reduces osteoclast activation.


Histology and mechanical testing of bone accrual induced by Wnt mimetics.

a, b Tibias were isolated after 28 days, formalin fixed, embedded in methyl methacrylate, sectioned longitudinally, and stained. Representative images of tibial sections from each treatment group at a 4× scale bar 1 mm and b 10× scale bar 200 µm stained with Goldner’s trichrome (mineralized bone is colored blue) and c 40× scale bar 50 µm stained for tartrate resistant acid phosphatase (TRAP). Femurs collected at study termination (day 28) were analyzed for bone strength at the mid-shaft femur by a standard 3-point bending test and quantified for d average femur stiffness and e average ultimate load to failure. a–c Representative images from n = 10 samples/group having similar results. d, e Bar graphs represent mean values ± SD. Statistical significance was determined by one-way ANOVA, where ****p < 0.0001. The following letters define groups (a = Vehicle, b = Anti-Sclerostin, c = 18R5-DKK1c). d a vs. b, p ≤ 0.0001; a vs. c, p ≤ 0.0001. e a vs. b, p ≤ 0.0001; a vs. c, p ≤ 0.0001. n = 10 mice/group.


Direct activation of Wnt signaling has the potential to enhance bone accrual and repress bone resorption in disease settings. One of the challenges in engineering a therapeutic that directly activates Wnt signaling is the existence of 19 Wnt ligands and multiple possible combinations of Wnts with receptors, including Frizzled 1–10 (Fzd1–10) and low-density lipoprotein receptor-related protein 5 and 6 (Lrp5, Lrp6). In addition, Wnt proteins are highly modified post-translationally. Palmitoleoylation on a conserved serine residue is essential for Wnt interaction and signaling through their receptors. Partly due to this post-translational palmitoleoylation, Wnts are hydrophobic, making them difficult to express and purify. The anti-Fzd antibody 18R5 is a Fzd1,2,5,7,8 binder, and the Lrp binder DKK1c binds to both Lrp5 and 6. To resolve these challenges associated with Wnts, a Wnt mimetic linking the Fzd binder (18R5 in a single-chain variable fragment [scFv] format) and the Lrp binder (DKK1c) into a single polypeptide chain (18R5-DKK1c) has been previously shown to have the potential to activate Wnt/β-catenin signaling in various tissue systems, including liver, bone, and intestine.


In vitro characterization of Wnt mimetic molecules.

a Diagram of the tetravalent bi-specific Wnt mimetic formats. VHH domain is represented by the open oval symbol and IgG by the black-filled oval symbol. b–e Wnt signaling STF activity is represented by relative luminescence units (RLU). b The dose-dependent STF activities of the four different VHH-IgG fusion mimetics shown in a. c Dose-dependent STF activities of the FA-L5, FA-L6, FB-L5, FB-L6, and WNT3A in HEK293 cells. d Dose-dependent STF activities of FA-L5, FA-L6, FB-L5, FB-L6, and WNT3A in C3H10T1/2 cells. e Dose-dependent STF activities of L5-Fc, L6-Fc, and WNT3A in HEK293 cells. Data are representative of three independent experiments performed in triplicates and are shown as mean ± SD.



However, the functional impact of this molecule in in vivo systems remained to be clearly demonstrated. They have also recently developed a platform for potent and selective Wnt mimetic generation and engineered water soluble and easily manufacturable IgG-based bi-specific molecules that bind to specific Fzd(s) and Lrp5 or 6. They identified multimerization of Fzd and Lrp, with optimal stoichiometry of two Fzds and one or two Lrps, as a requirement for maximal Wnt/β-catenin activation.



Wnt mimetic antibodies robustly stimulate bone accrual in naive young and older mice.

a Serum concentrations of FA-L6 antibody after intraperitoneal administration, as measured by ELISA over a 21-day period. Each time point represents the average assessed in three animals, with error bars representing the standard deviation of the mean. b–e 8-week-old C57BL6/J female mice were injected bi-weekly for 28 days intraperitoneally with increasing concentrations of FA-L6 antibody and compared to a negative control antibody (anti-GFP IgG 1 mg/kg), and a positive control (bi-weekly subcutaneous administration of 10 mg/kg anti-sclerostin antibody). b Longitudinal DEXA measurements of whole-body BMD quantified for days 0, 7, 14, and 21. c Longitudinal micro-CT of proximal tibial bone volume/tissue volume (BV/TV) quantified on days 0, 7, 14, and 21. d Endpoint distal femoral BV/TV and e mid-shaft cortical thickness were quantified at day 28. f Whole-body BMD by DEXA in 16-week-old C57BL6/J female mice injected bi-weekly for 14 days with FA-L6 (10 mg/kg) or FB-L5 (1 or 10 mg/kg) antibody, or a negative control antibody (anti-β-gal 10 mg/kg), or positive control (anti-sclerostin antibody 25 mg/kg). g, h 1-year-old C57BL6/J female mice were injected bi-weekly for 28 days intraperitoneally with 1 and 10 mg/kg of FA-L6 or FB-L5 antibody or a negative control antibody (anti-β-gal) or positive control (bi-weekly subcutaneous administration of 25 mg/kg anti-sclerostin antibody). g Longitudinal whole-body BMD over time and h day 28 endpoint distal femoral BV/TV measurements are shown. a Graph represents mean ± SD, n = 3 mice. b–h Bar graphs represent mean values ± SD. Statistical significance was determined by two-way ANOVA for b, c, g and one-way ANOVA for all others (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001) where the following letters define groups (a = Vehicle, b = Anti-Sclerostin, c = FA-L6 (0.3 mg/kg), d = FA-L6 (1 mg/kg), e = FA-L6 (10 mg/kg), f = FB-L5 (10 mg/kg)). b Day 7, a vs. d, p = 0.041; Day 14, a vs. b, p = 0.032, a vs. c, p = 0.0412, a vs. d, p ≤ 0.0001; Day 21, a vs. b, p = 0.0019, a vs. c, p = 0.0092, a vs. d, p = 0.0024. c Day 7, a vs. c, p ≤ 0.0001; a vs. d, p ≤ 0.0001. Day 14, a vs. c, p ≤ 0.0001; a vs. d, p ≤ 0.0001. Day 21, a vs. c, p ≤ 0.0001; a vs. d, p ≤ 0.0001. d a vs. b, p = 0.0183, a vs. c, p = 0.0141, a vs. d, p = 0.0002. e a vs. b, p = 0.001, a vs. d, p = 0.0401. f a vs. b, p = 0.007, a vs. e, p ≤ 0.0001, a vs. f, p ≤ 0.0001. g Day 14, a vs. f, p ≤ 0.0001; Day 21, a vs. b, p ≤ 0.0001, a vs. e, p ≤ 0.0001, a vs. f, p ≤ 0.0001; Day 28, a vs. b, p ≤ 0.0001, a vs. e, p ≤ 0.0001, a vs. f, p ≤ 0.0001. h a vs. b, p ≤ 0.0001, a vs. e, p ≤ 0.0001, a vs. f, p ≤ 0.0001. n = 8 mice/group.


In the studies presented here, they first examined whether 18R5-DKK1c can be delivered systemically to induce bone accrual in vivo. To test this hypothesis, they delivered 18R5-DKK1c through adeno-associated virus (AAV) in a mouse model and measured longitudinal changes in bone accrual using a variety of analyses. In parallel, they engineered tetravalent bi-specific antibody-based molecules with different Fzd and Lrp binders and specificities. They then examined the efficacy of these recombinant Wnt mimetic proteins on bone building in various disease models, including osteoporosis, aging, and long bone fracture. They found that these Wnt mimetics had robust bone-building effects, thus supporting their continued development as therapeutic antibodies to modulate the Wnt pathway for tissue regeneration1.


Wnt mimetic antibodies reverse bone loss associated with ovariectomy induced osteoporosis.

a Schematic of experimental design. 4-week-old C57BL6/J female mice were surgically ovariectomized (OVX). Seven months later, DEXA was used to confirm bone loss in the OVX groups prior to treatment initiation. Mice were injected bi-weekly with intraperitoneal administration of FA-L6 10 mg/kg or FB-L5 10 mg/kg antibody, or a negative (anti-β-gal) or positive control (bi-weekly subcutaneous injection of anti-sclerostin antibody 25 mg/kg) for 28 days. b Longitudinal DEXA measurements were collected and whole-body BMD quantified for days 0, 7, 14, 21, and 28 and c whole femur BMD on day 28. Bone strength at the mid-shaft femur was analyzed by a standard 3-point bending test and quantified for d average ultimate load to failure. e Representative micro-CT reconstructions of day 28 L4 lumbar vertebra (images from n = 8 samples/group having similar results) and f quantification of L4 BV/TV. Compressive loading was applied to L5 vertebrae to quantify g max load and h stiffness. Graphs represent mean values ± SD, n = 8 mice/group, except Vehicle + OVX, where n = 3. Statistical significance was determined by two-way ANOVA for b and one-way ANOVA for all others b–d and f–h, where the following letters define groups (a = OVX + Vehicle, b = OVX + Anti-Sclerostin, c = OVX + FA-L6, d = OVX + FB-L5) and *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. b Day 7, a vs. d, p = 0.0135. Day 14, a vs. c, p = 0.0014, a vs. d, p ≤ 0.0001. Day 21, a vs. b, p = 0.0023, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001. Day 28, a vs. b, p = 0.0005, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001. c, a vs. b, p = 0.0427, a vs. c, p = 0.0023, a vs. d, p = 0.0156. d, a vs. c, p = 0.0164, a vs. d, p = 0.0098. f a vs. b, p = 0.0002, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001. g a vs. b, p = 0.0111, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001. h a vs. b, p = 0.0174, a vs. c, p = 0.0056, a vs. d, p ≤ 0.0001.


Combination studies of Wnt mimetic antibodies with currently available therapeutics.

a 12-week-old C57BL6/J female mice were injected bi-weekly for 28 days with intraperitoneal administration of FB-L5 (10 mg/kg) antibody and compared to a negative control anti-β-gal (10 mg/kg) and positive control bi-weekly subcutaneous administration of 25 mg/kg anti-sclerostin antibody or a combination of anti-sclerostin antibody and FB-L5. A 2-week washout period then followed (day 42 assessment). Longitudinal DEXA measurements were collected and whole-body % BMD change quantified for days 0, 14, 28, and 42, normalized to baseline. At day 28, combination treatment induced statistically significant increases in BMD compared to individual treatments. At day 42, combination treatment maintained the bone accrual gained during the first 4 weeks of treatment and remained statistically significant compared to individual treatments. b 12-week-old C57BL6/J female mice were injected bi-weekly for 28 days with intraperitoneal administration of FB-L5 (10 mg/kg) antibody and compared to a negative control anti-β-gal (10 mg/kg). From day 28 through day 42, animals were treated with either vehicle or alendronate at 1 or 4 mg/kg. Longitudinal DEXA measurements were collected and whole-body % BMD quantified for days 0, 14, 28, and 42, normalized to baseline. a, b Graphs represent mean values ± SD, n = 8 mice/group. Statistical significance determined by two-way ANOVA (**p < 0.01, ****p < 0.0001), where the following letters define groups, a (a = anti-β-Gal, b = Anti-Sclerostin, c = FB-L5, d = FB-L5 + Anti-Sclerostin) and b (a = anti-β-Gal + Vehicle, b = FB-L5 + Vehicle, c = FB-L5 + Alendronate (1 mg/kg), d = FB-L5 + Alendronate (4 mg/kg)). a Day 14, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001, b vs. c, p ≤ 0.0001, b vs. d, p ≤ 0.0001. Day 28, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001, b vs. c, p ≤ 0.0001, b vs. d, p ≤ 0.0001, c vs. d, p ≤ 0.0001. Day 42, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001, b vs. c, p = 0.0031, b vs. d, p ≤ 0.0001, c vs. d, p ≤ 0.0001. b Day 14, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001. Day 28, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001. Day 42, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001, a vs. d, p ≤ 0.0001, b vs. c, p = 0.0023, b vs. d, p ≤ 0.0001.


Wnt mimetic antibodies increase bone accrual and bone repair when provided 2 weeks after fracture.

a Schematic of experimental design. 12-week-old C57BL6/J female mice were surgically fractured adhering to the Einhorn fracture model. Animals did not receive any treatments for 2 weeks to allow for a normal cartilaginous callus to form. b Callus formation was visually confirmed by x-ray (top panel) at day 14; representative images (n = 8 samples/group having similar results) demonstrate a small, but visible, callus with no cortical fusion. Mice were injected bi-weekly (BIW) for 21 days with intraperitoneal administration of FB-L5 (10 mg/kg), FA-L6 (10 mg/kg) and compared to a negative control anti-β-gal and bi-weekly subcutaneous administration of anti-sclerostin (30 mg/kg) antibody. Representative x-ray images demonstrate the large amount of cortical fusion and radiographic contrast present in the callus and marrow cavity of bones treated with FA-L6 and FB-L5 (bottom panel) at day 28. c Femurs were isolated at day 35 and scanned by µCT. Measurements were made according to the left panel, where red dotted lines represent 4.2 mm, or 400 slices by µCT, centered on the fracture line with image analysis shown for tissue area vs. bone area, and representative images shown for each treatment group. d Callus tissue volume, bone tissue volume, BV/TV, and bone mineral content/mm (BMC/mm) were measured for all animals at day 35. d Graphs represent mean values ± SD, n = 8 mice/group. Statistical significance was determined by one-way ANOVA (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001), where the following letters define groups, (a = Vehicle, b = FA-L6, c = FB-L5). Callus tissue volume, a vs. b, p ≤ 0.0001, a vs. c, p ≤ 0.0001. Bone tissue volume, a vs. b, p = 0.0001, a vs. c, p = 0.0022. BV/TV, a vs. b, p = 0.001, a vs. c, p = 0.0116. BMC/mm, a vs. b, p = 0.001, a vs. c, p = 0.0117.


  1. Fowler, T. W. et al. (2021). Development of selective bispecific Wnt mimetics for bone loss and repair. Nature communications, 12 (1). doi:10.1038/s41467-021-23374-8