Inhibition of GDF8 (Myostatin) accelerates bone regeneration in diabetes mellitus type 2

Abstract

Metabolic diseases like diabetes mellitus cause bone healing deficiencies. We found significant impairment of bone regeneration, osteogenic differentiation and proliferation in diabetic bone. Moreover recent studies suggest a highly underestimated importance of GDF8 (Myostatin) in bone metabolism. Our goal was to analyze the role of GDF8 as a regulator of osteogenic differentiation, proliferation and bone regeneration. We used a murine tibial defect model in diabetic (Lepr^db-/-) mice. Myostatin-Inhibitor Follistatin was administered in tibial bony defects of diabetic mice. By means of histology, immunohistochemistry and QRT-PC osteogenesis, differentiation and proliferation were analyzed. Application of Myostatin-inhibitor showed a significant improvement in diabetic bone regeneration compared to the control group (6.5 fold, p < 0.001). Immunohistochemistry revealed a significantly higher proliferation (7.7 fold, p = 0.009), osteogenic differentiation (Runx-2: 3.7 fold, p = 0.011, ALP: 9.3 fold, p < 0.001) and calcification (4.9 fold, p = 0.024) in Follistatin treated diabetic animals. Therapeutical application of Follistatin, known for the importance in muscle diseases, plays an important role in bone metabolism. Diabetic bone revealed an overexpression of the catabolic protein Myostatin. Antagonization of Myostatin in diabetic animals leads to a restoration of the impaired bone regeneration and represents a promising therapeutic option.

Figure 1

Expression of Myostatin and Activin A is highly upregulated in uninjured diabetic bone as compared to WT (shown in arbitrary units). (A) qPCR of uninjured (Sham) and injured postoperative (3dpo) mice showed increased expression of Myostatin in uninjured diabetic and injured diabetic and WT bone. Follistatin (Foll.) treated diabetic bone showed no significant difference compared to diabetic uninjured and injured bone. Follistatin treatment did not alter Myostatin expression compared to control animals (diabetic bone with PBS). (B) qPCR showed increased level of Activin A in diabetic bone compared to WT bone. Follistatin treatment did not alter Activin A expression compared to control animals (diabetic bone with PBS). Results are shown as means ± SEM. P-value: * < 0.05; ** < 0.01; ** < 0.001 (two sample t-test).

Figure 2

Reduced SMAD2/3 expression and increased RUNX-2 transcription is mediated by inhibition of Myostatin in vitro. (Left) mASC from Lepr^db−/− Black 6 mice were cultured for 7 days in osteogenic differentiation medium without or with either Follistatin Activin neutralizing Antibody (Activin-AB) or Myostatin neutralizing antibody (Myo.-AB). (Right) Quantification of immunofluorescence positive pixels for Runx-2 (red) showed increased Runx-2 levels in Follistatin and Myo.-AB treated groups compared to only differentiation medium. Myostatin positive pixels (red) were reduced in all treatment groups compared to the standard differentiation medium group. Significantly decreased Activin A positive levels were observed in the Activin-AB and Myo.-AB group but not in the Follistatin treated group compared to the standard differentiation medium group. SMAD 2/3 positive levels were reduced in the Follistatin and Myo.-AB group but not in the Activin-AB group compared to the standard differentiation medium group. Scale bar: 10 μm. Results are shown as means ± SEM. P-value: * < 0.05; ** < 0.01; *** < 0.001; (two sample t-test).

Figure 3

Local application of Follistatin inhibits Myostatin and GDF 11 and downstream targets in vivo. (Left) Myostatin (red), GDF 11 (green), Activin A (red), downstream targets and BMP 7 downstream target SMAD 1 (green) were detected by immunofluorescence for p38 (red) and SMAD 2/3 (red) (3 days postoperatively). (Right) Quantification of immunofluorescence positive pixels revealed that Follistatin treatment in diabetic mice significantly reduced Myostatin and GDF 11 levels and downstream targets p38 and SMAD 2/3 compared to elevated levels in diabetic animals without treatment. Blockage of Myostatin and downstream targets with Follistatin showed reduced levels of p38 and SMAD 2/3 compared to wildtype animals, while no difference in SMAD 1 levels was detected upon Follistatin treatment. Follistatin application does not significantly decrease Activin A levels in diabetic mice. Results are shown as means ± SEM. P-value: *** < 0.001, (two sample t-test). Scale bar: 20 μm. Foll. = Follistatin. Results are shown as means ± SEM. P-value: * < 0.05; ** < 0.01; *** < 0.001; (two sample t-test).

Figure 4

Follistatin enhanced osteogenic differentiation and proliferation. (Left) Osteogenic differentiation was detected by immunofluorescence for ALP (green) and RUNX-2 (red), proliferation by immunofluorescence for PCNA (red). (Right) Quantification of immunofluorescence positive pixels revealed that osteogenic differentiation is significantly increased in Follistatin treated diabetic defects compared to the control group db⁻/db⁻ and similar to WT mice ALP and RUNX-2 activity. Follistatin treatment in diabetic mice revealed a significant enhancement of proliferation compared to diabetic control animals and reached level of wildtype animals. Scale bar ALP + RUNX-2: 20 μm, PCNA: 10 μm. Foll. = Follistatin. Results are shown as means ± SEM. P-value: * < 0.05; ** < 0.01; *** < 0.001; (two sample t-test).

Figure 5

Follistatin enhances calcification of osteogenic differentiated mASC in vitro. Alizarin red staining of 28 days osteogenic differentiated mASCs showed increased mineralization in Follistatin treated cells compared to diabetic control. Scale bar: 20 μm. Results are shown as means ± SEM. P-value: * < 0.05; ** < 0.01; *** < 0.001; (two sample t-test).

Figure 6

Follistatin enhanced diabetic bone regeneration – histology. (A) Aniline blue staining of db⁻/db⁻ control mice showed impaired new osteoid formation compared to WT control mice, while diabetic defects treated with Follistatin showed improved osteoid formation compared to diabetic control mice. WT mice treated with Follistatin showed a slight but not significant improvement. (B) Extracted osteoid formation area. (C) Quantification of aniline blue positive pixels (6 histological sections per sample) revealed that osteoid formation is significantly increased in Follistatin treated diabetic defects compared to the control group db⁻/db⁻ and similar to osteoid formation of WT mice 7dpo (7 days postoperatively). Follistatin treatment in WT mice did not showed significantly improved osteoid regeneration. (D) Goldner staining of db⁻/db⁻ control mice revealed impaired osteoid formation compared to WT, while diabetic defects treated with Follistatin showed improved osteoid formation. (E) Quantification of acid green positive pixels. (F + G) Immunohistochemistry for Osteocalcin demonstrated reduced Osteocalcin positive cells in diabetic animals on 7dpo, while local application of Follistatin restored the Osteocalcin positive cell number to the WT group level. (H) Calculation of Ob.S/Tb. Area showed diminished osteoblast number in db⁻/db⁻ compared to wildtype mice. The treatment with Follistatin enhanced the number significantly. (I) Calculation of OA/TA revealed reduced osteoid regeneration in diabetic animals compared to WT animals, while the application of Follistatin increases the regeneration compared to diabetic condition. Scale bar A + D: 200 μm. Scale bar F: 50 μm. Cb indicates cortical bone. dpo = days postoperatively. Ctrl = Control. Foll. = Follistatin. Results are shown as means ± SEM. P-value: * < 0.05; ** < 0.01; *** < 0.001; (two sample t-test).

Figure 7

Follistatin enhances diabetic bone regeneration - radiography. (A) Representative digital radiography of mouse tibiae 28 days post operatively. Yellow box indicates the site of the defect. (B) Measurement of luminance at the defect site in the region of interest (40 × 20 Px) revealed a decrease of bone density in diabetic control tibiae compared to WT tibiae. Administration of Follistatin increased the bone density significantly compared to diabetic control animals. (B) Statistical analysis Scale bar: 5 mm. Results are shown as means ± SEM. P-value: * < 0.05; ** < 0.01; (two sample t-test).