Cellular hypoxia suppressed MSC proliferation but enhanced osteogenic differentiation
The hypoxia simulated by CoCl2 treatment significantly suppressed MSC proliferation. One day of hypoxia (i.e., 1 day of CoCl2 treatment followed by 6 days of incubation without CoCl2) did not show an impact on MSC proliferation after 1, 3 or 5 days of culture, but 1.34-fold inhibition was observed on day 7 (Fig. 1a). Interestingly, hypoxia for 3, 5 and 7 days showed similar inhibitory effect on cell proliferation after 5 and 7 days of culture (Fig. 1a).
Osteogenic gene expression, ALP staining and ALP activity in MSCs were analyzed on day 7. Hypoxia for 3 days respectively upregulated Col1α1, Runx2, Alp, Osx, Opn, Ocn and Vegf gene expression by 3.12-, 3.35-, 4.12-, 14.29-, 8.35-, 12.1- and 2.61-fold compared to the control group (Fig. 1b–j). Hypoxia for 5 days enhanced only Ocn and Vegf gene expression (respectively by 9.07- and 1.75- fold compared to the control group). Hypoxia for 1 day enhanced Vegf gene expression by 2.05-fold compared to the control group, but did not affect the expressions of other osteogenic markers (Fig. 1h). Interestingly, continuous simulation of hypoxia for 7 days did not affect all the expressions of all the osteogenic markers tested (Fig. 1b–j).
Hypoxia for 3 days yielded the strongest ALP and alizarin red staining (Fig. 2a and c). Similarly, hypoxia for 3 days enhanced ALP activity by 2.92- fold compared to the control group (Fig. 2c). Quantification of the mineralized matrix showed that hypoxia for 3 and 5 days respectively promoted matrix mineralization by 1.18-, and 1.09-fold compared to the control group (Fig. 2d).
COL1α1, OSX, RUNX2, and ALP protein expressions were analyzed using western blot assay on day 7. Hypoxia for 1, 3, 5 and 7 days respectively enhanced COL1α1 protein expression by 1.50-, 1.55-, 1.41- and 1.47- fold (Fig. 2e and Additional file 1: Figure S1A). Hypoxia for 1, 3, 5 and 7 days respectively enhanced RUNX2 protein by 2.67-, 4.05-, 2.97- and 2.29- fold (Fig. 2e and Additional file 1: Figure S1A). Hypoxia for 1, 3, 5 and 7 days respectively enhanced ALP protein expression by 1.81-, 4.97-, 2.68- and 1.91- fold (Fig. 2e and Additional file 1: Figure S1A). Similarly, hypoxia for 1, 3, 5 and 7 days respectively enhanced OSX protein expression by 1.84-, 2.71-, 2.25- and 2.08- fold (Fig. 2e and Additional file 1: Figure S1A). Hypoxia for 3 days showed the highest effect on osteogenic marker protein expression, which concurs with the results for mRNA expression.
STAT3 inhibitor reversed hypoxia-induced STAT3 phosphorylation and osteogenic differentiation
Cellular hypoxia enhanced HIF-1α expression by 1.81-fold. The STAT3 inhibitor did not alter hypoxia-induced HIF-1α expression (Fig. 3a and Additional file 1: Figure S2A). Hypoxia did not affect the total STAT3 expression (Fig. 3a). Hypoxia robustly enhanced (5.46-fold) STAT3 phosphorylation and the STAT3 inhibitor reduced this effect by 3.10-fold (Fig. 3a and Additional file 1: Figure S2B).
Since hypoxia for 3 days showed the highest effect on the osteogenic differentiation of MSCs, we choose this culture condition to analyze the effect of STAT3 inhibitor on hypoxia-induced osteogenic differentiation. STAT3 inhibitor respectively suppressed hypoxia-induced Col1α1, Runx2, Alp, Osx, Opn, Ocn and Vegf gene expression by 6.13-, 4.87-, 5.67-, 6.56-, 4.31-, 5.41- and 2.63-fold (Fig. 3b–h). STAT3 inhibitor alone did not affect the expression of osteogenic genes compared to the control group (Fig. 3b–h). STAT3 inhibitor reduced hypoxia-induced ALP protein expression and ALP activity (5.38-fold; Fig. 4a and c). STAT3 inhibitor strongly reduced (2.37-fold) hypoxia-induced matrix mineralization (Fig. 4b and d). Similarly, STAT3 inhibitor reduced matrix mineralization by 2.08- and 4.51-fold respectively compared to the results for the CoCl2 + inhibitor and control groups.
Western blot data analysis showed that STAT3 inhibitor respectively reduced hypoxia-induced COL1α1, RUNX2, ALP and OSX protein expressions by 4.56-, 1.67-, 1.34- and 1.78-fold (Fig. 3e and Additional file 1: Figure S3A–D). STAT3 inhibitor reduced OSX protein expression by 1.89-fold compared to the control group (Additional file 1: Figure S3D). However, STAT3 inhibitor did not affect the expressions of the other osteogenic proteins tested compared to the control group. RUNX2, ALP and OSX protein expression in STAT3 inhibitor group were respectively suppressed by 3.23-, 3.02- and 2.35-fold compared to the hypoxia + STAT3 inhibitor group (Additional file 1: Figure S3B–D).
Hypoxia enhanced bone regeneration and STAT3 inhibitor impaired this effect
Histological images showed more newly formed bone in the bone defect area at week 3 in the CoCl2 group compared to the images for the control, hypoxia + STAT3 inhibitor, and STAT3 inhibitor groups (Fig. 5a). Similarly, the bone defect area was filled with newly formed bone in the CoCl2 group at week 5. However, a clear bone defect gap was observed in the control, hypoxia + STAT3 inhibitor, and STAT3 inhibitor groups (Fig. 5a).
Hypoxia upregulated Hif-1α and Alp mRNA expression in bone defect femora and STAT3 inhibition reversed this effect
To investigate the possible interaction between hypoxia and STAT3 signaling during osteogenesis and bone defect healing, we analyzed Hif-1α and Alp mRNA expression in mice femoral bone defects treated with CoCl2 and/or STAT3 inhibitor. Hif-1α and Alp mRNA expression were upregulated in the femurs of all the bone defect groups compared to the results for the blank control group (Fig. 5b and c). CoCl2-induced hypoxia further upregulated Hif-1α and Alp expression by 1.81- and 2.77-fold, respectively (Fig. 5b and c). STAT3 inhibitor reduced hypoxia-induced Hif-1α and Alp expression by 1.15- and 2.30-fold, respectively (Fig. 5b and c). The STAT3 inhibitor did not affect Hif-1α expression but suppressed the Alp expression by 1.31-fold compared to the control group (Fig. 5c).
CoCl2-simulated hypoxia promoted bone defect healing and STAT3 inhibitor reversed this effect
μ-CT and X-ray images showed that CoCl2 promoted femoral bone defect healing at week 3 and 5 compared to the control group (Fig. 6a and Additional file 1: Figure S4). Interestingly, the STAT3 inhibitor reversed hypoxia-induced bone defect healing at week 3 and 5 (Fig. 6a and Additional file 1: Figure S4). Moreover, STAT3 inhibitor reduced bone defect healing compared to the control, CoCl2 and CoCl2 + STAT3 inhibitor groups (Fig. 6a and Additional file 1: Figure S4).
Similar effects of CoCl2 and STAT3 inhibitor were shown by the newly formed bone trabecular parameters at week 3 and week 5 (Fig. 6b–e). CoCl2 treatment enhanced BV/TV and Conn D levels by 1.51- and 2.44-fold, respectively, at week 3 compared to the control. STAT3 inhibitor reduced the CoCl2-induced impact on BV/TV and Conn D levels by 1.37- and 1.64-fold, respectively (Fig. 6b and d). Similarly, STAT3 inhibitor reduced BV/TV and Conn D levels by 1.28- and 1.27-fold, respectively, compared to the control group at week 3 (Fig. 6a and d). CoCl2 treatment enhanced BV/TV, Tb. N and Conn D levels by 1.49-, 1.45- and 1.46-fold, respectively, at week 5 compared to the control group (Fig. 6b–e). STAT3 inhibitor reduced the CoCl2-induced impact on BV/TV and Conn D levels by 1.38- and 1.31- fold, respectively, at week 5 (Fig. 6b–e). Moreover, STAT3 inhibitor reduced Conn D levels by 2.08-fold compared to the control (Fig. 6b–e), and reduced BV/TV, Tb. N, and Conn D levels by 1.49-, 1.25- and 2.27-fold, respectively, compared to those for the CoCl2 + inhibitor group at week 5 (Fig. 6b–e). Hypoxia suppressed Tb. Sp levels by 1.38-fold at week 5 compared to the control group (Fig. 6e). STAT3 inhibitor reversed hypoxia-mediated suppression at week 5 (Fig. 6e). Moreover, the STAT3 inhibitor group enhanced Tb. Sp levels by 1.36- and 1.30-fold at week 5 compared to the control and CoCl2 + inhibitor groups, respectively (Fig. 6e).
Hypoxia upregulated HIF-1α, pSTAT3 and ALP protein expression in the bone defect area and STAT3 inhibitor reversed this effect
Immunohistochemistry images showed stronger immunostaining (brown and light brown color) of HIF-1α, p-STAT3 and ALP at week 3 compared to the staining at week 5 in all the groups tested (Fig. 7a–c). The CoCl2-treated group showed stronger immunostaining of HIF-1α, p-STAT3 and ALP compared to all other groups at week 3 and 5.
Quantitative analysis of the immunohistochemistry of tissue section showed 3.32- and 1.74-fold higher expression of HIF-1α in the hypoxic group compared to the control group at week 3 and 5, respectively (Fig. 7a and d). STAT3 inhibitor did not affect CoCl2-induced HIF-1α expression at week 3 and 5. The CoCl2 group showed 1.87- and 2.85-fold higher expression of pSTAT3 compared to the control group at week 3 and 5, respectively (Fig. 7a and d). STAT3 inhibitor reversed the hypoxia-induced pSTAT3 upregulation at week 3 and 5. Similarly, the CoCl2-group showed 2.02- and 1.97-fold higher expression of ALP compared to the control group at week 3 and 5, respectively (Fig. 7a and d). STAT3 inhibitor reduced the hypoxia-induced ALP expression by 1.73- and 1.70-fold at week 3 and 5, respectively.