GNAS1 and PHD2 short-interfering RNA support bone regeneration in vitro and in an in vivo sheep model

Abstract

Background: Our ability to guide cells in biomaterials for in vivo bone repair is limited and requires novel strategies. Short-interfering RNA (siRNA) allows the regulation of multiple cellular pathways. Core binding factor alpha 1 (Cbfa1) and hypoxia-inducible factor 1 (HIF-1) pathways can be modulated to direct bone formation via siRNA against guanine nucleotide-binding protein alpha-stimulating activity polypeptide 1 (siGNAS1) and prolyl hydroxylase domain-containing protein 2 (siPHD2), respectively.

Questions/purposes: We determined whether the administration of siGNAS1 and siPHD2 in mesenchymal stem cells (MSCs) promotes osteogenic phenotype, the dose-dependent effects of siGNAS1 on MSC differentiation to osteogenic phenotype, and whether the two siRNAs promote bone formation in vivo.

Methods: siRNAs were administered to MSCs at Day 0, and protein expression of bone-specific markers was assessed at Days 1, 2, and 4 (n = 3/group/time point). In an in vivo model using seven sheep, chambers containing silk fibroin-chitosan (SFCS) scaffolds with siRNA were implanted over the periosteum and harvested at Days 7, 21, 36, and 70 (n = 4/group/time point, except at Day 70 [n = 2]) to assess bone formation.

Results: siGNAS1 promoted collagen I and osteopontin expression, whereas siPHD2 had no effect in vitro. Dose-dependent effects of siGNAS1 on ALP expression were maximal at Day 1 for 10 μg/mL and Day 4 for 100 μg/mL. In vivo, by Day 70, mean bone volume increased compared to Day 7 for siGNAS1-SFCS (47.8 versus 1.8 mg/mL) and siPHD2-SFCS (61.3 versus 1.5 mg/mL).

Conclusions: Both siPHD2 and siGNAS1 support bone regeneration in vivo, whereas only siGNAS1 regulates bone phenotype in MSCs in vitro.

Fig. 1A–B

(A) A schematic of the Cbfa1 pathway shows the action of siGNAS1. The blocking of the GNAS1 gene via siGNAS1 results in the activation of the Cbfa1 pathway and expression of proteins such as osteopontin, collagen I, and others. GTP = guanosine triphosphate; GDR = guanosine diphosphate; cAMP = cyclic adenosine monophosphate. (B) A schematic shows the hypoxia pathway regulated by HIF-1 (HIF-1α and HIF-1β) and targeted to regulate angiogenesis by blocking of PHD2 via siPHD2, which allows the binding of HIF-1α to HIF-1β and expression of VEGF, EPO, SDF-1, and iNOS.

Fig. 2A–B

Flow diagrams show the setup of the (A) in vitro and (B) in vivo studies. The in vitro study involved the transfection of MSCs with siGNAS1 and siPHD2 and evaluation of cell proliferation, collagen expression, osteopontin expression, and dose-dependent ALP expression. The in vivo study involved embedding either siGNAS1 or siPHD2 or both in SFCS scaffolds and implanting them over periosteum engrafted on the latissimus dorsi muscle to fabricate ectopic bone.

Fig. 3A–E

Photographs illustrate the surgical procedure for the in vivo sheep surgery. (A) Two pieces of periosteum (12–15 cm; width approximately 1–1.5 cm) were harvested from sheep ribs using a scalpel and periosteal elevator dissection, (B) cut into eight to nine 3-cm pieces, and (C) autografted over the latissimus dorsi muscle of the same sheep at 3 cm apart. (D) PMMA chambers, one chamber each for pure SFCS scaffold, bone graft, and empty as controls and two chambers each for siGNAS1-SFCS, siPHD2-SFCS, and siGNAS1-siPHD2-SFCS, were created and (E) implanted on top of the grafted periosteum on the latissimus dorsi muscle using PDMS to suture and 5.0 PROLENE^® suture.

Fig. 4

Cellular proliferation was measured using MTT assay for MSCs transfected with siGNAS1, siPHD2, siGNAS1-siPHD2, and no siRNA. MSC proliferation decreased with the administration of siRNA within 24 hours whereas there was an increase in cell number without any siRNA administration. Comparison among groups with time shows differences: * = p < 0.05, ** = p < 0.01, and *** = p < 0.001 compared to Day 4 in the same group; α = p < 0.05 and αα = p < 0.01 compared to GNAS1; β = p < 0.05, ββ = p < 0.01, and βββ = p < 0.001 compared to no siRNA within the respective time point. Values are expressed as mean ± SEM.

Fig. 5

Collagen I expression was determined for MSCs transfected with siGNAS1, siPHD2, siGNAS1-siPHD2, siLamin, and no siRNA using Western blot. Maximal collagen I expression was observed at Day 2 in siGNAS1-treated cells. Comparison among groups with time shows differences: *** = p < 0.001 compared to Day 1; ## = p < 0.01 compared to Day 4. Values are expressed as mean ± SEM.

Fig. 6

Osteopontin expression was determined for MSCs transfected with siGNAS1, siPHD2, siGNAS1 and siPHD2, siLamin, and no siRNA using Western blot. Osteopontin expression was maximal within 24 hours of siGNAS1 administration and decreased with time by Day 4, reaching the same expression level as siPHD2. The expression of osteopontin remained low for siPHD2-administered cells over the 4-day measurements. Comparison among groups with time shows differences: * = p < 0.05, ** = p < 0.01, and *** = p < 0.001 compared to siGNAS1 within the respective time point; αα = p < 0.01 and ααα = p < 0.001 compared to Day 1 siGNAS1; β = p < 0.05 compared to Day 2 siGNAS1. Values are expressed as mean ± SEM.

Fig. 7

ALP assay was used to determine the effect of siGNAS1 concentrations at 10, 50, and 100 μg/mL and compared to siLamin (10, 50, and 100 μg/mL), scaffold only (no siRNA), and no scaffold (no siRNA) controls. The maximal ALP expression was observed for 10 and 50 μg/mL at Day 1, which decreased with time. The ALP expression for 100 μg/mL increased from Day 1 to Day 4. Comparison among groups shows differences: * = p < 0.05, ** = p < 0.01, *** = p < 0.001 compared to 10 μg/mL siGNAS1; ααα = p < 0.001 and αα = p < 0.01 compared to 50 μg/mL GNAS1; β = p < 0.05 compared to scaffold only; $$ = p < 0.01 and $ = p < 0.05 compared to 100 μg/mL siGNAS1; @@ = p < 0.01 compared to 10 μg/mL siLamin; ### = p < 0.01 compared to Day 1 for 10 μg/mL siGNAS1; ++ = p < 0.01 and + = p < 0.05 compared to Day 1 for 50 μg/mL siGNAS1. Values are expressed as mean ± SEM.

Fig. 8

Micro-CT imaging of bone regeneration in pure SFCS scaffolds, siGNAS1-SFCS, siPHD2-SFCS, and siGNAS1-siPHD2-SFCS scaffolds at Days 7, 21, 36, and 70 after implantation shows bone formation in white. Bone formation is observed at the periosteum-chamber interface as early as Day 21 in SFCS, siPHD2-SFCS, and siGNAS1-siPHD2-SFCS. Bone filled into the chamber from Day 36 to Day 70 for siGNAS1-SFCS, siPHD2-SFCS, and siGNAS1-siPHD2-SFCS.

Fig. 9

H&E staining of a 4-μm section of tissue harvested at Day 21 shows heterotopic bone formation at the muscle-periosteum interface with mineralizing bone in reddish-pink surrounding the newly forming marrow spaces in the middle.

Fig. 10

Movat’s pentachrome staining of 4-μm sections of tissue harvested at Days 21 and 36 shows mineralizing bone in deep yellowish-brown and mineralizing silk fibroin fibrils in pinkish-red. At Day 36, black india ink perfusion shows the vasculature in black for siGNAS1, siPHD2, and siGNAS1-siPHD2.

Fig. 11

Bone volume was plotted against time to assess bone formation for SFCS scaffolds modified with siGNAS1, siPHD2, and siGNAS1-siPHD2. There was an overall increase in bone volume from Day 7 to Day 70 for siRNA-embedded scaffolds. For the siGNAS1 group, *** = p < 0.0001 compared to Day 70; for the siPHD2 group, αα = p < 0.01 compared to Day 36 and ββ = p < 0.01 compared to Day 70; for the siGNAS1-siPHD2 group, ## = p < 0.01 compared to Day 70. Values are expressed as mean ± SEM.

Fig. 12

Bone mass density was plotted against time for SFCS scaffolds modified with siGNAS1, siPHD2, and siGNAS1-siPHD2. There was no difference in bone mass density among the groups. Values are expressed as mean ± SEM.